Datasets:

Basticooler

/

pubmed_12B_sampled

like 0

Effects of a virtual gender swap on social and temporal decision-making Mounting evidence has demonstrated that embodied virtual reality, during which physical bodies are replaced with virtual surrogates, can strongly alter cognition and behavior even when the virtual body radically differs from one's own. One particular emergent area of interest is the investigation of how virtual gender swaps can influence choice behaviors. Economic decision-making paradigms have repeatedly shown that women tend to display more prosocial sharing choices than men. To examine whether a virtual gender swap can alter gender-specific differences in prosociality, 48 men and 51 women embodied either a same-or different-gender avatar in immersive virtual reality. In a betweensubjects design, we differentiated between specifically social and non-social decision-making by means of a virtually administered interpersonal and intertemporal discounting task, respectively. We hypothesized that a virtual gender swap would elicit social behaviors that stereotypically align with the gender of the avatar. To relate potential effects to changes in self-perception, we also measured implicit and explicit identification with gendered (or gender-typical) traits prior to and following the virtual experience, and used questionnaires that assessed the strength of the illusion. Contrary to our hypothesis, our results show that participants made less prosocial decisions (i.e., became more selfish) in different-gender avatars, independent of their own biological sex. Moreover, women embodying a male avatar in particular were more sensitive to temptations of immediate rewards. Lastly, the manipulation had no effects on implicit and explicit identification with gendered traits. To conclude, while we showed that a virtual gender swap indeed alters decision-making, genderbased expectancies cannot account for all the task-specific interpersonal and intertemporal changes following the virtual gender swap.Embodied cognition proposes a strong reciprocal link between bodily and cognitive processes. While such a relation was initially proposed primarily for the physical body, it has increasingly been extended to the subjectively perceived body as well, which is known to be plastic and shaped by both sensory inputs and prior expectancies (e.g., Dijkermann and Lenggenhager 1 ). In line with this idea, mounting evidence has demonstrated that experimentally induced temporary changes in the sense of one's own body can alter a broad range of cognitive and affective processes, including social cognition (for a review, see Maister et al. 2 ). This relation has often been investigated with multisensory stimulation paradigms, such as the rubber hand illusion 3 , and more recently also in virtual reality (VR). The precise sensory control of the virtual body, coupled with synchronous visuotactile and/ or visuomotor feedback between the participant's own body and the avatar, make VR a powerful tool to induce a strong temporary sense of embodiment, even in the presence of salient dissimilarities, such as different skin tones 4 or gender 5,6 . It has been suggested that such alterations of embodiment induce a so-called Proteus effect, where the embodiment of an avatar elicits behaviors and cognition concomitant with attributes and stereotypes of the embodied avatar 7 . This effect is believed to result from a deindividuation process in virtual environments, during which "virtual cues take precedence over physical cues" (Yee et al. 8 , p. 3). Along these lines, virtual body swap studies have demonstrated that embodiment of avatars that visually diverge from one's physical characteristics can exert strong influences on cognition. For example, embodiment of an elderly 9 or dark-skinned 4 person can reduce negative stereotypes towards such outgroup members in young or light-skinned adults, respectively. Mounting evidence has demonstrated that embodied virtual reality, during which physical bodies are replaced with virtual surrogates, can strongly alter cognition and behavior even when the virtual body radically differs from one's own. One particular emergent area of interest is the investigation of how virtual gender swaps can influence choice behaviors. Economic decision-making paradigms have repeatedly shown that women tend to display more prosocial sharing choices than men. To examine whether a virtual gender swap can alter gender-specific differences in prosociality, 48 men and 51 women embodied either a same-or different-gender avatar in immersive virtual reality. In a betweensubjects design, we differentiated between specifically social and non-social decision-making by means of a virtually administered interpersonal and intertemporal discounting task, respectively. We hypothesized that a virtual gender swap would elicit social behaviors that stereotypically align with the gender of the avatar. To relate potential effects to changes in self-perception, we also measured implicit and explicit identification with gendered (or gender-typical) traits prior to and following the virtual experience, and used questionnaires that assessed the strength of the illusion. Contrary to our hypothesis, our results show that participants made less prosocial decisions (i.e., became more selfish) in different-gender avatars, independent of their own biological sex. Moreover, women embodying a male avatar in particular were more sensitive to temptations of immediate rewards. Lastly, the manipulation had no effects on implicit and explicit identification with gendered traits. To conclude, while we showed that a virtual gender swap indeed alters decision-making, genderbased expectancies cannot account for all the task-specific interpersonal and intertemporal changes following the virtual gender swap. Embodied cognition proposes a strong reciprocal link between bodily and cognitive processes. While such a relation was initially proposed primarily for the physical body, it has increasingly been extended to the subjectively perceived body as well, which is known to be plastic and shaped by both sensory inputs and prior expectancies (e.g., Dijkermann and Lenggenhager 1 ). In line with this idea, mounting evidence has demonstrated that experimentally induced temporary changes in the sense of one's own body can alter a broad range of cognitive and affective processes, including social cognition (for a review, see . This relation has often been investigated with multisensory stimulation paradigms, such as the rubber hand illusion [bib_ref] Rubber hands 'feel' touch that eyes see, Botvinick [/bib_ref] , and more recently also in virtual reality (VR). The precise sensory control of the virtual body, coupled with synchronous visuotactile and/ or visuomotor feedback between the participant's own body and the avatar, make VR a powerful tool to induce a strong temporary sense of embodiment, even in the presence of salient dissimilarities, such as different skin tones 4 or gender [bib_ref] The effect of gender body-swap illusions on working memory and stereotype threat, Peck [/bib_ref] [bib_ref] Inducing and mitigating stereotype threat through gendered virtual body-swap illusions, Peck [/bib_ref]. It has been suggested that such alterations of embodiment induce a so-called Proteus effect, where the embodiment of an avatar elicits behaviors and cognition concomitant with attributes and stereotypes of the embodied avatar [bib_ref] The proteus effect: The effect of transformed self-representation on behavior, Yee [/bib_ref]. This effect is believed to result from a deindividuation process in virtual environments, during which "virtual cues take precedence over physical cues" (Yee et al. . Along these lines, virtual body swap studies have demonstrated that embodiment of avatars that visually diverge from one's physical characteristics can exert strong influences on cognition. For example, embodiment of an elderly 9 or dark-skinned 4 person can reduce negative stereotypes towards such outgroup members in young or light-skinned adults, respectively. www.nature.com/scientificreports/ Furthermore, avatar embodiment might not only change the perception of others but also the self, as the identification with a child-like virtual body, for example, results in stronger self-association of child-like personality traits [bib_ref] Illusory ownership of a virtual child body causes overestimation of object sizes..., Banakou [/bib_ref] , as well as in a shift towards a more child-like pitched voice [bib_ref] Embodiment in a child-like talking virtual body influences object size perception, self-identification,..., Tajadura-Jiménez [/bib_ref]. Gender swap illusions have similarly prompted users to engage in stereotype-relevant behaviors, such as conforming to gender-stereotyped language norms when embodying gendered avatars [bib_ref] Virtual gender identity: The linguistic assimilation to gendered avatars in computer-mediated communication, Palomares [/bib_ref] or engaging more in killing versus healing behaviors when embodying male versus female avatars [bib_ref] Do men heal more when in drag, Yee [/bib_ref]. In one study, female participants were less sensitive to the effects of stereotype threat (i.e., being told that women tend to perform worse than men on a working memory task they were about to complete) when embodying a male compared to a female avatar [bib_ref] The effect of gender body-swap illusions on working memory and stereotype threat, Peck [/bib_ref]. On the other hand, male participants underperformed when embodying a female avatar [bib_ref] Inducing and mitigating stereotype threat through gendered virtual body-swap illusions, Peck [/bib_ref]. Another study reported that when being caressed in intimate areas by a person of the same-gender, heterosexual female and male participants' ratings of pleasantness and erogeneity were higher during virtual embodiment of a body of the different-gender as compared to a same-gender toucher [bib_ref] Wearing same-and opposite-sex virtual bodies and seeing them caressed in intimate areas, Mello [/bib_ref]. Tacikowski et al. [bib_ref] Fluidity of gender identity induced by illusory body-sex change, Tacikowski [/bib_ref] demonstrated that a few minutes of immersion in a gender swap illusion resulted in altered implicit and explicit identification with gendered (or gender-typical) traits, particularly in a shift towards a more balanced identification with both genders, as well as concomitant decreases in gender-stereotyped beliefs about oneself. While most of these paradigms examined changes during or immediately following the illusion, a few studies have specifically assessed the longevity of cognitive alterations following virtual embodiment, and determined that some virtual embodiment manipulations induced changes that persisted for at least one week [bib_ref] Virtual embodiment of white people in a black virtual body leads to..., Banakou [/bib_ref]. Findings such as these demonstrate that alterations in cognition following transient modifications of embodiment can potentially persist even outside the laboratory. Therefore, a more finetuned understanding of how virtual body swaps can affect cognitive affective processes will prove informative not only for basic research, but may hold also practical, consumer-level relevance for assessing potential consequences of spending extended time as an avatar in virtual environments. One major theme of economic games surrounds the notion that humans care about and are motivated by the interests of others. As a major building block of prosocial behavior, generosity (i.e., sharing behaviors) has been incorporated into numerous economic models as a measure of social preferences [bib_ref] A theory of reciprocity, Falk [/bib_ref]. It should be noted that we here use the term "generosity" specifically as a description of the willingness to incur a relative disadvantage in order to benefit someone else. A repeatedly documented source of heterogeneity in generosity has been linked to the differential gender-specific social preferences of men and women [bib_ref] Sex differences in social behavior: Comparing social role theory and evolutionary psychology, Eagly [/bib_ref]. While prosocial and generous behaviors have characteristically been ascribed to women (where they are stereotypically perceived as a favorable trait), more selfish behaviors have in turn been generally characterized as a typecast male trait [bib_ref] Same behavior, different consequences: Reactions to men's and women's altruistic citizenship behavior, Heilman [/bib_ref]. Several behavioral economic studies have corroborated robust gender differences in social decision-making 20 , while a recent neuropharmacological study by Soutschek et al. [bib_ref] The dopaminergic reward system underpins gender differences in social preferences, Soutschek [/bib_ref] further substantiated such gender-specific prosocial preferences on a neural level by showing increased responses in the striatum, a component of the neural reward system, to prosociality in women. To add to the fundamental research on the contribution of altered bodily self experiences to cognitive processes, this study aimed to alter participants' gender identity using an immersive virtual gender swap illusion (i.e., using avatars that are different to the self in gender stereotypical physical features), and to examine the effects of this virtual gender swap on social decision-making. Examining such cognitive alterations in digital spaces not only contributes to the fundamental research on bodily self-consciousness, but further holds real-world implications, especially as an ever-increasing number of social interactions occur in digital environments. To this purpose, healthy female and male participants embodied a typified avatar representative of either the "female" or "male" gender in a virtual environment, in which they completed an interpersonal (generosity, as a measure of prosociality) and intertemporal (delay discounting) discounting task. Based on several empirical findings evincing that women tend to make more prosocial decisions 19-23 , we expected participants embodying a female avatar to make more prosocial choices than those embodying a male avatar, independent of their own biological sex. An intertemporal discounting task was chosen as a control task, since research has previously demonstrated no gender differences in intertemporal discounting, and other studies have similarly used intertemporal discounting as a control task [bib_ref] The dopaminergic reward system underpins gender differences in social preferences, Soutschek [/bib_ref]. To more directly relate potential effects of virtual embodiment to changes in self-perception, we measured implicit and explicit identification with gendered (or gender-typical) traits, as well as voice pitch before, during, and after avatar embodiment with the hypothesis that the different-gender avatar would reduce gender identification (Tacikowski et al. [bib_ref] Fluidity of gender identity induced by illusory body-sex change, Tacikowski [/bib_ref] and alter voice pitch 11 towards its typified gender. # Methods ## Participants. We recruited a female sample (n = 52) and a male sample (n = 52). Male and female participants were recruited in separate calls (one call specifically looking for "women", i.e., participants that identified as female, and the other call specifically looking for "men", i.e., participants that identified as men) via the student mailing lists of the Economics and Psychology Departments at the University of Zurich, as well as through personal contacts. While we did not specifically ask them about their biological sex, the call for participants was directed towards students with clearly defined gender identity. They were assigned to the gender-swap group (experimental groups) or no gender-swap group (control groups) by sequential counterbalancing, i.e., they were alternatingly allocated to the experimental (different-gender) or the control (same-gender) group as they signed up for the study. This resulted in two experimental groups, referred to as Fem-Male (female participants embodying a male avatar) and Male-Fem (male participants embodying a female avatar), and two control groups, referred to as Fem-Fem (female participants embodying a female avatar) and Male-Male (male participants embodying a male avatar). One female participant and four male participants had to be excluded due to technical issues, resulting in a final sample of n = 26 (23.5 ± 2.3 years) in Fem-Male, n = 25 (23.5 ± 4.7 years) in Fem-Fem, n = 23 (23.7 ± 3.3 years) in Male-Fem, and n = 25 (23.0 ± 2.7 years) in Male-Male. Inclusion criteria were good health (self-assessed and self-reported; we explicitly called for "healthy women/men" in the advertise- Equipment and virtual environment. Participants wore an HTC Vive Pro (Vive™), a head-mounted display with a resolution of 1440 × 1600 pixels per eye and a diagonal field of view of 110°, four Vive Trackers respectively placed on the left and right wrists and ankles, and a Vive Controller attached to the front of the midsection. The trackers and controller served as tracking points to enable tracker-to-body-matching with the avatar. The virtual environment was implemented in the Unity 2018.2 game engine (see https:// unity. com/). The Steam VR package (Steam®) was used to functionally combine the Vive kit with Unity for full body tracking. To enable inverse kinematics (in order to synchronize real movements with virtual movements), we used the Final IK VR Inverse Kinematics package, an animation tool provided by Unity RootMotion. The virtual environment was designed to simulate the real environment; therefore, the virtual room strongly resembled the actual testing room [fig_ref] Figure 1: Virtual environment and avatars [/fig_ref] in which the experiment took place. The main furniture consisted of a desk with a computer, which was used to complete the two discounting tasks. Since real-time mirror reflections have been shown to enhance virtual embodiment [bib_ref] The contribution of real-time mirror reflections of motor actions on virtual body..., González-Franco [/bib_ref] , we added two mirrors to the virtual environment. One large, full body mirror was placed on the wall [fig_ref] Figure 1: Virtual environment and avatars [/fig_ref] and one smaller mirror displaying only the upper body was placed directly behind the computer, so that participants could see the avatar's virtual reflections while completing the tasks on the virtual computer screen. ## Tasks and measures. Interpersonal and intertemporal discounting tasks. Following Soutschek et al. [bib_ref] The dopaminergic reward system underpins gender differences in social preferences, Soutschek [/bib_ref] , we adapted the tasks from Jones and Rachlin 25 to assess generosity and delay discounting. In the interpersonal discounting task, participants chose between a generous and a selfish choice option. If they chose the generous option, they shared a reward of 15 Swiss francs (7.50 each) with a person of varying social distance, whereas the selfish option offered a varying amount (7.50 to 15.50 Swiss francs) to participants. The typically larger amount provided by the selfish option to the participant came at the expense of the other person not receiving any financial reward. Six different social distances (1, 5, 10, 20, 50 and 100) were defined beforehand in a social distance questionnaire. The scale represented the social environment of participants and ranged from 0 to 100, where 0 represented "yourself ", 100 represented a "stranger seen randomly on the street", and 50 "someone you have seen before, but whose name you do not know". They were then asked to name specific persons from their social environment (to whom they did not have a negative relation) for distances 1, 5, 10 and 20. In the intertempo- www.nature.com/scientificreports/ ral discounting task, participants chose between a smaller-sooner or larger-later choice option. Smaller-sooner options varied from 0 to 16 Swiss francs and were paid "today" (i.e., the day of participation). The larger-later option was 16 Swiss francs with varying temporal distances, i.e., to be paid in 1, 10, 20, 50, 90 or 180 days after their participation. To promote honest and accurate decisions according to their true preferences, participants were informed that one trial for each task would be drawn at random at the end of the experiment, and paid out according to their decision. They were also truthfully informed that the other person (either a person they named or a stranger, depending on the social distance of the drawn trial) would be paid the amount specified in the drawn trial. Implicit association task (IAT). The IAT measures the categorization speed and accuracy in conditions involving verbal or pictorial stimuli and is therefore used to assess automatic association strength between two concepts 26 . The Self-Concept IAT combines the contrast of "self " versus "other" with a contrast of interest [bib_ref] Using the implicit association test to measure self-esteem and self-concept, Greenwald [/bib_ref]. In the current study, the contrast of interest was "female" versus "male" to which participants had to associate pictorial stimuli of female and male faces. A positive d score indicates faster sorting of "female" with "self " and "male" with "other", whereas a negative d score indicates faster sorting of "female" with "other" and "male" with "self ". We used the d score as a measure of implicit self-association with a gender. The IAT was implemented in PsyToolkit [bib_ref] PsyToolkit: A software package for programming psychological experiments using Linux, Stoet [/bib_ref] [bib_ref] PsyToolkit: A novel web-based method for running online questionnaires and reaction-time experiments, Stoet [/bib_ref]. Voice recordings. Participants' voices were recorded using a Zoom H6 Hand Recorder (Zoom Corporation, Japan) with an SGH-6 Shotgun Microphone Capsule at a sampling rate of 44.1 kHz at 24 bits. Each participant was stationed at a similar distance to the microphone, indicated by a mark on the virtual floor during the recording (as well as by a mark on the actual floor for the experimenter to confirm the appropriate location). The recording was started manually prior to beginning and upon completion of each task. Each task consisted of five repetitions of nine different made-up two syllable words that appeared on the VR headset, resulting in a total of 45 words. Each item was spelled with four letters and appeared on the screen for 1000 ms, followed by a pause of 500 ms. The applied procedure for the voice recordings is similar to those used in previous studies [bib_ref] Embodiment in a child-like talking virtual body influences object size perception, self-identification,..., Tajadura-Jiménez [/bib_ref] [bib_ref] Virtual embodiment of white people in a black virtual body leads to..., Banakou [/bib_ref] [bib_ref] Body ownership causes illusory self-attribution of speaking and influences subsequent real speaking, Banakou [/bib_ref]. Gender adjectives. We asked participants to rate how strongly they associate themselves with nine adjectives representing typically "feminine" or "masculine" categories 31,32 on a visual analogue scale (VAS). "Female" terms were placed on the left VAS pole and "male" terms on the right, so that the choices on the VAS (ranging from 0 "female" to 1 "male") could be calculated as a measure of explicit self-association with gender. The word pairs were: modest-dominant, social-egoistic, tender-rough, emotional-rational, weak-strong, soft-hard, romanticrealistic, flexible-stubborn, and anxious-unafraid. To measure participants' SVO, we implemented nine game items suggested in Van Lange et al. [bib_ref] Development of prosocial, individualistic, and competitive orientations: Theory and preliminary evidence, Van Lange [/bib_ref] into the virtual environment. For each item, participants made a choice between three different combinations of outcomes to self and other (e.g., option A: 480 points to self and 80 points to other, B: 540 points to self and 280 points to other, C: 500 points to self and 100 points to other). Each of the three combinations corresponded to one of the three SVO. Participant SVO was determined based on if choices corresponded to the same SVO for at least six out of nine items. The SVO task was implemented in VR and participants used the Vive's eye camera to pick their choices. Participants classified as prosocial are assumed to assign a positive value to both their own and others' outcomes, while those classified as individualist exclusively assign weight to outcomes for themselves, and those classified as competitor prioritize the relative advantage of their own outcomes compared to those of others [bib_ref] Social value orientation and intelligence: A test of the goal prescribes rationality..., Van Lange [/bib_ref]. Our aim was to use the SVO as a baseline measure of and control measure for prosociality between the different groups. Movement instructions. Audio instructions requesting participants to move around the virtual room by following predetermined points and to explore their "new" body in the mirrors were played through the Vive's integrated earphones. The experimenter's pre-recorded female voice instructed participants to perform a standardized set of movements, such as to wave to themselves or to alternately lift their legs in front of the mirrors. These tasks were done to call attention to the visuomotor synchrony between the virtual and physical body, and therefore to maximize the sense of embodiment of the virtual avatar (for a transcript of the audio instructions, see the supplementary materials). ## Measure of embodiment. To assess the strength of the embodiment illusion, participants answered an Embodiment Questionnaire, which was adapted from Peck et al. [bib_ref] The effect of gender body-swap illusions on working memory and stereotype threat, Peck [/bib_ref] Participants answered on a virtual VAS ranging from 0 (strongly disagree) to 1 (strongly agree). Debriefing questions. At the end of the experiment, participants answered the following debriefing questions, again on a virtual VAS ranging from 0 (strongly disagree) to 1 (strongly agree). Procedure. The general procedure is illustrated in . After signing the informed consent form, participants completed a Self-Concept IAT on a desktop computer. They then filled out the social distance questionnaire, which was later used for the interpersonal discounting task. In the next step, participants put on the trackers and the head-mounted display, and were asked to complete the first voice recording tasks. Words were presented to them through the head-mounted display and participants read them out loud while being recorded. Subsequently, they performed the gender adjectives task, which was followed by a virtual implementation of the SVO task. After completion, the virtual SVO scene changed into a virtual room, where a typified "female" or "male" avatar was calibrated to the participant's height by matching the physical trackers to the avatar's limbs [fig_ref] Figure 1: Virtual environment and avatars [/fig_ref]. After the instructed movements and exploring their virtual body in the mirror, participants com- www.nature.com/scientificreports/ pleted the first decision task (counterbalanced order within group of either the interpersonal or intertemporal discounting task) on a virtual gesture-reactive screen monitor. A second embodiment induction audio followed, instructing participants to return to the virtually displayed computer, where they repeated the voice recording task. Next, they completed the second discounting task (either intertemporal or interpersonal, depending on the counterbalancing). For each trial in both tasks, participants chose one of the two options by placing their hand over it on the virtual touch screen and then confirming their decision via the continue-button. Following the last trial of the second task, the scene changed to black and participants filled out the embodiment questionnaire, which was implemented and presented virtually through the head-mounted display. After completion, they repeated the gender adjectives task, answered a few debriefing questions, and then completed the voice recording task a third time. Following completion of the voice recording task, participants were asked to take off the headset. Finally, a post virtual embodiment Self-Concept IAT, which was identical to the IAT administered prior to the virtual embodiment, was repeated on the desktop. The entire experiment lasted approximately 75 min. At the end, participants drew two random numbers from 1 to 54 by clicking on a button. The numbers corresponded to one trial in each decision task. Participants were paid according to their decision in that trial. We followed the procedure Soutschek and colleagues 21 used to pay the participants. Interpersonal decision task: If participants' decision was to share in the random trial, we gave them 7.50 Swiss francs in cash and either sent the other 7.50 Swiss francs to the other person via letter or went to the cafeteria to give the money to a stranger (in case the participant shared with someone of a social distance of 50 or 100). If participants chose to keep the money, they were paid the trial-specific amount in cash. Intertemporal decision task: If participants chose the now-option on the randomly selected trial, then they received the trial-specific amount in cash. If they chose the larger later option, we sent them 16 Swiss francs after the corresponding delay via letter mail. Data processing and analysis. Behavioral and questionnaire data were recorded in Unity and preprocessed in R (RStudio, Version 1.2.5033). The statistical analysis of the behavioral data was performed with Matlab R2016 (MathWorks, Natick, MA) and IBM SPSS Statistics 24. For both the interpersonal and the intertemporal decision task, we computed hyperbolic discount functions that captured the discounted subjective value of generous and later rewards as a function of social distance and temporal delay. We fitted hyperbolic discount functions to the individual indifference values in these tasks. Each indifference value was defined as the point at which a participant chooses the generous vs. selfish reward or delayed vs. immediate reward with an equal probability of 50%. We computed indifference values separately with logistic regressions of choices on amount of the selfish reward option (interpersonal discounting task) or amount of the immediate reward option (intertemporal discounting task) separately for each social distance or temporal delay. The beta coefficients of each logistic regression allowed us to determine the selfish or immediate amount at which an individual was indifferent between choice options for a given social distance or delay. Finally, we fitted discount functions to these indifference values (using the function lsqcurvefit in Matlab) for each participant, assuming standard hyperbolic discount functions: where SV social and SV delay are the discounted values of the prosocial and delayed reward options, respectively, whereas k social and k delay represent participant-specific constants measuring the degree of social and intertemporal discounting. In social discounting, a larger k social reflects more selfish decisions with increasing social distance, while in intertemporal discounting, a larger k delay corresponds to more choices of the immediate reward option with longer delays. For the statistical analysis, the estimated parameters were log-transformed to normalize the skewed distributions of the parameters. We note that models of social discounting often include also a free intercept parameter which is sensitive to generosity towards close social others [bib_ref] The dopaminergic reward system underpins gender differences in social preferences, Soutschek [/bib_ref] [bib_ref] Social discounting, Jones [/bib_ref] [bib_ref] Brain stimulation reveals crucial role of overcoming selfcenteredness in self-control, Soutschek [/bib_ref]. However, such a twoparameter model explained the observed data less well (median AIC = 11.4) than the one-parameter discount model (median AIC = 11.0). We therefore report only the results for the best-fitting one-parameter discount model. We also analyzed the behavior in interpersonal and intertemporal decision tasks with model-free mixed generalized linear models (MGLMs). For the interpersonal decision task, we regressed binary choices (0 = selfish option, 1 = prosocial option) on fixed-effects predictors for Sex (female vs. male), Group (experimental vs. control), Social distance, Selfish reward, and all interaction terms. As random effects, we modelled random slopes for Social distance, Selfish reward, and the interaction term in addition to participant-specific random intercepts. Analogously, in the intertemporal decision task, choices (0 = immediate option, 1 = delayed option) were predicted by fixed-effect predictors for Sex, Group, Reward magnitude, Delay, and all interaction terms. We also included participant-specific random intercepts and random slopes for Reward magnitude, Delay, and the interaction term. IAT d scores were computed according to the procedure suggested in Greenwald et al. [bib_ref] Understanding and using the Implicit Association Test: I. An improved scoring algorithm, Greenwald [/bib_ref] The sense of ownership over the virtual body was assessed by Embodiment Questions (Q1 + (1 -Q2) + Q3)/3, the sense of agency by Q4 and sense of embodying a body of the avatar's gender by Q5. The voice recordings of each condition (pre, during, and post) were manually cut to the actual start and end of the task. Due to technical issues during the experimental procedure and noisy recordings, the voice recordings of two female and nine male participants had to be excluded. The recordings were processed with Praat 38,39 in order to extract the mean and standard deviation of the pitch for each repetition. Data were tested for normality and then integrated accordingly. The Aligned Rank www.nature.com/scientificreports/ Transformation Analysis of Variance (ART ANOVA) is a non-parametric approach to factorial ANOVA that enables the analysis of main as well as interaction effects [bib_ref] The aligned rank transform for nonparametric factorial analyses using only anova procedures, Wobbrock [/bib_ref]. We calculated ART ANOVAs to evaluate the results of the questionnaires and gender adjectives because the data were not normally distributed. # Results Discounting tasks. Interpersonal discounting task. To assess whether the body swap manipulation altered the generosity of female and male participants, we analyzed the log-transformed individually estimated discount factors k social with a 2 (Sex: biological sex male or biological sex female) × 2 (Group: experimental (different-gender avatar) or control (same-gender avatar)) ANOVA. This analysis yielded a significant main effect of Group, F(1, 95) = 4.65, p = 0.03, η p 2 = 0.047, indicating steeper social discounting in the different-gender (mean log-k social = − 2.10) than in the same-gender avatar (mean log-k social = − 1.85) group [fig_ref] Figure 3: Interpersonal discounting task [/fig_ref]. Contrary to our hypothesis, however, we observed no evidence that the impact of the gender-swap manipulation differed between female and male participants, F(1, 95) = 0.32 , p = 0.57, η p 2 = 0.003. Therefore, swapping into a different-gender (experimental group) rather than a same-gender (control group) avatar rendered behavior more selfish, independent of the participants' or the avatar's gender. This finding was further substantiated by the MGLM regressing binary choices (0 = selfish choice, 1 = prosocial choice) on predictors for Sex, Group, Social Distance, Selfish Reward, and the interaction effects [fig_ref] Table 1: SVO classification [/fig_ref]. Participants made more selfish choices with increasing social distance, β = − 3.14, t(77) = 13.13, p < 0.001, and with increasing magnitudes of the selfish reward option, β = -1.62, t(94) = 7.63, p < 0.001. They were less sensitive to the magnitude of the selfish reward in the experimental groups compared with control groups, β = 0.66, t(85) = 2.22, p = 0.03 [fig_ref] Figure 3: Interpersonal discounting task [/fig_ref]. Taken together, these findings suggest that choices are less generous in different-gender avatars than in same-gender avatars. Intertemporal discounting task. A 2 (Sex) × 2 (Group) ANOVA on log-transformed hyperbolic discount parameters k delay revealed no evidence for significant main effects of Sex or Group, both F < 1, both p > 0.58, η p 2 = 0.047. However, we observed a significant Sex × Group interaction, F(1, 95) = 5.29, p = 0.01, η p 2 = 0.067, suggesting that the gender-swap had dissociable effects for female and male participants: Female participants engaged in greater delay discounting (i.e., chose more smaller sooner rewards) when they swapped into male bodies than into female bodies, t(46) = 2.06, p = 0.04, Cohen's d = 0.61, while we observed no significant effects for male participants, t(49) = 1.59, p = 0.12, Cohen's d = 0.45 [fig_ref] Figure 4: Intertemporal discounting task [/fig_ref]. This result is supported by the findings of a MGLM, revealing a significant Sex × Group interaction, β = − 1.65, t(73) = 2.03, p = 0.046 . The effect of the gender swap on choices was stronger for female than male participants. Interestingly, we also observed a significant Group × Reward magnitude interaction, β = 0.96, t(115) = 2.11, p = 0.04, such that participants were less sensitive to immediate rewards (indicated by a less steep logistic curve describing the relationship between choices www.nature.com/scientificreports/ and reward magnitude) in the different-gender than in the same-gender group [fig_ref] Figure 4: Intertemporal discounting task [/fig_ref]. Taken together, female participants chose smaller sooner rewards more often when making intertemporal choices in a male compared to a female avatar, and both were less sensitive to immediate rewards in the body of a different-gender avatar. IAT. According to the pre IAT d score, participants overall showed little (± 0.2) automatic association for their own gender before the experiment. To examine whether implicit self-association changed, we analyzed the d scores with a 2 (Time: pre and post) × 2 (Sex) × 2 (Group) repeated measures ANOVA, with Sex and Group as between-participant variables and IAT d scores as a within-participant variable. The results showed a significant effect for Time, F(1, 95) = 4.14, p < 0.05, η p 2 = 0.042, indicating that the pre scores differed from the post scores, as well as a significant Time × Sex interaction, F(1, 95) = 7.09, p < 0.01, η p 2 = 0.069, reflecting that d scores differed between female and male participants. However, the Time × Group interaction was not significant, F(1, 95) = 0.20, p = 0.66, η p 2 = 0.002, nor was the Time × Sex × Group interaction, F(1, 95) = 0.07, p = 0.79, η p 2 = 0.000. In contrast to our expectations, participants in the different-gender avatar thus did not seem to align implicit associations between their gender self-concept and gender of the avatar after the body swap. While the main effect of Sex is expected (i.e., that female participants show positive d scores, i.e., faster sort "self " with female, and male participants show negative d scores, i.e., faster sort "self " with male"), we assume that the main effect of Time merely reflects a repetition effect that commonly occurs when taking the IAT. Mean IAT d scores are summarized in [fig_ref] Figure 3: Interpersonal discounting task [/fig_ref]. ## Gender adjectives. To assess explicit associations with gender, we asked participants to attribute gender adjectives to themselves. To verify the measures, we first tested whether the pre measure differed between female and male participants, irrespective of which avatar they later embodied. As expected, female participants initially rated themselves as more female, Mdn = 0. SVO task. We used the SVO task to assess baseline prosociality in the four groups of participants [fig_ref] Table 1: SVO classification [/fig_ref]. There was no significant evidence for unequal distributions of social value orientations among the four groups, Pearson's Chi-squared test, χ 2 (9,99) = 16.87, p = 0.051, suggesting that baseline differences in prosociality should not explain group differences, even though the results almost reached significance. Furthermore, we note that 42. An explorative 2 (Sex) × 2 (Group) ART ANOVA did not reveal a significant main effect for Sex, F(1, 95) = 0.01, p = 0.926, η p 2 = 0.001, but did for Group, F(1, 95) = 6.12, p < 0.05, η p 2 = 0.061. Participants who swapped into a different-gender avatar felt more like their decisions were influenced by the virtual body than those who swapped into a samegender avatar [fig_ref] Figure 6: Debriefing questions [/fig_ref]. The Sex × Group interaction was not significant, F(1, 95) = 1.20, p = 0.275, η p 2 = 0.013. Perceived similarity to physical body. On average, participants did not seem to perceive the virtual body to look similar to their physical body, Mdn = 0.17, IQR = 0.28. Again, we conducted an explorative 2 (Sex) × 2 (Group) ART ANOVA that revealed no significant main effect for Sex, F(1, 95) = 0.17, p = 0.681, η p 2 = 0.002, but did for Group, F(1, 95) = 32.12, p < 0.001, η p 2 = 0.253. Participants who swapped into different-gender avatars reported lower perceived similarity to their physical appearance than participants who swapped into a same-gender avatar [fig_ref] Figure 6: Debriefing questions [/fig_ref]. The Sex × Group interaction was not significant, F(1, 95) = 0.47, p = 0.493, η p 2 = 0.005. ## Liking of the virtual body. Participants were neutral about the virtual avatar's looks [fig_ref] Figure 6: Debriefing questions [/fig_ref] # Discussion The current study investigated the influence of a virtual gender swap on interpersonal and intertemporal discounting tasks, as well as on implicit and explicit identification with gendered (or gender-typical) traits. We expected that embodiment of a female avatar would stimulate generosity, while embodiment of a male avatar would increase selfish choices on the interpersonal discounting task, independent of participants' own biological sex. Initially having included the intertemporal discounting task as a control measure, in line with Soutschek et al. [bib_ref] The dopaminergic reward system underpins gender differences in social preferences, Soutschek [/bib_ref] , we expected the virtual gender swap to exert no discernible influence on delay discounting. The results suggest three main findings: First, regardless of their biological sex, participants made more selfish choices in the interpersonal discounting task when they embodied a different-gender avatar (i.e., female participants in a male avatar or vice versa). Second, women in different-gender avatars in particular chose options providing smaller but sooner rewards more often in the intertemporal discounting task, thereby demonstrating increased delay discounting. Third, there was no evidence that the different-gender swap illusion in our setup altered either implicit or explicit identification with gendered (or gender-typical) traits. People make more selfish choices during gender swap. In line with the Proteus effect 7 , the phenomenon in which embodiment of virtual surrogates can elicit behaviors effects that reflect the stereotyped identity characteristics of the embodied avatar, we hypothesized that participants would adjust their decision-making on the two discounting tasks according to the gender-stereotypical characteristics of the avatar's gender. Specifically, we expected participants embodying a female avatar to more often select generous options, while those embodying a male avatar were presumed to engage in more selfish decision-making. Contrary to this hypothesis, however, we did not find a main or interaction effect of the avatar's gender on the interpersonal discounting task. Although a recent meta-analysis reported generally reliable small to medium effect sizes for the Proteus effect [bib_ref] Avatar characteristics induce users' behavioral conformity with small-to-medium effect sizes: A meta-analysis..., Ratan [/bib_ref] , other studies reported no or even opposite effects. For example, Sherrick et al. [bib_ref] The role of stereotypical beliefs in gender-based activation of the proteus effect, Sherrick [/bib_ref] found that participants who adopted the role of a male (versus female) protagonist in an interactive fictional story were less likely to select masculine behaviors, but would instead display behaviors stereotypically associated with being female. However, studies focusing on the Proteus effect often differ in their methodological implementations, ranging from the type of virtual embodiment (i.e., immersive VR or two-dimensional desktop avatars) to the choice of behavioral tasks. Furthermore, many potentially moderating factors, such as the strength of belief in stereotypes or identification with the surrogate avatar could effectively alter behavior as well. Therefore, further research examining user-avatar dynamics within the context of different experimental paradigms will help in providing more detailed insight and further credence to the specific particulars of the Proteus effect. The data are separated by Sex and Group; median and interquartile ranges are displayed. Ownership: ART ANOVA revealed a significant Sex × Group interaction, driven by female participants, who experienced less ownership over their virtual bodies than male participants when swapping into different-gender avatars. Agency: ART ANOVA revealed significant effect for Group: Participants who swapped into a different-gender avatar reported lower agency than those who swapped into a same-gender avatar. Gender of Avatar: ART ANOVA revealed significant effect for Group: Participants who swapped into a different-gender avatar reported lower sense of the avatar's gender than those who embodied a same-gender avatar; significant difference for Group in agency and sex of avatar. www.nature.com/scientificreports/ Despite the absence of the hypothesized main effect of the avatar's gender on interpersonal discounting in our study, we did find an effect of different-gender avatar on generosity. Specifically, both female and male participants chose more selfish options when they embodied a different-gender avatar. In line with this, the same participants were less sensitive to the immediate amount they could receive in intertemporal decisions, which converges with the increased selfish behavior: We assume that the less selfish behavior in the same-gender avatar groups lead to a stronger focus on the amount when deciding to share versus not to share. Though the current study did not experimentally manipulate or individually correlate potential modulators, it is possible that the reduced sense of agency or increased sense of dissimilarity experienced by participants in the different-gender avatar could have elicited more selfish decisions. Although we can currently only speculate about the underlying mechanisms, the reduced sense of agency reported by the different-gender avatar participants could have constituted a driving factor for choice behaviors on the interpersonal discounting task. The sense of being in control of our own actions implies an inherent sense of personal responsibility that rests on voluntary choice, which may in turn promote moral preferences. Previous studies have demonstrated that both a stronger sense of ownership and agency are associated with strengthened moral identity [bib_ref] Experience of agency and sense of responsibility, Moretto [/bib_ref] , while performance of immoral actions or moral disengagement are conversely associated with lower perceived agency [bib_ref] Social cognitive theory: An agentic perspective, Bandura [/bib_ref] [bib_ref] Only giving orders? An experimental study of the sense of agency when..., Caspar [/bib_ref]. Furthermore, several studies have examined the link between the sense of responsibility and sense of agency retrospectively [bib_ref] Metacognition of agency, Metcalfe [/bib_ref] [bib_ref] Awareness of action: Inference and prediction, Moore [/bib_ref]. Metcalfe and Greene 49 suggest that participants are very good at judging whether their agency is high or not, and subsequently assess their performance in a task accordingly: They found that the sense of responsibility for both bad and good outcomes in the task was decreased when the sense of agency was low and increased when it was high. Therefore, the reduced generosity on the interpersonal discounting task of participants in the different-gender avatar condition could potentially be explained by diminished personal responsibility that accompanied the decreased sense of agency reported in these groups. However, one could also expect an effect in the other direction, where participants who behaved more selfishly in the interpersonal task may have subsequently experienced a diminished sense of agency as a coping method to distance themselves from their behavior, which would be an interesting question for future research. In addition to the diminished sense of agency, participants in the different-gender avatar also reported larger perceived dissimilarity to the virtual body, which could further account for the reduced generosity observed. Previous research has proposed that the so-called "online disinhibition effect" 51 , which maintains that use of a pseudonym or an avatar, rather than one's real name or picture, in online (or digital) environments increases perceived anonymity. This perceived anonymity in turn may create a compartmentalized "online self ", which can perceivably exist in an imaginative online dimension that stands separate from the offline (in-person) self. Furthermore, the behavior of this online identity might differ quite strongly from a person's offline behavior, including decreases in personal responsibility and moral cognitive processing. According to , acting under the guise of a virtual identity can promote a dissociative anonymity that encourages individuals to separate their digital behavior from their in-person characteristics, thereby facilitating a disinhibition of behaviors that would i.e., participants who swapped into a different-gender avatar felt more like their decisions were influenced by the virtual identity than those who swapped into a same-gender avatar. Similarity: Explorative ART ANOVA revealed a significant effect for Group, participants who swapped into different-gender avatars reported lower perceived similarity to their physical appearance than those who swapped into a same-gender avatar. Liking: No significant effects were found. www.nature.com/scientificreports/ normally be more suppressed by the in-person self. Therefore, participants who embodied the perceived dissimilarity between the virtual different-gender avatar and the physical own body may have facilitated an increased sense of dissociative anonymity that could have concomitantly reduced the sense of personal responsibility (i.e., averting it to the virtual different-gender body), thereby providing a guise to engage in more selfish behaviors. It could further be speculated that this form of altered self-identification could have additionally temporarily modulated the perceived degree of closeness and the relations towards the specific individuals that participants reported on their respective social distance lists. Therefore, the interpersonal relations to these people could have shifted, which might have encouraged greater social discounting behaviors. Lastly, it should be noted that we did not replicate the previously reported differences between men and women in the interpersonal discounting task, even when they embodied their own gender 21 , which could suggest that participants might use different strategies for this task in embodied VR, as compared to a non-mediated environment. However, our control SVO measure also did not reveal gender differences in our sample's baseline prosociality. Accordingly, it is reasonable to expect that men and women of same-gender avatars in our study would similarly not differ in their social discounting behaviors. Women in male avatars chose smaller sooner options more often. Although our initial hypothesis stated no gender differences in intertemporal decision-making, it should be noted that the empirical findings surrounding gender differences on this task are somewhat mixed, as another study found that men rather displayed stronger preferences for immediate (i.e., smaller sooner) over larger later rewards than women 52 . Our findings evince partial support for this alternate view; specifically, we found that women tended to prefer immediate over delayed rewards when embodying a male avatar, although no such differences were found in men. While gender differences have been described in potentially related psychological constructs 20 , such as impulsivity, only very few studies have looked at the effect of sex in the intertemporal discounting paradigm. Contrary to Soutschek et al. [bib_ref] The dopaminergic reward system underpins gender differences in social preferences, Soutschek [/bib_ref] , who found no delay discounting differences between men and women, Dittrich and Leipold 52 reported that female participants were more patient, whereas male participants tended to choose the immediate payment more often. This held true especially if the interest rate of the larger later option was neither too high nor too low. Our results showing that women in male avatars more often chose smaller sooner rewards are therefore in line with previous findings of gender-specific effects on this task. As predicted, women in male virtual bodies in our study therefore seemed to follow the Proteus effect, whereby they assimilated stereotypically male characteristics (e.g., impulsivity) to engage in greater delay discounting behaviors. In line with our interpretations of the interpersonal discounting task, we here similarly assume that the lower sense of agency and ownership in the women embodying a male avatar could have elicited a decreased sense of responsibility [bib_ref] Experience of agency and sense of responsibility, Moretto [/bib_ref]. However, it is unclear why we did not observe men in a female avatar to choose larger later rewards more frequently. Furthermore, although we did not administer any baseline control measures to account for potential gender differences in the intertemporal discounting task, it is plausible that the men and women in our sample, similarly to the SVO task, may not have differed in potentially contributing moderating psychological constructs, such as self-control. Therefore, the observed effect specific to women in male avatars may hint at a particularly instrumental effect of virtual embodiment on delay discounting in this group. Lastly, it is worth noting that many participants in the current study were students recruited from the Economics Department. Research has demonstrated that economics students can show systematically different behaviors than non-economic students, such as demonstrating strong self-interest and acting more selfishly [bib_ref] Social preferences in the lab: A comparison of students and a representative..., Cappelen [/bib_ref] [bib_ref] Selfish and indoctrinated economists?, Frey [/bib_ref]. Although intertemporal discounting cannot directly be compared to selfish decision-making, it is nevertheless possible that economics students possessed either prior knowledge of the intertemporal discounting task or exhibit moderating characteristics that could have substantially influenced their choices. Gender swap does not alter identification with gendered traits. Previous research suggested a change in implicit and explicit identification with gendered (or gender-typical) traits after a gender swap illusion [bib_ref] Fluidity of gender identity induced by illusory body-sex change, Tacikowski [/bib_ref]. Furthermore, Mello et al. [bib_ref] Wearing same-and opposite-sex virtual bodies and seeing them caressed in intimate areas, Mello [/bib_ref] reported that swapping into a virtual body of the opposite gender led heterosexual women and men to rate intimate touch as more pleasant and erogenous when it came from a toucher of the same sex, suggesting that gender-specific cognitive representations of the (sexual) self in heterosexual participants changed during the swap. In the current setup, however, we did not find any effect of the gender swap on either explicit traits (as measured using self-attribution of gender stereotyped adjectives) or implicit genderidentification (as measured by the IAT). Furthermore, results also did not provide evidence for a change in participants' voices, which stands in contrast to previous studies that had demonstrated voice changes following body swaps [bib_ref] Illusory ownership of a virtual child body causes overestimation of object sizes..., Banakou [/bib_ref] [bib_ref] Embodiment in a child-like talking virtual body influences object size perception, self-identification,..., Tajadura-Jiménez [/bib_ref]. It is important to note that the absence of an effect in our study cannot simply be explained by a lack of illusory embodiment. First, although ownership ratings were rather low, they were comparable to those reported by Peck et al. [bib_ref] The effect of gender body-swap illusions on working memory and stereotype threat, Peck [/bib_ref] , who observed a Proteus effect in female participants performing a working memory task in same-gender versus different-gender avatars. Second, the agency ratings in our sample were medium to high. Third, and most importantly, participants strongly agreed that they felt embodied in the avatar's gender. This suggests that, indeed, participants transiently embodied both the different-gender and the same-gender avatar. We therefore believe that the lack of effect on implicit and explicit measures of gender identity might rather be linked to differences in our design compared to previous studies. To the best of our knowledge, there is only one study that directly assessed the effects of a gender swap illusion on implicit and explicit gender identity [bib_ref] Fluidity of gender identity induced by illusory body-sex change, Tacikowski [/bib_ref]. Using an explicit gender adjectives task and an auditory version of the IAT, the authors reported a more balanced identification with both genders and less gender-stereotypical beliefs about one's own personality characteristics. However, a few important differences should be considered between this recent study and our study. First, the setup used by Tacikowski et al. [bib_ref] Fluidity of gender identity induced by illusory body-sex change, Tacikowski [/bib_ref] consisted of a video-based gender swap paradigm, in which participants viewed the physical body of another person through the head-mounted display. www.nature.com/scientificreports/ Furthermore, they additionally employed visuotactile stimulation (i.e., synchronicity between the felt touch on the own body and seen touch on the video-based virtual body) to strengthen the embodiment illusion. In contrast, our setup was based on a computer-generated gender swap (virtual avatar), where the illusion was induced by visuomotor synchrony (real-time full body tracking) and enhanced with audio instructions. Virtual avatars do not (yet) emulate physical bodies in such detail that they could be comparable to physical bodies; therefore, it is plausible that the more naturalistic video-based setup used by Tacikowski et al. [bib_ref] Fluidity of gender identity induced by illusory body-sex change, Tacikowski [/bib_ref] could have enhanced the perceived effects of the gender swap. Second, their design was a within-subject design, in which participants experienced both genders, each stimulated both synchronously (to induce embodiment) and asynchronously (to prevent embodiment). While not empirically confirmed, such repeated conditions could have made participants more aware of the study's purpose, which may then have facilitated the change in explicit identification with gendered (or gender-typical) traits. This potential issue is the reason we did not use a within participant design. Third, while design choices would not necessarily justify the implicit measure changes in the study by Tacikowski et al. [bib_ref] Fluidity of gender identity induced by illusory body-sex change, Tacikowski [/bib_ref] , it is important to note that they measured the IAT during rather than after the gender swap, which might be a reason for the different findings. Although some studies have demonstrated sustained changes in implicit biases on the IAT [bib_ref] Virtual embodiment of white people in a black virtual body leads to..., Banakou [/bib_ref] , the few studies that have examined the longevity of such changes had employed IAT measures focused on social cognition, rather than self-related cognition. Finally, we did not find any effects of gender swap on voice pitch. Although previous studies have shown such effects, methodological design choices here too differed in that they concurrently presented auditory vocal cues in multimodal settings [bib_ref] Body ownership causes illusory self-attribution of speaking and influences subsequent real speaking, Banakou [/bib_ref]. In particular, participants heard a voice (the avatar's) while synchronously seeing the avatar's lips (in a virtual mirror) move in correspondence with the spoken utterances (lip sync). Changes in the participants' vocal pitch were associated with visuomotor synchrony of the body together with vocal vibrotactile cues [bib_ref] Body ownership causes illusory self-attribution of speaking and influences subsequent real speaking, Banakou [/bib_ref] and with visuomotor synchrony of the body and a modulation of participants' own voice 11 , but not with non-vocal visuotactile stimulation to the body [bib_ref] Embodiment in a child-like talking virtual body influences object size perception, self-identification,..., Tajadura-Jiménez [/bib_ref]. In contrast, no concurrent auditory-vocal cues were presented in our study. Our findings thus suggest that embodiment of a stereotyped body might not be enough to result in pitch shifts, but that additional multisensory stimulation using vocal cues seem necessary. Further research is needed to fully investigate the link between vocal cues and illusory embodiment, in particular, whether a subjective link between the embodied avatar and the heard voice is necessary to result in pitch shifts, and whether the auditory feedback is sufficient or the audiovisual lip sync is important. # Limitations and outlook. There are some important limitations that should be considered. In line with Peck et al. [bib_ref] The effect of gender body-swap illusions on working memory and stereotype threat, Peck [/bib_ref] [bib_ref] Inducing and mitigating stereotype threat through gendered virtual body-swap illusions, Peck [/bib_ref] , the current study simplifies gender to a binary concept that does not necessarily correspond to the gender identity and expression on a non-binary continuum in reality. Additionally, participants embodied generic, typified avatars during the swap and did not experience a particularly strong embodiment illusion. Personalized avatars and a high degree of immersion (i.e., matching the shape of the avatar's 3D body to the participant's physical body) that increase the similarity between the virtual and physical body could enhance the sense of ownership and agency over the virtual body [bib_ref] The sense of embodiment in virtual reality, Kilteni [/bib_ref] [bib_ref] The Impact of avatar personalization and immersion on virtual body ownership, presence,..., Waltemate [/bib_ref]. Even though non-personalized male and female avatars have previously elicited the Proteus effect [bib_ref] The effect of gender body-swap illusions on working memory and stereotype threat, Peck [/bib_ref] [bib_ref] Inducing and mitigating stereotype threat through gendered virtual body-swap illusions, Peck [/bib_ref] , future studies incorporating advanced technological possibilities could examine the influence of personalized avatars in comparison to generic ones. Increasing similarity and immersion between the avatar and the physical body could potentially lead to increased liking of the avatar's appearance, and therefore to a strengthened embodiment illusion. On the flip side, we hypothesized that the sense of dissimilarity between the own and virtual body could have reduced participants' sense of agency and therefore influenced decision-making. We argued that embodying the avatar of a different gender enhanced this sense of dissimilarity; however, future research should examine whether this effect is specific to gender differences, or whether embodying a virtual body that differs in other characteristics (e.g., age) could similarly affect decision-making. Although we used the SVO as a measure of prosociality, we did not administer a non-VR based baseline measurement of participants' behavior in the interpersonal and intertemporal decision tasks to which we could have compared their behavior during the virtual body swap. Additionally, it should be noted that the interpersonal and intertemporal discounting tasks were both administered during the virtual embodiment session, while the implicit and explicit gender identification measures were assessed prior to and following the VR session, rather than during the virtual embodiment. Although studies have demonstrated effects of VR embodiment over an extended time (e.g., one week 16 ) following virtual embodiment, future studies assessing such tasks both during the virtual embodiment session, as well as assessment of the longevity of such effects, would be very informative. # Conclusion The present research provides a first investigation of how a virtual gender swap affects social and delay discounting in women and men. Our results demonstrate that both women and men tend to engage in more selfish behaviors when virtually swapping genders, and that women embodying male avatars engage in more delay discounting. Together, the current study suggests that a virtual gender swap can selectively and differentially affect the choice behavior of women and men. Technological advancements have increased the availability of virtual online worlds. While two-dimensional interactive spaces have been commonplace in our daily lives for several decades, the recent proliferation of consumer-level head-mounted displays and VR systems are paving the way for the significantly prominent role of immersive VR in our interpersonal lives. As an increasing number of individuals look to VR as a novel tool for advancing user experiences in health, education, and entertainment, understanding how embodied virtual identity manipulations impact behavior is of particular importance. ## Data availability The data is available on the Open Science Framework, see https:// osf. io/ 2czqk/. Received: 2 March 2021; Accepted: 16 July 2021 [fig] Figure 1: Virtual environment and avatars. (A) Virtual (top) and real (bottom) room in which participants moved around and the experiment took place. (B) Typified female (top) and male (bottom) avatar moving in front of the full body displaying mirror. Scientific Reports | (2021) 11:15376 | https://doi.org/10.1038/s41598-021-94869-z [/fig] [fig] SVO task: People differ in how they evaluate outcomes in relation to themselves versus others. Messick and McClintock 33 described three distinct Social Value Orientations (SVO): prosocial, individualist, and competitor. [/fig] [fig] Figure 3: Interpersonal discounting task. (A) Log-transformed estimated discount factors k social (log-k) by Sex and Group, median and interquartile ranges are displayed. A larger k reflects more selfish decisions with increasing social distance. ANOVA on log-ks revealed a significant difference for Group: Participants in different-gender avatars made less prosocial decisions with increasing distance than participants in same-gender avatars. (B) The probability of choosing a prosocial option as a function of z-transformed magnitude of selfish reward is displayed. MGLM on predictors Sex, Group, Social Distance, and Selfish Reward, revealed a significant Group × Selfish Reward magnitude interaction: Participants in different-gender avatars were less sensitive to the magnitude of the selfish reward than same-gender avatars.Scientific Reports | (2021) 11:15376 | https://doi.org/10.1038/s41598-021-94869-z [/fig] [fig] Figure 4: Intertemporal discounting task. (A) Log-transformed estimated discount factors k delay (log-k) by Sex and Group, median and interquartile ranges are displayed. A larger k reflects more immediate choices with longer delay. ANOVA revealed a significant Sex × Group interaction, which was driven by female participants: they chose more smaller sooner rewards when they swapped into male bodies than into female bodies. (B) The probability of choosing a delayed option is displayed as a function of the magnitude of the immediate reward. MGLM revealed a significant Group × Reward magnitude interaction: Participants in different-gender avatars were less sensitive to the magnitude of the immediate reward than same-gender avatars as indicated by the less steep logistic curve. effect of Group on social discount rates remains significant when we statistically controlled for baseline differences in SVO,F(1, 92) = 4.92, p = 0.03, η p 2 = 0.051. Measure of embodiment. Sense of ownership. Overall, the sense of ownership was low to moderate among all participants, Mdn = 0.43, IQR = 0.27. A 2 (Sex) × 2 (Group) ART ANOVA revealed no effect of Sex, F(1, 95) = 1.61, p = 0.21, η p 2 = 0.017, nor Group F(1, 95) = 2.33, p = 0.131, η p 2 = 0.024, but a significant interaction of Sex × Group F(1, 95) = 4.76, p < 0.05, η p 2 = 0.048. Female participants experienced less ownership over their virtual bodies than male participants when swapping into different-gender avatars (Fig. 5). Sense of agency. Participants reported moderate to high agency, Mdn = 0.59, IQR = 0.36. A 2 (Sex) × 2 (Group) ART ANOVA yielded no significant main effect for Sex, F(1, 95) = 0.17, p = 0.68, η p 2 = 0.002. There was a significant effect of Group, F(1, 95) = 12.01, p < 0.01, η p 2 = 0.113. Participants who swapped into a different-gender avatar reported a significantly lower sense of agency over their virtual body than participants who embodied the same-gender avatar (Fig. 5). The Sex × Group interaction was not significant, F(1, 95) = 0.70, p = 0.406, η p 2 = 0.007. Sense of embodying a body of the gender of the avatar. Participants strongly felt that they embodied a body of the avatar's gender, Mdn = 0.81, IQR = 0.35. An ART ANOVA showed no significant main effect for Sex, F(1, 95) = 1.90, p = 0.171, η p 2 = 0.020, but did for Group, F(1, 95) = 29, p < 0.001, η p 2 = 0.235. Participants of both sexes in the experimental group reported lower embodiment of the different-gender avatar than the control participants who embodied a same-gender avatar (Fig. 5). The Sex × Group interaction was not significant, F(1, 95) = 0.35, p = 0.555, η p 2 = 0.004. Voice recordings. We analyzed the recordings with 3 (Time: pre, post, and during) × 2 (Sex) × 2 (Group) mixed and repeated measures ANOVA, with Sex and Group as between-participant variables and Pitch (F0) as a within-participant variable. The assumptions of sphericity were violated as indicated by Mauchly's Test, χ 2 (19.82) = 0.79, p < 0.001; therefore, degrees of freedom were corrected using Huynh-Feldt estimates of sphericity, ε = 0.84. The results showed no main effect for Time, F(1.68, 140.98) = 2.65, p = 0.083, η p 2 = 0.031, and revealed no Time × Group interaction, F(1.68, 140.98) = 0.25, p = 0.744, η p 2 = 0.003, Time × Sex interaction, F(1.68, 140.98) = 0.32, p = 0.688, η p 2 = 0.004, nor a Time × Sex × Group interaction, F(1.68,1 40.98) = 0.18, p = 0.794, η p 2 = 0.002. As expected, the mean pitch frequency was higher in female compared to male participants (as reflected by the between-participant effect of Sex, F(1,84) = 294.30, p < 0.001, η p 2 = 0.778). However, participants did not seem to differentially change their voice after swapping into the body of a different-gender or a same-gender avatar, as we found no significant between-participant effect for Group F(1, 84) = 0.08, p = 0.755, η p 2 = 0.001, nor a Sex × Group interaction, F(1,84) = 0.33, p = 0.570, η p 2 = 0.004. Debriefing questions. Influence of the virtual body's identity. The agreement on whether the virtual avatar's identity influenced participants' decisions on the tasks was generally low, Mdn = 0.25, IQR = 0. [/fig] [fig] Figure 5: Embodiment questionnaire. Ownership, sense of agency and embodiment of avatar gender ratings. [/fig] [fig] Figure 6: Debriefing questions. Debriefing question ratings. The data are separated by Sex and Group; median and interquartile ranges are displayed. Identity: Explorative ART ANOVA revealed a significant effect for Group, [/fig] [table] Table 1: SVO classification. SVO social value orientation by van Lange et al.34 . [/table]

Stem cell secretome derived from human amniotic fluid affords neuroprotection in an ischemic model Human amniotic fluid stem cells (hAFSCs) are growing in interest; yet, little is understood about their secretome and neuroprotective actions in different diseases, including stroke. When stem cells are grown in vitro, they release an array of cytokines and growth factors that can stimulate neuroprotective processes. Furthermore, administering secretome rather than cells may be a safer route for patients who are at risk for rejection, promoting innate restorative processes. Current literature implicates that the miRNA contents of such secretome, more specifically exosomes, may regulate the effectiveness of secretome administration. In this review, we explore what factors may promote pro-survival and pro-apoptotic pathways after the administration of hAFSCs-derived secretome in ischemic models. S troke remains a significant cause of death and chronic disability across the world. The two major subcategories of stroke are ischemic and hemorrhagic. Ischemic stroke results from occlusion of an artery that supplies blood to the brain,while hemorrhagic stroke is caused by a ruptured artery or an unusual vascular design.Recent data have shown that the majority of stroke incidences are ischemic in nature.A disruption in the cerebral blood supply has devastating effects, such as oxidative stress, tissue damage, inflammation, and eventually death of resident neurons.Currently, stroke treatment is limited to tissue plasminogen activator (tPA), which has a narrow therapeutic window.Furthermore, tPA has other limitations such as cerebral ischemia/reperfusion (I/R) injury, which can cause severe disability and even mortality.Due to the unfavorable outcomes of current treatment options like I/R injury, treatments that target secondary inflammation pathways, warrant the need to be investigated. ## In vitro methods provide insight into molecular actions In vitro mechanisms are useful in understanding the cellular and molecular actions of pathologies like stroke. Although not complete, in vitro models are critical for the translation of a treatment to a clinical setting. Furthermore, they allow for a lost-cost solution to develop in vivo models, which further improve the clinical translation. O x y g e n a n d g l u c o s e d e p r i v a t i o n models (OGD) are an accurate way to simulate cerebral ischemia and cell damage that occur during stroke in vitro.This model is accomplished by placing cells in a glucose-free medium while simultaneously prohibiting oxygen.Further, the OGD model undergoes reperfusion in normal oxygen conditions.Amniotic Fluid: A Source of Regenerative Properties Current research in regenerative medicine suggests that stem cells from amniotic fluid may be promising. In adult stem cells, epigenetic changes may be preserved, thus limiting their use for specific applications.However, fetal stem cells are less differentiated when collected, and could be used more broadly. Furthermore, there are limited ethical concerns since the amniotic fluid is collected during amniocentesis and cesarean section.To further elucidate the therapeutic potential of amniotic fluid cells, signal transduction pathways triggered by human amniotic fluid stem cells (hAFSCs)-derived secretome in an I/R in vitro model were studied by Western blot analysis. In addition, the expression of miRNA within the exosomes of the conditioned medium was also studied. hAFSCs-derived secretome was observed to initiate pro-survival and anti-apoptotic mechanisms.After microRNA analysis in the exosomes, it was found that 16 miRNAs were overexpressed and involved in the management of signaling pathways.The pathways relating to I/R, including neurotrophin signaling, and those associated with neuroprotection and neuronal death, were of particular interest.The evidence compiled proposes that hAFSCs-conditioned medium commences neuroprotective actions within in vitro ischemia models. These observed effects can be achieved through adjusting and initiated pro-survival processes, partially, due to the secreted miRNAs. ## The role of secretome and exosomes in neuroprotection The amniotic fluid serves as a protective liquid that surrounds the fetus providing support and critical nutrients to aid in the development of the embryo.The fluid consists of main water, cells, and other chemical elements.The cells are primarily fetus-derived (respiratory, epithelial, urinary, and intestinal tract), as well as connective tissues and amniotic membranes. In addition, amniotic fluid possesses other cell subtypes such as amniotic, fibroblastic, and epithelioid, differing in prevalence depending on the age of gestation.Amniotic fluid mesenchymal stem cells (AFMSCs) are of particular interest given their potential for therapeutic applications. AFMSCs have shown the ability to differentiated broadly toward chondrogenic, osteogenic, and adipogenic lineages, consequently presenting as a viable candidate for regenerative and therapeutic methods.However, there remains a concern on the ability of engrafted stem cells to develop into the specific cell type of the damaged area. The mechanism of action has been reported broadly,yet some of the exogenous cells that persist into the regenerated tissue do not completely explain the regenerative effects. Potentially, the insufficient integration may be due to a degree of differentiation that took place in vitro before transplant. Furthermore, it is worth noting that only some of the cell populations may undergo differentiation. This group of cells may elicit poor immune-activating responses but may recruit nearby progenitor cells to repair the injured tissue. Several studies support this theory, highlighting the protective ability of the conditioned medium in the regeneration of damaged areas where chronic inflammation persists.Arising from previously described theories, efforts have been targeted towards stem cells secretome, and the effects it can have on stimulating regenerative pathways in damaged areas.Within conditioned medium, there exists extracellular vesicles (EV) and other soluble factors. More recently, EV have been implicated in upregulating mesenchymal stem cells (MSCs) regenerative effects,however, this mechanism is not fully understood. Exosomes are a subcategory of microvesicles, that have gained interests due to their ability to communicate from cell-to-cell, store biological information, serve as biomarkers, and their potential for regeneration and protection of the nervous system.Exosomes are the result of fusing with the plasma membrane and releasing its contents outside of the cell.The contents of an exosome usually consist of miDNA, DNA, proteins, carbohydrates, and lipids.Since exosomes are relatively simple in structure and small in size, they can cross the blood-brain barrier and present novel approaches for treatment delivery in cerebral injuries. Therefore, exosomes provide another therapeutic route besides cell transplanatation. ## Mitochondria direct apoptosis Deficits relating to I/R injuries can be traced back to the mitochondria. For example, when mitochondria are unable to produce a strong proton gradient, their ability to produce energy suffers and thus affect other organelles like the endoplasmic reticulum in calcium uptake. Furthermore, oxidative stress can result from changes in the mitochondria leading to a building up of reactive oxygen species.As a result, irregular mitochondria often initiate pro-apoptotic factors in the cytosol or nucleus leading to apoptosis. ## Human amniotic fluid stem cells-secretome initiates pro-survival and pro-apoptotic pathways Following MSCs or MSCs-CM administration, a decrease in apoptosis around the injury site, upregulation of pro-inflammatory factors, enhanced axonal growth, and neurogenesis and amelioration of neurological deficits have been shown [ ].Yet, it is the contents of the secretome that determine the therapeutic capacity.In the CM, a myriad of regenerative properties and growth-stimulating factors exist, which originate from stem cells. Furthermore, proteomic studies demonstrate the presence of several cytokines and growth factors in CM.In fact, MSCs have an unparalleled ability to produce effective exosome-producing cells.Given such evidence, paired with the paracrine hypothesis, which poses that stem cells elicit a positive effect by stimulating resident cells to deliver bioactive molecules and EV, the use of secretome may be more advantageous than MSCs, especially in regard to safety. Recent studies have shown that the release of the secretome from MSCs could attenuate neurological impairments observed in several neurodegenerative diseases such as traumatic brain injury, Parkinson's disease, stroke, and Alzheimer's disease.Conditioned media and secretome demonstrate the capacity to upregulate neurotrophins like BDNF, a vascular endothelial growth factor. Neurorestoration is initiated under the release of nerve growth factors that improve neuronal projections.Several factors support the use of hAFSCs for safe and effective treatment in stroke including highly proliferative, low tumorgenicity, immunogenicity, and anti-inflammatory activity.It was found that CM and more specifically, the exosomal component served an important role in fostering cell survival.Increased survival was accompanied by suppressing cell-death pathways (p75/JNK) and upregulating pro-growth pathways (BDNF/TrkB).In addition, pro-survival pathways PI3K/Akt and Erk5 were increased.miRNA Contents Determines Therapeutic Capacity Within exosomes are proteins, lipids, and regulatory RNAs that can be transferred and alter surrounding cell metabolism.Interestingly, among MSC-derived exosome proteins, higher levels of mBDNF were observed. Previously, several studies demonstrated that exosomes contained various neuronal proteins, all capable of passing the blood-brain barrier.Furthermore, it has been shown that high levels of mBDNF were found in both exosomal fractions and in soluble CM.In addition to previously discussed exosomal contents, numerous miRNAs have been found that can initiate neurorestoration. miRNAs that have been shown to have neuroprotective effects (miR-146a-5p, miR-154-5p, miR-22-3p, miR-23a-3p, miR-27a-3p, miR-29a-3p, and miR-31-5p) were also found in the hAFSC-derived exosomes studied. In particular, miR-146a was previously demonstrated to dampen inflammationand promote new formations of oligodendrocyteswithin the injured perinatal brain. In promoting the growth of nervous tissue, miR-154 was observed in EV derived from astrocytes.In addition, miR-154 has demonstrated to seek out DKK2, causing Wnt signaling activation and β-catenin upregulation,both important pathways in synaptic maintenance and neuronal survival.MiR-22-3p, 23a, and 27a all have roles in inhibiting apoptosis, thus providing neuroprotection.Recent studies suggest that miRNAs are dynamic in their response to ischemic conditions and are present in neuroprotection, such as TNF, Hippo, and PI3K.From the miRNAs that afford neuroprotection, several genes were identified that are involved in modulating apoptosis and inflammation.In the HIF-1 pathway, a systemic deactivating mechanism can be seen from the experimentally based miRNAs and target genes. However, apoptosis, neurotrophins, and PI3K-Akt pathways, all expected to be mostly inactive, were observed taking part in upregulating or downregulating various processes such as apoptosis, proliferation, and inflammation. However, this response is undeviating with the pleiotropic and adaptive nature of miRNAs, since they can regulate the expression of many genes at the epigenetic level.Despite the evidence, more investigation is needed to validate the role miRNAs have in regulating neurogenic pathways following ischemic injury. Furthermore, elucidating possible molecular mechanisms will aid in explaining the protective role miRNAs play in the central nervous system. Finally, it is certainly important to study how hAFSC-derived conditioned media behaves in vivo, as there are many other cells in the central nervous system besides neurons. ## Acknowledgement The figure was created using Adobe Photoshop. ## Financial support and sponsorship Nil. : A graphical abstract demonstrating the capacity of Human amniotic fluid stem cells-derived media to promote neuronal survival after exposed to ischemic conditions Conflicts of interestThere are no conflicts of interest.

The estimation of probability distribution for factor variables with many categorical values For all these simulation settings, the sample spaces are composed of three factor variables ∈ { 1 , 2 , ⋯ , } for = 1,2,3, where denotes the number of categories of . ## Joint uniform distributions Each variable is assumed to have two hidden super categories, which means ∈ { (1) , (2) } where ( Simulation setting (1): Perfectly dependent uniform distribution. Three variables are perfectly dependent and = 0.7 for all . That is, Pr [ 1 × 2 × 3 ] = 0.7/N (∏ (1) ) for the combinations of fine categories 1 × 2 × 3 ∈ ∏ (1) , Pr [ 1 × 2 × 3 ] = 0.3/ N (∏ (1) ) for 1 × 2 × 3 ∈ ∏ (2) , and 0 otherwise, where N (- ) denotes the number of unique combination of fine categories that a certain space can compose. Simulation setting (2): Independent uniform distribution. Three variables are independent, and 1 = 0.9, 2 = 0.8 and 3 = 0.7. That is, for examples, for the combinations of fine categories 1 × 2 × 3 ∈ ∏ (2) , Pr [ 1 × 2 × 3 ] = 0.006/N (∏ (1) ). Simulation setting (10), (11) and (12): Independent uniform distribution with multiple super categories. Three variables are independent, and each variable has 5, 10 and 20 super categories for simulation setting (10), (11) and (12), respectively. The probabilities for super categories are randomly assigned for each implementation: Pr( ( ) ) = (0,1) ∑ Pr( ( ) ) ⁄ , where a denotes an index for a super category. ## Normal distributions In this simulation, a trivariate joint normal distribution is assumed. In order to generate the simulation data, we assume that the marginal populations of the simulation data follow normal distributions with the hidden orders of the categories. We discretize the continuous normal distribution, and the probability that is assigned to each combinatorial cell is assumed to be the hidden true probability of the normal distribution with factor variables. Simulation setting (5): Independent normal distribution. The random sample data are generated from the distribution of 1~N (0,1), 2~N (0,1) and 3~N (0,1). ## Other distributions Simulation setting (6): Additive exponential distribution. Each variable follows the distribution respectively, 1~E xp(1), 1~E xp(1), 2~E xp(1), 2~1 + 1 and 3~2 + 2 . Simulation setting ( ## Supplementary table [fig] Simulation setting 3: : Normal distribution with low correlation. The random sample data are generated from the distribution of ( Simulation setting (4): Normal distribution with high correlation. The random sample data are generated from the distribution of ( [/fig] [table] Table A: Computation time for the simulation study cases. [/table]

Identification of novel high-impact recessively inherited type 2 diabetes risk variants in the Greenlandic population [bib_ref] New genetic loci link adipose and insulin biology to body fat distribution, Shungin [/bib_ref] ## Esm [fig] Figure 1: QQ-plot (a) and Manhattan plot (b) for recessive association analyses of variants on the MetaboChip with type 2 diabetes ESM Fig. 2. Statistical power simulation results for a recessive variant applying an additive model based association test (dotted lines) or a recessive model based association test (solid lines) [/fig] [table] Table 5: Association of ITGA1 rs870992 and LARGE1 rs16993330 with type 2 diabetes and metabolic traits in published genome-wide association studies performed using an additive genetic model Effects are for ITGA1 rs870992 G-allele and for LARGE1 rs16993330 A-allele. Nominally significant p-values are shown in bold. Additive genetic model GWASs were queried through online available summary statistics. Quantile is the fraction of p-values in the specific set of results, which has lower p-value than the queried variant. Type 2 diabetes results were from a recent study by the DIAGRAM Consortium[1] (http://diagram-consortium.org/). Glycemic traits were from MAGIC GWAS [2] (https://www.magicinvestigators.org/). Anthropometric traits were from GIANT Consortium[4][5][6] (http://portals.broadinstitute.org/collaboration/giant/index.php/GIANT_consortium), while lipid results were from GLGC[8] (http://lipidgenetics.org/). IVNT, inverse-normalized transformation. [/table]

Facemasks and ferrous metallurgy: improving gasification reactivity of low-volatile coals using waste COVID-19 facemasks for ironmaking application The global pandemic response to COVID-19 has led to the generation of huge volumes of unrecyclable plastic waste from single use disposable face coverings. Rotary hearth furnaces can be used to recover Zn and Fe from non-recyclable steelmaking by-product dusts, and waste plastic material such as facemasks could be utilized as a supplementary reductant for the rotary hearth furnace (RHF), but their fibrous form makes milling and processing to appropriate sizing for RHF application extremely challenging. A scalable method of grinding facemasks to powder by melting and mixing with Welsh coal dust reported herein provides a solution to both environmental challenges. The melt-blended PPE/coal dust shows a dramatically improved CO 2 gasification reactivity (E a = 133-159 kJmol −1 ) when compared to the untreated coal (E a = 183-246 kJmol −1 ), because of improved pore development in the coal during the pyrolysis stage of heating and the catalytic activity of the CaO based ash present in the facemask plastic. The results are promising for the application of waste facemasks in recycling steelmaking by-product dusts in rotary hearth furnaces and may also be suitable for direct injection to the blast furnace subject to further study. The use of plastic waste in ironmaking and steelmaking to reduce demand for extraction of fossil fuel reductants like coal is a promising area of study and application; however, to-date the two key areas where plastics have seen application at commercial scale for steelmaking have been the inclusion of a small component of waste plastic in the coke oven process [bib_ref] CO 2 reduction potentials by utilizing waste plastics in steel works, Sekine [/bib_ref] and injection at the tuyere of the BF [bib_ref] Conversion of injected waste plastics in blast furnace, Babich [/bib_ref]. The rotary hearth furnace (RHF) is an emerging technology for the separation of volatile metals (Zn, Pb etc.) from steelmaking by-product dusts that are too high in volatile metals for direct recycling into the ironmaking process [bib_ref] Development prospect of rotary hearth furnace process in China, Zhang [/bib_ref] [bib_ref] Features of FASTMET process, Tsutsumi [/bib_ref]. Self-reducing agglomerates prepared from ferrous by-product dusts and a carbon source (coal, coke breeze, blast furnace dust) are charged into a rotating turntable furnace and heated to 1200-1300 °C for a period between 10 and 30 min. The metal oxides in the agglomerate such as Fe and Zn are reduced via carbothermic reduction to yield pellets of direct reduced iron (DRI) and a separated recyclable secondary oxide dust containing the volatile metal components [bib_ref] Pyrometallurgical removal of zinc from basic oxygen steelmaking dust-A review of best..., Stewart [/bib_ref]. Previous studies have indicated the CO 2 gasification reaction is the rate determining step in the reduction process of cold bonded self-reducing iron-carbon agglomerates [bib_ref] Rate-determining steps for reduction in magnetite-coal pellets, Coetsee [/bib_ref] although with extremely high reactivity carbon sources heat transfer effects become more significant [bib_ref] Rate of reduction of ore-carbon composites: Part II. Modeling of reduction in..., Fortini [/bib_ref]. An increased gasification reactivity of the carbon source in self-reducing agglomerates for the RHF can therefore lead to improved reduction rates and therefore improved productivity. Studies into the inclusion of waste plastics into self-reducing agglomerates as a feedstock for RHFs have yielded promising results, showing good reactivity and reduced CO 2 emission as a result of the hydrogen content of the plastic [bib_ref] Reduction mechanism of iron oxide-carbon composite with polyethylene at lower temperature, Murakami [/bib_ref]. Unfortunately, a key issue in the inclusion of waste plastic in self-reducing agglomerates is particle sizing, where fine particle sizes (~ 115 µm) are required to form a composite pellet with adequate strength for processing [bib_ref] Use of iron ore fines in cold-bonded self-reducing composite pellets, Nikai [/bib_ref]. Milling of fibrous plastic materials such as facemasks to appropriate sizing is extremely challenging due to their ductility and relatively low softening temperature. Furthermore, the low density of conventionally milled plastic waste makes storage and transport challenging. Herein we report a new scalable process in which entire plastic facemasks are blended with a Welsh ground coal injection (GCI) coal and the subsequent change in the CO 2 gasification reactivity of the material is explored. Gasification reactivity of the material was measured using thermogravimetric techniques, as this allowed for direct observation of the reaction without simultaneous metal reduction reactions occurring. If the gasification reactivity of low volatile content coal can be substantially improved by pre-treatment with waste facemask plastic, it may be applicable as a means of economically increasing the productivity of RHF plants through reduced pellet hold times. # Results Disposable, non-medical grade, facemasks (Changzhou Huangshi Packaging Printing Co. Ltd.) were coarsely chopped into squares using scissors without any prior disassembly. The manufacturer specifies the non-metallic composition of the masks is non-woven plastic (60 wt.%), meltblown plastic (30 wt.%), spandex (7 wt.%) and polyolefin resin (3 wt.%). The masks also contained a small steel wire (ca. 6 wt.% of the total mask) designed to improve the seal against the wearers nose, this wire was not removed during initial processing to better replicate an upscaled process where that would not be feasible. A photographic image of the components is shown in [fig_ref] Figure 1: Components of the disposable mask [/fig_ref] , while scanning electron microscopy (SEM) images of the layers allowed for better visualization of their fibrous nature. The dense interlocking fibres of the filtration layer can be observed in [fig_ref] Figure 1: Components of the disposable mask [/fig_ref] , compared with the comparatively less tightly woven fibres in the external (hydrophobic splash resistant) and internal (comfort) layers of the mask in [fig_ref] Figure 1: Components of the disposable mask [/fig_ref] ,d, respectively. Single use facemasks such as the ones used in this study are typically made of polypropylene fabric. The FTIR spectra of untreated facemask [fig_ref] Figure 1: Components of the disposable mask [/fig_ref] is in reasonable agreement with Aragaw's work 22 characterizing facemask plastic, with the exception that a broad absorbance that was ascribed to cellulose O-H bonding is not observed in the masks used in this study. It appears that the masks used in this work contained no paper component, which given the tremendous variation in manufacturing processes for face coverings in the COVID-19 era and lack of widespread standardization seems likely. Pulverization of the masks, even by cryogenic ball milling, was found to be totally ineffective. In a similar manner, physically mixing the cut pieces of mask with coal (or charcoal), as per the approach of Wang et al. [bib_ref] Efficient utilization of waste plastics as raw material for metallic iron and..., Wang [/bib_ref] , did not create a homogeneous mixture. Dankwah et al. [bib_ref] Recycling mixed plastics waste as reductant in ironmaking, Dankwah [/bib_ref] have previously reported that by melting it was possible to physically pelletize iron oxide with a plastic powder. This was done by pulverizing plastic after it had been separately heated to 300 °C and embrittled via quenching. This suggested that the correct physical form for the plastic could be obtained through melt processing. However, in order to obtain a granular material, suitable for producing cold bonded briquettes melt processing was performed by mixing the waste facemasks with coal fines. In the process, coarsely cut untreated facemasks [fig_ref] Figure 2: Materials at various stages of the milling process [/fig_ref] were added to GCI coal fines and charcoal [fig_ref] Figure 2: Materials at various stages of the milling process [/fig_ref] in a 20:80 weight ratio, and physically mixed [fig_ref] Figure 2: Materials at various stages of the milling process [/fig_ref]. The resulting mixture was heat treated at 250 °C in a laboratory oven (in air) for one hour and then removed. While still hot, the mixture was stirred allowing the molten plastic to wet the carbon source and form a coarse powder [fig_ref] Figure 2: Materials at various stages of the milling process [/fig_ref]. Once cooled to room temperature, the coarse powder was ball milled for 5 min at 500 rpm, whereby the powder fully passed a < 215 μm mesh, except for fragments of nose wire that remained in the mixture. These fragments may easily be recovered via screening or magnetic separation or can be left in place as an Fe source in a BF. The texture of the GCI coal/facemask sample after melt processing is that of a granular powder, which makes it suitable to be stored and transported, in particular via a conveyor belt, without creating an airborne powder hazard. TGA-DTG shows melting of the facemask occurs at 167 °C in air [fig_ref] Figure 3: TGA-DTG curve for [/fig_ref] without significant decomposition. The lack of decomposition upon heating to 250 °C is confirmed by the FT-IR of the facemask after heating to this temperature [fig_ref] Figure 1: Components of the disposable mask [/fig_ref] , which shows no change. This indicates that the process of heating the facemask with GCI coal (or charcoal) simply involves the melting of the plastic with the carbon. An important observation is that the oxidative degradation mass loss of the facemask plastic begins at T > 250 °C, therefore this is the www.nature.com/scientificreports/ highest realistic temperature at which the milling procedure can be performed without substantial degradation of the mask material. The morphology of the resultant melt-blend materials was compared to that of the untreated carbon materials by SEM as shown in [fig_ref] Figure 4: SEM images of samples of carbon and carbon/facemask melt blends [/fig_ref]. The GCI coal samples [fig_ref] Figure 4: SEM images of samples of carbon and carbon/facemask melt blends [/fig_ref] shows the flaky structure typical of raw coals with limited porosity [bib_ref] Pore structure characterization of different rank coals using gas adsorption and scanning..., Nie [/bib_ref]. The charcoal sample [fig_ref] Figure 4: SEM images of samples of carbon and carbon/facemask melt blends [/fig_ref] appears significantly more porous as the material retains some of the cellular structure of the wood which it is derived from. Comparing the untreated coal and charcoal to the samples treated with facemask plastic [fig_ref] Figure 4: SEM images of samples of carbon and carbon/facemask melt blends [/fig_ref] ,d, respectively) it can be seen that the plastic readily wets and coats the coal particles while in the molten phase. Further to this, the mixing step disperses the molten polymer across the coal surface. In order to determine the suitability of the coal/facemask melt blended product as an alternative reductant, TGA analysis has been performed under suitable conditions, in this case a CO 2 atmosphere. Generally, CO 2 gasification is the rate determining step in iron reduction within self-reducing agglomerates because it is so endothermic and kinetically slow. This is because the oxygen partial pressure is very low in the pellet bed of the RHF and gasification of carbon to carbon monoxide within the pellets by CO 2 dominates over direct solid-solid reduction of metal oxides by fixed carbon. TGA shows combustion of the facemask occurs at 382 °C in air [fig_ref] Figure 3: TGA-DTG curve for [/fig_ref] ; as may be expected this decomposition temperature is increased to 447 °C under inert atmosphere [fig_ref] Figure 2: Materials at various stages of the milling process [/fig_ref]. In both cases a second smaller decomposition occurs at 700 °C leaving a residue of 3.06% (4.84% under Ar). As decomposition occurred under both oxidizing and inert conditions it is likely that this is a carbonate decomposition. Under the CO 2 atmosphere [fig_ref] Figure 3: TGA-DTG curve for [/fig_ref] the second decomposition step observed did not occur until 955 °C suggesting that a high partial pressure of CO 2 suppressed the decomposition step, supporting the observation that this is a carbonate decomposition. The ash content following oxidation at 1000 °C as determined in [fig_ref] Figure 3: TGA-DTG curve for [/fig_ref] was 2.20%, and by subtracting this value from the 4.84% mass remaining after heating under inert conditions to 1000 °C suggests that 2.64% of the total mass of the facemasks is available as fixed carbon (the solid, combustible, carbonaceous residue that remains after heating coal or organic material to high temperatures, after the pyrolysis of all volatile matter has taken place). The results suggest very little fixed carbon remains after the pyrolysis process for facemask plastic under inert conditions, in good agreement with pyrolysis tests performed by Jung et al. on the polypropylene www.nature.com/scientificreports/ layer of N95 masks [bib_ref] Valorization of disposable COVID-19 mask through the thermo-chemical process, Jung [/bib_ref]. Fixed carbon content is commonly used as a comparative metric in iron production, as it is effectively a measure of the solid carbon available for reduction of oxides at ironmaking reaction temperatures. Analysis of the ash residue left from combustion under oxidizing conditions by SEM-EDX [fig_ref] Figure 3: TGA-DTG curve for [/fig_ref] indicates it to be primarily calcium(II) oxide (CaO) and calcium carbonate (CaCO 3 ) which is consistent with the use of calcium carbonate as a filler to polypropylene and to enhance rheological properties during the manufacturing process [bib_ref] Shear yield behavior of calcium carbonate-filled polypropylene, Wang [/bib_ref] [bib_ref] Effect of filler treatments on rheological behavior of calcium carbonate and talc-filled..., Samsudin [/bib_ref]. This supports the observation that the mass loss at ~ 700 °C observed in [fig_ref] Figure 3: TGA-DTG curve for [/fig_ref] is decomposition of CaCO 3 . Calcium carbonate is commonly used within steelmaking as a supplementary flux [bib_ref] A review of binders in iron ore pelletization, Eisele [/bib_ref] , so a CaO based ash from the facemask plastic would not present a major issue for steelmaking application from the standpoint of the introduction of far more undesirable elements in ash such as K, Na and Zn. To replicate the oxidative behaviour of chars in the RHF the TGA of facemask blends was collected under an experimental program in which the initial rapid ramping period from ambient temperature to 900 °C occurs under inert atmosphere. After a stabilization period of 10 min at 900 °C, to allow for samples to equilibrate, the gas flow was switched to CO 2 to initiate the gasification reaction and the samples were ramped to 1200 °C at various rates between 1-10 °Cmin −1 . The TGA plots of GCI coal/20 wt.% facemask blend, in comparison with that of the untreated GCI coal are shown in [fig_ref] Figure 3: TGA-DTG curve for [/fig_ref]. The comparable plots for charcoal and charcoal/facemask blend are shown in [fig_ref] Figure 3: TGA-DTG curve for [/fig_ref]. The GCI coal samples shows pyrolysis of volatile organic compounds beginning at 382 °C with a mass loss of 13.8%, which is likely a complex mixture of volatile organic compounds present in the coal. However, for the facemask treated samples of GCI coal a far greater pyrolysis mass loss can be observed (29.1%) which is related to the simultaneous pyrolysis of the facemask plastic alongside the coal volatiles. Similarly, the facemask treated charcoal shows a greater mass loss compared to the untreated material at pyrolytic temperatures, but the difference is less stark (29.6% versus 20.6%). This observation suggests that the plastic has volatilized, and the products have departed from the reaction zone of the analyser long before CO 2 oxidation is initiated. www.nature.com/scientificreports/ It can be seen from [fig_ref] Figure 3: TGA-DTG curve for [/fig_ref] that the rate of CO 2 oxidation of GCI coal was substantially increased by meltblending with the facemask, with the untreated coal fully converting at 1155 °C compared to 993 °C for the facemask treated material. Interestingly, the treated charcoal performance is essentially unchanged compared to the untreated material, except for additional ash content. The samples completely reacted at 936 °C (untreated charcoal) and 917 °C (charcoal/facemask blend), respectively. SEM images of the materials following pyrolysis at 500 °C are shown in [fig_ref] Figure 4: SEM images of samples of carbon and carbon/facemask melt blends [/fig_ref]. Little difference in the structure is observed between the charcoal [fig_ref] Figure 4: SEM images of samples of carbon and carbon/facemask melt blends [/fig_ref] and facemask treated charcoal [fig_ref] Figure 4: SEM images of samples of carbon and carbon/facemask melt blends [/fig_ref] , both materials still exhibit the cellular structure and large pores characteristic of charcoal. In contrast, the pyrolyzed GCI coal sample treated with facemask plastic [fig_ref] Figure 4: SEM images of samples of carbon and carbon/facemask melt blends [/fig_ref] appears significantly more amorphous than the GCI coal that was not treated [fig_ref] Figure 4: SEM images of samples of carbon and carbon/facemask melt blends [/fig_ref]. This change in surface morphology suggests that the action of plastic pyrolyzing from the surface of the coal during heat treatment is altering the surface structure of the coal. This textural change on the surface of the material may be responsible for the differences observed in CO 2 reactivity. Surface area analysis of the samples provides important insight into the mechanism of the increased reactivity observed in the GCI coal. Firstly, it is seen that pre-pyrolysis the samples mixed with facemasks have a dramatically increased surface area. We propose this is due to the adhesion of polymer particles to either the coal or charcoal particles creating a rough surface on the carbon particles. Following pyrolysis, however, the increased porosity observed in the pre-pyrolysis charcoal samples is eliminated and the charcoal and facemask treated charcoal exhibit essentially identical surface areas . For the GCI coal however, it appears the mechanical and/or chemical action of the plastic pyrolysis has had a roughening effect on the relatively non-porous surface of the material leaving the post pyrolysis facemask treated sample significantly more porous than the sample without facemask plastic addition, in agreement with observations made using SEM in [fig_ref] Figure 4: SEM images of samples of carbon and carbon/facemask melt blends [/fig_ref]. A similar trend is seen in pore volume analysis performed using DFT, as the pyrolyzed GCI coal sample with 20 wt.% added facemask shows substantially higher pore volume than the untreated sample. Pore development of coals as volatile compounds are lost from the material is a critical factor in gasification reactivity [bib_ref] Investigating char agglomeration in blast furnace coal injection, Sexton [/bib_ref] , and thus surface area measurements can be useful for estimating reactivity. It is important to note, from the work of Sexton et al., that this relationship between surface area and CO 2 activity was not found to be absolute and that the intrinsic chemical structure of the char material plays a substantial role in CO 2 gasification reactivity. The chemical composition of the ash in the material also plays a key role: alkali and alkaline earth metals such as Na, K and Ca are extremely catalytically significant in CO 2 gasification of coals [bib_ref] Pore structure characterization of different rank coals using gas adsorption and scanning..., Nie [/bib_ref] [bib_ref] Volatilisation of alkali and alkaline earth metallic species during the gasification of..., Quyn [/bib_ref]. In order to gain further insight into the behaviour of the coal/facemask blend with regard to its potential application in RHF, the kinetics of the CO 2 gasification step were investigated in detail. In this regard, CO 2 gasification is an extremely complicated chemical process, many different chemical compounds in the complex starting material are reacting simultaneously and the mechanism of reaction likely changes with the degree of conversion; however, model-free kinetics allow for estimation of Arrhenius parameters without the assumption of a specific reaction mechanism and has previously been utilized to characterize the kinetics of pyrolysis of waste facemasks [bib_ref] Pyrolysis kinetic behaviour and TG-FTIR-GC-MS analysis of coronavirus face masks, Yousef [/bib_ref]. Methods such as the Friedman method 32 , Flynn-Wall-Ozawa method (FWO) [bib_ref] General treatment of the thermogravimetry of polymers, Flynn [/bib_ref] [bib_ref] A new method of analyzing thermogravimetric data, Ozawa [/bib_ref] , and the Kissinger Akahira Sunose method (KAS) [bib_ref] Reaction kinetics in differential thermal analysis, Kissinger [/bib_ref] are all commonly used iso-conversional methods of determining kinetic factors, primarily activation energy (E a ). Boudouard reaction energy for CO 2 gasification of carbon materials has been shown to vary substantially with conversion degree [bib_ref] Non-isothermal thermogravimetric kinetic analysis of the thermochemical conversion of human faeces, Fidalgo [/bib_ref]. For this reason, the Friedman method was selected as the FWO and KAS methods are derived from systems where activation energy does not vary with reaction progress. As changes in mechanism can be factored into the model free method used, CO 2 is presumed to be in large excess, and the reverse Boudouard reaction is assumed to be negligible due to the high purge flow rate through the thermogravimetric analyser. As such the CO 2 gasification of biomass/char/coal can be simplified to Eq. (1). The general kinetic expression for decomposition gas-solid reactions is described in Eq. (2) [bib_ref] Reaction kinetics in differential thermal analysis, Kissinger [/bib_ref] , where t is time, k(T) is the temperature dependent kinetic constant, f(α) is the expression of the reaction model, dα/dt is the rate of conversion with respect to time and α is the conversion degree as described in Eq. (3), where m o is the samples initial mass before reaction initiation, m t is the mass of the sample at time t and m f is the final mass of the sample after the reaction has abated. (1) C fix (s) + CO 2 g → 2CO g (2) dα/dt = k(T)f(α) . Surface area (BET) and pore volume (DFT) analysis for pre-and post-pyrolysis of samples carbon and carbon/facemask samples. a 20 wt.% facemask added to the carbon source. b 500 °C in argon. ## Sample ## Surface area (m 2 g −1 ) Pore volume (cm 3 g −1 ) Pre-pyrolysis (2), a general expression to calculate kinetic factors E a and A is given in Eq. (4), where A is the preexponential factor (min −1 ), E a is the activation energy of the conversion (kJmol −1 ) and T is absolute temperature (K). The expression as proposed by Friedman involves taking the natural logarithm of Eq. (4) to yield Eq. (5) [bib_ref] Kinetics of thermal degradation of char-forming plastics from thermogravimetry, Friedman [/bib_ref] [bib_ref] Carbon gasification in self-reducing mixtures, Bagatini [/bib_ref] [bib_ref] Kinetic study and thermal decomposition behavior of lignite coal, Heydari [/bib_ref]. The values of E a with respect to conversion degree α can be determined the plot of ln(dα/dt) against 1000/T at a constant conversion value for several experiments with varying heating rate β (°Cmin −1 ), where the slope of each iso-conversional line corresponds to − E a /R. A typical plot of α versus T for GCI coal and the GCI coal/ facemask melt blend is shown in ,b, respectively, while the analogous plots for the charcoal samples are given in Charcoal is seen to be significantly more reactive to CO 2 than GCI coal, this is likely related to the material's intrinsic high degree of porosity as observed in BET and SEM analyses [fig_ref] Figure 1: Components of the disposable mask [/fig_ref]. By plotting dα/dt against T for the samples, the difference in maximum rates between the samples is clearer and is shown in -h. The gasification of charcoal appears to take place in a single stage whereas the twin peaks observed in the GCI coal and GCI coal with 20 wt% added facemask suggests a more complicated multi-step mechanism of CO 2 gasification. Charcoal and facemask treated charcoal show extreme similarity in terms of CO 2 reactivity, with the DTG signals appearing nearly identical. This is in stark contrast to the GCI coal samples where maximal rate and reaction completion temperature are significantly improved (lowered) by addition of facemask plastic. Friedman plots of ln(dα/dt) against 1000/T at iso-conversions of α = 0.1-0.9 for each sample are shown in and calculated activation energies for each of the samples based on the iso-conversional line slopes in Activation energies for CO 2 gasification of charcoal ranged from 133 to 147 kJmol −1 compared with 136 to 142 kJmol −1 for facemask treated charcoal in reasonable agreement with Dai et al. 's calculated CO 2 gasification activation energies for various biomass chars [bib_ref] Kinetic analysis of CO 2 gasification of biochar and anthracite based on..., Dai [/bib_ref]. The tight range of activation energies at different stages of conversion suggests a consistent gasification mechanism for both treated and untreated charcoal. The results imply that the treatment with facemask plastic for charcoal has little to no effect on CO 2 gasification reactivity. Activation energies for GCI coal ranged from 184 to 246 kJmol −1 , with an apparent shift in the reaction mechanism at higher levels of conversion, where a sharp increase in E a is observed for α = 0.9. These results are broadly similar to calculated activated energies for Indian coals [bib_ref] Thermogravimetric and model-free kinetic studies on CO 2 gasification of low-quality, high-sulphur..., Das [/bib_ref] and coal chars 37 but the range of reported values for the Boudouard reaction is extremely wide in literature, due to its dependence on structural factors in the material. GCI coal treated with facemask plastic showed a marked decrease in activation energy across the entire conversion range when compared with the untreated material, with values spanning 134-160 kJmol −1 . That the addition of facemask plastic to coal can increase its reactivity so decidedly, even though the plastic has long since decomposed and volatilized before the reaction gas was introduced to the thermogravimetric analyser implies that either, the ash residue of the facemasks is acting in a catalytic manner to accelerate CO 2 gasification or that the action of the plastic in the material vaporizing and diffusing away from the solid is introducing some form of chemical or structural change in the material that is leading to increased CO 2 gasification reactivity . # Discussion The observation that charcoal is not seen to increase in reactivity with the addition of facemask plastic, but the GCI coal sample does, has two potential explanations. The first is that the far higher initial porosity of the facemask treated charcoal sample led to the facemask treatment doing little to increase porosity further, and therefore have no tangible effect on CO 2 gasification reactivity. Secondly, if the deposition of catalytic CaO ash following pyrolysis of the facemask plastic is the dominant mechanism of increased CO 2 reactivity, this may be masked due to the high ash content of the charcoal (18.7%) having such a significant catalytic effect on CO 2 reactivity that the further small addition of facemask ash has no tangible effect. It is likely there is a combination of catalytic and pore development factors responsible for the observed increase in reactivity of facemask treated GCI coal. Nevertheless, the CO 2 gasification reactivity of a sample of a low volatile content coal was dramatically improved by the addition of waste facemask plastic, as determined through non-isothermal thermogravimetric measurements. This was not observed in a substantially more porous, higher volatile content and higher ash content charcoal material. The results suggest a potentially synergistic mechanism where the deposition of CaO based ash following the pyrolysis of the facemask plastic, as well as a physical change in the structure of the coal causing increased porosity both play a role in increasing the reactivity of the material in the CO 2 gasification reaction [bib_ref] Synergistic effects of CaO and MgO on ash fusion characteristics in entrained-flow..., Zhang [/bib_ref]. Further study is required to fully parse these two competing mechanistic explanations. This increase in CO 2 gasification reactivity in a comparatively low volatile content coal is significant for the purposes of iron reduction in a RHF, where the gasification reactivity of the carbon is seen to dominate Fe reduction kinetics during much of the reduction reaction [bib_ref] Rate-determining steps for reduction in magnetite-coal pellets, Coetsee [/bib_ref]. This is especially significant because the carbon and hydrogen content of the facemask plastic is also available for reduction of iron oxides, meaning not only www.nature.com/scientificreports/ to emulate the conditions of the raceway of a BF and attributed this increase to increasing particle surface area and additional heat supplied by combustion of coal volatiles [bib_ref] Synergistic effects of CaO and MgO on ash fusion characteristics in entrained-flow..., Zhang [/bib_ref]. We propose that the action of dispersing molten facemask plastic across the surface of low volatile content coal is acting in a manner similar to intrinsic volatile matter within the coals, opening pores in the material and increasing the reactive surface area of the particles. An overall schematic for the process is shown in [fig_ref] Figure 7: Proposed process schematic showing the effect of the polymer fabric of facemasks... [/fig_ref]. [formula] (3) α = m 0 − m t m 0 − m f (4) dα dt = f (α)Aexp − E a RT (5) ln dα dt = ln[Af (α)] − E [/formula] The results are favourable for the purposes of increasing CO 2 reactivity of low-quality carbonaceous materials for RHF processing of steelmaking by-product sludges. Increased char reactivity may lead to improved iron and zinc reduction rates in self-reducing agglomerates and the addition reducing gases provided by the facemask plastic would also contribute additional reductant; however, through exploring the properties of the facemask treated GCI coal for RHF application it became apparent that the material may be a good candidate for direct injection at the BF tuyere as well. The injection of coal through the tuyeres into the blast furnace has become extremely commonplace in the global steel industry as a means of reducing demand for expensive and environmentally costly coke in iron production [bib_ref] Opportunities to improve the utilisation of granulated coals for blast furnace injection, Steer [/bib_ref] [bib_ref] High rate coal injection of 218kg/t at Fukuyama No. 4 blast furnace, Maki [/bib_ref]. BF coal injection rates have been steadily increasing for decades, from 60-80 kgtHM −1 in the 1980s to over 200 kgtHM −1 in some plants [bib_ref] Maximum rates of pulverized coal injection in ironmaking blast furnaces, Nomura [/bib_ref]. An extremely important consideration when selecting candidate coals for injection to the BF is the reactivity of the char produced following injection. Coal that enters the furnace via a tuyere experiences oxidizing conditions and an extremely high heating rate as the material enters the raceway. Any coal particles that are not completely reacted at this stage leave the raceway region as a partially burnt-out char where gasification under the prevailing gas condition of the furnace becomes dominant. In order to achieve the maximum possible injection rate, it is critical that the injected material reacts and gasifies in as short a time period as possible [bib_ref] Optimization strategies for pulverized coal injection into the blast furnace, Schott [/bib_ref] , de Lourdes Ilha Gomez et al. stipulated that "The more reactive a char is, the better is the effect in the replacement of coke" with respect to pulverised coal injection and the BF [bib_ref] Thermal analysis evaluation of the reactivity of coal mixtures for injection in..., De Lourdes Ilha [/bib_ref]. Char that is not completely consumed during PCI is obviously economically unfavourable as it is underutilized for the purposes of iron reduction, but much more significantly, unburnt material can accumulate within the furnace burden. This can have substantial negative effects on burden permeability, furnace stability, temperature distribution, coke erosion and char carryover [bib_ref] Evaluation of petrology and reactivity of coal blends for use in pulverized..., Osório [/bib_ref]. Injection rates to a BF can be limited by a number of other factors such as S and ash input but one of the key reactivity criteria that has an established effect on burden permeability is the insufficient gasification of chars formed after injection to the furnace [bib_ref] Investigating char agglomeration in blast furnace coal injection, Sexton [/bib_ref]. Therefore, the improved gasification reactivity exhibited by the facemask treated GCI coal may have potential application in improving coal reactivity during BF injection and therefore present a method of making cheaper, less reactive coals more attractive for injection or potentially increasing the maximum injection rate without negative effects on furnace operation. Globally, plastic injection systems are also emerging capable of injecting plastic to the furnace, reducing fossil fuel requirements and sparing plastic waste from landfill or entering the ocean [bib_ref] Recycling of waste plastic packaging in a blast furnace system. NKK, Yoji [/bib_ref]. These modified plants are capital intensive, however, requiring specialist support plant and tuyere modifications to be capable of handling the extreme heterogeneity of a recycled plastic feed 48 , subject to further trials the method of melt-blending described herein may present a way of introducing plastics to a BF with minimal plant modification. As waste facemask plastic is unfortunately, readily available in the wake of the pandemic of SARS-CoV-2 and its variants, its application in the iron and steelmaking industry is a positive avenue of disposal to prevent material being introduced into rivers and oceans while displacing environmentally damaging fossil fuels from the steelmaking process. Further study into applications of the material is required to assess whether the improved CO 2 gasification reactivity observed in this thermogravimetric study would indeed translate into improved reduction performance in an RHF. Drop tube furnace trials designed to replicate the unique chemical conditions of the BF tuyere will elucidate whether the improved char reactivity exhibited in this study may translate into the BF environment and make the material a suitable auxiliary BF injectant. # Methods Disposable, non-medical grade face coverings were used for the purpose of this study. Masks produced by the Changzhou Huangshi Packaging Printing Co. Ltd. were coarsely chopped into squares using scissors without any prior disassembly. Charcoal was purchased from Fisher Scientific, and the coal used was a typical BF injection coal. Both materials were dried at 90 °C for 12 h and pulverized and sieved to < 215 μm. Proximate analysis Thermogravimetric experiments were carried out using a TA Instruments SDT Q600 analyser. Experiments were performed on 20 mg ± 0.5 mg of sample, using alumina sample crucibles. For non-isothermal tests, samples were ramped to reaction temperature (900 °C) under protective cover of argon (100 cm 3 min −1 ) at a rate of 20 °C min −1 and held isothermally for a period of 10 min at reaction temperature. The purge gas was then changed to CO 2 at a rate of 120 cm 3 min −1 to initiate the reaction before ramping the analysis temperature from 900 °C to 1200 °C at a range of heating rates (1, 5, 8 and 10 °C min −1 ) [bib_ref] Carbon dioxide absorption by polyethylenimine-functionalized nanocarbons: A kinetic study, Andreoli [/bib_ref]. Surface area analysis was performed using a Quantachrome Nova 2000e analyzer using N 2 as an adsorbate gas. Around 130 mg of sample was degassed under vacuum at 130 °C overnight. BET surface area was calculated in the 0.09237-0.24399 P/P 0 range for each of the samples. FTIR measurements were taken using a Thermofisher Scientific Nicolet iS10. Scanning electron microscopy was performed on a Hitachi TM3030. Carbon and sulphur were determined via combustion analysis using an ELTRA CS500 C/S Analyzer. Ball milling of samples was performed in a Fritsch Pulverisite 6 ball mill at 500 rpm. Bulk pyrolysis experiments were performed using a carbolite MTF 12/38/400 controlled atmosphere tube furnace heated to 500 °C at a rate of 5 °C min −1 under the protection of 1 dm 3 min −1 pureshield argon to heat a pellet of material (16 mm diameter, 2.5 g, 975 bar, 60 s residence time) produced using a Retsch PP 25 Pellet Press. Samples were cooled in the furnace to ambient temperature for pulverizing and further analysis. [fig] Figure 1: Components of the disposable mask. (a) A photograph of a fully disassembled mask with the component pieces labelled. (b) SEM image of the untreated mask outer layer. (c) SEM image of the untreated mask filtration layer. (d) SEM image of the untreated mask inner layer. [/fig] [fig] Figure 2: Materials at various stages of the milling process. (a) Coarsely cut untreated facemasks, (b) GCI coal, (c) cut facemasks and GCI coal before heat treatment and (d) facemasks and GCI coal after 1 h at 250 °C and before milling showing the morphology. [/fig] [fig] Figure 3: TGA-DTG curve for (a) untreated facemasks under a 100 cm 3 min −1 flow air, (b) untreated facemasks under a 100 cm 3 min −1 flow carbon dioxide. DTG signals are denoted as dashed lines. TGA measurements to 900 °C under Ar and to 1200 °C under CO 2 atmospheres for (c) GCI coal (solid line) compared to GCI coal/20 wt.% facemask blend (dashed line) and (d) Charcoal (solid line) compared to charcoal/20 wt.% facemask blend (dashed line). Scientific Reports | (2022) 12:2693 | https://doi.org/10.1038/s41598-022-06691-w [/fig] [fig] Figure 4: SEM images of samples of carbon and carbon/facemask melt blends. (a) GCI coal, (b) GCI coal with 20 wt.% facemask plastic, (c) charcoal, and (d) charcoal with 20 wt% added facemask plastic. (e) GCI coal after pyrolysis to 500 °C, (f) GCI coal with 20 wt.% facemask plastic after pyrolysis to 500 °C, (g) charcoal after pyrolysis to 500 °C, and (h) charcoal with 20 wt% added facemask plastic after pyrolysis to 500 °C. Scientific Reports | (2022) 12:2693 | https://doi.org/10.1038/s41598-022-06691-w [/fig] [fig] Figure 5, Figure 6: Plots of α and dα/dt against T (°C) for different heating rates. (a) GCI coal, α against T. (b) GCI coal melt-blended with 20 wt.% facemask material, α against T. (c) Charcoal, α against T. (d) Charcoal melt-blended with 20 wt.% facemask material, α against T. (e) GCI coal, dα/dt against T. (f) GCI coal melt-blended with 20 wt.% facemask material, dα/dt against T. (g) Charcoal, dα/dt against T. (h) Charcoal melt-blended with 20 wt.% facemask material, dα/dt against T. Scientific Reports | (2022) 12:2693 | https://doi.org/10.1038/s41598-022-06691-w www.nature.com/scientificreports/ that a small amount of required coal is displaced, but also that the utilization of the coal still required is more efficient. Steer et al. described a close relationship between increasing volatile matter content of injection coals and reduced burnout time (and therefore improved combustion reactivity) using a drop tube furnace (DTF) Friedman plots and calculated activation energy (E a ) against α for the combustion of carbon and carbon/facemask melt blends in CO 2 atmosphere. (a) Friedman plot for GCI coal. (b) Friedman plot for GCI coal/facemask melt blend. (c) Activation energy against α for GCI coal and GCI coal/facemask melt blend. (d) Friedman plot for charcoal. (e) Friedman plot for charcoal/facemask melt blend. (f) Activation energy against α for charcoal and charcoal/facemask melt blend. Scientific Reports | (2022) 12:2693 | https://doi.org/10.1038/s41598-022-06691-w [/fig] [fig] Figure 7: Proposed process schematic showing the effect of the polymer fabric of facemasks on a coal particle as it undergoes heat treatment. [/fig]

EPISPOT: An epigenome-driven approach for detecting and interpreting hotspots in molecular QTL studies We present EPISPOT, a fully joint framework which exploits large panels of epigenetic annotations as variant-level information to enhance molecular quantitative trait locus (QTL) mapping. Thanks to a purpose-built Bayesian inferential algorithm, EPISPOT accommodates functional information for both cis and trans actions, including QTL hotspot effects. It effectively couples simultaneous QTL analysis of thousands of genetic variants and molecular traits with hypothesis-free selection of biologically interpretable annotations which directly contribute to the QTL effects. This unified, epigenome-aided learning boosts statistical power and sheds light on the regulatory basis of the uncovered hits; EPISPOT therefore marks an essential step toward improving the challenging detection and functional interpretation of trans-acting genetic variants and hotspots. We illustrate the advantages of EPISPOT in simulations emulating real-data conditions and in a monocyte expression QTL study, which confirms known hotspots and finds other signals, as well as plausible mechanisms of action. In particular, by highlighting the role of monocyte DNase-I sensitivity sites from >150 epigenetic annotations, we clarify the mediation effects and cell-type specificity of major hotspots close to the lysozyme gene. Our approach forgoes the daunting and underpowered task of one-annotation-at-a-time enrichment analyses for prioritizing cis and trans QTL hits and is tailored to any transcriptomic, proteomic, or metabolomic QTL problem. By enabling principled epigenome-driven QTL mapping transcriptome-wide, EPISPOT helps progress toward a better functional understanding of genetic regulation. # Introduction Molecular datasets and annotation databases are growing in size and in diversity. In particular, genetic data are now routinely collected along with gene, protein, or metabolite level measurements and analyzed in molecular quantitative trait locus (QTL) studies, with the aim of unravelling the regulatory mechanisms underlying common diseases. However, these studies present additional complexities compared to classical genome-wide association studies (GWASs). First, they entail a very different statistical paradigm: while GWASs consider a single or a few related clinical traits, molecular QTL studies typically involve hundreds or thousands of molecular traits, regressed on hundreds of thousands of genetic variants. Second, they need to accommodate two types of genetic control: a variant may affect molecular products of genes in its vicinity (cis action) or products of remote genes (trans action), where the latter mode of control is typically much weaker and, hence, harder to uncover than the former. In particular, pleiotropic or hotspot genetic variants may exert weak trans effects on many molecular traits. The current mapping practice only partially embraces the features of QTL studies. Indeed, widely used marginal screening approaches 1,2 suffer from a large multiplicity burden and tend to lack statistical power as they do not exploit the regulation patterns shared by the molecular entities, whereas joint modeling approachesare often limited by the computational burden implied by the exploration of high-dimensional spaces of candidate variants and traits. To manage this tension between scalable inference and comprehensive joint modeling, we recently proposed a variational inference approach, called ATLASQTL,which explicitly borrows information across thousands of molecular traits controlled by shared pathways and offers a robust fully Bayesian parametrization of hotspots; its increased sensitivity and that of earlier related models have been demonstrated in different molecular QTL studies.In complement to the actual mapping task, biologists increasingly try to capitalize on the wealth of available epigenetic annotation sources to infer the functional potential of genetic variants. The standard strategy uses epigenetic marks mostly for prioritization of hits derived from marginal screening: it consists in looping through all the loci with statistically significant associations and, for each locus, inspecting marks to decide on ''a most promising'' functional candidate genetic variant among all those in linkage disequilibrium (LD). This approach has the following disadvantages. First, publicly available databases nowadays contain several hundred epigenetic annotations. Preselecting just a few may involve omitting others that are relevant, which may bias the conclusions. Second, even if a comprehensive inspection were feasible, the degrees of relevance of the annotations may be very uneven and may depend on the conditions, cell types, tissues, and even genomic regions considered, so it is unclear how to weight each contribution. In response to this, a number of model-based approaches leveraging epigenetic annotations have been proposed over the past decade, whether for genome-wide association studies (e.g., iBMU,bfGWAS, or fine mapping (e.g., PAIN-TOR, . Despite this extensive development, no existing method provides a solution to our problem, namely, modeling the functional enrichment of trans-QTLs and hotspots, a task which is substantially more complex and elusive than for the functional enrichment of cis-QTLs or GWA signals for a series of related phenotypes. All available modeling tools are designed for genetic mapping with oneor a few 12 traits at a time, while trans-QTL and hotspot mapping requires considering thousands of traits simultaneously. It is also worth noting that many approaches accommodate only small numbers of candidate annotations by computational or statistical stability constraints,or take as input GWA summary statistics rather than individual-level data, thereby not benefiting from the added statistical power obtained from jointly modeling the latter, along with the functional information.Our work enables large-scale inference for cis-and trans-QTL regulation using whole panels of external epigenetic annotations and argues that the epigenome can serve both to increase statistical power for QTL mapping and to shed light on the biology underlying the uncovered genetic map in a systematic manner. Specifically, it couples a fully Bayesian QTL mapping strategy, in which all loci and molecular traits are analyzed jointly, with a principled leveraging of epigenetic information by treating this information as complementary predictor-level data that may inform the probability of genetic variants to be involved in QTL associations. As successfully demonstrated in the context of genetic mapping with clinical traits, suitable use of epigenetic information can boost the detection of weak associations and help in discriminating genuine signals from spurious ones caused by LD or other confounding factors.Our modeling framework, called EPISPOT, directly infers the role of sparse sets of annotations-from hundreds of candidate functional annotations-in the activation of both cis and trans mechanisms affecting hundreds to thousands of molecular traits. Importantly, it combines this epigenome-driven feature with a flexible hotspot modeling feature inspired from our previous work,thereby offering a unified toolkit to refine the detection of hotspots, aided by the epigenetic information at hand. The base version of EPISPOT assesses the action of the annotations uniformly for the full set of analyzed transcripts. However, for cases where a sensible partition into subsets of co-expressed molecular traits (modulesis available, we also develop a module version of EPISPOT, which accommodates module-specific epigenetic action by estimating the contribution of the epigenetic marks to the QTL associations in each module. Our take is that fully joint modeling is paramount to borrow information across loci, epigenetic marks, and molecular traits with complex dependences, but this requires careful algorithmic considerations to ensure scalable inference while retaining accuracy. EPISPOT implements an adaptive and parallel variational expectation-maximization (VBEM) algorithm, augmented with a simulated annealing scheme which effectively explores the multimodal parameter spaces induced by highly structured data. This optimization routine is purposely tailored to the analysis of genetic data with strong LD blocks, for which the inclusion of the epigenetic data has the greatest impact. Our framework also constitutes an effective tool for interpreting (1) the detected trans-acting and hotspot variants based on their overlap with the selected epigenetic marks and (2) the molecular traits under genetic control in light of these marks. This additional purpose of EPISPOT is key given that elucidating the mechanisms of action of hotspots is often as challenging as mapping them in the first place. Indeed, there is accumulating evidence that most genetic variants acting in trans lie in intergenic regions,where functional roles are difficult to decipher. Moreover, the massive trans-gene networks under genetic control are thought to be subject to subtle interplays, and researchers are often left with a variety of possible strategies to try to understand the interacting pathways between the genotype and underlying disease endpoints.These strategies range from hypothesis-driven bottom-up approaches that start from isolated mechanisms and try to generalize them (e.g., based on cis-mediation hypotheses) to agnostic top-down approaches that directly model the whole system in view of teasing apart its fundamental components (e.g., based on graphical modeling approaches).Our approach provides an alternative anchor toward decoding the complex networks controlled by hotspots, namely via the epigenetic marks found to be informative for the genetic mapping. EPISPOT is not targeted at genome-wide discovery but at effecting refined QTL mapping and hotspot prioritization, based on genomic regions-hereafter called candidate loci-harboring SNPs thought to be involved in QTL regulation. A crucial distinction with the existing enrichment approaches is that the candidate loci do not correspond to a previously determined list of QTL hits but are whole genomic regions, which can involve hundreds of genetic variants (most of them with no QTL activity). EPISPOT exploits shared epigenetic signals across these regions to then select QTL hits with an increased statistical power. Importantly, fruitful applications of EPISPOT, which can successfully decipher part of the molecular regulation machinery, require problems where the signal-to-noise and density of epigenetic/QTL signals are sufficient. In this work, we will describe extensive simulation experiments to highlight the benefits of using epigenetic information when available for a panel of regulation scenarios, and we will question the conditions under which inference is adequately powered to leverage this information. We will therefore formulate guidelines for practical use and provide a software implementation of EPISPOT along with documented code for the data-generation procedure used in the simulation experiments. Another key component of the present paper concerns illustrating and exploiting the advantages of EPISPOT in real molecular QTL conditions. We will conduct and discuss the findings of a thorough monocyte expression QTL (eQTL) study leveraging a panel of annotations, including DNase-I sensitivity sites identified in different tissues and cell types, Ensembl gene annotations, and chromatin state data from ENCODE. In particular, by pinpointing contextrelevant marks in a hypothesis-free manner, EPISPOT will allow us to disentangle key mechanisms pertaining to the lysozyme pleiotropic activity of chromosome 12-an activity which, although reported in several studies, is so far left unexplained in terms of its functional and mediation processes. Obtaining such evidence without EPISPOT would involve the daunting task of evaluating the enrichment of candidate eQTL hits in each individual epigenetic mark; this would also have no guarantee of success since one-ata-time inspection strategies are deprived of the enhanced statistical power obtained with a unified joint epigenome/ QTL mapping strategy. # Material and methods ## Two-level hierarchical regression model We consider a Bayesian model linking three data sourceswith two levels of hierarchy. The bottom level parametrizes the QTL effects and the top level parametrizes the epigenetic modulations of the primary QTL effects. Specifically, the bottom level hierarchy uses a series of conditionally independent spike-and-slab regressions to model the regulation of q molecular traits by p candidate genetic variants or single-nucleotide polymorphisms (SNPs) for n samples: [formula] y t b t ; t t $ N n Xb t ; t À1 t I n À Á ; t ¼ 1; .; q; b st jg st ; s 2 ; t t $ g st N 0; s 2 t À1 t À Á þ 1 À g st ð Þ d 0 ; s ¼ 1; .; p; (Equation 1) [/formula] where y ¼ ðy 1 ; .; y q Þ is an n3q matrix of centered responses (molecular traits) and X ¼ ðX 1 ; .; X p Þ is an n3p matrix of centered candidate predictors for them (SNPs). Here, d 0 is the Dirac distribution and to each regression parameter b st corresponds a binary latent parameter g st taking value 1 if and only if SNP s is associated with trait t. Taking the posterior means of the latent parameters g st then yields marginal posterior probabilities of inclusion (qtl-PPIs,, prðg st ¼ 1jyÞ, from which Bayesian false discovery rate (FDR) estimates can be obtained. Moreover, the precision parameters t t and s À2 are assigned diffused Gamma priors. The top-level hierarchy parametrizes the effects of the epigenetic marks on the QTL probability of association via a secondstage probit regression on the probability of effects: [formula] g st jq s ; z t ; x $ Bernoulli F z t þ q s þ V T s x À Á È É ; q s $ N 0; s 2 0s À Á ; z t $ N n 0 ; t 2 0 À Á ; x l jr l $ r l N 0; s 2 À Á þ 1 À r l ð Þ d 0 ; [/formula] r l $ Bernoulli u l ð Þ; l ¼ 1; .; r; [formula] (Equation 2) [/formula] where Fð $Þ is the standard normal cumulative distribution function and V ¼ ðV 1 ; .; V r Þ is a p3r matrix of (centered) predictorlevel covariates (epigenetic marks). The epigenetic marks therefore represent external annotations that directly annotate the SNPs, rather than sample-specific annotations. Although prior information on the relevance of the marks for the QTL control can be accommodated if desirable, this is not required, as the use of a sparse prior on the mark effects x allows incorporating a large number of marks even though only a fraction may be responsible for genetic activity. In particular, if none of the marks are relevant, the QTL mapping will not suffer any bias from modeling the candidate marks (see simulation studies hereafter). Moreover, similarly as for the QTL effects, mark selection is easily achieved using posterior probabilities of inclusion, prðr l ¼ 1jyÞ, corresponding to the posterior means of the binary latent inclusion indicators r l (epi-PPIs,. This typically yields a sparse subset of marks, whose biological interpretation may help in understanding the mechanisms of action of the SNPs involved in the QTL associations. ## A parametrization tailored to the detection of hotspots In addition to embedding the predictor-level regression for the epigenetic effects, the top-level probit model in Equation 2 also decouples the contributions of the predictors (SNPs) and the responses (molecular traits), namely, by involving a response-specific parameter, z t , which adapts to the sparsity level linked with each response y t and a predictor-specific parameter, q s , which encodes modulations of the probability of association according to the overall effect of each predictor X s . Parameter q s has a central role in pleiotropic molecular QTL settings as it represents the propensity of each predictor to be associated with multiple responses, i.e., its propensity to be a hotspot. Its Gaussian prior specification ensures closed-form updates, which is critical to the efficiency of the algorithm on large datasets. It also conveniently permits using a local-scale representation (via s 0s ) to prevent overshrinkage of large hotspot signals; see our previous work on the hierarchical modeling of hotspots, from which this formulation is borrowed.Here, the value of s 0s is set by empirical Bayes, and so are the epigenetic effect hyperparameters u l and s. The values of the hyperparameters n 0 and t 0 are chosen to induce sparsity, by specifying a prior expectation and a prior variance for the number of predictors associated with each response (supplemental material and methods). Hence, the EPISPOT model (Equations 1 and 2) borrows information across the three types of entities (epigenetic marks, SNPs, and molecular traits) in a unified manner, while providing interpretable posterior quantities, in particular qtl-PPIs and epi-PPIs, for the selection of each type of variable. It leverages the epigenome for two complementary purposes: (1) to enhance statistical power for QTL and hotspot mapping and (2) to shed light on the biology underlying the genetic control, via the inspection of the selected marks. ## A modification for module-specific epigenetic contributions The machinery of genetic control is complex and it is unlikely that the action of the epigenome on QTL regulation will uniformly affect the transcriptome. In particular, different groups of molecular traits may be governed by different functional mechanisms, involving different sets of epigenetic marks, to different degrees. When a partition into modules of genes (proteins or metabolites for pQTL or mQTL analyses, respectively) likely to be co-regulated is available to the analyst, it can be provided as input to the method which will then infer the annotation effects in a module-specific fashion, based on the following modification of top-level Equation 2: [formula] g st jq m;s ; z t ; x m $ Bernoulli F z t þ q m;s þ V T s x m À Á È É ; [/formula] q m;s $ N 0; s 2 0m;s ; z t $ N n 0 ; t 2 0 À Á ; x m;l r m;l $ r m;l N 0; s 2 [formula] m À Á þ 1 À r m;l À Á d 0 ; [/formula] r m;l $ Bernoulli u m;l ð Þ; l ¼ 1; .; r; [formula] (Equation 3) [/formula] where m˛M is a module of traits, with M a partition of f1; .; qg and mHt. Parameter x m then represents the epigenetic contribution of the r marks for the QTL associations involving the traits from module m. The hotspot parameter q m;s also accounts for the module structure: it represents the propensity of SNP s to be associated with few or many traits from module m. This encodes module-specific pleiotropic levels and also reflects the fact that a SNP controlling a given trait in a module is more likely to be also associated with related traits from the same module compared to traits outside the module. (B) Graphical representation for the two-level hierarchical model. The shaded nodes are observed, and the others are inferred. The toplevel regression corresponds to the top plate; the probability of association is decoupled into a trait-specific contribution, z t , a SNP-specific contribution with a ''hotspot propensity parameter'' q s and an epigenome-specific contribution, x l , where V l is the vector gathering the observations of predictor-level epigenetic covariate l for all candidate SNP predictors X s , s ¼ 1;.;p. Parameter b st models the effect between SNP X s and trait y t , and g st and r l are binary latent indicators for the QTL associations and epigenetic mark involvement, respectively. Parameter s models the typical size of QTL effects and t À1 t models the residual variability of trait y t . (C) Posterior output. Selection of epigenetic marks with a role in QTL regulation is carried out using the posterior probabilities of inclusion (epi-PPIs), prðr l ¼ 1jyÞ, l ¼ 1; .; r (bottom left) and selection of associated SNP-trait pairs (aided by the marks) is carried out using the posterior probabilities of inclusion (qtl-PPIs), prðg st ¼ 1jyÞ, s ¼ 1; .; p; t ¼ 1; .; q (bottom right). The hotspot Manhattan plot (top) reports the number of traits associated with each SNP (''hotspot size''), after using a selection threshold on the qtl-PPIs (e.g., FDR-based). (D) EPISPOT workflow. Candidate loci and molecular traits are obtained from a preliminary screening or from existing databases and supplied as input to the method along with epigenetic marks at the variants harbored by the loci. The algorithm is used with or without the module option depending on whether the traits are gathered into modules or not (M-EPISPOT in gray, resp. EPISPOT in blue). The output consists of sets of associated variants and traits, QTL hotspots, and epigenetic marks relevant to the primary QTL associations for given significance thresholds. It is then interpreted to generate mechanistic hypotheses about the functional processes underpinning the QTL associations. The corresponding version of the algorithm-implementing Equations 1 and 3-is hereafter called M-EPISPOT when an explicit distinction with the base, module-free version-implementing Equations 1 and 2-is needed. Different approaches, based on some prior state of knowledge, on specific optimization methods, or both, will typically yield complementary definitions of modules. In some instances, there will be obvious biological reasons backing up the obtained grouping; in others, no clear partitioning will emerge, in which case the analyst may choose to use the module-free version of the model. As there is no generic strategy for forming modules, it is important to understand the impact of such choices on inference. In particular, from a modeling point of view, a given module should ideally comprise co-regulated molecular traits, i.e., traits with shared genetic control, triggered by common epigenetic mechanisms. The top-level regression (Equation 3) will then represent the possible epigenetic effects underlying the functional mechanisms in the module, and module-specific epi-PPIs will be useful to select the marks involved in the regulation of each module. In particular, shared signals will be best leveraged when the molecular traits controlled by a given SNP belong to a same module. The simulation studies and the eQTL analysis will provide practical recommendations as well as analyses of sensitivity to module misspecification. ## A scalable purpose-built algorithm The hierarchical model described above couples two levels of spike-and-slab regression, which accommodate three large spaces of SNPs, molecular traits and epigenetic marks, with possibly thousands of variables each. Careful algorithmic strategies are therefore critical to ensure that inference is accurate and scalable. To meet both requirements, we implement an adaptive variational expectation-maximization (VBEM) algorithm and augment it with a simulated annealing procedure that efficiently explores the highly multimodal variable spaces formed by data with strong dependence structures. VBEM algorithms were introduced by Blei et al.in the context of Dirichlet allocation modeling. In short, they iterate between optimizing empirical Bayes estimates (in our case for the hotspot propensity and epigenetic effect hyperparameters) and running a variational algorithm for the remaining parameters, given the updated empirical Bayes estimates. We present hereafter the algorithm in its general module-based form (M-EPISPOT); omitting the index m and taking M ¼ 1 gives the base version with no module partitioning (EPISPOT). Let v ¼ ðb; t; g; s 2 ; q; z; x; rÞ denote the parameters for Equations 1 and 3, and let h ¼ ðh 1 ; .; h M Þ denote the second-stage model hyperparameters, with h m ¼ ðs 2 0m ; s 2 m ; u m Þ for module m ¼ 1; .; M. We propose estimating h via an empirical Bayes procedure, by finding where qðvÞ is the variational density for pðvjy; b hÞ for a current estimate b [formula] b h ¼ arg max h [ ðh; yÞ; (Equation 4 [/formula] h and E q ð $Þ is the expectation with respect to qðvÞ. More precisely, it initializes the parameter and hyperparameter vectors v ð0Þ and h ð0Þ , and alternates between the E-step, [formula] q ðtÞ ¼ arg max q L À q; h ðtÀ1Þ Á ; [/formula] using the variational algorithm for obtaining q ðtÞ at iteration t, and the M-step, [formula] h ðtÞ ¼ arg max h L À q ðtÞ ; h Á ; [/formula] until convergence of h ðtÞ . In our case, the updates for the M-step are obtained analytically by setting to zero the first derivative of Lðq ðtÞ ; hÞ with respect to each component of h. This only requires computing and differentiating the joint likelihood term E q logpðy; vjhÞ in Equation 5, as the entropy term ÀE q logqðvÞ is a function of h ðtÀ1Þ and is constant with respect to h. Variational inference is typically orders of magnitude faster than classical Markov chain Monte Carlo inference 5,6,22 for comparisons on GWA and molecular QTL models. Some computational cost is added for VBEM algorithms as each E-step requires running the variational algorithm until convergence. Moreover, the two regression levels of our Equations 1 and 2 or Equations 1 and 3 necessitate the exploration of a very large parameter space, which is complex and time consuming for any type of inference. We consider two strategies to overcome this burden. First, we substantially reduce the runtime of the within-EM variational runs by using an adaptive stopping criterion, namely, starting with a large tolerance and dynamically decreasing it according to the convergence state of the overall EM algorithm. The second strategy applies to the module version of our algorithm: the specification in Equation 3 suggests that its hyperparameters may be estimated reasonably well by restricting the VBEM scheme to subproblems corresponding to each module, i.e., applying Equations 1 and 2 to the subsets of responses y m separately for obtaining the corresponding empirical Bayes estimates h m , m ¼ 1;.;M. In addition to accelerating hyperparameter estimation for each module (as the model is much smaller), this has the advantage of allowing parallelization across modules. Once all module hyperparameters are estimated, they are inserted into Equations 1and 3 and variational inference is run on the entire dataset. Strong posterior multimodality can be induced by dense genotyping panels with marked LD structures, whereby the inclusion of epigenetic information is particularly beneficial to disentangle the genetic contributions. To robustly infer signals from problems with strong data dependence structures, we augment all variational schemes with a simulated annealing routine.Annealing introduces a so-called temperature parameter to index the variational distributions and control the level of separation between their modes, thereby easing the progression to the global optimum. In practice, we start with a temperature T 0 to flatten the posterior distribution and sweep most local modes away, and we then lower it at each iteration, until the original multimodal distribution, called the cold distribution, is reached. Finally, to ensure stable inference, our routine excludes redundant SNPs and marks (i.e., displaying perfect collinearity with other SNPs/marks) prior to the run. Moreover, constant marks or marks that concern less than a given proportion of SNPs (default 5%) are also discarded before the analysis as insufficiently informative. A sketch of the algorithm and the full derivation of the annealed VBEM updates are in the supplemental material and methods. The algorithm is implemented as a publicly available R package with Cþþ subroutines (see Web Resources). Both the EPISPOT and M-EPISPOT versions run within seconds to few hours depending on the numbers of loci, molecular traits, and epigenetic marks (see the runtime profiling in the supplemental material and methods). We also provide simulation studies that demonstrate the robustness of EPISPOT to different degrees of LD and the benefit of coupling VBEM inference with simulated annealing in case of strong LD (supplemental material and methods). ## Recommended use EPISPOT is a refining tool for the detection and interpretation of QTL and hotspot effects. It is meant to be used for joint analysis of preselected genomic regions (candidate loci) and transcripts believed to be under genetic control. Different approaches can be considered to obtain loci of interest. Public databases can be employed to form loci of given size around previously identified hits, provided this information is available for the condition, tissue, or cell type at hand. An alternative approach is based on a preliminary application of ATLASQTL 5 or another screening method, ideally on an independent dataset. If no independent dataset is available to the analyst, useful research hypotheses may still be obtained by running the prescreening step on the same dataset, prior to running EPISPOT. However, results should then be considered as exploratory, since this procedure interrogates the same data twice, which is subject to overfitting. The effectiveness of EPISPOT for detecting and exploiting the relevant epigenetic marks for QTL mapping depends on multiple conditions that have a coordinated effect on statistical power. The number of loci analyzed must be reasonably large to hope for the marks to be sufficiently represented at causal loci. These loci must also be densely genotyped or imputed to ensure that the causal SNPs, and the epigenetic marks they may fall into, are included in the analysis. The frequency of each relevant mark among causal SNPs, as well as the strength of its contribution to initiating the QTL effects and the quality of the mark annotation also play a role, as do the degree of co-regulation of traits by the same SNPs, the sample size of the analyzed dataset, the individual effect sizes of QTL associations, and the correlation structures among marks, traits, and SNPs (LD). We examine the impact of these different parameters in a series of simulation studies described in the results and in the supplemental material and methods. Were these conditions not sufficiently met for EPISPOT to borrow information across the loci and learn the mark contributions, the QTL mapping would not benefit from further level of information provided by the marks (no mark selected) but it would nonetheless benefit from the joint analysis of SNPs and traits. Notably, the sparse modeling of the marks implies that the inclusion of marks, were these insufficiently informative, has no risk of deteriorating the QTL mapping (see the ''Null scenario'' section in the supplemental material and methods); this is a major advantage of our method. # Results ## Data generation and simulation set-up The series of simulation studies presented in the next sections have the dual purpose of (1) illustrating the effectiveness of EPISPOT in learning from the epigenome when the epigenetic annotations at hand are sufficiently informative (first simulation study), and (2) evaluating the method in weakly informative scenarios (second simulation study) or scenarios where the module partition supplied to M-EPISPOT is misspecified (third simulation study). We simulate data so as to best emulate molecular QTL regulation and the role of the epigenome in triggering this regulation; the general data-generation procedure is detailed in the supplemental material and methods and we further tailor it to each simulation experiment in their dedicated sections. Here, for simplicity, we represent the presence or absence of a mark at each SNP using a binary variable. In real case scenarios, all types of continuous annotations can be considered without modification since they are encoded as predictors in the second-level regression framework employed by EPISPOT, hence with no distributional assumption. We use the following terminology when referring to the simulated association patterns: d an ''active SNP'' has at least one association with a molecular trait d an ''active locus'' involves at least one active SNP d an ''active trait'' has at least one association with a SNP d an ''active module'' contains at least one trait involved in QTL associations d an ''active mark'' triggers at least one SNP-trait QTL association d the ''hotspot size'' is the number of traits associated with a given hotspot SNP. We benchmark our approach against two representative state-of-the-art methods for QTL mapping, namely, the fully joint Bayesian QTL method ATLASQTL, 5 which is also tailored to the modeling of hotspots but does not accommodate the epigenetic marks, and the widely used marginal screening approach MATRIXEQTL, 2 which tests each SNP-trait pair one-by-one and does not involve any epigenetic information. ## A first illustration We first describe the type of posterior output produced by EPISPOT and its performance in a simple problem where no modules are involved, i.e., the active epigenetic marks exert their influence on all associated SNP-trait pairs. We simulate 32 datasets with an average of 600 molecular traits, r ¼ 500 candidate epigenetic marks and 60 candidate loci, each comprising an average of 20 real SNPs for 413 subjects. These are initial choices are meant to reflect plausible scenarios encountered in real applications, after preselecting candidate loci and candidate traits likely to be controlled by these loci. A subset of 100 SNPs are active (between 0 and 3 per locus; seeand their QTL effects are triggered by r 0 ¼ 3 active marks. This is a strong assumption, which permits a direct illustration of our algorithm in a simple setting but, since it may be unrealistic, we will only use it as a starting point for the more complex numerical experiments that follow. To help interpretability in the context of the simulations, we also generate marks with positive effects only, i.e., inducing QTL activity and not repressing it (supplemental material and methods). Moreover, the large number of candidate marks and the low number of active marks are used to illustrate the ability of EPISPOT to discriminate sparse subsets of relevant marks from whole panels of marks (most of which with no contribution to the QTL effects). The QTL signals are relatively weak: for any given trait, the cumulated QTL effects are responsible for at most 25% of its total variance. Many active SNPs are hotspots; across all 32 replicates, the active SNPs are associated with a number of traits ranging from 1 (isolated QTL association) to 96 (large hotspot), with an average of 27 active traits per active SNP. All these choices will be varied in the subsequent simulation experiments; for an extensive comparison over a grid of scenarios, see the supplemental material and methods.shows that EPISPOT could clearly discriminate the three active marks contributing to the QTL associations from the remaining r À r 0 ¼ 497 inactive marks. The partial receiver operating characteristic (ROC) curves also show that it outperforms ATLASQTL in terms of selecting associated SNP-trait pairs and hotspots. It is unsurprising given that ATLASQTL does not use any predictor-level information, yet it nevertheless confirms that EPISPOT can effectively exploit the marks to enhance the estimation of the primary QTL associations. MATRIXEQTL performs poorly compared to the two joint approaches EPISPOT and ATLASQTL, which is expected since, by design, it does not exploit the shared association signals across traits. We checked that EPISPOT and ATLASQTL display similar performance under simulation scenarios with no active mark: their 95% confidence intervals for the standardized partial area under the curve (pAUC) overlap, i.e., (0.74, 0.78) and (0.76, 0.79) for ATLASQTL, resp. EPISPOT (supplemental material and methods). This further supports the observation that the improvement of EPISPOT seen inis attributable to an effective use of the three informative marks and not to other intrinsic differences between the two models; more evidence on this is provided in the next simulation experiment. Performance under varying degrees of epigenome involvement Effectiveness in QTL mapping is subject to a number of interdependent factors pertaining to (1) the sparsity of the studied QTL network and magnitude of the QTL effects, (2) the amount of information contained in the data at hand, and (3) the ability of the statistical approach to interrogate the data, i.e., by both leveraging and being robust to the dependence structures within and across genetic variants and molecular traits. When it comes to exploiting the epigenome to enhance statistical power, an additional level of complexity is introduced for determining the impact of the above factors on the analysis, and new questions arise as to whether the signal present in the data is sufficient to inform inference on the location of the relevant epigenetic marks and of the QTL associations potentially triggered by these marks. In the previous simulation experiment, we generated data under the simplifying assumption that all QTL associations were induced by the epigenome, and to a degree to which the relevant marks would be detectable, as evidenced by the high epi-PPIs for the active marks and the power gained from leveraging this signal. Here, we focus on evaluating how the level of involvement of the epigenome in QTL activity impacts the detection of QTL effects and of the marks responsible for these effects. We consider a series of QTL problems, each generated by replicates of 32, for a grid of response numbers and degrees of involvement of the epigenome in activating QTL control. More precisely, we simulate data with a number of traits sampled from a Poisson distribution with mean l ¼ 200; 400; 600; 800; 1000, or 1,600 and 60 loci with 20 SNPs each and involving 100 active SNPs in total. We vary the proportion of active SNPs whose activity is triggered by epigenetic marks from p epi ¼ 0 (all QTL associations simulated independently of the action of the epigenome) to p epi ¼ 1 (all QTL associations simulated as the result of the action of the epigenome); see the supplemental material and methods for the data-generation details. The typical pleiotropic pattern simulated is displayed infor the different choices of p epi and problems with an average of l ¼ 600 traits.also shows the performance for the selection of QTL effects in terms of standardized pAUC. It provides two separate layers of information: first, it illustrates again how EPISPOT is able to leverage the epigenetic marks to improve QTL mapping, and more so when the number of active SNPs triggered by these marks increases (top to bottom rows) since EPISPOT is then able to effectively borrow information across the mark-activated SNPs. This underlines the need for the relevant epigenetic marks to be sufficiently represented at causal variants so that the analyzed data are informative about their involvement. It is therefore advised to use a reasonably large number of loci thought to be active and dense SNP panels (e.g., imputed SNPs, see the eQTL case study section), so the active SNPs are more likely to be included. Second, it shows that the joint modeling of all traits permits exploiting shared signals across these traits, thereby also improving statistical power, as reflected by the increased pAUCs for problems with larger numbers of traits in. This is particularly true in the presence of co-regulated molecular traits, a special case of which is the regulation of these traits by a single hotspot.also indicates that, when the epigenetic signal is moderate to large (p epi ¼ 0.4, 0.6, 0.8, or 1), EPISPOT is able to pick the active epigenetic marks from a large number of candidate marks, while setting the epi-PPIs of the inactive marks to zero. However, when the signal is weak (p epi ¼ 0.2), the active marks are barely detected, as expected. Importantly, though, in the null scenario where the epigenome plays no role (p epi ¼ 0), modeling the r ¼ 500 inactive marks does not deteriorate the performance (supplemental material and methods). These experiments also suggest that annotations which are more likely to trigger QTL associations at numerous causal SNPs, such as cell-type-specific enhancers, could have increased opportunities to be picked up and leveraged. This may imply that the QTL mapping would benefit more from the use of general annotations than from that of more specific types of marks, such as ChIP-seq binding sites of transcription factors, which may display a lower degree of sharing between hotspots. Further investigations on real datasets would need to confirm this. However, as there is no intrinsic limitation on the number of candidate annotations supplied to EPISPOT, nothing prevents the analyst from using both general and more specific annotations, and letting the model select the annotations which are sufficiently informative. Finally, the quality of the mark annotation will have a similar impact on performance. We show in complementary simulations (supplemental material and methods) that EPISPOT will not take full advantage of the epigenome if the supplied annotations are of poor quality: the QTL mapping performance declines with the level of noise in the annotations, but EPISPOT remains superior to alternative approaches for which no annotation information is supplied. We also tested the impact of other data scenarios on the ability of EPISPOT to detect and utilize the marks for improving QTL mapping. More precisely, we ran simulations for a grid of configurations, varying: the number of active SNPs, the average QTL effect sizes, the degree of co-regulation of the traits and the hotspot sizes; see section ''QTL mapping performance for a grid of simulated data scenarios'' of the supplemental material and methods. These experiments show that (1) these parameters have a coordinated effect on statistical power, and (2) thanks to its flexible hierarchical representation, EPISPOT is very effective at taking advantage of shared functional patterns, yielding a substantial mapping performance gain. ## Inferring module-specific epigenetic action The simulation experiments presented next focus on evaluating M-EPISPOT, i.e., the module version of the algorithm which models module-specific epigenetic effects. They illustrate how statistical power and interpretability are enhanced when the structure underlying epigenomedriven QTL associations is exploited. They also evaluate the robustness of inference when misspecified module partitions are supplied to M-EPISPOT. This is particularly important given the uncertainty that often surrounds the definition of modules, as reflected by fact that different co-expression inferential tools often produce different module specifications. We start with a simple example involving 60 concatenated loci of average size 40 SNPs and two modules of 50 The crosses indicate hotspots whose activity is triggered by the epigenome and the circles indicate hotspots whose activity is independent of the epigenome. Right: Average epi-PPIs, as inferred by EPISPOT for the simulated scenarios with an average of 600 traits. simulated traits each. In the first module (m 1 ), the traits are largely co-regulated by hotspots whose activity is imputable to the epigenome. In the second module (m 2 ), only few traits are involved in isolated QTL associations, with no implication of the epigenome.illustrates the corresponding simulated QTL pattern restricted to the active SNPs, for the first data replicate. We evaluate the performance of M-EPISPOT with the following settings: ## Mark selection average posterior probabilities of inclusion 1. The oracle case, where we assume the simulated module partition M ¼ fm 1 ; m 2 g to be known and provided it as input to M-EPISPOT; 2. the module-free case, where we perform inference with the base model EPISPOT which does not exploit the module partition; 3. a series of intermediate cases, where the module partition supplied to M-EPISPOT is misspecified, i.e., module m 1 is contaminated with 10; 20; 30 or 40 traits from module m 2. This mimics a real data scenario whereby the assignment of some traits to modules is difficult. The ROC curves ofshow that leveraging information about the underlying module partition can improve significantly the detection of QTL effects. They also confirm the intuition that the impact of misspecified partitions on performance is a function of the degree of misspecification: for a given specificity, the power decreases smoothly with the number of inactive traits from module m 2 contaminating module m 1 . hotspot in a single module permits maximizing the opportunities to learn the epigenetic contribution to the QTL activity by borrowing strength across co-regulated traits. It is advised to make use of prior information on pleiotropy when available in order to avoid splitting hotspotcontrolled networks of traits into distinct modules. The second simulation experiment considers a more general setting with 5 modules of average size 50. It compares ATLASQTL, EPISPOT, and M-EPISPOT with the oracle module partition supplied and M-EPISPOT with a contaminated module partition supplied, i.e., where a fifth of the traits in the simulated modules are randomly re-assigned to the other modules.leads to a conclusion similar to that of the previous example: the idealized scenario of the oracle module partition provided to M-EPISPOT yields the best performance, followed, in order, by the more realistic case of the contaminated partition, the EPISPOT run (with no module information) and finally, the ATLASQTL run which does not make use of any epigenetic information. Importantly, the fact that the module-free version EPISPOT outperforms ATLASQTL indicates that even when the module structure is not employed, the method is still able to leverage the epigenome in order to improve the QTL mapping.also shows how the marks responsible for the activation of the different modules are correctly recovered by M-EPISPOT. An inspection of these separate sets of marks provides a refined level of interpretability for a module-specific understanding of the genetic control. We will see in the eQTL analysis presented next how this can be particularly helpful to shed light on the mechanistic action of trans hotspots, when such hotspots are thought to control gene modules in a context-specific way. An epigenome-driven monocyte eQTL case study In this section, we take advantage of EPISPOT in a targeted eQTL study to refine the detection and characterization of genetic regulation in monocytes. Specifically, we analyze two independent datasets with transcript levels measured in CD14 þ monocytes. Our study workflow is described in : we discover active loci in a prescreening step using the joint hotspot QTL mapping approach ATLASQTL 5 in the first dataset (n ¼ 413 samples, and we then leverage the epigenome using EPISPOT in the second dataset (CEDAR cohort, n ¼ 286 samplesfor an in-depth analysis of the genetic activity in the preselected loci. The epigenetic information consists of a panel of 168 annotation variables, compiling DNase-I sensitivity sites from different tissues and cell types, Ensembl gene annotations, and chromatin state data from ENCODE. These variables display strong correlation structures within annotation types, as well as within tissues and cell types at a finer granularity level . Details about the prescreening step, as well as the epigenetic, genetic, and expression datasets are given in the supplemental material and methods, and the eQTL associations for the prescreen-ing and subsequent analyses are listed in Tables S1, S2, S3, and S4. In this case study, we concentrate our attention on the following key finding revealed by the prescreening step: chromosome 12 is highly pleiotropic, notably around the gene LYZ (MIM: 153450). This gene encodes lysozyme, a highly conserved enzyme with peptidoglycan-lytic activity that is robustly expressed in monocytes. The LYZ locus has already been reported as pleiotropic using several monocyte datasets, 27-29 but its functional role remains unclear. We will therefore exploit the epigenetic annotations within EPIS-POT to shed light on the mechanisms of action of this locus as well as of other surrounding cis-and trans-acting loci. Importantly, while our discussion will mainly concentrate on a few pleiotropic loci of interest, EPISPOT will be applied on a whole collection of loci from chromosome 12, which display QTL signal according to the ATLASQTL prescreening at 5% FDR. By borrowing strength across all the loci (learning from hotspot signals, as well as isolated cis and trans signals), EPISPOT will infer the epigenetic contributions to the QTL activity of the different regions. The LYZ-region pleiotropy defines two modules of transcripts A total of 977 eQTL associations, involving 350 unique SNPs on chromosome 12 and 430 unique transcripts genome-wide, were identified at FDR 5% from the AT-LASQTL prescreening analysis of the first dataset. When mapped to the CEDAR dataset, the ATLASQTL eQTLs corresponded to 195 independent loci, expected to involve distinct eQTL signals and comprising a total of p ¼ 1,540 SNPs (see and data-preparation details in the supplemental material and methods). As highlighted in the second simulation study (section ''performance under varying degrees of epigenome involvement''), supplying a dense panel of SNPs (here imputed SNPs) to EPISPOT is important to ensure a sufficient representation of the relevant epigenetic marks among the analyzed SNPs. We also mapped the prescreened transcripts to the CEDAR dataset. The LYZ-region pleiotropy defines two natural modules of transcripts, based on whether they are associated with SNPs in the vicinity of LYZ (<1 Mb from it) or not, and further augmenting these modules with highly correlated transcripts (supplemental material and methods). This module partition is driven by the following biological consideration: the peculiar pleiotropic QTL activity arising from the LYZ region may be triggered by specific epigenetic influences, which may differ from those triggering isolated (scattered) cis or trans effects outside the LYZ region; to reflect this, the modules are hereafter referred to as the pleiotropic module and the scattered module, respectively . The correlation structure within and across the two modules supports this partitioning . Namely, it indicates a strong co-expression of transcripts within the pleiotropic module, suggesting a dense network of genes whose connections may be attributed in large part to the shared QTL control exerted by the LYZ hotspots. Conversely, the transcripts in the scattered module display A B C D . Overview of the monocyte eQTL case study (A) Workflow for the monocyte eQTL case study. Candidate loci from chromosome 12 and transcripts are obtained from a preliminary prescreening in the first dataset 25 using the joint eQTL mapping approach ATLASQTL 5 with a permutation-based Bayesian false discovery rate (FDR) of 5% for selecting pairs of associated SNP-transcript. The analysis is then performed in the second dataset (CEDAR).EPISPOT and M-EPISPOT select associated SNP-transcript pairs, QTL hotspots, and epigenetic marks relevant to the primary QTL associations. This output is then interpreted as a whole to generate hypotheses about the mechanisms of action underlying these associations. (B) Correlation of the analyzed transcripts according to their module membership. The ''pleiotropic module'' displays a strong dependence pattern, reflecting dense connections in the network controlled by the hotspots; the traits in the ''scattered module'' are mostly uncorrelated, which is unsurprising given that they are mainly controlled via isolated cis mechanisms. (C) Correlation of the epigenetic annotations supplied to the method. All variables are binary, except the distance to the closest transcription start site (TSS) which is not included in the heatmap. Only the labels of the marks retained by M-EPISPOT are displayed; a heatmap with the full labels is provided in the supplemental material and methods. The majority of the marks are DNase-I hypersensitivity sites (DHSs) in different tissues and cell types. They tend to cluster together on the top left 4/5 of the heatmap, and DHSs in similar tissues and cell types also form subgroups. The remaining marks relate to gene structures and genome segmentation annotations. The labels indicated on the right are in gray and black depending on whether they were selected by M-EPISPOT as relevant for the pleiotropic, resp. scattered module. The þ and À indicate positive, resp. negative effects of the marks, i.e., their triggering or repressive action on the primary QTL effects. Their relevance is discussed in the main text and in the supplemental material and methods. (D) Hotspot sizes (i.e., number of associated transcripts per SNP) as inferred by M-EPISPOT. Only the active SNPs (i.e., associated with R 1 transcripts) are displayed. The gray and black colors indicate the module membership of the controlled transcripts. The numbers in parentheses refer to the discussion of the main text. little co-expression, which is unsurprising given that they tend to be involved in isolated QTL effects (most transcripts are controlled by distinct genetic variants). Overall comparison of methods and replication rates We next refined our understanding of the eQTL structure in this region using the CEDAR dataset. To assess the sensitivity of inference to this module partition, we compared the results of the module-based algorithm, M-EPISPOT, with those of the base algorithm, EPISPOT, i.e., with no module provided as input. Moreover, to highlight the benefits of using epigenetic information, we also confronted these two runs with an ATLASQTL analysis of the same data. We employed the same settings for all three runs to set common grounds for comparison. In particular, we used a same permutation-based Bayesian FDR threshold of 5% for declaring QTL associations and supplemental material and methods). Importantly, the simulated annealing scheme implemented as part of the EPISPOT algorithm is specifically designed to handle the strong LD structures present in the dense SNP panel data and the block correlation structures among transcript levels and epigenetic marks ; an illustration for different degrees of LD is given in the supplemental material and methods. In the CEDAR dataset, the M-EPISPOT analysis of the two modules (q ¼ 283þ191 transcripts) and the 195 candidate loci (p ¼ 1,540 SNPs) identified 514 eQTL associations, involving a total of 267 unique transcripts and 82 unique loci. In terms of independent replication of the prescreening hits, this corresponds to rates of 78.2% and 55.8% for the cis and trans QTL associations, respectively. Using ATLASQTL instead of M-EPISPOT on the CEDAR data yielded 262 unique active transcripts and 80 unique active loci, with slightly lower cis and trans replication rates, namely 77.9% and 54.9%, respectively. Similar observations were obtained for EPISPOT. Given the well-known difficulty to validate trans effects and the relatively small sample size of the CEDAR dataset (n ¼ 289), these appreciable independent replication rates may result from the efficient joint modeling of all transcripts and SNPs achieved by M-EPIS-POT, EPISPOT, and ATLASQTL. A focus on two susceptibility loci We next discuss two examples of pleiotropic loci. First, not only does M-EPISPOT confirm the LYZ pleiotropic activity , but it also uncovers associations of this locus with four additional genes compared to the ATLASQTL run, namely, COPZ1 (MIM: 615472), DPY30 (MIM: 612032), KLHL28, and OSTC (MIM: 619023). The EPISPOT run (with no module partitioning) reports the exact same list as ATLASQTL, also missing the above four genes. The second example is a pleiotropic locus uncovered by M-EPISPOT and for which only isolated effects were detected at the prescreening stage . This locus is located 32 Mb downstream to the LYZ locus and entails a hotspot of size 52 in the gene body of GNPTAB (MIM: 607840), namely, rs10860784 (r 2 ¼ 0:001 with the lead hotspot rs10784774 of the LYZ locus). The trans network formed by the controlled transcripts has not been previously described and neither has any trans-acting effect involving rs10860784 (up to proxies using r 2 > 0:8). However, rs10860784 is known to be cis-acting on DRAM1 (MIM: 610776) (located 98 Kb downstream) in multiple tissues,an association which M-EPISPOT also confirms using a looser FDR of 15%. Moreover, the UK Biobank PheG-WAS also reported 31 a strong association between this SNP and height (MIM: 606255) (p ¼ 1.47 3 10 À14 ). The module-free version EPISPOT run also finds a trans network for the exact same SNP, yet slightly smaller, as it involves 31 transcripts at FDR 5%; ATLASQTL finds no signal. This example suggests that the added value of epigenome-driven inference is particularly striking for the detection of weak trans signals. Indeed, a comparison of the estimated QTL effects attributable to rs10860784 with those attributable to LYZ pleiotropic locus shows that the former are significantly smaller in magnitude compared to the latter (t test p < 2 3 10 À16 ). The selected epigenetic annotations reveal possible genetic mechanisms of action The above figures suggest that the M-EPISPOT and EPIS-POT runs allow for more powerful QTL mapping compared to ATLASQTL. This probably results from their ability to leverage the epigenetic marks, as we next discuss. For each module, M-EPISPOT identifies a subset of epigenetic annotations with a potential to induce or inhibit the Number of eQTL associations discovered by the ATLASQTL prescreening (chromosome 12) and by each of the three (M-EPISPOT, EPISPOT, and ATLASQTL) analyses of the CEDAR data, along with the replication rates for the associations discovered at the prescreening stage. All analyses use an FDR threshold of 5%. The numbers of samples n, SNPs p, and transcripts q are indicated for each dataset. Full lists of eQTL associations for the different methods are provided in Tables S1, S2, S3, and S4. QTL associations (depending on the sign of the posterior mean of each annotation effect); these annotations are highlighted in . For instance, DNase-I hypersensitivity sites (DHS) in fibroblasts and epithelial cells of different tissues tend to promote the QTL effects. Interestingly, DHS in CD14 þ monocytes are found to be enhancers of eQTL effects in both the M-EPISPOT and EPISPOT runs, with epi-PPI > 0.99. The two runs also estimate a negative effect of the distance to transcription start sites (TSSs, epi-PPI > 0.99), in line with the frequently reported decay in abundance of eQTL signals with the distance to TSS.These last two observations are helpful to interpret the uncovered QTL signals, as we next discuss. CD14 þ cell DHS: Hints to a monocyte-specific pleiotropic activity in LYZ We first focus on the LYZ pleiotropic region. Previous studies have highlighted distinct lead hotspots around LYZ, 33 yet none provided a functional characterization that would allow a clear prioritization of one variant over another. The lead hotspots revealed by the M-EPISPOT and EPISPOT runs are intergenic variants, rs10784774 (size 154) and rs2168029 (size 109, r 2 ¼ 0:89 with rs10784774; see . They differ from the lead hotspot flagged by the ATLASQTL run, namely, rs1384 (size 149, r 2 ¼ 0:99 with rs10784774). We next examine the possible biology behind these candidates, starting with the ATLASQTL top hotspot. The fact that rs1384 is located within the 3 0 UTR of LYZ may suggest a trans action mediated by LYZ. This hypothesis is plausible given that the locus associates with LYZ in all M-EPISPOT, EPISPOT, and ATLASQTL runs and that GTEx also reported this cis association in whole blood and different tissues. Conversely, regressing out the effect of LYZ on the expression matrix does not explain away the hotspot effects (the size of the top hotspot in the LYZ locus is only marginally reduced: 134 versus 154 in the original M-EPISPOT analysis, supplemental material and methods). This does not rule out LYZ expression initiating the formation of the hotspot, but the downstream consequential changes in expression are too complex to simply regress out in a linear manner, and so only reduced mediation is observed. The monocyte-specific DHS annotation selected by M-EPISPOT for the pleiotropic module suggests a complementary scenario. Namely, the pleiotropic activity of the locus may be triggered by cell-type-specific enhancers in open chromatin regions, which are known to be key players in activating the transcription in trans.This hypothesis of monocyte-specific pleiotropy would also explain why no hotspot was reported so far in cell types and tissues other than monocytes.To investigate this further, we performed an additional enrichment analysis using the multiple tissue-and cell-type histone modification marks of the ENCODE catalog: we found that the two sets of genes associated with the M-EPISPOT's lead hotspots rs10784774 and rs2168029, respectively, are enriched in H3K27ac enhancers, again in CD14 þ monocytes only, which further supports cell-type-specific activation. One notable gene in this group is the transcription factor CREB1 (MIM: 123810), which has previously been suggested as a putative mediator of the LYZ pleiotropic network.Notably, regressing out the effect of CREB1 on the expression matrix substantially reduces the pleiotropy of the locus (the size of the top hotspot in the LYZ locus is 36, versus 154 in the original M-EPISPOT analysis). Moreover, the connectivity of the transcript conditional independence network is also markedly lower (supplemental material and methods). It seems most plausible, however, that the trans-mediation effect by CREB1 may be preceded by a cis effect on LYZ or an isoform-specific effect. This possibility is supported by a strong divergent allele-specific correlation between LYZ and CREB1, which we observed when conditioning on the genotype of the lead hotspot rs10784774 (supplemental material and methods). This indicates an indirect cis-trans-cis mediation of the trans network by LYZ-mediated CREB1 expression differentially feeding back onto LYZ, an observation replicated in both datasets analyzed. Notably, scanning SNP effects on transcription factor binding motifs identifies putative divergence in CREB1 binding dependent upon allelic carriage at rs10784774, in keeping with the allele-specific correlation observation (supplemental material and methods). While our analyses of residual values cannot completely resolve this, such a feedback circuit might explain why the effect of regressing for CREB1 is greater than the effect of regressing for LYZ. Finally, it has previously been noted that EP300 (MIM: 602700), a binding partner of CREB1, shows allelic effect on LYZ expression,although this in an opposing manner to that observed for CREB1 alone, and importantly, the effect size of the EP300 association is markedly less than that for CREB1. In total, these observations lend further weight to allele-specific regulation via rs10784774, although, given that CREB1 and EP300 may form components of multi-protein complexes, the fine mechanistic details of this regulation fall outside the scope of this publication. We further explored whether the two sets of genes associated with either rs10784774 and rs2168029 were enriched in transcription factor binding sites (TFBS) using the ENCODE data in K562 cells. We found a profound enrichment of a number of TBFS, including ATF3, CREB1, and c-Myc . The networks of transcription factors for rs10784774 and rs2168029 are similar, indicating conserved regulatory networks, although unlike with rs10784774, rs2168029 does not overlap a CREB1 binding site and therefore would not be proposed to feedback here. Interestingly, ATF transcription factors are CREB-binding proteins, in line with the CREB1-mediation hypothesis, but the strong enrichment for many other transcription factors suggests that the same loci can be targeted by different processes and the co-occupancy of these loci in primary monocytes may resolve this further, although is important to note that, unlike the very significant association between LYZ and CREB1, there is no association between LYZ and ATF3 expression, so we can discount this gene playing a role in this genomic circuit. The c-Myc transcription factor is involved in cell division and has broad transcriptional consequences,which is sensible given the large pleiotropy observed at the LYZ locus, for rs10784774 and rs2168029. Consistent with this, the UK Biobank data further reveal strong associations of these two SNPs with monocyte counts and other myeloid cell counts.Although by no means conclusive, these observations corroborate the context specificity of the trans effects controlled by the LYZ locus, and indeed may be more representative of other unresolved trans loci across the genome that, while of potential high biological importance, lack the pleiotropic effect of the LYZ locus. They also suggest that the epigenome-driven EPISPOT runs found promising candidate hotspots, whose presumed mechanisms of action on the massive LYZ gene network would merit experimental follow up. Distance to TSSs: Examples of cis and hotspot signals shared across cell types Another interesting result concerns the negative effect of the annotation coding the distance to TSSs, this time for transcripts belonging to the scattered module. As active transcripts in this module are mostly involved in cis associations, the module specificity of this annotation aligns with the previous observation that the distance to TSSs associates with an enrichment of cis eQTLs.Moreover, an empirical assessment of this enrichment in our dataset shows that the SNPs selected with M-EPISPOT are on average significantly closer to TSSs compared to SNP subsets of the same size randomly drawn within the analyzed loci (p ¼ 0.017). Such an enrichment is unsurprising and actually also present in the EPISPOT and ATLASQTL results, but the importance of the distance to TSS is nevertheless made explicit by the selection of the TSS variable by both EPISPOT and M-EPISPOT. For instance, three candidate hotspots, rs10876864, rs11171739 (r 2 ¼ 0:94 with rs10876864), and rs705699 (r 2 ¼ 0:86 with rs10876864), located 13 Mb upstream of the LYZ locus, are representative of this enrichment as they are within a TFBS, a 5 0 UTR and an exon, respectively -iii). Our ATLASQTL prescreening and EPISPOT analyses find that they control a small network of size 11 involving transcripts mapping to the cis gene RPS26 (MIM: 603701) and other distal genes, including IP6K2 (MIM: 606992) on chromosome 3. This locus has been linked with several autoimmune diseases 38-41 including type 1 diabetes (MIM: 222100), where evidence exists that RPS26 transcription does not mediate the disease association.Interestingly, previous studies have reported the RPS26 cis effect as an isolated association in monocytes. The trans activity, in particular on IP6K2, was unknown in monocytes, but is known in B and T cells.This suggests that it has so far gone unnoticed in monocytes using standard univariate mapping approaches, but our fully joint, annotation-driven method has enabled its detection. Moreover, unlike the monocyte-specific LYZ pleiotropic locus discussed above, this locus is an example of trans-hotspot eQTL present in several cell types. The genomic location also aligns with the observation that eQTLs common to multiple cell types or tissues tend to be closer to TSSs compared to eQTLs only detectable in a single cell type or tissue. # Discussion Large panels of epigenetic marks are nowadays collected along with genetic data and employed as part of different modeling approaches, whether for single-trait association studies or fine mapping.However, their use to enhance molecular QTL mapping remains mostly heuristic. Thanks to its hypothesis-free mark selection routine which is fully integrated within a joint QTL mapping framework, EPIS-POT can identify the relevant epigenetic marks from thousands of candidates, while also directly refining estimation in large molecular QTL studies. Specifically, EPISPOT brings important modeling and algorithmic contributions. First, it implements a flexible hierarchical model which enables parametrizing both cis and trans actions on thousands of molecular traits, whereas existing epigenome-based approaches are limited to GWAS or cis QTL mapping for one or a handful of traits.Second, it is both fully joint and scalable, accounting for all epigenetic marks, genetic variants and molecular levels, and their shared signals, in a single modeling framework. Third, it combines this information to perform an automated selection of the epigenetic marks relevant to the QTL effects of the problem at hand, thereby providing direct insight into the functional basis of the signals. Fourth, its crafted annealed variational algorithm ensures a robust exploration of complex parameters spaces, such as induced by candidate SNPs in high LD, corresponding to scenarios for which the use of epigenetic information is particularly beneficial. Finally, EPISPOT allows for module-specific learning of the epigenetic action. We showed in a series of simulation experiments emulating epigenome-driven QTL problems that EPISPOT effectively scales to large datasets, while retaining the accuracy necessary for a powerful QTL mapping. We demonstrated that our method was not only able to pinpoint the correct marks with high posterior probability, but that it could also leverage these marks to improve the detection of weak QTL signals. In particular, we saw that the spike-and-slab representation of the epigenome contribution ensures that the irrelevant epigenetic marks are effectively discarded as ''noise,'' so panels with hundreds of candidate marks can be considered without the risk of worsening inferences. This allows skipping the delicate process of pre-filtering marks, whose practical grounds are often blurry and disconnected from the QTL dataset under consideration. Moreover, although in a strict sense epigenetic marks represent a subset of functional annotations, it is possible to interpret this terminology more loosely and supply other types of annotations or scores that may carry information about the involvement of SNPs in QTL regulation. Our work attaches special importance to acknowledging the complexity of the learning task (selection of hotspots, pairwise QTL associations between variants and molecular traits, selection of epigenetic marks relevant to these QTL associations) and possible biological scenarios (pattern of regulation, importance of the epigenome in this regulation, dependence structures among variants, marks and molecular traits, and between them). Our simulations examined under what conditions inference is well powered to leverage the epigenetic information and evaluated the sensitivity to different input choices, in particular when gene modules are provided. Importantly, our method is not meant to be used as a black box to fish genetic variants involved in trans regulation and their epigenetic roots, but rather is predicated on a careful analysis design that takes into account the dataset, the biological question of interest, and the expected statistical power. Further assessments for specific problem settings (sparsity levels, association patterns, and epigenetic control) can be made using the code provided online (see EPISPOT and ECHOSEQ in web resources). Finally, we showed how our simulation studies prefigured the efficiency of EPISPOT in a large monocyte eQTL study (high replication in an independent sample, previously unreported pleiotropic loci, refined list of candidate lead hotspots). We further illustrated how the EPISPOT posterior output can be used to both select interpretable annotations underlying the QTL activity and reduce the range of hypotheses about the functional mechanisms involved, particularly for hotspots. We also showed how the localized nature of QTL activity could be accounted for when inferring annotations in a module-specific fashion using M-EPISPOT (the monocyte-specific enhancer activity affecting the pleiotropic module, the enrichment of QTL hits closer to TSSs affecting the scattered module). Altogether, this thorough case study demonstrates that QTL analyses may largely benefit from the use of rich complementary data sources annotating the primary genotyping data, provided principled joint approaches are used to capture shared association patterns. EPISPOT offers perspectives for robust and interpretable molecular QTL mapping, toward a better understanding of the functional basis of genetic regulation. Thanks to its efficient annealed VBEM algorithm with adaptive and parallel schemes, it enables information sharing across epigenetic marks, genetic variants, and molecular traits governed by complex regulatory mechanisms, at scale. In particular, its use of selection indicators in a spike-andslab framework allows for a systematic identification of sparse sets of epigenetic annotations which are directly relevant for the QTL regulation of the problem at hand. We envision holistic approaches such as EPISPOT to be increasingly adopted in an age where large molecular datasets and annotation information become widely available. EPISPOT is applicable to any type of molecular QTL problem, involving genomic, proteomic, lipidomic, or metabolomic levels, but also to genome-wide association with several clinical endpoints. In particular, exploiting the epigenome to build finer maps of hotspots across the genome holds great promises, as these master regulators are likely to be triggered by tissue-and cell-type-specific epigenetic functions. ## Data and code availability Fairfax et al.provide gene expression in CD14 þ monocytes and genotyping data from individuals with European ancestry. The raw expression data were generated with Illumina Hu-manHT-12 v4 arrays and downloaded from ArrayExpress 45 (accession E-MTAB-2232), while the raw genotyping data were generated by Illumina HumanOmniExpress-12 arrays and have been deposited at the European Genome-Phenome Archive (accessions: EGAD00010000144 and EGAD00010000520). The expression data are freely available, but the genotyping data require a data access agreement, as detailed in Fairfax et al.and https://www.well.ox.ac.uk/research/research-groups/julian-knightgroup/research-projects/data-access. The CEDAR datasetconsists of gene expression data from CD14 þ monocytes and genotyping data from individuals with European ancestry. The raw expression data were generated with Illumina HumanHT-12 v4 arrays and downloaded from ArrayExpress 45 (accession: E-MTAB-6667), while the raw genotyping data were generated by Illumina HumanOmniExpress-12 v1_A arrays and downloaded from ArrayExpress (accession: E-MTAB-6666). Both the expression and genotyping data are freely available. Both studies were approved by the local human research ethic committees, namely, the Oxfordshire Research Ethics Committee (COREC reference 06/Q1605/55)and the University of Liège Academic Hospital Ethics Committee.Participants provided informed written consent, and all procedures were conducted in accordance with the Declaration of Helsinki. All statistical analyses were performed using the R environment (v.3.6.1) 46 and the synthetic datasets were generated using the freely available R package ECHOSEQ (v.0.3.0). The R package EPIS-POT implements the method. ## Supplemental information

Protease-associated import systems are widespread in Gram-negative bacteria Thank you very much for submitting your Research Article entitled 'Proteaseassociated import systems are widespread in Gram-negative bacteria' to PLOS Genetics. Your manuscript was fully evaluated at the editorial level and by independent peer reviewers. The reviewers appreciated the attention to an important problem, but raised some substantial concerns about the current manuscript. Based on the reviews, we will not be able to accept this version of the manuscript, but we would be willing to review again a muchrevised version. We cannot, of course, promise publication at that time.As you will see from their comments, the three reviewers agreed that this is an interesting study that is clearly written and that it presents new insight into this protease-associated import system. However the three reviewers agreed also that the manuscript is missing a bit more experimental and biological results to support the structural work and its resultant conclusions. Each of the reviewers pointed to missing experiments.We thank Prof. Buchrieser and Prof. Casadesús for their acknowledgement of the importance and interest of our work and for the constructive suggestion to improve our manuscript, based on the reviewer comments. We feel that we have fully addressed these suggestions and the concerns of the reviewers in the revised version of the manuscript, either through additional experimental work or through reference to relevant studies in the literature. Specifically:For instance the authors should confirm that that the Ab does not bind to similar region of FusC,While E. coli does not contain FusC, we have shown that our PqqL antibody is specific, with no band observed for PqqL in a western blot of E. coli ΔpqqL.They should determine if (separate, non-polar) mutations in YddB and PqqL impair the ability of E. coli to grow in low-iron, as this would be indicative of a role in iron import,We have shown that the loss of YddB and PqqL does not impair the growth of E. coli in low-iron, when grown under lab conditions. cellular fractionation, immunoblotting should be done for YddB as it was for PqqL (Fig 4B)and/or microscopy for surface localization.Due to our inability to raise and antibody to YddB, this experiment was not possible. However, we now reference a number of studies that detect YddB in the outer membrane of E. coli. In addition, we draw attention to the structural features of YddB that indicate outer membrane localisation.Furthermore the Western blots should be quantifiedWe have now quantified all western blots presented and crystallographic statistics after anisotropy correction for YddB, especially cc1/2 and I/s and the validation reports for all the structures, need to be ## Provided. We have revised the crystallographic data for the study and now present all relevant/requested statistics. ## Comments to the authors: Reviewer #1: Grinder and colleagues are investigating the structure and function of the uncharacterised proteins YddB and PqqL from Ecoli that show some similarity to the FusC and FusA proteins that are involved in iron acquisition from ferredoxin. FusA and FusC encode for an outer membrane receptor that uptakes ferredoxin whereas FusC is a peptidase that releases iron from ferredoxin. Using sequence analysis they show that YddB and PqqL show similarity to the Fus system. Cell growth assays under limiting iron conditions shows that the proteins are upregulated. They also solved the crystal structure of both proteins and they show very similar fold to the Fus proteins suggesting that they are likely distant homologues. Functional data, revealed that YddB has protease activity. Overall, they provide good evidence that these systems are well conserved among proteobacteria. We thank the reviewer for the positive assessment of our work and constructive comments regarding the manuscript. Some issues that need to be addressed: 1. The authors show that under limit iron conditions, the PqqL and YddB are over expressed. Considering the high degree of similarity between PqqL and FusC, can they be confident that the Ab does not bind to similar region of FusC. A control experiment whereas a knockout of PqqL in the presence of Abs should be performed to strengthen this claim. They should also provide information on how the Abs were raised. Possibly the reviewer misunderstood here. FusC is present in Pectobacterium spp., but not in E. coli. E. coli possesses PqqL/YddB instead of FusC/FusA, not in addition to FusC/FusA. Therefore, the E. coli cell extracts tested in this work to not contain FusC. Keeping this in mind, we agree with the reviewer that in order to confirm that the PqqL antibody we raised is specifically detecting PqqL, a western blot of extracts from an ΔpqqL strain is a good control. We have now performed this experiment, which confirms that no band corresponding to PqqL is present in the ΔpqqL strain and included it in , in the revised manuscript. We have included the details of PqqL antibody production in the methods under 'Reagents and Antisera generation' of the revised manuscript. 2. The authors have nicely shown that PqqL displays protease activity using a peptide screen. Since they claim that this system is homologous to FusC, why not perform the activity in the presence of ferredoxin? Do the identified peptide sequences much the ferredoxin sequence that they could map them on? We have now tested the activity of PqqL against a panel of small iron containing proteins, including plant ferredoxin, human ferredoxins I and II and a number of globins. Despite its activity in the peptide assay, PqqL exhibited no activity towards these proteins, suggesting they are not the substrate of YddB/PqqL. These data are shown in of the revised manuscript. This finding is consistent with the high specificity of FusC towards its ferredoxin substrate shown with work published by[1]. Interestingly, the peptides we identified as being cleaved by FusC in our peptide assay do not correspond to the sequence of the FusC cleavage sites in ferredoxin. It is worth noting that the two sites on ferredoxin cleaved by FusC also do not share the same sequence being, EAG*IDL and QSF*LDD. As discussed in this paper and our previous work, we feel that the cleavage mechanism of this family of proteins is likely to be complex and the sequence specificity of these proteases may vary based on the conformation of both substrate and protease. Clearly further investigation of the mechanisms of cleavage of these proteins will yield exciting results. 3. I am concerned with the high Rmerge values for all data sets at low resolution. Is it possible that the data suffer from pseudo symmetry? Is the redundancy for PqqL full length really 125? Despite its long use as a standard for judging the quality of X-ray diffraction data, Rmerge is not a good measure of data quality as it increases relative to data redundancy. To avoid arbitrarily discarding useful high resolution data, it was recommended by experts in data processing that it be abandoned as means of determining where to truncate diffraction data. For data collected in this study we have employed a dual cutoff of CC1/2 >0.5, I/sigma =/>1.5. The higher than expected Rmerge values result from the redundancy of the data and the inclusion of these weak reflections, no pseudo symmetry detected in our data. The redundancy for PqqL is indeed 125. At the time of collection of these data, we didn't have sufficient phasing power to solve the structure by molecular replacement. We attempted extensive derivatisation with heavy atoms and tried to grow selenomethione labelled crystals but with no success. The high redundancy of the data was intended to maximise the anomalous signal from the Zn we hypothesized to be present in the enzyme active site. Ultimately the phasing power from this Zn atom also proved insufficient, however all the data collected was of high quality so we elected to keep it in the final dataset. As outlined in the methods, the structure was eventually solved by crystallising a trypsinised fragment of the protein (roughly the first half), solving the structure of that by MR, and using this structure for molecular replacement of the full length protein. The authors should provide crystallographic statistics after anisotropy correction for YddB, especially cc1/2 and I/s. In the initial processing of the YddB data we applied anisotropic correction to the data after it had been scaled and merged. This does not allow for reporting of cc1/2 and I/s for the anisotropic correction. In response to the reviewer's comment, we reprocessed the original data in XDS and scaled it using XSCALE. We submitted these unmerged data to the anisotropy server, however we were unhappy with the quality of the maps resulting from the data processed this way. As such, we have elected to process the data to a new resolution cutoff of 2.4 angstrom and not to apply anisotropy correction. As the anisotropy of these crystals isn't extreme, we are discarding a minimum of useful data this way and the maps are of high quality. We have re-refined our structure with these data and amended the PDB deposition. The crystallography data statistics table has been updated accordingly. In the revision they should provide the validation reports for all the structures. They should also list Ramachandran statistics. These data have been included with the current submission. ## Reviewer #2: Grinter et al. recently discovered and characterized a unique class of protein import systems dedicated iron uptake from ferredoxin in Pectinobacterium. The system consists of an outer membrane TonB-dependent porin FusA and a periplasmic protease FusC. Here, by using a combination of bioinformatic, biochemical and structural analyses, these authors show the presence of functionally analogous and structurally similar systems in a range of proteobacteria. To support their bioinformatics analysis, they determined the structure of a related system from Escherichia coli comprising the outer membrane component YddB and a periplasmic protease PqqL. They show that PqqL is induced upon iron limitation in E. coli, supporting the role of the system in iron scavenging from an iron-containing protein. PqqL structure determination and its comparison to that of FusC reveals a protein composed of two domains connected by a short linker. These domains adopt a closed conformation in the presence of substrate and an extended one in its absence. This conformational transition is thoroughly characterized by SAXS and molecular dynamics analysis in PqqL and FusC. The authors also determine the substrate specificity of PqqL and demonstrate that it is rather narrow, in line with its role in cleavage of a specific substrate. The study is original, well executed and the article is very clearly written. The study reveals important information on this new class of systems, by showing that many proteobacteria have the capacity to take up proteins from the environment. The study therefore provides a basis for a vast field of research that might reveal other biological functions of these protein import systems. We thank the reviewer for their interest in our work and their positive comments regarding our manuscript. We agree that this study forms the basis for a field of research on protein import in Gram-negative bacteria. Minor comments: 1. The PqqL western blots appear to be nonlinear and there is a clear difference in protein and control levels between the two strains in . The recent guidelines require that the linear range of detection be determined and the Western blots be quantified to support the claim that there is more PqqL in urine than in the presence of BiP. The loaded samples correspond to how many bacteria? The western blotting for the original study was visualised using X-ray film, a number of exposures were collected for each blot. SurA bands from different film exposure lengths were shown for CFT073 and BW25113 in the original figure, leading to the apparent differences in the levels of this control. Quantification from these original blots was difficult due to the overlap of the SurA control band and the lower non-target band. Therefore, we repeated this experiment and visualised the blots using a digital CCD-camera detection. To ensure the expression levels we observed were representative, we performed 3 biological replicates of this experiment. We quantified these blots ensuring that band intensity did not saturate the detector and have presented these data in a new experiment. Over the course of these experiments, while PqqL protein levels in E. coli CFT073 were generally higher than BW25113 when grown in urine, they were variable. Thus we have revised the claim that CFT073 differentially expressed PqqL in urine from the current version of the manuscript. We now only make the general claim that PqqL levels are higher under iron limiting conditions. 4.8x10 7 cells were used for each sample in these blots, after normalisation by absorbance at OD600 nm, serial dilution and colony counting was performed to determine cell numbers. We have also quantified the intensity of bands from the original localisation blot we performed from scanned X-ray film. These new data are presented in . be depicted with a different color for better contrast. In addition, the color of the zoomed area in should be same as in . is now orange. ## I suggest that the ion in ## This change has been made to the manuscript, the zn ion in ## Could the authors describe what was their positive control in the fret assay for peptide specificity? There isn't a positive control in this assay as it screens for the presence or absence of peptidase activity. Wells containing pools of peptide that aren't cleaved upon addition of the protease don't emit a FRET signal, while those in which peptides are cleaved do. The activity of the protease against individual peptides from the original pool is then confirmed in an individual FRET assay.. Lines 133-139. The authors advance a claim that bacteria containing these protein import systems tend to associate with plant or animal hosts, a claim immediately contradicted by their presence in marine bacteria. It may be better to avoid any general claims at this point as too little is known about their functions or "specific lifestyles". As for most TonB-dependent transporters, these systems are likely to promote uptake of scarce nutrients from the environment. We agree with the reviewer that stating these systems are generally associated with plant and animal hosts is over interpretation of the available data. We have revised the manuscript (lines 132-139) with the following text to moderate this claim: 'Genome metadata was mined to determine the environment from which the bacteria had been isolated, showing they adopt a variety of lifestyles, which tend to correlate with sequence cluster . For example, members of the clusters defined by Pectobacterium and Hemophillus sequences adopt a commensal or pathogenic relationships with plant or animal hosts, while members of the Marinomonas, Marinobacter and Pseudoalteromonas clusters were isolated from marine or other environmental samples ## Corrected Reviewer #3: Previous work done with Pectobacterium spp. had shown that the outer membrane protein FusA and the periplasmic protein FusC proteins conjoin to import and then degrade ferredoxin as a means toward iron assimilation. In this very interesting follow-up, Grinter et al show that i) gene clusters (proteins) related to FusA/FusC exist in many types of Proteobacteria, ii) the structure of the E. coli protein YddB is similar to that of FusA, and iii) E. coli PqqL is a periplasmic protease that is induced by lowiron growth conditions and is structurally similar to FusC. The MS is very well written and interesting. The structural biology work that was done is especially impressive. The MS' conclusions are generally appropriate. Thus, the findings here have implications for many Gram-negative's, including both plant and animal pathogens. However, the MS would benefit from the inclusion of more "biology" (points 1 & 2) and genetic analysis (point 3), in order to strengthen the conclusions made. We thank the reviewer for their interest in our work and the positive suggestions regarding our manuscript. We have incorporated the suggested changes into the revised version. Major points 1. Given that the homologs of YddB and PqqL are involved in iron assimilation and that the current study finds PqqL to be more highly expressed in low-iron, there should be some attempt to determine whether YddB-PqqL promotes iron assimilation in E. coli. (That past work by others had shown that YddB is important in systemic infection by a strain of UPEC does not alone make this point.) It is true, as the authors mention in their Discussion, that the substrate for the system need not be the same as that of the Pectobacterium system (i.e., ferredoxin); however, at the least, the authors should determine if (separate, non-polar) mutations in YddB and PqqL impair the ability of E. coli to grow in low-iron, as this would be indicative of a role in iron import. The fact that PqqL is hyper-expressed in LB containing the iron chelator BP strongly suggests that the proteins are needed under these growth conditions. It might be necessary to mutate yddB and pqqL in a strain that is lacking siderophore in order to clearly / dramatically see a role for YddB and PqqL. These experiments are worthwhile even if they do not reveal a link to growth in low-iron, as this would provide evidence for an import that is rather distinct from the Pectobacterium system. We agree with the reviewer that the specific role that YddB/PqqL plays in iron acquisition is a fascinating question. We have performed the following experiments based on the reviewer's suggestions and included the data in of the current form of the manuscript: These data and included in and lines 164-165 of the revised manuscript: [formula] - [/formula] 'As with FusA, YddB possesses a 22-stranded β-barrel fold, the pore of which is occluded by a globular N-terminal plug-domain. This fold is characteristic of the integral outer membrane TBDT family and like these proteins YddB possesses a hydrophobic transmembrane region .' Secondly, YddB belongs to the TonB-dependent transporter (TBDT) family, which is a well characterised superfamily of outer membrane proteins. YddB's closest characterised homologue the TBDT FusA has been shown to act as a conduit of ferredoxin and related bacteriocins (pectocins) across the outer membrane . The β-barrel architecture of TBDTs, shared by YddB, is characteristic of outer-membrane proteins, with inner membrane proteins being characteristically alpha-helical. As implied by the name the of TBDT family, these transporters depend of energy provided by the TonB-ExbBD complex. This complex resides in the inner membrane and the TonBcomponent spans the periplasm to interact with the TBDT in the outer membrane. This interaction provides energy from the proton motive force to drive substrate import through the TBDT and across the outer membrane. Thus it is highly unlikely that a protein with this architecture would be located anywhere other than the outer membrane. Thirdly, experimental data from a number of studies confirms the outer membrane localisation of YddB. Utilising a proteomics approach, a studies by and , showed YddB to be present in outer membrane fraction from E. coli cells. Moreover, three studies characterising the membrane protein content of E. coli outer membrane vesicles identified YddB as present in OMVs. These studies provide strong experimental evidence for the outer membrane localisation of YddB. We have included the following text on lines 149-151: 'YddB belongs to the outer membrane localised TBDT family and has been detected in the outer membrane and in outer membrane vesicles of E. coli in a number of studies .' And added references to these articles to provide the reader with a clear picture of the evidence supporting YddB's outer membrane localisation. 3. Lines 171, 182-184. Given that levels of PqqL are increased in low-iron growth conditions, it should be determined and discussed whether the yddB/pqqL operon is iron-regulated and Fur-regulated. Basic qRT-PCR can determine if the genes are iron-regulated, and sequence analysis should be able to identify a putative Fur box. (The fact that the operon was shown by others to be upregulated in urine does not alone make this point.). Following on point 2, immunoblotting could then confirm whether YddB levels are also influenced by iron levels.

Vitamin D Status and Its Determinants in Mexican Pregnant Women from a Rural and an Urban Area: A Comparative Study Citation: Chávez-Courtois, M.; Godínez-Martínez, E.; Muñoz-Manrique, C.; Negrete-Martínez, V.; González-Leyva, C.P.; Tolentino-Dolores, M.; Suárez-Rico, B.; Estrada-Gutierrez, G.; Perichart-Perera, O. Vitamin D Status and Its Determinants in Mexican Pregnant Women from a Rural and an Urban Area: A Comparative Study. Int. J. Environ. Res. Public Health 2021, 18, 4571. https://doi.org/10.3390/ # Introduction Vitamin D is a liposoluble vitamin and hormone that is synthesized in the skin, and it is normally related to bone health. Vitamin D is commonly known as the sun vitamin. Estimations indicate that sunlight exposure contributes to 90% of vitamin D production through the conversion of 7-dehydrocholesterol in the skin to pre-vitamin D, which isomerizes to cholecalciferol (Vitamin D 3 ). Exposure to Ultraviolet B (UVB) radiation (wavelengths from 290 nm to 315 nm) stimulate the conversion process. Thus, factors such as the geographic location (latitude), the season, time of the day, pollution, and even cloudiness influence vitamin D production. Besides the geographical situation, other factors such as skin pigmentation, diet, health status, obesity, clothing, use of sunscreen, and physical activity play a major role in vitamin D synthesis in the skin. The type of mobility of pregnant women in daily life influences sunlight exposure and, thus, the vitamin D concentration. Day-to-day mobility within a space of social coexistence is determined by territorial and socioeconomic characteristics. Food intake barely contributes to the body requirements of vitamin D since only a few foods contain this nutrient and they are not widely consumed. Main sources coming from food in the form of vitamin D 3 are codfish liver oil, salmon, beef liver, egg yolk, cheese, and fortified foods (margarines and some dairy). To a lesser extent, mushrooms contain the D 2 form (ergocalciferol) of the vitamin. In some countries, most of the vitamin D is obtained through nutritional supplements. Previously, the relevance of vitamin D was focused on the contribution to calciumphosphate homeostasis, which promotes healthy growth and reduces the risk of bone fractures. Nonetheless, vitamin D has been recently implicated in cellular and neuromuscular growth, immunological modulation, tissue inflammation reduction, and tumor suppression. Vitamin D and calcium participate in bone mineralization, bone accretion, and growth during the development of the fetus, therefore increasing the requirement of these nutrients during pregnancy. The dietary intake recommendation for vitamin D during pregnancy according to the Institute of Medicine of the National Academies (Food and Nutrition Board) is 600 IU/day (15 mcg/day)The Royal College of Obstetricians and Gynaecologists (RCOG) recommend 400 IU/day (10 mcg/day) for all pregnant women and 1000 IU/day (25 mcg/day) for high risk women. In Mexico, the dietary intake recommendation for pregnant women is 200 IU/day. The major circulating form of vitamin D is 25-hydroxyvitamin D (calcidiol, 25(OH)D), and its measurement is clinically used for the assessment of vitamin D status. Even though the cut-off points for vitamin D deficiency (VD-D) and insufficiency (VD-I) have not been well established, most experts agree that an adequate vitamin D status may be defined as 25(OH)D ≥ 30 ng/mL, and a deficient state is usually considered as 25(OH)D < 20 ng/mL. Globally, it has been reported that 54% of pregnant women have a VD-D. VD-D during pregnancy is associated with a higher risk of preeclampsia, gestational diabetes, preterm birth, low birth weight, among other complications. In Mexico, a cross-sectional study in mother-child binomials showed a prevalence of VD-D in 61% of pregnant women during the third trimester and in 98% of babies. A recent report from a cohort of Mexican healthy pregnant women showed that only 39% of them had adequate 25(OH)D concentrations during the third trimester of gestation, and 20% had a deficient status. The location and socioeconomic factors (activity, type of job, type of transportation) of pregnant women could lead to differences in sunlight exposure. Living in an urban or rural (countryside) area may affect 25(OH)D concentrations. Studies comparing maternal vitamin D status between women living in rural areas and women in urban areas are scarce. Some studies done in Vietnam and in Pakistan have reported higher 25(OH)D concentrations in women from rural areas compared with women from urban areas. In a study comparing pregnant women from an urban zone and rural zone in Mongolia and women in Boston, women living in rural provinces of Mongolia showed higher seasonadjusted 25(OH)D concentrations than their counterparts living in the capital (urban areas). No studies in women from Latin America or of Hispanic origin have evaluated vitamin D status by location or other socioeconomic factors. The aim of the present study was to compare the vitamin D status between women living in rural areas and women living in urban areas, and evaluate the effect of other sociodemographic factors. # Materials and methods This study derives from two different cohorts of pregnant women with singleton pregnancies. The design was a case-control study nested within two cohorts. Two groups were studied: one from an urban area (Mexico City) and another from a rural area (Cuetzalan, Puebla). The inclusion criteria were healthy adult women and singleton pregnancy. The authors declare that all the investigations were conducted according to the guidelines of the Declaration of Helsinki of 1975. ## Women from an urban area-mexico city Women from this group were selected from the prospective cohort OBESO (for its Spanish acronym "Origen Bioquímico y Epigenético del Sobrepeso y la Obesidad") Biochemical and Epigenetic Origin of Overweight and Obesity) carried out in the Instituto Nacional de Perinatología (INPer) Mexico City, Mexico. The study was approved in March 2017 by the INPer Ethics and Research Committee (Registry No. 3300-11402-01-575-17). The women that volunteered to participate signed an informed consent letter. All women were recruited from January 2017 to January 2020 during the first trimester of pregnancy at the Fetal Maternal Medicine Department. Only adult women with a singleton pregnancy were included. Every woman received regular prenatal care in the INPer facilities, without any modification to the usual clinical interventions. Assessments were performed during the second (18 to 22.6 weeks) and third (28 to 34.6 weeks) trimesters in the Nutrition Unit. Gestational age was calculated in every visit according to the fetal ultrasound screening at the first trimester. Height and weight were measured using a steady digital stadiometer (model 264, SECA, Hamburg, Germany) and bioimpedance equipment (model 230, Inbody, Seoul, Korea), respectively, in order to calculate the pregestational body mass index (pBMI). Women were classified as normal weight (pBMI > 18.5 < 24.9), overweight (pBMI ≥ 25), or obese (pBMI ≥ 30), according to the WHO parameters. Demographic information (education, occupation, and socioeconomic status) and clinical data (parity) were obtained. Maternal blood sample acquisition was recorded considering two categories: spring/summer and autumn/winter seasons. The socioeconomic status was classified as high/medium or low, according to the INPer parameters. Women were categorized as nulliparous (with no born child) or multiparous (at least one born child). Any micronutrient supplementation prescribed by an obstetrician-gynecologist or healthcare professional was independent of the present study. Dosage of prescribed vitamin D3 (IU/d) was calculated for each trimester according to the use of vitamin D supplements or any other multivitamin that could provide vitamin D. A previous analysis of this cohort showed that 89% of women received vitamin D supplementation throughout pregnancy or at some point in pregnancy; 50% of them received 500 IU/day. ## Women from a rural area-cuetzalan, puebla, mexico Indigenous pregnant women from this group derive from the study: "Aplicación y evaluación del modelo sociocultural para prevenir muertes maternas, en Cuetzalan del Progreso, Puebla", which was approved by the INPer Ethics and Research Committee in August 2017 (Registry No. 2017-1-55). All women belong to the Cuetzalan del Progreso municipality located in the Northern mountain range of Puebla, Mexico. The women that arrived between March and June 2018 at a rural clinic or medical unit in six out of the eight Town Councils that constitute the municipality were invited to participate in the latter study. The women that volunteered to participate received a verbal and written description of the study details and signed an informed consent letter afterward. For the data acquisition of the indigenous women, there was an approach to each clinic at the Town Councils and to the Cuetzalan General Hospital. A socioeconomic survey was applied to each woman. This instrument has been used before in the same population. Height and weight were measured using a digital floor scale (model 803, SECA, Hamburg, Germany) and a portable stadiometer (model 213, SECA, Hamburg, Germany), respectively. Pregestational weight was inquired, and the pBMI was calculated; pBMI classification was performed according to the WHO parameters (same procedure previously described). All pregnant women were supposed to receive a multivitamin with 200 IU/d of vitamin D as part of their prenatal care. Individual supplemented vitamin D doses were not recorded. A fasting blood sample was obtained from each woman and later centrifuged to obtain the serum. Serum samples were properly labeled and frozen at −70 - C in the Cuetzalan General Hospital. Transfer of the samples to the Institute was performed in controlled-temperature containers. For this study, adolescent women were excluded. All adult women that participated in the Cuetzalan study were included. ## Subset selection To avoid skewing caused by confounding factors and to achieve more homologous groups, pairs were selected from the urban study to match all women in the rural study. The following criteria were used to select women from the urban group: (1) Age: women with a maximum age difference of two years were selected from each group; (2) pBMI diagnosis: women in the same pBMI category (normal, overweight, or obese) were selected from each group, considering a maximum difference of two units in the raw pBMI value; (3) Trimester: Women within the same period assessment, whether second or third trimester, were selected from each group. Pairs were selected in a consecutive manner, using the number from the study ID. ## Vitamin d (25-oh-d) analysis Sample analysis was carried out in the Nutrition and Bioprogramming Department of the Institute. Serum concentrations of 25(OH)D were analyzed via ELISA (chemiluminescence) (Architect, Abbott, Longford, Ireland). The 25(OH)D calibration curve should be run in duplicate of 6 points. Range of calibration values: 0.0 ng/mL-160.0 ng/mL using a 4 parameter logistic curve fit data calculation method to generate the calibration curve. A single replicate of each of the different concentration controls must be run to assess the assay calibration. An acceptable coefficient of variation was considered as <5%. An insufficient status was considered when serum concentrations were <30 ng/mL, and a deficient status was considered when concentrations were <20 ng/mL. # Statistical analysis The 25(OH)D concentrations were assessed using graphics and simple association patterns with variables of interest. These association patterns were evaluated using a Student t-test, chi-square, linear regression models, and multinomial, depending on the response variable. According to the simple association patterns, some variables were collapsed into secondary categories. For the marital status variable, the categories were married, free union, and single or divorced. Categories of the occupation variable were a full-time job, partial job, student, and housewife. We defined a full-time job and partial job as following: (1) a full-time job as a paid job in agreement with the workday journey within a fixed schedule and a specific space (secretary, warehouse assistant, nurse, teacher, among others.); (2) partial job as any activity related to an informal job (street food sale, handcraft sale). Finally, for the education variable, the categories were: no education or primary (elementary and middle school), incomplete secondary (truncated high school), complete secondary (concluded high school), and higher education (Bachelor's degree). The date of the last menstrual period (LMP) was recorded for all women to assess the amount of time of sunlight exposure before pregnancy and at an early stage of pregnancy. This was established due to significant differences in the serum concentration of vitamin D according to the season of the year in which the blood samples were acquired. Based on the latter, the LMP reported by the pregnant women was considered as the date of the beginning of pregnancy and therefore classified as the following: winter season, from 21 December to 19 March; spring season, from 20 March to 19 June; summer season, from 20 June to 21 September; and autumn season, from 22 September to 20 December. Using multinomial logistic regression models, the sociodemographic characteristics of pregnant women with a higher probability of being situated in the categories of VD-D or VD-I were evaluated and compared to the probability of having vitamin D sufficiency. Both the variables with a significance level p < 0.01 in the simple association patterns, as well as the variables with solid evidence in the literature related to 25(OH)D concentration, were included in the model. Statistical analyses were carried out using the statistical program STATA (v. 12). # Results The final sample of the rural study was 149 adult pregnant women, so 149 pairs from the urban study were selected. A total of 298 women were studied. Sociodemographic characteristics of the participants are shown and arranged by location in. Women from the rural area were younger and had a lesser level of education compared to the urban counterpart, and a higher proportion had a partial job (p < 0.01). Moreover, significant differences were found in the reported season of LMP, where very few women in the rural group had their LMP in spring (p < 0.01). There were no differences in the pBMI nor in parity by location (p > 0.05). An overall 25(OH)D blood measurement was recorded in 23 (7.8%) women in the first trimester, in 136 (45.6%) women in the second trimester, and in 139 (46.6%) women in the third trimester. The mean 25(OH)D concentration during pregnancy was 26.7 ± 9.4 ng/mL, and this fluctuated according to certain characteristics. These measurements had no difference according to the location, but a trend of higher 25(OH)D concentrations was found in women from the rural area compared to women in the urban area (27.5 ng/mL vs. 25.8 ng/mL, p = 0.13). There was a higher 25(OH)D concentration in the third trimester compared to the first one (p = 0.02). Women whose LMP occurred during spring showed lower concentrations compared to the women whose LMP occurred during winter (p < 0.01). When the women were categorized by marital status, those who lived in a domestic partnership had higher 25(OH)D concentrations compared to married women (p < 0.01). Neither pBMI nor vitamin D supplementation during pregnancy had an impact on concentrations (p > 0.05). A total of 102 (34.2%) women showed adequate 25(OH)D concentrations (≥30 ng/mL). In contrast, 196 (65.8%) women showed VD-D or VD-I. A trend for higher frequency of VD-D was observed the urban group (n = 105) compared with the rural group (n = 91) (X 2 : 2.92, p = 0.08). Adequate status was observed in 38.8% of women from the rural area and in 29.5% of women in the urban area. All women in the rural study were prescribed a multivitamin with 200 IU/d. In the urban study, 89% of women had a supplement recommendation during pregnancy; the median dose prescribed in this study was 500 IU/d.shows the prevalence of VD-D and VD-I according to clinical and sociodemographic factors. As shown in, having a full-time job was associated with a trend of a lower probability of presenting VD-D compared to being a housewife (OR = 0.26; 95%CI = 0.06-1.16; p = 0.08). This same effect was observed between the third and first trimesters of pregnancy (OR = 0.21; 95%CI = 0.04-1.11; p = 0.07). In the case of women whose LMP occurred during spring (OR= 5.7; 95%CI = 1.53-21.07; p < 0.001) and during summer (OR = 3.3; 95%CI = 1.01-10.84; p = 0.04), there was a five-and two-fold increase, respectively, in the probability of having VD-D compared to women whose LMP occurred during winter. Factors associated with a reduced probability of VD-I included having a partial job compared to being a housewife (OR = 0.24; 95%CI = 0.06-0.99; p = 0.05) and being in the third trimester of pregnancy compared to being in the first one (OR = 0.22; 95%CI = 0.05-1.01; p = 0.05). An overall 25(OH)D blood measurement was recorded in 23 (7.8%) women in the first trimester, in 136 (45.6%) women in the second trimester, and in 139 (46.6%) women in the third trimester. The mean 25(OH)D concentration during pregnancy was 26.7 ± 9.4 ng/mL, and this fluctuated according to certain characteristics. These measurements had no difference according to the location, but a trend of higher 25(OH)D concentrations was found in women from the rural area compared to women in the urban area (27.5 ng/mL vs. 25.8 ng/mL, p = 0.13). There was a higher 25(OH)D concentration in the third trimester compared to the first one (p = 0.02). Women whose LMP occurred during spring showed lower concentrations compared to the women whose LMP occurred during winter (p < 0.01). When the women were categorized by marital status, those who lived in a domestic partnership had higher 25(OH)D concentrations compared to married women (p < 0.01). Neither pBMI nor vitamin D supplementation during pregnancy had an impact on concentrations (p > 0.05). A total of 102 (34.2%) women showed adequate 25(OH)D concentrations (≥30 ng/mL). In contrast, 196 (65.8%) women showed VD-D or VD-I. A trend for higher frequency of VD-D was observed the urban group (n = 105) compared with the rural group (n = 91) (X 2 : 2.92, p = 0.08). Adequate status was observed in 38.8% of women from the rural area and in 29.5% of women in the urban area. All women in the rural study were prescribed a multivitamin with 200 IU/d. In the urban study, 89% of women had a supplement recommendation during pregnancy; the median dose prescribed in this study was 500 IU/d.shows the prevalence of VD-D and VD-I according to clinical and sociodemographic factors. # Discussion The present study demonstrates that a high prevalence of vitamin D deficiency (25(OH)D < 20 ng/mL) and insufficiency (25(OH)D < 30 ng/mL) was found in Mexican pregnant women from both urban (VD-D: 37.6%) and rural areas (VD-D: 29.5%). VD-D during pregnancy represents a worldwide health issue, which has been widely reported. Studies in developing countries manifest a prevalence of VD-D ranging from 51.3% to 100%. In women from the American continent, a prevalence between 42% and 72% of vitamin D during pregnancy has been reported. A cross-sectional study carried out in Mexico showed that 61% of pregnant women had VD-D during the third trimester, a higher proportion than reported in our study. A previous report from the OBESO cohort showed that in pregnant women receiving prenatal care, the prevalence of VD-D in the first trimester was 37%, decreasing to 20% in the third trimester, mainly related to increased supplementation during pregnancy. So, even though women from both areas were exposed to supplementation, probably the doses were not enough to achieve an adequate vitamin D status. Not even the higher doses than women from the urban group took (50% of them received 500 IU/d) were sufficient to prevent a deficient status. This is very relevant considering that VD-D has been associated with a significantly higher risk of developing GDM, preeclampsia, preterm birth, and delivering a small-forgestational-age newborn. According to our study, it is possible that women in urban areas have higher risks of developing adverse perinatal outcomes. On the other hand, dietary intake recommendations in Mexico and in other countries appear to be very low (200 IU/d), reducing the probability of pregnant women receiving higher doses of vitamin D supplementation. It is worth noting that even though there were no significant differences in vitamin D concentration between the groups of urban and pregnant women in the rural area, we found a tendency to a higher concentration of vitamin D in the countryside women compared to the urban counterpart. A likely explanation could be the type of daily mobility linked to the occupation of each group of women according to the location. Interestingly, a "sort of protection" against VD-D and VD-I was observed in women that have a partial job, meaning that most of the activities are carried out in the street (food or handcraft sale). Countryside women showed a significantly greater proportion of having partial jobs, and they were probably prone to higher sunlight exposure as a consequence. Partial jobs, as a means of daily mobility, are not limited to the movement within an area, but it forms as such to allow the development of a lifestyle and day-to-day solutions that are culturally and socially codified. In this matter, daily mobility for urban women is usually done through public transportation, and the most common activities are going to work, carrying children to and from school, and attending to the places of medical service. In pregnant women from rural areas, daily mobility involves having to walk to either sell their goods, to wash clothes riverside, to deliver food at the husband's workspace, or to go to the places for medical service, and as such, this group may be more exposed to sunlight. In both groups of women, daily mobility entails a certain degree of sunlight exposure, although it is also influenced by geoclimatic conditions. The latitude of the locations where the studies were carried out is similar; for Mexico City, it is 19 - 25 N, and for Cuetzalan it is 20 - 06 N. Nonetheless, Cuetzalan has a hot semi-arid climate (18-26 - C), which implies that this group of countryside pregnant women are more exposed to sunlight in comparison to women located in Mexico City, which has a temperate climate (16 - C). It is plausible to declare that the daily mobility type is crucial to understand the differences in vitamin D concentration in pregnant women that live in distinct zones of the same territory. According to the previous, three components have to be considered: day-to-day activities, space where these are performed, and the location and geoclimatic conditions. Women living in rural areas appear to have higher sunlight exposure and higher 25(OH)D concentrations, which may result in a lower risk for many adverse perinatal outcomes. There are no previous studies in Mexico that compare vitamin D status in pregnant women living in urban or rural zones, and there are only a few studies globally. In countries such as Vietnam, Pakistan, and Mongolia, higher 25(OH)D concentrations have been reported in pregnant women living in rural areas. In a study comparing women from three different areas of Mongolia (one urban, two rural) and women from Boston, significantly lower 25(OH)D concentrations were observed in Mongolian women. However, higher concentrations were observed in women from the two rural provinces of Mongolia (15.2 and 15.3 ng/mL) compared to women in the capital (urban) (13.2 ng/mL). In another study done on non-pregnant and pregnant women in Vietnam, those living in a rural province had significantly higher 25(OH)D concentrations than women from an urban city (34 ng/mL vs. 31.2 ng/mL, respectively). Another important factor in Mexican women is the high prevalence of deficient intake of vitamin D. According to data from the National Nutrition and Health Survey in Mexico, 97% of non-pregnant women living in urban areas and 97% of those living in rural regions do not meet the dietary intake recommendation for vitamin D. In the present study, as in many others, greater concentrations of vitamin D were observed during the third trimester compared to the first trimester. Longitudinal studies have proven that the vitamin D concentration increases during pregnancy progression. There is also the possibility that women received more vitamin D supplementation later in pregnancy, considering that many women may start prenatal care very late. Likewise, there were differences according to the season in our study. All women whose LMP occurred in spring showed the lowest vitamin D concentrations and the highest prevalence of VD-D and VD-I. This could mean that these women were in their second or third trimester of pregnancy during autumn and winter. It has been widely reported that these two seasons can reduce vitamin D concentrations due to an association with sunlight scarcity in some countries. Finally, in a previous report of the OBESO cohort, in Mexican women, the 25(OH)D concentrations decreased by −1.85 ng/mL (95% CI: −2.99 to −0.72 ng/mL) in women during autumn/winter, compared to the spring/summer season. One of the main strengths of this study was that we included two cultural and socially distinct populations, considering several sociodemographic variables. We were able to compare two paired groups reducing skewing and confounding factors in the associative models. We identify within the limitations that serum samples of women in the rural group were retrieved solely during spring. This forced us to analyze the seasonal effect using the LMP as a reference. An important limitation was the lack of data about vitamin D intake, including food sources and supplement doses. The cross-sectional design is another weakness. # Conclusions The prevalence of vitamin D deficiency and insufficiency is high in Mexican pregnant women from both urban and rural areas, even though both groups were exposed to supplementation. A trend towards higher 25(OH)D concentrations was observed in women living in a rural area compared to women living in an urban area. Daily factors related to mobility, such as the type of job and the place where it is performed, can affect maternal vitamin D status. Perinatal health services should assess risk factors for vitamin D deficiency in pregnant women, including lifestyle and socioeconomic factors. Informed Consent Statement: Informed consent was obtained from all subjects involved in the study. # Data availability statement: The data presented in this study are available on request from the corresponding author. The data are not publicly available because this study derives from two larger studies that are ongoing. ## Conflicts of interest: OPP is a consultant of the Danone Institute and the Nestle Nutrition Institute in Mexico. No conflict of interest exists regarding this study and this manuscript.

Coexisting Coronary and Carotid Artery Disease – Which Technique and in Which Order? Case Report and Review of Literature # Introduction Coexisting carotid artery stenosis and coronary artery disease (CAD) is frequent, with prevalence of significant carotid lesions in patients undergoing coronary artery bypass graft (CABG) reported as high as 8% to 14%. [bib_ref] Coronary bypass and carotid endarterectomy: does a combined approach increase risk? A..., Borger [/bib_ref] In patients undergoing carotid endarterectomy (CEA), the prevalence of CAD is 40% to 50%. [bib_ref] Simultaneous hybrid revascularization by carotid stenting and coronary artery bypass grafting: the..., Versaci [/bib_ref] There is a risk of stroke in patients undergoing CABG (1.4%-3.8%) [bib_ref] Perioperative stroke, Selim [/bib_ref] and a risk of myocardial infarction in patients undergoing CEA (0%-2%). [bib_ref] Periprocedural myocardial infarction after carotid endarterectomy and stenting: systematic review and meta-analysis, Boulanger [/bib_ref] There is currently no consensus in treatment guidelines on the sequence of revascularization. Treatment strategies include (1) Combined or synchronous surgery, where CABG and CEA are performed in the same procedure or anesthetic setting; [bib_ref] Simultaneous hybrid revascularization by carotid stenting and coronary artery bypass grafting: the..., Versaci [/bib_ref] Staged surgeries, consisting of CABG with subsequent CEA or CEA with subsequent CABG; (3) Hybrid procedures, which can be synchronous or staged, and consist of CABG with carotid artery stenting (CAS) or PCI with CEA and (4) Transcatheter procedures combining PCI with carotid artery stenting. The etiology of stroke which occurs with CABG is multifactorial and includes the following: hypoperfusion due to hypotension in the presence of a severely stenotic carotid artery, micro-embolization from an ulcerated carotid plaque, as well as macro-embolization from ascending aorta atherosclerosis. In addition, many risk factors for stroke co-exist in CABG patients. 5 ## Case report A 55-year-old woman with insulin-dependent diabetes mellitus, hypertension, hyperlipidemia, peripheral vascular disease and severe triple vessel CAD status post 3-vessel stenting presented with anterior NSTEMI. Her EKG showed new T-wave inversions in V1-V4. A transthoracic echocardiogram demonstrated normal ejection fraction. There were no neurologic symptoms. Coronary angiography demonstrated triple vessel disease, with severe LAD stenosis (shown in [fig_ref] Figure 1: Coronary angiography during unstable angina demonstrated triple vessel disease with multiple prior... [/fig_ref] as well as high grade LCX and RCA lesions. In preparation for surgery, screening carotid doppler sonography showed severe stenosis of the right internal carotid artery (RICA). Carotid angiography showed 90% stenosis of the RICA (shown in [fig_ref] Figure 2: Carotid angiography prior to CABG showing the right internal carotid artery with... [/fig_ref] and 50% stenosis of the left internal carotid artery (LICA). It was decided to proceed with single CABG utilizing LIMA to the LAD and to defer stenting of the other coronary lesions. The plan was to treat the carotid disease after myocardial revascularization and resolution of active myocardial ischemia. Ten days after CABG, the patient again presented with recurrent severe unstable angina and repeat coronary angiogram showed a kink and insertion site stenosis in the LIMA which had been inserted proximal to the LAD lesion (shown in [fig_ref] Figure 3: Coronary angiography after LIMA-LAD coronary artery bypass surgery [/fig_ref]. PCI of the LIMA/LAD was performed with excellent results and complete relief of symptoms (shown in [fig_ref] Figure 4: Coronary angiography after LIMA-LAD coronary artery bypass surgery [/fig_ref]. Attention was then shifted to the carotid lesion, and a stent was placed in the RICA with excellent angiographic results (shown in [fig_ref] Figure 5: Treatment of internal carotid stenosis after resolution of angina symptoms with self-expanding... [/fig_ref]. She experienced hypotension for approximately 12 hours as a result of baroreceptor response requiring low dose vasopressors. There was no myocardial ischemia during this period. ## 2 ## Clinical medicine insights: cardiology and "coronary artery bypass graft" and combinations of the above. Literature was reviewed and selected based on several factors such as number of patients involved, journal published, and ability to answer clinical question. ## Screening for carotid artery stenosis in patients undergoing cabg Who should be screened? The ACC/AHA Guidelines for CABG (2011) [bib_ref] ACCF/AHA guideline for coronary artery bypass graft surgery: a report of the..., Hillis [/bib_ref] state that carotid artery duplex scanning is reasonable in selected patients who are considered to have high risk features (e.g., age >65 years, presence of a carotid bruit, left main coronary artery stenosis, peripheral artery disease, hypertension, smoking, diabetes mellitus, history of cerebrovascular disease (transient ischemic attack, stroke), history of cervical carotid disease) (Class IIa, Level of Evidence: C). These recommendations are based on a few observational studies. ## Options for screening for carotid artery disease The initial screening test is carotid duplex sonography which is widely available and has excellent sensitivity (94%) and specificity (92%) to detect carotid artery stenosis at 60% to 99%. 7 ## 3 When duplex sonography cannot be obtained, is equivocal or when additional anatomical information is needed, for example, for when an intervention is planned, CT-Angiography (CTA) or MR-Angiography (MRA) are other non-invasive diagnostic test options. Grading of the stenosis is most often based on the North American Symptomatic Carotid Endarterectomy Trial (NASCET) criteria. However, MRA often overestimates the degree of stenosis and assessment of calcified lesions with CTA is limited. Both CTA and MRA present difficulties in distinguishing subtotal and complete arterial occlusion. [bib_ref] SVS guideline on the management of patients with extracranial carotid and vertebral..., Brott [/bib_ref] When non-invasive tests are inconclusive and additional anatomic detail is needed, catheter-based contrast angiography can be considered. In practice, duplex sonography and CTA or MRA are most often sufficient for adequate estimation of stenosis, and catheter-based angiography is used with the option of therapeutic intervention in the form of carotid artery stenting. 9 ## Is there a benefit in screening? Multiple clinical trials comparing CEA with medical therapy in patients with atherosclerotic stenosis of extracranial carotid arteries have favored carotid revascularization (namely NASCET 10 and ECST 11 in symptomatic carotid disease and ACST 12 and ACAS 13 in asymptomatic carotid disease). A systematic review by Naylor et al [bib_ref] Carotid artery disease and stroke during coronary artery bypass: a critical review..., Naylor [/bib_ref] found an association between the degree of carotid stenosis and stroke risk in CABG patients, ranging from >2% in patients with no significant carotid disease up to 11% in patients with carotid occlusion (see [fig_ref] Table 1: Association between degree of carotid stenosis and stroke risk in patients undergoing... [/fig_ref]. Interestingly, in Naylor's review, 14 50% of stroke sufferers did not have significant carotid disease and in 60% the anatomic territory on CT scan or autopsy did not correspond to the region supplied by the affected carotid stenosis. ## Who should undergo carotid revascularization? Several factors have been shown to favor combined carotid and coronary revascularization, including (but not limited to) severe carotid artery disease, unfavorable morphological characteristics of the carotid lesion (eg, ulceration), presence of related symptoms and a history of TIA or stroke. [bib_ref] ACCF/AHA guideline for coronary artery bypass graft surgery: a report of the..., Hillis [/bib_ref] Patients with a history of TIA or stroke and severe carotid artery stenosis have a higher risk of post-CABG stroke and will likely benefit from revascularization. [bib_ref] Carotid endarterectomy for carotid stenosis in patients selected for coronary artery bypass..., Hejri [/bib_ref] The ACC guidelines comment that timing and sequence of carotid and coronary revascularization is based on the absence or presence of clinical symptoms. [bib_ref] ACCF/AHA guideline for coronary artery bypass graft surgery: a report of the..., Hillis [/bib_ref] On the other hand, CABG alone can be performed safely in patients with asymptomatic unilateral carotid stenosis as revascularization offers no significant reduction in risk of stroke or death in these individuals. [bib_ref] ACCF/AHA guideline for coronary artery bypass graft surgery: a report of the..., Hillis [/bib_ref] Carotid endarterectomy versus carotid artery stenting CEA remains the standard of care for most patients with severe extracranial carotid disease. There are multiple clinical trials comparing carotid artery stenting and carotid endarterectomy with short-term and long-term follow up results, for example, the SAPPHIRE trial from 2004, [bib_ref] Protected carotid-artery stenting versus endarterectomy in high-risk patients, Yadav [/bib_ref] the European trials EVA-3S in 2008 and 2014 and most recently the CREST trial in 2010 and 2016. [bib_ref] Stenting versus endarterectomy for treatment of carotid-artery stenosis, Brott [/bib_ref] [bib_ref] Long-term results of stenting versus endarterectomy for carotid-artery stenosis, Brott [/bib_ref] A systematic review of short term results of these trials showed an increased risk of stroke with CAS compared to CEA (RR 1.45; 95% CI 1.06-1.99; I 2 = 40%) but a decreased risk of periprocedural MI (RR, 0.43; 95% CI, 0.26-0.71; I 2 = 0%). [bib_ref] A systematic review and meta-analysis of randomized trials of carotid endarterectomy vs..., Murad [/bib_ref] Despite the higher risk of periprocedural stroke, carotid artery stenting has been accepted as a viable alternative to carotid endarterectomy in high risk patients and has gained popularity in recent years after CREST publication. A study by showed an increase of the utilization of CAS in patients older than 70 years from 11.9% in the pre-CREST to 13.8% in the post-CREST era (P = .005). However, there are no randomized controlled trials comparing CAS and CEA in patients who require coronary revascularization. [fig_ref] Table 2: Clinical and anatomic features that determine a high risk for complications from... [/fig_ref] show high risk features for CEA and CAS as well as advantages and disadvantages of CAS compared to CEA. Carotid endarterectomy and CABG can be performed simultaneously or in a staged approach (CEA then CABG, CABG then CEA). In general, simultaneous or synchronous procedures are reserved for patients with both acute coronary syndrome and cerebrovascular symptoms. [bib_ref] ACCF/AHA guideline for coronary artery bypass graft surgery: a report of the..., Hillis [/bib_ref] There are no randomized controlled trials that determine whether a simultaneous or staged approach is favorable, but there have been multiple systematic reviews and meta-analyses over the years with comparable outcomes. Most recently, Tzoumas et al 27 ascertained a significantly higher risk for stroke (OR 1.51, 95% CI 1.34-1.71, I 2 = 0%) and death (OR 1.33, 95% CI 1.01-1.75, I 2 = 47.8%) but a lower risk of MI (OR 0.15, 95% CI 0.04-0.61, I 2 = 0%) with the simultaneous compared to staged approach. An earlier review [bib_ref] A systematic review of outcomes following staged and synchronous carotid endarterectomy and..., Naylor [/bib_ref] showed similar results, which suggested that simultaneous procedures had an overall higher risk for death and any stroke (8.7%, 95% CI 7.7-9.8) compared to staged CEA and CABG (6.1%, 95% CI 2.4-9.2), but none of the comparisons reached statistical significance. However, when the risk of periprocedural MI was subsequently included in the analysis, the risk of overall complications was higher with the simultaneous approach compared to the staged approach (simultaneous = 11.5% [95% CI 10.1-12.9], staged CEA + CABG = 10.2% [95% CI 7.4-13.1]). This meta-analysis also analyzed outcomes for reverse staged CABG + CEA, meaning performing CABG first and then CEA, and found that there is a higher risk of stroke compared to simultaneous CABG + CEA or staged CEA then CABG. ## Cas and cabg Combining carotid artery stenting and CABG has gained popularity. These procedures can also be done simultaneously or staged. When done simultaneously, the operating room needs to be equipped with an angiography system and an experienced proceduralist should perform the procedure. When CAS is performed prior to CABG, there is need for initiation of dual antiplatelet therapy (DAPT), which should be continued for a minimum of 30 days (preferably 3 months) after CAS. [bib_ref] SVS guideline on the management of patients with extracranial carotid and vertebral..., Brott [/bib_ref] In that case, CABG needs to be deferred for 4 to 5 weeks, which is problematic in patients with active myocardial ischemia. No randomized controlled trials have compared simultaneous CAS + CABG to staged CAS then CABG. A review and meta-analysis by Paraskevas et al. [bib_ref] Carotid stenting prior to coronary bypass surgery: an updated systematic review and..., Paraskevas [/bib_ref] of 2727 patients, of whom 80% were neurologically asymptomatic, has shown similar mortality of staged CAS + CABG (4.8%, CI [3.3-6.8]) and simultaneous (meaning same-day) CAS + CABG (4.5% The higher risk of MI in staged CAS and CABG can be atttributed to reflex hypotension and bradycardia after CAS, which may worsen myocardial perfusion. Therefore, in patients which myocardial ischemia, CABG may be performed prior to CAS or same-day procedures can be considered. The authors noted that the study groups were largely heterogenous regarding the risk profile of patients, the length of dual antiplatelet therapy (DAPT), the use of cerebral protection devices and the use of off-pump bypass. The authors found significantly lower rates of in-hospital death in CAS + CABG (1.9% vs 4.4% in combined CEA + CABG and 3.8% in staged CEA + CABG, P < .01), but no statistically significant difference for stroke rates between all three groups. ## Comparison of cea and cas in cabg They also determined that CAS was generally performed in a higher risk cohort: in older patients, in patients with symptomatic carotid disease and in patients with a higher number of cardiovascular comorbidities such as hypertension, diabetes and chronic renal failure. To better analyze the outcomes over time the authors divided the study group into two time periods . The greatest improvement in outcomes over time was seen in the CAS + CABG group, the unadjusted rates of death (3.5% vs 1.0%; P = .02), stroke (5.2% vs 1.7%; P < .01) and death or stroke (7.0% vs 2.7%; P < .01) were significantly lower in later years (2009-2012) than in earlier years . The improved outcomes in CAS + CABG over time can likely be explained by the more widespread use of embolic protection devices and more experienced interventionalists. In addition, the overall number of procedures for CEA was decreasing over the time of the investigation, while the procedural rates for CAS remained fairly stable. [bib_ref] Comparison of trends and inhospital outcomes of concurrent carotid artery revascularization and..., Feldman [/bib_ref] ## Cea and pci As previously mentioned, there is a high risk of MI in patients undergoing CEA before CABG (10.2% in patients undergoing CEA then CABG according to . In a trial conducted by Illuminati et al, [bib_ref] Long-term results of a randomized controlled trial analyzing the role of systematic..., Illuminati [/bib_ref] 426 patients without history or symptoms of CAD as well as normal EKG and echocardiogram, were randomized to undergo coronary angiography before CEA. There was a significantly reduced risk of MI in those with coronary angiography and intervention after the mean follow up of 3.5 years (HR.078; 95% CI, 0.024-0.256; P < .001) as well as a higher survival in the intervention group (95.6 ± 3.2% in the intervention group vs. 89.7 ± 3.7% in the non-intervention group, log-rank = 6.35, P = .01). Mortality specifically related to MI was also significantly lower in the intervention group (n = 0/216) compared to the non-intervention group (n = 6/210, P = .01). The authors concluded that in patients with asymptomatic coronary artery disease, systematic coronary angiography followed by selective PCI or CABG prior to CEA significantly reduces the incidence of late MI and increases long-term survival. However, the need for DAPT after PCI delays carotid intervention with CEA. ## Cas and pci The risk of periprocedural stroke during PCI is low at 0.3% to 0.38%, but associated with increased mortality. [bib_ref] Characteristics of cerebrovascular accidents after percutaneous coronary interventions, Dukkipati [/bib_ref] [bib_ref] Stroke complicating percutaneous coronary interventions: incidence, predictors, and prognostic implications, Fuchs [/bib_ref] There is no data that suggest an increased risk of stroke in patients with concurrent carotid and coronary artery disease undergoing PCI. Hence, screening for carotid artery disease in patients undergoing PCI is not recommended. A prospective, multicenter, non-randomized study conducted in 239 patients with concomitant carotid and coronary artery disease treated with staged or simultaneous carotid artery stenting and percutaneous coronary intervention showed outcomes similar to hybrid or surgical procedures. The reported rate of death, myocardial infarction, and stroke at long-term follow-up (median 520 days) was 4.2%, 2.1%, and 3.8%, respectively. [bib_ref] Early and long-term outcomes after combined percutaneous revascularization in patients with carotid..., Tomai [/bib_ref] This goes along with other studies [bib_ref] Mid-term outcome after carotid artery stenting depends on presence of coronary artery..., Hofmann [/bib_ref] [bib_ref] Safety and efficacy of elective carotid artery stenting in high-risk patients, Shawl [/bib_ref] which concluded that simultaneous and staged CAS and PCI can be performed safely in patients with known carotid and coronary artery disease, with the timing and sequence of the procedures determined by the patient's anatomy and symptoms. The combination of PCI and CAS is particularly important in patients with high surgical risk. [bib_ref] Safety and efficacy of elective carotid artery stenting in high-risk patients, Shawl [/bib_ref] In contrast to hybrid procedures, [bib_ref] Revascularization of carotid stenosis before cardiac surgery, Das [/bib_ref] in patients with recent angiographic intervention requiring dual antiplatelet therapy, both PCI and CAS can be performed safely without interruption of the DAPT. [bib_ref] Early and long-term outcomes after combined percutaneous revascularization in patients with carotid..., Tomai [/bib_ref] Our case demonstrates a successful approach with CAS after PCI. # Conclusion Coexisting carotid and coronary artery disease is common, and the timing and sequence of myocardial and carotid revascularization is controversial. While many meta-analyses and reviews compare the rates of death, stroke and MI in different approaches, attention should be shifted to the individual patient: Which vascular bed is symptomatic? Which approach does the anatomy allow? What are the individual risk factors? [bib_ref] ACCF/AHA guideline for coronary artery bypass graft surgery: a report of the..., Hillis [/bib_ref] ## Clinical medicine insights: cardiology Our patient underwent myocardial revascularization before carotid stenting because she presented with unstable angina and was asymptomatic neurologically. This approach is consistent with recent guidelines, which suggest treating the more symptomatic region first. [bib_ref] ACCF/AHA guideline for coronary artery bypass graft surgery: a report of the..., Hillis [/bib_ref] Advances in technology allow a minimally invasive approach in selected patients with appropriate anatomy. Recently, there is discussion whether screening for carotid disease in patients scheduled for CABG is beneficial. [bib_ref] Does preoperative carotid stenosis screening reduce perioperative stroke in patients undergoing coronary..., Masabni [/bib_ref] [bib_ref] Managing patients with symptomatic coronary and carotid artery disease, Naylor [/bib_ref] About 90% of carotid revascularizations are done in neurologically asymptomatic patients and the risk of stroke in neurologically asymptomatic patients with a unilateral 70% to 99% stenosis undergoing an isolated CABG is extremely low. In addition, combining carotid and coronary revascularization overall increases morbidity and mortality, with a composite risk of death, stroke, or myocardial infarction within 30 days of combined CABG and carotid revascularization of 9% to 11% according to a review by Naylor. [bib_ref] Managing patients with symptomatic coronary and carotid artery disease, Naylor [/bib_ref] The development of hypotension following carotid artery stenting may result in worsening myocardial ischemia in patients with active coronary lesions. The literature used in this review consists of meta-analyses and reviews of retrospective studies or small case studies. There are several limitations, including very heterogenous study populations, whether patients were symptomatic or not, the timing of staged procedures ranging from days to several months and the use of DAPT. Furthermore, it is unclear in most of the retrospective trials what the exact morphology of the lesions were. Randomized controlled trials are needed to further identify the ideal revascularization strategy in different patient groups. Our case demonstrates successful approach by re-vascularizing the more symptomatic vascular bed first. [fig] Figure 2: Carotid angiography prior to CABG showing the right internal carotid artery with severe stenosis at the ostium (arrow). [/fig] [fig] Figure 3: Coronary angiography after LIMA-LAD coronary artery bypass surgery. There is a severe kink in the LIMA anastomosis which was inserted proximal to the LAD lesion (arrow). This resulted in recurrent severe unstable angina. [/fig] [fig] Figure 4: Coronary angiography after LIMA-LAD coronary artery bypass surgery. This shows the result after percutaneous coronary intervention of LIMA-LAD. A drug eluting stent spans the kink and the lesion. [/fig] [fig] Figure 1: Coronary angiography during unstable angina demonstrated triple vessel disease with multiple prior stents and LAD with diffuse disease and severe stenosis (arrow) which was thought to be the culprit lesion. [/fig] [fig] Figure 5: Treatment of internal carotid stenosis after resolution of angina symptoms with self-expanding stent placement with cerebral protection device. [/fig] [table] Table 1: Association between degree of carotid stenosis and stroke risk in patients undergoing CABG. According to Naylor et al.14 [/table] [table] Table 2: Clinical and anatomic features that determine a high risk for complications from carotid endarterectomy (CEA).25 HIGH RISk fEATURES foR CEA [/table] [table] Table 3: Plaque [/table] [table] Table 4: Advantages [/table]

Strong Selection at MHC in Mexicans since Admixture Mexicans are a recent admixture of Amerindians, Europeans, and Africans. We performed local ancestry analysis of Mexican samples from two genome-wide association studies obtained from dbGaP, and discovered that at the MHC region Mexicans have excessive African ancestral alleles compared to the rest of the genome, which is the hallmark of recent selection for admixed samples. The estimated selection coefficients are 0.05 and 0.07 for two datasets, which put our finding among the strongest known selections observed in humans, namely, lactase selection in northern Europeans and sickle-cell trait in Africans. Using inaccurate Amerindian training samples was a major concern for the credibility of previously reported selection signals in Latinos. Taking advantage of the flexibility of our statistical model, we devised a model fitting technique that can learn Amerindian ancestral haplotype from the admixed samples, which allows us to infer local ancestries for Mexicans using only European and African training samples. The strong selection signal at the MHC remains without Amerindian training samples. Finally, we note that medical history studies suggest such a strong selection at MHC is plausible in Mexicans. ## Author summary Whether or not there exists recent selection since admixture in Latinos has been a subject of debate. To detect selection signal, a method uniquely applicable to recently admixed samples is local ancestry analysis. We infer local ancestry of admixed samples (in our study, Mexicans), and look for regions where the average ancestry of one ancestry component significantly deviates from its genome-wide average. Inferring local ancestry requires training samples that represent the genuine ancestral source populations. One major concern for previously detected selection signals in Latinos via local ancestry analysis is the inaccuracy of Amerindian training samples. This is partly due to large genetic differences among Amerindian tribes and partly due to the difficulty in obtaining Amerindian training samples. We developed a new method which allows us to learn Amerindian ancestral # Introduction In 1492 Columbus discovered America. Europeans, led by the Spaniards, and armed with horses, wheels, germs, and steel, rapidly conquered the New World, and promptly Africans were brought there as slave labor. During the past 500 or so years, three populations-Amerindians, Europeans, and Africans-have occupied the same space and time, albeit asymmetrically, and were genetically admixing. Twenty generations later, the majority of the people inhabiting Central America, Caribbean Islands, and South America, such as Mexicans, Puerto Ricans, and Columbians have become an admixture of the three continental ancestral populations. These recently admixed populations are of great interest for modern genetic studies [bib_ref] Latino Populations: A Unique Opportunity for the Study of Race, Genetics, and..., Burchard [/bib_ref]. In 2007, Tang and colleagues analyzed a small cohort of Puerto Rican samples and reported three regions that are under strong recent selection [bib_ref] Recent Genetic Selection in the Ancestral Admixture of Puerto Ricans, Tang [/bib_ref]. Using their then state-of-the-art local ancestry inference software Saber [bib_ref] Reconstructing Genetic Ancestry Blocks in Admixed Individuals, Tang [/bib_ref] , Tang and colleagues discovered in Puerto Rican samples genomic regions whose mean local ancestries (averaged over individuals) significantly deviated from the genome-wide average-a hallmark of recent selection for admixed samples. Price and colleagues cautioned that the strong selection discovered by Tang and colleagues might be artifacts and they provided three arguments [bib_ref] Long-Range LD Can Confound Genome Scans in Admixed Populations, Price [/bib_ref]. First, Saber only models linkage disequilibrium (LD), the non-independence of genetic markers in a population, between adjacently markers and thus may produce unreliable local ancestry estimates in regions that harbor long-range LD. It was noted that all three loci under selection that Tang and colleagues reported are within the long-range LD regions. Second, the Amerindian training samples used by Tang and colleagues, which are Maya and Pima samples from human genetic diversity panel (HGDP) [bib_ref] Worldwide Human Relationships Inferred from Genome-Wide Patterns of Variation, Li [/bib_ref] , is an inaccurate ancestral population for Puerto Ricans, which might produce artifacts in local ancestry inference. Third, Price and colleagues analyzed a larger sample using their software Ancestry-Mapand did not discover the deviation of local ancestry reported by Tang and colleagues. We would like to make the following comments. First, the AncestryMap uses the so called ancestry informative markers (AIMs) to infer local ancestry; because that AIMs are sparse and that ancestry informative haplotypes may not contain sufficient number of AIMs, the statistical method underlying AncestryMap is evidently under-powered in detecting local ancestry compared to those that attempt to model haplotypes, particularly more recent model-based methods such as HapMix [bib_ref] Sensitive Detection of Chromosomal Segments of Distinct Ancestry in Admixed Populations, Price [/bib_ref] and ELAI [bib_ref] Detecting Structure of Haplotypes and Local Ancestry, Guan [/bib_ref]. Therefore, negative results from AncestryMap cannot convincingly refute positive findings by Tang and colleagues. blackSecond, the long-range LD, if properly modeled, will benefit the local ancestry inference, because in regions that harbor long-range LD there are more markers in sync to define population specific haplotypes. Although Saber [bib_ref] Reconstructing Genetic Ancestry Blocks in Admixed Individuals, Tang [/bib_ref] has difficulty with long-range LD, more recent model-based methods, such as ELAI [bib_ref] Detecting Structure of Haplotypes and Local Ancestry, Guan [/bib_ref] , can benefit from long-range LD. Third, inaccurate Amerindian training samples is a challenge in studying local ancestry of Latinos. Amerinidan training samples are rarely found in the public domain; the ones that are available, such as Maya and Pima samples from HGDP [bib_ref] Worldwide Human Relationships Inferred from Genome-Wide Patterns of Variation, Li [/bib_ref] , have small sample sizes and many samples have non-neglegible European ancestries [bib_ref] Ancestral Components of Admixed Genomes in a Mexican Cohort, Johnson [/bib_ref]. In this study we analyzed two datasets whose subjects are of Mexican descent, which we obtained from the database of genotype and phenotype (dbGaP). Our primary motivation is to follow up with selection findings in an early study [bib_ref] Detecting Structure of Haplotypes and Local Ancestry, Guan [/bib_ref] , which discovered signatures of recent selection in HapMap3Mexican samples based on a departure of local ancestry from the global average. Our second motivation is to report a method that can overcome the technical challenge presented by inaccurate Amerindian training samples when analyzing local ancestry of Latinos. We devised a novel method to infer local ancestry which allows us to discard Amerindian samples and instead learn Amerindian haplotypes from Mexican samples. The strong selection in the MHC region in Mexicans was confirmed in our study. # Results We applied for access and downloaded two GWAS datasets of Mexican descent from the dbGaP. One is the Viva La Familiar obesity-diabetes familial risk study (henceforth Viva) that contains 815 individuals from 261 families [bib_ref] Novel Genetic Loci Identified for the Pathophysiology of Childhood Obesity in the..., Comuzzie [/bib_ref]. The other is the Mexican hypertriglyceridemia study (henceforth Lipid) that contains 2229 unrelated individuals [bib_ref] Genomic study in Mexicans identifies a new locus for triglycerides and refines..., Weissglas-Volkov [/bib_ref]. After stringent QC (see Materials and Methods), we applied ELAI to infer local ancestry of each individual. ELAI outperforms other competing state-of-the-art methods in local ancestry inference [bib_ref] Detecting Structure of Haplotypes and Local Ancestry, Guan [/bib_ref]. It is also convenient to use as it does not require phasing for either training samples or cohort samples, nor does it require recombination map or global admixture proportions as inputs. Public resources, such as HapMap and 1000 Genomes projects, contain high quality European and African haplotypes, but not Amerindian haplotypes; this makes ELAI even more attractive than others in analyzing Mexican samples. ## Patterns of global and local ancestry In VIVA the global ancestry proportions (that is, the admixture proportions) for Amerindian, European, and African components are 0.484,0.452, and 0.064 respectively. In Lipid the numbers are 0.552,0.409, and 0.039. Compared to Viva, Lipid has a higher Amerindian ancestry proportion and lower European and African ancestry proportions. The sampling location is likely to account for the difference: participants in Lipid were recruited in Mexico City, Mexico, whereas participants in Viva were recruited in Houston, Texas. For each ancestry component, there are substantial variations among individuals (see two triangular plots in [fig_ref] Fig 1: Global ancestry proportions and principal components [/fig_ref]. For both datasets, the topological resemblance between the triangular plot and the principal component (PC) plot is remarkable. The relative positions of the Mexican outlier individuals are well matched, and an African American individual accidentally recruited in Viva is rather obvious. This suggests that ELAI estimates are sensible, and that using PC to derive admixture proportions has some merits [bib_ref] Principal Components-Based Assignment of Ancestry along Each Chromosome in Individuals with Admixed..., Brisbin [/bib_ref]. blackIt is believed that using East Asians as additional proxy to Amerindian training samples may improve the local ancestry inference of Latinos, because Amerindians are genetically more similar to East Asians. Our experience suggests, however, that this practice has little impact, and the PC plots, in which Chinese separate from Amerindians inconsistently in two datasets, seem to corroborate our experience. We computed at each marker the average dosages separately for each ancestral component by averaging that component over all individuals. blackThe average ancestry dosages were computed differently for Viva to account for relatedness in the sample (see Materials and Methods). For Viva, the sample standard deviation (ssd) of average dosages for Amerindian, European, and African components are 0.046,0.043, and 0.024 respectively. The largest deviations, measured by the ssd of average dosages for each ancestry, are 5.4,4.8, and 9.9. The locus whose African average dosage is 9.9 ssd above the mean is inside the MHC region, and under the normal approximation, a 9.9 ssd corresponds to a p-value of 2 × 10 −23 , which surpasses any reasonable significant threshold for a genome-wide analysis (in GWAS such a significant threshold is 5 × 10 −8 ). The same region inside MHC was again identified as significant in Lipid; the largest deviation of African average dosages is 14.8 ssd above the mean, which corresponds to a p-value of 3 × 10 −49 . The region identified in MHC is the same region identified by analyzing HapMap3 Mexican samples [bib_ref] Detecting Structure of Haplotypes and Local Ancestry, Guan [/bib_ref]. In that study, a region on chromosome 8 was also identified as border-line significant in Amerindian average dosages. In both Viva and Lipid, however, this region was not replicated. ## Different european and african training samples and their effects on local ancestry inference We used HapMap3 Utah Residents with Northern and Western European Ancestry (CEU) as European training samples; Yoruba in Ibadan, Nigeria (YRI), from west Africa, as African To produce a triangular plot, note that each individual associates a triplet of ancestry proportions (x, y, z) such that x + y + z = 1, and a unique point can be determined such that within an equilateral triangle its distances to three edges are x, y and z. (c) and (d) are PC plots for Viva and Lipid respectively. The PC plots shown are mirror images of the original as indicated by "-" sign in labels. TSI−MKK−MAYA. We also combined all training samples to perform inference (CEU+TSI −YRI+MKK−MAYA). blackThe genome-wide pattern of local ancestry is consistent for different sets of training samples (S1 using TSI as training samples are compensated for by sufficiency of European average dosages. We extracted 8679 SNPs in the extended MHC region, 25-35Mb on chromosome 6, from European and African training samples, and ran ELAI using two upper clusters without specifying the population label, which is essentially haplotype-based structure analysis [bib_ref] Detecting Structure of Haplotypes and Local Ancestry, Guan [/bib_ref]. One admixture component was arbitrarily chosen to make comparison, and the admixture component was averaged over 10 EM runs (after adjusting for label-switching across EM runs). The violin plots in [fig_ref] Fig 3: Comparison between different European and African training samples [/fig_ref] show that TSI is indeed more diverse than CEU at the MHC, MKK is more diverse than YRI, and MKK is the most diverse among four non-admixed populations, which agrees with the theory of east African origin of modern humans [bib_ref] The application of molecular genetic approaches to the study of human evolution, Cavalli-Sforza [/bib_ref]. Recently admixed African Americans (ASW) were included for sanity check of the haplotype-based structure inference. ## Amerindian training samples and how its removal affects local ancestry inference Next we turn to Amerindian training samples. The 1000 Genomes admixture analysis group used a collection of Amerindian samples [bib_ref] A Genomewide Admixture Mapping Panel for Hispanic/Latino Populations, Mao [/bib_ref] different from the Maya and Pima from HGDP that we used, but we had difficulty in obtaining that data. Moreover, a practical concern is that any specific choice of Amerindian training samples will be subject to suspicion of inaccuracy. To test the robustness of our inference against different Amerindian training samples, we elected to remove Amerindian training samples and used only European and African training samples to perform inference-but of course we kept the setting of three ancestral populations. ELAI can function with the absence of one training population as long as there are enough genetic components of that ancestry in the cohort samples. Because Mexicans have a large Amerindian ancestry proportion, when Amerindian training samples are missing, ELAI is still able to learn Amerindian ancestral haplotypes relatively easily from Mexican samples as long as the sample size is large. The same is true for European training samples, but it becomes more difficult if African training samples are missing. To borrow an analogy from next-generation sequencing, a large number of Mexican samples and a high ancestry proportion to local ancestry inference is analogous to a high coverage of sequencing reads to variant call. The recommended practice in an early version of ELAI is to split a large dataset into small subsets. Doing so not only improves computational efficiency on a computer cluster, but also allows ELAI to jointly fit training and cohort datasets. blackIt is evident [bib_ref] Practical Issues in Imputation-Based Association Mapping, Guan [/bib_ref] [bib_ref] Genotype imputation for genome-wide association studies, Marchini [/bib_ref] ] that a cluster model becomes less fit to the training samples in the presence of an overwhelmingly large number of cohort samples, which undermines the performance of local ancestry inference (or imputation). Recall that removing Amerindian training samples requires a large number of cohort samples jointly fitting the model with training samples-we are seemingly in a quandary. The solution is rather simple. In parameter estimation of the two-layer model underlying ELAI [bib_ref] Detecting Structure of Haplotypes and Local Ancestry, Guan [/bib_ref] , we can arbitrarily adjust relative weights between cohort and training samples without changing the expected ancestral allele (haplotype) frequency estimates. In other words, we can take an arbitrarily large number of cohort samples and down weight their contribution to parameter estimation. When the training samples are available, the weighting ensures the model fits to training samples sufficiently; otherwise, the ancestral alleles are estimated exclusively by cohort samples, and the weight cancels out in the parameter estimation as long as we assign equal weight to all cohort samples. black(The technical details can be found in Materials and Methods.) Thus, the weighting allows us to take the extreme measure of removing Amerindian training samples. We implemented the weighting scheme and applied it to both datasets. We combined CEU and TSI as European training samples and YRI and MKK as African training samples. [fig_ref] Fig 4: Comparison of estimations with and without Amerindian training samples [/fig_ref] demonstrates, using both Viva and Lipid datasets, the difference, or lack of it, in the estimated African average dosages with and without Amerindian training samples. Comparing the Amerindian average dosages, however, the estimates without Amerindian training samples are higher than that with. The mean differences are 0.09 for Viva and 0.08 for Lipid. This is not too surprising considering 1) Maya and Pima samples have some European ancestral components (PC plots in [fig_ref] Fig 1: Global ancestry proportions and principal components [/fig_ref] ; and 2) Maya and Pima samples may be imperfect representatives of the Amerindian source populations for Mexicans, and learning Amerindian ancestry components from a large number of cohort samples may provide a better fit. Our results shall eliminate concerns of possible artifacts caused by inaccurate Amerindian training samples. ## Strong selection at the mhc region If purely by chance, it is very unlikely that Amerindians share more alleles with Africans at MHC than the rest of the genome at such a significant level; that the pathogens from the Old world are often lethal to the native inhabitants of the New World seems to argue against such a peculiar sharing. The effect of the population bottleneck and the drift do not distinguish the MHC from the rest of the genome [bib_ref] The Genetic Structure of Admixed Populations, Long [/bib_ref]. If selection happened in Africans before admixture, one would expect to see such selection signals in African Americans, which are not there [bib_ref] Genome-wide Scan of 29,141 African Americans Finds No Evidence of Directional Selection..., Bhatia [/bib_ref]. Therefore, it is safe to assume that the African average dosages in Mexicans rose from the genome-wide mean p 0 , blackwhich is a proxy dosage before selection at MHC, to the inferred [bib_ref] Genome-wide Scan of 29,141 African Americans Finds No Evidence of Directional Selection..., Bhatia [/bib_ref] , which provides a lowerbound estimate of s compared to recursion formula for both dominance and additive models (see Materials and Methods). [fig_ref] Table 1: Estimates of selection coefficient s under different models [/fig_ref] summarizes the estimates of selection coefficient under different models; the lower-bound estimates are s = 0.05 for Viva and s = 0.07 for Lipid. Both estimates indicate a very strong selection, on par with the lactase selection in northern Europeans (0.09-0.19) [bib_ref] Genetic Signatures of Strong Recent Positive Selection at the Lactase Gene, Bersaglieri [/bib_ref] and the sickle-cell trait in Africans (0.05-0.18) [bib_ref] The first arrival time and mean age of a deleterious mutant gene..., Li [/bib_ref]. To understand how many SNPs have contributed to the selection signal in MHC, we assigned a phenotypic value to each individual based on their African ancestry dosage at the identified region in MHC (detailed in Materials and Methods), regressed out six leading principal component and admixture proportions, and performed the single-SNP association test using BIMBAM [bib_ref] Practical Issues in Imputation-Based Association Mapping, Guan [/bib_ref]. At a very liberal threshold of log 10 Bayesfactor > 10, we discovered 1700 SNPs in the extended MHC region to be genome-wide significant (S3 Considering the high correlation among SNPs in the region, we next performed multi-SNP analysis using a Bayesian variable selection regression procedure implemented in the software piMASS [bib_ref] Bayesian variable selection regression for genome-wide association studies, and other large-scale problems, Guan [/bib_ref]. piMASS implements a Markov chain Monte Carlo (MCMC) procedure to sample the posterior distribution of model space (SNP sets) under sparse and shrinkage priors. The output contains posterior probability of association (PPA) for each SNP, which roughly reflects how often the SNP is being selected in an additive model. We ran piMASS using all markers from chromosome 6 of Lipid with 10,000 burn-in steps and 1 million sampling steps. Two independent runs were conducted. In both runs, the proportion of variation explained (the narrow sense heritability) estimates had the same posterior mean of 0.88, with ssd of 0.015 and 0.017 respectively. The posterior mean model sizes (the number of SNPs in the model sampled) were 93±10.7 and 83±7.1 respectively (mean ± ssd). The two runs had 126 and 116 SNPs with PPA >0.1; among them, 60 SNPs overlapped, and the union contained 182 SNPs. We removed these 182 SNPs and reran local ancestry inference of chromosome 6. The pattern of the local ancestry was essentially unaffected. blackThese exercises suggest that the observed selection signal is driven by a large number of SNPs and their constitutional haplotypes. # Discussion In this paper we analyzed two existing GWAS datasets of Mexican subjects and demonstrated that the MHC region is under strong recent selection in Mexicans. Because Viva contains related individuals, we split individuals into non-overlapping subsets, each containing 40-50 unrelated individuals; performed local ancestry inference separately for each subset; and aggregated them to compute the average dosages. This practice produced congruent results as our combined analysis. In Lipid, samples were assigned case-control labels according to their triglyceride levels. The results presented in the paper ignored the case-control status. We analyzed cases and controls separately, and the results were highly congruent to that of the combined analysis. We also analyzed African American samples in HapMap3 and did not find any region under selection, which agrees with a recent study [bib_ref] Genome-wide Scan of 29,141 African Americans Finds No Evidence of Directional Selection..., Bhatia [/bib_ref]. This serves as a negative control for ELAI. We devised a model fitting technique to introduce weighting into parameter estimation, which makes it possible to infer local ancestry of Mexicans using only European and African training samples. This rids us of the concern that the detected selection signals in Mexicans are artifacts produced by inaccurate Amerindian training samples. A previous study detected selection in 1000 genomes Mexican samples through local ancestry analysis [bib_ref] Detecting Structure of Haplotypes and Local Ancestry, Guan [/bib_ref]. Bhatia and colleagues questioned the plausibility of that finding; they argued that if signals were there, the 1000 genomes admixture analysis group would have found it [bib_ref] Genome-wide Scan of 29,141 African Americans Finds No Evidence of Directional Selection..., Bhatia [/bib_ref]. We took this opportunity to investigate why the 1000 genomes admixture analysis group failed to detect the strong selection at the MHC region in Mexicans. We simulated genotypes using a demographic model that mimic the out-of-Africa migration events [bib_ref] Calibrating a coalescent simulation of human genome sequence variation, Schaffner [/bib_ref] , performed forward simulations to mimic admixture and selection at three linked loci (details in Materials and Methods), and inferred local ancestry. The 1000 genomes used consensus call from four programs: HapMix [bib_ref] Sensitive Detection of Chromosomal Segments of Distinct Ancestry in Admixed Populations, Price [/bib_ref] , LAMP-LD [bib_ref] Fast and accurate inference of local ancestry in Latino populations, Baran [/bib_ref] , RFMix [bib_ref] RFMix: A Discriminative Modeling Approach for Rapid and Robust Local-Ancestry Inference, Maples [/bib_ref] , and MultiMix [bib_ref] Multiway Admixture Deconvolution Using Phased or Unphased Ancestral Panels, Churchhouse [/bib_ref]. The publicly available version of HapMix was designed exclusively for two-way admixture, and the extended version used to analyze the 1000 Genomes data was not available to us [bib_ref] An integrated map of genetic variation from 1,092 human genomes, Consortium [/bib_ref]. Thus it was excluded from our analysis. MultiMix performed poorly despite our best effort and was excluded as well. For both LAMP-LD and RFMix we used the same parameter settings as those used in the 1000 Genomes admixture analysis group [bib_ref] An integrated map of genetic variation from 1,092 human genomes, Consortium [/bib_ref]. Both LAMP-LD and RFMix require phased training samples, and RFMix also requires phased cohort samples. (ELAI works with diplotypes.) When supplied with true phasing, both LAMP-LD and RFMix works well, on par with ELAI. We then introduced 2% switch-errors into cohort haplotypes and training haplotypes that mimic Amerindians, 1% switch-errors into European and African training samples. LAMP-LD is robust to switch-errors, but RFMix under-performs [fig_ref] Fig 4: Comparison of estimations with and without Amerindian training samples [/fig_ref] It is worthwhile to note that MHC is notoriously hard to phase, and phasing for admixed samples at MHC is even more challenging as it requires the phasing algorithm to correctly identify local ancestry-a catch-22 for RFMix. We were surprised at the worse-than-the-expected performance of RFMix in the presence of switch-errors (S4C Further investigation revealed that its window size parameter has a sweet-spot (S4D When using the best window size RFMix performed on par with ELAI [fig_ref] Fig 4: Comparison of estimations with and without Amerindian training samples [/fig_ref] blackGoing back to the question why the 1000 genomes admixture analysis group failed to detect the signal, our simulation studies suggested that the democratic strategy adopted by 1000 genomes admixture analysis group, which used consensus calls from four methods to identify local ancestry, was perhaps not optimal. The simulation studies prompted us to use LAMP-LD and RFMix to analyze chromosome 6 of Viva and Lipid data. We phased the Maya and Pima samples from HGDP using SHAPEIT [bib_ref] A linear complexity phasing method for thousands of genomes, Delaneau [/bib_ref] , which were used in combination with CEU and YRI haplotypes as training datasets. LAMP-LD was then applied to infer local ancestry of Viva and Lipid datasets. We then phased the Viva and Lipid datasets, and RFMix was applied to infer their local ancestry. Reassuringly, both LAMP-LD and RFMix discovered the signal of selection at MHC (S5 The MHC region influences susceptibility and resistance to a broad range of infectious agents such as viruses, bacteria, and parasites. It is sensible to observe more alleles of African ancestry at MHC in Mexicans if those alleles confer selective advantages in the presence of certain infectious agents. The European conquerors brought to America European and African diseases such as smallpox, measles, and typhus. Spaniards imposed an urbanized life style and farming practice on native people. A sudden increase in local population concentration, displacement, social upheaval, food shortages, and stress made them much vulnerable to infectious diseases. An estimated 5-8 million native people perished in a smallpox epidemic alone in early 1500s. Nevertheless, after "difficult struggles of the formative period," the acceptance and enthusiasm of the new life emerged from the persistence of the old; for a brief period a "fusion of European and Mesoamerican cultures seemed ready to emerge". But severe drought hit and lethal pandemic broke out. The epidemic, called "huey cocoliztli," was symptomatically different from those imported from the Old World; some medical historians suspect it was a hemorrhagic fever caused by arenavirus carried by rodents. It first broke out in 1545 and lingered until 1815 [bib_ref] Large epidemics of hemorrhagic fevers in Mexico 1545-1815, Acuna-Soto [/bib_ref]. The epidemic selectively targeted native people, and 90% of the population perished in a few generations [bib_ref] Large epidemics of hemorrhagic fevers in Mexico 1545-1815, Acuna-Soto [/bib_ref]. This sustained epidemic harbors plenty of opportunities for strong selection at MHC, which fits our analysis. Once again history left its mark in genomes for posterity [bib_ref] A Genetic Atlas of Human Admixture History, Hellenthal [/bib_ref]. ## Materials and methods datasets The first dataset, Viva La Familia obesity-diabetes familial risk study (dbGaP Study Accession: phs000616.v1.p1), contains 858 genotyped individuals [bib_ref] Novel Genetic Loci Identified for the Pathophysiology of Childhood Obesity in the..., Comuzzie [/bib_ref]. Among them, 815 Mexicans children from 261 families were genotyped with Illumina HumanOmni 1-v1.0 BeadChips, and the remaining 43 children were genotyped on HumanOmni 2.5-8v1 BeadChips. We chose to analyze the 815 samples that were typed on the same chip. Study participants in Viva La Familiar study were recruited in Houston, Texas. The second dataset, Mexican hypertriglyceridemia study (dbGaP Study Accession: phs000618.v1.p1), contains 2229 samples with 1117 cases and 1112 controls, where the case-control status was ascertained based on an individual's serum triglyceride level [bib_ref] Genomic study in Mexicans identifies a new locus for triglycerides and refines..., Weissglas-Volkov [/bib_ref]. Note that although there were 4350 study samples reported in the paper, the dbGaP contains only 2229 that were genotyped with Illumina Human610-Quad BeadChipsstage 1 of the GWAS. The rest samples were only typed on selected 1200 SNPs-stage 2. Study participants in this study were recruited in Mexico City. We call the first dataset Viva and the second Lipid. ## Data quality control We removed all A/T, C/G SNPs whose potential allele flipping between different datasets cannot be identified without additional information. A SNP was removed if it was missing in one of the datasets, either training or cohort. We also removed SNPs whose missing proportion was larger than 5%. Although we realized that the Hardy-Weinburg disequilibrium test is not appropriate for admixed samples, we used it anyway to remove SNPs whose HWD test p-values <10 −6 . It is understood that this practice errs toward the safe side by eliminating possibly good SNPs. Finally, we obtained the cluster plots for each SNP, devised a simple algorithm to assign quality scores to each SNP cluster plot, and visually inspected those SNPs whose score indicated low quality. We removed those SNPs that contained a fourth cluster, or whose clusters were not distinct (examples of such cluster plots can be found in [bib_ref] Detecting Local Haplotype Sharing and Haplotype Association, Xu [/bib_ref]. We were particularly stringent to conduct such SNP quality control at the MHC region. Of the two GWAS datasets we obtained from dbGaP, Viva contains SNP cluster information, but Lipid does not. In the end, we had 352,754 SNPs from Viva and 479,757 SNPs from Lipid. The low number of SNPs in Viva reflected small number of overlapping SNPs between the Illumina HumanOmni 1-v1.0 and the Illumina 650Y arrays, the latter of which was used by the HGDP study that generates the Maya and Pima genotypes used as Amerindian training samples. ## Local ancestry inference We used ELAI [bib_ref] Detecting Structure of Haplotypes and Local Ancestry, Guan [/bib_ref] for local ancestry inference, which has been demonstrated to outperform competing methods such as HapMix [bib_ref] Sensitive Detection of Chromosomal Segments of Distinct Ancestry in Admixed Populations, Price [/bib_ref] and LAMP-LD [bib_ref] Fast and accurate inference of local ancestry in Latino populations, Baran [/bib_ref]. ELAI implements a two-layer cluster model and the model is fitted via the EM algorithm. The upper-layer clusters are parameterized to represent haplotypes from ancestral populations, and the lower-layer clusters contemporary haplotypes. The two-layer model was motivated by approximating the coalescent with recombination. It directly applies to diplotypes and automatically integrates out phase uncertainty. It can also estimate the recombination rates between markers, and hence doesn't require recombination map as an input. Thus, the requirement for running ELAI is minimal-just genotypes and marker positions. To run ELAI, one needs to provide training samples. We used European and African samples from HapMap3 and Maya and Pima samples from HGDP as default training samples (or reference panels, or source populations). ELAI is a cluster-based model and we wanted to specify numbers of clusters. The number of upper-layer clusters represents the number of source populations and we set it as 3; the number of lowerlayer clusters was set as 15. Extensive simulations demonstrated that this parameter setting performs well [bib_ref] Detecting Structure of Haplotypes and Local Ancestry, Guan [/bib_ref]. Lastly, we needed to specify number of admixing generations and we used 20. All ELAI results were averaged over 10 independent EM runs of 20 steps each, unless noted. ## Compute average ancestry dosages Lipid data contains unrelated individuals, and we treated an individual as unit and the computation is straightforward. Viva data contains 261 unrelated families. Each family contains 1-8 children, with majority of families (242) having 2-4 children. To account for relatedness in Viva data, we treat a family instead of an individual as unit, and computed the average dosages in the following manner: first we obtained family ancestral dosages by averaging over family members, and then we averaged over families to obtain overall average dosages. ## Assign different weights to training and cohort samples The two-layer model and the details of model fitting using EM algorithm can be found in [bib_ref] Detecting Structure of Haplotypes and Local Ancestry, Guan [/bib_ref]. Here we show how to estimate θ, the allele frequency associated with the cluster which emits the observed data. To simplify notation and presentation, we assume observing haplotypes instead of diplotypes. The weighting scheme can be applied to mixed sample that contains both haplotypes and diplotypes. To update parameters in each EM step, we take derivative of the expected full data log likelihood with respect to a parameter we want to update, say x 2 ξ, and solve for x to obtain updates. Z (i) is the latent state of haplotype h (i) , which contains two components, one for each layer of clusters. The expectation in Eq (1) is with respect to the posterior probability of latent states, conditioning on ξ à , which is the collection of parameters of the two-layer model estimated from the previous iteration, and ξ is the collection of parameters to be estimated. At marker m, write q ij ¼ P s pðZ m ¼ ðs; jÞjh ðiÞ m ; x à Þ, which is the marginal posterior probability of h ðiÞ m emitted from cluster j. Let T k ¼ fi : h ðiÞ m ¼ kg for k = 0,1. Take the derivative with respect to θ mj , which is the allele frequency associated with cluster j, to get [formula] À1 1 À t j X i2T 0 q ij þ 1 t j X i2T 1 q ij ¼ 0;ð2Þ [/formula] and solve to get [formula] t j ¼ P i2T 1 q ij P i2T 0 q ij þ P i2T 1 q ij ;ð3Þ [/formula] which can be thought as estimates of θ mj with equal weight 1. To apply differential weights, we split T k into training sample T ðtÞ k and cohort sample T ðcÞ k : For training sample we assign a weight w t and for cohort sample w c . Eq (3) is generalized to [formula] t j ¼ w t P i2T t 1 q ij þ w c P i2T c 1 q ij w t P i2T t 0 q ij þ P i2T t 1 q ij þ w c P i2T c 0 q ij þ P i2T c 1 q ij :ð4Þ [/formula] Let w t ) w c , then cohort samples contribute very little to t j when training samples are present. This is often desirable because the q ij estimates of training samples are more reliable, which is especially true in the context of imputation [bib_ref] Practical Issues in Imputation-Based Association Mapping, Guan [/bib_ref]. When training samples are missing, the first terms of both nominator and denominator on the right hand side which involve w t disappear and Eq (4) reduces to Eq (3). Using simulated data (described below), we fit the ELAI model using two training samples of European and African, discarding the Amerindian training samples. The African ancestral dosages were used to compare the inferred values and the truth. The results demonstrated that the weighting samples works well for selection coefficients of 0.02 and 0.05, and showed a bias for selection coefficient of 0.10, but the biased estimates were conservative for the purpose of detecting selection (S6 Fig). ## Define phenotype for association test We defined a marker set A that contained markers whose African average dosages were greater than 0.30. This threshold was 13 sample standard deviations away from the mean (in Lipid dataset), and the resulting markers formed a consecutive region within MHC. We assigned each individual a phenotypic value obtained by averaging African ancestry dosages over markers in A. ## Compute selection coefficient Let s be the selection coefficient, and f n (s) denote allele frequency at the n-th generation which is a function of s. Here the allele is referred to as a class of population specific alleles. Assume that the population size is constant but infinite so that we have a deterministic model. and f 20 (s) and we want to find s. Because f n (s) is a monotone function of s, we perform intervalbisection search to numerically solve for s. We start with an interval , such that f 20 (a)< f 20 (s)<f 20 (b), we evaluate y ¼ f 20 [formula] aþb 2 À Á ; [/formula] if y > f 20 (s), we set b = y; otherwise we set a = y. We repeat this procedure until y−f 20 (s)2(−, ) for a small . Note that to apply the recursion formulae, the input f 0 (s) and f 20 (s) have to be allele frequencies, which are half of the allele dosages for humans. We call the model defined by recursion f n+1 (s) = f n (s)(1+s) the simple model. It is easy to check that for dominance model we have f nþ1 ðsÞ ¼ f n ðsÞð1þsÞ 1þð2Àf n ðsÞÞf n ðsÞ s < f n ðsÞð1 þ sÞ; and for additive model we have f nþ1 ðsÞ ¼ f n ðsÞð1þsþf n ðsÞ sÞ 1þ2f n ðsÞ s < f n ðsÞð1 þ sÞ: Therefore the simple model produces a lower-bound estimate of s for both dominance and additive models. Let f 0 (s) = p 0 and f 20 (s) = p 1 , we have p 1 = p 0 (1+s) [bib_ref] Genome-wide Scan of 29,141 African Americans Finds No Evidence of Directional Selection..., Bhatia [/bib_ref] , and therefore the simple model estimate of selection coefficient is s = exp(log(p 1 /p 0 )/20) − 1. ## Simulate admixed samples under selection We used a population genetics model that mimics the out-of-Africa migration events to simulate a 3 Mb region of three source populations that mimic Amerindian, European, and African [bib_ref] Calibrating a coalescent simulation of human genome sequence variation, Schaffner [/bib_ref]. After setting aside 200 haplotypes from each source population as training haplotypes, we used the remaining haplotypes to simulate three-way admixed individuals by a one-pulse model [bib_ref] Mapping of disease-associated variants in admixed populations, Shriner [/bib_ref]. Specifically, we randomly selected 50,000 haplotypes from the three source populations using proportions of 50%, 45%, and 5%, mimicking the admixture proportion of Mexicans. We split 3 Mb into three segments, and assigned at two splitting points recombination hotspots. At each hotspot, we assumed equal recombination probability of 0.1,0.2, and 0.5 per generation. We sampled two haplotypes with replacement and introduced possible crossover events at hotspots to produce two new haplotypes. We repeated the pairing and crossover 25,000 times to produce 50,000 haplotypes for the next generation. The admixture simulation was done for 20 generations. To simulate selection, we designated the mid-section as the locus under selection, and assumed selection coefficients of 0.02,0.05, and 0.10 per generation. The alleles under positive selection were those 5% from the source population that mimicked Africans. After 20 generations, we randomly chose 2,000 admixed haplotypes, pairing them to form 1,000 diplotypes as cohort samples. We used two sizes of mid-section: 0.5 Mb and 1 Mb. A small mid-section produces a more challenging problem. To investigate how switch-errors affected local ancestry inference for different methods, in addition to perfect phasing situation, we also introduce 2% phasing errors into Amerindian training samples and the cohort samples, and 1% phasing errors to European and African training samples. To do so, at randomly selected heterozygous marker, from left to right we crossed-over two haplotypes. There are 9 combinations of crossover probability (0.1,0.2, and 0.5) and selection coefficients (0.02,0.05, and 0.10) for two sizes of the mid-section. This plot is for mid-section of size 1 Mb. The mid-section harbors alleles under selection, and a smaller size produces a more challenging problem. Plots also compare effects of phasing errors (2% for cohort and the Amerindian training sample and 1% for the other two training samples). RFMix underperformed after phasing errors were introduced in (b). Compared to RFMix, LAMP-LD was less sensitive to phasing errors. ELAI was unaffected by phasing errors. Parameters for LAMP-LD: window size = 100, number of HMM states = 25; for RFMix: window size = 0.1 cM, which approximately contains 100 SNPs. Both parameter settings were used by the 1000 Genomes admixture analysis group. B) Comparison between LAMP-LD, RFMix, and ELAI under different simulation conditions, Part II. The same simulation setup as in [fig_ref] Fig 4: Comparison of estimations with and without Amerindian training samples [/fig_ref] mid-section of size 0.5 Mb. C) RFMix performance with different switch-errors. This is the same dataset as used in S4B the legends, the number before the plus sign is the switch-error for cohort and the Amerindian training sample, and the number after is for the European and African training samples. D) RFMix performance with different choices of window size. The mean absolute deviation was computed from the same dataset that was used in [fig_ref] Fig 4: Comparison of estimations with and without Amerindian training samples [/fig_ref] simulation parameter settings (recombination probability and selection strength). The switch-errors were 2% for cohort and the Amerindian training sample and 1% for the other two training samples. We used centi-Morgan (cM) to measure the window size, which is the unit used by RFMix. In our simulations, 1 cM contains roughly 1000 SNPs. E) RFMix performs well with the optimal window size. This is the same dataset as used in S4B On each plot the main text displays the simulation parameters with C for crossover probability and S for selection coefficient. For example, C = 0.2, S = 0.05 means crossover probability is 0.2 and selection coefficient is 0.05. (PDF) ## Supporting information [fig] Fig 1: Global ancestry proportions and principal components. (a) and (b) are triangular plots for Viva and Lipid respectively. [/fig] [fig] Fig 2: African average dosages. Plot shows all 22 autosomes for two GWAS datasets. The spike at MHC region on chromosome 6 is rather striking in both datasets. The blue lines are the genome-wide mean of average dosages; the gray lines are mean ± 4ssd (ssd stands for sample standard deviation). [/fig] [fig] Fig 3: Comparison between different European and African training samples. The comparison was performed with chromosome 6 of Lipid dataset. African (a) and European (b) average dosages for five sets of training samples shown in legend, where ALL means CEU+TSI−YRI+MKK−MAYA. (c) The difference of estimated European average dosages of Mexicans between two European training samples (see main text for explanation). (d) The violin plots of structure analysis of five HapMap3 populations, where ASW denotes Americans from the Southwest, an African American population. On each violin plot, gray dot denotes the median and black dot the mean. doi:10.1371/journal.pgen.1005847.g003 [/fig] [fig] Fig 4: Comparison of estimations with and without Amerindian training samples. (a) African average dosages of Viva. (b) Amerindian average dosages of Viva. (c) African average dosages of Lipid. (d) Amerindian dosages of Lipid. We combined CEU and TSI as European training samples, and YRI and MKK as African training samples. doi:10.1371/journal.pgen.1005847.g004 value of p 1 at MHC in the past 20 generations, and it is selection at work. blackA selection coefficient s can be computed via a simple model p 1 = p 0 × (1+s) [/fig] [fig] S1: Table. Summary statistics for different sets of training samples for autosomes of the Lipid dataset. ALL means CEU+TSI−YRI+MKK−MAYA, and ssd means sample standard deviation. (PDF) Fig. Average dosages with different European and African training samples. A) Average dosages for Amerindian (blue), European (red), and African (green) ancestries for Viva (top) and Lipid (bottom) datasets with training samples of CEU−YRI−MAYA. B) Average dosages for Amerindian (blue), European (red), and African (green) ancestries for Lipid dataset with training samples of CEU−MKK−MAYA (top) and TSI−MKK−MAYA (bottom). C)Average dosages for Amerindian (blue), European (red), and African (green) ancestries for Lipid dataset with training samples of TSI−YRI−MAYA (top) and CEU+TSI−YRI+MKK−MAYA (bottom). (PDF) S2 Fig. Comparison of Amerindian average dosages. The five sets of training samples are shown in the legend, where ALL means CEU+TSI−YRI+MKK−MAYA. The comparison was performed with chromosome 6 of Lipid dataset. (PDF) S3 Fig. SNPs associated with enrichment of African local ancestry in Lipid dataset. Bayes factors (BF) were computed using BIMBAM. The horizontal blue line is log 10 BF = 10. (PDF) S4 Fig. Simulation studies to evaluate LAMP-LD and RFMix. A) Comparison between LAMP-LD, RFMix, and ELAI under different simulation conditions, Part I. [/fig] [table] Table 1: Estimates of selection coefficient s under different models. p 0 is the genome-wide mean of African average dosages; p 1 is the peak African average dosage at MHC. [/table]

Aging and the (Chemical) Senses: Implications for Food Behaviour Amongst Elderly Consumers Publisher's Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.Abstract: The growing aging population are increasingly suffering from the negative health consequences of the age-related decline in their senses, especially their chemical senses. Unfortunately, however, unlike for the higher senses of vision and hearing, there is currently nothing that can be done to bring back the chemical senses once they are lost (or have started their inevitable decline). The evidence suggests that such chemosensory changes can result in a range of maladaptive food behaviours, including the addition of more salt and sugar to food and drink in order to experience the same taste intensity while, at the same time, reducing their overall consumption because food has lost its savour. Here, though, it is also important to stress the importance of the more social aspects of eating and drinking, given the evidence suggesting that a growing number of older individuals are consuming more of their meals alone than ever before. Various solutions have been put forward in order to try to enhance the food experience amongst the elderly, including everything from optimising the product-intrinsic food inputs provided to the remaining functional senses through to a variety of digital interventions. Ultimately, however, the aim has to be to encourage healthier patterns of food consumption amongst this rapidly-growing section of the population by optimising the sensory, nutritional, social, and emotional aspects of eating and drinking. An experimental dinner with the residents of one such home where nostalgic-flavoured healthy ice-creams were served is described. # Introduction The chemical senses, namely smell (olfaction), taste (gustation), and the trigeminal sense, just like the higher spatial senses of vision, audition, and touch (e.g., [bib_ref] Sensory processes and age effects in normal adults, Corso [/bib_ref] [bib_ref] Aging in the eighties, impaired senses for sound and light in persons..., Havlik [/bib_ref] [bib_ref] Central somatosensory conduction time from 10 to 79 years, Hume [/bib_ref] start their inevitable decline as we age (e.g., [bib_ref] Uniformity of olfactory loss in aging, Cain [/bib_ref] [bib_ref] Development of taste perception in humans: Sensitivity and preference throughout the life..., Cowart [/bib_ref] [bib_ref] Relationships between taste and smell across the adult life span, Cowart [/bib_ref] [bib_ref] Influence of age and age-related diseases on olfactory function, Doty [/bib_ref] [bib_ref] Performance on an odor detection and identification test as a predictor of..., Koskinen [/bib_ref] [bib_ref] Aging in the olfactory system, Mobley [/bib_ref] [bib_ref] Prevalence of olfactory impairment in older adults, Murphy [/bib_ref] [bib_ref] Aging and sensory senescence, Nusbaum [/bib_ref] [bib_ref] Variability of olfactory threshold and its role in assessment of aging, Stevens [/bib_ref] [bib_ref] Assessment of chemosensory functioning in aging: Subjective and objective procedures, Weiffenbach [/bib_ref] [bib_ref] Chemical senses in aging, Weiffenbach [/bib_ref]. Unfortunately, however, there are currently no remediation devices, such as glasses and hearing aids, that can be used to make up for the loss of the chemical senses [bib_ref] Taste and smell perception affect appetite and immunity in the elderly, Schiffman [/bib_ref] [bib_ref] The skin as a medium for sensory substitution, Spence [/bib_ref]. This is especially problematic given a growing aging population, with more people than ever before living to an age where impairments to their chemical senses starts to become much more noticeable. To give some sense of the emerging problem, according to the National Institute on Aging, National Institutes of Health, the number of those living into their eighties has grown exponentially over the preceding 40 years. ## Sensory decline in aging Our sensory and perceptual functions start their inevitable decline at different ages and at very different rates (see [bib_ref] Cognitive tasks in several modalities, Axelrod [/bib_ref] [bib_ref] Sensory functioning, Baker [/bib_ref] [bib_ref] Does sensory function decline independently or concomitantly with age? Data from the..., Gadkaree [/bib_ref]. For example, while tactile, auditory and visual perceptual abilities show clear signs of decline by the time we reach middle age, taste and smell sensitivity do not show any marked deterioration until we reach 60-70 years of age, whereupon aging takes a greater toll on smell than on taste [bib_ref] Sensory processes and age effects in normal adults, Corso [/bib_ref] [bib_ref] Chemical senses in aging, Weiffenbach [/bib_ref] [bib_ref] Effects of ageing on smell and taste, Boyce [/bib_ref] [bib_ref] Smell identification ability: Changes with age, Doty [/bib_ref] [bib_ref] Perspectives on population-based epidemiological studies of olfactory and taste impairment, Hoffman [/bib_ref] [bib_ref] Decline in taste and odor discrimination abilities with age, and relationship between..., Kaneda [/bib_ref]. These age-related declines in sensory processing reflect the consequences of both peripheral physical and central neural degeneration (such as the increasing opacity of the lens of the eye, and the general reduction in the population of nerve cells), as well as the cognitive decline associated with a general loss of flexibility of mental processing in older individuals (see also [bib_ref] Educating your nose, Cain [/bib_ref] [bib_ref] Sensory-specific impairment among older people. An investigation using both sensory thresholds and..., Cavazzana [/bib_ref] [bib_ref] Global sensory impairment in older adults in the United States, Correia [/bib_ref]. Some commentators have even suggested that this cognitive inflexibility may actually provide older people with the "comfort and security in seeing and hearing events in the accustomed way" [bib_ref] Sensory processes and age effects in normal adults, Corso [/bib_ref] , p. 96; see also [bib_ref] Taste and smell perception in aging, Weiffenbach [/bib_ref]. The decline in the sense of smell with aging [bib_ref] Decline in taste and odor discrimination abilities with age, and relationship between..., Kaneda [/bib_ref] [bib_ref] Bars, P. Taste and aging, Bartoshuk [/bib_ref] [bib_ref] Chemical senses and aging: Taste versus smell, Stevens [/bib_ref] is likely to exert a more pronounced detrimental effect on multisensory flavour perception than any loss of taste (gustation), given figures suggesting that as much as 75-95% of what we think we taste, we actually smell. Here, though, it is important to highlight the potentially important distinction between orthonasal smell (as when we sniff food) and the retronasal release of volatile-rich aromas pulsed out from the back of the nose when we swallow and masticate [bib_ref] Optimal directional volatile transport in retronasal olfaction, Ni [/bib_ref] [bib_ref] Taste-smell confusions" and the duality of the olfactory sense, Rozin [/bib_ref]. Interestingly, while most studies of the decline in olfactory sensitivity with aging have assessed orthonasal olfaction, it has been suggested that the findings do not necessarily provide a reliable prediction of retronasal experience, especially when dealing with complex foods [bib_ref] Performance on an odor detection and identification test as a predictor of..., Koskinen [/bib_ref] [bib_ref] Olfactory dysfunction and related nutritional risk in free-living elderly women, Duffy [/bib_ref] [bib_ref] Measurement of sensitivity to olfactory flavour: Application in a study of aging..., Duffy [/bib_ref] [bib_ref] Correspondence between three olfactory tests and suprathreshold odor intensity ratings, Koskinen [/bib_ref]. Research looking specifically at the age-related loss in the sense of taste (gustation) has revealed a decline in sensation for the majority of basic taste qualities [bib_ref] Ageing and taste, Methven [/bib_ref] [bib_ref] Sweet and sour discrimination abilities of elderly people compared to those of..., Mingioni [/bib_ref] [bib_ref] Taste perception with age: Generic or specific losses in supra-threshold intensities of..., Mojet [/bib_ref]. That said, there is also some older evidence to suggest that people's sensitivity to specific individual taste qualities may also change differentially across the lifespan. For example, the sensitivity to sweetness is sometimes maintained while the sensitivity to saltiness has been reported to deteriorate significantly (e.g., [bib_ref] Taste: Robust across the age span?, Bartoshuk [/bib_ref]. One of the key problems for those hoping to optimise the design of food and drink products for the elderly is that the variability in sensitivity to olfactory stimuli across the population tends to increase as we age [bib_ref] Evidence for different patterns of chemosensory alterations in the elderly population: Impact..., Sulmont-Rossé [/bib_ref]. What this means, in practice, is that while some elderly individuals may be functionally anosmic, others may retain a level of olfactory functioning that is not much different from their younger counterparts. Another point to stress here is that, if anything, the decline in chemosensory function in the elderly often appears to be more apparent when assessed with pure tastants/olfactory stimuli than when assessed with real food stimuli (e.g., see [bib_ref] Sweet and sour discrimination abilities of elderly people compared to those of..., Mingioni [/bib_ref] [bib_ref] Taste perception with age: Generic or specific losses in supra-threshold intensities of..., Mojet [/bib_ref]. Here, it would certainly also be interesting to know more about whether or not orthonasal and retronasal olfactory abilities decline at the same rate in the elderly [bib_ref] Performance on an odor detection and identification test as a predictor of..., Koskinen [/bib_ref] [bib_ref] Individual differences in retronasal odor responsiveness: Effects of aging and concurrent taste, Flaherty [/bib_ref]. Intriguingly, those who have lost the ability to taste, for example, in the case of herpes zoster oticus in the case of the psychophysicist Pfaffmann (see [bib_ref] Taste loss due to herpes zoster oticus: An update after 19 months, Pfaffmann [/bib_ref] , report remarkably little loss of taste sensation. Similarly, Brillat-Savarinreported on the case of soldiers whose tongues had been cut out in the Algerian war also reporting little loss of sensation as far as the flavour of food and drink were concerned. By contrast, those of us who have had a head cold know only too well the profound loss of taste that often results when olfactory inputs are not available. Meanwhile, in a study of nearly 2000 people ranging from 5 to 99 years in age, Doty et al. [bib_ref] Smell identification ability: Changes with age, Doty [/bib_ref] reported that the ability to identify smells peaked between the ages of 20 and 40 years, and started to decline thereafter. In fact, over half of the 65-80 year olds tested by Doty and his colleagues, and more than three-quarters of those over 80 years of age, exhibited major impairments in olfactory processing, with many exhibiting a clinical deficit in their ability to sense, known as anosmia (see also the National Geographic Smell Survey, [bib_ref] National geographic smell survey: Effects of age are heterogenous, Wysocki [/bib_ref] , for a much larger survey of the effects of age on smell; and [bib_ref] Uniformity of olfactory loss in aging, Cain [/bib_ref]. Crucially, recent increases in life expectancy mean that more people than ever before are currently suffering from age-related impairments in their ability to taste and smell. To give some idea of the magnitude of this change in chemosensory perception, it is predicted that by the year 2025 more than a billion people around the world will be over 60 years of age [bib_ref] Food quality report from noninstitutionalized aged, Stevens [/bib_ref]. Research directly comparing age-related declines in olfactory and gustatory sensitivity suggests that olfactory losses tend to start earlier and to be more severe than those seen for taste (e.g., [bib_ref] Relationships between taste and smell across the adult life span, Cowart [/bib_ref]. Moreover, it is important to note that the severity and nature of these age-related declines in chemosensory functioning show marked variability across individuals, and also vary markedly as a function of the particular measures of olfactory perceptual ability that happen to be used (e.g., [bib_ref] Assessment of chemosensory functioning in aging: Subjective and objective procedures, Weiffenbach [/bib_ref]. The decline in chemosensory functioning presumably helps to explain why it is that so many elderly individuals complain that food lacks flavour (given that much of the flavour of food actually comes from its smell; [bib_ref] Influence of age and age-related diseases on olfactory function, Doty [/bib_ref]. The growing popularity of pungent spices, such as chilli pepper, black pepper, and ginger in food has also been attributed in part to the effects of aging [bib_ref] Pepper potency and the forgotten flavor sense, Lawless [/bib_ref] [bib_ref] Nutrition and taste and smell deficits: A risk factor or an adjustment, Ferris [/bib_ref]. Similarly the provision of a variety of seasonings, such as butter, tomato ketchup, lemon, parsley, mayonnaise, etc., was shown to exert a positive effect on meal enjoyment and food intake in one study conducted in a nursing home [bib_ref] Improving meal context in nursing homes. Impact of four strategies on food..., Divert [/bib_ref] [bib_ref] Increases in energy, protein and fat intake following the addition of sauce..., Appleton [/bib_ref]. ## Decline in salivary function with aging Although it is not often mentioned in the literature, it is important to note that saliva also plays a key role in helping us to masticate and swallow food (see [bib_ref] Mouth-watering: The influence of environmental and cognitive factors on salivation and gustatory/flavour..., Spence [/bib_ref] , for a review). It also plays an important role in our ability to experience the taste and flavour of food too [bib_ref] Saliva and flavor perception: Perspectives, Canon [/bib_ref] [bib_ref] Main effects of human saliva on flavour perception and the potential contribution..., Muñoz-González [/bib_ref]. Importantly, salivary function has been reported to decline with increasing age and hence this will also interfere with multisensory flavour perception in the elderly (see [bib_ref] Effect of oral and nasal chemoreception on parotid gland secretion, Chauncey [/bib_ref] [bib_ref] Flow rates of resting whole and stimulated parotid saliva in relation to..., Percival [/bib_ref] [bib_ref] Age, sex and contralateral differences in the volumes of human submandibular salivary..., Scott [/bib_ref] [bib_ref] Salivary flow decreases in healthy elderly people independently of dental status and..., Vandenberghe-Descamps [/bib_ref] [bib_ref] Taste and salivary function, Weiffenbach [/bib_ref] [bib_ref] Impact of aging on human salivary gland function: A community-based study, Yeh [/bib_ref]. The effect of aging on the gut should not be forgotten either [bib_ref] Food intake and ageing-the role of the gut, Parker [/bib_ref]. ## Unhealthy eating habits in the elderly The declining chemosensory abilities that have been extensively documented in the elderly can all too easily lead to unhealthy eating habits [bib_ref] Living arrangements and dietary patterns of older adults in the United States, Davis [/bib_ref] [bib_ref] Old and alone: Barriers to healthy eating in older men living on..., Hughes [/bib_ref] , as the latter increase their intake of salt and sugar to make up for their inability to taste these ingredients in food at lower concentrations (e.g., [bib_ref] Taste and smell perception affect appetite and immunity in the elderly, Schiffman [/bib_ref] [bib_ref] Effect of olfactory deficits on nutritional status: Does age predict persons at..., Ferris [/bib_ref]. According to Stevens, Cain, Demarque, and Ruthruff [bib_ref] On the discrimination of missing ingredients: Aging and salt flavour, Stevens [/bib_ref] , older individuals may need to add as much as two or three times more salt to perceive the same intensity in a tomato soup as those who are younger. Shockingly, this figure increased to twelve times for those older individuals who were on five or more medications, which turns out to be the majority of them (see also [bib_ref] Influence of age and age-related diseases on olfactory function, Doty [/bib_ref]. Given the negative health consequences of the overconsumption of salt (e.g., hypertension), this likely represents a very serious issue, and one that needs to be tackled by those hoping to optimise food delivery amongst the elderly. Note here only that according to observational data from the Framingham study in North America, the lifetime risk of developing hypertension in those who are 55-65 years of age is 90% [bib_ref] Residual lifetime risk for developing hypertension in middle-aged women and men: The..., Vasan [/bib_ref]. Researchers argue that many of the most serious problems faced by the elderly stem from age-related changes in their sense of smell and to a lesser extent taste, rather than from the age-related declines affecting any of the other senses. This is all the more unfortunate given that many older people claim that eating and drinking represent the last remaining pleasures in their lives [bib_ref] Psychophysiological effect of odor, Manley [/bib_ref]. Or, as the famous French gastronome Jean Anthelme Brillat-Savarin put it almost two centuries ago that: "The pleasures of the table, belong to all times and all ages, to every country and to every day; they go hand in hand with all our other pleasures, outlast them, and remain to console us for their loss.", p. 14). The fundamental point here is that the provision of acceptable food needs to be recognised as an important factor determining the quality of life for many older individuals, especially those who find themselves living in care facilities [bib_ref] Effects of relaxing music on agitation during meals among nursing home residents..., Goddaer [/bib_ref] , p. 150; [bib_ref] Resident and staff ratings of foodservices in long-term care: Implications for autonomy..., West [/bib_ref]. ## Malnutrition-an increasingly common problem amongst the elderly Many older people, especially those who find themselves in hospital or else in a care home setting, all too often fail to eat sufficiently to maintain their weight (e.g., [bib_ref] Changes during aging and their association with malnutrition, Amarya [/bib_ref] [bib_ref] Protein-caloric malnutrition in the nursing home, Rudman [/bib_ref] [bib_ref] Weight loss and metabolic changes in dementia, Wang [/bib_ref]. It is therefore critically important that we introduce strategies to encourage increased consumption of nutritionally-balanced foods and so avoid the malnutrition that is so often reported amongst the elderly (see also [bib_ref] Nutrition management in nursing homes, Coulston [/bib_ref]. Indeed, the importance of establishing robust solutions for long-term care residents has been highlighted by a number of authors in recent decades (e.g., [bib_ref] Enhancing the dining experience in long-term care, Dory [/bib_ref] ; see also [bib_ref] Quality of care in nursing homes: From the resident's perspective, Pearson [/bib_ref]. Meanwhile, according to another recent report from the National Health Service here in the UK, providing elderly patients with an extra meal a day halved their chances of dying while in hospital. This presumably assumes that they eat that meal, and do not return it to the kitchen untouched, as is unfortunately so often the case. One-on-one support with eating at mealtimes has been shown to be especially effective in increasing patients' consumption of food in the hospital setting, but the associated staff costs normally limit the uptake of this kind of solution. The metallic taste that many patients report while undergoing treatment for cancer would also appear to be an effective appetite suppressant (see [bib_ref] The mystery of "metal mouth" in chemotherapy, Reith [/bib_ref]. While cancer can strike at any age, the risk increases markedly with age [bib_ref] Age and cancer risk: A potentially modifiable relationship, White [/bib_ref] , thus meaning that the problem of metallic taste may be more prevalent amongst this age group. ## Enhancing the sensory appeal of food and drink amongst the elderly One route to increasing the sensory/perceptual interest of food and drink amongst older populations is to enhance the stimulation provided by the remaining functional senses [bib_ref] Perception of chemosensory stimuli and related responses to flavored yogurts in the..., Koskinen [/bib_ref] [bib_ref] Flavor enhancement as a tool for increasing pleasantness and intake of a..., Koskinen [/bib_ref]. It has, for example, been suggested that increasing the trigeminal input by incorporating more pepper pungency, heat from chili, and ginger, etc., may help to prevent foods from becoming too bland as an individual's gustatory and olfactory function declines [bib_ref] Pepper potency and the forgotten flavor sense, Lawless [/bib_ref] ; though see also [bib_ref] Reduction of odor and nasal pungency associated with aging, Stevens [/bib_ref] ; and [bib_ref] Interactions between texture and trigeminal stimulus in a liquid food system: Effects..., Forde [/bib_ref] , on the effects of interactions between texture and trigeminal stimulus in a liquid food system on the preferences of elderly consumers). That said, the evidence to date demonstrating increased consumption of taste/flavourenhanced foods amongst the elderly is mixed (see [bib_ref] The changing role of the senses in food choice and food intake..., Boesveldt [/bib_ref] , for a review of research to date). When thinking about how to enhance the design of food and drink experiences amongst the elderly, it is important to note that much of our multisensory flavour experience is determined by our flavour expectations that are built on the basis of associative learning as a result of our prior food experiences. The latter tend to be set by sight, orthonasal smell (sniffing), and, on occasion, sound (think here only of the sizzle of the steak on the hot plate, or the ding of the microwave;and touch cues-as we feel the softness of the fruit, or the heat emanating from the outer surface of our coffee cup (e.g.,. It has been suggested that very often, we live in the world of our flavour expectations, only occasionally checking on the taste of what we are actually consuming (see [bib_ref] The generation of sensory expectation by external cues and its effect on..., Deliza [/bib_ref] [bib_ref] Sensory expectations based on product-extrinsic food cues: An interdisciplinary review of the..., Piqueras-Fiszman [/bib_ref] , for reviews). Should the taste experience be pretty much as we expected, then we largely live in the world of our flavour expectations. If, however, there is too much of a divergence between expectations and experience, then this may very well lead to a negatively-valenced disconfirmation of expectation response [bib_ref] Effects of product beliefs on product perception and liking, Schifferstein [/bib_ref]. At the same time, however, it is also important to recognize how mental imagery sometimes help to fill in the gaps in our perception of what we expect to taste and smell (see [bib_ref] Multimodal mental imagery, Nanay [/bib_ref] [bib_ref] Crossmodal mental imagery, Spence [/bib_ref]. The key point here though is that the visual attributes of those food products designed specifically for the elderly should not be neglected. That is, it is not enough to simply think about flavour enhancement as the only solution to get elderly individuals to eat more/better [bib_ref] The changing role of the senses in food choice and food intake..., Boesveldt [/bib_ref]. It has long been suggested that meals should be made more colourful and sonically interesting for elderly and hospitalized individuals (e.g.,. Further, beyond enhancing the multisensory perception of the food itself, one should not neglect the importance of the colour of the plateware on which it is served (see [bib_ref] Background colour & its impact on food perception & behaviour, Spence [/bib_ref] for a review of the impact of plateware colour on taste and consumption). It has been reported that some older individuals may struggle to distinguish the food from the plate visually, especially when pallid white institutional foods are served against the background of the ubiquitous large round white American plate. Interesting here, therefore, is research showing that simply by switching to high-contrast coloured (e.g., red or blue) plateware and glassware, a dramatic increase in the consumption of food can be achieved, at least in the short term, amongst older patients and care home residents (e.g., [bib_ref] Visual contrast enhances food and liquid intake in advanced Alzheimer's disease, Dunne [/bib_ref] [bib_ref] Impact of three different plate colours on short-term satiety and energy intake:..., Akyol [/bib_ref] [bib_ref] Impact of type, size and shape of plates on hospital patients' perceptions..., Hannan-Jones [/bib_ref]. Thinking more carefully about the presentation of the food can also help to encourage greater consumption amongst the elderly [bib_ref] Improved meal presentation increases food intake and decreases readmission rate in hospitalized..., Navarro [/bib_ref]. Many older individuals, including those who have lost their teeth, are often fed pureed meals as they can find it hard to deal with solid foods [bib_ref] Measuring eating capability, liking and difficulty perception of older adults: A textural..., Laguna [/bib_ref]. Unfortunately, however, such texturally monotonous foods have lost most of their sensorial interest and hence may result in undernutrition. What is more, the absence of textural cues make it much harder for people to identify vegetables too (see [bib_ref] The role of smell, taste, flavour and texture cues in the identification..., Van Stockkom [/bib_ref]. Note here also how many elderly people suffer from a more general difficulty in identifying foods [bib_ref] Food recognition in the elderly, Schiffman [/bib_ref]. This is potentially important because people are less likely to consume those foods that they struggle to identify. Japanese researchers have developed a headset that older people in this situation can wear that presents mastication-like sounds elicited by jaw movements [bib_ref] The effect of a crunchy pseudo-chewing sound on perceived texture of softened..., Endo [/bib_ref]. It has even been suggested that different sounds might be used to represent different food textures, thus providing an additional sensory cue. Preliminary findings suggest that this might provide an effective means of adding some sonic interest to mealtimes for such individuals, it remains to be seen whether there will be widespread uptake of such high-tech solutions, especially amongst elderly and hospitalized individuals. ## Ice-cream as an effective vehicle for nutrient delivery in the elderly Ice-cream is often noted as being a popular food amongst many older individuals (e.g., [bib_ref] Dietary patterns and associated factors among the elderly, Ferreira [/bib_ref] [bib_ref] Using ice-cream as an effective vehicle for energy/nutrient delivery in the elderly, Spence [/bib_ref]. Indeed, its unique sensory properties have been highlighted as one of the reasons why so many people find that they still have space for this highly-desirable food even when they are otherwise full at the end of the meal [bib_ref] Dynamic contrast: A sensory contribution to palatability, Hyde [/bib_ref] ; see also [bib_ref] Cold pleasure. Why we like ice drinks, ice-lollies and ice cream, Eccles [/bib_ref]. It may be the case that in some older individuals, the oral-somatosensory cues that are provided by the cold temperature of the ice-cream, as well as perhaps the fatty/creamy mouthfeel characteristics help to provide agreeable sensory stimulation from food in those who may otherwise be suffering from marked olfactory loss or else may even be functionally anosmic (e.g., [bib_ref] National geographic smell survey: Effects of age are heterogenous, Wysocki [/bib_ref]. Working with this idea, and against the common conception of ice-cream as an unhealthy (and possibly childish/necessarily indulgent) food, chef Jozef Youssef of Kitchen Theorycreated a series of ice-creams using a range of healthier ingredients (including pureed vegetables and meal replacement powders such as Huel; [bib_ref] Using ice-cream as an effective vehicle for energy/nutrient delivery in the elderly, Spence [/bib_ref]. Furthermore, by using a Pacojet machine to make the ice-cream, it was possible to deliver a deliciouslysmooth texture without the necessity of adding cream. Unfortunately, however, the high price of the latter machines (c. £5000 for a new model) will likely limit the uptake of this item of modernist culinary technology by most of those providing food for the elderly. The intervention of chefs and neuroscientists to help recover pleasure lost due to sensory loss was also investigated by the Roca Brothers in Spain in 2019 (see, for a couple of press reports). While the concept of savoury ice-creams is currently unfamiliar to many Western consumers, they are nevertheless popular in Japan, as well as in the context of many modernist restaurants around the world [bib_ref] Assessing the long-term impact of the molecular gastronomy movement on haute cuisine, Spence [/bib_ref]. What is more, savoury ice-creams were once popular in Europe and presumably beyond during the 18th and 19th centuries (e.g., see [bib_ref] The Cheese). Food & Wine, 23, Campbell-Schmitt [/bib_ref]. Appropriately-designed (i.e., nutritionally-balanced) ice-creams may therefore provide an excellent vehicle for the delivery of protein and other elements necessary for a balanced diet in elderly populations who might otherwise be suffering from poor nutrition [bib_ref] Influence of soy protein isolate on physical and sensory properties of ice..., Akesowan [/bib_ref] [bib_ref] The effect of soy protein concentrate on the physical, chemical and sensory..., Dervisoglu [/bib_ref] [bib_ref] Increasing the protein content of ice cream, Patel [/bib_ref] ; see also [bib_ref] Effect of calcium chloride addition on ice cream structure and quality, Costa [/bib_ref] [bib_ref] Novel meat-enriched foods for older consumers, Farouk [/bib_ref]. Here, it is worth noting that much the same approach to nutritionally-enhanced ice-cream has also been proposed previously in the case of cancer patients [bib_ref] Adapted ice cream as a nutritional supplement in cancer patients: Impact on..., Casas [/bib_ref]. That said, when introducing novel ice-cream flavours, one has to be careful not to trigger a negatively-valenced 'disconfirmation of expectation' response amongst consumers, elderly or otherwise, who may initially be unfamiliar with such savoury flavours in the context of ice-cream [bib_ref] The role of expectancy in sensory and hedonic evaluation: The case of..., Yeomans [/bib_ref] ; see also. Chef Jozef Youssef and his team created a range of savoury ice-creams with various flavours chosen to elicit positive nostalgia amongst older individuals. Given the UK base for this particular intervention, the meal incorporated Heinz cream of tomato soup, prawn cocktail, and bone marrow ice cream flavours (see [fig_ref] Figure 1: The menu created especially for the nostalgic flavours ice-cream-focused dining concept created... [/fig_ref] for the menu from the event). These dishes were served in the context of a multisensory environment that was itself designed to trigger positive nostalgia. For example, traditional visual designs were projected onto the dining table along with retro food labels matching the flavour of the dish the aged diners, and their carers, were currently eating using projection mapping (see [fig_ref] Figure 2: Still images from the Denville Hall dining concept [/fig_ref]. Meanwhile, Vera Lynn, one of the most popular vocalists during the war years, and such-like was presented over the loudspeakers. Several recent studies have demonstrated that ambient soundscapes influence both the sensory-discriminative and hedonic experience of ice-cream and gelati [bib_ref] Listening to music can influence hedonic and sensory perceptions of gelati, Kantono [/bib_ref] [bib_ref] The effect of music on gelato perception in different eating contexts, Kantono [/bib_ref] [bib_ref] Background soundscapes influence the perception of ice-cream and on electrophysiological measures, Xu [/bib_ref] [bib_ref] Changes in flavour, emotion and electrophysiological measurements when chocolate ice cream is..., Xu [/bib_ref]. Although no quantitative data were obtained, the qualitative reports of the various residents and supporters of the Denville Hall residential home for aged actors who supported this particular culinary exercise were, on the whole, very positive. Indeed, the authors hope to collect quantitative data to support the approach outlined here (namely using healthy ice-cream as a vehicle to enhance food consumption behaviours amongst the elderly) in the near future. Until such time, however, the findings reported here should be treated as merely anecdotal. These dishes were served in the context of a multisensory environment that was itself designed to trigger positive nostalgia. For example, traditional visual designs were projected onto the dining table along with retro food labels matching the flavour of the dish the aged diners, and their carers, were currently eating using projection mapping (see [fig_ref] Figure 2: Still images from the Denville Hall dining concept [/fig_ref]. Meanwhile, Vera Lynn, one of the most popular vocalists during the war years, and such-like was presented over the loudspeakers. Several recent studies have demonstrated that ambient soundscapes influence both the sensory-discriminative and hedonic experience of ice-cream and gelati [bib_ref] Listening to music can influence hedonic and sensory perceptions of gelati, Kantono [/bib_ref] [bib_ref] The effect of music on gelato perception in different eating contexts, Kantono [/bib_ref] [bib_ref] Background soundscapes influence the perception of ice-cream and on electrophysiological measures, Xu [/bib_ref] [bib_ref] Changes in flavour, emotion and electrophysiological measurements when chocolate ice cream is..., Xu [/bib_ref]. Although no quantitative data were obtained, the qualitative reports of the various residents and supporters of the Denville Hall residential home for aged actors who supported this particular culinary exercise were, on the whole, very positive. Indeed, the authors hope to collect quantitative data to support the approach outlined here (namely using healthy ice-cream as a vehicle to enhance food consumption behaviours amongst the elderly) in the near future. Until such time, however, the findings reported here should be treated as merely anecdotal. ## Music and soundscapes to enhance meal times amongst agitated seniors Another relatively-simple low-cost intervention to enhance food behaviour at mealtimes is to use music, or ambient soundscapes, to help relax those individuals who might otherwise be too agitated to eat. This is apparently a common problem amongst many psychiatric patients as well as a growing number of those older individuals who are suffering from Alzheimers/dementia (e.g., [bib_ref] Effects of relaxing music on agitation during meals among nursing home residents..., Goddaer [/bib_ref] [bib_ref] Dinner music: Does it affect the behavior of psychiatric patients?, Courtright [/bib_ref] [bib_ref] Mealtime problems in a continuing-care hospital for the elderly, Davies [/bib_ref] [bib_ref] Influence of dinner music on food intake and symptoms common in dementia, Ragneskog [/bib_ref]. Intriguingly, back in the 1970s, a number of psychiatric hospitals in North American would apparently play 'Sea gulls . . . Music for rest and relaxation'for just this reason [bib_ref] Prevalence and characteristics of persons with dependency on feeding at institutions for..., Sandman [/bib_ref]. ## Hunger and forgetting to eat Research conducted with individuals suffering from amnesia suggests that it is external cues that often trigger the initiation of meal consumption in the absence of awareness/memory of the meals that may just have been consumed [bib_ref] What causes humans to begin and end a meal? A role for..., Rozin [/bib_ref]. At the same time, however, many of those older individuals living alone may simply forget to eat [bib_ref] Health and social factors affecting the food choices and nutritional intake of..., Wylie [/bib_ref] , because they lack the robust external (or exogenous) hunger cues, such as the kitchen aromas of food cooking that may play an important part in encouraging the rest of us that it is time to eat. ## Music and soundscapes to enhance meal times amongst agitated seniors Another relatively-simple low-cost intervention to enhance food behaviour at mealtimes is to use music, or ambient soundscapes, to help relax those individuals who might otherwise be too agitated to eat. This is apparently a common problem amongst many psychiatric patients as well as a growing number of those older individuals who are suffering from Alzheimers/dementia (e.g., [bib_ref] Effects of relaxing music on agitation during meals among nursing home residents..., Goddaer [/bib_ref] [bib_ref] Dinner music: Does it affect the behavior of psychiatric patients?, Courtright [/bib_ref] [bib_ref] Mealtime problems in a continuing-care hospital for the elderly, Davies [/bib_ref] [bib_ref] Influence of dinner music on food intake and symptoms common in dementia, Ragneskog [/bib_ref]. Intriguingly, back in the 1970s, a number of psychiatric hospitals in North American would apparently play 'Sea gulls…Music for rest and relaxation'for just this reason [bib_ref] Prevalence and characteristics of persons with dependency on feeding at institutions for..., Sandman [/bib_ref]. ## Hunger and forgetting to eat Research conducted with individuals suffering from amnesia suggests that it is external cues that often trigger the initiation of meal consumption in the absence of awareness/memory of the meals that may just have been consumed [bib_ref] What causes humans to begin and end a meal? A role for..., Rozin [/bib_ref]. At the same time, however, many of those older individuals living alone may simply forget to eat [bib_ref] Health and social factors affecting the food choices and nutritional intake of..., Wylie [/bib_ref] , because they lack the robust external (or exogenous) hunger cues, such as the kitchen aromas of food cooking that may play an important part in encouraging the rest of us that it is time to eat. In order to try and address the latter problem, Prof. Spence was involved as a consultant in a project a few years ago designed to try and help older individuals, specifically early-stage Alzheimers/dementia patients who might otherwise need to be hospitalized due to undernutrition [bib_ref] Improvements in nutritional intake and quality of life among frail homebound older..., Gollub [/bib_ref] , to retain their independence living at home for a little longer. The idea behind the 'Ode', as it is called, was to release familiar meal time-specific and culturally/age group-appropriate ambient food scents into the home three times a day. The hope was that this might help those who might otherwise forget to eat, to eat. The results of a small study suggested the efficacy of this award-winning plug-in food scent alarm clock device. The six food aromas developed for the launch included fresh orange juice, cherry Bakewell tart, homemade curry, pink grapefruit, beef casserole, and Black Forest gateau. In order to try and address the latter problem, Prof. Spence was involved as a consultant in a project a few years ago designed to try and help older individuals, specifically early-stage Alzheimers/dementia patients who might otherwise need to be hospitalized due to undernutrition [bib_ref] Improvements in nutritional intake and quality of life among frail homebound older..., Gollub [/bib_ref] , to retain their independence living at home for a little longer. The idea behind the 'Ode', as it is called, was to release familiar meal time-specific and culturally/age group-appropriate ambient food scents into the home three times a day. The hope was that this might help those who might otherwise forget to eat, to eat. The results of a small study suggested the efficacy of this award-winning plug-in food scent alarm clock device. The six food aromas developed for the launch included fresh orange juice, cherry Bakewell tart, homemade curry, pink grapefruit, beef casserole, and Black Forest gateau. They were specifically chosen to be representative of food aromas that were likely to be familiar to those in the target age group (though see also [bib_ref] Decreased discrimination of food odors in the elderly, Schiffman [/bib_ref]. The results of a smallscale 10-week pilot study involving fifty people with dementia, along with their families, revealed that more than half of those who used the device ended up maintaining their weight, or else showing a slight increase, as compared to an expected decline in weight that is so often seen in this group (e.g.,. ## Age-related decline in multisensory integration and attention Flavour is undoubtedly one of the most multisensory of our everyday experiences [bib_ref] Multisensory flavour perception, Spence [/bib_ref] , potentially engaging each and every one of our senses. Hence, over and above any loss of sensitivity in the individual senses that either help to set our flavour expectations, or else contribute directly to our flavour experiences, one might also ask what role, if any, a more central loss of multisensory integration, or information-processing abilities [bib_ref] Information-processing rates in the elderly, Cerella [/bib_ref] [bib_ref] The rise and fall in information processing rates over the life span, Cerella [/bib_ref] , might have for food perception/behaviour amongst the elderly. While there has been little research directly targeting this question with respect to the chemical senses, one can find a multitude or answers in the case of the higher spatial senses [bib_ref] Cognitive neuroscience of aging: Contributions of functional neuroimaging, Cabeza [/bib_ref]. For example, Laurienti, Burdette, Maldjian, and Wallace [bib_ref] Enhanced multisensory integration in older adults, Laurienti [/bib_ref] have argued that multisensory integration may actually become more important as we age. They suggest that multisensory integration can, in some sense at least, help to make up for the loss of sensitivity and responsiveness of the individual senses as they start their inevitable decline (see also [bib_ref] Aging gracefully: Compensatory brain activity in high-performing older adults, Cabeza [/bib_ref]. In their study, Laurienti and colleagues had groups of younger and older participants make speeded detection responses to a random sequence of auditory, visual, and audiovisual target stimuli. While unisensory response latencies were shown to slow with increasing age, multisensory RTs were similarly fast in both age groups. At the same time, however, other researchers have argued that the ability to integrate multisensory cues declines with increasing age (see also [bib_ref] Multisensory integration mechanisms during aging, Freiherr [/bib_ref] [bib_ref] Multisensory processes in old age, Laurienti [/bib_ref]. In particular, the increased risk of falling that has been documented amongst the elderly has been put down, at least in part, to a failure to appropriately integrate vestibular and visual cues [bib_ref] Impaired multisensory integration predisposes the elderly people to fall: A systematic review, Zhang [/bib_ref]. The authors are not, however, aware of any research that has specifically addressed the question of whether there is any central impairment affecting the multisensory integration of the flavour senses (namely, retronasal olfaction, gustation, and possibly also trigeminal inputs) with advancing years. Here, it is worth stressing that it is not only the neural sites of multisensory integration that differ between the chemosensory and the spatial senses, but also the very nature of the rules governing that integration (see [bib_ref] Extending the study of visual selective attention to a multisensory world, Spence [/bib_ref] , for a review). So, for example, spatial co-location, and attention have been reported to play more of a modulatory role over the integration of auditory, visual, and tactile stimuli than would appear to be the case for the integration of the flavour senses. Visual (colour) cues, in particular, exert a significant modulatory effect over the sensory-discriminative and hedonic aspects of tasting (see [bib_ref] Does food color influence taste and flavor perception in humans?, Spence [/bib_ref] , for reviews). The visual dominance over taste/flavour perception as well as the central importance of visual cues for driving our food selection behaviours both need to be recognised (see. Hence, one might consider whether the presentation of more brightly-coloured foods/dishes could be used to help stimulate the appetite amongst the elderly [bib_ref] Taste and smell perception affect appetite and immunity in the elderly, Schiffman [/bib_ref] ; though see also [bib_ref] Sensory dominance and multisensory integration as screening tools in aging, Murray [/bib_ref] , on the changing patterns of sensory dominance with aging). Note here also how the presentation of a visually-stimulating array of produce colours also fits with contemporary nutritional guidelines [bib_ref] The importance of the visual aesthetics of colours in food at a..., Paakki [/bib_ref] [bib_ref] Visual attractiveness depends on colorfulness and color contrasts in mixed salads, Paakki [/bib_ref]. Attention plays a key role in both the phenomenon of oral referral [bib_ref] Oral referral: On the mislocalization of odours to the mouth, Spence [/bib_ref] [bib_ref] The role of attention in the localization of odors to the mouth, Stevenson [/bib_ref] [bib_ref] Modality-specific neural effects of selective attention to taste and odor, Veldhuizen [/bib_ref] and multisensory flavour perception more generally [bib_ref] Attending to the chemical senses, Spence [/bib_ref]. Furthermore, in younger people, the research clearly shows that increasing the perceptual load of a visual task lowers taste/flavour perception [bib_ref] Leaving a flat taste in your mouth: Task load reduces taste perception, Van Der Wal [/bib_ref] ; see also [bib_ref] Temporarily loading visual attention induces a prolonged loss of olfactory awareness, Forster [/bib_ref]. Hence, one danger might be that the loss of flexibility of attentional allocation/switching, in the elderly [bib_ref] Watching television while eating increases energy intake. Examining the mechanisms in female..., Braude [/bib_ref] might mean that the TV dinner has an even more detrimental effect on the perception of food consumption-related sensory cues than on the rest of us. ## Dining as a fundamentally social activity However, over and above any perceptual decline, a large part of the poor food consumption behaviours that have been evidenced in older populations are likely to result, at least in part, from the increasingly isolated living that many older people face, and the consequent lack of social interaction that have been documented amongst older age groups in many countries. To give some sense of the problem in relation to food consumption, in Japan, where people live longer than in most other places, it has been estimated that 24% of pensioners eat the majority of their meals alone [bib_ref] Combined effects of eating alone and living alone on unhealthy dietary behaviors,..., Tani [/bib_ref]. Importantly, here in the UK, eating meals alone has recently been reported to be the biggest lifestyle cause of unhappiness [bib_ref] Eating alone and depression in older men and women by cohabitation status:..., Tani [/bib_ref]. Recognizing this growing social issue, there is an emerging interest in the field of digital commensality, with a number of researchers trying to bring back the enjoyment of eating by means of digital technologies [bib_ref] Digital commensality: On the pros and cons of eating and drinking with..., Spence [/bib_ref]. While acknowledging the lack of familiarity with contemporary digital technologies amongst many in this age group currently, there has nevertheless been widespread interest in the possibility of using robotic dining assistants to interact with elderly patients, simultaneously monitoring their consumption, and ideally nudging them toward healthier patterns of food consumption. Assisted eating at mealtimes in hospitals also helps here, though, as has already been mention, the cost implications cannot be ignored. This kind of approach recognizes the fact that both physiological and psychological factors play an important role in controlling eating in later life [bib_ref] Eating in late life: Physiological and psychological factors, Busse [/bib_ref]. Of course, the social solutions in this space need not be digital, with other commentators advocating for lunch clubs. Of perhaps more widespread relevance, given the rapid rise of at home food delivery services, there seems to be an unrecognized opportunity amongst the larger providers of food direct to the home (be it take-away or ready-made meals, or even meal kits, from the likes of Deliveroo, Uber Eats, or Blue Apron) to connect those individuals who may be living, (cooking) and eating alone over the internet with other like-minded individuals who find themselves in the same position [bib_ref] Digital commensality: On the pros and cons of eating and drinking with..., Spence [/bib_ref]. There is also scope to enhance the provision of meals for one, too. # Conclusions A wide body of evidence now points toward the conclusion that the rapidly-growing aging population are engaging in a variety of unhealthy eating behaviours. On the one hand, this takes the form of adding excessive amounts of sugar and salt to taste [bib_ref] On the discrimination of missing ingredients: Aging and salt flavour, Stevens [/bib_ref] , while, at the same time, not eating enough to maintain a healthy weight. Indeed, clinical malnutrition is not uncommon amongst hospitalised elderly individuals, not to mention those living in care homes [bib_ref] Changes during aging and their association with malnutrition, Amarya [/bib_ref] [bib_ref] Protein-caloric malnutrition in the nursing home, Rudman [/bib_ref] [bib_ref] Nutrition management in nursing homes, Coulston [/bib_ref]. While sensory decline, specifically amongst the chemical senses, is undoubtedly partly to blame, the absence of social interaction is also likely to be a core part of the problem [bib_ref] Digital commensality: On the pros and cons of eating and drinking with..., Spence [/bib_ref]. Indeed, one of the key challenges for the future is how to deal with the sensory underload so often experienced by those of advanced years, many of whom are living alone and eating alone far more often than they would like. Optimizing food design to stimulate the remaining functional senses (e.g., [bib_ref] Research prospects in nutrition and the chemical senses in aging, Murphy [/bib_ref] [bib_ref] Improving Food Sensory Quality with and for Elderly Consumers, Sulmont-Rossé [/bib_ref] and harnessing various digital technologies to increase the opportunity for distributed social interaction at mealtimes (and beyond) are all likely going to help manage the situation in the years ahead [bib_ref] Digital commensality: On the pros and cons of eating and drinking with..., Spence [/bib_ref]. Looking to the future, it will likely become increasingly important to consider the changing role of the senses in food choice and food intake across the lifespan [bib_ref] The changing role of the senses in food choice and food intake..., Boesveldt [/bib_ref]. Furthermore, the idea of transgenerational design will likely become increasingly important too [bib_ref] The aging process: A challenge for design, Haigh [/bib_ref] , especially given the increasing value of the silver dollar [bib_ref] The silver sensory experience-A review of senior consumers' food perception, liking and..., Doets [/bib_ref]. The importance of marketing foods effectively to this growing demographic should not be underestimated either [bib_ref] Marketing to older adults: An updated overview of present knowledge and practice, Moschis [/bib_ref]. As stressed in this review, one promising vehicle for the delivery of nutritional requirements in at least some elderly individuals may be nutritionally-enhanced ice-creams (see [bib_ref] Using ice-cream as an effective vehicle for energy/nutrient delivery in the elderly, Spence [/bib_ref] ; see also [bib_ref] Novel flavours paired with glutamate condition increased intake in older adults in..., Dermiki [/bib_ref]. Such foods optimise food-related stimulation of the remaining functional senses, though the possibility of 'disconfirmation of expectation' given the novelty of such unusual formulations/flavours needs to be carefully thought through [bib_ref] The role of expectancy in sensory and hedonic evaluation: The case of..., Yeomans [/bib_ref]. Although there is as yet limited research into this approach, the preliminary evidence at least looks promising [bib_ref] Using ice-cream as an effective vehicle for energy/nutrient delivery in the elderly, Spence [/bib_ref]. Ultimately, it is only by optimizing the sensory, social, nutritional, and emotional aspects of food and eating design that we will be able to provide a satisfactory food environment in the years ahead. Further, given the many reports of chemosensory loss constituting one of the most common symptoms of COVID-19 [bib_ref] Acute nasal dryness in COVID-19, Navarra [/bib_ref] , a loss that has been reported to affect not only smell and taste, but also the trigeminal sense [bib_ref] More than smell-COVID-19 is associated with severe impairment of smell, taste, and..., Parma [/bib_ref] , one might wonder to what extent some of the sensory strategies outlined here to deal with the sensory losses experienced by the elderly might be relevant to providing enhanced food experiences for the worryingly large number of those suffering from long COVID. Indeed, according to the results of the latest report, approximately one-third of patients suffering from long COVID reported impaired chemosensory function three months after infection [bib_ref] Mysteries of COVID Smell Loss Finally Yield Some Answers, Sutherland [/bib_ref]. Author Contributions: J.Y. and his team at Kitchen Theory developed the dining concept and recipes that were incorporated in the menu. C.S. wrote the literature review and integrated comments from J.Y. in revision. All authors have read and agreed to the published version of the manuscript. Funding: This research received no external funding. # Data availability statement: There is no data to share, though chef J.Y. woud be happy to share the recipes on request. [fig] Figure 1: The menu created especially for the nostalgic flavours ice-cream-focused dining concept created for Denville Hall. [/fig] [fig] Figure 2: Still images from the Denville Hall dining concept. [/fig]

Solubility, Stability, and Avidity of Recombinant Antibody Fragments Expressed in Microorganisms Solubility of recombinant proteins (i.e., the extent of soluble versus insoluble expression in heterogeneous hosts) is the first checkpoint criterion for determining recombinant protein quality. However, even soluble proteins often fail to represent functional activity because of the involvement of non-functional, misfolded, soluble aggregates, which compromise recombinant protein quality. Therefore, screening of solubility and folding competence is crucial for improving the quality of recombinant proteins, especially for therapeutic applications. The issue is often highlighted especially in bacterial recombinant hosts, since bacterial cytoplasm does not provide an optimal environment for the folding of target proteins of mammalian origin. Antibody fragments, such as single-chain variable fragment (scFv), single-chain antibody (scAb), and fragment antigen binding (Fab), have been utilized for numerous applications such as diagnostics, research reagents, or therapeutics. Antibody fragments can be efficiently expressed in microorganisms so that they offer several advantages for diagnostic applications such as low cost and high yield. However, scFv and scAb fragments have generally lower stability to thermal stress than full-length antibodies, necessitating a judicious combination of designer antibodies, and bacterial hosts harnessed with robust chaperone function. In this review, we discuss efforts on not only the production of antibodies or antibody fragments in microorganisms but also scFv stabilization via (i) directed evolution of variants with increased stability using display systems, (ii) stabilization of the interface between variable regions of heavy (V H ) and light (V L ) chains through the introduction of a non-native covalent bond between the two chains, (iii) rational engineering of V H -V L pair, based on the structure, and (iv) computational approaches. We also review recent advances in stability design, increase in avidity by multimerization, and maintaining the functional competence of chimeric proteins prompted by various types of chaperones. # Introduction Antibodies are widely used for medical applications such as disease diagnosis and therapy. Valuable pharmaceutical properties of antibodies such as high affinity to their target molecules have led to them becoming constituted as key materials not only in antibody-based biosensors, which offer the promise of in-depth target detection capacitybut also in antibody-based proteomics, which provides insights into cancer disease states via cancer biomarker discovery. In addition to the intrinsic property, industrial applications require high productivity and long shelflife from thermal stress, organic solvents, and other stresses than physiological conditions. However, production of full-length antibodies cost extremely high, as they are typically expressed in mammalian cell lines such as Chinese hamster ovary (CHO) or NS0 murine myeloma cell linesdue to N-glycan heterogeneity among different speciesand also the complex disulfide bond pattern; hence, the biopharmaceutics industry has devoted immense resources on its production processes. Instead, the single-chain variable fragment (scFv,, a rational polypeptide design, consisting only of variable regions from heavy (V H ) and light (V L ) chains, joined together by a linker, not only maintains antigen binding capacitybut also can easily be produced in prokaryotes, such as Escherichia coli (E. coli) or Brevibacillus choshinensis (B. chosinensis), along with high yield, which keeps the cost of production low. Despite the advantages of scFvs, they have a few drawbacks that limit their therapeutic potential, such as (i) deteriorated stability because of their propensity to readily aggregate under thermal stress; (ii) a short serum half-life of <1 day compared to 3 weeks for full-length immunoglobulin G (IgG)1, IgG2, and IgG4 antibodies; and (iii) reduced affinity compared to the full-length antibody counterpart. Therefore, scFv format is suitable for limited cases, such as macular degeneration or blood-related diseases. Immunoglobulin G, the most abundant monoclonal antibody (mAb) isotype in serum is composed of two antigen binding fragments (Fab) and one homodimeric fragment crystallizable (Fc) domain that contribute to the overall stability of the molecule;. Since Fab of an IgG becomes more sensitive to the heat denaturation when Fc region is removed, researchers have tried to engineer Fab to stabilize the interactions between constant heavy 1 (C H 1) and constant light (C L ) chains in order to obviate the need for using mammalian host cells for the expression of full-length antibodies because of N-glycan on the Fc region. This requires immense resources such as expensive media, facilities to maintain germfree conditions, and time. However, limited successes have been made. Further elimination of C H 1-C L pair in Fab, resulting in fragment variable (Fv), significantly discounts thermodynamic stability. This is presumably due to the unnatural exposure of the lower V L and V H regions, flanking C H 1 and C L , where hydrophobic interaction used to contribute to the stability as a whole as well as the absence of the contribution of C H 1, which controls the assembly of heavy and light chains of the whole IgG molecule. The only light-heavy intermolecular disulfide bond in native IgG antibodies on the residues Cys220 in C H 1 and Cys214 in C L of Fab region in canakinumab; PDB ID of 5BVJ) contributes to the thermodynamic stability of the whole Fab fragment. In addition, intramolecular disulfide bonds in both the V H and V L regions is critical in the thermodynamic stability because elimination of them significantly enhanced the propensity of scFv aggregation. In this article we review efforts on increasing expression yield as well as protein stability of antibody fragment and recent diverse designs of antibody fragments. ## Production of antibody or antibody fragments in bacteria To reduce the cost of production of antibodies, researchers in both academia and industry put enormous efforts on elevating expression yield of IgG antibody or its fragment by (i) engineering expression plasmids, i.e., rhamnose-inducible expression systemor comprehensive optimization via high-throughput screening, (ii) engineering global sigma factor RhoD, which regulates more than 1,000 gene expressions, and (iii) devising bacterial strains capable of forming disulfide bonds in cytoplasm such as CyDisCoor SHuffle. Despite the advantages, drawbacks limiting its potential are (i) the low stability of scFvs, known for their propensity to readily aggregate under thermal stress, (ii) absence of glycosylation machinery, (iii) lack of efficient secretory mechanism as compared to yeast or animal cells, functionally limited protein trafficking machinery from the cytoplasm to the periplasmic space or to the outside of the cells, and (iv) overproduction of acetic acid byproduct during fermentation. ## Engineering intrinsic stability of scfvs: directed evolution, rational design, and computational approaches Antibody fragments can be expressed in several compartments in E. coli: mostly as inclusion bodies in the cytoplasm, or as soluble forms displayed on (i) the inner membrane, (ii) in the periplasmic space, (iii) on the outer membrane, and/or (iv) outside the bacterium, facilitated by various signal sequences, such as outer membrane protein A (OmpA), pectate lyase B (PelB), or new lipoprotein A (NlpA). To overcome the drawbacks of scFvs, which comprise only V H and V L antigen-binding domains, to reduce the protein size in order to increase protein production but maintain high target molecule affinity, researchers have engineered scFvs with resistance to aggregation and enhanced intrinsic stability of antibody fragments.summarizes the engineering efforts. ## Directed evolution Greg Winter et al., utilized phage display directed evolution methodology to isolate V H variants that are more resistant to heat denaturation. They further engineered V H and identified a key residue, Arg28 in V H that renders resistance to heat and acid aggregation. Daniel Christ's group at the Garvan Institute of Medical Research selected critical residues for antigen binding in both V H and V L and constructed a phage library introducing aspartate or glutamate in those residues to screen for heat resistance. The isolated variants resulted in not only enhanced biophysical property but also structural conservation. Dane Wittrup's group at MIT devised a yeast surface display system to isolate scFv variants with high affinity to antigen and increased stability by constructing yeast mutant libraries, expressing scFv on the cell surface, followed by successive rounds of flow cytometry sorting. Brian Inc. used sequence-and structure-based analyses to devise a high-throughput screening methodology that measure scFv extracellularly expressed by E. coli. This screening methodology resulted in enhanced melting temperature (Tm) by 14 - C and additional Tm improvement by 12 - C through combination of the resulting variants. ## Rational design Unlike the presence of intermolecular C H 1-C L disulfide bond, there is not one in the native V H -V L. Instead of placing a linker between V H and V L (scFv increating non-native disulfide bond between V H and V L via substituting amino acid residues in both framework 2 (FR2) in V H and FR4 in V L (dsFv; disulfide-stabilized Fv inled to indistinguishable specificity to antigen and similar cytotoxic activity when fused with exotoxin but exhibited superior protein stability at 37 - C, compared to scFv counterpart. Similarly, substitution of Val84 in V H to aspartic acid led to not only improved periplasmic production by 25-fold but also decreased the rate of thermally induced aggregation reaction. In another study, introduction of Cys44 in V H and Cys100 in V L in anti-aflatoxin B(1) scFv resulted in improved stability and resistance to protein aggregation. Introduction of the disulfide bond to anti-FGF2 diabody (ds-diabody, see section "multimerization" for diabody) also improved biological activity. This is presumably due to its lower propensity to the open state of V H -V L pair, in contrast to the presence of both assembled and disassembled state in case of scFv where V H and V L domains are simply connected by a G 4 S linker. These results indicate that bridging V H and V L by establishing intermolecular disulfide bond formation via cysteine residue incorporation can be a decent strategy for Fv stabilization. Another recent approach incorporating closed state of V H -V L pair is a cyclization of scFv using an enzyme sortase A, which ligate the pair, leading to cyclic scFv: this methodology markedly suppressed aggregation tendency without affecting affinity to antigen. Alexey Lugovskoy et al., at Merrimack Pharmaceuticals, Inc. showed that both essential and non-essential tyrosine residues for antigen binding in either CDR or FR can improve the biophysical property of scFv . An-Suei Yang et al., at the National Defense Medical Center, Taipei, elucidated the nature of intra-and inter-hydrophobic domains of scFv: the former is flexible and indirectly affects antigen binding, as opposed to the latter affecting antigen binding directly. Robin Curtis's group at the University of Manchester investigated the aggregation propensity of argininerich scFv under denaturing condition: substitution of arginine residues in scFv with lysine significantly reduced aggregation. This diverse knowledge of protein nature in scFv may assist engineers with consensus-based design of antibody fragment for generating stabilizing mutations to pre-existing scFvsor bispecific antibody fragments. ## Computational approaches Andreas Plückthun's group at University of Zurich designed a stabilized scFv from human V H germline sequences by analyzing hydrophobic core, pairing of hydrogen bonds, clusters of charge, and packing of β-sheets, leading to reduction of G = 20.9 kJ/mol as well as improvement of scFv expression yield by 4-fold. Furthermore, they could stabilize scFv by CDR grafting to more stable framework, using a structure-based analysis. Computational web servers, such as Prediction of Immunoglobulin Structure (PIGS)or Web Antibody Modeling (WAM), made computational modeling of antibody variable regions possible. Importantly, recent advances in structural knowledge and computational protein modeling such as RosettaDesign accelerated antibody design toward improved antigen affinity as well as physicochemical properties. The homology modeling provides with guidance on not only prediction on the conformations of CDR loops but also V H -V L orientations via energy calculations such as antibody-antigen docking, comparing with known crystal structures. For example, a computational homology modeling significantly improved resistance of scFvs to heat inactivation FIGURE 2 | A crystal structure of fragment antigen binding (Fab) region of canakinumab, a human monoclonal antibody neutralizing IL-1β (PDB ID: 5BVJ). V H regions are colored in light green; C H 1 in light blue; V L in orange; C L in cyan. Cysteine residues involved in disulfide bonds are partially depicted in red ball and stick model. The only intermolecular disulfide bonds between heavy and light chain of IgG molecule is present on the cysteine residues in C H 1 (Cys220) and C L (Cys214), respectively. The positions of the cysteine residues vary among different V H and V L due to the differences in CDR length. by supercharging the protein through energy calculations. In addition, combinatorial engineering approach, including CDR grafting onto stable frameworks, V H -V L interface stabilization, and in vitro somatic hypermutation significantly increased thermal stability of full-length antibody by 10 - C, compared to the native IgG1 antibody. Recent advances in computational methodologies for both antibody sequencing and backbonesand for multistage processing of antibody engineering, capitalizing on computational design and experimental validation cycles, have enabled successful de novo antibody engineering, such as anti-influenza antibodies. Of note, Georgiou and Ellington at the University of Texas at Austin used the Rosetta modeling programto predict amino acid substitutions for anti-HA33 scFv stabilization and confirmed a melting temperature increase of 4.5 - C by antigen-binding enzyme-linked immunosorbent assay (ELISA) after thermal stress for 2 h at 70 - C. ## A fusion partner Tagging of anti-FGFR3 scFv with a solubilizing partner small ubiquitin-related modifier (Sumo), followed by removal of the Sumo protein using Sumo protease, enabled over 95% purity with the yield of 4 mg/L bacterial culture. The resulting anti-FGFR3 scFv has exhibited complete biological activity . Another example is an "Fv-clasp, " where scFv was fused with anti-parallel coiled coil structure, SRAH domain of human Mst1 with scFv. In addition, introduction of disulfide bond to the Fv greatly enhanced thermal stability and tendency of crystallization. This is presumably due to the shielding of hydrophobic residues exposed in Fv, according to the X-ray crystallography. Alternatively, approaches based on the chaperone function of RNAs could also be considered. Fusion with an RNA-interaction domain (RID) greatly enhances the solubility (i.e., the ratio of soluble versus insoluble expression in heterogeneous hosts) and the overall yield of soluble proteins, harnessed with unique properties of RNAs as chaperone (chaperna; chaperone + RNA), although this approach has not yet been documented for recombinant antibody fragments. ## Engineering the host cell: chaperone coexpression or genome-level screening The folding of nascent polypeptides is often assisted by molecular chaperones, although their utility in recombinant expression has been documented only in limited cases . The stability problem associated with V H -V L pair in the scFv molecule can be circumvented through assistance with the chaperone of a pairing vehicle. Coexpression of the chaperone Skp enhanced binding activity of anti-TLH scFv by 3-4 fold, relative to the native counterpart, expressed in E. coli. ## Multimerization Diabodies not only render bivalency but also enhance stability of Fv by linker design (Le. Introducing covalent bonds between V H 1 and V L 2 as well as V H 2 and V L 1 enables bispecific binding capacity of the two distinct scFv in one molecule (diabody in. One example in clinic is blinatumomab (BLINCYTO, R 2014), a bispecific scFv for CD19 and CD3, which functions as a T lymphocyte engager to cancerous B lymphocytes for the treatment of acute lymphoblastc leukemia. In addition, constructing anti-HER3 trivalent scFv using SpyCatcher ligase system enhanced affinity by 12-fold as compared to a monomeric anti-HER3 counterpart. Another general approach of antibody fragments is the utilization of targeting ligands on nanoparticles in nanomedicine. Several antibody fragment-based nanoparticles are under clinical trials, including Erbitux-EDVS pac , which is a bacteria-derived mini-cell nanoparticle targeting EGFR currently under clinical phase II. Multimerization of scFvs as nanoparticles, using self-assembling scaffolds via chaperna approachholds promise for further enhancing the avidity and thermostability of recombinant antibody fragments. # Conclusion Up to March 2020, the US Food and Drug Administration (FDA) approved eight antibody fragments as drugs, six of which are produced from E. coli (75%,Microorganisms are favorable expression hosts for antibody fragments, such as scFvs or Fab, in therapeutic applications, because of the low production cost and lack of a carbohydrate chain. However, despite these advantages, scFvs expressed in bacteria have neither comparable stability relative to native full-length antibodies nor a comparable production yield of ∼1 g/L in bioreactorsrelative to mammalian hosts, that is, >10 g/L in CHO cells. Therefore, scientists and engineers in both academia and industry put extensive efforts on increasing production yield as well as protein stability of scFv expressed in bacteria. To obtain improved yield various bacterial expression systems have been developed in terms of vector systems or engineered strains with engineered chaperone molecules. The low intrinsic solubility and stability of native scFv protein with a relatively shorter shelf-life is a bottleneck for industrial application. To overcome the disadvantages there have been enormous research attempts on stability design via site-directed mutagenesis, generation of non-natural covalent bonds between the heavy and light variable chains, rational design, and recently computer-based engineering or chimeric approaches. Engineering of scFv with respect to increasing the stability lowers both the kinetic complexity in folding process and subsequently the propensity to aggregate into non-functional form. Folding into soluble, functional form with the desired level of avidity is often aided by exploiting the chaperone function of naturally existing molecular chaperones or artificial solubilizing tags. Besides thermodynamic aspects on overall stability, due consideration should be given to the kinetic aspects in de novo folding pathway for designer antibody fragments toward improved solubility, thermal stability, and productivity. # Author contributions TK and BS designed and wrote the manuscript. Both authors agreed to be accountable for the content of the work. # Funding This work was supported by grants from the Korean government (NRF-2020R1F1A1072124, NRF-2018M3A9H4079358, and HI20C0144).

Geographical, Temporal and Environmental Determinants of Bryophyte Species Richness in the Macaronesian Islands Species richness on oceanic islands has been related to a series of ecological factors including island size and isolation (i.e. the Equilibrium Model of Island Biogeography, EMIB), habitat diversity, climate (i.e., temperature and precipitation) and more recently island ontogeny (i.e. the General Dynamic Model of oceanic island biogeography, GDM). Here we evaluate the relationship of these factors with the diversity of bryophytes in the Macaronesian region (Azores, Madeira, Canary Islands and Cape Verde). The predictive power of EMIB, habitat diversity, climate and the GDM on total bryophyte richness, as well as moss and liverwort richness (the two dominant bryophyte groups), was evaluated through ordinary least squares regressions. After choosing the best subset of variables using inference statistics, we used partial regression analyses to identify the independent and shared effects of each model. The variables included within each model were similar for mosses and liverworts, with orographic mist layer being one of the most important predictors of richness. Models combining climate with either the GDM or habitat diversity explained most of richness variation (up to 91%). There was a high portion of shared variance between all pairwise combinations of factors in mosses, while in liverworts around half of the variability in species richness was accounted for exclusively by climate. Our results suggest that the effects of climate and habitat are strong and prevalent in this region, while geographical factors have limited influence on Macaronesian bryophyte diversity. Although climate is of great importance for liverwort richness, in mosses its effect is similar to or, at least, indiscernible from the effect of habitat diversity and, strikingly, the effect of island ontogeny. These results indicate that for highly vagile taxa on oceanic islands, the dispersal process may be less important for successful colonization than the availability of suitable ecological conditions during the establishment phase. # Introduction The Equilibrium Model of Island Biogeography (EMIB) states that, other things being equal, area and geographic isolation are the two main factors determining extinction and immigration rates, which in turn regulate the level of species richness that is reached at a dynamic equilibrium [bib_ref] An equilibrium theory of insular zoogeography, Macarthur [/bib_ref] ,. Although many hypotheses have been proposed to explain the role of area in species richness patterns, in its original formulation the EMIB postulated the effect of area per se referring specifically to demographic processes (i.e. smaller areas support smaller populations that are hence more prone to species extinctions). Despite its importance in the development of ecology and biogeography, the EMIB has been criticized for the lack of ability of its simple mechanisms to account for variations in species richness (e.g. [bib_ref] Is a new paradigm emerging for oceanic island biogeography, Heaney [/bib_ref]. In fact, models based on additional factors have been suggested to also account for island diversity, including energy, [bib_ref] Species-energy theory -an extension of species-area theory, Wright [/bib_ref] , habitat diversity [bib_ref] A model for the species-area-habitat relationship, Triantis [/bib_ref] or island ontogeny in the particular case of oceanic archipelagos [bib_ref] A general dynamic theory of oceanic island biogeography, Whittaker [/bib_ref]. In essence, the models considering energy relate the amount of available resources with the possibility of maintaining higher population sizes and therefore more species [bib_ref] Species-energy theory -an extension of species-area theory, Wright [/bib_ref]. The variety of resource types (e.g. habitat diversity) would also promote the coexistence of more species by diminishing interspecific competition and increasing sympatric speciation through ecological space partitioning (see [bib_ref] Island species richness increases with habitat diversity, Hortal [/bib_ref]. The ontogenetic evolution of the island itself may affect its carrying capacity and hence species richness, because the variations in area and structural complexity occurring during the island's life cycle influence both typical immigrationextinction dynamics and diversification by in situ speciation (i.e. the General Dynamic Model of oceanic island biogeography or GDM) [bib_ref] A general dynamic theory of oceanic island biogeography, Whittaker [/bib_ref] , [bib_ref] A general dynamic theory of oceanic island biogeography: extending the MacArthur-Wilson theory..., Whittaker [/bib_ref]. The effects of energy, habitat diversity and island ontogeny on species richness have been typically examined using surrogates such as actual evapotranspiration or other climatic factors (e.g. [bib_ref] A global model of island biogeography, Kalmar [/bib_ref] , topographic variables or habitat classifications (e.g. [bib_ref] The roles of island area per se and habitat diversity in the..., Ricklefs [/bib_ref] , and the maximum geological age dated for islands (e.g. [bib_ref] Effect of island geological age on the arthropod species richness of Azorean..., Borges [/bib_ref] , [bib_ref] Time, area and isolation: factors driving the diversification of Azorean arthropods, Borges [/bib_ref] , respectively. Numerous studies have evaluated some of these models for a wide variety of taxa and archipelagos, either confirming or rejecting their predictions (e.g. [bib_ref] Biogeographical determinants of lichen species diversity on islets in the West-Estonian Archipelago, Juriado [/bib_ref] - [bib_ref] Accounting for data heterogeneity in patterns of biodiversity: an application of linear..., Patiñ O [/bib_ref]. Although all these factors are known to affect island species richness, few attempts have been made to assess their comparative importance within a single evaluation (but see [bib_ref] A global model of island biogeography, Kalmar [/bib_ref] , [bib_ref] Drivers of diversity in Macaronesian spiders and the role of species extinctions, Cardoso [/bib_ref] , [bib_ref] Global diversity of island floras from a macroecological perspective, Kreft [/bib_ref]. This may be due to the fact that most predictors are often correlated and therefore it is difficult to separate their true influence on species richness through common statistical techniques [bib_ref] Island, archipelago and taxon effects: mixed models as a means of dealing..., Bunnefeld [/bib_ref] (see also [bib_ref] Escaping the trap of low sample size in island biogeography, Hortal [/bib_ref]. In addition, generalizations about the importance of the processes underlying these predictors depend on the idiosyncratic characteristics of both islands (e.g. the range of variation in area, isolation or elevation) and taxa (e.g. dispersal ability or life cycle). For example, the influence of isolation on immigration depends on the dispersal ability of the taxon, which in turn limits the probability of in situ speciation [bib_ref] Time, area and isolation: factors driving the diversification of Azorean arthropods, Borges [/bib_ref] , [bib_ref] How do different dispersal modes shape the species-area relationship? Evidence for between-group..., Aranda [/bib_ref] ,. Similarly, the influence of environmental heterogeneity (habitat or climatic diversity) on the successful establishment of species varies according to their physiological and ecological tolerances (i.e. niche breadth), eventually determining the shape of richnessenvironment relationships (see [bib_ref] Scale and species richness: towards a general, hierarchical theory of species diversity, Whittaker [/bib_ref]. Bryophytes -which encompass hornworts, liverworts and mosses -are unique among land plants because: (i) the gametophyte is the dominant phase of the life cycle comprising the leafy or thalloid plants; and (ii) the sporophyte, which consists mainly of a short-lived small ''capsule'', is always attached to and dependent on the gametophyte. Both singularities make the two generations of the life cycle to contribute significantly to the dispersal and establishment processes. In addition, contrary to seed plants and ferns, they lack complex vascular tissues and developed a poikilohydric strategy that allows them to absorb water over their whole surface by capillarity, being able to remain metabolically inactive when dry conditions exist. Furthermore, bryophytes are characterized by extremely low levels of endemism in oceanic floras (see [bib_ref] The anagenetic world of spore-producing plants, Patiño [/bib_ref] for review], which is thought to be a consequence of the high dispersal ability of the group [bib_ref] Dispersal, diversity and evolution of the Macaronesian cryptogamic floras, Vanderpoorten [/bib_ref]. Despite these interesting features, bryophytes have received relatively little attention in island biogeography studies compared to other plant groups (but see [bib_ref] How do different dispersal modes shape the species-area relationship? Evidence for between-group..., Aranda [/bib_ref] , [bib_ref] Bryophyte flora of the Canary Islands: an updated compilation of the species..., González-Mancebo [/bib_ref] - [bib_ref] Bryophyte Island Biogeography -A study in lake Manapouri, Tangney [/bib_ref]. Most of these works include only one archipelago (but see [bib_ref] Accounting for data heterogeneity in patterns of biodiversity: an application of linear..., Patiñ O [/bib_ref] , [bib_ref] How do different dispersal modes shape the species-area relationship? Evidence for between-group..., Aranda [/bib_ref] or do not consider all the above-mentioned factors, and in particular climate (but see [bib_ref] Characteristic pattern of species diversity on the Canary Islands, Steinbauer [/bib_ref]. Also, the effect of climate on large-scale species richness gradients has been occasionally analyzed in spore-dispersed plants [bib_ref] Contrasting environmental and regional effects on global pteridophyte and seed plant diversity, Kreft [/bib_ref] , being mostly studied indirectly through its correlation with latitude and altitude (e.g. [bib_ref] Species richness of vascular plants, bryophytes, and lichens along an altitudinal gradient..., Grytnes [/bib_ref] - [bib_ref] A global comparative analysis of elevational species richness patterns of ferns, Kessler [/bib_ref]. In the present study we examine the role of geographical, temporal and environmental factors on the between-island variation of bryophyte species richness in the Macaronesian Region (i.e. Azores, Madeira, Canary Islands and Cape Verde). Specifically, we evaluate four non-exclusive hypotheses under the following premises: H 1 . The Equilibrium Model of Island Biogeography (EMIB) should not significantly account for the variation in species richness of bryophytes, or its effect should be negligible. We expect that geographic isolation will not have a significant effect on immigration rates since bryophytes have the potential to disperse long distances by spores [bib_ref] Diversity, dispersal and biogeography of bryophytes (mosses), Frahm [/bib_ref] , [bib_ref] Oceanic islands are not sinks of biodiversity in spore-producing plants, Hutsemékers [/bib_ref]. In spite of some discrepancies [bib_ref] Accounting for data heterogeneity in patterns of biodiversity: an application of linear..., Patiñ O [/bib_ref] , [bib_ref] How do different dispersal modes shape the species-area relationship? Evidence for between-group..., Aranda [/bib_ref] , [bib_ref] Species richness patterns and metapopulation processes-evidence from epiphyte communities in boreo-nemoral forests, Löbel [/bib_ref] , the dispersal ability of bryophytes should in turn limit the influence of area per se because the high rescue effect from surrounding source populations would minimize species extinctions. H 2 . The General Dynamic Model of oceanic island biogeography (GDM) should not be of high relevance for bryophytes because the effect of area in species richness should be minimized with increasing dispersal ability and also because former studies suggested that time per se appears to have little support in predicting species richness in the group [bib_ref] Accounting for data heterogeneity in patterns of biodiversity: an application of linear..., Patiñ O [/bib_ref]. [formula] H 3 . [/formula] Habitat diversity (HD) should have a significant effect on species richness because bryophyte communities are known to show significant degrees of compositional turnover between different habitats [bib_ref] Characteristic pattern of species diversity on the Canary Islands, Steinbauer [/bib_ref] , [bib_ref] Bryophyte community composition and habitat specificity in the natural forests of Terceira, Gabriel [/bib_ref] - [bib_ref] Physiological Ecology: Michigan Technological University and the International Association of Bryologists, Glime [/bib_ref]. H 4 . Precipitation and temperature (CLIMATE) should have a strong effect on species richness, since sexual reproduction and photosynthesis in bryophytes are highly dependent on water availability, and optimal growth occurs with moderate temperatures. # Results Univariate regressions between dependent variables (S TOT , S M and S L ) and all the considered predictors showed similar results for mosses and liverworts . From the predictors representing the EMIB only area (A) was significantly related to moss species richness variation. In the case of GDM, both the linear and quadratic functions of time were not statistically significant for any of the groups. For the HD hypothesis, however, most variables were correlated with species richness of mosses and liverworts, being highly significant in the former group. Regarding the CLIMATE hypothesis, orographic mist layer (MistL) accounted for the highest proportion of data variability in both S M and S L . The negative effect of maximum temperature (T MAX ) on liverwort species richness was also remarkable. Temperature seasonality (T S ), although showing a lower correlation, was statistically significant for both mosses and liverworts. The subset of variables included in the best model for each hypothesis was also similar between mosses and liverworts [fig_ref] Table 2: Multiple regression results showing the best subset of predictors for each considered... [/fig_ref]. The EMIB never exceeded 22% of explained variance, being weakly or even marginally significant, while GDM was statistically significant for both groups of bryophytes, particularly in mosses. Here note that while time alone is a poor predictor of species richness , when included in a model with area (A) the two variables account for as much as 64% and 36% of data variation in moss and liverwort species richness, respectively [fig_ref] Table 2: Multiple regression results showing the best subset of predictors for each considered... [/fig_ref]. HD seems to be particularly important for mosses, although CLIMATE was the model with the highest explanatory capacity for all groups (up to 77%). Given that the above mentioned differences between mosses and liverworts cannot be discerned when considering all species together (S TOT ), henceforth we will focus on comparing the main findings for both taxonomic groups separately. Results from partial regressions including the hypotheses that seem to better explain species richness (see [fig_ref] Table 2: Multiple regression results showing the best subset of predictors for each considered... [/fig_ref] indicated that the combined 'GDM+CLIMATE' model explained most of the variation in moss and liverwort richness (87.0% and 91.1%, respectively), followed by 'HD+CLIMATE' (71.5% and 79.8%) and 'HD+GDM' (71.1% and 43.6%) . However, there were contrasting differences in the independent and shared effects of these hypotheses between both groups. In mosses, shared effects were very high in all pairwise model combinations (ranging between 48.9-60.6%), while pure effects were much lower . Hence, when combined with the other hypotheses, CLIMATE alone explained around 11.0-19.2% of S M , GDM between 10.5-18.9% and HD no more than 3.5%. In the case of liverworts, however, around half of the variability in species richness was accounted exclusively by CLIMATE, while the independent effects of both GDM and HD were weaker (up to 19% and 0.4% of explained variance, respectively). Note that pairwise comparisons between GDM and HD hypotheses showed a relatively higher contribution of the former over the latter in both taxonomic groups. The spatial analyses evidenced that the above models account for almost all spatially-structured variation in the data. Model residuals were not significantly correlated with latitude in any case except for the HD model in liverworts (Spearman r = 0.57, p = 0.01). In fact, there were only three statistically significant autocorrelation values in model residuals (in the case of GDM for mosses and of GDM and CLIMATE for liverworts; see Tables S3.1, S3.2 in File S3). Further, archipelago idiosyncrasies seem not to have affected model estimates, since SAR models were consistent with OLS regressions (File S3) and the statistical significance of the parameters for the variables included in OLS regressions remained similar. Once the spatial structure was taken into account, the predictive ability of SAR models (R 2 ) increased only slightly (compare .3 in File S3 with [fig_ref] Table 2: Multiple regression results showing the best subset of predictors for each considered... [/fig_ref]. # Discussion Although interpreting patterns of species diversity is a recurrent issue in island biogeography, the number of studies examining different factors altogether is surprisingly low (e.g. [bib_ref] A global model of island biogeography, Kalmar [/bib_ref] , [bib_ref] Drivers of diversity in Macaronesian spiders and the role of species extinctions, Cardoso [/bib_ref] , [bib_ref] Disentangling the effects of area, energy and habitat heterogeneity on boreal forest..., Honkanen [/bib_ref] , [bib_ref] Direct and indirect effects of area, energy and habitat heterogeneity on breeding..., Jonsson [/bib_ref]. For instance, Kreft et al. [bib_ref] Global diversity of island floras from a macroecological perspective, Kreft [/bib_ref] related worldwide patterns of vascular plant island species richness to geographic, topographical and climatic characteristics, filtering also by the geological origin of islands. However, these authors did not account for the different intra-archipelago relationships expected when crossing biogeographical regions [bib_ref] Island, archipelago and taxon effects: mixed models as a means of dealing..., Bunnefeld [/bib_ref] and their meta-analyses did not disentangle the combined and independent effect of each factor. Our findings suggest that climate and habitat are the most relevant factors in . Univariate regressions explaining the variation in species richness of all Macaronesian bryophytes (S TOT ), mosses (S M ) and liverworts (S L ) as a function of the predictors chosen for the Equilibrium Model of Island Biogeography (EMIB), the General Dynamic Model (GDM), the Habitat Diversity model (HD) and the Climatic Model (CLIMATE). ## All bryophytes (s tot ) Mosses explaining bryophyte richness differences on oceanic islands. These suggest that the factors operating during species establishment may be relatively more important than dispersal during the colonization process, at least in the case of the Macaronesian region. [formula] (S M ) Liverworts (S L ) R 2 F R 2 F R 2 F EMIB A (+) 0. [/formula] ## The role of climate and habitat in bryophyte diversity Our results show that the variation of bryophyte species richness between the Macaronesian islands can be relatively well predicted by climatic conditions, particularly with those favouring higher humidity. In this respect, it is well known that water availability is an important component of several key ecophysiological processes in bryophytes. One could assume that water availability is of major importance during the colonization of oceanic islands by bryophytes. Since long-distance dispersal (LDD) is mostly driven by spores (see below), water becomes essential since both sexual reproduction and spore germination depends on it. Once the populations are effectively established on an island, the availability of water at adequate growing temperatures becomes crucial to achieve a positive net photosynthetic rate over time, otherwise the plants enter dormancy upon drying (reviewed by [bib_ref] Climate responses and limits of bryophytes: comparisons and contrasts with vascular plants, Proctor [/bib_ref]. Consequently, the time of survival when an individual colony or shoot remains dry is also dependent on the periods with mild temperatures. It follows that the role of water availability (and secondarily temperature) to maintain populations and ultimately avoid local species extinctions could be the main reason to explain the relevance of the considered climatic variables in our study. Contrary to most vascular plants, bryophytes regulate water uptake mainly by capillarity and since they are usually small, their survival may be more restricted by the frequency rather than by the volume of rainfall [bib_ref] Microclimate, light adaptation and desiccation tolerance of epiphytic bryophytes in two Venezuelan..., León-Vargas [/bib_ref]. Actually, dew or cloud water deposition is often sufficient to remoisten most bryophyte species. In agreement with this idea, our results show that orographic mist layer (MistL), together with lower values of minimum precipitation (Pmin) and maximum temperature (Tmax) correlate with higher values of species richness. Such climate is typical of the Azorean archipelago and Madeira island but also, to a lesser extent, of La Gomera and La Palma in the Canaries [bib_ref] Bryophyte flora of the Canary Islands: an updated compilation of the species..., González-Mancebo [/bib_ref]. The importance of air humidity is clearly evident in (sub-) tropical rainforests [bib_ref] Exploring the effect of host tree identity on epiphyte bryophyte communities in..., Patiño [/bib_ref] , especially in canopy epiphytes (see [bib_ref] Bryophyte cover on trees as proxy for air humidity in the tropics, Karger [/bib_ref] , whose species diversity may indeed be comparable to some Azorean islands where more than 25 species can occur in plots of only 30 cm630 cm [bib_ref] Bryophyte community composition and habitat specificity in the natural forests of Terceira, Gabriel [/bib_ref]. However, the impact of climate must be broadly analyzed, taking into account the landscape conditions where the species grow [bib_ref] Can the effects of climate change on British bryophytes be distinguised from..., Bates [/bib_ref]. Our results show high correlations of habitat diversity with bryophyte diversity, providing an indirect signal for a certain degree of habitat specialization in Macaronesian bryophytes, in agreement with evidences that closely related species from this group may coexist sympatrically in separate niche spaces [bib_ref] Niche theory and practice: Bryophyte studies, Slack [/bib_ref]. This implies that the high dispersal ability of bryophytes does not lead necessarily to habitat generalism, as for other taxa with strong vagility (e.g. [bib_ref] The roles of island area per se and habitat diversity in the..., Ricklefs [/bib_ref]. Strikingly, habitat diversity seems to be as important as climate in our case, particularly for the Macaronesian moss flora. However in this study, most of the variables representing habitat diversity are eminently topographical and hence somehow correlated with island ontogeny (i.e. time) as well as with climatic predictors resulting in strong shared effects and making it difficult to disentangle their specific influence on species richness (see . In fact, the explanatory power of this factor remains similar even when we consider the number of ecological zones (EZ) as a surrogate of habitat diversity, probably because the major vegetation formations in Macaronesia are strongly structured in altitudinal belts (see File S1); the number of ecological zones is highly correlated with both sdELEV (r = 0.94) and ELEV (r = 0.92), thus being a surrogate for mesoscale climate gradients as well as for habitat diversity per se (see [bib_ref] Species richness can decrease with altitude, but not with habitat diversity, Hortal [/bib_ref]. The contrasting patterns found in liverworts and mosses could at first be related with their distinct ability to produce sexual and asexual diaspores, but no differences in the expression of several life-history traits between the two groups were detected in a suite of oceanic archipelagos [bib_ref] Baker's law and the island syndromes in bryophytes, Patiñ O [/bib_ref]. The apparently higher climatic sensitivity of liverworts compared to mosses could then be understood by their lower desiccation tolerance [bib_ref] Invited essay: New frontiers in bryology and lichenology -The nature and distribution..., Wood [/bib_ref] , especially notable in leafy liverworts due to their life-form traits [bib_ref] Ecological classifications of bryophytes and lichens, During [/bib_ref]. This is mirrored by the high sensitivity of the group to human-induced disturbances [bib_ref] Effect of forest clearcutting on subtropical bryophyte communities in waterfalls, on dripping..., Patiño [/bib_ref]. Perhaps due to this, mosses can be found in a comparatively wider range of landscapes, including grasslands and other man-made habitats, while liverworts seem to be more dependent on sheltered habitats like forests as compared with other open landscapes. In fact, liverwort richness at the Azorean native forests is higher than that of mosses above 600 m a.s.l. [bib_ref] Bryophyte community composition and habitat specificity in the natural forests of Terceira, Gabriel [/bib_ref]. Further studies are necessary to confirm such higher habitat specificity across the latitudinal gradient provided by the Macaronesian archipelagos. ## Effects of island isolation, area and time Long-distance dispersal (LDD) is a rare and stochastic event [bib_ref] Long-distance dispersal of plants, Nathan [/bib_ref] , although its prevalence over long time periods may be common [bib_ref] Is a new paradigm emerging for oceanic island biogeography, Heaney [/bib_ref]. In the case of bryophytes, LDD may occur only occasionally because spore production seems to be highly constrained due to unsuccessful sexual reproduction and even when occurring, spore release typically falls within the first tens of meters [bib_ref] The fate of the missing spores -Patterns of realized dispersal beyond the..., Lönnell [/bib_ref] (but see [bib_ref] Spore rain in relation to regional sources and beyond, Sundberg [/bib_ref]. Hence, asexual propagation is often the most frequent way of dispersion [bib_ref] Diversity, dispersal and biogeography of bryophytes (mosses), Frahm [/bib_ref]. This argument is also supported by an increasing number of molecular evidences showing that population connectivity at local or even landscape scales is hindered by dispersal limitation (e.g. [bib_ref] Macroecological patterns of genetic structure and diversity in the aquatic moss Platyhypnidium..., Hutsemékers [/bib_ref] , [bib_ref] Finescale spatial genetic structure of a liverwort (Barbilophozia attenuata) within a network..., Korpelainen [/bib_ref]. In line with this rationale, significant shifts in life-history traits towards decreased sporophyte production and increased production of specialized asexual diaspores on oceanic islands recently pointed to a global loss of LDD ability in oceanic bryophyte floras [bib_ref] Baker's law and the island syndromes in bryophytes, Patiñ O [/bib_ref]. It must be acknowledged, however, that spore production may be overlooked when fieldwork has not been sufficiently intensive (e.g. [bib_ref] Southern Scandinavian lowland populations of Rhytidium rugosum (Hedw.) Kindb. (Bryophyta, Rhytidiaceae) differ..., Hedenä S [/bib_ref]. The ability of taxa to undergo LDD via asexual propagules (usually larger than 50 mm) is still not clearly understood and, contrary to classical studies with spores [bib_ref] Experimental studies on trans-oceanic long-range dispersal of moss spores in the southern..., Van Zanten [/bib_ref] , [bib_ref] Experimental dispersal geography of neotropical liverworts, Van Zanten [/bib_ref] , little is known about the maximum distances that vegetative propagules might actually travel (cf. [bib_ref] Dispersal potential of spores and asexual propagules in the epixylic hepatic Anastrophyllum..., Pohjamo [/bib_ref]. Simulation studies have recently pointed out that above a diameter of 20 mm wind dispersal of microbes between continents becomes increasingly unlikely, and it does not occur at all for those of 60 mm diameter [bib_ref] Modelling the effect of size on the aerial dispersal of microorganisms, Wilkinson [/bib_ref]. Hence, once airborne, bryophyte spores can be virtually transported large distances by wind, so other establishment impediments such as edge colonization or gene surfing [bib_ref] Founder takes all: density-dependent processes structure biodiversity, Waters [/bib_ref] and niche specialization [bib_ref] The founder space race: a response to Waters et al, Buckley [/bib_ref] could be as limiting factors as dispersal per se. Here we are assuming that LDD is likely achieved by spores because they are resilient, microscopic and released in several millions, with some of them eventually being able to colonize an island in the very long range. This mechanism might particularly apply to archipelagos such as the Macaronesian islands [bib_ref] Oceanic islands are not sinks of biodiversity in spore-producing plants, Hutsemékers [/bib_ref] , which are relatively less isolated and exhibited a higher connectivity with the continental sources in the past, due to the presence of a higher number of emerged islands that remain today as submersed seamounts [bib_ref] A reconstruction of Palaeo-Macaronesia, with particular reference to the long-term biogeography of..., Fernández-Palacios [/bib_ref]. Our results are in line with these arguments showing that the limited contribution of the Equilibrium Model of Island Biogeography in Macaronesian bryophytes can be justified at least by the negligible effect of geographic isolation we expected. Different studies have shown similar results in bryophytes [bib_ref] Accounting for data heterogeneity in patterns of biodiversity: an application of linear..., Patiñ O [/bib_ref] , [bib_ref] Colonization of Sphagnum on land uplift islands in the Baltic Sea: time,..., Sundberg [/bib_ref] , [bib_ref] The barriers to oceanic island radiation in bryophytes: insights from the phylogeography..., Vanderpoorten [/bib_ref] , as well as in other organisms that disperse passively by spores, for which wind connectivity seems to be more important than geographic proximity between land masses [bib_ref] Wind as a long-distance dispersal vehicle in the Southern Hemisphere, Muñ Oz [/bib_ref]. By contrast, the effect of area on bryophyte species richness has been either supported or not in different studies, both in the case of islands and isolated patches on fragmented landscapes (e.g. [bib_ref] How do different dispersal modes shape the species-area relationship? Evidence for between-group..., Aranda [/bib_ref] , [bib_ref] Species richness patterns and metapopulation processes-evidence from epiphyte communities in boreo-nemoral forests, Löbel [/bib_ref] , [bib_ref] Colonization of Sphagnum on land uplift islands in the Baltic Sea: time,..., Sundberg [/bib_ref] , [bib_ref] Bryophyte species richness on insular boulder habitats: The effect of area, isolation,..., Kimmerer [/bib_ref]. Our results show that the General Dynamic Model -which is ultimately an extension of that originally proposed by the EMIB-still exerts significant predictive power over island richness after accounting for other factors, particularly in the case of mosses (cf. [bib_ref] Accounting for data heterogeneity in patterns of biodiversity: an application of linear..., Patiñ O [/bib_ref]. The variation of island area through time is not expected to affect the speciation process in this taxon because of its high dispersal potential, but both immigration and extinction are probably influenced by the changes in habitat diversity through the island's ontogeny, as predicted by this theory [bib_ref] A general dynamic theory of oceanic island biogeography, Whittaker [/bib_ref]. The few studies that have evaluated the effects of island age on whole floras of spore-dispersed plants show a comparatively lower predictive power of this variable than area and habitat diversity [bib_ref] Accounting for data heterogeneity in patterns of biodiversity: an application of linear..., Patiñ O [/bib_ref] , [bib_ref] Characteristic pattern of species diversity on the Canary Islands, Steinbauer [/bib_ref] , [bib_ref] Colonization of Sphagnum on land uplift islands in the Baltic Sea: time,..., Sundberg [/bib_ref]. At this point, however, one could argue that it is very difficult to disentangle the effect of HD and GDM hypotheses because time is implicitly accounting for the changes in island topography which, in turn, is also correlated with area. In fact, we observe strong shared effects between both hypotheses although, as we mentioned above, this could be related with the variables chosen to represent habitat diversity. Hence, what we may interpret from our results is that neither time nor area (nor topography) alone explain as much variability in the data as both factors together (around 40% or even 60% in liverworts and mosses, respectively). In spite of its importance, the precise relationship between area and species richness is still under debate. Several alternative hypotheses have been proposed to explain the importance of this relationship on extinction, colonization and speciation rates or even stochastic processes, among which its correlation with habitat diversity seems to apply in different taxa (see [bib_ref] Island species richness increases with habitat diversity, Hortal [/bib_ref] and references therein). Yet, separating the effects of these factors is statistically challenging, and comparatively fewer studies have analyzed the influence of the 'effective area' over species richness, that is, the relevance of area of suitable habitats [bib_ref] Drivers of diversity in Macaronesian spiders and the role of species extinctions, Cardoso [/bib_ref] , [bib_ref] Disentangling the effects of area, energy and habitat heterogeneity on boreal forest..., Honkanen [/bib_ref]. ## Some notes of caution Determining which predictors are the most ecologically meaningful for a particular taxon is always difficult to scrutinize by the available statistical techniques, particularly because the obtained relationships depend basically on (i) the geographic extent of the study region, (ii) the dataset (i.e. sample size) and (iii) the existence of multicollinearity among predictors. These problems are exacerbated in the case of island biogeography due to the combination of small sample sizes and different relationships between species richness and island characteristics among different archipelagos [bib_ref] Escaping the trap of low sample size in island biogeography, Hortal [/bib_ref]. However, our objective here was to choose the best combination of variables to represent each hypothesis -i.e. to account for data variation in species richness -rather than selecting the individual predictors that are most biologically important. In spite of this, we mentioned this issue and, by comparing also with analyses including the Cape Verde archipelago (File S1), we could say that at least orographic mist layer together with temperature seem to be important for these organisms (see [fig_ref] Table 2: Multiple regression results showing the best subset of predictors for each considered... [/fig_ref].1 and S1.2 in File S1). Few studies have proved experimentally the influence of mist layer because quantitative data on air humidity are often hard to obtain (e.g. [bib_ref] Exploring the effect of host tree identity on epiphyte bryophyte communities in..., Patiño [/bib_ref] , [bib_ref] Local climatic conditions and zonation of vegetation on Madeira, Sjögren [/bib_ref] , hence using typically indirect measures like bryophyte cover itself [bib_ref] Bryophyte cover on trees as proxy for air humidity in the tropics, Karger [/bib_ref]. Further investigation is however required to confirm the role of mist precipitation by incorporating an actual proxy (e.g. frequency or volume of mist precipitation), testing its independent and combined effect with annual precipitation. Obviously, widening the spatial extent of the study entails an overall higher contribution of climate -in both groups of bryophytes -due to a stronger latitudinal gradient .1, S1.2 and .1 in File S1). Nevertheless, habitat variables have a notably contribution explaining the richness of mosses .1, S1.2 in File S1), while showing a high shared variance with climate in File S1). Nonetheless, the consistency between the results obtained with ordinary least squares models and spatial autoregressive regressions indicate that differences in the spatial positioning of islands are not conditioning the main patterns found at the Macaronesian extent. # Conclusions Our findings indicate that large-scale variations in the Macaronesian bryophyte diversity are highly influenced by environmental factors but also, at least in the case of mosses, by factors related with the island ontogeny. This could be the case for other taxa with high dispersal ability. We have also shown strongly different macroecological patterns between mosses and liverworts, reinforcing the idea that not only dispersal ability, but also different ecophysiological responses of these two evolutionarily distinct lineages, are probably shaping the distribution of species diversity. Our results evidence the seeming importance of climate, in particular orographic mist layer, for liverwort diversity, while in mosses this factor has a similar or, at least, indiscernible effect to that of habitat or even the geologic ontogeny. These results point to a presumably large relevance of the establishment process on the island diversity of spore-dispersed plants. Future studies, using broader spatial extents are required to generalize these conclusions. # Methods ## Area of study The five Macaronesian archipelagos (Azores, Madeira, Selvagens, Canaries and Cape Verde) lie in the North Atlantic Ocean, covering a maximum latitudinal extension of almost 3000 km. Although all these archipelagos have a volcanic origin, the geographical characteristics of the islands vary widely in both size (ranging from 3 km 2 to more than 2000 km 2 ) and isolation (oscillating from less than 100 km to about 1800 km in distance to the nearest continent). Their maximum geological age also differs significantly among islands, from the youngest island of Pico in the Azores (less than half a million years) to the Selvagem Grande in the Selvagens archipelago that dates back to the Miocene (27 Ma). It is known, however, that a much older and interconnected ''Palaeo-Macaronesia'' existed during the Paleocene (60 Ma), most of which remains today as submersed seamounts [bib_ref] A reconstruction of Palaeo-Macaronesia, with particular reference to the long-term biogeography of..., Fernández-Palacios [/bib_ref]. There are also evident climatic differences between archipelagos along the large latitudinal gradient they form, from the temperate oceanic conditions of the Azores to the Mediterranean climate of Madeira, Selvagens and Canary Islands. The most extreme conditions for bryophyte survival occur in Cape Verde, where tropical arid climate prevails, and in the Canarian islands of Lanzarote and Fuerteventura, all showing desert affinities due to the Sub-Saharan influence. Despite such contrasting island features, there are some biotic elements shared between most archipelagos among which the evergreen laurel forests (or laurisilva) are probably the best representative example for Azores, Madeira and the Canaries. Although most of these forest areas were highly reduced after the Pleistocene glaciations and current human activity, they present the optimal habitat conditions for attaining maximum levels of bryophyte diversity (e.g. [bib_ref] Bryophyte vegetation in the Azores Islands, Sjögren [/bib_ref]. ## Data compilation We calculated total species richness (S TOT ) per island using recent checklists updated with some relevant references (see further details in [bib_ref] How do different dispersal modes shape the species-area relationship? Evidence for between-group..., Aranda [/bib_ref]. As the species checklist from Selvagens and some Cape Verde islands cannot be considered reliable [bib_ref] How do different dispersal modes shape the species-area relationship? Evidence for between-group..., Aranda [/bib_ref] , we focus our analyses on the main islands (n = 19) of Azores, Madeira and the Canaries (but see also File S1, where we show additional analyses including some Cape Verde islands that could be comparable in terms of inventories). We also run separate analyses using the species richness of the two dominant groups of mosses (S M ) and liverworts (S L ) because they normally present different ecophysiological responses [bib_ref] Physiological Ecology: Michigan Technological University and the International Association of Bryologists, Glime [/bib_ref]. In total, our database included all the 729 bryophyte species recorded in the four Macaronesian archipelagos. Out of these, 505 are mosses, 218 liverworts and 6 hornworts. Fifteen predictor variables representing geography, time, habitat diversity and climate were used to evaluate the relevance of the four hypotheses formulated above (see File S2 for further details on computation and data sources). For H 1 (i.e. EMIB), we compiled data on island area (A), distance to mainland (D M ) and distance to the closest island (D I ). We also calculated the neighbour index (N) for all the islands as proposed by Kalmar and Currie [bib_ref] A global model of island biogeography, Kalmar [/bib_ref] , to account for the combined effects of the area and distance of nearby islands. For H 2 (i.e. GDM), apart from area we obtained the time elapsed since island formation (T) for each island. For H 3 (i.e. HD), we used the number of main ecological zones in the islands (EZ) as well as three topographical surrogates: maximum elevation (ELEV), standard deviation of elevation (sdELEV) and diversity of slopes (SLOPEdiv). Finally, for H 4 (i.e. CLIMATE), we used six variables accounting for extreme, average and intraannual variation of precipitation and temperature that are of particular importance for bryophyte distribution: maximum temperature of warmest month (T MAX ), precipitation of driest quarter (P MIN ), temperature seasonality (T S ), precipitation seasonality (P S ), annual precipitation (P ANN ) and an index of horizontal precipitation as surrogate of orographic mist layer (MistL). ## Statistical analyses We used ordinary least squares (OLS) regressions to evaluate the influence of the different predictors on bryophyte species richness. Prior to the analyses, all these variables were standardized to zero mean and one standard deviation to remove the effect of different measurement scales. We first explored the univariate relationships between the considered dependent variables (S TOT , S M and S L ) and each one of the 15 predictors. We evaluated both linear and curvilinear (quadratic) relationships selecting for subsequent analyses the function that maximized the explained variance in the response variable. In the case of GDM we included both the linear (T) and quadratic (TT 2 ) functions of time as suggested in the literature [bib_ref] A general dynamic theory of oceanic island biogeography, Whittaker [/bib_ref] , [bib_ref] A general dynamic theory of oceanic island biogeography: extending the MacArthur-Wilson theory..., Whittaker [/bib_ref]. We then selected the best subset of predictors representing each hypothesis through multiple regressions, using the Akaike information criterion corrected for small sample sizes (AIC C ) to compare alternative models. Lower AIC C values indicate a compromise between higher model fit and lower complexity, so for each hypothesis we chose as best model the one with the minimum AIC C. In the case of GDM, we applied directly the ATT 2 model proposed by [bib_ref] A general dynamic theory of oceanic island biogeography, Whittaker [/bib_ref] that assumes an unimodal response of species richness to time, also accounting for the positive and monotonic influence of area. We additionally reported conventional statistics (R 2 and Ftest) for all the obtained models. To ascertain whether the combined effects of the different models selected previously contribute to increase significantly the explained variance, we made pairwise comparisons between the best models chosen for each hypothesis. To separate the single and combined effects of the different models, we used partial regression analyses. Unlike standard regression methods, this technique allows disentangling the proportion of explained variance that can be attributed exclusively to one set of factors once the effect of other sets has been controlled for, assuming that combined effects reflect the shared variance that cannot be unequivocally attributed to any of the individual sets of predictors. In order to examine whether archipelago' idiosyncrasies nonstrictly related with the abiotic island characteristics here considered may be disrupting our interpretation of species richness patterns, we followed three alternative but complementary approaches. As the low number of observations hinders the use of a qualitative variable reflecting the different archipelagos as a fixed factor, we took advantage of the strong latitudinal gradient in the location of the different Macaronesian archipelagos to examine if a spatial pattern remains in our models. To do this we correlated model residuals with latitude, also evaluating the statistical significance of Moran's I autocorrelation values using Monte Carlo methods (see [bib_ref] Spatial autocorrelation and red herrings in geographical ecology, Diniz-Filho [/bib_ref]. We also run simultaneous autoregressive regressions (SAR) taking explicitly into account the spatial coordinates in the analysis as a connectivity matrix [bib_ref] Coefficient shifts in geographical ecology: an empirical evaluation of spatial and non-spatial..., Bini [/bib_ref]. All analyses were performed in SAM v4.0 [bib_ref] SAM: a comprehensive application for Spatial Analysis in Macroecology, Rangel [/bib_ref] (available at www.ecoevol.ufg.br/sam/). ## Supporting information File S1 Additional analyses incorporating the Cape Verde archipelago. [fig] File: S2 Abiotic characteristics of the Macaronesian islands considered, and species richness of all bryophytes (S TOT ), mosses (S M ) and liverworts (S L ). (DOCX) S3 Spatial autocorrelation of model residuals and spatial autoregressive regressions (SAR) alternative to typical ordinary least squares (OLS) regression models. (DOCX) Author Contributions Conceived and designed the experiments: SCA RG PAVB AMCS EBDA JP JH JML. Performed the experiments: SCA RG. Analyzed the data: SCA JML. Contributed reagents/materials/analysis tools: SCA RG EBA JP. Contributed to the writing of the manuscript: SCA RG PAVB AMCS EBDA JP JH JML. [/fig] [table] Table 2: Multiple regression results showing the best subset of predictors for each considered model (EMIB, GDM, HD and CLIMATE) to explain the between-island variation in species richness of Macaronesian bryophytes.The best subset of variables that were chosen using the lowest sample size-corrected Akaike information criterion (AIC C ) is shown in brackets. Adjusted R 2 values and its statistical significance according to the F-test are also shown. Model acronyms and variable codes as inTable 1. doi:10.1371/journal.pone.0101786.t002 [/table]

Budesonide orodispersible tablets for induction of remission in patients with active eosinophilic oesophagitis: A 6‐week open‐label trial of the EOS‐2 Programme # Introduction Eosinophilic oesophagitis (EoE) is a chronic, immune-mediated, organ-restricted disease, characterized by symptoms of oesophageal dysfunction and an eosinophil-predominant inflammation of the oesophagus. [bib_ref] Guidelines on eosinophilic esophagitis: evidencebased statements and recommendations for diagnosis and management..., Lucendo [/bib_ref] Over the last 2 decades, the incidence and prevalence of EoE have constantly increased, especially in Western countries. [bib_ref] Systematic review with meta-analysis: the growing incidence and prevalence of eosinophilic oesophagitis..., Navarro [/bib_ref] EoE is now regarded as the most common cause of dysphagia and bolus impaction [bib_ref] The clinical predictors of aetiology and complications among 173 patients presenting to..., Sengupta [/bib_ref] [bib_ref] Food impaction: etiology over 35 years and association with eosinophilic esophagitis, Lenz [/bib_ref] and the second leading cause of chronic oesophagitis after gastro-oesophageal reflux disease (GORD). [bib_ref] Epidemiology and natural history of eosinophilic esophagitis, Dellon [/bib_ref] In adult EoE patients, the predominant symptoms are chronic dysphagia, food impaction and chest pain. [bib_ref] Guidelines on eosinophilic esophagitis: evidencebased statements and recommendations for diagnosis and management..., Lucendo [/bib_ref] EoE is a chronicprogressive disease and, if left untreated, is usually associated with persistence of symptoms and inflammation. [bib_ref] Natural history of primary eosinophilic esophagitis: a follow-up of 30 adult patients..., Straumann [/bib_ref] In addition, ongoing eosinophilic inflammation may lead to oesophageal remodeling resulting in fibrosis with possible stricture formation and functional damage in a large proportion of patients. [bib_ref] Delay in diagnosis of eosinophilic esophagitis increases risk for stricture formation in..., Schoepfer [/bib_ref] [bib_ref] The natural course of eosinophilic esophagitis and long-term consequences of undiagnosed disease..., Warners [/bib_ref] [bib_ref] Natural history of eosinophilic esophagitis: a systematic review of epidemiology and disease..., Shaheen [/bib_ref] Health-related quality of life (HRQoL) is substantially impaired in EoE patients by causing emotional distress and restricting social activities. [bib_ref] Systematic review: health-related quality of life in children and adults with eosinophilic..., Lucendo [/bib_ref] Therefore, treatment of active EoE is mandatory. [bib_ref] Guidelines on eosinophilic esophagitis: evidencebased statements and recommendations for diagnosis and management..., Lucendo [/bib_ref] [bib_ref] AGA Institute and the Joint Task Force on Allergy-Immunology Practice Parameters Clinical..., Hirano [/bib_ref] Swallowed topical corticosteroids (STC) are an established firstline pharmacologic treatment option for patients with EoE. [bib_ref] Guidelines on eosinophilic esophagitis: evidencebased statements and recommendations for diagnosis and management..., Lucendo [/bib_ref] [bib_ref] AGA Institute and the Joint Task Force on Allergy-Immunology Practice Parameters Clinical..., Hirano [/bib_ref] In the absence of approved formulations for EoE therapy, STCs (mainly fluticasone or budesonide), originally developed for airway administration in patients with asthma or hospital pharmacy made viscous solutions, have been used in several EoE trials which confirmed the rate of clinico-histological remission was 58% compared to 0% with placebo. The corresponding figures for histological remission and clinical remission after 6 weeks of treatment were 93% versus 0% and 58% versus 14%, respectively. [bib_ref] Efficacy of budesonide orodispersible tablets as induction therapy for eosinophilic esophagitis in..., Lucendo [/bib_ref] Here, we report the results of an open-label induction of remission treatment in a large prospective cohort of adult patients with active EOE, which served as a feeding arm for the further double-blind maintenance treatment (EOS-2) within the phase 3 programme. [bib_ref] Orodispersible budesonide tablets for the treatment of eosinophilic esophagitis -a review of..., Miehlke [/bib_ref] # Methods ## Study design and conduct The EOS-2 study was a randomized, double-blind, placebo- ## Therapy At baseline and at weeks 2 and 4, eligible patients received study medication for the next period. The orodispersible tablet was administered twice daily and was placed on the tip of the tongue and pressed gently against the hard palate until it had completely disintegrated by contact with saliva, whose production was stimulated by the slight effervescence of the study medication. [bib_ref] Orodispersible budesonide tablets for the treatment of eosinophilic esophagitis -a review of..., Miehlke [/bib_ref] ## Safety and tolerability Physical examinations were performed during screening and at week 6. Vital signs, concomitant medications and adverse events were recorded, and general laboratory tests and urinalysis were performed. Serum morning cortisol (8-9 AM) levels were measured at baseline and week 6. Tolerability was classified independently by the patient and the investigator at week 6. ## Study endpoints The ## Statistical analyses Descriptive statistics were used to summarize data, including incidences of adverse events. Analyses were performed using SAS ® # Results ## Patient enrolment and baseline characteristics As this open-label arm was used as a second feeding arm for the double-blind maintenance phase of the trial, which was primarily fed by patients rolling over from the EOS-1 trial after having brought ## Clinical efficacy Clinico and , indicating that BOT provides an optimal oesophageal targeting. Moreover, the peak eosinophil count was dramatically decreased by 89.0% from baseline confirming that BOT 1 mg BID was able to induce remission even in severely inflamed cases (see The modified EREFS total score significantly decreased from baseline to week 6 by −3 points, which was mainly driven by the inflammatory subscore . Deep endoscopic remission was achieved in 53.6% of patients, which was in line with the endoscopist's overall global assessment of "no endoscopic signs of EoE' (55.8%; . HRQoL as measured by modified SHS and EoE-QoL-A scores was moderately impaired at baseline . HRQoL improved significantly from baseline to week 6 (LOCF) based upon the disease specific EoE-QoL-A questionnaire and all its subscores. This was in line with the generic modSHS instrument showing significant improvement in mean scores from baseline to week 6 for all four dimensions. ## Safety and tolerability Overall, BOT 1 mg BID was well tolerated in this study. A total of 60 patients (33.1%) experienced 95 adverse drug reactions ( inflammation is undoubtedly the key driver for fibrosis and stricture formation and most likely also increases the risk of future food impaction. [bib_ref] Delay in diagnosis of eosinophilic esophagitis increases risk for stricture formation in..., Schoepfer [/bib_ref] [bib_ref] The natural course of eosinophilic esophagitis and long-term consequences of undiagnosed disease..., Warners [/bib_ref] [bib_ref] Natural history of eosinophilic esophagitis: a systematic review of epidemiology and disease..., Shaheen [/bib_ref] The remission rates observed in this trial, both for the composite endpoint as well as for its two components, were highly consistent with those of the previous phase 2 and 3 trials of BOT in adults with active EoE. [bib_ref] A randomised, double-blind trial comparing budesonide formulations and dosages for short-term treatment..., Miehlke [/bib_ref] remission, defined as '0' eos/hpf, was observed in 84.5% of patients in this trial, which was consistent with the rate of 89% in the previous double-blind phase 3 trial. 14 As shown in the previous double-blind phase 2 and phase 3 trials, BOT 1 mg BID had similar anti-inflammatory effects in the entire oesophagus, independently of severity, localization or extent of inflammation, again underlining an optimal oesophageal targeting. [bib_ref] A randomised, double-blind trial comparing budesonide formulations and dosages for short-term treatment..., Miehlke [/bib_ref] [bib_ref] Efficacy of budesonide orodispersible tablets as induction therapy for eosinophilic esophagitis in..., Lucendo [/bib_ref] The consistently high efficacy of BOT may be explained by the unique way of delivery of budesonide to the oesophagus. Once BOT is placed on the tongue, it stimulates the production of saliva via its Note: All intra--group comparisons were performed using 2-sided, one-sample t-test, except for comparison of peak eos count, for which a 2-sided Wilcoxon signed-rank test was used. concluded that BOT currently appears to be the most effective drug therapy for EoE in adults. As in the previous double-blind phase 3 trial, [bib_ref] Efficacy of budesonide orodispersible tablets as induction therapy for eosinophilic esophagitis in..., Lucendo [/bib_ref] On endoscopy, treatment with BOT 1 mg BID resulted in a significant improvement of the total modified EREFS mainly attributable to the 'inflammatory signs' subscore. As stated above, the EREFS fibrosis subscore was low at baseline and did not change significantly after 6 weeks. Notably, deep endoscopic remission (defined as modified EREFS subscores: fixed rings = '0' or '1', exudates = '0', furrows = '0' and edema = '0') was achieved in more than half of the patients. We In summary, BOT 1 mg BID is a highly effective and safe therapy for induction of disease remission in adult patients with active EoE. ## Acknowledgement This study was funded by Dr. Falk Pharma GmbH, Freiburg, Germany. ## Conflict of interest Stephan Miehlke reports receiving consulting fees from Celgene, Dr. [fig] version 9. 3: SAS Institute Inc.) according to the intention-to-treat principle. Missing data at week 6 were replaced using the lastobservation-carried-forward (LOCF) method. Adverse events were classified by using the Medical Dictionary for Regulatory Activities, 24 version 19. [/fig] [fig] Figure 3: where patients with baseline peak eosinophilic counts of even more than '1.000' eos/mm 2 hpf achieved deep histological remission with '0' eos/hpf after 6-week BOT 1 mg BID treatment.Based on the PatGA, 84.5% of patients reported a drop from ≥4 points prior treatment to ≤2 points at week 6 (LOCF) on the NRS related to the overall assessment of their EoE severity (Figure 4a), with an absolute mean (SD) change of −4 (2.1) points. This was in line with the analogue assessment on EoE by the physician showing an absolute mean (SD) change of −5 (2.2) points from baseline to week 6 (LOCF) as well as with the course of clinical remission defined by symptom resolution (NRS ≤2 for both dysphagia and odynophagia, or by EEsAI-PRO ≤20, respectively (Figure 4b,c), which all showed the quick clinical response to BOT 1 mg BID. Clinical remission, defined as a score of EEsAI-PRO ≤20, was observed in 49.2% of patients (Table 2), with a significant absolute mean (SD) change of −29 (21.4) points. [/fig] [fig] DISCUSSION: This prospective open-label multicentre trial confirmed that a 6-week treatment with BOT 1 mg BID is highly effective and safe for in-duction of clinical and histological remission in a large adult patient population with active EoE. The composite endpoint used in this study is regarded as the major therapeutic target of any induction therapy of EoE and is therefore recommended by recent European and American guidelines. 1,11 Histological remission is regarded particularly important since chronic eosinophilPRO (score 0-100), mean (SD); lower values indicate less disease activity 52 (17) Patient's global assessment of EoE activity (NRS 0-10), mean (SD); lower values indicate less disease activity 6 (1.5) Physician's global assessment of EoE activity (NRS 0-10), mean (SD); lower values indicate less disease activity 6 (1. [/fig] [fig] F I G U R E 3: Individual pre-and post-treatment (with BOT 1 mg BID) peak eos/mm 2 hpf counts and median group values with IQR. BID, twice daily; BOT, budesonide orodispersible tablet; eos, eosinophils; hpf, high power field (400x); IQR, interquartile range; [/fig] [fig] a: Deep endoscopic remission (modified EREFS subscores: fixed rings = '0' or '1', exudtes = '0', furrows = '0', nd edem = '0').b Deep clinicl remission (symptoms severity of '0' points on ech 0-10 NRS for dysphgi nd odynophgi, respectively on ech dy in the week prior to week 6. effervescence chrcteristics over severl minutes. During this period, BOT completely disintegrtes leding to slow relese of budesonide into the sliv, which is continuously swllowed in smll volumes.15 The mucodhesive properties of sliv my led to n enhnced exposure nd prolonged contct time of budesonide to the oesophgel mucos.The prticulr mode of drug delivery with BOT might explin the high histologicl remissions rtes in contrst to other corticosteroids suspensions which re usully ingested vi single swllow of reltively lrge volume. For exmple, in the recent ORBIT tril, phse 3 rndomized plcebo-controlled tril including 318 dolescent nd dult EoE ptients, 12-week tretment with 2 mg BID of budesonide orl suspension (10 ml t concentrtion of 0.2 mg/ml), specificlly developed for tretment of EoE, resulted in histologicl remission rte (≤6 eos/hpf) of 53.1%.25 Obviously, comprison of such trils conducted in different ptient popultions re not permitted nd hed-to-hed comprisons mong different topicl corticosteroids re lcking. However bsed on the vilble evidence, the ltest network met-nlysis26 [/fig]

Estimation of the solubility parameters of model plant surfaces and agrochemicals: a valuable tool for understanding plant surface interactions Background: Most aerial plant parts are covered with a hydrophobic lipid-rich cuticle, which is the interface between the plant organs and the surrounding environment. Plant surfaces may have a high degree of hydrophobicity because of the combined effects of surface chemistry and roughness. The physical and chemical complexity of the plant cuticle limits the development of models that explain its internal structure and interactions with surface-applied agrochemicals. In this article we introduce a thermodynamic method for estimating the solubilities of model plant surface constituents and relating them to the effects of agrochemicals. Results: Following the van Krevelen and Hoftyzer method, we calculated the solubility parameters of three model plant species and eight compounds that differ in hydrophobicity and polarity. In addition, intact tissues were examined by scanning electron microscopy and the surface free energy, polarity, solubility parameter and work of adhesion of each were calculated from contact angle measurements of three liquids with different polarities. By comparing the affinities between plant surface constituents and agrochemicals derived from (a) theoretical calculations and (b) contact angle measurements we were able to distinguish the physical effect of surface roughness from the effect of the chemical nature of the epicuticular waxes. A solubility parameter model for plant surfaces is proposed on the basis of an increasing gradient from the cuticular surface towards the underlying cell wall. Conclusions: The procedure enabled us to predict the interactions among agrochemicals, plant surfaces, and cuticular and cell wall components, and promises to be a useful tool for improving our understanding of biological surface interactions.Plant surfaces play a major role in protection against multiple potential biotic and abiotic stress factors[1]. To adapt to these multiple functions, the plant epidermis has developed various characteristics, including specialised cell types such as trichomes or stomata[2]. Epidermal cells are surrounded by a cell wall, which plays a crucial structural and physiological role in plant development and survival[3].Differentiation and maintenance of the epidermis are essential for plant growth and survival and require continuous cross-talk between epidermal cells and their immediate environment[2]. Epidermal cells also provide mechanical support by adhering strongly to each other via a strengthened cell wall, which is usually noticeably thicker on the external surface. In addition to the asymmetrical deposition of cell wall material, epidermal cells secrete a lipid-rich cuticle specifically into the thickened external cell wall matrix [bib_ref] Epidermis: the formation and functions of a fundamental plant tissue, Javelle [/bib_ref]. Therefore, the cuticle may be considered a cutinised cell wall, emphasizing the heterogeneous nature of this layer and its interconnection with the cell wall beneath [bib_ref] The biophysical design of plant cuticles: an overview, Domínguez [/bib_ref]. The main protective role of the cuticle is related to the prevention of uncontrolled exchange of water and gases between the plant and the surrounding environment [bib_ref] Uptake of hydrophilic solutes through plant leaves: current state of knowledge and..., Fernández [/bib_ref]. The functional relevance of the cuticle to plant growth and survival is evidenced by the significant commitment of epidermal cells to cuticle production [bib_ref] Sealing plant surfaces: cuticular wax formation by epidermal cells, Samuels [/bib_ref]. The cuticle is made of a bio-polymer matrix, waxes that are deposited on to (epicuticular) or intruded into (intracuticular) this matrix, and variable amounts of polysaccharides and phenolics [bib_ref] The biophysical design of plant cuticles: an overview, Domínguez [/bib_ref] [bib_ref] Building lipid barriers: biosynthesis of cutin and suberin, Pollard [/bib_ref]. It is an asymmetric membrane [bib_ref] Leaf cuticles behave as asymmetric membranes. Evidence from the measurement of diffusion..., Tyree [/bib_ref] generally comprising three distinct layers from the outer to the inner side of the organ, namely: (i) the epicuticular wax layer, (ii) the "cuticle proper" containing waxes and cutin and/or cutan, and (iii) the "cuticular layer" composed of cutin and/or cutan and a high polysaccharide content [bib_ref] The fine structure of the plant cuticle, Jeffree [/bib_ref]. Waxes commonly constitute 20 to 60% of the cuticle mass and are complex mixtures of straight chain aliphatics [bib_ref] Sealing plant surfaces: cuticular wax formation by epidermal cells, Samuels [/bib_ref]. Wax composition and structure can vary among different species, organs, states of development, and environmental and stress conditions during growth [bib_ref] The impact of water deficiency on leaf cuticle lipids of Arabidopsis, Kosma [/bib_ref] [bib_ref] The fruit cuticles of wild tomato species exhibit architectural and chemical diversity,..., Yeats [/bib_ref]. The mechanisms of epicuticular wax formation and regeneration have been assessed in some studies [bib_ref] Ultrastructure and recrystallization of plant epicuticular waxes, Jeffree [/bib_ref] and it has been proposed that cuticular transpiration is the driving force behind wax movement through the cuticle [bib_ref] Movement and regeneration of epicuticular wax through plant cuticles, Neinhuis [/bib_ref] [bib_ref] Self assembly of epicuticular waxes on living plant surfaces imaged by atomic..., Koch [/bib_ref]. The cuticle matrix is commonly made of cutin, which is a biopolymer formed by a network of inter-esterified, hydroxyl-and hydroxy-epoxy C [bib_ref] Linear and branched poly(ω-hydroxyacid) esters in plant cutins, Graça [/bib_ref] and/or C 18 fatty acids [bib_ref] Biopolyester membranes of plants: cutin and suberin, Kolattukudy [/bib_ref]. At least six different types of cuticular ultrastructures have been identified by transmission electron microscopy (TEM) [bib_ref] The fine structure of the plant cuticle, Jeffree [/bib_ref] , but their relationship to cutin monomer composition remains unclear [bib_ref] Building lipid barriers: biosynthesis of cutin and suberin, Pollard [/bib_ref] [bib_ref] Linear and branched poly(ω-hydroxyacid) esters in plant cutins, Graça [/bib_ref]. The formation of cutinsomes, which are spherical nanoparticles resulting from the self-assembly of cutin hydroxyacid monomers in a polar environment, has been demonstrated; cutinsomes have been proposed as building units of the bio-polyester cutin [bib_ref] Structural characterization of polyhydroxy fatty acid nanoparticles related to plant lipid biopolyesters, Heredia-Guerrero [/bib_ref]. While cutin is depolymerised and solubilised upon saponification, cuticles from some species contain a non-saponifiable and non-extractable polymer known as cutan, which yields a characteristic series of long chain n-alkenes and n-alkanes upon flash pyrolysis [bib_ref] Characterization and biosynthesis of non-degradable polymers in plant cuticles, Villena [/bib_ref]. Cutin has been found to be the only polymer present in the cuticles of many fruits and leaves of Solanaceae and Citrus species [bib_ref] The fine structure of the plant cuticle, Jeffree [/bib_ref] , while different proportions of cutin and cutan have been determined in cuticular membranes extracted from leaves [bib_ref] Characterization and biosynthesis of non-degradable polymers in plant cuticles, Villena [/bib_ref] and fruits such as peppers, apples or peaches [bib_ref] Spectroscopic characterization of aliphatic moieties in four plant cuticles, Johnson [/bib_ref] [bib_ref] New insights into the properties of pubescent surfaces: peach fruit as model, Fernández [/bib_ref]. Major differences in surface topography have been observed in different species and organs, but three hierarchical levels of structuring may occur in association with: (i) the general shape of epidermal cells, (ii) cuticular folds, and (iii) epicuticular wax crystals [bib_ref] Quantitative assessment to the structural basis of water repellency in natural and..., Wagner [/bib_ref]. For example, the presence of papillae [bib_ref] Superhydrophobicity in perfection: the outstanding properties of the lotus leaf, Ensikat [/bib_ref] or trichomes [bib_ref] New insights into the properties of pubescent surfaces: peach fruit as model, Fernández [/bib_ref] can have a major effect on surface topography and wettability at the microscale level. Also, increased surface roughness and surface hydrophobicity have been reported owing to the occurrence of nano-scale structures provided by epicuticular wax crystals [bib_ref] Superhydrophobicity in perfection: the outstanding properties of the lotus leaf, Ensikat [/bib_ref] [bib_ref] Natural and biomimetic artificial surfaces for superhydrophobicity, self-cleaning, low adhesion, and drag..., Bhushan [/bib_ref]. Different degrees of wettability of leaves from various species have been reported by measuring water contact angles (e.g., [bib_ref] Quantitative assessment to the structural basis of water repellency in natural and..., Wagner [/bib_ref] [bib_ref] The effects of superficial wax on leaf wettability, Holloway [/bib_ref] [bib_ref] Pattern of leaf surface wetness in some important medicinal and aromatic plants..., Pandey [/bib_ref] [bib_ref] Leaf wettability decreases along an extreme altitudinal gradient, Aryal [/bib_ref]. In addition, phyllosphere-related factors such as the deposition of aerosols or microorganisms can lead to plant surface heterogeneity [bib_ref] Stomatal penetration by aqueous solutions -an update involving leaf surface particles, Burkhardt [/bib_ref] , especially in urban or polluted habitats [bib_ref] Tree leaf wettability as passive bio-indicator of urban habitat quality, Kardel [/bib_ref]. However, non-wettable surfaces have been observed to accumulate particles more slowly than wettable ones [bib_ref] Seasonal changes of leaf surface contamination in beech, oak, and ginkgo in..., Neinhuis [/bib_ref]. Recently Fernández et al. [bib_ref] New insights into the properties of pubescent surfaces: peach fruit as model, Fernández [/bib_ref] estimated the surface free energy, polarity and work of adhesion of a model pubescent surface and proposed the implementation of membrane science approaches to exploring the physical-chemical properties of plant surfaces. It has been suggested that the cuticle acts as a "solution-diffusion" membrane for the diffusion of some solvents and solutes [bib_ref] Cuticular water permeability and its physiological significance, Kerstiens [/bib_ref] [bib_ref] Transport of lipophilic non-electrolytes across the cuticle, Riederer [/bib_ref]. To analyse the permeability of the plant cuticle to solutes and solvents, both the solubility and diffusivity of the compounds must be taken into consideration. While diffusivity is a kinetic parameter associated with the molecular size of a compound in relation to the structure of the matrix, solubility is a thermodynamic parameter that indicates the affinity of a given chemical for the cuticle. Therefore, and as a preliminary step towards the evaluation of plant cuticle permeability, we have analysed for the first time the solubility of model plant surfaces and chemical constituents in relation to agrochemicals of commercial significance, following a thermodynamic approach. Prediction of solubility parameters is commonly used, for example, in the design and fabrication of polymeric membranes [bib_ref] Study on surface modification by surface-modifying macromolecules and its applications in membrane..., Khayet [/bib_ref] [bib_ref] Preferential surface segregation of homopolymer and copolymer blend films, Khayet [/bib_ref] , in the coating industryand also in pharmacology [bib_ref] The use of solubility parameters in pharmaceutical dosage form design, Hancock [/bib_ref]. However, with the exception of the human skin [bib_ref] Three dimensional solubility parameters and their use in characterising the permeation of..., Gröning [/bib_ref] [bib_ref] Influence of membrane-solvent-solute interactions on solute permeation in skin, Dias [/bib_ref] , this procedure has not so far been applied to estimating the properties of biological surfaces. As model plant surfaces, peach and pepper fruits were selected since they contain alkanes as major wax constituents but have significantly different surface topographies. Juvenile Eucalyptus globulus leaves, which are covered with a dense layer of nano-tubes and contain β-diketones as dominant waxes, were also evaluated for comparison. For model plant surfaces, cuticular constituents and agrochemicals, the following hypotheses were tested: (i) is it possible to predict the solubility of plant surface constituents and the affinity of agrochemicals for plant surfaces? and (ii) can solubility parameters be used to estimate the properties of the plant cuticle? # Materials and methods # Plant material The plant materials analysed correspond to intact, undamaged mature peaches (Prunus persica (L.) Batsch. cv. 'Calrico'), red bell peppers (Capsicum annum L. cv. 'Genil') and juvenile Eucalyptus leaves (Eucalyptus globulus Labill. ssp. globulus). ## Epicuticular waxes, cutin monomers and cell wall polysaccharides The properties of the major wax constituents present in Eucalyptus leaves, bell peppers and peach fruits were used for calculating the solubility parameters [fig_ref] Figure 1: Molecular structures of the cuticular constituents evaluated [/fig_ref] , [fig_ref] Table 1: Chemical formula and molar volume of the dominant epicuticular waxes extracted from... [/fig_ref]. Alkanes are the dominant class of compounds covering the surface of the peach fruits analysed [bib_ref] New insights into the properties of pubescent surfaces: peach fruit as model, Fernández [/bib_ref]. Alkanes are also the dominant class of wax compounds extracted from pepper fruits, followed by triterpenoids such as αor β-amyrin [bib_ref] Composition of the surface waxes from bell pepper and eggplant, Bauer [/bib_ref] [bib_ref] Characterization of physiological and biochemical factors associated with postharvest water loss in..., Kissinger [/bib_ref] [bib_ref] Fruit cuticle lipid composition and fruit postharvest water-loss in an advanced backcross..., Parsons [/bib_ref]. Beta-diketones are the dominant class of wax compounds in juvenile Eucalyptus leaves, but n-nonacosane, heptadecan-2-one and n-hexacosanal are also present in significant concentrations [bib_ref] Variation in leaf waxes of the Tasmanian Eucalyptus species -I. Subgenus Symphyomyrtus, Li [/bib_ref] [bib_ref] Epicuticular wax of juvenile Eucalyptus leaves and headspace analysis of leaf volatiles, Wirthensohn [/bib_ref] [bib_ref] Genetic resistance of Eucalyptus globulus to autumn gum moth defoliation and the..., Jones [/bib_ref] [bib_ref] Epicuticular waxes and plant primary metabolites on the surfaces of juvenile Eucalyptus..., Steinbauer [/bib_ref]. To estimate the solubility parameter range of the cuticle matrix, calculations were carried out with model cutin monomers, which have commonly been identified in plant cuticle monomer analyses [bib_ref] Polyesters in higher plants, Kolattukudy [/bib_ref] [bib_ref] Cutin from plants, Kolattukudy [/bib_ref]. The selected ω-hydroxy-fatty acids are: 16-hydroxy-hexanodecanoic acid, 10,16-dihydroxy-hexanodecanoic acid, 9,10-epoxy-18-hydroxyoctadecanoic acid, and 9,10,18-trihydoxy-octadecanoic acid [fig_ref] Table 1: Chemical formula and molar volume of the dominant epicuticular waxes extracted from... [/fig_ref]. Maximal and minimal solubility parameter values were estimated per monomer according to the potential formation of ester bonds. The solubility parameter of cellulose, a biopolymer formed from unbranched, unsubstituted (1,4)-β-D-glucan chains [bib_ref] Heterogeneity in the chemistry, structure and function of plant cell walls, Burton [/bib_ref] , was evaluated by estimating the properties of the D-glucose monomer. The solubility parameter range of pectins was assessed by analysing the structure of homogalacturonans based on α-1-4 linked, D-galacturonic acid [fig_ref] Table 2: Characteristics of the chemicals used for estimation of solubility parameters [/fig_ref] ; [bib_ref] Heterogeneity in the chemistry, structure and function of plant cell walls, Burton [/bib_ref] [bib_ref] The composition and structure of plant primary cell walls, O&apos;neill [/bib_ref]. ## Chemicals Several compounds with different properties and degrees of complexity were selected for calculation of solubility parameters, [fig_ref] Table 2: Characteristics of the chemicals used for estimation of solubility parameters [/fig_ref] [fig_ref] Table 2: Characteristics of the chemicals used for estimation of solubility parameters [/fig_ref]. ## Microscopy Gold-sputtered intact Eucalyptus adaxial leaves, peach and pepper fruit surfaces were examined with a Hitachi S-3400 N (Tokyo, Japan) and a Philips XL30 (Eindhoven, The Netherlands) scanning electron microscope (SEM). For TEM observations of Eucalyptus leaf tissue, approximately 1 mm 2 sections were cut with a scalpel and fixed in 2% paraformaldehyde plus 2% glutaraldehyde (both from Electron Microscopy Sciences (EMS), Hatfield, USA) for 6 h in ice-cold phosphate buffer (pH 7.2). Samples were subsequently washed five times in phosphate buffer, kept at 4°C overnight, and post-fixed with 1% osmium tetraoxide (TAAB Laboratories, Berkshire, UK) and 1.5% potassium ferrocyanide (Sigma-Aldrich, Munich, Germany) in distilled water ## Contact angle measurements and prediction of solubility parameters Advancing contact angles of drops of double-distilled water, glycerol and diiodomethane (both 99% purity, Sigma-Aldrich) were measured at 25°C using a CAM 200 contact angle meter (KSV Instruments Ltd., Helsinki, Finland). Contact angles were measured on intact Eucalyptus adaxial leaf, peach and pepper fruit surfaces (30 repetitions). The plant surfaces analysed were collected from fruits and leaves previously observed by SEM. No materials that could affect contact angle measurements (e.g., salt deposits or microorganisms) were found to be deposited on them. Two μL drops of each liquid were deposited on to the plant surfaces with a manual dosing system holding a 3 mL syringe (0.5 mm diameter needle). Side view images of the drops were captured at a rate of 10 frames s -1 . Contact angles were automatically calculated by fitting the captured drop shape to the one calculated from the Young-Laplace equation. For the three plant surfaces evaluated, the total surface free energy, including its three components (i.e. the Lifshitz-van der Waals (LW), acid (+) and base (−) components), was calculated in addition to the surface polarity and work of adhesion [bib_ref] New insights into the properties of pubescent surfaces: peach fruit as model, Fernández [/bib_ref]. ## Prediction of solubility parameters The solubility parameter of each plant surface analysed, δ θ , was calculated from the following equation [bib_ref] Surface modification of polyvinylidene fluoride pervaporation membranes, Khayet [/bib_ref] : [formula] δ θ ¼ e c ð Þ 1 2 =ð1Þ [/formula] where e c (MJ m -3 ) is the cohesive energy density, which is related to the surface free energy, γ s , (mJ m -2 ) as follows: [formula] e c ¼ γ s 0:75 2 3 =ð2Þ [/formula] The solubility parameter of a material can be calculated from either the cohesive energy (Eqn. 1) or the molar attraction constant, F ((MJ/m 3 ) 1/2 mol -1 ), as: [formula] δ ¼ F υð3Þ [/formula] where v is the molar volume (cm 3 mol -1 ) of the molecule. The solubility parameter has three components taking into account the interactions due to dispersion forces (δ d ), polar forces (δ p ) and hydrogen (H)-bonding (δ h ), and it is expressed as: [formula] δ ¼ ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffi δ 2 d þ δ 2 p þ δ 2 h qð4Þ [/formula] According to van Krevelen and Hoftyzer, the solubility parameter components can be predicted from group contributions, using the following equations: [formula] δ d ¼ X F di v ð5Þ δ p ¼ ffiffiffiffiffiffiffiffiffiffiffiffi ffi X F 2 pi q v ð6Þ δ h ¼ ffiffiffiffiffiffiffiffiffiffiffiffiffi X E hi v sð7Þ [/formula] where F di and F pi are the molar attraction constants of the dispersion and polar components, respectively, E hi is the H-bonding energy and v is the molar volume. The contributions of the functional groups present in the chemicals and plant structural compounds analysed to the solubility parameter components are shown in . From the solubility parameter components, the total solubility parameter (δ) can be calculated from , and is hereafter named δ m for agrochemicals, δ wax for epicuticular waxes and δ nm for cutin and polysaccharide monomers. Finally, to evaluate the affinity of a polymer for a solventor the affinity of an agrochemical for a given plant surface, the following equation was used: [formula] Δδ wax¼ ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffi δ m À δ wax ð Þ 2 qð8Þ [/formula] Moreover, to study the affinities of the agrochemicals for plant surfaces as derived from the solubility parameter calculated from contact angle measurements (δ θ ), the following equation was applied: [formula] Δδ θ¼ ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi δ m À δ θ ð Þ 2 qð9Þ [/formula] The results from Equations 8 and 9 imply that the lower the values of Δδ wax and Δδ θ , the higher the affinity between agrochemical and plant surface. # Results ## Surface topography and hydrophobicity The contact angles (in°) of water, glycerol and diiodomethane with plant surfaces are (mean ± standard deviation): 142.6 ± 6.7, 136.5 ± 11.2 and 84.0 ± 7.0 for Eucalyptus leaves; 83.4 ± 4.7, 68.6 ± 9.2 and 60.8 ± 6.2 for pepper fruits; and 134.2 ± 7.0, 130.9 ± 7.0 and 55.7 ± 3.9 for peach fruits (see [fig_ref] Figure 3: Contact angle measurements on intact, adaxial Eucalyptus [/fig_ref] as an illustration of the measurements). The topographies of the plant materials analysed are shown in [fig_ref] Figure 4: Scanning electron micrographs of intact plant surfaces [/fig_ref]. The adaxial surface of juvenile Eucalyptus leaves is densely covered with a network of wax nano-tubes, which can be clearly identified as such at higher magnifications [fig_ref] Figure 4: Scanning electron micrographs of intact plant surfaces [/fig_ref] and G). In contrast, the pepper fruit surface is covered with a pattern of epidermal cells [fig_ref] Figure 4: Scanning electron micrographs of intact plant surfaces [/fig_ref] and E) and epicuticular waxes with no clear structure, yielding a smooth and rather flat surface. The peach surface is densely covered with conspicuous trichomes (approximately 1 mm Contributions of structural groups present in the selected molecules to the solubility parameter componentStructural group [fig_ref] Figure 4: Scanning electron micrographs of intact plant surfaces [/fig_ref]. Given the dense and thick network of micro-(trichomes) and nano-(epicuticular waxes) tubes covering peach fruits and Eucalyptus leaves, respectively, we could not determine their roughness by atomic force microscopy. Adaxial Eucalyptus leaf surfaces are almost super-hydrophobic and the peach fruit surface is also very hydrophobic, while the pepper fruit surface is more wettable. [formula] F di ((MJ/m 3 ) 1/2 mol -1 ) F pi ((MJ/m 3 ) 1/2 mol -1 ) E hi (J/ [/formula] ## Surface free energy, polarity, work of adhesion and solubility parameter Measurement of contact angles with water, glycerol and diiodomethane enabled several plant surface properties to be calculated . The total surface free energies of peach and pepper fruits are similar and significantly higher than that measured for Eucalyptus leaves (approximately 32.2 versus 17.4 mJ m -2 , respectively). In all cases, there is a major contribution of the Lifshitz-van der Waals component, while the acid-base component is more significant in pepper fruits. Peach and Eucalyptus surfaces have higher contributions from electron acceptor interactions, whilst electron donor interactions predominate in pepper. The lowest and the highest surface polarities correspond to peach and pepper fruits, respectively. The work of adhesion for water and glycerol is much higher in pepper fruits (81.2 mJ m -2 ) than peach fruits and Eucalyptus leaves (between 15.0 and 22.1 mJ m -2 ; . However, the work of adhesion for diiodomethane lies within a similar range to that for peach and pepper fruits, and is significantly lower in Eucalyptus leaves. Concerning the solubility parameters (δ θ ) of the three plant materials analysed , Eucalyptus leaves exhibit a significantly lower value (10.6 MJ 1/2 m -3/2 ) than pepper and peach fruit surfaces (approximately 17 MJ 1/2 m -3/2 ). ## Solubility parameter of epicuticular waxes The solubility parameters of the most abundant epicuticular waxes (δ wax ) of Eucalyptus leaves, pepper and peach fruits are shown in [fig_ref] Table 5: Total solubility parameter [/fig_ref]. The dominant class of compounds in both pepper and peach fruit waxes is n-alkanes, which have a δ wax around 16 MJ 1/2 m -3/2 for the most abundant compounds reported (C 23 to C 31 n-alkanes). However, it is remarkable that such compounds lack polar (δ p ) and H-bonding (δ h ) solubility parameter components. While n-alkanes are the most abundant waxes in peach, a relatively abundant class of triterpenoids (e.g., α,β-amyrin) can be found in pepper fruits. Although the presence of such waxes will not significantly modify δ wax (16.7 MJ 1/2 m -3/2 , for α,β-amyrin), they could potentially facilitate interactions due to polar forces and H-bonding owing to the presence of a functional alcohol group. With regard to Eucalyptus leaves, the dominant class of compounds is β-diketones, which have a δ wax of approximately 16.6 MJ 1/2 m -3/2 and non-zero δ p and δ h components owing to the presence of ketone functional groups. Other wax classes often found in Eucalyptus leaves are n-alkanes (chiefly n-nonacosane), alkanals (aldehydes) such as hexacosanal, and ketones (e.g., heptadecan-2-one). These kinds of waxes and some others not included in [fig_ref] Table 2: Characteristics of the chemicals used for estimation of solubility parameters [/fig_ref] (data not shown) were found to have δ wax values ranging between 16 and 17 MJ 1/2 m -3/2 , but in contrast to n-alkanes they all have contributions from δ p and δ h because of the presence of aldehyde, ketone and/or alcohol functional groups. For peach and pepper fruits, the values of δ θ and δ wax are within the same range (16 to 17 MJ 1/2 m -3/2 ). However, a significantly lower δ θ value (10.6 MJ 1/2 m -3/2 ) was determined for Eucalyptus leaf surfaces than the δ wax calculated for β-diketones (approximately 16.6 MJ 1/2 m -3/2 ). This may be attributed to the nano-structure of the Eucalyptus leaf surface, which decreases the δ θ value in association with a high degree of surface roughness and hydrophobicity, as shown for various synthetic and natural materials [bib_ref] Superhydrophobic surfaces and emerging applications: non-adhesion, energy, green engineering, Nosonovsky [/bib_ref]. ## Solubility parameters of agrochemicals The total solubility parameters (δ m ) and solubility parameter components of the selected molecules are shown in [fig_ref] Table 6: Total solubility parameters [/fig_ref]. The water-soluble compounds urea and sorbitol have high δ m values and major contributions from δ p and especially δ h . A similarly high δ m value was determined only for the non-systemic fungicide chlorothalonil, which also has the highest δ d value of all the compounds considered. Surface free energy per unit area, Lifshitz van der Waals component (γ LW ), acidbase component (γ AB ) with the contributions of electron donor (γ -) and electron acceptor (γ + ) interactions, total surface free energy (γ), surface polarity (γ AB γ -1 ), solubility parameter (δ θ ) and work of adhesion (for water, W w ; glycerol, W g ; diiodomethane, W d ) of adaxial Eucalyptus leaf, pepper and peach fruit surfaces Surface free energy and its components (mJ m -2 ) Work of adhesion (mJ m -2 ) Insecticides (esfenvalerate, fenoxycarb, α-cypermethrin and formetanate) and the fungicide flutolanil have δ m values ranging between 20 to 24 MJ 1/2 m -3/2 , with a major contribution from the δ d component. In contrast, a higher δ m was determined for the fungicide chlorothalonil. The results also indicate that compounds with different chemical structures such as flutolanil, fenoxycarb and esfenvarelate can have similar δ m values. [formula] Sample γ LW γ - γ + γ AB γ γ AB γ -1 (%) δ θ (MJ 1/2 m -3/2 ) W a. [/formula] The δ m values of the three non-ionic surfactants are between 19.5 and 21.4 MJ 1/2 m -3/2 and the differences among them are chiefly associated with the values of the δ d and δ p components. The major difference between Genapol X-80 and Brij 35 is related to δ p (8.1 and 5.9 MJ 1/2 m -3/2 , respectively). The alkyl-phenol surfactant Triton X-100 lies between the values calculated for the two alkyl ethoxylates (δ m = 20.2 MJ 1/2 m -3/2 ). ## Affinity of agrochemicals for plant surfaces Results concerning the affinity of agrochemicals for plant surfaces in relation to δ θ and δ wax are shown in [fig_ref] Figure 5: Affinity of agrochemicals for Eucalyptus leaf [/fig_ref]. The affinities of chemicals for the dominant epicuticular waxes present in Eucalyptus leaves, pepper and peach fruits are within a similar range [fig_ref] Figure 5: Affinity of agrochemicals for Eucalyptus leaf [/fig_ref] , B and C, light grey bars). The n-alkanes present in pepper and peach fruit surfaces provide a lower affinity for agrochemicals than the β-diketones of the Eucalyptus leaf. The compounds with the lowest affinity for the epicuticular waxes covering the three plant materials analysed (i.e., those with the highest Δδ wax values) are urea and sorbitol. Regarding the plant protection active ingredients, the highest affinity for epicuticular waxes was calculated for esfenvalerate and flutolanil, followed by fenoxycarb and formetanate. The compounds α-cypermethrin and chlorothalonil have higher Δδ wax values and hence a lower affinity for the dominant waxes. Genapol X-80 is the surfactant with the highest affinity for epicuticular waxes, followed by Triton X-100. The range of affinities of the agrochemicals for pepper and peach fruit surfaces based on contact angle measurements (Δ δ θ ) is similar to the range predicted from the dominant epicuticular waxes (Δδ wax ). In contrast, lower affinities of agrochemicals for the Eucalyptus leaf surface were estimated in relation to contact angle measurements, which take into account the combined effects of surface chemistry and roughness [fig_ref] Figure 5: Affinity of agrochemicals for Eucalyptus leaf [/fig_ref] , dark grey bars). ## Solubility parameter gradient across the plant surface The total solubility parameters (δ mn ) and solubility parameter components of model cutin monomers and cell wall polysaccharides are shown in [fig_ref] Table 7: Total solubility parameters [/fig_ref]. The values estimated for free ω-hydroxy-fatty acids, D-glucose and D-galacturonic acid are higher than those calculated after monomer esterification or formation of glycosidic bonds, mainly because the H-bonding component is lower. The δ mn values of the monomers were estimated by determining the possible minimum and maximum values in association with the potential occurrence of more than two ester bonds (e.g., on 9,10,18-tryhidroxy-octadecanoic acid). The presence of free hydroxyl groups will raise the value of δ h and hence increase the total solubility parameter. This can be observed, for example, by contrasting 10,16-oxy-hexanodecanoate (three ester bonds) with 10-oxy, 16-hydroxy-hexanodecanoate (two ester bonds); the δ mn values are 17.9 and 19.4 MJ 1/2 m -3/2 , respectively. Similarly, the monomer 9,12,18-oxyoctadecanoate (four ester bonds; 17.7 MJ 1/2 m -3/2 ) has a lower δ mn value than 9-oxy, 18,12-dihydroxy-octadecanoate (two ester bonds; 20.0 MJ 1/2 m -3/2 ). The cutin monomers derived from 16-hydroxy-hexanodecanoic and 9,10-epoxy-18-hydroxy-octadecanoic acid can only form two ester bonds and have the lowest δ mn parameter. The total solubility parameters determined for polymerised D-glucose and Dgalacturonic acid (around 32.2 MJ 1/2 m -3/2 ) are remarkably above the range assessed for waxes and cutin, chiefly because of the major contributions of the δ p and δ h components. Methylation of the carboxylic group in the pectic compound leads to only a slight increase of the δ mn parameter (data not shown). By analysing the solubility parameter results estimated for the different materials covering epidermal cells and for the dominant epicuticular waxes present in the three species used for contact angle determinations, a solubility parameter gradient from the outer to the inner side of the cuticle can be expected as depicted in [fig_ref] Figure 6: Solubility parameter gradient model for the cuticle and cell wall covering the... [/fig_ref]. The epicuticular wax layer, which is in direct contact with the atmosphere, has the lowest solubility parameter value and may often lack polar and H-bonding components, as observed for n-alkanes. Cutin monomers usually form the cuticular matrix and their degree of polymerization can alter the total solubility parameter of the biopolymer (ranging between 17 and 20 MJ 1/2 m -3/2 ). In the more external cuticular layers there are variable proportions of waxes and cutin (i.e., in the epicuticular wax layer and cuticle proper). Polysaccharides are present in the cuticular layer in direct contact with the cell wall. Therefore, according to current views on the composition of the materials (cuticle and cell wall) covering epidermal plant cells, a solubility parameter gradient is established from the external and more hydrophobic epicuticular wax layer towards the more hydrophilic internal cell wall. # Discussion In this study, a procedure for predicting the interactions among different structural plant surface constituents and agrochemicals, based on the estimation of solubility parameters, has been introduced for the first time in a plant science context. While prediction of solubility parameters is commonly used in membrane science [bib_ref] Study on surface modification by surface-modifying macromolecules and its applications in membrane..., Khayet [/bib_ref] and also in pharmacology [bib_ref] The use of solubility parameters in pharmaceutical dosage form design, Hancock [/bib_ref] , this procedure has not so far been applied to estimating the surface properties of biological materials, with the exception of a few studies focused on human skin [bib_ref] Three dimensional solubility parameters and their use in characterising the permeation of..., Gröning [/bib_ref] [bib_ref] Influence of membrane-solvent-solute interactions on solute permeation in skin, Dias [/bib_ref]. The permeability of a compound through a plant cuticle is the product of its solubility, which is a thermodynamic parameter reflecting the degree of interaction between that compound and the plant cuticle, and its diffusivity through the matrix of the plant cuticle, which is a kinetic parameter associated with the molecular size of the compound and the structure of the matrix. This study is only focused on analysing for the first time the solubilities (not the permeabilities) of model plant surfaces and chemical constituents in relation to agrochemicals of commercial significance, adopting a thermodynamic perspective. Two different approaches have been followed to assess the solubility parameters of plant surfaces. One is based on contact angle measurements (δ θ ), which reflect both physical and chemical effects associated with the topography and composition of the surface. The other is limited to considering the nature of the dominant epicuticular waxes covering the surface (δ wax ). The three materials selected are good examples of the diversity of plant surface structures in relation to the variability of cell shapes and micro-and nano-structures on the cell surfaces. In addition, the similar chemical composition of the dominant waxes in pepper and peach fruits in contrast to the β-diketones prevailing in juvenile Eucalyptus leaves offers another interesting aspect to evaluate in the plant surfaces investigated. From an agrochemical viewpoint, substances with different activities, polarities, and degrees of complexity were assessed as model compounds. ## Plant surface properties The adaxial Eucalyptus leaf surface is almost superhydrophobic and has a high degree of nano-roughness conferred by the dense network of wax nano-tubes [bib_ref] Epicuticular wax structure and regeneration on developing juvenile Eucalyptus leaves, Wirthensohn [/bib_ref]. However, the very hydrophobic surface of the highly pubescent peach variety analysed has a high degree of micro-roughness provided by the trichome network. In contrast, the pepper fruit surface has a smooth topography and it is more wettable than the other plant surface samples. The determination of contact angles of three liquids enabled the three distinct plant materials to be compared by a novel approach [bib_ref] New insights into the properties of pubescent surfaces: peach fruit as model, Fernández [/bib_ref]. Consequently, the three plant surfaces were characterised in terms of surface free energy, polarity, work of adhesion and solubility parameter. Fruit surfaces have the highest surface free energies and solubility parameter values, in contrast to the Eucalyptus leaf surface. However, the peach skin is the surface with the lowest polarity, which implies the lowest degree of potential polar and H-bonding interactions among the three surfaces analysed. The pepper surface has a significantly higher work of adhesion for water than the peach fruit and the Eucalyptus leaf surface. This indicates that water drops falling on the pepper fruit surface will be retained, in contrast to the repulsion of the drops falling on to peach fruit and especially Eucalyptus leaf surfaces. Therefore, the behaviour of the plant surfaces evaluated may bring some ecophysiological advantage to the plant organs and related species, which future studies should investigate further [bib_ref] New insights into the properties of pubescent surfaces: peach fruit as model, Fernández [/bib_ref]. Furthermore, the proposed tools based on contact angle measurements of the three liquids may be useful for investigating plant surface dynamics during the growing season or as affected by plant biotic or abiotic stress factors. For example, epicuticular wax erosion in association with environmental pollution [bib_ref] Tree leaf wettability as passive bio-indicator of urban habitat quality, Kardel [/bib_ref] or the deposition of aerosols and microorganisms [bib_ref] Stomatal penetration by aqueous solutions -an update involving leaf surface particles, Burkhardt [/bib_ref] could increase the degree of heterogeneity and wettability of the surface, ultimately affecting plant-water relationships [bib_ref] The exclusion of ambient aerosols changes the water relations of sunflower (Helianthus..., Pariyar [/bib_ref]. ## Solubility parameters of cuticular components and agrochemicals The solubility parameters of the dominant waxes in the analysed surfaces lie within the range 16.0 to 16.7 MJ 1/2 m -3/2 . Such a range is representative of a wide number of wax compounds such as β-diketones, n-alkanes, amyrins, and a few other additional compounds having aldehyde, ketone, and alcohol functional groups. The model cutin monomers evaluated have higher solubility parameters than waxes, varying between 19 and 22 MJ 1/2 m -3/2 for the free ω-hydroxy-fatty-acids and possibly falling to 17 MJ 1/2 m -3/2 after esterification of all functional hydroxyl groups. The model cellulose and pectin monomers have considerably higher solubility parameters than other cuticle constituents, which supports previous observations on the hydrophobicity of the cuticle as compared to the cell wall [bib_ref] Epidermis: the formation and functions of a fundamental plant tissue, Javelle [/bib_ref]. The presence of wax compounds such as n-alkanes renders the plant surface apolar. Despite the smooth topography and wettability of the pepper fruit, its n-alkane coating will only lead to the occurrence of dispersive interactions with surface-deposited materials and liquids, a phenomenon that can also be expected for peach fruits. The different numbers of EO units in the surfactants evaluated had only a slight effect on the total solubility parameter. A correlation between surfactant solubility parameters and HLB or critical micelle concentrations has been reported [bib_ref] Micellar properties of non-ionic surfactants in relation to their solubility parameters, Samaha [/bib_ref] [bib_ref] Hydrophilic-lipophilic balance, solubility parameter, and oil-water partition coefficient as universal parameters of..., Schott [/bib_ref]. The surfactant with the lowest solubility parameter (Genapol X-80) was recorded as having the lowest surface tension (approximately 27 mJ m -2 at 0.1%) while the highest surface tension (around 45 mJ m -2 at 0.1%) and solubility parameter were estimated for Brij 35. It must be noted that water has a total solubility parameter of 47.9 MJ 1/2 m -3/2 [bib_ref] General principles governing dissolution of materials in solvents. 4.1 Simple solvent characteristics, Senichev [/bib_ref] and that all chemicals except urea and sorbitol, which can be supplied at concentrations above 1%, were applied at approximately 0.1% concentrations. However, it can be expected that the chemicals assessed will interact with epicuticular waxes once sprayed on to plant surfaces as aqueous solutions, especially if surface-active agents are applied to improve contact between the solid and the liquid. ## Affinities of agrochemicals for plant surfaces The affinities of agrochemicals for plant surfaces were evaluated on the basis of contact angle measurements and by considering the chemical structures of the dominant epicuticular waxes. According to Greenhalgh et al. [bib_ref] Solubility parameters as predictors of miscibility in solid dispersions, Greenhalgh [/bib_ref] , compounds with a solubility parameter difference (Δδ) below 7 MJ 1/2 m -3/2 are likely to be miscible, while chemicals with a Δδ higher than 10 MJ 1/2 m -3/2 are likely to be immiscible. The agrochemicals and plant surfaces selected in this study were found to have Δδ values ranging between 2.5 MJ 1/2 m -3/2 (the highest affinity) and 26.6 MJ 1/2 m -3/2 (no affinity). This indicates that such plant surfaces have a high affinity for some agrochemicals (Genapol X-80, flutolanil, esfenvalerate, Triton X-100, fenoxycarb, Brij 35, formetanate), which can readily penetrate into the plant organ (leaf or fruit in this case), and less affinity for other compounds (α-cypermethrin, sorbitol, urea and chlorothalonil). The plant protective chemicals with the highest affinities for plant surfaces were found to be those with the lowest total solubility parameters. The high affinities of esfenvalerate, flutolanil and fenoxycarb for the surfaces evaluated make these compounds more prone to cuticular uptake and sorption into plant tissues. Since such plant protection products can be sprayed on to the leaves and fruits of agro-forest species, the estimation of solubility parameters could be used as a complementary tool for pesticide risk assessment. Therefore, the compounds with the lowest toxicity risk will be those with lower affinities for plant surfaces. The proposed methodology may be useful for improving the performance of foliar sprays of e.g., plant protection products, herbicides and fertilisers, taking into account their mode of action (e.g., systemic or contact) and the surface properties of the target organism (e.g., the plant, or surface pathogens and pests). The surfactants selected in this study also have high affinities for plant surfaces, especially in the case of Genapol X-80. Surfactant solutions sprayed on to foliage have often been observed to be phytotoxic [bib_ref] Reduction of non-ionic surfactant phytotoxicity by divalent cations, Uhlig [/bib_ref] , which may be associated with the high solubility of some of these compounds in plant surfaces. The lower affinities of agrochemicals for the Eucalyptus leaf estimated from contact angle measurements are due to the major roughness provided by the wax nano-tubes that densely cover the surface. These nano-tubes are also responsible for the high degree of hydrophobicity of the material [bib_ref] Superhydrophobicity in perfection: the outstanding properties of the lotus leaf, Ensikat [/bib_ref] [bib_ref] Natural and biomimetic artificial surfaces for superhydrophobicity, self-cleaning, low adhesion, and drag..., Bhushan [/bib_ref]. In contrast, such differences between the affinities predicted from contact angle measurements and epicuticular wax chemistry were not observed for the peach and pepper fruit surfaces. The micro-scale roughness provided by the dense layer of trichomes covering the peach surface markedly increases the hydrophobicity of that surface, but seems to have limited effect on the solubility parameter. ## Solubility parameter gradient model To gain insight into the characteristics of the cuticle and the cell wall by calculating solubility parameters, additional estimations were made for common cuticular matrix constituents and cell wall polysaccharides. To our knowledge, this is the first time in which the polar, dispersive and H-bonding properties of plant cuticular and cell wall constituents have been interpreted in quantitative terms. Owing to the properties of the dominant epicuticular waxes present in the three analysed plant materials, it is concluded that the solubility parameter increases with increasing depth from the epicuticular wax surface towards the cell wall. The solubility parameters determined for the model cellulose and pectin compounds are much higher than those for cutin monomers and epicuticular waxes but are still far away from the value of water. Given the ubiquitous presence of waxes, cutin and polysaccharides in the layers covering plants' epidermal cells [bib_ref] Cutin from plants, Kolattukudy [/bib_ref] , and assuming that the chemical constituents selected are representative of a wide range of species with hydrophobic surface properties and within the same range of potential alternative chemical components, a solubility parameter gradient is observed for a model plant surface, which can be applied to e.g., adaxial and abaxial leaf, fruit, flower or trichome surfaces that are covered with a cuticle. On the basis of thermodynamic principles, compounds with a low surface free energy (i.e., a low solubility parameter) will tend to migrate from the plant cell wall towards the epicuticular wax layer in order to decrease the Gibbs free energy [bib_ref] Surface modification of polyvinylidene fluoride pervaporation membranes, Khayet [/bib_ref]. This could be an alternative and/or complementary hypothesis to explain the migration of cuticular material (waxes and cutin) towards the air/plant interface, in contrast to cuticular transpiration as a driving force [bib_ref] Movement and regeneration of epicuticular wax through plant cuticles, Neinhuis [/bib_ref] [bib_ref] Self assembly of epicuticular waxes on living plant surfaces imaged by atomic..., Koch [/bib_ref] [bib_ref] Structural characterization of polyhydroxy fatty acid nanoparticles related to plant lipid biopolyesters, Heredia-Guerrero [/bib_ref]. As noted by Scherbatskoy & Tyree [bib_ref] Kinetics of exchange of ions between artificial precipitation and maple leaf surfaces, Scherbatskoy [/bib_ref] , cuticular polymers contain polar and ionisable substituents, providing the cuticle with polar hydrophilic regions and ion exchange capacity. The topography and chemistry of epicuticular waxes generally provide a lower solubility parameter than the one prevailing in the cuticle proper and principally in the cuticular layer, where significant amounts of polysaccharides are present. The lack of polar and H-bonding functional groups in the dominant epicuticular waxes and a higher degree of monomer esterification in the cuticle matrix will tend to lower the total solubility parameter of the membrane by reducing the polar and H-bonding components. The presence of cutan can also decrease the solubility parameter of the cuticle matrix to some extent. According to Jeffree [bib_ref] The fine structure of the plant cuticle, Jeffree [/bib_ref] , cutanization (i.e., the gradual formation of cutan in the cuticle matrix) of the cuticular layer, as reported for the Clivia minata leaf cuticle [bib_ref] Development of plant cuticles fine structure and cutin composition of Clivia miniata..., Riederer [/bib_ref] , can arise from a maturation process involving the progressive modification of the previously deposited cutin and any embedded polysaccharides and waxes. The author suggests that the progressive reduction in reactivity of all components of the cuticular layer during cutanisation indicates that all types of polar functional groups are systematically eliminated during this maturation phase. In order to calculate the solubility parameters, the properties of the trichomes covering the peach skin surface were considered. In this case the cuticular matrix is composed exclusively of cutin, in contrast to the high percentage of cutan in shaved cuticular membranes [bib_ref] New insights into the properties of pubescent surfaces: peach fruit as model, Fernández [/bib_ref]. According to the cutan hypothesis proposed by Jeffree [bib_ref] The fine structure of the plant cuticle, Jeffree [/bib_ref] , the cuticular matrix of the peach fruit has a lower solubility parameter than cutin as the dominant cuticular matrix bio-polymer. However, the cuticular domain of the peach fruit cuticle will also have contributions from cutin and polysaccharides, which will gradually raise the total solubility parameter of the membrane as it comes closer to the cell wall. The cuticular matrix of pepper is mainly made of cutin [bib_ref] Fruit cuticle lipid composition and fruit postharvest water-loss in an advanced backcross..., Parsons [/bib_ref] , but an insoluble fraction likely to be cutan has also been identified [bib_ref] Spectroscopic characterization of aliphatic moieties in four plant cuticles, Johnson [/bib_ref]. No information is currently available on the composition of the Eucalyptus globulus leaf cuticle matrix, but the major reduction in solubility parameter associated with the nano-scale roughness of the Eucalyptus surface supports the occurrence of the solubility parameter gradient shown in [fig_ref] Figure 6: Solubility parameter gradient model for the cuticle and cell wall covering the... [/fig_ref]. While most aerial plant surfaces are believed to be covered with a cuticle based on a cutin and/or cutan matrix, which contains variable amounts of waxes and polysaccharides, trials with more hydrophilic surfaces and different materials should be carried out in the future to estimate the solubility parameters of plant surface chemical constituents quantitatively. # Conclusions A novel method for predicting the interactions between plant surface structural constituents, plant surfaces and agrochemicals has been introduced, which was useful for predicting the solubilities of plant surface constituents and the affinities of agrochemicals for plant surfaces. Calculation of the solubility parameters of plant surface constituents led us to observe a solubility parameter gradient established from the cuticular surface towards the wall covering epidermal cells. Comparison of solubility parameters between cuticular and cell wall components will be helpful for clarifying the structure and development of the cuticle from an ontological viewpoint and also for establishing a relationship between the chemical composition and structure of the cuticular membrane, which is currently lacking. The methodology should also be of interest for multiple biological purposes and could help us understand surface phenomena on multiple biological materials. [fig] Figure 1: Molecular structures of the cuticular constituents evaluated. [/fig] [fig] 2 h: They were then rinsed in distilled water (3 × 10 min) and dehydrated in an acetone: water (v/v) series of 50, 60, 70, 80, 90, 95 (2 × 10 min each) and 100% (3 × 10 min). The tissues were successively immersed in 1:3 (2 h), 1:1 (2 h) and 1:3 (3 h) Spurr's resin:acetone (v/v) solutions and kept overnight in pure Spurr's resin (TAAB Laboratories). The samples were finally placed in moulds and were incubated at 60°C for three days. Ultrathin sections were stained with uranyl acetate (20 min) and lead citrate (4 min; both chemicals from EMS) and were examined by TEM (Jeol JEM-1010, Tokyo, Japan). [/fig] [fig] Figure 2: Molecular structures of the agrochemicals selected for calculation of solubility parameters. [/fig] [fig] 20: μm thick), which provide a high degree of micro-scale roughness (Figure 4C, F and I) in contrast to the nano-scale surface roughness of Eucalyptus leaves ( [/fig] [fig] Figure 3: Contact angle measurements on intact, adaxial Eucalyptus (A,C,E) leaf surfaces and pepper fruit surfaces (B,D,F). Drops of: (A,B) water, (C,D) glycerol, and (E,F) diiodomethane. [/fig] [fig] Figure 4: Scanning electron micrographs of intact plant surfaces. Eucalyptus adaxial leaf surfaces: (A) ×400, (D) ×1,000, and (G) ×9,000. Pepper fruit surfaces: (B) ×200, (E) ×400, and (H) ×1,000. Peach fruit surfaces: (C) ×100, (F) ×500, and (I) ×1,300. [/fig] [fig] Figure 5: Affinity of agrochemicals for Eucalyptus leaf (A), pepper (B) and peach fruit (C) surfaces. Light grey bars represent the calculated Δδ wax , based on the solubility of agrochemicals in relation to the most abundant epicuticular wax compound (δ wax ), and dark grey bars refer to Δδ θ (± standard errors), calculated from contact angle measurements (δ θ ). [/fig] [fig] Figure 6: Solubility parameter gradient model for the cuticle and cell wall covering the epidermis of plants. The three cuticular layers are indicated as: EWL (epicuticular wax layer), CP (cuticle proper) and CL (cuticular layer). From the EWL (−) to the cell wall (+) there is a gradual solubility parameter increase. The TEM micrograph corresponds to a transverse section of a juvenile, adaxial Eucalyptus globulus leaf surface (×80,000). [/fig] [table] Table 1: Chemical formula and molar volume of the dominant epicuticular waxes extracted from Eucalyptus leaves, pepper and peach fruits and of common cutin monomers found in plant cuticles [/table] [table] Table 2: Characteristics of the chemicals used for estimation of solubility parameters [/table] [table] Table 5: Total solubility parameter (δ wax ) and solubility parameter components of the most abundant epicuticular waxes of Eucalyptus leaves, pepper and peach fruits [/table] [table] Table 6: Total solubility parameters (δ m ) and solubility parameter components of agrochemicals [/table] [table] Table 7: Total solubility parameters (δ mn ) and solubility parameter components of common cutin monomers and cell wall constituents [/table]

Success stories about severe pneumonia caused by Panton–Valentine leucocidin-producing Staphylococcus aureus ClindamycinStaphylococcus aureusPanton-Valentine leukocidin a b s t r a c tWe describe three cases of community-acquired necrotizing pneumonia which were caused by Panton-Valentine leucocidin-producing strains of Staphylococcus aureus (one of them methicillin sensitive). All cases were successfully treated without any sequelae for the patients due to the prompt initiation of adequate antimicrobial therapy. High suspicion toward this fatal pathogen was the key to the successful outcome of the patients.A previously healthy 18-year-old boy, a Panama resident traveling in Europe, was admitted to the ICU of our hospital suffering from pneumonia and severe respiratory distress. He was transferred from another hospital where he presented with high fever, weakness and productive cough. He had http://dx. # Introduction Community-acquired pneumonia due to Staphylococcus aureus is an infrequent but potentially lethal infection. [bib_ref] for the EMERGENCY ID NET Study Group. Prevalence of methicillin-resistant Staphylococcus aureus..., Moran [/bib_ref] [bib_ref] Incidence, characteristics and outcome of patients with severe community acquired -MRSA pneumonia, Vardakas [/bib_ref] It is often associated with the production of Panton-Valentine leucocidin (PVL), which is responsible for extensive tissue necrosis and a high mortality rate. [bib_ref] Involvement of Panton-Valentine leukocidin-producing Staphylococcus aureus in primary skin infections and pneumonia, Lina [/bib_ref] Herein we present three cases of necrotizing pneumonia, one of them caused by a methicillin susceptible strain, all of which had a favorable outcome. ଝ The work was performed at Attikon University Hospital. been hospitalized for three days and received oseltamivir, ceftriaxone and moxifloxacin. He reported having a flu-like syndrome some days earlier. At ICU admission he had a temperature of 38 - C, a respiratory rate of 40 per min, a heart rate of 120 beats/min and he was hemodynamically stable. Thorax auscultation revealed crackles over the lower and middle, right lung fields, while chest X-ray (CXR) showed patchy alveolar opacities in the right lung. Blood gases on admission showed: pH 7.31, pCO 2 : 50 mmHg, pO 2 : 65 mmHg in a non-rebreathing mask. Laboratory findings revealed a white blood count (WBC) of 1220 cells/L (59% neutrophils) and a CRP of 306 mg/L. The Acute Physiology and Chronic Health Evaluation (APACHE) II severity score was 21. His respiratory distress gradually increased and 24 h later he was intubated and needed inotropic support. The new CXR revealed that alveolar type infiltrates were spread in both lungs. Urine antigens for Legionella pneumophila type 1 and Streptococcus pneumoniae were negative. A bronchoscopy was performed and the Gram stain of the BAL revealed the presence of numerous Gram-positive cocci. The treatment was switched to linezolid and moxifloxacin. Cultures from BAL specimens yielded a PVL, methicillin sensitive S. aureus (MSSA) while blood cultures were negative. Test for HIV was negative. Hydrocortisone was added to the therapeutic regimen because of the septic shock and filgrastim was given for the leukopenia. Clindamycin was also added in order to diminish the toxin production by the PVL strain. [bib_ref] Effect of antibiotics on Staphylococcus aureus producing Panton-Valentine leukocidin, Dumitrescu [/bib_ref] On day 8 the fever subsided but his course was complicated by a spontaneous pneumothorax, which necessitated the insertion of a pleural catheter. Subsequently he suffered an episode of bacteremic ventilator-associated pneumonia by a carbapenem-resistant Klebsiella pneumoniae, which was successfully treated with colistin and tigecycline. The patient was extubated on day 21 and he was finally discharged on day 27. ## Case 2 A 34-year-old woman was transferred to our ICU from the Department of Obstetrics and Gynecology due to severe respiratory failure six days after in vitro fertilization. She had no past medical history. The embryo-transfer was complicated by Ovarian Hyperstimulation Syndrome. On admission the APACHE II score was 26 and blood gases values were pH: 7.33, pO 2 : 80 mmHg, pCO 2 : 38 mmHg on facemask providing FiO 2 of 40%. Auscultation revealed diffuse crackles. During the next 6 h her temperature rose to 39.5 - C and she presented productive cough with dark sputum and shortness of breath at rest, so she was admitted to the ICU. CXR on ICU admission revealed an infiltrate of the left middle lung field and bilateral pleural effusions. Laboratory tests showed a WBC of 3700/L, CRP of 298 mg/L, procalcitonin of 36 ng/mL, fibrinogen of 507 mg/dL, and hypogammaglobulinaemia, while human chorionic gonadotropin (hCG) levels were compatible with pregnancy. Sputum Gram stain revealed Gram-positive cocci and urinary antigens for L. pneumophila type 1, and S. pneumoniae and pharyngeal aspirate for Influenzae A and B were negative. Treatment with intravenous linezolid and clindamycin was immediately started because the severity of the patient's condition and the isolation of Gram-positive cocci in the sputum supported the clinical suspicion of community-acquired staphylococcal pneumonia. Nevertheless, the patient worsened and became hemodynamically unstable requiring intubation and vasopressors. She was also given hydrocortisone and activated protein C. The blood cultures and the tracheal aspirate grew MRSA. The patient showed gradual hemodynamic improvement and was weaned from vasopressors after three days, while hCG values declined rapidly returning to nonpregnant range. A chest CT scan was performed and showed bilateral infiltrates with multiple cavitating lesions, left lung abscess formation and bilateral pleural effusions. Culture of pleural fluid developed also MRSA. Her course was complicated by nosocomial bacteremia and she was submitted to a tracheostomy. Gas exchange as well as lung mechanics gradually stabilized and she was finally weaned from mechanical ventilation. She was discharged to the ward on day 37 and she had a long rehabilitation period. ## Case 3 A 33-year-old male with no significant past history was admitted to the ICU of our hospital because of acute respiratory failure due to a lower respiratory tract infection. He presented to the Emergency Department several days earlier complaining for chest pain, high temperature (39 - C) and progressive dyspnea with blood-stained sputum. The CXR revealed the presence of bilateral pulmonary infiltrates. After his admission to the internal medicine department his situation progressively deteriorated and two days later he was intubated and transferred to the ICU. His treatment initially included ceftriaxone and clarithromycin but because of the severity of his situation a staphylococcal pneumonia was suspected and the antibiotic regimen was empirically modified to linezolid and moxifloxacin. On ICU admission he presented hemodynamic instability and had to be resuscitated with intravenous fluids and norepinephrine. Initial laboratory data obtained in the ICU showed WBC: 17.890/L (neutrophils: 82%), CRP: 250 mg/L. Arterial blood gases analysis showed respiratory acidosis and severe hypoxemia (pH: 7.11, pCO 2 : 91 mmHg, pO 2 : 56 mmHg -in assist control ventilation, FiO 2 : 100%). Blood cultures and sputum specimens taken in the ward grew MRSA. The blood cultures remained positive after seven days of treatment with linezolid and daptomycin. Clindamycin were added to the regimen. Two transesophageal echos performed in fifteen days time did not reveal the presence of vegetations. Of note, HIV and tests for other respiratory pathogens were negative. The patient's course was complicated by the development of spontaneous pneumothorax and empyema of the right lung [fig_ref] Figure 1 -: C/T scan of the chest showing right pneumothorax as well as bilateral... [/fig_ref]. He had to be submitted to lung decortication. MRSA was also isolated from the culture of the pleural fluid. On day 23 he was submitted to a tracheostomy. His situation ameliorated, the tracheostomy was closed and he was transferred to the ward on day 40. He was discharged from the hospital one month later. It should be noted that one of the herein described cases has been previously published. 5 ## In vitro evaluation of s. aureus clinical isolates Seven isolates were available for further testing. Susceptibilities to various antimicrobial agents were evaluated by an automated system (BD Phoenix automated microbiology system; BD Diagnostic Systems, Sparks, MD) [fig_ref] Table 1 -: Minimum inhibitory concentrations [/fig_ref]. All isolates were found to carry the lukF-PV and lukS genes coding for Panton-Valentine leukocidin (PVL), 6 while all isolates of cases 2 and 3 were found to also carry the mecA gene in type IVc/e staphylococcal cassette chromosome mec element (SCCmec). [bib_ref] Multiplex PCR strategy for subtyping the staphylococcal cassette chromosome mec type IV..., Milheiric [/bib_ref] The PFGE patterns of SmaI-digested DNA of all isolates of cases 2 and 3 were identical. MLST typing performed only to blood isolates linked PFGE pattern with the well-described ST80 clone that seems to be spreading through Europe. Isolate m4442B of case 1 was lost to subculture and was not submitted to MLST typing. # Discussion S. aureus is a commonly recognized cause of infections in humans. These infections can become life-threatening due to the ability of the microorganism to produce virulence factors including adhesion factors and toxins. PVL is a toxin produced by S. aureus strains (carrying the lukS-PV and lukF-PV genes) often associated to severe skin and soft tissue infections but also causing necrotizing pneumonia with a high mortality. The presence of this toxin has been detected among MRSA as well as MSSA strains. The usual patient with PVL-associated S. aureus pneumonia is a child or a young otherwise healthy adult. A viral illness can precede pneumonia while leucopenia attributed to the toxin is not an uncommon finding. A specific disease entity has also been ascribed to PVL-positive S. aureus-associated respiratory infection including a rapidly progressing hemorrhagic and necrotic pneumonia which can cause Acute Respiratory Distress Syndrome (ARDS) possibly leading to septic shock and multi-organ failure. We describe three patients with severe pneumonia caused by Panton-Valentine-producing S. aureus. In two cases bacteremia was also present. Diagnosis of influenza was not confirmed in any case while prodrome flu-like symptoms were present in only one patient. Both MRSA isolates belonged to ST80, which is the most common clone causing infection in the community as well as in the hospital in Greece. [bib_ref] Spread of Staphylococcus aureus clinical isolates carrying Panton-Valentine leukocidin genes during a..., Chini [/bib_ref] Among bloodstream infections caused by MRSA, the hospitalassociated clone ST239 and sporadic cases of ST30, ST97 and ST398 9 follow ST80. One of the study isolates was MSSA producing PVL. Although rare, the production of PVL toxin by MSSA strains has been reported previously. [bib_ref] Involvement of Panton-Valentine leukocidin-producing Staphylococcus aureus in primary skin infections and pneumonia, Lina [/bib_ref] Three similar cases due to MRSA strains have also been described in Greece, two of them with an unfavorable outcome. [bib_ref] Community-acquired pneumonia and bacteremia due to methicillin-resistant Staphylococcus aureus carrying Panton-Valentine-leukocidin gene..., Balis [/bib_ref] [bib_ref] Community-acquired methicillin-resistant Staphylococcus aureus carrying Panton-Valentine leukocidin genes; a lethal cause of..., Magira [/bib_ref] The mortality of PVL-positive CAP lies between 45% and 56%. [bib_ref] Multiplex PCR strategy for subtyping the staphylococcal cassette chromosome mec type IV..., Milheiric [/bib_ref] [bib_ref] Spread of Staphylococcus aureus clinical isolates carrying Panton-Valentine leukocidin genes during a..., Chini [/bib_ref] The severity of the disease is similar whether the isolate is a methicillin-resistant strain or not because it is associated to other factors such as multi-organ failure, ICU admission, mechanical ventilation, leukopenia, necrotizing pneumonia, shock, disseminated intravascular coagulation, rash, ARDS, airway bleeding and development of complications. [bib_ref] Comparison of community-acquired pneumonia due to methicillin-resistant and methicillin-susceptible Staphylococcus aureus producing..., Vardakas [/bib_ref] [bib_ref] Factors predicting mortality in necrotizing community-acquired pneumonia caused by Staphylococcus aureus containing..., Gillet [/bib_ref] Our patients presented some of these factors such as ICU treatment, mechanical ventilation due to pneumonia-induced ARDS and need for inotropic support while all of them developed several complications. Two cases had leucopenia on admission, which is considered to be inversely associated with survival. On the contrary, another unfavorable prognostic factor, airway bleeding, was absent possibly contributing to the favorable outcome. The optimal treatment of this often lethal disease has not been established. Since the detrimental effects of the infection are mainly caused by the production of PVL rather than the microorganism itself, an effective treatment should target first at the eradication of S. aureus but additionally at diminishing the effects of the toxin. Drainage of the suppurative collection, if possible, is suggested in order to remove PVL containing tissues. Furthermore certain antimicrobial agents are able to reduce the production of PVL. Such molecules include clindamycin, linezolid and rifampicin. 14 These antimicrobial agents are effective in reducing PVL production even at suboptimal concentrations in necrotic tissues. have shown that if other agents such as beta-lactams are used and levels above the MIC (minimum inhibitory concentrations) are not achieved in the infected tissues, PVL secretion may be enhanced. 14 Moreover upregulation of PVL toxin observed with nafcillin 15 could also result with flucloxacillin when used for treating MSSA. [bib_ref] Diagnosis and treatment of Panton-Valentine leukocidin (PVL)-associated Staphylococcal pneumonia, Morgan [/bib_ref] As for clindamycin additional advantages of the drug include its ability (a) to decrease Toxic Shock Syndrome Toxin-1 production 17 and (b) to stop production of alpha-toxin by translational inhibition. [bib_ref] Effects of subinhibitory concentrations of antibiotics on alpha toxin gene expression of..., Ohlsen [/bib_ref] Combined use of clindamycin and linezolid has been described as having successful results on PVL toxin and patients' survival in two of three cases of necrotizing pneumonia. [bib_ref] Prompt and successful toxin targeting treatment of three patients with necrotizing pneumonia..., Rouzic [/bib_ref] On the other hand vancomycin, a traditional drug for the treatment of MRSA, has no impact on exotoxin formation. [bib_ref] Pleuropulmonary complications of Panton-Valentine leukocidin positive community acquired methicillin-resistant Staphylococcus aureus: importance..., Micek [/bib_ref] Polyvalent human intravenous immunoglobulin (IVIg) is an adjunctive therapy for serious PVL-associated infections. It acts by neutralizing the PVL pore formation and by inhibiting the cytotoxic effects of PVL on polymorphonuclear cells as shown by Gauduchon et al. [bib_ref] Neutralization of Staphylococcus aureus Panton-Valentine leukocidin by intravenous immunoglobulin in vitro, Gauduchon [/bib_ref] Although the result appears to be dose-dependent, the usually reported dose does not exceed 2 g/kg. [bib_ref] Diagnosis and treatment of Panton-Valentine leukocidin (PVL)-associated Staphylococcal pneumonia, Morgan [/bib_ref] The use of IVIg has been reported to have a benefit on patients' survival in sporadic cases. [bib_ref] Prompt and successful toxin targeting treatment of three patients with necrotizing pneumonia..., Rouzic [/bib_ref] [bib_ref] Antitoxin treatments for necrotizing pneumonia due to Panton-Valentine leukocidin-secreting Staphylococcus aureus, Libert [/bib_ref] [bib_ref] Panton-Valentine leukocidin-producing Staphylococcus aureus infections: report of 4 French cases, Salliot [/bib_ref] All our patients received a combination of linezolid and clindamycin agents that provide both antimicrobial action and anti-toxin effect. In the second case we used IVIg as adjunctive therapy, a factor probably contributing to patients recovery. Of note, although the use of linezolid is only recommended for hospital-acquired MRSA pneumonia, it has the advantage of achieving high concentrations in the epithelial lining fluid of the lung [bib_ref] Efficacy of linezolid compared to vancomycin in an experimental model of pneumonia..., Martinez-Olondris [/bib_ref] and it is a reasonable first choice for community-acquired MRSA pneumonia as well, particularly in the suspicion of PVL for the aforementioned reasons. Panton-Valentine associated pneumonia remains a disease entity with a high fatality rate. A high degree of awareness is necessary in order to initiate a proper and aggressive treatment even in the absence of a preceding flu-like syndrome. As the identification of the toxin is not always feasible or quickly performed, empirical choice of antimicrobials is of paramount importance. # Funding No grant or financial support was received for the study. [fig] Figure 1 -: C/T scan of the chest showing right pneumothorax as well as bilateral consolidations and characteristic bullae. [/fig] [table] Table 1 -: Minimum inhibitory concentrations (MICs) of 15 antimicrobials against S. aureus study isolates. [/table]

Sorption of Nickel(II) on a Calcareous Aridisol Soil, China: Batch, XPS, and EXAFS Spectroscopic Investigations ## 2 ## Si 1. xrd, ft-ir, sem, tem and xps analysis Calcareous aridisol (CA) soil was characterized by Fourier Transform Infrared (FT-IR) spectroscopy, X-ray diffraction (XRD), scan electron microscopy (SEM), and transmission electron microscopy (TEM). FT-IR measurement was mounted on a Bruker EQUINOX55 spectrometer in KBr pellet at room temperature. XRD pattern of the CA soil was obtained from the Panalytical X'Pert PRO equipment with a rotation anode using Cu Kα radiation (λ = 0.15406 nm). XRD device was operated at 40 kV and 80 mA, and the measurements were carried out in the range of 2 0 ≤ 2θ ≤ 70 0 . SEM images were collected using a Hitachi S-4800 type instrument. Prior to analysis, the CA soil was sprinkled onto adhesive carbon taps supported on metallic disks. TEM images were taken of each sample using a Philips CM10 type microscope at 200 kV. A Rh-Co grid was dipped in a glue solution and representative sample was distributed over the grid. The batch filtered samples were dried at T = 25 ± 1 °C under N2 conditions before the XPS analysis. The XPS data were obtained with an ESCALab220i-XL electron spectrometer from VG Scientific using 300 W Mg Kα radiation. The pressure in the analysis chamber was maintained below 3×10 -9 mbar. The binding energies of carbon species, i.e., C 1s line at 284.8 eV, was used to correct the observed binding energies for surface charging. Fig. S1 showed the typical SEM and TEM images of the CA soil. CA soil has sheet-like amorphous aggregates and a layer structure, which results in the rough surface of the CA soil . The special structure of CA soil indicates that it owns strong adsorption ability and high capacity for metal ions. Form , it is clearer that CA soil exhibits irregular layer structure and platy shape, and the irregular size is about 200-500 nm. The FT-IR spectrum of the CA soil is shown in [fig_ref] 3, Figure: S1 SEM and TEM images of the CA soil [/fig_ref]. ## Si 2. exafs samples preparation, data collection and analysis ## Si 3. effect of temperature on ni(ii) sorption on the ca soil From the thermodynamic data (see [fig_ref] Table S2: Thermodynamic parameters of Ni [/fig_ref] [formula]     ) / ( (1) RT ΔH R ΔS logK 0 0 d   (2) [/formula] The Gibbs free energy change (ΔG o ) was calculated from the equation: [formula] 0 0 0 S T H G      (3) [/formula] where V is the solution volume (mL) and m is the solid mass (g); Cs is the amount of Ni(II) adsorbed on per weight unit solid after equilibrium (mol/g); and Ce is the [fig] 3, Figure: S1 SEM and TEM images of the CA soil. (a): SEM, and (b): TEM. [/fig] [fig] Fig: S2 XRD pattern of the CA soil. C: Chlorite, Ca: Calcite, F: Feldspar, G: Gypsum, H: Hydromica and Q: Quartz. [/fig] [fig] For: EXAFS analysis, NiO powder was purchased from Sinopharm Chemical Regent Co. The standard sample of Ni(II)aq was obtained from dissolving higher purity nickel powder (Sinopharm Chemical Regent Co.) using nitric acid. The Ni(OH)2(s) was prepared by adding 550 mL 30% ammonia to 500 mL of 1.0×10 -3 mol/L Ni(NO3)2 solution. The mixture solutions were vigorously stirred and purged N2; after 2 hours, the suspension was centrifuged and washed with Milli-Q water in six cycles, shock-freezed in liquid N2 and freeze-dried. Ni-Al LDH sample was prepared by controlled hydrolysis according to reference(4). The amounts of Ni and Al (added as Ni(NO3)2· 9H2O and Al(NO3)3· 9H2O) were adjusted to give Ni/Al ratio(3:1), then 400 mL of both solutions were combined in a 500 mL flask under vigorous stirring. The pH was gradually raised to 6.9 by discontinuous addition of 2.5 mol/L NaOH and then kept constant for 5 hours using an auto-titrator. The precipitation was washed with Milli-Q water and dried as before. The Ni(II) adsorbed CA soil samples at different conditions were conducted using 500 mL vessels with 10.0 g/L CA soil suspension, 0.01 or 0.1 mol/L NaClO4 and 1.71 mmol/L Ni(II) stock solution, and 6 added 0.1 and 0.01 mol/L NaOH solution to balance the acidity of Ni(II) solution. The base solution was added firstly, and then the Ni(II) stock solution was introduced as: 10-50 μL increments of the Ni(II) stock solution were introduced into the suspension under constant stirring to disperse the small aliquot of Ni(II) solution. Periods of a few minutes between the increments were chosen to avoid the local concentration of Ni(II) exceeding the solubility limit of Ni(OH)2(s), while allowing the completion ofNi(II) addition in reasonable delay(5). Finally, the suspension was allowed to equilibrate for 2 days. The samples were recovered by filtration using 0.25 µm membrane, and sealed into the polyethylene bag for EXAFS measurement. It was noted that all the metal solutions were purged with N2 and NaOH solution was freshly prepared to minimize the uptake of carbonate.A series set of Ni K-edge X-ray absorption spectra at 8333 eV were recorded at the Shanghai Synchrotron Radiation Facility (SSRF, China). The beam line energy of X-ray was equipped with a fixed-exit double-crystal Si (111) monochromator, the electron beam energy in the storage ring was 3.5 GeV, and a maximum current of 220 mA was used. The ionization chambers with N2 or Ar atmosphere were used to collect the Ni K-edge spectra in fluorescence mode at ambient temperature. A 29-element pixel high purity Ge solid-state detector was used to collect the fluorescence signal.Prior data collection, the energy was calibrated to the first inflection point on the K absorption edge of a Ni foil standard (E0 = 8333 eV). Solid samples were loaded into individual acrylic holder, whereas a special quartz container was applied to hold the aqueous standards. The data were collected at the Ni K-edge over the energy range of 8233-9322 eV.Addition and normalization of X-ray absorption spectra, extraction of EXAFS oscillations and data analysis were performed with ATHENA and ATERMIS interfaces to the IFFEFIT software. The EXAFS oscillations were isolated from the raw, averaged data by removal of the pre-edge background, approximated by a first-order polynomial. The energy axis (eV) was converted to photoelectron wave vector units (Å -1 ) by assigning the origin, E0, to the first inflection point of the absorption edge. The resulting (k) functions were weighted with k 3 to compensate for 7 the dampening of the EXAFS amplitude with increasing k and were Fourier transformed to obtain radial structure functions (RSFs). The amplitude reduction fact, (S0)2 , was fixed at 0.85. A good fit was determined on the basis of the minimum residual error (Rf). The Debye-Waller factor (σ 2 ) and energy shift (ΔE0) were allowed to vary during this optimization. The theoretical backscatter phases and amplitudes used in data analysis were calculated with the scattering code FEFF 7.0 using the crystal structures of Ni(OH)2 (6), NiO (7) and NiAl2O4(8). [/fig] [table] Table S1: Some chemical and physical properties of the CA soil [/table] [table] Table S2: Thermodynamic parameters of Ni(II) sorption on the CA soil. [/table]

Effects of the Usage of l-Cysteine (l-Cys) on Human Health This review summarizes recent knowledge about the use of the amino acid L-Cysteine (L-Cys) through diet, nutritional supplements or drugs with the aim to improve human health or treat certain diseases. Three databases (PubMed, Scopus, and Web of Science) and different keywords have been used to create a database of documents published between 1950 and 2017 in scientific journals in English or Spanish. A total of 60,885 primary publications were ultimately selected to compile accurate information about the use of L-Cys in medicine and nutritional therapies and to identify the reported benefits of L-Cys on human health. The number of publications about the use of L-Cys for these purposes has increased significantly during the last two decades. This increase seems to be closely related to the rise of nutraceutical industries and personalized medicine. The main evidence reporting benefits of L-Cys usage is summarized. However, the lack of accurate information and studies based on clinical trials hampers consensus among authors. Thus, the debate about the role and effectiveness of supplements/drugs containing L-Cys is still open. # Introduction L-cysteine (L-Cys) is a non-essential amino acid and thus is one of the building blocks required for the synthesis of proteins. It contains sulfur in the form of a thiol group (-SH) at the end of its side chain [fig_ref] Figure 1: Chemical structure [/fig_ref]. The -SH group is responsible for the high reactive capacity of the amino acid, and therefore is responsible for many of its biological functions in human beings. L-Cys is the amino acid establishing disulfide bridges, a type of covalent bond that plays a fundamental role in the folding and stabilization of the tertiary structure of the proteins, thereby supporting their biological activities [bib_ref] Biologically important thiol-disulfide reactions and the role of cyst(e)ine in proteins: An..., Fahey [/bib_ref] [bib_ref] Oxidation Resistance of the Sulfur Amino Acids: Methionine and Cysteine, Bin [/bib_ref]. The presence of conserved Cys in protein motifs found in all organisms could indicate that this feature was harnessed in early evolution to support enzyme catalysis, transcriptional regulation, protein folding, and 3-dimensional structure [bib_ref] The Cysteine Proteome. Free Radic, Go [/bib_ref] [bib_ref] Glutathione homeostasis and redox-regulation by sulfhydryl groups, Meyer [/bib_ref]. Cys is synthesized from methionine (an essential amino acid) thanks to two chemical reactions [bib_ref] Oxidation Resistance of the Sulfur Amino Acids: Methionine and Cysteine, Bin [/bib_ref]. The first of these reactions is a transmethylation reaction, from which homocysteine is obtained as product. Then, homocysteine is transformed into cysteine through a transsulfuration reaction [bib_ref] Metabolism of sulfur-containing amino acids, Stipanuk [/bib_ref]. Cysteine can be assimilated through different pathways depending on the needs of the cells, giving rise to sulfur compounds [bib_ref] Metabolism of sulfur-containing amino acids, Stipanuk [/bib_ref] [bib_ref] Cysteine concentration regulates cysteine metabolism to glutathione, sulphate and taurine in rat..., Stipanuk [/bib_ref]. The main metabolite obtained by the assimilation of Cys is sulfinate, a molecule that is metabolized to give rise to sulfinylpyruvate and pyruvate, or hipotaurine and taurine [fig_ref] Figure 2: Summary of cysteine metabolism[8] [/fig_ref]. On the one hand, taurine is a fairly abundant molecule at the intracellular level. Although the biological role of taurine is not at all clear, some studies suggest that it may be a compound involved in the nervous system, being especially necessary for brain development [bib_ref] Open questions concerning taurine with emphasis on the brain, Oja [/bib_ref]. This suggestion is supported by the fact that high levels of taurine have been found in the fetal brain [bib_ref] Taurine enhances proliferation and promotes neuronal specification of murine and human neural..., Pasantes-Morales [/bib_ref]. In addition, taurine displays many other biological roles, such as the marking of certain toxic intermediaries or the regulation of intracellular calcium levels. On the other hand, the sulfinylpyruvate produced during metabolism of Cys may undergo oxidation to sulfate, which is later used in the synthesis of 3′-phosphoadenosine-5′-phosphosulfate. There are many other important reactions occurring during the metabolism of Cys that allow the synthesis of thiocysteine or the transfer of sulfur between molecules. The human genome encodes about 214,000 Cys-coding sequences [bib_ref] The Cysteine Proteome. Free Radic, Go [/bib_ref]. Thus, the presence of this amino acid in proteins is significant. The reactivity and diverse functions of Cys are mirrored by a spectrum of susceptibilities and dysfunctions of their respective proteins, resulting in central roles for the Cys proteome in development, signal transduction, biologic defenses, aging, and disease [bib_ref] The Cysteine Proteome. Free Radic, Go [/bib_ref] [bib_ref] Glutathione homeostasis and redox-regulation by sulfhydryl groups, Meyer [/bib_ref] [bib_ref] Chemistry and proteomics of cysteine modifications in redox biology, Kim [/bib_ref]. Due to these important implications, L-Cys has attracted much attention of medical researchers and professionals worldwide (mainly in the USA and Europe). Over the past few years, important biological functions of L-Cys in human beings have been described for the first time [bib_ref] The Cysteine Proteome. Free Radic, Go [/bib_ref] [bib_ref] Glutathione homeostasis and redox-regulation by sulfhydryl groups, Meyer [/bib_ref] [bib_ref] Metabolism of sulfur-containing amino acids, Stipanuk [/bib_ref] [bib_ref] Cysteine metabolism and metal toxicity, Quig [/bib_ref] [bib_ref] Redox Signaling Regulated by Cysteine Persulfide and Protein Polysulfidation, Kasamatsu [/bib_ref]. At the same time, the potential applications of this compound at the industrial level have been increasing, mainly in fields related to pharmaceuticals, medicines, and nutraceuticals. As examples of these applications, L-Cys plays an important role in the food industry (it is used as flavoring or as a chelating agent) [bib_ref] Effect of silver nanomaterials on the activity of thiol-containing antioxidants, Zhou [/bib_ref] , pharmaceutical industry (it is part of several drug formulas, for instance in drugs used to reduce the levels of acetaldehyde in the oral cavity) [bib_ref] Buccal tablets containing cysteine and chlorhexidine for the reduction of acetaldehyde levels..., Juliano [/bib_ref] and the cosmetics industry (L-Cys as part of skin or hair care formulas) [bib_ref] Anti-inflammatory and anti-bacterial properties of tetramethylhexadecentyl succiniyl cysteine (TSC): A skin-protecting cosmetic..., Fernández [/bib_ref]. Cysteine can be assimilated through different pathways depending on the needs of the cells, giving rise to sulfur compounds [bib_ref] Metabolism of sulfur-containing amino acids, Stipanuk [/bib_ref] [bib_ref] Cysteine concentration regulates cysteine metabolism to glutathione, sulphate and taurine in rat..., Stipanuk [/bib_ref]. The main metabolite obtained by the assimilation of Cys is sulfinate, a molecule that is metabolized to give rise to sulfinylpyruvate and pyruvate, or hipotaurine and taurine [fig_ref] Figure 2: Summary of cysteine metabolism[8] [/fig_ref]. On the one hand, taurine is a fairly abundant molecule at the intracellular level. Although the biological role of taurine is not at all clear, some studies suggest that it may be a compound involved in the nervous system, being especially necessary for brain development [bib_ref] Open questions concerning taurine with emphasis on the brain, Oja [/bib_ref]. This suggestion is supported by the fact that high levels of taurine have been found in the fetal brain [bib_ref] Taurine enhances proliferation and promotes neuronal specification of murine and human neural..., Pasantes-Morales [/bib_ref]. In addition, taurine displays many other biological roles, such as the marking of certain toxic intermediaries or the regulation of intracellular calcium levels. On the other hand, the sulfinylpyruvate produced during metabolism of Cys may undergo oxidation to sulfate, which is later used in the synthesis of 3 -phosphoadenosine-5 -phosphosulfate. There are many other important reactions occurring during the metabolism of Cys that allow the synthesis of thiocysteine or the transfer of sulfur between molecules. Cysteine can be assimilated through different pathways depending on the needs of the cells, giving rise to sulfur compounds [bib_ref] Metabolism of sulfur-containing amino acids, Stipanuk [/bib_ref] [bib_ref] Cysteine concentration regulates cysteine metabolism to glutathione, sulphate and taurine in rat..., Stipanuk [/bib_ref]. The main metabolite obtained by the assimilation of Cys is sulfinate, a molecule that is metabolized to give rise to sulfinylpyruvate and pyruvate, or hipotaurine and taurine [fig_ref] Figure 2: Summary of cysteine metabolism[8] [/fig_ref]. On the one hand, taurine is a fairly abundant molecule at the intracellular level. Although the biological role of taurine is not at all clear, some studies suggest that it may be a compound involved in the nervous system, being especially necessary for brain development [bib_ref] Open questions concerning taurine with emphasis on the brain, Oja [/bib_ref]. This suggestion is supported by the fact that high levels of taurine have been found in the fetal brain [bib_ref] Taurine enhances proliferation and promotes neuronal specification of murine and human neural..., Pasantes-Morales [/bib_ref]. In addition, taurine displays many other biological roles, such as the marking of certain toxic intermediaries or the regulation of intracellular calcium levels. On the other hand, the sulfinylpyruvate produced during metabolism of Cys may undergo oxidation to sulfate, which is later used in the synthesis of 3′-phosphoadenosine-5′-phosphosulfate. There are many other important reactions occurring during the metabolism of Cys that allow the synthesis of thiocysteine or the transfer of sulfur between molecules. The human genome encodes about 214,000 Cys-coding sequences [bib_ref] The Cysteine Proteome. Free Radic, Go [/bib_ref]. Thus, the presence of this amino acid in proteins is significant. The reactivity and diverse functions of Cys are mirrored by a spectrum of susceptibilities and dysfunctions of their respective proteins, resulting in central roles for the Cys proteome in development, signal transduction, biologic defenses, aging, and disease [bib_ref] The Cysteine Proteome. Free Radic, Go [/bib_ref] [bib_ref] Glutathione homeostasis and redox-regulation by sulfhydryl groups, Meyer [/bib_ref] [bib_ref] Chemistry and proteomics of cysteine modifications in redox biology, Kim [/bib_ref]. Due to these important implications, L-Cys has attracted much attention of medical researchers and professionals worldwide (mainly in the USA and Europe). Over the past few years, important biological functions of L-Cys in human beings have been described for the first time [bib_ref] The Cysteine Proteome. Free Radic, Go [/bib_ref] [bib_ref] Glutathione homeostasis and redox-regulation by sulfhydryl groups, Meyer [/bib_ref] [bib_ref] Metabolism of sulfur-containing amino acids, Stipanuk [/bib_ref] [bib_ref] Cysteine metabolism and metal toxicity, Quig [/bib_ref] [bib_ref] Redox Signaling Regulated by Cysteine Persulfide and Protein Polysulfidation, Kasamatsu [/bib_ref]. At the same time, the potential applications of this compound at the industrial level have been increasing, mainly in fields related to pharmaceuticals, medicines, and nutraceuticals. As examples of these applications, L-Cys plays an important role in the food industry (it is used as flavoring or as a chelating agent) [bib_ref] Effect of silver nanomaterials on the activity of thiol-containing antioxidants, Zhou [/bib_ref] , pharmaceutical industry (it is part of several drug formulas, for instance in drugs used to reduce the levels of acetaldehyde in the oral cavity) [bib_ref] Buccal tablets containing cysteine and chlorhexidine for the reduction of acetaldehyde levels..., Juliano [/bib_ref] and the cosmetics industry (L-Cys as part of skin or hair care formulas) [bib_ref] Anti-inflammatory and anti-bacterial properties of tetramethylhexadecentyl succiniyl cysteine (TSC): A skin-protecting cosmetic..., Fernández [/bib_ref]. The human genome encodes about 214,000 Cys-coding sequences [bib_ref] The Cysteine Proteome. Free Radic, Go [/bib_ref]. Thus, the presence of this amino acid in proteins is significant. The reactivity and diverse functions of Cys are mirrored by a spectrum of susceptibilities and dysfunctions of their respective proteins, resulting in central roles for the Cys proteome in development, signal transduction, biologic defenses, aging, and disease [bib_ref] The Cysteine Proteome. Free Radic, Go [/bib_ref] [bib_ref] Glutathione homeostasis and redox-regulation by sulfhydryl groups, Meyer [/bib_ref] [bib_ref] Chemistry and proteomics of cysteine modifications in redox biology, Kim [/bib_ref]. Due to these important implications, L-Cys has attracted much attention of medical researchers and professionals worldwide (mainly in the USA and Europe). Over the past few years, important biological functions of L-Cys in human beings have been described for the first time [bib_ref] The Cysteine Proteome. Free Radic, Go [/bib_ref] [bib_ref] Glutathione homeostasis and redox-regulation by sulfhydryl groups, Meyer [/bib_ref] [bib_ref] Metabolism of sulfur-containing amino acids, Stipanuk [/bib_ref] [bib_ref] Cysteine metabolism and metal toxicity, Quig [/bib_ref] [bib_ref] Redox Signaling Regulated by Cysteine Persulfide and Protein Polysulfidation, Kasamatsu [/bib_ref]. At the same time, the potential applications of this compound at the industrial level have been increasing, mainly in fields related to pharmaceuticals, medicines, and nutraceuticals. As examples of these applications, L-Cys plays an important role in the food industry (it is used as flavoring or as a chelating agent) [bib_ref] Effect of silver nanomaterials on the activity of thiol-containing antioxidants, Zhou [/bib_ref] , pharmaceutical industry (it is part of several drug formulas, for instance in drugs used to reduce the levels of acetaldehyde in the oral cavity) [bib_ref] Buccal tablets containing cysteine and chlorhexidine for the reduction of acetaldehyde levels..., Juliano [/bib_ref] and the cosmetics industry (L-Cys as part of skin or hair care formulas) [bib_ref] Anti-inflammatory and anti-bacterial properties of tetramethylhexadecentyl succiniyl cysteine (TSC): A skin-protecting cosmetic..., Fernández [/bib_ref]. Medicines and pharmaceuticals based on natural products are currently in an important period of expansion worldwide. Not only patients but also professionals in medicines, pharmaceuticals, and cosmetics look for new formulas and less aggressive therapies, in many cases integrating compounds/drugs of biological origin [bib_ref] Nitrogenus compounds of interest in clinical nutrition, Fontana [/bib_ref]. Thus, "personalized medicine", based on drugs of natural origin as well as clinical nutrition, has experienced a boom in Western countries during the last few decades [bib_ref] Nitrogenus compounds of interest in clinical nutrition, Fontana [/bib_ref]. Besides, advances in biotechnology have promoted the design of a huge number of drugs and supplements for the prevention and the treatment of various diseases [bib_ref] Nitrogenus compounds of interest in clinical nutrition, Fontana [/bib_ref] [bib_ref] Trends in alternative medicine use in the United States, 1990-1997: Results of..., Eisenberg [/bib_ref]. L-Cys is a compound widely used in the development of numerous drugs so far. However, the number of clinical trials testing the effects of L-Cys on human health and wellness is currently scarce. So, the impact of L-Cys usage (as part of drugs or as dietary supplements) on human health is a controversial issue. This original work is based on a systematic review of the scientific studies published on L-Cys using different quality criteria to select the accurate and useful information. The analysis presented here sheds light on the role and potential positive effects of its use in the prevention and treatment of diseases, as well as in the improvement and maintenance of health status. ## Objectives The main objective is to carry out an exhaustive, accurate, and rigorous analysis of the information available in scientific databases to identify the effects of the usage of L-Cys in human health. The analysis considers the use/administration of L-Cys as part of the formulations of drugs, nutraceuticals, or through food supplementation. This objective can be divided into two sub-objectives:To compile information about the use of L-Cys in medicines and nutritional therapies (including clinical trials); (2) To identify the benefits reported by L-Cys on human health. # Results ## Compilation of bibliographic sources of interest After a general search through Google Scholar, the major search engines PubMed, Scopus, and Web of Science were used to compile information about L-Cys available up to December 2017. This search indicated that the number of original papers identified through PubMed was significantly higher than the other two databases mentioned. Consequently, PubMed was the database ultimately chosen to identify and select the documents of interest for this systematic review. A total of 126,919 studies were thereby identified using the keyword "L-cysteine" (122,195 of them in English). The analysis of the documents reveals that the earliest publications focusing on L-Cys can be traced back to the first third of the 20th century. Nevertheless, the number of scientific publications concerning L-Cys has mainly increased during the last four decades . Medicines and pharmaceuticals based on natural products are currently in an important period of expansion worldwide. Not only patients but also professionals in medicines, pharmaceuticals, and cosmetics look for new formulas and less aggressive therapies, in many cases integrating compounds/drugs of biological origin [bib_ref] Nitrogenus compounds of interest in clinical nutrition, Fontana [/bib_ref]. Thus, "personalized medicine", based on drugs of natural origin as well as clinical nutrition, has experienced a boom in Western countries during the last few decades [bib_ref] Nitrogenus compounds of interest in clinical nutrition, Fontana [/bib_ref]. Besides, advances in biotechnology have promoted the design of a huge number of drugs and supplements for the prevention and the treatment of various diseases [bib_ref] Nitrogenus compounds of interest in clinical nutrition, Fontana [/bib_ref] [bib_ref] Trends in alternative medicine use in the United States, 1990-1997: Results of..., Eisenberg [/bib_ref]. L-Cys is a compound widely used in the development of numerous drugs so far. However, the number of clinical trials testing the effects of L-Cys on human health and wellness is currently scarce. So, the impact of L-Cys usage (as part of drugs or as dietary supplements) on human health is a controversial issue. This original work is based on a systematic review of the scientific studies published on L-Cys using different quality criteria to select the accurate and useful information. The analysis presented here sheds light on the role and potential positive effects of its use in the prevention and treatment of diseases, as well as in the improvement and maintenance of health status. ## Objectives The main objective is to carry out an exhaustive, accurate, and rigorous analysis of the information available in scientific databases to identify the effects of the usage of L-Cys in human health. The analysis considers the use/administration of L-Cys as part of the formulations of drugs, nutraceuticals, or through food supplementation. This objective can be divided into two subobjectives:To compile information about the use of L-Cys in medicines and nutritional therapies (including clinical trials); (2) To identify the benefits reported by L-Cys on human health. # Results ## Compilation of bibliographic sources of interest After a general search through Google Scholar, the major search engines PubMed, Scopus, and Web of Science were used to compile information about L-Cys available up to December 2017. This search indicated that the number of original papers identified through PubMed was significantly higher than the other two databases mentioned. Consequently, PubMed was the database ultimately chosen to identify and select the documents of interest for this systematic review. A total of 126,919 studies were thereby identified using the keyword "L-cysteine" (122,195 of them in English). The analysis of the documents reveals that the earliest publications focusing on L-Cys can be traced back to the first third of the 20th century. Nevertheless, the number of scientific publications concerning L-Cys has mainly increased during the last four decades . In total, 92% of the identified studies were descriptions of the physico-chemical properties of L-Cys or its biological roles. From the total, 8304 (6.54%) studies correspond to systematic reviews and 1638 (1.29%) are clinical trials. A total of 60,885 primary publications (47.97% of the total identified) were ultimately selected to comprise the database used in this research. These publications were selected considering the selection and exclusion criteria indicated in the materials and methods section, and combining the previous keyword (L-cysteine) with other terms of interest for this study ("L-cysteine & human health"; "L-cysteine & pharmacology"; "L-cysteine & food processing"; "L-cysteine & nutritional therapy"; "L-cysteine & nutrition") [fig_ref] Table 1: Number of publications included in the database for this work [/fig_ref]. ## Bibliometric and bibliographic analysis The increasing interest worldwide for natural compounds and their application in medicines, cosmetics, pharmaceuticals, and nutrition has impacted L-Cys research too. As an example, it is important to highlight that the number of papers reporting positive results from the usage of L-Cys in medicines and nutritional therapies has significantly increased during the last decade. As is displayed in [fig_ref] Table 1: Number of publications included in the database for this work [/fig_ref] , the United States and United Kingdom are the main countries from which the selected contributions originate (about 30% each), followed by Japan. Italy, France, and Spain are the European countries with the highest number of studies in this field at the time of writing this manuscript. The bibliometric analysis of the selected information states that the greatest number of scientific-technical publications were within the field of pharmacology (52,873) [fig_ref] Table 1: Number of publications included in the database for this work [/fig_ref]. It is very striking that, despite the great scientific and technical production on L-Cys and its potential uses, there is no author/institution and/or journal that significantly highlights the topic. Therefore, it could be concluded that it is a topic of great interest analyzed from very different perspectives and for multiple applications. The increase in the number of publications about L-Cys during the last two decades is meaningful and relevant . To understand it, some reflections must be made about the recent boom in "personalized medicine" globally as well as about nutritional habits in developed countries. Interest in personalized medicine increased significantly between the end of the 20th century and the beginning of the 21st century as a result of: (i) a great discontent with traditional treatments in certain segments of the population (some traditional treatments in occidental and oriental medicine are unable to cure or relieve symptoms in many cases); (ii) personalized treatments being highly demanded in developed countries; and (iii) the rejection of the population regarding chemically synthesized drugs becoming significant. In addition, the recognition of nutrition as a key factor for maintaining and restoring health has also contributed to the general interest for compounds of natural origin in general, and for L-Cys and other amino acids in particular. In total, 92% of the identified studies were descriptions of the physico-chemical properties of L-Cys or its biological roles. From the total, 8304 (6.54%) studies correspond to systematic reviews and 1638 (1.29%) are clinical trials. A total of 60,885 primary publications (47.97% of the total identified) were ultimately selected to comprise the database used in this research. These publications were selected considering the selection and exclusion criteria indicated in the materials and methods section, and combining the previous keyword (L-cysteine) with other terms of interest for this study ("L-cysteine & human health"; "L-cysteine & pharmacology"; "L-cysteine & food processing"; "Lcysteine & nutritional therapy"; "L-cysteine & nutrition") [fig_ref] Table 1: Number of publications included in the database for this work [/fig_ref]. ## Review of clinical trials on effects of ## Bibliometric and bibliographic analysis The increasing interest worldwide for natural compounds and their application in medicines, cosmetics, pharmaceuticals, and nutrition has impacted L-Cys research too. As an example, it is important to highlight that the number of papers reporting positive results from the usage of L-Cys in medicines and nutritional therapies has significantly increased during the last decade. As is displayed in [fig_ref] Table 1: Number of publications included in the database for this work [/fig_ref] , the United States and United Kingdom are the main countries from which the selected contributions originate (about 30% each), followed by Japan. Italy, France, and Spain are the European countries with the highest number of studies in this field at the time of writing this manuscript. The bibliometric analysis of the selected information states that the greatest number of scientifictechnical publications were within the field of pharmacology (52,873) [fig_ref] Table 1: Number of publications included in the database for this work [/fig_ref]. It is very striking that, despite the great scientific and technical production on L-Cys and its potential uses, there is no author/institution and/or journal that significantly highlights the topic. Therefore, it could be concluded that it is a topic of great interest analyzed from very different perspectives and for multiple applications. The increase in the number of publications about L-Cys during the last two decades is meaningful and relevant . To understand it, some reflections must be made about the recent boom in "personalized medicine" globally as well as about nutritional habits in developed countries. Interest in personalized medicine increased significantly between the end of the 20th century and the beginning of the 21st century as a result of: (i) a great discontent with traditional treatments in certain segments of the population (some traditional treatments in occidental and oriental medicine are unable to cure or relieve symptoms in many cases); (ii) personalized treatments being highly demanded in developed countries; and (iii) the rejection of the population regarding chemically synthesized drugs becoming significant. In addition, the recognition of nutrition as a key factor for maintaining and restoring health has also contributed to the general interest for compounds of natural origin in general, and for L-Cys and other amino acids in particular. ## Review of clinical trials on effects of l-cysteine on human health Special attention has been paid to research based on recently published clinical trials to identify the possible advantages and disadvantages arising from the use of this amino acid in the context of medicines and nutritional supplements. A total of 1638 clinical trials were selected using the keyword combinations [fig_ref] Table 1: Number of publications included in the database for this work [/fig_ref]. In total, 1431 of these studies showed significant evidence (both negative and positive) on the uses of the L-Cys alone or in combination with other compounds such as vitamin D ## -cysteine on human health Special attention has been paid to research based on recently published clinical trials to identify the possible advantages and disadvantages arising from the use of this amino acid in the context of medicines and nutritional supplements. A total of 1638 clinical trials were selected using the keyword combinations [fig_ref] Table 1: Number of publications included in the database for this work [/fig_ref]. In total, 1431 of these studies showed significant evidence (both negative and positive) on the uses of the L-Cys alone or in combination with other compounds such as vitamin D or glycine. To make the reading of this work more understandable, the benefits of the usage of L-Cys in human beings have been summarized and classified into two tables as follows: summarizes the main effects of the usage of L-Cys (alone) on human beings; displays the benefits of its usage for human beings when the amino acid is combined with other molecules such as vitamin D or glycine. Comments on negative effects of the usage of L-Cys are also summarized after . Regarding the use of L-Cys as part of drug formulas or nutritional supplements, the following benefits have been reported : antioxidant power, regulation of the mucolytic function, strengthening of the hair, improvement of the functions of the immune system, protection and detoxification of the liver, promotion or elimination of heavy metals, prevention of heart disease, diabetes prevention, delay of aging, and the protection of the digestive system. Some studies suggest that there is no clear effect (neither positive nor negative) of L-Cys on human health whilst some recent works state that L-Cys could behave as a negative modulator on GABAergic neurotransmission [bib_ref] Negative modulation of the GABAA ρ1 receptor function by l-cysteine, Beltrán González [/bib_ref]. . Summary of the potential effects of L-Cys (alone) supported by studies based on clinical trials. ## Examples of l-cysteine usage effects Ref. Nutritional therapy in children with severe edematous malnutrition Restoration of the rate of synthesis and the concentration of glutathione during the first phase of treatment [bib_ref] Cysteine supplementation improves the erythrocyte glutathione synthesis rate in children with severe..., Badaloo [/bib_ref] Scarring of the cornea after a photoreactive keratectomy Reduced average time of scarring [bib_ref] Role of cysteine in corneal wound healing after photorefractive keratectomy, Meduri [/bib_ref] Nutritional therapy in Ictus patients Reduced risk of cardiovascular accident [bib_ref] Dietary cysteine and other amino acids and stroke incidence in women, Larsson [/bib_ref] Hair care Reduced hair loss and increased hair strengthening abilities [bib_ref] The efficacy of drug therapy in structural lesions of the hair and..., Petri [/bib_ref] Protection of digestive system Reduction in the concentration of acetaldehyde by avoiding exposure in cases of achlorhydria [bib_ref] Reducing carcinogenic acetaldehyde exposure in the achlorhydric stomach with cysteine, Linderborg [/bib_ref] Treatment chronic inflammation Increased antioxidant status [bib_ref] Clinical and nutritional benefits of cysteine-enrich protein supplements, Mcpherson [/bib_ref] Prevention of upper digestive tract cancer and breast cancer Decrease of acetaldehyde in saliva or it can be used as part of metabolic starvation therapy [bib_ref] Removal of acetaldehyde from saliva by a slow-release buccal tablet of L-cysteine, Salaspuro [/bib_ref] [bib_ref] Nonessential amino acid metabolism in breast cancer, Geck [/bib_ref] Indicator for the control of cardiovascular diseases Pro-inflammatory signaling [bib_ref] Cysteine/cystine redox signalling in cardiovascular disease. Free Radic, Go [/bib_ref] Treatment of erythropoietic porphyria Photosensitivity improvement [bib_ref] Long-term treatment of erythropoietic protopotphyria with cysteine, Mathews-Roth [/bib_ref] Treatment of type-2 diabetes Control of glycaemia and vascular inflammation [bib_ref] L-cysteine supplementation as an adjuvant therapy for type-2 diabetes, Jain [/bib_ref] [bib_ref] Cysteine and hydrogen sulphide in the regulation of metabolism: Insights from genetics..., Carter [/bib_ref] [bib_ref] Vitamin d and L-cysteine levels correlate positively with GSH and negatively with..., Jain [/bib_ref]. Summary of the potential effects of L-Cys (combined with other compounds) supported by studies based on clinical trials. ## Composition of the mixture examples of usage effects Ref. ## L-cysteine + glycine Nutritional therapy in elderly HIV + patients Improved oxidation of carbohydrates, insulin sensitivity and body composition [bib_ref] Effect of Increasing glutathione with cysteine and glycine supplementation on mitochondrial fuel..., Nguyen [/bib_ref] [bib_ref] Deficient synthesis of glutathione underlies oxidative stress in aging and can be..., Sekhar [/bib_ref] Treatment of oxidative stress during aging Increased synthesis of glutathione and decreases oxidative stress levels L-Cysteine + glycine + dithreonine Treatment of hypostatic ulcer Reduced pain and improved degree of ulcer healing [bib_ref] L-cysteine, glycine and di-threonine in the treatment of hypostatic leg ulceration: A..., Harvey [/bib_ref] L-Cysteine + vitamin D Treatment of patients with type-2 diabetes Increased levels of glutathione and decreased levels of triglycerides [bib_ref] Vitamin d and L-cysteine levels correlate positively with GSH and negatively with..., Jain [/bib_ref] L-Cysteine + basic fibroblast growth factor (bFCF) Treatment of corneal epithelium after photoreactive keratectomy in patients affected by myopia Reduced time of resurfacing corneal. [bib_ref] Effect of the combination of basic fibroblast growth factor and cysteine on..., Meduri [/bib_ref] L-Cysteine + theanine Improvement of well-trained athletes' performance Restoration of the attenuation of the activity of Natural Killer cells [bib_ref] Cystine and theanine supplementation restores high-intensity resistance exercise-induced attenuation of natural killer..., Kawada [/bib_ref] Although some prominent clinical trials have tested the effects of L-Cys, most of them have analyzed the effects of N-acetyl-L-cysteine (NAC) instead [bib_ref] N-acetylcysteine: Pharmacological considerations and experimental and clinical applications, Cotgreave [/bib_ref] [bib_ref] The chemistry and biological activities of N-acetylcysteine, Samuni [/bib_ref]. This compound [fig_ref] Figure 4: Chemical structure of N-acetyl-L-cysteine [/fig_ref] is a precursor of L-Cys that promotes glutathione biosynthesis [bib_ref] A minireview on N-acetylcysteine: An old drug with new approaches, Dhouib [/bib_ref]. It acts directly as a scavenger of free radicals, mainly oxygen radicals. Consequently, it is a powerful antioxidant useful for treating several disorders that result from the generation of free oxygen radicals. Additionally, it is a highly efficient mucolytic drug promoting tenacious mucous discharges [bib_ref] A Review on Various Uses of N-Acetyl Cysteine, Mokhtari [/bib_ref]. The main effects of its usage in human beings are summarized in [fig_ref] Table 4: Summary of the potential effects of NAC supported by studies based on... [/fig_ref]. It acts as an anti-relapse agent in abstinent subjects [bib_ref] N-acetylcysteine for treating cocaine addiction-A systematic review, Echevarria [/bib_ref] [bib_ref] N-acetylcysteine reverses cocaine-induced metaplasticity, Moussawi [/bib_ref] In brief, NAC supplementation has exerted favorable effects on vascular health, muscle strength, bone density, cell-mediated immunity, preservation of cognitive function, or marking systemic inflammation [bib_ref] Multiple clinical applications, Millea [/bib_ref] [bib_ref] N-acetylcysteine for treating cocaine addiction-A systematic review, Echevarria [/bib_ref] [bib_ref] The effect of N-acetylcysteine (NAC) on human cognition-A systematic review, Skvarc [/bib_ref] [bib_ref] N-acetylcysteine in the treatment of psychiatric disorders: Current status and future prospects, Minarini [/bib_ref] [bib_ref] N-acetylcysteine in depressive symptoms and functionality: A systematic review and meta-analysis, Fernandes [/bib_ref] [bib_ref] Basic and clinical pharmacology, Arakawa [/bib_ref]. Although positive effects of NAC have been reported for treatments of patients with acute liver failure [bib_ref] Effects of N-acetylcysteine on cytokines in non-acetaminophen acute liver failure: Potential mechanism..., Stravitz [/bib_ref] , other studies on acute liver failure and liver surgery suggest patients randomized to postoperative NAC received no benefit [bib_ref] Nottingham HPB Surgery Group. Current evidence for the use of N-acetylcysteine following..., Kemp [/bib_ref] [bib_ref] Effect of N-acetylcysteine on liver recovery after resection: A randomized clinical trial, Grendar [/bib_ref] [bib_ref] N-acetylcysteine administration does not improve patient outcome after liver resection, Robinson [/bib_ref]. Finally, NAC is part of the paracetamol formula (acetaminophen). When paracetamol is taken in large quantities, a minor metabolite called N-acetyl-p-benzoquinone imine (NAPQI) accumulates within the body. It is normally conjugated by glutathione, but when taken in excess, the body's glutathione reserves are not sufficient to deactivate the toxic NAPQI. This metabolite is then free to react with key hepatic enzymes, thereby damaging liver cells. This may lead to severe liver damage and even death by acute liver failure. This phenomenon is commonly named paracetamol poisoning [bib_ref] Oral and Intravenous Acetylcysteine for Treatment of Acetaminophen Toxicity: A Systematic Review..., Green [/bib_ref] [bib_ref] Evidence for the changing regimens of acetylcysteine, Chiew [/bib_ref] [bib_ref] Management of paracetamol overdose: Current controversies, Kozer [/bib_ref]. # Discussion The review of all selected documents, and particularly the clinical trials, has enabled identification of the effects of the usage of L-Cys in human health and wellness, either by its application individually or combined with other compounds. The clearly demonstrated benefits on human health are as follows: -Antioxidant role. L-Cys acts as a precursor for the synthesis of glutathione, which is an important antioxidant. The reduced form of glutathione plays a fundamental role in the defense of the organism against damage caused by oxidative stress [bib_ref] El glutatión y su asociación con las enfermedades neurodegenerativas, la esquizofrenia, el..., Martínez-Sámano [/bib_ref]. This property is due to its ability to neutralize reactive particles that can cause damage to cells and tissues. Thus, diet supplementation with L-Cys restores the synthesis of glutathione in cases in which it has been ## Examples of usage effects Ref. Treatment of methamphetamine-dependent patients Methamphetamine dependence decreases [bib_ref] The efficacy of N-acetylcysteine in the treatment of methamphetamine depended: A double-blind..., Mousavi [/bib_ref] Performance of athletes undergoing strenuous physical training Redox equilibrium and adaptation processes improve [bib_ref] Effect of N-acetylcysteine on cycling performance after intensified training, Slattery [/bib_ref] Treatment of Thalassemia Oxidative stress and DNA damage decrease [bib_ref] N-acetylcysteine supplementation reduces oxidative stress and DNA damage in children with β-thalassemia, Ozdemir [/bib_ref] [bib_ref] Role of iron in inducing oxidative stress in thalassemia: Can it be..., Rachmilewitz [/bib_ref] Protection against the carcinogenic effect of tobacco Modulation of biomarkers associated with cancer [bib_ref] Effects of oral administration of N-acetyl-L-cysteine: A multi-biomarker study in smokers, Van Schooten [/bib_ref] Treatment of bacterial meningitis Antioxidant role [bib_ref] N-acetyl-L-cysteine as a therapeutic option in bacterial meningitis, Klein [/bib_ref] [bib_ref] Multiple clinical applications, Millea [/bib_ref] Treatment against influenza virus Proliferation of the virus is inhibited [bib_ref] Multiple clinical applications, Millea [/bib_ref] [bib_ref] Antioxidant therapy as a potential approach to severe influenza-associated complications, Uchide [/bib_ref] Mucolytic expectorant and treatment of respiratory tract infections The viscosity decreases and facilitates the removal of mucus [bib_ref] The chemistry and biological activities of N-acetylcysteine, Samuni [/bib_ref] [bib_ref] The effect of N-acetylcysteine on biofilms: Implications for the treatment of respiratory..., Blasi [/bib_ref] Specific antidote for acetaminophen overdose Regeneration of glutathione levels [bib_ref] Multiple clinical applications, Millea [/bib_ref] [bib_ref] Intravenous acetylcysteine for indications other than acetaminophen overdose, Bass [/bib_ref] [bib_ref] Management of acetaminophen toxicity, Larsen [/bib_ref] [bib_ref] Assessment of the clinical use of intravenous and oral N-acetylcysteine in the..., Blackford [/bib_ref] Cardiovascular complications in patients with diabetes Treatment of type-2 diabetes Attenuation of cardiovascular complications [bib_ref] Prevention of diabetes-induced cardiovascular complications upon treatment with antioxidants, Xu [/bib_ref] [bib_ref] A review on the possible molecular mechanism of action of N-acetylcysteine against..., Lasram [/bib_ref] [bib_ref] Effect of N-acetyl-L-cysteine on insulin resistance caused by prolonged free fatty acid..., Pereira [/bib_ref] Prevention of cardiovascular diseases Reduction of plasma concentrations and homocysteine levels [bib_ref] Oral N-acetylcysteine reduces plasma homocysteine concentrations regardless of lipid or smoking status, Hildebrandt [/bib_ref] Treatment of chronic hepatitis C Increase in glutathione and improvement in response to treatment with interferon [bib_ref] Therapy of hepatitis C: Other options, Bonkovsky [/bib_ref] [bib_ref] Hepatic inflammation and progressive liver fibrosis in chronic liver disease, Czaja [/bib_ref] Treatment of patients with acute liver failure Reduced IL-17 levels [bib_ref] Effects of N-acetylcysteine on cytokines in non-acetaminophen acute liver failure: Potential mechanism..., Stravitz [/bib_ref] [bib_ref] Nottingham HPB Surgery Group. Current evidence for the use of N-acetylcysteine following..., Kemp [/bib_ref] [bib_ref] Effect of N-acetylcysteine on liver recovery after resection: A randomized clinical trial, Grendar [/bib_ref] [bib_ref] N-acetylcysteine administration does not improve patient outcome after liver resection, Robinson [/bib_ref] Treatment of nephropathic cystinosis Reduced oxidative stress and improved renal function [bib_ref] N-acetyl-cysteine is associated to renal function improvement in patients with nephropathic cystinosis, De Faria Guimaraes [/bib_ref] Treatment of noise-induced hearing loss Protective effect. Hearing loss is reduced [bib_ref] Comparison of the effects of N-acetyl-cysteine and ginseng in prevention of noise..., Doost [/bib_ref] Treatment of cocaine addiction It acts as an anti-relapse agent in abstinent subjects [bib_ref] N-acetylcysteine for treating cocaine addiction-A systematic review, Echevarria [/bib_ref] [bib_ref] N-acetylcysteine reverses cocaine-induced metaplasticity, Moussawi [/bib_ref] In brief, NAC supplementation has exerted favorable effects on vascular health, muscle strength, bone density, cell-mediated immunity, preservation of cognitive function, or marking systemic inflammation [bib_ref] Multiple clinical applications, Millea [/bib_ref] [bib_ref] N-acetylcysteine for treating cocaine addiction-A systematic review, Echevarria [/bib_ref] [bib_ref] The effect of N-acetylcysteine (NAC) on human cognition-A systematic review, Skvarc [/bib_ref] [bib_ref] N-acetylcysteine in the treatment of psychiatric disorders: Current status and future prospects, Minarini [/bib_ref] [bib_ref] N-acetylcysteine in depressive symptoms and functionality: A systematic review and meta-analysis, Fernandes [/bib_ref] [bib_ref] Basic and clinical pharmacology, Arakawa [/bib_ref]. Although positive effects of NAC have been reported for treatments of patients with acute liver failure [bib_ref] Effects of N-acetylcysteine on cytokines in non-acetaminophen acute liver failure: Potential mechanism..., Stravitz [/bib_ref] , other studies on acute liver failure and liver surgery suggest patients randomized to postoperative NAC received no benefit [bib_ref] Nottingham HPB Surgery Group. Current evidence for the use of N-acetylcysteine following..., Kemp [/bib_ref] [bib_ref] Effect of N-acetylcysteine on liver recovery after resection: A randomized clinical trial, Grendar [/bib_ref] [bib_ref] N-acetylcysteine administration does not improve patient outcome after liver resection, Robinson [/bib_ref]. Finally, NAC is part of the paracetamol formula (acetaminophen). When paracetamol is taken in large quantities, a minor metabolite called N-acetyl-p-benzoquinone imine (NAPQI) accumulates within the body. It is normally conjugated by glutathione, but when taken in excess, the body's glutathione reserves are not sufficient to deactivate the toxic NAPQI. This metabolite is then free to react with key hepatic enzymes, thereby damaging liver cells. This may lead to severe liver damage and even death by acute liver failure. This phenomenon is commonly named paracetamol poisoning [bib_ref] Oral and Intravenous Acetylcysteine for Treatment of Acetaminophen Toxicity: A Systematic Review..., Green [/bib_ref] [bib_ref] Evidence for the changing regimens of acetylcysteine, Chiew [/bib_ref] [bib_ref] Management of paracetamol overdose: Current controversies, Kozer [/bib_ref]. # Discussion The review of all selected documents, and particularly the clinical trials, has enabled identification of the effects of the usage of L-Cys in human health and wellness, either by its application individually or combined with other compounds. The clearly demonstrated benefits on human health are as follows: -Antioxidant role. L-Cys acts as a precursor for the synthesis of glutathione, which is an important antioxidant. The reduced form of glutathione plays a fundamental role in the defense of the organism against damage caused by oxidative stress [bib_ref] El glutatión y su asociación con las enfermedades neurodegenerativas, la esquizofrenia, el..., Martínez-Sámano [/bib_ref]. This property is due to its ability to neutralize reactive particles that can cause damage to cells and tissues. Thus, diet supplementation with L-Cys restores the synthesis of glutathione in cases in which it has been compromised, thus improving the redox balance and promoting the reduction of oxidative stress. In addition, the elimination of free radicals may also be associated with certain benefits, such as in the case of reduced healing time following certain surgical procedures (photorefractive keratectomy, for instance) [bib_ref] Role of cysteine in corneal wound healing after photorefractive keratectomy, Meduri [/bib_ref]. Besides, the antioxidant role of L-Cys is also related to reduced risk of a cerebrovascular accident [bib_ref] Dietary cysteine and other amino acids and stroke incidence in women, Larsson [/bib_ref] and reduced noise-induced hearing loss [bib_ref] Comparison of the effects of N-acetyl-cysteine and ginseng in prevention of noise..., Doost [/bib_ref]. -Mucolytic function. NAC causes a decrease in the viscosity of mucus secretion, thus facilitating their elimination. Bronchial secretions contain high concentrations of mucoproteins. The decrease in viscosity promoted by NAC is mainly due to the breakage of the disulfide bonds of the mucoproteins, resulting in a fragmentation of the chains from the mucins, immunoglobulins and serum albumin present in the mucous secretion [bib_ref] The chemistry and biological activities of N-acetylcysteine, Samuni [/bib_ref]. -Strengthening of hair. Blends fortified with L-Cys help to strengthen hair. Keratin is one of the most abundant proteins in the skin and the hair, and contains high amounts of L-Cys as building blocks. L-Cys forms disulfide bridges, which provide strength and rigidity to keratin. Consequently, the use of blends fortified with L-Cys promotes the repair of structural lesions and slows down hair loss experienced by patients affected by certain disorders (diffuse alopecia, for instance [bib_ref] The efficacy of drug therapy in structural lesions of the hair and..., Petri [/bib_ref]. -Regulation of the activity of the immune system. It has been reported that L-Cys can regulate immune system activity by promoting changes in levels of production of its effector molecules, as is the case of IL-17. IL-17 is a cytokine produced mainly by some T-cells, called Th17 cells, which act on epithelial cells and fibroblasts, as well as on other cells of the immune system. It has also been demonstrated that the administration of NAC can significantly improve transplant-free survival in patients affected by acute liver failure not related to acetaminophen when administered during the early stages of hepatic encephalopathy. This effect is due to the regulation of IL-17 production, which is closely related to the progression of the encephalopathy [bib_ref] Effects of N-acetylcysteine on cytokines in non-acetaminophen acute liver failure: Potential mechanism..., Stravitz [/bib_ref]. -Protection of the digestive system. Excessive alcohol consumption is considered a risk factor for the development of cancer in the upper gastrointestinal tract, because of exposure to acetaldehyde, which is carcinogenic to humans. In this sense, L-Cys intake reduces the concentration of acetaldehyde in saliva, thus decreasing the exposure of the gastrointestinal tract to this compound and, consequently, the risk of cancer [bib_ref] Removal of acetaldehyde from saliva by a slow-release buccal tablet of L-cysteine, Salaspuro [/bib_ref]. -Reducing the risk of stroke. Some studies suggest that consumption of certain amino acids, among which L-Cys, may be related to certain cardiovascular benefits, such as reduced arterial stiffness or reduced blood pressure, thereby fighting some risk factors related to vascular accidents in healthy women [bib_ref] Prevention of diabetes-induced cardiovascular complications upon treatment with antioxidants, Xu [/bib_ref]. Despite the high volume of evidence describing positive impacts of L-Cys on human health, it is worth noting that the number of works describing no effects or negative effects is also significant [bib_ref] Negative modulation of the GABAA ρ1 receptor function by l-cysteine, Beltrán González [/bib_ref] [bib_ref] An increased need for dietary cysteine in support of glutathione synthesis may..., Mccarty [/bib_ref] [bib_ref] Mechanisms of L-cysteine neurotoxicity, Janáky [/bib_ref]. Negative effects of L-Cys derivatives on human health have been reported [bib_ref] Nephrotoxicity of halogenated alkenyl cysteine-S-conjugates, Nagelkerke [/bib_ref] [bib_ref] Toxic, halogenated cysteine S-conjugates and targeting of mitochondrial enzymes of energy metabolism, Cooper [/bib_ref]. Thus, although most of the studies on L-Cys highlight its role in the homeostasis of redox status, some studies suggest that redox modulation is not involved during L-Cys actions and that L-Cys might act as a competitive antagonist of GABAA ρ1 receptors for instance [bib_ref] Negative modulation of the GABAA ρ1 receptor function by l-cysteine, Beltrán González [/bib_ref]. Also, some in vivo studies have shown that several S-conjugates are nephrotoxic and that the toxicity is associated with β-lyase-dependent bioactivation [bib_ref] Nephrotoxicity of halogenated alkenyl cysteine-S-conjugates, Nagelkerke [/bib_ref] ; in other cases, toxic effects of L-Cys on the nervous system have also been reported (oxidized L-Cys derivatives or compounds such as cysteine alpha-carbamate caused neuronal degeneration) [bib_ref] Mechanisms of L-cysteine neurotoxicity, Janáky [/bib_ref]. So, the controversy surrounding the benefits of supplements or drugs containing L-Cys is still open, mainly regarding their use for some specific treatments. In fact, many of the studies have been done in vitro and translational studies are still scarce. The following considerations can be underlined, taking into account this controversy and the results obtained from this review: (i) According to the results analyzed here, NAC is preferentially used instead of L-Cys. Besides, L-Cys is usually administrated as part of a formula containing other compounds, such as glycine, vitamin D, bFCF or theanine. So, the effects reported in these studies cannot be directly attributed solely to the L-Cys molecule. (ii) Most of the studies done at the time of writing this review are based on laboratory tests and cellular lines. (iii) The number of studies on the effects of L-Cys on human health based on clinical trials is limited and in many cases these are very preliminary studies in which the studied population does not reflect a standard population (considering aspects such as age, sex, etc.). (iv) Several important details related to L-Cys and NAC doses and metabolism remain unknown. Optimal L-Cys doses and safety concentration ranges according to some pathologies are far from known. In addition, optimization of the assimilation of L-Cys and its derivatives, when applied through nutritional supplements, is poorly described. # Materials and methods ## Search strategy and information processing First, several searches of generic character in the "Google Scholar" portal (https://scholar.google.es/) were carried out. In this search, several important secondary sources were identified which have been helpful for the design of this systematic review and for writing the introduction section. The second step in this work was the realization of a comprehensive bibliometric/bibliographic review through the major search engines available, PubMed, Scopus, and Web of Science (which are connected to specialized databases on the subject covered by this study). This search indicated that the number of original papers identified through PubMed was significantly higher than the other two databases mentioned. Consequently, PubMed was the database ultimately chosen to identify and select the documents of interest for this work. The search date was between January 1950 and December 2017 and included texts published in scientific journals in English and Spanish. The research questions were as follows: What role does L-cysteine play in pharmacology, human health, nutrition, and food processing? What is the effect of L-Cys on human health? The information retrieval system "boolean" was used to identify the works of interest for this review [bib_ref] Information retrieval: An overview of system characteristics, Wiesman [/bib_ref]. The keywords used were as follows: "L-cysteine"; "L-cysteine & human health"; "L-cysteine & pharmacology"; "L-cysteine & food processing"; "L-cysteine & nutritional therapy"; "L-cysteine & nutrition". These keywords were previously identified through the "MeSH" (Medical Subject Heading) database as descriptors for the realization of this work. The advanced search form was used in the PubMed database to identify the documents of interest. Additionally, the following options were selected: "Title/abstract", "article", "review". At a more advanced stage of the study, special attention was paid to clinical trials (most of them published over the last 25 years) to identify the main uses and possible effects of L-Cys in its application within the field of medicine. ## Data extraction and selection of relevant studies A total of 126,919 articles were identified but 60,885 were ultimately selected to constitute the database for this research. The selection criteria were as follows: primary sources, review/original articles, clinical trials, and studies with objective data on direct correlations between the use of L-Cys and effects on human health (adults, adolescents, and children). In all the selected studies, the direct correlations were established when L-Cys was naturally part of the food intake or as part of nutritional supplements. The exclusion criteria were as follows: articles of non-specific systematic review, meta-analyses of humans, or other non-human clinical trials with non-representative sample size. Each of the selected articles has been analyzed by the three authors with backgrounds in biology, biochemistry and nutrition. The classic scheme proposed by Vilanova has been used to assure the quality of the selection criteria. Thus, the following questions were considered to evaluate each of the identified works (quality criteria): is it related to the research objectives of this review? Is the methodology clear and objective? Is the study 100% reproducible? Is the sample size coherent? Is the sample well-defined? Is the sample representative? Is the hypothesis clearly stated? # Conclusions Over the past few years, several benefits have been attributed to L-Cys and NAC, contributing to their becoming useful compounds within the food, pharmaceuticals, medicines or cosmetics industries. A considerable number of cysteine-rich products that are mainly used in hair and skin care have been commercialized. More recently, L-Cys has emerged as an important molecule in the food supplements industry, a sector that is in full growth, possibly due to increasingly widespread concern in the populations of developed countries to address nutritional deficiencies caused by new habits of life and the inability, on certain occasions, to follow a balanced diet. Thus, nutritional therapy and functional foods have involved L-Cys as part of treatment of several pathologies such as cirrhosis (in which L-Cys biosynthesis is compromised) [bib_ref] Nutritional care for patients with liver cirrhosis, Aceves-Martins [/bib_ref] , or simply to prevent cancer or promote good health [bib_ref] Functional foods and their role in cancer prevention and health promotion: A..., Aghajanpour [/bib_ref] [bib_ref] L-Cysteine metabolism and its nutritional implications, Yin [/bib_ref]. Although several potential benefits have been attributed to L-Cys or NAC, only a few of these benefits have been clearly demonstrated. Some of these examples are (i) the use of these molecules as a mucolytic expectorant; (ii) positive effects on cirrhosis treatment or positive contributions in the maintenance of redox homeostasis in cells and in cancer treatment (through GSH synthesis). Another important aspect to point out is the potential use of L-Cys in processes related to food conservation and processing. L-Cys shows antioxidant, chelating, and flavoring properties which could be useful in food industries [bib_ref] Effect of silver nanomaterials on the activity of thiol-containing antioxidants, Zhou [/bib_ref] [bib_ref] Disulphide bond formation in food protein aggregation and gelation, Visschers [/bib_ref]. However, few applications of this amino acid are already available on the market (and only in developed countries). In summary, although the available literature on L-Cys and NAC is abundant, several key questions remain unaddressed. Consequently, more effort must be made in the near future to establish clear and direct connections between these two molecules and benefits on human health. These connections should be made not only based on in vitro analysis but also based on clinical trials carried out under quality standards and good practices as has been recently recommended [bib_ref] The need for registration of preclinical studies, Pavlica [/bib_ref]. [fig] Figure 1: Chemical structure (a) and structural model (b) of an L-cysteine molecule (adapted from[1]). [/fig] [fig] Figure 2: Summary of cysteine metabolism[8]. [/fig] [fig] 1: Number of publications included in the database for this work (Database: PubMed). * %: calculated over the total number of publications compiled from each combination of keywords.KeywordsTotal of Publications Publications in the Last 5 Years (%)* [/fig] [fig] Figure 4: Chemical structure of N-acetyl-L-cysteine (NAC). [/fig] [table] Table 4: Summary of the potential effects of NAC supported by studies based on clinical trials. [/table]

Application-aware deadline constraint job scheduling mechanism on large-scale computational grid Recently, computational Grids have proven to be a good solution for processing large-scale, computation intensive problems. However, the heterogeneity, dynamics of resources and diversity of applications requirements have always been important factors affecting their performance. In response to these challenges, this work first builds a Grid job scheduling architecture that can dynamically monitor Grid computing center resources and make corresponding scheduling decisions. Second, a Grid job model is proposed to describe the application requirements. Third, this paper studies the characteristics of commercial interconnection networks used in Grids and forecast job transmission time. Fourth, this paper proposes an application-aware job scheduling mechanism (AJSM) that includes periodic scheduling flow and a heuristic application-aware deadline constraint job scheduling algorithm. The rigorous performance evaluation results clearly demonstrate that the proposed application-aware job scheduling mechanism can successful schedule more Grid jobs than the existing algorithms. For successful scheduled jobs, our proposed AJSM method is the best algorithm for job average processing time and makespan. Grid job scheduling mechanisms aim to effectively exploit the benefits of Grids' idle computational resources by mapping jobs to appropriate Grid computing centers. This is a well-known NP-complete problem, in the general case, that can exhibit a huge search space of possible scheduling solutions[6][7][8]. The problem increases in complexity when the Grid computational resources are heterogeneous, dynamic, and even the load on Grid computing centers varies with time.Most classical Grid job scheduling strategies, including immediate mode and batch mode, are based on the assumption that the Grid resources provided by Grid computing centers are constant in a relatively long period[5,6]. However, for some actual Grid systems, such as the China National Grid, resources are seriously affected by the Grid computing center's local systems and Grid systems, and the number of resources changes dynamically [9]. For example, the number of computational nodes provided to the China National Grid by the Changsha National Supercomputing Center may go from 254 to 120 in space of a few minutes.On the other hand, Grid resources are heterogeneous not only in terms of hardware, such as multicore and manycore type, speed, network capacity, storage, and more, but also in terms of software type, license, version, and so on [10]. The Grid applications also have diverse resource requirements [11][12][13]. For example, automotive crash simulation analysis needs multidisciplinary finite element solver RADIOSS and pre-processing software HyperMesh, and CPU+GPU coordinated parallel computing [11,12]. These application requirements are not provided by all Grid computing centers. Furthermore, the Grid computing center's corresponding software license and number of available computing nodes may not meet application requirements. Some Grid computing centers choose not to support certain specific applications because of security issues, performance impact, or business strategies. Additionally, Grid application data transmission from the job submission point to the Grid computing center is a major challenge. This is owing to the fact that most Grid systems are connected by commercial interconnect networks, the communication bandwidth of which is highly affected by the environments. These application-aware issues are worth further investigation for the job scheduling mechanism.Motivated by these challenges, this paper designs and evaluates an application-aware job scheduling mechanism (AJSM) for Grids. The major contributions of this work are multifold and can be summarized as follows:- First, this paper constructs a Grid job scheduling architecture, including a Job Queue, Scheduler, Grid Resources Monitor, Network Prediction, Job Dispatch, and Grid environments, which can dynamically obtain Grid computing centers' idle resources and make job scheduling decisions.- Secondly, this paper studies the communication characteristics of Grid computing centers based on commercial interconnection networks, and adopt an ARIMA model to forecast data transfer bandwidth and job transmission time.- Thirdly, this paper builds a Grid job model to accommodate application requirements, and normalize the heterogeneous Grid computing resources as a standard multicore and manycore computational node model. This paper also formulates the application-aware deadline constraint job scheduling problem as a linear programming problem.- Fourthly, this paper proposes an application-aware job scheduling mechanism (AJSM), which mainly consists of periodic scheduling flow and a heuristic job scheduling algorithm. The heuristic scheduling algorithm first tries to find Grid computing centers that can satisfy Application-aware deadline constraint job scheduling mechanism on large-scale computational grid PLOS ONE | https://doi.org/10.Application-aware deadline constraint job scheduling mechanism on large-scale computational grid PLOS ONE | https://doi.org/10.Application-aware deadline constraint job scheduling mechanism on large-scale computational grid PLOS ONE | https://doi.org/10. # Introduction Computational Grids are a platform that can share, select, and aggregate geographically distributed heterogeneous idle computing resources to achieve vast computation and storage capabilities [bib_ref] Dynamic and Adaptive Fault Tolerant Scheduling With QoS Consideration in Computational Grid, Haider [/bib_ref]. In recent years, Grid techniques have been widely used in solving computation intensive problems in physics, genetics, astronomy, civil engineering, and among others [bib_ref] A grid architecture for scalable monitoring and enhanced dependable job scheduling, Bellavista [/bib_ref] [bib_ref] A Hierarchical Reliability-Driven Scheduling Algorithm in Grid Systems, Tang [/bib_ref]. The China National Grid, which consists of large Supercomputing centers, provinces or university computing nodes, is one example of such a computational Grid examples . Many large-scale computation intensive jobs, such as rice genome-wide association analysis, community earth system models, and large airliner CFD (Computational Fluid Dynamics) checks and auxiliary design, have utilized this Grid. To achieve the promised high computing jobs software requirements. Then, the ARIMA job transmission time is applied, and the algorithm schedules job to corresponding Grid computing center. - Finally, performance evaluation are conducted and the experimental results show that our proposed AJSM algorithm can successful schedule more Grid jobs than MGA, Min-Min. The AJSM method also outperforms existing algorithms in terms of job average processing time and makespan. The rest of the paper is organized as follows: related works are summarized in Section 2. The computational Grid model, job scheduling architecture, and job model are described in Section 3. In Section 4, this paper provides a Grid job transmission time prediction method based on an ARIMA model. The paper presents Grid heterogeneous computing nodes, scheduling attributes, and problems in Section 5. Section 6 proposes an application-aware job scheduling mechanism. Performance evaluation is given in Section 7, where the performance of AJSM is assessed in comparison with two similar algorithms. Finally, this paper summarize the contributions and comment on the future directions of this work in Section 8. # Related works Many effective heuristic and meta-heuristic Grid scheduling algorithms have been proposed to obtain near-optimal solutions, such as MET (Minimum Execution Time), Min-Min, Max-Min, and XSufferage [bib_ref] Budget-constraint Stochastic Task Scheduling on Heterogeneous Cloud Systems, Tang [/bib_ref]. The Min-Min heuristic algorithm tries to schedule job with overall minimum execution finish time. In contrast to Min-Min, the heuristic Max-Min algorithm chooses the job and Grid center pair with the maximum minimum execution finish time. Min-Min and Max-Min have been extended to adapt to different Grid job scheduling solutions. For example, Vaaheedha and Nazreen proposed a MiM-MaM algorithm, which combines Min-Min and Max-Min to overcome their drawbacks [bib_ref] MiM-MaM: A new task scheduling algorithm for grid environment, Vaaheedha [/bib_ref]. Bioinspired meta-heuristic algorithms are another class of scheduling mechanisms applied effectively to Grid [bib_ref] Improved auto control ant colony optimization using lazy ant approach for grid..., Tiwari [/bib_ref] [bib_ref] Enhancing the genetic-based scheduling in computational grids by a structured hierarchical population, Kołodziej [/bib_ref] [bib_ref] Maximizing availability for task scheduling in computational grid using genetic algorithm, Prakash [/bib_ref] [bib_ref] A genetic algorithm for independent job scheduling in grid computing, Younis [/bib_ref] [bib_ref] Observing the effect of interprocess communication in auto controlled ant colony optimization-based..., Tiwari [/bib_ref] [bib_ref] Cooperation and profit allocation in two-echelon logistics joint distribution network optimization, Wang [/bib_ref]. Liu et al. extended conventional particle swarm optimization particles' positions and velocities from real vectors to fuzzy matrices. This scheduling method can dynamically generate an optimal schedule solution [bib_ref] Scheduling job on computational grids using a fuzzy particle swarm optimization algorithm, Liu [/bib_ref]. In work [bib_ref] Observing the effect of interprocess communication in auto controlled ant colony optimization-based..., Tiwari [/bib_ref] , the authors applied an automatically controlled ant colony optimization (ACO) method to Grid job scheduling, which effectively processes the effect of interprocess communication and optimizes the turnaround time of the job. Tiwari and Vidyarthi introduced lazy ants into the Grid job scheduling ACO and obtained a good balance between diversification and convergence of the search process [bib_ref] Improved auto control ant colony optimization using lazy ant approach for grid..., Tiwari [/bib_ref]. This algorithm not only produces a good solution for the given objectives but also reduces the time complexity of the algorithm. In paper [bib_ref] Enhancing the genetic-based scheduling in computational grids by a structured hierarchical population, Kołodziej [/bib_ref] , the authors enhanced a genetic algorithm's main branching operations, and implemented a Grid job scheduling method which can simultaneously optimize two objectives: makespan and flowtime. Considering the Grid resources availability, Prakash and Vidyarthi proposed a Grid scheduling technique based on a genetic algorithm [bib_ref] Maximizing availability for task scheduling in computational grid using genetic algorithm, Prakash [/bib_ref]. Younis and Yang proposed an improved genetic algorithm (MGA) that adopts a new mutation procedure to solve grid independent job scheduling problem [bib_ref] A genetic algorithm for independent job scheduling in grid computing, Younis [/bib_ref]. Tang et al. proposed a hybrid algorithm combining Genetic Algorithm and Simulated Annealing Algorithm to search optimal solution in designing reasonable departure schedule [bib_ref] A hybrid algorithm for urban transit schedule optimization, Tang [/bib_ref]. However, these scheduling strategies can not effectively deal with Grid application requirements. The application-aware Grid job scheduling problem reported in the literature was proposed by Hu and Veeravalli [bib_ref] Requirement-Aware Scheduling of Bag-of-Tasks Applications on Grids with Dynamic Resilience, Hu [/bib_ref] , whose RAPAR and RAKAR algorithms addressed the scheduling of applications with heterogeneous processing requirements on a Grid. Paper [bib_ref] Scheduling big data applications within advance reservation framework in optical grids, Abouelela [/bib_ref] considered the geographically distributed data feature of Big Data applications and proposed an advance reservation scheduling framework in optical Grid. Xu and Yang proposed a heuristic multiobjective scheduling algorithm to optimize both Grid users' applications and Grid resource providers' incentives, such as cost [bib_ref] An incentive-based heuristic job scheduling algorithm for utility grids, Xu [/bib_ref]. Other heuristic scheduling optimization techniques are conventional k-means cluster scheduling [bib_ref] A novel approach for initializing the spherical K-means clustering algorithm, Duwairi [/bib_ref] , cost-driven partial critical paths scheduling [bib_ref] Cost-Driven Scheduling of Grid Workflows Using Partial Critical Paths, Abrishami [/bib_ref] , dynamic programming [bib_ref] A fuzzy-based customer clustering approach with hierarchical structure for logistics network optimization, Wang [/bib_ref] [bib_ref] Profit distribution in collaborative multiple centers vehicle routing problem, Wang [/bib_ref] , K-Percent Best(KPA) [bib_ref] Immediate mode scheduling in grid systems, Xhafa [/bib_ref] , rank-based hybrid scheduling (RBHS) [bib_ref] A rank-Based hybrid algorithm for scheduling data-and computation-Intensive jobs in grid environment, Abdoli [/bib_ref]. This paper will consider the hardware, software, and job transmission time of Grid application requirements, and make optimization scheduling decisions to improve computational Grid performance. ## Computational grid and job models This section describes the target computational Grid, job scheduling architecture, and job model used in our study. ## Large-scale computational grid This paper studies the China National Grid, which consists of many geographically distributed heterogeneous computing resources, including 2 main centers, 6 National Supercomputing centers, and 11 common centers [fig_ref] Fig 1: China National Grid [/fig_ref]. The Supercomputing center of the Chinese Academy of Sciences is one of the main centers and is responsible for managing the whole Grid. The National Supercomputing centers are Wuxi, Changsha, Jinan, Guangzhou, Shenzhen, and Tianjin. All of the National Supercomputing centers have powerful computing capability, with resources such as the Sunway TaihuLight and Tianhe-2, the top 2 supercomputers in a recent TOP500 list . The centers of the Grid are connected by ChinaNet or CerNet, which have heterogeneous public commercial interconnection bandwidth and delay. Each Grid computing center GC i provides many parallel computational software packages, such as Molecular Massively Parallel Simulator(LAMMPS), CPMD, GAMES, MPI, SANSYS, Application-aware deadline constraint job scheduling mechanism on large-scale computational grid RADIOSS, HyperMesh, and so on. This paper uses PS(GC i ) to denote the set of available software. Each software has attributes: software name, software id, license, and version. The symbols TN(GC i ), AN(GC i ), and AS(GC i ) denote the total number of computational nodes, available computational nodes, and available computational storage of the Grid computing center GC i , respectively. This paper uses the symbol MM(GC i ) to indicate that the computational node can work as a multicore and manycore model. This architecture assumes that all applications or jobs, along with their software, computing nodes, execution time, deadlines, storage, and so on, provided by user, are submitted to the main center by a web interaction interface. All jobs are inserted into the job linked list queue and can be periodically scheduled by the Scheduler, which is a scheduling decision module, according to the requirements of the application, Grid network prediction, and the dynamic Grid environments. The module Grid Resources Monitor can periodically collect Grid computing centers running jobs, available computing nodes, cores, storage, network bandwidth, delay, and so on. The resources of Grid computing centers change dynamically with local and Grid job assignment, job operation completion, resource failure, and safety maintenance. Therefore, in the scheduling architecture, Grid computing centers will report their resource status to the main Grid center at an interval of 4 minutes. Network Prediction is used to dynamically forecast future network communication conditions among the main center and other Grid computing centers. Job Dispatch can dispatch jobs to the corresponding computing center according to scheduling decisions. ## Job scheduling architecture ## Grid application model This paper only considers the scheduling of bag-of-tasks (BoT) or parameter-sweep applications (jobs) on a large-scale distributed computational Grid. Therefore, the jobs A 1 , A 2 , � � �, A N are assumed to be independent and atomic. Examples of these Grid jobs include Monte Carlo simulations [bib_ref] Random number generators for large-scale parallel Monte Carlo simulations on FPGA, Lin [/bib_ref] , tomographic reconstructions, rice genome-wide association analysis [bib_ref] Genome-wide Association Analyses Reveal the Genetic Basis of Stigma Exsertion in Rice, Zhou [/bib_ref] , and data mining algorithms. They are frequently used in fields such as astronomy, bioinformatics, high energy physics, and many others. In our application model, each job A i has requirements, such as, software (including version and license), number of computational nodes, manycore demand, and so on. Furthermore, the job also has characteristics of size, arrival time, execution time, deadline, and more. The Grid application notations and their meanings used throughout this paper are listed in [fig_ref] Table 1: Grid job A i characteristics [/fig_ref]. ## Grid job transmission ## Grid data transfer characteristics The performance of data transfer between the main Grid computing center GC 1 and other Grid computing centers GC i changes with time. This is owing to the fact that most Grids, such as the China National Grid, are interconnected by multi-commercialized internet and not by a dedicated interconnected network. For example, the Changsha National Supercomputing Grid center has China Telecom and China Unicom Internet as its ISPs, and the quality of service from each is different to the other. Another reason is that the commercialized internet is greatly affected by the network environment. Therefore, data transfer bandwidths vary with time. shows a data transfer bandwidth variance curve between the China National Grid main Grid computing center (Supercomputing Center of Chinese Academy of Sciences) and the Changsha National Supercomputing Grid center. From we can conclude that the data-transfer bandwidth is a set of values of a variable during a consecutive time series. This non-stationary time series can be forecasted by many existing prediction techniques, such as ARIMA model [bib_ref] Local Storage-Based Consolidation With Resource Demand Prediction and Live Migration in Clouds, Zhang [/bib_ref] , Hidden Markov Model [bib_ref] Inferring driving trajectories based on probabilistic model from large scale taxi GPS..., Tang [/bib_ref] , auto-regressive [bib_ref] Wind farm micro-siting based on auto-regressive wind prediction, Qiao [/bib_ref] , and so on. G. Zhang et al. proved that the ARIMA is one of most suitable prediction models for server workload, resource, and communication network with high efficiency and low time complexity [bib_ref] Local Storage-Based Consolidation With Resource Demand Prediction and Live Migration in Clouds, Zhang [/bib_ref]. Therefore, this paper uses ARIMA model to forecast Grid job data transmission time among Grid computing centers. ## Job transmission time prediction The ARIMA model is the combination of Auto Regressive (AR) and Moving Average (MA) models, and was developed by Box and Jenkins. Generally, ARIMA is model as ARIMA(p, d, q), which has the following concise form [formula] � p ðBÞr d x t ¼ y q ðBÞe t :ð1Þ [/formula] where x t is the prediction of Grid data-transfer bandwidth at time t, B is the backward shift [formula] operator, ϕ p (B) is the Auto Regressive operator defined as ϕ p (B) = 1 − ϕ 1 B − ϕ 2 B 2 − � � � − ϕ p B p , and r d = (1 − B) d is the dth order of difference operator. e t is the normally distributed error at period t, θ q (B) = 1 − θ 1 B − θ 2 B 2 − � � � − θ q B q . The ARIMA model uses previous time series data-transfer bandwidths x t−1 , x t−2 , � � � to forecast x t . [/formula] In this paper, the time period is set as 5s. Therefore, at time period t, the Grid can transfer 5x t M data from the main Grid center to the corresponding computing center. The data-transfer bandwidth x t also can be iteratively used to forecast the next time series x t+1 , x t+2 , x t+3 , � � �. Thus, for Grid job A i , the data transmission prediction time DPT(A i , GC k ) from the main Grid center to the Grid computing center GC k can be expressed as [formula] ( SiðA i Þ ¼ 5 P x tþs ; s 2 0; 1; � � � ; r; DPTðA i ; GC k Þ ¼ 5 � ðr þ 1Þ:ð2Þ [/formula] where r is the max data transfer periods. ## Problem formulation ## Heterogeneous computational node normalization The computation capacity of Grid computing centers is naturally heterogeneous. For example, the Tianhe-2 supercomputer in the Guangzhou National Supercomputing Center has 17920 computational nodes, each node has 2 Intel Xeon E5-2692v2 12C 2.2GHz processors and 3 Xeon Phi 57 . The Dawning Nebulae supercomputer in the Shenzhen National Supercomputing Center has 2560 computational nodes, each node has 2 Intel Xeon 6C 2.66GHz processors and 1 NVidia C2050 GPU . Therefore, an important task for job scheduling is to standardize the heterogeneous Grid computing center computation capacity. There are many research work to address heterogeneity from engineering disciplines. Zou et al. applied a generalized finite mixture of negative binomial (NB) models with K mixture components to solve heterogeneous data in empirical Bayes estimation [bib_ref] Empirical Bayes estimates of finite mixture of negative binomial regression models and..., Zou [/bib_ref]. Fan et al. use deep learning method to virtualize heterogeneous radio into normalized resources [bib_ref] A Radio Resource Virtualization-Based RAT Selection Scheme in Heterogeneous Networks, Fan [/bib_ref]. These methods are very effective for solving the corresponding problems, but they are not suitable for our proposed periodic scheduling mechanism because of their high time complexity. In the following, we propose a simple and efficient heterogeneous computational node normalization method. In this paper, we adopt 2 CPUs, which have 6 cores at 2.0GHz, as the computational node standardization multicore capacity. Here, systems let GMS(GC i ) and GMC(GC i ) denote the speed and cores of the Grid computing center GC i CPU, respectively. GMN(GC i ) is the CPU number of the Grid computing center GC i computational node. Therefore, the standardization multicore capacity GSC(GC i ) of the Grid computing center GC i computational node is [formula] GSCðGC i Þ ¼ GMNðGC i Þ � GMCðGC i Þ � GMSðGC i Þ 2 � 6 � 2:0 :ð3Þ [/formula] For the computational node manycore capacity, this paper adopts the NVIDIA Tesla C2050, which has 448 cores and a computational capacity of 515.0GFlops, as the standardization capacity. The single core capacity among manycores, such as NVIDIA, Xeon Phi, SW26010, and so on, is heterogeneous. Therefore, this paper gives a heterogeneity ϕ for manycores other than NVIDIA. For example, the manycore heterogeneity ϕ of Xeon Phi to NVIDIA is ϕ = 2.3. Here, this paper also defines MCC(GC i ) as the manycore computational capacity of Grid computing center GC i . The computational node standardization manycore capacity MSC (GC i ) is defined as [formula] MSCðGC i Þ ¼ � � MCCðGC i Þ 515:0 :ð4Þ [/formula] ## Scheduling attributes To facilitate the presentation of the proposed application-aware constraint job scheduling algorithm, it is necessary to introduce some definitions and assumptions. Let ET(A i , GC k ) denote the execution time of job A i on Grid computing center GC k , such that: [formula] ETðA i ; GC k Þ ¼ EtðA i Þ MINfGSCðGC k Þ; MSCðGC k Þg JmðA i Þ is true EtðA i Þ GSCðGC k Þ Otherwise: 8 > > > < > > > :ð5Þ [/formula] where ET(A i , GC k ) is the maximum execution time between multicore and manycore processors on a computational node when the application manycore requirement Jm(A i ) is true. Otherwise, the application A i only uses the multicore of the computational node. The job A i execution finish time JFT(A i , GC k ) on Grid computing center GC k is the sum of the scheduling point, job transmission prediction time, and job execution time, and can be defined as follows [formula] JFTðA i ; GC k Þ ¼ sPoint j þ DPTðA i ; GC k Þ þ ETðA i ; GC k Þ:ð6Þ [/formula] where sPoint j is the system periodic scheduling point with interval 120s (2 minutes according to scheduling architecture module Grid Resources Monitor). In fact, the system periodic scheduling point sPoint j is the current scheduling time, such as 13: 47: 12, and the next scheduling point sPoint j+1 will be 13: 49: 12. Thus, job A i 's actual processing time JPT(A i , GC k ) is the difference between its execution finish time and arrival time. This paper expresses it as [formula] JPTðA i ; GC k Þ ¼ JFTðA i ; GC k Þ À AtðA i Þ:ð7Þ [/formula] On the contrary, the job scheduling strategies are constrained by application software and hardware requirements. Each Grid computing center provides an application software set PS(GC i ), and the software license and version must satisfy the job requirements. That is to say that the license li(sf k ) and version vs(sf k ) for application software sf k 2 PS(GC i ) must be higher than job A i 's software Sw(A i ) requirements: license Sl(A i ) and version Sv(A i ). That is, [formula] SwðA i Þ ¼ sf k^s f k 2 PSðGC i Þ; liðsf k Þ � SlðA i Þ; vsðsf k Þ � SvðA i Þ: 8 > > > < > > > :ð8Þ [/formula] The Grid computing center GC k must satisfy job A i hardware requirements, such as manycore support, available computational nodes, and available storage and can be expressed as [formula] JnðA i Þ � ANðGC k Þ; [/formula] JsðA i Þ � ASðGC k Þ; [formula] MMðGC i Þ ¼ true if JmðA i Þ is true: 8 > > > < > > > :ð9Þ [/formula] Generally, jobs are also expected to be completed before their deadline. That is, [formula] JFTðA i ; GC k Þ � DlðA i Þ:ð10Þ [/formula] # Problem statement This section sets X i = 1 if job A i is scheduled on Grid computing center GC k , and X i = 0 if job A i is rejected by the system and the Grid system can not find a suitable Grid computing center GC k to accomplish its execution. Therefore, the total processing time of jobs TPT can be expressed as [formula] TPT ¼ X m i¼1 X i JFTðA i ; GC k Þ:ð11Þ [/formula] Here, this paper outlines the main scheduling objectives used in this study. The first performance objective is the average processing time APT, which is the average of all jobs actual processing time and is defined as [formula] APT ¼ P m i¼1 X i JFTðA i ; GC k Þ P m i¼1 X i :ð12Þ [/formula] where m is the total number of jobs in the Grid system including many scheduling point jobs. The other scheduling objective is to try to degrade the job rejection ratio JobRej, which is defined as [formula] JobRej ¼ m À P m i¼1 X i m 100%:ð13Þ [/formula] This paper tries to minimize both the average processing time and job rejection ratio. This optimization scheduling problem can be expressed as ( Minimize APT and JobRej; Subject to Eqð8Þ; Eqð9Þ; Eqð10Þ: ## Application-aware job scheduling mechanism The proposed application-aware job scheduling mechanism (AJSM) tries to periodically schedule jobs by using an application-aware deadline constraint job scheduling algorithm. The following sub sections will describe the main ideas. ## The periodic scheduling flow This section proposes an application-aware periodical scheduling flow, as shown in [fig_ref] Fig 4: The application-aware periodic job scheduling flow [/fig_ref] The Grid job scheduling mechanism first initializes system parameters, such as the scheduling point periodSch = 0, the Grid computing centers' software, the total number of computational nodes TN(GC i ), and so on. The Grid computing centers' heterogeneous computational nodes are then normalized according to Section heterogeneous computational node normalization. Next, the Network Prediction and Grid Resources Monitor module are adopted to periodically collect Grid computing centers and network information, which are used in the later scheduling decision. The interval of periodic scheduling is set to 4 minutes according to the Grid Resources Monitor module. Lastly, the application-aware deadline constraint job scheduling algorithm is responsible for scheduling all jobs submitted by users in each period. ## Application-aware deadline constraint job scheduling algorithm Our proposed application-aware deadline constraint job scheduling algorithm first needs to find Grid computing centers that can satisfy job (or application) software requirements. This process is outlined in Algorithm 1, which attempts to find the set of available Grid computing centers Avc(A i ) for each job. The set Avc(A i ) must satisfy Eq (8) The application-aware deadline constraint job scheduling algorithm is formalized in Algorithm 2. The goal of this algorithm is to the deliver job that has the minimum execution finish time with the application requirements and deadline constraints on the Grid. To achieve this goal, the algorithm first uses the Grid computing centers search algorithm to find the job's available Grid computing centers Avc(A i ). Next, for any unscheduled jobs, our proposed algorithm uses the ARIMA forecast transmission time DPT(A i , GC k ) and computes the job's minimum execution finish time on its available Grid computing centers (Steps 6-8). If the computing resource demands and the job's deadline constraint are met, the Grid computing center is put into job's schedulable set (Steps 9-11). Steps 13-18 try to find a Grid computing center with the minimum execution finish time for job A i . If there is no Grid computing center that can run job A i , job A i will be inserted into the next scheduling point queue until the system rejects it. Lastly, this algorithm assigns the job to the Grid computing center with minimum JFT(A i , GC k ) for all job and Grid computing center pairs, and updates Grid resource and job scheduling queue information. Algorithm 2: Application-aware deadline constraint job scheduling algorithm. Compute job execution time ET(A i , GC k ) (Eq (5)); 7 Use ARIMA forecast DPT(A i , GC k ) (Eq (2)); 8 Compute job execution finish time JFT(A i , GC k ) (Eq (6)); 9 if Eqs (9) and (10) are satisfied then [bib_ref] Requirement-Aware Scheduling of Bag-of-Tasks Applications on Grids with Dynamic Resilience, Hu [/bib_ref] Put GC k into job A i 's schedulable set. 11 end Find A i and GC k pair with minimum JFT(A i , GC k ); 21 Assign job A i to Grid computing center GC k ; 22 Update Grid center AN(GC k ) and AS(GC k ); 23 Remove job A i from job queue. 24 end ## Time complexity The time complexity of job scheduling algorithms is usually expressed in terms of the number of jobs N, the Grid computing centers W, and the maximum number of software packages Z. The time complexity of this application-aware deadline constraint job scheduling algorithm is analyzed as follows: The application searching algorithm can be done in time O(NWZ). In fact, the time complexity of the ARIMA prediction method is much higher than steps 6 and 8-11. Here, this paper assumes that the ARIMA prediction ## Performance evaluation To assess the performance of proposed AJSM, this paper developed a discrete-event simulation Grid environment based on GridSim [bib_ref] An enhanced load balancing mechanism based on deadline control on GridSim, Hao [/bib_ref]. This paper compares the AJSM algorithm with a baseline traditional scheduling strategies Min-Minand a recently new meta-heuristic algorithm MGA [bib_ref] A genetic algorithm for independent job scheduling in grid computing, Younis [/bib_ref] to understand its effectiveness on Grids. The performance metrics chosen for the comparison are the all jobs total processing time TPT in Eq (11), average processing time APT in Eq (12), makespan, and job rejection ratio JobRej in Eq [bib_ref] Taxi Trips Distribution Modeling Based on Entropymaximizing Theory: A Case Study in..., Tang [/bib_ref]. Here, makespan is the maximum job finish time for all jobs and defined as makespan ¼ Max i¼1;2;���;m fX i JFTðA i ; GC k Þg: ð15Þ The Min-Min algorithm begins with computing the set of minimum completion time for each unmapped Grid jobs (or applications) on all Grid computing centers. Then, the job with the overall minimum completion time is chosen and allocated to the corresponding Grid computing center. Last, the newly mapped job is removed from unmapped Grid job set and the process repeats until all jobs are scheduled. The Min-Min is a traditional and widely used scheduling algorithm that has been adopted by many research works as a reference object or evaluation benchmark [bib_ref] Budget-constraint Stochastic Task Scheduling on Heterogeneous Cloud Systems, Tang [/bib_ref] [bib_ref] MiM-MaM: A new task scheduling algorithm for grid environment, Vaaheedha [/bib_ref] [bib_ref] A genetic algorithm for independent job scheduling in grid computing, Younis [/bib_ref]. The improved genetic algorithm (MGA) starts with an initial population, which is generated by seeding the population with one individual generated by Min-Min, and the other individuals generated randomly. Then, the following steps: selection, crossover, and mutation operators are applied. The key of this MGA algorithm is that its mutation operator uses the concept of swap and transfer to alter individuals [bib_ref] A genetic algorithm for independent job scheduling in grid computing, Younis [/bib_ref]. This is an effectively and newly scheduling strategy that we choose to compare with our proposed mechanism. ## Experimental settings and environments In the following experiments, this paper simulates 20 Grid computing centers with different characteristics, such as number of computational nodes, application software set, and storage, while each node has multicore (CPU, core, speed), manycore (capacity, heterogeneity ϕ), and memory characteristics. The main parameters of the simulated computing resources are listed in [fig_ref] Table 2: The settings of simulated Grid computing center [/fig_ref]. The first 10 Grid computing centers (GC 1 , � � �, GC 10 ) are derived from the China National Grid [4], where their total number of nodes is up to 74626. The other 10 Grid computing centers (GC 11 , � � �, GC 20 ) are small servers with same configuration. Here, the Grid computing center GC 1 is set as the main center, and the network communication among GC 1 and other centers is a dynamical generated uniformly distribution between 100M and 50G. The available computational nodes are divided into three categories according to their properties; The first is busy, and the number of available computational nodes is randomly generated as [0.5%-3%] of their total nodes, such as GC 4 , GC 5 , GC 8 , GC [bib_ref] Enhancing the genetic-based scheduling in computational grids by a structured hierarchical population, Kołodziej [/bib_ref] ; the second has medium resources available with [3%-10%], such as GC 1 , GC 2 , GC 9 , GC 12 ; the third has resource availability of [10%-50%]. In the simulations, the Grid applications (or jobs) and their application software come from the field of natural science and engineering. Examples include automobile frame stiffness analysis, bridge wind characteristics numerical simulation, mesoscale numerical weather forecast, large airliner CFD check and auxiliary design, and more. These jobs characteristics are derived from the Parallel Workloads Archive HPC2N traceand China National Grid real-world applications . lists three jobs characteristics as an example. Grid applications submitted by the user vary from 960 to 2880 with 240 steps, and the scheduling periods are set as 60 and 120 (meaning 4 and 8 hours). Application-aware deadline constraint job scheduling mechanism on large-scale computational grid ## Job transmission prediction results As job transmission time is an important factor in job execution finish time, this paper will evaluate our prediction method based on the ARIMA model in the first experiments. This paper tests the above applications among the Grid main computing center and other centers. Network communication historical data are retrieved from the China National Grid. [fig_ref] Table 4: The experimental results of job transmission time prediction [/fig_ref] lists 10 applications transmission prediction time and their actual time. From [fig_ref] Table 4: The experimental results of job transmission time prediction [/fig_ref] , this paper can conclude that our proposed prediction method is effective for 7 jobs, with an error ratio lower than 10% in all 10 applications. # Experimental results In the second experiments, this paper first compares the performance of AJSM, MGA, and Min-Min with 60 scheduling periods; the experimental results are shown in [fig_ref] Fig 5: Performance impact of jobs with 60 scheduling points [/fig_ref] From [fig_ref] Fig 5: Performance impact of jobs with 60 scheduling points [/fig_ref] , this paper can conclude that the job rejection ratio of AJSM is much lower than the other two algorithms. For the average rejection ratio, AJSM significantly outperforms MGA by 85.3%, and Min-Min by 87.5%. This improvement is due to the fact that the AJSM approach is an application-aware algorithm, which can adaptively search Grid computing centers that satisfy jobs software and hardware requirements. Whereas, MGA and Min-Min do not comprehensively consider the computation intensive Grid applications' requirements, especially for their computing software characteristics. Thus, some jobs scheduled by MGA and Min-Min can not execute on the corresponding Grid computing center and are rejected by the Grid systems, regardless of the existence of other Grid computing centers that can execute those jobs. In contrast, jobs rejected by AJSM are mainly due to the Grid systems lacking a Grid computing center that can meet their software, hardware, and deadline constraints. Therefore, Our proposed algorithm ASJM is more successful than MGA, Min-Min in scheduling Grid jobs. This paper also observe from [fig_ref] Fig 5: Performance impact of jobs with 60 scheduling points [/fig_ref] that as the number of jobs increases, the job rejection ratio of AJSM, MGA, Min-Min all increase too. This is mainly due to the fact that as the number of jobs increase, the system workload increases and results in the operation of Grid computing centers with low processing capacity. Therefore, more jobs get rejected as their execution finish times are beyond the deadline constraint. For low workloads, such as jobs that are 960, 1200, or 1440, there are only a few rejected jobs for the AJSM approach. However, as the number of jobs increases, the growth rate of the AJSM job rejection ratio is more than that of the MGA and Min-Min job rejection ratios. For high system workload, such as the number of jobs exceeds 2880, 5000, the job rejection ratio of AJSM may close to that of MGA. The main reason is that the deadline restriction becomes the key element of job rejection. [fig_ref] Fig 5: Performance impact of jobs with 60 scheduling points [/fig_ref] and 5(b) plot the job total processing time and average processing time of the three algorithms when the number of jobs increases from 960 to 2880. [fig_ref] Fig 5: Performance impact of jobs with 60 scheduling points [/fig_ref] reveals that the AJSM job total processing time is more than that of MGA and Min-Min. This is a reasonable experimental phenomenon for AJSM handling more jobs, which results in a greater total processing time and lower job rejection ratio. This performance improvement manifests mainly in the average processing time of [fig_ref] Fig 5: Performance impact of jobs with 60 scheduling points [/fig_ref] , where AJSM exceeds MGA by 6.9% and Min-Min by 5.4%, for the average experimental results. The experimental results for comparison metric makespan are shown in [fig_ref] Fig 5: Performance impact of jobs with 60 scheduling points [/fig_ref] , where the AJSM outperforms MGA, Min-Min by an average of 6.2%, 5.4%, respectively. This is mainly due to the fact that our proposed ASJM strategy adopts two key techniques: job transmission time prediction based on the ARIMA model and heterogeneous Grid computing node resource normalization, which can give a more accurate job execution finish time. Therefore, our proposed ASJM is better than MGA, Min-Min in terms of average processing time, makespan, and job rejection ratio. From [fig_ref] Fig 5: Performance impact of jobs with 60 scheduling points [/fig_ref] this paper can also conclude that Min-Min outperforms MGA in terms of average processing time and makespan, and MGA is better than Min-Min in term of job rejection ratio. The improvements of AJSM over MGA and Min-Min could also be concluded from [fig_ref] Fig 6: Performance impact of jobs with 120 scheduling points [/fig_ref] , which shows the simulation experimental results with 120 scheduling periods. The AJSM algorithm significantly outperforms MGA by 91.6%, Min-Min by 92.3%, in term of job rejection ratio, respectively. Moreover, AJSM is also better than MGA by 10.7%, Min-Min by 5% in term of average processing time, and MGA by 9.7%, Min-Min by 5.8% in term of makespan. On the other hand, the average processing time and job rejection ratio of AJSM algorithm are superior to those of the experimental results with 60 scheduling periods. This is mainly due to the fact that the Grid systems' workload with 120 scheduling periods is lower than the workload with 60 scheduling periods, and the AJSM can find a more optimal Grid computing center with the minimum execution finish time. ## Conclusions and future work In this paper, our main objective was to effectively deal with Grid application software, hardware, and deadline requirements. Therefore, this paper first built a Grid job scheduling architecture that can periodically make job scheduling decisions. This paper then used an ARIMA model to forecast job transmission times. Next, this paper normalized the Grids' heterogeneous computing nodes and formulated the application-aware deadline constraint job scheduling problem as a linear programming problem. Lastly, an AJSM scheduling mechanism was proposed to solve this problem with low time complexity. The comparison studies demonstrated that our proposed AJSM can successful schedule more Grid jobs than MGA, Min-Min. This is mainly due to the fact that the AJSM algorithm has a lower job rejection ratio than MGA and Min-Min. For successful scheduled jobs, AJSM scheduling mechanism also outperforms existing algorithms: MGA, Min-Min in terms of job average processing time and makespan. Future studies in this area are twofold. First, we shall extend the Grid job transmission time prediction using an artificial neural network. Second, we plan to build a more precise job requirements model to describe Grid applications. Supporting information S1 File. Experimental results dataset S1_File.docx. (DOCX) [fig] Fig 1: China National Grid. https://doi.org/10.1371/journal.pone.0207596.g001 [/fig] [fig] Fig 2: depicts the large-scale computational Grid job scheduling architecture. [/fig] [fig] Fig 4: The application-aware periodic job scheduling flow. https://doi.org/10.1371/journal.pone.0207596.g004 [/fig] [fig] Fig 5: Performance impact of jobs with 60 scheduling points. (a) Total Processing Time; (b) Average Processing Time; (c) Makespan; (d) Job Rejection Ratio. https://doi.org/10.1371/journal.pone.0207596.g005 [/fig] [fig] Fig 6: Performance impact of jobs with 120 scheduling points. (a) Total Processing Time; (b) Average Processing Time; (c) Makespan; (d) Job Rejection Ratio. https://doi.org/10.1371/journal.pone.0207596.g006 [/fig] [table] Table 1: Grid job A i characteristics. [/table] [table] Table 2: The settings of simulated Grid computing center. [/table] [table] Table 4: The experimental results of job transmission time prediction. [/table]

Internet-based support for self-management strategies for people with COPD–protocol for a controlled pragmatic pilot trial of effectiveness and a process evaluation in primary healthcare ## Measurement of fidelity of patients and providers needs to be clarified Since this is pragmatic trial, how will the exposure to usual provider encounters and counseling on self-management, outside of the study, be assessed and taken into consideration in the analyses? ## Reviewer Linda Nici, M.D. Providence Veterans Affairs Medical Center and Brown University, Providence, Rhode Island, USA REVIEW RETURNED 11-Apr-2017 ## General comments The manuscript, "Internet-based support for evidence based selfmanagement strategies for people with COPD -a controlled pragmatic pilot trial of effectiveness and a process evaluation in the primary health care", by Nyberg and colleagues, is a pragmatic controlled pilot trial to evaluate the feasibility and effectiveness of an internet-based COPD self-management program from both the patient and the providers perspective. The study is well described and has the potential for real impact in this arena, however, there are some issues that need to be addressed and discussed, which are not clear in the manuscript's present form: 1-Why are patients completing self-assessments at home? The literature supports having these items done in the presence of the staff such that they are clearly the answers of the patient, not the caregivers and that there are no issues with comprehension or completeness. 2-There is no discussion of the criteria for the COPD patient enrollment-i.e. severity of disease, documentation by PFT, some element of matching by disease severity, age, gender, comorbidities. While I realize this is a pilot study, your outcome measures may be heavily influenced if there is a broad range of disease severity or if underlying respiratory disease is not clear. 3-It is not totally clear to me why the HCP is trained using the webbased program? Is it your intention for them to interact with the patient face-to-face as the study moves forward such that they are actually reinforcing the web-based information? If so, will this be standardized. If some patients see their provider more times than others and material is reinforced, this may bias outcomes from the web based product alone. 4-Is there any plans to videotape the interactions between provider and patient for internal consistency? ## Reviewer ## Chris burtin Hasselt University, Belgium REVIEW RETURNED 12-Apr-2017 ## General comments Dr Nyberg et al submitted a protocol of a trial which will investigate feasibility, acceptance and effectiveness of a web-based program to support patients with COPD and health professionals in light of selfmanagement processes. This trial will be a controlled pilot trial, exploring numerous quantitative and qualitative outcomes, that will be used to design an adequately powered randomized controlled trial in the future. The methodology appears sound and scientifically correct. I have a few remarks and comment about the manuscript/methodology. The title and introduction suggest that this project concerns selfmanagement strategies (as a comprehensive term), but in the methods section it becomes clear that self-management strategies offered by the COPD-web tool are largely related to physical activity and exercise. As self-management is more comprehensive than physical activity and exercise, I am wondering whether this is the correct term to use in this protocol. It is remarkable that the type of health professionals that will be recruited are not further specified. One would assume that different professions need different approaches in terms to optimize their delivery of self-management strategies to patients. Which kind of health professionals will be included? It will be recorded to what extent health care professionals implement the web-tool in their daily clinical practice. On the other hand, the health care professionals are expected to include the patients. Am I correct that the same individuals are both research subject (as health care provider that will use the platform) and researchers (who will include patients for this trial)? In light of that, will the proportion of the target population (patients) that will receive the intervention adequately reflect clinical reality? Will there be any interaction between health care professionals and patients in light of self-management strategies; in other words will the professional assist the patient in his self-management and/or using the COPD-web tool throughout time or will he/she only introduce the patient to it? The timeline of assessments is a difficult to interpret because time lines of both target groups (patients and health care professionals) are mixed. It might be easier to understand when these are separated. According to the timeline (p8), this trial has already started in 2016. ## Version 1 -author response Reviewer 1 What does evidence based self-management strategies mean? Authors reply: We have revised the title and removed the word evidence based as this might be ambiguous. The self-management strategies presented at the COPD-web are to most parts evidence based and recommended in national and international guidelines i.e. strategies to promote increased level of PA, strategies during exacerbations, strategies in case of malnutrition, strategies for smoking cessation etc. However, as the COPD-web as a strategy to support the self-management not is evidence based, we believe that removing "evidence based" might be more clear. Further, we have added a figure which show the content on the website. An example of a webpage viewed by patients and that viewed by providers would be very informative. Authors reply: As the COPD-web is made up of several sub-pages and subsub-pages focusing on different topics such as About COPD (How does the lungs work?, Facts about COPD, How do one get the diagnosis? Excacerbations) Self-care and treatment (Smoking, Physical Activity, Register physical activity, Action plan, Follow the disease, Nutrition, Tips for making everyday life easier, Breathing and coughing techniques, etc) we don't believe that an example of the webpage would contribute with added value. However, in order to provide more information about the content on the COPD-web, we have added a map of the content referred to on page 8. More detail on the content would strengthen this paper. How do users interact with COPDweb over 12 months? Is there dynamic content, recurrent content or all new content? Authors reply: We have added some information about how we continuously add new material to the COPD-web on page 9. Today, we do not have an answer to the question about how the users will interact with the COPD-web over time. However, as a part of the process evaluation, we will collect user data from the web site and carry out qualitative interviews at 3 and 12 months and this data will hopefully contribute with information on this issue. A glaring omission is self-management of acute exacerbations-home meds of antibiotics or prednisone, when to call provider etc. This needs to be included. Authors reply: We agree with you and this information is on the COPD-web. This information has been added in the manuscript on page 8-9 and . What happens to patients in the study if they get hospitalized or have an acute exacerbation? Authors reply: Considering the study, nothing happens. This study will not capture the participants use of health care services, neither the frequency of exacerbations. There are a lot of questionnaires. What is the burden on the patient? Is this feasible to administer so many? Authors reply: We realize that there are a lot of questionnaires. However, as this is a pilot study, one of the aims is to evaluate the feasibility (research question: Are the intervention and the study procedures acceptable and feasible from the perspective of patients with COPD and health professionals?) and we might probably decrease the number of questionnaires in the full study. Seven days of activity monitoring will likely yield 5 days, so a weekend may not be captured. Would recommend at least 10 days of monitoring to capture weekend days too. Authors reply: Thank you for your recommendation which we will consider carefully. We have also added a reference (Demeyer 2014) in which at least 4 days is recommended. However, according to our experience from previous studies, the participants have worn the activity monitor during 7 full days and a weekend day has consequently been monitored. Moreover, as this is a pilot study, we will evaluate the feasibility of this method for measuring the participants' physical activity and consider your recommendation of at least 10 days. How will the qualitative interviews be analyzed? Authors reply: As stated on page 20, the qualitative interviews will be analyzed using qualitative content analysis. We have elaborated the description slightly in order to clarify which of the process evaluation components that will be analyzed using inductive and deductive qualitative content analysis respectively (ses also below). Measurement of fidelity of patients and providers needs to be clarified Authors reply: We will assess the fidelity considering health professionals' delivery of the intervention. The fidelity will be assessed using the qualitative interviews and the observations performed during the health professional -patient encounter when the patient is introduced to the COPD-web. We have elaborated the description of the analyzed of the qualitative data slightly in order to clarify which of the process evaluation components that will be analyzed using inductive and deductive qualitative content analysis respectively (page 20). Thank you for this useful comment. Since this is pragmatic trial, how will the exposure to usual provider encounters and counseling on self-management, outside of the study, be assessed and taken into consideration in the analyses? Authors reply: This will to some extent be captured in the qualitative interviews. However, exposure to usual provider encounters will always have a potential influence in clinical studies and since we have a control group we hope to make up for this. Reviewer 2 1-Why are patients completing self-assessments at home? The literature supports having these items done in the presence of the staff such that they are clearly the answers of the patient, not the caregivers and that there are no issues with comprehension or completeness. Authors reply: To have the participants to complete the self-assessment tool in the presence of the staff might be an ideal situation. However, due to the situation in the primary care in Sweden today with a shortage of health professionals and a constant time pressure, this would not be possible. As the participants will be recruited at six primary care units, it will neither be possible for us to administer this task. However, the feasibility of this procedure will be evaluated. 2-There is no discussion of the criteria for the COPD patient enrollment-i.e. severity of disease, documentation by PFT, some element of matching by disease severity, age, gender, co-morbidities. While I realize this is a pilot study, your outcome measures may be heavily influenced if there is a broad range of disease severity or if underlying respiratory disease is not clear. Authors reply: We have chosen a pragmatic approach which, as suggested by Zwarenstein et al 2008, focus on how the intervention work when used in normal practice. This means that the trial is designed to the needs of those who meet the patients in their daily clinical practice. As comorbidities are very common among persons with COPD, a consequence of the pragmatic approach is that also these persons are included in the study. We have clarified this in the discussion section. 3-It is not totally clear to me why the HCP is trained using the web-based program? Is it your intention for them to interact with the patient face-to-face as the study moves forward such that they are actually reinforcing the web-based information? If so, will this be standardized. If some patients see their provider more times than others and material is reinforced, this may bias outcomes from the web based product alone. Authors reply: In the present study, the health professionals are the ones that introduce the COPDweb to the patients when they visit the primary care. The introduction follow a standardized routine described in Box 1. In line with the pragmatic approach the intervention should be "applied flexibly as it would be in normal practice" (Zwarenstein 2008). Consequently, in the present study, there might be patients who see their providers more times and have the material reinforced. However, also in the control group, some patients might see their provider more than once. 4-Is there any plans to videotape the interactions between provider and patient for internal consistency? Authors reply: No, there are no such plans. Videotaping the interaction would be very interesting for several reason. Unfortunately, to date we do not presently have the resources to videotape or to analyse such data. This will be considered for future studies. The title and introduction suggest that this project concerns self-management strategies (as a comprehensive term), but in the methods section it becomes clear that self-management strategies offered by the COPD-web tool are largely related to physical activity and exercise. As selfmanagement is more comprehensive than physical activity and exercise, I am wondering whether this is the correct term to use in this protocol. Authors reply: Thank you for your important comment. We realize that we have focused too much on physical activity and physical training in the description of the COPD-web. We have now elaborated the description of the content of the COPD-web and also added a map of the content of the patient section of the COPD-web (page 8-9 and . It is remarkable that the type of health professionals that will be recruited are not further specified. One would assume that different professions need different approaches in terms to optimize their delivery of self-management strategies to patients. Which kind of health professionals will be included? Authors reply: The COPD-web is not designed for any particular profession but is thought to be used by the person (could be a nurse, doctor, physiotherapist, etc) meeting a person with COPD in clinical practice. In line with the pragmatic approach in the present study, the intervention i.e. the COPD-web, is applied flexible in order to increase the applicability in ordinary clinical practice. It will be recorded to what extent health care professionals implement the web-tool in their daily clinical practice. On the other hand, the health care professionals are expected to include the patients. Am I correct that the same individuals are both research subject (as health care provider that will use the platform) and researchers (who will include patients for this trial)? In light of that, will the proportion of the target population (patients) that will receive the intervention adequately reflect clinical reality? Authors reply: The health care professionals included in the study will be informed to include (ask for participation) all patients with COPD attending the clinic during the study period i.e. they will not be informed to contact patients especially for this study. As the patients will be recruited during an ordinary visit to their clinic, the target population cannot more reflect clinical reality. For example, in Sweden patients with COPD in the primary care have annual visits to their primary care clinic to perform a follow-up of their COPD. If this annual visit occurs during the intervention period, the patient will be asked to participate. However, the reason for the visit could be something completely else. As long as the reason for the visit is due to their COPD and that they visit their primary care clinic during the intervention period, they will be asked to participate. Will there be any interaction between health care professionals and patients in light of selfmanagement strategies; in other words will the professional assist the patient in his self-management and/or using the COPD-web tool throughout time or will he/she only introduce the patient to it? Authors reply: In line with the pragmatic approach in the present study, the intervention i.e. the COPD-web, is applied flexible i.e. the health professional is allowed to use the COPD-web as he/she consider suitable as long as the routine for introduction is carried out. This has been elaborated in the manuscript (page 10). The timeline of assessments is a difficult to interpret because time lines of both target groups (patients and health care professionals) are mixed. It might be easier to understand when these are separated. Authors reply: Separating the timeline would mean that several parts of the data collection would be described in two different tables, which might be even more difficult to interpret. We have tried to ease the understanding of the time line by adding information about the source of data collection on each row. We hope that this will facilitate the interpretation. According to the timeline (p8), this trial has already started in 2016.

Neurosurgical enhanced recovery after surgery ERAS for geriatric patients undergoing elective craniotomy: A review Population aging is an unprecedented, multifactorial, and global process that poses significant challenges to healthcare systems. Enhanced recovery after surgery (ERAS) protocols aim to optimize perioperative care. The first neurosurgical ERAS protocol for elective craniotomy has contributed to a shortened postoperative hospital stay, accelerated functional recovery, improved patient satisfaction, and reduced medical care cost in adult patients aged 18 to 65 years compared with conventional perioperative care. However, ERAS protocols for geriatric patients over 65 years of age undergoing cranial surgery are lacking. In this paper, we propose a novel ERAS protocol for such patients by reviewing and summarizing the key elements of successful ERAS protocols/guidelines and optimal perioperative care for geriatric patients described in the literature, as well as our experience in applying the first neurosurgical ERAS protocol for a quality improvement initiative. This proposal aimed to establish an applicable protocol for geriatric patients undergoing elective craniotomy, with evidence addressing its feasibility, safety, and potential efficacy. This multimodal, multidisciplinary, and evidence-based ERAS protocol includes preoperative, intraoperative, and postoperative assessment and management as well as outcome measures. The implementation of the current protocol may hold promise in reducing perioperative morbidity, enhancing functional recovery, improving postoperative outcomes in geriatric patients scheduled for elective craniotomy, and serving as a stepping stone to promote further research into the advancement of geriatric patient care.Abbreviations: ERAS = enhanced recovery after surgery, IV = intravenous, LOS = length of stay, PCA = patient-controlled analgesia, POD = postoperative day, PONV = postoperative nausea and vomiting, TIVA = total intravenous anesthetic, VAS = visual analog scale. # Introduction The world has witnessed unprecedented challenges due to an aging population and an increasing burden of healthcare. According to census data, in western countries, 1 in 5 people will be over 65 years old by 2030, and the 85-and-older will make up approximately 2% to 3% of the population. Moreover, in 2030, there will be over 200 million people over 65 years and 66 million over 80 years in China.Owing to the age-associated demographic trends and associated healthcare costs, there is a need to preoptimize the perioperative care of geriatric patients undergoing cranial neurosurgical procedures. Enhanced recovery after surgery (ERAS) or fast-track surgery protocols, first conceptualized by been developed rapidly and applied widely in clinical practice of multiple surgical specialties. [bib_ref] Enhanced recovery protocols for adults undergoing colorectal surgery: a systematic review and..., Greer [/bib_ref] [bib_ref] A systematic review of patient reported outcomes and patient experience in enhanced..., Jones [/bib_ref] [bib_ref] Evidence-based surgical care and the evolution of fast-track surgery, Kehlet [/bib_ref] [bib_ref] Enhanced recovery after surgery: a review, Ljungqvist [/bib_ref] [bib_ref] Enhanced recovery after surgery (ERAS) for spine surgery: a systematic review, Dietz [/bib_ref] With the optimization of perioperative care, ERAS protocols have been shown to benefit patients with shorter hospital length of stay (LOS), better postoperative functional status, lower perioperative complications, higher patient satisfaction, and cheaper healthcare costs. [bib_ref] Enhanced recovery protocols for adults undergoing colorectal surgery: a systematic review and..., Greer [/bib_ref] [bib_ref] A systematic review of patient reported outcomes and patient experience in enhanced..., Jones [/bib_ref] [bib_ref] Evidence-based surgical care and the evolution of fast-track surgery, Kehlet [/bib_ref] [bib_ref] Enhanced recovery after surgery: a review, Ljungqvist [/bib_ref] [bib_ref] Enhanced recovery after surgery (ERAS) for spine surgery: a systematic review, Dietz [/bib_ref] Not until recently, did our group propose the first neurosurgical ERAS protocol for adult patients undergoing elective craniotomy and conduct the first randomized controlled trial to validate its efficacy and safety. [bib_ref] Safety and efficacy of a novel neurosurgical enhanced recovery after surgery protocol..., Wang [/bib_ref] Compared with conventional perioperative care, the neurosurgical ERAS protocol is associated with shortened LOS and accelerated functional recovery. [bib_ref] Safety and efficacy of a novel neurosurgical enhanced recovery after surgery protocol..., Wang [/bib_ref] In the secondary analysis, decreased postoperative nausea and vomiting (PONV), absorbable skin suture, and shortened LOS are independent predictors for a higher patient satisfaction; while aspects including information transfer, professional support, shared responsibility and active participation, readiness for discharge, and follow-up are patient experience themes to be addressed to improve the quality of care. [bib_ref] Neurosurgical enhanced recovery after surgery (ERAS) programme for elective craniotomies: are patients..., Liu [/bib_ref] Furthermore, in the subgroup of glioma patients, implementation of the ERAS protocol is associated with improved health-related quality of life (i.e., higher functioning and lower symptom burden) during the follow-up of up to 6 months after surgery. [bib_ref] Impact of neurosurgical enhanced recovery after surgery (ERAS) program on health-related quality..., Liu [/bib_ref] However, these encouraging results were confined to adult craniotomy patients aged 18 to 65 years, whereas elderly patients aged >65 years were excluded. Indeed, evidence supporting the applicability and efficacy of ERAS protocols for the elderly group is mostly limited to colorectal surgery. [bib_ref] A systematic review of the intervention components, adherence and outcomes of enhanced..., Fagard [/bib_ref] Due to the lack of a neurosurgical ERAS protocol developed for geriatric patients undergoing elective craniotomy, we herein propose a novel ERAS protocol for such patients based on a review of successful ERAS protocols and optimal perioperative care of geriatric patients in the current up-to-date medical literature, in addition to our experience in implementing the ERAS protocol as a constant quality improvement program. # Methods Similar to our previously published ERAS protocols for elective cranial and intraspinal tumor surgery, [bib_ref] Safety and efficacy of a novel neurosurgical enhanced recovery after surgery protocol..., Wang [/bib_ref] [bib_ref] Enhanced recovery after intraspinal tumor surgery: a single-institutional randomized controlled study, Liu [/bib_ref] a literature review of published ERAS protocols for all surgical specialties was first performed to retrieve the key elements. Additional perioperative risk factors for complications and morbidities in geriatric patients, as well as corresponding prophylaxis and interventions, have also been identified in the literature. Panel discussions were held by the Neurosurgical ERAS Working Group, which consists of medical and ancillary staff from neurosurgery, anesthesiology, inpatient and operative nursing, and other services, including nutrition, psychiatry, physiotherapy, and rehabilitation. Key elements of the workflow were organized into 3 chronological time periods: preoperative, intraoperative, and postoperative. ## Eras protocol ## Preoperative evaluation and management The purpose of preoperative evaluation and management is to preoptimize geriatric patients physically, mentally, and functionally prior to the scheduled craniotomy, which includes patient and family counseling; comprehensive geriatric assessment incorporating functional, nutritional, and mental statuses; comorbidity; respiratory preparation; smoking and alcohol abstinence; antithrombotic prophylaxis; preoperative oral carbohydrate loading; antimicrobial prophylaxis; and discharge planning [fig_ref] Table 1: Continued [/fig_ref]. The timeline of these measures varies from weeks to 1 day prior to surgery, depending on the time required for risk assessment and intervention, as well as each individual's medical condition. Moreover, some interventions may continue after surgery or even after discharge to maximize benefits. 3.1.1. Patient and family counseling. The importance of educating the patient and family about surgical expectations is well acknowledged, as it improves patient preparedness, satisfaction, and outcome. [bib_ref] Enhanced recovery after surgery for oncological craniotomies, Hagan [/bib_ref] [bib_ref] Guidelines for perioperative care in elective colorectal surgery: enhanced recovery after surgery..., Gustafsson [/bib_ref] Detailed instruction on the ERAS protocol is necessary to offer a roadmap for the patient and family to motivate active participation. An ERAS handbook is provided at least 1 week before the scheduled surgery to allow the patient to read and ask questions. [bib_ref] Neurosurgical enhanced recovery after surgery (ERAS) programme for elective craniotomies: are patients..., Liu [/bib_ref] Additional demonstrations are provided in multiple ways (verbal, video, printed, and iPad-mediated) during hospitalization. 3.1.2. Functional status. Preoperative physical conditioning contributes to enhanced functional capacity, improved quality of life, shortened LOS, and reduced perioperative complications in patients undergoing elective orthopedic, spine, cardiac, and abdominal surgery, including the geriatric surgical population. [bib_ref] Effect of total-body prehabilitation on postoperative outcomes: a systematic review and meta-analysis, Mina [/bib_ref] [bib_ref] The effects of preoperative exercise therapy on postoperative outcome: a systematic review, Valkenet [/bib_ref] [bib_ref] Optimizing functional exercise capacity in the elderly surgical population, Carli [/bib_ref] The level of preoperative functional status could be augmented with prehabilitation, which compromises a progressive exercise program guided by physiotherapists for 2 to 4 weeks before surgery. ## Nutrition. The patient's preoperative nutritional status is related to perioperative morbidity and LOS, and is a modifiable risk factor to improve clinical outcomes. [bib_ref] Postoperative complications in gastrointestinal cancer patients: the joint role of the nutritional..., Bozzetti [/bib_ref] [bib_ref] Pre-optimization of spinal surgery patients: development of a neurosurgical enhanced recovery after..., Ali [/bib_ref] Although the majority of adult craniotomy patients (18-65 years) had a good nutritional status and required no nutritional intervention preoperatively, [bib_ref] Safety and efficacy of a novel neurosurgical enhanced recovery after surgery protocol..., Wang [/bib_ref] this may not apply to the geriatric population. Age is a well-documented risk factor for malnutrition due to physiological and anatomical changes, chronic diseases, medication use, and dietary and psychosocial habits. [bib_ref] Nutritional risk factors for postoperative complications in Brazilian elderly patients undergoing major..., Dos Santos Junqueira [/bib_ref] Preoperative nutritional status assessment is crucial for geriatric neurosurgical patients. Patients with a body mass index <18.5, >24, or with a low albumin level are recommended to receive nutritional consultation and intervention. [bib_ref] Pre-optimization of spinal surgery patients: development of a neurosurgical enhanced recovery after..., Ali [/bib_ref] [bib_ref] Nutritional risk factors for postoperative complications in Brazilian elderly patients undergoing major..., Dos Santos Junqueira [/bib_ref] 3.1.4. Mental status. The patient's mental status is also a predictor of functional outcomes after surgery. [bib_ref] A systematic review of bio-psychosocial risk factors for an unfavourable outcome after..., Boer [/bib_ref] Therefore, preoperative evaluation of hospital anxiety and depression may help screen patients for potentially beneficial psychiatric intervention. ## Management of comorbidities. Comorbidities such as diabetes, congestive heart failure, coronary artery disease, and chronic steroid use have been shown to increase the risk of perioperative complications in elective craniotomy patients. [bib_ref] Identification of preoperative and intraoperative risk factors for complications in the elderly..., Johans [/bib_ref] [bib_ref] Frequency and predictors of complications in neurological surgery: national trends from 2006..., Rolston [/bib_ref] No single comorbidity precludes elective craniotomy, however, geriatric patients with known comorbidities should undergo appropriate preoperative assessment and specialty consultation to obtain optimal control prior to surgery. 3.1.6. Respiratory preparations. Airway risk assessment is performed with respect to age, smoking history, pulmonary function, body mass index, past and concomitant cardiopulmonary diseases, and comorbidity.All geriatric patients are encouraged to undergo breathing and exercise training, [bib_ref] A nurse-driven enhanced recovery after surgery (ERAS) nursing program for geriatric patients..., Li [/bib_ref] while high-risk patients should receive inhalation treatment with mucolytics and expectorants.3.1.7. Smoking and alcohol abuse. Both smoking and alcohol are well-recognized risk factors for postoperative complications, including difficulty in wound healing, surgical site infection, pulmonary complications, and general infections, which result in intensive care unit admission, prolonged LOS, higher reoperation rates, and greater costs. [bib_ref] Wound healing and infection in surgery. The clinical impact of smoking and..., Sorensen [/bib_ref] [bib_ref] Preoperative alcohol consumption and postoperative complications: a systematic review and meta-analysis, Eliasen [/bib_ref] Short-term preoperative smoking cessation for 3 to 4 weeks could significantly reduce the risk of postoperative complications. [bib_ref] Wound healing and infection in surgery. The clinical impact of smoking and..., Sorensen [/bib_ref] [bib_ref] Short-term preoperative smoking cessation and postoperative complications: a systematic review and meta-analysis, Wong [/bib_ref] Similarly, intensive preoperative alcohol abstinence for 4 to 8 weeks, which may involve pharmacological interventions for relapse prophylaxis and withdrawal symptoms, could decrease postoperative comorbidity. [bib_ref] Perioperative alcohol cessation intervention for postoperative complications, Egholm [/bib_ref] [bib_ref] Preoperative alcohol cessation prior to elective surgery, Oppedal [/bib_ref] Therefore, all geriatric neurosurgical patients are required to abstain from smoking and alcohol for at least 4 weeks prior to surgery. 3.1.8. Antithrombotic prophylaxis. Current evidence supports the use of mechanical prophylaxis of venous thromboembolism with intermittent pneumatic compression and graduated compression stockings. [bib_ref] Perioperative thromboprophylaxis in patients with craniotomy for brain tumours: a systematic review, Salmaggi [/bib_ref] Timing and duration of mechanical prophylaxis is critical, which should be started as early as possible preoperatively and continued until discharge to ensure maximal benefits. On the other hand, the benefits of chemoprophylaxis in reducing venous thromboembolism rates should be cautiously weighed against the increased risk of major bleeding. [bib_ref] Perioperative thromboprophylaxis in patients with craniotomy for brain tumours: a systematic review, Salmaggi [/bib_ref] The most recent systematic review concluded that chemoprophylaxis is effective and safe in patients undergoing elective cranial or www.md-journal.com [fig_ref] Table 1: Continued [/fig_ref] ERAS protocol for geriatric patients undergoing elective craniotomy. spinal surgery without increasing major or minor bleeding events. [bib_ref] Chemical venous thromboembolism prophylaxis in neurosurgical patients: an updated systematic review and..., Khan [/bib_ref] Yet, the robustness of the conclusion for the elderly subpopulation and the optimal time of chemoprophylaxis remain unvalidated and warrant further investigation. Owing to safety concerns, patients in China are instructed and monitored to receive mechanical prophylaxis starting from admission until discharge. 3.1.9. Preoperative oral carbohydrate loading. Instead of traditional prolonged fasting, modern guidelines advocate a shortened period of preoperative fasting, with the allowance of clear liquids up to 2 hours and solids up to 6 hours prior to surgery.Ample evidence has proved the role of preoperative carbohydrate loading in attenuating metabolic stress response (e.g., insulin resistance and protein breakdown) and improving clinical outcomes (e.g., pulmonary function, handgrip strength, and LOS) without increasing complication rates. [bib_ref] A meta-analysis of randomised controlled trials on preoperative oral carbohydrate treatment in..., Awad [/bib_ref] [bib_ref] Preoperative carbohydrate treatment for enhancing recovery after elective surgery, Smith [/bib_ref] [bib_ref] A randomized controlled study of preoperative oral carbohydrate loading versus fasting in..., Liu [/bib_ref] Additionally, no difference in complication rates, peak glycemic levels, or insulin requirements is found between diabetic patients who receive and those who do not receive preoperative carbohydrates. [bib_ref] Preoperative carbohydrate loading in diabetic patients within an enhanced recovery after surgery..., Lee [/bib_ref] For geriatric patients without gastrointestinal mobility disorder, preoperative oral carbohydrate is integrated into the routine procedure. ## Discharge planning. A predefined discharge criterion is used for patients undergoing elective neurosurgery, which includes adequate pain control, adequate oral nutrition, no fever, independent mobility, and a safe discharge destination. [bib_ref] Safety and efficacy of a novel neurosurgical enhanced recovery after surgery protocol..., Wang [/bib_ref] [bib_ref] Enhanced recovery after intraspinal tumor surgery: a single-institutional randomized controlled study, Liu [/bib_ref] Besides following the robust criteria, an appropriate timing for discharge planning is of vital importance to secure early discharge in the ERAS protocol and improve patient satisfaction. [bib_ref] Neurosurgical enhanced recovery after surgery (ERAS) programme for elective craniotomies: are patients..., Liu [/bib_ref] [bib_ref] The importance of preoperative information for patient participation in colorectal surgery care, Aasa [/bib_ref] Therefore, discharge planning is initiated at the time of patient education to establish patient and family expectations and to increase their sense of security on early discharge, which also includes the availability of a variety of services and benefits upon discharge from the inpatient stay. A timely and responsive follow-up strategy could enhance patient experience through improved communication and access. [bib_ref] Neurosurgical enhanced recovery after surgery (ERAS) programme for elective craniotomies: are patients..., Liu [/bib_ref] Furthermore, discharge planning and home follow-up in hospitalized elders could contribute to fewer readmissions and lower costs. [bib_ref] Comprehensive discharge planning and home follow-up of hospitalized elders: a randomized clinical..., Naylor [/bib_ref] ## Intraoperative management Intraoperative management primarily focuses on the day of surgery, the goal of which is to optimize surgical and anesthetic procedures and minimize stress response. The associated measures included microinvasive surgery, anesthetic protocol, hypothermia avoidance, goal-directed fluid balance, scalp incision anesthesia, absorbable sutures, and restrictive drains [fig_ref] Table 1: Continued [/fig_ref]. ## Microinvasive surgery. The concept of microinvasive surgery involves not only less invasive approaches (e.g., keyhole surgery and endoscopic endonasal surgery), [bib_ref] Transcranial minimally invasive neurosurgery for tumors, Garrett [/bib_ref] but also maximal neuroprotection and functional preservation with the aid of multimodal neuronavigation, electrophysiologic monitoring, awake craniotomy, and intraoperative MRI. Surgery duration and estimated blood loss are established risk factors for postcraniotomy complications in the elderly. [bib_ref] Identification of preoperative and intraoperative risk factors for complications in the elderly..., Johans [/bib_ref] [bib_ref] Risk factors for postoperative systemic complications in elderly patients with brain tumors...., Asano [/bib_ref] Neurosurgeons and anesthesiologists should make efforts to minimize the length of the procedure and blood loss. ## Anesthetic protocol. For neurosurgical patients, though no significant difference is found between total intravenous anesthetic (TIVA) and inhalational anesthetic of sevoflurane regarding time to emergence from anesthesia and brain relaxation, TIVA seems to be moderately preferable in terms of a lower risk of PONV. [bib_ref] Intravenous versus inhalational techniques for rapid emergence from anaesthesia in patients undergoing..., Prabhakar [/bib_ref] [bib_ref] Postoperative nausea and vomiting after craniotomy: an evidence-based review of general considerations,..., Uribe [/bib_ref] [bib_ref] Comparison of propofol and volatile agents for maintenance of anesthesia during elective..., Chui [/bib_ref] Use of isoflurane in craniotomy patients is not recommended due to delayed emergence from anesthesia and decreased brain relaxation. [bib_ref] Intravenous versus inhalational techniques for rapid emergence from anaesthesia in patients undergoing..., Prabhakar [/bib_ref] TIVA or combined intravenous (IV)-inhalation anesthesia with short-acting agents could be applied according to institutional practice patterns and anesthetist preference. ## Hypothermia avoidance. Inadvertent perioperative hypothermia is a common and detrimental phenomenon that is associated with prolonged postanesthetic recovery and postoperative complications, including surgical site infection, delayed wound healing, pressure ulcer, bleeding, and cardiovascular events. [bib_ref] Active body surface warming systems for preventing complications caused by inadvertent perioperative..., Madrid [/bib_ref] [bib_ref] A systematic review of intraoperative warming to prevent postoperative complications, Scott [/bib_ref] Active warming has a beneficial effect in reducing complication rates, blood transfusion need, and medical costs. [bib_ref] Active body surface warming systems for preventing complications caused by inadvertent perioperative..., Madrid [/bib_ref] [bib_ref] A systematic review of intraoperative warming to prevent postoperative complications, Scott [/bib_ref] Awareness of the consequences of hypothermia is indispensable, and avoiding hypothermia is crucial for all surgical patients. According to ERAS protocols, forced-air or electric heating pads, together with warmed irrigation and IV fluids, are applied to all neurosurgical patients. [bib_ref] Safety and efficacy of a novel neurosurgical enhanced recovery after surgery protocol..., Wang [/bib_ref] [bib_ref] Enhanced recovery after intraspinal tumor surgery: a single-institutional randomized controlled study, Liu [/bib_ref] 3.2.4. Goal-directed fluid balance. During cranial surgery, anesthesiologists face the dilemma of restricting IV fluids to prevent brain swelling and maintain hemodynamic stability. To meet the dual goals of satisfactory exposure and hemodynamic stability, strategies for cardiac output-guided hemodynamic management and fluid balance with intraoperative goal-directed fluid restriction have been explored, which are associated with decreased intensive care unit stay, costs, and postoperative complications. [bib_ref] A systematic review and meta-analysis on the use of preemptive hemodynamic intervention..., Hamilton [/bib_ref] [bib_ref] Goal-directed fluid restriction during brain surgery: a prospective randomized controlled trial, Luo [/bib_ref] This strategy has become standard practice with consistent implementation of neurosurgical ERAS protocols, [bib_ref] Safety and efficacy of a novel neurosurgical enhanced recovery after surgery protocol..., Wang [/bib_ref] [bib_ref] Enhanced recovery after intraspinal tumor surgery: a single-institutional randomized controlled study, Liu [/bib_ref] and is applicable to geriatric patients. ## Local incision anesthesia. Placement of the pin head holder and scalp incision are the most painful stages of craniotomy. Scalp block or local incision anesthesia can reduce the hemodynamic response to painful stimuli, as well as postoperative pain, analgesic consumption, and PONV. [bib_ref] Safety and efficacy of a novel neurosurgical enhanced recovery after surgery protocol..., Wang [/bib_ref] [bib_ref] Enhanced recovery after intraspinal tumor surgery: a single-institutional randomized controlled study, Liu [/bib_ref] [bib_ref] Which one is more effective for analgesia in infratentorial craniotomy? The scalp..., Akcil [/bib_ref] [bib_ref] Effect of scalp blocks with levobupivacaine on recovery profiles after craniotomy for..., Hwang [/bib_ref] These are already routine procedures in several centers and are mandatory for all geriatric neurosurgical patients. ## Absorbable suture. In addition to achieving satisfactory wound healing, absorbable sutures offer cosmetic advantages and reduce discomfort of suture removal. [bib_ref] Absorbable intradermal closure of elective craniotomy wounds, Paolini [/bib_ref] [bib_ref] Skin closure in vascular neurosurgery: a prospective study on absorbable intradermal suture..., Pereira [/bib_ref] Absorbable sutures are a key component of our ERAS protocols and are associated with a decreased postoperative LOS. [bib_ref] Safety and efficacy of a novel neurosurgical enhanced recovery after surgery protocol..., Wang [/bib_ref] [bib_ref] Enhanced recovery after intraspinal tumor surgery: a single-institutional randomized controlled study, Liu [/bib_ref] Dura, muscle, and subcutaneous tissue are closed with interrupted absorbable sutures, and the skin incision is closed with an intradermal running suture. ## Restrictive surgical site drainage. Avoiding surgical site drainage is another distinguishing component of the ERAS protocol. If drain placement is deemed necessary by the chief surgeon, it is removed as early as possible after surgery, mostly within 24 to 48 hours. This measure is safe and effective in reducing postoperative discomfort associated with drains, promoting early ambulation, and shortening the LOS without increasing the rate of surgical complications. [bib_ref] Safety and efficacy of a novel neurosurgical enhanced recovery after surgery protocol..., Wang [/bib_ref] [bib_ref] Enhanced recovery after intraspinal tumor surgery: a single-institutional randomized controlled study, Liu [/bib_ref] ## Postoperative management Postoperative management aims to reduce postoperative morbidity and maximize patients' physical conditioning to promote functional recovery after surgery. This consists of nonopioid analgesia, PONV management, postoperative diet, restrictive IV fluids, urinary catheter removal, early ambulation, delirium prevention, and glycemic control [fig_ref] Table 1: Continued [/fig_ref]. ## Nonopioid analgesia. The incidence, magnitude, and duration of postoperative pain experienced by craniotomy patients have been underestimated for decades, and undertreated pain remains a problem. [bib_ref] Acute and chronic pain following craniotomy: a review, De Gray [/bib_ref] What is worse, acute postoperative pain becomes chronic pain in approximately 20% to 60% of patients after craniotomy. [bib_ref] Acute and chronic pain following craniotomy: a review, De Gray [/bib_ref] Patient-controlled analgesia (PCA) with morphine is an effective method in craniotomy patients, however, there are risks of PONV, sedation, respiratory depression, ileus, urinary retention, and cardiovascular events, especially in geriatric patients. [bib_ref] Acute and chronic pain following craniotomy: a review, De Gray [/bib_ref] [bib_ref] Comparison of the analgesic efficacy and respiratory effects of morphine, tramadol and..., Sudheer [/bib_ref] In our previous neurosurgical ERAS protocols, the traditional regimen of postoperative morphine PCA is abandoned, instead, we have applied a standardized and nonopioid analgesia strategy based on repeated assessments of pain visual analog scale (VAS) scores. [bib_ref] Safety and efficacy of a novel neurosurgical enhanced recovery after surgery protocol..., Wang [/bib_ref] [bib_ref] Enhanced recovery after intraspinal tumor surgery: a single-institutional randomized controlled study, Liu [/bib_ref] No analgesia or minimal oral nonopioid is prescribed for patients reporting mild pain (VAS score 1-3); regular oral or IV nonopioid is given in cases of moderate pain (VAS score 4-6); opioid is reserved for patients with severe pain (VAS score 7-10) while morphine PCA is only used in refractory cases. [bib_ref] Safety and efficacy of a novel neurosurgical enhanced recovery after surgery protocol..., Wang [/bib_ref] [bib_ref] Enhanced recovery after intraspinal tumor surgery: a single-institutional randomized controlled study, Liu [/bib_ref] This strategy has achieved satisfactory pain control with significantly decreased opioid consumption and PCA use. The shift from morphine PCA to nonopioid analgesia can not only mitigate the potential side effects of opioids, such as nausea, sedation, and dizziness, but also enhance postoperative mobility. [bib_ref] A comprehensive multimodal pain treatment reduces opioid consumption after multilevel spine surgery, Mathiesen [/bib_ref] Additionally, local incision anesthesia and active management of PONV also contribute to improved pain control. [bib_ref] Safety and efficacy of a novel neurosurgical enhanced recovery after surgery protocol..., Wang [/bib_ref] [bib_ref] A comprehensive multimodal pain treatment reduces opioid consumption after multilevel spine surgery, Mathiesen [/bib_ref] As an adjunct to general anesthesia, intraoperative dexmedetomidine, ketamine, and lidocaine have been used for acute postcraniotomy pain control. [bib_ref] Enhanced recovery after surgery for oncological craniotomies, Hagan [/bib_ref] Of note, Dexmedetomidine is associated with reduced PONV in addition to reduced postoperative pain and analgesic consumption. [bib_ref] Effect of intraoperative dexmedetomidine on post-craniotomy pain, Peng [/bib_ref] Lastly, preemptive analgesia with medications such as nonsteroidal anti-inflammatory drugs, acetaminophen, and gabapentinoids is another strategy to prevent postoperative pain by inhibiting autonomic hyperactivity, which also reduces PONV. [bib_ref] Guidelines for perioperative care in elective colorectal surgery: enhanced recovery after surgery..., Gustafsson [/bib_ref] [bib_ref] Preemptive analgesia after lumbar spine surgery by pregabalin and celecoxib: a prospective..., Kien [/bib_ref] [bib_ref] Administration of preemptive analgesia by diclofenac to prevent acute postcraniotomy headache, Simon [/bib_ref] Taken together, a multimodal nonopioid analgesia strategy as a continuum pre-,intra-,and postoperatively is proposed for geriatric neurosurgical patients. ## Ponv management. The high prevalence of PONV in craniotomy patients (ranged 20%-70%) and the increased risks of intracranial bleeding, brain edema, and aspiration associated with PONV signify the importance of appropriate PONV management, which consists of perioperative risk assessment, management of surgical and anesthesia-related risk factors, and multimodal pharmacological interventions targeting different chemoreceptors in the vomiting center. [bib_ref] Postoperative nausea and vomiting after craniotomy: an evidence-based review of general considerations,..., Uribe [/bib_ref] [bib_ref] Postoperative nausea and vomiting in patients after craniotomy: incidence and risk factors, Latz [/bib_ref] The most effective and commonly used medications are 5-HT3 antagonists, glucocorticoids, NK1 receptor antagonists, or their combination. [bib_ref] Postoperative nausea and vomiting after craniotomy: an evidence-based review of general considerations,..., Uribe [/bib_ref] According to our previous successful ERAS protocols, [bib_ref] Safety and efficacy of a novel neurosurgical enhanced recovery after surgery protocol..., Wang [/bib_ref] [bib_ref] Enhanced recovery after intraspinal tumor surgery: a single-institutional randomized controlled study, Liu [/bib_ref] active PONV prophylaxis with dexamethasone and tropisetron is given to patients with a PONV risk score≥3 according to the PONV Simplified Risk Assessment Scale. [bib_ref] A simplified risk score for predicting postoperative nausea and vomiting: conclusions from..., Apfel [/bib_ref] Postoperative treatment with repeated tropisetron is given to patients reporting a PONV VAS score ≥5, and a combination of tropisetron and droperidol, or promethazine is given to refractory cases. [bib_ref] Safety and efficacy of a novel neurosurgical enhanced recovery after surgery protocol..., Wang [/bib_ref] [bib_ref] Enhanced recovery after intraspinal tumor surgery: a single-institutional randomized controlled study, Liu [/bib_ref] 3.3.3. Postoperative diet. Early restoration of oral nutrition is a foundation of all ERAS protocols. [bib_ref] Early enteral nutrition within 24h of colorectal surgery versus later commencement of..., Andersen [/bib_ref] Given the nature of cranial surgery where the operative site is independent of the gastrointestinal tract, as well as the use of nonopioid analgesia and active PONV management, early diet advancement is encouraged for all craniotomy patients except for those remaining comatose for a prolonged period. [bib_ref] Safety and efficacy of a novel neurosurgical enhanced recovery after surgery protocol..., Wang [/bib_ref] [bib_ref] Enhanced recovery after surgery for oncological craniotomies, Hagan [/bib_ref] Oral free fluids are permitted as early as 4 hours after surgery; light diet or a polymeric nutritional supplement drink is given 8 hours after surgery, semiliquid/solid diet is allowed 12 to 24 hours after surgery, and restoration of ordinary solid diet is achieved within 24 to 48 hours after surgery. [bib_ref] Safety and efficacy of a novel neurosurgical enhanced recovery after surgery protocol..., Wang [/bib_ref] [bib_ref] Enhanced recovery after intraspinal tumor surgery: a single-institutional randomized controlled study, Liu [/bib_ref] This strategy is well tolerated by the neurosurgical patients and is associated with improved functional status compared with the conventional perioperative care which roughly doubles time to diet advancement. [bib_ref] Safety and efficacy of a novel neurosurgical enhanced recovery after surgery protocol..., Wang [/bib_ref] [bib_ref] Enhanced recovery after intraspinal tumor surgery: a single-institutional randomized controlled study, Liu [/bib_ref] In addition, chewing gum is an convenient, economical and harmless way for amelioration of postoperative ileus and recovery of gastrointestinal function. [bib_ref] Chewing gum for postoperative recovery of gastrointestinal function, Short [/bib_ref] In keeping with the evidence, an early diet advancement regimen combined with chewing gum is provided for geriatric neurosurgical patients. Supplementary immunonutrition should be considered for malnourished patients with cancer. [bib_ref] Enhanced recovery after surgery for oncological craniotomies, Hagan [/bib_ref] [bib_ref] Guidelines for perioperative care in elective colorectal surgery: enhanced recovery after surgery..., Gustafsson [/bib_ref] 3.3.4. Urinary catheter removal. Prolonged use of urinary catheters is associated with an increased risk of urinary tract www.md-journal.com infection in surgical patients. [bib_ref] Using NSQIP to investigate SCIP deficiencies in surgical patients with a high..., Trickey [/bib_ref] Considering the preventable nature of hospital-acquired urinary tract infection, evidencebased preventive intervention is to remove the urinary catheter on postoperative day (POD1) or as early as possible according to daily assessment of the need for maintaining the catheter. [bib_ref] Enhanced recovery after surgery for oncological craniotomies, Hagan [/bib_ref] [bib_ref] Using NSQIP to investigate SCIP deficiencies in surgical patients with a high..., Trickey [/bib_ref] Early removal of the urinary catheter within 6 hours is feasible in neurosurgical patients, which promotes mobility and shortens postoperative LOS without increasing the rates of urinary retention. [bib_ref] Safety and efficacy of a novel neurosurgical enhanced recovery after surgery protocol..., Wang [/bib_ref] [bib_ref] Enhanced recovery after intraspinal tumor surgery: a single-institutional randomized controlled study, Liu [/bib_ref] Moreover, implementing a 2-to 3-hour voiding schedule immediately after catheter removal, using bladder scanning, and encouraging early and aggressive mobilization may help to decrease the risk of urinary retention. Attempts to remove the urinary catheter within POD1 are recommended for geriatric neurosurgical patients, and regular assessment is required to maintain the catheter to ensure removal as early as possible. ## Early mobilization and ambulation. Early mobilization and ambulation is an integral and proven strategy to prevent muscle loss, improve cardiopulmonary function, and reduce deep vein thrombosis risk and insulin resistance. [bib_ref] Enhanced recovery after surgery: a review, Ljungqvist [/bib_ref] [bib_ref] Computerized tomographic determination of human thigh components. The effects of immobilization in..., Ingemann-Hansen [/bib_ref] [bib_ref] Adherence to early mobilisation: key for successful enhanced recovery after liver resection, Yip [/bib_ref] Adherence to early mobilization and reduction in associated complications are keys for successful ERAS protocols. [bib_ref] Adherence to early mobilisation: key for successful enhanced recovery after liver resection, Yip [/bib_ref] With adequate pain control and PONV management, offbed ambulation is achievable for neurosurgical patients on POD1. [bib_ref] Safety and efficacy of a novel neurosurgical enhanced recovery after surgery protocol..., Wang [/bib_ref] [bib_ref] Enhanced recovery after intraspinal tumor surgery: a single-institutional randomized controlled study, Liu [/bib_ref] Postoperative fear of movement is addressed and actively management by the doctors, nurses, physiotherapists and psychiatrists. Patients were instructed and monitored by physiotherapists to start in-bed limb exercises 6 hours after surgery and ambulation within 24 hours after surgery. ## Delirium prevention. Given the high prevalence of postoperative delirium in the elderly and the lack of first-choice treatment, prevention is the best strategy. Intraoperative use of a depth of anesthesia monitor and sedation as light as possible are recommended. [bib_ref] The role of anesthesia in the prevention of postoperative delirium: a systematic..., Orena [/bib_ref] Additionally, avoiding medications such as anticholinergics, antipsychotics, and benzodiazepines whenever possible is vital in elderly patients, considering the risk of causing/ worsening dementia, delirium, and falls.Furthermore, ERAS components of prehabilitation, early oral feeding, early mobilization, restrictive use of drains and invasive tubes, as well as regular assessment and nursing care all contribute to a lower risk of delirium. [bib_ref] Postoperative cognitive dysfunctioncurrent preventive strategies, Kotekar [/bib_ref] [bib_ref] Postoperative cerebral dysfunction in the elderly: diagnosis and prophylaxis, Benhamou [/bib_ref] ## Outcome measures Outcome measures included objective and subjective patient-reported outcomes [fig_ref] Table 2: Outcome measures [/fig_ref]. Hospital LOS is the most commonly used outcome measure in ERAS protocols as a proxy for functional status and a key criterion for evaluating a successful ERAS protocol. [bib_ref] Adherence to early mobilisation: key for successful enhanced recovery after liver resection, Yip [/bib_ref] [bib_ref] Systematic review of outcomes used to evaluate enhanced recovery after surgery, Neville [/bib_ref] Other objective outcomes include postoperative morbidity and mortality, surgical and nonsurgical complications, reoperation and readmission rates, functional recovery status (such as pulmonary function, muscle strength, nutritional parameters, diet advancement, and mobility), analgesic consumption, and healthcare costs. Patient-reported outcomes have been suggested as the gold standard for assessing the clinical efficacy of surgical interventions and patient-perceived quality of care, and thus are preferable outcome measures for ERAS protocols. [bib_ref] Neurosurgical enhanced recovery after surgery (ERAS) programme for elective craniotomies: are patients..., Liu [/bib_ref] [bib_ref] Pre-optimization of spinal surgery patients: development of a neurosurgical enhanced recovery after..., Ali [/bib_ref] [bib_ref] Adherence to early mobilisation: key for successful enhanced recovery after liver resection, Yip [/bib_ref] [bib_ref] Systematic review of outcomes used to evaluate enhanced recovery after surgery, Neville [/bib_ref] The most frequently used and important outcomes are patient-reported symptom status outcomes, such as pain, PONV, fatigue, anxiety and depression, general health perceptions and quality of life, and patient satisfaction and comfort. # Discussion With the successful development and implementation of the neurosurgical ERAS protocol for elective neurosurgical patients in clinical trials, [bib_ref] Safety and efficacy of a novel neurosurgical enhanced recovery after surgery protocol..., Wang [/bib_ref] [bib_ref] Neurosurgical enhanced recovery after surgery (ERAS) programme for elective craniotomies: are patients..., Liu [/bib_ref] [bib_ref] Enhanced recovery after intraspinal tumor surgery: a single-institutional randomized controlled study, Liu [/bib_ref] the proven benefits of reducing LOS, accelerating postoperative functional recovery, and improving patient satisfaction have called into attention the constant and universal application of the ERAS protocol in everyday practice. Compared with other surgical specialties, neurosurgery, especially cranial surgery, lags behind in the era of ERAS, and perioperative care continues to resemble conventional methods in many centers. [bib_ref] Enhanced recovery after surgery: a review, Ljungqvist [/bib_ref] [bib_ref] Enhanced recovery after surgery (ERAS) for spine surgery: a systematic review, Dietz [/bib_ref] [bib_ref] Pre-optimization of spinal surgery patients: development of a neurosurgical enhanced recovery after..., Ali [/bib_ref] As the most populous country in the world, China has the largest aging population, which is confronting healthcare providers with greater challenges in managing multiple comorbidities associated with advanced age and improving the quality of perioperative care. This multimodal, multidisciplinary, and evidence-based ERAS protocol for geriatric patients undergoing elective craniotomy aimed to establish a feasible, applicable, reasonable, and standard pattern of practice that involves and engages healthcare providers from multiple services. Under the coordination and cooperation of the Neurosurgical ERAS Working Group, cumulative benefits contributed by individual components of the ERAS protocol may translate into significant improvements in patient recovery after cranial surgery. Moreover, given the multistep nature of an ERAS protocol, both patient and provider compliance are decisive factors for success or failure. A systemic audit of process adherence and outcomes is a useful tool for assessing impact and enhancing compliance. [bib_ref] Safety and efficacy of a novel neurosurgical enhanced recovery after surgery protocol..., Wang [/bib_ref] [bib_ref] Enhanced recovery after intraspinal tumor surgery: a single-institutional randomized controlled study, Liu [/bib_ref] [bib_ref] Enhanced recovery after surgery for oncological craniotomies, Hagan [/bib_ref] # Conclusion The current proposal for a multimodal, multidisciplinary, and evidence-based ERAS protocol for geriatric patients undergoing elective craniotomy is feasible and crucial for optimizing perioperative care in this patient population. This approach is generalizable from the previous accomplishment of a neurosurgical ERAS protocol for adult patients ages 18 to 65 years and tailored to the geriatric population. Implementation of the current protocol may hold promise in reducing perioperative morbidity, enhancing functional recovery, shortening LOS, and serving as a stepping stone to guide future efforts to improve geriatric patient care. [fig] Follow: -up Home + clinic follow-up Timely and responsive follow-up with social media cellphone/website app + outpatient clinic visit Functional status, pain medication use, late-onset complications, quality of life Audit Standardized audit and feedback Data and safety monitoring to document process/outcome data, assess impact and encourage compliance BMI = body mass index, GCS = graduated compression stockings, GDFR = goal-directed fluid restriction, ICU = intensive care unit, IPC = intermittent pneumatic compression, IV = intravenous, NSAIDs = nonsteroidal anti-inflammatory drugs, PCA = patient-controlled analgesia, POD = postoperative day, PONV = postoperative nausea and vomiting, TIVA = total intravenous anesthetic, VAS = visual analog scale. [/fig] [table] Table 2: Outcome measures. DVT = deep vein thrombosis, EORTC QLQ-C30/BN20 = European Organisation for Research and Treatment of Cancer Quality of Life Core Questionnaire 30/Brain Cancer Module, KPS = Karnofsky performance status, LOS = length of stay, PONV = postoperative nausea and vomiting, SSI = surgical site infection, UTI = urinary tract infection, VAS = visual analog scale, VTE = venous thromboembolism. Medicine [/table]

Myrrh for treatment of severe vulvar edema in ovarian hyperstimulation syndrome☆ A B S T R A C TBackground: Severe vulvar edema is a rare entity occurring with ovarian hyperstimulation syndrome. This edema can be incapacitating; causing pain and limited patient mobility. With the usual conservative approach, vulvar edema can take several days to resolve. Aim: The aim of this case report is to describe the use of local myrrh for the management of severe vulvar edema associated with ovarian hyperstimulation syndrome. Case Presentation: 29-year-old female with severe vulvar edema associated with ovarian hyperstimulation syndrome. Conclusion: Local myrrh application for severe vulvar edema in ovarian hyperstimulation syndrome resulted in substantial improvement, and with further studies, myrrh could be used as an option for the management of vulvar edema. # Introduction Ovarian hyperstimulation syndrome is an iatrogenic complication of assisted reproductive technologies [bib_ref] Preventing ovarian hyperstimulation syndrome: guidance for the clinician, Humaidan [/bib_ref]. With a reported incidence of 1% to 5% in its moderate and severe forms, it can be life-threatening with 1.9% cases requiring hospitalization [bib_ref] Preventing ovarian hyperstimulation syndrome: guidance for the clinician, Humaidan [/bib_ref] [bib_ref] Oocyte number as a predictor for ovarian hyperstimulation syndrome and live birth:..., Steward [/bib_ref]. Although there is no agreed upon definition, the disorder is characterized by marked cystic ovarian enlargement, fluid shifts from intravascular compartment resulting in ascites and hydrothorax, electrolyte imbalance, hemoconcentration, hypercoagulation, and impaired renal perfusion [bib_ref] Preventing ovarian hyperstimulation syndrome: guidance for the clinician, Humaidan [/bib_ref] [bib_ref] Oocyte number as a predictor for ovarian hyperstimulation syndrome and live birth:..., Steward [/bib_ref]. Risk factors for ovarian hyperstimulation syndrome include young age, history of ovarian hyperstimulation, or polycystic ovarian syndrome [bib_ref] Preventing ovarian hyperstimulation syndrome: guidance for the clinician, Humaidan [/bib_ref]. Risk reducing strategies include coasting, cycle cancellation, cryopersavation of embryos, low-dose hCG instead of FSH, GnRH agonist for final follicular maturation instead of hCG, and cabergoline [bib_ref] Cabergoline for the prevention of ovarian hyperstimulation syndrome: systematic review and metaanalysis..., Leitao [/bib_ref]. Cabergoline is particularly useful as it does not seem to decrease implantation rate and can be used in all high risk women [bib_ref] Cabergoline for the prevention of ovarian hyperstimulation syndrome: systematic review and metaanalysis..., Leitao [/bib_ref]. The first case of vulvar edema associated with severe ovarian hyperstimulation syndrome was described in 1995 by Coccia et al. [bib_ref] Massive vulvar edema in ovarian hyperstimulation syndrome. A case report, Coccia [/bib_ref]. Their patient had massive vulvar edema associated with fissures [bib_ref] Massive vulvar edema in ovarian hyperstimulation syndrome. A case report, Coccia [/bib_ref]. Her management was conservative with human albumin, lactated Ringer's, and heparin [bib_ref] Massive vulvar edema in ovarian hyperstimulation syndrome. A case report, Coccia [/bib_ref]. The vulva was back to normal after one week of treatment which consisted of ice packs, topical hydrocortisone, and twice daily topical gentamicin [bib_ref] Massive vulvar edema in ovarian hyperstimulation syndrome. A case report, Coccia [/bib_ref]. Our patient also had severe vulvar edema that was painful and prevented her from mobilization. She was conservatively treated with local application of ice packs and Myrrh. It was surprising that the edema which was deferring the patient from moving around easily was resolved in two days. It's quite common for herbal medicine to be used in Saudi Arabia [bib_ref] Risks of Myrrh usage in pregnancy, Al-Jaroudi [/bib_ref]. People still strongly believe in the role of herbs with Myrrh being one of the most commonly used herb in the Arabian Peninsula [bib_ref] Risks of Myrrh usage in pregnancy, Al-Jaroudi [/bib_ref]. Myrrh is consituited of volatile essential oils, sesquiterpenes, and a water soluble gum [bib_ref] Risks of Myrrh usage in pregnancy, Al-Jaroudi [/bib_ref]. In large doses it may be unsafe and may affect the kidney and the heart [bib_ref] Risks of Myrrh usage in pregnancy, Al-Jaroudi [/bib_ref]. It poses analgesic, anti-inflammaty, antiseptic, antifungal, antispasmodic, astringent, carminative, emmenagogue, expectorant, and anti-hyperlipidemic effects [bib_ref] Risks of Myrrh usage in pregnancy, Al-Jaroudi [/bib_ref]. We are reporting a case of severe vulvar edema treated with local myrrh application. Patient approved the use of the picture for educational and research purposes. Institutional review board approval was obtained to report the case. ## Case report Our patient is a 29-year-old, G2P0+ 1 who was referred to our On her arrival to the emergency department, patient was complaining of diffuse abdominal distension, vomiting, and severe shortness of breath. She was also complaining of severe vulvar swelling limiting her mobility [fig_ref] Figure 1: Vulvar edema after beginning of treatment with partial improvement [/fig_ref]. On assessment patient was in distress, abdomen was tense and distended, fluid thrill and shifting dullness were positive. Patient was admitted to the gynecology ward as a case of severe ovarian hyperstimulation syndrome. She was closely observed with daily measurement of her weight, abdominal girth and daily labs including complete blood counts with platelet, coagulation screen, kidney and liver profiles. She was monitored for input and output with strict charting. She was also given analgesia, low molecular weight heparin, insulin and folic acid. Ultrasound showed a uterus that is normal in size, shape and texture. Right ovary measured 15 * 9.8 * 12 cm with multiple cysts, and the left ovary measured 16 * 9.4 * 12 cm. Additionally, marked ascites was reported. Patient had low albumin and her electrolytes were normal. An 8 French catheter abdominal drain was inserted by the interventional radiology under ultrasound guidance. After the drainage of ascitic fluid, the patient started to improve. Her breathing was better, and her abdomen became soft but remained distended. The total amount of fluid drained was seven liters. Her vulvar edema [fig_ref] Figure 1: Vulvar edema after beginning of treatment with partial improvement [/fig_ref] was managed with ice packs, and local myrrh which was dissolved in warm water and applied locally. There was complete resolution of edema in two days. Patient was also kept in Trendelenburg position to help decrease the swelling. She gradually improved and was able to mobilize. Ultrasound [fig_ref] Figure 2: Ultrasound image of patient's right ovary [/fig_ref] was repeated during her admission and showed two small intrauterine gestational sacs, and enlarged ovaries. She was discharged home after one week in a stable condition. # Discussion A dangerous complication of ovarian stimulation is ovarian hyperstimulation syndrome (OHSS) [bib_ref] GnRH antagonist rescue protocol combined with cabergoline versus cabergoline alone in the..., Fouda [/bib_ref]. It consists of cystic enlargement of ovaries, increased capillary permeability and fluid shift into the third space [bib_ref] GnRH antagonist rescue protocol combined with cabergoline versus cabergoline alone in the..., Fouda [/bib_ref]. Consequently, OHSS leads to hemoconcentration, hypoxia and hemodynamic instability [bib_ref] Is severe OHSS associated with adverse pregnancy outcomes? Evidence from a case-control..., Haas [/bib_ref]. The syndrome is activated through an increase in hCG that causes an increase in vascular endothelial growth factor, which then stimulates the vascular endothelium [bib_ref] Oocyte number as a predictor for ovarian hyperstimulation syndrome and live birth:..., Steward [/bib_ref]. Although vascular endothelial growth factor (VEGF) has a major role in the development of OHSS, other factors may play a role including VEGF's soluble receptor sFlt-1, other cytokines and growth factors [bib_ref] Is severe OHSS associated with adverse pregnancy outcomes? Evidence from a case-control..., Haas [/bib_ref]. In an attempt to reduce the risk of ovarian hyperstimulation syndrome, a number of measures have been attempted including the use of GnRH antagonist protocol, coasting and Cabergoline [bib_ref] GnRH antagonist rescue protocol combined with cabergoline versus cabergoline alone in the..., Fouda [/bib_ref]. Coasting, in which gonadotropin administration is stopped, is the most often used [bib_ref] GnRH antagonist rescue protocol combined with cabergoline versus cabergoline alone in the..., Fouda [/bib_ref]. It is important to keep in mind that a longer duration of coasting (more than three days) is associated with a decrease in the number of oocytes collected, implantation and clinical pregnancy rate [bib_ref] GnRH antagonist rescue protocol combined with cabergoline versus cabergoline alone in the..., Fouda [/bib_ref]. In 1996, Luxman et al. described nine cases of unilateral vulvar edema following paracentesis in patients with severe ovarian hyperstimulation syndrome [bib_ref] Unilateral vulvar edema associated with paracentesis in patients with severe ovarian hyperstimulation..., Luxman [/bib_ref]. They believed that a fistulous tract was created by the needle when the lower abdomen was the puncturing site [bib_ref] Unilateral vulvar edema associated with paracentesis in patients with severe ovarian hyperstimulation..., Luxman [/bib_ref]. Ascitic fluid was forced due to increased intraabdominal pressure into the subcutaneous tissue, leading to unilateral vulvar edema [bib_ref] Unilateral vulvar edema associated with paracentesis in patients with severe ovarian hyperstimulation..., Luxman [/bib_ref]. The patients were treated conservatively and the edema was resolved within 10 days [bib_ref] Unilateral vulvar edema associated with paracentesis in patients with severe ovarian hyperstimulation..., Luxman [/bib_ref]. Another case of bilateral vulvar edema following bilateral paracentesis was described by Vavillis et al. [bib_ref] Postparacentesis bilateral massive vulvar edema in a patient with severe ovarian hyperstimulation..., Vavilis [/bib_ref]. Our case describes bilateral severe vulvar edema in a patient with severe ovarian hyperstimulation syndrome. The edema was severe enough to cause pain and prevent the patient from ambulation. Our patient was treated conservatively and was given myrrh to apply directly to the edematous vulva. This greatly helped in accelerating the resolution of the vulvar edema, and by the time of discharge, she was ambulating without difficulty. We believe that myrrh, due to its astringent properties, can be used to help accelerate the resolution of vulvar edema in patients. Further studies are needed to assess the benefits of such treatment for similar cases of OHSS. # Conclusion Myrrh's use has shown a beneficial effect in the acceleration of resolution of vulvar edema and may with further studies be used to treat more patients. # Disclosure We have nothing to disclose and the patient consented to the publishing of the image and material. There were no funds received. [fig] Figure 1: Vulvar edema after beginning of treatment with partial improvement. [/fig] [fig] Figure 2: Ultrasound image of patient's right ovary. [/fig] [fig] Figure 3: Ultrasound image of patient's left ovary. [/fig]

Clinical aspects of TP53 gene inactivation in diffuse large B-cell lymphoma Background: The knowledge about specific mechanisms generating TP53 dysfunction in diffuse large B-cell lymphoma is limited. The aim of the current study was to comprehensively explore TP53 gene variability resulting from somatic mutations, promoter methylation, and allelic imbalance in tumorous tissue of diffuse large B-cell lymphoma (DLBCL). Methods: DNA samples from 74 patients with DLBCL were used. Genomic DNA was isolated from paraffin blocks of lymph nodes or from extranodal biopsies of tumors by the phenol-chloroform extraction method with guanidine. Analysis of coding sequences of the TP53 gene was based on Sanger's direct sequencing method. The methylation status of the TP53 promoter was analyzed using by methylation-specific PCR on bisulfite-converted DNA. Assessment of the detected mutations was carried out in the IARC TP53 Database and the TP53 UMD mutation database of human cancer. Results: The mutations in regions coding for the DNA-binding domain were prevalent (95%). In the analyzed sample of patients, codons 275, 155, 272, and 212 were hotspots of mutations in the TP53 gene. In addition, functionally significant intron mutations (IVS6-36G > C and IVS5 + 43G > T) were detected. Instances of TP53 promoter methylation were observed only in a few samples of diffuse large B-cell lymphoma tissue. Furthermore, loss of heterozygosity was revealed only in the subgroup of patients with altered status of the gene (mutations were detected in five patients and promoter methylation in one case). Conclusions: Thus, the results suggest that there are two sequential events in the formation of diffuse large B-cell lymphoma in at least some cases. The first event is mutation or methylation of the TP53 promoter, leading to appearance of a cell with increased risk of malignant transformation. The second event is the loss of an intact allele of the gene; this change is necessary for tumorigenesis. We identified TP53 mutation patterns in a Russian cohort of patients with de novo DLBCL who were treated with R-CHOP and R-CHOP-like regimens and confirmed that TP53 mutation status is a valuable prognostic biomarker. # Introduction Diffuse large B-cell lymphoma (DLBCL) is characterized by diffuse proliferation of atypical large lymphocytes containing a vesicular nucleus, prominent nucleoli, and basophilic cytoplasm. DLBCL occurs in one third of cases of non-Hodgkin's lymphoma among adults: up to 25-30% in developed countries and 30-40% in developing countries; these statistics make it one of the most frequent types of lymphoma in the world [bib_ref] Genetics and diffuse large B-cell lymphoma, Niroula [/bib_ref]. An important mechanism underlying the development of DLBCL is the genetic instability of lymphoid cells as part of normal maturation of B cells; this instability can lead to precancerous genetic lesions. As a result, a disturbance of B-cell homeostasis with unregulated proliferation, differentiation blockage, and B-cell immortalization occurs at one of the stages of lymphoid-cell maturation [bib_ref] Genetic susceptibility to lymphoma, Skibola [/bib_ref]. Genetic factors that disrupt DNA repair or apoptosis may increase the risk of precancerous events. The lesions in the B-lymphocyte genome that had not been repaired or had not been eliminated by apoptosis may be modulated in the future by environmental influences, epigenetic factors (hypo-or hypermethylation), concomitant (autoimmune) diseases, and/or genetic polymorphism and may promote further tumorigenesis [bib_ref] Dysfunction of the TP53 tumor suppressor gene in lymphoid malignancies, Xu-Monette [/bib_ref]. The protein p53 is a nuclear phosphoprotein playing a crucial role in rapid elimination of damaged and potentially dangerous cells [bib_ref] Massively regulated genes: the example of TP53, Hollstein [/bib_ref]. Its tumor-suppressing function results from participation in such processes as cell cycle control, DNA repair, apoptosis, aging, and autophagy through transcription-dependent and -independent mechanisms [bib_ref] The significance of TP53 in lymphoid malignancies: mutation prevalence, regulation, prognostic impact..., Cheung [/bib_ref]. Lymphocytes under stress tend to go through p53-dependent apoptosis, in contrast to other cell types, which undergo cell cycle arrest as well as p53-independent apoptosis or necrosis under stressful conditions [bib_ref] The role of p53 in determining sensitivity to radiotherapy, Gudkov [/bib_ref]. For this reason, dysfunction of the TP53 gene is a basis for initiation and progression of lymphoproliferative disorders [bib_ref] The role of p53 in determining sensitivity to radiotherapy, Gudkov [/bib_ref] [bib_ref] TP53 in hematological cancer: low incidence of mutations with significant clinical relevance, Peller [/bib_ref]. An increase in genetic instability that promotes further tumor progression and allows malignant cells to escape immunosurveillance and therapeutic interventions has been observed in B lymphocytes with an inactivated TP53 gene. Acceleration of the pace of polyclonal evolution of B cells-with various genetic abnormalities, such as changes in chromosome numbers, chromosomal rearrangements, gene mutations, and amplification of some regions of the genome-takes place under conditions of p53 dysfunction [bib_ref] Multifaced p53: variety of forms, functions, tumor-supressive and oncogenic activities, Kopnin [/bib_ref]. Dysfunction of the p53 protein may be due to disturbances in the structure of the gene, changes in the transcription process and stability of mRNA or malfunction of post-translational modifications or of interactions of the p53 protein. Probably, molecular mechanisms involving DNA and leading to dysfunction of p53 include gene mutations, promoter methylation, allelic imbalance, and genetic polymorphism [bib_ref] Dysfunction of the TP53 tumor suppressor gene in lymphoid malignancies, Xu-Monette [/bib_ref]. # Materials and methods The aim of this study was to comprehensively describe the frequency of promoter methylation and that of loss of heterozygosity as well as the frequency, diversity, and functional significance of mutations in coding and intron regions of the TP53 gene among patients with DLBCL in Novosibirsk, Russia. ## Study population The study population included 74 patients with DLBCL (35 men and 39 women), aged 21-78 years (52.8 ± 14.3, mean ± SD), who were admitted to Novosibirsk Hematological Center during 2012-2015. As many as 91% of these patients had advanced (III-IV) stages of the disease and two-thirds of them had a poor prognosis according to the International Prognostic Index (IPI). All patients underwent 6-8 cycles of R-CHOP-21 and R-CHOP-like regimens. The [fig_ref] Table 1: Clinical features of DLBCL patients at the time of diagnosis [/fig_ref]. presents summarizing patients characteristics. ## Genomic dna isolation Genomic DNA was isolated from paraffin blocks of lymph nodes or from extranodal biopsies of tumors by the phenol-chloroform extraction method with guanidine. The tissue sections containing at least 70-80% of tumor cells were chosen for analysis. ## Tp53 gene sequencing and mutation analysis A prescreening of mutations was not performed. Analysis of coding sequences of the TP53 gene (from exon 3 to exon 10) and of adjacent intron regions was carried out by Sanger's direct sequencing method, according to the IARC protocol (2010 update) . At the first stage, single fragments of DNA were produced by PCR, with the genomic DNA as a template. The obtained amplicons were desalted and cleaned up from unincorporated primers and deoxynucleotide triphosphates on microcolumns with Sephadex ТМ G-50 resin (GE Healthcare Bio-Sciences AB). The sequencing of samples was carried out by the method of capillary electrophoresis on a Hitachi 3500 Genetic Analyzer (Applied Biosystems) with the BigDye® Terminator v.3.1 Kit (Applied Biosystems). Analysis of the sequencing results and alignment and comparison of the obtained data with a reference sequence were conducted in software packages Chromas, SeqScape v.2.7, and Sequence Scanner. Assessment of biological significance of the detected mutations was carried out in the IARC TP53 Database and the TP53 UMD mutation database of human cancer [bib_ref] Datadriven unbiased curation of the TP53 tumor suppressor gene mutation database and..., Edlund [/bib_ref] [bib_ref] Impact of mutant p53 functional properties on TP53 mutation patterns and tumor..., Petitjean [/bib_ref]. In addition, bioinformatic analysis of mutations was conducted in SIFT, Mut_ass and on-line program Polymorphism Phenotyping 2 (PolyPhen-2) [bib_ref] Predicting functional effect of human missense mutations using PolyPhen-2, Adzhubei [/bib_ref]. To evaluate the biological significance of substitutions in introns, the NetGene2 software was employed [bib_ref] Prediction of human mRNA donor and acceptor sites from the DNA sequence, Brunak [/bib_ref]. ## Methylation-specific pcr The bisulfite conversion of DNA samples was performed by means of the EZ DNA Methylation Kit (Zymo Research, USA). Three hundred to 500 ng of DNA was used per reaction. Analysis of the methylation status of the TP53 promoter was carried out by methylation-specific PCR on bisulfite-converted DNA in two microtubes with primers specific to methylated and unmethylated alleles, in accordance with the method described above [bib_ref] Polymorphisms and DNA methylation of gene TP53 associated with extraaxial brain tumors, Almeida [/bib_ref]. For methodological reasons, detection of the methylation status of the TP53 promoter was performed on tumor samples from 69 patients with DLBCL by the methylation-specific PCR method [fig_ref] Figure 1: The analysis of promoter methylation status in the TP53 gene [/fig_ref]. # Microsatellite analysis Assessment of the loss of heterozygosity of TP53 was carried out at microsatellite locus D17S796 by a PCR method [bib_ref] Frequent loss of heterozygosity on chromosomes 3p and 17p without VHL or..., Grebe [/bib_ref]. In this analysis, 24 pairs of samples of normal and tumorous tissue from patients with DLBCL were used [fig_ref] Figure 2: The analysis of microsatellite instability in locus D17S796 [/fig_ref]. # Statistical analysis A comparison of type frequencies of nucleotide substitutions in TP53 in DLBCL between the examined sample of patients and data in the IARC TP53 mutation database was performed by statistical methods: Pearson's χ 2 test and Fisher exact test. Clinical and laboratory features were compared using the Fisher exact test. Differences were considered statistically significant at p < 0.05. # Results ## Mutations in coding and intron sequences of the tp53 gene Overall, 33 mutations were revealed: 21 in coding and 12 in intron sequences of TP53 . The following distribution of mutations was observed [fig_ref] Table 2: General characteristics of sequencing results [/fig_ref] : 1 (3%) mutation causing a defect of RNA splicing, 11 (33%) intron mutations with an unknown effect, 12 (37%) missense mutations, 6 (18%) sense mutations, 2 (6%) nonsense mutations, and 1 (3%) frameshift mutation in the TP53 gene. Except for A189Pfs, all these mutations (96.9%) were single-nucleotide substitutions, 5 (15.6%) of which were mutations of type GC > AT in CpG islands. Substitutions GC > AT constituted 34.4%, GC > CG 3.1%, GC > TA 9.4%, AT>GC 12.5%, AT>CG 12.5%, and AT>TA substitutions represented 12.5%; these All mutations in the coding sequences of TP53 that were identified in our sample of patients with DLBCL had been described earlier in the IARC TP53 mutations database [bib_ref] Datadriven unbiased curation of the TP53 tumor suppressor gene mutation database and..., Edlund [/bib_ref] in other oncological diseases, and these mutations (with the exception of р.A307A) are located in exons 5-8 coding for the DNA-binding domain p53. In the examined sample, 4 (6.8%) patients had multiple mutations, and some findings were revealed repeatedly (each in two cases) (in coding sequences: p.W146R, p.T155I, p.V272E, and p.R213Х; in intron regions: IVS7 + 31G > С, IVS9 + 12Т > С, and IVS8 + 10С > А; see [fig_ref] Table 1: Clinical features of DLBCL patients at the time of diagnosis [/fig_ref]. Evaluation of biological significance of all the revealed missense mutations of TP53 in our patients with DLBCL yielded the following results [fig_ref] Table 3: The results of functional analysis of missense mutations of ТР53 [/fig_ref] : All three prognostic software tools (PolyРhen-2, SIFT and Mut_ass) classified mutations p.L130F, p.T155I, p.R196Q, p.G244S, p.V272E, and p.A276V (which lead to appearance of a functionally inactive protein) as dangerous, probably pathogenic substitutions, or substitutions with a high/moderate degree of danger. In contrast, mutations р.W146R and p.G293R (slightly decreasing the activity of p53) and mutation p.R156C (hyperactive p53) were regarded as non-pathogenic, neutral substitutions or substitutions with a low/moderate degree of danger. Undoubtedly, mutations p.R213Х and p.A189Pfs have biological significance because each causes emergence of a nonfunctional truncated protein. It is more complicated to evaluate the effect of sense mutations detected in our sample of patients because such mutations are considered synonymous substitutions, i.e., keeping the sense of a codon. According to the prognosis of TP53 Mutant Assessor, among the sense mutations, substitution р.A307A is noteworthy because codon 307 is near the end of an exon and potentially may be located in a splicing site of an RNA molecule [bib_ref] The TP53 website: an integrative resource Centre for the TP53 mutation database..., Leroy [/bib_ref]. The functional effects of the majority of intronic mutations revealed in our group of patients with DLBCL are unknown. One of the biologically significant mutations of the TP53 gene (IVS6-36G > C) is located in intron 6 of the gene. According to the TP53 UMD mutation database, in human cancer, this mutation means changes that influence splicing [bib_ref] Datadriven unbiased curation of the TP53 tumor suppressor gene mutation database and..., Edlund [/bib_ref]. An in vitro experiment indicates that this substitution in the absence of a change in gene coding sequence leads to the survival of cells after chemotherapy and inhibits apoptosis for a long period [bib_ref] Elevated frequency and functional activity of a specific germ-line p53 intron mutation..., Lehman [/bib_ref]. According to NetGene2 prognosis, in the group of patients with DLBCL in our study, among the substitutions detected within introns, mutation IVS5 + 43G > Т caused formation of an additional acceptor site of splicing, which is absent under normal conditions. It may lead to inclusion of a part of intron 5 in the sequence of mRNA, premature formation of a stop codon at position 189, and the synthesis of a truncated p53 protein lacking functional activity. Besides, IVS4-30Т > С is outstanding among intron mutations because the alternative gene promoter is placed in intron 4 of TP53. This promoter takes part in the synthesis of the delta133 isoform, which is expressed in lymphoid tissue under normal conditions [bib_ref] p53 isoforms regulate aging-and tumor-associated replicative senescence in T lymphocytes, Mondal [/bib_ref]. ## Analysis of allelic imbalance and tp53 methylation status The frequency of TP53 promoter methylation in the sample of 69 patients with DLBCL was 4 (5.8%). Promoter methylation frequency did not significantly differ in the subgroups with the mutant and wild-type gene sequence [1 (4.2%) out of 24 vs 3 (6.7%) out of 45, р = 0.5663). Detection of the loss of heterozygosity of TP53 using microsatellite marker D17S796 was performed on tumor samples from 24 patients with DLBCL, among which 13 patients had mutations, and 11 patients did not have changes in TP53 sequence. Six (25%) cases of loss of heterozygosity were revealed, among which 5 (83.3%) cases of loss of heterozygosity were detected by the sequencing method in DLBCL tissue samples with mutations in exons 5-8 and adjacent intron regions of the TP53 gene. Gene promoter methylation was uncovered in one case of the loss of heterozygosity in a DLBCL tissue sample. ## Tp53 mutation status and clinical features of dlbcl The study cohort was divided into two subgroups: with and without functional TP53 gene mutations (TP53mut and TP53wt, accordingly). The TP53mut subgroup comprised 12 patients: patients with p.L130F, p.T155I, р.A307A, p.R196Q, p.G244S, IVS6-36G > C and p.A276V mutations, two patients with p.R213Х mutation, two patients with p.V272E mutation and one patient with multiple mutations (p.T155, p.A189Pfs and IVS5 + 43G > T). The clinical features of patient's subgroups are compared and summarized in [fig_ref] Table 1: Clinical features of DLBCL patients at the time of diagnosis [/fig_ref]. The analysis demonstrate that TP53mut correlated with B-symptoms (P = 0.016), splenomegaly (P = 0.044) and bone marrow involvement (P = 0.028), as well as IPI score of > 2 (P = 0.018). DLBCL patients with TP53wt tended to had complete remission more often (P = 0.066, [fig_ref] Table 1: Clinical features of DLBCL patients at the time of diagnosis [/fig_ref] and had better overall survival (OS) (P = 0.026, [fig_ref] Figure 4: Overall Survival of DLBCL patients with TP53mut and TP53wt status [/fig_ref] compared with DLBCL patients with TP53mut. The 5-year OS was 69.4% for patients with TP53wt versus 41.7% for those with TP53mut DLBCL. The median OS of DLBCL patients with TP53mut was 20 months. In contrast, median OS of DLBCL patients with TP53wt was not achieved. Univariate analysis showed, that extranodal foci, IPI score of > 2 and TP53 mutations predicted decrease OS of DLBCL patients. Multivariate analysis showed that IPI score were the only prognostic factors that independently predicted worse OS of DLBCL patients treated with R-CHOP and R-CHOP-like regiments. Patients with IPI score of > 2 had a three times hazard for OS (P = 0.005) compared with patients with IPI score of ≤2 [fig_ref] Table 4: Univariate and multivariate analysis of OS predictors of patients with DLBCLAbbreviations [/fig_ref]. Because of the small cohort of DLBCL patients with other TP53 aberrations in our study, we do not present the analysis of prognostic and predictive impact of loss of heterozygosity and methylation. # Discussion Analysis of literature data has shown that deletion of 17p13.1 leading to the loss of heterozygosity of TP53 has been registered at different frequencies in various studies: from 30.4 to 42% according to Chinese authors [bib_ref] Prognostic value of P53 aberrations in diffuse large Bcell lymphoma, Lu [/bib_ref] , from 40 to 50.4% in the Czech population [bib_ref] Prognostic impact of p53 aberrations for R-CHOP-treated patients with diffuse large B-cell..., Stefancikova [/bib_ref] [bib_ref] Detailed mapping of chromosome 17p deletions reveals HIC1 as a novel tumor..., Stocklein [/bib_ref] , 30.4% in the Arab population [bib_ref] The p53 mutation/deletion profle in a small cohort of the Omani population..., Tamimi [/bib_ref] , and 22.5% in the Austrian population [bib_ref] Diffuse large B-cell lymphomas with plasmablastic/plasmacytoid features are associated with TP53 deletions..., Simonitsch-Klupp [/bib_ref]. The lowest frequency of 17p13.1 deletion (22.2%) is reported in the International DLBCL Rituximab-CHOP Consortium Program Study [bib_ref] Mutational profile and prognostic significance of TP53 in diffuse large B-cell lymphoma..., Xu-Monette [/bib_ref] , combining the samples of patients from 16 hematological centers in the USA, Switzerland, Holland, Germany, Italy, and Spain. The most actively studied topic on TP53 gene variability in DLBCL is the analysis of its coding sequences revealing the presence of mutations. It has been shown that mutation frequency in the TP53 gene in DLBCL is 20% or higher [bib_ref] Structural profiles of TP53 gene mutations predict clinical outcome in diffuse large..., Young [/bib_ref]. In 1990, the IARC TP53 mutation database was created for documenting the mutations in this gene and contains information on more than 30,000 somatic and 700 germ-line mutations at present [bib_ref] Detailed mapping of chromosome 17p deletions reveals HIC1 as a novel tumor..., Stocklein [/bib_ref].Research in this database has revealed that in DLBCL, more than 120 mutations of TP53 have been described; 95% of them are single-nucleotide substitutions: 88% are missense mutations, 7% are nonsense mutations, and 5% are frame-shift mutations. More than 95% of mutations have been detected in exons 5-8, but mutations in exons 9-11 have not been described. The following codons are hotspots of mutations in TP53 in relation to DLBCL (in the order of decreasing frequency) [fig_ref] Figure 1: The analysis of promoter methylation status in the TP53 gene [/fig_ref] , 249, and 297 (http://p53.iarc.fr/DownloadDataset.aspx). Nonetheless, the prevalence of so-called hotspot mutations may change depending on the type of cancer and the ethnic origin of patients [bib_ref] A functional screen for germ line p53 mutations based on transcriptional activation, Frebourg [/bib_ref]. Comparative analysis of TP53 mutations indicates that their diversity and frequencies may significantly vary too, depending on the population under study [bib_ref] p53 mutation analysis in breast tumors by a DNA microarray method, Tennis [/bib_ref]. There is no information about a Russian population in the current version of the IARC TP53 mutation database. The vast majority of studies on the role of changes in the nucleotide sequence of TP53 have focused on exons 5-8. The intron regions have hardly been researched. Nevertheless, these sequences may potentially influence not only splicing of RNA but also gene expression by disturbing the processing autoregulation, normal post-transcriptional mRNA modifications, and post-translational protein modifications [bib_ref] Elevated frequency and functional activity of a specific germ-line p53 intron mutation..., Lehman [/bib_ref] [bib_ref] TP53 is frequently altered by methylation, mutation, and/or deletion in acute lymphoblastic..., Agirre [/bib_ref]. The relation between hypermethylation of the TP53 promoter and downregulation of gene transcription has been revealed in some tumors [bib_ref] Dysfunction of the TP53 tumor suppressor gene in lymphoid malignancies, Xu-Monette [/bib_ref] [bib_ref] Reduction of p53 gene expression in human primary hepatocellular carcinoma is associated..., Pogribny [/bib_ref] [bib_ref] Methylation in the p53 promoter is a supplementary route to breast carcinogenesis:..., Kang [/bib_ref]. Despite thorough exploration of this gene's methylation in cancer, this topic has been insufficiently studied in hematological cancers and hardly addressed in lymphoproliferative diseases [bib_ref] Dysfunction of the TP53 tumor suppressor gene in lymphoid malignancies, Xu-Monette [/bib_ref]. For example, methylation of the TP53 promoter is observed in one-third of patients with acute lymphoblastic leukemia [bib_ref] Methylation of CpG dinucleotides and/or CCWGG motifs at the promoter of TP53..., Garcia-Delgado [/bib_ref] and in one-fifth of patients with chronic lymphocytic leukemia [bib_ref] Dysfunction of the TP53 tumor suppressor gene in lymphoid malignancies, Xu-Monette [/bib_ref]. There is only anecdotal evidence on the frequency of methylation of the TP53 promoter in DLBCL [bib_ref] Presence of simian virus 40 DNA sequences in diffuse large B-cell lymphomas..., Amara [/bib_ref]. Comprehensive characterization of TP53 gene variability in DLBCL has not been carried out yet. The aim of this study was to comprehensively describe the frequency of promoter methylation and that of loss of heterozygosity as well as the frequency, diversity, and functional significance of mutations in coding and intron regions of the TP53 gene among patients with DLBCL in Novosibirsk, Russia. The diversity of single-nucleotide substitutions detected in tumor samples from our group of patients with DLBCL did not significantly differ from that in the IARC TP53 mutation database. Mutations in regions encoding the DNA-binding site were predominant (95%). Mutation p.G293R is the only revealed missense substitution that does not affect the functionally significant DNA-binding domain of p53. Searches in the IARC TP53 mutation database showed that all the functionally significant mutations detected in our study population had been described previously in a wide range of human cancers. Codons 196 and 213 in various cancers and codon 244 in hematological cancers in general and in DLBCL in particular are hotspots of mutations in TP53 (http://p53.iarc.fr/DownloadDataset.aspx). Moreover, Li-Fraumeni syndrome (http://p53.iarc.fr/ DownloadDataset.aspx) has been described, which is characterized by the development of cancer in different parts of the body because of a germline mutation, p.R213Х, р.G244S, p.L130F, or p.T155, similar to the mutations described in our study. In this study, mutations were not detected in the majority of codons [fig_ref] Figure 1: The analysis of promoter methylation status in the TP53 gene [/fig_ref] , 249, and 297, except codon 244) for which most of the mutations in TP53 have been described in DLBCL in the IARC TP53 mutation database. Codons 275, 155, 272, and 212 were found to be hotspots of mutations in the analyzed sample of patients. Review of the published literature [bib_ref] Lack of correlation between p53-dependent transcriptional activity and the ability to induce..., Kakudo [/bib_ref] [bib_ref] Characterization of the p53 mutants ability to inhibit p73 beta transactivation using..., Monti [/bib_ref] [bib_ref] Inactive fullleng thе p53 mutants lacking dominant wild-type p53 inhibition highlight loss..., Dearth [/bib_ref] [bib_ref] Аssessment of the transcriptional activity of p53 improves the prediction of recurrence..., Dekairelle [/bib_ref] about the consequences of mutations in the TP53 gene showed that each of these mutations may have multidirectional effects on different functions of p53. These effects may be tentatively subdivided into the consequences for protein structure, its biochemical properties, and biological activity and, in some cases, lead to the emergence of new functions of p53, absent in the wild-type protein. Thus, each of the mutant forms of the protein is a unique product of a mutation and may combine an increase and/or decrease in a particular type of p53 activity and may alter its structure or generate new properties. The emergence of a functionally inactive protein p53 in the analyzed group of patients with DLBCL was caused by one of missense mutations-p.L130F, p.T155I, p.R196Q, p.G244S, p.V272E, or p.A276V-together with mutation p.A189Pfs, leading to frameshift mutations, nonsense substitution p.R213Х, or splicing mutation IVS6-36G > C. All the above events in the coding part of the genethat affect the sequence carrying information about highly conserved sites of the DNA-binding domain of p53-were fixed during evolution (in phylogeny) and occur in most isoforms of p53 and in the structure of homologous proteins p63 and p73. Mutations p.R213Х and р.G244S as well as p.V272E have already been described in DLBCL [bib_ref] Mutations in the DNA-binding codons of TP53, which are associated with decreased..., Young [/bib_ref] [bib_ref] Loss of the p53 tumor suppressor activity is associated with negative prognosis..., Stefancikova [/bib_ref] , and р.T155I (detected in our study) has been previously detected in tumor samples from some patients with chronic lymphocytic leukemia and is known to be associated with poor prognosis and a weak response to treatment [bib_ref] TP53 mutation and survival in chronic lymphocytic leukemia, Zenz [/bib_ref]. Among cases of hematological cancers, mutation p.V272E has been described in Burkitt's lymphoma, Hodgkin's lymphoma, and B-cell non-Hodgkin's lymphoma (http://p53.iarc.fr/DownloadDataset.aspx). All these findings are suggestive of selection of p.L130F, p.T155I, p.R196Q, p.G244S, p.V272E, p.A276V, p.R213Х, and p.A189Pfs at various stages of cancer progression [bib_ref] Lack of correlation between p53-dependent transcriptional activity and the ability to induce..., Kakudo [/bib_ref] [bib_ref] Characterization of the p53 mutants ability to inhibit p73 beta transactivation using..., Monti [/bib_ref] [bib_ref] Inactive fullleng thе p53 mutants lacking dominant wild-type p53 inhibition highlight loss..., Dearth [/bib_ref] [bib_ref] Isolation of temperature-sensitive p53 mutations from a comprehensive missense mutation library, Shiraishi [/bib_ref] [bib_ref] Tumour p53 mutations exhibit promoter selective dominance over wild type p53, Monti [/bib_ref] [bib_ref] Р53 mutants can often transactivate promoters containin gap21 but not Bax or..., Campomenosi [/bib_ref] ; therefore, their detection in tumor samples from patients with DLBCL is not coincidental. Our findings indicate that only two missense substitutions (p.W146R and p.G293R)-among all the revealed cases in the analyzed group of patients with DLBCLdid not significantly influence the function of p53. In contrast, p.R156C led to the appearance of a hyperactive mutant protein. Out of all the mutations detected in our study, only two may influence splicing of RNA. These include samesense substitutions р.A307A and IVS6-36G > C. The functional significance of IVS6-36G > C has been proved in an in vitro experiment [bib_ref] Elevated frequency and functional activity of a specific germ-line p53 intron mutation..., Lehman [/bib_ref]. According to the prediction of TP53 Mutant assessor (release 1.00, 2012), р.A307A is also located at a splicing site of RNA [bib_ref] The TP53 website: an integrative resource Centre for the TP53 mutation database..., Leroy [/bib_ref]. Even though in sense mutations, the new codon continues encoding the same amino acid, it is believed that this type of mutations may change splicing, transcription, and/or stability of RNA [bib_ref] Datadriven unbiased curation of the TP53 tumor suppressor gene mutation database and..., Edlund [/bib_ref]. The functional effects of most intron and sense mutations detected in our group of patients with DLBCL (but not discussed) remain unknown. It is likely that intron mutations may influence not only RNA splicing but also gene expression control, by causing gain-or loss-of-function regulatory elements in TP53 region, thereby creating or disrupting binding sites for certain DNA-sequence-specific transcription factors that interfere with normal activation or autoregulation of TP53, and possibly transcriptional regulation of other potential downstream genes [bib_ref] Biology and clinical implications of the 19q13 aggressive prostate cancer susceptibility locus, Gao [/bib_ref]. Furthermore, intron mutation IVS4-30Т > С is noteworthy because an alternative gene promoter is placed in intron 4 of TP53 and takes part in the synthesis of delta133 and delta160 isoforms of p53 [bib_ref] p53 isoforms regulate aging-and tumor-associated replicative senescence in T lymphocytes, Mondal [/bib_ref]. Thus, whether these newly identified intron mutations of TP53 are functional warrants further investigation. Some published reports about low frequencies of TP53 promoter methylation in DLBCL were verified in our group of patients. The frequency of TP53 promoter methylation in the analyzed group of patients with DLBCL was 5.8% and did not differ significantly from subgroups with mutant (4.2%) and wild-type (6.7%) gene structure or from the findings of K. Amara and coauthors (3.7%) [bib_ref] Presence of simian virus 40 DNA sequences in diffuse large B-cell lymphomas..., Amara [/bib_ref]. Analysis of the loss of heterozygosity of TP53 was performed on 24 tumor samples of DLBCL in our study population, and 13 of them were found to have mutations, whereas 11 did not have changes in the sequences of TP53. Our investigation of microsatellite marker D17S796, which is located near TP53, detected 6 (25%) cases of loss of heterozygosity, in agreement with the literature data [bib_ref] Mutational profile and prognostic significance of TP53 in diffuse large B-cell lymphoma..., Xu-Monette [/bib_ref]. Furthermore, loss of heterozygosity was observed only in the subgroup of patients with a alteration of the TP53 gene (mutations were revealed in 5 tumor samples, and promoter methylation in one tumor), which accounted for 6 (42.9%) out of 14 versus 0% in the group of 10 patients with the intact gene (р = 0.0223). Our comprehensive assessment of gene variability indicates that the dysfunction of p53 in DLBCL may emerge via a two-hit mechanism. According to this model, two sequential events are necessary for transformation of a normal B cell into a cancerous cell during carcinogenesis of at least some cases of DLBCL. The first event is a mutation or methylation in the TP53 promoter, giving rise to a cell with increased risk of malignant transformation. The second event is the loss of an intact allele of TP53 in the cell; this change is necessary for tumorigenesis. Thus, in DLBCL, our results proved the selection of functionally significant mutations in the gene regions encoding the DNA-binding domain of p53. In the analyzed sample of patients with DLBCL, the location of hotspots of mutations (in contrast to the set of single-nucleotide substitutions) is different from the data listed in the IARC TP53 Mutation database. The presence of important intronic and samesense substitutions was demonstrated and confirmed the importance of studying the noncoding gene regions adjacent to exons and of bioinformatic analysis of the uncovered synonymous substitutions. In total, our results show that TP53 mutation status is a prognostic factor that stratifies DLBCL patients treated with R-CHOP and R-CHOP-like regimens. This observation is a further supports the crucial role of p53 in death of tumor cells and tumor suppression. In Russian cohort of de novo DLBCL patients, we show that TP53mut are correlated with B-symptoms, splenomegaly and bone marrow involvement, as well as adverse IPI prognostic groups. Nonetheless, DLBCL patients with TP53wt tended to had complete remission more often (P = 0.066). It is known, that despite the addition of rituximab to therapy, TP53 mutation is an independent prognostic factor that predicts poor survival in patients with DLBCL [bib_ref] Structural profiles of TP53 gene mutations predict clinical outcome in diffuse large..., Young [/bib_ref]. To eliminate the possible impact of TP53mut on OS, we also performed survival analysis for patients study cohort and found that patients with TP53wt had significantly better OS (P = 0.026). These data are in agreement with a recent study [bib_ref] Mutational profile and prognostic significance of TP53 in diffuse large B-cell lymphoma..., Xu-Monette [/bib_ref]. Our results show that together with extranodal foci and IPI score of > 2 TP53mut predicted decrease OS of DLBCL patients. In fact, TP53 mutation status is an independent prognostic factor in patients with DLBCL treated with R-CHOP [bib_ref] Structural profiles of TP53 gene mutations predict clinical outcome in diffuse large..., Young [/bib_ref]. But multivariate analysis show that the impact on survival DLBCL patients of IPI score of > 2 is more pronounced than the impact of TP53mut. This might be because of the small cohort of DLBCL patients with TP53mut in our study. # Conclusions In conclusion, in the present study, we identified TP53 mutation patterns in a Russian cohort of patients with de novo DLBCL who were treated with R-CHOP and R-CHOP-like regimens and confirmed that TP53 mutation status is a valuable prognostic biomarker. Therefore, therapeutic approaches targeting the inactivated TP53 pathway may further improve clinical outcomes of patients with DLBCL. Because of the small cohort of DLBCL patients with other TP53 aberrations in our study we do not present the analysis of prognostic and predictive impact of LOH and methylation. However, the comprehensive analysis of TP53 status gives us better insights into the possible mechanisms behind participation of this gene's variability in the pathogenesis of DLBCL. It was shown here that dysfunction of p53 in DLBCL may emerge according to the two-hit principle. [fig] Figure 1: The analysis of promoter methylation status in the TP53 gene: the results of methylation-specific PCR with flanking primers; electrophoresis in an 8% polyacrylamide gel. M: PCR with primers specific to a methylated allele (length of the product: 166 bp); NМ: PCR with primers specific to an unmethylated allele (length of the product: 170 bp), 1-5: the ID numbers of cases; 1, 2, and 4: normal; 3 and 5: methylation; K: control DNAs results did not significantly differ from the data in the IARC TP53 mutation database(Fig. 2). [/fig] [fig] Figure 2: The analysis of microsatellite instability in locus D17S796: the results of PCR with flanking primers, electrophoresis in an 8% polyacrylamide gel (length of the product: 144-174 bp); B: DNA from blood, T: DNA of a tumor tissue, 1-3: the ID numbers of cases; 1, 3: loss of heterozygosity; 2: normalFig. 3 The sequenced fragments of DNA containing mutations of TP53. a: homozygote of p.L130F, b: a heterozygote of p.W146R, c1: a homozygote of p.T155I, c2: a heterozygote of p.T155I, d: a homozygote of p.R156C, e: a homozygote of р.V157 V, f: a homozygote of р.H179H, g: a homozygote of p.A189Pfs, h: a homozygote of p.R196Q, i: a heterozygote of p.R213Х, j: a heterozygote of p.G244S, k: a heterozygote of p.L252 L, l: a heterozygote of p.V272E, m: a heterozygote of р.V272 V, n: a heterozygote of p.A276V; o: a homozygote of p.G293R, p: a heterozygote of p.G302G, q: a heterozygote of р.A307A, r: a homozygote of IVS4-30Т > С, s: a homozygote of IVS5 + 43G > T, t: a heterozygote of IVS5-17Т > С, u: a heterozygote of IVS6-36G > C, v: a heterozygote of IVS7 + 31G > С, w: a heterozygote of IVS8 + 10С > А, x: a heterozygote of IVS8 + 20A > G, y: a heterozygote of IVS8 + 37A > G, and z: a heterozygote of IVS9 + 12Т > С [/fig] [fig] Figure 4: Overall Survival of DLBCL patients with TP53mut and TP53wt status [/fig] [table] Table 1: Clinical features of DLBCL patients at the time of diagnosis (n = 74) [/table] [table] Table 2: General characteristics of sequencing results [/table] [table] Table 3: The results of functional analysis of missense mutations of ТР53 [/table] [table] Table 4: Univariate and multivariate analysis of OS predictors of patients with DLBCLAbbreviations: S-LDH serum lactate dehydrogenase, IPI International Prognostic Index, HR hazard ratio, CI confidence interval [/table]

Closed reduction and percutaneus Kirschner wire fixation for the treatment of dislocated calcaneal fractures: surgical technique, complications, clinical and radiological results after 2–10 years Introduction To reduce complications, a minimally invasive technique for the treatment of dislocated intraarticular fractures of the calcaneus was used. Therefore previously described closed reduction and internal Wxation techniques were combined and modiWed. Materials and methods Sixty-seven out of 92 calcaneal fractures could be retrospectively evaluated with an average follow-up time of 5.7 years (minimum 2-10 years follow-up). For radiographic evaluation, plain radiographs and CT scans were obtained. The Zwipp score was used for clinical evaluation. Sanders type II, III and IV fractures were diagnosed. Results Length of surgery averaged 61 min (range 20-175 min). The incidence of subtalar arthritis was correlated to the severity of fracture. Böhler's angle was restored in 70.1% (47 of 67) of the cases. On the last follow-up evaluation the average Zwipp score was 130 points (range 48-186 points). The majority (77.7%) of patients were content with their treatment result. The rate of signiWcant complications was 6.5%.Discussion Compared to open techniques the presented minimally invasive technique showed comparable results with a low rate of serious complications and is a viable alternative for the treatment of intraarticular, dislocated calcaneal fractures. # Introduction A fracture of the calcaneus allowed to heal in improper anatomical position leads to static and dynamic malfunctions of the whole foot with consequent limited load bearing capacity and walking ability [bib_ref] Injuries of the talus and calcaneus, Brunner [/bib_ref] [bib_ref] Intra-articular fracture of the calcaneus. ClassiWcation, assessment and surgical procedures, Zwipp [/bib_ref]. The associated pain leads to a signiWcant impairment in quality of life. The goal of therapy for calcaneal fractures is the elimination of pain and restoration of walking ability for patients with normal foot shape and the ability to wear normal footwear. At present, there are multiple operative procedures for the treatment of calcaneal fractures. The procedures can be diVerentiated by approach, implant type and whether the treatment is one-or two-stages. Recently, open procedures using internal Wxation have been favored for surgical therapy of the calcaneus [bib_ref] Therapy of dislocated calcaneus joint fracture with the AO calcaneus plate, Boack [/bib_ref] [bib_ref] The surgical treatment of calcaneus fractures, Brunner [/bib_ref] [bib_ref] Early wound complications of operative treatment of calcaneus fractures: analysis of 190..., Folk [/bib_ref] [bib_ref] Calcaneus fractures. Open reduction and internal Wxation, Rammelt [/bib_ref] [bib_ref] Operative treatment in 120 displaced intraarticular calcaneal fractures, Sanders [/bib_ref] [bib_ref] Treatment of intraarticular calcaneal fractures in adults. A treatment algorithm, Thermann [/bib_ref] [bib_ref] Open reduction and internal Wxation of calcaneal fractures with a low proWle..., Thordarson [/bib_ref]. A possible complication of an open procedure is the disturbance of wound healing with skin and soft tissue necrosis [bib_ref] Therapy of dislocated calcaneus joint fracture with the AO calcaneus plate, Boack [/bib_ref] and the possibility of cutaneous Xaps [bib_ref] The surgical treatment of calcaneus fractures, Brunner [/bib_ref] [bib_ref] Management of soft-tissue complications of the lateral approach for calcaneal fractures, Cavadas [/bib_ref] [bib_ref] Early wound complications of operative treatment of calcaneus fractures: analysis of 190..., Folk [/bib_ref] [bib_ref] Calcaneus fractures. Open reduction and internal Wxation, Rammelt [/bib_ref] [bib_ref] Operative treatment in 120 displaced intraarticular calcaneal fractures, Sanders [/bib_ref]. In addition to posttraumatic arthritis in the lower ankle and adjoining joints, there are reports of osteitis of the calcaneus [bib_ref] Therapy of dislocated calcaneus joint fracture with the AO calcaneus plate, Boack [/bib_ref] [bib_ref] The surgical treatment of calcaneus fractures, Brunner [/bib_ref] [bib_ref] Operative treatment in 120 displaced intraarticular calcaneal fractures, Sanders [/bib_ref]. Advanced osteitis of the calcaneus can require a calcanectomy [bib_ref] Calcaneus fractures. Open reduction and internal Wxation, Rammelt [/bib_ref] or amputation of the lower leg [bib_ref] Early wound complications of operative treatment of calcaneus fractures: analysis of 190..., Folk [/bib_ref] [bib_ref] Operative treatment in 120 displaced intraarticular calcaneal fractures, Sanders [/bib_ref]. In an eVort to reduce the complications that can occur with an open procedure, we have combined and modiWed previously described closed reduction and internal Wxation techniques [bib_ref] Treatment of severe fractures of the calcaneus by reposition and percutaneous wire..., Poigenfurst [/bib_ref] [bib_ref] A new treatment method for the calcaneus fracture, Westhues [/bib_ref] [bib_ref] Treatment of petrous bone fractures by internal Wxation, Wondrak [/bib_ref] to create a minimally invasive technique. The aim of this study was to evaluate the clinical and radiographic results of our minimally invasive surgical treatment of intraarticular fractures of the calcaneus at 2-10 years postoperatively. We then compared our results to the results of both open and minimally invasive surgical techniques found in the literature. # Material and methods ## Patients A total of 88 patients with 92 closed, dislocated and intraarticular fractures of the calcaneus were consecutively treated with a minimally-invasive technique developed at our institution by modifying and combining the procedures of Westhues [bib_ref] A new treatment method for the calcaneus fracture, Westhues [/bib_ref] , Poigenfürst and Buch [bib_ref] Treatment of severe fractures of the calcaneus by reposition and percutaneous wire..., Poigenfurst [/bib_ref] , and Wondrák [bib_ref] Treatment of petrous bone fractures by internal Wxation, Wondrak [/bib_ref]. All the surgeries were performed without the use of bone graft. The average age at time of calcaneal fracture was 46.1 years (range 18-82 years) and most patients were male (71.6%). The cause of fracture was a fall from varying heights in 75 (85.2%) of cases and a motor vehicle accident in 13 (14.8%) of cases. From this group of patients, 63 (71.6%) patients with 67 fractures were available for retrospective examination with an average follow-up time of 5.7 years (range 2-10 years). Twenty-Wve patients were unavailable for followup examination; 12 patients who satisWed with the result of the surgery did not wish to participate in the study due to age or unwillingness to travel to the hospital, 7 patients were at an unknown address and 6 patients deceased. ## Examination All the medical records, radiographs, pre-and post-operative computed tomography scans were available for the entire study group. Traumatic soft tissue damage was determined using the classiWcation method of Tscherne and Oestern [bib_ref] A new classiWcation of soft-tissue damage in open and closed fractures, Tscherne [/bib_ref]. In addition to radiographic evaluation using the calcaneus lateral, calcaneus axial and Broden views at 20° and 40° [bib_ref] Roentgen examination of the subtaloid joint in fractures of the calcaneus, Brodèn [/bib_ref] , all the patients obtained a bilateral calcaneal preoperative CT scan for fracture classiWcation and surgical planning. Fractures were evaluated by the classiWcation scale of Sanders et al. [bib_ref] Operative treatment in 120 displaced intraarticular calcaneal fractures, Sanders [/bib_ref] [bib_ref] Displaced intra-articular fractures of the calcaneus, Sanders [/bib_ref]. The duration of surgery, as well as the time the X-ray image intensiWer used, were noted as operative data. For postoperative evaluation of the reduction, radiographs of the calcaneus in lateral, axial, Broden 20° and 40° [bib_ref] Roentgen examination of the subtaloid joint in fractures of the calcaneus, Brodèn [/bib_ref] views were obtained. At the last follow-up evaluation, a CT scan of both calcanei was obtained to examine the geometry of the calcaneus and to evaluate arthritic changes in the lower ankle joint. For a reconstruction of the calcaneal posterior facet to be considered satisfactory a joint deviation of ·2 mm had to be seen radiographically on the postoperative radiographs. For the evaluation of the clinical results, the Zwipp score was used. # Statistical analysis For statistical analysis the Fisher's exact test and the Chisquare test were used. Perioperative treatment and surgical technique After a calcaneal fracture was diagnosed, the lower leg was evaluated. If the soft tissue was in good condition, primary surgery performed. In cases of severe swelling with potential soft tissue damage NSAIDs, local cryotherapy and active movement exercises determined the course of therapy. Once the soft tissue was in good condition, surgery was performed. The patient was placed prone on the traction table under general anaesthesia or spinal anaesthesia without arrest of blood supply. After application of the calcaneus wire for traction in the dorso-cranial plane, the varus or valgus malaligment of the back foot was corrected with the wire positioned orthogonally to the longitudinal axis of the calcaneus. To achieve correct placement, the surgeon pulled at the traction bow with the knee bent along the longitudinal axis of the calcaneus. The subsequent traction in the plantar plane was performed with the leg stretched. Once placed in the proper orientation, the traction bow was connected to the retaining jig of the traction table (traction of 20 kg). The surgeon then held the heel of the patient with both hands and applied compression both medially and laterally with his thenar muscle to reduce the main medial and lateral fragments. With the use of the X-ray image intensiWer, the restoration of Böhler's angle and the reduction of the posterior facet were veriWed. A Steinmann pin was put with the pointed end into a universal chuck and the other end into the dorso-lateral calcaneus beneath the posterior facet with a stab incision . For reduction of the posterior facet, the traction wire was used as a hypomochlion . The anatomical restoration of the posterior facet was veriWed with use of the Brodén radiographic views (20° and 40°). The reduction of the calcaneus was veriWed by examining the lateral radiographic view [fig_ref] Figure 2: Intraoperative lateral radiograph of a dislocated calcaneal fracture before reduction [/fig_ref]. The fragments were Wxed with percutaneus Kirschner wires, which were arranged conically into the talus and/or cuboid [fig_ref] Figure 2: Intraoperative lateral radiograph of a dislocated calcaneal fracture before reduction [/fig_ref]. All of the main fragments to be reduced were held in place by the Kirschner wires and the arch of foot was held until the bone healed [bib_ref] Treatment of petrous bone fractures by internal Wxation, Wondrak [/bib_ref]. The wires were bent above the skin level, the wire traction removed and the stab incision closed with sutures. Postoperatively, a dorsal lower leg splint was applied for immobilization and the lower portion of the operative side was elevated. On the Wrst postoperative day, the patient was mobilized on crutches with no weight bearing on the operative side. Depending on pain levels, the patient began with active dorsal extension and plantar Xexion in the upper ankle joint. The radiographic controls, (upper ankle joint anterior and lateral, calcaneus axial Brodén views [bib_ref] A new classiWcation of soft-tissue damage in open and closed fractures, Tscherne [/bib_ref] were performed immediately postoperative, and again 2 and 8 weeks postoperative. After the fracture was healed, the Kirschner wires were removed. Partial weight bearing began after the eighth postoperative week at 20 kg and was increased up to full weight bearing by the twelfth postoperative week [fig_ref] Figure 3: Clinical example [/fig_ref]. # Results Fifty-Wve of the 88 (62.5%) patients suVered exclusively from calcanal fractures. Thirty-three of the 88 (37.5%) patients had multiple injuries. Twenty of the 33 (61%) patients with multiple injuries had local co-injuries such as fractures of the upper ankle joint, tarsus and front-foot. In 83 of the 92 calcaneal fractures, the soft tissue injury was graded as 1° or 2°; in 9 fractures, the soft tissue injury was graded as 3°. In eight cases, including three patients with multiple injuries, surgery was performed immediately on the day of the trauma. Length of surgery averaged 61 min (range 20-175 min) and screening time averaged 115 s (range 20-454 s). To obtain proper retention of the fracture, between 4 and 7 Kirschner wires were used. The Kirschner wires were removed with or without local anaesthesia at an average of 10 weeks (range 7-15 weeks) as an outpatient procedure. Full weight bearing was achieved at an average of 15 weeks postoperative. ## Sanders classiwcation Sanders type II, III and IV fractures were diagnosed [fig_ref] Table 1: Patients with immediate postoperative reconstruction of the posterior facet by Sanders fracture... [/fig_ref]. All the patients had a joint deviation of at least 2 mm. Patients with type I non-dislocated fractures were treated conservatively and were therefore not included in this study. Reconstruction of the posterior facet was veri-Wed radiographically in [bib_ref] Calcaneal fractures: indirect reduction and external Wxation, Mcgarvey [/bib_ref]. Regardless of fracture type, The Steinmann pin is placed into the dorso-lateral calcaneus above the traction wire (left side). For the reduction of the posterior facet, the traction wire is used as a hypomochlion (right side). The hematome is drained though the stab incision. Thirty (44.8%) patients considered their treatment outcome very good or good, 28 (41.8%) as satisfactory and suYcient and 9 (13.4%) as insuYcient. Thirty-seven (55.3%) of the 67 patients had no pain while full weight bearing or could walk at least 4 h without pain at last follow-up evaluation. Nine (13.4%) patients had constant pain. Thirty-three (49.3%) patients had a range of motion in the upper ankle joint identical to that of the non-aVected side. Thirty-four (50.7%) patients had a restricted range of motion of up to 15° and more than half (58.2%) of the patients had achieved more than 75% of their total range of motion in the lower ankle joint. Forty-three (64.2%) of 67 patients were able to wear normal shoes while 5 (7.4%) used shoes with an unroll aid. Nineteen (28.3%) patients had obtained orthopaedic shoes on their own at last follow-up. Two (3%) of 67 patients had arthritis of the upper ankle joint at last follow-up. Independent of fracture type 33 (50.2%) patients had arthritic changes in the lower ankle joint and 14 (20.89%) cases had arthritic changes in the calcaneocuboidal joint. Böhler's angle was restored in 70.1% (47 of 67) of the cases. A total of 85.1% (57 of 67) cases had a reduction in calcaneal height to within 10% when compared to the non-operative side. Additionally, a reduction in the length of the calcaneus occurred within 10 in 94% (63 of 67) of cases when compared to the non-operative side. Thirty-eight (56.8%) of the 67 patients had a widening of the calcaneus to within 10% compared to the opposite side. ## Complications Of the 92 surgically treated calcaneus fractures, 76 (82.6%) healed without complications. In nine (9.8%) cases, superWcial skin infections, perforations of the Kirschner wires and bone dystrophy occurred and healed without any further complications. SigniWcant complications occurred in six (6.5%) cases: three cases had osteitis of the calcaneus, one case had dislocation of the fracture requiring revision surgery and two cases had peroneal tendon impingement. Three cases of osteitis healed through conservative therapy with oral antibiotics and the two patients with impingement of the peroneal tendon refused any operative intervention. Using our method, disturbance of wound healing with skin and soft tissue necrosis requiring operative intervention was not observed. Additionally, no lower leg amputations and no total or partial calcanectomies had to be performed. # Statistical analysis Using the Chi-square test a statistical signiWcant correlation was found between the fracture type and the incidence of subtalar arthritis (p < 0.0268). Using the same test there was no statistical signiWcant correlation found between the fracture type and quality of reduction (p > 0.6522), between fracture type and clinical result (p > 0.3204) and between fracture type and Böhler's angle (p > 0.5488). However, a statistical signiWcant correlation was found between the clinical result and subtalar arthritis (p < 0.0013) and between the incidence of subtalar arthritis and quality of reduction when using the Fisher`s exact test (p < 0.0003). # Discussion To correct deformations of the calcaneus, spare the soft tissue and lower the complication rate, indirect and less invasive reduction and Wxation techniques to treat calcaneal fractures have been developed . Besides techniques with percutaneous reduction and internal K-wire Wxation [bib_ref] Conservative treatment of calcaneus fracture versus repositioning and percutaneous bore wire Wxation...., Buch [/bib_ref] [bib_ref] Treatment of severe fractures of the calcaneus by reposition and percutaneous wire..., Poigenfurst [/bib_ref] [bib_ref] Clinical and radiologic late results following internal wire osteosyntheses of the calcaneus...., Russe [/bib_ref] [bib_ref] Minimally-invasive treatment of intra-articular fractures of the calcaneum, Stulik [/bib_ref] percutaneous reduction techniques with external Wxation are described in the literature [bib_ref] External Wxation for displaced intra-articular fractures of the calcaneum, Magnan [/bib_ref] [bib_ref] Calcaneal fractures: indirect reduction and external Wxation, Mcgarvey [/bib_ref] [bib_ref] Percutaneous treatment of displaced intra-articular calcaneal fractures, Schepers [/bib_ref] [bib_ref] Management of displaced intraarticular calcaneal fractures by using external ring Wxation, minimally..., Talarico [/bib_ref]. In our department we have combined and modiWed previously described closed reduction and internal Wxation techniques [bib_ref] Treatment of severe fractures of the calcaneus by reposition and percutaneous wire..., Poigenfurst [/bib_ref] [bib_ref] A new treatment method for the calcaneus fracture, Westhues [/bib_ref] [bib_ref] Treatment of petrous bone fractures by internal Wxation, Wondrak [/bib_ref] to create a minimally invasive technique. This method used the lower joint surface of the talus as a guide for remodeling the posterior facet and reconstruction of the calcaneus. A multiple of diVerent evaluation scores makes a comparison of the clinical results with published results diYcult [bib_ref] Clinical rating systems for the ankle, hindfoot, midfood, hallux and lesser toes, Kiotka [/bib_ref] [bib_ref] Operative treatment in 120 displaced intraarticular calcaneal fractures, Sanders [/bib_ref] [bib_ref] DiYculties in evaluating follow-up outcome in calcaneus fracture managed with plate osteosynthesis, Schuh [/bib_ref]. Using the Zwipp score, 61.3% of the cases (41 of 67) treated by our modiWed method had good or very good results regardless of fracture type. In 38.7% (26 of 67) of cases, a satisfactory or bad result was achieved. After open reduction and internal Wxation with a plate, Rammelt et al. [bib_ref] Calcaneus fractures. Open reduction and internal Wxation, Rammelt [/bib_ref] and Boack et al. [bib_ref] Therapy of dislocated calcaneus joint fracture with the AO calcaneus plate, Boack [/bib_ref] reported good and very good results in 65 and 66% of cases, respectively with the use of the §200 Zwipp score. Boack et al. [bib_ref] Therapy of dislocated calcaneus joint fracture with the AO calcaneus plate, Boack [/bib_ref] presented satisfactory or bad clinical results in 34% of cases also using the Zwipp score. Sanders et al. [bib_ref] Displaced intra-articular fractures of the calcaneus, Sanders [/bib_ref] showed that under application of a lateral approach followed by internal Wxation with a plate, the worst clinical results occurred with type IV fractures. In this study the worst clinical results were achieved in the treatment of type IV fractures although within our cases no statistical signiWcant correlation between fracture type and clinical result could be demonstrated. Patient satisfaction is an essential criterion for the successful treatment of calcaneus fractures. With a comparable follow-up period of 5.4 years, Thermann et al. [bib_ref] Treatment of intraarticular calcaneal fractures in adults. A treatment algorithm, Thermann [/bib_ref] reported that 48.3% of patients viewed their treatment outcome as good or very good after open reduction of their calcaneus fractures, 37.3% of patients judged their treatment outcome as satisfactory and suYcient and 14.4% judged their outcome as insuYcient. Our treatment method had comparable results with 44.8% of patients considering their treatment outcome very good or good, 41.8% as satisfactory and suYcient and 13.4% as insuYcient. At last follow-up evaluation, 64.2% of patients in this study were able to wear normal shoes while 28.4% required orthopaedic shoes. Comparably, in the study by Therman et al. [bib_ref] Treatment of intraarticular calcaneal fractures in adults. A treatment algorithm, Thermann [/bib_ref] , 68.7% of the patients wore normal shoes and 16.8% required orthopaedic shoes. Overall, satisfactory mobility of the joints adjoining the calcaneus was achieved with our surgical technique. In approximately half of our patients, complete range of motion in the upper ankle joint was achieved and in the lower ankle joint, approximately 60% of patients had no or low movement restriction. Similar mobility in the upper and lower ankle joints has been reported in the literature after open reduction and osteosynthesis with a plate [bib_ref] The surgical treatment of calcaneus fractures, Brunner [/bib_ref] [bib_ref] Treatment of intraarticular calcaneal fractures in adults. A treatment algorithm, Thermann [/bib_ref]. However, Buch [bib_ref] Wire osteosynthesis of the calcaneal fracture, Buch [/bib_ref] reported a worse range of motion in the upper and lower ankle joint after performing a percutaneus wire osteosynthesis in 100 calcaneal fractures with a varus or valgus deviation of the back foot occurring in half of the cases. In this study, malpositions of the back foot of more than 5° were diagnosed in 6 (9%) cases of the 67 cases. Clinically relevant changes in the axis of the front foot were not observed. The radiological evaluation of the joints adjoining the calcaneus showed arthritic changes in the upper ankle joint in two of our cases. Half of our cases showed arthritic changes in the lower ankle joint and in approximately a Wfth of our evaluated cases, arthritic changes were found in the calcaneocuboidal joint. In our study, 66.7% of type-II fractures and 98% of type-III/IV fractures showed arthritic changes in the lower ankle joint. These results are similar to those presented by Thermann et al. [bib_ref] Treatment of intraarticular calcaneal fractures in adults. A treatment algorithm, Thermann [/bib_ref] , who showed arthritic changes in the lower ankle joint in 65.2% of type-II fractures and in 81.7% of type III/IV fractures. Regardless of fracture type, anatomical restoration of the joint surface or a joint deviation of up to 2 mm of the posterior facet was achieved postoperatively in 82.1% of cases. A total of 86.7% of type-II fractures and 72.8% of type-III/ IV fractures achieved anatomical restoration of the posterior facet. These results are comparable to those presented in the literature with open reduction and internal Wxation with plates or screws [bib_ref] Therapy of dislocated calcaneus joint fracture with the AO calcaneus plate, Boack [/bib_ref] [bib_ref] Operative treatment in 120 displaced intraarticular calcaneal fractures, Sanders [/bib_ref] [bib_ref] Treatment of intraarticular calcaneal fractures in adults. A treatment algorithm, Thermann [/bib_ref]. In this study, the average length of the surgery was 61 min. This is considerably less than the average length of surgeries utilizing open reduction and internal Wxation of 139 and 168 min [bib_ref] Therapy of dislocated calcaneus joint fracture with the AO calcaneus plate, Boack [/bib_ref] [bib_ref] DiYculties in evaluating follow-up outcome in calcaneus fracture managed with plate osteosynthesis, Schuh [/bib_ref]. In 9.8% of cases, complications such as superWcial skin infections, perforation of the Kirschner wires and bone dystrophy healed with conservative therapy or after removal of the Kirschner wires. This rather insigniWcant rate of complications is similar to those described in the literature after operative treatment of calcaneal fractures utilizing a minimally invasive technique [bib_ref] Conservative treatment of calcaneus fracture versus repositioning and percutaneous bore wire Wxation...., Buch [/bib_ref] [bib_ref] Clinical and radiologic late results following internal wire osteosyntheses of the calcaneus...., Russe [/bib_ref] [bib_ref] Minimally-invasive treatment of intra-articular fractures of the calcaneum, Stulik [/bib_ref]. The rate of signiWcant complications in this study was 6.5%. In two cases, an impingement of the peroneal tendon occurred. One case had a dislocation of the fragments requiring revision surgery. Three cases had infections of the calcaneus and were related to a traumatic soft tissue injury that healed with conservative therapy. With open reduction, skin and soft tissue necrosis [bib_ref] Therapy of dislocated calcaneus joint fracture with the AO calcaneus plate, Boack [/bib_ref] with possible cutaneous Xaps have been reported in scattered cases [bib_ref] The surgical treatment of calcaneus fractures, Brunner [/bib_ref] [bib_ref] Management of soft-tissue complications of the lateral approach for calcaneal fractures, Cavadas [/bib_ref] [bib_ref] Early wound complications of operative treatment of calcaneus fractures: analysis of 190..., Folk [/bib_ref] [bib_ref] Calcaneus fractures. Open reduction and internal Wxation, Rammelt [/bib_ref] [bib_ref] Operative treatment in 120 displaced intraarticular calcaneal fractures, Sanders [/bib_ref]. Folk et al. [bib_ref] Early wound complications of operative treatment of calcaneus fractures: analysis of 190..., Folk [/bib_ref] demonstrated that in 25% of cases, wound complications required an additional surgery after open reduction of calcaneus fractures 84% of the time. Abidi et al. [bib_ref] Wound-healing risk factors after open reduction and internal Wxation of calcaneal fractures, Abidi [/bib_ref] reported wound healing disturbances in 33% of cases following open reduction and internal Wxation of calcaneal fractures. There have also been reports in the literature of partial or total calcanectomies [bib_ref] Calcaneus fractures. Open reduction and internal Wxation, Rammelt [/bib_ref] , including amputations, as a result of calcaneal osteitis after open reduction [bib_ref] Early wound complications of operative treatment of calcaneus fractures: analysis of 190..., Folk [/bib_ref] [bib_ref] Operative treatment in 120 displaced intraarticular calcaneal fractures, Sanders [/bib_ref]. Sanders et al. [bib_ref] Operative treatment in 120 displaced intraarticular calcaneal fractures, Sanders [/bib_ref] considered a lower leg amputation if a patient suVered continual osteitis. In the current study no disturbances of wound healing, skin or soft tissue necrosis that required microsurgical intervention were observed. In this study no statistical signiWcant correlation could be found between fracture type and the ability to restore Böhler`s angle. However, regardless of fracture type Stulik et al. [bib_ref] Minimally-invasive treatment of intra-articular fractures of the calcaneum, Stulik [/bib_ref] showed similar results using a comparable minimally invasive technique while Thermann et al. [bib_ref] Treatment of intraarticular calcaneal fractures in adults. A treatment algorithm, Thermann [/bib_ref] had worse results following open reduction with internal Wxation. In our study, there was a slight reduction in the height and length of the operated calcaneus when compared to the non-operated side. With our closed reduction and internal Wxation technique, more than half of the cases had a widening of the calcaneus of more than 10% when compared to the opposite side, which is an unsatisfactory result. At the same time only two patients had an impingement of the peroneal tendon that the patients considered tolerable. In summary, we presented a minimally invasive technique for the treatment of intraarticular, dislocated calcaneus fractures and were able to produce results comparable to open techniques with a lower rate of serious complications. In the majority of cases, an almost identical Böhler angle and geometry of the calcaneus was achieved when compared to the opposite side at the time of last follow-up. Simple removal of the Kirschner wires and shorter surgery time decrease patient stress and must be recognized as an advantage of this minimally invasive technique. Thus, we feel that our minimally invasive technique is a viable alternative for the treatment of intraarticular, dislocated calcaneal fractures. Open Access This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the original author(s) and source are credited. [fig] Figure 2: Intraoperative lateral radiograph of a dislocated calcaneal fracture before reduction (left side). After reduction the fragments are Wxed with percutaneus Kirschner wires (right side) 41 (61.2%) of the 67 patients had a very good or good result, 24 (35.8%) patients had a satisfactory result and 2 (3%) patients had a bad result. The worst clinical results occurred with type IV fractures [/fig] [fig] Figure 3: Clinical example. a Preoperative lateral radiograph of an intraarticular, displaced calcaneal fracture (left side). The preoperative CT scan shows a Sanders Type II AC fracture on the coronal reconstruction (right side). b Lateral radiograph taken 2.5 years postoperatively.The patient had a Zwipp score of 166 points (left side). Anatomical reconstruction of the posterior facet was achieved (right side). This coronal CT image shows the anatomical reconstruction of the posterior facet with no joint deviation [/fig] [table] Table 1: Patients with immediate postoperative reconstruction of the posterior facet by Sanders fracture classiWcation [/table]

Aspirin inhibits proliferation of gemcitabine-resistant human pancreatic cancer cells and augments gemcitabine-induced cytotoxicity Aim: To investigate whether aspirin is able to augment gemcitabine-induced cytotoxicity in human pancreatic cancer cells. Methods: Two gemcitabine-insensitive human pancreatic cancer cell lines, PANC-1 and Capan-1, were used. Cells were treated with either aspirin or gemcitabine alone or both of them. Cell growth and apoptosis were determined by MTT assay, Annexin V or Hoechest 33258 staining. Cell cycle distribution was examined by flow cytometry. Western blot with specific phosphorylated protein antibodies was used to detect the activation of protein kinase. RT-PCR and Western blot were applied to assess the transcription and protein level for cyclin D1 and Bcl-2. Results: Aspirin alone significantly inhibits the proliferation of PANC-1 cells by causing cell cycle arrest at G 1 phase. Aspirin potentiates the anti-survival effect of gemcitabine as well as its pro-apoptotic effect in PANC-1 cells, although aspirin per se does not trigger apoptosis. Aspirin inhibits GSK-3β activation and suppresses the expression of its downstream gene products (cyclin D1 and Bcl-2), which are implicated in proliferation, survival and chemoresistance of pancreatic cancer. The effects of aspirin on Capan-1, were similar to that on PANC-1. Conclusion: Our results suggest that aspirin inhibits the proliferation of gemcitabine-resistant pancreatic cancer cells and augments the antisurvival effect of gemcitabine, probably by suppressing the activity of GSK-3β and its downstream gene products. # Introduction Pancreatic cancer remains a fatal disease with a 5-year survival rate less than 5% [bib_ref] Epidemiology and prevention of pancreatic cancer, Lowenfels [/bib_ref]. Besides surgery and radiation, chemotherapy regimens often fail to improve the outcome of pancreatic cancer patients. Currently, gemcitabine (difluorodeoxycytidine) appears to be the best chemotherapeutic agent in the treatment of advanced pancreatic cancer [bib_ref] Role of NF-kappaB and Akt/PI3K in the resistance of pancreatic carcinoma cell..., Arlt [/bib_ref]. However, even with this drug, the objective tumor response rate is less than 10%, and the impact of the drug on survival is minor [bib_ref] Nuclear factor-kappaB and IkappaB kinase are constitutively active in human pancreatic cells,..., Li [/bib_ref]. Moreover, gemcitabine is associated with drug resistance and highly toxic for tumor cells as well as normal cells [bib_ref] Role of NF-kappaB and Akt/PI3K in the resistance of pancreatic carcinoma cell..., Arlt [/bib_ref] [bib_ref] Preclinical toxicology (subacute and acute) and efficacy of a new squalenoyl gemcitabine..., Reddy [/bib_ref]. Thus, there is a need for novel strategies involving less toxic agents that can sensitize pancreatic cancer cells to chemotherapy. The transcription factor nuclear factor-κB (NF-κB) plays a major role in promoting gemcitabine resistance [bib_ref] Role of NF-kappaB and Akt/PI3K in the resistance of pancreatic carcinoma cell..., Arlt [/bib_ref] , because NF-κB mediates transcription of a serial of proliferation and antiapoptosis genes, such as cyclin D1 and Bcl-2. Recent evidence indicates that glycogen synthase kinase-3β (GSK-3β) positively regulates NF-κB-mediated gene transcription and cell survival [bib_ref] Inhibition of glycogen synthase kinase-3 activity leads to epigenetic silencing of nuclear..., Ougolkov [/bib_ref] [bib_ref] Glycogen synthase kinase-3beta participates in nuclear factor kappaB-mediated gene transcription and cell..., Ougolkov [/bib_ref]. It is reported that pancreatic cancer cells contain a pool of active GSK-3β and that pharmacologic inhibition of GSK-3β kinase activity using small molecule inhibitors or genetic depletion of GSK-3β by RNA interference influences NF-κB-mediated gene (such as cyclin D1, Bcl-2) transcription, leading to decreased cancer cell proliferation and survival. Together, this evidence suggests GSK-3β as a potential therapeutic target in the treatment of pancreatic cancer. Therefore, agents that block GSK-3β activation could reduce chemoresistance to gemcitabine and perhaps be used in combination with gemcitabine as a novel therapeutic regimen for pancreatic cancer. Aspirin (acetylsalicylic acid, ASA), the traditional nonsteroid anti-inflammatory drug (NSAID), is one such agent that is nontoxic to humans, which has become one of the most commonly utilized therapeutic drugs all over the world since its introduction into modern medicine in 1897 [bib_ref] The role of aspirin-triggered lipoxins in the mechanism of action of aspirin, Gilroy [/bib_ref]. Dihlmann [bib_ref] Reduction of beta-catenin/ T-cell transcription factor signaling by aspirin and indomethacin is..., Dihlmann [/bib_ref] suggested that ASA was able to induce phosphorylation/ inactivation of GSK-3β in several colon cancer cell lines. Nevertheless, the effect of ASA on GSK-3β activity in pancreatic cancer cells is never, to our knowledge, investigated before. In addition, although previous studies report that ASA is capable of suppressing pancreatic cancers growth in vitro and in vivo [bib_ref] Nuclear factor kappa B activation is a potential target for preventing pancreatic..., Sclabas [/bib_ref] , the exact function and the underlying mechanism of ASA on pancreatic cancer remain to be further explored. Thus, the goal of this study was to investigate the impact of ASA on the growth of human pancreatic cancer cells. Additionally, we investigate whether ASA can potentiate the gemcitabine-induced cytotoxicity in pancreatic cancer cells in vitro. # Materials and methods Cell culture and drug treatment Human pancreatic cancer cell line PANC-1, Capan-1 were obtained from Shanghai Cell Bank and maintained in DMEM (Invitrogen, Grand Island, NY) supplemented with 10% fetal bovine serum (FBS, Invitrogen), 100×10 3 U/L penicillin, and 100×10 3 U/L streptomycin in 5% CO 2 at 37 °C. Human pancreatic cancer cell line PANC-1, Capan-1 were chosen as in vitro model, because they are considered relatively resistant to many chemotherapeutic regimens [bib_ref] K-ras oncogene silencing strategy reduces tumor growth and enhances gemcitabine chemotherapy efficacy..., Rejiba [/bib_ref]. ASA (Sigma, St Louis, Mo) was dissolved in DMSO (Sigma) and diluted with DMEM medium to a final concentration of 1% DMSO. The pH value of the ASA-containing medium was adjusted to 7.2 with 2.8% NaHCO 3 (Shanghai Sangon Biological Engineering Technology & Services Co, Ltd, China). Vehicle was treated with an equivalent volume of 10% FBS-medium with 1% DMSO. Gemcitabine (difluorodeoxycytidine, Lilly, Bad Homburg, Germany) was stored at 4 °C and dissolved in PBS on the day of use. Cell growth and survival assays All assays were carried out in quintuple of three separate experiments. Cell growth was tested by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide (MTT, Sigma) assay. Apoptosis was evaluated with Annexin V-fluoroisothiocyanate apoptosis detection kit according to the instruction of the manufacturer (Sigma) and analyzed with use of a EPICS ALTRA flow cytometer (Beckman Coulter, Fullerton, CA) and CellQuest software as previously described [bib_ref] EXEL-0862, a novel tyrosine kinase inhibitor, induces apoptosis in vitro and ex..., Pan [/bib_ref]. Apoptosis was observed by Hoechst 33258 staining as described [bib_ref] 17-Beta-estradiol alters Jurkat lymphocyte cell cycling and induces apoptosis through suppression of..., Jenkins [/bib_ref]. Apoptotic cells were characterized by morphological alteration as condensed nuclei and cell shrinkage. Necrosis was assayed using the CytoTox96 non-radioactive cytotoxicity assay kit (Promega, Madison, WI), which quantifies cell death and cell lysis, based on the measurement of lactate dehydrogenase (LDH) activity released from the cytosol of damaged cells into the supernatant. # Cell cycle analysis Cell cycle was assessed as previously described [bib_ref] Increased expression of NFkappa B subunits in human pancreatic cancer cells, Chandler [/bib_ref] with minor modifications. Briefly, cells were plated in parallel in 35-mm 2 culture plates at a concentration of 8×10 5 cells per plate. After 24 h of serum starve, cells were exposed to 10% FBS-medium with/without 4 mmol/L ASA for various durations and then were harvested by trypsinization, washed in cool PBS twice and fixed in 75% ethanol overnight in 4 °C. After that, cells were incubated in solution with DNA-binding dye propidium iodide (PI, 50 g/L), RNase (4×10 3 kU/L), NaF (0.3 g/L) and sodium citrate (1 g/L) for 30 min at 37 °C in the dark. Finally, red fluorescence from 488 mm laser-excited PI in every cells was analyzed by EPICS ALTRA flow cytometer (Beckman Coulter, Fullerton, CA) using a peak fluorescence gate to discriminate aggregates. The percentage of cells in G 0 /G 1 , S and G 2 /M was determined from DNA content histograms by Multicycle for windows (Phoenix Flow Systems, San Diego, CA). Preparation of nuclear extracts PANC-1 cells were incubated with different concentrations of ASA for 24 h, followed by preparation of nuclear extracts using nuclear extract kit (Pierce, IL) according to the manufacturer's instructions. In brief, about 3×10 6 cells per sample were washed with ice-cold PBS/phosphate inhibitors, scraped, and collected by centrifugation at 500×g for 5 min. The pellets were suspended in 500 μL of hypotonic buffer, incubated on ice for 15 min and centrifuged at 14 000×g for 30 s at 4 °C. The supernatant (cytosolic extract) was removed and the pellet (nuclear fraction) was suspended in 50 μL of complete lysis buffer and incubated on ice for 30 min with frequent mixing. Finally the suspension was centrifuged at 14 000×g for 10 min at 4 °C and the supernatant (nuclear extract) was subjected to Western blot analysis. # Western blot analysis Western blot was performed as previously described [bib_ref] Increased expression of NFkappa B subunits in human pancreatic cancer cells, Chandler [/bib_ref]. The following antibodies were used: antibody against PCNA (1:15 000, Cell Signaling Technology, Beverly, MA), GAPDH, cyclin D1, Bcl-2, GSK-3β, phosphor-GSK-3β-Ser9, Akt, phosphor-Akt-Ser473 (1:1000, Cell Signaling Technology), phosphor-PP2A-Tyr307 (1:1000, Santa Cruz Biotechnology, Santa Cruz, CA), PP2A (1:1000, Millipore, Billerica, MA), β-actin (1:1000, Thermo Scientific IHC, Fremont, CA) and C23 (also as designated nucleolin, 1:1000, Santa Cruz Biotechnology). ## Reverse transcription-polymerase chain reaction (rt-pcr) Total RNA was isolated with TRIzol reagent (Gibco-BRL) according to the manufacturer's instructions. Complementary DNA was synthesized from 1 μg of total RNA by reverse transcription using the Superscript TM II reverse transcriptase kit (Gibco-BRL). Sequence of the PCR primers: GAPDH: 5'-CCA-CCCATGGCAAATTCCATGGCA-3' (sense primer), 5'-TCTA-GACGGCAGGTCAGGTCCACC-3' (antisense primer). Cyclin D1: 5'-GTCACACTTGATCACTCTGG-3' (sense primer), 5'-TGGCCATGAACTACCTGGA-3' (antisense primer). Bcl-2: 5'-GTGGAGGAGCTCTTCAGGGA-3'(sense primer), 5'-CGGTGCTTGGCAATTAGTGG-3' (antisense primer). The PCR conditions contained an initial cDNA synthesis reaction ## Statistics Data are presented as mean±SEM for three separate experiments. P<0.05 was considered significant using Student's t test. # Results ## Asa inhibits the growth of panc-1 cells in vitro Initially, we determined the effect of ASA on the growth of PANC-1 cells using MTT assay. As the growth curve shown, ASA treatment attenuates the growth rate of PANC-1 cells time-and dose-dependently [fig_ref] Figure 1: ASA inhibits growth of PANC-1 cells [/fig_ref]. Compared with the untreated cells, 4 mmol/L ASA is sufficient to inhibit the cell growth by about 40% (P<0.01) at 24 h, and the inhibitory effect of ASA becomes more significant at 72 h. Clearly, ASA alone is able to slow down the growth of PANC-1 cells. ## Asa decreases proliferation instead of inducing apoptosis or necrosis in panc-1 cells The reduction in growth of PANC-1 cells in response to ASA could be explained either by increased cell death or by reduced cell proliferation. Thus, the level of proliferating cell nuclear antigen (PCNA), an established index for proliferation cells, was firstly assessed to determine cell proliferation in PANC-1 cells [bib_ref] Suppression of pancreatic carcinoma growth by activating peroxisome proliferator-activated receptor gamma involves..., Dong [/bib_ref]. Western blot analysis clearly shows that PCNA protein expression undergoes a down-regulation change in a time-dependent manner after exposure to ASA [fig_ref] Figure 2: Effect of ASA on PANC-1 cells proliferation, apoptosis and necrosis [/fig_ref]. Secondly, Anexin V, Hoechst staining and LDH examination were used to investigate whether treatment of ASA causes apoptosis or necrosis. There was no increase of Anexin V positive cells after ASA treatment during 24 h to 72 h [fig_ref] Figure 2: Effect of ASA on PANC-1 cells proliferation, apoptosis and necrosis [/fig_ref]. Meanwhile, Apoptotic bodies are not observed in either [fig_ref] Figure 2: Effect of ASA on PANC-1 cells proliferation, apoptosis and necrosis [/fig_ref]. And there is not any statistically significant difference in the mean LDH absorbance between ASA-treated and untreated cells during the time course of 24 h to 72 h [fig_ref] Figure 2: Effect of ASA on PANC-1 cells proliferation, apoptosis and necrosis [/fig_ref]. Hence, cytotoxicity is not contributing to the reduction in cell growth. Therefore, we strongly believe that the inhibitory effect of ASA on PANC-1 growth is based on the reduction of proliferation instead of induction of cell death. ASA causes cell cycle arrest at G 1 phase and decreases S phase fraction in PANC-1 cells To determine the mechanism involved in the reduction of cell proliferation, we analyzed the cell cycle distributions of PANC-1 cells treated by ASA. ASA hampers the cell cycle progression by arresting up to three quarters of cancer cells at G 1 phase, and by decreasing S phase fraction by about 40% in 24 h compared with the control [fig_ref] Figure 3: Cell cycle distributions in ASA-treated PANC-1 cells [/fig_ref]. Additionally, no subdiploid (sub-G 0 /G 1 ) peak in DNA content histograms is obtained by flow cytometry, which further supports the findings mentioned above that no apoptosis is induced by ASA. ASA augments the effect of gemcitabine on cell survival and induction of apoptosis in PANC-1 cells To determine the effect of ASA on gemcitabine-induced cytotoxicity, PANC-1 cells were treated with gemcitabine in the presence of ASA. Only minor reduction of cell survival was observed with the treatment of gemcitabine alone at the dose of 0.2 μmol/L to 20 μmol/L in 24 h [fig_ref] Figure 4: ASA augments the anti-survival and pro-apoptosis effects of gemcitabine on PANC-1 cells [/fig_ref]. Obviously, human pancreatic PANC-1 cells are not sensitive to gemcitabine treatment, which is consistent with the previous report [bib_ref] K-ras oncogene silencing strategy reduces tumor growth and enhances gemcitabine chemotherapy efficacy..., Rejiba [/bib_ref]. However, ASA significantly promotes gemcitabineinduced cytotoxicity, which is dependent upon concentrations of ASA and gemcitabine (P<0.05). Similar results were observed with Hoechst staining. Data from Hoechst assays [fig_ref] Figure 4: ASA augments the anti-survival and pro-apoptosis effects of gemcitabine on PANC-1 cells [/fig_ref] show that ASA pronouncedly increased the apoptotic effect of gemcitabine, whereas gemcitabine alone, has minimal apoptotic effect on the PANC-1 cells. ## Asa inhibits gsk-3β activation in panc-1 cells To explore the molecular mechanism underlying the above effect of ASA on PANC-1 cells, we investigated whether the effect of ASA is associated with the inhibition of GSK-3β activation. Using Western blot analysis of total cell proteins, we observed that ASA dose-dependently increased phosphorylated level of GSK-3β at Ser9, which represents the inactive form of GSK-3β kinase, while has no effect on the total level of [fig_ref] Figure 5: Inactivation of GSK-3β by ASA [/fig_ref]. To exactly verify whether this alteration takes place in the nucleus, we replicated the experiment using nuclear extracts. Interestingly, the change of GSK-3β activation in nuclear fractions followed a similar pattern to that in total extracts [fig_ref] Figure 5: Inactivation of GSK-3β by ASA [/fig_ref]. Thus, we identified that GSK-3β is inactivated by ASA in the nucleus of PANC-1 cells. To further elucidate the signaling pathways that inactivate GSK-3β, we examined the effect of ASA on Akt and PP2A, which are both implicated in the regulation of GSK-3β activity [bib_ref] GSK-3beta acts downstream of PP2A and the PI3-kinase-Akt pathway, and upstream of..., Lin [/bib_ref]. In untreated PANC-1 cells, phosphorylated Akt at Ser437 is marginal. After treatment with ASA for 24 h, a very strong band of the phosphorylated Akt appeared [fig_ref] Figure 5: Inactivation of GSK-3β by ASA [/fig_ref]. Therefore, ASA stimulates Akt activation via protein phosphorylation at Ser437 provided that phosphorylation of the Ser437 residue is required for a maximal Akt activation. On the other hand, ASA causes little change on the inactivated phosphorylation status of PP2A at Tyr307. ## Asa downregulates the expression of cyclin d1 and bcl-2 We then assessed the expression of cyclin D1, which plays important role in tumor cell proliferation and cell cycle progression from G 1 phase to S phase; and Bcl-2, which is involved in tumor survival and chemoresistance in pancreatic cancer cells [bib_ref] Curcumin potentiates antitumor activity of gemcitabine in an orthotopic model of pancreatic..., Kunnumakkara [/bib_ref] , because both of the genes can be regulated by GSK-3β in pancreatic cancer cells [bib_ref] Glycogen synthase kinase-3beta participates in nuclear factor kappaB-mediated gene transcription and cell..., Ougolkov [/bib_ref]. RT-PCR analysis [fig_ref] Figure 6: ASA downregulates the expression of cyclin D1 and Bcl-2 in PANC-1 cells [/fig_ref] indicates that mRNA level of cylcin D1 and Bcl-2 are significantly suppressed by ASA in a dose-dependent manner in 24 h. Western blotting [fig_ref] Figure 6: ASA downregulates the expression of cyclin D1 and Bcl-2 in PANC-1 cells [/fig_ref] shows that ASA also reduces the protein level of cyclin D1 and Bcl-2. ASA inhibits growth, decreases S phase accumulation, augments the effect of gemcitabine and represses GSK-3β activation in Capan-1 cells To test whether the effects of ASA in PANC-1 cells are cellspecific or not, another human pancreatic cancer cell line Capan-1 cells were used. ASA inhibits cell growth dose-and time-dependently in Capan-1 cells as in PANC-1 cells. Moreover, 4 mmol/L ASA decreases S phase accumulation time-dependently. Furthermore, augment of gemcitabine-induced apoptosis by ASA was also seen in Capan-1 cells as in PANC-1 cells. Mechanically, after treatment with ASA for 24 h, the levels of GSK-3β and Akt phosphorylation displayed the same alteration panel as in PANC-1 cells. These findings demonstrate that ASA may have a broad therapeutic potential in human pancreatic cancer cells. # Discussion The aim of this study is to examine the effect of ASA on the growth of pancreatic cancer cells and determine whether ASA, the famous NSAID, can sensitize the cells to gemcitabine. We find that ASA alone inhibits the proliferation of human pancreatic cancer cells by hampering cell cycle progressing, and that ASA enhance the apoptotic effects of gemcitabine; these effects of ASA may be associated with suppression of the ASA is a first-line medication in pancreatic cancer pain control, intend ed to keep patients comfortable without resorting to opioids [bib_ref] Aspirin and pancreatic cancer pain, Moertel [/bib_ref] [bib_ref] Pilot study of nitric oxide-donating aspirin in patients with pancreatic cancer pain, Iconomou [/bib_ref]. Because high concentrations (1-5 mmol/L) of ASA are achievable in vivo by oral administration of ASA at 4-10 g/d during treatment of rheumatic disorder and arthritis [bib_ref] Aspirin protects Caco-2 cells from apoptosis after serum deprivation through the activation..., Ricchi [/bib_ref] , accumulating literature is evaluating additional, COX-independent, biological activities for these high doses of ASA. Previous studies suggest that ASA is able to induce apoptosis in pancreatic cancer cells [bib_ref] Nuclear factor kappa B activation is a potential target for preventing pancreatic..., Sclabas [/bib_ref]. However, in our work, we demonstrated that ASA, even at its relative high dose (4 mmol/L), hardly causes apoptosis in PANC-1 cells. This was evidenced by the Annexin V and Hoechest 33258 staining and flow cytometry assays. The pronounced inhibition of growth of pancreatic cancer cells by ASA, is mainly due to the reduction of proliferation and retardation of cell cycle progression. Gemcitabine alone only has a marginal effect on cell growth and apoptosis in PANC-1 cell line, which fits well with previous finding that PANC-1, Capan-1 cells are considered relatively resistant to many chemotherapeutic regimens [bib_ref] Role of NF-kappaB and Akt/PI3K in the resistance of pancreatic carcinoma cell..., Arlt [/bib_ref] [bib_ref] K-ras oncogene silencing strategy reduces tumor growth and enhances gemcitabine chemotherapy efficacy..., Rejiba [/bib_ref]. However, ASA was shown to significantly enhance the apoptotic effect of gemcitabine in these cells. This finding is not contra-dictory to the previous report that combination of gemcitabine and celecoxib did not demonstrate significant improvement in patients with advanced pancreatic cancer [bib_ref] Gemcitabine plus celecoxib in patients with advanced or metastatic pancreatic adenocarcinoma: results..., Dragovich [/bib_ref]. Because PANC-1 and Capan-1 cells are cyclooxygenase-2-negative, most bioactivities of ASA on this cell line were thought to be COX-2independent. The mechanism by which ASA exerts above bioactivity may involve the suppression of activity of GSK-3β. Ougolkov [bib_ref] Aberrant nuclear accumulation of glycogen synthase kinase-3beta in human pancreatic cancer: association..., Ougolkov [/bib_ref] have reported that GSK-3β aberrantly accumulates in human pancreatic cancer cells and its accumulation in cell nuclear is associated with its kinase activity and tumor differentiation. GSK-3β positively regulates the activity of NF-κB, while NF-κB per se plays a pivotal role in promoting gemcitabine resistance in pancreatic cancer [bib_ref] Role of NF-kappaB and Akt/PI3K in the resistance of pancreatic carcinoma cell..., Arlt [/bib_ref] , so it is very likely that the inactivation of GSK-3β by ASA can sensitize the cells to gemcitabine. This hypothesis was evidenced by the finding that LiCl, the pharmacological inhibitor of GSK-3β, can significantly sensitize PANC-1 or Capan-1 cells to gemcitabine-induced apoptosis (Supplementary . Additional studies are needed to determine the necessary role of GSK-3β in the effect of ASA on pancreatic cancer cells and provide evidence that this mechanism occurs also in vivo. Our results also display that ASA inhibits the expression of GSK-3β downstream genes, Bcl-2 and cyclin D1. Of note, the anti-apoptotic Bcl-2 plays an important role in the development of many chemo- therapy resistances in cancer cells. It impairs the mitochondrial apoptotic function by neutralizing the proapoptotic Bcl-2 family members such as Bax and Bak [bib_ref] Expression of bcl-2 in small cell lung carcinoma cells, Ikegaki [/bib_ref]. Thereby, Bcl-2 may be the effecter molecule of ASA to regulate gemcitabine-induced cytotoxicity and apoptosis. Cyclin D1 is overexpressed in human pancreatic cancer tissue and inversely correlated with patient survival [bib_ref] Increased cyclin D1 in human pancreatic cancer is associated with decreased postoperative..., Kornmann [/bib_ref]. It governs the checkpoint of G 1 to S phase progression [bib_ref] The regulation of cyclin D1 degradation: roles in cancer development and the..., Alao [/bib_ref] and is always responsible for the aberrant cell cycle in tumor cells [bib_ref] The cell cycle: a review of regulation, deregulation and therapeutic targets in..., Vermeulen [/bib_ref]. In our work, suppression of cylcin D1 expression by ASA seems to contribute to the cell cycle retardation and proliferation reduction in PANC-1 cells. Admittedly, GSK-3β can also regulate cyclin D1 at the posttranslational level by phosphorylating cyclin D1 on T286 and inducing its rapid turnover [bib_ref] The regulation of cyclin D1 degradation: roles in cancer development and the..., Alao [/bib_ref]. In the present work, we display that ASA significantly inactivates GSK-3β and suppresses cyclin D1 at the transcription level in pancreatic cancer cells, which is, to our knowledge, never reported before. Akt and PP2A pathways are implicated in the activation of GSK-3β. PP2A dephosphorylates and activates GSK-3β [bib_ref] Integrin alpha 2 beta 1 promotes activation of protein phosphatase 2A and..., Ivaska [/bib_ref] while Akt can inactivate GSK-3β by phosphorylating GSK-3β at Ser9 [bib_ref] Inhibition of glycogen synthase kinase-3 by insulin mediated by protein kinase B, Cross [/bib_ref]. Alternatively, PP2A can act upstream of Akt pathway to indirectly regulate GSK-3β signaling [bib_ref] GSK-3beta acts downstream of PP2A and the PI3-kinase-Akt pathway, and upstream of..., Lin [/bib_ref]. However, our results demonstrate that only Akt activation is involved in the inactivation of GSK-3β by ASA in PANC-1 cells, which takes place independently of PP2A pathway. Concerns may be aroused that Akt activation renders the resistance of pancreatic carcinoma against anticancer drugs [bib_ref] Survival signalling by Akt and eIF4E in oncogenesis and cancer therapy, Wendel [/bib_ref] [bib_ref] Aspirin reduces the outcome of anticancer therapy in Meth A-bearing mice through..., Di Palma [/bib_ref]. In fact, Akt does not seem to be involved in gemcitabine resistance of human pancreatic carcinoma cell lines. It is reported that neither did the basal Akt activity correlate with the sensitivity towards gemcitabine treatment, nor did the inhibition of Akt by LY294002 alter gemcitabine-induced apoptosis [bib_ref] Role of NF-kappaB and Akt/PI3K in the resistance of pancreatic carcinoma cell..., Arlt [/bib_ref]. Thus ASA is not likely to confer chemoresistance to the current clinical regimen for pancreatic carcinomas. In conclusion, our results show that ASA inhibits proliferation and potentiates the apoptosis-inducing effect of gemcitabine in pancreatic cancer cells, probably by inhibiting activation of GSK-3β and the expression of its regulated targets. Thus, the traditional agent ASA may prove to be a novel candidate to be used in combination with gemcitabine for the chemotherapy of pancreatic carcinoma. [fig] Figure 1: ASA inhibits growth of PANC-1 cells. Cells were treated with various concentrations of ASA and cell growth was determined using the MTT assay at 24, 48, and 72 h. Results represent the mean absorbance at 570 nm of three different experiments with quintuple wells. SEM bars are not included for clarity. c P<0.01 compared with untreated cells at all time points. [/fig] [fig] Figure 2: Effect of ASA on PANC-1 cells proliferation, apoptosis and necrosis. (A) PANC-1 cells were treated with 4 mmol/L ASA for various time and cell lysates were subjected to Western blot using anti-PCNA and anti-GAPDH antibodies. (B) AnnexinV/propidium iodide double staining for apoptosis assay. (C) Fluorescent staining of nuclei by Hoechst 33258 in PANC-1 cells treated with or without 4 mmol/L ASA for 24 h. (D) PANC-1 cells were incubated with 4 mmol/L ASA for the indicated time and cultured medium was collected for LDH release assay. Results represent the mean of three different experiments with quintuple wells. [/fig] [fig] Figure 3: Cell cycle distributions in ASA-treated PANC-1 cells. PANC-1 cells were treated with or without 4 mmol/L ASA for various time and cell cycle analysis was performed using the fluorescent cytometry. [/fig] [fig] Figure 4: ASA augments the anti-survival and pro-apoptosis effects of gemcitabine on PANC-1 cells. (A) Cells were treated with various concentrations of gemcitabine in the presence/absence of indicated concentrations of ASA for 24 h and cell growth was measured by MTT assay. Data are expressed as percentage of cell growth observed in untreated cultures. Results represent the mean of three different experiments with quintuple wells. b P<0.05 compared with untreated cells; e P<0.05 compared with gemcitabine alone-treated cells at corresponding concentrations. (B) Cells were treated with various concentrations of ASA in combination with 0.2 μmol/L gemctitabine for 24 h. Apoptosis was then determined by Hoechst 33258 assay. b P<0.05 compared with 0 [/fig] [fig] Figure 6: ASA downregulates the expression of cyclin D1 and Bcl-2 in PANC-1 cells. PANC-1 cells were treated with various concentrations of ASA for 24 h and harvested for RT-PCR (A) or Western blotting analysis (B). [/fig] [fig] Figure 5: Inactivation of GSK-3β by ASA. Cells were treated with various concentrations of ASA for 24 h. Total extracts (A, C) or nuclear extracts (B) were obtained and subjected to Western blotting analysis. GAPDH and C23 were loaded to normalized total and nuclear protein. GSK-3β and downregulation of cyclin D1 and Bcl-2. These findings extend our understanding of the function and molecular basis of ASA on the dismal disease. [/fig] [fig] Figure 7: (A) Capan-1 cells were treated with various concentrations of ASA and cell growth was determined using the MTT assay at 24, 48, and 72 h. Results represent the mean absorbance at 570 nm of three different experiments with quintuple wells. SEM bars are not included for clarity. c P<0.01 compared with untreated cells at corresponding time points. (B) Capan-1 cells were treated with or without 4 mmol/L ASA for various time and cell cycle analysis was performed using the fluorescent cytometry. (C) Cells were treated with various concentrations of ASA in combination with 0.2 μmol/L gemctitabine for 24 h. Apoptosis was then determined by Hoechst 33258 assay. b P<0.05 compared with 0.2 μmol/L gemcitabine-treated cells. (D) The effect of ASA on Akt/GSK-3β signaling pathway in Capan- [/fig]

Bilateral Hypoplasia of the Medial and Lateral Menisci Discoid and abnormally attached menisci are common congen ital abnormalities of the meniscus 1,2) . In contrast, hypoplasia of the meniscus is very rare. To date, it has been reported alone 36) , associated with other abnormalities of the knee joint5,7,8), and in patients with partial deficiency of the meniscus 9) . To our knowl edge, however, there have been no reports describing hypoplasia of the medial and lateral menisci in both knee joints.Hypoplasia of the meniscus is a very rare congenital abnormality, with only a few cases reported to date. A 9yearold girl visited our hospital due to lateral knee pain following a hyperextension injury to the left knee. Magnetic resonance imaging showed hypoplasia of the medial and lateral menisci, as well as a posterior horn tear of the lateral meniscus, in both knee joints. To our knowledge, this is the first report of a patient with hypoplasia of the medial and lateral menisci in both knee joints. ## Case report A 9yearold girl visited our hospital due to lateral knee pain that developed one week after a hyperextension injury of her left knee joint during exercise. Physical examination revealed lateral joint line tenderness and positive results for the McMurray test in both knee joints. The Lachman test, anterior and posterior drawer and pivot shift tests, and maneuvers for rotator instabil ity were all negative. The range of motion (ROM) was no flexion contracture and a further flexion of 120° with mild pain in her left knee. Plain radiographs showed no specific findings and the notch width index was 0.234. Magnetic resonance imaging (MRI) of both knee joints showed an almost complete absence of the anterior and posterior horns of the medial meniscus, except for the peripheral portion, hy poplastic anterior horns and tears in the posterior horns of the lateral meniscus in both knees [fig_ref] Figure 1: Magnetic resonance imaging of the left [/fig_ref]. The patient did not com plain of pain or discomfort in her right knee, but did complain of pain and limited motion in her left knee. We, therefore, decided to perform arthroscopic surgery only on her left knee. At arthroscopy, the medial meniscus showed hypoplastic changes (approximately 5 mm in width and 1-2 mm in height). The anterior horn of the lateral meniscus showed hypoplastic changes, while the posterior horn was connected to the menisco femoral ligament with fibrous tissue. This fibrous tissue extended and covered the popliteal hiatus. Both the anterior and posterior cruciate ligaments were intact and there was no chondral pathol ogy medially or laterally. Arthroscopic partial meniscectomy was performed for the horizontal tear of the posterior horn of the lateral meniscus [fig_ref] Figure 2: Arthroscopic findings of the left knee joint [/fig_ref]. Postoperatively, the patient was allowed to bear weight and per form knee motions as tolerated. Two weeks after the operation, she restored full active and passive ROM without pain. At 4year followup, she had no symptoms or abnormal objective find ings for both knees and she had returned to full activity. # Discussion The characteristic shapes of both the medial and lateral me nisci are observed early in prenatal development, maturing along with the capsule and coronary and cruciate ligaments. No abrupt changes in development occur after birth. Rather, gradual changes occur, including a decrease in vascularity, progressing from the center to the peripheral margins; a growth in size; and changes in the configuration of the menisci 1) . The association of other anomalies in the knee with hypoplastic menisci may be due to the common mesenchymal origin of these structures [bib_ref] Development of the menisci of the hu man knee joint: morphological changes..., Clark [/bib_ref] [bib_ref] Bilateral hypo plasia of the medial meniscus, Monllau [/bib_ref]. How ever, in the patient described here, the only anomalies were in the menisci. The most common congenital abnormality of the meniscus is a discoid meniscus, which is more frequent in the lateral than the medial meniscus [bib_ref] Development of the menisci of the hu man knee joint: morphological changes..., Clark [/bib_ref] [bib_ref] Congenital absence of the menisci and cruciate ligaments of the knee: a..., Tolo [/bib_ref]. Other abnormalities have been reported, including abnormal attachments of the meniscus, hypoplasia and partial deficiency 39) . These abnormalities were all incidentally diagnosed in patients with symptoms associated with trauma or other congenital anomalies. Thus, their exact incidence is un known [bib_ref] Bilateral hypo plasia of the medial meniscus, Monllau [/bib_ref] [bib_ref] Bilateral hypoplastic lateral meniscus, Ohana [/bib_ref]. Congenital hypoplasia of the medial meniscus has been report ed in three patients [bib_ref] The hypoplastic, hypermobile meniscus, Pfeil [/bib_ref] [bib_ref] Congenital hypoplasia of the medial meniscus: a report of two cases, Twyman [/bib_ref] , and bilateral hypoplasia of the medial me niscus with an Outerbridge grade III chondral lesion of the me dial femoral condyle in one patient 2) . One patient was reported to have partial deficiency of the lateral meniscus 9) and another was found to have bilateral congenital absence of the anterior cruciate ligament and the internal menisci of the knee [bib_ref] Bilateral congeni tal absence of the anterior cruciate ligament and the internal..., Dejour [/bib_ref]. Another patient was found to have congenital absence of both menisci and both cruciate ligaments with radial club hands and congenital throm bocytopenia [bib_ref] Congenital absence of the menisci and cruciate ligaments of the knee: a..., Tolo [/bib_ref]. Bilateral hypoplastic lateral menisci with an osteo chondral lesion in the lateral tibial plateau were reported in one patient [bib_ref] Bilateral hypoplastic lateral meniscus, Ohana [/bib_ref] , with another having congenital hypoplasia of the lateral meniscus and anterior cruciate ligament associated with osteo chondritis dissecans of the medial femoral condyle [bib_ref] Osteochondritis disse cans of the medial femoral condyle associated with congeni tal..., Mitsuoka [/bib_ref]. Osteochon dritis dissecans of the medial femoral condyle associated with hypoplastic medial and partially deficient lateral menisci was also reported [bib_ref] Osteochondritis disse cans of the medial femoral condyle associated with malfor mation..., Beyzadeoglu [/bib_ref]. However, we have been unable to find a description of bilateral hypoplastic medial and lateral menisci. We think that future research efforts are necessary to detect the incidence and clinical course in patients with hypoplastic menisci. ## Conflict of interest No potential conflict of interest relevant to this article was re ported. [fig] Figure 1: Magnetic resonance imaging of the left (A-D) and right (E-H) knee joints. (A, E) Hypoplastic anterior and posterior horns of the medial menis ci. (B, F) Hypoplastic anterior horns and posterior horn tears of the lateral menisci. (C, D, G, H) Hypoplasia of both menisci and lateral meniscus tears. [/fig] [fig] Figure 2: Arthroscopic findings of the left knee joint. (A, B) Hypoplastic anterior and posterior horns of the medial meniscus. (C) Hypoplastic anterior horn of the lateral meniscus. (D, E) Abnormal attachment of the posterior horn of the lateral meniscus. The posterior horn was connected to a menis cofemoral ligament with fibrous tissue. (F) Extension of the fibrous tissue covering the popliteal hiatus. [/fig]

Sound out the impaired perfusion: Photoacoustic imaging in preclinical ischemic stroke Acoustically detecting the optical absorption contrast, photoacoustic imaging (PAI) is a highly versatile imaging modality that can provide anatomical, functional, molecular, and metabolic information of biological tissues. PAI is highly scalable and can probe the same biological process at various length scales ranging from single cells (microscopic) to the whole organ (macroscopic). Using hemoglobin as the endogenous contrast, PAI is capable of label-free imaging of blood vessels in the brain and mapping hemodynamic functions such as blood oxygenation and blood flow. These imaging merits make PAI a great tool for studying ischemic stroke, particularly for probing into hemodynamic changes and impaired cerebral blood perfusion as a consequence of stroke. In this narrative review, we aim to summarize the scientific progresses in the past decade by using PAI to monitor cerebral blood vessel impairment and restoration after ischemic stroke, mostly in the preclinical setting. We also outline and discuss the major technological barriers and challenges that need to be overcome so that PAI can play a more significant role in preclinical stroke research, and more importantly, accelerate its translation to be a useful clinical diagnosis and management tool for human strokes. KEYWORDS photoacoustic imaging, ischemic stroke, photoacoustic microscopy, photoacoustic computed tomography, functional brain imaging, blood oxygenation, brain perfusion Frontiers in Neuroscience 01 frontiersin.org # Introduction Ischemic stroke occurs with a vessel blockage of blood flow in a certain region of the brain. This blood flow interruption is often a result of embolic or thrombotic occlusion of an artery within the brain. The interruption of blood flow in the brain can cause tissue damage, such as neuronal injury and death, due to a lack of vital nutrient delivery to the brain tissue. Thus, timely detection and removal of the blood clot and recanalization of the vessel are critical for acute ischemic stroke treatment [bib_ref] Treatment of acute ischemic stroke, Brott [/bib_ref]. The first-line acute stroke treatment is mechanical or pharmacologic reperfusion therapy [bib_ref] Therapeutic strategies for the treatment of stroke, Green [/bib_ref] , and use of a clinical imaging modality such as magnetic resonance angiography (MRA) or computed tomography angiography (CTA) that can visualize the occluded vessel. These imaging modalities come at a high cost and require the use of exogenous contrast agents. Furthermore, these clinical modalities often provide only the information of blood perfusion, but lack the ability to measure blood oxygenation, the presence of specific biomolecules, or tissue viability. Thus, MRA and CTA are less used in preclinical studies of ischemic stroke. Instead, preclinical studies of ischemic stroke have heavily relied on the use of histology and behavioral testing as measures of experimental treatment efficacy. However, histology requires sacrifice of the animals and thus can only be used as an endpoint measure, and behavioral tests cannot provide concrete anatomical or physiological information. To provide a more efficient and quantitative approach to measuring stroke outcomes, various biomedical imaging technologies have recently been developed and tested for preclinical ischemic stroke studies. Photoacoustic imaging (PAI) has emerged as a popular biomedical imaging modality over the past twenty years due to its intrinsic ability to combine optical contrast with acoustic detection [bib_ref] Photoacoustic imaging in biomedicine, Xu [/bib_ref] [bib_ref] A practical guide to photoacoustic tomography in the life sciences, Wang [/bib_ref]. The fundamental imaging principle of PAI is shown in . A short laser pulse excites the sample, leading to the absorbers within the target to heat up via photothermal effect. This temperature rise results in a thermoelastic expansion and subsequently outwardly propagating ultrasound waves. In principle, any molecule absorbing light qualifies as a potential contrast in PAI, allowing for a plethora of both endogenous and exogenous contrast agents [bib_ref] Contrast agents for molecular photoacoustic imaging, Weber [/bib_ref]. Of the endogenous category, hemoglobin is the most commonly used contrast for PAI due to its abundance in biological tissues and its relatively strong optical absorption in the visible and nearinfrared (NIR) light domain. With hemoglobin used as the primary contrast, PAI has been widely used as an angiographic imaging modality that produces blood vessel images in the deep tissues. While the spatial scale of PAI varies greatly from microscopic to macroscopic depending on the implementation, the contrast origin remains the same at all scales. This makes PAI a powerful research tool to investigate the same biological phenomena at vastly different scales. Photoacoustic imaging (PAI) has two major implementations: photoacoustic microscopy (PAM) and photoacoustic computed tomography (PACT) . PAM can be further divided into optical-resolution PAM (OR-PAM) and acoustic-resolution PAM (AR-PAM). OR-PAM differentiates itself from AR-PAM based on the tight optical focusing of the excitation light. AR-PAM relies on diffuse optical excitation, while the resolution depends on the acoustic focusing of the ultrasound transducer. OR-PAM produces higher resolution images than AR-PAM but has a superficial penetration depth (∼1 mm). The macroscopic PAI implementation, PACT, utilizes diffuse light excitation and parallel acoustic detection with ultrasound transducer arrays to create tomographic images. PACT generally has a greater penetration depth (>1 cm) than PAM implementations, but lower spatial resolution (>300 µm). Due to the high flexibility of ultrasound transducers and methods of diffuse light illumination, PACT has many different implementations. The geometry of the transducer elements, scanning pattern, central frequency and bandwidth, excitation laser wavelength and delivery method, and reconstruction technique all contribute to the resultant image quality. Depending on the design requirements of the imaging system, an optimal combination of system components can be selected. For example, while a hemispherical ultrasound transducer can produce higher spatial-resolution with minimal limited-view artifacts, it usually provides a smaller field of view than a planar or ring-shaped array with the same number of elements [bib_ref] Review of cost reduction methods in photoacoustic computed tomography, Fatima [/bib_ref]. Furthermore, choosing a high-frequency transducer will result in high resolution images at the cost of imaging depth and increased sensitivity to the speed of sound heterogeneity. The experimental challenges of acquiring data should also be considered when designing a PACT system. For example, brain imaging with a full-ring array can be more experimentally difficult than with a hemispherical or linear-array due to the need for acoustic coupling between the head and transducer. The PACT implementation for preclinical stroke study can be optimized by considering the desired spatial and temporal resolution, field of view, detection sensitivity, cost, and animal models. While PAI has been used in several applications in brain research [bib_ref] Photoacoustic brain imaging: From microscopic to macroscopic scales, Yao [/bib_ref] , imaging acute ischemic stroke is of particular interest, due to both the blood-sensitive nature of PAI and the severity and prevalence of ischemic stroke [bib_ref] Acute ischemic stroke, Van Der Worp [/bib_ref]. Over 795,000 people per year in the US suffer stroke, approximately 80% of which are ischemic stroke [bib_ref] Heart disease and stroke statistics-2022 update: A report from the American heart..., Tsao [/bib_ref]. Stroke remains a leading cause of disability in the US, including language difficulty and cognitive deficits, with motor impairment being the most common complication after stroke. Many stroke survivors struggle with daily activity and have poor quality of life. With its ability to provide both structural and functional information of the brain vessels, PAI can be a useful imaging tool in the setting of ischemic stroke. Ischemic stroke causes disrupted blood flow to a region of the brain, resulting in reduced vessel density and blood oxygenation, both of which can be readily detected by using PAI at different length scales without using any exogenous contrast agents [bib_ref] Photoacoustic tomography of blood oxygenation: A mini review, Li [/bib_ref]. Thus, PAI of ischemic stroke has been extensively investigated Fundamental principle of PAI. and validated [bib_ref] In vivo flow cytometry of circulating clots using negative photothermal and photoacoustic..., Galanzha [/bib_ref] [bib_ref] Vessel segmentation analysis of ischemic stroke images acquired with photoacoustic microscopy, Soetikno [/bib_ref] [bib_ref] Noninvasive label-free detection of circulating white and red blood clots in deep..., Juratli [/bib_ref] [bib_ref] Combined ultrasound and photoacoustic imaging of blood clot during microbubble-assisted sonothrombolysis, Das [/bib_ref] [bib_ref] Labelfree high frame rate imaging of circulating blood clots using a dual..., Das [/bib_ref]. Using PAI as an alternative to MRA or CTA in evaluating post-stroke reperfusion could provide many benefits for the patients, although there are significant technical obstacles that need to be overcome. This concise review is divided into two major sections. In the first section, we introduce the current achievements of PAI in ischemic stroke study at multiple length scales, ranging from micro vessels in small animal brains to major vessels in human brains. In the second section, we discuss the major technological challenges still faced in PAI of ischemic stroke, both in light and sound, and we present potential solutions that may translate PAI into a more powerful technology for neuroscience research. ## Applications of photoacoustic imaging in preclinical ischemic stroke research The works presented in this review were mostly identified using the Google Scholar search engine by November 2022. The major key words used in the search engine included ischemic stroke, photoacoustic imaging, photoacoustic microscopy, photoacoustic computed tomography, and brain imaging. Over 300 papers were reviewed, wherein the relevance, novelty, and quality of the imaging systems were the main criteria when considering representative papers for inclusion. Representative applications of PAI in ischemic stroke research in the last decade is summarized in [fig_ref] TABLE 1: Summary of major publications on ischemic stroke using PAI in the past... [/fig_ref]. The included papers were chosen to represent ischemic stroke studies from single-vessel to large-animals, showing the wide range of imaging scales of PAI. Among these studies, three stroke models have been used in PAI studies: photothrombotic (PT) stroke, temporary and permanent middle cerebral artery occlusion (tMCAO and pMCAO) [bib_ref] Rodent models of focal stroke: Size, mechanism, and purpose, Carmichael [/bib_ref] [bib_ref] Acute ischemic stroke: Overview of major experimental rodent models, pathophysiology, and therapy..., Durukan [/bib_ref]. PT and pMCAO stroke models represent permanent ischemic stroke injury while the tMCAO model represents ischemic stroke with reperfusion. ## Imaging mini-stroke at single-vessel level Taking advantage of the high spatial resolution of PAI, especially PAM, researchers have been able to investigate occlusion of a small vessel within the brain. For example,used both PAM and PACT to visualize singlevessel ischemic stroke using a PT stroke model in rats, as shown in. The PT stroke model uses a laser beam to activate a photoactive dye (e.g., Rose Bengal) and cause coagulation restricted to the illuminated region, allowing for a small beam to produce a targeted single-vessel occlusion in vivo. Both PA modalities were used to evaluate the performance of an experimental ischemic stroke treatment. The treatment was a photothermal-activatable liposome carrying tissue plasminogen activator (tPA), which once activated (by laser irradiation) could dissolve a blood clot and induce recanalization. The PA imaging was performed before and after stroke, and after tPA treatment, showing the healthy flow, impaired flow, and restored flow, respectively. This study displayed the feasibility of both PAM and PACT for ischemic stroke treatment monitoring, an exciting application of the imaging technology. In another study,applied high-speed PAM to monitor the occurrence of mini-stroke at the capillary level in mouse models. Using a microelectromechanical system (MEMS) scanning mirror in conjunction with a high pulse repetition rate laser, they were able to image microvasculature at high spatial resolution (∼3 µm lateral and ∼15 µm axial) and high temporal resolution (400 Hz B-scan rate over 3 mm scanning range). They were able to observe the occlusion location and to measure the changes in microvascular blood flow resulting from the photothrombosis. Interestingly, their data demonstrated a vastly reduced blood flow velocity in the parent vessel, with a reversed blood flow in one of the branching daughter vessels, as shown in. ## Multi-scale imaging of ischemic stroke on small-animal models Small model animals, such as mice and rats, are commonly used in PAI technological development and biomedical applications. PAI is well suited for small-animal imaging of ischemic stroke because of the relatively thin skull, which is one of the predominant difficulties of PA and ultrasoundbased brain imaging. Although not negligible, the relatively thin skull can partially mitigate the acoustic aberration and optical scattering [bib_ref] Impacts of the murine skull on high-frequency transcranial photoacoustic brain imaging, Liang [/bib_ref] [bib_ref] 3D Monte Carlo simulation of light distribution in mouse brain in quantitative..., Tang [/bib_ref]. This is beneficial for achieving high imaging quality by OR-PAM [bib_ref] Optical-resolution photoacoustic microscopy of ischemic stroke, Hu [/bib_ref] [bib_ref] Highspeed label-free functional photoacoustic microscopy of mouse brain in action, Yao [/bib_ref] , AR-PAM [bib_ref] Noninvasive, in vivo imaging of the mouse brain using photoacoustic microscopy, Stein [/bib_ref] [bib_ref] Model-based optical and acoustical compensation for photoacoustic tomography of heterogeneous mediums, Park [/bib_ref] , and PACT [bib_ref] High-speed three-dimensional photoacoustic computed tomography for preclinical research and clinical translation, Lin [/bib_ref] of mouse and rat brains. OR-PAM can produce the high-resolution (a few micrometers) images of ischemic stroke, such as. However, OR-PAM often requires removing the scalp, thinning the skull, or sometimes even installing a cranial window, which makes the modality invasive and more difficult for longitudinal studies of ischemic stroke on the same animals. OR-PAM also has superficial penetration depth (∼1 mm), since it relies primarily on the tight focusing of ballistic photons, which is undesirable for ischemic stroke studies because the injured tissue is generally not constrained to the superficial regions of the brain. On the other hand, AR-PAM utilizes diffuse photons while maintaining relatively tight acoustic focusing. This allows for improved penetration depth to ∼3-5 mm at the cost of spatial resolution (tens of micrometers) relative to OR-PAM.shows representative images of pre and post ischemic stroke utilizing and AR-PAM system. Of the three major implementations of PAI, PACT is the most promising for studying ischemic stroke on the whole-brain level in small-animal models, with a spatial resolution of a few hundred micrometers. Both [bib_ref] In vivo photoacoustic imaging dynamically monitors the structural and functional changes of..., Lv [/bib_ref] and we used PACT to visualize PT stroke in mice both structurally and functionally. Many small-animal PAI studies have allowed better understanding of the neuroprotective processes and hemodynamics that occur in response to ischemic stroke [bib_ref] Assessment of neurovascular dynamics during transient ischemic attack by the novel integration..., Liu [/bib_ref] [bib_ref] Functional and oxygen-metabolic photoacoustic microscopy of the awake mouse brain, Cao [/bib_ref] , and the ability to assess potential treatments for ischemic stroke. ## Deep imaging of ischemic stroke on large-animal models Photoacoustic imaging (PAI) has been successfully applied for both ex vivo and in vivo studies of ischemic stroke on large animal models. Large animal models better simulate the conditions of human brains, namely the size of the brain and thickness of the skull. Therefore, the application of PAI of ischemic stroke on large animals is a prerequisite to human stroke research. Hariri et al. tested the feasibility of transfontanelle PAI using an ex vivo sheep brain, with an opening in the skull mimicking the fontanelle [bib_ref] Functional photoacoustic tomography for neonatal brain imaging: Developments and challenges, Hariri [/bib_ref]. The results showed that functional PAI through the fontanelle of a neonate brain is possible. Later, showed the feasibility of neonatal functional PAI using an in vivo pig model . In a follow-up study, as shown in, [bib_ref] Validation of noninvasive photoacoustic measurements of sagittal sinus oxyhemoglobin saturation in hypoxic..., Kang [/bib_ref] used a linear-array PAI of single vessel occlusions. (A) PAM (top row) and PACT (bottom row) images of a blood vessel before PT stroke, after PT stroke, and after experimental tPA treatment, acquired at 720nm. (B) PAM images of a blood vessel before and after PT stroke acquired at 532nm with corresponding blood flow velocity calculations on the left. P indicates the parent vessel of the clot, with D1 and D2 indicating two branching daughter vessels. transducer to image neonatal pigs within 2 h following a PT stroke surgery. They found significant changes in both blood perfusion and oxygenation between injured and non-injured regions of the brain, demonstrating that PAI can be used for detecting perinatal ischemic stroke. Perinatal ischemic stroke occurs in 1 in 2300 term infants, which is 17 times more common than later in childhood or beyond [bib_ref] Perinatal ischemic stroke, Nelson [/bib_ref] , and is a particularly compelling and clinically relevant application for PAI. The relatively small head size, the thinner bone, and the soft fontanelles of infants all help PAI achieve higher spatial resolution and larger penetration depth. These recent advances in PAI of ischemic stroke in large-animal models show a promising trend of technology advancement. Nevertheless, far fewer large-animal studies exist than small-animal studies so far, mostly due to the experimental difficulty and technology limitation. As a prerequisite to human stroke studies using PAI, more large-animal studies must be performed for extensive validation. ## Clinical translation of photoacoustic imaging for human brain imaging Ultimately PAI can be developed for use in human brain imaging, but there are several technical challenges for example, strong acoustic aberration of the thick human skull (∼6-8 mm) and the strong optical attenuation. There have been no published reports utilizing PAI to monitor ischemic stroke in humans yet. [bib_ref] Acoustic impact of the human skull on transcranial photoacoustic imaging, Liang [/bib_ref] simulated the acoustic distortion caused by the human skull using different implementations of PACT. They found that the reflections, refractions, and mode-conversions of the human skull can introduce severe signal distortion and reconstruction artifacts in PACT, greatly reducing the resolution and quantitative accuracy of the image reconstructions. As shown in [fig_ref] FIGURE 5 MRI: MRA, and PACT images of the brain of a hemicraniectomy patient [/fig_ref] , Na et al. imaged hemicraniectomy patients using MRA and PACT [bib_ref] Massively parallel functional photoacoustic computed tomography of the human brain, Na [/bib_ref]. The results showed that the MRA and PACT can provide Multi-scale PAI for small-animal stroke research. (A) OR-PAM oxygen saturation of hemoglobin (sO 2 ) image of mouse brain before stroke, 330s and 411s after stroke (skull removed). (B) AR-PAM image of mouse brain before and after stroke (skull removed) [bib_ref] In vivo imaging of hemodynamics and oxygen metabolism in acute focal cerebral..., Deng [/bib_ref]. (C) (Top) Baseline and post stroke MRI and corresponding PACT images of mouse brain. Infarct region in MRI corresponds with increased deoxygenated blood in PACT (scalp and skull intact) [bib_ref] In vivo photoacoustic imaging dynamically monitors the structural and functional changes of..., Lv [/bib_ref]. (Bottom) Corresponding triphenyltetrazolium chloride (TTC) stained coronal slices and PACT sO 2 coronal images of mouse brain. Deoxygenated region in sO 2 image corresponds with infarct region identified by TTC (scalp and skull intact) (Menozzi et al.). well-matched structural images of the human brain vasculature at depths up to 10 mm under the cortical surface. Functional MRI (fMRI) and functional PACT were also performed to monitor the brain activities in the motor and language regions in response to certain activities, such as finger tapping and passive listening. Although it has not been applied for monitoring ischemic stroke on humans yet, this recent study demonstrated that high-quality human brain imaging can be achieved with PAI technology. ## Technological challenges and solutions for photoacoustic imaging All the above achievements in PAI of ischemic stroke have reflected the strong momentum of this exciting technology and its great potential in preclinical stroke research. However, to maximize its impact, there are still major technical challenges Frontiers in Neuroscience 06 frontiersin.org Structural and functional PACT images of a neonatal pig in vivo with PT stroke . in PAI that need to be addressed through innovative solutions in laser engineering, data science, mathematics, and chemistry. These challenges can be broadly divided into two categories: optical challenges and acoustic challenges, as summarized in . ## Optical challenges Limited penetration depth of light due to strong optical attenuation Skin, bone, and brain tissue are three of the most scattering biological tissues in the 500-900 nm optical wavelength range [bib_ref] Optical properties of biological tissues: A review, Jacques [/bib_ref] , making light penetration deep into the brain (>1 cm depth) a difficult task. To avoid tissue damage, there exist limitations on the maximum optical energy density that can be delivered to the surface of the skin, usually guided by the laser safety standard by the American National Standards Institute (ANSI). Thus, simply increasing the excitation light energy is not a preferred option for PAI to improve its penetration depth. A few engineering solutions have been developed to improve optical penetration depth while maintaining the laser safety. Major technological challenges of PAI in ischemic stroke research. Technological challenges are shown in red boxes with existing and potential solutions shown in green boxes. the use of internal illumination in PAI. Laser pulses delivered within the body cavity, generally via catheter or miniaturized probe, bypass the highly absorbing layers of skin and bone, and thus can often be closer to the target being imaged. Internal illumination PAI has shown great promise for improving optical penetration depth in the animal brains through phantom experiments and in vivo imaging, by delivering light through the oral cavity [bib_ref] In vivo deep brain imaging of rats using oral-cavity illuminated photoacoustic computed..., Lin [/bib_ref] [bib_ref] Internal-illumination photoacoustic computed tomography, Li [/bib_ref]. In a different approach, [bib_ref] Photoacoustic tomography through a whole adult human skull with a photon recycler, Nie [/bib_ref] improved optical penetration depth by increasing the efficiency of light delivery using a photon recycler. The photon recycler was a highly reflective surface which re-delivered photons that had been scattered out of the skin surface back into the brain. Wavefront engineering is another potential solution for improving optical penetration depth while maintaining high resolution. This allows focusing light into a deeper region of a scattering medium [bib_ref] Wave front engineering from an array of thin aperture antennas, Kang [/bib_ref] [bib_ref] Wavefront shaping and its application to enhance photoacoustic imaging, Yu [/bib_ref]. Optimizing the excitation wavelength can further improve optical penetration depth in PAI. For example, microwave excitation to image a whole Rhesus monkey brain [bib_ref] Rhesus monkey brain imaging through intact skull with thermoacoustic tomography, Xu [/bib_ref]. However, it should be noted that changing the excitation wavelength may change the source of contrast to other biomolecules such as water. ## Limited optical contrast of brain functions In the NIR excitation domain, lipids, hemoglobin, melanin, and water are the primary endogenous sources of PAI contrast [bib_ref] Photoacoustic microscopy, Yao [/bib_ref]. While hemoglobin can be used to probe the brain's hemodynamics, these exogenous contrasts fall short of providing more information about the brain's functional and molecular status, such as the neuroactivities and inflammation. Ultimately, ischemic stroke causes critical damage to the neurons in the infarct region. To precisely quantify the effect and severity of stroke, the health and function of the neurons must be evaluated, which cannot be achieved by measuring blood oxygenation alone. Many exogenous PAI contrast agents have been developed in order to resolve this issue [bib_ref] Semiconducting polymer nanoparticles as photoacoustic molecular imaging probes in living mice, Pu [/bib_ref] [bib_ref] Sensitivity of photoacoustic microscopy, Yao [/bib_ref] [bib_ref] Photoacoustic imaging: Contrast agents and their biomedical applications, Fu [/bib_ref] [bib_ref] Recent advances in photoacoustic contrast agents for in vivo imaging, Upputuri [/bib_ref]. Exogenous contrast agents often have stronger optical absorption and thus provide stronger PA signals and larger penetration, as well as target certain physiological events, particularly neuronal activities. In an exciting example, [bib_ref] A near-infrared genetically encoded calcium indicator for in vivo imaging, Shemetov [/bib_ref] , using an engineered NIR genetically encoded calcium indicator, simultaneously imaged hemodynamics and neuronal activity in mouse brains with PAM and fluorescence microscopy. Such targeted contrast provides a powerful new pathway for photoacoustic signal generation in the brain and can also be applied to provide more comprehensive information of the brain impaired by ischemic stroke . ## Inaccurate quantitative analysis due to wavelength-dependent light attenuation Functional PAI often requires multiple excitation wavelengths to spectrally unmix the concentrations of several absorbing molecules that collectively contribute to the PA signals. Assuming a total of n molecules contributed to the PA signals, linear unmixing methods require at least a total of n images to be taken at different wavelengths, in order to solve a system of linearly independent equations with n unknowns. However, this commonly used linear model assumes an important oversimplification: the optical fluence at a given depth in a target strongly depends on the wavelength of light used to illuminate the target [bib_ref] The challenges for quantitative photoacoustic imaging, Cox [/bib_ref]. This is inherently due to the wavelength-dependent optical absorption and scattering coefficients of biological tissue. Therefore, linear unmixing cannot produce accurate quantification of molecular concentrations in deep tissues. Alternatively, measurements of relative quantities can still be made and produce informative results. For example, Matsumoto et al. distinguished between arteries and veins by comparing deoxygenated blood in veins to nearby oxygenated blood in arteries [bib_ref] Visualising peripheral arterioles and venules through high-resolution and large-area photoacoustic imaging, Matsumoto [/bib_ref]. Although wavelength-dependent optical fluence is a pervasive problem for functional PAI, more accurate solutions do exist. Using PAI as a functionally complementary modality to MRI or computed tomography (CT) could allow for fluencecorrection by simulating photon transport in tissues, using the anatomical information from MRI and CT [bib_ref] MCML-monte carlo modeling of light transport in multi-layered tissues, Wang [/bib_ref] [bib_ref] Coupling 3D Monte Carlo light transport in optically heterogeneous tissues to photoacoustic..., Jacques [/bib_ref]. Furthermore, iterative (or model-based) reconstruction methods, such as the algorithm proposed by [bib_ref] Model-based optical and acoustical compensation for photoacoustic tomography of heterogeneous mediums, Park [/bib_ref] , can map both the optical fluence and speed of sound in the heterogeneous media [bib_ref] Model-based optical and acoustical compensation for photoacoustic tomography of heterogeneous mediums, Park [/bib_ref] , allowing for more accurate image reconstruction and spectral unmixing. ## Acoustic challenges Strong acoustic impact of the skull Reflection, refraction, and mode conversion of acoustic waves occur at boundaries of brain tissue and skull with drastically different acoustic impedance, resulting in signal loss of upward of 75% [bib_ref] Acoustic impact of the human skull on transcranial photoacoustic imaging, Liang [/bib_ref] , which is related to the density and speed of sound of the medium [bib_ref] Basic physics of ultrasound imaging, Aldrich [/bib_ref]. In ultrasound imaging, reflection is desirable during the forward propagation of the acoustic waves because it provides the image contrast. In photoacoustic imaging, however, reflection of acoustic waves is normally undesirable; the reflections reduce signal strength and introduce artifacts in the image reconstruction [bib_ref] Photoacoustic-guided focused ultrasound (PAFUSion) for identifying reflection artifacts in photoacoustic imaging, Singh [/bib_ref] [bib_ref] Reflection artifact identification in photoacoustic imaging using multi-wavelength excitation, Nguyen [/bib_ref] [bib_ref] Impacts of the murine skull on high-frequency transcranial photoacoustic brain imaging, Liang [/bib_ref]. The refraction of photoacoustic waves propagating from the target to the transducer result in distortions in the reconstructed image, primarily because most reconstruction algorithms, such as the ubiquitous delay and sum (DAS) method, assume linear ultrasound transmission in a homogeneous medium [bib_ref] Universal back-projection algorithm for photoacoustic computed tomography, Xu [/bib_ref]. Furthermore, ultrasound propagation through the skull results in significant shear waves due to mode conversion, leading to further distortions in image reconstruction [bib_ref] Longitudinal and shear mode ultrasound propagation in human skull bone, White [/bib_ref]. The most practical method for reducing the impact of the skull in PAI is to use low-frequency detection. Lower ultrasonic frequencies show less attenuation and thus greater penetration than higher frequencies, allowing for more signal to bypass the skull with less distortion. However, the use of lower frequencies also results in a reduction of spatial resolution. Co-registered anatomical information, such as from MRI or CT, could similarly improve PA reconstructions by providing segmented skull geometry [bib_ref] Photoacoustic tomography in absorbing acoustic media using time reversal, Treeby [/bib_ref]. This would allow the reconstruction (e.g., time reversal method) to correct for acoustic absorption, refraction, and speed-of-sound inhomogeneities between tissue layers. Skull clearing cranial windows have also shown promise in smallanimal PAI to reduce both the optical and ultrasonic attenuation, improving the overall vascular brain images [bib_ref] Chronic cranial window for photoacoustic imaging: A mini review, Wang [/bib_ref]. However, these are invasive procedures, limiting their use to preclinical studies. ## Incomplete target structure due to limited detection view An exact reconstruction of the initial pressure distribution in PAI requires a detector that can completely surround the object being imaged, or provide a detection view of 4π. In practice, this is very challenging to implement, leading to the well-known limited-view problem in PAI that manifests as incomplete structure and reconstruction artifacts. The limited-view problem is augmented in brain imaging due to the size and shape of the head, adding distortions and artifacts to the image reconstruction. Deep-learning and modelbased approaches have been applied to reduce the limited-view artifacts [bib_ref] Limited-view and sparse photoacoustic tomography for neuroimaging with deep learning, Guan [/bib_ref] [bib_ref] A new deep learning network for mitigating limited-view and under-sampling artifacts in..., Zhang [/bib_ref] [bib_ref] Compensating for visibility artefacts in photoacoustic imaging with a deep learning approach..., Godefroy [/bib_ref] [bib_ref] Deep learning for biomedical photoacoustic imaging: A review, Gröhl [/bib_ref]. [bib_ref] A generative adversarial network for artifact removal in photoacoustic computed tomography with..., Vu [/bib_ref] demonstrated the use of a generative adversarial network to reduce the limited-view artifacts using simulated, phantom, and in vivo data. This approach allows for the reduction of the limited-view artifacts without changing the imaging system design. Even so, enlarging the detection geometry is the most effective way to mitigate the limited-view problem. For example, a 2D ring detection geometry or a hemispherical detection geometry are reported to produce superior wholebrain imaging, due to larger detection angles [bib_ref] Whole-body photoacoustic imaging techniques for preclinical small animal studies, Kye [/bib_ref]. However, using these detection arrays can introduce a higher cost of the system and also require a more complicated experimental setup. Alternatively, multiple acquisitions could be acquired at varying angles and later integrated during the image reconstruction, which, however, may result in a decreased temporal resolution and thus increased susceptibility to motion artifacts. This approach was used by Zhang et al. to image a whole mouse brain using a linear-array, in which 39 limitedview images were combined to create an enhanced image . A new approach for bypassing the limitedview problem is to track sparsely distributed highly absorbing exogenous contrast as point targets as they flow through the blood vessels, though this method results in a longer scanning time as a large number of frames are needed to accumulate sufficient contrast agents in the field of view . ## Limited spatial resolution due to the band-limited ultrasound detection Photoacoustic signals have an inherently broad bandwidth and are highly related to the target sizes, ranging from below 1 MHz to over tens of MHz [bib_ref] A simulation study on photoacoustic signals from red blood cells, Saha [/bib_ref]. In most PAI systems, piezo-based ultrasound transducers used in traditional ultrasound imaging are adapted as the acoustic sensor, usually due to the broad availability and convenience. However, piezobased ultrasound transducers are not optimized to sense photoacoustic signals as the elements often have both narrow acceptance angles and narrow bandwidths. Photoacoustic imaging systems require ultrasound sensing that is more suited to the broad-band nature of a photoacoustic signal [bib_ref] Overview of ultrasound detection technologies for photoacoustic imaging, Manwar [/bib_ref]. An emerging solution is optical acoustic sensors for PAI, including amplitude-modulation-based and interferometerbased sensors. A number of studies have shown the optical acoustic sensors can usually achieve higher detection sensitives per unit sensor area, broader detection bandwidths, and larger acceptance angles, compared with traditional piezoelectric sensors [bib_ref] Characterization of a broadband all-optical ultrasound transducer-from optical and acoustical properties to..., Hou [/bib_ref] [bib_ref] A transparent broadband ultrasonic detector based on an optical micro-ring resonator for..., Li [/bib_ref] [bib_ref] Optical detection of ultrasound in photoacoustic imaging, Dong [/bib_ref]. Although optical sensors show great promise, they are mostly used as a single-element detector for either PAM or PACT , which has substantially limited imaging speed. Arrays of these sensors are necessary for fast PACT, but they are often difficult to fabricate with high uniformity, particularly of the resonant frequencies [bib_ref] A review of transparent sensors for photoacoustic imaging applications, Ren [/bib_ref]. Furthermore, resonance-based optical sensors tend to be unstable due to thermally induced resonant-frequency shifts. Improving the stability and fabrication process for optical sensors will be an enormous step forward for the field. # Conclusion and discussion The blood-sensitive nature of PAI makes it well suited for investigating ischemic stroke. By combining rich optical contrast with deep ultrasonic detection, PAI can be a safe alternative or complement to existing imaging modalities for studying vasculature. Furthermore, with high spatial scalability and tunable optical contrast, PAI allows for both microscopic imaging at single blood vessel level and macroscopic imaging at tissue level, providing ample metrics to monitoring ischemic stroke disease progress in timely fashion. PAI is already a proven tool in preclinical ischemic stroke studies, particularly in small-animal stroke models. Both PAM and PACT have been used to investigate the hemodynamic processes in ischemic stroke (major publications shown in [fig_ref] TABLE 1: Summary of major publications on ischemic stroke using PAI in the past... [/fig_ref]. The ability to monitor the vasculature and hemodynamics in small animals both non-invasively and longitudinally allows precise evaluation of the safety and efficacy of new treatments. As PAI technology continues to improve, more groundbreaking discoveries of the pathophysiological processes in ischemic stroke are expected in the near future. The advancement of PAI technology in the past decade has allowed the technology to be more widely applied in imaging ischemic stroke. PAM technology has seen improvement in imaging speed, field-of-view, resolution, and sensitivity . These improvements have come from the application of novel high-speed multi-spectra light sources [bib_ref] High-speed wide-field photoacoustic microscopy using a cylindrically focused transparent high-frequency ultrasound transducer, Chen [/bib_ref] , deep learning imaging enhancement [bib_ref] Sounding out the hidden data: A concise review of deep learning in..., Dispirito [/bib_ref] , innovative scanning configurations [bib_ref] Photoacoustic microscopy: Principles and biomedical applications, Liu [/bib_ref] , and exogenous contrast agents [bib_ref] Molecular photoacoustic contrast agents: Design principles & applications, Borg [/bib_ref]. These developments have allowed for better quantification and visualization of the hemodynamics of ischemic stroke. PACT technology has also improved in image reconstruction algorithms [bib_ref] A survey of computational frameworks for solving the acoustic inverse problem in..., Poudel [/bib_ref] , depth of penetration, imaging speed, and resolution [bib_ref] Spatial resolution in photoacoustic computed tomography, Tian [/bib_ref]. These advances have culminated in recent years to the use of PACT in brain imaging in large animal models and, for the first time, in humans. As PAI technology continues to develop in the coming decade, PAI-enabled advances in the understanding and treatment of ischemic stroke are expected to follow. Although PAI has many existing clinical applications, clinical use of the technology for ischemic stroke treatment guidance and reperfusion assessment is yet to be explored. Robust solutions to the technological challenges presented in human brain imaging are a prerequisite for establishing PAI as a practical clinical tool. Promising technological advances have already been reported in attempts to solve the current problems, including the development of PA contrast agents and optical ultrasound sensors, the use of model-based image reconstructions and deep learning image enhancement, and innovative detection and illumination schemes. More rigorous evaluations of the PAI technologies on large animal stroke models are necessary before moving on to human testing. [fig] FIGURE 5 MRI: MRA, and PACT images of the brain of a hemicraniectomy patient(Na et al., 2022). (Left) Representative transverse MRI slice with fMRI data overlaid. (Middle) PACT vasculature of the brain. (Right) Corresponding MRA vasculature of the brain. Scalp vessel (Vs), cortical vessel (Vc), and superficial temporal arteries (STA) were used for comparison between MRA and PACT. [/fig] [table] TABLE 1: Summary of major publications on ischemic stroke using PAI in the past decade. [/table] [table] Table 2: summarizes reported illumination techniques for improving penetration depth in PAI of brains. One solution is [/table]

Phosphoproteomic analysis of Methanohalophilus portucalensis FDF1T identified the role of protein phosphorylation in methanogenesis and osmoregulation Methanogens have gained much attention for their metabolic product, methane, which could be an energy substitute but also contributes to the greenhouse effect. One factor that controls methane emission, reversible protein phosphorylation, is a crucial signaling switch, and phosphoproteomics has become a powerful tool for large-scale surveying. Here, we conducted the first phosphorylationmediated regulation study in halophilic Methanohalophilus portucalensis FDF1 T , a model strain for studying stress response mechanisms in osmoadaptation. A shotgun approach and MS-based analysis identified 149 unique phosphoproteins. Among them, 26% participated in methanogenesis and osmolytes biosynthesis pathways. Of note, we uncovered that protein phosphorylation might be a crucial factor to modulate the pyrrolysine (Pyl) incorporation and Pyl-mediated methylotrophic methanogenesis. Furthermore, heterologous expression of glycine sarcosine N-methyltransferase (GSMT) mutant derivatives in the osmosensitive Escherichia coli MKH13 revealed that the nonphosphorylated T68A mutant resulted in increased salt tolerance. In contrast, mimic phosphorylated mutant T68D proved defective in both enzymatic activity and salinity tolerance for growth. Our study provides new insights into phosphorylation modification as a crucial role of both methanogenesis and osmoadaptation in methanoarchaea, promoting biogas production or reducing future methane emission in response to global warming and climate change.Methane is considered to be an energy substitute for petroleum 1,2 , but the emission of biologically produced methane is also a critical factor in the greenhouse effect and results in extreme climate events 3,4 . Methanogenic archaea are exceptional in the unusual type of metabolism that they exhibit, having the ability to gain energy from reducing CO, CO 2 , formate, methanol, methylamines, or acetate as energy and carbon sources for growth 5,6 . Thus, methanogens have received much attention because they play a pivotal role in both the recycling of carbon compounds to useful resources and the maintenance of the global carbon flux on Earth 3,6 . This dual role underscores the importance of regulating both the methanogenesis process and the stress response to environmental changes for methane emission. Methanogens can utilize three types of methanogenic pathways for energy conservation: CO 2 -reduction, aceticlastic reactions, and methyl-group reduction (methylotrophic) pathways [bib_ref] Metabolic, phylogenetic, and ecological diversity of the methanogenic archaea, Liu [/bib_ref]. The order Methanosarcinales has the widest substrate range for methanogenesis. They are widely distributed in marine and freshwater sediments, anaerobic sewage digestors, and animal gastrointestinal tracts [bib_ref] Metabolic, phylogenetic, and ecological diversity of the methanogenic archaea, Liu [/bib_ref] [bib_ref] Physiology and biochemistry of the methane-producing Archaea, Hedderich [/bib_ref]. The methanogen used in this study, Methanohalophilus portucalensis FDF1 T , belongs to the Methanosarcinales order and is cultivated from a naturally hypersaline environment [bib_ref] Isolation and characterization of Methanohalophilus portucalensis sp-nov and DNA reassociation study of..., Boone [/bib_ref]. The strain FDF1 T utilizes only methanol, monomethylamine (MMA), dimethylamine (DMA), and trimethylamine (TMA) as carbon and energy sources for growth through the methylotrophic methanogenesis pathway. Intriguingly, a non-canonical amino acid, pyrrolysine (Pyl), can be incorporated by an in-frame amber codon (UAG) of three distinct methylamine methyltransferases, TMA methyltransferase (mttB), DMA methyltransferase (mtbB), and MMA methyltransferase (mtmB), to initiate methanogenesis from methylamines [bib_ref] Comparative genomics highlights the unique biology of Methanomassiliicoccales, a Thermoplasmatales-related seventh order..., Borrel [/bib_ref] [bib_ref] Functional context, biosynthesis, and genetic encoding of pyrrolysine, Gaston [/bib_ref]. The utilization of Pyl is restricted to methyltransferases or other Pyl-containing proteins existing only in a few methanogenic archaea and bacteria. The importance of Pyl in methylamine-dependent methanogenesis was shown when deletion of the amber suppressor tRNA Pyl (pylT) gene in Methanosarcina acetivorans disabled methane production from precursor MMA, DMA, or TMA 12 . Despite the functional roles of the critical components involved in the methanogenesis pathway that have been reported, there is still surprisingly little known about their regulatory networks in methanogens. Importantly, methane produced by halophilic methanogens contributes significantly to the carbon mineralization in marine and hypersaline environments where large amounts of the greenhouse gas methane are stored [bib_ref] Methane hydrate -a major reservoir of carbon in the shallow geosphere, Kvenvolden [/bib_ref]. Furthermore, TMA in these habitats is an important methane precursor for the methylotrophic methanogens, since TMA precursors are constantly provided by the degradation of osmolytes, like glycine betaine (betaine), for osmoregulation to cope with high external salinity [bib_ref] Methanogenesis from methylated amines in a hypersaline algal mat, King [/bib_ref] [bib_ref] Metabolism of trimethylamine, choline, and glycine betaine by sulfate-reducing and methanogenic bacteria..., King [/bib_ref]. Betaine is a type of quaternary ammonium compound that can equilibrate unbalanced osmolarity and is a common osmoprotectant in prokarya and eukarya [bib_ref] Glycinebetaine protects plants against abiotic stress: mechanisms and biotechnological applications, Chen [/bib_ref] [bib_ref] Glycinebetaine: an effective protectant against abiotic stress in plants, Chen [/bib_ref] [bib_ref] Osmoadaptation in bacteria and archaea: common principles and differences, Roesser [/bib_ref] [bib_ref] Bacterial osmoadaptation: the role of osmolytes in bacterial stress and virulence, Sleator [/bib_ref]. The model methanogen M. portucalensis FDF1 T has been reported to possess the capability for de novo synthesis of betaine through three-step methylation from precursor glycine via glycine sarcosine N-methyltransferase (GSMT) and sarcosine dimethylglycine N-methyltransferase (SDMT) [bib_ref] Effects of substrate and potassium on the betaine-synthesizing enzyme glycine sarcosine dimethylglycine..., Lai [/bib_ref] [bib_ref] Characterization and regulation of the osmolyte betaine synthesizing enzymes GSMT and SDMT..., Lai [/bib_ref] [bib_ref] Characterization of osmolyte betaine synthesizing sarcosine dimethylglycine N-methyltransferase from Methanohalophilus portucalensis, Chen [/bib_ref]. It is known that intracellular salt concentration regulates the expression of GSMT and SDMT in strain FDF1 T , and the amount of monovalent ions modulates the activity of rate-limited enzyme GSMT [bib_ref] Characterization and regulation of the osmolyte betaine synthesizing enzymes GSMT and SDMT..., Lai [/bib_ref]. Furthermore, drought and salt tolerance in the arabidopsis model also increased in response to heterogeneously expressed GSMT and SDMT from strain FDF1 . Despite these studies on key regulatory factors contributing to osmoadapation, there is still little evidence explaining how strain FDF1 T has an immediate on-off switch in the betaine synthesis process to respond to environmental changes. Reversible protein phosphorylation is the most common cellular mechanism to regulate many physiological and adaptational processes. The first identified methanogen phosphoprotein with a known function was the methyltransferase activation protein from Methanosarcina barkeri, which converts methanol to methane via the methylotrophic methanogenesis pathway [bib_ref] Purification and properties of an enzyme involved in the ATP-dependent activation of..., Daas [/bib_ref]. Furthermore, the archaeal two-component system was found to be involved in the regulation of methanogenesis in Methanosaeta harundinacea [bib_ref] Characterization of an archaeal two-component system that regulates methanogenesis in Methanosaeta harundinacea, Li [/bib_ref]. As genome sequences continue to accumulate, it is apparent that methanoarchaea also possess protein kinases and phosphatases, but only a few of them and their protein substrates have been defined by basic biochemical and genetic approaches [bib_ref] Archaeal protein kinases and protein phosphatases: insights from genomics and biochemistry, Kennelly [/bib_ref] [bib_ref] The serine, threonine, and/or tyrosine-specific protein kinases and protein phosphatases of prokaryotic..., Shi [/bib_ref] [bib_ref] Protein Ser/Thr/Tyr Phosphorylation in the Archaea, Kennelly [/bib_ref]. Therefore, this report will be of particular use in delineating the physiological processes for methanogenesis and salt stress tolerance by displaying the global phosphorylation network from M. portucalensis FDF1 T . In this study, we provided a genome-wide, shotgun and phosphorylation site-specific investigation in a halophilic methanogen through MS-based systematic phosphoproteomic analysis. Both Ser/Thr/Tyr and His/ Asp phosphorylation-based signaling systems were involved in diverse biological processes, especially in the single-carbon energy metabolism for methane production and osmolyte biosynthetic pathways. Although a previous study in the E. coli system had uncovered the ability to overcome hyper-salinity stress via expression of the osmolyte betaine synthesizing enzymes, GSMT and SDMT, from strain FDF1 T 21 , we further undertook a phosphosite mutagenesis analysis of GSMT (MPF_0823) to clarify its enzymatic activity and how its osmoregulatory function could be modulated by phosphorylation modification. # Results Establishment of the phosphoproteome from M. portucalensis FDF1 T . The phosphorylation-mediated cellular signaling network in methylotrophic halophilic M. portucalensis FDF1 T was acquired through high accuracy LC-MS/MS analysis to derive a global and site-specific phosphoproteomic data set. In order to analyze a comprehensive phosphoproteome, protein extracts prepared from mid-exponential phase cultures were subjected to a combination of gel-free and gel-based approaches. In addition, TiO 2 -based HAMMOC was applied for efficient phosphopeptide enrichment [bib_ref] Phosphopeptide enrichment by aliphatic hydroxy acid-modified metal oxide chromatography for nano-LC-MS/ MS..., Sugiyama [/bib_ref] [bib_ref] Protocol for micro-purification, enrichment, pre-fractionation and storage of peptides for proteomics using..., Rappsilber [/bib_ref] , and then analyzed in duplicates, as summarized in [fig_ref] Figure 1: A representative MS/MS spectrum and classification of the identified phosphoproteins in M [/fig_ref]. In total, we identified 308 unique phosphopeptides originating from 149 phosphoproteins with high confidence and a FDR of less than 1.0%. The distribution of class I phosphosites [bib_ref] Global, in vivo, and site-specific phosphorylation dynamics in signaling networks, Olsen [/bib_ref] on Ser, Thr, Tyr, Asp, and His with a probability higher than 75.0% is listed in Supplementary Table 1. Following the typical protocols to prepare phosphopeptides, 11.8% of the phosphosites we identified were His and Asp residues. These phosphosites are thought to be mediated through two-component systems. [fig_ref] Figure 1: A representative MS/MS spectrum and classification of the identified phosphoproteins in M [/fig_ref] exemplifies the manually annotated MS/MS spectra (full results provided in [fig_ref] Figure 2: Schematic illustration of phosphorylation events in methanogenesis from TMA metabolic pathways in... [/fig_ref] Classification of Methanoarchaeal Phosphoproteins. The identified phosphoproteins in M. portucalensis FDF1 T were categorized by cellular localization and biological function to gain insight into their functional roles [fig_ref] Figure 1: A representative MS/MS spectrum and classification of the identified phosphoproteins in M [/fig_ref]. Various data mining techniques using GO annotations provided useful information on the different classes, as summarized in Supplementary Table 2. Of the 149 identified phosphoproteins, 146 phosphoproteins were successfully categorized into different cellular compartments, including cytoplasm (53%), membrane (19%), ribosome (17%), nucleoid (9%), proteasome (1%), and exosome (1%) [fig_ref] Figure 1: A representative MS/MS spectrum and classification of the identified phosphoproteins in M [/fig_ref] and Supplementary . As for biological function, 88.7% of the identified phosphoproteins could be assigned to 14 functional classes [fig_ref] Figure 1: A representative MS/MS spectrum and classification of the identified phosphoproteins in M [/fig_ref] and Supplementary . Among them, many phosphoproteins were involved in pathways responsible for the control of key physiological processes, such as replication, transcription, translation, DNA repair systems, thermosome, proteasome, chaperone systems, osmoadaptation, and S-layer protein synthesis. To facilitate the integration of the phosphoproteome data set and uncover the functional relevance of the identified phosphoproteins, they were mapped into their complex physiological pathways as shown in [fig_ref] Figure 3: Manual annotation and MS/MS analysis reveal Pyl incorporation at amber codons and... [/fig_ref]. Methylotropic halophilic methanogen M. portucalensis FDF1 T can metabolize TMA via methyltransferase systems to provide electrons for reducing additional molecules to methane, thereby generating the electrochemical gradient for ATP synthesis [bib_ref] The genome sequence of Methanohalophilus mahii SLP(T) reveals differences in the energy..., Spring [/bib_ref] [bib_ref] The evolution of A-, F-, and V-type ATP synthases and ATPases: reversals..., Cross [/bib_ref]. Notably, 20% of the total identified phosphoproteins were classified in methane production for energy gain categories, such as methyltransferase systems (7%), methanogenesis (8%), and electron transport and the ATP synthase complex (5%) [fig_ref] Figure 1: A representative MS/MS spectrum and classification of the identified phosphoproteins in M [/fig_ref]. Moreover, the proteins involved in osmolyte biosynthesis (6%) were associated with phosphorylation-mediated regulation, which could be essential for M. portucalensis FDF1 T to accumulate osmolytes in order to balance cell turgor under hypersaline conditions [fig_ref] Figure 1: A representative MS/MS spectrum and classification of the identified phosphoproteins in M [/fig_ref]. The novel observation of these unique phosphorylation events advances our understanding of signal The MS/MS spectrum acquired in the Orbitrap mass spectrometer for a threoninephosphorylated peptide (VLDVApTGTGFNSVR), in which one phosphorylated site on Thr-68 was identified from GSMT (MPF_0823). Sequence-informative fragmentation ions are summarized on the peptide sequences and the matched "y" and "b" ions are annotated in red and blue, respectively. Phosphorylation site-specific ions are indicated with "p". Fragment ion signals corresponding to additional neutral loss of NH 3 are indicated by asterisks. The identified phosphoproteins were grouped by (b) cellular locations and (c) biological function based on GO terms. The biological functions related to methanogenesis and osmoadaptation are highlighted as bold blue and orange, respectively. Hypothetical proteins are grouped into the "Unknown" category and the percental distribution is given. Scientific RepoRts | 6:29013 | DOI: 10.1038/srep29013 transduction in methanogenic archaea, especially for single carbon metabolisms in methanogenesis and osmotic adaptation, suggesting their functions in controlling numerous intracellular signaling and regulatory pathways. Phosphorylation in the methylotrophic methanogenesis pathway. In this study, the only carbon and nitrogen source for M. portucalensis FDF1 T growth was TMA. Ten methyltransferases initiating methylotrophic methanogenesis to convert methylamines to methyl groups were phosphorylated (Supplementary . Also phosphorylated were the proteins participating in the last methane-forming step leading up to an overall transfer of the methyl group from TMA to methane, including the methyl-CoM reductase McrAGB complex (MPF_0920, 0921, and 0924) and heterodisulfide reductase HdrD (MPF_0946). In energy-yielding processes generating a proton gradient for driving ATP synthesis, we found that five distinct ATP synthases, two membrane-bound subunits from the F 420 H 2 dehydrogenase complex, and one electron transport RnfC could be phosphorylated . Likewise, we found ten phosphoproteins in carbohydrate metabolic processes such as glycolysis, gluconeogenesis, and the reductive tricarboxylic acid cycle. [fig_ref] Figure 2: Schematic illustration of phosphorylation events in methanogenesis from TMA metabolic pathways in... [/fig_ref] is a schematic drawing of the phosphorylation events, illustrating in greater detail these phosphoproteins participating in TMA utilization for methanogenesis and energy metabolism. These unprecedented findings suggest that protein phosphorylation globally regulates the initiation of methyl transfer reactions for the metabolism of TMA, which produces methane and generates a sufficiently positive redox potential. Pyrrolysine in phosphorylated methylamine methyltransferases. Interestingly, several phosphorylation sites were located in the C-terminal sequence downstream of the amber (UAG) codon, which encodes pyrrolysine (Pyl) during protein synthesis [fig_ref] Figure 3: Manual annotation and MS/MS analysis reveal Pyl incorporation at amber codons and... [/fig_ref] and [fig_ref] Figure 4: The transmethylation activity of MpGSMT was regulated via Thr phosphorylation [/fig_ref]. The peptide with the amber (UAG)-encoded pyrrolysyl-residue incorporated into MttB (MPF_1478) was measured by mass spectrometry as shown in [fig_ref] Figure 3: Manual annotation and MS/MS analysis reveal Pyl incorporation at amber codons and... [/fig_ref] and Supplementary Table 4, demonstrating that Pyl can be naturally synthesized and incorporated into proteins. Furthermore, the pylTSBCD genes for Pyl synthesis are located within a cluster that contains the gene for methylamine-specific methyltransferase MtbA (MPF_0343) [fig_ref] Figure 3: Manual annotation and MS/MS analysis reveal Pyl incorporation at amber codons and... [/fig_ref]. Intriguingly, the pyrrolysyl-tRNA synthetase (MPF_0547/PylS) identified for Pyl-tRNA Pyl formation included multiple phosphosites at Ser-390/Ser-391 and possibly Tyr-384/Thr-385 on the C-terminal tail of the predicted structure [fig_ref] Figure 4: The transmethylation activity of MpGSMT was regulated via Thr phosphorylation [/fig_ref] , which is the core-binding surface for the tRNA Pyl acceptor helix [bib_ref] Pyrrolysyl-tRNA synthetase-tRNA(Pyl) structure reveals the molecular basis of orthogonality, Nozawa [/bib_ref] [bib_ref] Structure of Desulfitobacterium hafniense PylSc, a pyrrolysyl-tRNA synthetase, Lee [/bib_ref] [bib_ref] Crystallographic studies on multiple conformational states of active-site loops in pyrrolysyl-tRNA synthetase, Yanagisawa [/bib_ref]. Taking into account the importance of Pyl being present in the active site of monomethylamine methyltransferase MtmB in M. barkeri 37 , our findings imply that the formation of the Pyl-tRNA Pyl may be regulated by protein phosphorylation, thereby influencing the enzymatic integrity of methylamine methyltransferases. Phosphorylation in salt stress response. Unlike bacteria, most archaea lack rigid outer envelopes, like the peptidoglycan, for salt resistance [bib_ref] Cell wall polymers in Archaea (Archaebacteria), Kandler [/bib_ref]. They instead rely on strategies like accumulating small molecular osmolytes to overcome the high turgor pressure. M. portucalensis FDF1 T synthesizes glycine betaine de novo as a preferred osmolyte for turgor adjustment [bib_ref] Characterization and regulation of the osmolyte betaine synthesizing enzymes GSMT and SDMT..., Lai [/bib_ref]. We found that most proteins participating in the betaine uptake and synthesis pathways were phosphorylated (Supplementary , including the glycine betaine BtaABC transporter [fig_ref] Figure 3: Manual annotation and MS/MS analysis reveal Pyl incorporation at amber codons and... [/fig_ref] and enzymes involved in the methionine transmethylation cycle and betaine biosynthesis [fig_ref] Figure 5: Phosphosite mutation of MpGSMT affects methyltransferase activity and growth rate at elevated... [/fig_ref]. Because GSMT was a protein known to be the rate-limiting enzyme in the production of intermediate substrate sarcosine for further betaine synthesis 21 , we investigated the phosphorylation of GSMT. Previous research had uncovered only four phosphorylated serine residues in the GSMT ortholog in rat hepatocytes, Glycine N-methyltransferase (GNMT) [bib_ref] Identification of phosphorylation sites in glycine N-methyltransferase from rat liver, Luka [/bib_ref]. We identified one of those phosphosites in GSMT as well as nine more. The multiple phosphosites observed on GSMT included six unambiguous phosphosites with a localization probability ≥ 75% (Ser-46, Thr-68, Thr-70, Tyr-169, Ser-178, and Ser-179) and four putative residues scoring < 75% (Ser-74, Asp-170, Asp-174, and Tyr-177) (Supplementary . Mammalian GNMT exhibits strong structural similarity with the highly phosphorylated M. portucalensis GSMT (MpGSMT) , suggesting that there are more phosphorylation sites to be uncovered in GNMT, which may underscore the importance of GNMT to a further extent than the four previously identified serine residues. To better understand how cells maintain osmotic balance via protein phosphorylation, we focused on phosphorylated MpGSMT and examined the roles of the identified phosphorylation sites. The effects of MpGSMT phosphothreonine on methyltransferase activity and salt stress response. Sequence alignment and predicted MpGSMT structure indicate that the first four phosphorylated sites (Ser-46, Thr-68, Thr-70, and Ser-74) are located in the SAM binding region, and the remaining phosphosites (Tyr-169, Asp-170, Asp-174, Tyr-177, Ser-178, and Ser-179) are located in the lid structure for substrate glycine or sarcosine binding [fig_ref] Figure 4: The transmethylation activity of MpGSMT was regulated via Thr phosphorylation [/fig_ref]. Notably, phosphosites Thr-68 and Thr-70 are situated in a highly conserved GxG motif of class I methyltransferases for the interaction with the methionine group of SAM [bib_ref] SAM (dependent) I AM: the S-adenosylmethionine-dependent methyltransferase fold, Martin [/bib_ref] [fig_ref] Figure 4: The transmethylation activity of MpGSMT was regulated via Thr phosphorylation [/fig_ref]. Intriguingly, phosphosite Thr-68 was highly conserved in prokaryotic GSMT, but the residue was replaced with a nonphosphorylatable Cys in eukaryotic GNMT, hinting at the possible regulation of MpGSMT mediated by threonine phosphorylation. Based on energy minimization of the catalytic environment of MpGSMT modeling structures [fig_ref] Figure 4: The transmethylation activity of MpGSMT was regulated via Thr phosphorylation [/fig_ref] , the distance between phenolic hydroxyl Tyr-26 and the sulfur atom on SAM indicate suitable geometry to form a hydrogen bond with the potential to inhibit the enzymatic reaction in the T68D phospho-mimicking mutant, but not in the other two modeling structures (WT and T68A). We therefore examined the transmethylation activities by site-specific mutagenesis to determine the essential phosphorylated-threonine residues to be Thr-68 and Thr-70. While WT-and T68A-MpGSMT retained their GMT and SMT enzymatic activities, transmethylation activities were almost completely abrogated in all dephospho-(T70A and T68A/T70A) and phospho-mimetic isoforms (T68D, T70D, and T68D/T70D) [fig_ref] Figure 5: Phosphosite mutation of MpGSMT affects methyltransferase activity and growth rate at elevated... [/fig_ref]. These results implied that Thr-68, but not Thr-70, could have a regulatory role for M. portucalensis in response to betaine biosynthesis through phosphorylation/de-phosphorylation. Additionally, circular dichroism (CD) analysis showed that the secondary structures of T68A and T68D mutant proteins had no conformational change (Supplementary , indicating that these point mutations did not affect their protein structure. To further examine the biological effects of MpGSMT isoforms on osmoadaptation, we measured the growth of the osmosensitive strain E. coli MKH13 with heterologous expression of Mpgsmt-sdmt (WT) and mutant Mpgsmt-sdmt (T68A-or T68D-GSMT) genes. Media containing 0 mM, 500 mM or 700 mM NaCl were prepared in solid form for salt shock and liquid form for salt adaptation growth tests. The growth patterns of E. coli MKH13 with WT-, T68A-or T68D-GSMT were similar under the non-saline condition [fig_ref] Figure 5: Phosphosite mutation of MpGSMT affects methyltransferase activity and growth rate at elevated... [/fig_ref]. Interestingly, E. coli MKH13 expressing T68A-GSMT had a fast-growth phenotype, adapting faster in 500 mM and 700 mM NaCl, while the strain expressing T68D-GSMT displayed a slow-growth phenotype [fig_ref] Figure 5: Phosphosite mutation of MpGSMT affects methyltransferase activity and growth rate at elevated... [/fig_ref]. In other words, expression of T68A-GSMT in E. coli MKH13 conferred higher tolerance to salt stress than the expression of WT-GSMT [fig_ref] Figure 5: Phosphosite mutation of MpGSMT affects methyltransferase activity and growth rate at elevated... [/fig_ref]. Conversely, the strain expressing T68D-GSMT required a longer lag period (45 hours) to overcome 700 mM NaCl salt stress [fig_ref] Figure 5: Phosphosite mutation of MpGSMT affects methyltransferase activity and growth rate at elevated... [/fig_ref]. Similarly, the salt shock growth tests using solid media showed that the strain expressing T68A-GSMT possessed higher salt stress tolerance as NaCl concentration increased from 0 to 500 mM NaCl [fig_ref] Figure 5: Phosphosite mutation of MpGSMT affects methyltransferase activity and growth rate at elevated... [/fig_ref]. However, none of the constructs could protect the cells against salt shock on solid media with 700 mM NaCl. Taken altogether, our results support the hypothesis that Thr-68 phosphorylation in MpGSMT could regulate the dual catalytic activities of GMT and SMT in GSMT while playing a leading role in osmoprotection. # Discussion Microbial methanogenesis is important in the carbon cycle, impacting climate change or contributing to renewable energy. A fair amount of research attention in methanogenic ecosystems or biosynthetic functions has therefore approached this issue [bib_ref] Metabolic, phylogenetic, and ecological diversity of the methanogenic archaea, Liu [/bib_ref] [bib_ref] Genomic and phenotypic differentiation among Methanosarcina mazei populations from Columbia River sediment, Youngblut [/bib_ref] [bib_ref] Taxonomic, phylogenetic, and ecological diversity of methanogenic Archaea, Garcia [/bib_ref] [bib_ref] Metabolic versatility in methanogens, Costa [/bib_ref]. In this study, we provided additional insights into the unique metabolic pathways and environmental adaptation correlated with functionally meaningful phospho-regulatory events in the halophilic methanogen M. portucalensis FDF1 T . Through further investigation by site-directed mutagenesis, we clarified the role of MpGSMT phosphorylation in regulating methyltransferase activities and salt tolerance. The 149 phosphoproteins identified in this study approximately 7.0% of all the ORFs encoded in M. portucalensis FDF1 T genome (nearly 2131 encoded proteins), compared with 2.5% of the halophilic archaeon Halobacterium salinarum [bib_ref] Ser/Thr/Tyr protein phosphorylation in the archaeon Halobacterium salinarum-a representative of the third..., Aivaliotis [/bib_ref]. Both archaeal phosphoprotemic analyses revealed that phosphorylated proteins were involved in almost every cellular process, though unique physiological traits such as methanogenesis and osmolyte biosynthesis were uncovered in M. portucalensis [fig_ref] Figure 1: A representative MS/MS spectrum and classification of the identified phosphoproteins in M [/fig_ref]. We observed that many phosphoproteins were classified as translation and protein repair [fig_ref] Figure 1: A representative MS/MS spectrum and classification of the identified phosphoproteins in M [/fig_ref] and , possibly promoting protein homeostasis to adapt to fluctuations in external osmotic pressure. This phenomenon is similarly observed in eukaryotes to buffer normal fluctuations in cellular state and maintain cellular homeostasis via phosphorylation-mediated regulation of translation or chaperone proteins [bib_ref] Cellular stress responses: cell survival and cell death, Fulda [/bib_ref]. Methanogens in the Methanosarcinaceae family have been found in a wide variety of anaerobic environments where methane is produced from the widest substrate range; methylamines and methanol are generally used to generate energy [bib_ref] Methanococcoides vulcani sp. nov., a marine methylotrophic methanogen that uses betaine, choline..., L&apos;haridon [/bib_ref]. The enzymes involved in methylotrophic methanogenesis and the coupled electron transfer were phosphorylated [fig_ref] Figure 2: Schematic illustration of phosphorylation events in methanogenesis from TMA metabolic pathways in... [/fig_ref] , suggesting that methylotrophic methanogenesis might be globally regulated by protein phosphorylation. On the other hand, the formation of a multienzyme complex is known to be essential for demethylating substrates. All of the identified phosphosites from MtmC, MtbC, MttC, MtaA, and MtbA were located within the C-terminal domain with a Rossmann fold structure in C subunits or within the N-terminal segment capping the active center in A subunits , which is the interface between an A subunit and the core complex of the BC subunits [bib_ref] Structure of the corrinoid:coenzyme M methyltransferase MtaA from Methanosarcina mazei, Hoeppner [/bib_ref] [bib_ref] Insight into the mechanism of biological methanol activation based on the crystal..., Hagemeier [/bib_ref]. Likewise, some phosphosites on MtaB, MtmB, and MttB were located in the helical layer, which surrounds the TIM barrel structure to interact with the partner C subunit and potentially the A subunit [bib_ref] Insight into the mechanism of biological methanol activation based on the crystal..., Hagemeier [/bib_ref]. Taken together, these findings suggest that phosphorylation may influence the stability of methyltransferase quaternary structural complexes, and in turn alter their enzymatic activities in methylotrophic methanogenesis, but further biological validation is required to confirm the significance of our findings. Notably, phosphorylated MttB was found to possess an in-frame amber codon with Pyl, and the pyrrolysyl-tRNA synthetase PylS (MPF_0547) for Pyl-tRNA Pyl formation and Pyl incorporation was found to be phosphorylated [fig_ref] Figure 3: Manual annotation and MS/MS analysis reveal Pyl incorporation at amber codons and... [/fig_ref]. Building upon the known importance of Pyl incorporation in methanogenesis [bib_ref] A new UAG-encoded residue in the structure of a methanogen methyltransferase, Hao [/bib_ref] [bib_ref] The trimethylamine methyltransferase gene and multiple dimethylamine methyltransferase genes of Methanosarcina barkeri..., Paul [/bib_ref] , our results not only demonstrated that M. portucalensis FDF1 T is a Pyl-utilizing archaea but also suggested the possibility that protein phosphorylation plays an upstream regulatory role in facilitating methylamine metabolism. Osmolyte biosynthesis pathways and protein refolding processes are energy consuming systems and are required to respond immediately to osmotic stresses [bib_ref] Bioenergetic aspects of halophilism, Oren [/bib_ref]. Twelve units of ATP are required to generate one SAM, and three units of SAM are capable of synthesizing one unit of the osmolyte betaine, consuming 36 units of ATP in total [bib_ref] Extreme halophiles synthesize betaine from glycine by methylation, Nyyssola [/bib_ref]. From a bioenergetic point of view, phospho-regulatory mechanisms might provide a more energy-efficient strategy to maintain osmotic equilibrium by rapidly switching protein activity on or off. Results from the enzymatic activities and salinity tolerance assay in E. coli MKH13 [fig_ref] Figure 5: Phosphosite mutation of MpGSMT affects methyltransferase activity and growth rate at elevated... [/fig_ref] demonstrated that the Thr-68-dephosphorylated MpGSMT was constitutively active in adapting to osmotic fluctuations. Furthermore, we present the substrate-assisted catalysis model using MpGSMT_WT, T68A, and T68D mutants to illustrate that the phosphorylation-simulated mutant T68D has the potential to form a hydrogen bond between the hydroxyl group of Y26 and the sulfur atom of SAM [fig_ref] Figure 4: The transmethylation activity of MpGSMT was regulated via Thr phosphorylation [/fig_ref]. Because the position of residue Y26 in the predicted MpGSMT structure is a highly conserved homology of the Y21 residue of GNMT (PDB:1nbh) , the catalytic center of both methyltransferases may be similar [bib_ref] Methyl transfer in glycine N-methyltransferase. A theoretical study, Velichkova [/bib_ref] [bib_ref] Catalytic mechanism of glycine N-methyltransferase, Takata [/bib_ref]. Accordingly, we further propose a transmethylation mechanism for the nonphosphorylated MpGSMT through the electron transferring from Tyr-26 to attack the amine group of glycine, followed by the methyl group attacking the sulfur atom of SAM, which may lead to methyl transfer from SAM to glycine, thus generating methylated glycine (sarcosine) and S-adenosylhomocysteine (SAH) . Conversely, we speculate that phosphorylation on Thr-68 shortens the distance between Y26 and SAM [fig_ref] Figure 4: The transmethylation activity of MpGSMT was regulated via Thr phosphorylation [/fig_ref] and in turn retards the transmethylation reaction, as evidenced by the T68D mutant causing an inactive form of MpGSMT [fig_ref] Figure 5: Phosphosite mutation of MpGSMT affects methyltransferase activity and growth rate at elevated... [/fig_ref]. These results indicate that threonine phosphorylation negatively regulates MpGSMT for energy conservation. It should be noted that GNMT is considered a tumor suppressor of human hepatocellular carcinoma, and the position of the phosphorylated residues in the GNMT tertiary structure is likely to affect the protein's conformation and activity [bib_ref] The multi-functional roles of GNMT in toxicology and cancer, Yen [/bib_ref]. For instance, phosphosite Ser-71 of GNMT may affect the microenvironmental net charge of SAM binding pocket, and Ser-182 could modulate the tetramer switching to the dimer [bib_ref] Identification of phosphorylation sites in glycine N-methyltransferase from rat liver, Luka [/bib_ref]. More surprising is the fact that phosphorylated GNMT significantly increases enzyme activity [bib_ref] Identification of phosphorylation sites in glycine N-methyltransferase from rat liver, Luka [/bib_ref] , in contrast to phosphorylated MpGSMT, which presented a methyltransferase-inactive phenotype [fig_ref] Figure 5: Phosphosite mutation of MpGSMT affects methyltransferase activity and growth rate at elevated... [/fig_ref]. Intriguingly, the phosphosites Ser-74 and Asp-174 identified in MpGSMT corresponding to rat GNMT Ser-71 and Ser-182 were evolutionarily conserved , while the quaternary structure of MpGSMT is regulated via potassium ion concentration but not mediated by protein phosphorylation [bib_ref] Characterization and regulation of the osmolyte betaine synthesizing enzymes GSMT and SDMT..., Lai [/bib_ref]. Furthermore, the highly conserved phosphosite Thr-68 in prokaryotic GSMT was replaced by Cys-65 in eukaryotic GNMT [fig_ref] Figure 4: The transmethylation activity of MpGSMT was regulated via Thr phosphorylation [/fig_ref] , revealing that GNMTs in eukarya possessing constitutive activity might be due to the non-phosphorylatable residue Cys-65. It may therefore be regulated through another mechanism modulating the ratio of SAM/SAH in cells. Collectively, our recent findings revealed a tight correlation between the phospho-regulation and methyltransferase activities of MpGSMT, especially as an energy-efficient strategy in response to osmotic regulation. In conclusion, we present the global phosphorylation-mediated cellular signaling networks in halophilic methanogen M. portucalensis FDF1 T to highlight the importance of phospho-regulation in methanogenesis and the betaine biosynthesis process. In particular, we demonstrated that the FDF1 T is a pyrrolysine-utilizing archaea and hypothesized that methane production via the methyltrophic methanogenesis may be regulated by protein phosphorylation in both protein translation and the methylamine metabolic process. In addition, this is the first study to elucidate the importance of Thr phosphorylation in prokaryotic GSMT, and clarifying that Thr-68 phosphorylation plays a distinct role in enzymatic activity and salt tolerance. This report advances our knowledge of the underlying mechanisms in the modulation of methanogenesis and osmotic adaptation and potentiates improved methane yields or minimized methane emission, both crucial in addressing global warming and climate change. # Methods Growth conditions and protein extraction. The bacterial strains used for cloning and heterologous expression were E. coli DH5α and E. coli BL21(DE3)RIL (Novagen, Madison, WI), respectively. The archaeal strain Methanohalophilus portucalensis strain FDF1 T (= DSM 7471) [bib_ref] Isolation and characterization of Methanohalophilus portucalensis sp-nov and DNA reassociation study of..., Boone [/bib_ref] , was used in this study to derive the phosphoproteome. M. portucalensis FDF1 T were cultured in medium containing 120 g L −1 NaCl and 20 mM trimethylamine as a sole carbon and energy source, as described in the literature [bib_ref] Distribution of compatible solutes in the halophilic methanogenic archaebacteria, Lai [/bib_ref]. The harvested mid-exponential phase (OD 540 at 0.5) cell pellets were resuspended in fresh lysis buffer (25 mM ammonium bicarbonate, PhosSTOP phosphatase inhibitor mixture tablets (Roche), 6 M urea, and 2 M thiourea) and disrupted by sonication on ice. The protein concentration was quantified by the Bradford protein assay (Bio-Rad). ## Phosphopeptide preparation and nanolc-ms/ms analysis. In order to compile a comprehensive phosphoproteome data set, the total protein was treated with trypsin in both gel-based and gel-free processes following procedures described in the literature [bib_ref] Phosphoproteomic analysis reveals the effects of PilF phosphorylation on type IV pilus..., Wu [/bib_ref]. Phosphopeptides from the tryptic peptides were enriched by custom-made HAMMOC tips, which were prepared using 0.5 mg TiO 2 beads (GL Sciences, Tokyo, Japan) packed into 10-μ L C 8 -StageTips, as described previously [bib_ref] Phosphopeptide enrichment by aliphatic hydroxy acid-modified metal oxide chromatography for nano-LC-MS/ MS..., Sugiyama [/bib_ref] [bib_ref] Protocol for micro-purification, enrichment, pre-fractionation and storage of peptides for proteomics using..., Rappsilber [/bib_ref]. The peptide mixtures were analyzed by online nanoflow liquid chromatography tandem mass spectrometry (LC-MS/MS) on a nanoAcquity system (Waters, Milford, MA) coupled to an LTQ-Orbitrap Velos hybrid mass spectrometer (Thermo Scientific) equipped with a PicoView nanospray interface (New Objective). Detailed detection conditions are described in Supplementary Information. strain FDF1 T genome sequence were previously predicted by Glimmer 2.13 60 , GeneMark 2.4, and GeneMark. hmm 2.1 61 and annotated with the RefSeq Microbial Genomes database 62 using BLASTP in standard settings (E-value < 10 −5 , identity > 40%, and matched length > 30%). The detailed search criteria are described in Supplementary Information. The identified phosphoproteins matched to protein sequences in M. portucalensis strain FDF1 T were reported and listed in . The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium 63 via the PRIDE partner repository with the dataset identifier PXD002024. The detailed bioinformatics analyses used to further classify each identified protein are described in , and all results were compiled into a data set provided in . ## Ms/ms Cloning, expression, and purification of recombinant GSMT and mutant proteins. The Mpgsmt (GenBank: AEG64703) (EC 2.1.1.156) was cloned into the pET28a expression vector (Novagen) [bib_ref] Characterization and regulation of the osmolyte betaine synthesizing enzymes GSMT and SDMT..., Lai [/bib_ref] and then used as a template for site-directed mutagenesis [bib_ref] An efficient one-step site-directed and site-saturation mutagenesis protocol, Zheng [/bib_ref]. Residues Thr-68 and Thr-70 of MpGSMT were substituted with Ala to mimic a non-phosphorylated state, and with Asp to mimic a phosphorylated state. The specific primers are listed in . The pET28a-Mpgsmt construct and its mutant derivatives were transformed into E. coli BL21(DE3)RIL for heterologous expression as described in literature 21 . Methyltransferase activity assay. MpGSMT exhibiting GMT and SMT activities were determined in the forward direction, measuring sarcosine and dimethylglycine formation from glycine (Sigma) and sarcosine (Sigma) substrates. A modified acid-washed charcoal method was used to detect methyltransferase activities following the standard protocol described in literature Salt tolerance test. The plasmid pUHE21-Mpgsmt-sdmt co-expressing MpGSMT/SDMT was initially introduced into the osmolyte uptake mutant E. coli MKH13 as described in literature [bib_ref] Characterization and regulation of the osmolyte betaine synthesizing enzymes GSMT and SDMT..., Lai [/bib_ref]. This plasmid was also used as a template for the point mutation of MpGSMT at residue Thr-68 to Ala or Asp for mimicking dephosphorylation and phosphorylation, respectively. The MKH13 cells harboring plasmid pUHE21-Mpgsmt-sdmt were defined as wild type (WT), while cells carrying the plasmid with a mutated version of Mpgsmt at Thr-68 were indicated as mutant T68A or T68D. For growth tests, three E. coli strains (WT, T68A, and T68D) were grown in M9 minimal medium without NaCl for at least three generations, and then subcultured in different NaCl concentrations in solid and liquid medium to investigate the influence of salt shock and salt adaptation, respectively. To assess the effects of salt shock, the overnight cultures of the three strains were diluted to OD 600 of 1.0 using M9 medium without NaCl, and then used for further serial dilution to be dotted in equal volumes of each dilution, ranging from 10 0 to 10 −7 , on M9 agar plates containing 0, 500, or 700 mM NaCl. All the conditions were carried out in triplicate and incubated at 37 °C. To assess the effects of salt adaptation, the three strains were acclimated in M9 liquid medium containing 0, 500, or 700 mM NaCl for three generations. The cells in 0 mM NaCl were inoculated at an initial OD 600 of 0.01 for growth curves. In order to shorten the lag period prior to initiating growth in high concentrations of 500 or 700 mM NaCl, the initial OD 600 value was fixed at 0.03. The growth rates of each strain in different salt conditions were determined in duplicate experiments. [fig] Figure 1: A representative MS/MS spectrum and classification of the identified phosphoproteins in M. portucalensis FDF1 T . (a) [/fig] [fig] Figure 2: Schematic illustration of phosphorylation events in methanogenesis from TMA metabolic pathways in M. portucalensis FDF1 T . Integrated phosphoproteome data mapped into methanogenic and housekeeping pathways for cells to utilize TMA, produce methane, and gain energy, as shown in different color arrows. The carbon atoms of the methanogenesis are labeled in red "C". The shaded boxes with gene IDs (MPF_ numbers) show the phosphorylated enzymes in this study. [/fig] [fig] Figure 3: Manual annotation and MS/MS analysis reveal Pyl incorporation at amber codons and downstream phosphorylation. (a) The completely sequenced M. portucalensis genome revealed the operon (cyan) that synthesizes Pyl and incorporates it into methyltransferases. Most databases truncate methylamine methyltransferases at the amber codon, but our analysis indicated the incorporation of a Pyl residue (white line) and downstream phosphorylation on Ser-408/Ser-450 and Ser-391 in four methyltransferases (orange). (b) LC-MS/MS analysis confirmed the Pyl incorporation in a tryptic peptide (residues 320-338). The pyrrolysine residue "O" with a mass of 237.3 Da was detected at position 334 of TMA methyltransferase MttB (MPF_1478). [/fig] [fig] Figure 4: The transmethylation activity of MpGSMT was regulated via Thr phosphorylation. (a) Partial sequence alignment of the MpGSMT with its homologous regions from prokaryotic GSMT and eukaryotic GNMT. The amino acid sequences of GSMT were captured from M. portucalensis FDF1 T , Aphanothece halophytica, Halorhodospira halochloris, and Synechococcus sp. WH8102, and the sequences of GNMT from Danio rerio, Homo sapiens, Rattus norvegicus, and Sus scrofa. The conserved residues are highlighted in black, while the strongly similar residues are shown in gray background. Arrows indicate the phosphorylated residue Thr-68, Thr-70, and Ser-74 located in SAM binding motif. The red arrow marks the position of phospho-residue that regulate activity in MpGSMT. (b) The phosphorylation sites mapped to MpGSMT predicted structure according to the rat GNMT (PDB: 1nbh) as template, which was solved with SAM and acetate mimicking glycine in catalytic pocket. The lid moiety (167-219) was shown in green ribbon and central domain containing SAM binding motif (43-166 and 220-263) was marked in blue. The phosphosites were highlighted in sticks. (c) An expanded view of catalytic pocket of wild-type GSMT, T68A, and T68D mutants. The Tyr-26 located near the SAM-binding site was labeled in yellow stick. The distances between Tyr-26 and SAM were labeled in dashed lines. The possible hydrogen bond was denoted in solid line with the distance of 2.9 Å. Scientific RepoRts | 6:29013 | DOI: 10.1038/srep29013 Supplementary [/fig] [fig] Figure 5: Phosphosite mutation of MpGSMT affects methyltransferase activity and growth rate at elevated osmolarity. (a) The relative GMT and SMT activities of recombinant WT MpGSMT and various mutant forms. The methyltransferase activities were performed by modified acid-washed charcoal method under 1.0 M of KCl with 0.5 M glycine or sarcosine as the substrate. All the data points were averaged by triplicate experiments and displayed as percentages relative to wild-type MpGSMT. The effect of salt adaptation was assayed by the growth of E. coli MKH13 containing WT or mutant (T68A and T68D) MpGSMT co-expressing with SDMT in M9 minimal media supplemented with (b) 0 mM, (c) 500 mM, or (d) 700 mM NaCl. Growth rate was monitored by measuring 600 nm optical densiometry (OD 600 ) over 24 h at 37 °C. E. coli strain MKH13 without a recombinant GSMT 21 had the same growth rate as the T68D mutant. The effect of salt shock shown in the inset of (b,c) was measured by drop tests of serious dilution from 10 0 to 10 −7 of overnight cultures on agar plates, but cells failed to grow on agar plate with (d) 700 mM NaCl. Insect in (d) indicates the growth curve of T68D mutant with a large time-scale over 48 hours. [/fig]

The Management of Nonseminomatous Testicular Cancer # Introduction Despite the fact that testis cancer accounts for only 1% of all male malignancies, it is the most common solid malignancy affecting males between the ages of 15 to 35 years. The management of germ cell tumors (GCT) of the testis has proved to be a model of success among solid tumors. This is due to tremendous advances during the past 2 decades, which have changed the once dismal prognosis for patients with advanced disease. [bib_ref] Treatment of testicular cancer: a new and improved model, Einhorn [/bib_ref] Testicular cancer has, in fact, become one of the most curable of all solid neoplasms. Thirty years ago testicular cancer accounted for 11.4% of all cancer deaths in the 25 to [bib_ref] GJ: Adjunctive surgery after chemotherapy for nonseminomatous germa cell tumors: recommendations for..., Toner [/bib_ref] year age group, with an overall 5 year survival rate of 64%. [bib_ref] D: Clues from the natural history and results of treatment supporting the..., Oliver [/bib_ref] The most recent five-year survival rate for GCT of the testis in the United States was recently reported as being in excess of 90%. [bib_ref] Testicular germ cell cancer, Bosl [/bib_ref] Improved prospects for cure and long-term disease free survival rate, are related to a better understanding of the natural history of testicular tumors, improved staging methods and surgical techniques, as well as to the introduction of effective platinum based combination chemotherapy. [bib_ref] Prognostic features of primary and secondary germ cell tumors, Levin [/bib_ref] Modifications to the surgical management of testis cancer have significantly reduced morbidity associated with the classical full bilateral retroperitoneal lymph node dissection. Template and nerve sparing retroperitoneal node dissections are now routinely employed in low stage nonseminomatous germ cell tumors (NSGCT), providing the patient with the benefit of accurate pathological staging combined with reduced morbidity. In addition, nerve sparing techniques have also been used in dissections for residual disease following chemotherapy. [bib_ref] P: Nerve sparing post-chemotherapy retroperitoneal lymph node dissection for advanced testicular cancer, Coogan [/bib_ref] Chemotherapy for advanced GCT has undergone a number of important breakthroughs during the past three decades. In the 1960's the first major chemotherapeutic advance in metastatic NSGCT was described; a combination of chlorambucil, methotrexate and dactinomycin produced a response rate of 30%. [bib_ref] Effects of combined drug therapy on metastatic cancer of the testis, Li [/bib_ref] By the 1970's combination therapies such as vinblastine and bleomycin had been shown to produce objective responses; 57% complete response rates and 45% long-term disease-free survival rates were reported. [bib_ref] Continuous intravenous bleomycin (NSC-125066) therapy with vinblastine (NSC-49842) in stage III testicular..., Samuels [/bib_ref] The discovery of cisplatin as a highly effective anti-neoplastic agent in 1965, [bib_ref] Inhibition of cell division in E. Coli by electrolysis products from a..., Rosenberg [/bib_ref] and its later use in germ cell tumors of the testis (GCT), was another important breakthrough. In 1974 the combination of cisplatin, vinblastine and bleomycin was investigated by the Indiana group. [bib_ref] Cis-diaminedichloroplatinum, vinblastine and bleomycin combination chemotherapy in disseminated testicular cancer, Einhorn [/bib_ref] Of the 47 patients in the initial phase II study, 100% achieved a response rate. Thirty five (74%) patients achieved complete remission, the remaining patients achieved a partial remission. In 1983 the combination of bleomycin, etoposide, and cisplatin (BEP) was first described by a group of investigators at the Royal Marsden Hospital. [bib_ref] Treatment of disseminated germ cell tumors with cisplatin, bleomycin and either vinblastine..., Willaims [/bib_ref] This combination was later compared in randomized studies to cisplatin, vinblastine and bleomycin (PVB) and found to be both less toxic and more effective. The use of BEP therefore became a milestone in the treatment of GCT of the testis and has become a widely used treatment for advanced disease. [bib_ref] The treatment of metastatic germ-cell testicular tumors with beleomycin, etoposide and cisplatin..., Peckham [/bib_ref] [bib_ref] Treatment of disseminated germ cell tumors with cisplatin, bleomycin and either vinblastine..., Williams [/bib_ref] [bib_ref] Two drug therapy in patients with metastatic germ cell tumors, Bajorin [/bib_ref] Since the late 1980's investigators have distinguished between low (good) risk and high (poor) risk disease, based on the diverse prognosis of patients with advanced GCT. [bib_ref] Evaluation of optimal duration of chemotherapy in favourable-prognosis disseminated germ cell tumors:..., Einhorn [/bib_ref] [bib_ref] BEP versus VIP in intermediate risk patients with disseminated nonseminomatous testicular cancer..., Stoter [/bib_ref] In those patients with good risk disease, the objective is to reduce treatment toxicity while maintaining high rates of complete response. In this regard three cycles of BEP were found to be as effective as four cycles of BEP, in patients with good risk disease. 14 On the other hand, studies in poor risk patients have focused on improving both response rates and survival. Numerous phase II and phase III studies have not led to very significant improvements in outcome in poor risk patients, who are fortunately rarely encountered today. [bib_ref] H: Randomised study of cisplatin dose intensity in poor risk germ cell..., Nichols [/bib_ref] [bib_ref] Early intensified chemotherapy with autologous bone marrow transplantation in first line treatment..., Chevreau [/bib_ref] ## Principles of surgical management of testis cancer The principles which underlie the modern surgical treatment of GCT are based on the fact that testis cancer spreads in a predictable and stepwise fashion, with the notable exception of choriocarcinoma. The spermatic cord contains four to eight lymphatic channels that ascend into the retroperitoneum and fan out medially into the retroperitoneal lymph node chain. The first echelon of lymph nodes draining the right testis is located within the interaortocaval nodes at the level of the second vertebral body. The first echelon of nodes draining the left testis is located in the para-aortic region in an area bounded by the renal vein superiorly, aorta medially, ureter laterally and the origin of the inferior mesenteric artery inferiorly. Following spread to either the left or right primary echelon of nodes, subsequent spread may occur in a retrograde fashion to the common, external and inguinal nodes, or cephalad via the cysterna chyli, thoracic duct and supraclavicular nodes. Despite the predictability of lymph node metastasis in testis cancer, there is a failure rate of 5% with distant metastasis following node negative retroperitoneal lymphadenectomy (RPL).This is probably due to the fact that testicular lymphatics may very occasionally bypass retroperitoneal lymph nodes altogether and communicate directly with the thoracic duct. In a review of 104 consecutive cases of stage II NSGCT, Donohue et al, made a number of important observations confirming the predictability of lymphatic spread in testicular cancer. [bib_ref] R: Distribution of nodal metastases in nonseminomatous testis cancer, Donohue [/bib_ref] He reported that suprahilar lymph node spread was rare in stage IIA disease, unlike stage IIB disease, where suprahilar spread was not uncommon. In low stage disease the absence of contralateral nodal involvement was noted. In addition, in this series gonadal vein involvement was noted in a significant number of cases in both low and high stage II disease. These observations have important implications for the surgical management of testis cancer. Lymphatics of the epididymis drain into the external iliac chain; therefore, in locally extensive disease, epididymal involvement may be associated with positive pelvic nodes. Testis cancer that involves the scrotum may result in inguinal node metastases; in addition, prior scrotal surgery or retrograde spread from extensive retroperitoneal involvement may also cause inguinal metastases. Scrotal violation in the setting of orchiectomy for testis cancer has generally been condemned as compromising patient prognosis and for this reason, inguinal orchiectomy with high ligation of the spermatic cord has been the standard of care in the initial management of testis cancer for almost 100 years. [bib_ref] Sarcoma of the testicle, conclusion based upon 114 cases, Keber [/bib_ref] However in a recent review regarding the implications of scrotal violation in testis cancer, Capeluto et al reported on a meta-analysis of 1,182 patients, scrotal violation had occurred in 206 patients. They found no statistical differences in distant recurrence or survival, implying that scrotal violation may not necessarily carry a worse prognosis. Distant spread of testis cancer occurs most commonly to the pulmonary region, with intraparenchymal pulmonary involvement. Subsequent spread is to the liver, viscera, brain and bone, with bony secondaries being encountered late in the course of the disease. [bib_ref] Metastasis from testicular carcinoma, Johnson [/bib_ref] In summary, with the notable exception of choriocarcinoma, testicular cancer generally spreads in a predictable pattern. This has led to the development of new surgical techniques, which provide accurate pathological staging while at the same time providing therapeutic benefit. These techniques are also associated with reduced morbidity when compared to the classic full retroperitoneal lymphadenectomy (RPL). ## Prediction of metastatic potential In general, patients with clinical stage I NSGCT have a 20 -30% incidence of micrometasteses in retroperitoneal lymph nodes. In addition follow-up studies of these patients have shown that visceral metastases occur in approximately 10%. Therefore, in theory approximately 70% of patients with clinical stage I NSGCT could be cured by radical orchiectomy alone. Studies to investigate this dilemma were initiated in the United Kingdom in 1979, and in Denmark in 1980. [bib_ref] Surveillance alone versus radiotherapy after orchiectomyfor clinical stage I nonseminomatous testicular cancer, Rorth [/bib_ref] [bib_ref] Medical Research Council prospective study of surveillance for stage I testicular cancer, Read [/bib_ref] These studies showed that the relapse rate after orchiectomy alone was approximately 30%. More than 90% of relapses occur within the first 2 years after orchiectomy; median time to relapse was 4 to 5 months after orchiectomy. Less than 10% of relapses occur later than 2 years after orchiectomy. Fifty percent of relapses occur in the retroperitoneum and 30% of relapses occur without elevation of tumor markers. One third of relapses occur in the lung and 12% of relapses occur with elevation of tumor markers as the only sign of tumor recurrence. Several studies have investigated prognostic factors that predict relapse. A study performed by the MRC in the United Kingdom described four prognostic factors; 24 vascular invasion, lymphatic invasion, the presence of embryonal carcinoma and the absence of yolk sac tumor. A prognostic index was subsequently created based on these factors. Patients with four positive prognostic factors were found to have a 46% risk of relapse compared with 21% for those with 2 factors and 16% for those with one factor. No relapses occurred in patients without any risk factors. [bib_ref] Histopathology in the prediction of relapse in patients with stage I testicular..., Freedman [/bib_ref] In addition, primary tumor extent has also been shown to have prognostic significance. [bib_ref] Prognostic features of primary and secondary germ cell tumors, Levin [/bib_ref] Clinical stage I NSGCT is therefore no longer managed as a homogeneous disease, rather treatment may now be individualized according to the likelihood of positive retroperitoneal lymph nodes. Patients with low risk of having positive nodes have the option of entering an observation protocol and thus may be spared the morbidity of adjuvant therapy. However, despite prognostic criteria, it is still not possible to establish highly accurate distinctions between patients who should and those who should not receive adjuvant therapy. In the MRC study in the United Kingdom the high risk group constituted less than 25% of all patients and most importantly less than half of all relapses. [bib_ref] Medical Research Council prospective study of surveillance for stage I testicular cancer, Read [/bib_ref] In a recent report, the long term outcome results of a prospective surveillance trial for clinical stage I NSGCT were described. With a median follow-up of 11.3 years, 74.3% of patients remained disease free, while the remainder experienced relapse. All relapses occurred within 2 years, the median time to relapse was 5 months. Significant indicators of relapse during surveillance were a predominant embryonal carcinoma histology and vascular invasion. In patients with neither of these two risk factors, the relapse rate was 12%, none of these patients died of their disease. 26 ## Staging testis cancer There are both clinical and pathological staging systems for NSGCT [fig_ref] TABLE 1 CLINICAL: STAGING SYSTEMS FOR TESTICULAR CANCER [/fig_ref]. Most systems in use today are a variation on the one proposed by Boden and Gibb in 1951.Low stage disease refers to clinical stages I (A) and IIA (B1) disease, whereas high stage disease encompasses stages IIB (B2), IIC (B3), and III (C) disease. This distinction between low and high stage disease is important, as it determines, to a large degree, the adjuvant therapy that the patient receives. While the clinical staging system is based predominantly on CT scan and chest X-Ray, pathologic staging is based on the surgical findings and the surgical specimens obtained at radical orchiectomy and RPL. The staging error associated with clinical staging is therefore removed by pathologic staging. ## American joint committee on cancer tnm classification skinner Primary tumor (T) -extent of the primary tumor is classified after radical orchiectomy pTX -Primary tumor cannot be evaluated i.e. radical orchiectomy not yet performed. pT0 -No evidence of primary tumor (histologic scar in the testis) pTIS -Carcinoma in situ, i.e. pre-invasive stage pT1 -Tumor limited to the testis, including the rete testis pT2 -Tumor invades beyond the tunica albuginea or into the epididymis pT3 -Tumor invades the spermatic cord pT4 -Tumor invades the scrotum A: Confined to the testis Regional lymph nodes (N) NX -Regional lymph nodes cannot be assessed N0 -No regional lymph node metastasis N1 -Metastasis in a single lymph node N2 -Metastasis in a single lymph node 2 cm or 2cm, no extranodal extension less in greatest dimension, or multiple lymph nodes, none more than 5 cm in greatest dimension N3 -Metastasis in a lymph node more than [bib_ref] P: Nerve sparing post-chemotherapy retroperitoneal lymph node dissection for advanced testicular cancer, Coogan [/bib_ref] ## Retroperitoneal lymphadenectomy (rpl) in gct of the testis ## Rpl-i This dissection is performed on patients with no clinical signs of spread to the retroperitoneum, and no surgically visible disease, i.e. clinical stage I disease. The surgical technique employed is either a template or nerve sparing technique. Therefore sympathetic nerves responsible for antegrade ejaculation are preserved, resulting in normal ejaculatory function in the vast majority of patients undergoing RPL-I. ## Rpl-ii This procedure is performed on patients with low volume clinically demonstrable disease, or visible disease at the time of surgery, i.e. clinical stage IIA or IIB disease. The surgical boundaries are generally wider than in RPL-I, usually bilateral above the inferior mesenteric artery and in most cases bilateral below the mesenteric artery as well. The lumbar sympathetics and the hypogastric plexus are carefully preserved, resulting in preservation of ejaculation in experienced hands in over 95% of patients undergoing RPL-II. ## Rpl-iii This procedure is performed on patients who have failed primary therapy, either primary chemotherapy, or primary retroperitoneal node dissection and who have either retroperitoneal recurrences or a persistent retroperitoneal mass detected radiographically (usually by CT scan). In selected cases nerve sparing techniques maybe employed. 5 ## Management of clinical stage i disease - SURVEILLANCE VS ADJUVANT THERAPY The decision to place a patient on a surveillance protocol or to opt for adjuvant therapy in the form of either RPL-I or chemotherapy depends on a number of factors, such as risk factors for disease relapse, patient compliance and local expertise with respect to performance of the surgical procedure. - ADVANTAGES OF RPL-I 3. RPL-I is both a staging and a therapeutic procedure. 4. Follow-up after RPL-I is significantly less intensive than patients followed on surveillance protocols and therefore less expensive. 5. Anxieties, which inevitably accompany the uncertainties associated with a surveillance protocol, are eliminated by surgical intervention. ## Management of clinical stage ii disease The most appropriate therapy for patients with clinical stage II NSGCT is controversial. Debate centers around the role of RPL-II vs chemotherapy and also around the role of post RPL chemotherapy. There is general agreement that clinical stages IIC and III disease be managed initially with chemotherapy, and that surgery should be reserved for residual retroperitoneal masses. The initial management of clinical stages IIA and IIB disease, however, remains controversial. At the Brigham and Womenís Hospital most patients who present with clinical stage IIA disease undergo RPL-II. Furthermore, patients with < 3cm of nodal involvement also undergo RPL-II, while chemotherapy is generally recommended for those patients with > 3cm of nodal involvement. - ADJUVANT CHEMOTHERAPY FOLLOWING RPL-II The role of adjuvant chemotherapy following RPL-II vs. an expectant approach has been addressed by a number of investigators. [bib_ref] Is adjuvant chemotherapy necessary for patients with stage BI testicular cancer, Richie [/bib_ref] [bib_ref] Adjuvant chemotherapy in selected patients with pathologic stage II NSGCT, Pizzocaro [/bib_ref] Richie reviewed 39 patients with pathologic stage II disease who were carefully followed post RPL-II. Relapses occurred in only 8% of patients, all of whom were rendered disease free with 3 to 4 cycles of chemotherapy. [bib_ref] Is adjuvant chemotherapy necessary for patients with stage BI testicular cancer, Richie [/bib_ref] In this study therefore immediate post RPL-II chemotherapy would have benefited only a small number of patients. In contrast, a randomized multi-center study reported a 48% relapse rate for patients with pathologic stage II disease, following RPL-II. [bib_ref] Immediate adjuvant chemotherapy versus observation with treatment at relapse in pathological stage..., Williams [/bib_ref] These two studies highlight the fact that results from institutions where RPL is performed frequently, may not parallel results from large cooperative trials in which some centers contribute relatively small numbers of patients. Following RPL-II the decision to administer two cycles of chemotherapy depends on a number of factors: tumor volume, number of positive nodes, presence of vascular invasion in the primary tumor and histology of the primary tumor. Expectant management is appropriate in patients with fewer than 6 positive nodes (all of which are less than 2cm in diameter), with no positive margins. For patients with more than 6 positive nodes or high volume disease, two cycles of chemotherapy are given following RPL-II. [bib_ref] Adjuvant chemotherapy for metastatic stage NSGCT of the testis, Weisbach [/bib_ref] This protocol has achieved excellent remission rates. For patients who develop a recurrence while being treated expectantly, chemotherapy is employed at that stage, once again with excellent remission rates. ## Surgical management of advanced nsgct: rpl-iii RPL-III in patients with advanced NSGCT provides essential information for the future management of the patient: - Evaluates the response to chemotherapy. - Removes viable GCT, thus achieving a complete response to chemotherapy and surgery. - Directs the need for further therapy. Approximately 30% of patients with high volume metastatic NSGCT will experience a partial remission following cisplatin based chemotherapy and are therefore candidates for RPL-III, in addition to those patients who recur in the retroperitoneum following an incomplete node dissection for low stage disease. Partial remission is defined as normal tumor markers in combination with radiographic evidence of disease in the chest, abdomen, mediastinum, neck or elsewhere. ## Persistent or recurrent retroperitoneal mass: histopathology and implications. Early retrospective studies revealed that RPL-III defines three subsets of patients based on histopathological analysis of the resected specimen: 40% -necrosis/fibrosis; 40% -adult teratoma; 20%residual NSGCT. [bib_ref] Cytoreductive surgery for advanced nonseminomatous germ cell tumors of the testis, Brenner [/bib_ref] [bib_ref] Surgical resection in disseminated testicular cancer following cytoreduction, Einhorn [/bib_ref] Therefore approximately 60% of patients with evidence of a residual mass on postchemotherapy imaging studies will either have viable cancer or teratoma. A more recent study showed that the likelihood of malignancy in postchemotherapy resected tumor was 13%, with the remainder of tumor specimens containing teratoma or necrosis. [bib_ref] GJ: Adjunctive surgery after chemotherapy for nonseminomatous germa cell tumors: recommendations for..., Toner [/bib_ref] The finding of only necrotic/fibrotic tissue implies that no further treatment is required. [bib_ref] The management of patients with advanced germ cell tumors, Law [/bib_ref] However patients with viable GCT require an additional two cycles of chemotherapy. Fox et al 36 reported a series of 27 patients who received additional cisplatin-based chemotherapy whereas 7 did not. Disease free survival was 70% in the chemotherapy group, while all 7 patients who did not receive additional chemotherapy relapsed. The rationale to resect residual teratoma is multi-factorial: firstly, by resecting residual teratoma the growing teratoma syndrome is avoided, in which indolent growth causes complications from local progression; 37 secondly, malignant transformation of mature teratoma to sarcoma and adenocarcinoma which is resistant to chemotherapy has been well described; [bib_ref] I: Teratoma following cisplatin based chemotherapy for nonseminomatous germ cell tumors. A..., Loehrer [/bib_ref] thirdly, there is a risk of recurrent viable GCT in teratoma, with no difference in incidence between mature and immature teratoma. 39 ## Rpl-iii: indications and patient selection Despite the well described hazards of residual viable GCT in the retroperitoneum following chemotherapy, as well as recurrent retroperitoneal GCT following RPL-I or RPL-II, the indications for RPL-III are controversial and there are no clear guidelines. The presence of elevated tumor markers, however, remains the only generally accepted contraindication to adjunctive surgery in patients after chemotherapy. [bib_ref] Current perspectives on the role of adjunctive surgery in combined modality treatment..., Bajorin [/bib_ref] The management of patients whose tumor markers have normalized following chemotherapy ranges from observation, irrespective of the results of imaging studies, to surgery for all patients. [bib_ref] Nonseminomatous germ cell testicular tumors: residual masses after chemotherapy, Levitt [/bib_ref] [bib_ref] Post-chemotherapy lymph node histology in radiologically normal patients with metastatic nonseminomatous testicular..., Fossa [/bib_ref] A residual mass is usually detected by CT scan following chemotherapy; however, the definitions of a normal CT scan differ from one institution to another. Definitions include: no visible mass, lymph nodes no larger than 1 cm and lymph nodes less than 1.5 cm to 2 cm in diameter. [bib_ref] The growing teratoma syndrome, Logothesis [/bib_ref] In view of the well described complications related to RPL-III, patient selection plays an important role in excluding those patients who are unlikely to have viable disease. Donohue et al [bib_ref] Management of retroperitoneal recurrences: seminoma and nonseminoma, Donohue [/bib_ref] found that in cases where no teratomatous elements were detected in the primary tumor and a greater than 90% decrease was seen in the volume of retroperitoneal disease by sequential CT scans, viable germ cell tumor or teratoma was not seen in the resected specimen. Fossa et al 42 reported on 37 patients who underwent RPL-III despite negative CT scans following chemotherapy; teratoma was found in 30% while 3% were found to have NSGCT. The volume of teratoma was small, and may thus not be clinically significant to justify a blanket policy of RPL-III despite negative CT scan findings. Toner et al 34 reported on 185 patients who underwent RPL-III after chemotherapy; in this series, no factors could be found that could reliably predict patients who had residual viable malignancy or teratoma in the retroperitoneum. In view of this RPL-III was recommended for all patients with residual abnormalities on imaging of the retroperitoneum, and in all patients with an initial mass of greater than 3 cm regardless of the post chemotherapy radiological findings. [bib_ref] Cytoreductive surgery for advanced nonseminomatous germ cell tumors of the testis, Brenner [/bib_ref] [bib_ref] Surgical resection in disseminated testicular cancer following cytoreduction, Einhorn [/bib_ref] An international data set was recently collected comprising data from six study groups. Logistic regression analysis was used to estimate the probability of necrosis and the ratio of cancer and mature teratoma, in patients who were treated with cisplatin-based chemotherapy for metastatic NSGCT, and who obtained normal tumor markers prior to resection. Multivariate combination of predictors yielded reliable models, which discriminated necrosis well from other histologies, but which discriminated cancer only reasonably well from mature teratoma. [bib_ref] Prediction of residual retroperitoneal mass histology after chemotherapy for metastatic NSGCT: multivariate..., Steyerberg [/bib_ref] While there are statistical correlationís with various factors such as degree of tumor shrinkage, size of residual mass, pre-chemotherapy tumor markers and teratomatous elements in the primary specimen, the risk of a false negative prediction is approximately 20%. [bib_ref] GJ: Adjunctive surgery after chemotherapy for nonseminomatous germa cell tumors: recommendations for..., Toner [/bib_ref] [bib_ref] Post-chemotherapy lymph node histology in radiologically normal patients with metastatic nonseminomatous testicular..., Fossa [/bib_ref] Complications of RPL III RPL-III is a technically very demanding procedure, which sometimes entails a combined abdominal and thoracic approach. Loss of ejaculation has been well documented in patients undergoing RPL-III. In a recent study 25.9% of patients were reported to have absence of ejaculation following RPL-III. In this study, the volume of tumor resected was higher in those patients who had absence of ejaculation than in those patients with preservation of ejaculation. In addition, absence of ejaculation was reported significantly more often in patients with right paracaval/interaortocaval tumor (34.5%), than in those with left para-aortic tumor (16.7%). [bib_ref] Complications of post-chemotherapy resection of retroperitoneal residual tumor mass in patients with..., Gels [/bib_ref] Nerve sparing techniques have been successfully employed in a subset of these patients. [bib_ref] P: Nerve sparing post-chemotherapy retroperitoneal lymph node dissection for advanced testicular cancer, Coogan [/bib_ref] In a recent review of 472 patients who underwent RPL-III, 93 (19.7%) were subjected to nerve sparing procedures, mean size of residual disease on CT scan prior to the procedure was 6.3 cm (range 1 to 19). In this subset of patients 76.5% reported normal ejaculation and ten pregnancies were conceived. Testis cancer recurred in 6 patients but no tumor was retroperitoneal. Therefore, nerve sparing RPL-III appears to be viable in a subset of patients with limited volume and lateralized disease, without compromising cancer surgery outcomes, while preserving ejaculation. [bib_ref] P: Nerve sparing post-chemotherapy retroperitoneal lymph node dissection for advanced testicular cancer, Coogan [/bib_ref] Occasionally, resection of the infrarenal vena cava is necessary and has been well described. [bib_ref] Resection of the inferior vena cava or intraluminal vena caval tumor thrombectomyduring..., Donohue [/bib_ref] [bib_ref] Vascular considerations in post chemotherapy retroperitoneal lymph node dissection: part I -vena..., Donohue [/bib_ref] In addition, aortic tube grafts may be necessary to replace aortic segments which are damaged by subadvential dissection. [bib_ref] Vascular considerations in post chemotherapy retroperitoneal lymph node dissection: part II, Donohue [/bib_ref] The complication rate is therefore directly related to the complexity of the dissection; in addition, complications related to the cumulative toxicity of chemotherapeutic agents are of particular relevance to patients undergoing RPL-III following chemotherapy. Of particular relevance to post-chemotherapy patients is the association of restrictive lung disease in patients who have received bleomycin as part of their chemotherapy protocol. Bleomycin induced pulmonary toxicity is associated with a mortality rate which approaches 50%. In a recent report, Baniel et. al. [bib_ref] Complications of post-chemotherapy retroperitoneal lymph node dissection, Baniel [/bib_ref] reviewed the surgical morbidity of 603 patients who underwent RPL after primary chemotherapy for stage II and III testis cancer. The complication rate was 20.7% and the mortality rate was 0.8%. Pulmonary complications were the most frequent cause of severe morbidity: 6 patients developed adult respiratory distress syndrome, the underlying cause of which was a combination of bleomycin induced pulmonary toxicity, together with large volume retroperitoneal and pulmonary disease resected in this subset of patients. ## Chemotherapy for patients with advanced nsgct While chemotherapeutic trials were being conducted, it became clear that prognoses varied and that patients should therefore be divided into low and high risk categories. For those patients predicted to have a more favorable outcome i.e. good risk, the goals have been to maintain high cure rates while reducing treatment related toxicity. [bib_ref] Testicular germ cell cancer, Bosl [/bib_ref] [bib_ref] Comparison of criteria for assigning germ cell tumor patients to good risk..., Bajorin [/bib_ref] On the other hand the goal of treating poor risk patients has been to improve the proportion of patients achieving a complete response, while at the same time achieving tolerable treatment side effects. Using data from the cisplatin, vinblastine, bleomycin (PVB) studies, the Indiana University staging system was developed [fig_ref] TABLE 3 INDIANA: UNIVERSITY STAGING SYSTEM FOR DISSEMINATED DISEASE [/fig_ref]. This classification uses clinical parameters to divide patients with disseminated testis cancer into three stages -minimal, moderate and advanced disease. While minimal and moderate disease is good risk and is associated with high complete response rates to cisplatin based chemotherapy and low risk of relapse, advanced disease is poor risk disease and therefore has a lower complete response rate and a higher rate of relapse. Although the Indiana University system is easy to use it does not take into account the important role that the tumor markers HCG (human chorionic gonadotropin), AFP (alpha fetoprotein) and LDH (lactate dehydrogenase) play in this disease. [bib_ref] Comparison of criteria for assigning germ cell tumor patients to good risk..., Bajorin [/bib_ref] Bosl et al developed one of the first validated models to mathematically categorize patients into prognostic groups. [bib_ref] Multivariate analysis of prognostic variables in patients with metastatic testicualr cancer, Bosl [/bib_ref] The variables in the model include the number of metastatic sites as well as HCG and LDH. Several other models have been published affirming the principles that the Bosl model described. [bib_ref] Comparison of criteria for assigning germ cell tumor patients to good risk..., Bajorin [/bib_ref] The International Germ Cell Cancer Collaborative Group recently published a clinically based prognostic classification for GCT in which the following adverse factors were identified: mediastinal primary site; degree of elevation of alpha-fetoprotein (AFP), human chorionic gonadtrophin (hCG) and lactate dehydrogenase (LDH); presence of nonpulmonary metasteses such as liver, bone and brain. Integration of these factors produced the following groupings: good prognosis, comprising 60% of GCT with a 91% 5- [fig_ref] TABLE 4 INTERNATIONAL: GERM CELL CONSENSUS CLASSIFICATION FOR NSGCT [/fig_ref]. Hopefully the widespread use of a standardized prognostic model will result in more patients benefiting from chemotherapeutic regimens with fewer side effects. [bib_ref] Prognostic factors for metastatic germ cell cancers treated with platinum basec chemotherapy:..., Mead [/bib_ref] ## Chemotherapy in good risk patients The first prospective trials in good risk patients compared the VAB-6 regimen (vinblastine, dactinomycin, bleomycin, cyclophosphamide and cisplatin) with a two drug regimen of etoposide plus cisplatin (EP). [bib_ref] A randomized trial of etoposide + cisplatin versus vinblastine + bleomycin +..., Bosl [/bib_ref] The results of this trial demonstrated equivalence but with a toxicity favoring EP. In long term follow-up no excess of late relapse was found in patients treated with 4 cycles of EP. 14 Thus 4 cycles of EP with cycles repeated every 21 -28 days has become one of the international standards for good risk disease. Einhorn et al compared 4 cycles of BEP (bleomycin, etoposide, cisplatin) to 3 cycles of BEP. One hundred eighty-four patients were entered into this trial. Overall 99% of patients with minimal advanced stage disease and 95% of patients with moderate advanced stage disease achieved disease free status. Only 5.6% of patients relapsed and the overall disease free survival was the same in both arms of this study, which demonstrated that elimination of the fourth cycle of BEP reduced toxicity without affecting survival. [bib_ref] Evaluation of optimal duration of chemotherapy in favourable-prognosis disseminated germ cell tumors:..., Einhorn [/bib_ref] In a trial conducted by the Eastern Cooperative Oncology Group (ECOG), 178 good risk patients were randomized to receive three cycles of EP either with or without bleomycin. [bib_ref] Randomised trial of etoposide and cisplatin versus etoposide and carboplatin in patients..., Loehrer [/bib_ref] Myelosuppression was slightly greater in the BEP arm (57% versus 45%); no pulmonary toxicity was seen in either group. This trial was stopped early because of a statistically significant increase in the number of adverse events (defined as death, persistent or progressive cancer, or relapse) in patients not receiving bleomycin. Of the 86 patients treated with BEP, 95% achieved a disease-free status and 86% were continuously disease free, compared with 90% and 70% respectively in the EP arm. This data supports the necessity of including bleomycin when three cycles of cisplatin based chemotherapy are used. Carboplatin was chosen as a substitute for cisplatin in a trial performed by Bajorin et al., because of its better toxicity profile, i.e. fewer renal, otologic and neurologic toxicities. 57 A durable complete response was achieved in only 76% of patients treated with etoposide and carboplatin (EC) compared to a durable complete response rate of 87% in patients treated with EP. [bib_ref] The management of patients with advanced germ cell tumors, Law [/bib_ref] This trial therefore confirmed the efficacy of four cycles of EP and also showed that carboplatin cannot be substituted for cisplatin in this subset of patients. ## Chemotherapy in poor risk patients Clinical trials involving patients with advanced disease have focused on enhancing response rates and improving survival. Poor risk patients are fortunately rare and usually present with primary mediastinal GCT. [bib_ref] Prognostic factors in metastatic disease, Droz [/bib_ref] Marked elevation of one or more tumor markers (beta-human chorionic gonadotropin, alphafetoprotein and lactate dehydrogenase) is a common reason for inclusion into the poor risk group. In addition the presence of liver, bone or brain metastases as well as increasing bulk of disease are also used as inclusion criteria for the poor prognostic group. [bib_ref] Prognostic factors in metastatic disease, Droz [/bib_ref] [bib_ref] Prognostic factors for metastatic germ cell cancers treated with platinum-based chemotherapy: the..., Mead [/bib_ref] Although there are a large number of phase II studies for the treatment of poor risk patients, none of them has been sufficiently encouraging to be considered as a standard therapy. Several trials have been conducted using conventional dose chemotherapy. Ifosfamide in combination with etoposide and cisplatin (VIP) has been compared to BEP. [bib_ref] P: Phase three study of cisplatin (P) plus etoposide (VP-16) with either..., Loehrer [/bib_ref] VIP proved to be more toxic but equivalent to BEP with respect to efficacy. Impressive results have been reported from single center studies describing the use of cyclic alternating chemotherapy regimens. These regimens expose the GCT to a wide variety of chemotherapeutic drugs in the hope of preventing drug resistance. The results however have not been validated by randomized trials. Due to the fact that high dose chemotherapy given as second or third line therapy was associated with a very high morbidity and mortality rate, the role of chemotherapy with autologous bone marrow transplantation (ABMT) was investigated. [bib_ref] Dose intensive chemotherapy in refractory germ cell cancer: A phase I/II trial..., Nichols [/bib_ref] [bib_ref] Phase II trial of high dose carboplatin and etoposide with autologous bone..., Motzer [/bib_ref] Twenty eight poor risk patients were entered into a trial in which high dose chemotherapy was given to those patients whose tumor markers exhibited a prolonged decline following standard dose therapy. When the results in this subset of patients were compared to the prior experience in poor risk patients, there appeared to be a survival benefit. [bib_ref] Scrotum and testis, Rowland [/bib_ref] A French cooperative group reported their experience with high dose chemotherapy followed by ABMT. Patients were randomized to receive either conventional or high dose chemotherapy in combination with ABMT. However, 28% of patients randomized to receive high-dosed therapy did not receive this treatment and no statistical differences were apparent between the 2 groups of patients. [bib_ref] Early intensified chemotherapy with autologous bone marrow transplantation in first line treatment..., Chevreau [/bib_ref] Growth factor and peripheral blood stem cell support have been used in combination with dose intensification of all cycles of chemotherapy. In a study involving 77 poor risk patients, dose escalation of the chemotherapy regimen was combined with granulocyte-macrophage-colony stimulating factor (GM-CSF). Overall 67% of patients achieved progression-free survival. 64 ## Chemotherapy for relapsed testis cancer For the 20% to 30% of patients with advanced GCT, who do not achieve complete response to first line chemotherapy, second line salvage chemotherapy remains an option. [bib_ref] Treatment of testicular cancer: a new and improved model, Einhorn [/bib_ref] [bib_ref] Second line chemotherapy with vinblastine, ifosfamide and cisplatin after initial chemotherapy with..., Einhorn [/bib_ref] The current standard chemotherapy that serves as a basis for comparison is vinblastine, ifosfamide and cisplatin (VeIP). [bib_ref] Second line chemotherapy with vinblastine, ifosfamide and cisplatin after initial chemotherapy with..., Einhorn [/bib_ref] As second line therapy 33% to 69% of patients achieved a complete response with ifosfamide based regimens, which was found to be durable in approximately 50% of patients. [bib_ref] Cisplatin, etoposide and ifosfamide salvage therapy for refractory or relapsing germ cell..., Harstrick [/bib_ref] In view of this, ifosfamide is generally considered to be a standard first line salvage therapy. Approximately 25% to 50% of patients who relapse following chemotherapy have been shown to achieve a complete response to high dose carboplatin and etoposide with and without oxazophosphorine (cyclophosphamide, ifosfamide) and ABMT. [bib_ref] High dose chemotherapy for resistant germ cell tumors. Recent advances and future..., Motzer [/bib_ref] Forty three percent of patients achieved a complete response, 27% remained alive and disease free at 8 months of follow-up. Favorable pretreatment prognostic factors for survival to treatment with conventional dose salvage therapy include a prior complete response to cisplatin based chemotherapy and a testicular primary site. Prognostic factors that predict an inferior long term disease free survival include an initial incomplete response to cisplatin based chemotherapy and a primary mediastinal site of disease. [bib_ref] High dose chemotherapy for resistant germ cell tumors. Recent advances and future..., Motzer [/bib_ref] These poor prognostic factors predict a very poor outcome and the role of standard chemotherapy is thus limited. Aggressive surgery may offer a reasonable alternative as may radiotherapy, in addition high dose chemotherapeutic regimens may have a role. [bib_ref] The management of patients with advanced germ cell tumors, Law [/bib_ref] # Conclusions Advances in the management of GCT of the testis have led to a multimodal approach which has resulted in the vast majority of patients being cured of their disease. More than 90% of patients with newly diagnosed GCT will be cured. RPL not only provides accurate data for pathological staging, but is also therapeutic in low stage disease. Moreover RPL plays an essential role in the management of high stage disease with respect to both directing further patient management and providing long-term disease free survival following complete resection of residual teratoma. Precise surgical technique however is mandatory to ensure disease free survival. Patients presenting with advanced disease are stratified into good, intermediate and poor risk groups. Therapy for good risk disease consists of either four cycles of EP or three cycles of BEP; resulting complete response rates are high. On the other hand poor risk patients require intensive and specialized management and should be referred to an appropriate center with expertise in this area. ABMT or stem cell support combined with high dose chemotherapeutic regimens (VeIP) can be successfully used to treat patients who relapse following first line chemotherapy. However patients who relapse with poor prognostic features do not respond well to high dose chemotherapy, and may thus require aggressive surgery or radiotherapy. [table] TABLE 1 CLINICAL: STAGING SYSTEMS FOR TESTICULAR CANCER [/table] [table] TABLE 2 PATHOLOGICAL: STAGING [/table] [table] 1: Thirty percent of patients with clinical stage I NSGCT have pathologic stage II disease. 2. Accurate pathologic staging allows for earlier definition of further treatment options. [/table] [table] TABLE 3 INDIANA: UNIVERSITY STAGING SYSTEM FOR DISSEMINATED DISEASE: MINIMAL AND MODERATE STAGES CONSTITUTE GOOD RISK DISEASEyear survival rate; intermediate prognosis, comprising 26% of GCT with a 79% 5-year survival rate; and poor prognosis, comprising 14% of GCT with a 48% 5-year survival rate53 [/table] [table] TABLE 4 INTERNATIONAL: GERM CELL CONSENSUS CLASSIFICATION FOR NSGCT: [/table]

In-situ localization and biochemical analysis of bio-molecules reveals Pb-stress amelioration in Brassica juncea L. by co-application of 24-Epibrassinolide and Salicylic Acid Lead (Pb) toxicity is a major environmental concern affirming the need of proper mitigation strategies. In the present work, potential of combined treatment of 24-Epibrassinolide (24-EBL) and Salicylic acid (sA) against pb toxicity to Brassica juncea L. seedlings were evaluated. seedlings pre-imbibed in eBL (0.1 mM) and SA (1 mM) individually and in combination, were sown in Pb supplemented petri-plates (0.25, 0.50 and 0.75 mM). Various microscopic observations and biochemical analysis were made on 10 days old seedlings of B. juncea. The toxic effects of Pb were evident with enhancement in in-situ accumulation of pb, hydrogen peroxide (H 2 o 2 ), malondialdehyde (MDA), nuclear damage, membrane damage, cell death and polyamine. Furthermore, free amino acid were lowered in response to pb toxicity. the levels of osmoprotectants including total carbohydrate, reducing sugars, trehalose, proline and glycine betaine were elevated in response to pb treatment. soaking treatment with combination of 24-EBL and SA led to effective amelioration of toxic effects of Pb. Reduction in Pb accumulation, reactive oxygen content (Ros), cellular damage and GsH levels were noticed in response to treatment with 24-EBL and SA individual and combined levels. The contents of free amino acid, amino acid profiling as well as in-situ localization of polyamine (spermidine) was recorded to be enhanced by co-application of 24-EBLand SA. Binary treatment of 24-EBL and SA, further elevated the content of osmoprotectants. The study revealed that co-application of combined treatment of 24-EBL and SA led to dimination of toxic effects of Pb in B. juncea seedlings.Contamination of environment with heavy metals is one of the major concerns of the environmentalist in developing and developed countries. Un-controlled addition of heavy metals to the soil has led to far reaching effects on agriculture, as a result of effect on food safety, economic value and uptake by plants and humans 1 . Various metals and metal oxides nanoparticles have been reported to be deleterious for plants 2 . Heavy metal toxicity in plants hampers growth, efficacy of photosynthetic apparatus, senescence and functioning of specific enzymes 3-9 . Lead (Pb) is considered to be one of the most abundant environmental pollutants and enters the environment through anthropogenic addition and consequently cause contamination of biocoenosis and biotopes 10,11 . Once Pb reaches the interior of roots, it gets accumulated in root cells or is translocated to the aerial parts 12-14 . Due to highly toxic nature of Pb, variable symptoms are observed in effected plants including necrosis, chlorosis, growth inhibition, senescence and enhanced generation of reactive oxygen species (ROS) such as hydrogen peroxide (H 2 O 2 ), superoxide anion (O 2 − ), hydroxyl ion (HO), singlet oxygen (O) and nitric oxide (NO) etc 15,16 . To circumvent metal toxicity, plants have developed discrete strategies by which toxic metal ions are effluxed, retained in roots or are transported to the other parts of plant 3 . These strategies are broadly classified into avoidance and detoxification mechanisms 17 . Plants synthesize certain antioxidative enzymes such as superoxide dismutase (SOD), catalase (CAT), guaiacol peroxidase (POD), ascorbate peroxidase (APOX) and glutathione reductase (GR) and non-enzymatic antioxidants (glutathione, cysteine, ascorbic acid, tocopherol etc.)18,19. A few avoidance mechanisms include enhanced accumulation of metal chelating compounds, phenolic compounds and osmoprotectants 20 . More recently, use of exogenous application of plant growth regulators (PGRs) to provide protection to plants against oxidative stress has gained attention8,9,21,22. Various phytohormones including auxins (AUX), gibberellins (GBs), ethylene (ET), brassinosteroids (BRs), jasmonic acid (JA) and salicylic acid (SA) have been studied for their positive potential to promote growth and elevate tolerance of plants to heavy metal toxicity23,24.Numerous studies have reported positive potential of BRs as stress protective agents25,26. They have been reported to alleviate metal stress in yellow mustard 27 , raddish 28 , cucumber 29 tomato 30 , Indian mustard 31 and maize 32 . Similarly, participation of SA in adaptation of plants to wide array of stresses is largely documented 33,34 . Exogenous supplementation with SA to metal stressed plants led to growth promotion and improved photosynthetic efficacy 35,36 , reduced ROS levels 37 and altered osmolyte levels 38,39 . It has been studied for its anti-stress potential in tobacco 22 , wheat 40 , potato 41 and rice 42 . In response to various environmental cues BRs interplays with other plant hormones to regulate plethora of attributes of growth and developmental processes in plants 43 . Interplay between BRs and SA has been reported to counteract various stresses 44,45 , including viral infection 46 , fungal infection 47 , salt and temperature stress 45 .Adequate amount of foregoing studies have been carried out on amelioration of heavy metal induced toxicity by application of BRs and SA individually. In our previous studies we determined the effects of combination of 24-EBL and SA on some physiological and antioxidative characteristics of B. juncea seedlings grown hydroponically and under field conditions 48-50 . The present study further extended into the few more biochemical parameters including evaluation of oxidative stress, amino acid levels and osmolytes contents and histochemical analysis and is an attempt to better understand the interactive effect of 24-EBL and SA in heavy metal stress amelioration.Materials and MethodsPlant Material. Indian mustard (Brassica juncea L., var. RLC-1) seeds were obtained from Punjab Agriculture University, Ludhiana, Punjab, India. The procured seeds were surface sterilized by rinsing for 1 minute in 0.01% mercuric chloride (HgCl 2 ), followed by washing with double distilled water. The seeds were then soaked in different solutions of hormones i.e. EBL (0.1 mM), SA (1 mM) and EBL + SA (0.1 mM + 1 mM) for 6 hr. The seeds were sown in autoclaved petri-plates (10 cm, diameter), lined with Whatmann no. 1 filter paper. Various Pb concentrations were prepared in Hoagland's medium (half strength medium was prepared following the method of Cowgill and Milazzo 51 ). The petri-plates labelled as control were supplied with only Hoagland's nutrient medium. The three different Pb (NO 3 ) 2 (lead nitrate) concentration i.e. 0.25 mM, 0.50 mM and 0.75 mM, were selected on the basis of IC 50 (50% inhibitory concentration). The seedlings were then raised in seed germinator and were provided with 25 ± 0.5 °C temperature, 175 µmol m −2 s −1 light intensity, 16 hr photoperiod and 80-90% relative humidity. The seedlings were harvested after 10 days of growth. 3 replicates of each treatment were taken for further analysis. pb Localization. The roots of B. juncea L. seedlings were stained using Pb specific flouroscent probe Leadmium TM Green AM dye (Invitogen, ThermoFisher Scientific). The roots were placed in Na 2 -EDTA (disodium ethylene diamine tetra acetic acid) (20 mM) for 15 minutes and kept at room temperature. The sections were then washed with double distilled water three times for 10 min each. The stock solution of probe was prepared by adding 50 µL dimethyl sulfoxide directly to the vial of the dye, followed by diluting it 1:10 with 0.85% NaCl (sodium chloride) [bib_ref] A new method to detect cadmium uptake in protoplasts, Lindberg [/bib_ref]. The roots were then immersed in this diluted stock for 2 hr at 37 °C. It was followed by washing three times with 0.85% NaCl. The roots were then visualized under confocal laser scanning microscope (Nikon AIR). The observations were made at the excitation wavelength of 488 nm and emission wavelength of 590 nm. Oxidative Damage. The oxidative damage was evaluated by estimating contents of superoxide anion, H 2 O 2 and MDA spectrophotopmetrically and by in-situ localization of H 2 O 2 , nuclear damage, membrane damage, glutathione (GSH), MDA and cell viability employing confocal laser scanning microscope and visible compound microscope. Superoxide anion content. The level of superoxide anion was determined by following the method of Wu, et al. [bib_ref] Alleviation of salt stress in citrus seedlings inoculated with mycorrhiza: changes in..., Wu [/bib_ref]. The seedling extract was prepared in 50 mM potassium phosphate buffer (PPB) with pH 7.8 and supernatant was obtained by centrifugation at 13, 000 × g for 15 minute at 4 °C. To 0.5 ml of the supernatant 0.5 ml PPB, 0.1 ml of 10 mM hydroxylamine hydrochloride was added. The reaction mixture was incubated for 30 minutes at 25 °C, followed by addition of equal volumes of 7 mM 1-napthylamine and 58 mM of 3-aminobenzenesulphonic acid and was incubated for another 20 minutes. The absorbance of the reaction mixture was read at 530 nm. A standard equation was derived to calculate the content of superoxide anion and standard used was sodium nirite (NaNO 2 ). H 2 O 2 estimation. H 2 O 2 content was estimated by following method of Velikova, et al. [bib_ref] Oxidative stress and some antioxidant systems in acid rain-treated bean plants, Velikova [/bib_ref]. The seedling extract was prepared by homogenizing 100 mg fresh plant samples in 1.5 ml of tri-chloroacetic acid (0.1%). The supernatant was obtained by centrifugation of extract at 12,000 × g at 4 °C for 15 minutes. To 0.4 ml of supernatant, 400 µL of MDA content. The level of MDA was determined by the protocol of Heath and Packer [bib_ref] Photoperoxidation in isolated chloroplasts, Heath [/bib_ref]. Fresh seedlings were homogenized in tri-chloroacetic acid (0.1%) and were centrifuged for 20 minutes at 4 °C at 13, 000 × g. To the supernatant, 0.5% thiobarbituric acid and tri-chloroacetic acid (20%) were added. The reaction mixture was then kept in water bath at 95 °C for about 95 °C for 30 minutes, followed by cooling the mixture on an ice bath. The absorbance of the reaction mixture was read at 532 nm and 600 nm and content was calculated using extinction coefficient i.e. 155/ mM/ cm. H 2 O 2 localization. The method given by Ortega-Villasante, et al. [bib_ref] Rapid alteration of cellular redox homeostasis upon exposure to cadmium and mercury..., Ortega-Villasante [/bib_ref] was followed for localization of H 2 O 2 which was done with 2 1 -7 1 -dichloroflourcien diacetate and was incubated for 30 minutes. The roots were then washed three times for 5 minutes with double distilled water and were mounted on a glass slide. The excitation wavelength was 488 nm and emission wavelength was 530 nm respectively. Visualization of membrane damage. For visualization of membrane damage, propidium iodide (PI) a fluorescent adduct was used. A 50 µM solution of PI was prepared according to the method of Gutierrez-Alcala, et al. [bib_ref] Glutathione biosynthesis in Arabidopsis trichome cells, Gutierrez-Alcala [/bib_ref]. The root samples were dipped in PI fluorescent probe for 15 minutes, followed by washing with double distilled water and were then mounted on glas slide. The excitation wavelength was 535 nm and emission wavelength was 617 nm respectively. Nuclear damage. The nuclear damage in the root cells was examined by employing a fluorescent dye i.e. 4,6-diamino-2-phenylindole (DAPI) and was prepared following the method of Callard, et al. [bib_ref] Novel Molecular Markers for Late Phases of the Growth Cycle of Arabidopsis..., Callard [/bib_ref]. The dye was prepared by dissolving 0.1 gm of DAPI in 100 ml of phosphate buffer saline (PBS). The roots were mounted on glass slide and visualized at excitation wavelength of 358 nm and emission wavelength of 461 nm respectively. MDA localization. MDA was visualized by employing schiff 's reagent (a visible dye) following the protocol of Wang and Yang 59 . The roots of 10 days old seedlings were excised and dipped in schiff 's reagent for 15 minutes and was followed by washing with 0.5% potassium metabisulphite prepared in 0.05 M HCl. The washed roots were mounted on glass slide and visualized under visible compound microscope. Cell Viability. The method of Romero-Puertas, et al. [bib_ref] Cadmium-induced subcellular accumulation of O 2 − and H 2 O 2..., Romero-Puertas [/bib_ref] was used for analyzing cell viability of root cells. The non-viable cells were localized by 0.25% Evan's blue dye. The roots were dipped in visible probe and were kept at room temperature for 10 minutes, followed by washing with double distilled water. The prepared slides were visualized under visible compound microscope. Quantitative and Qualitative estimation of Amino Acids. The quantitative estimation of total free amino acids was done by using spectrophotometer, while qualitative profiling of different amino acids was done by employing Amino acid analyzer (Shimadzu, Nexera X 2 ). Total free amino acid content. Total free amino acid content was estimated by following the method of Lee and Takahashi 62 . 100 mg of dried plant sample was extracted by dipping the plant material in 5 ml of 80% ethanol (extractant). The samples were then centrifuged for 20 minutes at 4 °C at 2000 × g, followed by addition of 3.8 ml of ninhydrin reagent to 0.2 ml of supernatant. The reaction mixture was boiled for 12 minutes in a water bath. After boiling, the sample were cooled at room temperature followed by reading the absorbance at 570 nm. The betaine hydrochloride was used as standard for preparation of standard curve. Amino acid profiling. The samples for amino acid profiling were prepared by method of Iriti, et al. [bib_ref] Induction of Resistance to Gray Mold with Benzothiadiazole Modifies Amino Acid Profile..., Iriti [/bib_ref] with minor modifications. 1 gm of fresh seedlings was crushed in 5 ml of 80% methanol. The samples were then centrifuged at 10,000 × g at 4 °C for 20 minutes. To 1 ml of supernatant, 1 ml of sulphosalicylic acid (6%) was added followed by centrifugation at 10,000 × g at 4 °C for 20 minutes. The prepared samples were filtered through 0.22 µm syringe filter. The 1 µL of this sample was injected in the sample vials of amino acid analyzer and were quantified. polyamine (spermidine) localization. Synthesis of Probe. The fluorescent probe employed for localization of spermidine was synthesized by following already reported procedure [bib_ref] Spermidine induced aggregation of terphenyl derivative: An efficient probe for detection of..., Tejpal [/bib_ref]. Reaction mixture comprising of formyl bonic acid, potassium carbonate, Pd (0) in 60 ml of dioxane-H 2 O and tetrabromo 18-crown-6/tetrabromobennzene were kept in N 2 atmosphere for 24 hrs and was continuously stirred at 80 °C. After 24 hrs, when the reaction was complete, the mixture was removed under pressure, to procure a residue to which water was added. Extraction of the aqueous layer was carried out by using 20 ml chloroform and the process was repeated thrice. A solid residue was obtained from the chloroform fraction by washing the fraction with water followed by drying by using sodium sulphate and finally was distilled under low pressure. Solid residue was then subjected to column chromatography to obtain pure compound by using ethyl acetate as an effluent. The probe was then www.nature.com/scientificreports www.nature.com/scientificreports/ crystallized by employing ethanol. The synthesized compound was characterized by 1 H NMR, 13 C NMR and Mass spectroscopic studies. Confocal imaging. A 5 mM concentration of probe (molecular weight-815.2) was prepared in dimethyl sulfoxide (DMSO). Localization of spermidine was done by following the method proposed by Singh et al. [bib_ref] Hexaphenylbenzene appended AIEE active FRET based fluorescent probe for selective imaging of..., Singh [/bib_ref] with slight modification. The roots of 10 days old seedlings of B. juncea were excised and incubated in the fluorescent probe for 30 minutes, followed by washing twice with phosphate buffer saline. The washed roots were then mounted on glass slides and visualized under confocal microscope. osmoprotectants. Osmoprotectant viz. total carbohydrate, reducing sugar, trehalose, glycine betaine and proline content were estimated employing double beam UV-Vis spectrophotometer. Total carbohydrate content. Total carbohydrate content was assessed by method of Hodge and Hofreiter 66 . 100 mg of dried plant samples were dipped in 2.5 N HCl and boiled in a water bath for 3 hr. The samples were then cooled and neutralized by addition of sodium carbonates. The sample volume was made upto 100 ml, followed by centrifugation for 20 minutes at 4 °C at 13, 000 × g. To 1 ml of above extract 4 ml of Anthrone reagent and was heated for 8 minutes. The samples were then cooled and observations were made at 630 nm. Glucose was used as standard. Reducing sugar estimation. The protocol proposed by Miller 67 was used for estimation of reducing sugar levels. 0.1 gm of dried plant material was extracted with 80% ethanol. 3 ml of 3,5-dinitrosalicylic acid (DNSA) was added to 3 ml of plant extract. DNSA reagent was prepared by dissolving 50 mg sodium sulphite, 1 gm DNSA and 200 mg of phenol crystals in 100 ml NaOH (1%) and the reaction mixture was stored at 4 °C. 1 ml of 40% potassium sodium tartarte was added to the reaction mixture. The absorbance of sample was read at 510 nm. Standard glucose concentrations were used for preparation of graph for estimation of reducing sugar levels. Trehalose Content. The method proposed by Trevelyan and Harrison 68 were used for estimation of trehalose levels. 500 mg of dried plant material was crushed in 80% ethanol, followed by centrifugation at 5000 × g for 10 minutes at 4 °C. 2 ml of tri-chloroacetic acid (0.5 M) and 4 ml of Anthrone reagent was added to 0.1 ml of supernatant. The absorbance of yellow green color was read at 620 nm. Trehalose was used as standard for preparation of graph of absorbance vs. standard trehalose concentration. Glycine Betaine. Estimation of glycine betaine levels was done by following the protocol of Grieve and Grattan 69 . 0.5 gm of dried plant material was extracted with 5 ml of 0.05% of toluene and distilled water mixture. The reaction mixture was incubated in dark for 24 hr and was filtered. 2 N HCl and 0.1 ml of PI were added to 0.5 ml of extract. The samples were incubated for one and half hr in an ice bath. To the above mixture, 2 ml of ice cold water and 10 ml of 1,2-dichloroethane was added. Upper layer was discarded and absorbance of lower layer was read at 365 nm. Proline content. Proline content was assessed by employing method of Bates, et al. [bib_ref] Rapid determination of free proline for water-stress studies, Bates [/bib_ref]. 0.5 gm of fresh seedlings were crushed in 3% of sulphosalicylic acid, followed by centrifugation at 13, 000 × g 4 °C for 20 minutes. 2 ml of ninhydrin reagent (1.56 gm of ninhydrin was dissolved in glacial acetic acid and 6 M ortho-phosphoric acid and was warmed and stored at 4 °C), was added and warmed for 1 hr in water bath. To stop the reaction the test tubes were immediately shifted to ice-bath, followed by addition of 4 ml toluene and shook for 30-40 seconds vigorously. The absorbance of toluene layer was read at 520 nm. statistical Analysis. Data obtained was statistically analyzed by self-coded MS Excel software. The data is presented in the form of mean ± standard deviation (SD). The data was also subjected to two-way analysis of variance (ANOVA) and tukey's test (honestly significant difference, HSD). Multiple Linear Regression (MLR) was employed to analyze the response of independent variables i.e. Pb, 24-EBL and SA. β regression coefficient values implied relative effect of independent variables X 1 -Pb, X 2 -24-EBL and X 3 -SA. Following equation was used to evaluate the response: [formula] = + + + Y a b X b X b X 1 1 2 2 3 3 [/formula] where,Y is parameter analyzed X 1 , X 2 , X 3 are independent variables i.e. Pb, 24-EBL and SA b 1 , b 2 , b 3 are partial regression coefficient due independent variables β 1 , β 2 , β 3 are regression coefficient values. # Results pb localization. The localization of Pb metal ions by Leadmium TM Green AM dye showed green fluorescenece when visualized under confocal microscope. It was observed that the intensity of fluorescence was higher in Pb (0.75 mM) treated seedlings when compared to control. Metal treated seedlings pre-imbibed in 24-EBL and SA, alone and in combination lowered the intensity of green fluorescence. Combined treatment of 24-EBL and SA were found to be most effective [fig_ref] Figure 1: Confocal laser scanning micrographs showing imaging of effect of pre-treatment t with... [/fig_ref] [fig_ref] Figure 1: Confocal laser scanning micrographs showing imaging of effect of pre-treatment t with... [/fig_ref]. Membrane damage. PI fluorescent probe was used to observe membrane damage in B. juncea root cells. PI dye has red fluorescence and the intensity of red color was enhanced in Pb (0.75 mM) treated plants in contrast to control seedlings implying enhanced membrane damage. Pre-treatment of seedlings with 24-EBL, SA and their combination led to reduction in membrane damage as indicated by lowered intensity of red fluorescence [fig_ref] Figure 1: Confocal laser scanning micrographs showing imaging of effect of pre-treatment t with... [/fig_ref]. Nuclear Damage. Nuclear damage was assessed by employing DAPI and has blue fluorescence. The intensity of fluorescence was maximum in Pb (0.75 mM) treated seedlings. 24-EBL and SA alone and in combination showed relative lower intensity of blue color as compared to 0.75 mM Pb treated seedlings. Whereas, 24-EBL + SA combined treatment were the most effective [fig_ref] Figure 2: Confocal laser scanning [/fig_ref]. MDA visualization. The spectrophotometric results of MDA content were further confirmed by is visualization by schiff 's reagent using visible compound microscope. The Pb (0.75 mM) stressed roots showed higher intensity of pink color of schiff 's reagent in comparison to control seedlings. Priming of seedlings with 24-EBL and SA and their combination showed reduction in MDA levels as compared to metal (0.75 mM Pb) treated seedlings as suggested by lowered intensity of pink color [fig_ref] Figure 2: Confocal laser scanning [/fig_ref]. Cell Viability. The damaging effect of Pb (0.75 mM) was also assessed by using Evan's blue dye to visualize cell viability. Lowered viability of cells was observed in (0.75 mM) Pb treated seedlings when compared to control seedlings as implied by darkly stained cells. The viability of 24-EBL, SA and 24-EBL + SA treated seedlings was enhanced as indicated by lowered intensity of Evan's blue [fig_ref] Figure 2: Confocal laser scanning [/fig_ref]. www.nature.com/scientificreports www.nature.com/scientificreports/ Glutathione tagging. The results of glutathione visualization revealed decline in accumulation of glutathione in 0.75 mM Pb stressed seedlings as evident by reduced fluorescence of MCB dye in comparison to control seedlings. Glutathione levels were enhanced in response to 24-EBL, SA and 24-EBL + SA pre-treatment as indicated by enhanced blue fluorescence [fig_ref] Figure 3: Confocal laser scanning micrographs showing imaging of effect of pre-treatment with EBL... [/fig_ref]. Quantitative and Qualitative amino acids levels. Total free amino acid level. The endogenous levels of free amino acids were lowered by 60.49% in 0.75 mM Pb treated seedlings in contrast to un-treated seedlings. Priming of seeds with 24-EBL, SA and their combination led to elevation in free amino acid content by 90.26%, 16.24% and 124.10% respectively, in contrast to 0.75 mM Pb treated seedlings [fig_ref] Figure 4: Effect of pre-treatment with EBL and SA on [/fig_ref]. Amino acid profiling. A total of 21 amino acids were detected as mentioned in [fig_ref] Table 2: Effect of pre-sowing treatment with combination of EBL and SA on contents... [/fig_ref]. Pb metal treatment resulted in decline in levels of specific amino acid including glutamine, GABA, methionine and leucine in comparison to control seedlings. Elevations in content of amino acids were observed when pre-treated with 24-EBL and SA individually as well as in combination. Levels of serine, glutamine, histidine, β-Alanine, GABA, methionine, isoleucine and leucine were further enhanced by co-application of 24-EBL and SA. polyamine (spermidine) imaging. In-situ localization of spermidine in roots of B. juncea was done using chemical probe, which showed blue fluorescence. It was observed that, 0.75 mM Pb treated plants had higher levels of polyamine when compared to control seedlings as indicated by darkly stained cells. Individual as well as combined application of 24-EBL and SA led to further elevation in accumulation of polyamines as evidenced by darkly stained cells [fig_ref] Figure 3: Confocal laser scanning micrographs showing imaging of effect of pre-treatment with EBL... [/fig_ref]. osmoprotectants. Total Carbohydrates Content. Pb (0.75 mM) treatment led to significant enhancement (195.60%) in total carbohydrate content in comparison to control seedlings. Seedlings treated with 24-EBL and SA individually and in combination led to further enhancement in levels of total carbohydrates by 179.9%, 17.48% and 207.80% respectively. Combined treatment of 24-EBL and SA was most effective in enhancing total carbohydrate levels [fig_ref] Figure 4: Effect of pre-treatment with EBL and SA on [/fig_ref]. . Effect of pre-treatment with combination of 24-EBL and SA on superoxide anion (µg g −1 FW), H 2 O 2 (µM g −1 FW) and MDA (mM g −1 FW) content in 10 days old B. juncea seedlings exposed to Pb. * Data is presented as mean ± SD. Two-way ANOVA, Tukey's test and MLR analysis was performed * and ** designated significant at P ≤ 0.05 and P ≤ 0.01 respectively. Statistical letters are mentioned HSD. www.nature.com/scientificreports www.nature.com/scientificreports/ Reducing sugar content. The seedlings raised under 0.75 mM Pb treatment had elevated levels of reducing sugars by 51.25% in comparison to control. Seedlings primed with in 24-EBL and SA individual treatment led to increase in reducing sugar levels by 222.40% and 146.20% respectively in contrast to 0.75 mM Pb treatment. Co-application of 24-EBL and SA maximally enhanced reducing sugar levels by 246.25% in comparison to metal (0.75 mM) Pb treated seedlings [fig_ref] Figure 4: Effect of pre-treatment with EBL and SA on [/fig_ref]. Trehalose content. Trehalose levels were significantly elevated in response to Pb treatment by 219.42% in contrast to control seedlings. Priming with 24-EBL, SA and 24-EBL + SA resulted in enhancement in trehalose levels by 28.72%, 8.06% and 140.80% respectively in comparison to 0.75 mM Pb treated seedlings [fig_ref] Figure 4: Effect of pre-treatment with EBL and SA on [/fig_ref]. www.nature.com/scientificreports www.nature.com/scientificreports/ Glycine betaine content. Content of glycine betaine was enhanced by 64.15% in Pb (0.75 mM) treated seedlings in contrast to control seedlings. Priming of seeds with 24-EBL and SA individual treatment further enhanced the levels of glycine betaine by 21.07% and 10.92% respectively in comparison to 0.75 mM Pb treated seedlings. Pb (0.75 mM) treated seedlings pre-imbibed in combined treatment of 24-EBL and SA led to most significant elevation i.e. by 40.61% respectively in glycine betaine content [fig_ref] Figure 4: Effect of pre-treatment with EBL and SA on [/fig_ref]. Proline level. Significant elevation in proline content was recorded in 0.75 mM Pb treated seedlings (237.91%) in comparison to control seedlings. Further increase in proline levels were recorded in 24-EBL, SA and 24-EBL + SA primed seedlings by 50.60%, 31.51% and 76.31% respectively when compared to 0.75 mM Pb treated seedlings [fig_ref] Figure 4: Effect of pre-treatment with EBL and SA on [/fig_ref]. # Discussion Pb is a non-essential element that prominently perturbs plants physiology. It is a less available and low solubility metal which occurs in phosphates, nitrates and sulphate forms [bib_ref] Hg(II) removal from water by chitosan and chitosan derivatives: A review, Miretzky [/bib_ref]. It is a extremely toxic metal and exerts certain critical effects on the physiological and biochemical attributes in plants [bib_ref] The alleviative effects of salicylic acid on the activities of catalase and..., Song [/bib_ref]. In-situ localization of Pb ions in roots revealed enhancement in Pb content in response to increment in metal concentration. Similar elevation in Pb www.nature.com/scientificreports www.nature.com/scientificreports/ content was reported in Oryza sativa [bib_ref] Lead (Pb) Toxicity; Physio-Biochemical Mechanisms, Grain Yield, Quality, and Pb Distribution Proportions..., Ashraf [/bib_ref] , Nicotiana tabaccum 74 , Medicago sativa [bib_ref] Characterisation of lead-induced stress molecular biomarkers in Medicago sativa plants, Hattab [/bib_ref] and Triticum aestivum [bib_ref] LIB spectroscopic and biochemical analysis to characterize lead toxicity alleviative nature of..., Tripathi [/bib_ref]. The bulk of Pb in soil is absorbed by roots of plants and bind to the carboxyl group of mucilage uronic acids [bib_ref] Measurement of Pb 2+ , Cu 2+ and Cd 2+ binding with..., Morel [/bib_ref]. A very small fraction of Pb is available for plants to accumulate due to the fact that it forms strong complexes with colloidal particles or organic matter. The endodermis of the root cells act as a barrier for movement of Pb from roots to shoots. This may be a possible reason for higher accumulation of Pb in roots when compared to shoots [bib_ref] Lead toxicity induces lipid peroxidation and alters the activities of antioxidant enzymes..., Verma [/bib_ref]. Plant growth regulators (PGRs) are well studied for their anti-metal stress activities [bib_ref] Minimising toxicity of cadmium in plants-role of plant growth regulators, Asgher [/bib_ref]. In present study pre-treatment of seedling with EBL, SA and their combination showed reduced Pb metal localization in root cells. This is attributed to BR stimulated activation of antioxidative defense system, lowered ROS content and improvement in growth of plants [bib_ref] The phytohormone crosstalk paradigm takes center stage in understanding how plants respond..., Kohli [/bib_ref] [bib_ref] Castasterone assisted accumulation of polyphenols and antioxidant to increase tolerance of B...., Yadav [/bib_ref]. Several reports suggest BR-induced reduction in metal upake such as Al in Phaseolus aureus [bib_ref] Brassinolide amelioration of aluminum toxicity in mungbean seedling growth, Abdullahi [/bib_ref] , Ni in Brassica juncea [bib_ref] Isolation and characterization of 24-Epibrassinolide from Brassica juncea L. and its effects..., Kanwar [/bib_ref] , Cr in Raphanus sativus [bib_ref] Effect of 28-homobrassinolide on antioxidant defence system in Raphanus sativus L. under..., Sharma [/bib_ref] and Cu in Brassica juncea [bib_ref] Effect of 24-epibrassinolide on seed germination, seedling growth and heavy metal uptake..., Sharma [/bib_ref]. Similarly, SA application to B. juncea seedlings resulted in lowering of Pb ion deposition in root cells. Transportation and accumulation of several metals have been reported to be affected by exogenous application of SA. Decline in metal content i.e. Cd in Zea mays [bib_ref] Salicylic Acid and Sodium Salicylate Alleviate Cadmium Toxicity to Different Extents in..., Gondor [/bib_ref] , Pb in B. juncea [bib_ref] Effect of salicylic acid on growth and oxidative metabolism of Brassica juncea..., Goel [/bib_ref] , As in Raphanus sativus [bib_ref] Exploring the role of salicylic acid to attenuate cadmium accumulation in radish..., Raza [/bib_ref] and Ni in Nicotiana tabaccum [bib_ref] Salicylic Acid Mitigates Pb Stress In Nicotiana Tabacum, Halim [/bib_ref] was reported in SA pre-treated plants. Reduced uptake of Pb in Chlorella vulgaris cells has been reported under 24-EBL treatment by Bajguz [bib_ref] Effect of brassinosteroids on nucleic acids and protein content in cultured cells..., Bajguz [/bib_ref]. He suggested that Pb metal in combination with EBL results in stimulation of phytochelatin de-novo synthesis. Similarly, Kaur et al. [bib_ref] Growth, photosynthetic activity and oxidative stress in wheat (Triticum aestivum) after exposure..., Kaur [/bib_ref] suggested that SA treatment leads to reduced metal uptake, possibly due to its ability to reduce oxidative stress and increased membrane stability [bib_ref] Growth, photosynthetic activity and oxidative stress in wheat (Triticum aestivum) after exposure..., Kaur [/bib_ref]. Exposure of B. juncea plants to Pb in the present study resulted in enhanced synthesis of ROS which further caused cellular damage in plants. Pb treatment resulted in elevation in levels of superoxide anion, H 2 O 2 and MDA levels. ROS production in response to Pb stress is well documented [bib_ref] Genotoxicity evaluation in workers occupationally exposed to lead, Grover [/bib_ref] [bib_ref] Heavy metals toxicity in plants: An overview on the role of glutathione..., Yadav [/bib_ref]. H 2 O 2 is easily converted to hydroxyl ion (free radical) by fenton reaction and is membrane permeable and diffusible [bib_ref] Singlet oxygen and photo-oxidative stress management in plants and algae, Ledford [/bib_ref]. The enhanced production of ROS might also enhance the tolerance mechanism in plants 92 . Asada-Halliwell pathway, with the co-ordination of enzymatic and non-enzymatic antioxidants, also plays an instrumental role in the removal of H 2 O 2 . Histological studies carried out also supported the biochemical responses and affirmed enhancement in total ROS, H 2 O 2 and MDA levels. Furthermore, cellular damage and loss of viability of cells caused by oxidative burst was also observed. The results of present study corroborated with the previous studies conducted on Ocimum tenuiflorum [bib_ref] Effect of chromium accumulation on photosynthetic pigments, oxidative stress defense system, nitrate..., Rai [/bib_ref] , Vigna radiata [bib_ref] Speciation dependant antioxidative response in roots and leaves of sorghum (Sorghum bicolor..., Shanker [/bib_ref] and Zea mays [bib_ref] Jasmonic acid induced changes in physio-biochemical attributes and ascorbate-glutathione pathway in Lycopersicon..., Bali [/bib_ref]. BRs application resulted in reduction in ROS levels and hence aided in lowering the toxic effects of Pb. Similar findings of lowered ROS production were reported by Sharma and Bhardwaj 83 in Zea mays plants by in Raphanus sativus, by Yadav, et al. [bib_ref] Castasterone assisted accumulation of polyphenols and antioxidant to increase tolerance of B...., Yadav [/bib_ref] and by in B. juncea. Reduction in Pb induced MDA, H 2 O 2 and superoxide anion content by BRs pre-treatment is attributed to their ability to enhance activity of antioxidative defense system and hence scavenging of free radicals. Interplay between SA signaling networks and other signaling cascade has been studied to regulate ROS stimulated ## Amino acid content (µg −1 g of fw) Aspartic www.nature.com/scientificreports www.nature.com/scientificreports/ cell death [bib_ref] Reactive oxygen species and hormonal control of cell death, Overmyer [/bib_ref]. The observation was also affirmed with histological studies which showed reduction in ROS content, nuclear and membrane damage and altered cell viability. Sulfur is one of the most imperative elements that is incorporated in several biomolecules including amino acids and antioxidants [bib_ref] Antioxidant Effects of Sulfur-ContainingAmino Acids, Atmaca [/bib_ref]. GSH is one of the sulfur containing antioxidant. Present work further revealed reduction in levels of glutathione in response to Pb toxicity. Our findings corroborated the studies of Okamoto, et al. [bib_ref] Antioxidant Modulation in Response to Metal-Induced Oxidative Stress in Algal Chloroplasts, Okamoto [/bib_ref] who reported lowered GSH in Gongauloux polyedea exposed to Pb stress. The decline in GSH levels is due to lowered de-novo production of GSH as well as over-utilization of GSH in the synthesis of phytochelatins, pre-requisite for chelation of toxic metal ions [bib_ref] Pb-Stress Induced Oxidative Stress Caused Alterations in Antioxidant Efficacy in Two Groundnut, Nareshkumar [/bib_ref]. This antioxidant fortifies protein oxidation and acts as a substrate for production of glutathione peroxidase and glutathione-S-transferase [bib_ref] Interactions between biosynthesis, compartmentation and transport in the control of glutathione homeostasis..., Noctor [/bib_ref]. In addition, heavy metals can alter the active sites of the enzymes, thus rendering them nonfunctional [bib_ref] Effects of heavy metal contamination upon soil microbes: lead-induced changes in general..., Sobolev [/bib_ref]. Exogenous application of combination of 24-EBL and SA resulted in enhancement in levels of GSH in the present study. Exogenous application of BRs to stressed plants elevated the contents of glutathione in Lycopersicon esculentum exposed to Cd and Pb 30 , Cicer arietinum exposed to Cd [bib_ref] 28-Homobrassinolide protects chickpea (Cicer arietinum) from cadmium toxicity by stimulating antioxidants, Hasan [/bib_ref] and Vigna radiata exposed to Al 105 . The possible reasons for BR induced increase in the activities of antioxidative enzymes might be due to their ability to regulate the transcription and/or translation of genes that further mediate the activation, or de novo synthesis of antioxidative enzymes [bib_ref] Effect of brassinosteroids on nucleic acids and protein content in cultured cells..., Bajguz [/bib_ref]. Similarly, application of SA was also reported to enhance endogenous levels of GSH in Medicago sativa plants under Mg stress. SA primed plants had enhanced tolerance to Cd in Triticum aestivum plants which was largely co-related to enhanced GSH synthesis. Furthermore, several studies suggested regulation of various enzymes of AsA-GSH (ascorbate-glutathione) cycle by SA application [bib_ref] Cadmium Toxicity and Alleviating Effects of Exogenous Salicylic Acid in Iris hexagona, Han [/bib_ref] [bib_ref] Salicylic acid enhances antioxidant system in Brassica juncea grown under different levels..., Parashar [/bib_ref]. Another possible reason for enhanced activity of antioxidative defense system might be due to BR Signaling Kinase 1 (BSK 1) which is a positive regulator of SA accumulation in stressed plants and eventually results in amelioration of oxidative damage [bib_ref] Role of brassinosteroid signaling in modulating Tobacco mosaic virus resistance in Nicotiana..., Deng [/bib_ref]. Amino acids specifically S containing amino acids have anti-stress and antioxidant properties. These organic amino acids include glutathione and methionine [bib_ref] The Sulfur-Containing Amino Acids: An Overview, Brosnan [/bib_ref]. In the present study, the levels of free amino acids and other amino acids was lowered in response to Pb stress. Amino acids have metal chelating, antioxidant and signaling ability under metal stress [bib_ref] The significance of amino acids and amino acid-derived molecules in plant responses..., Sharma [/bib_ref]. The probable reason for lowered amino acid levels might be due to enhanced chelation of bivalent metal ions with free amino acid which consequently lowers available amino acids 111 . Co-application of BRs and SA resulted in elevation in free amino acid levels. Our findings corroborated with the study carried out by Li, et al. [bib_ref] Lead tolerance mechanism in Conyza canadensis: subcellular distribution, ultrastructure, antioxidative defense system,..., Li [/bib_ref] who reported enhancement in amino acids content in response to 24-EBL application. Similarly, application of SA resulted in elevation in levels of amino acids specifically proline, glycine betaine and glutathione in the present study. Few specific amino acids such as proline and glycine betatine also act as osmoprotectants [bib_ref] Phytohormones and their metabolic engineering for abiotic stress tolerance in crop plants, Wani [/bib_ref]. Combined application of 24-EBL and SA led to enhancement in osmolyte levels in the present study. Similar elevation in glycine betaine levels were observed in response to CO, Fe and Zn stress [bib_ref] Assessment of Osmolyte Accumulation in Heavy Metal Exposed Salvinia natans, Dhir [/bib_ref]. Glycine betaine levels are known to alter the membrane permeability of stressed plant cells [bib_ref] Role of glutathione and polyadenylic acid on the oxidative defense systems of..., Khattab [/bib_ref]. Enhanced levels of proline and glycine betaine resulted in further elevation in antioxidative capacities of berries [bib_ref] Regulating the secondary metabolism in grape berry using exogenous 24-epibrassinolide for enhanced..., Xi [/bib_ref]. Proline content significantly elevated by BRs application is well documented [bib_ref] 24-Epibrassinolide regulates photosynthesis, antioxidant enzyme activities and proline content of Cucumis sativus..., Fariduddin [/bib_ref] [bib_ref] Effect of 28-homobrassinolide on antioxidant defence system in Raphanus sativus L. under..., Sharma [/bib_ref] [bib_ref] Changes induced by Cu 2+ and Cr 6+ metal stress in polyamines,..., Choudhary [/bib_ref] [bib_ref] Epibrassinolide ameliorates Cr (VI) stress via influencing the levels of indole-3-acetic acid,..., Choudhary [/bib_ref]. SA also resulted in further enhancement in proline and glycine betaine levels and they aid in detoxification of elevated ROS content, maintenance of membrane stability and enzyme activities [bib_ref] Exogenously-sourced ethylene increases stomatal conductance, photosynthesis, and growth under optimal and deficient..., Iqbal [/bib_ref]. It was observed by Parmar et al. [bib_ref] Structural and functional alterations in photosynthetic apparatus of plants under cadmium stress, Parmar [/bib_ref] , that the main reason for increase in levels of proline content is associated with increase in synthesis of new amino acids under heavy metal stress. Elevation in concentration of osmolytes is considered as an important marker indicating metal stress and have important role in stress mitigation. In-situ localization of polyamines (spermidine) revealed elevation in polyamines accumulation in Pb stressed seedlings. Polyamines specifically spermine and spermidine have antioxidant properties and protect DNA against oxidative damage 120 . Similar to our findings, Groppa, et al. [bib_ref] Polyamines and heavy metal stress: the antioxidant behavior of spermine in cadmium-and..., Groppa [/bib_ref] reported elevation in polyamines content in response to Cu and Cd toxicity. Another report suggests similar decline in polyamine accumulation in Lacuta sativa plants under Zn stress 122 . Co-application of 24-EBL and SA significantly elevated in-situ deposition of polyamines in root cells of B. juncea in the present study. In our earlier study, Choudhary, et al. [bib_ref] Interaction of Brassinosteroids and Polyamines Enhances Copper Stress Tolerance in. Raphanus Sativus, Choudhary [/bib_ref] reported positive crosstalk between BRs and polyamines in regulating Cu stress in Raphanus sativus plants. It was further suggested that synergistic interplay between BRs and polyamines (spermidine) led to enhancement in levels of other polyamines such as putrecine and spermine. Similarly, exogenous application of SA resulted in higher level of polyamines in Zea mays [bib_ref] Priming seed with salicylic acid increases grain yield and modifies polyamine levels..., Szalai [/bib_ref] and Lycopersicon esculentum [bib_ref] Comparison of polyamine metabolism in tomato plants exposed to different concentrations of..., Takács [/bib_ref]. Sugars and sugar polyols accumulate in response to stress and they act as an imperative biomarker and osmoregulator [bib_ref] Synthesis of Conjugated Polymers for Organic Solar Cell Applications, Cheng [/bib_ref]. Levels of total carbohydrate, reducing sugars and trehalose were significantly increased in response to Pb treatment. Sugars play a vital role in modulating plants osmotic balance and membrane stability in stressed plants. It has been hypothesized that heavy metal toxicity might hinder the metabolic pathway of carbohydrates or it might play a role in the accumulation of photoassimilates because of reduced loading of veins [bib_ref] Vein loading in seedlings of Phaseolus vulgaris exposed to excess cobalt, nickel,..., Rauser [/bib_ref]. Plethora of reports suggest enhancement in sugar levels due to Pb toxicity in B. juncea [bib_ref] Combined effect of 24-epibrassinolide and salicylic acid mitigates lead (Pb) toxicity by..., Kohli [/bib_ref] and Raphanus sativus [bib_ref] Amelioration of lead toxicity in radish (Raphanus sativus L) plants by brassinolide, Anuradha [/bib_ref]. It was further suggested by Bhushan and Gupta 130 that enhanced levels of total carbohydrates and reducing sugar could be due to meddling of Pb ions with transportation through endodermis into plant cells causing severe toxicity. Pre-treatment with 24-EBL and SA led to further increase in content of sugars. Similar to amino acids, sugars also have stress mitigating properties via enhanced sequestration of ROS molecules as well as activation of antioxidant defense system [bib_ref] Interaction effects of nitric oxideand salicylic acid in alleviating salt stress of..., Dong [/bib_ref]. It has been reported that sugars acts as osmolytes and provide protection to cells from metal toxicity [bib_ref] Effect of arsenic on behaviour of enzymes of sugar metabolism in germinating..., Jha [/bib_ref] [bib_ref] Brassinosteroids-induced systemic stress tolerance was associated with increased transcripts of several defencerelated..., Li [/bib_ref]. These sugars not only help as osmolytes, participate in stabilization of cellular membranes and maintenance of turgor but also act as signaling molecules [bib_ref] AMF-induced tolerance to drought stress in citrus: A review, Wu [/bib_ref]. # Conclusion The present study indicates that co-application of 24-EBL and SA to B. juncea L. plants may enhance their potential to overcome Pb stress. The positive interpay between 24-EBL and SA led to elevation in total free amino acids, content of various amino acids (glutathione, proline, methionine etc.), which act as antioxidants and metal chelating compounds. These metabolites scavenge ROS and also result in reduced lipid peroxidation, nuclear and www.nature.com/scientificreports www.nature.com/scientificreports/ membrane damage. From the results, it is further concluded that 24-EBL and SA combined treatment has a significant role in regulating the balance of total carbohydrates, reducing sugars and osmoprotectants like trehalose, glycine betaine and proline which are involved in osmoregulation of plant cells. The co-application of 24-EBL and SA reduced Pb accumulation in plants that consequently results in lowering oxidative stress. Therefore, combined treatment of 24-EBL and SA can counter adverse effects of Pb toxicity through regulating various physiological processes. Further insight into mechanisms of 24-EBL and SA crosstalk pertaining to metal stress amelioration might provide better understanding of stress tolerance strategies in plants. [fig] Figure 1: Confocal laser scanning micrographs showing imaging of effect of pre-treatment t with EBL and SA on: (a) Pb localization, (b) H 2 O 2 content and (c) membrane damage in roots of 10 days old B. juncea seedlings under Pb stress. Scale bar = 100 μm, (CN = control, Pb = lead, 24-EBL = 24-epibrassinolide, SA = salicylic acid). [/fig] [fig] Figure 2: Confocal laser scanning (a) and visible (b,c) micrographs showing imaging of effect of pre-treatment with EBL and SA on: (a) nuclear damage, (b) MDA content and (c) cell viability in roots of 10 days old B. juncea seedlings under Pb stress. Scale bar = 100 μm (confocal microscope), (CN = control, Pb = lead, 24-EBL = 24-epibrassinolide, SA = salicylic acid). [/fig] [fig] Figure 3: Confocal laser scanning micrographs showing imaging of effect of pre-treatment with EBL and SA on: (a) glutathione content and (b) polyamine (spermidine) content in roots of 10 days old B. juncea seedlings under Pb stress. Scale bar = 100 μm, (CN = control, Pb = lead, 24-EBL = 24-epibrassinolide, SA = salicylic acid). [/fig] [fig] Figure 4: Effect of pre-treatment with EBL and SA on: (A) free amino acid, (B) total carbohydrate, (C) reducing sugar, (D) trehalose, (E) glycine betaine and (F) proline (µg g −1 DW) content in roots of 10 days old B. juncea seedlings under Pb stress, (CN = control, 24-EBL = 24-epibrassinolide, SA = salicylic acid). [/fig] [table] Table 2: Effect of pre-sowing treatment with combination of EBL and SA on contents of amino acids (µg g −1 FW) in 10 days old B. juncea seedlings under Pb stress. [/table]

Cytokines as Biomarkers and Their Respective Clinical Cutoff Levels Cytokines, including interleukins, interferons, tumor necrosis factors, and chemokines, have a variety of pro-and anti-inflammatory effects in the body through a number of biochemical pathways and interactions. Stimuli, actions, interactions, and downstream effects of cytokines have been investigated in more depth in recent years, and clinical research has also been conducted to implicate cytokines in causal patterns in certain diseases. However, particular cutoffs of cytokines as biomarkers for disease processes have not been well studied, and this warrants future work to potentially improve diagnoses for diseases with inflammatory markers. A limited number of studies in this area are reviewed, considering diseases correlated with abnormal cytokine profiles, as well as specific cutoffs at which cytokines have been deemed clinically useful for diagnosing those diseases through Receiver Operator Characteristics modeling. In light of studies such as those discussed in this review, cytokine testing has the potential to support diagnosis due to its lack of invasiveness and low cost, compared to other common types of testing for infections and inflammatory diseases. # Introduction Cytokines and Their Role in Human Biology and Disease Processes. Cytokines are small, nonstructural proteins, including interleukins, chemokines, interferons, and tumor necrosis factors, which have a multitude of pleiotropic effects in various organs [bib_ref] Historical insights into cytokines, Dinarello [/bib_ref]. They are released in a number of paracrine, autocrine, or endocrine pathways and have been implicated in a variety of infections and immune system-affecting disorders by both proinflammatory and anti-inflammatory mechanisms. Cytokines which have proinflammatory effects include interferon-(IFN-) , interleukin-(IL-) 17, IL-1 , and tumor necrosis factor-(TNF-) [bib_ref] Roles of pro-and anti-inflammatory cytokines in the pathogenesis of SLE, Su [/bib_ref] [bib_ref] Increased pro-inflammatory cytokines (TNF-and IL-6) and anti-inflammatory compounds (sTN-FRp55 and sTNFRp75) in..., Pinto [/bib_ref] , and those with anti-inflammatory effects include IL-10, IL-4, and IL-1ra [bib_ref] Roles of pro-and anti-inflammatory cytokines in the pathogenesis of SLE, Su [/bib_ref] [bib_ref] Regulation by anti-inflammatory cytokines (IL-4, IL-10, IL-13, TGF ) of interleukin-8 production..., Marie [/bib_ref]. However, the distinction between pro-and antiinflammatory cytokine effects is not always entirely clear: pathway interactions play a major role, as individual and a combination of several cytokines can contribute to upregulation or downregulation of other cytokines and certain cytokines can have both pro-and anti-inflammatory effects [bib_ref] Pro-versus anti-inflammatory cytokines: myth or reality, Cavaillon [/bib_ref]. Two of the most important cytokine effector pathways are the JAK-STAT and NF-B pathways. These pathways are activated by cytokine ligands and are also regulated by and stimulate further release of cytokines. When a cytokine binds to a receptor in the JAK-STAT pathway, the receptor dimerizes and JAKs are activated. The JAK proteins then phosphorylate and activate the receptor as well as the STATs which are now associated with the activated receptor. This allows the STATs to dimerize and travel to the nucleus to regulate gene expression [bib_ref] Interleukin-6-type cytokine signalling through the gp130/Jak/STAT pathway, Heinrich [/bib_ref]. In the canonical NF-B pathway, I B proteins inhibit the B protein, in the absence of ligand, preventing B from activating transcription of genes involved in inflammation and stress responses. When cytokines are present, they serve as ligands to bind to and activate IKK complexes, which then phosphorylate I B proteins, targeting them for degradation and therefore freeing NF-B transcription factors to locate to the nucleus [bib_ref] Shared principles in NF-B signaling, Hayden [/bib_ref]. However, these pathways differ upon interaction of cytokines. When IL-12 is a ligand for the JAK-STAT pathway in macrophages, a proinflammatory response of Th1 is produced, but when IL-10 is present, IL-12's proinflammatory effects are downregulated [fig_ref] Figure 1: IL-12 activation of JAK-STAT pathway [/fig_ref]. IL-10 has similar effects on other proinflammatory cytokines, reducing or even terminating their inflammatory responses [bib_ref] IL-10, an inflammatory/inhibitory cytokine, but not always, Conti [/bib_ref]. In studies of TNFas a ligand in mast cells for the NF-B pathway, the result is a proinflammatory cascade [bib_ref] IL-10, an inflammatory/inhibitory cytokine, but not always, Conti [/bib_ref]. Yet when that same ligand is present along with IL-10 in the pathway, this becomes muted most likely due to inhibition of NF-B, yielding a smaller inflammatory response . Similar to IL-10, IL-6 also is involved in various pathways which alter, or are altered by, other cytokines. IL-6 helps upregulate IL-21 for improved activity of CD4+ cells, suppresses IFN-signaling to aid in T-cell differentiation, and regulates TGF-mediation of CD4+ differentiation, demonstrating the ability of multiple cytokines to differentially regulate a single pathway [bib_ref] The pro-and anti-inflammatory properties of the cytokine interleukin-6, Scheller [/bib_ref]. Further, differences in types of signaling in these pathways can also affect whether the cytokines behave in a proinflammatory or anti-inflammatory matter. IL-6 is typically thought of as proinflammatory, though it can also have anti-inflammatory effects depending on how its receptors and signaling receptor proteins interact. IL-6 signaling in the JAK-STAT pathway can behave proinflammatorily when the signaling receptor protein and IL-6 receptor are located on the same cell (classic signaling), yet the same cytokine behaves anti-inflammatorily when IL-6 binds to a soluble IL-6 receptor which activates the membrane-bound signaling receptor protein (transactivation) [bib_ref] The pro-and anti-inflammatory properties of the cytokine interleukin-6, Scheller [/bib_ref]. In a related way, IL-10 is most often thought of as an anti-inflammatory cytokine, though depending on its target cell and concentration, it can also act to promote inflammation. Activation of certain STAT proteins (STAT3 in the JAK-STAT pathway, which has anti-inflammatory effects) and inhibition of NF-B (through suppression of IKK complexes) by IL-10 play a role in inhibiting the immune response [bib_ref] Interleukin-10: new perspectives on an old cytokine, Mosser [/bib_ref]. However, IL-10 has also been found to exert proinflammatory effects, stimulating immune cells including B cells and cytotoxic T cells, in high concentration [bib_ref] IL-10, an inflammatory/inhibitory cytokine, but not always, Conti [/bib_ref]. Due to their vast pro-and anti-inflammatory effects, cytokines have been implicated in various disease processes. However, it is often difficult to use cytokines as diagnostic tools, although recent studies have investigated cytokine clinical diagnostic cutoffs. The limited amount of work in this area warrants future investigation to confirm diagnostic guidelines and explore guidelines for not-yet-studied cytokine-disease associations. ## The current state of disease diagnosis ## Defining normal and abnormal levels of cytokines. It is particularly challenging to evaluate cytokines' diagnostic International Journal of Inflammation 3 ability due to the difficulty of establishing "normal" versus "abnormal" cytokine levels. Cytokines vary greatly among individuals, and their release and subsequent effects can differ based upon activating signals, specific cell targets, and physiological factors including stress, fitness level, and feeding state [bib_ref] Conceptual and methodological issues relevant to cytokine and inflammatory marker measurements in..., Zhou [/bib_ref]. Cytokines also can vary in different physiological locations and environments, and thus studies that measure cytokines in abnormal and normal circumstances must only compare results with other studies of the same biological fluid (e.g., serum, amniotic fluid, and pleural fluid). Furthermore, only few studies have been conducted to investigate cytokine levels in healthy subjects, and there have been a limited number of variables explored when considering healthy subjects' cytokine profiles [bib_ref] Serum cytokine profiles in healthy young and elderly population assessed using multiplexed..., Kim [/bib_ref] [bib_ref] Cytokine levels in the serum of healthy subjects, Kleiner [/bib_ref] [bib_ref] Normal physiological levels of human cytokines using Bio-Plex Pro Cytokine Assays, Chapman [/bib_ref]. Therefore, most studies have established differing "normal" cytokine profiles based on the characteristics of their study populations and the modes of cytokine measurement. The factors are as follows: (a) variation exists in what is considered to be a "normal" cytokine profile; (b) few conclusions have been drawn across studies to define normal cytokine levels; and (c) a variety of factors contribute to cytokine release and action. Hence, these studies often define cytokine levels only within the population of interest. Most studies do not use preestablished cutoffs for cytokine reference values but rather consider the median or mean cytokine levels of healthy subjects in a defined set of population to be "normal." This mean or median cytokine level for a particular study is then used as a reference cutoff to identify comparatively abnormal cytokine levels in diseased patients; that is, levels are only considered abnormal if they differ from the mean of that population by approximately 2 standard deviations, though this is not diagnostically reliable. Therefore, it is necessary to conduct a greater number of studies within various populations, controlling the aforementioned factors, in order to establish normal cytokine levels and therefore have a more uniform reference for "abnormal" cytokine levels for use in supplementing clinical diagnosis. A number of bioassay and immunoassay methods are used in clinical practice to detect cytokines currently, though immunoassays are most often used because of their specificity for individual cytokines [bib_ref] The clinical usefulness of the measurement of cytokines, Bienvenu [/bib_ref]. Common immunoassay techniques include enzyme-linked immunosorbent assays (ELISA), multiplex arrays, bead-based assays, and the recently studied immunosensing method [bib_ref] Conceptual and methodological issues relevant to cytokine and inflammatory marker measurements in..., Zhou [/bib_ref]. Cytokines can also be measured indirectly, using mRNA transcripts, though these are not always an indication of cytokine activity and instead represent the potential for cytokine production. Therefore, direct protein detection uses are most often used for improved interpretation of physiologic cytokine activity. Direct immunodetection assays vary in consistency and reliability, cost, required time, ease of use, throughput, and sample volumes required, among other factors. Currently, the most popular way to detect cytokines is through ELISA, which involves immobilization of proteins and their detection using antibodies, either directly or indirectly [bib_ref] Conceptual and methodological issues relevant to cytokine and inflammatory marker measurements in..., Zhou [/bib_ref]. Although this method is commonly used, it is time-consuming and permits analysis of only one cytokine at a time. A method which has improved upon these drawbacks is the multiplex array, a similar protein detection method which permits measurement of multiple cytokines at once [bib_ref] ELISA and multiplex technologies for cytokine measurement in inflammation and aging research, Leng [/bib_ref]. A more recent approach to detecting cytokines is known as immunosensing, which transduces antigen-antibody interactions into electrical signals, though this technique is not yet well studied [bib_ref] Recent advances in cytokine detection by immunosensing, Liu [/bib_ref]. As such methods for cytokine detection improve, the complexity of cytokine interactions and effects can be more accurately portrayed. ## Establishing clinical cutoffs using receiver Operating Characteristics Analysis. Due to lack of conclusiveness regarding normal and abnormal cytokine levels, particular clinical cytokine cutoffs for disease states are difficult to establish. In other words, cutoffs for normal and abnormal cytokines in disease states need to be more well-established in order to more accurately support and distinguish between diagnoses, as well as estimate prognoses. A common way of evaluating diagnostic accuracy of individual biomarkers is by using Receiver Operating Characteristics Analysis (ROC analysis), which plots the true positive rate ("benefits") versus the false positive rate ("costs") of a particular disease at different cutoffs of the implicated biomarker [bib_ref] An introduction to ROC analysis, Fawcett [/bib_ref]. This type of analysis indicates the levels of the biomarker that are most diagnostically useful, allowing the ruling out of disease, through sensitivity values, or allowing disease to be essentially confirmed, through specificity values. An optimal cutoff can be established by identifying the point on the ROC curve with the highest sensitivity and specificity, for maximum diagnostic discriminatory ability [bib_ref] Understanding receiver operating characteristic (ROC) curves, Fan [/bib_ref]. A limited number of studies have utilized ROC analysis to evaluate diagnostic utility of cytokines for particular disease states. The majority of conditions for which cytokine associations and their clinical cutoffs have been explored can be divided into three categories: infections and postoperative infections, inherited and chronic diseases, and obstetric and gynecological conditions. These associations and clinical cutoffs will be discussed below, along with their implications for clinical testing and future work. ## Role of cytokines and their clinical cutoffs in infection diagnosis and postprocedural infection diagnosis Correlations of abnormal cytokine profiles with infection, including tuberculosis, pneumonia, and systemic inflammatory response state (SIRS), have been investigated in various studies. Hospital-acquired and postoperative infections were also found to have associations with abnormal cytokine profiles, including neonatal sepsis, periprosthetic joint infection, and postcervical neck dissection infection. More importantly, beyond simple associations between abnormal cytokine profiles and these infections, a limited number of studies have also investigated clinical cytokine cutoffs to support diagnosis, which is unprecedented in various other disease states. effusions are known to have significantly higher pleural effusion cytokine levels than non-Tb patients [bib_ref] Assay of pleural fluid interleukin-6, tumuor necrosis factor-alpha and interferon-gamma in the..., Wong [/bib_ref]. The study established cutoffs with considerable predictive and diagnostic value for Tb with pleural effusions: 4000 pg/mL for IL-6 (sensitivity: 90.6%; specificity: 76.5%), 4 pg/mL for TNF-(sensitivity: 90.6%; specificity: 79.4%), and 60 pg/mL for IFN-(sensitivity: 100%; specificity: 100%). The following year, Sharma and Banga (2004) specifically investigated IFNas a predictor of Tb pleural effusions [bib_ref] Diagnostic utility of pleural fluid IFN-in tuberculosis pleural effusion, Sharma [/bib_ref]. In this study population, the IFN-levels of Tb patients were significantly higher than healthy individuals (1480 pg/mL versus 3 pg/mL, resp.), and it was determined that the best cutoff of pleural fluid IFN-to predict Tb pleural effusion was 138 pg/mL (AUC 95.4%, sensitivity: 90.2%, specificity: 97.3%). The study also examined IFN-in peritoneal fluid ascites to determine cytokine clinical cutoffs for Tb and discovered an optimal cutoff for IFN-in peritoneal ascites of 112 pg/mL (AUC 99.0%, sensitivity: 97%, specificity: 97%). However, IFNassays using both pleural effusions [bib_ref] Diagnostic utility of pleural fluid IFN-in tuberculosis pleural effusion, Sharma [/bib_ref] and peritoneal ascites [bib_ref] Diagnostic accuracy of ascitic fluid IFN-and adenosine deaminase assays in the diagnosis..., Sharma [/bib_ref] are considerably expensive, which may render this method less clinically useful and practical. [bib_ref] Diagnostic accuracy of cytokine levels (TNF-, IL-2 and IFN-) in bronchoalveolar lavage..., Küpeli [/bib_ref] investigated a less expensive method of testing for cytokines as diagnostic biomarkers of Tb, using serum and bronchoalveolar lavage fluid (BALF) [bib_ref] Diagnostic accuracy of cytokine levels (TNF-, IL-2 and IFN-) in bronchoalveolar lavage..., Küpeli [/bib_ref]. This study found a TNF-cutoff of 17.6 pg/mL in serum and BALF (sensitivity: 73%; specificity: 76%) for distinguishing patients with smearnegative Tb from healthy subjects. However, [bib_ref] Diagnostic accuracy of cytokine levels (TNF-, IL-2 and IFN-) in bronchoalveolar lavage..., Küpeli [/bib_ref] did not find significant differences between smearnegative Tb patients and non-Tb groups when distinguishing using IL-12 and IFN-and therefore did not evaluate clinical diagnostic cutoffs for these cytokines [bib_ref] Diagnostic accuracy of cytokine levels (TNF-, IL-2 and IFN-) in bronchoalveolar lavage..., Küpeli [/bib_ref]. Most recently, [bib_ref] Diagnostic role of inflammatory and anti-inflammatory cytokines and effector molecules of cytotoxic..., Shu [/bib_ref] proposed that IFN-cannot be used alone and demonstrated that, in order to improve the IFN-model, DcR3 and TNF-sR1 should be included when developing an ROC curve for predicting Tb pleural effusion, suggesting that additional factors must be considered when using cytokines as diagnostic tools [bib_ref] Diagnostic role of inflammatory and anti-inflammatory cytokines and effector molecules of cytotoxic..., Shu [/bib_ref]. ## Tuberculosis, pneumonia, and sirs Cytokines were also investigated in ventilator-associated pneumonia (VAP), a disease with similar clinical signs and symptoms to Tb. A 2009 study found that VAP can be predicted and diagnosed using only serum levels of IL-6, as other cytokines did not prove to be associated or predictive of VAP [bib_ref] Systemic inflammatory response and increased risk for ventilator-associated pneumonia: a preliminary study, Ramírez [/bib_ref]. To distinguish between patients who did and did not subsequently develop VAP, a baseline IL-6 cutoff value of 198 pg/mL was determined (sensitivity: 71%; specificity: 78%). To distinguish patients with confirmed VAP versus suspected VAP, an IL-6 cutoff for the disease state was established to be 620 pg/mL (sensitivity: 71%; specificity: 89%). Morris et al. identified two additional serum cytokines, which can be used to predict VAP, IL-1 and IL-8, among others, including IL-6, TNF-, and IL-10 [bib_ref] Diagnostic importance of pulmonary interleukin-1 and interleukin-8 in ventilator-associated pneumonia, Morris [/bib_ref]. Clinical cutoff values of IL-1 and IL-8 for diagnosing VAP were 10 pg/mL (sensitivity: 94%; specificity: 64%) and 2000 pg/mL (sensitivity: 81%; specificity: 83%), respectively. Although these studies suggest a particular cytokine profile for identifying VAP, it is still necessary to distinguish the condition from Tb due to their similar symptoms and clinical signs. The previously mentioned studies can identify VAP, or Tb, though they cannot necessarily distinguish between the two. Su et al. (2010) investigated cytokines that may be able to distinguish between the two conditions to confirm diagnosis of either Tb or pneumonia, finding IFN-and IL-12 to yield considerable results through ROC analysis [bib_ref] Identification of cytokines in whole blood for differential diagnosis of tuberculosis versus..., Su [/bib_ref]. When cells were stimulated with ESAT-6, a cutoff of a 3.59% change in IFN-was found to yield sensitivity, specificity, and accuracy of over 80% for diagnosing Tb (90.4% AUC). When diagnosing pneumonia in cells stimulated with LPS, a cutoff of 3.59% change in IFN-produced sensitivity, specificity, and accuracy of 80% (89.1% AUC), and a cutoff of 2.08% change in IL-12 produced sensitivity of 80%, specificity of 78.9%, and accuracy of 79.4% (85.2% AUC). These results suggest that pneumonia and Tb may be distinguished diagnostically by cytokine responses of IFN-and IL-12 upon cell stimulation by different reagents and also suggest such laboratory testing to rapidly support diagnosis of these diseases in light of the slow rate at which other clinical features (such as bacterial infection) may be measured and observed. Another related form of inflammation with respect to cytokines is the limited published work regarding SIRS state. This is an important area of research to explore because systemic inflammation can be caused by a variety of factors, including trauma, and may lead to other previously discussed infections (such as pneumonia). In a distinctive study, [bib_ref] Correlation between IL-6 levels and the systemic inflammatory response score: can an..., Giannoudis [/bib_ref] found that IL-6 was predictive of a SIRS state at all points following hospital admittance for trauma (femoral shaft fracture) [bib_ref] Correlation between IL-6 levels and the systemic inflammatory response score: can an..., Giannoudis [/bib_ref]. At days 0 and 1 after admittance, a cutoff of 20,000 pg/mL IL-6 diagnosed a SIRS state (83% sensitivity; 75% specificity), and an IL-6 level above 300 pg/mL in SIRS patients was correlated with larger risks of complications, including pneumonia and death. Therefore, diagnosing a SIRS state might become more feasible, if more work is done to investigate these cutoffs, which will help prevent further complications that result from such an inflammatory state. ## Neonatal sepsis. A more well-studied phenomenon with respect to cytokine clinical cutoff values is neonatal sepsis, which is a bacterial bloodstream infection that typically appears in infants within twenty-four hours of birth. A litany of studies has connected neonatal sepsis to abnormal cytokine profiles, though only a small number have investigated specific cytokine cutoffs for diagnosing neonatal infection. Yet compared to other conditions, neonatal sepsis has been more thoroughly studied with respect to cytokine cutoffs and thus is an excellent example for the direction in which other studies can be conducted in the future. Three cytokines, IL-6, TNF-, and IL-1 , were investigated and strongly implicated as diagnostic tools in neonatal sepsis [bib_ref] Diagnosis of late onset neonatal sepsis with cytokines, adhesion molecule, and C-reactive..., Ng [/bib_ref]. The study established cutoffs facilitating diagnosis on day one of infection: IL-6: 31 pg/mL; TNF-: 17 pg/mL; IL-1 : 1 pg/mL. Combination of multiple markers enhanced accuracy of the tests, particularly when IL-6 and TNFwere combined (sensitivity: 95%; specificity: 84%). Timing and duration of infection upon testing seemed to affect diagnostic capabilities of these tests as well: IL-6 appeared to have greater diagnostic value on the day of infection onset (sensitivity: 89%; specificity: 96%), and TNF-had greater diagnostic value on the day after infection onset (sensitivity: 82%; specificity: 93%). Notably, optimal calculated values for clinical cutoffs differed considerably from cutoff values recommended by the manufacturer of the serum test, which suggests a need for future similar work to confirm cytokine cutoffs in establishing detection limits. Recent studies have also established highly sensitive and specific IL-6 cutoffs for diagnosing neonatal sepsis [bib_ref] Diagnostic value of cytokines and c-reactive protein in the first 24 hours..., Laborada [/bib_ref] [bib_ref] Evaluation of serum interleukins-6, 8 and 10 levels as diagnostic markers of..., Boskabadi [/bib_ref]. It was identified that initial neonatal sepsis onset (within the first 24 hours of infection) was best diagnosed with a combination of IL-6 and CRP detection, with a cutoff of 18 pg/mL for IL-6 and 10 pg/mL for CRP (sensitivity: 89%; specificity: 73%) [bib_ref] Diagnostic value of cytokines and c-reactive protein in the first 24 hours..., Laborada [/bib_ref]. However, the same study also found that one day after sepsis onset, CRP alone was a better predictor of infection, warranting future work. Another study, by [bib_ref] Utility of interleukin-12 and interleukin-10 in comparison with other cytokines and acute-phase..., Sherwin [/bib_ref] , reported a variety of cytokines using ROC analysis, including Il-1 , IL-6, IL-12, IL-8, IL-10, and TNF- [bib_ref] Utility of interleukin-12 and interleukin-10 in comparison with other cytokines and acute-phase..., Sherwin [/bib_ref]. This study found that IL-12 was the most promising source of diagnostic aid, particularly in confirming neonatal sepsis, and that IL-6 was not as effective in predicting the condition. Evaluating use of IL-12 in diagnosis of neonatal sepsis yielded an optimal cutoff of 75 pg/mL (AUC 74%; sensitivity: 28%; specificity: 98%). IL-10 was the next most promising diagnostic aid, with the same cutoff but different test characteristics (sensitivity: 17%; specificity: 99%). In these results, one must consider low sensitivities for these tests, which imply that these cytokine biomarkers can be more accurately used to confirm diagnosis of neonatal sepsis, rather than confirming its absence. Boskabadi et al. conducted a variety of studies regarding clinical cytokine cutoffs for diagnosing neonatal sepsis: they found IL-8 at a cutoff of 60 pg/mL to be predictive of the disease [bib_ref] Serum interleukin 8 level as a diagnostic marker in late neonatal sepsis, Boskabadi [/bib_ref]. In subsequent studies by the same group, IL-10 was reported to have the greatest predictive value at a cutoff of 14 pg/mL [bib_ref] Evaluation of serum interleukins-6, 8 and 10 levels as diagnostic markers of..., Boskabadi [/bib_ref] [bib_ref] Early diagnosis of late neonatal sepsis by measuring interleukin 10: a case..., Boskabadi [/bib_ref]. Among IL-6, IL-8, and IL-10, both IL-6 and IL-8 were found to predict neonatal sepsis at cutoffs of 10.85 pg/mL (sensitivity: 92.5%; specificity: 97.6%) and 60.05 pg/mL (sensitivity: 93.7%; specificity: 65%), respectively [bib_ref] Evaluation of serum interleukins-6, 8 and 10 levels as diagnostic markers of..., Boskabadi [/bib_ref]. IL-6 predicted mortality due to neonatal sepsis, with a higher cutoff of 78.2 pg/mL (sensitivity: 85%; specificity: 76%). As evidenced by these studies, IL-6, IL-8, and IL-10 may be the most promising cytokine biomarkers to diagnose neonatal sepsis once more narrow ranges in cutoffs across studies are established. IL-6 is of special interest, due to its potential ability to both diagnose neonatal sepsis and predict mortality due to the condition, as well as its relatively uniform range in cutoffs across these preliminary studies. ## Surgical site infections: periprosthetic joint infection and postcervical neck dissection infection. A variety of other infection pathologies have been connected with cytokines, though very few have been established as predicted by particular cytokine cutoffs. Among these, periprosthetic joint infection (PJI) is a risk after revision surgery for shoulder arthroplasty and is considerably difficult to diagnose due to a lack of biomarkers to distinguish between septic and aseptic outcomes. Cytokines in synovial fluid may be able to aid in diagnosis of PJI, based on correlative studies, though few have evaluated particular cytokine cutoffs for the condition. [bib_ref] Synovial fluid interleukin-6 as a predictor of periprosthetic shoulder infection, Frangiamore [/bib_ref] found that IL-6 in synovial fluid could help predict PJI after revision surgery for shoulder arthroplasty, with an optimal cutoff of 359.3 pg/mL (sensitivity: 87%; specificity: 90%) [bib_ref] Synovial fluid interleukin-6 as a predictor of periprosthetic shoulder infection, Frangiamore [/bib_ref]. The most recent study of Frangiamore et al. in this area (2016) also investigated IL-1 and IFN-as potential diagnostic biomarkers for the disease, at cutoffs of 8.26 pg/mL (AUC 92% and 91%, resp.) and 34 pg/mL [bib_ref] Synovial cytokines and the MSIS criteria are not useful for determining infection..., Frangiamore [/bib_ref]. Perhaps more importantly, IL-1 and IL-6 had the highest sensitivities as diagnostic tools (90% and 81%, resp.), and both also had the greatest decrease between explantation and reimplantation demonstrating infection resolution (12.4-fold decrease for IL-1 ; 11.2-fold decrease for IL-6), implicating IL-1 and IL-6 as potentially the most applicable for usage in diagnosis/prognosis of PJI. It is also important to note that all cytokines measured did not show significant diagnostic utility or sensitivity to rule out infection before reimplantation. Cytokine assays have the potential to be significant in PJI diagnosis, as synovial fluid cytokine testing may be more sensitive than current testing available for diagnosing PJI, which may promise better patient outcomes. In a similar vein, another study demonstrated statistical differences between patients with and without surgical site infection (SSI) after cervical neck dissection on postsurgical drainage fluid cytokine levels, including IL-1 , IL-2, IL-6, IL-8, and TNF- [bib_ref] Early modification in drainage of interleukin-1 and tumor necrosis factor-best predicts surgical-site..., Candau-Alvarez [/bib_ref]. IL-10 was the only cytokine not associated with SSI. TNF-and IL-1 showed the greatest diagnostic efficacy of the cytokines associated with SSI: sensitivities and specificities were 100% and 87.88% for TNF-at a cutoff of 14.5 pg/mL (on day 1 of infection for TNF-) and 83.33% and 78.79% for IL-1 at a cutoff of 115 pg/mL (on day 3 of infection for IL-1 ). IL-2 levels were also indicative of disease on days 1 and 3 above levels of 6.5 pg/mL, and IL-6 levels were indicative on day three, above levels of 3,300 pg/mL, though at lower sensitivity and specificity values. This study emphasizes the importance of temporally sensitive evaluation of cytokines, due to their variation in release and concentration over time, along with the promise of potential cytokine biomarkers for infection prediction and diagnosis at various stages of infection. ## Recovery from procedure for left ventricular dysfunction and Prognosis of Heart Failure. Cytokine cutoffs may also be established to evaluate prognosis of particular diseases with inflammatory indications, including left ventricular dysfunction (LVD), treated by percutaneous coronary intervention (PCI) and its often associated subsequent heart failure, or acute myocardial infarction (AMI). Left ventricular dysfunction is a risk factor for poor prognosis following acute myocardial infarction (AMI) [bib_ref] Serum concentrations of interleukin-4 and interferon-gamma in relation to severe left ventricular..., Szkodzinski [/bib_ref] , which has been associated with an imbalance in pro-and anti-inflammatory markers, and therefore some studies have sought to investigate potential correlations between these conditions and cytokines for improved prognosis evaluation. [bib_ref] Serum concentrations of interleukin-4 and interferon-gamma in relation to severe left ventricular..., Szkodzinski [/bib_ref] investigated cytokines predicting LVD before and after PCI to evaluate possible cytokine predictors [bib_ref] Serum concentrations of interleukin-4 and interferon-gamma in relation to severe left ventricular..., Szkodzinski [/bib_ref]. Measurements of serum cytokines before and after PCI showed high diagnostic value of IL-4 both before and immediately after PCI, whereas IFN-measurements only before PCI were of diagnostic aid. IL-4 at levels above 15 pg/mL before PCI indicated LVD, and IL-4 at slightly higher levels of 17.2 pg/mL after PCI indicated LVD. IFN-predicted LVD before PCI with a cutoff of 0.3 pg/mL. [bib_ref] Relationship between preimplant interleukin-6 levels, inflammatory response, and early outcome in patients..., Caruso [/bib_ref] attempted to further this evaluation of LVD, identifying potential cytokine cutoffs to predict prognosis of LVD patients in those who require left ventricular assist devices (LVAD) [bib_ref] Relationship between preimplant interleukin-6 levels, inflammatory response, and early outcome in patients..., Caruso [/bib_ref]. Among LVD patients, many require a LVAD to aid in preventing the development of multiorgan failure (MOF). Aside from the inflammatory indications of LVD, unique inflammatory profiles also appear upon implantation of the assist device. In this study, white blood cell count, indicating infection, and MOF severity were associated with IL-6 at levels above 8.3 pg/mL. This cutoff of IL-6 also indicated longer ICU stay, longer hospitalization, poor early outcome, and higher levels of release of other proinflammatory molecules, such as IL-8, which can help evaluate prognosis of patients with LVAD long-term. These studies indicate the potential for particular interleukins to aid in diagnosis of LVD, regardless of surgical intervention, and in prognosis of LVD, in patients with LVAD. ## Role of cytokines and their clinical cutoffs in evaluation of severity and prognosis for chronic conditions Aside from infections and diseases with inflammatory indications, cytokines may also be useful in assessing prognoses of inherited and often chronic conditions, including Alzheimer's Disease, neurological outcomes following cardiac arrest (causing prolonged hypoxia), cancer, lupus nephritis, and lymphohistiocytosis. Conditions such as these have been correlated with elevated proinflammatory cytokine levels in previous work, though few have investigated particular cytokine cutoffs for predicting severity and prognosis of these diseases. ## Alzheimer's disease and neurological outcomes in cardiac Arrest Patients. Alzheimer's Disease (AD) has been well established as a disease which is caused by inflammatory processes, and cytokines have accordingly been implicated in its etiology. However, cutoffs of cytokines in diagnosing AD have not been well studied. A recent study in a Malaysian population demonstrated the possibility of establishing a particular cutoff for cytokines and chemokines in AD [bib_ref] 115 * IP-10 and IL-13 as potentially new, non-classical blood-based cytokine biomarkers..., Dayana [/bib_ref]. In the blood, IP-10 (a chemokine) and IL-13 were both associated with AD: IL-13 was 18-fold lower in AD patients than in healthy controls, and IL-13 was 9-fold lower in European AD patients when compared to healthy controls. Another study also found that anti-inflammatory IL-13, as well as IP-10, was downregulated in AD patients compared to controls [bib_ref] Peripheral cytokines, C-X-C motif ligand10 and interleukin-13, are associated with Malaysian Alzheimer's..., Hasni [/bib_ref]. An IL-13 cutoff value of 9.315 pg/mL was determined (sensitivity and specificity 100%) to diagnose AD, in accordance with the anti-inflammatory properties of IL-13. The suggestion of such a specific and sensitive test for AD using cytokines is significant to the diagnostic field of this disease: standard methods for diagnosing Alzheimer's are costly, as they require highly trained specialist physicians and expensive imaging methods. Further, a method that relies purely on serum measurements of cytokines would facilitate diagnosis of AD, particularly in poorer nations. In addition to poor neurological outcomes resulting from AD, such outcomes can also result from cardiac arrest due to prolonged oxygen deprivation, and these effects may be identified by cytokine cutoffs. [bib_ref] The cutoff values of intrathecal interleukin 8 and 6 for predicting the..., Oda [/bib_ref] measured cytokines in cerebrospinal fluid (CSF) in patients 6 months following cardiac arrest and evaluated patients according to the Glasgow Outcome Score (GOS) in order to evaluate poor neurological outcome as indicated by CSF cytokines [bib_ref] The cutoff values of intrathecal interleukin 8 and 6 for predicting the..., Oda [/bib_ref]. IL-8 and IL-6 were found to be significantly higher in subjects who had experienced cardiac arrest, and both cytokines were found to be correlated with poor neurological outcome. Cytokine cutoffs for predicting poor neurological outcome following cardiac arrest were 1423 pg/mL for IL-8 and 2708 pg/mL for IL-6 (both with sensitivity of 100% and specificity of 86%). These studies suggest that, following future work, evaluation of poor neurological prognosis may be possible in cases of both AD and cardiac arrest through serum and CSF cytokine levels, respectively. ## Gastric cancer. Gastric cancer is another condition characterized by an abnormal cytokine profile. Various studies have linked IL-6 to gastric cancer, though few studies have investigated clinical cutoffs of IL-6 for this condition. Most studies of cytokines in cancer, as in diagnosis of other conditions using cytokines, set a cutoff of a particular percentile rather than considering ROC analysis for diagnostic efficacy. However, despite considerable precedent, a 2005 study found the optimal diagnostic cutoff of IL-6 for gastric cancer to be 1.97 pg/mL (sensitivity 81.8%; specificity: 66.7%) [bib_ref] Clinical significance of interleukin-6 (IL-6) in the spread of gastric cancer: role..., Ashizawa [/bib_ref]. This corresponded with an accuracy of 77.1% as well as significantly lower patient survival rates and positive immunohistochemical staining of IL-6 in cancer cells. A slightly higher serum cutoff of this cytokine was identified as having the potential to diagnose preoperative gastric cancer and to evaluate prognoses after surgical tumor removal: 6.77 pg/mL serum IL-6 (sensitivity 85.7%) [bib_ref] Clinical significances of preoperative serum interleukin-6 and C-reactive protein level in operable..., Kim [/bib_ref]. However, the consensus between the two studies regarding the role of IL-6 in predicting gastric cancer suggests that the correct cytokine for diagnostic use has been identified and that it will just be a matter of time to establish a narrower diagnostic cutoff range with improved sensitivity. ## Lupus nephritis and lymphohistiocytosis. Lupus nephritis and lymphohistiocytosis are chronic conditions of immune deficiency. Lupus nephritis often accompanies systemic lupus erythematous (SLE) and is characterized by inflammation of the kidney. Lupus nephritis has been associated with immune markers, including cytokines, though few studies have established specific clinical cutoffs for these markers. [bib_ref] Urinary monocyte chemotactic protein-1 and transforming growth factor-in systemic lupus erythematosus, Torabinejad [/bib_ref] investigated TGF-, among other noncytokine immune molecules, and evaluated its effect on diagnosis of lupus nephritis in patients with SLE; the study established a cutoff point of 54.2 pg/mL TGF-(sensitivity: 71.4%; specificity: 95.6%) [bib_ref] Urinary monocyte chemotactic protein-1 and transforming growth factor-in systemic lupus erythematosus, Torabinejad [/bib_ref]. Despite low sensitivity, the test was highly specific and can therefore be important in obviating the need for additional diagnostic testing, such as renal biopsies for certain patients. If further studies are conducted to establish a cutoff with higher sensitivity, this cytokine may be used diagnostically. A 2015 study determined cutoffs of cytokines IL-17 and IL-6 to diagnose lupus nephritis in patients with SLE [bib_ref] The role of serum IL-17 and IL-6 as biomarkers of disease activity..., Galil [/bib_ref]. Activity of these cytokines was present at a significant level, both during active disease periods and during remission. Optimal cutoffs for diagnosing active forms of lupus nephritis using IL-6 and IL-17 were 12.3 pg/mL (AUC 93%) and 19.7 pg/mL (AUC 95%), respectively, whereas optimal cutoffs for diagnosing lupus nephritis in remission were higher, at 20.8 pg/mL (AUC 80%) and 27 pg/mL (AUC 82%), respectively. Accordingly, these biomarkers may be used not only to diagnose lupus nephritis, but also to predict remission of the disease. Hemophagocytic lymphohistiocytosis (HLH) is an inherited immune deficiency resulting from dysfunction of T and natural killer cells. [bib_ref] Original article: diagnostic accuracy of a specific cytokine pattern in hemophagocytic lymphohistiocytosis..., Xu [/bib_ref] considered IFN-, TNF-, IL-10, IL-6, IL-4, and IL-2 as potential biomarkers of this disease in pediatric febrile patients [bib_ref] Original article: diagnostic accuracy of a specific cytokine pattern in hemophagocytic lymphohistiocytosis..., Xu [/bib_ref]. In patients with HLH, the median level of IFN-was 1088.5 pg/mL, and the median level of IL-10 was 623.5 pg/mL. IL-6 was elevated to a lesser extent, with a median level in these HLH patients of 51.1 pg/mL, though among HLH patients with sepsis, as expected, a higher median IL-6 level was detected (244.6 pg/mL). Upon ROC analysis, IFN-had a sensitivity of 94.4% and specificity of 97.2% for diagnosing HLH at a 100 pg/mL cutoff. IL-10 did not provide considerable diagnostic utility individually, but when IFN-and IL-10 were considered at cutoffs of >75 pg/mL and >60 pg/mL, respectively, specificity increased to 98.9% and the sensitivity increased to 93%. However, IL-6, though it moderately increased in these patients, did not appear to be diagnostically useful. Despite its lack of diagnostic relevance, IL-6 levels may help contribute to the characterization of the profile of pediatric febrile HLH patients in its moderate level, among higher levels of IFN-and IL-10. It will become necessary to consider nonfebrile patients in future studies due to the possibility of the fever state contributing to the inflammatory process, though this study and the previous studies of lupus nephritis may be useful at gaining a preliminary understanding of the particular levels of cytokines at which the body responds to immune deficiencies. ## Role of cytokines and their clinical cutoffs in obstetric and gynecological conditions The third category of conditions where cytokines can be evaluated as biomarkers is obstetric and gynecological conditions characterized by inflammation, including endometriosis, ovarian abnormalities, ectopic pregnancy, miscarriage, preterm premature rupture of membranes (PROM), and preterm delivery. Such conditions have been significantly correlated with cytokines in various studies, though recently, and perhaps more importantly, clinical cutoffs are being established for diagnosis as well. ## Endometriosis and ovarian Cancer. Endometriosis is a condition characterized by a variety of inflammatory biomarkers, though diagnosis typically involves a pelvic exam or imaging. Cytokines have the potential to aid in diagnosis of this condition, reducing the need for costly imaging techniques or uncomfortable physical examinations. Of a number of proinflammatory cytokines, two have been established as potentially useful for diagnostics, IL-6 and TNF-, which were found to be correlated with endometriosis and to be effective in diagnosing endometriosis in serum and peritoneal fluid as well as in menstrual effluents. [bib_ref] Prediction of endometriosis with serum and peritoneal fluid markers: a prospective controlled..., Bedaiwy [/bib_ref] found IL-6 and TNF-to be able to distinguish between patients with and without endometriosis with high sensitivity and specificity, at respective cutoffs of 2 pg/mL (sensitivity 90%; specificity 67%) and 15 pg/mL (sensitivity 100%; specificity 89%) [bib_ref] Prediction of endometriosis with serum and peritoneal fluid markers: a prospective controlled..., Bedaiwy [/bib_ref]. [bib_ref] Interleukin-6, interleukin-1 , and tumor necrosis factor in menstrual effluents as biomarkers..., Tortorella [/bib_ref] found significantly higher levels of IL-6, IL-1 , and TNF-in menstrual effluents of women diagnosed with endometriosis compared to healthy subjects, though they found limited diagnostic capability of each of these cytokines individually [bib_ref] Interleukin-6, interleukin-1 , and tumor necrosis factor in menstrual effluents as biomarkers..., Tortorella [/bib_ref]. However, when combined (IL-6 combined with TNFor IL-6 combined with IL-1 ), the cytokines provided diagnostic sensitivity of nearly 100% when at least one cytokine was above the diagnostic value and provided a near 100% specificity when both cytokines in a combination exceeded their cutoff values: IL-6, 8,968 pg/mL; TNF-, 203 pg/mL. Logistic regressions demonstrated the combination of IL-6 and TNF-to be significantly predictive of chronic endometriosis diagnosis, confirming their use as diagnostic tools, as suggested by ROC analysis. It is important to note, however, that the IL-6 and TNF-cutoffs differ greatly between the two studies, warranting future work to narrow this range of diagnostic cutoffs. More studies in this area have the potential to avoid unnecessary invasive diagnostic procedures for endometriosis by providing the capacity for a test of either serum, peritoneal fluid, or menstrual effluent, to diagnose or eliminate the probability of disease. Ovarian cancer has also been associated with a variety of cytokines, though it has only been investigated with respect to diagnostic utility. [bib_ref] Peritoneal fluid cytokines and the differential diagnosis of benign and malignant ovarian..., Chechlinska [/bib_ref] measured a number of cytokines (VEGF, IL-6, bFGF, IL-8, and M-CSF) in peritoneal fluid of untreated ovarian cancer patients and those with benign ovarian tumors [bib_ref] Peritoneal fluid cytokines and the differential diagnosis of benign and malignant ovarian..., Chechlinska [/bib_ref]. Although IL-6, VEGF, and CA 125 all were elevated in ovarian cancer patients, and bFGF and M-CSF were decreased in these patients, only IL-6 and VEGF were significantly elevated in stagees I and II cancer patients. For these two significantly associated cytokines, two cutoff levels were tested: 400 pg/mL and 1200 pg/mL. At these two cutoffs, IL-6 had a sensitivity of 92% and 84%, respectively, and a specificity of 60% and 87%, respectively. At a cutoff of 400 pg/mL, VEGF had a sensitivity of 90% and a specificity of 80%. Accordingly, measurements of these two cytokines in peritoneal fluids may be useful for identifying malignant versus benign ovarian tumors upon first diagnosis. [bib_ref] Multiplexed immunobead-based cytokine profiling for early detection of ovarian cancer, Gorelik [/bib_ref] considered a wider range of cytokines in serum, including IL-6, IL-8, EGF, VEGF, MCP-1, and cancer antigen-125, due to evaluated significance in correlations [bib_ref] Multiplexed immunobead-based cytokine profiling for early detection of ovarian cancer, Gorelik [/bib_ref]. These cytokines were analyzed through classification tree analysis to determine their optimal combinations to diagnose ovarian cancer. A tree combining all of the aforementioned cytokines, excepting MCP-1, was generated to diagnose ovarian cancer versus controls with sensitivity 84% and specificity 95%; a tree combining CA-125, G-CSF, IL-6, EGF, and VEGF was generated to distinguish ovarian cancer from benign 8 International Journal of Inflammation ovarian tumors with sensitivity 86.5% and specificity 93%. Both of these studies suggest IL-6 and VEGF as possible diagnostic biomarkers for ovarian cancer with reasonable sensitivities and specificities for clinical applications. ## Ectopic pregnancy and miscarriage. Ectopic pregnancy has been correlated with proinflammatory cytokines, including IL-2R, IL-6, IL-8, and TNF-, and these cytokines have been investigated as diagnostic biomarkers with specific cutoffs. These four cytokines were measured in serum levels of women with miscarriage, normal intrauterine pregnancy, and ectopic pregnancy (EP). One study found that higher IL-8, IL-6, and TNF-were present in women with EP when compared to women with miscarriage and normal pregnancy, though only IL-8 yielded a cutoff, of 40 pg/mL, for predicting ectopic pregnancy with notable sensitivity and specificity (82.4% and 81.8%, resp.) [bib_ref] IL-6, IL-8, and tumor necrosis factor alpha in patients with ectopic pregnancy, Soriano [/bib_ref]. Another study found potential utility of serum IL-6 and IL-8 in diagnosing ectopic pregnancy, as compared to normal intrauterine pregnancy and miscarriage [bib_ref] Diagnostic significance of IL-6 and IL-8 in tubal ectopic pregnancy, Rajendiran [/bib_ref]. However, this study only found IL-6 to yield remarkable sensitivity and specificity for ectopic pregnancy diagnosis at a cutoff of 26.48 pg/mL and did not find IL-8 to be significantly predictive of the condition at any cutoff. Based on these findings, it is possible that within a larger study, IL-6, IL-8, or a combination of the two may be determined useful as noninvasive markers for predicting ectopic pregnancy. ## Preterm prom and preterm Delivery. Preterm PROM is characterized by rupture of the amniotic sac before the fetus is carried to term and is often associated with inflammation due to microbial invasion, as indicated by amniotic fluid cytokine levels obtained through amniocentesis. The associated inflammation and microbial invasion in this condition have been most conclusively correlated with proinflammatory IL-6 in amniotic fluid, though the specific cutoffs of this biomarker suggest the disease have only been recently and preliminarily investigated. [bib_ref] Bedside assessment of amniotic fluid interleukin-6 in preterm prelabor rupture of membranes, Kacerovsky [/bib_ref] sought to evaluate the diagnostic utility of IL-6 in diagnosing microbial invasion of the amniotic cavity in PROM pregnancies; a cutoff of 1000 pg/mL IL-6 in amniotic fluid was determined to be the optimal diagnostic level for identifying microbial amniotic invasion (sensitivity 50%, specificity 95%) [bib_ref] Bedside assessment of amniotic fluid interleukin-6 in preterm prelabor rupture of membranes, Kacerovsky [/bib_ref]. The study also identified a 1000 pg/mL cutoff in PROM patients for identifying histological chorioamnionitis in combination with microbial amniotic invasion (sensitivity 60%, specificity 94%). However, amniocentesis is extremely invasive, prompting [bib_ref] Vaginal fluid IL-6 concentrations as a point-of-care test is of value in..., Musilova [/bib_ref] to test cytokine levels in vaginal fluid of pregnant women in an attempt to predict preterm PROMrelated inflammation in a less invasive manner [bib_ref] Vaginal fluid IL-6 concentrations as a point-of-care test is of value in..., Musilova [/bib_ref]. IL-6 was determined to be associated with microbial invasion of the amniotic cavity, as well as intra-amniotic inflammation and microbial-associated intra-amniotic inflammation, and was able to predict these conditions at a cutoff of 2500 pg/mL with high sensitivity and specificity: 53% and 89%, 74% and 91%, and 100% and 90%, respectively. These consistently higher specificity values suggest the ability of IL-6 to distinguish between pregnant women with and without PROM-related inflammation. Like preterm PROM, preterm delivery has also been linked to IL-6, though preterm delivery has additionally been correlated with IL-8. These cytokines have primarily been measured in cervicovaginal fluid to evaluate their diagnostic utility due to ease of access and limited cost. [bib_ref] Diagnostic accuracy of cervicovaginal interleukin-6 and interleukin-6:albumin ratio as markers of preterm..., Woodworth [/bib_ref] investigated IL-6 in cervicovaginal fluid to predict early delivery, finding a 35% sensitivity value and 91% specificity value at a 250 pg/mL [bib_ref] Diagnostic accuracy of cervicovaginal interleukin-6 and interleukin-6:albumin ratio as markers of preterm..., Woodworth [/bib_ref]. Odds ratios supported this finding, testing IL-6 as a predictor of preterm delivery, with statistical significance ( = 0.0001). The IL-6 : albumin ratio provided significance on a lower confidence interval and did not yield a similarly high sensitivity or specificity. Nonetheless, IL-6 levels are suggested to be predictive of preterm delivery within 14 days as an inexpensive and limitedly invasive diagnostic method. [bib_ref] Te value of interleukin-8 and interleukin-6 in cervical secretions as predictors of..., Gandevani [/bib_ref] also found IL-6 to predict preterm delivery, though Gandevani et al. found IL-8 to be predictive of this condition [bib_ref] Te value of interleukin-8 and interleukin-6 in cervical secretions as predictors of..., Gandevani [/bib_ref]. Fourfold and nearly fivefold increases in IL-6 and IL-8 concentrations were found in preterm versus normal deliveries, respectively. Cutoffs of these cytokines for predicting early preterm delivery were determined as 751.25 pg/mL IL-8 and 157 pg/mL IL-6 (sensitivities of 89% and specificities of 83% and 78%, resp.). Likewise, [bib_ref] Maternal serum cytokines in the prediction of preterm labor and response to..., Shahshahan [/bib_ref] found results suggesting IL-6 and IL-8 to be diagnostic markers for preterm labor and delivery [bib_ref] Maternal serum cytokines in the prediction of preterm labor and response to..., Shahshahan [/bib_ref]. However, Shahshahan and Hashemi (2014) measured the two cytokines in serum of pregnant women with and without preterm labor, rather than in cervicovaginal fluid, finding IL-6 to have a lower optimal cutoff for predicting preterm labor of 37.8 pg/mL and IL-8 to have a lower optimal cutoff of 9.5 pg/mL. Cytokine levels were also evaluated to predict response to tocolytic therapy in pregnant women with preterm delivery, with cutoffs of 45 pg/mL for IL-6 and 171 pg/mL for IL-8. These studies suggest the possibility of both IL-6 and IL-8 as predictors of preterm delivery and the response of such pregnancies to tocolytic therapy, though further work must be conducted to narrow the range of cutoffs for diagnostic utility. # Conclusion ## Future of cytokine research in the diagnostic field. Many of the studies mentioned above are unprecedented in their search for cytokine cutoffs to aid in diagnosis of particular conditions, as many studies simply search for correlations between cytokines and disease states without evaluating specific clinical cutoffs. Although cytokines are implicated in various disease states and are dependent upon a variety of pathways, studies such as those reported in this review demonstrate the possibility of using cytokines or combinations of cytokines, in addition to other factors, to diagnose various immunologically implicated conditions with greater ease and accuracy, and at a lower cost. The establishment of clinical cytokine cutoffs for various conditions may be facilitated by future improvement in our understanding of normal cytokine profiles. To better understand cytokine levels and to establish diagnostic cutoffs in disease states, it is important to first characterize normal cytokine levels in various populations and to understand how International Journal of Inflammation 9 cytokines interact and modulate in various biochemical pathways in healthy individuals. Particularly, if lack of research in certain diseases' areas prevents the use of cytokines alone in diagnosis, gaining a better understanding of normal cytokine profiles will also allow confirmation and supplementation or reconsideration, of diagnoses made through other diagnostic methods and clinical features. Further, as normal and diagnostic cutoffs for these biomarkers become elucidated, cytokines may be considered to improve patient outcomes when definitive diagnosis is not possible with clinical features alone. Many diseases have very similar clinical features, and therefore establishment of pathological cytokine profiles may aid in differentiation of such conditions. Due to the limited number of studies investigating effectiveness of particular cytokine cutoffs as diagnostic tools, further studies must be conducted to narrow diagnostic ranges. Such a narrowing of diagnostic ranges will facilitate confirmation or rejection of diagnoses suggested by other clinical features to ultimately improve patient outcomes. [fig] Figure 1: IL-12 activation of JAK-STAT pathway. IL-12 activates a series of phosphorylation events which permit STAT4 molecules to dimerize and translocate to the nucleus, upregulating a proinflammatory cascade. IL-10 has an inhibitory effect on this response transcriptionally. [/fig]

Candidatus Bartonella mayotimonensis and Endocarditis We describe a new Bartonella species for which we propose the name Candidatus Bartonella mayotimonensis. It was isolated from native aortic valve tissue of a person with infective endocarditis. The new species was identifi ed by using PCR amplifi cation and sequencing of 5 genes (16S rRNA gene, ftsZ, rpoB, gltA, and internal transcribed spacer region). The patient lived alone on a farm in Iowa, USA, and had not had recent exposure to animals. However, he had observed murine fecal droppings in his house and mice on the farm. He had had a house cat for 18 years until its death a few years before his hospitalization and had intermittent contact with cats when he visited his daughter. Serum immunoglobulin G titers were positive for B. henselae and B. quintana (>1,024). Oral doxycycline and rifampin were prescribed for treatment of presumed Bartonella endocarditis. Gentamicin was not administered because of development of the acute renal dysfunction. Two weeks later, he underwent aortic valve and aortic root replacement. Results of gram staining, acid-fast staining, fungal staining, anaerobic bacterial culture, aerobic bacterial culture, mycobacterial culture, and fungal culture on resected aortic valve tissue were negative for Bartonella species. PCR performed at the Mayo Clinic on resected aortic valve tissue detected part of the citrate synthase gene (gltA) of Bartonella species. However, the melting temperature was not characteristic of B. quintana or B. henselae [bib_ref] Bivalvular Bartonella henselae prosthetic valve endocarditis, Vikram [/bib_ref]. Oral doxycycline and rifampin were continued for 12 weeks after aortic valve resection. The patient was well and had no signs of relapsing infection at a follow-up visit 11 months after valve surgery. Aortic valve tissue and serum were tested at the Unité des Rickettsies, Marseille, France. B. quintana Oklahoma, B. henselae Houston (ATCC 49882), B. vinsonii subsp. berkhoffi i (URBVAIE25), B. vinsonii subsp. arupensis (ATCC 700727), and B. alsatica (CIP 105477 T) strains were used for immunofl uorescent assays and Western blotting [bib_ref] First isolation of Bartonella alsatica from a valve of a patient with..., Raoult [/bib_ref]. Valve tissue was injected into human endothelial cells in a shell vial assay and onto Columbia 5% sheep blood agar plates and incubated at 37°C in an atmosphere of 5% CO 2 as described [bib_ref] First isolation of Bartonella alsatica from a valve of a patient with..., Raoult [/bib_ref]. A Bartonella species was detected in a shell vial by immunofl uorescence after 15 days of culture; identifi cation was confi rmed by PCR. DNA was extracted from valve specimen and injected cells by using the QIAamp Tissue Kit (QIAGEN, Hilden, Germany). DNA was used as a template in a genus Bartonella Lightcycler assay with primers and a Taqman probe specifi c for the internal transcribed spacer (ITS) gene [bib_ref] First isolation of Bartonella alsatica from a valve of a patient with..., Raoult [/bib_ref] and in standard PCR assays specifi c for the 16S rRNA, ITS, rpoB, gltA, and ftsZ genes (8). Sequences from both DNA strands were determined twice for all PCR products. These products were resolved in an ABI 3100 automated sequencer (PerkinElmer, Waltham, MA, USA). Sterile water was used as a negative control in each assay. Percentages of similarity among sequences were determined by using MEGA 2.1 software [bib_ref] MEGA2: molecular evolutionary genetics analysis software, Kumar [/bib_ref]. Phylogenetic relationships among Bartonella strains were inferred from concatenated sequences by using MEGA 2.1 software [bib_ref] MEGA2: molecular evolutionary genetics analysis software, Kumar [/bib_ref]. Surgically resected aortic valve tissues were fi xed in formalin, embedded in paraffi n, and sectioned to a thickness of 5 μm. Sections were stained with periodic acid-Schiff, Giemsa, Gram, Grocott-Gomori methenamine silver, and Warthin-Starry stains. Immunohistochemical analysis was performed by using a procedure described elsewhere [bib_ref] Quantitative analysis of valvular lesions during Bartonella endocarditis, Lepidi [/bib_ref] and polyclonal antibody against B. vinsonii at a dilution of 1:1,000. Serum samples showed immunoglobulin G endpoint titers of 50 against all Bartonella species tested by immunofl uorescent assay. Western blot results were positive and characteristic of Bartonella infection [fig_ref] Figure 1: A [/fig_ref]. Results of PCR (Bartonella genus Lightcycler assay and standard PCR for cardiac valve) and cell culture were positive, and amplifi cation products of the expected size were obtained. Among known validated species, sequences obtained shared 99.1% (1, [fig_ref] Figure 2: Phylogenetic tree showing the position of Candidatus Bartonella mayotimonensis among members of... [/fig_ref]. Sequences of gltA, 16S rDNA, ftsZ, ITS, and rpoB were deposited in GenBank under accession nos. FJ376732-FJ376736. Histologic analysis of resected aortic valve showed infective endocarditis with vegetation containing microorganisms that stained with Warthin-Starry and Giemsa. Warthin-Starry staining showed darkly stained bacilli consistent with Bartonella species [fig_ref] Figure 1: A [/fig_ref]. Results of staining with periodic acid-Schiff, Gram, and Grocott- Gomori methenamine silver were negative. Immunohistochemical analysis detected bacteria in valvular vegetations in a location superimposable with that detected by Warthin-Starry staining [fig_ref] Figure 1: A [/fig_ref]. # Conclusions We isolated a new Bartonella species and propose that it be named Candidatus Bartonella mayotimonensis to recognize the contributing institutions (Mayo Clinic and Hôpital de la Timone, Marseille, France). This is the seventh Bartonella species documented to cause infective endocarditis in humans. The reservoir of Candidatus B. mayotimonensis has yet to be determined. Different Bartonella species have been isolated from a variety of mammals, and each species is highly adapted to its reservoir host [bib_ref] Bartonella quintana in domestic cat, La [/bib_ref] [bib_ref] Recommendations for treatment of human infections caused by Bartonella species, Rolain [/bib_ref]. The domestic cat is the primary mammalian reservoir for B. henselae [bib_ref] Bacteremia, fever, and splenomegaly caused by a newly recognized Bartonella species, Eremeeva [/bib_ref]. Other Bartonella species have been found in mammalian hosts, including rats (B. [bib_ref] Bartonella vinsonii subsp. berkhoffi i as an agent of afebrile blood culture-negative..., Roux [/bib_ref] [bib_ref] Bartonella koehlerae, a new cat-associated agent of culturenegative human endocarditis, Avidor [/bib_ref] [bib_ref] First isolation of Bartonella alsatica from a valve of a patient with..., Raoult [/bib_ref] [bib_ref] Recommendations for treatment of human infections caused by Bartonella species, Rolain [/bib_ref] [bib_ref] Bartonella infection in animals: carriership, reservoir potential, pathogenicity, and zoonotic potential for..., Breitschwerdt [/bib_ref] [bib_ref] Bartonella spp. in pets and effect on human health, Chomel [/bib_ref]. Our patient had direct exposure to mice on his farm and also had intermittent contact with cats while visiting his daughter. Additional investigations are needed to determine the reservoir(s) and vector(s) for this novel bacterium. The immunofl uorescent assay, the current serologic method for diagnosis of Bartonella infection, does not distinguish among Bartonella species. Only Western blot analysis and cross-adsorption enable serologic identifi cation of species. PCR and culture are critical when a Bartonella species is identifi ed for the fi rst time as a human pathogen. Newly encountered Bartonella strains should be considered a new species if a 327-bp gltA fragment shares <96.0% sequence similarity with those of validated species, and if an 825-bp rpoB fragment shares <95.4% sequence similarity with those of validated species as reported in the current case [bib_ref] Gene-sequence-based criteria for species defi nition in bacteriology: the Bartonella paradigm, Scola [/bib_ref]. This case reinforces the hypothesis that any Bartonella species can cause human infection, including culture-negative endocarditis. Candidatus B. mayotimonensis should be added to the list of human pathogens that can cause culture-negative endocarditis. [fig] Figure 1: A) Western blot of serum sample from patient infected with Candidatus Bartonella mayotimonensis. Left lane, Molecular mass standard; lane 1, Bartonella quintana; lane 2, B. henselae; lane 3, B. elizabethae; lane 4, B. vinsonii subsp. berkhoffi i; lane 5, B. alsatica. Values on the left are in kilobases. B) Numerous darkly stained bacilli consistent with Bartonella species organized in clusters in the valvular vegetation (Warthin-Starry stain; original magnifi cation ×400). C and D) Bacteria detected by immunohistochemical analysis of an extracellular location inside the valvular vegetation (polyclonal antibody against B. vinsonii subsp. berkhoffi i, Warthin-Starry stain and hematoxylin counterstain; original magnifi cation ×100 in C and ×400 in D). [/fig] [fig] Figure 2: Phylogenetic tree showing the position of Candidatus Bartonella mayotimonensis among members of the genus Bartonella based on comparisons of concatenated sequences of the 16S rRNA gene, the citrate synthase gene gltA, the RNA polymerase β-subunit gene rpoB, the cell division gene ftsZ, and the 16S-23S rRNA internal transcribed spacer region sequences. The tree was constructed by using the neighbor-joining method and a maximum-likelihood-based distance algorithm. Numbers on branches indicate bootstrap values derived from 500 replications. [/fig]

Subtrochanteric Fracture of the Femur Accompanying Pre-existing Ipsilateral Osteoarthritis of the Hip Successfully Treated with Intramedullary Nailing in the Lateral Decubitus Position: A Case Report # Introduction A subtrochanteric fracture of the femur accompanying pre-existing osteoarthritis of the ipsilateral hip (OA/STF) is rare [bib_ref] Subtrochanteric femur fracture below an arthritic hip: opinion: open reduction and internal..., Egol [/bib_ref] [bib_ref] Subtrochanteric femur fracture below an arthritic hip: opinion: open reduction and internal..., Koval [/bib_ref]. In general, intramedullary nailing is the preferred method of treatment of subtrochanteric femur fractures [bib_ref] Prevention of complications after treatment of proximal femoral fractures, Archdeacon [/bib_ref] [bib_ref] Use of an interlocked cephalomedullary nail for subtrochanteric fracture stabilization, French [/bib_ref]. However, arthritic deformities, such as the medial migration of the head and the shortening of the neck, complicate the correct placement of the entry point and the insertion of the nail along the long femoral axis. The entry point is often placed too far laterally when a fracture table with a perineal post is used. As a result, the fracture is reduced in varus malalignment [bib_ref] A critical analysis of the eccentric starting point for trochanteric intramedullary femoral..., Ostrum [/bib_ref] , which must be avoided to prevent nonunion [bib_ref] Prevention of complications after treatment of proximal femoral fractures, Archdeacon [/bib_ref] [bib_ref] Subtrochanteric fracture non-unions with implant failure managed with the, Giannoudis [/bib_ref]. Our aim was to report a case of OA/STF and demonstrate the superiority of the lateral decubitus position for avoiding varus malreduction in intramedullary nailing. ## Case presentation A 72-year-old female presented to the emergency room with the complaint of left thigh pain and the inability to walk after tumbling. She had been suffering from osteoarthritis of the left hip but was able to walk with a cane and sit on a couch. She had also received the diagnosis of osteoporosis and had been taking alendronate 35 mg weekly for six years. The initial radiograph in the emergency room showed a non-comminuted subtrochanteric fracture of the left femur with cortical thickness and the beaking of the lateral cortex at the fracture site. The fracture line Open Access Case was transverse on the lateral side and oblique on the medial side. The radiographic findings and the weakness of the force that caused the fracture satisfied the criteria for an atypical fracture of the femur. The radiograph showed Kellgren-Lawrence grade 4 osteoarthritis of the left hip as well, with superomedial migration, the external rotation of the femoral head, and the shortening of the femoral neck . ## Figure 1: preoperative radiography A subtrochanteric fracture of the left femur was seen. Cortical thickness, beaking of the lateral cortex, and a non-comminuted fracture met the criteria for an atypical fracture of the femur. The radiograph also showed osteoarthritis of the left hip with superomedial migration, external rotation of the femoral head, and shortening of the femoral neck, suggesting a difficulty in inserting a femoral nail. Osteosynthesis with intramedullary nailing was planned. However, the deformity of the hip joint appeared to preclude correct nail insertion. Furthermore, whenever we used a fracture table in the past, we often observed that the perineal post of the fracture table impeded the adduction of the proximal fragment of the subtrochanteric fracture while the whole leg was being adducted, resulting in varus malalignment. On the other hand, we were aware of easy access to the entry point via the standard lateral decubitus position without using the perineal post. For these reasons, we decided to fix the fracture with a nail in the lateral decubitus position. Total hip arthroplasty with fracture fixation was not chosen because of its invasiveness, which was inappropriate for the rather mild, pre-injury symptoms due to osteoarthritis, and because of the patient's wish to preserve the femoral head. In surgery, a flat, radiolucent operating table was used. The whole injured leg was sterilized and draped, and the hip was slightly flexed. The C-arm was placed at the ventral side of the patient. Although the proximal fragment was externally rotated, the entry point was located easily by adjusting the C-arm to obtain a correct A-P view. A guide pin was inserted in line with the long axis of the proximal fragment to the level of the fracture without any interference by the torso . ## Figure 4: postoperative radiography The fracture was anatomically reduced and fixed with a reconstruction nail. The operating time was 122 minutes. The total blood loss was 86 ml. Toe-touch weight bearing was initiated soon and full weight bearing was allowed in five weeks. A solid union was confirmed by radiograph after 10 months [fig_ref] FIGURE 5: Solid union was confirmed 10 months postoperatively 2018 Iga et al [/fig_ref]. She was able to walk with a cane as she did before her injury without a worsening of the pain. Alendronate was discontinued after the injury and replaced with vitamin D supplements. No anabolic agent, such as teriparatide, was administered. # Discussion Subtrochanteric fractures of the femur complicated by osteoarthritis of the ipsilateral hip are uncommon and only found in a few reports [bib_ref] Subtrochanteric femur fracture below an arthritic hip: opinion: open reduction and internal..., Egol [/bib_ref] [bib_ref] Subtrochanteric femur fracture below an arthritic hip: opinion: open reduction and internal..., Koval [/bib_ref]. However, they may become more common as the number of atypical subtrochanteric femoral fractures caused by the widespread use of bisphosphonates increases [bib_ref] Atypical femoral fractures related to bisphosphonate treatment, Koh [/bib_ref]. Understanding the proper surgical technique is necessary despite the rarity of this condition. Surgical treatment for OA/STF is controversial. Osteosynthesis with an intramedullary nail is indicated for younger patients with mild pre-existing symptoms due to osteoarthritis. In such cases, arthroplasty is not preferred because of extensile exposure, increased morbidity, and an increased risk of revision surgery [bib_ref] Subtrochanteric femur fracture below an arthritic hip: opinion: open reduction and internal..., Koval [/bib_ref]. However, osteosynthesis without addressing the arthritic hip is questionable, especially for the elderly, since reconstruction using a cable-plate followed by arthroplasty with a long stem provides early functional recovery. Other reasons against using osteosynthesis with an intramedullary nail are the difficulty in gaining a correct entry point and the possibility of malunion, which may make future long-stemmed arthroplasty complicated [bib_ref] Subtrochanteric femur fracture below an arthritic hip: opinion: open reduction and internal..., Egol [/bib_ref]. In general, either osteosynthesis or arthroplasty can be selected, depending on the patient's condition. We feel that osteosynthesis may be preferable for patients with mild pre-injury symptoms regardless of the radiographic severity of the osteoarthritis. Osteosynthesis with both intramedullary nailing and plating has satisfactory results. However, considering the frequent delayed union or nonunion observed in atypical femoral fractures, intramedullary nailing is preferable due to its mechanical superiority and endochondral repair. Plating for atypical femoral fractures is not recommended because of its unacceptably high failure rate [bib_ref] Atypical femoral fractures related to bisphosphonate treatment, Koh [/bib_ref]. When operating on a subtrochanteric fracture of the femur with intramedullary nailing, the entry point needs to be chosen correctly to avoid malreduction. If the entry point is misplaced only a few millimeters laterally, the fracture is angulated into varus, as the nail is inserted distally even if it was well-aligned before nail insertion [bib_ref] A critical analysis of the eccentric starting point for trochanteric intramedullary femoral..., Ostrum [/bib_ref]. Varus malalignment increases the natural tension-compression stresses on the subtrochanteric region and disturbs fracture union as well [bib_ref] Prevention of complications after treatment of proximal femoral fractures, Archdeacon [/bib_ref]. In fact, varus malalignment was observed in most cases of subtrochanteric nonunion with a breakage of the intramedullary nail [bib_ref] Subtrochanteric fracture non-unions with implant failure managed with the, Giannoudis [/bib_ref]. Therefore, varus malreduction must be avoided. However, subtrochanteric fractures are often reduced in varus [bib_ref] Use of an interlocked cephalomedullary nail for subtrochanteric fracture stabilization, French [/bib_ref] , especially when a patient lies supine on a fracture table where the injured leg has to be adducted in order to insert the nail [bib_ref] Prevention of complications after treatment of proximal femoral fractures, Archdeacon [/bib_ref] [bib_ref] Traction table-related complications in orthopaedic surgery, Flierl [/bib_ref]. However, the proximal fragment tends not to follow this movement because it is blocked by the perineal post. It should be noted that a perineal post is positioned adjacent to the fracture, not to the hip joint, and works as a pivot point for varus angulation [bib_ref] Prevention of complications after treatment of proximal femoral fractures, Archdeacon [/bib_ref]. Additionally, the working space is limited by the torso, especially in obese patients. Without enough space, a guide pin or a nail tends to be inserted in the distal-medial direction rather than along the long axis of the proximal fragment [bib_ref] Traction table-related complications in orthopaedic surgery, Flierl [/bib_ref]. To make space, the torso is often moved to the contralateral side, but then the proximal fragment may be abducted by hip abductor muscles. Under these conditions, the entry point tends to be misplaced laterally. Thus, using a fracture table in intramedullary nailing for a subtrochanteric fracture poses the risk of varus malreduction. In addition, if there is any arthritic change in the hip joint such as central migration or collapse of the head or shortening of the neck, the entry point will be located more medially than in a normal hip joint, and access to the correct entry point will be limited [bib_ref] Subtrochanteric femur fracture below an arthritic hip: opinion: open reduction and internal..., Egol [/bib_ref] [bib_ref] Subtrochanteric femur fracture below an arthritic hip: opinion: open reduction and internal..., Koval [/bib_ref]. Therefore, surgical positioning using some other means besides a fracture table with a perineal post is essential in intramedullary nailing for OA/STF. Intramedullary nailing in the lateral decubitus position on a standard radiolucent table for subtrochanteric fractures of the femur reportedly has satisfactory results [bib_ref] Prevention of complications after treatment of proximal femoral fractures, Archdeacon [/bib_ref] [bib_ref] Closed intramedullary nailing of the femur in the lateral decubitus position, Bishop [/bib_ref] [bib_ref] Lateral decubitus positioning for intramedullary nailing of the femur without the use..., Carr [/bib_ref]. In terms of coronal plane reduction, subtrochanteric fractures may sometimes drift into the valgus before reduction due to gravity [bib_ref] Closed intramedullary nailing of the femur in the lateral decubitus position, Bishop [/bib_ref] , indicating that the adduction of the proximal fragment has not been blocked as might appear on the fracture table where the tendency for varus malreduction may be apparent. Also, this position allows for sufficient working space for inserting the nail. In the lateral position, the distal fragment should be 'flexed' in accordance with the flexion of the proximal fragment by the iliopsoas to obtain a sagittal plane reduction. This position allows the guide pin and nail to be inserted without interference by the torso . ## Figure 6: inserting an intramedullary nail with the lateral decubitus position The injured hip is flexed in order to obtain sufficient working space for inserting the nail. A guide pin or a nail can be inserted without interference by the torso. This position also helps to reduce sagittal deformity. Note the setting of the fluoroscopy. Both the piriformis fossa and trochanteric entry nail can be used with the lateral position [bib_ref] Closed intramedullary nailing of the femur in the lateral decubitus position, Bishop [/bib_ref] [bib_ref] Lateral decubitus positioning for intramedullary nailing of the femur without the use..., Carr [/bib_ref]. However, approaching the trochanteric entry point appears to be more feasible in OA/STF with a shortened neck since the anatomical relationship between the greater trochanter and femoral shaft is usually normal. Access to the entry point is easy and varus malreduction can be avoided. The lateral position also facilitates sagittal plane reduction [bib_ref] Prevention of complications after treatment of proximal femoral fractures, Archdeacon [/bib_ref] [bib_ref] Closed intramedullary nailing of the femur in the lateral decubitus position, Bishop [/bib_ref] [bib_ref] Lateral decubitus positioning for intramedullary nailing of the femur without the use..., Carr [/bib_ref] , allows the avoidance of complications caused by the fracture table [bib_ref] Closed intramedullary nailing of the femur in the lateral decubitus position, Bishop [/bib_ref] [bib_ref] Lateral decubitus positioning for intramedullary nailing of the femur without the use..., Carr [/bib_ref] , and improves accessibility to the assistant [bib_ref] Closed intramedullary nailing of the femur in the lateral decubitus position, Bishop [/bib_ref]. Although manual traction by an assistant is usually successful [bib_ref] Closed intramedullary nailing of the femur in the lateral decubitus position, Bishop [/bib_ref] [bib_ref] Lateral decubitus positioning for intramedullary nailing of the femur without the use..., Carr [/bib_ref] , the Shantz screw, a femoral distractor, or limited open reduction is occasionally required [bib_ref] Prevention of complications after treatment of proximal femoral fractures, Archdeacon [/bib_ref] [bib_ref] Closed intramedullary nailing of the femur in the lateral decubitus position, Bishop [/bib_ref] [bib_ref] Lateral decubitus positioning for intramedullary nailing of the femur without the use..., Carr [/bib_ref]. Rotational reduction can be achieved by rotating the leg and can be assured by a comparison of the true anteroposterior and lateral C-arm images of the uninjured side [bib_ref] Closed intramedullary nailing of the femur in the lateral decubitus position, Bishop [/bib_ref]. Some tricks for the C-arm setting to display accurate anteroposterior and lateral images of the proximal and distal femur should be noted [bib_ref] Closed intramedullary nailing of the femur in the lateral decubitus position, Bishop [/bib_ref] [bib_ref] Lateral decubitus positioning for intramedullary nailing of the femur without the use..., Carr [/bib_ref]. The insertion of distal interlocking can be accomplished by adjusting the C-arm. # Conclusions In intramedullary nailing for subtrochanteric fractures of the femur, varus malreduction must be avoided to prevent nonunion. If the fracture is accompanied by pre-existing osteoarthritis of the ipsilateral hip joint, the entry point will inevitably become misplaced laterally and result in varus, especially when a fracture table with a perineal post is used. In such cases, the lateral decubitus position provides easy access to the entry point and allows accurate reduction. # Additional information disclosures Human subjects: Consent was obtained by all participants in this study. IRB of Tokyo Metropolitan Tama Medical Center issued approval 30-48. Conflicts of interest: In compliance with the ICMJE uniform disclosure form, all authors declare the following: Payment/services info: All authors have declared that no financial support was received from any organization for the submitted work. Financial relationships: All authors have declared that they have no financial relationships at present or within the previous three years with any organizations that might have an interest in the submitted work. Other relationships: All authors have declared that there are no other relationships or activities that could appear to have influenced the submitted work. [fig] FIGURE 5: Solid union was confirmed 10 months postoperatively 2018 Iga et al. Cureus 10(7): e3081. DOI 10.7759/cureus.3081 5 of 8 [/fig]

Three cases of Charles Bonnet Syndrome in patients with advanced glaucomatous visual field loss but preserved visual acuity Purpose: To describe three cases of Charles Bonnet syndrome (CBS) in glaucoma patients with preserved visual acuity. Methods: Three glaucoma patients who had taken part in a recent CBS study were interviewed about their hallucinations. The patients underwent macular optical coherence tomography (OCT) of both eyes. The visual function was evaluated with visual field measurement (Humphrey visual field analyser) and visual acuity testing (Snellen scale). Results: All three patients had preserved visual acuity (≥0.5 in both eyes) and at least one eye with advanced visual field defect (Mean Deviation worse than À12.00 decibel). They all reported vivid visual hallucinations with insight into the unreal nature of the hallucinations. Conclusion: Charles Bonnet syndrome can occur in glaucoma despite preserved visual acuity. Awareness of this relation is desirable among clinicians, as it will improve communication with patients. # Introduction Charles Bonnet syndrome (CBS) is characterized by vivid and complex visual hallucinations in otherwise mentally healthy individuals with visual impairment [bib_ref] Complex visual hallucinations. Clinical and neurobiological insights, Manford [/bib_ref] [bib_ref] The elephant in the room: understanding the pathogenesis of Charles Bonnet syndrome, Carpenter [/bib_ref]. Charles Bonnet first described CBS in 1769; however, it is still widely mistaken for a psychological disorder due to limited knowledge about CBS among both medical practitioners and patients [bib_ref] An examination of the relationship between low vision and Charles Bonnet syndrome, Gilmour [/bib_ref] [bib_ref] Hallucinations experienced by visually impaired: Charles Bonnet Syndrome, Pang [/bib_ref]. Thus, patients are generally reluctant to report it for fear of being perceived as mentally ill [bib_ref] Hallucinations experienced by visually impaired: Charles Bonnet Syndrome, Pang [/bib_ref]. There are various theories explaining the pathogenesis of CBS, but the main hypothesis comprises the deafferentation theory, which suggests that hallucinations are caused by increased activity in the occipital cortex, secondary to reduced sensory input from the eyes (ffytche 2009; [bib_ref] The prevalence and clinical characteristics of Charles Bonnet Syndrome in Danish patients..., Singh [/bib_ref]. Charles Bonnet syndrome can affect anyone with decreased visual function [bib_ref] Complex visual hallucinations. Clinical and neurobiological insights, Manford [/bib_ref] [bib_ref] Complex visual hallucinations in the visually impaired: the Charles Bonnet Syndrome, Menon [/bib_ref] [bib_ref] Charles Bonnet syndrome in Asian patients in a tertiary ophthalmic centre, Tan [/bib_ref] , but it has also been reported after prolonged blindfolding [bib_ref] Visual hallucinations during prolonged blindfolding in sighted subjects, Merabet [/bib_ref]. Many studies suggest that CBS is common among patients with age-related macular degeneration (AMD) [bib_ref] Visual hallucinations in patients with macular degeneration, Holroyd [/bib_ref] [bib_ref] Charles Bonnet syndrome in age-related macular degeneration: the nature and frequency of..., Khan [/bib_ref] [bib_ref] What associates Charles Bonnet syndrome with age-related macular degeneration?, Vojnikovic [/bib_ref] [bib_ref] The prevalence and clinical characteristics of Charles Bonnet Syndrome in Danish patients..., Singh [/bib_ref]. A recent meta-analysis by [bib_ref] Prevalence of Charles Bonnet syndrome in patients with age-related macular degeneration: systematic..., Niazi [/bib_ref] reported a prevalence of 15.8% in AMD patients. Charles Bonnet syndrome is traditionally described in patients with low visual acuity [bib_ref] Charles Bonnet syndrome in glaucoma patients with low vision, Nesher [/bib_ref] [bib_ref] Complex visual hallucinations in the visually impaired: the Charles Bonnet Syndrome, Menon [/bib_ref] [bib_ref] Charles Bonnet syndrome in age-related macular degeneration: the nature and frequency of..., Khan [/bib_ref] [bib_ref] An examination of the relationship between low vision and Charles Bonnet syndrome, Gilmour [/bib_ref]. There are a few reports on CBS in patients with preserved visual acuity [bib_ref] Charles Bonnet syndrome in Asian patients in a tertiary ophthalmic centre, Tan [/bib_ref] [bib_ref] Charles Bonnet syndrome in patients with glaucoma and good acuity, Madill [/bib_ref] , and a review and case series by [bib_ref] Isolated visual hallucinations and the Charles Bonnet syndrome: a review of the..., Gold [/bib_ref] suggests that reduced acuity is not an essential requirement for CBS diagnosis. However, as far as we know, only one previous case report describes CBS in four glaucoma patients with preserved visual acuity [bib_ref] Charles Bonnet syndrome in patients with glaucoma and good acuity, Madill [/bib_ref]. In this report, we describe three cases of CBS in patients with advanced glaucomatous visual field defect but maintained visual acuity. These patients are under clinical care at the glaucoma outpatient department at Sk ane University Hospital in Malm€ o, Sweden. All three patients had taken part in a recent CBS study carried out at the same department and gave their written informed consent to this case report. The patients had undergone automated visual field measurement of both eyes using 24-2 Swedish interactive thresholding algorithm (SITA) programme on the Humphrey Field Analyser (Carl Zeiss Meditec, Dublin, CA, USA). The macula was examined with Swept-Source Optical Coherence Tomography (SS-OCT) (Triton, Topcon, Medical Systems, Tokyo, Japan) at the study visit, and the best-corrected visual acuity (BCVA) was measured using a standard Snellen chart. All three patients had insight into the unreal nature of their hallucinations and had no neurological, psychological [bib_ref] Hallucinations in schizophrenia and Parkinson's disease: an analysis of sensory modalities involved..., Llorca [/bib_ref] or hearing disabilities [bib_ref] Auditory hallucinations in adults with hearing impairment: a large prevalence study, Linszen [/bib_ref] that could cause hallucinations. None of the patients experienced auditory hallucinations. ## Patient 1 A 73-year-old male patient with bilateral primary open-angle glaucoma since 2014, arterial hypertension and diabetes mellitus. The patient reported no history of serious head trauma (other than a mild concussion 40 years ago). The BCVA in the right eye was 0.9 and 0.7 in the left eye. The VF of the right eye showed an advanced visual field defect (VFD) stretching from paracentral superior to the nasal and inferior part of the VF [fig_ref] Figure 1: Visual field defects of three patients with open-angle glaucoma and preserved visual... [/fig_ref]. The mean deviation (MD) was À18.00 decibel (dB) and the visual field index (VFI) 44% in this eye. The left eye had an even more advanced VFD with a concentric visual field loss with only a small central island and a minor temporal part remained [fig_ref] Figure 1: Visual field defects of three patients with open-angle glaucoma and preserved visual... [/fig_ref] , a MD of À26.39 dB and a VFI of 19%. The SS-OCT scanning of the maculae exposed no pathology at all. The patient had undergone selective laser trabeculoplasty (SLT) in both right eye and left eye a couple of years before and was currently on threesubstance ocular hypotensive drug (OHD) for his glaucoma. This patient reported having intermittent visual hallucinations consisting of colourful paintings, blackbirds flying around his room and children running and playing. He reported seeing hundreds of birds flying around his room in the mornings. He also reported seeing children running towards him. The children would suddenly appear in front of him, making him stop walking in order to avoid stepping on the children, but they would quickly disappear making him realize that it was a hallucination. These hallucinations lasted a few seconds. They came in periods, not every day and were usually not associated with unpleasant feelings. ## Patient 2 An 86-year-old female patient with primary open-angle glaucoma (POAG), senile incipient cataract and arterial hypertension. She had no history of head trauma. The BCVA in the right eye was 0.8 and 0.5 in the left eye. Both eyes showed advanced concentric VFD with only small central apertures left [fig_ref] Figure 1: Visual field defects of three patients with open-angle glaucoma and preserved visual... [/fig_ref]. The MD was À28.53 dB and VFI 13% in the right eye. The MD in the left eye was À30.27 and VFI 4%. The OCT revealed subfoveal neurosensory detachment in the right eye and drusenoid maculopathy in the left eye. The patient had undergone SLT treatment in her right eye previously and was on two-substance OHD for her glaucoma. She reported seeing a woman with long boots walking in front of her. She first experienced this hallucination while taking a walk with her husband. She mentioned telling her husband, how strange it was that the woman had been walking in front of them all the time but found out that her husband did not see any woman in front of them. The patient also reported seeing the face of a monkey, placed on her grandchild's face, sitting on the sofa in front of her. Other visual hallucinations that the patient reported consisted of an unknown man in a trench coat. The patient also experienced hallucinations consisting of small colourful duvets with grape and clover prints. These hallucinations lasted for many hours during the day, occurring both in-and outdoors, which was disturbing for the patient. She reported having these hallucinations (the woman in long boots and the face of a monkey) for 5 months until they disappeared in December 2017. However, some of the hallucinations could be transient and intermittent, and they decreased or disappeared during the winter and came back around April when it got lighter. The patient had been experiencing this sporadic reappearance of hallucinations for 3 years in a row. For instance, she had been recently experiencing the hallucination of the man in the black trench coat walking in front of her. ## Patient 3 A 65-year-old female patient with POAG and no history of head trauma. The BCVA in the right eye was 0.8 and 0.7 in the left eye. The VF in the right eye showed a superior arcuate scotoma and preserved central and inferior visual field [fig_ref] Figure 1: Visual field defects of three patients with open-angle glaucoma and preserved visual... [/fig_ref]. The MD was À11.75 dB and the VFI 72% in this eye. The VFD in the left eye comprised of both superior and inferior arcuate scotoma [fig_ref] Figure 1: Visual field defects of three patients with open-angle glaucoma and preserved visual... [/fig_ref] , a MD of À13.44 dB and a VFI of 66%. No pathology was found on the OCT scan. She received both Argon laser trabeculoplasty (ALT) and SLT in her right eye and SLT in her left eye many years ago and underwent trabeculectomy in both eyes a couple of years ago. She was currently using a onesubstance OHD in her left eye only. She reported seeing blackbirds, black cats and geometrical structures. These hallucinations started more than 3 years ago. She experienced these symptoms a few times per month. They mostly occurred in the afternoons and lasted a few seconds, sometimes associated with unpleasant feelings. However, she had learned not to pay attention to the hallucinations, which made the experience less disturbing. # Discussion The hallucinations in the three cases presented consisted of visual images such as people, animals, plants or paintingsa finding, which is consistent with other studies [bib_ref] Isolated visual hallucinations and the Charles Bonnet syndrome: a review of the..., Gold [/bib_ref] [bib_ref] Charles Bonnet syndrome in Asian patients in a tertiary ophthalmic centre, Tan [/bib_ref]. Charles Bonnet syndrome patients are usually aware of the unreal nature of their hallucinations, but due to the stigma and fear associated with hallucinations and lack of awareness among both medical practitioners and patients, CBS is an underreported condition [bib_ref] Hallucinations experienced by visually impaired: Charles Bonnet Syndrome, Pang [/bib_ref]. Charles Bonnet syndrome (CBS) is also traditionally associated with low vision [bib_ref] Complex visual hallucinations in the visually impaired: the Charles Bonnet Syndrome, Menon [/bib_ref] [bib_ref] Charles Bonnet syndrome in age-related macular degeneration: the nature and frequency of..., Khan [/bib_ref] [bib_ref] An examination of the relationship between low vision and Charles Bonnet syndrome, Gilmour [/bib_ref] [bib_ref] The prevalence and clinical characteristics of Charles Bonnet Syndrome in Danish patients..., Singh [/bib_ref] , which increases the risk of overlooking CBS in patients with maintained visual acuity. All patients in this report had preserved corrected visual acuity, but advanced to severe visual field defects, which might lead to hallucinations according to the deafferentation theory. These patients were eventually aware of the unreal nature of the hallucinations, but due to the fear of being labelled as dement or mentally ill, the patients in case one and three had not reported having hallucination to their physicians previously. The patient in case two reported severe distress initially as a result of her hallucinatory experiences. She could not contact her own doctor about the hallucinations due to the fear of being regarded to be mentally ill, but she had contacted the department of ophthalmology and talked to a nurse about her experience, without getting any explanation or information on why she might be experiencing these symptoms. She had later searched the Internet for information about her symptoms and mentioned that to an ophthalmologist, who then confirmed her theory about CBS. # Conclusion It is important to be aware of the fact that CBS can also occur in patients with advanced glaucoma but preserved visual acuity. CBS is a differential diagnosis of many diseases with the symptom of hallucinations. Hence, it is crucial that not only ophthalmologists, but also other clinicians are aware of this condition. Patients experiencing CBS can initially be frightened as a result of their hallucinations, not knowing why they are having those. They are also afraid to seek help due to the stigma associated with hallucinations. Proper information about CBS and reassurance can alleviate these patients' distress and is thereby an important part of the therapy. Especially in glaucoma patients with advanced visual field defect, ophthalmologists should be aware of CBS in order to be able to give proper information to this group of patients, which can reduce confusion and unnecessary suffering. [fig] Figure 1: Visual field defects of three patients with open-angle glaucoma and preserved visual acuity reporting CBS symptoms. VA = Visual acuity (measured with Snellen scale); CBS = Charles Bonnet syndrome. [/fig]

Applications of Chitosan in Surgical and Post-Surgical Materials The continuous advances in surgical procedures require continuous research regarding materials with surgical applications. Biopolymers are widely studied since they usually provide a biocompatible, biodegradable, and non-toxic material. Among them, chitosan is a promising material for the development of formulations and devices with surgical applications due to its intrinsic bacteriostatic, fungistatic, hemostatic, and analgesic properties. A wide range of products has been manufactured with this polymer, including scaffolds, sponges, hydrogels, meshes, membranes, sutures, fibers, and nanoparticles. The growing interest of researchers in the use of chitosan-based materials for tissue regeneration is obvious due to extensive research in the application of chitosan for the regeneration of bone, nervous tissue, cartilage, and soft tissues. Chitosan can serve as a substance for the administration of cell-growth promoters, as well as a support for cellular growth. Another interesting application of chitosan is hemostasis control, with remarkable results in studies comparing the use of chitosan-based dressings with traditional cotton gauzes. In addition, chitosan-based or chitosan-coated surgical materials provide the formulation with antimicrobial activity that has been highly appreciated not only in dressings but also for surgical sutures or meshes. # Introduction Improvement in health care systems has led to a notable increase in expected lifetimes and standard of living [bib_ref] Merits of exercise therapy before and after major surgery, Hoogeboom [/bib_ref]. It is a truism that surgery is nowadays an important factor in improving people's health. Differences in the state of surgical care across countries and among people from various socioeconomic standings have been clearly related to inequalities in levels of death and disability. Even easily treatable problems become serious illnesses when safe and accessible surgical care is not available. Thus, it is estimated that in 2010, around 16.9 million people died due to a lack of access to surgical care. Therefore, surgical care must be understood to be a central part of health systems and strategies to improve the standard of living. Access to surgical care, as well as the quality of surgical conditions and post-surgical follow-up, is important since complications from surgical procedures can also lead to serious disabilities. Post-operative wound complications, such as wound bleeding, serum fluid accumulation, or wound dehiscence, are quite common and may derive from surgeryor patient-related causes [bib_ref] Improving wound healing and preventing surgical site complications of closed surgical incisions:..., Scalise [/bib_ref]. Another frequent and serious complication is surgical site infections-associated with 2-11% of surgical interventions-which cause an increase in treatment costs and mortality [bib_ref] Surgical site infections-Review of current knowledge, methods of prevention, Kolasiński [/bib_ref]. Proper post-operative antibiotic prophylaxis can significantly reduce the rate of infection in both clean and contaminated wounds. Hemorrhagic and thrombotic complications can also occur during or after a surgical procedure. They can be caused by patient-related hemostatic functional problems or by improper closure of [fig_ref] Figure 2: Figure 2 [/fig_ref]. Schematic representation of the commonly used fabrication methods for producing chitosan-based scaffolds-freeze-drying, freeze gelation, salt leaching, electrospinning, and 3D printing. Reprinted with permission from Saravanan et al. [bib_ref] Chitosan based biocomposite scaffolds for bone tissue engineering, Saravanan [/bib_ref]. Copyright 2016, Elsevier. ## Scaffolds Scaffolds are porous solids with controlled geometry and microstructures [fig_ref] Figure 3: Macroscopic and microscopic images of scaffolds fabricated by different methods in bone... [/fig_ref]. They provide extracellular support for cell proliferation and can also serve as a template, for example, to guide tissue regeneration [bib_ref] Chitosan-gelatin scaffolds for tissue engineering: Physicochemical properties and biological response of buffalo..., Thein-Han [/bib_ref]. Biocompatible polymers such as chitosan are considered for the manufacture of implantable scaffolds since they generally ensure absorption/degradability capacity and the absence of toxicity [bib_ref] Chitosan and Its Potential Use as a Scaffold for Tissue Engineering in..., Rodríguez-Vázquez [/bib_ref]. A described method for the fabrication of chitosan-based scaffolds is the lyophilization of chitosan gels and solutions [bib_ref] Porous chitosan scaffolds for tissue engineering, Madihally [/bib_ref]. This technique makes it possible to control the average pore diameterwhich varies from 1 to 250 μm-by means of freezing conditions. An alternative is the manufacturing of chitosan scaffolds by 3D printing, a technique that allows us to obtain systems with a tightly controlled shape and structure [bib_ref] 3D Printed Chitosan Composite Scaffold for Chondrocytes Differentiation, Sahai [/bib_ref]. Finally, it is also possible to obtain self-assembled scaffolds. For this, it is necessary to use a second raw material that, when combined with chitosan, spontaneously forms a scaffold structure. An example is the manufacture of hybrid scaffolds of chitosan and sericin; the positive charges present in the chitosan structure react with the negative charges of aspartic and glutamic acids that are present in the serine structure. This combination also improves cell adhesion and porosity, maintaining the biocompatibility of the system. Schematic representation of the commonly used fabrication methods for producing chitosanbased scaffolds-freeze-drying, freeze gelation, salt leaching, electrospinning, and 3D printing. Reprinted with permission from Saravanan et al. [bib_ref] Chitosan based biocomposite scaffolds for bone tissue engineering, Saravanan [/bib_ref]. Copyright 2016, Elsevier. ## Scaffolds Scaffolds are porous solids with controlled geometry and microstructures [fig_ref] Figure 3: Macroscopic and microscopic images of scaffolds fabricated by different methods in bone... [/fig_ref]. They provide extracellular support for cell proliferation and can also serve as a template, for example, to guide tissue regeneration [bib_ref] Chitosan-gelatin scaffolds for tissue engineering: Physicochemical properties and biological response of buffalo..., Thein-Han [/bib_ref]. Biocompatible polymers such as chitosan are considered for the manufacture of implantable scaffolds since they generally ensure absorption/degradability capacity and the absence of toxicity [bib_ref] Chitosan and Its Potential Use as a Scaffold for Tissue Engineering in..., Rodríguez-Vázquez [/bib_ref]. A described method for the fabrication of chitosan-based scaffolds is the lyophilization of chitosan gels and solutions [bib_ref] Porous chitosan scaffolds for tissue engineering, Madihally [/bib_ref]. This technique makes it possible to control the average pore diameterwhich varies from 1 to 250 µm-by means of freezing conditions. An alternative is the manufacturing of chitosan scaffolds by 3D printing, a technique that allows us to obtain systems with a tightly controlled shape and structure [bib_ref] 3D Printed Chitosan Composite Scaffold for Chondrocytes Differentiation, Sahai [/bib_ref]. Finally, it is also possible to obtain self-assembled scaffolds. For this, it is necessary to use a second raw material that, when combined with chitosan, spontaneously forms a scaffold structure. An example is the manufacture of hybrid scaffolds of chitosan and sericin; the positive charges present in the chitosan structure react with the negative charges of aspartic and glutamic acids that are present in the serine structure. This combination also improves cell adhesion and porosity, maintaining the biocompatibility of the system. ## Sponges Sponges are porous solid systems, similar to scaffolds, but with a different manufacturing process. For the manufacture of chitosan sponges, the polymer is dissolved in an acidic or saline aqueous solution. A surfactant, usually sodium dodecyl sulfate, is then added while stirring at high speed to obtain a foam. A pore-forming agent may be incorporated into the foam at this point in the process. Finally, the system is lyophilized to obtain the chitosan sponge [fig_ref] Figure 4: Schematic representation of the fabrication process of chitosan sponge and the interaction... [/fig_ref] [bib_ref] One-step fabrication of chitosan sponge and its potential for rapid hemostasis in..., Fan [/bib_ref]. The use of chitosan sponges in post-surgical procedures as hemostatic systems has been deeply studied [bib_ref] Using absorbable chitosan hemostatic sponges as a promising surgical dressing, Huang [/bib_ref]. These devices are valued for their biodegradability and antimicrobial activity and their ability to absorb large volumes of fluids. In addition, they can also be used as reservoirs for the release of antibiotics, such as doxycycline, thus improving their antibacterial activity [bib_ref] Antibacterial Activity and Drug Release of Chitosan Sponge Containing Doxycycline Hyclate, Phaechamud [/bib_ref]. Different modifications have been made to chitosan to improve the properties of these sponges. For example, hydrophobically modified chitosan sponges showed improved bleeding control compared to unmodified chitosan [bib_ref] Determination of efficacy of novel modified chitosan sponge dressing in a lethal..., Castro [/bib_ref] ; thiol-modified chitosan also showed excellent hemostatic performance [bib_ref] Antibacterial and Hemostatic Thiol-Modified Chitosan-Immobilized AgNPs Composite Sponges, Wu [/bib_ref] , and alkylated chitosan sponges were able to rapidly absorb large volumes of water and blood [bib_ref] Microchannelled alkylated chitosan sponge to treat noncompressible hemorrhages and facilitate wound healing, Du [/bib_ref]. It is also possible to develop mixed sponges, combining chitosan with other polymers, the combination of chitosan and gelatin being the most studied [bib_ref] Chitosan/gelatin composite sponge is an absorbable surgical hemostatic agent, Lan [/bib_ref] [bib_ref] Healing of skin wounds with a chitosan-gelatin sponge loaded with tannins and..., Lu [/bib_ref]. ## Sponges Sponges are porous solid systems, similar to scaffolds, but with a different manufacturing process. For the manufacture of chitosan sponges, the polymer is dissolved in an acidic or saline aqueous solution. A surfactant, usually sodium dodecyl sulfate, is then added while stirring at high speed to obtain a foam. A pore-forming agent may be incorporated into the foam at this point in the process. Finally, the system is lyophilized to obtain the chitosan sponge [fig_ref] Figure 4: Schematic representation of the fabrication process of chitosan sponge and the interaction... [/fig_ref] [bib_ref] One-step fabrication of chitosan sponge and its potential for rapid hemostasis in..., Fan [/bib_ref]. The use of chitosan sponges in post-surgical procedures as hemostatic systems has been deeply studied [bib_ref] Using absorbable chitosan hemostatic sponges as a promising surgical dressing, Huang [/bib_ref]. These devices are valued for their biodegradability and antimicrobial activity and their ability to absorb large volumes of fluids. In addition, they can also be used as reservoirs for the release of antibiotics, such as doxycycline, thus improving their antibacterial activity [bib_ref] Antibacterial Activity and Drug Release of Chitosan Sponge Containing Doxycycline Hyclate, Phaechamud [/bib_ref]. Different modifications have been made to chitosan to improve the properties of these sponges. For example, hydrophobically modified chitosan sponges showed improved bleeding control compared to unmodified chitosan [bib_ref] Determination of efficacy of novel modified chitosan sponge dressing in a lethal..., Castro [/bib_ref] ; thiol-modified chitosan also showed excellent hemostatic performance [bib_ref] Antibacterial and Hemostatic Thiol-Modified Chitosan-Immobilized AgNPs Composite Sponges, Wu [/bib_ref] , and alkylated chitosan sponges were able to rapidly absorb large volumes of water and blood [bib_ref] Microchannelled alkylated chitosan sponge to treat noncompressible hemorrhages and facilitate wound healing, Du [/bib_ref]. It is also possible to develop mixed sponges, combining chitosan with other polymers, the combination of chitosan and gelatin being the most studied [bib_ref] Chitosan/gelatin composite sponge is an absorbable surgical hemostatic agent, Lan [/bib_ref] [bib_ref] Healing of skin wounds with a chitosan-gelatin sponge loaded with tannins and..., Lu [/bib_ref]. ## Meshes Other devices widely used nowadays in surgery are meshes, replacing sutures in many cases, such as in hernia repair or pelvic floor construction. They are flexible networks formed by crosslinked fibers. Polypropylene meshes are the most common, being a resistant, economical, and non-resorbable material. However, there are still some cases of rejection due to foreign-body reactions and infections of the area [bib_ref] An In Vivo biocompatibility study of surgical meshes made from bacterial cellulose..., Piasecka-Zelga [/bib_ref]. These polypropylene-based meshes have been modified by including chitosan on their surface, which improves biocompatibility and antimicrobial properties, thus accelerating the healing process [bib_ref] Nanodiamond-chitosan functionalized hernia mesh for biocompatibility and antimicrobial activity, Saha [/bib_ref]. This synergy needs to be explored more deeply; another study evaluated the coating of polypropylene meshes with different concentrations of chitosan, but when they were implanted in rats, the adhesion and histopathological parameters were not modified [bib_ref] The effect of a chitosan coating on the adhesive potential and tensile..., Altınel [/bib_ref]. It is also possible to obtain meshes of other materials. For example, a poly(N-isopropylacrylamide)/chitosan hydrogel mesh was able to form a swelling-resistant structure with improved adhesive properties [bib_ref] Hydrogel-mesh composite for wound closure, Gao [/bib_ref]. [bib_ref] One-step fabrication of chitosan sponge and its potential for rapid hemostasis in..., Fan [/bib_ref]. Copyright 2020, IOP Publishing. ## Meshes Other devices widely used nowadays in surgery are meshes, replacing sutures in many cases, such as in hernia repair or pelvic floor construction. They are flexible networks formed by crosslinked fibers. Polypropylene meshes are the most common, being a resistant, economical, and non-resorbable material. However, there are still some cases of rejection due to foreign-body reactions and infections of the area [bib_ref] An In Vivo biocompatibility study of surgical meshes made from bacterial cellulose..., Piasecka-Zelga [/bib_ref]. These polypropylene-based meshes have been modified by including chitosan on their surface, which improves biocompatibility and antimicrobial properties, thus accelerating the healing process [bib_ref] Nanodiamond-chitosan functionalized hernia mesh for biocompatibility and antimicrobial activity, Saha [/bib_ref]. This synergy needs to be explored more deeply; another study evaluated the coating of polypropylene meshes with different concentrations of chitosan, but when they were implanted in rats, the adhesion and histopathological parameters were not modified [bib_ref] The effect of a chitosan coating on the adhesive potential and tensile..., Altınel [/bib_ref]. It is also possible to obtain meshes of other materials. For example, a poly(N-isopropylacrylamide)/chitosan hydrogel mesh was able to form a swelling-resistant structure with improved adhesive properties [bib_ref] Hydrogel-mesh composite for wound closure, Gao [/bib_ref]. ## Membranes Similar to meshes, membranes or films can be manufactured. They are also flexible polymeric layers, but with a continuous structure instead of a fiber network. The standardized manufacturing technique for chitosan membranes involves the preparation of an acid solution of chitosan, which is freeze-dried after being poured into a mold. After this process, the membrane is treated with alkali to displace the acid used to dissolve the polymer and facilitate the polymerization of chitosan. Finally, the system is dried to obtain chitosan membranes [bib_ref] Dense chitosan surgical membranes produced by a coincident compression-dehydration process, Dooley [/bib_ref]. Some authors have replaced the lyophilization step with the immersion of the membrane in liquid nitrogen for 10 s [bib_ref] Guided bone regeneration with tripolyphosphate cross-linked asymmetric chitosan membrane, Ma [/bib_ref]. Chitosan acetate films were developed for biomedical applications. The influence of the inclusion of glycerol, oleic acid, and a mixture of them as plasticizers was evaluated. All of them became biodegradable films suitable for skin recovery [bib_ref] Chitosan composite films. Biomedical applications, Cárdenas [/bib_ref]. Membranes can also be manufactured as multilayer films, allowing different materials to be combined and providing different properties to the membrane. For example, multilayer chitosan/poly(L-lactic acid) membranes combine the biocompatibility and cell-growth promotion of chitosan with the mechanical [fig_ref] Figure 4: Schematic representation of the fabrication process of chitosan sponge and the interaction... [/fig_ref]. Schematic representation of the fabrication process of chitosan sponge and the interaction of its molecules. Reprinted with permission from Fan et al. [bib_ref] One-step fabrication of chitosan sponge and its potential for rapid hemostasis in..., Fan [/bib_ref]. Copyright 2020, IOP Publishing. ## Membranes Similar to meshes, membranes or films can be manufactured. They are also flexible polymeric layers, but with a continuous structure instead of a fiber network. The standardized manufacturing technique for chitosan membranes involves the preparation of an acid solution of chitosan, which is freeze-dried after being poured into a mold. After this process, the membrane is treated with alkali to displace the acid used to dissolve the polymer and facilitate the polymerization of chitosan. Finally, the system is dried to obtain chitosan membranes [bib_ref] Dense chitosan surgical membranes produced by a coincident compression-dehydration process, Dooley [/bib_ref]. Some authors have replaced the lyophilization step with the immersion of the membrane in liquid nitrogen for 10 s [bib_ref] Guided bone regeneration with tripolyphosphate cross-linked asymmetric chitosan membrane, Ma [/bib_ref]. Chitosan acetate films were developed for biomedical applications. The influence of the inclusion of glycerol, oleic acid, and a mixture of them as plasticizers was evaluated. All of them became biodegradable films suitable for skin recovery [bib_ref] Chitosan composite films. Biomedical applications, Cárdenas [/bib_ref]. Membranes can also be manufactured as multilayer films, allowing different materials to be combined and providing different properties to the membrane. For example, multilayer chitosan/poly(L-lactic acid) membranes combine the biocompatibility and cell-growth promotion of chitosan with the mechanical strength of poly(L-lactic acid) [bib_ref] Chitosan/poly(L-lactic acid) multilayered membrane for guided tissue regeneration, Ku [/bib_ref]. HemCon ® is a commercialized chitosan-based film that is recommended for preventing blood loss, thanks to its hemostatic properties, while providing an antibacterial barrier that avoids wound infection [bib_ref] A Special Report on the Chitosan-based Hemostatic Dressing: Experience in Current Combat..., Wedmore [/bib_ref]. Similarly, CELOX TM is a chitosan hemostatic dressing that has shown equivalence to HemCon ® in the control of bleeding [bib_ref] Comparison of ChitoFlex ® , CELOX™, and QuikClot ® in Control of..., Devlin [/bib_ref]. Chitosan membranes can also be used in surgical procedures in combination with polypropylene mesh; they provide an antiadhesive barrier that prevents the adhesion of peritoneal tissue to the mesh [bib_ref] Use of chitosan membrane associated with polypropylene mesh to prevent peritoneal adhesion..., Paulo [/bib_ref]. There are also references to the use of chitosan membranes in guided tissue regeneration [bib_ref] Chitosan as a barrier membrane material in periodontal tissue regeneration, Xu [/bib_ref]. A collagen film impregnated with a layer of chitosan has been used as a barrier membrane for managing periodontal furcation, resulting in excellent biological acceptance and low gingival recession [bib_ref] Management of periodontal furcation defects by guided tissue regeneration using collagen-Chitosan as..., Harikumar [/bib_ref]. Chitosan membranes have also been developed for guided bone regeneration; these devices have antimicrobial properties and are capable of inducing angiogenesis, thus promoting bone regeneration in in vivo studies [bib_ref] Asymmetric Collagen/chitosan Membrane Containing Minocycline-loaded Chitosan Nanoparticles for Guided Bone Regeneration, Ma [/bib_ref]. ## Hydrogels Hydrogels have also been studied for surgical applications. These systems consist of a liquid phase, which generally comprises 90% of the formulation, trapped in a solid phase that gives the gel its structure [bib_ref] Chitosan-based biomaterials for tissue engineering, Croisier [/bib_ref]. This water content makes these systems highly biocompatible, and their soft consistency prevents damage to surrounding tissues. Chitosan hydrogels show similar mechanical properties to connective tissues, which favors tissue regeneration [bib_ref] Effectiveness of chitosan scaffold in skin, bone and cartilage healing, Oryan [/bib_ref]. Three types of chitosan hydrogels for surgical applications are described in the literature: physically associated hydrogels based on the crosslinking of chitosan chains through hydrogen bonds or electrostatic interactions; coordination complex crosslinked hydrogels, which require metal ions to form covalent bonds with chitosan, making them less suitable for biomedical applications; and chemically crosslinked hydrogels that undergo irreversible gelation via covalent bonds but require modifications to the chitosan structure [bib_ref] Structure and interactions in chitosan hydrogels formed by complexation or aggregation for..., Berger [/bib_ref] [bib_ref] Preparation of Chitosan Gel Beads by Ionotropic Molybdate Gelation, Dambies [/bib_ref]. Chitosan hydrogels have been studied for different applications, such as hemorrhage control, dental and bone regeneration, and treatment of wound infections [bib_ref] Fabrication of Hydroxypropyl Chitosan/Soy Protein Isolate Hydrogel for Effective Hemorrhage Control, Zhao [/bib_ref] [bib_ref] Microporous methacrylated glycol chitosan-montmorillonite nanocomposite hydrogel for bone tissue engineering, Cui [/bib_ref] [bib_ref] Cefuroxime conjugated chitosan hydrogel for treatment of wound infections, Pawar [/bib_ref]. It is worth mentioning the possibility of developing thermosensitive hydrogels, which are capable of forming a gel after administration in response to increased temperature [bib_ref] Glycerophosphate-based chitosan thermosensitive hydrogels and their biomedical applications, Zhou [/bib_ref]. For example, a thermosensitive chitosan/gelatin hydrogel was developed for the sustained release of stem cells in therapeutic angiogenesis [bib_ref] Sustained release of adipose-derived stem cells by thermosensitive chitosan/gelatin hydrogel for therapeutic..., Cheng [/bib_ref]. ## Nanofibers and nanoparticles Another formulation prepared from chitosan is nanofibers. The electrospinning technique allows the fabrication of chitosan nanofibers with a diameter of a few nanometers and, therefore, a large specific surface [bib_ref] Effectiveness of chitosan scaffold in skin, bone and cartilage healing, Oryan [/bib_ref]. Porous meshes or scaffolds can be manufactured from the developed nanofibers [bib_ref] Chitosan nanofiber scaffold improves bone healing via stimulating trabecular bone production due..., Ho [/bib_ref] [bib_ref] Enhanced effects of electrospun collagen-chitosan nanofiber membranes on guided bone regeneration, Guo [/bib_ref]. Chitosan nanoparticles have also been developed for surgical applications. These formulations can be included in another system to enhance their activity. For example, melatonin-loaded lecithin-chitosan nanoparticles have been included in a hydrogel for topical administration as a wound-healing promoter. In vivo studies have shown the induction of angiogenic and fibroblast proliferation after the administration of nanoparticles [bib_ref] Melatonin loaded lecithin-chitosan nanoparticles improved the wound healing in diabetic rats, Correa [/bib_ref]. Another example is the minocycline-loaded nanoparticles developed by Ma et al., which were included in a collagen/chitosan membrane for guided bone regeneration [bib_ref] Asymmetric Collagen/chitosan Membrane Containing Minocycline-loaded Chitosan Nanoparticles for Guided Bone Regeneration, Ma [/bib_ref]. ## Surgical and post-surgical applications of chitosan-based devices ## Nerve regeneration Nowadays, biomaterials are being widely studied for nerve regeneration. Among them, molecules of natural origin, such as collagen, silk, and chitosan, can be highlighted [bib_ref] Materials for peripheral nerve repair constructs: Natural proteins or synthetic polymers?, Gregory [/bib_ref]. Chitosan has several properties that make it especially promising for this application. Its biocompatibility, its antibacterial activity, and its ability to interact with regeneration-associated cells and the neural environment, promoting axonal regeneration and less neuroma formation, are some of these properties [bib_ref] Materials for peripheral nerve repair constructs: Natural proteins or synthetic polymers?, Gregory [/bib_ref] [bib_ref] Relevance and Recent Developments of Chitosan in Peripheral Nerve Surgery, Boecker [/bib_ref]. Chitosan has gained special importance in the field of regeneration of peripheral nerves. Among other reasons, this is due to the ability to simulate the multilayer structure of these nerves in nerve conduits designed with artificial tissue, thanks to the physical and chemical properties of this polysaccharide [bib_ref] Relevance and Recent Developments of Chitosan in Peripheral Nerve Surgery, Boecker [/bib_ref]. However, it must be kept in mind that the degree of acetylation of chitosan can influence the mechanical properties and degradation time of the conduit and, consequently, the regeneration process [bib_ref] Relevance and Recent Developments of Chitosan in Peripheral Nerve Surgery, Boecker [/bib_ref]. Nerve-guiding conduits could help overcome problems of donor nerve availability and secondary injuries associated with autograft, which is the treatment of choice for peripheral nerve injuries [bib_ref] Situ Fabrication of Nerve Growth Factor Encapsulated Chitosan Nanoparticles in Oxidized Bacterial..., Wei [/bib_ref]. Along these lines, in 2015, the FDA approved a chitosan-based nerve conduit under the tradename Reaxon ® Nerve Guide for bridging gaps of up to 26 mm in peripheral nerves [bib_ref] Natural-Based Biomaterials for Peripheral Nerve Injury Repair, Fornasari [/bib_ref]. It is a flexible nerve guide, the positive surface charge of which establishes an electrostatic interaction with negatively charged biomolecules and cellular components, contributing to nerve regeneration. Its hydrogel wall favrors the transport of oxygen and nutrients to the injured nerve, thus creating an optimal environment for Schwann cells. In addition, this nerve guidance inhibits the growth of fibroblastic tissue and scar formation. Despite the above, researchers continue working to achieve the "ideal nerve tube", which, according to Bąk et al., would have the following properties: "biodegradability, porosity and permeability of the tube wall, presence of an inner scaffold made of fibers or filaments, the capacity to sustain cell livability and promote cell migration, the ability to secrete growth factors and electrical conductivity". A large number of research works related to the use of chitosan in nerve regeneration can be found in the literature. Some of them are collected in . . Studies of chitosan-based systems for nerve regeneration. ## Chitosan form/system treated nerve main results reference Chitosan-selenium biodegradable nanocomposite conduit Sciatic Number and diameter of myelinated fibers significantly higher against chitosan [bib_ref] Fabrication and transplantation of chitosan-selenium biodegradable nanocomposite conduit on transected sciatic nerve:..., Dolkhani [/bib_ref] Chitosan tubes with different degrees of acetylation Sciatic Intermediate degree of acetylation as the best choice in terms of degradation and regeneration efficacy [bib_ref] Chitosan tubes of varying degrees of acetylation for bridging peripheral nerve defects, Haastert-Talini [/bib_ref] Laser-activated chitosan Posterior tibial Good functional recovery and tensile strength with laser-activated chitosan [bib_ref] Nerve transection repair using laser-activated chitosan in a rat model, Bhatt [/bib_ref] Double-layer composite hydrogel conduit based on chitosan Sciatic Significant regeneration against chitosan hollow conduit and repair ability comparable to autologous transplantation when loaded with 7,8-dihydroxyflavone [bib_ref] Multifunctional Double-Layer Composite Hydrogel Conduit Based on Chitosan for Peripheral Nerve Repairing, Deng [/bib_ref] Chitosan gel absorbed into Spongostan ® Facial Positive effect of chitosan gel in nerve healing and better results when combined with platelet-rich plasma [bib_ref] Effects of chitosan and platelet-rich plasma on facial nerve regeneration in an..., Şahin [/bib_ref] Corrugated chitosan-film-enhanced chitosan nerve guides Median Accelerated functional recovery and thicker myelin sheats against other nerve guides [bib_ref] Two-Chambered Chitosan Nerve Guides with Increased Bendability Support Recovery of Skilled Forelimb..., Dietzmeyer [/bib_ref] Aligned chitosan nanofiber hydrogel grafted with peptides as conduit filler ## Sciatic Enhanced nerve regeneration, secretion of neurotrophic factors, vascular penetration, and functional recovery than other conduitsChitosan functionalized magnetic nanoparticles Sciatic Nerve outreach without surgical intervention and better functional outcome versus without treatmentPolycaprolactone/chitosanhydroxyapatite hybrid implants Peripheral Possibility of controlling the diffusion of oxygen and nutrients and invariable mechanical properties for up to 28 days [bib_ref] Fabrication and Characterization of Polycaprolactone/Chitosan-Hydroxyapatite Hybrid Implants for Peripheral Nerve Regeneration, Nawrotek [/bib_ref] The conduits currently available for nerve regeneration are hollow tubes, which are generally associated with poor recovery and difficulty in nerve extension due to scar formation. For this reason, research has focused on the inclusion of fillers (such as some of the examples in and growth factors or the Schwann cells in them to enhance the regeneration process [bib_ref] Effect of Local Administration of Verapamil Combined with Chitosan Based Hybrid Nanofiber..., Mohajer [/bib_ref] [bib_ref] The electrostimulation and scar inhibition effect of chitosan/oxidized hydroxyethyl cellulose/reduced graphene oxide/asiaticoside..., Zheng [/bib_ref]. Chitosan is commonly used for the development of different vehicles to transport drugs or cells for nerve regeneration. For instance, verapamil-filled chitosan/polycaprolactone hybrid nanofiber conduits were tested in rats for regeneration of the transected sciatic nerve, suggesting a beneficial effect of calcium channel blockers on nerve recovery [bib_ref] Effect of Local Administration of Verapamil Combined with Chitosan Based Hybrid Nanofiber..., Mohajer [/bib_ref]. Regarding the enrichment of chitosan-based systems with cells or growth factors, Liu et al. demonstrated the efficacy of combining chitosan nerve conduits with nerve growth factor microspheres for the repair of injured facial nerves in rabbits [bib_ref] Chitosan conduits combined with nerve growth factor microspheres repair facial nerve defects, Liu [/bib_ref] , and Zhao et al. proposed a sponge-based hydroxypropyl chitosan/soy protein isolate composite conduit with bone marrow mesenchymal-stem-cell-derived Schwann cells for sciatic nerve regeneration in rats [bib_ref] Hydroxypropyl Chitosan/Soy Protein Isolate Conduits Promote Peripheral Nerve Regeneration, Zhao [/bib_ref]. It is worth mentioning that chitosan has also been studied for the treatment of spinal cord injuries. Wang et al. formulated valproic-acid-labeled chitosan nanoparticles to combine the beneficial effects of valproic acid and chitosan nanoparticles for injured spinal cords [bib_ref] Valproic acid-labeled chitosan nanoparticles promote recovery of neuronal injury after spinal cord..., Wang [/bib_ref]. Among the findings of this study were the promotion of functional recovery and tissue repair and the improvement of the integrity of the blood-spinal cord barrier. In contrast, Yang et al. developed a chitosan scaffold composed of graphene oxide that was transplanted into rats [bib_ref] Graphene oxide-composited chitosan scaffold contributes to functional recovery of injured spinal cord..., Huang [/bib_ref]. The results indicated that the scaffold promoted nerve cell growth, neuron migration, and tissue regeneration, in addition to showing better functional recovery than other chitosan scaffolds. ## Bone regeneration Bone is an organ that can heal itself, thanks to a continuous remodeling process. However, this is only effective for small lesions (<8 mm); larger defects require bone substitutes and surgery [bib_ref] The role of natural polymers in bone tissue engineering, Guo [/bib_ref]. Autografts, allografts, surgical reconstruction, and metal implants represent the traditional options for bone reconstruction, with autologous bone grafting being the treatment of choice [bib_ref] Chitosan as a Bone Scaffold Biomaterial, Kozusko [/bib_ref]. Nevertheless, they are associated with issues such as the risk of disease transmission, rejection, or the need for repeated surgeries [bib_ref] Chitosan in Surface Modification for Bone Tissue Engineering Applications, Abinaya [/bib_ref]. For this reason, tissue-engineered bone grafts are being studied as possible substitutes. According to Venkatesan et al., the implanted material must be biocompatible, osteoconductive, highly porous, and with good mechanical properties [bib_ref] Chitosan Composites for Bone Tissue Engineering-An Overview, Venkatesan [/bib_ref]. Furthermore, since bone is a highly vascularized tissue, implant vascularization is important for successful bone regeneration [bib_ref] The role of natural polymers in bone tissue engineering, Guo [/bib_ref]. Given their biocompatibility and biodegradability, natural polymers are widely used materials for bone grafting. Among them, chitosan, collagen, silk fibroin, gelatin, cellulose, alginate, and starch can be highlighted [bib_ref] The role of natural polymers in bone tissue engineering, Guo [/bib_ref]. Chitosan is an excellent candidate for bone reconstruction as it has antimicrobial properties, can generate porous structures suitable for cell growth and osteoconduction, and promotes osteoblast and mesenchymal cell proliferation and neovascularization in vivo. Furthermore, its structure is similar to glycosaminoglycans, a component of the bone extracellular matrix [bib_ref] The role of natural polymers in bone tissue engineering, Guo [/bib_ref] [bib_ref] Chitosan Composites for Bone Tissue Engineering-An Overview, Venkatesan [/bib_ref] [bib_ref] A review of chitosan and its derivatives in bone tissue engineering, Logithkumar [/bib_ref]. Chitosan has been formulated in different systems for bone tissue engineering, including scaffolds, sponges, hydrogels, micro-nanospheres, and membranes, among others [bib_ref] The role of natural polymers in bone tissue engineering, Guo [/bib_ref] [bib_ref] Chitosan-Based Scaffold for Mineralized Tissues Regeneration, Sukpaita [/bib_ref]. According to Sukpaita et al., 3D biomaterial scaffolds are one of the three main pillars of bone tissue engineering, along with osteogenic stem cells and bioactive molecules such as growth factors or drugs. These scaffolds can be defined as implants capable of accelerating the formation of new bone integrated into the host bone without causing adverse reactions [bib_ref] Chitosan-Based Scaffold for Mineralized Tissues Regeneration, Sukpaita [/bib_ref]. A large number of studies on chitosan-based scaffolds for this application can be found in the literature. It is usual to develop systems that combine chitosan with other compounds to improve properties such as osteoconductivity and mechanical properties and thus obtain materials that mimic natural bone as much as possible [bib_ref] Chitosan Composites for Bone Tissue Engineering-An Overview, Venkatesan [/bib_ref] [bib_ref] Bio-Functionalized Chitosan for Bone Tissue Engineering, Brun [/bib_ref]. Among these reinforcing components, inorganic compounds such as hydroxyapatite stand out. Soriente et al. manufactured chitosan/hydroxyapatite composite scaffolds by the sol-gel method and subsequent lyophilization [bib_ref] Chitosan/hydroxyapatite nanocomposite scaffolds to modulate osteogenic and inflammatory response, Soriente [/bib_ref]. In these scaffolds, hydroxyapatite nanoparticles were embedded in the chitosan matrix. They observed that the higher the content of the inorganic component in the scaffold, the better the osteogenic differentiation of mesenchymal cells towards osteoblasts and the greater the anti-inflammatory response. The combination of chitosan with other polymers to obtain composite materials also frequently improves its applicability to bone regeneration [bib_ref] A review of chitosan and its derivatives in bone tissue engineering, Logithkumar [/bib_ref]. Guo et al. manufactured electrospun nanofiber membranes made from collagen and chitosan [bib_ref] Enhanced effects of electrospun collagen-chitosan nanofiber membranes on guided bone regeneration, Guo [/bib_ref]. The results of this work showed that these membranes had higher tensile strength, a more stable rate of degradation, and better in vivo results in repairing calvarial bone defects than electrospun membranes based only on collagen. Other strategies that have been proposed to improve the characteristics of chitosanbased systems for the repair of bone defects are the following: ## - Inclusion of nanohydroxyapatite/chitosan microspheres in chitosan membranes, with better results than pure chitosan membranes in terms of mechanical properties [bib_ref] Reinforced chitosan membranes by microspheres for guided bone regeneration, Huang [/bib_ref]. - Incorporation of halloysite chitosan-modified nanotubes in thermosensitive hydrogels of chitosan/glycerophosphate, thus improving the mechanical properties and proliferation of the encapsulated stem cells with respect to the hydrogel [bib_ref] Injectable chitosan hydrogel embedding modified Halloysite nanotubes for bone tissue engineering, Kazemi-Aghdam [/bib_ref]. - Functionalization of chitosan to give rise to derivatives such as sulfated methacrylate chitosan, which is associated with vascularization [bib_ref] Effect of sulfated chitosan hydrogel on vascularization and osteogenesis, Jiang [/bib_ref] , and peptide-functionalized chitosan, which promotes a greater adhesion and proliferation of osteoblasts than chitosan [bib_ref] Bio-Functionalized Chitosan for Bone Tissue Engineering, Brun [/bib_ref]. Despite what has been described above, research works on systems based solely on chitosan for bone regeneration can also be found in the literature. For instance, Sukul et al. formulated sponges based on chitosan with different degrees of deacetylation and molecular weights and determined the influence of these parameters on the adhesion, growth, and differentiation of human osteoblasts [bib_ref] In Vitro biological response of human osteoblasts in 3D chitosan sponges with..., Sukul [/bib_ref]. Chitosan-based systems loaded with drugs, stem cells, or growth factors for bone regeneration have also been developed. Based on the osteogenic potential of simvastatin, Murali et al. formulated fatty-acid-modified electrospun chitosan membranes loaded with the drug [bib_ref] Simvastatin loaded chitosan guided bone regeneration membranes stimulate bone healing, Murali [/bib_ref]. Regarding the incorporation of stem cells, stem cells derived from human urine have been loaded onto hybrid scaffolds composed of biphasic sponges of calcium phosphate and chitosan [bib_ref] Urine-derived stem cells loaded onto a chitosan-optimized biphasic calcium-phosphate scaffold for repairing..., Liu [/bib_ref]. Finally, porous chitosan scaffolds were loaded with protein growth factors for the repair of tibial defects in rabbits [bib_ref] Protein growth factors loaded highly porous chitosan scaffold: A comparison of bone..., Nandi [/bib_ref]. A specific application of chitosan within bone regeneration is periodontal bone repair. According to Xu et al., "periodontitis is a progressive infectious inflammatory disease, which leads to alveolar bone resorption and loss of periodontal attachment" [bib_ref] Metformin-loaded β-TCP/CTS/SBA-15 composite scaffolds promote alveolar bone regeneration in a rat model..., Xu [/bib_ref]. Alveolar bone resorption causes serious and irreversible damage that cannot normally be repaired by physiological mechanisms. For this reason, bone grafts are often used to treat bone defects associated with periodontitis. One of the problems with conventional therapy is the proliferation of adjacent tissues into the defect before the migration of osteoblasts; this hinders bone regeneration and causes drawbacks such as implant encapsulation. Thus, guided bone regeneration (GBR) has emerged as a tool to prevent the ingrowth of these surrounding tissues, being one of the most effective methods in alveolar bone regeneration. Among the systems studied for GBR therapy, membranes can be highlighted. Due to its properties, such as mucoadhesiveness, antimicrobial activity, and biocompatibility, chitosan has enormous potential for the development of GBR barrier membranes [bib_ref] Chitosan as a barrier membrane material in periodontal tissue regeneration, Xu [/bib_ref]. An example of the success of resorbable barrier membranes based on biopolymers is the work developed by Tamburaci et al. [bib_ref] Development of Si doped nano hydroxyapatite reinforced bilayer chitosan nanocomposite barrier membranes..., Tamburaci [/bib_ref]. They fabricated a novel bilayer membrane composed of an upper layer of chitosan/polyethylene oxide, obtained by electrospinning, and a microporous sublayer of nanohydroxyapatite particles doped with chitosan/silica, obtained by lyophilization [fig_ref] Figure 5: Stereomicroscopy images of bilayer membranes [/fig_ref]. The results showed that the first layer acted as a barrier, preventing the unwanted proliferation of fibroblasts, and the second layer participated in osteogenic activity at the site of the bone defect. The incorporation of drugs in chitosan-based formulations represents a common strategy for periodontal bone repair. Specifically, collagen/chitosan membranes containing minocycline-loaded chitosan nanoparticles and minocycline-controlled release hydroxyapatite/chitosan composites have been studied to prevent infections associated with this type of bone regeneration [bib_ref] Asymmetric Collagen/chitosan Membrane Containing Minocycline-loaded Chitosan Nanoparticles for Guided Bone Regeneration, Ma [/bib_ref] [bib_ref] Controlled release of minocycline in hydroxyapatite/chitosan composite for periodontal bone defect repair, Gao [/bib_ref]. In addition, based on the anti-inflammatory properties of metformin, composite scaffolds combining β-tricalcium phosphate, chitosan, and mesoporous silica SBA-15 have been loaded with this drug to repair alveolar bone defects in rats [bib_ref] Metformin-loaded β-TCP/CTS/SBA-15 composite scaffolds promote alveolar bone regeneration in a rat model..., Xu [/bib_ref]. Periodontal regeneration includes not only the regeneration of alveolar bone but also the regeneration of cementum, periodontal ligament, and gingiva [bib_ref] Chitosan as a barrier membrane material in periodontal tissue regeneration, Xu [/bib_ref]. Thus, Varoni et al. proposed a three-layer scaffold based on chitosan for multi-tissue periodontal healing [bib_ref] Chitosan-Based Trilayer Scaffold for Multitissue Periodontal Regeneration, Varoni [/bib_ref]. This system consisted of two porous compartments formed by chitosan of low and medium molecular weights, respectively, crosslinked with genipin and subsequently lyophilized, and a third compartment obtained by the electrochemical deposition of chitosan. The first and second compartments were intended for bone and gingival repair, while the third aimed to regenerate the periodontal ligament. Mar. Drugs 2022, 20, x FOR PEER barrier, preventing the unwanted proliferation of fibroblasts, and the second layer participated in osteogenic activity at the site of the bone defect. The incorporation of drugs in chitosan-based formulations represents a common strategy for periodontal bone repair. Specifically, collagen/chitosan membranes containing minocycline-loaded chitosan nanoparticles and minocycline-controlled release hydroxyapatite/chitosan composites have been studied to prevent infections associated with this type of bone regeneration [bib_ref] Asymmetric Collagen/chitosan Membrane Containing Minocycline-loaded Chitosan Nanoparticles for Guided Bone Regeneration, Ma [/bib_ref] [bib_ref] Controlled release of minocycline in hydroxyapatite/chitosan composite for periodontal bone defect repair, Gao [/bib_ref]. In addition, based on the anti-inflammatory properties of metformin, composite scaffolds combining β-tricalcium phosphate, chitosan, and mesoporous silica SBA-15 have been loaded with this drug to repair alveolar bone defects in rats [bib_ref] Metformin-loaded β-TCP/CTS/SBA-15 composite scaffolds promote alveolar bone regeneration in a rat model..., Xu [/bib_ref]. Periodontal regeneration includes not only the regeneration of alveolar bone but also the regeneration of cementum, periodontal ligament, and gingiva [bib_ref] Chitosan as a barrier membrane material in periodontal tissue regeneration, Xu [/bib_ref]. Thus, Varoni et al. proposed a three-layer scaffold based on chitosan for multi-tissue periodontal healing [bib_ref] Chitosan-Based Trilayer Scaffold for Multitissue Periodontal Regeneration, Varoni [/bib_ref]. This system consisted of two porous compartments formed by chitosan of low and medium molecular weights, respectively, crosslinked with genipin and subsequently lyophilized, and a third compartment obtained by the electrochemical deposition of Finally, Yan et al. combined an enzymatically solidified chitosan hydrogel, the gelation of which was a pH-dependent process based on the enzymatic hydrolysis of urea with periodontal ligament cells for the repair of periodontal defects in rats [bib_ref] Periodontal Tissue Regeneration Using Enzymatically Solidified Chitosan Hydrogels With or Without Cell..., Yan [/bib_ref]. Although the cells did not lead to the expected result, cell-free hydrogels showed great potential for this purpose in terms of functional ligament length. It is worth mentioning that chitosan has also been used for two particular purposes within periodontal regeneration: augmentation of the maxillary sinus floor and treatment of periodontitis/peri-implantitis. Maxillary sinus floor augmentation is a surgery that consists of lifting the sinus membrane and placing a bone graft under it to increase bone support prior to a dental implant [bib_ref] Maxillary Sinus Floor Augmentation to Enable One-Stage Implant Placement by Using Bovine..., Barbu [/bib_ref]. An example of the use of chitosan in the development of materials for this procedure is the work of Li et al., who prepared an injectable thermosensitive hydrogel composed of chitosan, β-sodium glycerophosphate disodium salt hydrate, and gelatin and loaded it with erythropoietin as a minimally invasive tool for maxillary sinus floor augmentation [bib_ref] Injectable thermosensitive chitosan/gelatinbased hydrogel carried erythropoietin to effectively enhance maxillary sinus floor..., Li [/bib_ref]. The system turned quickly into a gel at body temperature, allowing easy handling and a rapid formation of the final formulation once administered, and promoted new bone formation by intramembranous osteogenesis in vivo, thanks to the growth factor. Periodontitis is defined as a pathology that causes inflammation of the connective tissue and loss of bone tissue surrounding the teeth; it is called peri-implantitis in the case of a dental implant. They are usually associated with plaque biofilm dysbiosis [bib_ref] Treatment of residual pockets using an oscillating chitosan device versus regular curettes..., Hussain [/bib_ref] and bacterial biofilms on the implant surface [bib_ref] Peri-implantitis is not periodontitis: Scientific discoveries shed light on microbiome-biomaterial interactions that..., Kotsakis [/bib_ref] , respectively. Treatment includes mechanical debridement and infection control with proper oral hygiene and patient care, up to surgery in the most severe cases [bib_ref] Treatment of residual pockets using an oscillating chitosan device versus regular curettes..., Hussain [/bib_ref] [bib_ref] Supportive treatment following peri-implantitis surgery; an RCT using titanium curettes or chitosan..., Koldsland [/bib_ref]. Research in this field pursues new non-surgical treatments that are more effective than those currently available. The potential of chitosan for bone regeneration and its antimicrobial properties, among others, make it very interesting for the development of different systems for non-surgical periodontal treatment. Thus, a chitosan brush seated on an oscillating dental handpiece has been tested in a clinical trial for the debridement of residual pockets in the treatment of periodontitis compared to regular curettes. The results of this study indicated the potential of the chitosan brush for this application [bib_ref] Treatment of residual pockets using an oscillating chitosan device versus regular curettes..., Hussain [/bib_ref]. Zhou et al. fabricated hyaluronic acid/chitosan composite hydrogels loaded with dexamethasone for the treatment of peri-implantitis. These systems showed sustained drug release and adequate antibacterial and anti-inflammatory effects for the intended purpose [bib_ref] Preparation and characterization of antibacterial and anti-inflammatory hyaluronic acid-chitosandexamethasone hydrogels for peri-implantitis..., Zhou [/bib_ref]. ## Cartilage regeneration and viscosupplementation Adult articular cartilage is an avascular tissue, which hinders its self-healing and leads to surgical procedures such as microfracture for its repair. However, these treatments are not always as successful as expected, so other tools are being studied for this purpose, such as biomaterials [bib_ref] Strong and Elastic Chitosan/Silk Fibroin Hydrogels Incorporated with Growth-Factor-Loaded Microspheres for Cartilage..., Min [/bib_ref]. Its similarity with glycosaminoglycan and hyaluronic acid makes chitosan an interesting option for cartilage repair [bib_ref] Chitosan scaffolds for osteochondral tissue regeneration, Abarrategi [/bib_ref]. Moreover, chitosan is a polymer that shows great potential for viscosupplementation, which, according to Comblain et al., "is a process that aims to restore the normal rheological properties of synovial fluid" [bib_ref] Chitosan: A promising polymer for cartilage repair and viscosupplementation, Comblain [/bib_ref]. Hyaluronic acid, which is frequently used for viscosupplementation, has the disadvantage of remaining in the joint cavity for a short period. Based on what was described above, different works can be found in the literature on the use of chitosan for the repair of chondral and osteochondral defects (the latter involves a joint injury in which both the cartilage and the underlying bone are affected [bib_ref] 14-Bone tissue engineering, Santin [/bib_ref] and viscosupplementation. Some of them are described below. Abarrategi et al. studied the influence of molecular weight, deacetylation degree, and calcium content on the properties of chitosan for osteochondral tissue regeneration [bib_ref] Chitosan scaffolds for osteochondral tissue regeneration, Abarrategi [/bib_ref]. Chitosan-based porous scaffolds with different values of these parameters were manufactured and implanted in rabbits with osteochondral knee defects. Chitosan scaffolds with lower molecular weight, a lower degree of deacetylation, and intact mineral content turned out to be the most suitable among those evaluated for the regeneration of cartilage and bone in these defects. Akmeşe et al. compared a chitosan-based liquid scaffold with a hyaluronic-acid-based soft scaffold for the repair of osteochondral defects of the talus [bib_ref] Comparison of Chitosan-Based Liquid Scaffold and Hyaluronic Acid-Based Soft Scaffold for Treatment..., Akmeşe [/bib_ref]. Although both cell-free scaffolds proved to be effective in treating these lesions in terms of cartilage repair, no differences could be established between them. Meanwhile, Hoemann et al. demonstrated the utility of chitosan-glycerol phosphate/blood clot implants for cartilage repair in microfracture defects in sheep [bib_ref] Chitosan-Glycerol Phosphate/Blood Implants Improve Hyaline Cartilage Repair in Ovine Microfracture Defects, Hoemann [/bib_ref]. These implants offered greater retention in the defect walls than normal clots at one hour postoperatively, which was attributed to the adhesive and thrombogenic properties of chitosan. Six months after the operation, the implants showed better results in terms of hyaline cartilage regeneration than microfracture alone. In addition, chitosan has also been used for the development of systems that load bioactive molecules, such as growth factors (for example, transforming growth factor-β1 [bib_ref] Synergetic integrations of bone marrow stem cells and transforming growth factor-β1 loaded..., Zheng [/bib_ref] [bib_ref] Enhanced efficacy of transforming growth factor-β1 loaded an injectable cross-linked thiolated chitosan..., Zhang [/bib_ref] and stem cells (such as synovial mesenchymal [bib_ref] Chitosan hydrogel/3D-printed poly(ε-caprolactone) hybrid scaffold containing synovial mesenchymal stem cells for cartilage..., Li [/bib_ref] or adipose-derived stem cells [bib_ref] Repair of an articular cartilage defect using adiposederived stem cells loaded on..., Zhang [/bib_ref] for cartilage regeneration. Regarding the use of chitosan for viscosupplementation, Scognamiglio et al. formulated hydrogels based on lactose-modified chitosan crosslinked with boric acid for the treatment of osteoarthritis [bib_ref] A hydrogel system based on a lactose-modified chitosan for viscosupplementation in osteoarthritis, Scognamiglio [/bib_ref]. One of the different compositions studied showed rheological properties similar to commercial products, so it was chosen for further characterization. This formulation was biocompatible and showed greater resistance to degradation than the hyaluronic-acid-based samples, which translated into greater stability after administration. The selected hydrogel was also associated with the inactivation of ROS, which are involved in osteoarthritis. ## Soft tissue regeneration Soft tissue engineering can be described as the use of biomaterials for the restoration of the biological function of defective or damaged soft tissues, such as skin, mucosa, or cornea, among others [bib_ref] Regulating Preparation of Functional Alginate-Chitosan Three-Dimensional Scaffold for Skin Tissue Engineering, Zhu [/bib_ref]. Wound healing occurs naturally through a number of biological processes, including inflammation and cell proliferation. Soft tissue engineering can improve the efficacy of these processes. In the case of severely damaged tissue that cannot be repaired by organic responses, biomaterials can also replace that tissue [bib_ref] Optimization, characterization, and efficacy evaluation of 2% chitosan scaffold for tissue engineering..., Kumar [/bib_ref]. These biomaterials must be biocompatible, have antibacterial activity to protect the damaged tissue from infection, absorb liquids, and allow gas permeation. Moreover, they must have optimal mechanical strength and should provide a favorable microenvironment for new cells [bib_ref] Regulating Preparation of Functional Alginate-Chitosan Three-Dimensional Scaffold for Skin Tissue Engineering, Zhu [/bib_ref] [bib_ref] Optimization, characterization, and efficacy evaluation of 2% chitosan scaffold for tissue engineering..., Kumar [/bib_ref]. Chitosan has several advantages for soft tissue regeneration. First of all, it meets all the requirements mentioned above for biomaterials intended for soft tissue engineering. It also has other characteristics, including its slow rate of degradation by different enzymes into oligomers and, ultimately, into N-acetylglucosamine, a common amino sugar that induces and enhances the natural healing process due to its anti-inflammatory properties [bib_ref] Chitosan for Gene Delivery and Orthopedic Tissue Engineering Applications, Raftery [/bib_ref]. Moreover, chitosan has antifungal, antimicrobial, immunostimulant, and hemostatic properties, features that are related to the optimal restoration of damaged tissue [bib_ref] Effectiveness of chitosan scaffold in skin, bone and cartilage healing, Oryan [/bib_ref]. Chitosan can be formulated in hydrogels and in 2D and 3D materials, and each of them has different possibilities. To begin with, hydrogels are soft, flexible systems that can be applied to mimic soft tissue without damaging adjacent structures. Two-dimensional materials include thin films or porous membranes, which can be easily used to cover small to large damaged surfaces (such as skin wound healing), protecting them from external agents and enhancing natural healing. Lastly, 3D scaffolds can be manufactured with tunable porosities and enhance regeneration by inducing matrix restoration, allowing the diffusion of nutrients and gases in the regeneration of soft systems [bib_ref] Effectiveness of chitosan scaffold in skin, bone and cartilage healing, Oryan [/bib_ref]. Thanks to the aforementioned properties of chitosan, it has been widely explored for soft tissue regeneration . . Some recent advances in chitosan-based materials for soft tissue regeneration. ## Chitosan form/system purpose main results reference Chatechol-conjugated chitosan patch Oral mucositis Mucoadhesive patches with enhanced healing properties through sustained release of triamcinolone acetonide. [bib_ref] Chitosan oral patches inspired by mussel adhesion, Ryu [/bib_ref] Carboxymethyl chitosan/alginate-plantamajoside hydrogel Burn wound skin Reduces inflammation, increases collagen deposition, promotes cell migration and proliferation, and accelerates skin scald repair. [140] Cotton fabrics coated with carboxymethyl chitosan Damaged skin Antibacterial properties against S. aureus and E. coli and accelerated reepithelization. [bib_ref] In Vivo assessment of the durable, green and in situ bio-functional cotton..., Shaheen [/bib_ref] Poloxamer 407/methylcellulose chitosan thermosensitive gel Cornea damage Good spreading ability, mucoadhesion, and ocular biocompatibility; accelerated corneal healing. [bib_ref] Hybrid thermosensitive-mucoadhesive in situ forming gels for enhanced corneal wound healing effect..., Fathalla [/bib_ref] Chitosan patches Tympanic perforation More effective than spontaneous healing in tympanic regeneration. [bib_ref] Development of water-insoluble chitosan patch scaffold to repair traumatic tympanic membrane perforations, Kim [/bib_ref] [bib_ref] Tympanic Membrane Regeneration Using a Water-Soluble Chitosan Patch, Kim [/bib_ref] Chitosan-coated filaments Surgical suture Significant reduction of biofilm formation. [bib_ref] Chitosan-Coated Surgical Sutures Prevent Adherence and Biofilms of Mixed Microbial Communities, Prabha [/bib_ref] ## Skin and mucosa wound healing Madrazo-Jiménez et al. conducted a clinical trial to evaluate the effectiveness of a chitosan hydrogel containing allantoin, dexpanthenol, and chlorhexidine as a wound healing system [bib_ref] The effects of a topical gel containing chitosan, 0.2% chlorhexidine, allantoin and..., Madrazo-Jimenez [/bib_ref]. The use of this gel significantly improved the healing process in the buccal mucosa after the extraction of third molars. As an example of chitosan-based 2D materials, Zaitun Hasibuan et al. mixed cellulose nanofibers, chitosan, and silver nanoparticles in a wound-healing dressing. Based on the results of the study, the authors were able to conclude that this system is a non-hemolytic material with strong antibacterial properties, making it an interesting candidate for future in vivo studies. Ruprai et al. also developed 2D structures as adhesives. These systems, based on chitosan and loaded with L-DOPA, were obtained by lyophilization and showed a photochemical binding capacity to tissue when treated with green light. Interestingly, this approach overcomes one of the main limitations of glues and patches, as the porous structure of freeze-dried chitosan allows the efficient permeation of gases and substances, thus enhancing the wound healing process [bib_ref] Porous chitosan adhesives with L-DOPA for enhanced photochemical tissue bonding, Ruprai [/bib_ref]. Another approach was proposed by Zhu et al., who developed flurbiprofen-grafted chitosan/alginate composite-based 3D scaffolds for subcutaneous implantation [bib_ref] A hydrogel system based on a lactose-modified chitosan for viscosupplementation in osteoarthritis, Scognamiglio [/bib_ref]. These scaffolds were prepared by the lyophilization of a previously obtained gel and then grafted with crosslinkers and flurbiprofen. The obtained scaffold exhibited a homogeneous porous structure with optimal mechanical properties. In addition, the presence of flurbiprofen was synergized with the structure and characteristics of the scaffold, enhancing cell adhesion and proliferation. ## Cornea damage Corneal damage is the cause of vision loss in 10 million people yearly; 12 million people have already gone blind. Corneal damage leads to the opacity of the cornea due to the formation of scar tissue. Nowadays, synthetic or allogenic implants are the most widely used options for their treatment, but adverse effects such as high rejection rates frequently occur, and the shortage of donors generates long waiting lists. In the search for alternatives, chitosan has emerged as a good option [bib_ref] Polysaccharides, as biological macromolecule-based scaffolding biomaterials in cornea tissue engineering: A review, Dong [/bib_ref] [bib_ref] Exosomes-loaded thermosensitive hydrogels for corneal epithelium and stroma regeneration, Tang [/bib_ref]. Although chitosan has a number of advantageous properties, the films obtained by the solvent casting method have poor mechanical properties. According to this, Tayebi et al. prepared chitosan/poly-ε-caprolactone composite films loaded with chitosan nanoparticles as a vehicle for corneal endothelial cells. The authors found that chitosan nanoparticles improved biocompatibility and surface properties without affecting the transparency of the systems, which is a key requirement for this application. Corneal epithelial cells were also able to adhere to the membrane, survive, and proliferate adequately. This membrane could, therefore, be an option for the future in the treatment of corneal damage. Another innovative approach was described by Feng et al., who developed transparent thermogelling scaffolds using oligoethylene glycol-based dendronized chitosan [bib_ref] Thermo-Gelling Dendronized Chitosans as Biomimetic Scaffolds for Corneal Tissue Engineering, Feng [/bib_ref]. Interestingly, the gel point of these systems can be easily adjusted. These transparent hydrogels were shown to be capable of promoting the migration and proliferation of keratocytes, with a positive effect on corneal regeneration in a rabbit animal model [fig_ref] Figure 6: Schematic illustration for the formation of dendronized chitosan [/fig_ref]. ## Gastric ulcer Gastric ulcer can be described as a pathological condition that occurs when the gastric epithelium is damaged by excessive inflammatory erosion, usually leading to epigastric pain and gastric mucosa bleeding. Different factors can increase the risk of gastric ulcers, such as the excessive consumption of alcohol or non-steroidal anti-inflammatory drugs, as well as bacterial infections such as Helicobacter pylori [bib_ref] Protective Effects of Chitosan-Bilirubin Nanoparticles Against Ethanol-Induced Gastric Ulcers, Huang [/bib_ref] [bib_ref] Chitosan-albumin based core shell-corona nano-antimicrobials to eradicate resistant gastric pathogen, Niaz [/bib_ref]. In order to improve the actual treatments for gastric ulcers, Maeng et al. designed an endoscopic chitosan hydrogel loaded with epithelial growth factor, given that it binds to a receptor that initiates a series of responses that accelerate tissue regeneration [bib_ref] Endoscopic application of EGF-chitosan hydrogel for precipitated healing of GI peptic ulcers..., Maeng [/bib_ref]. The authors evaluated this gel using in vitro (wounded cell monolayer) and in vivo (GI ulcers in rabbits and micro-pigs) models of ulcers. Histological analysis allowed them to conclude that this formulation significantly reduces the time required for the complete healing of gastric ulcers. Another innovative approach was described by Feng et al., who developed transparent thermogelling scaffolds using oligoethylene glycol-based dendronized chitosan [bib_ref] Thermo-Gelling Dendronized Chitosans as Biomimetic Scaffolds for Corneal Tissue Engineering, Feng [/bib_ref]. Interestingly, the gel point of these systems can be easily adjusted. These transparent hydrogels were shown to be capable of promoting the migration and proliferation of keratocytes, with a positive effect on corneal regeneration in a rabbit animal model [fig_ref] Figure 6: Schematic illustration for the formation of dendronized chitosan [/fig_ref]. ## Gastric ulcer Gastric ulcer can be described as a pathological condition that occurs when the gastric epithelium is damaged by excessive inflammatory erosion, usually leading to epigastric pain and gastric mucosa bleeding. Different factors can increase the risk of gastric ulcers, such as the excessive consumption of alcohol or non-steroidal anti-inflammatory drugs, as well as bacterial infections such as Helicobacter pylori [bib_ref] Protective Effects of Chitosan-Bilirubin Nanoparticles Against Ethanol-Induced Gastric Ulcers, Huang [/bib_ref] [bib_ref] Chitosan-albumin based core shell-corona nano-antimicrobials to eradicate resistant gastric pathogen, Niaz [/bib_ref]. In order to improve the actual treatments for gastric ulcers, Maeng et al. designed an endoscopic chitosan hydrogel loaded with epithelial growth factor, given that it binds to a receptor that initiates a series of responses that accelerate tissue regeneration [bib_ref] Endoscopic application of EGF-chitosan hydrogel for precipitated healing of GI peptic ulcers..., Maeng [/bib_ref]. The authors evaluated this gel using in vitro (wounded cell monolayer) and in vivo (GI ulcers in rabbits and micro-pigs) models of ulcers. Histological analysis allowed them to conclude that this formulation significantly reduces the time required for the complete healing of gastric ulcers. ## Chronic tympanic membrane perforation Most tympanic perforations heal spontaneously in 7 to 10 days, thanks to epithelial migration, fibroblastic activity, and vascular proliferation. However, cases in which healing does not occur spontaneously within 3 months are called chronic tympanic perforations, which require intervention for recovery [bib_ref] Animal models of chronic tympanic membrane perforation: In response to plasminogen initiates..., Wang [/bib_ref]. Although there are surgeries with high efficacy (>90%), there are some drawbacks, such as the high cost or the risk of anesthesia. For this reason, new biomaterials for non-invasive applications have recently been studied, among which chitosan stands out: it is highly biocompatible and allows cell adhesion and wound healing. In addition, it has good mechanical properties and is antibacterial [bib_ref] Regeneration of Chronic Tympanic Membrane Perforation Using an EGF-Releasing Chitosan Patch, Seonwoo [/bib_ref]. For this reason, chitosan has been shown to be effective in improving the healing of tympanic perforations [bib_ref] Chitosan patch scaffold for repair of chronic safe tympanic membrane perforation, Shehata [/bib_ref]. In this context, Kim et al. prepared 3D chitosan porous scaffolds and evaluated their structural and mechanical properties, as well as biocompatibility and healing effects, in an in vivo model [bib_ref] A Healing Method of Tympanic Membrane Perforations Using Three-Dimensional Porous Chitosan Scaffolds, Kim [/bib_ref]. These scaffolds allowed the ## Chronic tympanic membrane perforation Most tympanic perforations heal spontaneously in 7 to 10 days, thanks to epithelial migration, fibroblastic activity, and vascular proliferation. However, cases in which healing does not occur spontaneously within 3 months are called chronic tympanic perforations, which require intervention for recovery [bib_ref] Animal models of chronic tympanic membrane perforation: In response to plasminogen initiates..., Wang [/bib_ref]. Although there are surgeries with high efficacy (>90%), there are some drawbacks, such as the high cost or the risk of anesthesia. For this reason, new biomaterials for non-invasive applications have recently been studied, among which chitosan stands out: it is highly biocompatible and allows cell adhesion and wound healing. In addition, it has good mechanical properties and is antibacterial [bib_ref] Regeneration of Chronic Tympanic Membrane Perforation Using an EGF-Releasing Chitosan Patch, Seonwoo [/bib_ref]. For this reason, chitosan has been shown to be effective in improving the healing of tympanic perforations [bib_ref] Chitosan patch scaffold for repair of chronic safe tympanic membrane perforation, Shehata [/bib_ref]. In this context, Kim et al. prepared 3D chitosan porous scaffolds and evaluated their structural and mechanical properties, as well as biocompatibility and healing effects, in an in vivo model [bib_ref] A Healing Method of Tympanic Membrane Perforations Using Three-Dimensional Porous Chitosan Scaffolds, Kim [/bib_ref]. These scaffolds allowed the regeneration of the perforated tympanic membrane through cell migration, epidermal connective tissue, and mucosal restoration. However, to improve the properties of chitosan scaffolds, Seonwoo et al. prepared and evaluated chitosan scaffolds loaded with epithelial growth factor [bib_ref] Regeneration of Chronic Tympanic Membrane Perforation Using an EGF-Releasing Chitosan Patch, Seonwoo [/bib_ref]. This factor may enhance healing by aiding the migration of fibroblasts, endothelial cells, and vascular cells. Their results allowed the authors to demonstrate that the loaded scaffold improved the ability of raw chitosan in terms of cell viability and in vitro wound-healing rate. This approach is not as effective as surgical methods, but it is an alternative option for patients with less severe tympanic perforations, or it could even be combined with surgery to improve results. ## Sutures Sutures, medical devices intended to stabilize wounds to accelerate their healing, have traditionally been made from non-bioabsorbable substances, although bioabsorbable materials have gained attention in recent years. These devices can be made up of a monofilament or a set of twisted or braided filaments. There are some requirements for sutures, including good mechanical properties and high biocompatibility [bib_ref] Effect of Squid Cartilage Chitosan Molecular Structure on the Properties of Its..., Tan [/bib_ref]. Among the most frequent problems attributable to the sutures used today, the frequent infection of the tissues adjacent to the sutured wounds has recently attracted the attention of researchers. There are several strategies proposed to reduce the incidence of infections, such as the incorporation of an antibacterial drug (such as triclosan) or the use of antibacterial materials [bib_ref] Development and characterization of antibacterial braided polyamide suture coated with chitosan-citric acid..., Debbabi [/bib_ref]. It is at this point that the use of chitosan is of interest since, in addition to its known antibacterial, hemostatic, and bioabsorption properties, it has also been shown to be capable of improving and accelerating the healing process. Two possible strategies have been proposed: the design of chitosan fibers, alone or in combination with other polymers and drugs, and the coating of fibers with chitosan. As an example of the former, Tan et al. prepared electrospun monofilaments with chitosan of different molecular weights and deacetylation degrees [bib_ref] Effect of Squid Cartilage Chitosan Molecular Structure on the Properties of Its..., Tan [/bib_ref]. These devices were then characterized through various techniques, including surface morphology, mechanical properties, swelling ratio and in vitro degradation, and cytotoxicity, among others. The authors found out that chitosan with a molecular weight of 1.2 × 10 6 g/mol and a degree of deacetylation of 85% allowed them to manufacture fibers with outstanding properties in terms of mechanical strength and biocompatibility. Another approach was proposed by Costa da Silva et al., who loaded N-acetyl-Dglucosamine into chitosan filaments through the wet-spinning method. N-acetyl-Dglucosamine has proven effective in accelerating the recovery of the epithelium. Several characterization techniques were applied, including SEM, evaluation of mechanical properties and in vitro biodegradations, drug release, and cytotoxicity. These fibers rapidly degraded due to the presence of N-acetyl-D-glucosamine, although the mechanical properties were somewhat worse compared to pure chitosan-based filaments. However, its mechanical performance was still better than that required by the U.S. pharmacopeia. Furthermore, drug release was maintained and biocompatibility was verified. Regarding chitosan-coated fibers, Mohammadi et al. prepared chitosan/hyaluronic acid-coated nylon fibers using the layer-by-layer technique [bib_ref] Improving physical and biological properties of nylon monofilament as suture by Chitosan/Hyaluronic..., Mohammadi [/bib_ref]. The two polymers selected for the coating were chosen due to their ability to form polyelectrolyte complexes. The authors found that coating with the polyelectrolyte complex leads to an improved growth rate of Vero cells on the coated filament. This, coupled with the antibacterial activity of the coating, makes this approach a good option for the development of next-generation sutures. ## Hemostasis Hemostasis is the biological process of maintaining blood fluidity in intact blood vessels and causing blood to coagulate when a vessel is damaged to prevent bleeding [bib_ref] Diagnostic strategies in disorders of hemostasis, Hrachovinová [/bib_ref]. This process may be impaired in some diseases, such as hereditary hemophilia, where a coagulation factor deficiency prevents the body from stopping bleeding [bib_ref] Laboratory Monitoring in Emicizumab-Treated Persons with Hemophilia A, Müller [/bib_ref]. In surgical procedures, it is common to make incisions that cause tissue or blood vessels to rupture. Control of bleeding is vital in these situations to ensure patient survival. For this, the procoagulant effect of chitosan has been studied. Although the mechanisms of chitosan to stop bleeding are still uncertain, it was evaluated how the physicochemical properties of chitosan and the changes in its structure modify this characteristic. Thus, it was concluded that chitosan with a lower degree of deacetylation and higher molecular weight improves blood coagulation. Moreover, chitosan derivatives obtained by treating the polymer with acids show an enhanced procoagulant effect [bib_ref] Investigation of the Effects of Molecular Parameters on the Hemostatic Properties of..., Hu [/bib_ref]. Considering this, different biomaterials based on chitosan have been developed to be used in the control of hemorrhage in surgical procedures. An example of this is chitosan hemostatic sponges, which combine the anticoagulant activity of chitosan with the high porosity and fluid absorption capacity of this device. Absorbable chitosan-based sponges of different degrees of deacetylation were fabricated and tested in vivo in a liver hemorrhage model. From the results, sponges based on chitosan with a degree of deacetylation of 40% were the ones that showed the best hemostatic effect. The blood and tissue compatibility of the device was also notable. The biodegradability of the system was remarkable, with visible degradation in the first week after implantation and the complete absence of the sponge after 4 weeks [bib_ref] Using absorbable chitosan hemostatic sponges as a promising surgical dressing, Huang [/bib_ref]. Sponges can also be prepared from a combination of chitosan and other polymers. The combination with gelatin is frequently used as these polymers can combine their advantages. Thus, a chitosan/gelatin sponge was prepared to be used as coagulation-promoting material and compared with sponges based on only one of these polymers. Composites were evaluated in rabbits for the coagulation of ear artery and liver injuries. The coagulation process was accelerated with the use of these sponges since platelet aggregation was promoted, which reveals the potential of these sponges to be used in surgical hemorrhages [bib_ref] Chitosan/gelatin composite sponge is an absorbable surgical hemostatic agent, Lan [/bib_ref]. Some modifications of chitosan have been synthesized, particularly for use as procoagulant materials. Cheng et al. prepared a marine collagen peptide grafted with carboxymethyl chitosan and then manufactured sponges from this material. Both in vitro and in vivo tests revealed the procoagulant effect of these devices, which could be useful for hemorrhage control [bib_ref] Marine collagen peptide grafted carboxymethyl chitosan: Optimization preparation and coagulation evaluation, Cheng [/bib_ref]. Du et al. manufactured sponges based on microchanneled alkylated chitosan. The main advantage of this modification is the increased blood absorption capacity due to the increased porosity of the developed system. When evaluated in vivo in rat and pig liver perforations, the use of this sponge was shown to be superior in hemostatic capacity to other commercial gauzes and to gelatinbased sponges [bib_ref] Microchannelled alkylated chitosan sponge to treat noncompressible hemorrhages and facilitate wound healing, Du [/bib_ref]. Another example of sponges based on chitosan derivatives is the thiol-modified chitosan sponges developed by Wu et al. These authors included silver nanoparticles in the prepared system to increase the antibacterial activity of the device. As a result, the sponge has both hemostatic and antimicrobial performance, accelerating wound healing and avoiding infections [bib_ref] Antibacterial and Hemostatic Thiol-Modified Chitosan-Immobilized AgNPs Composite Sponges, Wu [/bib_ref]. The efficacy of chitosan sponges has also been compared to chitosan fiber dressings. Both systems are highly biocompatible, but while the sponges offer a higher fluid absorption capacity, chitosan fiber dressings show the most rapid hemostatic effect. Moreover, this device also serves as an inhibitor of microbial proliferation [fig_ref] Figure 7: Antimicrobial activity evaluation [/fig_ref]. The advantages of chitosan dressings compared to traditional cotton surgical gauzes have been verified by different authors, the antimicrobial and procoagulant characteristics of chitosan being the ones that give this material promising properties for its use as a first-line dressing in the management of surgical hemorrhages [bib_ref] Procoagulant and Antimicrobial Effects of Chitosan in Wound Healing, Wang [/bib_ref]. Another study evaluated the efficacy of chitosan-based dressings in controlling hemostasis after dental extractions in patients receiving oral antithrombotic therapy. Again, the efficacy of this material was compared to the use of cotton gauze dressings. Not only was the mean time for hemostasis significantly reduced with chitosan-based dressings, but the incidence of adverse events-dry socket and pus discharge-was also notably lower [bib_ref] A comparative evaluation of the effectiveness of chitosan-based dressing and conventional method..., Seethamsetty [/bib_ref]. The combination of fibrin glue and chitosan dressings has also been explored to improve the hemostatic process in patients with hemophilia. The conclusion of the study was that, although the use of these systems is not always necessary to achieve hemostasis, they have a particularly useful synergistic effect in surgical procedures with excessive bleeding [bib_ref] Local fibrin glue and chitosan-based dressings in haemophilia surgery, Rodriguez-Merchan [/bib_ref]. It is not only sponges and dressings that have been developed as chitosan-based systems for hemostasis control. Logun et al. developed a hydrophobically modified chitosan foam, which was tested in rats with non-lethal liver excisions. A chitosan-based foam, a hydrophobically modified chitosan-based foam, and a control using fibrin sealant were evaluated. The expandable foam was injected into the damaged area, and the animals were observed for 6 weeks, after which they were sacrificed and the injury sites visualized microscopically. The developed systems proved to be safe and well-tolerated by animals. Both chitosan and chitosan-modified foams exhibited faster biodegradability after application than fibrin sealant and minimal tissue adhesion [bib_ref] Expanding Hydrophobically Modified Chitosan Foam for Internal Surgical Hemostasis: Safety Evaluation in..., Logun [/bib_ref]. Chitosan-based tampons have also been developed as a strategy to prevent vaginal bleeding after loop electrosurgical excision procedures. The presence of hemorrhages in the 2 weeks following the surgical procedure was evaluated and compared to the use of a general tampon. Bleeding, as well as vaginal discharge and abdominal pain, was significantly reduced in the group of women who used the chitosan tampon, proving the effectiveness of this material in wound healing. Another formulation that is being evaluated for hemostasis is chitosan gels, with several examples of this application in the literature. For example, the use of a chitosan gel as a coagulation promoter after endoscopic sinus surgery has been extensively studied, both in animal and human models [bib_ref] The Efficacy of a Novel Chitosan Gel on Hemostasis after Endoscopic Sinus..., Valentine [/bib_ref] [bib_ref] The Efficacy of a Novel Chitosan Gel on Hemostasis and Wound Healing..., Valentine [/bib_ref] [bib_ref] Effect of a Chitosan Gel on Hemostasis and Prevention of Adhesion After..., Chung [/bib_ref]. This gel not only improves hemostasis but also prevents adhesion between tissues. A modified biodegradable chitosan hydrogel developed by Fang et al. was able to achieve a rapid sol-gel transition after injection and improved adhesion to biological tissues. In addition, the self-contracting property of this hydrogel provides an advantage for wound healing over other chitosan-based hydrogels [bib_ref] A chitosan hydrogel sealant with self-contractile characteristic: From rapid and long-term hemorrhage..., Fang [/bib_ref]. Finally, it is worth mentioning that it is not only chitosan that has been studied to prepare these hemostatic hydrogels but also its combination with other compounds. As an example, a hydroxypropyl chitosan/soy protein isolate hydrogel was developed by Zhao et al. for hemorrhage control. This formulation was able to effectively reduce bleeding in an animal model [bib_ref] Fabrication of Hydroxypropyl Chitosan/Soy Protein Isolate Hydrogel for Effective Hemorrhage Control, Zhao [/bib_ref]. Additionally, chitosan has the ability to open cell tight junctions, which may be useful in improving drug delivery. This feature can also be applied to hemostatic control by including hemostatic drugs in the chitosan-based formulation. For example, the inclusion of tranexamic acid in a mixed chitosan-dextran gel applied after endoscopic sinus surgery was used to control bleeding [bib_ref] Should chitosan and tranexamic acid be combined for improved hemostasis after sinus..., Bartley [/bib_ref]. Mar. Drugs 2022, 20, x FOR PEER REVIEW 18 of 29 systems is not always necessary to achieve hemostasis, they have a particularly useful synergistic effect in surgical procedures with excessive bleeding [bib_ref] Local fibrin glue and chitosan-based dressings in haemophilia surgery, Rodriguez-Merchan [/bib_ref]. It is not only sponges and dressings that have been developed as chitosan-based systems for hemostasis control. Logun et al. developed a hydrophobically modified chitosan foam, which was tested in rats with non-lethal liver excisions. A chitosan-based foam, a hydrophobically modified chitosan-based foam, and a control using fibrin sealant were evaluated. The expandable foam was injected into the damaged area, and the animals were observed for 6 weeks, after which they were sacrificed and the injury sites visualized microscopically. The developed systems proved to be safe and well-tolerated by animals. Both chitosan and chitosan-modified foams exhibited faster biodegradability after application than fibrin sealant and minimal tissue adhesion [bib_ref] Expanding Hydrophobically Modified Chitosan Foam for Internal Surgical Hemostasis: Safety Evaluation in..., Logun [/bib_ref]. Chitosan-based tampons have also been developed as a strategy to prevent vaginal bleeding after loop electrosurgical excision procedures. The presence of hemorrhages in the 2 weeks following the surgical procedure was evaluated and compared to the use of a general tampon. Bleeding, as well as vaginal discharge and abdominal pain, was significantly reduced in the group of women who used the chitosan tampon, proving the effectiveness of this material in wound healing. Another formulation that is being evaluated for hemostasis is chitosan gels, with several examples of this application in the literature. For example, the use of a chitosan gel as a coagulation promoter after endoscopic sinus surgery has been extensively studied, both However, it is not only surgeries that require adequate hemostatic control. When a patient suffers successive punctures in a certain area, as can occur in chronic hemodialysis patients, they may suffer from acquired coagulopathy and excessive clotting time after needle removal. Chitosan-based systems have also been explored for these situations. Thus, Misgav et al. evaluated the potential of chitosan-based pads to shorten the time needed to control bleeding in these patients. The application of the pad on the puncture site proved to be effective in significantly reducing the time required to seal the hemorrhage-it was reduced from 18.5 to 3 min in arterial access and from 13.2 to 2.8 min in a venipuncture site [bib_ref] The hemostatic efficacy of chitosan-pads in hemodialysis patients with significant bleeding tendency, Misgav [/bib_ref]. The InnoSeal hemostatic pad is a commercially available catecholamine-chitosan pad, capable of significantly reducing hemostasis time after cardiac catheterization compared to other available pads [bib_ref] Randomized trial comparing radial hemostasis techniques; catechol conjugated chitosan pad (InnoSEAL) versus..., Pathan [/bib_ref]. Similarly, the Clo-Sur Plus Radial TM pad is another chitosan-based hemostatic pad evaluated for sealing bleeding after transradial arterial access. Although its efficacy was first compared with mechanical compression, it was concluded that the combination of both strategies is the most appropriate resource for bleeding control, with the chitosan pad being responsible for minimizing complications at the access site [bib_ref] Comparison of Hemostasis Times with a Chitosan-Based Hemostatic Pad (Clo-SurPlus Radial™) vs..., Roberts [/bib_ref]. ## Other surgical applications of chitosan Although the use of chitosan for tissue regeneration and homeostasis control has been deeply explored, the innate properties of this polymer have made it a candidate for several alternative applications. As mentioned before, it has a slightly antimicrobial activity that becomes highly useful in surgically implemented devices. Titanium coating with chitosan by covalent bond was performed through a coupling agent (triethoxysilylpropyl succinic anhydride). This chitosan-coated titanium alloy proved to be active in reducing the growth of Staphylococcus aureus and Escherichia coli on the surface of the device, which could be useful in preventing infections after implantations of prostheses or devices made of this material [bib_ref] Chitosan coating as an antibacterial surface for biomedical applications, D&apos;almeida [/bib_ref]. In addition, chitosan can enhance the antimicrobial activity of drugs. With this in mind, nanofibers with a Nylon-6 core and a chitosan/polyethylene oxide shell structure, including two antimicrobial compounds, were fabricated; 5-chloro-8-quinolol was incorporated in the shell and poly(hexanide) in the core. Surgical meshes based on nanofibers were developed from this material with the purpose of use as meshes for hernias. This material showed greater in vitro activity against Staphylococcus aureus and Pseudomonas aeruginosa compared to the use of free drugs [bib_ref] Nylon-6/chitosan core/shell antimicrobial nanofibers for the prevention of mesh-associated surgical site infection, Keirouz [/bib_ref]. Similarly, implantable chitosan sponges loaded with cefuroxime, ciprofloxacin, and vancomycin were evaluated in terms of their antibacterial potential. It was observed that the high aqueous solubility of vancomycin made the system highly hydrophilic; thus, the sponge underwent rapid degradation, rendering it useless as an implantable device. However, the sponges loaded with cefuroxime and ciprofloxacin achieved a sustained release of the drug, with high tissue concentrations of the active ingredient and lower plasma levels of antibiotics [bib_ref] Chitosan sponges as a sustained release carrier system for the prophylaxis of..., Pawar [/bib_ref]. This did not rule out the inclusion of vancomycin in chitosan-based biomaterials. Foster et al. managed to develop a vancomycin-loaded chitosan film, intended to prevent infections at surgical sites. The combination with chitosan was shown to improve the antimicrobial efficacy of the antibiotic. Furthermore, the adhesive properties of chitosan were enhanced through laser irradiation of the film after implementation. This adhesiveness would be highly useful in facilitating wound healing. In fact, the use of chitosan films as a surgical adhesive was deeply studied by this research group. The authors demonstrated the efficacy of this system as an adhesive for ovine intestine regeneration [bib_ref] Chitosan adhesive for laser tissue repair: In Vitro characterization, Lauto [/bib_ref]. Subsequently, the laser-activated chitosan thin film developed by these researchers (SurgiLux), with positive results tested both in vitro and in vivo on several tissues, was approved by the FDA for use as a post-surgical adhesive [bib_ref] A Chitosan Based, Laser Activated Thin Film Surgical Adhesive, Foster [/bib_ref]. Different strategies to improve the adherence of chitosan biomaterials to biological tissues have also been evaluated. For example, the use of microbial transglutaminase was able to accelerate the binding of chitosan to different biological tissues (cardiac, dermal, and hepatic) and to polydimethylsiloxane-based devices [bib_ref] Direct Bonding of Chitosan Biomaterials to Tissues Using Transglutaminase for Surgical Repair..., Fernandez [/bib_ref]. Similarly, the mixture of chitosan with oxidized dextran was used to prepare a biocompatible and biodegradable injectable adhesive, with an adhesiveness of 4-5 times higher than that of fibrin glue [bib_ref] A novel injectable tissue adhesive based on oxidized dextran and chitosan, Balakrishnan [/bib_ref]. Finally, it is noteworthy that despite the proven adhesive properties of chitosan, the chemical modification of this polymer has also been studied to reduce its adhesiveness. This is also useful in post-surgical applications as these chitosan-based antiadhesive materials could prevent peritoneal adhesion after abdominal surgery [bib_ref] The Processing of Chitosan and Its Derivatives and Their Application for Postoperative..., Zhu [/bib_ref]. Aussel et al., who developed chitosan hydrogels for small-diameter vascular grafts, studied another application of chitosan materials. The mechanical properties of the fabricated device proved to be strong enough to withstand the intended application [bib_ref] In Vitro Mechanical Property Evaluation of Chitosan-Based Hydrogels Intended for Vascular Graft..., Aussel [/bib_ref]. In addition, the implantation of two chitosan tubes as carotid grafts in sheep demonstrated that this polymer could be sutured without breaking, maintaining arterial pressure without flow obstruction [bib_ref] Chitosan-based hydrogels for developing a small-diameter vascular graft: In Vitro and in..., Aussel [/bib_ref]. have evaluated the use of hydroxyapatite-chitosan patches for the obliteration of the mastoid cavity. It was assessed in the tympanic cavity of rats and was shown to be superior to homologous cartilage and bone cement in mastoid obliteration [bib_ref] In Vivo Study of Mastoid Obliteration Using Hydroxyapatite-Chitosan Patch, Kang [/bib_ref]. Another curious application of chitosan is the use of chitosan/carbon nanotubes in hemoperfusion, as studied by Zong et al. Nanocomposite beads based on these nanotubes were fabricated, and their bilirubin adsorption capacity was evaluated. These chitosan-based materials showed their potential to be used in blood purification [bib_ref] Preparation of chitosan/amino multiwalled carbon nanotubes nanocomposite beads for bilirubin adsorption in..., Zong [/bib_ref]. Finally, it is worth mentioning that the use of chitosan has also reached diagnostic applications. Ghosh et al. developed a derivative of chitosan through the crosslinking of iodinated 2,5-dimethoxy-2,5-dihydrofuran, thus becoming a radiopaque polymer used to manufacture microspheres with in vivo contrast properties, suitable for use in clinical diagnostics of the gastrointestinal tract [bib_ref] Chitosan Derivatives Cross-Linked with Iodinated 2,5-Dimethoxy-2,5-dihydrofuran for Non-Invasive Imaging, Ghosh [/bib_ref]. ## Concluding remarks Chitosan is one of the most studied biopolymers for surgical applications due to its biodegradability, biocompatibility, and absence of toxicity. Its characteristic pH-dependent solubility provides the molecule with positive charges due to the protonation of amine in acidic media, which favors the formation of polyelectrolyte complexes and allows its union with mucosal tissues. The versatility of this polymer is reflected in the great variety of formulations and devices that have been manufactured, highlighting, in particular, scaffolds, sponges, membranes, and hydrogels. The developed materials take advantage of the natural bacteriostatic, fungistatic, hemostatic, and analgesic properties of chitosan. The use of chitosan-based systems-mainly scaffolds-has been deeply explored for tissue regeneration. The clearest example is bone regeneration, where chitosan scaffolds provide biodegradable support for cell growth. Similarly, nerve tissue, cartilage, and various soft tissues can also be regenerated using chitosan-based systems. This application of chitosan would be really promising as it could allow, or speed up, the regeneration of tissues that cannot heal themselves. Another interesting application of chitosan in the manufacture of surgical material is chitosan-based dressings for hemostasis control. Several studies have shown chitosan's improved ability to control bleeding and prevent infection compared to traditional cotton surgical gauze. Therefore, it is expected that, in the future, chitosan may be the material of choice for the development of these gauzes. The antimicrobial properties of chitosan have also been used to manufacture surgical meshes or sutures that could effectively prevent post-surgical infections, which, nowadays, are a frequent complication that increases treatment costs and mortality. Based on the multiple applications that derive from the innate properties of this polymer, it is expected that chitosan will soon be the material of choice for surgical applications. [fig] Figure 2: Figure 2. Schematic representation of the commonly used fabrication methods for producing chitosanbased scaffolds-freeze-drying, freeze gelation, salt leaching, electrospinning, and 3D printing. Reprinted with permission from Saravanan et al. [36]. Copyright 2016, Elsevier. [/fig] [fig] Figure 3: Macroscopic and microscopic images of scaffolds fabricated by different methods in bone tissue engineering. (A-D) show the macroscopic images of scaffolds prepared by freeze-drying, electrospinning, the sol-gel method, and 3D-bioprinting, respectively, and (E-H) show the corresponding SEM images of the scaffolds. Reprinted with permission from Soundarya et al. [42]. Copyright 2018, Elsevier. [/fig] [fig] Figure 4: Schematic representation of the fabrication process of chitosan sponge and the interaction of its molecules. Reprinted with permission from Fan et al. [/fig] [fig] Figure 5: Stereomicroscopy images of bilayer membranes (1×, 2×) and SEM images of chitosan/PEO nanofiber coated porous layer surface (a-c) with 250×, 1000× and 2500× magnifications; cross-sectional view of bilayer structure (d-f) with 250×, 500× and 10,000× magnifications. Reprinted with permission from Tamburaci et al.[114]. Copyright 2021, Elsevier. [/fig] [fig] Figure 6: Schematic illustration for the formation of dendronized chitosan (DC) hydrogels and the application for corneal stromal defects. (A) Cartoon presentation of DCs featured with radial amphiphilicity and their self-assembly in water to form fibrous bundles and the instant formation of hydrogels via heating around physiological temperature. (B) Injection of DC solution into the corneal stromal defects of a rabbit model, in situ formation of a hydrogel filler, and defects repairing. Reprinted with permission from Feng et al.[152]. Copyright 2021, American Chemical Society. [/fig] [fig] Figure 7: Antimicrobial activity evaluation. (A) In vitro. (B) Adenosine triphosphate (ATP) assay of microbial proliferation in patients with surgical wounds, * p < 0.05. Reprinted from [167] under the terms of the Creative Commons Attribution License 4.0 (copyright 2019 Wang et al.; doi:10.3390/polym11111906). [/fig] [fig] Author: Contributions: Conceptualization, F.N.-P., A.M.-I., R.C.-L. and M.D.V.; writing-original draft preparation, F.N.-P., A.M.-I. and R.C.-L.; writing-review and editing, R.R.-C. and M.D.V. All authors have read and agreed to the published version of the manuscript. [/fig]

Relationship of increased aurora kinase A gene copy number, prognosis and response to chemotherapy in patients with metastatic colorectal cancer BACKGROUND: Increased Aurora kinase A gene copy number (AURKA-CN) has been reported in metastatic colorectal cancer (mCRC), with unknown relationship to clinical outcome. We correlated increased AURKA-CN in mCRC tumours with KRAS mutation status, overall and progression-free survival (OS, PFS). METHODS: Sixty-one mCRC tumours were analysed for AURKA-CN using q-PCR, and KRAS mutation status by direct sequencing. Expression of AURKA protein was analysed by immunohistochemistry. Cox-proportional hazard method, Kaplan -Meier curves and log-rank statistics were used to estimate and compare the hazard ratios and median survival between the groups. RESULTS: In all, 68% of tumour exhibited high AURKA-CN, and 29% had a KRAS mutation, without correlation between the two. Patients with high AURKA-CN tumours had longer median OS (48.6 vs 18.8 months, P ¼ 0.01), with stronger trend among KRAS wild-type tumours (median OS not reached vs 18.8 months, P ¼ 0.003). Progression-free survival was longer on first-line or secondline chemotherapy among patients with KRAS wild-type and high vs low AURKA-CN (first: 17.6 vs 5.13 months, P ¼ 0.04; second: 10.4 vs 5.1 months, P ¼ 0.01). AURKA-CN level did not affect outcomes among patients with KRAS mutant tumours. CONCLUSION: Increased AURKA-CN is common in mCRC tumours and is associated with longer OS and longer PFS during chemotherapy, particularly in KRAS wild-type tumours. Aurora kinases (Aurora A, B and C) are important regulatory proteins of the mitotic process [bib_ref] The centrosome-associated Aurora/Ipllike kinase family, Goepfert [/bib_ref] [bib_ref] Functional significance of Aurora kinase A in centrosome amplification and genomic instability, Sen [/bib_ref]. Due to their crucial function in cell division, these proteins have been extensively studied in many cancers for their role in carcinogenesis and as potential treatment targets [bib_ref] Aurora B expression correlates with aggressive behaviour in glioblastoma multiforme, Zeng [/bib_ref] [bib_ref] Expression of Aurora A (but not Aurora B) is predictive of survival..., Nadler [/bib_ref] [bib_ref] Quantitation of Aurora kinase A gene copy number in urine sediments and..., Park [/bib_ref] [bib_ref] Effects of DNAzymes targeting Aurora kinase A on the growth of human..., Qu [/bib_ref] [bib_ref] Aurora kinase B is a predictive factor for the aggressive recurrence of..., Tanaka [/bib_ref] [bib_ref] Inhibition of Aurora-A suppresses epithelial-mesenchymal transition and invasion by downregulating MAPK in..., Wan [/bib_ref]. Aurora kinase A (AURKA) (also known as Aurora-2, BTAK/STK15) regulates mitotic entry, centrosome maturation, bipolar spindle assembly, chromosome alignment, cytokinesis and exit from mitosis [bib_ref] Drosophila Aurora-A is required for centrosome maturation and actin-dependent asymmetric protein localization..., Berdnik [/bib_ref] [bib_ref] Aurora-A and an interacting activator, the LIM protein Ajuba, are required for..., Hirota [/bib_ref] [bib_ref] Aurora-A kinase maintains the fidelity of early and late mitotic events in..., Marumoto [/bib_ref] [bib_ref] Phosphorylation of CDC25B by Aurora-A at the centrosome contributes to the G2-M..., Dutertre [/bib_ref]. Aberrations in the function of Aurora kinases can result in abnormal cell division and aneuploidy due to losses or gains of whole chromosomes . Amplification of AURKA has been shown to induce the formation of a multipolar mitotic spindle, which results in abnormal chromosome alignment and cell division [bib_ref] Aurora-A overexpression reveals tetraploidization as a major route to centrosome amplification in..., Meraldi [/bib_ref]. Resultant genomic instability, aneuploidy and hyperploidy can promote tumour development. In colorectal cancer (CRC), AURKA protein overexpression and amplification have been frequently observed. [bib_ref] A homologue of Drosophila aurora kinase is oncogenic and amplified in human..., Bischoff [/bib_ref] were the first to report overexpression of AURKA mRNA in 450% of CRC tumours. later reported overexpression of AURKA protein by immunohistochemistry (IHC) in 19% of CRC samples of various disease stages. In a multivariate analysis, AURKA protein overexpression was associated with chromosomal instability (identified as loss of heterozygosity in 2p, 5q, 17q and 18q) but did not correlate with clinical outcomes. [bib_ref] Aurora kinase expression in colorectal adenocarcinoma: correlations with clinicopathological features, p16 expression,..., Lam [/bib_ref] identified AURKA overexpression by IHC in 48.5% of early-stage CRC samples. This finding was associated with well/moderately differentiated tumours (P ¼ 0.04), tumours of the distal colon (P ¼ 0.01) and non-mucinous histology (P ¼ 0.001). However, no association with clinical outcomes was detected. The group led by [bib_ref] High copy amplification of the Aurora-A gene is associated with chromosomal instability..., Nishida [/bib_ref] was the first to demonstrate an increase in AURKA gene copy number (AURKA-CN) in 29% of a small sample of colorectal tumours. This study demonstrated a strong correlation between increased AURKA-CN and chromosomal instability, and no association between increased AURKA-CN and KRAS mutation status. Recently, the group led by [bib_ref] Copy number increase of aurora kinase A in colorectal cancers: a correlation..., Zhang [/bib_ref] reported increased AURKA-CN in 32% of CRC tumour samples and particularly in higher stage tumours, suggesting that AURKA may have a role in tumour progression. To date, the finding of increased AURKA-CN has not been correlated with clinical outcomes of patients with CRC. In recent years, the KRAS pathway has been proven to have an important predictive role in the treatment of CRC [bib_ref] Cetuximab monotherapy and cetuximab plus irinotecan in irinotecanrefractory metastatic colorectal cancer, Cunningham [/bib_ref] [bib_ref] Open-label phase III trial of panitumumab plus best supportive care compared with..., Van Cutsem [/bib_ref] [bib_ref] ) K-ras mutations and benefit from cetuximab in advanced colorectal cancer, Karapetis [/bib_ref] [bib_ref] Randomized, phase III trial of panitumumab with infusional fluorouracil, leucovorin, and oxaliplatin..., Douillard [/bib_ref] [bib_ref] Efficacy according to biomarker status of cetuximab plus FOLFOX-4 as first-line treatment..., Bokemeyer [/bib_ref]. Anti-EGFR therapy is effective and approved for the management of metastatic CRC (mCRC) without KRAS mutations. Earlier work by our group has demonstrated synergistic lethality achieved by inhibition of both the Aurora and the KRAS pathways on CRC tissue samples . The goal in this study was to assess the frequency of increased AURKA-CN in archival tumour tissues of patients with metastatic CRC (mCRC) and correlate this finding with progression-free survival (PFS), overall survival (OS), and KRAS mutation status. # Patients and methods ## Patient samples We analysed 61 consecutive mCRC tumour samples submitted between 2006 and 2009 to the Fox Chase Cancer Center (FCCC) molecular genetic facility for evaluation of KRAS mutation status, with remaining tissue available for AURKA-CN evaluation. Most of the patients had their tumour samples submitted by their treating oncologist for KRAS mutation status evaluation after the test became routinely used in clinical practice . Twenty patients in this cohort were enrolled on an institutional phase II study evaluating the combination of capecitabine, oxaliplatin, cetuximab and bevacizumab in the front line metastatic setting . These patients submitted tissue samples for future research at time of study enrollment, which was used for this analysis. Overall survival data were available for 59 out of 61 patients. Fifty-three patients had received chemotherapy treatment at FCCC and were evaluable for PFS. Progression-free survival was defined as the time that elapsed from treatment initiation until evidence of progressive disease or death. For the purpose of this analysis, deintensification of therapy for adverse events, or planned treatment breaks were considered as same line of therapy. ## Evaluation of aurka-cn We utilised a quantitative genomic PCR method described by [bib_ref] High copy amplification of the Aurora-A gene is associated with chromosomal instability..., Nishida [/bib_ref] to evaluate the AURKA gene copy number. In all, 20 ng of genomic DNA purified from paraffin-embedded tissue (PET) sections using the QIAamp DNA Mini Kit (QIAGEN, Valencia, CA, USA) was used. A real-time PCR method was used to determine copy number alterations in the AURKA gene at the Fox Chase Genomics and Genetics Facility using an ABI Prism 7900 sequence detection system (Applied Biosystems, Foster City, CA, USA). The primer and probe sequences for genomic real-time PCR for each of the genes were as follows: the Aurora A forward primer: 5 0 -TCTTTTATAGAAATGTGTGGAAGTTCCT-3 0 ; reverse primer: 5 0 -CAATAAAAAAGTACAGACGCATAAACCA-3 0 ; probe: 5 0 -CTGT CCTTAGAAATAACCACTAC-3 0 . All probes were labelled with FAM as the reporter dye and TAMRA as the quencher. Each PCR amplification reaction was performed in triplicate to ensure accurate results. The RNAse-P used was the endogenous reference gene, for which a set primer probe was purchased from Applied Biosystems. Genomic DNA isolated from DLD1 or HCT116 cell lines served as negative controls for AURKA gene amplification. DNA from the Caco-2 cell line provided a positive control as the AURKA gene locus in this cell line has been reported as four-fold amplified [bib_ref] High copy amplification of the Aurora-A gene is associated with chromosomal instability..., Nishida [/bib_ref]. ## Kras mutation status evaluation Tumour samples were tested for the presence of a KRAS mutation on codons 12 and 13. DNA analyses were performed within the Fox Chase Clinical Molecular Genetics Laboratory. Extraction, isolation and purification of DNA were performed from formalin-fixed PET suitable for molecular analysis using the WaxFree DNA extraction kit (WF-100; TrimGen, Sparks, MD, USA). Ten to fifteen fresh cut unstained slides were used for analysis. DNA (B10 ng) was amplified by PCR using primer sequences located on either side of the region of the coding exon of interest. PCR products were detected by agarose gel electrophoresis. Mutations were detected by sequencing of the purified PCR amplified product (BigDye Terminator v.1.1 Cycle Sequencing Kit; Applied Biosystems) and evaluated by capillary electrophoresis (ABI 3100; Applied Biosystems). ## Ihc evaluation of aurka phosphorylation Activation of AURKA occurs through autophosphorylation of a threonine-288 residue. We evaluated the hyperactive state of the AURKA by IHC with antibodies specific to the autophosphorylated threonine-288 residue (Aurora A phospho-T288, Abcam (Cambridge, MA, USA), rabbit primer, Cat. AB58494) [bib_ref] The mitotic serine/threonine kinase Aurora2/AIK is regulated by phosphorylation and degradation, Walter [/bib_ref]. Immunohistochemical staining was performed on 5 mm sections. After deparaffinisation and rehydration, sections were subjected to heat-induced epitope retrieval by immersion in a 0.01 M citrate buffer (pH 6.0). Endogenous peroxidase activity was blocked for 15 min in 3% hydrogen peroxide in methanol. Nonspecific binding was blocked by treatment with a blocking reagent (Protein Block Serum-Free; Dako, Carpinteria, CA, USA) for 30 min at room temperature. Immunodetection was performed with the Dako Envision þ system. The antigen -antibody immunoreaction was visualised using 3-3 0 -diaminobenzidine as the chromogen. The slides were washed, counterstained with haematoxylin, dehydrated with alcohol, cleared in xylene and mounted. Patient samples that were shown previously to express high levels of phospho-T288 Aurora A (pAurora A) were used as positive controls. The negative control was performed by replacing the primary antibody with negative control rabbit IgG. Evaluation of the staining intensity and positive cell numbers of pAurora A was performed by a pathologist (KC). Tumours with an intermediate or strong nuclear pAurora A immunoreaction in 410% of the tumour cells were scored as pAurora positive. Lack of staining or staining o10% of tumour cells was scored as negative [bib_ref] Aurora kinase expression in colorectal adenocarcinoma: correlations with clinicopathological features, p16 expression,..., Lam [/bib_ref]. ## Clinical data collection A clinicopathological database was created and included demographic information, stage at presentation, treatment regimens, response to therapies and survival. Clinical data were obtained through medical record review using the electronic medical record at FCCC. All patient data were coded and all identifiers were removed before analysis. The study was approved by FCCC's Institutional Review Board. # Statistical analysis Sixty-one tissue samples of mCRC patients at FCCC were included in this analysis. The distribution of AURKA-CN was examined. Full survival information was available for 59 patients. The distribution of AURKA-CN was examined. There are outliers with very high values. To avoid the large influence from the outliers while studying the association between AURKA-CN and survival, we delineated the values by quartile. Kaplan -Meier curves were constructed and log-rank test analysis showed a significant association (results not shown). After confirming this association, we conducted an analysis to search for the most sensitive cutoff for the definition of 'high' AURKA-CN based on our samples. From the Kaplan -Meier curves stratified by quartiles, we found that only the group with AURKA-CN in the first quartile showed a large separation from the remainder of the cohort. The distribution of AURKA-CN was then summarised based on the percentile of the values and used to create data sets with various cutoff points. We examined the data sets with cutoffs from the 25th percentile to 50th percentile. For each data set, Kaplan -Meier curves were created and compared using log-rank statistics and Cox-proportional hazard model. Considering our outcome was longitudinal and traditional area under the ROC curve analyses do not taken into account the varied follow-up time, we utilised the C index [bib_ref] Evaluating the yield of medical tests, Harrell [/bib_ref] [bib_ref] Multivariable prognostic models: issues in developing models, evaluating assumptions and adequacy, and..., Harrell [/bib_ref] which is a member of the 'Kendall family' of rank parameters and is constantly used to estimate the concordance probability with censored data. Like AUC, a value of 0.5 implies complete discordance. Higher values suggest higher concordance between the data and the predicted values from the model. This statistical analysis enabled us to determine the best cutoff points by comprehensively considering hazard ratios (HRs), P-values, separation between Kaplan-Meier curves and C index. After we determined the best cutoff points, we compared the baseline demographics and treatment using a t-test or w 2 -test depending on whether the variable under consideration was continuous or categorical. As the groups were well balanced, no further adjustment for confounding factors was performed. [fig_ref] Table 1: Patient and treatment characteristics Abbreviations [/fig_ref] summarises the characteristics and treatment of patients whose tumours were included in this cohort. The majority of patients in this analysis had initially presented with metastatic disease and were treated with FOLFOX (infusional and bolus 5-fluoruracil, leucovorin and oxaliplatin) plus bevacizumab. All characteristics were well balanced between high and low AURKA-CN cohorts [fig_ref] Table 1: Patient and treatment characteristics Abbreviations [/fig_ref]. The use of various chemotherapeutic agents was also well balanced between the groups with the exception of bevacizumab, which was used more commonly in the low AURKA-CN group in the second-line setting. # Results ## Patient and treatment characteristics ## Determination and frequency of high aurka-cn and kras Of the 61 tissue samples obtained for AURKA-CN analysis, 62% originated from primary tumours, while 38% originated from metastatic sites. This distribution was similar between the high AURKA-CN group and the low AURKA-CN group (P ¼ 0.925). In the analysis of AURKA-CN, we viewed any value 42 as representing abnormal expression of the AURKA gene. We thus created multiple data sets using values selected from percentile tables. These data sets were each analysed using various cutoff points to determine the one most predictive of clinical outcomes. An OS analysis demonstrated statistically significant differences in OS using multiple cutoff values o3.0 (Table 2; . The percentile analysis demonstrated a cutoff of 2.6 (defining approximately the 30th percentile; [fig_ref] Table 2: Survival analysis utilising various AURKA-CN cutoff values for the entire cohort [/fig_ref] as the most sensitive and this was selected as the cutoff value used to define high vs low AURKA-CN for further analyses. By these criteria, 42 of 61 (68.85%) samples expressed increased AURKA-CN. AURKA-CN values ranged from 0.122 to 23.73 (mean 4.875, s.d. 3.96). A KRAS mutation was identified in 18 of 61 samples (29.5%). The most common mutation noted was in codon 12 (16 samples, 89%) with two patient tumours having a mutation in codon 13. No correlation was noted between KRAS mutation status and the presence of high or low AURKA-CN (P ¼ 0.81; [fig_ref] Figure 2: Frequency of KRAS mutations by AURKA-CN [/fig_ref]. ## Aurka-cn, kras and pfs Complete treatment information was available for 53 out of 61 patients, while eight patients were lost to follow-up or elected to receive treatment at an outside institution. The median PFS on first-line chemotherapy was 11.4 months for the full cohort. For patients with high AURKA-CN tumours, the median PFS on firstline chemotherapy was 11.5 months vs 7.7 months for patients with low AURKA-CN tumours (HR ¼ 0.56, 95% CI: 0.28 -1.1, P ¼ 0.094; . Among patients with KRAS wild-type tumours, those with high AURKA-CN tumours had prolonged PFS compared with patients with low AURKA-CN tumours (HR ¼ 0.43, 95% CI: 0.19 -0.94, P ¼ 0.04; . Conversely, PFS did not differ by AURKA-CN expression among patients with KRAS mutant tumours (HR ¼ 1.06, 95% CI: 0.28 -3.93, P ¼ 0.93), although the Thirty-three patients in our cohort who received second-line chemotherapy were evaluable for PFS. The median PFS on secondline chemotherapy was 9.1 months for the full cohort. No statistically significant difference in PFS was noted among patients with high vs low AURKA-CN tumours in the group overall (10.4 months vs 7.7 months; HR ¼ 0.54, 95% CI: 0.24 -1.19, P ¼ 0.13; . However, as with response to first-line treatment, patients with high AURKA-CN and KRAS wild-type status had an improved PFS compared with patients with low AURKA-CN and KRAS wild-type status (10.4 months vs 5.1 months; HR ¼ 0.28, 95% CI: 0.11 -0.74, P ¼ 0.01; . We next performed an exploratory analysis to determine if the relationship of AURKA-CN to PFS was more pronounced in patients receiving the anti-EGFR antibody cetuximab. [fig_ref] Table 3: Progression-free survival on first-line chemotherapy by KRAS mutation status and receipt of... [/fig_ref] outlines median PFS and HRs for high and low AURKA-CN tumours by KRAS mutational status and the use of cetuximab. Interestingly, among patients with KRAS wild-type tumours, high AURKA-CN appeared to have the largest association with outcome among patients who did not receive cetuximab. However, small numbers preclude any definitive conclusion. ## Aurka-cn, kras and os The median follow-up of patients in this cohort was 742 days (range 173 -3229 days). When using quartiles, one quartile increase of AURKA-CN reduced the hazard by 0.63 (95% CI: 0.40 -0.97, P ¼ 0.04). The C index is the highest at a cutoff of 2.813. However, the P-values for Cox model and log-rank tests were barely significant. The C index for cutoff at 2.6 is very similar while the HR is much lower. Thus, we decided to use a cutoff of 2.6 for the rest of the analysis. A significantly longer OS was noted among patients with high vs low AURKA-CN (median OS 48.6 months for patients with high AURKA-CN tumours compared with 18.8 months for patients with low AURKA-CN tumours, HR ¼ 0.28, 95% CI: 0.10 -0.73, P ¼ 0.01; . In all, 1-year and 2-year survival were also longer among patients with high AURKA-CN tumours compared with those with low AURKA-CN tumours (1 year: 92.5% vs 82.2% Po0.001; 2 year: 80.9% vs 29.9%, Po0.001). Similarly to PFS, the longer OS for patients with high AURKA-CN tumours was particularly pronounced in the KRAS wild-type . Among patients with KRAS wildtype tumours, the median OS was not reached for high AURKA-CN compared with 18.8 months for the group with low AURKA-CN tumours (HR ¼ 0.14, 95% CI: 0.038 -0.514, P ¼ 0.003). Similarly, increased 1-year and 2-year survival was seen among those patients with KRAS wild-type tumours and high AURKA-CN compared with patients with KRAS wild-type tumours and low AURKA-CN (1 year: 96% vs 75% Po0.0001; 2 year: 93% vs 28% Po0.0001). No difference in survival was noted by AURKA-CN status in the KRAS mutated population , HR ¼ 0.75, 95% CI: 0.18 -2.98, P ¼ 0.68). ## Ihc evaluation of aurka hyperactivity Immunohistochemical evaluation of phosphorylated AURKA protein was performed on 27 available tumour samples (20 with high AURKA-CN and 7 with low AURKA-CN). Among the seven tumours with low AURKA-CN, none were positive for phosphorylated T-288. However, among the 20 tumours with high AURKA-CN, 9 were found to have positive staining (P ¼ 0.06 for difference between high and low AURKA-CN tumours). A representative image of positive and negative phosphorylated AURKA staining is provided in . # Discussion Although five classes of systemic agents are available for the treatment of advanced CRC, the majority of patients are not cured and median survival has plateaued near 2 years [bib_ref] Bevacizumab in combination with oxaliplatin-based chemotherapy as first-line therapy in metastatic colorectal..., Saltz [/bib_ref] [bib_ref] A randomized phase IIIB trial of chemotherapy, bevacizumab, and panitumumab compared with..., Hecht [/bib_ref] [bib_ref] Chemotherapy, bevacizumab, and cetuximab in metastatic colorectal cancer, Tol [/bib_ref]. Thus, the search for novel targets such as AURKA is of critical importance. Better understanding of the role of the Aurora kinases and their activity in CRC would support the development of new treatment modalities targeting this pathway. Previous groups have studied the frequency of increased AURKA-CN in CRC samples using various cutoff values for defining high AURKA-CN. We performed a novel statistical analysis incorporating the HRs, P-values and Kaplan -Meier survival curves to determine the most sensitive cutoff value. Immunohistochemistry staining for phosphorylated AURKA in metastatic colorectal cancer samples. Activated state of the Aurora A kinase was evaluated using specific antibodies to the phosphorylated threonine-288 residue (Aurora A phospho-T288). Positive staining (410% nuclear staining) is displayed on the left and negative staining is displayed on the right. Our analysis demonstrated the most sensitive cutoff value to be 2.6, which resulted in identification of 68% of mCRC samples as having high AURKA-CN. The groups led by [bib_ref] High copy amplification of the Aurora-A gene is associated with chromosomal instability..., Nishida [/bib_ref] and [bib_ref] Copy number increase of aurora kinase A in colorectal cancers: a correlation..., Zhang [/bib_ref] reported increased AURKA-CN in B30% of CRC tumours. The differences in frequency may simply reflect alternate methodology and cutoffs. The two aforementioned studies were conducted in an Asian patient population with all stages of disease. Differences in patient population and stage of disease studied may thus also account for differences in AURKA-CN frequency. The role of Aurora kinase in tumourigenesis is well documented, although its utility as a prognostic marker is still under investigation in many cancers. [bib_ref] Expression of Aurora A (but not Aurora B) is predictive of survival..., Nadler [/bib_ref] reported a high expression of Aurora A by AQUA to be a poor prognostic marker in patients with breast cancer. Similarly overexpression of Aurora kinase B was found to be a poor prognostic marker in non-small cell lung cancer [bib_ref] Overexpression of aurora B kinase (AURKB) in primary non-small cell lung carcinoma..., Smith [/bib_ref]. In contrast, our report in mCRC is the first demonstrating an association between high AURKA-CN and longer OS. A few groups have previously been unable to demonstrate a correlation between AURKA protein expression by IHC and clinical outcome. Similarly to our findings, Lam and colleagues have demonstrated higher detection of the AURKA protein by IHC in low-grade (well or moderately differentiated) CRC samples compared with high-grade (poorly differentiated) tumours. Therefore, high AURKA-CN may be a marker of a less aggressive biology with improved clinical outcome. Alternatively, increased AURKA-CN may predict for improved benefit from chemotherapy by rendering the cells more susceptible, due to abnormal cell divisions. Our data demonstrating increased PFS for patients with high AURKA-CN tumours support this hypothesis. However, as our sample population did not include untreated patients, this hypothesis requires further evaluation. The improved OS for patients with high AURKA-CN was most pronounced in our study among patients with KRAS wild-type tumours. Furthermore, the improvement in PFS on first-line and second-line chemotherapy was most pronounced in this specific patient population. The small sample size of this subset analysis, while demonstrating statistical significance, limits definitive conclusion and requires further validation in an independent data set. We did not observe any correlation between the presence of KRAS mutations and AURKA-CN, suggesting these may represent two independent pathways in the biology of CRC. By combining these two biomarkers, we may be able to identify a subgroup of patients with mCRC who exhibit increased response to therapy and superior outcomes. The improved response of patients with KRAS wild-type tumours treated with anti-EGFR therapy has been well documented in the literature [bib_ref] Cetuximab monotherapy and cetuximab plus irinotecan in irinotecanrefractory metastatic colorectal cancer, Cunningham [/bib_ref] [bib_ref] ) K-ras mutations and benefit from cetuximab in advanced colorectal cancer, Karapetis [/bib_ref] [bib_ref] Cetuximab and chemotherapy as initial treatment for metastatic colorectal cancer, Van Cutsem [/bib_ref] [bib_ref] Randomized, phase III trial of panitumumab with infusional fluorouracil, leucovorin, and oxaliplatin..., Douillard [/bib_ref] [bib_ref] Randomized phase III study of panitumumab with fluorouracil, leucovorin, and irinotecan (FOLFIRI)..., Peeters [/bib_ref] [bib_ref] Efficacy according to biomarker status of cetuximab plus FOLFOX-4 as first-line treatment..., Bokemeyer [/bib_ref]. However in our study, among patients with KRAS wild-type tumours, the impact of high AURKA-CN level was most pronounced in patients who did not receive cetuximab. One unifying hypothesis may be that Aurora A overexpression confers sensitivity to chemotherapy and not to anti-EGFR therapy. Alternatively, the relatively small number of patients available for this subgroup analysis may preclude a meaningful assessment of this association. In-vitro work has demonstrated strong synergy between anti-EGFR drugs and novel agents targeting the Aurora pathway, which may potentially become a therapeutic approach . Further research to evaluate the response to EGFR-targeted therapy in patients with high and low AURKA-CN may further define the interaction between these two pathways. Many groups have studied AURKA protein expression by IHC in CRC and other cancers. In CRC, increased expression of AURKA protein by IHC has been noted in 19 -30% of patients [bib_ref] A homologue of Drosophila aurora kinase is oncogenic and amplified in human..., Bischoff [/bib_ref] [bib_ref] Aurora kinase expression in colorectal adenocarcinoma: correlations with clinicopathological features, p16 expression,..., Lam [/bib_ref]. In our study, IHC analysis was performed on approximately one third of the available samples, of which a third displayed positive staining (n ¼ 9; 33%). Our results demonstrated lack of staining of samples with low AURKA-CN, and positive staining in less than half of the samples with high AURKA-CN. Increased gene copy number is expected to result in increased activity of the product protein. Thus, the fact that most of the samples with high AURKA-CN were not found to have a positive stain suggests the possibility that despite high copy number AURKA may in fact be inactive in these cells due to posttranscriptional and post-translational regulation. Alternatively, this observation may raise questions regarding the sensitivity of IHC staining. IHC carries many challenges, most notably a variable pattern of staining and the presence of artifacts related to the fixation process. Differences in the preservation process between the samples may also result in various rates of protein degradation, which may affect the IHC protein expression. Furthermore, this assay is user dependent and may suffer from significant interreader variability. A PCR-based method is likely to be more accurate and sensitive in detecting increased activity of a pathway at the gene level. Our study does have several limitations. As a retrospective review, we were limited by specimen availability, which may have introduced unexpected bias into the results. However, by using all specimens available during a specific timeframe, we attempted to minimise this impact. The small sample size limits our ability to draw clear conclusions regarding some of the analyses conducted in the study, such as the interaction between AURKA-CN level and the use of cetuximab. We also did not evaluate a validation cohort to support our chosen cutoff for the analysis of AURKA-CN level. However, all cutoffs selected demonstrated improved clinical outcome for patients with high compared with low AURKA-CN. In summary, we demonstrated a high frequency of increased AURKA-CN in mCRC samples using a novel statistical methodology to evaluate the most appropriate cutoff for analysis. Moreover, our study is the first to demonstrate an association between high AURKA-CN and improved clinical outcome among patients with mCRC receiving chemotherapy, with a more pronounced association noted among patients with KRAS wild-type tumours. Additional study utilising tissue from larger randomised studies to distinguish the prognostic and predictive impact of AURKA-CN is warranted. Implications of these findings for future development of agents targeting Aurora kinase should be considered. This work is published under the standard license to publish agreement. After 12 months the work will become freely available and the license terms will switch to a Creative Commons Attribution-NonCommercial-Share Alike 3.0 Unported License. [fig] 13, Figure 1: AURKA-CN <2.813 AURKA-CN .2.813 Overall survival AURKA-CN cutoff 2AURKA-CN <3 AURKA-CN .3 Overall survival AURKA-CN cutoff 3.0 HR=0.46, 95% CI=0.19 -1.10, P =0.AURKA-CN <3.5 AURKA-CN .3.5 Overall survival AURKA-CN cutoff 3.5 HR=0.45, 95% CI=0.19-1.07, P =0.AURKA-CN <4 AURKA-CN .4 Overall survival AURKA-CN cutoff 4.0 HR=0.51, 95% CI=0.21-1.23, P =0.Kaplan -Meier survival estimates by AURKA-CN and various cutoff values. [/fig] [fig] Figure 2: Frequency of KRAS mutations by AURKA-CN. [/fig] [fig] Low 13, Figure 3: diagnosis of metastatic disease (months) KRAS wt/Low AURKA-CN KRAS wt/High AURKA-CN KRAS mt/Low AURKA-CN KRAS mt/High AURKA-CN PFS on first-line chemotherapy by KRAS and AURKA-CN KRAS wt: HR=0.43; P =0.04 KRAS mt: HR=1.06; metastatic disease (months) KRAS wt/Low AURKA-CN KRAS wt/High AURKA-CN KRAS mt/Low AURKA-CN KRAS mt/High AURKA-CN PFS on second-line chemotherapy by KRAS and AURKA-CN KRAS wt: HR=0.28; P =0.01 KRAS mt: HR=0.61; P =0.AURKA-CN -2.6 High AURKA-CN >2.6 Percent without progression PFS on second-line chemotherapy by AURKA-CN HR=0.54; P =0.Progression-free survival (PFS) of metastatic colorectal cancer patients receiving first-line and second-line chemotherapy by AURKA-CN and KRAS mutation status. (A) PFS on first-line chemotherapy by AURKA-CN; (B) PFS on first-line chemotherapy by AURKA-CN and KRAS mutation status; (C) PFS on second-line chemotherapy by AURKA-CN; and (D) PFS on second-line chemotherapy by AURKA-CN and KRAS mutation status. [/fig] [fig] Figure 4, Figure 5: diagnosis of metastatic disease (months) KRAS wt/low AURKA-CN KRAS wt/high AURKA-CN Overall survival of KRAS wt tumours by AURKA-CN HR=0.14; P =0.diagnosis of metastatic disease (months) Overall survival of KRAS mt tumours by AURKA-CN KRAS mt/low AURKA-CN KRAS mt/high AURKA-CN HR=0.75; P =0.Overall survival of metastatic colorectal cancer patients by AURKA-CN for entire cohort (A), the KRAS wild-type population (B) and the KRAS mutant population (C). Positive staining in colorectal cancer with high AURKA-CN Negative staining in colorectal cancer with low AURKA-CN [/fig] [table] Table 1: Patient and treatment characteristics Abbreviations: AA ¼ African American; CN ¼ copy number. [/table] [table] Table 2: Survival analysis utilising various AURKA-CN cutoff values for the entire cohort [/table] [table] Table 3: Progression-free survival on first-line chemotherapy by KRAS mutation status and receipt of cetuximab [/table]

The Effect of Normobaric Hypoxia in Middle- and/or Long-Distance Runners: Systematic Review # Introduction With an increase in altitude, the barometric pressure decreases exponentially, which leads to a progressive reduction in the ambient partial pressure of oxygen (PO 2 ), otherwise called hypobaric hypoxia (HH) [bib_ref] Effect of acute hypoxia on cognition: A systematic review and meta-regression analysis, Mcmorris [/bib_ref]. This hypoxia can be injurious to health, depending on the altitude attained, rate of ascent, duration at altitude, and physical activity among other factors. It is known that at altitudes of 2.438 to 3.048 m, there is an acute hypoxic ventilatory response (HRV) causing hypocapnia and respiratory alkalosis, which endanger blood-oxygen supply and brain blood flow 438 to 4572 m) [bib_ref] Hypoxic hypoxia at moderate altitudes: Review of the state of the science, Petrassi [/bib_ref]. However, hypoxia also has the ability to foster a series of adaptations, such as: increased capability for transporting oxygen per unit of blood; increased oxygen supply for a given cardiac output [bib_ref] Effect of acute hypoxia on cognition: A systematic review and meta-regression analysis, Mcmorris [/bib_ref] ; improved glycolytic enzymes, glucose transport and pH regulation [bib_ref] Exercise training in normobaric hypoxia in endurance runners. III. Muscular adjustments of..., Zoll [/bib_ref] [bib_ref] Molecular adaptations in human skeletal muscle to endurance training under simulated hypoxic..., Vogt [/bib_ref] or improved adaptations gained from resistance training [bib_ref] Hypoxia Increases Muscle Hypertrophy Induced by Resistance Training, Nishimura [/bib_ref] [bib_ref] Effects of resistance training combined with vascular occlusion or hypoxia on neuromuscular..., Manimmanakorn [/bib_ref]. ## Types of participants We included studies with middle-and/or long-distance runners. We included studies that recruited both men and women, and studies that recruited men and women separately. ## Language Articles published in English or Spanish were included. ## Publication date The search covered publication dates from the beginning until December 2021. ## Exclusion criteria Studies that did not consider measures of sports performance or haematological parameters, and studies that included athletes with illness or injury were excluded. [bib_ref] Exercise training in normobaric hypoxia in endurance runners. III. Muscular adjustments of..., Zoll [/bib_ref] ## Data sources and search strategy We searched in PubMed, SportDiscus, Cochrane Library, Scopus and PEDro database for relevant articles. Thus, we used only reliable, peer-reviewed databases, platforms and sources with search tools that allowed us to access the study dates, and thereby systematically identify studies. We checked the reference lists of all the included studies and systematic reviews for additional references. The terms used to search the database were: (normobaric hypoxia OR altitude) AND runners. ## Selection of studies Two review authors independently screened the titles and abstracts of all the retrieved references in Excel (Microsoft Excel 2018 for Windows). The full-text study reports were retrieved for all the citations that at least one review author considered potentially relevant. Two review authors independently screened the full-text articles and identified studies for inclusion; they also identified and recorded the reasons for excluding studies in the excluded studies characteristics. Any disagreements were resolved through discussion. The selection process is detailed in a PRISMA flow diagram [bib_ref] Preferred reporting items for systematic reviews and meta-analyses: The PRISMA statement, Moher [/bib_ref]. ## Data extraction and management We used a standardized piloted data collection form in Microsoft Excel 2018 for Windows and extracted the following study characteristics and outcome data: (i) Methods: study design; (ii) Participants: randomized number, study participants' mean age or age range, study location and setting, recruitment methods, inclusion and exclusion criteria, and type of endurance sport; (iii) Interventions: a description of the training intervention characteristics, the dose and duration of the training intervention; (iv) Outcomes: a description of the primary and secondary outcomes in the review that were reported in the trial, and a listing of other outcomes collected in the trial; (v) Notes: the trial funding and any notable conflicts of interest of the trial authors. Two review authors independently extracted the outcome data from the included studies into Microsoft Excel 2018 spreadsheets, and compared the data to identify any discrepancies in the data entries. Any disagreements were resolved by consensus. # Methodological quality assessment The methodological quality of each of the studies was carried out with PEDro Scale [fig_ref] Table 1: Methodological quality [/fig_ref] [bib_ref] Reliability of the PEDro Scale for Rating Quality of Randomized Controlled Trials, Maher [/bib_ref]. The maximum score was 7/10 [bib_ref] Intermittent normobaric hypoxia does not alter performance or erythropoietic markers in highly..., Julian [/bib_ref] [bib_ref] Psycho-physiological responses to perceptually-regulated interval runs in hypoxia and normoxia, Hobbins [/bib_ref] and the minimum was 3/10 [bib_ref] Eighteen days of living high, training low stimulate erythropoiesis and enhance aerobic..., Brugniaux [/bib_ref] [bib_ref] Effect of Acute Hypoxia on Cardiorespiratory Coherence in Male Runners, Uryumtsev [/bib_ref]. The remainder of the articles ranged from 4/10 to 6/10 in their score. The initial research comprised 158 studies. After removing duplications, 88 article titles and abstracts were reviewed. Following this process, 19 (full-text) were thoroughly read and, from these, 12 were finally included for the systematic review. The PRISMA flowchart illustrates the search and selection process [fig_ref] Figure 1: Preferred Reporting Items for Systematic Reviews and Meta-Analysis [/fig_ref]. ## Included studies Twelve studies made up this systematic review [fig_ref] Table 2: Study characteristics [/fig_ref]. It is important to highlight that some of these presented several intervention groups, and that for each of these groups, the procedures involved differences; for instance, in the time of day at which the hypoxia was administrated (night/day), in the position (repose and/or physical activity), or in the time and manner of administration of hypoxia. ## Included studies Twelve studies made up this systematic review [fig_ref] Table 2: Study characteristics [/fig_ref]. It is important to highlight that some of these presented several intervention groups, and that for each of these groups, the procedures involved differences; for instance, in the time of day at which the hypoxia was administrated (night/day), in the position (repose and/or physical activity), or in the time and manner of administration of hypoxia. Only eight of the scientific studies had a control group for their interventions [bib_ref] Intermittent normobaric hypoxia does not alter performance or erythropoietic markers in highly..., Julian [/bib_ref] [bib_ref] Eighteen days of living high, training low stimulate erythropoiesis and enhance aerobic..., Brugniaux [/bib_ref] [bib_ref] Effects of Intermittent Hypoxia on Running Economy, Burtscher [/bib_ref] [bib_ref] Exercise training in normobaric hypoxia in endurance runners. I. Improvement in aerobic..., Dufour [/bib_ref] [bib_ref] Effect of intermittent hypoxia on oxygen uptake during submaximal exercise in endurance..., Katayama [/bib_ref] [bib_ref] Changes in ventilatory responses to hypercapnia and hypoxia after intermittent hypoxia in..., Katayama [/bib_ref] [bib_ref] The effects of nightly normobaric hypoxia and high intensity training under intermittent..., Neya [/bib_ref] [bib_ref] Reproducibility of Performance Changes to Simulated Live High/Train Low Altitude, Robertson [/bib_ref]. The control groups of the studies [bib_ref] Intermittent normobaric hypoxia does not alter performance or erythropoietic markers in highly..., Julian [/bib_ref] [bib_ref] Eighteen days of living high, training low stimulate erythropoiesis and enhance aerobic..., Brugniaux [/bib_ref] [bib_ref] Effects of Intermittent Hypoxia on Running Economy, Burtscher [/bib_ref] [bib_ref] Effect of intermittent hypoxia on oxygen uptake during submaximal exercise in endurance..., Katayama [/bib_ref] [bib_ref] Changes in ventilatory responses to hypercapnia and hypoxia after intermittent hypoxia in..., Katayama [/bib_ref] did not receive the hypoxia at rest, while their corresponding intervention groups did. Furthermore, the control groups of Neya et al. [bib_ref] The effects of nightly normobaric hypoxia and high intensity training under intermittent..., Neya [/bib_ref] and Dufour et al. [bib_ref] Exercise training in normobaric hypoxia in endurance runners. I. Improvement in aerobic..., Dufour [/bib_ref] were physically active, but with an inspiratory pressure of 20.9%, both of which factors are considered normoxia conditions. ## - Study location Only eight of the scientific studies had a control group for their interventions [bib_ref] Intermittent normobaric hypoxia does not alter performance or erythropoietic markers in highly..., Julian [/bib_ref] [bib_ref] Eighteen days of living high, training low stimulate erythropoiesis and enhance aerobic..., Brugniaux [/bib_ref] [bib_ref] Effects of Intermittent Hypoxia on Running Economy, Burtscher [/bib_ref] [bib_ref] Exercise training in normobaric hypoxia in endurance runners. I. Improvement in aerobic..., Dufour [/bib_ref] [bib_ref] Effect of intermittent hypoxia on oxygen uptake during submaximal exercise in endurance..., Katayama [/bib_ref] [bib_ref] Changes in ventilatory responses to hypercapnia and hypoxia after intermittent hypoxia in..., Katayama [/bib_ref] [bib_ref] The effects of nightly normobaric hypoxia and high intensity training under intermittent..., Neya [/bib_ref] [bib_ref] Reproducibility of Performance Changes to Simulated Live High/Train Low Altitude, Robertson [/bib_ref]. The control groups of the studies [bib_ref] Intermittent normobaric hypoxia does not alter performance or erythropoietic markers in highly..., Julian [/bib_ref] [bib_ref] Eighteen days of living high, training low stimulate erythropoiesis and enhance aerobic..., Brugniaux [/bib_ref] [bib_ref] Effects of Intermittent Hypoxia on Running Economy, Burtscher [/bib_ref] [bib_ref] Effect of intermittent hypoxia on oxygen uptake during submaximal exercise in endurance..., Katayama [/bib_ref] [bib_ref] Changes in ventilatory responses to hypercapnia and hypoxia after intermittent hypoxia in..., Katayama [/bib_ref] did not receive the hypoxia at rest, while their corresponding intervention groups did. Furthermore, the control groups of Neya et al. [bib_ref] The effects of nightly normobaric hypoxia and high intensity training under intermittent..., Neya [/bib_ref] and Dufour et al. [bib_ref] Exercise training in normobaric hypoxia in endurance runners. I. Improvement in aerobic..., Dufour [/bib_ref] were physically active, but with an inspiratory pressure of 20.9%, both of which factors are considered normoxia conditions. ## - ## Study location The studies took place in different countries: four in Japan [bib_ref] Effects of Normobaric Hypoxia Equivalent to 2,000-m Altitude on Sleep and Physiological..., Hoshikawa [/bib_ref] [bib_ref] Effect of intermittent hypoxia on oxygen uptake during submaximal exercise in endurance..., Katayama [/bib_ref] [bib_ref] Changes in ventilatory responses to hypercapnia and hypoxia after intermittent hypoxia in..., Katayama [/bib_ref] [bib_ref] The effects of nightly normobaric hypoxia and high intensity training under intermittent..., Neya [/bib_ref] , two in Australia [bib_ref] Effectiveness of intermittent training in hypoxia combined with live high/train low. Graefe's..., Robertson [/bib_ref] [bib_ref] Reproducibility of Performance Changes to Simulated Live High/Train Low Altitude, Robertson [/bib_ref] , two in France [bib_ref] Eighteen days of living high, training low stimulate erythropoiesis and enhance aerobic..., Brugniaux [/bib_ref] [bib_ref] Exercise training in normobaric hypoxia in endurance runners. I. Improvement in aerobic..., Dufour [/bib_ref] , one in The United States [bib_ref] Intermittent normobaric hypoxia does not alter performance or erythropoietic markers in highly..., Julian [/bib_ref] , one in Russia [bib_ref] Effect of Acute Hypoxia on Cardiorespiratory Coherence in Male Runners, Uryumtsev [/bib_ref] , one in Austria [bib_ref] Effects of Intermittent Hypoxia on Running Economy, Burtscher [/bib_ref] and one in England [bib_ref] Psycho-physiological responses to perceptually-regulated interval runs in hypoxia and normoxia, Hobbins [/bib_ref]. ## - Sample size and years of studies The sample size of each of the studies had a variation of 7 to 29 subjects. The oldest studies dated from 2004 [bib_ref] Intermittent normobaric hypoxia does not alter performance or erythropoietic markers in highly..., Julian [/bib_ref] [bib_ref] Effect of intermittent hypoxia on oxygen uptake during submaximal exercise in endurance..., Katayama [/bib_ref] [bib_ref] Changes in ventilatory responses to hypercapnia and hypoxia after intermittent hypoxia in..., Katayama [/bib_ref] , and the more recent dated from 2020 [bib_ref] Effect of Acute Hypoxia on Cardiorespiratory Coherence in Male Runners, Uryumtsev [/bib_ref]. ## - Duration of the hypoxia programme The duration of the normobaric hypoxia exposure programme and the sessions per week were variable [fig_ref] Table 3: Characteristics of normobaric hypoxia [/fig_ref]. Three of the studies had a programme duration of one week [bib_ref] Psycho-physiological responses to perceptually-regulated interval runs in hypoxia and normoxia, Hobbins [/bib_ref] [bib_ref] Effects of Normobaric Hypoxia Equivalent to 2,000-m Altitude on Sleep and Physiological..., Hoshikawa [/bib_ref] [bib_ref] Changes in ventilatory responses to hypercapnia and hypoxia after intermittent hypoxia in..., Katayama [/bib_ref] , while the maximum duration was thirteen weeks, which occurred in only one of the cases [bib_ref] Effects of Intermittent Hypoxia on Running Economy, Burtscher [/bib_ref]. Regarding the sessions per week, half of the studies exposed the subjects to hypoxia every day [bib_ref] Eighteen days of living high, training low stimulate erythropoiesis and enhance aerobic..., Brugniaux [/bib_ref] [bib_ref] Effects of Normobaric Hypoxia Equivalent to 2,000-m Altitude on Sleep and Physiological..., Hoshikawa [/bib_ref] [bib_ref] Effect of intermittent hypoxia on oxygen uptake during submaximal exercise in endurance..., Katayama [/bib_ref] [bib_ref] Changes in ventilatory responses to hypercapnia and hypoxia after intermittent hypoxia in..., Katayama [/bib_ref] [bib_ref] The effects of nightly normobaric hypoxia and high intensity training under intermittent..., Neya [/bib_ref] [bib_ref] Reproducibility of Performance Changes to Simulated Live High/Train Low Altitude, Robertson [/bib_ref] , while the remainder ranged from two to five sessions per week. ## - participants The sporting or professional levels to which the different populations of runners in the different articles belonged were highly heterogeneous [fig_ref] Table 2: Study characteristics [/fig_ref] : 76 belonged to college teams, 18 were part of a local athletics team, 25 were national runners with the USA national team, 28 were considered elite, and there were 37 participants whose professional level was not specified. Mean VO 2 max (ml/min/kg) was not included in every study, while the maximum was recorded as 68.75 and the minimum was 59.65. All the participants were middle-or long-distance runners. Only three of the studies described the timing of the participants undergoing the hypoxia programme: Katayama et al. [bib_ref] Effect of intermittent hypoxia on oxygen uptake during submaximal exercise in endurance..., Katayama [/bib_ref] conducted their study seven weeks before the start of a championship, Hoshiwaka et al. [bib_ref] Effects of Normobaric Hypoxia Equivalent to 2,000-m Altitude on Sleep and Physiological..., Hoshikawa [/bib_ref] conducted theirs during the training season, while the study by Burstcher et al. [bib_ref] Effects of Intermittent Hypoxia on Running Economy, Burtscher [/bib_ref] took place two months after the end of the season. The ages of the participants ranged from 19.6 to 33.4 years. In terms of gender, five of studies included men and women [bib_ref] Intermittent normobaric hypoxia does not alter performance or erythropoietic markers in highly..., Julian [/bib_ref] [bib_ref] Psycho-physiological responses to perceptually-regulated interval runs in hypoxia and normoxia, Hobbins [/bib_ref] [bib_ref] Effects of Intermittent Hypoxia on Running Economy, Burtscher [/bib_ref] [bib_ref] Effectiveness of intermittent training in hypoxia combined with live high/train low. Graefe's..., Robertson [/bib_ref] [bib_ref] Reproducibility of Performance Changes to Simulated Live High/Train Low Altitude, Robertson [/bib_ref] , six included only men [bib_ref] Eighteen days of living high, training low stimulate erythropoiesis and enhance aerobic..., Brugniaux [/bib_ref] [bib_ref] Effect of Acute Hypoxia on Cardiorespiratory Coherence in Male Runners, Uryumtsev [/bib_ref] [bib_ref] Exercise training in normobaric hypoxia in endurance runners. I. Improvement in aerobic..., Dufour [/bib_ref] [bib_ref] Effect of intermittent hypoxia on oxygen uptake during submaximal exercise in endurance..., Katayama [/bib_ref] [bib_ref] Changes in ventilatory responses to hypercapnia and hypoxia after intermittent hypoxia in..., Katayama [/bib_ref] [bib_ref] The effects of nightly normobaric hypoxia and high intensity training under intermittent..., Neya [/bib_ref] and in one study the participants were only women [bib_ref] Effects of Normobaric Hypoxia Equivalent to 2,000-m Altitude on Sleep and Physiological..., Hoshikawa [/bib_ref]. - Intervention ## Intervention model Normobaric hypoxia was administered to participants in two states: resting or performing physical activities [fig_ref] Table 3: Characteristics of normobaric hypoxia [/fig_ref]. Some of the studies included several intervention groups in which participants underwent normobaric hypoxia at rest and/or normobaric hypoxia while undertaking physical activity. The study by Robertson et al. [bib_ref] Effectiveness of intermittent training in hypoxia combined with live high/train low. Graefe's..., Robertson [/bib_ref] presented two intervention groups: the first was administered hypoxia during physical activity only, while the second was administered the treatment in both states, i.e., at rest and during physical activity. Similarly, Neya et al. [bib_ref] The effects of nightly normobaric hypoxia and high intensity training under intermittent..., Neya [/bib_ref] established two intervention groups: the first group underwent hypoxia at rest, and the second group underwent hypoxia during physical activity. The remaining studies administered hypoxia at rest only [bib_ref] Intermittent normobaric hypoxia does not alter performance or erythropoietic markers in highly..., Julian [/bib_ref] [bib_ref] Eighteen days of living high, training low stimulate erythropoiesis and enhance aerobic..., Brugniaux [/bib_ref] [bib_ref] Effect of Acute Hypoxia on Cardiorespiratory Coherence in Male Runners, Uryumtsev [/bib_ref] [bib_ref] Effects of Intermittent Hypoxia on Running Economy, Burtscher [/bib_ref] [bib_ref] Effect of intermittent hypoxia on oxygen uptake during submaximal exercise in endurance..., Katayama [/bib_ref] [bib_ref] Changes in ventilatory responses to hypercapnia and hypoxia after intermittent hypoxia in..., Katayama [/bib_ref] [bib_ref] Reproducibility of Performance Changes to Simulated Live High/Train Low Altitude, Robertson [/bib_ref] , during physical activity only [bib_ref] Psycho-physiological responses to perceptually-regulated interval runs in hypoxia and normoxia, Hobbins [/bib_ref] [bib_ref] Exercise training in normobaric hypoxia in endurance runners. I. Improvement in aerobic..., Dufour [/bib_ref] and both at rest and during physical activity [bib_ref] Effects of Normobaric Hypoxia Equivalent to 2,000-m Altitude on Sleep and Physiological..., Hoshikawa [/bib_ref]. Among the studies administering hypoxia at rest, four did not explain the specific physical position of participants at the time of administration [bib_ref] Intermittent normobaric hypoxia does not alter performance or erythropoietic markers in highly..., Julian [/bib_ref] [bib_ref] Effects of Intermittent Hypoxia on Running Economy, Burtscher [/bib_ref] [bib_ref] Effect of intermittent hypoxia on oxygen uptake during submaximal exercise in endurance..., Katayama [/bib_ref] [bib_ref] Reproducibility of Performance Changes to Simulated Live High/Train Low Altitude, Robertson [/bib_ref]. In one group in the study by Neya et al. [bib_ref] The effects of nightly normobaric hypoxia and high intensity training under intermittent..., Neya [/bib_ref] and in the study by Brugniaux et al. [bib_ref] Eighteen days of living high, training low stimulate erythropoiesis and enhance aerobic..., Brugniaux [/bib_ref] , participants received hypoxia while sleeping. The hypoxia administrations lasted 10-12 h/session and 14 h/session, respectively. In two of the studies, participants were seated in a chair [bib_ref] Effect of Acute Hypoxia on Cardiorespiratory Coherence in Male Runners, Uryumtsev [/bib_ref] [bib_ref] Changes in ventilatory responses to hypercapnia and hypoxia after intermittent hypoxia in..., Katayama [/bib_ref]. Regarding the form of exposure, Urymstev et al. [bib_ref] Effect of Acute Hypoxia on Cardiorespiratory Coherence in Male Runners, Uryumtsev [/bib_ref] performed intermittent normobaric administration, with a duration of 10:10 per session, i.e., 10 min of normoxia followed by 10 min of hypoxia, while Katayama et al. [bib_ref] Changes in ventilatory responses to hypercapnia and hypoxia after intermittent hypoxia in..., Katayama [/bib_ref] administered 3 h continuously. In studies by Robertson et al. [bib_ref] Effectiveness of intermittent training in hypoxia combined with live high/train low. Graefe's..., Robertson [/bib_ref] (second group) and Hoshiwaka et al. [bib_ref] Effects of Normobaric Hypoxia Equivalent to 2,000-m Altitude on Sleep and Physiological..., Hoshikawa [/bib_ref] , participants were subjected to hypoxia in both states. The latter study administered normobaric hypoxia to subjects in the supine position for 7 h each night. In addition, it combined treadmill and cycloergometer exercises for physical activity. The second group in the study by Robertson et al. [bib_ref] Effectiveness of intermittent training in hypoxia combined with live high/train low. Graefe's..., Robertson [/bib_ref] also used the treadmill to exercise under hypoxic conditions for a duration of approximately one hour per session. However, for administering hypoxia at rest, the exact position used was not described. Subjects in the Dufour et al. [bib_ref] Exercise training in normobaric hypoxia in endurance runners. I. Improvement in aerobic..., Dufour [/bib_ref] and Hobbins [bib_ref] Psycho-physiological responses to perceptually-regulated interval runs in hypoxia and normoxia, Hobbins [/bib_ref] studies, and in the second intervention group of Neya et al. [bib_ref] The effects of nightly normobaric hypoxia and high intensity training under intermittent..., Neya [/bib_ref] and the first intervention group of Robertson et al. [bib_ref] Effectiveness of intermittent training in hypoxia combined with live high/train low. Graefe's..., Robertson [/bib_ref] underwent normobaric hypoxia while performing the physical activity, all using the treadmill to perform the activity. Physical Activity Programme during Exposure to Normobaric Hypoxia The second group in the study by Neya et al. [bib_ref] The effects of nightly normobaric hypoxia and high intensity training under intermittent..., Neya [/bib_ref] trained on a treadmill for 30 min each session. The intensity was of 80-90% of the maximum heart rate (HRmax) reached during the VO 2 max test at sea level before the intervention. The subjects started at 80% of their HRmax for the first 10 min, and gradually increased speed over the next 20 min until they reached 90% of their HRmax. The duration of the exercise in the physical activity programme of the Dufour et al. [bib_ref] Exercise training in normobaric hypoxia in endurance runners. I. Improvement in aerobic..., Dufour [/bib_ref] study changed according to the week of intervention. During weeks 3 and 6, the participants ran for two periods of 12 min each, while during weeks 4 and 7, the duration of each running period was increased to 16 min, and during weeks 5 and 8, the participants ran for two periods of 20 min each. Robertson et al. [bib_ref] Effectiveness of intermittent training in hypoxia combined with live high/train low. Graefe's..., Robertson [/bib_ref] presented two intervention groups. During the week they undertook one long-duration session, one moderate-duration session and two interval or high-intensity sessions. The intensity during each session is not described in two studies [bib_ref] Exercise training in normobaric hypoxia in endurance runners. I. Improvement in aerobic..., Dufour [/bib_ref] [bib_ref] Effectiveness of intermittent training in hypoxia combined with live high/train low. Graefe's..., Robertson [/bib_ref]. Hobbins et al. [bib_ref] Psycho-physiological responses to perceptually-regulated interval runs in hypoxia and normoxia, Hobbins [/bib_ref] suggested a HIIT (High Intensity Interval Training) programme during the hypoxia. This consisted of four running sessions of 4 min duration. Between each of the sessions there was a rest time of 3 min, also under normobaric hypoxia conditions. The intensity during the first 30 s of the HIIT was determined by the participant's favourite running speed; after 30 s, the participant indicated an increase or decrease in the speed by signalling up or down with their thumb. Each participant was in normobaric hypoxia for 28 min per session. The training programme chosen by Hoshiwaka et al. [bib_ref] Effects of Normobaric Hypoxia Equivalent to 2,000-m Altitude on Sleep and Physiological..., Hoshikawa [/bib_ref] consisted of five sets of 30 s on a cycloergometer, 4 min rest between exercises, six periods of 5 min running at a heart rate of 120-180 beats per minute, and finally, 30 min of cycloergometer exercise at 80 rpm. ## Duration of exposure to hypoxia per session and duration of the intervention The duration of hypoxia per session was variable among the different intervention groups, both at rest and during physical activity. Regarding the first group (at rest), two studies performed the exposure intermittently [bib_ref] Intermittent normobaric hypoxia does not alter performance or erythropoietic markers in highly..., Julian [/bib_ref] [bib_ref] Effect of Acute Hypoxia on Cardiorespiratory Coherence in Male Runners, Uryumtsev [/bib_ref]. Julian et al. [bib_ref] Intermittent normobaric hypoxia does not alter performance or erythropoietic markers in highly..., Julian [/bib_ref] conducted 5:5, i.e., 5 min of hypoxia followed by 5 min of normoxia for 70 min, 5 times per week for 4 weeks, while Urymtsev et al. [bib_ref] Effect of Acute Hypoxia on Cardiorespiratory Coherence in Male Runners, Uryumtsev [/bib_ref] followed a pattern of 10:10 min in each session. The latter article provides no information about either the number of sessions per week or the total duration. Among the studies in which hypoxia was undergone while sleeping, Neya et al. [bib_ref] The effects of nightly normobaric hypoxia and high intensity training under intermittent..., Neya [/bib_ref] applied hypoxia for 10-12 h over 29 nights, Bryniaux et al. [bib_ref] Eighteen days of living high, training low stimulate erythropoiesis and enhance aerobic..., Brugniaux [/bib_ref] followed a programme of 14 h per day for 18 days and Hoshiwaka et al. [bib_ref] Effects of Normobaric Hypoxia Equivalent to 2,000-m Altitude on Sleep and Physiological..., Hoshikawa [/bib_ref] administered hypoxia 7 h per day for 6 nights. In all cases, these sessions took place on consecutive nights. Regarding the studies [bib_ref] Effects of Intermittent Hypoxia on Running Economy, Burtscher [/bib_ref] [bib_ref] Effect of intermittent hypoxia on oxygen uptake during submaximal exercise in endurance..., Katayama [/bib_ref] [bib_ref] Changes in ventilatory responses to hypercapnia and hypoxia after intermittent hypoxia in..., Katayama [/bib_ref] [bib_ref] Reproducibility of Performance Changes to Simulated Live High/Train Low Altitude, Robertson [/bib_ref] in which all participants were seated, or in which the position was not specified in the study, participants had an average of 5 h of exposure per session, and the hypoxia programme lasted an average of 29.4 days. In relation to the groups that underwent hypoxia while they were performing physical activity, the mean duration of the sessions was 45 min and the hypoxia programme lasted a mean of 19.5 days. ## Altitude and hypoxia simulator device Hypoxia was generated using different altitude simulator devices. Face masks were used in 41.67% of studies [bib_ref] Effects of Intermittent Hypoxia on Running Economy, Burtscher [/bib_ref] [bib_ref] Exercise training in normobaric hypoxia in endurance runners. I. Improvement in aerobic..., Dufour [/bib_ref] [bib_ref] Effects of Normobaric Hypoxia Equivalent to 2,000-m Altitude on Sleep and Physiological..., Hoshikawa [/bib_ref] [bib_ref] Effect of intermittent hypoxia on oxygen uptake during submaximal exercise in endurance..., Katayama [/bib_ref] [bib_ref] Reproducibility of Performance Changes to Simulated Live High/Train Low Altitude, Robertson [/bib_ref] and hypoxic rooms were used in 58.63% [bib_ref] Intermittent normobaric hypoxia does not alter performance or erythropoietic markers in highly..., Julian [/bib_ref] [bib_ref] Psycho-physiological responses to perceptually-regulated interval runs in hypoxia and normoxia, Hobbins [/bib_ref] [bib_ref] Eighteen days of living high, training low stimulate erythropoiesis and enhance aerobic..., Brugniaux [/bib_ref] [bib_ref] Effect of Acute Hypoxia on Cardiorespiratory Coherence in Male Runners, Uryumtsev [/bib_ref] [bib_ref] Changes in ventilatory responses to hypercapnia and hypoxia after intermittent hypoxia in..., Katayama [/bib_ref] [bib_ref] The effects of nightly normobaric hypoxia and high intensity training under intermittent..., Neya [/bib_ref] [bib_ref] Effectiveness of intermittent training in hypoxia combined with live high/train low. Graefe's..., Robertson [/bib_ref]. The simulated altitudes varied widely between studies. In three studies [bib_ref] Intermittent normobaric hypoxia does not alter performance or erythropoietic markers in highly..., Julian [/bib_ref] [bib_ref] Eighteen days of living high, training low stimulate erythropoiesis and enhance aerobic..., Brugniaux [/bib_ref] [bib_ref] Effects of Intermittent Hypoxia on Running Economy, Burtscher [/bib_ref] , simulated altitudes changed during the intervention. Participants in the study by Julian et al. [bib_ref] Intermittent normobaric hypoxia does not alter performance or erythropoietic markers in highly..., Julian [/bib_ref] were exposed to different altitudes throughout the 4 weeks of hypoxia. The first week the fraction of inspired oxygen was 0.12, in the second week it decreased to 0.11 and in the two last weeks it fell to 0.10. Brugniaux et al. [bib_ref] Eighteen days of living high, training low stimulate erythropoiesis and enhance aerobic..., Brugniaux [/bib_ref] simulated an altitude of 2500 metres for the first six nights of the study, and 3000 metres for the following twelve nights. Similarly, Butscher et al. [bib_ref] Effects of Intermittent Hypoxia on Running Economy, Burtscher [/bib_ref] varied the altitude by between 15% and 11%, which equated to fluctuations from 3200 to 5500 metres, respectively. Additionally, several authors chose to vary the altitude and oxygen percentage depending on the presence or absence of physical activity. Hoshiwaka et al. [bib_ref] Effects of Normobaric Hypoxia Equivalent to 2,000-m Altitude on Sleep and Physiological..., Hoshikawa [/bib_ref] administered an inspired oxygen fraction (FiO 2 ) of 16.4% at rest and an FiO 2 = 14.4% in the active state, i.e., equivalent to 2000 and 3000 metres, respectively. Robertson et al. [bib_ref] Effectiveness of intermittent training in hypoxia combined with live high/train low. Graefe's..., Robertson [/bib_ref] made a similar distinction, stipulating 3000 metres when performing physical activity and 2200 metres when at rest. Neya et al. [bib_ref] The effects of nightly normobaric hypoxia and high intensity training under intermittent..., Neya [/bib_ref] , Dufour et al. [bib_ref] Exercise training in normobaric hypoxia in endurance runners. I. Improvement in aerobic..., Dufour [/bib_ref] and Robertson et al. [bib_ref] Reproducibility of Performance Changes to Simulated Live High/Train Low Altitude, Robertson [/bib_ref] stipulated 3000 metres for runners. Katayama et al. [bib_ref] Effect of intermittent hypoxia on oxygen uptake during submaximal exercise in endurance..., Katayama [/bib_ref] and Katayama administrated an inspired oxygen fraction of 12.3%. Urymtsev et al. [bib_ref] Effect of Acute Hypoxia on Cardiorespiratory Coherence in Male Runners, Uryumtsev [/bib_ref] did not specify the simulated altitude. ## Outcome measure Outcome measures were highly heterogeneous among the different included articles. This is because although all the studies administered normobaric hypoxia to middle-and long-distance runners, the objectives of some studies included highly specific outcomes. In this systematic review, we classified the outcome measures into two groups: haematological parameters [fig_ref] Table 4: Haematological parameters outcome [/fig_ref] and sport performance measures [fig_ref] Table 5: Sport performance outcome [/fig_ref]. INT 2 values increased after the end of the intervention (p < 0.05). However, after 2 weeks the values were the same as the initial values. INT 1 increased significantly after the end of the intervention (p < 0.05) These measures were executed by field trials or laboratory tests. We included time to exhaustion and 3000 m run tests in this group. ## Haematological parameters This section includes the following parameters: hemoglobin concentration, lactate concentration, percentage of reticulocytes, oxygen saturation, heart rate, maximal heart rate, percentage of hematocrit and erythropoietin values. # Discussion As far as we know, this is the first systematic review evaluating the changes produced by exposure to normobaric hypoxia in middle-and long-distance runners from different competition levels. In terms of exposure time, most studies used exposure times of 3 h or less; this type of exposure is called intermittent hypoxia [bib_ref] Combining Hypoxic Methods for Peak Performance, Millet [/bib_ref]. Currently, this appears to be the most commonly used type, as the articles with the oldest publication dates were those with the longest exposures (more than 12 h/day) [bib_ref] Eighteen days of living high, training low stimulate erythropoiesis and enhance aerobic..., Brugniaux [/bib_ref] [bib_ref] The effects of nightly normobaric hypoxia and high intensity training under intermittent..., Neya [/bib_ref] [bib_ref] Reproducibility of Performance Changes to Simulated Live High/Train Low Altitude, Robertson [/bib_ref]. The hypoxic room was the most commonly used hypoxia simulator. Studies with short exposure time (less than one hour) used a face mask [bib_ref] Psycho-physiological responses to perceptually-regulated interval runs in hypoxia and normoxia, Hobbins [/bib_ref] [bib_ref] Effect of Acute Hypoxia on Cardiorespiratory Coherence in Male Runners, Uryumtsev [/bib_ref] [bib_ref] Exercise training in normobaric hypoxia in endurance runners. I. Improvement in aerobic..., Dufour [/bib_ref]. This may be because a hypoxic room or tent is more comfortable for long exposure times. Hoshiwaka et al. [bib_ref] Effects of Normobaric Hypoxia Equivalent to 2,000-m Altitude on Sleep and Physiological..., Hoshikawa [/bib_ref] was the only study in this review to use higher altitudes during exercise than at rest. This is striking, as exercise sessions in hypoxia place greater stress, fatigue, immunosuppression and stress on the body than those performed in normoxia [bib_ref] Is Hypoxia Training Good for Muscles and Exercise Performance?, Vogt [/bib_ref] , and coupled with the increased altitude, cause an even greater stimulus to the body. ## Outcome measures It seems that it is not necessary to apply long exposure times [bib_ref] The effects of nightly normobaric hypoxia and high intensity training under intermittent..., Neya [/bib_ref] to obtain a significant increase in the time to exhaustion, with interventions of three hours or less being effective [bib_ref] Exercise training in normobaric hypoxia in endurance runners. I. Improvement in aerobic..., Dufour [/bib_ref] [bib_ref] Effect of intermittent hypoxia on oxygen uptake during submaximal exercise in endurance..., Katayama [/bib_ref] , and the treatment's effectiveness does not depend on the state of activity in which it is applied (at rest or exercise). The HIIT programme, carried out by Dufour et al. [bib_ref] Exercise training in normobaric hypoxia in endurance runners. I. Improvement in aerobic..., Dufour [/bib_ref] , increased this outcome measure in runners. In addition, Vallier et al. [bib_ref] Effects of physical training in a hypobaric chamber on the physical performance..., Vallier [/bib_ref] , reported that a HIIT programme under hypoxia conditions for 3 weeks also increased time to exhaustion in in elite triathletes. The study of Katayama et al. [bib_ref] Effect of intermittent hypoxia on oxygen uptake during submaximal exercise in endurance..., Katayama [/bib_ref] was the only one to show a trend of improvement in the 3000 m run, while Julian et al. [bib_ref] Intermittent normobaric hypoxia does not alter performance or erythropoietic markers in highly..., Julian [/bib_ref] showed no significant change. This may have been a consequence of using sessions that were too short in time (70 min), and in addition, they alternated between normoxic and hypoxia states, and thus did not achieve 70 min of hypoxia. Even so, some authors have achieved significant improvements in the 3000 m race time using hypobaric hypoxia [bib_ref] Intermittent Hypoxia Improves Endurance Performance and Submaximal Exercise Efficiency, Katayama [/bib_ref]. However, we know that hypobaric hypoxia is a more stressful stimulus than normobaric hypoxia [bib_ref] Effects of Normobaric Hypoxia on Matched-severe Exercise and Powerduration Relationship, Sousa [/bib_ref] , so these parameters should be tested using normobaric hypoxia to ascertain whether there are similar improvements in 3000 m run times. To increase haemoglobin concentration, studies agreed on the need for long exposure times [bib_ref] Live high, train low does not change the total hemoglobin mass of..., Ashenden [/bib_ref] [bib_ref] Effect of living in hypoxia and training in normoxia on sea level..., Rusko [/bib_ref]. However, there was no consensus on the exact time, with some claiming that 12-16 h/day for 25 days are necessary [bib_ref] Effect of living in hypoxia and training in normoxia on sea level..., Rusko [/bib_ref] , while others affirmed that 8-10 h/day are enough [bib_ref] Live high, train low does not change the total hemoglobin mass of..., Ashenden [/bib_ref]. This may be the reason why the studies by Julian et al. [bib_ref] Intermittent normobaric hypoxia does not alter performance or erythropoietic markers in highly..., Julian [/bib_ref] , Katayama et al. [bib_ref] Effect of intermittent hypoxia on oxygen uptake during submaximal exercise in endurance..., Katayama [/bib_ref] and Bustcher et al. [bib_ref] Effects of Intermittent Hypoxia on Running Economy, Burtscher [/bib_ref] did not increase haemoglobin concentration (using short exposures), whereas the studies in our review using exposures of 14 h/day or more achieved an increase after intervention [bib_ref] Effectiveness of intermittent training in hypoxia combined with live high/train low. Graefe's..., Robertson [/bib_ref] [bib_ref] Reproducibility of Performance Changes to Simulated Live High/Train Low Altitude, Robertson [/bib_ref]. In the studies that showed an improvement in the reticulocyte percentage [bib_ref] Intermittent normobaric hypoxia does not alter performance or erythropoietic markers in highly..., Julian [/bib_ref] [bib_ref] Effectiveness of intermittent training in hypoxia combined with live high/train low. Graefe's..., Robertson [/bib_ref] or in haematocrit [bib_ref] Effects of Intermittent Hypoxia on Running Economy, Burtscher [/bib_ref] , improvement was only seen during the days following the start of the intervention; however, with the passage of time (during the intervention or after the intervention had ended) these improvements were not sustained. Julian et al. [bib_ref] Intermittent normobaric hypoxia does not alter performance or erythropoietic markers in highly..., Julian [/bib_ref] increased the altitude week by week, and despite these variations in the stimulus, the increase in reticulocytes was not sustained after the end of the intervention. Oxygen saturation (SpO 2 ) decreased in three studies [bib_ref] Psycho-physiological responses to perceptually-regulated interval runs in hypoxia and normoxia, Hobbins [/bib_ref] [bib_ref] Effects of Normobaric Hypoxia Equivalent to 2,000-m Altitude on Sleep and Physiological..., Hoshikawa [/bib_ref] [bib_ref] Reproducibility of Performance Changes to Simulated Live High/Train Low Altitude, Robertson [/bib_ref]. Hoshiwaka et al. [bib_ref] Effects of Normobaric Hypoxia Equivalent to 2,000-m Altitude on Sleep and Physiological..., Hoshikawa [/bib_ref] intended to relate the effects of normobaric hypoxia to sleep parameters and, consequently, hypoxia occurred while the patients were sleeping. This study evinced that exposure to hypoxia causes a decrease in the SpO 2, and this in turn leads to lighter sleep. This finding adds to others in the literature [bib_ref] Changes in sleep quality of athletes under normobaric hypoxia equivalent to 2,000-m..., Hoshikawa [/bib_ref]. This may suggest that the application of a hypoxia programme while athletes sleep may affect the quality of their rest. However, we cannot relate this to a reduction in their sports performance, because in their study, Hoshiwaka et al. [bib_ref] Effects of Normobaric Hypoxia Equivalent to 2,000-m Altitude on Sleep and Physiological..., Hoshikawa [/bib_ref] did not take into account field or laboratory tests evaluating the participants' sports performance. The decrease in SpO 2 in the study by Hoshiwaka et al. [bib_ref] Effects of Normobaric Hypoxia Equivalent to 2,000-m Altitude on Sleep and Physiological..., Hoshikawa [/bib_ref] was related to an increase in heart rate during the first night after breathing oxygen-poor air compared to normoxia. However, when comparing the last night of hypoxia with the first night, there was a significant decrease, which derived from the effects of training on the autonomic nervous system, and not from the acclimatisation to hypoxia. Acclimatisation of the autonomic nervous system took place in six days in the study by Hoshiwaka et al. [bib_ref] Effects of Normobaric Hypoxia Equivalent to 2,000-m Altitude on Sleep and Physiological..., Hoshikawa [/bib_ref] (from the first to the sixth night of hypoxia), and the same was true for the study by Astorino et al. [bib_ref] Adaptations to high-intensity training are independent of gender. Graefe's Arch, Astorino [/bib_ref]. Of the articles included in this review, EPO values only increased in the study by Robertson et al. [bib_ref] Reproducibility of Performance Changes to Simulated Live High/Train Low Altitude, Robertson [/bib_ref] , whose duration of hypoxia administration was 14 h/day, and whose participants were elite middle-and long-distance runners. It increased on days 2 and 6 of exposure (in both block 1 and 2) but was not maintained on days 20 and 27 in either intervention block. In the studies by Katayama et al. [bib_ref] Effect of intermittent hypoxia on oxygen uptake during submaximal exercise in endurance..., Katayama [/bib_ref] and Julian et al. [bib_ref] Intermittent normobaric hypoxia does not alter performance or erythropoietic markers in highly..., Julian [/bib_ref] , 3 h/day and 70 min/session, respectively, were not enough to stimulate a sharp increase in the concentration of EPO. These results in the studies in this review are ratified by the literature, as many authors defend the idea that the use of daily exposure sessions shorter than 12 h does not produce any improvement in the production of red blood cells in athletes. [bib_ref] Live high, train low does not change the total hemoglobin mass of..., Ashenden [/bib_ref] [bib_ref] Effects of a 12-day "live high, train low" camp on reticulocyte production..., Ashenden [/bib_ref] [bib_ref] Living high-training low: Effect on erythropoiesis and maximal aerobic performance in elite..., Robach [/bib_ref] [bib_ref] Improved running economy in elite runners after 20 days of simulated moderate-altitude..., Saunders [/bib_ref]. ## Limitations, perspective and practical applications Limitations of this review include the fact that there was no control group in any of the studies. Another limitation is that not all studies in this systematic review included male and female participants. Six studies only included male runners. This may have an impact on the results, as the effects after hypoxia may be different depending on the sex of the participant. Only three studies took into account the time of the season when the hypoxia exposure took place. This is an important factor to consider in future research, as depending on the time of the season, participants may have a greater or lesser range of improvement, which may influence their athletic performance. In addition, future studies may consider the relationship between exposure to hypoxia during sleep and sports performance. Sleep parameters have been shown to be significantly altered, but it has not been demonstrated whether this has any impact on athletic performance. The strengths of this review are that it is focused on runners only, excluding other endurance athletes, such as cyclists, thus differentiating it from other studies. In addition, it includes a large number of haematological parameters and performance measures. This synthesis of information may help coaches who use hypoxia in their preparation. # Conclusions Short exposures (less than 3 h) to normobaric hypoxia significantly increase the time to exhaustion. The methodological quality of the studies that report this is 5/10 and 6/10 on the PEDro scale. This systematic review shows that studies using long periods of exposure to hypoxia (14 h or more) increased haemoglobin values, while short exposure times were ineffective. This assertion was made by several studies in this review, including the study with the highest score on the PEDro Scale (7/10). Altitudes and durations of exposure to hypoxia were highly heterogeneous in the included studies. These values will differ according to the haematological parameters to be improved. However, the most commonly used altitude is 3000 m. [fig] Figure 1: Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) diagram of the study screening process for examining the effect of intermittent normobaric hypoxia in runners. [/fig] [fig] Author: Contributions: Conceptualization, I.A.-C. and I.M.-G.-M.; methodology, I.A.-C. and I.M.-G.-M.; software, I.A.-C., I.M.-G.-M. and V.F.-L.; validation, I.A.-C., I.M.-G.-M. and V.F.-L., formal analysis, I.A.-C., I.M.-G.-M. and V.F.-L.; investigation; resources, I.A.-C. and I.M.-G.-M.; data curation, I.A.-C., I.M.-G.-M. and V.F.-L.; writing-original draft preparation, I.A.-C.; writing-I.A.-C., I.M.-G.-M. and V.F.-L.; visualization, I.A.-C., I.M.-G.-M. and V.F.-L.; supervision, I.A.-C., I.M.-G.-M. and V.F.-L.; project administration, I.M.-G.-M.; funding acquisition, I.M.-G.-M. All authors have read and agreed to the published version of the manuscript. [/fig] [table] Table 1: Methodological quality (PEDro scale). EC: Eligibility criteria; RA: Random allocation; CA: Concealed allocation; BC: Baseline comparability; BS: Blind subjects; BT: Blind therapists; BA: Blind assessors; AF: Adequate follow-up; ITTA: Intention-to-treat analysis; BGC: Between-group comparisons; PEAV: Point estimates and variability. [/table] [table] Table 2: Study characteristics. [/table] [table] Table 3: Characteristics of normobaric hypoxia. INT1: Group 1 of intervention; INT2: Group 2 of intervention; CONT: Control group; HR: Hypoxia at rest; HE: Hypoxia during exercise. [/table] [table] Table 4: Haematological parameters outcome. [/table] [table] Table 5: Sport performance outcome. [/table]

Nutrition-Related Practices and Attitudes of Kansas Skipped-Generation(s) Caregivers and Their Grandchildren Despite growing numbers, the nutrition practices and attitudes of skipped-generation(s) kinship caregivers regarding feeding the dependent children in their care have not been examined. In this qualitative study, transcriptions of semi-structured interviews with 19 female and four male skipped-generation(s) Kansas caregivers (ages 47 to 80, 92% non-Hispanic whites, 83% female, 78% grandparents and 22% great-aunt or great-grandparent caregivers; caring for a range of one to four children, ages three to 18, for an average of nine years) were content analyzed for how their nutrition-related practices and attitudes had changed since parenting the first time. Sub-themes regarding practices included: being more nutrition and food safety conscious now, and shifting their child feeding style. The children seemed to be adversely affected by an on-the-go lifestyle and the use of more electronics. Caregivers described their sources for child feeding advice as being based mostly on information from their mothers, physicians, and their past parenting experiences. Sub-themes for attitudes included opinions that nutrition and safe food handling are important and that nutritious food is expensive. They preferred printed or video nutrition education materials and wanted to receive information through organizations they trusted. This population could benefit from education on: infant, child, adolescent, and sports nutrition; feeding -picky eaters‖; healthful recipes, -fast foods‖ and packaged foods; quick, inexpensive meals and snacks low in fat, sugar, and salt; limiting sedentary time; family meals; using food thermometers; and intergenerational gardening and cooking. # Introduction The proportion of grandparents, or another older adult relative raising children, is increasing in the United States (U.S.). While many terms are used to describe this population, and we use them here interchangeably, -grandparent caregivers‖ are defined by the U.S. Census Bureau as -people who had primary responsibility for their co-resident grandchildren younger than 18‖; they numbered 2,617,580 in 2008. The lifetime prevalence of grandparents raising grandchildren is higher than that reported in any year's census data: grandparent caregiving that lasted for at least six months occurred for one in ten (10.9 percent) grandparents, based on 1992-1994 national data [bib_ref] America's grandparent caregivers: Who are they?, Fuller-Thomson [/bib_ref]. This percentage can be expected to be higher now, however, since the situation has become more common. The number of Kansan children who resided in a grandparent-headed household increased 43 percent between 1990 and 2000, which was 13 percent higher than the national average. Approximately 19,995 grandparents were responsible for their grandchild(ren) in Kansas in 2008. Not only is the trend of skipped generation households increasing, the commitment generally lasts for an extended time. Among Kansan grandparents responsible for their grandchildren, 22 percent had cared for them one to two years, while 46 percent had shouldered this responsibility for three or more years. Grandparent and other skipped-generation(s) kinship caregivers who provide parental care in the absence of the children's biological parents experience unique physical, financial, familial, social, emotional, psychological and legal challenges. The implications of these have been well-published [bib_ref] Determinants of role satisfaction among traditional and custodial grandparents, Hayslip [/bib_ref] [bib_ref] Grandparents raising their grandchildren: A review of the literature and suggestions for..., Hayslip [/bib_ref] [bib_ref] The assumption of caregiving: Grandmothers raising the children of the crack cocaine..., Roe [/bib_ref]. A collaboration of national organizations provides state-specific data and information about the range of support services, benefits, and policies that grandfamilies may need to fulfill caregiving roles. One concern for grandparent caregivers is health problems for themselves-including high rates of depression, poor self-rated health, and the frequent presence of multiple chronic health problems-and their grandchildren [bib_ref] America's grandparent caregivers: Who are they?, Fuller-Thomson [/bib_ref]. Another concern is financial hardship. Grandparent caregivers are more likely than -all families with related children under 18 years of age‖ to have incomes below the poverty level, with 19 percent meeting this criterion in 2008. Eating a nourishing diet is yet another challenge faced by many people in this population. Of all older adults with limited resources, grandparents with responsibility for their grandchildren are the most likely to have low or very low food security. Marginal food insecurity occurred in 30 percent of senior households with a grandchild but no adult child present, compared to 10 percent of households without a grandchild present. In a follow-up study, it was reportedthat, nationally, 15.4 percent of senior households with a grandchild, but no adult child, present were at-risk of hunger, or food insecure, compared to 5.2 percent of households without a grandchild present. In Kansas, the rate of food insecurity was higher than the national average, at 16.8 percent, for families with a grandchild, but no adult child, present, compared to 4.6 percent for senior households without a grandchild present. Related to these findings, many grandparents with insufficient financial resources who are raising their grandchildren were found to decrease their grandchildren's food portion sizes at times [bib_ref] Food security and health of grandparents raising grandchildren, Serrano [/bib_ref]. The interaction between generations in a household can affect the eating behaviors of both. For example, the presence of grandchildren influences grandparent caregivers' dietary intakes. The three main perceived influences on healthful eating and physical activity of 18 urban, mostly female (marital status not reported), African-American grandparent caregivers were the presence of the grandchildren, cultural foods, and financial issues [bib_ref] Grandparents raising grandchildren: A response to a nutrition and physical activity intervention, Kicklighter [/bib_ref]. Grandchildren in the home negatively influenced these grandparents from making healthful dietary changes, stemming from consideration of their grandchildren's tastes and food preferences, based on content analysis of the interviews. Also, cultural influences on diet and preferences for traditional foods, as well as the higher cost, were stated barriers to the caregivers eating more healthful diets. Similarly, familial practices and attitudes can have a significant impact on the children's dietary choices and psychosocial health [bib_ref] Parenting style and family meals: Cross-sectional and 5-year longitudinal associations, Berge [/bib_ref] [bib_ref] Demographic, familial and trait predictors of fruit and vegetable consumption by pre-school..., Cooke [/bib_ref]. For example, studies have shown that child feeding styles of parents that are either too controlling or too relaxed with child feeding can result in less healthful child dietary choices and weight status [bib_ref] Influence of mother's educational level on food parenting practices and food habits..., Vereecken [/bib_ref] [bib_ref] Development of eating behaviors among children and adolescents, Birch [/bib_ref]. In one study of multiple generation households, both parents and grandparents expressed concern over their children and grandchildren's unhealthful eating practices, and used multiple strategies to get family members to eat healthfully, but there was a lack of communication between all three generations at times concerning healthful nutrition behaviors [bib_ref] Intergenerational family conversations and decision making about eating healthfully, Kaplan [/bib_ref]. The few published studies to date regarding nutrition and grandparents raising grandchildren have focused on education. For instance, in a report advising family service workers on how to assist grandparent caregivers, one team recommended working with them to develop a list of topics they should discuss with their health care provider, including: getting information about managing the child(ren)'s weight, nutrition, and fitness level; discussing the challenges of trying to meet their grandchild's nutritional needs, and specifics related to meal planning, food budgets, and exercise options; and ensuring that their own specific health care needs were being met. Only three research reports were found, and all showed that nutrition education, provided in conjunction with overall wellness or physical activity programs, increased the nutrition knowledge of grandparents raising their grandchildren [bib_ref] Grandparents raising grandchildren: A response to a nutrition and physical activity intervention, Kicklighter [/bib_ref] [bib_ref] Promoting healthy lifestyles among grandparents raising grandchildren in Riverside County, Ganthavorn [/bib_ref] [bib_ref] A faith-based intergenerational health and wellness program, Duquin [/bib_ref]. Despite the rise in numbers of skipped-generation(s) households, the high prevalence of household food insecurity for custodial grandparents, and the interaction on eating behaviors between generations, the authors are not aware of any published research exploring nutrition-related practices or attitudes of skipped-generation(s) kinship caregivers living in the U.S. with regard to feeding their dependent children. Nutrition-related attitudes are opinions about particular eating, physical activity or food safety behaviors, and are associated with the adoption and maintenance of nutrition-related healthful habits [bib_ref] Food safety educational intervention positively influences college students' food safety attitudes, beliefs,..., Yarrow [/bib_ref] [bib_ref] The nutrition attitude survey: Associations with dietary habits, psychological and physical well-being,..., Hollis [/bib_ref]. Nutrition attitudes of individuals do not always correlate well with their reported nutrition practices, however [bib_ref] Influential factors of caregiver behavior at mealtime, Nahikian-Nelms [/bib_ref] [bib_ref] A national survey of practices and attitudes of primary-care physicians relating to..., Levine [/bib_ref]. Nutrition-related attitudes measure dietary preferences, perceptions about the role of food, and a person's views about the benefits and feasibility of adopting healthier habits [bib_ref] Influential factors of caregiver behavior at mealtime, Nahikian-Nelms [/bib_ref] [bib_ref] Nutrition knowledge and attitudes of hotel and restaurant management students, Hamm [/bib_ref]. These attitudes can be positive (such as -a low-fat diet is enjoyable‖) or negative (such as -eating healthfully is too much effort‖) [bib_ref] Influence of nutrition attitudes and motivators for eating on postpartum weight status..., Nuss [/bib_ref]. The purpose of our qualitative study was to investigate and uncover previously unarticulated nutrition-related practices and attitudes of grandparents or other skipped-generation(s) kinship caregivers and the dependent children in their care. Our goal for this report is to discuss what they perceived to be relevant issues related to adequately feeding the children for whom they were responsible, and in particular, how these nutrition-related practices and attitudes had changed from when they were parenting the first time. ## Experimental section ## Design The interviewer used a semi-structured interview guide format. In-depth, face-to-face qualitative interviewing was the method of choice for this study for several reasons: the exploratory nature of the research (guided conversations with open-ended questions elicited stories from the participants about their experiences with a variety of nutrition practices and their attitudes, in their own words), the sensitive nature of the research topic (interviews were typically conducted in their own homes), and the convenience to the caregivers (not having to write down their opinions reduced the burden associated with written surveys, and they did not have to travel anywhere or arrange for alternative child care). ## Instrument development Although the interview guide (see [fig_ref] Table 1: Main questions of the interview guide [/fig_ref] had six categories of questions, only the data from four portions are used in this report, reflecting the objectives regarding: nutrition-related attitudes, nutrition education, changes in their nutrition-related practices as second-time parents, and participant characteristics. The guide was developed by the authors and tested with a family systems specialist who was also a grandmother living with, but not responsible for, her grandchild. During instrument development, questions were checked for singularity and clarity, as recommended by Patton. ## Participants To participate in the study, caregivers had to be (a) 45 years of age or older, (b) living in Kansas, (c) able to speak English, and (d) previously a parent and now the sole or primary caregiver for their (great)grandchild(ren) or another skipped-generation(s) child relative under the age of 18 years, where neither biological parent was present. This study used a purposeful snowball sampling strategyto find information-rich key informants who met the participant criteria. Participants were recruited throughout the state through local grandparent support groups and referrals from K-State Research and Extension county faculty. Participant recruitment ended when no new sub-themes were being identified in new interviews. ## Procedures The study (number 4582) was approved on 3 March 2008 by the Kansas State University Committee on Research Involving Human Subjects. The interviewer contacted potential participants in person or via telephone to explain the study and answer questions. If they were interested, she then scheduled the in-person interview time. The face-to-face setting for each interview was selected by the participant. Participants were fully briefed and signed consent forms before the interviews. All interviews were recorded on audiotape with participants' permission. Participants were told at the onset of the interviews that they could refuse to answer any question without penalty. Regardless of the number or depth of questions answered, participants were offered a $15 per family incentive. All oral interviews were conducted, using the interview guide, during a two-month period by one master's level Registered Dietitian (second author, BJM), who had completed coursework and self-study in qualitative interviewing and research. The main questions were primarily open-ended. The protocol included the use of clarifying questions and follow-up probes designed to elicit details. # Data analysis Interviews were transcribed verbatim, using a word processor. The interviewer listening to the tapes while comparing them to the transcripts verified accuracy of the transcripts. Data analysis was carried out manually to find themes and sub-themes, both within each interview and between interviews. Principles described by Pattonguided the data analysis. All transcriptions were read several times to become fully familiar with the data. Transcribed quotes considered to represent the same concept were clustered together categorically, according to the researchers' primary questions within the interview guide and additional unanticipated categories. The categorized quotes were then coded according to key phrases. We looked for patterns and repetition of key phrases, which we then sorted and re-sorted according to content similarities, to allow the -emergence‖ of sub-themes of data within our main predetermined thematic scheme in line with the interview guide topics. Peer debriefing was used throughout the study for increased credibility. Descriptive statistics were obtained for demographic variables and screen time using Microsoft Excel (Microsoft Office, 2003). # Results ## Participants Participant characteristics are shown in [fig_ref] Table 2: Participant characteristics [/fig_ref]. All caregivers appeared to be comfortable discussing the questions, based on their responses and body language; interviews averaged one hour but ranged from a half hour to two hours in length. Twenty-three caregivers representing 19 households and living in 17 counties across Kansas were interviewed (four male and 19 female). Caregivers ranged in age from 47 to 80 years, with a mean of 62 years. The majority of the caregivers were married (70%) non-Hispanic whites (92%). Length of care ranged from less than a year to 18 years, with a mean of nine years. The majority (78%) of those interviewed were grandparent caregivers. The remaining 22% were either great-aunt or great-grandparent caregivers. Thirty-seven percent of the households reported that they currently received some form of governmental monetary assistance. Only 11% of households lived in urban areas (>150 residents per square mile, or rpsm), with the remainder living in very rural or frontier areas of Kansas (37% in frontier areas, <6.0 rpsm; 26% in rural areas, 6.0-19.9 rpsm; and 26% in densely-settled rural areas, 20.0-39.9 rpsm). Their dependent child relatives (whom we will also refer to as -grandchildren‖) were primarily non-Hispanic whites (56%) or of mixed descent (32%) and ranged in age from three to 18 years, with a mean of 12 years. Thirty-two percent were female. The average number of skipped-generation(s) children for which the caregivers were primarily responsible was one child, with a range from one to four children. The circumstances leading to the skipped-generation(s) care stemmed mostly from deemed parental inability because of child abuse or neglect, illegal activity, teen pregnancy, or mental health issues. Only three households (16%) reported they were providing care because of parental death. ## Themes The nutrition-related findings from this study of skipped-generation(s) households were divided into two over-arching themes reflecting the objectives of the study: nutrition-related practices and nutrition-related attitudes. Several sub-themes emerged from each theme, as illustrated with quotes that convey how the participants described their practices and attitudes. Caregivers had a wide variety of nutrition-related practices and attitudes. # Nutrition-related practices Within the nutrition-related practices theme, five sub-themes were identified from the interview data. Compared to when they were parenting the first time, skipped-generation(s) caregivers overall reported that they are more nutrition and food safety conscious and that their grandchildren are adversely affected by an on-the-go lifestyle and the use of more electronics. Caregivers have shifted their child feeding style. Their sources of child feeding advice are based mostly on information from their mothers, physicians, and their past parenting experiences. ## More nutrition and food safety conscious Caregivers described practices that showed they are more nutrition conscious with raising their second generation of children than their first. The most common of these practices included providing a more nutritious variety of foods in the diet, including increasing servings of fruits, vegetables, and milk. Children in the majority of households ate diets that included a variety of foods from all food groups. One food that I use more of with [my grandson] than I did with my kids: broccoli and cauliflower and more of the exotic fruits…. We're more apt to try kiwi fruit or try more things. Sometimes [he] will see something and ask about it, and if we can afford it, we'll get one and try it. (grandmother [bib_ref] Food security and health of grandparents raising grandchildren, Serrano [/bib_ref] In addition, seven households reported cooking foods from -scratch‖ more often now rather than from packages. Five households reported no change in this aspect of cooking. While five households described using more packaged foods now than with their first set of children, nevertheless all but a few of the caregivers who cook with more convenience foods also described eating more healthfully now than with their first set of children. When we were first starting out, money was tighter, so when we shopped it was always the bargains. And sometimes the bargains weren't always the healthiest choices…. [I use more packaged food now,] yes, it's quicker. I used to cook pretty much from scratch. We do eat a lot of vegetables…. It was a bad habit to bake, but we don't need all of that, so I quit. And we always have a dessert of some sort, but mostly fruit. (grandmother 3) Participants also described reading the Nutrition Facts labels on packaged products: 13 households did and five did not use them. Fat, sugar, and sodium content were the items on the label they looked at most often. In addition to their nutrition-related behaviors, participants described increasing many practices with food safety in mind. They mentioned a wide variety of actions, including checking expiration dates on food or leftovers, paying attention to food recalls, refrigerating or freezing perishable foods promptly after shopping and after eating meals, checking refrigerator temperatures, washing their hands, washing fresh produce, and keeping food areas clean during preparation. Although six of the households reported having a food thermometer, only one caregiver reported using it when cooking. We do the expiration date thing…. We buy food that doesn't sit for long periods of time. That way it keeps it from going bad or not as nutritional further down the road. (grandmother 1) Well, I make sure that the refrigerator's on the right temperature. I make sure things are covered and put in the refrigerator. I make sure that the meat is cooked. I'm very conscious about that. I make sure that the raw vegetables are washed. I'm pretty conscious about food safety. And I teach the kids that if you want some fruit, you wash it first. (grandmother 7) Caregivers commonly attributed their improved awareness to having: more time to plan meals and cook, more information about nutrition and food safety because it was more available and emphasized more in modern society, and, in some cases, more financial resources. Although caregivers reported many healthful changes in raising their second set of children, such as planning nutritious meals, these skipped-generation(s) households as a whole reported some less healthful changes because of the shift in society toward a busy, on-the-go lifestyle. Even though the caregivers and their grandchildren were more aware of nutrition, that knowledge at times did not result in the most healthful food selections. Some grandparents noted specific occasions for the increased availability and consumption of packaged products, pizza, -junk‖ foods, and -fast food‖ by the children in their care. Every once in a while I try to tell him [my grandson] he's eating too much junk but it doesn't do no good…. I think he eats too much junk food, like McDonald's every day…. ‗Cause he eats lunch out. The boys [my sons] didn't have any place to go, they had to eat at school, and our daughter had to eat lunch there. We didn't have McDonalds or Sonic or any of those things. (grandfather 9) We probably do more frozen pizzas occasionally ‗cause she [my great-niece] likes to get them. I do make them from scratch sometimes, too. If they have them on sale, she'll say, -Let's get a pizza.‖ And they're nice too, because you can pull them out and they're fast. (great-aunt 23) Some caregivers attributed their grandchildren's occasional poor food choices to having less time to eat healthfully and needing to eat something quick for breakfast or -fast food‖ when they were away from home. Fourteen of the 19 households had at least one child involved in team sports, which took a lot of time and was often coupled with selecting quick, less healthful foods. The households with teenage boys with driver's licenses seemed to eat the least healthfully. Because of the irregularity of these boys' schedules, they had an almost daily consumption of -fast foods‖ and high amounts of -junk‖ snack foods. They [our grandsons] really don't want to sit down for a good meal. All they want is junk food. Our kids didn't really eat junk food. It seems like they're [our grandsons] not here half the time. Caregivers in four households specifically mentioned that they used to grow their own vegetables, and some even produced their own meat, eggs, milk, and fruits, but now they buy them. We had a big garden… so I didn't buy a lot of stuff. I canned a lot. (grandmother 3) ## More electronics increase sedentary activities and purchases of advertised foods Caregivers noted the increase between generations in children's use of electronics, from hand-held video games to computers. In discussing the child(ren)'s favorite things to do, everyone described screen time as one of these activities. Responses regarding how much time was spent ranged from one caregiver reporting -very seldom‖, to several saying -a lot‖. Six households were unable to estimate specific amounts of time, and these were evenly split between -lots‖ and -a little‖. Thirteen households gave estimates of the specific range of time that the children spent with computers, videos, and television: one to four hours each day. The authors averaged these responses for the group, and concluded that the children spent at least 2.5 hours a day with electronics. Several caregivers described actively monitoring and limiting time spent, while others spoke of their grandchildren spending -too much time‖ in sedentary activities: watching television, playing video games, and playing on the computer. In addition, the use of electronics was described as affecting the types of food purchased in some households. After the youngsters watched advertisements, such as on television, they sometimes asked their caregivers to purchase those specific foods. All caregivers reported shifts in their child feeding styles that affected their child feeding practices. The next generation of children being raised by these caregivers tended to be -pickier eaters‖ compared to their earlier counterparts. Two different types of shifts in child feeding style were described by participants. More than half (11 households) shifted to being more relaxed and indulgent with the second generation. For example, these caregivers reported catering to the food preferences of their grandchildren, and a few struggled not to -spoil‖ the grandchildren by allowing too many -junk‖ foods. Another common practice noted was a shift from family meals together at the The second child feeding style shift was characterized as being more involved in feeding. The eight households in this group tended to be more protective of their second group of children, including more carefully monitoring child feeding. This group shifted to more scheduled meals at the table as a family, and they were more nutrition-conscious caregivers. But I'm more conscious with these kids than I was with my own kids, about nutrition and about everything else. I'm more cautious. What they eat, are they getting what they need? My kids had more snack foods than these kids do…. I have more time to plan meals, to prepare them…. Now we all have dinner together…. Sometimes they [my granddaughters] wouldn't even eat if one of us didn't say, -You will eat what's on your plate.‖… And I do see that they eat better than what my kids were allowed to. When my kids were growing up, if they were hungry when they came home from school, they'd grab something, and if they weren't hungry for dinner, no big deal. (grandmother 7) I think [I do] a little better [with feeding my grandson] than with the first group…. But I've learned a lot since then. I've learned so much more. (grandmother 17) ## Sources of child feeding advice The caregivers stated that over the years, their primary source of child feeding advice was information passed from the previous generation to the caregivers, particularly by their mothers. Additionally, they identified their family doctor or pediatrician as a main source of feeding advice when their first children were born or adopted. Many commented that there was much less emphasis on child feeding by society with their first set of children compared to their second set. When asked if they sought additional sources of nutritional advice for parenting the second time, most described relying primarily on their past parenting experiences, although a few mentioned governmental programs such as the Health Department, Cooperative Extension System, or the Women, Infants, and Children (WIC) program. My mom… she was the best mom in the world. (laughs) And she was such a good cook. # Nutrition-related attitudes Within the nutrition-related attitudes theme, three sub-themes were identified: the caregivers held opinions that nutrition and safe food handling were important; that nutritious food is expensive; and that population-specific nutrition education materials may or may not be helpful either to them or to others in their situation. ## Nutrition and safe food handling Attitudes toward nutrition and child feeding shifted, from simply serving meals in order to obtain energy to wanting to serve and eat a balanced diet now. All but two participants expressed at some point during their interviews their belief in the importance of giving this second generation a good start so that they would form life-long nutritious eating habits. Some expressed concerns about the recent rise in childhood obesity and other chronic diseases, such as diabetes, or mentioned having chronic conditions themselves, including diabetes and heart conditions, which made them more aware of the role of nutrition in health. Four caregivers used to view a -good‖ meal as one that included meat and potatoes, but now have the opinion that incorporating more variety is important, especially including more fruits and vegetables. Three female caregivers mentioned that this -meat and potatoes‖ attitude came primarily from their husbands. But trying to give them [my granddaughters] as good a start as possible nutritionally, and learning how to have pop [soft drinks] and stuff like that in moderation. You can have them, but you don't need them all the time. But we try to encourage fruits and vegetables, and try to get them on a track of good nutrition, so that possibly when they have their own kids, they'll do the same thing. And maybe that will extend to a healthy lifestyle. (grandfather 2) Caregivers believed that it was important to serve less fat, sugar, and salt in their households' diets. To do this, most thought that looking at food package labels was helpful, while others believed that it was unnecessary for them to read labels because they -already had a good feel for the nutritional value‖ of the products they buy. Of course, I look for sugar content. I will not buy anything that has hydrogenated oils. I say I won't buy-I will occasionally, but not on a regular basis. And if it has, it won't be up on the second and third . I feel that that's where a lot of our high cholesterol comes from and our health problems. And I do my best to stay away from that as my main things. The two main concerns the caregivers expressed about their grandchildren's eating habits were that the children: were not eating enough food overall; or were eating too much -fast food‖, sweets, sports drinks, energy drinks, high fat snacks, and other -junk‖ foods. He's [my grandson] so small and I want him to eat more than he eats, but he doesn't eat a lot at one time. Now sometimes he'll get into the junk food and he'll eat a lot of unnutritious food at one time. Regarding attitudes toward safe food handling, the caregivers were of the opinion that storing perishable food in a refrigerator or freezer is important, and also stressed the importance of many other practices. In contrast to their stated opinions that cooking foods properly is important to them, however, only one household reported checking internal temperatures of cooked foods on a regular basis. Rather, most believed that they could adequately tell whether meat was cooked long enough by the way it looked or smelled. I wash vegetables, that's for sure! I used to use the foam, but I'm kind of scared. The vegetables that you get around here, like spinach, says it's washed, but I usually wash it again. And chicken I let soak in salt water. I am too particular sometimes. (grandmother 1) ‗Cause back when I was growing up and when my kids were small, if you had leftovers, you should let them sit and cool. You shouldn't put hot stuff in the refrigerator. And I've learned since then that you don't do that…. I always did what my folks did…. I didn't use to think about it, but I do now. (grandmother 10) I make sure that my meat is cooked properly, just make sure it's done.… I had a meat thermometer, but I don't use that…. I just know if it's done if the chicken doesn't have blood against the bone. It's just something you pick up when you've cooked for so many years. (grandmother 8) ## Food and economics A commonly held food attitude among the caregivers was that eating nutritiously is expensive, but worth it. Most described making monetary sacrifices to eat more healthfully now. Caregivers found ways to eat nourishing foods while minimizing their food expenses. And economically, now days, in order to buy food that actually fills the kids up, you can't afford good stuff, because it's expensive to eat healthy. It's very, very expensive to eat healthy. And that's one of the main sacrifices that we've had to deal with, is groceries have become so much more expensive. Because we're paying a lot more attention to what we're buying. (grandfather 2) Especially with the fruit. If it's grapes and they're three dollars, I'll say, -We can't get it today, . We'll wait until it goes on sale.‖ And then I'll go get a can of fruit cocktail or peaches or pears, because I always know that I can get those cheaper. Price influenced what foods were bought and where they shopped for 16 of the households, while the remaining three did not mention price as being a factor in their food shopping decisions. As discussed in the section on caregivers being more nutrition and food safety conscious, some households had more financial resources for food purchases now, and explained that they could eat more healthfully with this generation of children because they had enough money to purchase foods, including more fresh fruits and vegetables. Nevertheless, even most of these caregivers valued keeping their budget in mind while shopping for food. These boys [my great-nephews] are probably being fed better than ours were, because just the difference in our jobs and what we could afford to buy. (great-aunt 16) ## Population-specific nutrition education materials Attitudes diverged regarding whether population-specific nutrition education materials would be helpful. However, positive opinions about the usefulness of such materials tailored to skipped-generation(s) caregivers, either to them or to others in their situation, were held by most. Caregivers who were open to the idea of nutrition education stated that the best ways to receive such information would be via: written means, such as brochures and newsletters; videos; having their grandchildren taught about it at school; and organizations, such as local grandparent support groups, Cooperative Extension System offices, and WIC clinics. They did not consider classes to be the best way to share nutrition education. Their suggestions for topics for educational materials focused on nutrition as it relates to children, adolescents, and sports; and on healthful recipes and snack ideas. I would say more education on what makes you healthy on the outside, ‗cause they're not going to care at this age what makes you healthy on the inside. (grandmother 4) Maybe not a class but a group that you could ask if there's a way to do it better or anybody got any suggestions of what would work if it's not working for me. Just to know that -Sally Smith‖ down the street is doing the same thing, who's raising her grandkids, so maybe I am doing something right. Maybe if you're raising your grandkids, you don't realize what's out there that wasn't available when you were raising your own kids. (grandmother 11) I think just a refresher on how many times a week they [children] should get fruits and vegetables, and maybe some alternative foods for snacks that are not high in sugar and are better for them. I think that you would get more people to read something like that than to attend a class. Just to be able to do that when a person had time to sit down and read it or whatever. (great-aunt [bib_ref] Development of eating behaviors among children and adolescents, Birch [/bib_ref] Four of the caregivers had negative attitudes toward nutrition education materials for themselves and others in their situation. They expressed beliefs that most grandparents are going to cook what's healthy, already have enough information, and would not be willing to change behaviors at this stage in their lives. Older age did not seem to be an indicator of whether the caregivers had negative attitudes. Mostly grandparents know by now how they're going to buy food. At my age, there's not a lot I haven't learned about food because I've been there and done that. (grandmother 3, age 68 years) I've been involved with my kids, my grandkids, and now my great-grandkids. But things have changed down through the years. Every ten years is different. And I think that if you don't have some guidelines, you can't keep up. (great-grandmother 14, age 80 years) # Discussion The participants in our study were married at the same percentage as the national grandparent caregiver population. Our population was a little older, more economically disadvantaged, more rural/frontier, made up of more females and more non-Hispanic whites, and had cared for the grandchildren longer than national and Kansas grandparent caregiver percentages. The caregivers we interviewed were eager to share their experiences, and the information richness of their stories provided great insight into the caregiving aspects of their lives. The circumstances begetting full-time caregiving, the non-nutrition-related challenges, and perceived rewards described by our participants, as reported elsewhere, were consistent with the findings of others [bib_ref] America's grandparent caregivers: Who are they?, Fuller-Thomson [/bib_ref] [bib_ref] The assumption of caregiving: Grandmothers raising the children of the crack cocaine..., Roe [/bib_ref] [bib_ref] Grandparent education, Chenoweth [/bib_ref]. Caregivers in the current study described practices resulting from being more nutrition and food safety conscious with their second group of children. Making more healthful food choices was explicitly related to the presence of the children, at least in some households. One of the largest barriers to eating a nutritious diet for older adults is not having social contact [bib_ref] Barriers to the consumption of fruits and vegetables among older adults, Sahyoun [/bib_ref] , but this problem would not apply to caregivers in skipped-generation(s) households. Our findings of improved nutrition attitudes and practices might not be entirely related to the presence of grandchildren, however, because previous research has shown that older adults tend to make different food choices as they age [bib_ref] Impact of aging on eating behaviors, food choices, nutrition, and health status, Drewnowski [/bib_ref] , including consuming higher amounts of fruits and vegetables than their younger counterparts [bib_ref] Barriers to the consumption of fruits and vegetables among older adults, Sahyoun [/bib_ref]. Also, because many of the caregivers we interviewed did not work outside the home, they were more available to monitor the children's food choices and environments, which also might have led to more healthful food and activity choices [bib_ref] What kids say they do and what parents think kids are doing:..., Moag-Stahlberg [/bib_ref]. The caregivers in our study highlighted an on-the-go lifestyle and children's use of more electronics as adversely affecting the children. Nutrition challenges resulting from an on-the-go lifestyle, such as occasions of increased consumption of packaged foods and -fast food‖ described by some of our participants, have been reported in a previous study of grandparent caregivers [bib_ref] Intergenerational family conversations and decision making about eating healthfully, Kaplan [/bib_ref]. The caregivers in our study were rightly concerned about the increased time spent using electronics, and thus more sedentary activity, for their grandchildren, because eight of the households in our study reported in excess of the American Academy of Pediatrics' recommendation of no more than one to two hours of quality television programming per dayand because of the previously-reported positive correlation between screen time and childhood obesity [bib_ref] Children's food consumption during television viewing, Matheson [/bib_ref]. Likewise, our caregivers' perceptions of the influence of televised food advertisements on their grandchildren's food requests are supported by previous research. Our finding that caregivers shifted their child feeding style to be either more controlling or more relaxed might have important health implications for the children. Research points to the importance of balanced familial control of child food choices. Children who grow up in households in which food choices are too controlled or restrictive tend to overeat the restricted foods in excess, even when they are not hungry, when able to do so [bib_ref] Influential factors of caregiver behavior at mealtime, Nahikian-Nelms [/bib_ref] [bib_ref] Parents' restrictive feeding practices are associated with young girls' negative self-evaluation of..., Fisher [/bib_ref]. Similarly, children who grow up in households in which food choice is unregulated tend to be more likely to consume large amounts of soft drinks and sweets, and fewer vegetables [bib_ref] Influence of mother's educational level on food parenting practices and food habits..., Vereecken [/bib_ref]. Skipped-generation(s) caregivers in the current study described relying mostly on information from their mothers, i.e., -tradition‖, their physicians, and their past parenting experiences for nutritional advice when they were raising their second set of children. In contrast, primary sources of child feeding advice for new parents may also be tradition, family, or a doctor [bib_ref] Prevention of childhood obesity: Sociocultural and familial factors, Bruss [/bib_ref] or they may actively seek current infant and child feeding recommendations [bib_ref] Online pediatric information seeking among mothers of young children: Results from a..., Bernhardt [/bib_ref]. Our caregivers may have believed that they had adequate infant and child feeding information, since they had already raised a child(ren), not realizing that while some nutrition-related -expert recommendations‖ have remained constant over time, others have changed dramatically. In addition, although many stated that they were relying on past parenting experiences, they also described practicing newer behaviors, such as checking Nutrition Facts labels. The seeming discrepancy can be explained if traditional sources of information are seen as a way of conveying either old or new information while using informal channels of communication, for example, via television, magazine stories, and social networks. With their newest generation of children, our participants reported that their nutrition-related attitudes had improved, along with their practices. For example, the higher value that they now place on eating healthfully led them to change their food selection behaviors, including serving a wider variety of foods and, especially, more fruits and vegetables. Remarkably, many described cooking more foods from -scratch‖ for their grandchildren, despite the increased numbers of packaged and convenience foods in grocery stores now. Our findings that caregivers found the fat, sugar, and sodium information on the Nutrition Facts labels to be the most helpful agree with findings from the 2007 Food and Health Survey of adults of all ages. Compared to younger consumers, older adults are less likely to change their shopping behaviors based on Nutrition Facts labels; however, nutrition labels have a larger effect on older adults' nutrition attitudes toward the products if they know more about nutrition and understand the labels [bib_ref] Age, product nutrition, and label format effects on consumer perceptions and product..., Burton [/bib_ref]. Thus, as their nutrition knowledge increases, and Nutrition Facts labels become more familiar to the older population, these labels might begin to affect older consumers' purchasing practices, as implicated in our findings. Our participants also described having both improved attitudes and practices towards handling food safely, compared to their opinions on this issue with their first set of children. For example, they now believe that refrigerating leftovers immediately after meals, instead of letting them cool on the counter, is important. From a historical perspective, -expert recommendations‖ for consumer practices have changed drastically over the caregivers' lifetimes, as a result of many factors, and these likely have helped increase the caregivers' awareness of the importance, and relevance, of home food safety practices. For instance, the food safety attitudes of older adults may be influenced by whether they subscribe to the -germ theory‖ of disease. Since the theory had not become popular in the U.S. until as late as 1920, many of our caregivers likely grew up without benefit of such microbiological information. A second example of dramatic change related to safe food handling practices that these caregivers have experienced is that the oldest participants in this study were at least 10 years old before any kind of electrical service began in Kansas. Because the attitudes of older adults toward food safety strongly correlate with their prevention practices [bib_ref] Use of the health belief model to examine older adults' food-handling behaviors, Hanson [/bib_ref] [bib_ref] Consumer knowledge of foodborne microbial hazards and food-handling practices, Altekruse [/bib_ref] , our study's findings regarding attitudes about which food handling practices were the most important have implications for the practices of the caregivers, and can inform future food safety education. For example, the commonly held opinion that meat can be cooked safely without checking its internal temperature could be used as the basis for population-specific education. We found that economic issues, and attitudes that healthful foods are expensive, strongly influenced our caregivers' food shopping behaviors. This finding is likely linked to the high (37 percent) number of the households in our study who were receiving monetary assistance that could help purchase food. In contrast, many of the monetary and food assistance programs that can assist these low-income households, such as medical assistance, emergency food providers, free and reduced-price school meal programs, the WIC program, and the Supplemental Nutrition Assistance Program, may not be familiar to eligible skipped-generation(s) caregivers. Our participants, who had fewer financial resources than the national average, yet who valued good nutrition, described buying less costly healthful foods, such as canned fruit instead of fresh. Attitudes that healthful foods are expensive have been reported for other caregiver studies [bib_ref] Grandparents raising grandchildren: A response to a nutrition and physical activity intervention, Kicklighter [/bib_ref] [bib_ref] Intergenerational family conversations and decision making about eating healthfully, Kaplan [/bib_ref]. Nutrition educators are advised to be mindful of both real and perceived financial barriers to making healthful dietary choices. In particular, because the attitude that -good food is expensive‖ is a recurring theme in older adult research, including international caregiver research [bib_ref] Influence of grandparents on eating behaviors of young children in Chinese three-generation..., Jingxiong [/bib_ref] , educators should stress the viability of making meals that are low in cost yet high in health benefits with this population. Nutrition education should result in behavior change, and by design, should be tailored to the audience it is intended for based on their assessed needs and interests [bib_ref] Tailoring nutrition education intervention programs to meet needs and interests of older..., Higgins [/bib_ref]. Caregivers in this study discussed wanting educational materials that focused on nutrition as it relates to children, adolescents, and sports; and on healthful recipes and snack ideas. In addition, based on their reported practices and attitudes compared to current expert nutrition recommendations, nutrition education for this population should include quick and inexpensive healthful meals that are low in fat, sugar, and salt; healthful -fast food‖ and packaged food options; the importance of checking the internal temperatures of meat when cooking; infant and child feeding; ways to feed -picky eaters‖; benefits of eating together as a family; tips to help children learn to cook and to limit their sedentary time; and intergenerational gardening and cooking. Cooking and eating inexpensive yet healthful meals as a family, along with being physically active, such as gardening together, would offer skipped-generation(s) household members daily opportunities to benefit in some areas where they may be experiencing difficulties. For example, eating and gardening with children are associated with frequent and uncontrived chances for relaxed communication and emotional connections with each other; a boost in decision-making skills, confidence and self-esteem; improved math, science, and language skills and general academic achievement of children; decreased likelihood of risky behaviors by the younger generation; and overall more positive familial and other social relationships [bib_ref] Kids jump into community gardening: What bounces out?, Krasny [/bib_ref]. Thus, providing caregivers in skipped-generation(s) households with comprehensive nutrition education would be expected to lead them to be healthier older adults and more effective second-time-around parents, and at the same time, help to ensure a healthier future for the children being cared for by this group. The caregivers in this study who were open to the idea of nutrition education were of the opinion that the best ways to receive information were via: written means, such as brochures and newsletters; videos; having their grandchildren taught about it at school; and organizations they trust, such as those we recruited from (local grandparent support groups and Cooperative Extension System offices) and WIC clinics. Published literature confirms that print and video sources were also preferred by other groups of older adults [bib_ref] Improving effectiveness of nutrition education resources for older adults, Higgins [/bib_ref] , including grandparents raising grandchildren [bib_ref] Grandparent education, Chenoweth [/bib_ref] , with printed newsletters being inexpensive yet effective. Support groups as one of the preferred ways to receive informational help and to relieve stress is congruent with findings from other skipped-generation caregiver studies [bib_ref] Grandparents raising their grandchildren: A review of the literature and suggestions for..., Hayslip [/bib_ref] [bib_ref] Grandparents raising grandchildren: A response to a nutrition and physical activity intervention, Kicklighter [/bib_ref] [bib_ref] Grandparents raising grandchildren: Stressors, social support, and health outcomes, Leder [/bib_ref]. Our research focused exclusively on skipped-generation(s) caregivers providing full-time surrogate care to dependent children in the absence of their parents. The results, however, also have implications for providing nutrition education to older adults who either regularly or occasionally provide supplemental care to the younger generation. For instance, many grandparents provide day care for their grandchildren, and, as such, are responsible for providing most of the meals and snacks eaten by the infants, children and teens under their supervision. The same may be true for older adults who make their living as paid day care providers. # Limitations This study was designed to explore and describe a broad scope of nutrition-related practices and attitudes, which limited the study depth into any one area. The sample in this study was not demographically representative of the population of grandparent caregivers as a whole, in racial/ethnic groups, gender, or residence. Additionally, a generally positive response bias is associated with self-selected community samples of middle-aged and older adults [bib_ref] Determinants of role satisfaction among traditional and custodial grandparents, Hayslip [/bib_ref]. While we used no objective measures for participants' truthfulness, they seemed to speak from their hearts and to believe strongly in what they stated; we have no reason to think that they falsified any descriptions shared with the interviewer. We did not interview enough people to learn about all situations, practices, and attitudes; a point of data -saturation‖ likely was not reached. Our findings should be considered preliminary work leading to more research in this area of study, and not be generalized. # Conclusions Nutrition is just one of the often-complicated challenges experienced by members of skipped-generation(s) households, although it is an important variable in health. This study provides beginning insight into the practices and attitudes of skipped-generation(s) kinship caregivers in their role as nutrition providers for the dependent children in their care. Based on content analysis of personal interviews, caregivers described being more nutrition and food safety conscious compared to when they were parenting the first time. Five changes in their practices included: serving a more nutritious variety of foods, reading Nutrition Facts labels, doing more cooking, storing foods properly, and keeping food preparation areas clean. The caregivers mentioned new challenges to their dependent children eating nutritionally-balanced meals because of shifts toward an on-the-go lifestyle. An increased use of electronics by the children increased their sedentary activity and family purchases of advertised foods. Many caregivers noted shifts in their child feeding styles, which influenced their child feeding practices. The two types of shifts were being more relaxed and indulgent with the second generation, or being more involved in feeding now. Our participants credited their child feeding knowledge primarily to information from their mothers, a doctor, or, with the newest generation, to their past experiences. Caregivers had an improved attitude toward nutrition and safe food handling, compared to the regard with which they held these practices with their first generation of children. They perceived economic issues as a challenge to selecting healthful diets. Participants recommended that if nutrition education materials were developed for this population, they should be distributed primarily using printed or video materials and wanted to receive information through organizations they trusted. Based on the results, in juxtaposition with current expert nutrition recommendations, the authors conclude that this population could benefit from education on nutrition as it relates to infant, child, adolescent, and sports nutrition; feeding -picky eaters‖; healthful recipes, -fast foods‖ and packaged foods; quick, inexpensive meals and snacks low in fat, sugar, and salt; limiting sedentary time; family meals; using food thermometers; and intergenerational gardening and cooking. The findings have implications for nutrition educators and other health professionals, who can help older skipped-generation(s) caregivers be healthier older adults and more effective parents, and help ensure a healthier future for the dependent children in their care. [fig] I: 've had to change the way think about food. And every once in a while, get hungry for good ole' comfort food. t's hard to give it up. think in this day and time, if kids grow up eating healthier, they won't have to go through that. (grandmother 12) think with a child in the house, you're more aware of what you're eating and serving. And think it's probably going to benefit [my husband] and in the long run, because we have the desire to be more healthier and to be around for her [my granddaughter] longer. (grandmother 20) [/fig] [table] Table 1: Main questions of the interview guide.Can you briefly describe the children for whom you are primarily responsible? What advice would you offer to other grandparent caregivers? What have been the biggest challenges to you as a grandparent caregiver? Briefly, what were the circumstances behind your becoming the primary caregiver of the child(ren)?What were the most helpful sources of nutrition advice when you parented the first time? What are the most helpful sources of nutrition advice now that you are parenting for the second time? What kinds of nutrition education have been helpful in the past or are currently helpful? Are there any nutrition-related topics you would have liked to have had information on in the past? Are there any nutrition-related topics you would like information on now or for the future? [/table] [table] Table 2: Participant characteristics. [/table]

Associations between trajectories of obesity prevalence in English primary school children and the UK soft drinks industry levy: An interrupted time series analysis of surveillance data BackgroundAU : Pleaseconfirmthatallheadinglevelsarerepresentedcorrectly: points (PPs) (95% confidence interval (CI): 1.1, 2.1), with greatest reductions in the two most deprived quintiles (e.g., there was an absolute reduction of 2.4 PP (95% CI: 1.6, 3.2) in prevalence of obesity in the most deprived quintile). In year 6 boys, there was no change in obesity prevalence, except in the least deprived quintile where there was a 1.6-PP (95% CI: 0.7, 2.5) absolute increase. In reception children, relative to the counterfactual, there were no overall changes in obesity prevalence in boys (0.5 PP (95% CI: 1.0, −0.1)) or girls (0.2 PP (95% CI: 0.8, −0.3)). This study is limited by use of index of multiple deprivation of the school attended to assess individual socioeconomic disadvantage. ITS analyses are vulnerable to unidentified cointerventions and time-varying confounding, neither of which we can rule out. # Conclusions Our results suggest that the SDIL was associated with decreased prevalence of obesity in year 6 girls, with the greatest differences in those living in the most deprived areas. Additional strategies beyond SSB taxation will be needed to reduce obesity prevalence overall, and particularly in older boys and younger children. ## Trial registration ## Isrctn18042742. Author summary Why was this study done? - In England, childhood obesity rates are high with around 10% of reception age children (4/5 years) and 20% of children in year 6 (10/11 years) recorded as living with obesity in 2020. - Children who are obese are more likely to suffer from serious health problems including high blood pressure, type 2 diabetes, and depression in childhood and in later life. - In March 2016, to tackle childhood obesity, the UK government announced there would be a soft drinks industry levy (SDIL) on manufacturers of soft drinks to incentivize them to reduce the sugar content of drinks. ## What did the researchers do and find? - We tracked changes in the levels of obesity in children in England from reception (ages 4/5 years) and year 6 (ages 10/11 years) over time between 2014 and 2020. This analysis involved comparing obesity levels 19 months following the SDIL with predicted obesity levels had the SDIL not happened according to gender of the child and school's area level of deprivation. - The UK SDIL was associated with an 8% relative reduction in obesity levels in girls aged 10/ 11 years, equivalent to prevention of 5,234 cases of obesity per year in girls aged 10/11 years, alone. Reductions were greatest in girls whose school was in the 40% most deprived areas. - No associations were found between the SDIL and changes in obesity levels in boys aged 10/11 years or younger children aged 4/5. # Introduction There is strong evidence that consumption of sugar-sweetened beverages (SSBs) increases the risk of serious diseases including type 2 diabetes, cardiovascular disease, dental caries, and obesity [bib_ref] The role of sugar-sweetened beverages in the global epidemics of obesity and..., Malik [/bib_ref] [bib_ref] Effect of sugar-sweetened beverages on oral health: a systematic review and meta-analysis, Valenzuela [/bib_ref]. Children and adolescents in the United Kingdom are particularly high consumers of added sugars [bib_ref] Changes in consumption of sugars by English adolescents over 20 years, Rugg-Gunn [/bib_ref] with consumption typically peaking at approximately 70 g/day in late adolescence, equivalent to over twice the recommended maximum intake of 30 g [bib_ref] What's on the Menu? Policies to Reduce Young People's Sugar Consumption, Griffith [/bib_ref]. SSBs are the primary source of free sugar in the diets of children and are associated with weight gain, obesity, and fatness in children [bib_ref] Sugar-sweetened beverages and weight gain in children and adults: a systematic review..., Malik [/bib_ref] [bib_ref] Effects of Soft Drink Consumption on Nutrition and Health: A Systematic Review..., Vartanian [/bib_ref] [bib_ref] Relation between consumption of sugar-sweetened drinks and childhood obesity: a prospective, observational..., Ludwig [/bib_ref]. Demographic patterns of SSB and added sugar consumption mirror each other with highest consumption in older children [bib_ref] What's on the Menu? Policies to Reduce Young People's Sugar Consumption, Griffith [/bib_ref] [bib_ref] Sugar-sweetened beverage consumption from 1998-2017: Findings from the health behaviour in school-aged..., Morgan [/bib_ref] , boys [bib_ref] Sugar-sweetened beverage consumption from 1998-2017: Findings from the health behaviour in school-aged..., Morgan [/bib_ref] [bib_ref] Sugar consumption among Canadians of all ages, Langlois [/bib_ref] , and children from lower socioeconomic groups [bib_ref] Consumption frequency of added sugars and UK children's dental caries, Hong [/bib_ref] [bib_ref] Sugars consumption in a low-income sample of British young people and adults, Ntouva [/bib_ref] [bib_ref] Sugar-sweetened beverage consumption from 1998-2017: Findings from the health behaviour in school-aged..., Morgan [/bib_ref]. Recently born cohorts of children are much more likely to have obesity than children from older cohorts such that 10-year-olds born after the 1980s are 2 to 3 times more likely to develop obesity than those born before the 1980s [bib_ref] How Has the Age-Related Process of Overweight or Obesity Development Changed over..., Johnson [/bib_ref]. The persistence of obesity from childhood into adulthood [bib_ref] Predicting adult obesity from childhood obesity: A systematic review and meta-analysis, Simmonds [/bib_ref] and its acute and chronic negative physical [bib_ref] Childhood obesity and its physical and psychological co-morbidities: a systematic review of..., Sanders [/bib_ref] [bib_ref] Being overweight or obese and the development of asthma, Lang [/bib_ref] [bib_ref] Musculoskeletal effects of obesity, Chan [/bib_ref] and mental [bib_ref] Childhood obesity and its physical and psychological co-morbidities: a systematic review of..., Sanders [/bib_ref] [bib_ref] Pediatric Obesity/Obesity Comorbidity Exploring the association between childhood and adolescent obesity and..., Quek [/bib_ref] health consequences in children has led to governments around the world focusing on preventive strategies to reduce obesity in early life. The World Health Organization recommends taxes on SSBs to reduce consumption of added sugars to improve health. OverAU : Pleasecheckandconfirmthattheedittotheparenthesesinthesentence} 50 jurisdictions have implemented taxes on soft drinks, although they differ in terms of how much tax is passed through to the consumer, the types of soft drink targeted and the structure of the tax (including banded structure [bib_ref] Evaluating the 2014 sugar-sweetened beverage tax in Chile: An observational study in..., Nakamura [/bib_ref] and taxes levied in terms of volume sold [bib_ref] The Effectiveness of sin food taxes: Evidence from Mexico, Aguilar [/bib_ref] or as a proportion of the price [bib_ref] Trends in beverage prices following the introduction of a tax on sugar-sweetened..., Alvarado [/bib_ref]. In March 2016, the UK government proposed a number of strategies, including a soft drinks industry levy (SDIL) on manufacturers, importers, and bottlers of SSBs, to reduce prevalence of obesity in childhood [bib_ref] Is Obesity Policy in England Fit for Purpose? Analysis of Government Strategies..., Theis [/bib_ref]. The two-tier SDIL, implemented in April 2018, differed from most other tax structures in that it was designed to incentivise manufacturers to reformulate higher sugar soft drinks to move them to a lower tax tier. Manufacturers and importers were subject to a charge of £0.24/litre on soft drinks containing �8 g of sugar per 100 ml, £0.18/litre on soft drinks containing between �5 to <8 g of sugar per 100 ml, and no levy on drinks containing <5 g sugar per 100 ml. Levy exempt drinks include milk, milk-based drinks, 100% fruit juice, and powders used to make drinks. As part of the broader health strategy for young people, the UK government indicated they would use revenues raised through the SDIL to fund physical education in schools and breakfast and after-school clubs. Evidence suggests that the UK SDIL led to substantial reformulation of the UK soft drinks market. The percentage of drinks containing >5 g sugar/100 ml fell from 49% to just 15% between September 2015 and February 2019, with reformulation accelerating after announcement of the UK SDIL [bib_ref] Impact of the announcement and implementation of the UK Soft Drinks Industry..., Scarborough [/bib_ref]. Overall, the UK SDIL was associated with a reduction in sugar purchased from soft drinks [bib_ref] Changes in soft drinks purchased by British households associated with the UK..., Pell [/bib_ref]. While the price of soft drinks increased following implementation of the SDIL, the levy was only partially passed on to the consumer. For example, in drinks containing between �5 to <8 g of sugar per 100 ml, approximately one-third of the levy was passed on [bib_ref] Impact of the announcement and implementation of the UK Soft Drinks Industry..., Scarborough [/bib_ref]. A number of modelling studies [bib_ref] The potential impact on obesity of a 10% tax on sugar-sweetened beverages..., Briggs [/bib_ref] [bib_ref] Modelled health benefits of a sugar-sweetened beverage tax across different socioeconomic groups..., Lal [/bib_ref] have predicted that the introduction of SSB taxes would lead to a modest reduction in obesity in children and adults at the population level, but no study to date has used empirical data to examine whether the response of the SSB industry to the UK SDIL was associated with a subsequent change in the prevalence of childhood obesity. A few studies have used empirical data to estimate associations between SSB taxes and weight-related outcomes in children and adolescents and have either shown no overall association [bib_ref] The Effectiveness of sin food taxes: Evidence from Mexico, Aguilar [/bib_ref] [bib_ref] Associations between state-level soda taxes and adolescent body mass index, Powell [/bib_ref] [bib_ref] The effects of soft drink taxes on child and adolescent consumption and..., Fletcher [/bib_ref] or small to modest associations in specific subgroups such as low-income householdschildren with higher body mass indices (BMIs) [bib_ref] Changes in Weight-Related Outcomes Among Adolescents Following Consumer Price Increases of Taxed..., Gračner [/bib_ref] or in adolescent girls but not boys [bib_ref] Changes in Weight-Related Outcomes Among Adolescents Following Consumer Price Increases of Taxed..., Gračner [/bib_ref]. Different findings from these discrete studies may be related to use of different outcome measures (in particular, one study relied on subjective measures of self-reported weight [bib_ref] Associations between state-level soda taxes and adolescent body mass index, Powell [/bib_ref] , differences in change in SSB prices achieved by taxes (some were associated with small average increases in prices of SSB (<5%) [bib_ref] Associations between state-level soda taxes and adolescent body mass index, Powell [/bib_ref] [bib_ref] The effects of soft drink taxes on child and adolescent consumption and..., Fletcher [/bib_ref] or differences in substitutions to high-calorie untaxed food [bib_ref] The Effectiveness of sin food taxes: Evidence from Mexico, Aguilar [/bib_ref] and drinks [bib_ref] The Effectiveness of sin food taxes: Evidence from Mexico, Aguilar [/bib_ref] [bib_ref] The effects of soft drink taxes on child and adolescent consumption and..., Fletcher [/bib_ref]. In this study, we use cross-sectional data on monthly prevalence of objectively assessed obesity in children when they enter (reception class; ages 4 to 5) and exit (year 6; ages 10 to 11) English primary schools to examine whether 19 months following the implementation of the UK SDIL there were changes in the trajectory of prevalence of obesity (1) overall and (2) by sex and deprivation. # Methods The study was registered (ISRCTN18042742) and the study protocol published. This study is reported as per the REporting of studies Conducted using Observational Routinelycollected health Data (RECORD) Statement (S1 Checklist). ## Data source We used population level data from the National Child Measurement Programme (NCMP). This surveillance programme began in 2006 and measures the height and weight of approximately 1 million children from English state-maintained primary schools in reception (ages 4 to 5 years) and year 6 (ages 10 to 11 years) annually, with the aim of monitoring national rates of overweight and obesity in children. LocalAU : PleasenotethatasperPLOSstyle; datatakesplura authorities oversee the data collection, and letters are sent to the parents of eligible children where they are informed about why the data are collected and how these are stored. There is also an opportunity to opt out of measurement. Approximately 99% of eligible schools (approximately 17,000 schools) take part each year and individual response rates are high with over 90% of eligible pupils taking part. Surveillance data provided by NCMP include prevalence of children with overweight or obesity by school class (reception or year 6), sex (male or female), school year (e.g., 2013/ 14), month of measurement, and the index of multiple deprivation (IMD) quintile of the location of the primary school that the child attends. The NCMP measures the height and weight of children in England throughout the academic school year (September to July); hence, there was no available data for the month of August when the long summer holiday takes place. IMD scores are commonly used in England as measures of multiple deprivation by considering seven distinct domains including income, employment, education, barriers to housing, health and disability, crime, and living environment. The BMI thresholds used to derive overweight and obesity prevalence values were based on the 85th and 95th centiles, respectively, of a reference sample of measures taken in the UK in 1990 taking account of height, weight, sex, and age, reflecting the definitions used by Public Health England for population surveillance. The study period was initially planned to end 2 years following the implementation of SDIL, but follow-up was curtailed in November 2019 (4 months prior to the proposed end date) to avoid any influence of potential household storing of food and drink in preparation for (i) the UK leaving the European Union (December 2019) and (ii) national lockdown because of the COVID-19 pandemic (March 2020)to avoid contamination with documented changes in weight status occurring in the pandemic. # Statistical analysis Interrupted time series (ITS) analyses were conducted to assess obesity prevalence in relation to the UK SDIL in children attending primary school reception or year 6 classes, overall and by sex and IMD quintile. The ITS used monthly data from September 2013 (study month 1) until November 2019 (study month 69), including the months of the SDIL announcement (March 2016; study month 29) and implementation (April 2018; study month 52). Generalised least squares (GLS) models were used. Autocorrelation in the time series was examined visually using plots of autocorrelation and partial autocorrelation and statistically using Durbin-Watson tests; an autocorrelation-moving average (ARIMA) correlation structure was used, with the order (p) and moving average (q) parameters chosen to minimise the Akaike information criterion (AIC) in each model. School holidays are reported to influence weight-related outcomes in school children. To take account of this and other key events in the academic calendar year that might impact weight, we used calendar months as a proxy. Following a standard data-driven approach, to identify which calendar months might predict significant changes in obesity prevalence, we ran a series of GLS models in which a single calendar month was added to the equation. After all, calendar months were tested individually; models were finalised by including all the months that showed significant changes in obesity prevalence. Adding all months as dummy variables was avoided to restrict the number of variables to those that were informative, to reduce error, and to increase the precision of our estimates. The months of September, October, June, and February were significant for reception class children, and September and July were significant for year 6 children. Models for year 6 and reception age children were examined separately because reception age children in England typically start school full time, a few weeks after older children have returned. Model specifications for year 6 and reception class children are included (S1 Text). Counterfactual scenarios were estimated based on pre-announcement trends (S1 Fig). Absolute and relative differences in prevalence of obesity between observed and counterfactual values were estimated at month 69 (November 2019). Confidence intervals were calculated from standard errors estimated using the delta method [bib_ref] A Note on the Delta Method, Oehlert [/bib_ref]. All statistical analyses were performed in R version 4.1.0. ## Sensitivity analysis 1: inclusion of two alternative interruption points The main analysis included a counterfactual based on the pre-announcement trend (i.e., a scenario where neither the announcement nor implementation happened); however, previous research suggests that reformulation of drinks began some months after the announcement of SDIL but before implementation [bib_ref] Impact of the announcement and implementation of the UK Soft Drinks Industry..., Scarborough [/bib_ref]. Therefore, as well as capturing the earliest possible time when reformulation could come into effect, in sensitivity analyses (S1 we used two alternative interruption points. First (sensitivity analysis 1a), we used a counterfactual based on the trend from September 2013 to November 2016 (equivalent to 8 months post-announcement and the point at which reformulation increased rapidly) [bib_ref] Impact of the announcement and implementation of the UK Soft Drinks Industry..., Scarborough [/bib_ref]. Second (sensitivity analysis 1b), we used a counterfactual based on the pre-implementation trend, i.e., from September 2013 to April 2018. ## Sensitivity analysis 2: combining overweight and obesity prevalence In addition to examining prevalence of obesity, the main analysis was repeated and broadened to examine trajectories of excess weight prevalence, in relation to the SDIL, using monthly measures of overweight in addition to obesity. summarises the mean obesity prevalence in the study period (i) before the SDIL announcement and (ii) after the SDIL announcement, in primary school children in reception and year 6, overall and by sex and IMD quintile. Highest levels of obesity were observed in the most deprived areas regardless of age and sex; pupils in schools from the most deprived IMD quintiles had nearly twice the prevalence of obesity as those in the least deprived IMD quintiles. # Results ## Changes in obesity prevalence in relation to sdil Unless stated otherwise below, all estimates of changes in prevalence of obesity are based on values from November 2019 with respect to the counterfactual scenario of no SDIL announcement or implementation having occurred. Across all year 6 children, there was a 0.8-percentage point (PP) (95% confidence interval (CI): 0.3, 1.3) absolute reduction or 3.6% (95% CI: 1.2, 5.9) relative reduction in obesity prevalence compared to the counterfactual (see . Year 6 children in schools from the most deprived IMD quintiles (IMD1 and 2) had the greatest (relative) reductions in obesity prevalence of 4.1% (95% CI: 1.8, 6.3) and 5.5% (95% CI: 3.3, 7.7), respectively; however, large differences between year 6 girls and boys were observed. In year 6 girls, there was an overall relative reduction in obesity prevalence of 8.0% (95% CI: 5.4, 10.5). Analysis by IMD revealed greatest reductions in the two most deprived IMD quintiles (1 and 2) of 9.0% (95% CI: 5.9, 12.1) and 11.0% (95% CI: 9.2, 12.7), respectively, where a clear break in trend was observed graphically some months following the SDIL implementation [fig_ref] Table 1: Mean obesity prevalence [/fig_ref]. In year 6 boys, there was no overall change in obesity prevalence and no obvious pattern in changes in prevalence by IMD quintile, although there was a large relative increase in obesity prevalence of 10.1% (95% CI: 4.3, 15.9) in the least deprived IMD quintile and a small reduction in prevalence of obesity in IMD2 of 3.30% (95% CI: 0.4, 6.2) [fig_ref] Fig 2: Prevalence [/fig_ref]. In reception children, compared to the counterfactual, there was no absolute change in obesity prevalence overall in girls (0.2 PP (95% CI: 0.8, −0.3)) and boys (0.5 PP (95% CI: 1.0, −0.1)). Examination by IMD and sex showed a consistent increase in prevalence of obesity, compared to the counterfactual, in the least deprived IMD groups in both girls (0.6 PP (95% CI: 1.2, 0.003)) [fig_ref] Fig 3: Prevalence [/fig_ref] and boys (0.6 PP (95% CI: 1.1, 0.1)) [fig_ref] Fig 4: Prevalence [/fig_ref] in reception class. When the interruption point was changed to December 2016 (8 months post-SDIL announcement, the point at which reformulation began, sensitivity analysis 1a), changes in obesity prevalence were consistent with the main findings, with reductions in obesity prevalence evident in year 6 children, specifically girls from schools in the most deprived areas (IMD 1 and 2) (S1 , and increases in obesity prevalence in year 6 boys from the least deprived areas (IMD 4 and 5). When the interruption point was changed to April 2018 (month of SDIL implementation, sensitivity analysis 1b, S2 findings varied from the main analysis, with an overall absolute increase in the prevalence of obesity in reception age children by 0.7 PP (95% CI: 0.1, 1.3). Compared to the counterfactual, there were few significant changes in obesity prevalence in the different year 6 groups, although reductions (e.g., [bib_ref] Effect of sugar-sweetened beverages on oral health: a systematic review and meta-analysis, Valenzuela [/bib_ref].8% (95% CI 5.7, 2.0) in year 6 girls from IMD 2) and increases (e.g., 3.8% (95% CI 0.2, 7.4) in boys in IMD4) were observed in some groups. Changes in prevalence of excess weight (overweight or obesity) in relation to the UK SDIL were comparable to the main findings on changes in trends in prevalence of obesity, . Absolute and relative changes in prevalence of obesity (95% CIs), compared to the counterfactual 1 , in reception and year 6 boys and girls, by IMD at 19 months post-implementation of the UK SDIL. with greatest reductions in excess weight observed in girls from schools in IMD quintiles 1 and 2 and no change in prevalence of excess weight overall in year 6 boys or reception age children (S3 . However, compared to the counterfactual scenario of no announcement or implementation, there was an observed absolute reduction in excess weight of reception age girls from the most deprived IMD (1) of 1.6 PP (95% CI 1.1, 2.1). ## Total population boys girls ## Interruption-sdil announcement pp change relative change (%) pp change relative change (%) pp change relative change (%) ## Reception # Discussion ## Summary of findings This is the first study that we are aware of that uses empirical data to examine changes in childhood obesity prevalence in England in relation to the UK SDIL. After accounting for prior trends in obesity, there was a 0.8-PP absolute reduction in year 6 children living with obesity, 19 months after the implementation of the SDIL. These reductions in year 6 children were predominantly driven by changes in girls, where there was a 1.6-PP absolute or 8.0% relative reduction in obesity prevalence. Assuming, based on our 2019 data, that there are 337,658 year 6 girls in England (of whom 18.4% have obesity), this reduction is equivalent to 5,234 averted cases of obesity in year 6 girls. Relative to the counterfactual, no overall change was observed in year 6 boys. We observed that for year 6 girls, reductions in obesity were greatest in the 40% most deprived IMD areas, with a 2.4-PP absolute or 9.0% relative reduction in the most deprived IMD quintile. Overall, the prevalence of obesity in reception class children was unchanged, compared to the counterfactual. ## Comparison with other studies and implications In this section, we draw on evidence from other studies and compare our findings with them, while also providing some potential explanations for our results and their implications. First, our findings are plausible since associations between SSB consumption and risk of obesity are well described in the literature [bib_ref] Sugar-sweetened beverages and weight gain in children and adults: a systematic review..., Malik [/bib_ref] [bib_ref] Effects of Soft Drink Consumption on Nutrition and Health: A Systematic Review..., Vartanian [/bib_ref] [bib_ref] Relation between consumption of sugar-sweetened drinks and childhood obesity: a prospective, observational..., Ludwig [/bib_ref]. Furthermore, a relationship between the UK SDIL and an overall reduction in sugar purchased from soft drinks across the population has previously been reported [bib_ref] Changes in soft drinks purchased by British households associated with the UK..., Pell [/bib_ref]. Several modelling studies have also predicted that SSB taxes are likely to be most effective at targeting sugar intake in children and younger adults [bib_ref] Health impact assessment of the UK soft drinks industry levy: a comparative..., Briggs [/bib_ref] [bib_ref] How well targeted are soda taxes?, Dubois [/bib_ref]. Second, the magnitude and pattern of associations in our results are consistent with recent findings from Mexico that report a modest reduction in overweight or obesity prevalence in adolescent girls (aged 10 to 18) with a 1.3-PP absolute decrease 2 years after a 10% SSB price increase (compared to a 1.6-PP absolute decrease observed in this study in 10-to 11-year-old girls 19 months after the levy was introduced) [bib_ref] Changes in Weight-Related Outcomes Among Adolescents Following Consumer Price Increases of Taxed..., Gračner [/bib_ref]. Moreover, similar to the findings of this study, no significant reductions in weight-related outcomes were observed in adolescent boys in Mexico. We note, however, that the tax implemented in Mexico is not directly comparable with the UK SDIL; in Mexico; the tax had a different design aimed at increasing the price to consumers resulting in 100% of the SSB tax being passed through to consumers, equating to a 14% increase in prices [bib_ref] Soda taxes and the prices of sodas and other drinks: Evidence from..., Grogger [/bib_ref] , and, importantly, the tax was included as a wider package of antiobesity measures, which included charging 8% on high-energy foods [bib_ref] The Effectiveness of sin food taxes: Evidence from Mexico, Aguilar [/bib_ref]. We note the importance of the finding that the tax in Mexico was more effective in girls who were heavier. Similar analysis was not possible here because we only had access to repeated cross-sectional data, which cannot be linked over time. Third, we found that reductions in obesity in relation to the levy were greatest in children who were older and from the most deprived areas. Previous studies have reported the same children are more likely to be higher consumers of SSBs [bib_ref] What's on the Menu? Policies to Reduce Young People's Sugar Consumption, Griffith [/bib_ref] [bib_ref] Sugar-sweetened beverage consumption from 1998-2017: Findings from the health behaviour in school-aged..., Morgan [/bib_ref] [bib_ref] Consumption frequency of added sugars and UK children's dental caries, Hong [/bib_ref] [bib_ref] Sugars consumption in a low-income sample of British young people and adults, Ntouva [/bib_ref] [bib_ref] Sugar-sweetened beverage consumption from 1998-2017: Findings from the health behaviour in school-aged..., Morgan [/bib_ref]. This suggests a possible dose-response gradient between consumption levels and effectiveness of the levy in reducing obesity. This also adds to the growing international evidence that SSB taxes may reduce inequalities in diet-related health outcomes. For example, some studies from other countries have shown that lower-income households were more likely to reduce their purchases or intake of sugar from SSBs following introduction of SSB taxes [bib_ref] After Mexico Implemented a Tax, Purchases of Sugar-Sweetened Beverages Decreased and Water..., Colchero [/bib_ref] , although this is not always the case [bib_ref] Evaluating the 2014 sugar-sweetened beverage tax in Chile: An observational study in..., Nakamura [/bib_ref] [bib_ref] The Impact of Sugar-Sweetened Beverage Taxes by Household Income: A Multi-City Comparison..., Barker [/bib_ref] [bib_ref] How do consumers respond to 'sin taxes'? New evidence from a tax..., Fichera [/bib_ref]. In this study, we also demonstrate that the UK SDIL is not associated with a change in obesity prevalence in children in the first year of primary school. This result is congruous with findings from a cohort of British children showing that SSB consumption at ages 5 or 7 are not related to adiposity at age 9 years [bib_ref] Is sugar-sweetened beverage consumption associated with increased fatness in children?, Johnson [/bib_ref]. Added sugars from drinks make up 30% of all added sugars in the diet of young children (aged 1 to 3 years), but this increases to more than 50% by late adolescence [bib_ref] What's on the Menu? Policies to Reduce Young People's Sugar Consumption, Griffith [/bib_ref]. The lower intake of sugars from soft drinks at very young ages may lower the potential of a tax on SSBs, making it harder to observe health effects at the population level. Fruit juices, which are not included in the levy, are thought to contribute similar amounts of sugar in young children's diets as SSBs and may explain why the levy alone is not sufficient to reduce weight-related outcomes in reception age children. In addition to drinks, confectionery, biscuits, desserts, and cakes are also important high-added sugar items, which are regularly consumed by young children and could be a target of additional obesity reduction strategies [bib_ref] What's on the Menu? Policies to Reduce Young People's Sugar Consumption, Griffith [/bib_ref]. While our finding that the SDIL had greater impacts on obesity prevalence in girls than boys is consistent with previous studies [bib_ref] Changes in Weight-Related Outcomes Among Adolescents Following Consumer Price Increases of Taxed..., Gračner [/bib_ref] , it is unclear why this might be the case, especially since boys were higher baseline consumers of SSBs [bib_ref] Sugar-sweetened beverage consumption from 1998-2017: Findings from the health behaviour in school-aged..., Morgan [/bib_ref]. One explanation is that there were factors (e.g., in food advertising and marketing) at work around the time of the announcement and implementation of the levy that worked against any associations of the SDIL among boys. There is evidence that soft drink manufacturers altered their marketing strategies in different ways in response to the SDIL including repackaging and rebranding products [bib_ref] Understanding Marketing Responses to a Tax on Sugary Drinks: A Qualitative Interview..., Forde [/bib_ref]. Numerous studies have found that boys are often exposed to more food advertising content than girls [bib_ref] Gender in food advertising to children: Boys eat first, Childs [/bib_ref] [bib_ref] Self-reported advertising exposure to sugar-sweetened beverages among US youth, Kumar [/bib_ref] [bib_ref] Television viewing and exposure to food-related commercials among European school children, associations..., Klepp [/bib_ref] , both through higher levels of TV viewing [bib_ref] Television viewing and exposure to food-related commercials among European school children, associations..., Klepp [/bib_ref] and through the way in which adverts are framed. Physical activity is often used to promote junk food, and boys, compared to girls, have been shown to be more likely to believe that energy-dense junk foods depicted in adverts will boost physical performanceand thus they are more likely to choose energy-dense, nutrient-poor products following celebrity endorsements. There is also evidence that girls tend to make healthier choices when it comes to diet (e.g., consuming more fruit and vegetables and less energy-dense foods) and other health behaviours (e.g., brushing teeth). One possibility for the observed differences between boys and girls may be that girls were more responsive to public health signalling arising from discussions around the SDIL or that they were more likely to choose drinks that had been reformulated to contain less sugar following the SDIL announcement. Even the strongest association of the SDIL among the most levy-responsive groups (e.g., year 6 girls) reflected only a dampening of the rate of increase in obesity prevalence compared to the counterfactual rather than a reversal in trends. This highlights that alongside the SDIL, additional evidence-informed obesity reduction strategies need to be in place to improve weight-related outcomes, especially in boys and younger children, as they enter primary school education. ## Strengths and limitations of the study This study makes use of a unique and well-powered ongoing nationally representative sample covering over 90% of children aged 4 to 5 and 10 to 11 years in state-run primary schools over the study period and tracks the prevalence of overweight and obesity in over 1 million school children annually. Obesity prevalence data were based on objective measures of height and weight rather than parental self-report, where there is a tendency to underestimate overweight [bib_ref] Factors Associated with Errors in Self-reports of Stature, Weight, and Body Mass..., Himes [/bib_ref]. The NCMP uses 85th and 95th centiles of the UK1990 growth reference to monitor overweight and obesity in children (accounting for age and sex), respectively [bib_ref] Establishing a standard definition for child overweight and obesity worldwide: international survey, Cole [/bib_ref]. However, other cut points are sometimes used, and there is some debate over whether this is the best measure of adiposity, particularly in younger children [bib_ref] Body composition data show that high BMI centiles overdiagnose obesity in children..., Wright [/bib_ref]. Parental consent in NCMP involves a selective opt-out, which is designed to increase participation rates. However, it has been suggested that girls with obesity are less likely to participate. This may have led to underestimation of the association between SDIL on obesity prevalence in girls. These effects are, however, likely to be small given that obesity levels in girls have not changed dramatically and participation in the sample overall remained high throughout our study period. Socioeconomic disadvantage was assessed using an area-level indicator (IMD) of the school that each child attended, a less sensitive measure than capturing socioeconomic disadvantage at the household level. However, there is a strong correlation between school-level IMD and the proportion of pupils eligible for free school meals, a measure of the number of children attending a school with a low household income,suggesting that the measure used here is a suitable proxy measure of household deprivation. Data on time trends of expected childhood weight loss in relation to diet interventions are sparse with studies not monitoring weight-related outcomes with regularity and from early in the intervention. This makes it particularly challenging to estimate how long from the SDIL announcement we would expect to observe changes in obesity prevalence in children. However, there is evidence that changes in energy balance in children can lead to rapid changes in weight loss, for example, seasonal differences in BMI are observed in school children, with weight gain typically occurring during the summer periods especially in children with overweight or obesity. Consistent with these observations, our statistical models and ITS graphs reveal spikes in obesity prevalence in the months following the summer holidays (September in reception and year 6 children, and October in children in reception) and dips in other months (e.g., in June and July) in some subgroups. These require further investigation that could contribute to understanding of seasonal variations in childhood obesity. Furthermore, our ITS graphs reveal that in some groups, there may be continued improvement in the longer term with a widening between counterfactual and observed values in, for example, year 6 girls (IMD 1, 2, and 5). The ITS approach used modelled counterfactuals on the obesity prevalence trends immediately prior to the SDIL announcement. Given that estimates of the overall difference between observed and counterfactual obesity prevalence can be sensitive to the time points at which the counterfactuals are modelled, as part of a sensitivity analysis, we included two extra interruption points. The first additional interruption was 8 months post-announcement of SDIL, a time when reformulation of SSBs was visibly starting to increase; here we observed very similar findings to the main analysis indicating that they are robust. The second additional interruption was assigned to the date of the SDIL implementation; using this model, we observed fewer significant changes in obesity prevalence compared to the counterfactual (for example, no significant difference was observed in year 6 girls overall). This finding may be explained by the fact that companies had already reformulated most of their products prior to the implementation date and trajectories of obesity prevalence had responded rapidly. Furthermore, examining trajectories of "excess weight" prevalence rather than prevalence of obesity as the outcome of interest led to findings broadly consistent with the main analysis. # Conclusions The UK SDIL was proposed by the UK government to tackle childhood obesity. The pattern of findings of this study suggests that the SDIL can contribute to reducing obesity prevalence in older primary school children. The SDIL announcement and implementation was associated with an overall relative decrease in obesity prevalence in year 6 girls aged 10 to 11 years of approximately 8% compared to the counterfactual scenario based on pre-announcement trends. These associations were even greater in girls from schools in the 40% most deprived areas, suggesting the SDIL could help to reduce inequalities in child obesity. Further obesity reduction policies are needed alongside taxes on SSBs to improve and reverse the current obesity prevalence in children. Changes in obesity prevalence compared to a counterfactual scenario based on trends prior to 8 months post-announcement. Absolute and relative changes in prevalence of obesity (95% CIs), compared to a counterfactual scenario 1 based on trends prior to 8 months post-announcement, overall and by IMD in reception and year 6 children, 19 months post-implementation of UK SDIL. CIAU : AbbreviationlistshavebeencompiledforthoseusedinS1 À S3Ta , confidence interval; IMD, index of multiple deprivation; SDIL, soft drinks industry levy. (DOCX) S2 Changes in obesity prevalence compared to a counterfactual scenario based on trends prior to the SDIL implementation. Absolute and relative changes in prevalence of obesity (95% CIs), compared to a counterfactual scenario 1 based on pre-SDIL implementation trends, overall and by IMD in reception and year 6 children, 19 months post-implementation of UK SDIL. CI, confidence interval; IMD, index of multiple deprivation; SDIL, soft drinks industry levy. (DOCX) S3 Changes in excess weight prevalence compared to a counterfactual scenario based on trends prior to the SDIL announcement. Absolute and relative changes in prevalence of excess weight (overweight or obesity) and 95% CIs, compared to a counterfactual scenario 1 , based on pre-SDIL announcement trends, overall and by IMD in reception and year 6 children, 19 months post-implementation of UK SDIL. CI, confidence interval; IMD, index of multiple deprivation; SDIL, soft drinks industry levy. (DOCX) ## Supporting information [fig] Table 1: Mean obesity prevalence (standard deviation) in the pre-and post-announcement periods of the UK SDIL, by school class, sex, and IMD quintiles. Mean (standard deviation) obesity prevalence in primary school children in reception 1 and year 6 [/fig] [fig] Fig 1: Prevalence (%) of obesity in year 6 girls (aged 10/11) between September 2013 and November 2019. Observed and modelled prevalence of obesity is shown by IMD quintile and overall. Dark blue points show observed data and dark blue lines (with grey shadows) shows modelled data (and 95% CIs) of obesity prevalence. The red line indicates the counterfactual line based on the pre-SDIL announcement trend (assuming the announcement and implementation had not occurred). The first and second dashed vertical lines indicate the announcement and implementation of the SDIL, respectively. CIAU : Abbreviatio , confidence interval; IMD, index of multiple deprivation; SDIL, soft drinks industry levy.https://doi.org/10.1371/journal.pmed.1004160.g001 [/fig] [fig] Fig 2: Prevalence (%) of obesity in year 6 boys (aged 10/11) between September 2013 and November 2019. Observed and modelled prevalence of obesity is shown by IMD quintile and overall. Dark blue points show observed data and dark blue lines (with grey shadows) shows modelled data (and 95% CIs) of obesity prevalence. The red line indicates the counterfactual line based on the pre-SDIL announcement trend (assuming the announcement and implementation had not occurred. The first and second dashed vertical lines indicate the announcement and implementation of the SDIL, respectively. NB: The scales used in Figs 2-4 differ to maximise resolution of the image. CI, confidence interval; IMD, index of multiple deprivation; SDIL, soft drinks industry levy. https://doi.org/10.1371/journal.pmed.1004160.g002 [/fig] [fig] Fig 3: Prevalence (%) of obesity in girls in reception class (aged 4/5) between September 2013 and November 2019. Observed and modelled prevalence of obesity is shown by IMD quintile and overall. Dark blue points show observed data and dark blue lines (with grey shadows) shows modelled data (and 95% CIs) of obesity prevalence. The red line indicates the counterfactual line based on the pre-SDIL announcement trends (assuming the announcement and implementation had not occurred). The first and second dashed vertical lines indicate the announcement and implementation of the SDIL, respectively. CI, confidence interval; IMD, index of multiple deprivation; SDIL, soft drinks industry levy.https://doi.org/10.1371/journal.pmed.1004160.g003 [/fig] [fig] Fig 4: Prevalence (%) of obesity in boys in reception class (aged 4/5) between September 2013 and November 2019.Observed and modelled prevalence of obesity is shown by IMD quintile and overall. Dark blue points show observed data and dark blue lines (with grey shadows) shows modelled data (and 95% CIs) of obesity prevalence. The red line indicates the counterfactual line based on the pre-SDIL announcement trends (assuming the announcement and implementation had not occurred). The first and second dashed vertical lines indicate the announcement and implementation of the SDIL, respectively. CI, confidence interval; IMD, index of multiple deprivation; SDIL, soft drinks industry levy.https://doi.org/10.1371/journal.pmed.1004160.g004 [/fig] [fig] S1: Checklist. RECORD checklist. RECORD, Reporting of studies Conducted using Observational Routinely-collected Data. (DOCX) Fig. Schematic diagram of the interrupted time series. Blue solid lines indicate observed data. Dashed red lines represent counterfactuals. Counterfactual for (1) main analysis based on obesity prevalence trends from 09/2013-03/2016; (2)sensitivity analysis (a) based on obesity prevalence trends from 09/2013-12/2016; and (3)sensitivity analysis (b) based on obesity trends from 09/2013-04/2018. (DOCX) Text. Model specifications for children in year 6 and reception class. (DOCX) [/fig]

The immuno-behavioural covariation associated with the treatment response to bumetanide in young children with autism spectrum disorder Bumetanide, a drug being studied in autism spectrum disorder (ASD) may act to restore gamma-aminobutyric acid (GABA) function, which may be modulated by the immune system. However, the interaction between bumetanide and the immune system remains unclear. Seventy-nine children with ASD were analysed from a longitudinal sample for a 3-month treatment of bumetanide. The covariation between symptom improvements and cytokine changes was calculated and validated by sparse canonical correlation analysis. Response patterns to bumetanide were revealed by clustering analysis. Five classifiers were used to test whether including the baseline information of cytokines could improve the prediction of the response patterns using an independent test sample. An immuno-behavioural covariation was identified between symptom improvements in the Childhood Autism Rating Scale (CARS) and the cytokine changes among interferon (IFN)-γ, monokine induced by gamma interferon and IFN-α2. Using this covariation, three groups with distinct response patterns to bumetanide were detected, including the best (21.5%, n = 17; Hedge's g of improvement in CARS = 2.16), the least (22.8%, n = 18; g = 1.02) and the medium (55.7%, n = 44; g = 1.42) responding groups. Including the cytokine levels significantly improved the prediction of the best responding group before treatment (the best area under the curve, AUC = 0.832) compared with the model without the cytokine levels (95% confidence interval of the improvement in AUC was [0.287, 0.319]). Cytokine measurements can help in identifying possible responders to bumetanide in ASD children, suggesting that immune responses may interact with the mechanism of action of bumetanide to enhance the GABA function in ASD.Translational Psychiatry (2022) 12:228 ; https://doi. # Introduction Autism spectrum disorder (ASD) affects about 1% children around the worldand can cause lifelong disability and elevate premature mortality [bib_ref] Premature mortality in autism spectrum disorder, Hirvikoski [/bib_ref]. Currently, no medication that can cure ASD or all of its core symptoms is available. The recent success of repurposing drugs for novel treatments in psychiatry has been highlighted [bib_ref] Placing old wine into new bottles: Successful repurposing of bumetanide for treatment..., Phillips [/bib_ref] , with one of the examples given being the use of bumetanide to improve the core symptoms in ASD [bib_ref] Bumetanide for autism spectrum disorder in children: a review of randomized controlled..., James [/bib_ref] [bib_ref] Symptom improvement in children with autism spectrum disorder following bumetanide administration is..., Zhang [/bib_ref] [bib_ref] A randomised controlled trial of bumetanide in the treatment of autism in..., Lemonnier [/bib_ref] [bib_ref] Effects of bumetanide on neurobehavioral function in children and adolescents with autism..., Lemonnier [/bib_ref] [bib_ref] Improved symptoms following bumetanide treatment in children aged 3 to 6 years..., Dai [/bib_ref]. The most frequent adverse events were hypokalemia, increased urine elimination, loss of appetite, dehydration and asthenia. The heterogeneity in the treatment effect of bumetanide among ASD patients was significant, ranging from 37.3% to 47.62% in the randomised clinical trials (RCTs) in China [bib_ref] Symptom improvement in children with autism spectrum disorder following bumetanide administration is..., Zhang [/bib_ref] [bib_ref] Improved symptoms following bumetanide treatment in children aged 3 to 6 years..., Dai [/bib_ref] and 51.80% [bib_ref] A randomised controlled trial of bumetanide in the treatment of autism in..., Lemonnier [/bib_ref] or from 26.3% to 45.2% [bib_ref] Effects of bumetanide on neurobehavioral function in children and adolescents with autism..., Lemonnier [/bib_ref] in the RCTs in France. Besides, there were also studies reported nonsignificant treatment effect of bumetanide for patients with ASD [bib_ref] Bumetanide for core symptoms of autism spectrum disorder (BAMBI): a single center,..., Sprengers [/bib_ref]. Understanding this heterogeneity is essential for its clinical applicability and requires further investigation of its underlying mechanism of action to achieve precision medicine for ASD children. The use of bumetanide as a potential drug to improve symptoms in ASD is based on a hypothesised pathoetiology of ASD, namely the delayed developmental switch of the gammaaminobutyric acid (GABA) functioning from excitatory to inhibitory [bib_ref] GABA system dysfunction in autism and related disorders: from synapse to symptoms, Coghlan [/bib_ref] [bib_ref] Dysfunction in GABA signalling mediates autism-like stereotypies and Rett syndrome phenotypes, Chao [/bib_ref] [bib_ref] Oxytocin-mediated GABA inhibition during delivery attenuates autism pathogenesis in rodent offspring, Tyzio [/bib_ref]. In the valproate and fragile X rodent models of autism, this GABA-switch can be facilitated by the reduction of intracellular chloride concentration, which is mediated by a sequential expression of the main chloride transporters, such as the potassium (K)-Cl co-transporters 2 (KCC2) and the importer Na-K-Cl cotransporter 1 (NKCC1) [bib_ref] Oxytocin-mediated GABA inhibition during delivery attenuates autism pathogenesis in rodent offspring, Tyzio [/bib_ref]. Therefore, bumetanide as an NKCC1 inhibitor has been tested for its ability to restore GABA function in ASD . However, these transporters can also be influenced by other molecules, such as cytokines, which are a number of small cell-signalling proteins closely interacting with each other to modulate the immune reactions. The cytokines have been implicated not only in brain development [bib_ref] The role of the innate immune system in psychiatric disorders, Jones [/bib_ref] , but also in GABAergic transmission [bib_ref] Prolonged interferon-gamma exposure decreases ion transport, NKCC1, and Na+-K+-ATPase expression in human..., Bertelsen [/bib_ref] [bib_ref] Environmental regulation of the chloride transporter KCC2: switching inflammation off to switch..., Pozzi [/bib_ref] [bib_ref] Maternal immune activation delays excitatory-to-inhibitory gamma-aminobutyric acid switch in offspring, Corradini [/bib_ref]. It has been reported that the interferon (IFN)-γ can decrease the levels of NKCC1 and the αsubunit of Na + -K + -ATPase, contributing to the restore of inhibitory GABA function [bib_ref] Prolonged interferon-gamma exposure decreases ion transport, NKCC1, and Na+-K+-ATPase expression in human..., Bertelsen [/bib_ref]. In mice subjected to maternal deprivation, the interleukin (IL)-1 has also been found to reduce the expression of KCC2, delaying the developmental switch of the GABA function and thereby possibly contributing to the pathophysiology of developmental disorders such as ASD [bib_ref] Environmental regulation of the chloride transporter KCC2: switching inflammation off to switch..., Pozzi [/bib_ref] [bib_ref] Maternal immune activation delays excitatory-to-inhibitory gamma-aminobutyric acid switch in offspring, Corradini [/bib_ref]. Therefore, a question naturally arises that whether the treatment effect of bumetanide for ASD can be affected by the immune responses in the patients. Indeed, compared with healthy controls, changes of the cytokine levels have already been reported in patients with ASD [bib_ref] The role of the immune system in autism spectrum disorder, Meltzer [/bib_ref] [bib_ref] Anti-inflammatory cytokines in autism spectrum disorders: A systematic review and meta-analysis, Saghazadeh [/bib_ref] [bib_ref] Serum cytokine levels in children with spectrum autism disorder: Differences in pro-and..., Kordulewska [/bib_ref] [bib_ref] Elevated immune response in the brain of autistic patients, Li [/bib_ref]. Recent meta-analyses showed that the levels of antiinflammatory cytokines IL-10 and IL-1 receptor antagonist (Ra) were decreased [bib_ref] Anti-inflammatory cytokines in autism spectrum disorders: A systematic review and meta-analysis, Saghazadeh [/bib_ref] , while proinflammatory cytokines IL-1β, IL-6 and anti-inflammatory cytokines IL-4, IL-13 were elevated in blood of patients with ASD [bib_ref] Serum cytokine levels in children with spectrum autism disorder: Differences in pro-and..., Kordulewska [/bib_ref]. The levels of IFN-γ, IL-6, tumour necrosis factor (TNF)-α, granulocyte-macrophage colony-stimulating factor (GM-CSF) and IL-8 were observed to be elevated [bib_ref] Elevated immune response in the brain of autistic patients, Li [/bib_ref] in postmortem brain tissues of ASD patients, and increased level of IFN-γ, monocyte chemotactic protein (MCP)-1, IL-8, leukaemia inhibitory factor (LIF) and interferon-gamma inducible protein (IP)-10 were found in another study [bib_ref] Neuroglial activation and neuroinflammation in the brain of patients with autism, Vargas [/bib_ref]. These widely spread changes suggest that the cytokine signalling in ASD may be better characterised by multivariate patterns of cytokines. In literatures, many associations had been reported between the levels of cytokines (e.g., MCP-1, IL-1β, IL-4, IL-6, etc.) and both core symptoms and adaptive functions in children with ASD [bib_ref] Associations of impaired behaviors with elevated plasma chemokines in autism spectrum disorders, Ashwood [/bib_ref] [bib_ref] Decreased transforming growth factor beta1 in autism: a potential link between immune..., Ashwood [/bib_ref] [bib_ref] Elevated plasma cytokines in autism spectrum disorders provide evidence of immune dysfunction..., Ashwood [/bib_ref]. Therefore, it has been suggested that cytokines may be used as biomarkers to identify different subsets within ASD. In each of these subsets the patients with ASD may share a commonly immune-related pathoetiology and therefore may have similar profiles of response to treatment [bib_ref] The immune system, cytokines, and biomarkers in autism spectrum disorder, Masi [/bib_ref]. Based on these previous findings, we analysed data acquired through the Shanghai Xinhua ASD registry, China, that began in 2016 to test the hypothesis that the immune activity of patients might help to identify the best responders to bumetanide in ASD. ## Materials and methods participants The ASD participants were recruited from the Shanghai Xinhua ASD registry at Shanghai Jiaotong University Medical School Affiliated Xinhua Hospital in Shanghai, China, including the participants from two previous registered clinical studies, i.e., CHICtr-OPC-16008336 and NCT03156153. The patients were diagnosed with ASD according to the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5). Diagnoses were confirmed with the Autism Diagnostic Observation Schedule (ADOS), and a Children Autism Rating Scale (CARS) total score of no less than 30. Exclusion criteria include liver and kidney dysfunction; a history of allergy to sulfa drugs; abnormal electrocardiography; genetic or chromosomal abnormalities; suffering from nervous system diseases (e.g., epilepsy, etc.). Comprehensive behavioural assessments and collections of clinical samples were performed for all patients. Between May 1st, 2018, to April 30th, 2019, a total of 90 ASD children, aged 3-10 years old, under a 3-month stable treatment of bumetanide without behavioural interventions and any concomitant psychoactive medications had both blood draws and behavioural assessments. Among these patients, 11 of them were further excluded due to the lack of the follow-up data at month 3. A group of 37 children, under 3-month stable treatment of placebo without behavioural interventions and any concomitant psychoactive medications had both blood draws and behavioural assessments. Therefore, the current analysis used a subsample of 116 young children with ASD, whose blood samples were available both before and after the treatment. The blood samples were sent in three batches (Discovery Set: n = 37 on December 4, 2019; Validation Set: n = 42 on May 22, 2019; and Control Set: n = 37 on January 5, 2022) to measure the serum levels of 48 cytokines for the immune response [fig_ref] Table 1: The demographic and clinical [/fig_ref] , and the clinical symptoms were assessed using CARS, ADOS and the Social Responsiveness Scale (SRS). Following the protocols of previous studies [bib_ref] Improved symptoms following bumetanide treatment in children aged 3 to 6 years..., Dai [/bib_ref] , bumetanide treatment consisted of two 0.5 mg tablets per day for three months, given at 8:00 a.m. and 4:00 p.m. The tablet size is 8 mm diameter x 2 mm thickness, which is quite small. Each time, the patient took half of a tablet, which was not difficult for most of the patients. However, the careers were recommended to grind the half-tablet into powder and give the powder in water, if necessary. Possible side effects were closely monitored during the treatment. Blood parameters (serum potassium and uric acid) were monitored via laboratory tests and symptoms (thirst, diuresis, nausea, vomiting, diarrhoea, constipation, rash, palpitation, headache, dizziness, shortness of breath, and any other self-reported symptoms) were telephone interviewed , and both of them were reported to the research team by telephone at 1 week and 1 month after the initiation of treatment and at the end of the treatment period. The cytokine levels of the children with gastrointestinal problems were compared with those without such problems . Behavioural assessments of CARS and ADOS and measurements of cytokine levels were performed at the baseline before the treatment and after the 3-month treatment. The behavioural assessment of SRS was used at the baseline only. The study was conducted in accordance with the provisions of the Declaration of Helsinki and Good Clinical Practice guidelines and was approved by the Ethics Committee of Xinhua Hospital affiliated to Shanghai Jiao Tong University School of Medicine. Written informed consent was obtained from the parent or legal guardian of each participant before sample collection. ## Measures Clinical assessments. The CARS was used to diagnose and evaluate the severity of clinical symptoms of ASD patients. The CARS consisted of 15 items rated on a 7-point scale from one to four; higher scores are associated with a higher level of impairment. Total scores can range from a low of 15 to a high of 60; scores below 30 indicate that the individual is in the non-autistic range, scores between 30 and 36.5 indicate mild to moderate autism, and scores from 37 to 60 indicate severe autism. We further categorised these items into three subscales [bib_ref] Domains of the childhood autism rating scale: Relevance for diagnosis and treatment, Dilalla [/bib_ref] : Social impairment, Negative emotionality and Distorted sensory response. ADOS was used as a supplement to gauge disease severity, and it contained total score items and four modules for assessment of Social interaction, Communication, Play, and Imaginative use of materials for individuals suspected of having ASD [bib_ref] The autism diagnostic observation schedule-generic: a standard measure of social and communication..., Lord [/bib_ref]. SRS identified a wide spectrum of deficits in reciprocal social behaviour, ranging from absent to severe, based on observations of a child's behaviour in naturalistic social settings, focused on the behaviour of a child or adolescent between the ages of 4 and 18 years. It was a 65-item questionnaire that is completed by teacher, a parent, and/or another adult caregiver. Scoring is on a four-point Likert Scale. Five subscales are also provided: Social Awareness (AWA), Social cognition (COG), Social Communication (COM), Social Motivation (MOT), and Autistic Mannerism (MANN) [bib_ref] Social responsiveness scale: SRS-2, Constantino [/bib_ref]. Cytokine levels. For each subject peripheral blood was collected, centrifuged at 2300 rpm for 10 min, and the plasma harvested. Plasma was aliquoted and stored at −70°C until cytokine analysis. A plane of 48 cytokines, chemokines, and growth factors were measured using the Bio-Plex multiple immunoassays (Bio-Plex®; BIO-RAD Laboratories, Inc.). Prior to setting up our assays, the Bio-Plex 200 System plane Reader instrument was calibrated according to manufacturer instructions. To prepared experimental samples, frozen plasma aliquots were passively thawed to room temperature and diluted four-fold in assay buffer (15 μL sample + 45 μL sample diluent HB). After preparation of capture bead mixture, and standards, the immunoassay was carried out on a 96-well plane. The experimental steps were in accordance with the instructions. Data acquisition was set to a 50bead count minimum per analyte per well. Unknown sample cytokine concentrations were processed and presented with Bio-plex Manager software using a standard curve derived from the known reference cytokine concentrations supplied by the manufacturer. A five-parameter model was used to calculate final concentrations and values were expressed in pg/ml. The sensitivity of this assay allowed the detection of cytokine concentrations within the following ranges: IFN-α2 3.6-3992.4 pg/ml; IFN-γ 0.9-14556. [bib_ref] Improved symptoms following bumetanide treatment in children aged 3 to 6 years..., Dai [/bib_ref] # Statistical analysis Data preprocessing. After excluding the cytokines whose values were less than the limit of detection, 35 cytokines were included in our analysis. Data from three batches, containing both the baseline and the change (the difference value from the baseline and the follow-up data) of the cytokine levels, were min-max normalised and log transformed separately. To adjust for the batch effect, we applied an empirical Bayes approach to the baseline of cytokine levels using the 'ComBat' parametric algorithm provided in the R-package 'sva' [bib_ref] Adjusting batch effects in microarray expression data using empirical Bayes methods, Johnson [/bib_ref]. Principle component analysis (PCA) was used to visualise the non-biological variation due to the batch effect and was repeated to confirm its adjustment . Before and after treatment, we compared the demographic parameters (i.e., sex proportion, age, body mass index [BMI]) and symptom severity (i.e., ADOS and CARS) between these three data sets using two-sided Mann-Whitney U-test, or Pearson's chi-squared test where applicable. Multivariate association analysis to characterise the immuno-behavioural covariation. First, the pairwise correlation between the CARS_total score and each of the 35 cytokine levels were assessed by the Spearman-rank correlation. The correlation between the change in the CARS_total score and the change in each of the 35 cytokine levels after the treatment was also tested. The falsediscovery rate (FDR) was used to correct for the multiple comparisons. Second, to uncover the multivariate association between the behavioural assessments and the cytokine levels, we employed the sparse canonical correlation analysis (sCCA) provided in a R-package 'sRDA' (version 1.0.0) [bib_ref] Sparse redundancy analysis of high-dimensional genetic and genomic data, Csala [/bib_ref]. Canonical correlation analysis (CCA) is a classical method for determining the relationship between two sets of variables. Given two data sets X 1 and X 2 of dimensions n × p 1 and n × p 2 , respectively, from n observations, CCA seeks the pairs of linear combinations (i.e., the canonical variables), one from the variables in X 1 and the other from the variables in X 2 , that are maximally correlated with each other. However, some variables may make negligible but non-zero contributions to the canonical variables. sCCA was developed to address this issue. sCCA applies an L 1 penalty to the canonical weights, which forces some of them to take a value of exactly zero. Mathematically, [formula] max w1;w2 w T 1 X T 1 X 2 w 2 subject to w 1 k k 2 ¼ 1; w 2 k k 2 ¼ 1; w 1 k k 1 c 1 ; w 2 k k 1 c 2 ; w 1 2 R p 1 ; w 2 2 R p 2 : [/formula] Here, c 1 and c 2 are assumed to fall within the bounds 1 c 1 ffiffiffiffiffi p 1 p and 1 c 2 ffiffiffiffiffi p 2 p , where p 1 and p 2 are the numbers of features in X 1 and X 2 , respectively. We refer to w 1 and w 2 as the canonical weights, and X 1 w 1 and X 2 w 2 as the canonical scores. Therefore, this algorithm could identify a linear combination of three CARS subscales (i.e., the behaviouralcomponent) that was significantly associated with another linear combination of a few cytokine levels (i.e., the cytokine-component). Meanwhile, the sparsity of this algorithm ensured only the key cytokines driving the behavioural association were selected in the immune component. In the Discovery Set, we explored the sCCA between CARS subscales and cytokine levels using the baseline data or using the changes between baseline and follow-up. The significance of an identified canonical correlation was assessed by 5000 permutations [bib_ref] Sparse redundancy analysis of high-dimensional genetic and genomic data, Csala [/bib_ref]. Only the significant canonical components were retained. The Validation Set was used to confirm these significant canonical components among patients with bumetanide treatment. The Control Set was used to test whether the association between the cytokine-component and the behaviouralcomponent was significant in the placebo group. Sensitivity analysis was conducted using the data from the Validation Set by re-evaluating the canonical correlation after controlling for the potential confounders, including age, sex, BMI and the canonical variables at the baseline. Using the cytokine-component score (x-axis) and the behavioural-component score (y-axis) established by sCCA, each patient could be mapped onto a two-dimensional, called the immuno-behavioural covariation plane, characterising the immuno-behavioural covariation in the response patterns to the bumetanide treatment among young children with ASD. Clustering analysis to identify the immuno-behavioural groups. We applied K-means, an unsupervised clustering algorithm, to identify the clusters of patients according to the immuno-behavioural covariation. The patients in each cluster (i.e., an immuno-behavioural group within ASD) had the similar canonical scores, suggesting they shared the similar patterns of response to bumetanide in both immune system and clinical behaviour. The cluster structures were first identified using the Discovery Set and then validated using the Validation Set. The optimal number of clusters was selected based on the elbow (maximum change) of the scree plot using the Hubert statistic implemented in the R-package 'NBclust' [bib_ref] NbClust: an R package for determining the relevant number of clusters in..., Charrad [/bib_ref]. To demonstrate the distinct patterns of response to bumetanide in the immuno-behavioural groups, we applied the one-sample t-test to the after-treatment changes of both the CARS subscales and the cytokines selected by the sCCA algorithm as described above. We also compared these changes among the above identified immunobehavioural groups using the Kruskal-Wallis rank sum test with the FDR correction for multiple comparisons. Prediction of the treatment response to bumetanide using the baseline information. To predict the response to bumetanide of a patient with ASD, we trained the classifiers for the immuno-behavioural groups identified above using the baseline information before the treatment. The baseline information included the 35 cytokine levels, 3 types of clinical assessment (CARS, SRS, ADOS) and 2 demographic parameters (sex and age). The classifiers included the Oblique Random Forest (ORF), Partial Least Squares (PLS), sparse Linear Discriminant Analysis (sLDA), Neural Networks (NN) and Support Vector Machine (SVM) as implemented in the R-package 'caret' with both feature selection and oversampling [bib_ref] Building predictive models in R using the caret package, Kuhn [/bib_ref]. The models were first trained using the Discovery Set, and their performances were compared using the Validation Set. The 95% confidence interval of the difference between the areas under the curves of a pair of models was constructed by 100 bootstraps. First, we tested whether including the cytokine levels at the baseline could improve the prediction of the immuno-behaviourally defined responders. The averaged performances of these five classifiers (i.e., the averaged area under the curve, AUC) were reported and compared. Second, we tested whether the behaviourally defined responders were more difficult to be predicted at the baseline compared with the immuno-behaviourally defined ones. In the previous clinical trials of bumetanide for ASD [bib_ref] Symptom improvement in children with autism spectrum disorder following bumetanide administration is..., Zhang [/bib_ref] [bib_ref] A randomised controlled trial of bumetanide in the treatment of autism in..., Lemonnier [/bib_ref] [bib_ref] Effects of bumetanide on neurobehavioral function in children and adolescents with autism..., Lemonnier [/bib_ref] , the proportion of patients who responded positively to the treatment was between 30% and 40%. Therefore, we divided the patients with ASD into two groups according to the ΔCARS_total (= the baseline CARS_totalthe follow-up CARS_total) with a cutoff of 2.5 (n = 22, 27.85% of the patients had ΔCARS_total >2.5) or 2 (n = 32, 40.51% of the patients had ΔCARS_total >2) points. # Results ## Participants Three data sets of children with ASD (n = 116) were used in the current study. Discovery Set (n = 37) had a mean age of 47 months (±17.35 months), 18.92% of whom were girls; Validation Set (n = 42) had a mean age of 54 months (±20.19 months), 23.81% of whom were girls; Control Set (n = 37) had a mean age of 50 months (±10.72 months), 13.16% of whom were girls. No significant difference in clinical characteristics or cytokine levels was identified between these three data sets [fig_ref] Table 1: The demographic and clinical [/fig_ref]. Changes after the administration of bumetanide Seventy-nine patients were treated with bumetanide for 3 months, and the CARS total score decreased after the treatment (effect size Cohen's d = 1.26, t 78 = 11.21, p < 0.001). The treatment effect showed no difference between the Discovery Set and the Validation Set (ΔCARS_total: mean(±SD): 1.54 (±1.40) vs. 1.90 (±1.34)). Consistent to the previous studies of the low-dose bumetanide for ASD, the side effects were rarely reported . No significant difference in the cytokine levels between the children with and without the gastrointestinal problems at the baseline . A number of cytokine levels were changed significantly after the treatment of bumetanide, but none of them was changed significantly after the treatment of placebo . No significant pairwise association could be identified in the Discovery Set, the Validation Set and the Control Set among four groups of variables, including the baseline CARS total score, the baseline cytokine levels, the change of CARS total score, and the changes of cytokine levels . ## Covariation between symptom improvement and cytokine changes Using the Discovery Set, we found a canonical correlation (r = 0.459; p < 0.001 by permutation; between 2 canonical components (i.e., cytokine-component and behavioural-component) by sCCA. The cytokine-component was a combination of the changes of the three cytokine levels, including the MIG, IFN-α2, IFNγ. The behavioural-component was a combination of the changes of three subscales of the CARS scores, including the social impairment score, negative emotionality score, distorted sensory response score. Applying these canonical weights to an independent data set (i.e., the Validation Set), we confirmed the correlation between the cytokine-component and the behavioural-component (r = 0.316; p = 0.012 by permutation; . In contrast, applying these canonical weights to the Control Set, we found that the association between cytokine-component and behavioural-component was not significant (r = 0.038, p = 0.821), which might indicate that the association between cytokine-component and behavioural-component was related to the drug effect. Sensitivity analysis Using the Validation Set, we found that the correlation between the canonical components identified above remained significant . At the baseline, we also found that the cytokines canonical scores were associated with other clinical assessments, including the SRS_total (r = Three distinct response patterns revealed by the immunobehavioural covariation Using both the CARS-and the cytokine-component scores, we found that the patients fell into three clusters and Comparing among these three groups of patients in both data sets , D and ; , we found that the best responding group (n = 17) had the greatest reduction in the CARS total score (ΔCARS_total: 3.32(±1.47), Hedge's g = 2.16, p < 0.001), which was most prominent in the social impairment score (ΔCARS_S: 2.15(±1.13), g = 1.81, p < 0.001) and increased cytokine levels (ΔMIG: g = 0.72, p = 0.012; ΔIFN-α2: g = 0.84, p = 0.006). The least responding group (n = 18) had the least reduction in the CARS total score (ΔCARS_total: 1.03(±0.96), g = 1.02, p < 0.001), which was most significant in the negative emotionality score (ΔCARS_N: 0.69(±0.55), g = 1.21, p < 0.001), and decreased in the least responding group (ΔMIG: g = −1.07, p < 0.001; ΔIFN-γ: g = −1.32, p < 0.001). The medium responding group had a significant decrease in both the CARS_total score and all subscales with a small effect size each, while the IFN-γ level decreased and the IFN-α2 level increased in this group . ## Baseline cytokine levels helped in identifying the treatmentresponders Training five different types of classifiers with Discovery Set and testing the performance using Validation Set, we found that including the cytokine levels at the baseline significantly improved the prediction accuracy for both the best responding group (AUC with/without the baseline cytokine levels = 0.768/ 0.646; improvement in AUC ¼ 0:122, 95% confidence interval (CI): (0.103,0.130); [fig_ref] Figure 3: ROC curve for prediction for the immuno-behaviourally defined responding group [/fig_ref] , B) and the least responding group ( AUC with/without the baseline cytokine levels = 0.698/0.618; improvement in AUC ¼ 0:080, 95%CI: (0.064,0.097); [fig_ref] Figure 3: ROC curve for prediction for the immuno-behaviourally defined responding group [/fig_ref]. Each of the five models showed a better performance after including the cytokine levels into the model . Furthermore, we found that the behaviourally defined responders according to the ΔCARS_total with a cutoff of 2.5 were more difficult to be predicted at the baseline (AUC ¼ 0:569) compared with the immuno-behaviourally defined responders ( AUC ¼ 0:768; ; . Similar results were found when using the threshold of 2 to behaviourally define the responders . # Discussion In this study, we observed a significant improvement of clinical symptoms with bumetanide treatment in children with ASD, and such improvement was associated with a pattern of changes in three cytokine levels, namely the IFN-γ, MIG and IFN-α2 (r = 0.459 in the Discovery Set and r = 0.316 in the Validation Set). These cytokine levels at the baseline could improve the prediction of the bumetanide responders compared with using the behavioural assessments alone, and the best predictor achieved an AUC of 0.83 in the independent test data set . The implications of these findings may be twofold: (1) a significant part of the clinical heterogeneity in the treatment effect of bumetanide for ASD is associated with the differences in the immune system of patients, and (2) the component score of cytokines had a potential to construct a blood signature for predicting and monitoring the bumetanide treatment in young children with ASD. Our finding of an immuno-behavioural covariation highlights the role of the immune system in the clinical effect of bumetanide in young children with ASD. IFN-γ, as a T helper cell 1 (Th1) cytokine with proinflammatory effects, was selected by the sCCA algorithm to be one of the three cytokines to form the canonical score that was associated with the improvement in CARS. Compared to controls, higher level of IFN-γ has been reported in the brain tissue [bib_ref] Elevated immune response in the brain of autistic patients, Li [/bib_ref] , cerebrospinal fluid (CSF) [bib_ref] Neuroglial activation and neuroinflammation in the brain of patients with autism, Vargas [/bib_ref] , plasma [bib_ref] A meta-analysis of pro-inflammatory cytokines in autism spectrum disorders: effects of age,..., Saghazadeh [/bib_ref] and peripheral blood mononuclear cell (PBMC) [bib_ref] Elevated cytokine levels in children with autism spectrum disorder, Molloy [/bib_ref] in ASD patients, and lower level has been observed in neonatal dried blood samples (n-DBSS) of ASD children [bib_ref] Neonatal levels of cytokines and risk of autism spectrum disorders: an exploratory..., Abdallah [/bib_ref]. Accumulating evidences support that IFN-γ can inhibit chloride secretion [bib_ref] Probiotics and commensals reverse TNF-alpha-and IFN-gamma-induced dysfunction in human intestinal epithelial cells, Resta-Lenert [/bib_ref] and down-regulate both the NKCC1 expression [bib_ref] Prolonged interferon-gamma exposure decreases ion transport, NKCC1, and Na+-K+-ATPase expression in human..., Bertelsen [/bib_ref] [bib_ref] Probiotics and commensals reverse TNF-alpha-and IFN-gamma-induced dysfunction in human intestinal epithelial cells, Resta-Lenert [/bib_ref] and the Na + -K + -ATPase expression [bib_ref] Prolonged interferon-gamma exposure decreases ion transport, NKCC1, and Na+-K+-ATPase expression in human..., Bertelsen [/bib_ref] , which had been implicated in the GABAergic dysfunction in ASD [bib_ref] GABA system dysfunction in autism and related disorders: from synapse to symptoms, Coghlan [/bib_ref] [bib_ref] The K+/Clco-transporter KCC2 renders GABA hyperpolarizing during neuronal maturation, Rivera [/bib_ref]. Indirectly, an animal study also showed that stimulation with high concentration of IFN-γ could increase the expression of IL-1β [bib_ref] Interrelation of inflammation and APP in sIBM: IL-1 beta induces accumulation of..., Schmidt [/bib_ref] , which is an inflammatory cytokine that can affect the expression of chloride transporters and delay the developmental switch of Sparse canonical correlation analysis. We carried out sparse canonical correlation analysis in A Discovery Set and B Validation Set. The canonical scores between CARS and cytokines, which were min-max normalised and log transformed, were significant related in both data sets. GABA signalling [bib_ref] Environmental regulation of the chloride transporter KCC2: switching inflammation off to switch..., Pozzi [/bib_ref]. Therefore, the immune system may interact with the mechanism of action of bumetanide to restore the GABA function in ASD. The cytokine-symptom association was identified in the changes after the treatment of bumetanide but not before the treatment, suggesting that bumetanide might interact with the cytokines and the changes of which contributed to the treatment effect of bumetanide. Animal studies showed a rapid brain efflux of bumetanide, but a number of clinical trials have shown a significant treatment effect for neuropsychiatric disorders, including ASD, epilepsy and depression [bib_ref] Off-label use of bumetanide for brain disorders: an overview, Kharod [/bib_ref] [bib_ref] Bumetanide treatment during early development rescues maternal separation-induced susceptibility to stress, Hu [/bib_ref]. These findings may suggest the possible systemic effects of bumetanide as a neuromodulator for these neuropsychiatric disorders. Considering its molecular structure, bumetanide has been recently identified by an in vitro screen of small molecules that can act as an anti-proinflammatory drug via interleukin inhibition [bib_ref] Small molecule therapeutics for COVID-19: repurposing of inhaled furosemide, Wang [/bib_ref]. This anti-proinflammatory activity of bumetanide might alter the blood levels of cytokines outside the brain-blood-barrier (BBB). In fact, it has already been reported that bumetanide reduced the Lipopolysaccharide-induced production of proinflammatory cytokines following a direct pulmonary administration in RAW264.7 cells and in lung-injured mice [bib_ref] Bumetanide attenuates acute lung injury by suppressing macrophage activation, Hung [/bib_ref]. These inflammatory signalling messengers may pass the BBB [bib_ref] Passage of cytokines across the blood-brain barrier, Banks [/bib_ref] and influence the neuronal chloride homoeostasis via, for example, altering the KCC2 expression [bib_ref] Maternal immune activation delays excitatory-to-inhibitory gamma-aminobutyric acid switch in offspring, Corradini [/bib_ref]. The plausibility of reducing inflammation to enhance the KCC2 expression has recently been discussed in a 2020 review [bib_ref] Environmental regulation of the chloride transporter KCC2: switching inflammation off to switch..., Pozzi [/bib_ref]. Indeed, we found that the best responding group (Hedge's g = 2.16 for the reduction of CARS total score) had the greatest increase in the cytokine-component score. In the contrary, the least responding group (g = 1.02) had the greatest decrease in the cytokine-component score. Taken together, these findings suggest that bumetanide may be a drug to inhibit NKCC1 and enhance KCC2 through its interplay with the cytokines inside and/or outside of the brain. It was disappointing that the phase 3 clinical studies assessing bumetanide in the treatment of ASD were terminated after the middle analysis, which showed no sign of effectiveness, led by two pharmaceutical companies Servier and Neurochlore [https://servier.com/en/communique/servier-andneurochlore-announce-the-main-results-of-the-two-phase-3-clinicalstudies-assessing-bumetanide-in-the-treatment-of-autism-spectrum- Differences in three immuno-behavioural groups. K-means cluster plot on the immuno-behavioural plane. K-means cluster analysis was carried out in Discovery Set (A) and the patients from the Validation Set were mapped to this immuno-behavioural plane (B). C Radar chart for the ratios of changes to baseline of CARS and cytokine levels in 3 immuno-behavioural groups. D Boxplot for the significant changes of CARS and cytokine levels in 3 immuno-behavioural groups. disorders-in-children-and-adolescents/]. To date, the detailed results of these Phase 3 trials have not been published, we could not learn more details about the operation and the outcomes. Considering the nature of heterogeneity in the aetiology and pathology of ASD, the development of biological first-line treatment may only be effective in certain subtypes. Therefore, an important direction of bumetanide treatment is to identify potential biomarkers for a high responsiveness to this drug, and thereby to identify those patients who are more likely to benefit from this treatment. Indeed, our findings of an association between cytokine levels and the symptom improvement after the bumetanide treatment demonstrated the potential of cytokine levels as blood biomarkers for precision medicine in ASD. Our findings may suggest that the identified canonical score of cytokines had a potential to construct a blood signature for predicting and monitoring the bumetanide treatment in young children with ASD. Accurately identifying patients who are likely to respond positively to bumetanide can facilitate the precision medicine for ASD. Our prediction model based on the cytokine levels before the treatment may provide a potentially new tool for the precision medicine of ASD. Given the inherit heterogeneity of ASD, it is of great clinical value to accurately identify the subgroup of ASD patients that are likely to respond positively to its medical treatments [bib_ref] Defining precision medicine approaches to autism spectrum disorders: concepts and challenges, Loth [/bib_ref]. Multiple factors, including both genetic and environmental factors, could contribute to the heterogeneity of ASD and its response to treatment [bib_ref] Oxytocin-mediated GABA inhibition during delivery attenuates autism pathogenesis in rodent offspring, Tyzio [/bib_ref] [bib_ref] The role of the immune system in autism spectrum disorder, Meltzer [/bib_ref] [bib_ref] Prenatal air pollution influences neurodevelopment and behavior in autism spectrum disorder by..., Frye [/bib_ref]. For example, prenatal insults including maternal infection and subsequent immunological activation during gestation may increase the risk of autism in the child [bib_ref] The role of the immune system in autism spectrum disorder, Meltzer [/bib_ref]. Increased exposure to air pollution during gestation was also associated with abnormalities in mitochondrial metabolism during childhood, which may also increase the risk for ASD [bib_ref] Defining precision medicine approaches to autism spectrum disorders: concepts and challenges, Loth [/bib_ref]. However, our findings suggested that using the cytokine levels improved the prediction of response to the bumetanide treatment for ASD in three folds: (1) The immuno-behavioural covariation enabled the identification of more homogenous subgroup of ASD in terms of response to bumetanide. Using five models and an independent test data set, we demonstrated consistent evidence that the responding group identified by the immuno-behavioural covariation could be significantly better predicted by the baseline information compared with the responder group defined by the CARS alone. (2) Combining the cytokine levels with clinical assessments of CARS, ADOS and SRS before treatment, we achieved a higher accuracy of 84.3% in identifying the ASD children that were likely to respond positively to the bumetanide treatment. (3) The blood cytokine levels are more easily accessible in clinical practice. There are several limitations of our study. Although we had two separate data sets to validate our findings, the sample sizes were limited. Previously a sex difference in the cytokinesymptom association had been reported, but we could not test such sex difference owing to the small numbers of girls with ASD in our sample [bib_ref] Cytokine levels and associations with symptom severity in male and female children..., Masi [/bib_ref]. Hence, future multi-centre, prospective studies with lager sample sizes are necessary to confirm the current findings. Second, the hypothesised molecular mechanism underlying the bumetanide treatment effect for ASD requires causal confirmation in animal studies. In summary, we identified an association between the changes of the cytokine levels and the improvements in symptoms after the bumetanide treatment in young children with ASD, and found that the treatment effect of bumetanide can be better characterised by an immuno-behavioural covariation. This finding may provide new clinically important evidence supporting the hypothesis that immune responses may interact with the mechanism of bumetanide to restore the GABAergic function in ASD. This finding may also have relevance for determining enriched samples of ASD children to participate in novel drug treatment studies of drugs with a similar mode of action to bumetanide, but with potentially greater efficacy and fewer side effects. . Changes in CARS and cytokine levels after the treatment in three immuno-behavioural groups. Best responding group (n immuno-behavioural covariation plane, the models were trained to predict the response to bumetanide for the children with ASD. As described in the main text, the models were trained using the Discovery Set, and tested using the Validation Set. The performances of the classification accuracy in the testing data set were reported in this A Models with the cytokine levels at the baseline for predicting patients with ASD in the best responding group. B Models without the cytokine levels at the baseline for predicting patients with ASD in the best responding group. C Models with the cytokine levels at the baseline for predicting patients with ASD in the least responding group. D Models without the cytokine levels at the baseline for predicting patients with ASD in the least responding group. [fig] Figure S3: No significant difference in the distribution of patients among these three clusters between the Discovery Set and the Validation Set. Besides, no significant difference in the baseline of CARS scores and cytokine levels among these 3 clusters. [/fig] [fig] Figure 3: ROC curve for prediction for the immuno-behaviourally defined responding group. The classifiers included the Oblique Random Forest (ORF) model, Partial Least Squares (PLS) model, Support Vector Machine (SVM) model, sparse Linear Discriminant Analysis (sLDA) model and Neural Networks (NN) model. Based on the [/fig] [table] Table 1: The demographic and clinical (mean(SD)) characteristics of three data sets.BMI body mass index, CARS Childhood Autism Rating Scale, ADOS the Autism Diagnostic Observation Schedule, SRS the Social Responsiveness Scale, CARS_total CARS total score, CARS_S CARS score on social impairment domain, CARS_N CARS score on negative emotionality domain, CARS_D CARS score on distorted sensory response domain, ADOS_S ADOS score on social interaction, ADOS_C ADOS score on communication, ADOS_P ADOS score on play, ADOS_I ADOS score on imaginative use of materials, SRS_AWA SRS score on social awareness, SRS_COG SRS score on social cognition, SRS_COM SRS score on social communication, SRS_MOT SRS score on social motivation, SRS_MANN SRS score on autistic mannerism, SRS_total SRS total score.Mann-Whitney U-test for non-normal features, while chi-square test for sex.Sample size for ADOS data in Discovery Set, Validation Set and Control Set are 36, 41 and 36. [/table]

Vinyl azides in organic synthesis: an overview Among organic azides, vinyl azides have attracted significant attention, because of their unique properties in organic synthesis, which led to reports of many types of research on this versatile conjugated azide in recent years. This magical precursor can also be converted into intermediates such as iminyl radicals, 2H-azirines, iminyl metal complexes, nitrilium ions, and iminyl ions, making this compound useful in heterocycle synthesis. # Introduction Vinyl azides are versatile building blocks in which a double bond is attached directly to the azide group.One reason for the great importance of vinyl azides in organic chemistry is the application of this compound to the synthesis of a large number of nitrogen-containing heterocycles, such as pyrazoles, pyrroles, imidazoles, thiazoles, triazoles, pyridines, quinolines, isoquinolines, and imidazo[1,2-a]pyridines [fig_ref] 1: Palladacycle 69 underwent migratory insertion with vinyl azide, and then release of... [/fig_ref].There are different types of cyclization pathway for vinyl azides, like the azirine pathway, carbodiimide pathway, and initial alkenyl group, which can be seen in Scheme 1.The versatility of vinyl azides in organic synthesis has long been known. Photolysis, thermolysis, cycloaddition, and reaction with nucleophiles and electrophiles are common processes that vinyl azides can go through. The double bond in this compound also causes the azide to be more reactive with other functional groups. It is found that the addition of an external radical species causes the production of an iminyl radical that undergoes numerous transformations [fig_ref] Figure 2: Conversion of vinyl azide to different intermediates [/fig_ref]. a-Substituted vinyl azides have been utilized as radical acceptors to produce a variety of molecules, including a-triuoro methylated ketones, a-triuoromethyl azines, keto sulfones, a-azido styrene, cyclic bamino ketones, N-unprotected enamines, enaminones, b-keto phosphine oxides, unsymmetrical ketones, and 2aryacetophenones.It has also been observed that vinyl azides can serve as multipurpose precursors for the synthesis of various units through diverse cleavages. Units generated from different cleavages are shown in [fig_ref] Figure 3: Vinyl azide as a multipurpose precursor [/fig_ref].Vinyl azide decomposition has been the subject of several investigations. Thermal decomposition of vinyl azides gives different products (depending on substituents): in most cases, internal vinyl azides give azirine intermediates.It is found that thermolysis and photolysis of internal vinyl azides produce azirines in good yields, and sometimes with a small amount of iminoketenes (Scheme 2). In rare cases, ketene imines are major products of this type of reaction.In Scheme 3 two mechanistic pathways for the transformation of vinyl azide to azirine (A / B) are shown. The vinyl azides can either release nitrogen to generate vinyl nitrene and then cyclize to the azirine (path a), or can decompose with simultaneous ring closure to generate B directly (pathway b).Smolinsky, who rst pyrolyzed vinyl azides to azirines, considered the possibility that vinyl azides would decompose to azirine via an unstable isotriazole intermediate (path C) (Scheme 4).A computational study on vinyl azide decomposition was also done by da Silva et al. in 2014. Electronic structural calculations showed that the decomposition of the s-cis conformer of vinyl azide leads to the generation of ketenimine via a single-step conversion: s-cis-CH 2 CHN 3 / CH 2 CNH + N 2 while the transformation of the s-trans conformer to acetonitrile happens in two steps: s-trans-CH 2 CHN 3 / cyc-CH 2 NCH + N 2 / CH 3 CN + N 2 . 6 [fig_ref] 1: Palladacycle 69 underwent migratory insertion with vinyl azide, and then release of... [/fig_ref] Examples of compounds produced by vinyl azides. Results of their calculations indicate that the s-cis (øCCNN = 0.0°) and s-trans (øCCNN = 180.0°) conformers of vinyl azide have analogous relative energetic stabilities. Furthermore, in an experimental microwave study, the s-cis conformer was declared to be a bit more stable than the s-trans by 0.460 kcal mol −1 .Because of the very small energy difference between these conformers, fast interconversion is anticipated.According to the calculations, the decomposition of vinyl azides in the singlet state can go through two different reaction pathways, as shown in Scheme 5, beginning from s-cis or s-transvinyl azide. Ketenimine (CH 2 CNH) is generated through singlestep conversion of the s-cis conformer. The transition state for the s-cis-CH 2 CHN 3 / CH 2 CNH reaction (2T) is 37.78 kcal mol −1 (47.92 kcal mol −1 at MP2/6-311++G(d,p) level) higher in energy than s-cis-CH 2 CHN 3. Meanwhile, conversion of the s-trans conformer to acetonitrile happens in two steps: s-trans-CH 2 CHN 3 / cyc-CH 2 NCH + N 2 / CH 3 CN + N 2 . The transition state between s-trans-CH 2 CHN 3 and cyc-CH 2 NCH-(2H-azirine) (5T) is 33.90 kcal mol −1 (41.89 kcal mol −1 at MP2/6-311++G(d,p) level) higher than s-trans-CH 2 CHN 3 , while the transition state between acetonitrile and 2H-azirine (7T) is 49.93 kcal mol −1 (68.55 kcal mol −1 at MP2/6-311++G(d,p) level) higher than 3Hazirine [fig_ref] Figure 4: Full energy profile of the IRC path calculated for the decomposition of... [/fig_ref].From the analysis of da Silva and coworker it can be deduced that N 2 , 2H-azirine, acetonitrile and ketenimine are generated from the decomposition of vinyl azide in the singlet state. It is Scheme 1 Cyclization pathways of vinyl azides. also known that the decomposition mechanism of organic azide includes a nitrene species corresponding to a TS localized at a point of intersystem crossing (ISC) between singlet ( 1 S) and triplet ( 3 S) surfaces.Vinyl azides can also be good precursors for the synthesis of nitrile compounds.In the synthesis of nitrile from terminal vinyl azide, the decomposition of vinyl azide rst produces iminoketene; then, the latter product is tautomerized to the more stable nitrile (Scheme 6).Vinyl nitrene is also proposed as an intermediate for the chemistry of vinyl azides. The former compound can be singlet or triplet [fig_ref] Scheme 5: Multistep reaction for the decomposition of vinyl azide in the singlet state [/fig_ref]. They can also act as a three-atom C-C-N synthon in cycloaddition reactions.In an investigation of the stability of vinyl azides, it is important to pay attention to the point that the azide group in vinyl azide increases the electron density at the b carbon atom. This means, that canonical structure 1C makes an extreme contribution to the overall stabilization of azide with the result that the order of the N 2 -N bond that cleaves on thermolysis is less than 1.5 (Scheme 7). Gerrit Labbe and coworkers showed that vinyl azides exhibit moderate energies of activation and low entropies of activation, consistent with the nitrene pathway.Todeschini and coworker, in 1978, compared the net charge of N 1 in cis and trans forms of vinyl azides. The net charge understood on N 1 (−0.388 in the trans form and −0.396 in the cis form) and on N 3 (−0.299 in trans and −0.294 in cis) conrm the suggestion that the resonance form of covalently bonded azides (R-N]N + ]N − ) is more stable in the trans conformation, while (R-N]N +^N ) is more stable in the cis conformation. 12 ## Formation of 5-membered heterocycles In 2015, Donthiri et al. proposed a novel strategy for the synthesis of substituted 1H-pyrroles 3 through the copper- catalyzed C(sp 3 )-H functionalization of ketones 2 with vinyl azides 1. Their method efficiently accessed a series of 2,3,5trisubstituted-1H-pyrroles in fair to excellent yields (50-93%).Pyrroles are considered as specic heterocycles due to their application in biological systems and their remedial activities. Pharmaceuticals such as potent blockers for potassiumcompetitive acid, anti-tumor agents, the leading cholesterollowering drug Lipitor, and a wide range of natural products are constituted with a pyrrole core structure.Pyrroles also serve as useful building blocks in the synthesis of bioactive compounds (Scheme 8).To explain the mechanism of this reaction, deoxy benzoin 2 in the presence of Cu(II) rst generated a radical intermediate 4, and Cu(II) transformed into Cu(I). Intermediate 4 in the presence of vinyl azides 1, gave iminyl radical intermediate 5. Next, the latter compound instantly underwent a radical addition and gave Cu(II) complex intermediate 6. In the nal step, abstraction of Cu(II) followed by-H shi, gave the nal product 3 (Scheme 9).In 2019, Jiang et al. synthesized substituted pyrroles 9 and 11 by utilizing vinyl azides 1 and terminal alkynes 8 and 10 in the presence of a nano copper catalyst. By using this strategy, the research group synthesized 2,5-disubstituted pyrroles 11 in poor to very good yields (13-83%) and 2,3,4-trisubstituted pyrroles 9 in fair to very good yields (50-82%).In this project, Jiang and coworkers observed switchable reactions by varying the substituent R of the terminal alkyne. This transition-metalcatalyzed C-C and C-N bond formation is a necessary tool in organic synthesis, allowing for the construction of basis backbones which are popular building blocks of active molecules in the material sciences (Scheme 10).To illustrate the mechanism, the vinyl azides 1 underwent thermal decomposition to generate 2H-azirine 12. In pathway A for the generation of the 2,5-disubstituted pyrrole 11, 2H-azirine rst underwent an SET reduction with CuNPs to give iminyl Cu(II) intermediate 13. This C-radical was trapped by terminal alkyne 10 through intermolecular addition to give intermediate 14. Then, the radical was captured by intramolecular copper(II) to generate copper adduct 15, which underwent reductive elimination to reproduce the active copper and gave the 3Hpyrroles. Eventually, 2,5-disubstituted pyrrole 11 was produced through the isomerization of the 3H-pyrrole (Scheme 11). In path B for the synthesis of 2,3,4-trisubstituted pyrrole 9, CuNPs rst reacted with the terminal alkyne 8 to generate copper acetylide 16. Then, 2H-azirine underwent a nucleophilic ring opening with copper acetylide 16, giving iminyl copper intermediate 17, which was protonated to generate the NH imine. In the next step, the iminyl copper intermediate 17 reacted with another 2H-azirine, generating intermediates 18 and 19, which are keto-enol tautomers. The intramolecular cyclization of 18 generated intermediate 20, which underwent protonolysis to release intermediate 21 along with copper. Further elimination of NH 3 and isomerization produced the desired product 9 (Scheme 11).In 2018, Liu et al. utilized a copper-catalyzed dark reaction with eosin Y to produce ene-g lactams 23 (Scheme 12). Eosin Y is a common organo-photocatalyst in visible-light photo-redox processes. It showed excellent catalytic activities for thermal redox reactions in the presence of a catalytic amount of Cu(OAc) 2 . With this catalytic system, ketene silyl acetals 22 and vinyl azides 1 combined through [3 + 2]-cycloadducts, resulting in the production of ene-g-lactams 23 in fair to excellent yields (57-96%).Ene-g-lactams are important synthons as they are broadly used in the synthesis and key structural elements of many bioactive natural products and medicinally relevant compounds.The electron-rich enamine scaffolds of ene-glactams, as highly versatile intermediates, could take part in a hetero-Diels-Alder 22 reaction or [3 + 2]-cyclization with a,bunsaturated carbonyl compounds, producing tetrahydropyranopyrazole or enantioenriched bicyclic g-lactams. Ene-glactams could also react with electrophilic aldehydes to produce fused bicycles.To explain the mechanism, the initial reaction between Cu(OAc) 2 and eosin Y would give an eosin Y radical cation, which would be capable of oxidizing the ketene silyl acetal through single electron transfer (SET) to a-ester radicals. Addition of a-ester radicals to vinyl azides 1 would give the iminyl radical and dinitrogen. The resulting iminyl radicals should aerward be reduced by low-valent Cu(I), thus generating iminyl anions and regenerating Cu(II). Eventually, the intramolecular nucleophilic substitution of iminyl anion to ester aer isomerization produced nal products 23 (Scheme 13).In 2019, Guo et al. reported Ni-catalyzed ring-opening/radical addition/ring-closing of cycloketone oxime esters and vinyl azides (Scheme 14). Their protocol resulted in fast access to cyano-alkylated 3,4-dihydro-2H-pyrroles 25 (ve-membered ring) in fair to good yields (56-77%).To describe the mechanism, Ni-mediated single-electron reduction of cyclobutanone oxime ester 24 rst generated iminyl radical 26, which underwent C-C bond cleavage and led to the formation of cyano-alkyl radical 27. The latter radical reacted with the C]C bond of vinyl azide, producing iminyl radical 28 by releasing N 2 . For aryl vinyl azides, iminyl radical 28 proceeded through 1,5-H-transfer to generate radical 29. Oxidation, cyclization, and then deprotonation of 29 gave desired product 25 (Scheme 15).In 2019, Zhang et al. achieved novel 2H-pyrrol-2-imine derivatives 32 in a one-pot process through a rhodiumcatalyzed reaction of vinyl azides 1 and two equivalent Scheme 11 Mechanism for the Cu-catalyzed synthesis of pyrroles 9 and 11 from vinyl azides 1 and alkyne. amounts of isocyanide 31 (Scheme 16). Their method offered facile access to 3-amino-5-aryl-2H-pyrrol-2-imines bearing different substitutions on the nitrogen in poor to excellent yields (16-96%). 25 2H-Pyrrole is a signicant structural motif present in a broad range of natural products and biologically active compounds.A plausible mechanism consists of two processes: Rh(I)catalyzed cross-coupling of vinyl azides 1 with the rst isocyanide 31, and the cyclization of the vinyl carbodiimide with the second isocyanide. Vinyl carbodiimide 36 was rst generated through an Rh-nitrene pathway with the release of N 2 . In the next step, thermal cyclization of the second isocyanide generated 2H-pyrrol-2-imine 32 as the product (Scheme 17).In 2014, Bairagi et al. came up with a new approach to synthesize imidazole heterocycles (Scheme 18). In this study, by utilizing copper-catalyzed C-H functionalization of pyridine and isoquinoline, the research group synthesized derivatives of 39 in poor to good yields (29-74%) and 41 in poor to good yields (48-76%).The development of an efficient strategy for the synthesis of azaheterocycles via direct functionalization of C-H bonds utilizing transition metal catalysts is a highly important subject in organic chemistry. Imidazo[1,2-a]pyridines 39 are aza heterocycles that have received a lot of attention due to their diverse biological activities.Scheme 19 provides plausible information about the mechanism of the reaction. Vinyl azides 1 rst underwent thermal decomposition to generate the 2H-azirine, which gave imi-Pyrroles are a signicant class of heterocyclic compound; they are also building blocks in many natural products, 32 synthetic pharmaceutical compounds, 15 and material science.Compared with previous acidic or basic conditions for poly-substituted pyrrole synthesis, the conditions were mild, neutral, and did not involve any additive. They resulted in Mechanism for the synthesis of cyano-alkylated 3,4dihydro-2H-pyrroles 25 from vinyl azide 1. the synthesis of 2,4-disubstituted pyrrole in poor to good yields (39-79%) and 3,4-disubstituted pyrrole in poor to very good yields (40-82%).For the Cu-catalyzed formation of 2,4-disubstituted pyrrole 47 (path A), Cu(OAc) 2 might rst be reduced by DMSO/EtOH to generate Cu(I) or through disproportionation to generate Cu(I). In the next step, radical intermediates 49 were produced through a denitrogenative decomposition of the vinyl azides 1. Then, the radical coupling of intermediate 49 with enol tautomers of the phenylacetaldehydes produced intermediates 50, which underwent nucleophilic attack on the aldehyde, leading to the addition of intermediates 51. Subsequent protonation gave intermediates 52, and dehydration of intermediates 52 produced 2,4-disubstituted pyrroles 47 (Scheme 21). For the Ni-catalyzed production of 3,4-disubstituted pyrrole 48 (path B), NiCl 2 rst promoted the decomposition of vinyl azides 1 to give 2H-azirines 56, which could not be generated in the presence of a Cu catalyst. In the next step, nucleophilic attack by the enol tautomers of phenylacetaldehydes generated intermediates 58. Ring-opening of intermediates 58 gave vemembered species 59. b-OH elimination gave the 2H-pyrroles 60 and Ni complexes, which reacted with the aldehyde to regenerate enol intermediate 57. In the nal step, 3,4-disubstituted pyrroles 48 were produced through tautomerization of intermediates 60 (Scheme 21).The eld of donor-acceptor cyclopropane chemistry has seen signicant resurgence since the early work of Wenkert, Danishefsky, and Reissig. In 2016, Kerr et al. a Mannich-style ring closure to produce 62. Option B involved an azirine formation by the thermolysis of the vinyl azide through a putative nitrene 63. Ring opening of the cyclopropane by nucleophilic nitrogen gave the same intermediate 65. Eventually, option C has the vinyl azide, itself, nucleophilically opening cyclopropane 61 to give intermediate 66.In 2018, Cui et al. synthesized diverse 2-arylindole 68 derivatives with the Pd(II)-catalyzed cyclization of anilines with vinyl azides 1 (Scheme 24).Indoles are a key structural unit in a large number of synthetic drugs and are widely found in natural products.Indole and its derivatives can also be used in the eld of skeletal modication. In this study, the research group developed a palladium-catalyzed highly regioselective cyclization reaction of vinyl azide and aniline by utilizing pyridine as a directing group to produce 2-arylindole derivatives in poor to very good yields (40-84%).To explain the plausible mechanism of this reaction, a chelation-assisted C-H bond activation rst took place via a concerted metalation-deprotonation process to generate the six-membered cyclopalladated intermediate 69. In the next step, two pathways were probably involved in the following transformation.Finally, they removed the 2-pyridyl directing group from compound 68. 2-Arylindole 68 was treated with methyl tri-uoromethanesulfonate (MeOTf) to generate a pyridinium intermediate and NH-free indole 76 in 89% yield (Scheme 26).In 2008, Narasaka et al. reached various derivatives of polysubstituted N-H pyrroles through the reaction of vinyl azides and 1,3-dicarbonyl compounds. They proposed two synthetic methods for the synthesis of tetra-and tri-substituted N-H pyrroles that resulted in the production of scaffolds of 80 with good to excellent yields (74-96%) and scaffolds of 81 in poor to very good yields (43-88%). In method (I) they used the reaction of vinyl azide 77 with 1,3-dicarbonyl 78 compounds that proceed through 1,2-addition of 1,3-dicarbonyl compounds to 2H-azirin, the research group used Cu(II)-catalyzed synthesis of pyrroles from vinyl azide 77 and ethyl acetoacetate 79 via the 1,4-addition reaction (Scheme 27).In method (I) they designed a plan to use vinyl azides 77 as precursors of 2H-azirines. When a mixture of vinyl azides 77 and acetylacetone 78 was heated in toluene at 85-100°C (depending on its derivatives), substituted pyrrole was obtained in good to excellent yields (74-96%) (Scheme 28).In method (II) since the Cu-catalytic reaction was done at 40-60°C, it was improbable that a 2H-azirine intermediate would be found in the reaction course. The reaction may be started by the 1,4-addition of copper enolate 85 to vinyl azide 77, the internal nitrogen of which is coordinated to copper. Simultaneous removal of dinitrogen obtained alkylidene amino copper 86 that underwent intramolecular nucleophilic attack on the carbonyl group, producing pyrrole with the extrusion of water (Scheme 29).In 2017, Jiang et al. proposed an approach to produce 2Himidazoles 91 from oxime acetates 90 and vinyl azides 1 under redox-neutral conditions (Scheme 30). They developed a novel method for the synthesis of 2H-imidazoles via iron-catalyzed [3 + 2]-annulation of oxime acetates with vinyl azides that resulted in a synthesis of scaffolds of 91 in fair to excellent yields (53-90%). 38 2H-Imidazoles are signicant motifs found in a broad range of natural products 39 as well as in organic synthesis. These compounds are used for detection of diradical intermediatesand could also be used as a chiral auxiliary for organic synthesis reactions.Oximes and their derivatives can be prepared by the reaction of hydroxylamine hydrochloride, ketone, and acetic anhydride.To explain the possible mechanism, imine anion intermediate 92 was rst generated through the reduction of oxime acetate 90 by Fe(II) with a two-step, single-electron-transfer operation. In the nal step, the desired product was obtained via a sequential oxidation deprotonation process (path B) (Scheme 31).In 2017, Ramachary et al. reported the synthesis of fully decorated N-vinyl-1,2,3-triazoles 98 scaffolds in good to excellent yields (50-95%) through an organocatalytic vinyl azidecarbonyl [3 + 2]-cycloaddition.Triazoles have an important role in pharmaceutical chemistry and their drug properties mainly depend on their aromatic or aliphatic substitution.Tazobactam, solithromycin, cefatrizine, and runamide are some examples of 1,2,3-triazole-based drugs (Scheme 32).Reaction of the aryl acetaldehydes/aryl acetones/ deoxybenzoines with catalyst DBU in DMSO (solvent) gave the stable enolate that on in situ treatment with vinyl-azides 1 produced 1,2,3-triazolines by [3 + 2]-cycloaddition (Scheme 33).In 2017, 1 by aryldiazonium salts 101 (Scheme 34). By this strategy, Tang and coworkers succeeded in the synthesis of diverse N 2substituted 1,2,3-triazoles in poor to very good yields (35-82%).1,2,3-Triazoles have played an important role in material science, pharmacological development, and chemical biology due to their privileged chemical structures and properties as ve-membered heterocycles. 1,2,3-Triazoles have also received interest as chelates, 46 auxiliaries for C-H activation,and systematic carbene precursors. To illustrate the mechanism of this reaction, the aryl cation 103 was rst delivered from aryldiazonium cation 101 with the release of nitrogen under acidic conditions. Copper-chelated complex 104 that was generated from vinyl azides 1 was attacked by aryl cation 103 to generate iminyl metal ions 106,In 2020, Weng et al. developed a novel approach for the synthesis of poly-substituted 5-triuoromethyl isoxazole 113 and its derivatives (Scheme 36). The research group, through denitrogenative cyclization of vinyl azides with triuoroacetic anhydride, synthesized diverse 5-triuoromethyl isoxazoles in poor to quantitative yields (25-99%).Nitrogen-and oxygencontaining heterocycles, due to the exhibition of biological and pharmaceutical activities, such as antimicrobial, antibiotic, anti-inammatory, and anticancer activities have been receiving a lot of attention in the past few years.In the same year, Weng et al. synthesized 5-diuoromethyl isoxazoles 115 and 5-chlorodiuoromethyl isoxazole 117 via the reaction between vinyl azides 1 and diuoroacetic anhydride 114 or chlorodiuoroacetic anhydride 116 (Scheme 37). Their method involved sequential diuoroacetylation of vinyl azide, followed by deprotonation, cyclization, and dinitrogen elimination, providing a suitable synthesis of 5-diuoromethyl in poor to good yields (19-75%) and 5-chlorodiuoromethyl isoxazole in poor to very good yields (35-86%). 51 Among organo-uorine molecules, diuoromethylated compounds, especially diuoromethylated heterocyclic structures, play a signicant and unique role in agricultural and medicinal chemistry. 52 Diuoromethyl (-CF 2 H) behaves in an isopolar and isosteric way to the hydroxyl (-OH) group and can act as a hydrogen donor via hydrogen bonding.In Scheme 38, diuoroacetylation of vinyl azides 1 with diuoroacetic anhydride 114 rst generated intermediate 118, which underwent deprotonation with a base to give intermediate 119. Isomerization of 119 to an alkoxy anion 120 followed by a 5-endo cyclization gave intermediate 121. In a nal step, intermediate 121 underwent a dinitrogen elimination to produce desired product 115.In 2015, Bi et al. introduced a new strategy to synthesize dihydrofuran-2(3H)-ones 124 via the synthesis of 4-ynamides 123 (Scheme 39). The nucleophilic addition of vinyl azides 1 to propargylic alcohols 122 in the presence of BF 3 $Et 2 O catalysis is an efficient method that produced 4-ynamides in fair to excellent yields (62-90%). Then, a Vilsmeier intramolecular cyclization of 4-ynamides 123 resulted in good to very good yields (73-84%) of dihydrofuran-2(3H)-ones 124, which was the rst report of the conversion of alkyne to dihydrofurane-2(3H)-ones by the use of a Vilsmeier reagent.In a plausible mechanism, electrophilic attack of BFbiomedical elds. In recent years, many drugs that are listed as best-selling drugs contain sulfur.For example, cefmenoxime, 56 a bis-S-heterocyclic, is an antibiotic with a high curative effect on special kinds of bacteria, and nocodazole, 57 a thiophene structure, is an anticancer drug that can inhibit cancer cell division. The research team in this work employed a concise synthetic mode to pick up the thiophene and thiazole ring in a one-pot method, in which S 8 was utilized as a two-sulfur donor.In 2015, Yu et al. reported selective access to 4-substituted 2aminothiazoles 138 and 4-substituted 5-thiocyano-2aminothiazoles 139 from potassium thiocyanate 137 and vinyl azides 1 switched by palladium and iron catalysts (Scheme 42). Yu and co-workers constructed diverse derivatives of 4substituted 2-aminothiazoles in poor to excellent yields (42-94%) and 5-thiocyano-2-aminothiazoles in poor to excellent yields (49-95%).2-Aminothiazoles and their derivatives are one of the most signicant aza-heterocycles broadly found in pharmaceutical compounds and natural products. The wide spectrum of biological activities shown by this structure include antimicrobial, 59 anticancer, 60 antiviral and antiprion, and anti-inammatory effects. The most widely utilized synthetic method to gain 2-aminothiazoles is the Hantzsch cyclocondensation of thiourea and a-halo carbonyl compounds.Many other methods to synthesize 2-aminothiazoles have also been expanded.For the Pd(OAc) 2 -catalyzed reaction (path A), the rst coordination of palladium(II) to the azide group gave palladium(II)azide complex 140, expanding the electrophilicity of pendant Scheme 43 Mechanism for iron-and palladium-catalyzed reaction of vinyl azides 1 and potassium thiocyanate 137. [fig_ref] Table 1: Synthesis of five-membered heterocycles through vinyl azides [/fig_ref] was prepared to provide a summary of the strategies the research groups utilized to synthesize ve-membered heterocycle derivatives in which the catalyst, solvent, temperature, yields, and MW irradiation are compared. ## Formation of 6-membered heterocycles In 2018, Guo et al. approached a great method to synthesize substituted phenanthridines 153 (Scheme 44). Guo's research group, by using metal-free, visible-light-promoted decarboxylative radical cyclization of N-acryloxy phthalimides 152 and vinyl azides 151 synthesized diverse phenanthridines in poor to very good yields (32-80%).Phenanthridines are in a signicant class of alkaloids that show remarkable biological activities and optoelectronic properties, so many efforts have been devoted to developing efficient strategies for the synthesis of these motifs.Photoexcitation of eosin Y by visible light generated excited eosin Y*. Oxidative quenching of eosin Y* by single electron transfer (SET) to an NHP ester gave radical anion 154 along with eosin Y .+ . A phthalimide anion and the corresponding alkyl radical 155, were generated through fragmentation of CO 2 from 154. Then, the phthalimide anion and alkyl radical 155 added to the C]C bond of 152, furnishing iminyl radical 156 by releasing N 2 . Iminyl radical 156 underwent an intramolecular cyclization to give intermediate 157, which was oxidized by eosin Y .+ to generate the corresponding carbocation 158. In the last step, deprotonation of intermediate 158 gave the desired product 153 (Scheme 45).In 2021, Wang et al. reported a novel method to synthesize 2,4-diaryl-6-triuoromethylated pyridines 160 through coppercatalyzed cyclization of CF 3 -ynones 159 and vinyl azides 1 (Scheme 46). Their procedure for this transformation led to the production of diverse 2,4-diaryl-6-triuoromethylated pyridines in poor to fair yields (36-65%).Pyridine derivatives are versatile precursors of different topoisomerase I inhibitors and anticancer agents.The pyridine skeletons containing tri-uoromethyl motifs, like 2-triuoromethylated pyridines, have emerged as core units in an increasing number of important agrochemicals and drugs.To explain the mechanism of Wang and coworkers' methods, the reaction of vinyl azides 1 with PPh 3 rst gave vinyl iminophosphorane 161 through the Staudinger-Meyer reaction. ThenIn 2019, Guo et al. proposed a new method by utilizing cycloketone oxime esters 24 and vinyl azides 1 to produce substituted phenanthridines 172 (Scheme 48). This reaction proceeded under mild and redox-neutral conditions with a wide substrate scope that resulted in the synthesis of diverse phenanthridines in fair to very good yields (63-81%).In the plausible mechanism, Ni-mediated single-electron reduction of cyclobutanone oxime ester rst generated iminyl radical 26, which underwent C-C bond cleavage producing cyanoalkyl radical 27. In the next step, radical 27, added to the C]C bond of vinyl azide, generated iminyl radical 28 by releasing N 2 . For biaryl vinyl azides, iminyl radical 28 underwent oxidative cyclization to afford the nal product 172 (Scheme 49).In 2016 yields (26-93%). Zhou's research group also utilized readily available sulfonyl chloride 183 as a sulfonylation reagent.In 2016, Liu et al. synthesized diverse triuoromethyl isoquinolines 187 by utilizing multicomponent cascade synthesis from alkynes 186 and vinyl azides 1 (Scheme 53). The research group used Togni's reagent 185 as a CF 3 radical supplier; they also used the Rh(III)-Cu(II) bimetallic system for their methods that resulted in the production of triuoromethyl isoquinolines in poor to very good yields (21-85%).The triuoromethyl group is highly important among uorine functional groups because of its remarkable potential for modulating a molecule's chemical, physical, and biochemical properties.Togni's reagent has been utilized to produce triuoromethylate compounds, usually in the presence of Cu salts as catalysts.The mechanism of this reaction was proposed as depicted in Scheme 54. In step 1, CF 3 radicals may rst be generated from Togni's reagent in the presence of Cu(II). The CF 3 radical may be added to vinyl azides 1 to generate hypothetical radical intermediate 188, which could then be trapped by Cu(II) to generate 189 Various substance scopes and specic regioselectivities were shown in this reaction. As regards the broad application of alicyclic[b]-fused pyridine, it is expected this Pd-catalyzed regioselective C-H activation will gain wide synthetic usage.In the plausible mechanism of the reaction, Pd(II)-catalyzed C-H activation with a,b-unsaturated hydrazone rst formed In the last step, a-carbolines 204 were produced through a 1,3-H shi and oxidative aromatization process.In 2018, Zhang et al. developed a new and efficient method by utilizing vinyl azides 1, isonitriles 31, and alkynes 213 in the presence of Rh as catalyst to synthesize various derivatives of pyridine and isoquinoline in poor to very good yields (17-80%) (Scheme 59).The cascade cyclization was further developed from alkynes to benzynes 214, which afforded aminoisoquinoline 215 as the product in poor to good yields (20-74%) (Scheme 60).Aer the generation of vinyl carbodiimide 36, from the reaction of vinyl azides 1 and isonitrile 31 two pathways were possible: direct electrocyclization or Rh(I)-catalyzed oxidative/ cyclization/reductive elimination (Scheme 61).This research group also used vinyl carbodiimides for the synthesis of different heterocycles. Other popular cyclic building blocks were employed to construct different azaheterocycles. When vinyl carbodiimide reacted with an allene 219, aminopyridine 220 was obtained in 74% yield. The reaction of vinyl carbodiimide with benzoquinone 221 led to the formation of aminoisoquinoline-5,8-dione 222 in 65% yield. The vinyl carbodiimide intermediate 36 could also go through an a,ainsertion reaction with another isocyanide to produce pyrrole-2imine 224 with 30% yield (Scheme 62).In 2016 that are widely utilized as solvents in organic synthesis and the chemical industry.Phenanthridines are an important class of alkaloids that, due to their remarkable biological activities and optoelectronic properties, have attracted the interest of chemists. The research group, in this protocol, presented a cyclization process involving the capture of an iminyl radical by the intramolecular aryl ring, which provided a distinctive idea for constructing the phenanthridine framework.To illustrate the mechanism, metal-mediated single electron transfer (SET) or thermal homolytic cleavage or oxidation of TBPB rst yielded the tert-butoxy radical. Subsequent hydrogen abstraction from toluene by the tert-butoxy radical created benzyl radical 290. Then, radical 230 added to the double bond of vinyl azide, producing an iminyl radical with the release of dinitrogen. Aer that, radical 230 underwent an intramolecular cyclization to create radical 231, which was converted to the corresponding carbocation by Cu(II) via single-electron oxidation with subsequent loss of H + , regenerating Cu(I) and producing desired product 226 (Scheme 64).In 2018, Jiang et al. synthesized 4-substituted quinolines 233 with vinyl azides 1 as dual synthons through C-N and C]C bond cleavage in fair to very good yields (51-84%). In this protocol, vinyl azides acted as dual synthons through C]C and C-N bond cleavage with the formation of two C]C bonds and one C]N bond (Scheme 65). 2 As a new strategy to further develop MCRs, the dual-synthon approach, which uses one reactant to obtain multiple fragments, has attracted a lot of attention.The mechanism of this reaction can be seen in Scheme 66. . This study focused on the use of cyclopropanols 241 as precursors of b-carbonyl radicals and reactions of vinyl azides. In this protocol, utilization of Mn(acac) 3 was found to be essential because it might play a dual role in the oxidation of cyclopropanol and dihydropyridine 250 (Scheme 67).The reaction was initiated by the addition of b-keto radical 245, which was produced by one-electron oxidation of 241 by Mn(III) to vinyl azide, generating iminyl radical 246 with the release of dinitrogen. Cyclization of iminyl radical 246 to an intramolecular carbonyl group would generate alkoxy radical 247 that can be reduced by Mn(II) and then protonated to give tetrahydropyridine 249. Dehydration of 249 and further oxidation producing dihydropyridine 250 would produce desired pyridine 242 (Scheme 68).Treatment of chiral bicyclic cyclopropanol 243 with vinyl azides 1 provided racemic 244. Generation of ring-expanded bketo radical 251 from bicyclic cyclopropnol followed by radical addition of 251 to vinyl azides 1 was probable in the reaction mechanism (Scheme 69).In 2015 the carbon-metal bond, and elimination led to nal product 254 (Scheme 71). 77 [fig_ref] Table 2: Synthesis of six-membered heterocycles through vinyl azides [/fig_ref] provides a summary of the strategies the researchers utilized to synthesize six-membered heterocycles derivatives in which the catalyst, solvent, temperature, yields, and MW irradiation have been compared. ## Others In 2020, Mukherjee et al. developed an efficient method for the enantioselective a-alkylation of amides using vinyl azides 1 as an enolate surrogate (Scheme 72). By this strategy, Mukherjee and coworkers succeeded in synthesizing derivatives of 264 in poor to good yields (40-75%). Among the unstabilized enolates utilized as nucleophiles in iridium-catalyzed asymmetric allylic alkylation reactions, amide enolates are the least explored. In this study, vinyl azides 1 were employed as amide enolate surrogates in Ir-catalyzed asymmetric allylic alkylation with branched allylic alcohols 263 as the allylic electrophile.In the possible mechanism, the nucleophilic addition of vinyl azides 1 to the more substituted terminal of an electrophilic p-allyl-Ir complex was expected to give branched imino diazonium ion 266, which can generate a mixture of (E)-and (Z)isomers. Schmidt-type rearrangement of (E)-266 through 1,2aryl migration followed by hydration of the resulting nitrilium ion 267 would then furnish the a-functionalized acetamide 264. A competitive pathway involving (Z)-266 favored 1,2-alkyl migration to generate an isomeric nitrilium ion 268 and nally N-homo-allylbenzamide derivatives 269. Controlling the geometry of 266 was critical to ensure the desired aryl migration. Another competitive pathway contained the hydrolysis of iminodiazonium ion 266 to the a-allyl ketone 270. The suppression of the latter two pathways has been the key to the success of this method (Scheme 73).In 2021, O. Terent'ev et al. reported the photo-redoxcatalyzed synthesis of N-unsubstituted enamino sulfones 272 from vinyl azides 1 and sulnates 271 (Scheme 74). The research group used ethanol as solvent and eosin Y as photocatalyst in combination with nitrobenzene as an electron shuttle. In this strategy, elimination of N 2 from vinyl azides allowed the energetically favorable production of enamine derivatives in poor to fair yields (38-68%). 80 Formation of N-unsubstituted enamine derivatives through radical addition to vinyl azides 1 is depicted in Scheme 75. [bib_ref] Terent'ev, Mulina [/bib_ref] In 2020, O. Terent'ev et al. reported a cerium(IV) ammonium nitrate promoted synthesis of O-phthalimide oximes 278 from N-hydroxy phthalimide 277 and vinyl azides 1 (Scheme 76). The disclosed protocol was based on the radical transformation of vinyl azides with the release of nitrogen and the formation of iminyl N-radicals that led to the formation of O-phthalimide oximes in fair to very good yields (54-88%). 81 N-oxyl radicals are broadly used in biological and organic chemistry and material science.In organic synthesis, more stable nitroxyl radicals are utilized as a catalyst for the oxidation of alcohol and free radical scavengers.The plausible mechanism started with the formation of a PINO radical 279 from NHPI 277 under the action of CAN, followed by addition to the terminal carbon of the C]C bond of vinyl azides 1. Nitrogen elimination from 280 occurred with the generation of iminyl radical 281. In the last step, radical 281 was intercepted by the PINO radical to form the nal product 278 (Scheme 77). [bib_ref] Terent'ev, Paveliev [/bib_ref] In 2020, Fan et al. proposed a new method to synthesize aamido ketone 283 through the cascade reaction of carboxylic acid 282 with vinyl azides 1 under catalyst-free conditions (Scheme 78). This strategy led to the formation of diverse aamido ketones in fair to very good yields (58-88%). 85 a-Amido ketone derivatives have attracted attention due to their being not only necessary motifs of a plethora of pharmaceutically active compounds but also essential intermediates that are broadly utilized in organic synthesis.To explain the mechanism, the vinyl azides 1 were rst decomposed under the reaction conditions to generate an azirine intermediate 12. Then, 12 reacted with carboxylic acid 282 to generate an azirine 284. Eventually, the unstable aziridine 284 through a thermal rearrangement produced a-amido ketone 283 (Scheme 79).In 2017, Yu et al. synthesized b-keto phosphine oxides 286 through Mn(III)-catalyzed phosphorylation of vinyl azides 1 in fair to excellent yields (65-92%).Organophosphorus compounds play an important role in material science, organic chemistry, and pharmaceuticals.In recent years, protocols between organo-phosphorous radicals and radical acceptors have been expanded for producing organo-phosphorous compounds, especially for the production of b-keto phosphine oxides (Scheme 80).The mechanism of this reaction might be initiated by the addition of phosphine radical 288, produced by one-electron oxidation of 287 by Mn(III) to vinyl aizide, generating iminyl radical 289 with the release of N 2 . Sequentially, the produced iminyl radical 289 was reduced by Mn(II) and then protonated to .The triuoromethyl group (CF 3 ) appears to have wide applications in pharmaceuticals, such as meoquine (antimalaria), 92 efavirenz (anti-HIV),and sorafenib (anti-cancer).In the possible mechanism, vinyl azides 1 rst complexed with Cu(II) to give an N-diazo enamine copper 294, which comfortably trapped electrophile 292 by nucleophilic addition to generateThe mechanism of this reaction can be seen in Scheme 85. First, TBHP underwent a single electron transfer (SET) to give t-BuOc radical and/or t-BuOOc radical by the action of copper ions. Then, benzoyl radical 299 was generated via the processes of deprotonation and oxidation of benzaldehyde 135 in the vicinity of the t-BuOc radical and/or t-BuOOc radical, which then attacked the terminal carbon of the C]C bond of a-phenyl vinyl azide, affording iminyl radical 300 with the release of nitrogen. The generated radical 300 then underwent isomerization to produce enaminyl radical 301, which recombined with radical 299 to give nal product 298.In 2022, Dandela et al. published the iodine-TBHP mediated synthesis of a-ketoamide 303 through the reaction of vinyl azides 1 and amines 302 in fair to very good yields (57-86%). This reaction can be seen in Scheme 86. 5 a-Ketoamides and their derivatives are present in a variety of natural products, biologically active molecules such as antitumor, anti-IBD,antiviral, 99 anti-HIV,and antibacterial drugs, and functional materials. The a-ketoamide moieties are also present in diverse pharmacologically interesting compounds, as shown in [fig_ref] Figure 7: Examples of biologically active molecules containing aketoamides [/fig_ref]. 5 In the possible mechanism, in the presence of I 2 , N-unsubstituted imine 306 would rst be generated from a-aryl vinyl azides, with the release of N 2 .In 2017, Bi et al. synthesized N-unprotected enamines 311, 312, and 313 through radical enamination of vinyl azides 1 in fair to excellent yields (63-91%), poor to fair yields (37-62%), and fair to very good yields (60-87%), respectively (Scheme 88). In this work, the research group focused on establishing the relationship between enamines and vinyl azides. It was found that an electron-withdrawing-group-generable radical induced enamination of vinyl azides, which resulted in different types of b-functionalized N-unprotected enamines. The research team eventually found this goal using acyl, nitro, and sulfonyl radicals, thereby providing a general way to access different types of b-functionalized N-unprotected enamines. To evaluate the scalability of this strategy, Bi and coworkers further transformed the b-functionalized primary enamines to many bioactive molecules as useful building blocks according to In 2019, Singh et al. developed a novel method for the synthesis of 3-oxoisoindoline-1-acetamides 323 (Scheme 90). This reaction progressed at ambient temperature with a broad range of 3-hydroxy isoindole-1-ones and vinyl azides and resulted in the production of 3-oxoisoindoline-1-actamideds in fair to excellent yields (67-97%).In modern pharmaceuticals and biologically active compounds, the amide functionality is omnipresent. It has also been found that small organic molecules containing methylene amide linkages are in many bioactive compounds, drug leads, and chemical probes.On the other hand, isoindolinones are the key structural unit present in numerous synthetically useful or naturally bioactive compounds, as shown in [fig_ref] Figure 8: Some biologically active isoindolinone derivatives [/fig_ref].In the mechanism of this reaction, N-acyl iminium salt intermediate 324 was rst produced from 3-aryl-3hydroxyisoindolineones, when BFIn 2020, Szpilman et al. reported the indium(III)-catalyzed reaction of indole 327 and vinyl azides 1, which led to the production of derivatives of 328 in fair to excellent yields (61-96%). In this strategy, another indole also reacted with 328 and produced 329 in fair to excellent yields (66-96%) (Scheme 92). This is the rst displacement of the azide group by a carbon nucleophile while keeping the vinyl part. Extraordinarily, the substitution of azide on the sp 2 carbon with the retention of the alkene system has not yet been disclosed. It would also be an attractive type of reactivity and would pioneer many new possibilities in chemistry.In such a process, it would be necessary to reversibly attach an electrophilic catalyst, e.g., a Lewis acid, at the a-position of vinyl azides 1, generating an electrophilic diazoimine species 331. The initial step would allow a nucleophile-like indole to attack the azide-bearing carbon in 331 with the generation of 333. E 2 elimination of indium trichloride and azide anion 335 would lead to the generation of charged species 334 and close the catalytic cycle. This step would likely be irreversible. In the nal step, the azide anion would abstract a proton from 334 to generate the product vinyl indole 328 and hydrazoic acid (Scheme 93).In 2021, Dong Xu et al. reported an efficient synthesis of bketo suldes 337 through an aryl-thiol azide coupling reaction (Scheme 94). Diverse b-keto suldes were produced through this method in poor to excellent yields (45-90%). Thiyl radicals have been well suited and broadly used in organic synthesis, but the reaction of vinyl azide with thiyl radical is rare.For the rst time in 1997, Montevecchi and coworkers studied the reaction of thiols 336 and vinyl azides 1. They found that the reaction of a-phenyl vinyl azides with aryl thiols 336 was fast, producing b-keto suldes 337 almost quantitatively. In this study, the authors proposed a radical-chain mechanism. First, aryl thiyl radical 338, produced through the spontaneous oxidation of aryl thiol by oxygen, added to b-vinylic carbon of aphenyl vinyl azides with subsequent nitrogen extrusion to generate b-sulfanyliminyl radical 339. Hydrogen abstraction from the next aryl thiol reproduced thiyl radical 338 and generated intermediate imine 340 and its tautomer 341 that were both hydrolyzed to desired product 337 (Scheme 95).In 2019, Fie Xu et al. developed an efficient method to obtain cyclic b-amino ketones 343 via visible-light photo-redox catalysis in poor to very good yields (18-82%). 108 b-Amino ketones are versatile synthetic building blocks in organic chemistry that can be converted into a range of benecial and valuable derivatives containing b-amino-alcohols. b-Amino carbonyl derivatives are attractive as key synthetic intermediates of a wide range of drugs and biologically active natural products (Scheme 96).To illustrate the mechanism of this reaction, the singleelectron transfer from the amine to visible-light-excited photocatalyst *Ru(bpy derivatives of 354 in fair to good yields (48-74%) (Scheme 98).Nitrile groups are present in the structures of many pharmaceuticals and bioactive natural products.Nitriles are also widely utilized as a directing group in C-H activation chemistry 112 and as versatile synthetic intermediates, especially as precursor to heterocycles 113 and functionalities at the carboxylic acid oxidation level.3-Azido-2-methylbut-3-en-2-ol 353 was considered an appropriate tool to achieve the cyanomethylation of radicals due to it encompassing two key design elements: (I) vinyl azides 1 that can act as masked cyanomethyl groups, and (II) dimethyl carbinol as a latent radical leaving group. With radical generation from a substrate through the oxidative quenching of an excited-state photo-redox catalyst (PC* / PC 1+ ), it was anticipated that reagent 355 would intercept open-shell species to initiate a cascade process via radical addition to the olen, producing adduct 356 that would release dinitrogen to produce iminyl radical 357. In the next step, fragmentation of iminyl radical 357 via a-C-C bond cleavage and ejection of the 2hydroxypropyl radical 358 was expected to produce nitrile functionality (Scheme 99).In 2022, Tang et al. synthesized N-aryl-(triuoromethyl sul-nyl) acetamides 362 in good to very good yields (73-88%), through S-triggered Schmidt-type rearrangement of vinyl azides 1 (Scheme 100).The triuoromethylthio (CF 3 S) functional group is very common in the structures of agrochemical 115 compounds, such as pronil, toltrazuril, triorex, and cefazafur, and medicinal compounds, improving physio-chemical properties and pharmacokinetics, owing to their electronnegativity and excellent lipophilicity.As depicted in Scheme 101, a Pd(0) catalyst rst activated vinyl azides 1 to give metal complex 363 via path a. Addition of CF 3 SOCl, in situ generated from the reaction of CF 3 SO 2 Na with BTC, to the palladium-activated intermediate 363 happened to form the C-S bond, giving imine species 367. On the other hand, path b suggested that the PdL complex might be inserted into the CF was prepared to provide a summary of the strategies researchers have utilized to synthesize other compounds through vinyl azides, in which catalyst, solvent, temperature, yield, and MW irradiation are compared. # Conclusion Vinyl azides as versatile synthons are applied for the synthesis of different types of compounds, such as cyclic, heterocyclic, and non-cyclic compounds. N-heterocycles are a group of nitrogencontaining compounds with a broad range of interesting biological and pharmaceutical applications. Due to their high and diverse reactivity, vinyl azides are promising candidates as precursors for the production of N-heterocycles. It is our belief that utilizing this three-atom synthon in new protocols for the synthesis of heterocyclic compounds and other pharmaceutical compounds will continue to progress in coming years. ## Conflicts of interest There are no conicts to declare. ## Notes and references [fig] Figure 2: Conversion of vinyl azide to different intermediates. [/fig] [fig] Figure 3: Vinyl azide as a multipurpose precursor.Scheme 2 Formation of azirine and iminoketene from internal vinyl azide. [/fig] [fig] Scheme 5: Multistep reaction for the decomposition of vinyl azide in the singlet state: (A) decomposition of s-cis, (B) decomposition of s-trans. For the transition states we show in parentheses the values of imaginary frequencies (in cm −1 ). [/fig] [fig] Figure 4: Full energy profile of the IRC path calculated for the decomposition of s-cis-vinyl azide (orange points) and s-trans-vinyl azide (purple points) at B3LYP/6-311++G(d,p) level. Scheme 6 Formation of nitril from terminal vinyl azide. [/fig] [fig] Figure 5: Singlet and triplet vinyl nitrene. Scheme 7 Resonance forms of vinyl azide. [/fig] [fig] Scheme 12: Synthesis of ene-g lactams 23 from vinyl azides 1. Scheme 13 Mechanism for the synthesis of ene-g lactams 23 from vinyl azides 1. Scheme 14 Synthesis of cyano-alkylated 3,4-dihydro-2H-pyrroles 25 through reaction of vinyl azides 1 and cycloketone oxime esters 24. [/fig] [fig] 1: Palladacycle 69 underwent migratory insertion with vinyl azide, and then release of TFA from 73 gave intermediate 74. Then, oxidative cyclization generated indoline intermediate 75, which transformed to product 68 via the elimination of HN 3 . Meantime, the catalytically active Pd(II) was regenerated by oxidation of Pd(0). Decomposition of vinyl azide gave 2H-azirine 12 by the acceleration of DABCO, which underwent migratory insertion into palladacycle 69 to generate intermediate 70. In the next step, reductive elimination of 70 produced intermediate 71 with simultaneous formation of Pd(0), which could be reoxidized to the active catalytic Pd(II). In the nal step, an intramolecular nucleophilic addition of intermediate 71 resulted in intermediate 72, which underwent further deamination, generating product 68 (Scheme 25). [/fig] [fig] 48: Nickel-catalyzed reaction of cycloketone oxime esters 24 and vinyl azides 1. Scheme 49 Mechanism for the nickel-catalyzed reaction of cycloketone oxime esters 24 and vinyl azides 1. Scheme 50 Palladium-catalyzed synthesis of isoquinoline derivatives 176 and 177. Scheme 51 Mechanism for palladium-catalyzed synthesis of isoquinoline derivatives 176. intermediate 195, along with the release of HOAc. Pyrolysis of vinyl azides 1 generated 2H-azirines which underwent C-N bond cleavage and reduction with Cu(I) to yield Cu(II) azaenolate 196 species. Next, protonation with HOAc gave iminyl cooper species 197. Also, trans-metalation with intermediate 195 generated iminyl palladium species 198, which underwent tautomerization to form Pd(II) complex 199. Aerward, the reductive elimination of 200 provided intermediate 201 and Pd(0) species. The Pd(II) catalyst was reproduced through reoxidation with Pd(0) by the Ag/Cu oxidant. Then, intramolecular condensation and subsequent aromatization with the release of NH 2 NHAc produced the pyridine product 194 (Scheme 56). 71 In 2020, Zhao et al. reported a one-pot and novel method to obtain a-carbolines 204 through rhodium/copper-catalyzed coupling-cyclization of O-alkenyl aryl isocyanides 203 with vinyl azides 1 (Scheme 57). In this method, the reactive vinyl carbodiimides, produced in situ from the coupling reaction of vinyl azides 1 with O-alkenyl aryl isocyanides 203, underwent intermolecular [4 + 2]-cycloaddition and provided a new method for the synthesis of poly-substituted a-carbolines in poor to excellent yields (40-74%). Among different functionalized aryl isocyanides, O-alkenyl aryl isocyanides have appeared as valuable precursors for the efficient synthesis of different fused azaheterocycles in the past decade. 72 Scheme 52 Synthesis of 6-(sulfonylmethyl) phenanthridines 184. Scheme 53 Rhodium-catalyzed synthesis of trifluoromethyl isoquinolines 187. Scheme 54 Plausible mechanism for the synthesis of trifluoromethyl isoquinoline 187. Scheme 55 Palladium-catalyzed synthesis of alicyclic[b]-fused pyridine 194 via C(sp 2 )-H activation. Scheme 56 Mechanism for the palladium-catalyzed synthesis of alicyclic[b]-fused pyridine 194 via C(sp 2 )-H activation. Scheme 57 Rhodium/copper-catalyzed coupling-cyclization of Oalkenyl aryl isocyanides 203 with vinyl azides.The proposed mechanism of this reaction is depicted in Scheme 58. First, aryl isocyanide 203 coordinated with [Cp*RhCl 2 ] 2 to generate intermediate 205, which coordinated with vinyl azides 1 to give 206. Then, intermediate 206 released N 2 to give nitrene intermediate 207, which underwent migratory insertion, and rhodium was released to produce vinyl carbodiimide 211. In the next step, intermediate 208, which was activated by copper, underwent an intermolecular [4 + 2]cycloaddition to give intermediate 209. [/fig] [fig] Scheme 63: Synthesis of substituted phenanthridines 226 and 228 through the copper-catalyzed oxidative cyclization of vinyl azides. Scheme 64 Mechanism for the synthesis of substituted phenanthridines 226 through copper-catalyzed oxidative cyclization of vinyl azides. Scheme 65 Vinyl azides as dual synthons for the synthesis of 4substituted quinolines 233. Scheme 66 Mechanism for the synthesis of 4-substituted quinolines 233. [/fig] [fig] Figure 6: Examples of natural products containing an octahydrocyclopenta[b]pyran motif. Scheme 71 Mechanism for the synthesis of pyran-based indeno[1,2-c] isochromene 254 scaffolds via vinyl azides. [/fig] [fig] 74: Photo-redox-catalyzed synthesis of N-unsubstituted enamino sulfones 272 through vinyl azides 1. Scheme 75 Radical addition to vinyl azides for the synthesis of Nunsubstituted enamine. Scheme 76 Synthesis of O-phthalimide oximes 278 from N-hydroxy phthalimide 277 and vinyl azides 1. Scheme 77 Mechanism for the synthesis of O-phthalimide oximes 278. Scheme 78 Synthesis of a-amido ketones 283 through the reaction of carboxylic acid 282 with vinyl azides 1. Scheme 79 Mechanism for the synthesis of a-amido ketones 283 through the reaction of carboxylic acid with vinyl azides. generate imine intermediate 291. The hydrolysis of 291 would produce the desired b-keto phosphonate or b-keto phosphine oxides 286 (Scheme 81). 87 In 2019, Tang et al. proposed a new method to synthesize bhydroxy-triuoromethyl ketone 293 (Scheme 82) through the copper-catalyzed aldol reaction of vinyl azides 1 with triuoromethyl ketone 292 in good to excellent yields [/fig] [fig] 80: Mn(III)-catalyzed synthesis of b-keto phosphine oxides 286. Scheme 81 Mechanism for the Mn(III)-catalyzed synthesis of b-keto phosphine oxides 286. Scheme 82 Copper-catalyzed aldol reaction of vinyl azides with trifluoromethyl ketone 292. Scheme 83 Mechanism for the copper-catalyzed aldol reaction of vinyl azides with trifluoromethyl ketone 292. Scheme 84 Synthesis of b-enaminones 298 via a coupling reaction of vinyl azides with aldehydes 135. [/fig] [fig] Figure 7: Examples of biologically active molecules containing aketoamides. Scheme 87 Mechanism for the synthesis of a-keto thioamide 303 from vinyl azides 1 and amines. Scheme 88 Synthesis of N-unprotected enamines 311, 312, and 313. [/fig] [fig] Scheme 89: Transformation of b-functionalized enamine. Scheme 90 Synthesis of 3-oxoisoindoline-1-acetamides 323 via vinyl azides. [/fig] [fig] Figure 8: Some biologically active isoindolinone derivatives. © 2023 The Author(s). Published by the Royal Society of Chemistry RSC Adv., 2023, 13, 990-1018 | 1013 [/fig] [table] Table 2: Synthesis of six-membered heterocycles through vinyl azides [/table]

Observation of an oxonium ion intermediate in ethanol dehydration to ethene on zeolite Zeolite-catalyzed dehydration of ethanol offers promising perspectives for the sustainable production of ethene. Complex parallel-consecutive pathways are proposed to be involved in the reaction network of ethanol dehydration on zeolites, where the initial step of ethanol dehydration is still unclear particularly for the favorable production of ethene at lower temperature. Here we report the observation of a triethyloxonium ion (TEO) in the dehydration of ethanol on zeolite H-ZSM-5 by using ex situ and in situ solid-state NMR spectroscopy. TEO is identified as a stable surface species on the working catalyst, which shows high reactivity during reaction. Ethylation of the zeolite by TEO occurs at lower temperature, leading to the formation of surface ethoxy species and then ethene. The TEO-ethoxide pathway is found to be energetically preferable for the dehydration of ethanol to ethene in the initial stage, which is also supported by theoretical calculations. thene is one of the most important commodity chemicals, which is currently produced by cracking processes from petroleum. Due to the large availability of bioethanol from renewable biomass sources and the dwindling of fossil resource 1 , the conversion of ethanol to ethene and higher hydrocarbons now is receiving increasing attention from both academia and industry [bib_ref] Recent advances in catalytic conversion of ethanol to chemicals, Sun [/bib_ref] [bib_ref] Dehydration of ethanol to ethylene, Zhang [/bib_ref]. Among of the heterogeneous catalysts studied, zeolites especially ZSM-5 or modified ZSM-5 are the promising catalysts as they can be tuned to exhibit higher activity than the traditional alumina-based catalysts [bib_ref] Steam reforming of ethanol for hydrogen production: Thermodynamic analysis including different carbon..., Diaz Alvarado [/bib_ref] [bib_ref] High effective dehydration of bioethanol into ethylene over nanoscale HZSM-5 zeolite catalysts, Bi [/bib_ref] [bib_ref] Mechanistic study of alcohol dehydration on gamma-Al 2 O 3, Roy [/bib_ref] [bib_ref] Correlating acid properties and catalytic function: a first-principles analysis of alcohol dehydration..., Janik [/bib_ref] [bib_ref] Direct conversion of bio-ethanol to isobutene on nanosized ZnxZryOz mixed oxides with..., Sun [/bib_ref] [bib_ref] Production of ethylene from hydrous ethanol on H-ZSM-5 under mild conditions, Phillips [/bib_ref] : the reaction temperature is lower and a low concentration of ethanol aqueous solution can be used. The formation of ethene is the first step in ethanol dehydration; subsequent polymerization, cracking and aromatization by the secondary reaction of ethene leads to the formation of longer-chain hydrocarbons [bib_ref] Insights into the reaction mechanism of ethanol conversion into hydrocarbons on H-ZSM-5, Van Der Borght [/bib_ref] , very similar to the formation of hydrocarbons in the conversion of methanol over zeolites [bib_ref] Unraveling the reaction mechanisms governing methanol-to-olefins catalysis by theory and experiment, Hemelsoet [/bib_ref] [bib_ref] Conversion of methanol to hydrocarbons: how zeolite cavity and pore size controls..., Olsbye [/bib_ref]. The detailed knowledge of the reaction mechanism of ethene formation in ethanol dehydration is important not only for understanding ethanol dehydration as a model reaction of alcohol conversion but also for the development of improved catalysts for light olefins production. The formation of ethene in the initial dehydration process has been experimentally [bib_ref] Catalytic consequences of hydroxyl group location on the rate and mechanism of..., Chiang [/bib_ref] [bib_ref] Diethyl ether cracking and ethanol dehydration: Acid catalysis and reaction paths, Phung [/bib_ref] [bib_ref] Kinetic modelling of the transformation of aqueous ethanol into hydrocarbons on a..., Gayubo [/bib_ref] and theoretically [bib_ref] Activation energies for the reaction of ethoxy species to ethene over zeolites, Kondo [/bib_ref] [bib_ref] DFT-based microkinetic modeling of ethanol dehydration in H-ZSM-5, Alexopoulos [/bib_ref] [bib_ref] Ethanol dehydration in HZSM-5 studied by density functional theory: evidence for a..., Kim [/bib_ref] [bib_ref] Catalytic dehydration of ethanol over post-treated ZSM-5 zeolites, Xin [/bib_ref] investigated. A parallel-consecutive route was proposed [bib_ref] Dehydration of ethanol to ethylene, Zhang [/bib_ref] : dehydration of ethanol to ethene and diethyl ether in parallel and a consecutive decomposition of diethyl ether to ethene. In the unimolecular dehydration of ethanol to ethene, it was proposed that the acidic proton transfers from the zeolite to the OH group on ethanol, which subsequently releases water to form an intermediate such as carbocation or an alkoxide bound to zeolite. Solid-state NMR and infrared spectroscopic studies [bib_ref] Activation energies for the reaction of ethoxy species to ethene over zeolites, Kondo [/bib_ref] [bib_ref] Formation and decomposition of surface ethoxy species on acidic zeolite Y, Wang [/bib_ref] [bib_ref] An ethoxy intermediate in ethanol dehydration on bronsted acid sites in zeolite, Kondo [/bib_ref] Diethyl ether (DEE) is concurrently produced with ethene, and the kinetic and spectroscopic studies revealed that intermolecular dehydration of ethanol leads to DEE via dimeric ethanol species or reaction of an ethoxy group with undissociated ethanol on zeolites [bib_ref] Catalytic consequences of hydroxyl group location on the rate and mechanism of..., Chiang [/bib_ref] [bib_ref] IR spectroscopy of neutral and ionic hydrogen-bonded complexes formed upon interaction of..., Zecchina [/bib_ref] [bib_ref] Calorimetric study of alcohol and nitrile adsorption complexes in H-ZSM-5, Lee [/bib_ref]. The following cracking of DEE produces ethene, which occurs preferably at lower reaction temperatures as compared to the unimolecular dehydration of ethanol to ethene [bib_ref] Diethyl ether cracking and ethanol dehydration: Acid catalysis and reaction paths, Phung [/bib_ref] [bib_ref] Conversion of ethanol in aqueous solution over ZSM-5 zeolites: Study of the..., Nguyen [/bib_ref]. Note that the ethoxy species was generally proposed to be the direct precursor to ethene. However, the generation of such species is still not well understood. Thus, although different reaction schemes have been proposed, the detailed mechanism of the initial step of ethanol dehydration remains elusive, particularly for the favorable production of ethene from DEE at lower temperature. Here we report the observation and identification of reaction intermediates in ethanol dehydration to ethene over zeolite H-ZSM-5. The ex situ and in situ solid-state NMR spectroscopy allows for exploration of the dehydration of ethanol at the initial stage. A triethyloxonium ion (TEO) is observed as a potential active intermediate on the working catalyst. We provide evidence for the involvement of TEO in the formation of surface ethoxy species and then ethene on the zeolite. The elementary reactions leading to the formation of ethene are identified by combining experiments and theoretical calculations, which suggests that a TEO-mediated reaction route is operative for the dehydration of ethanol to ethene. # Results Observation of TEO. H-ZSM-5 (Si/Al = 11.5, Zeolyst) was used for the dehydration of ethanol. The XRD and [bib_ref] In situ solid-state NMR study of methanol-to-gasoline chemistry in zeolite HZSM-5, Munson [/bib_ref] Al and 29 Si solidstate NMR spectra show structure information of the H-ZSM-5 catalyst (Supplementary [fig_ref] Figure 2: Identification of triethyloxoium ion by correlation NMR spectroscopy [/fig_ref]. The acidic property was examined by FT-IR and NH 3 -TPD (Supplementary . The reactions were conducted in a pulse-quench reactor [bib_ref] Pulse-quench catalytic reactor studies reveal a carbonpool mechanism in methanol-to-gasoline chemistry on..., Goguen [/bib_ref] , in which ethanol was pulse-injected at temperatures ranging from 160 to 350°C and allowed to react for 4 s under continuous He carrier gas flow before the reaction was thermally quenched by pulsing liquid nitrogen onto the catalyst bed. The reaction effluent was analyzed by GC [fig_ref] Figure 5: Proposed catalytic cycle for ethanol dehydration to ethene [/fig_ref]. Ethene is the only product at 160-200°C, indicating that the dehydration of ethanol to ethene is favorable at lower temperature. With increasing reaction temperature, the secondary reaction of ethene leads to the formation of C 3 + products including olefins and aromatics. shows the ex situ 13 C CP/MAS NMR spectra of trapped products from the dehydration of 13 C-1-enriched ethanol (CH 3 13 CH 2 OH, 99% 13 C) on H-ZSM-5 at 160-350°C for 4 s. Ethanol and DEE are observed as reflected by the strong signals at 62.5 and 69.0 ppm respectively, due to the 13 C-enriched methylene carbons; the methyl carbons produce the weak signals at 12-17 ppm. The adsorbed species formed at 200°C were further analyzed by two-dimensional (2D) 1 H-13 C HETCOR NMR experiment . A weak signal at 72.8 ppm that was seriously overlapped by the ethanol signal (69.0 ppm) in the 1D 13 C NMR spectra was clearly discerned in the 2D spectrum, which can be ascribed to the methylene carbon of ethoxy group. The formation of the alkoxy species was well established in methanol and ethanol dehydration on acidic zeolites. Hunger et al. [bib_ref] Formation and decomposition of surface ethoxy species on acidic zeolite Y, Wang [/bib_ref] reported the surface ethoxy species from ethanol dehydration on HY zeolite, which has a chemical shift of 72.6 ppm. Interestingly, a well-resolved signal appears at 85.0 ppm 90.0 ppm in the literatures [bib_ref] Formation and decomposition of surface ethoxy species on acidic zeolite Y, Wang [/bib_ref] [bib_ref] The Nature, structure, and composition of adsorbed hydrocarbon products of ambient temperature..., Stepanov [/bib_ref]. Since there is no any C 3+ species formed at this reaction condition, the oligomerization of ethene did not likely occur in our case. Furthermore, no methylene signals at 20.0 to 40.0 ppm can be observed that are supposed to be oligomerized compounds [bib_ref] In situ solid-state NMR study of methanol-to-gasoline chemistry in zeolite HZSM-5, Munson [/bib_ref]. Therefore, the 85.0 ppm signal should not be ascribed to the oligomeric alkoxy species. Instead, we proposed on the basis of its chemical shift the formation of alkyl-substituted oxonium ions. Munson and Haw reported the formation of trimethyloxonium ion (TMO) with a characteristic 13 C NMR signal at 80.0 ppm by the reaction of dimethyl ether on H-ZSM-5 [bib_ref] NMR observation of trimethyloxonium formation from dimethyl ether on zeolite HZSM-5, Munson [/bib_ref] , and they also confirmed that TMO was not an intermediate in the methanol to hydrocarbons conversion. Since DEE is readily formed and in equilibrium with ethanol , we assume that a triethyloxoium ion (TEO) may be formed. To gain insights into the structure of TEO, a 13 C-13 C J-based refocused INADEQUATE (Incredible Natural Abundance Double Quantum Transfer Experiment) [bib_ref] Through-bond carbon-carbon connectivities in disordered solids by NMR, Lesage [/bib_ref] MAS NMR experiment was performed, which provides an unambiguous identification of bond connectivity of carbon species. The INADEQUATE spectrum [fig_ref] Figure 2: Identification of triethyloxoium ion by correlation NMR spectroscopy [/fig_ref] of H-ZSM-5 reacted with 13 CH 3 13 CH 2 OH at 200°C for 4 s exhibits a clear correlation peak pair between the methylene carbon at 85.0 ppm and the methyl carbon at 12.0 ppm, confirming the formation of TEO that is composed of ethyl groups. Two other correlations are identified, showing the connectivity of the methylene carbons of DEE (69.0 ppm) and ethanol (62.5 ppm) to the corresponding methyl carbons at 12.8 and 17.0 ppm, respectively. The structure of TEO was also theoretically optimized in the H-ZSM-5 channel by periodic density functional theory (DFT) calculations . The 13 C chemical shift of the methylene carbon of TEO is predicted to be 84.6 ppm, in good agreement with the experimental value. Therefore, our NMR experiments and theoretical calculations provide solid evidence for TEO formation in the ethanol dehydration process. For the adsorbed species formed on H-ZSM-5 after reaction at 200°C for 4 s, the amount of TEO (85.0 ppm), ethoxy (72.8 ppm), DEE (69.0 ppm) and ethanol (62.5 ) was determined by 13 C NMR experiment to be 0.34, 0.46, 28.76 and 6.99 mmol/g, respectively. TEO and the ethoxy species are solely formed and located on Brønsted acid sites for charge compensation, while DEE and ethanol are trapped over zeolite via either chemical or physical adsorption. Considering the total Brønsted acid sites (about 1.03 mmol/g) on the catalyst, the amount (less than 0.23 mmol/g) of DEE adsorbed on Brønsted acid sites should be lower than that (0.34 mmol/g) of TEO at this reaction condition. At temperature above 300°C [fig_ref] Figure 2: Identification of triethyloxoium ion by correlation NMR spectroscopy [/fig_ref] , the 85.0 ppm signal is almost unobservable, which can be ascribed to the decomposition of TEO. In the meantime, the secondary reaction of ethene produces the dominating C 3+ hydrocarbons, reflected by the strong highfield signals below 30.0 ppm. The signals at 147.3-156.3, 243.6-255.7 and 47.6 ppm are characteristics of the cyclic cations such as dimethylcyclopentenyl and ethylcyclopentenyl ions [bib_ref] New insight into the hydrocarbon-pool chemistry of the methanol-to-olefins conversion over zeolite..., Wang [/bib_ref] [bib_ref] Experimental evidence on the formation of ethene through carbocations in methanol conversion..., Wang [/bib_ref] [bib_ref] Direct detection of supramolecular reaction centers in the methanol-to-olefins conversion over zeolite..., Wang [/bib_ref] , which are usually involved in the formation of aromatics. This implies that the aromatics-based cycle prevails in a similar route to the conversion of methanol to hydrocarbons [bib_ref] Conversion of methanol into hydrocarbons over zeolite H-ZSM-5: Ethene formation is mechanistically..., Svelle [/bib_ref] [bib_ref] On reaction pathways in the conversion of methanol to hydrocarbons on HZSM-5, Sun [/bib_ref]. Reactivity and intermediate role of TEO. In order to explore the reactivity of TEO, continuous-flow 13 CH 3 13 CH 2 OH conversion over H-ZSM-5 was investigated by in situ 13 C NMR spectroscopy (see methods section for details of the in situ NMR experiments). shows the real-time monitoring of formed surface species on H-ZSM-5 in the ethanol dehydration process. At a constant reaction temperature of 220°C , the appearance of 85.0 ppm signal after 200 s points to the formation of TEO. The broad signal centered at 58.5 ppm should come from the adsorbed ethanol, whose methyl group produces a high-field strong signal at 15.0 ppm. Note that the methylene group of DEE (69.0 ppm) is invisible in the in situ experiment. This is probably due to the strong adsorption of DEE on zeolite, leading to large anisotropic interactions (chemical shifts and nuclear dipoledipole couplings) which cannot be efficiently removed by the low magic spinning speed (2 kHz). However, the formation of DEE is evidenced by the 12.5 ppm signal due to its methyl group featured by high mobility. TEO is observable with increasing the reaction time up to 1200 s (20 min), indicative of its stability at the moderate reaction temperature. The GC analysis of the effluent products obtained from the reaction with time-on-stream shows the formation of ethene and a small amount of higher hydrocarbons . shows the in situ 13 C NMR spectra obtained at elevating temperatures. The TEO is observable over a wide temperature range. The formation of hydrocarbons prevails at higher temperatures as reflected by a set of high-filed signals (8.7-32.6 ppm) while the TEO is significantly reduced due to its decomposition, consistent with the ex situ NMR result . In comparison, the structurally similar species TMO formed on H-ZSM-5 was readily decomposed before the onset of hydrocarbon synthesis [bib_ref] Insitu solid-state NMRstudy of methanol-to-gasoline chemistry in zeolite HZSM-5, Munson [/bib_ref] [bib_ref] An in-situ solid-state NMR-study of the formation and reactivity of trialkylonium ions..., Munson [/bib_ref]. To confirm the intermediate role of TEO in the formation of ethene, the 13 C NMR spectra of trapped species obtained from the pulse-quench reactions of CH 3 13 CH 2 OH were recorded during the reaction course of 4~64 s at 250°C. shows the expanded spectra and their deconvolutions (see for the whole spectra), in which the formation and evolution of TEO (85.0 ppm) is evident. We know from above experiments that the ethoxy species (72.8 ppm) is simultaneously produced, which is confirmed by 2D 1 H-13 C HETCOR MAS NMR . Comparison of the integrated 13 C signal intensities shows that the ethoxy species keeps a similar evolution trend with TEO . Thus, the formation of surface ethoxy species is most likely related to TEO. Furthermore, TEO was deliberately prepared by an ion exchange of triethyloxonium tetrafluoroborate with H-ZSM-5. The formation of surface TEO on H-ZSM-5 is confirmed by its characteristic signal at 85.0 ppm in the 13 C NMR spectra . The concentration of TEO determined by [bib_ref] Catalytic consequences of hydroxyl group location on the rate and mechanism of..., Chiang [/bib_ref] C MAS NMR is about 1 mmol/g, comparable to the amount (1.03 mmol/g) of Brønsted acid site. Importantly, the isolation of stable TEO at room temperature allows us to unambiguously trace its reactivity and transformation on the catalyst. After heating, TEO was partially converted into ethoxy species (at 72.8 ppm) and DEE (at 69.0 ppm) at lower temperature of 80°C, while higher temperature (200°C) led to its complete conversion. We also obtained the activation energy for the formation of ethoxy species from TEO by measuring the rate constants at different temperatures using 13 C MAS NMR spectroscopy . The experimentally determined activation energy is ca.77.3 kJ/mol. It is worth noting that a large quantity of ethene was observed in the effluent product upon the consumption of TEO by GC . These results demonstrate a direct correlation of TEO to ethoxy species and ethene product. Mechanism of ethene formation via TEO. Our experimental observations indicate a strong potential of TEO serving as an active intermediate in the formation of ethene. A whole reaction network for ethanol dehydration to ethene was proposed [fig_ref] Figure 5: Proposed catalytic cycle for ethanol dehydration to ethene [/fig_ref] , which includes the previously theoretically predicted unimolecular and bimolecular processes [bib_ref] DFT-based microkinetic modeling of ethanol dehydration in H-ZSM-5, Alexopoulos [/bib_ref] [bib_ref] Catalytic dehydration of ethanol over post-treated ZSM-5 zeolites, Xin [/bib_ref] and our proposed TEOmediated routes. It consists of three types of mechanisms: direct ethanol dehydration to ethene (route 5, 6, 7 and 8), diethyl ether decomposition to ethene (route 3 and 4) and TEO-mediated ethene formation (route 1 and 2). DFT calculations were performed to explore the most favorable ethene formation route and the optimized transition state structures on the zeolite cluster model [fig_ref] Figure 2: Identification of triethyloxoium ion by correlation NMR spectroscopy [/fig_ref] are displayed in . Our calculations show that all the kinetic steps of ethanol dehydration are endothermic processes (Supplementary . The calculated Gibbs free energy barrier -18 for the energy diagrams), pre-exponential factor, and rate coefficient at 473 K of the forward and reverse reactions are tabulated in . The comparison of energy barriers and rate coefficients indicates that the direct dehydration processes (steps 9, 10, 15 and 16) are energetically less favorable for ethene formation, in agreement with the previous work [bib_ref] DFT-based microkinetic modeling of ethanol dehydration in H-ZSM-5, Alexopoulos [/bib_ref]. Compared with the direct ethanol dehydration routes (route 5, 6, 7 and 8), the two molecules mediated route 4 including dimer-mediated etherification (ΔG act(f) = 129.0 kJ/mol, k f = 5.54 × 10 −2 s −1 , step 4) and the following DEE decomposition to ethene (ΔG act(f) = 136.1 kJ mol −1 , k f = 9.24 × 10 −3 s −1 , step 11) is favorable for ethanol dehydration to ethene, in agreement with the previous report that the reaction path via DEE is preferentially involved in ethene formation at temperatures lower than 500 K 17 . It is reasonable to assume that TEO would be readily formed due to the facile formation of DEE. This is confirmed by a free energy barrier of ca. 94.1 kJ mol −1 for the formation of TEO from DEE (step 6), which is much lower than that of DEE decomposition routes (ΔG act(f) = 136.1 kJ/mol for step 11 in route 4 and ΔG act(f) = 155.9 kJ mol −1 for step 12 in route 3). We have experimentally observed that the ethoxy species and DEE can be produced from TEO on H-ZSM-5 . This can be understood by the fact that TEO ethylates the conjugate base site of zeolite to form a framework-bound ethoxy species with the release of a DEE, which shows the characteristic property of trialkyloxonium ions acting as an powerful alkylating agent. This process is calculated to have an activation energy of 73.6 kJ mol −1 (step 7, Supplementary the presence of TEO provides an energetically preferable process at lower reaction temperature. # Discussion The dehydration of ethanol to ethene over H-ZSM-5 has been investigated by combined experiments and DFT calculations. Our ex situ and in situ solid-state NMR data provide the evidence for the formation of TEO in the dehydration process. Generally, onium ions that are often isolable as salts are well recognized as reaction intermediates. It was reported that trimethyloxonium can be generated on H-ZSM-5 zeolite. However, mechanistic significance of trimethyloxonium is uncertain since its intermediate role for the initial C-C bond formation in methanol conversion was excluded based on experimental observations [bib_ref] In situ solid-state NMR study of methanol-to-gasoline chemistry in zeolite HZSM-5, Munson [/bib_ref] [bib_ref] NMR observation of trimethyloxonium formation from dimethyl ether on zeolite HZSM-5, Munson [/bib_ref]. Herein, TEO is formed as a stable surface species on the working catalyst in both continuous-flow and pulse-quench reactions. We calculated the host-guest interaction free energy of TEO confined in the H-ZSM-5 channel, which is as low as −360.2 kJ mol −1 at 473 K, indicating that the zeolite framework can stabilize TEO intermediate effectively. The implications of the observation of TEO in ethanol dehydration to ethene in present work are manifold. The ethoxy species is often experimentally observed during ethanol dehydration on zeolite, which is generally considered as the precursor to ethene in both processes of unimolecular dehydration of ethanol and bimolecular dehydration of ethanol followed by DEE decomposition. DEE is largely formed in the initial stage of dehydration of ethanol and its coverage on acid sites is estimated to be comparable to TEO at lower temperature (e.g., 473 K). The presence of TEO provides an energetically favorable mechanism to link DEE and ethoxy species, which is viable for the formation of ethene at low temperature. In the onium-ylide mechanism proposed for the C-C bond formation from C1 compound such as methanol and DME on zeolite, Olah et al. predicated the intermediate formation of ethyldimethyloxonium, which could undergo β-elimination to yield ethene and DME [bib_ref] Onium Ylide chemistry. 1. Bifunctional acid-base-catalyzed conversion of heterosubstituted methanes into ethylene..., Olah [/bib_ref]. Analogously, the direct deprotonation and decomposition of TEO could lead to ethene and DEE on zeolites (route 2), which, however, occurs with a relatively higher energy barrier. We have experimentally shown that TEO tends to alkylate basic site (Si-O − -Al) of zeolite to form surface ethoxy species, which in nature resembles the alkyloxonium salt featured by strong alkylating ability in organic synthesis. This result indicates the reactivity of alkyloxonium ion could be altered by the zeolite catalyst, which leads to different reaction routes. Although the alkyloxonium ion (i.e., a protonated alcohol) has long been proposed as a key intermediate in the dehydration of alcohol to produce alkenes and ethers particularly in strong liquid acid systems, the detailed characterizations of TEO in the ethanol dehydration process in this work provide an example of uncovering the important role of this kind of intermediate in the formation of ethene on zeolite catalysts, which may open avenues for further experimental and theoretical exploration of the oxonium ions chemistry in alcohol conversion. Methods X-ray power diffraction (XRD). The ammonium form ZSM-5 (Si/Al = 11.5, obtaind from Zeolyst) was calcined in air at 823 K for 5 h to obtain the proton form H-ZSM-5. The structure and crystalline nature of H-ZSM-5 zeolites were examined by X-ray diffractometer (X'Pert 3 Powder) using a CuKα radiation with a step of 0.02°at a respective voltage of 40 kV and a current of 40 mA. FT-IR of pyridine adsorption. The FT-IR of pyridine adsorption measurements were performed on a Bruker Tensor 27 spectrometer. The catalysts were first activated at 773 K under high vacuum (<10 −5 Pa) for 12 h. After cooling down to 313 K, a background spectrum was collected. Excessive amount of pyridine was then introduced to the infrared cell and held for 2 h to allow equilibrium. The residual pyridine was removed by vaccum. The FT-IR spectra of pyridine-adsorbed samples were measured at 313 K after evacuation at 373, 473, 573, 673 and 773 K, respectively. Temperature-programmed desorption of ammonia (NH 3 -TPD). The NH 3 -TPD measurement was performed using a FINESORB-3010 chemisorption instrument. Typically, 100 mg of sample was pre-treated at 423 K for 1 h under 30 sccm of helium gas. After the sample cooled down to 323 K, NH 3 was introduced and held for 1 h. The temperature was then elevated from 323 to 1023 K at a ramping rate of 10 K/min and the desorbed NH 3 was detected by a thermal conductivity detector (TCD). Catalytic testing. The H-ZSM-5 powder was pressed into pellets between 60-80 mesh. The pellets (0.2 g) were activated at 400°C in flowing helium for 1 h prior to the ethanol dehydration reactions. A pulse-quench reactor was used to quench the reaction by reducing the reaction temperature with liquid nitrogen within a very short period (<1 s) [bib_ref] Pulse-quench catalytic reactor studies reveal a carbonpool mechanism in methanol-to-gasoline chemistry on..., Goguen [/bib_ref]. Typically, when the reaction proceeded in a Reaction Gibbs free energy barrier (ΔG act , kJ mol −1 ), pre-exponential factor(A, s −1 ), and rate coefficient (k, s −1 ) at 473 K of the forward (f) and reverse reactions (r) for the elementary step associated with the eight ethene production routes pulse-quench reactor for a pre-set period, the reaction was thermally quenched by pulsing liquid nitrogen onto the catalyst bed, which was achieved by using highspeed valves controlled by GC computer. In each case, 10 μl of reactant was pulsed into the reactor (heated at 160-350°C) containing 0.2 g H-ZSM-5 and allowed to react for different time, before quenching by liquid nitrogen. The trapped surface species were analyzed by 13 C solid-state NMR spectroscopy and the effluent products were on line determined by GC-MS analysis. For the continuous flow reaction, ethanol with a weight hourly space velocity (WHSV) of 2 h −1 was reacted over the H-ZSM-5 (0.2 g) pellets in a fixed bed reactor at the temperature range of 140-260°C. Gas chromatography. The effluent products drawn from the flowing gas were analyzed quantitatively by online GC-FID chromatograph (Shimadzu GC-2010 plus) which equipped with a flame-ionization detector and a Supelco Supel-Q TM PLOT capillary column (30 m × 0.32 mm × 15 μm). The initial temperature programming started at 40°C (maintained for 2 min), followed by a rate of 5°C min −1 to 132.5°C and a rate of 10°C min −1 to the final temperature of 220°C. The retained products in the catalyst was directly analyzed by solid-state NMR spectroscopy (see the following). Preparation of triethyloxoium ion (TEO) on H-ZSM-5. A concentration of 0.2 g of H-ZSM-5 zeolite (Si/Al = 11.5) was dehydrated on a vacuum line (<10 −3 Pa) at 400°C for 12 h. After dehydration, the sample was allowed to cool down to room temperature for subsequent use. A concentration of 0.3 g triethyloxonium tetrafluoroborate (Aldrich) was added into 3 g dry dichloromethane solvent, and then the dehydrated zeolite sample was added into the solution. All the operations were carried out in a glovebox filled with pure N 2 . After ultrasonic treatment of the mixture in ice water bath at 0°C for 20 min, the mixtrue was filtrated and evacuated completely and the sample was dried at room tempearture for 5 h by vacuum. The obtianed triethyloxoium ion (TEO) exchanged ZSM-5 zeolite was denoted as TEO-ZSM-5. Solid-State NMR experiments. After the reaction was quenched, the pulsequench reactor containing the catalyst was sealed. The sealed reactor was then transferred to a glove box filled with pure N 2 and the catalyst was packed into to an NMR rotor for ex situ NMR measurements. To enhance the detection sensitivity, [bib_ref] Catalytic consequences of hydroxyl group location on the rate and mechanism of..., Chiang [/bib_ref] C labeled ethanol was used for 13 C NMR experiments in both the pulse-quench and the in situ reactions. All the ex-situ 13 C solid-state NMR spectroscopy experiments were carried out at 9.4 T on a Bruker Avance III-400 spectrometer, equipped with a 4 mm probe, with resonance frequencies of 399.33 and 100.42 MHz for 1 H and 13 C, respectively. The magic angle spinning rate was set to 10 kHz. For the 1 H → 13 C CP/MAS NMR experiments, the Hartmann-Hahn condition was achieved using hexamethylbenzene (HMB), with a contact time of 5 ms and a repetition time of 2 s. The hpdec 13 C MAS NMR experiments were performed using a 13 C 90-degree pulse length of 5 μs and a recycle delay of 5 s. The 13 C chemical shifts were referenced to HMB (a second reference to TMS). The [bib_ref] In situ solid-state NMR study of methanol-to-gasoline chemistry in zeolite HZSM-5, Munson [/bib_ref] Al MAS NMR spectra were also acquired on the same 4 mm probe by small-flip angle technique with a pulse length of 0.3 μs (<π/12) and a recycle delay of 1 s. The magic angle spinning rate was set to 10 kHz. The 27 Al chemical shifts were referenced to 1 M Al(NO 3 ) 3 aqueous solution (0 ppm). [bib_ref] Through-bond carbon-carbon connectivities in disordered solids by NMR, Lesage [/bib_ref] Si MAS NMR experiments were carried out at 7.1T on a Varian Infinity plus-300 spectrometer, with resonance frequencies of 299.78 and 70.11 MHz for 1 H and [bib_ref] Through-bond carbon-carbon connectivities in disordered solids by NMR, Lesage [/bib_ref] Si, respectively. Single-pulse [bib_ref] Through-bond carbon-carbon connectivities in disordered solids by NMR, Lesage [/bib_ref] Si MAS NMR spectra with high power proton decoupling were recorded on a 7.5 mm probe, using a π/2 pulse of 5 μs, a recycle delay of 80 s and a spinning rate of 4 kHz. The [bib_ref] Through-bond carbon-carbon connectivities in disordered solids by NMR, Lesage [/bib_ref] Si chemical shifts were referenced to kaolinite (−91.5 ppm). 2D 1 H -13 C CP HETCOR experiment was performed using a Avance III 800 spectrometer operating at a 1 H Larmor frequency of 800.36 MHz and a 4 mm HCN E-free probe at a spinning frequency of 12 kHz. The 1 H π/2-pulse length was 4.85 μs. 13 C magnetization was created using a cross-polarization (CP) ramp of magnitude 80 to 100% with a contacted time of 5 ms. 256 transients were co-added using a recycle delay of 2 s. A total of 256 t 1 FIDs were recorded at increments of 3.19 ms using the States-TPPI method to achieve sign discrimination in F1. 1 H SPINAL-64 40 decoupling was applied during the t 2 acquision with a RF-field amplitude of 51.5 kHz. 2D 13 C-13 C CP J-refocused INADEQUATE 29 MAS NMR experiment was carried out using a Avance III 800 spectrometer at a MAS speed of 8 kHz. The 1 H π/2-pulse length was 2.75 us. 13 C magnetization was created using a crosspolarization (CP) ramp of magnitude 80 to 100% with a contacted time of 7 ms. [bib_ref] Catalytic consequences of hydroxyl group location on the rate and mechanism of..., Chiang [/bib_ref] C π/2 and π pulses were 4.8 us and 9.6 µs, respectively. The J-evolution τ periods were rotor synchronized and set to 3.24 ms. 64 transients were co-added using a recycle delay of 3 s. A total of 160 t 1 FIDs were recorded at increments of 3.97 ms using the States-TPPI method to achieve sign discrimination in F1. 1 H SPINAL-64 decoupling with a RF-field amplitude of 90.91 kHz was employed after CP covering all t 1 , J-evolution and t 2 periods. The two frequency axes in the 2D spectra are used to assign the 13 C resonances corresponding to through-bond 13 C-13 C connectivities, which allows for unambiguous assignment and structure determination of organic compounds. All the in situ 13 C solid-state NMR experiments were performed at 11.7 T on a Bruker Avance III 500 spectrometer, equipped with a 7 mm H/X MASCAT probe , with resonance frequencies of 500.58 and 125.87 MHz for 1 H and 13 C, respectively. The magic angle spinning rate was set to 2 kHz. Prior to the in situ NMR experiments, 0.2 g of pre-dehydrated H-ZSM-5 was pressed into a 7 mm NMR rotor. An axial hole of 2.5 mm in diameter was made into the sample with a special tool, by which an uniform annular catalyst bed was formed in the rotor. After inserting the injection tube into the MAS rotor, the rotor was heated at 280°C for 1 h via the bearing gas, keeping helium (200 ml min −1 ) injected into the rotor, then cooled down to the reaction temperature. [bib_ref] Catalytic consequences of hydroxyl group location on the rate and mechanism of..., Chiang [/bib_ref] C labelled ethanol ( 13 CH 3 13 CH 2 OH, 98% 13 C, Sigma-Aldrich) was diluted to 50% (v/v) with ethanol in natural aboundance. The diluted 13 C labelled ethanol was fed into the NMR rotor by the helium carrier gas (weight hourly space velocity (WHSV) = 2.2 h −1 ) through a saturator. The reactants flow inside the rotor via the injection tube and pass through the catalyst from the bottom to the top and leave the rotor via an exhaust tube in the rotor cap. The 13 C CP/MAS NMR spectra were recorded with a contant time of 5 ms and a recycle delay of 1 s. 30 and 900 scans were accumulated for each spectrum at different time and temperature respectively. The 13 C chemical shifts were referenced to HMB (a second reference to TMS). The in situ 13 C MAS NMR spectra in were generated by the overlay of 1D NMR spectra as a function of reaction time or temperature. The correlation between 13 C resonances and the reaction parameters (time and temperature) was followed in the 2D maps, which facilitates the monitoring of the fate of the organic species during the time or temperature evolutions. The activation energy of the formation of ethoxy species from TEO on H-ZSM-5 was measured as follows. The sealed NMR rotor containing the TEO-ZSM-5 sample was heated at a specific temperature (from 328 to 348 K) for a period of time, and then the reaction was quenched by liquid N 2 . 13 C MAS NMR measurement was performed at room temperature. At such low reaction temperature, only ethoxy species and diethyl ether were produced on the TEO-ZSM-5. The concentration of TEO during the reaction was measured by normalized integrated NMR signal at 85 ppm. Finally, the temperature dependent rate constant was obtained and the activation energy was derived from the Arrhenius equation. Theoretical calculations. The host-guest interactions between the cations and zeolite framework would result in the 13 C chemical shift moving due to the redistributions of electronic environment invoked by the H-ZSM-5 confined pore. Thus, to unambiguously assign the experimental NMR results, the 13 C NMR chemical shifts of the cations are further calculated using the periodic boundary condition. The geometry optimizations were performed by using the CASTEP program 41 with the generalized gradient approximation (GGA) proposed by Perdew-Burke-Ernzerhof (PBE) functional. And the ultrasoft pseudo potential, fine plane wave cut-off energy (340 eV) and a default fine level Monkhorst-Pack Kpoint (1 × 1 × 1) were adopted to sample the Brillouin zone. DFT-D method [bib_ref] Long-range corrected hybrid density functionals with damped atom-atom dispersion corrections, Chai [/bib_ref] was used in the structure optimization and the NMR calculation to accurately describe the weak interaction in H-ZSM-5. During optimization, the 22T active site atoms and the adsorbed cation were relaxing to their equilibrium positions. All of the NMR shielding calculations were performed by the GIPAW method in the MS CASTEP-NMR code at the GGA/PBE level based on the optimized zeolite structures. All of the 13 C chemical shifts (δ 13 C) cal were derived using the CASTEP-NMR module available in the Materials Studio package based on the optimized structures of cation accommodation in the H-ZSM-5 zeolite. The fine Kpoint (1 × 1 × 1) and cut-off energy of 550 eV were employed, combined with core-valance interactions described by ultrasoft pseudo potential generated on-thefly. The 13 C calculated chemical shifts were further converted to (δ 13 C) cal values, which were referred to the absolute shielding of TMO, namely 80 ppm for the experimental values. For the ethene formation pathway calculation, H-ZSM-5 zeolites are represented by a 72T model (HSi71AlO179, 252 atoms), which were extracted from their crystallographic structural data (http://www.iza-structure.org/databases/). The 72T contains the complete double 10-MR intersection pores of H-ZSM-5 zeolite. The terminal Si-H was fixed at a bond length of 1.47 Å, oriented along the direction of the corresponding Si-O bond. The substituted Al atom was placed at the T12 site of the crystallographic position during structural optimization, whereas the proton was located at the O24 site. The previous works have demonstrated that the Si 12 -O 24 (H)-Al 12 Bronsted acid site located at the channel intersection of H-ZSM-5 zeolite with maximum accessibility for bulky reactants and transition states [bib_ref] DFT-based microkinetic modeling of ethanol dehydration in H-ZSM-5, Alexopoulos [/bib_ref] [bib_ref] Acidity differences between inorganic solids induced by their framework structure. a combined..., Brändle [/bib_ref] [bib_ref] Quantum chemical modeling of zeolite-catalyzed methylation reactions: toward chemical accuracy for barriers, Svelle [/bib_ref]. The 22T active site atoms (HSi21AlO25, 48 atoms) and the adsorbed hydrocarbon complex were treated as the high-layer (See [fig_ref] Figure 2: Identification of triethyloxoium ion by correlation NMR spectroscopy [/fig_ref] while the rest of the frameworks were treated as the low-layer. To retain the structural integrities of the modeled zeolite, partial structure optimizations of the 72T cluster were performed by relaxing the atoms in the the high-level layer while keeping the rest of atoms fixed at their crystallographic positions. All the TS structures are found by the QST 3 method in the Gaussian program. Then the IRC (Intrinsic Reaction Coordinate) method was used to determine the structures of the corresponding reactant and product. Then based on the imaginary vibrational model of the optimized TS, we adjusted the positions of the vibrational atoms slightly along the calculated reaction coordinate on the two directions toward the reactant and product, respectively, and finally optimized the resulting structures to the minimum structures. These methods have been widely employed in the previous theoretical work. A combined theoretical approach, namely ONIOM (ωB97xd/6-31 G(d,p): am1) was used for the geometry optimization of adsorption states and transition states (TS). The ωB97XD hybrid density function was developed to consider long-rangecorrected hybrid functional, which implicitly accounted for empirical dispersion and could describe long-range dispersion interactions well with respect to the traditional density functional theory methods [bib_ref] Long-range corrected hybrid density functionals with damped atom-atom dispersion corrections, Chai [/bib_ref]. This functional was also recently found to perform very well for the description of adsorption and reactions on zeolites. All energies report herein were predicted at the ωB97XD/6-31 G(d, p) level based on the optimized structures. The harmonic frequency calculations employing a partial Hessian vibrational analysis (PHVA) [bib_ref] Vibrational modes in partially optimized molecular systems, Ghysels [/bib_ref] , including the 22T high layer active acidic sites and organic species were performed to check whether the stationary points found exhibit the proper number of imaginary frequencies. In frequency calculations, besides the atoms in high level and organic fragments, the constraints of the zeolite framework were also kept as the same in geometry optimizations, so one negative frequency would be observed for transition state point and none for the corresponding reactant and product. The Gibbs free energies (G) at 423 K (T) were calculated from harmonic frequencies. [formula] G ¼ H À T S ¼ E þ ZPVE þ H vib þ H trans þ H rot ð Þ À T S:ð1Þ [/formula] It can be seen that Gibbs free energies (G) include contributions from electronic energies (E), zero-point vibrational energies (ZPVE), vibrational enthalpies (H vib ), translational (H trans ) and rotational enthalpies (H rot ) for reactant molecules and the effect of entropy energy (T × S), which can more accurately describe the mechanism of ethanol dehydration reaction. And then, the reaction rate constants (k) at 473 K were further obtained by transition state theory (TST): [formula] k ¼ A: exp À ΔH act RT ¼ k B T h : exp ΔS act R : exp À ΔH act RT ¼ k B T h : exp À ΔG act RT ;ð2Þ [/formula] where A is pre-exponential factor, k B is Boltzmann's constant, h is Planck's constant and T is the reaction temperature. All the geometry optimizations and frequency calculations were performed using the Gaussian 09 package. ## Data availability All the data supporting the findings of this study are available within the article and its supplementary information file or from the corresponding author upon reasonable request. Received: 10 November 2018 Accepted: 2 April 2019 [fig] Figure 2: Identification of triethyloxoium ion by correlation NMR spectroscopy. 2D 13 C-13 C INADEQUATE MAS NMR spectrum of trapped species obtained from the reaction of 13 CH 3 13 CH 2 OH on H-ZSM-| (2019) 10:1961 | https://doi.org/10.1038/s41467-019-09956-7 | www.nature.com/naturecommunications [/fig] [fig] Figure 3, Figure 4: in good agreement with our experimental value (77.3 kJ mol −1 ), strongly supporting the formation of ethoxy via TEO intermediate. The low free energy barrier of step 8 (107.6 kJ mol −1 ) suggests the transformation of ethoxy species to ethene can readily occur. Moreover, the calculated electronic energy barrier for the conversion of ethoxy species to ethene (127.6 kJ mol −1 , SupplementaryTable 2) is similar to the experimentally estimated value (ca. 122 kJ mol −1 ) for the decomposition of ethoxy species to ethene on H-ZSM-538 . The direct decomposition of TEO into ethene (step 13 in route 2) is also considered, which is however characterized by a higher free energy barrier (127.7 kJ mol −1 ). A full comparison of free energy barriers and rate coefficients of the rate-determination steps in the proposed reaction network indicates that route 1 (steps 4, 6, 7, 8) and route 2 (steps 4, 6, 13) involving TEO intermediate are the most favorable routes. Thus, concerning the overall kinetics during the ethanol dehydration to ethene, Real-time NMR monitoring of ethanol dehydration on zeolite. In situ 13 C MAS NMR spectra of 13 CH 3 13 CH 2 OH conversion on H-ZSM-5 with time on stream (a) and at elevating temperatures (b). Spectra (a) were recorded at every 30 s from 0 to 1200 s at 220°C and 30 scans were accumulated for each time slice. Spectra (b) were recorded at every 10°C from 160-260°C and 900 scans were accumulated for each temperature slice. The 1D spectra on the top are recorded at 840 s (a) and 250°C (bEvolution of TEO in the formation of ethene.13 C CP MAS NMR spectra of trapped species on H-ZSM-5 obtained from pulse-quench reactions of CH 3 13 CH 2 OH at 250 o C for different time (a) and 13 C MAS NMR spectra of TEO-ZSM-5 at room temperature (RT), heated at 80°C for 5 min, and at 200°C for 5 min (b). The deconvoluted components are indicated in color Si (1 0 ) [/fig] [fig] Figure 5: Proposed catalytic cycle for ethanol dehydration to ethene. The structures of transition states (TS) involved in these reaction steps are depicted in Supplementary Fig. 13. * indicates the adsorbed state | (2019) 10:1961 | https://doi.org/10.1038/s41467-019-09956-7 | www.nature.com/naturecommunications [/fig]

Clinical care in hepatocellular carcinoma: A mixed methods assessment of experiences and challenges of oncology professionals Introduction:Healthcare providers (HCPs) may face numerous dilemmas in optimally screening, diagnosing, and treating patients with, and/or at risk for, hepatocellular carcinoma (HCC). This study aimed to achieve a greater understanding of the challenges in HCC care which in turn could delineate HCP educational opportunities within this oncologic sub-specialty.Methods:A mixed-methods approach was used to identify practice gaps and clinical barriers experienced by US-based medical oncologists, hepatologists, oncology physician assistants, oncology nurse practitioners, and interventional radiologists involved in HCC care. The qualitative (semi-structured interview) and quantitative (survey) data collection approaches were deployed sequentially with findings subsequently triangulated.Results: A total of 214 HCPs participated in this study. Analysis revealed challenges related to screening and diagnosing HCC, specifically in applying appropriate screening guidelines, and the optimal use and decisions related to diagnostic imaging and biopsy. Issues related to treatment selection included the application of existing HCC guidelines in treatment decision-making, weighing risk/benefit ratios of various antineoplastics regimens (i.e., tyrosine kinase inhibitors-TKIs, immunotherapy agents, chemotherapy), sequencing therapies, potential toxicity management, and optimally educating patients about their HCC. Conclusion: These findings highlight the educational needs of those involved in HCC care and provide a starting point for clinicians to both reflect on their practice and identify opportunities to enhance communication within the HCC team and between provider and patient. There is an opportunity to optimize continuing professional development interventions that address the identified gaps in clinical practice specifically related to teamwork and interdisciplinary communication. | 3671 JACOBS et al. How to cite this article: Jacobs G, Boyle DA, El-Serag HB, et al. Clinical care in hepatocellular carcinoma: A mixed methods assessment of experiences and challenges of oncology professionals. # | introduction Primary liver cancer is a leading cause of cancer death worldwide. [bib_ref] Management and treatment of hepatocellular carcinoma with immunotherapy: a review of current..., Ghavimi [/bib_ref] Hepatocellular carcinoma (HCC) represents 75%-85% of primary hepatic malignancies. [bib_ref] Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for..., Bray [/bib_ref] Its occurrence has increased globally over the last decades and is projected to affect 1.4 million people by 2040. [bib_ref] International trends in hepatocellular carcinoma incidence, Petrick [/bib_ref] [bib_ref] Epidemiological projections of viral-induced hepatocellular carcinoma in the perspective of WHO global..., Vo Quang [/bib_ref] Risk factors such as hepatitis B and C, chronic alcohol consumption, and non-alcoholic fatty liver disease (NAFLD) are associated with HCC resulting from cirrhosis. [bib_ref] Recent advances in hepatocellular carcinoma therapy, Dutta [/bib_ref] [bib_ref] Disease burden of hepatocellular carcinoma: a global perspective, Sayiner [/bib_ref] Currently, NAFLD's limited diagnostic testing options and therapeutic modalities increase HCC's global prevalence. [bib_ref] Hepatocellular carcinoma and non-alcoholic fatty liver disease, Golabi [/bib_ref] While early diagnosis is associated with improved outcomes and disease control, HCC tends to be diagnosed at a later stage because of low disease awareness from general practitioners and patients, low levels of screening of atrisk patients by specialists, and the high associated cost of screening. [bib_ref] Diagnostic challenges in primary hepatocellular carcinoma: case reports and review of the..., Pazgan-Simon [/bib_ref] [bib_ref] Epidemiology and surveillance for hepatocellular carcinoma: new trends, Singal [/bib_ref] In particular, appropriate screening (i.e., with ultrasound, non-invasive cross-sectional abdominal imaging or alpha fetoprotein [AFP] marker testing) is underutilized for patients at risk of HCC. This often results in a diagnosis of advanced stage HCC when liver resection, transplantation, or ablation are no longer viable treatment options. [bib_ref] Hepatocellular carcinoma: a review, Balogh [/bib_ref] [bib_ref] Systemic treatment options in hepatocellular carcinoma, Rimassa [/bib_ref] Biopsy-guided diagnosis also has limitations due to access (e.g., obscured tumor location, presence of ascites) and concern for increased risk of bleeding. [bib_ref] Controversies in the management of hepatocellular carcinoma, Forner [/bib_ref] [bib_ref] When and how should we perform a biopsy for HCC in patients..., Russo [/bib_ref] The use of systemic therapies for patients with advanced HCC, especially those not candidates for resection, transplantation, or local ablative therapies, have recently increased. [bib_ref] Hepatocellular carcinoma treatment: hurdles, advances and prospects, Kumari [/bib_ref] [bib_ref] Stereotactic body radiation therapy vs. transarterial chemoembolization in inoperable Barcelona clinic liver..., Su [/bib_ref] However, toxicity prevalence has been associated with dose reduction or treatment interruption. [bib_ref] Hepatocellular carcinoma: a review, Balogh [/bib_ref] Due to the plethora of novel systemic therapy options for HCC, new challenges for HCPs have arisen related to staying updated with those rapidly evolving treatment options, and the ideal regimen sequencing. [bib_ref] Systemic therapy of advanced hepatocellular carcinoma, Galle [/bib_ref] The complexity and heterogeneity of HCC requires treatment selection to be highly individualized and, ideally, multidisciplinary in nature, integrating numerous indices such as tumor burden and stage, performance status, presence of comorbidities, quality of life (QoL), and patient preferences. [bib_ref] Heterogeneity of hepatocellular carcinoma on imaging, Rimola [/bib_ref] [bib_ref] New advances in the diagnosis and management of hepatocellular carcinoma, Yang [/bib_ref] However, such multidisciplinary settings are not uniformly available, and while palliative care is ideally integrated into care planning early in the patient trajectory, unfortunately, these services are currently underutilized for patients with HCC. [bib_ref] Palliative care in hepatocellular carcinoma, Laube [/bib_ref] As a cancer with increasing global incidence that has a propensity for late-stage diagnosis and currently lacks effective curative treatment options, there is a need to better understand and address potential challenges that HCC healthcare providers (HCPs) face. HCPs who are wellinformed and who more effectively leverage the knowledge and experience of their colleagues (and coordinate their care delivery as a member of the HCP team) are better able to inform the patient regarding their care pathway. The primary objective of this study was to identify and categorize team-based clinical practice gaps and challenges experienced by US-based medical oncologists (MOs), hepatologists (HEPs), oncology physician assistants (PAs), oncology nurse practitioners (NPs), and interventional radiologists (IRs) involved in the care of patients with HCC. The identification of these issues can provide a starting point for clinicians to reflect on their practice and for both clinicians and educators to recognize opportunities for learning and improvement. # | methods A mixed-methods sequential design combining a qualitative exploratory data collection phase with a quantitative validation phase was utilized. [bib_ref] Research design for mixed methods: a triangulation-based framework and roadmap, Turner [/bib_ref] The qualitative exploration consisted of 45-minute semi-structured interviews with open-ended questions to explore the HCP's perceptions about screening, diagnosis, and treatment of HCC. Additionally, participants were queried about the nature and quality of their inter-provider and patient-provider communication. Findings from this qualitative phase informed the development of a 20-min online survey (quantitative validation phase). The study was approved by Veritas IRB (QC, Canada), an international, independent ethical review board registered with the United States Department of Health and Human Services (DHHS). ## | recruitment Potential participants were initially identified using two separate ICC/ESOMAR guideline compliant panels.Each panel contains validated healthcare professionals who volunteered to receive study invitations via email. A subsample of each panel, restricted to the targeted professions and specialty, was sent an email invitation containing a secure link to an online screener.Responses provided to the online screener were used to determine eligibility (see next paragraph), and only eligible participants were presented the consent form. Those who consented to participate were K E Y W O R D S cancer management, hepatocellular carcinoma, liver cancer, medical oncology, screening redirected either to the interview availability form or the online survey. Participants were compensated in accordance with the nature of their participation (interview/survey) and their profession in alignment with best practices and ethics. [bib_ref] Differential payment to research participants in the same study: an ethical analysis, Persad [/bib_ref] ## | research criteria Eligibility criteria was established in accordance with the project focus and objectives through discussions between education experts (including GJ, PL, PM) and clinical oncology practitioners (DAB, HBE, RJL, SMS). Participants were required to have an active practice for a minimum of 3 years as a medical oncologist (MO), hepatologist (HEP), oncology PA or NP, or interventional radiologist (IR) in the United States, with a minimum monthly caseload of 10 HCC patients for MO/HEP, 3 for PA/NP, and 5 for IR. A purposive sampling methodology was used to ensure that participants represented multiple perspectives in terms of practice settings (i.e., academic, community), years of practice, and geographic location, (i.e., rural, suburban, urban). [bib_ref] Purposeful sampling for qualitative data collection and analysis in mixed method implementation..., Palinkas [/bib_ref] ## | data collection A non-exhaustive review of the literature, in tandem with consultations between education experts and clinical oncology practitioners, guided the development of the qualitative data collection instrument. The semi-structured interview guide included open-ended questions and probes for trained interviewers to elicit elaboration on reported professional challenges. The interviews (conducted March to May 2020) were recorded with participant consent and then transcribed. An online survey, informed by the qualitative findings, was subsequently developed and administered (August-September 2020). The survey contained 146 to 149 items (depending on profession) and was designed to evaluate participants' level of knowledge, skills, confidence, and agreement to statements specific to HCC care. A five-point Likert-type scale was used to quantify knowledge and skill levels (1 = no knowledge/ skill; 5 = expert knowledge/skill). Confidence items utilized a visual analogue scale (0 = not at all confident; 100 = highly confident). Agreement items included a fivepoint Likert-type scale (1 = strongly disagree; 5 = strongly agree). Each item also included the option to respond, "Not relevant in my current role." # | analysis Interview transcripts were coded and analyzed through an approach drawing from the tenets of directed content analysis [bib_ref] Directed qualitative content analysis: the description and elaboration of its underpinning methods..., Assarroudi [/bib_ref] and thematic analysis [bib_ref] Synthesising qualitative and quantitative evidence: a review of possible methods, Dixon-Woods [/bib_ref] with NVivo software (QSR International Pty Ltd, . The coding tree was developed a priori based on the interview guide structure and then refined as details emerged from the data. Quantitative data from the survey were analyzed through cross-tabulations, analysis of variance (ANOVA), and Kruskal-Wallis H tests with SPSS 26.0 software (IBM Corporation, Armonk, NY) to identify differences in knowledge, skills, confidence, and agreement levels according to profession, years of practice, and type of setting. The five-point Likert-type scale for knowledge/ skill items was recoded into 2 categories: "sub-optimal" (1 = none, 2 = basic, 3 = intermediate) and "optimal" (4 = advanced, 5 = expert)." When over 30% of a respondent sub-group reported "sub optimal" knowledge or skills, this was considered an indication of a gap or educational need. Confidence levels were determined to be "sub optimal when the mean was below 80. The five-point Likert scale for agreement items was recoded in 3 categories: "disagree/strongly disagree," "neither agree nor disagree," and "agree/strongly agree." Qualitative and quantitative data were then combined using a triangulation of sources (five professions), methods (qualitative, quantitative), and investigational perspectives (multidisciplinary interpretation between educational and clinical experts). [bib_ref] Research design for mixed methods: a triangulation-based framework and roadmap, Turner [/bib_ref] # | results A total of 214 HCPs participated in this study: MOs (n qual = 8; n quant = 48), HEPs (n qual = 8; n quant = 45), oncology PAs (n qual = 2; n quant = 19), oncology NPs (n qual = 6; n quant = 26), and IRs (n qual = 8; n quant = 44). The majority (62%) of survey respondents practiced in community settings, while qualitative phase participants were evenly split between academic and community settings. Participants were generally well distributed across the US, although the West region was slightly underrepresented. Demographic data of participants are detailed in . Mixed method analysis revealed three challenges reported by the HCC team related to screening and diagnosis and five challenges related to treatment selection and management. ## | screening and diagnosis 3.1.1 | Applying appropriate screening criteria/guidelines Quantitative data revealed that 40% of all HCPs "agreed/ strongly agreed" with the statement, "There is a lack of reliable guidelines on HCC screening," with variability in agreement according to profession . Suboptimal knowledge of the AASLD screening guidelines for HCC was reported by most -PAs (63%), NPs (69%), and IRs (65%) (Table 2-A). Regarding the NCCN screening guidelines for HCC, suboptimal knowledge was also reported by most-HEPs (58%), PAs (58%), NPs (50%), and IRs (65%) [fig_ref] Table 2 -: H/I/J [/fig_ref]. Similarly, suboptimal knowledge was also reported regarding the NCI PDQ HCC screening guidelines by the majority of HEPs (60%), PAs (74%), NPs (69%), and IRs (70%) [fig_ref] Table 2 -: H/I/J [/fig_ref]. Across all professional groups, an average of 78% of all HCPs "agreed/strongly agreed" that, "There is a lack of T A B L E 1 Description of sample by phase (qualitative and quantitative) and speciality awareness about HCC screening at the primary care level" , a concern that HCPs also expressed during the qualitative phase . PAs (47%), NPs (46%), and IRs (51%) reported suboptimal skills determining the need for ongoing surveillance following a first screening for HCC [fig_ref] Table 3 -: B [/fig_ref]. HCPs also reported suboptimal mean levels of confidence in developing an action plan based upon the results of screening: between (mean ± standard deviation; 0-100 scale) 64 ± 20 for PAs and 78 ± 19 for HEPs 3.1.2 | Using imaging for diagnosis MOs (34%), PAs (58%), and NPs (46%) reported suboptimal knowledge of contraindications for imaging modalities [fig_ref] Table 2 -: H/I/J [/fig_ref]. Qualitative data also revealed how patient-specific variables can make the diagnosis challenging [fig_ref] Table 2 -: H/I/J [/fig_ref]. Suboptimal skills interpreting imaging results with atypical presentations were reported by MOs (43%), HEPs (38%), PAs (79%), and NPs (92%) ( ## | biopsy decision-making Twenty-seven percent of MOs, 47% of PAs, and 54% of NPs reported suboptimal skills determining which cases necessitate a biopsy of a suspicious lesion/mass [fig_ref] Table 3 -: B [/fig_ref]. Suboptimal confidence when deciding if a biopsy of a suspicious lesion/mass is required was also reported by MOs (74 ± 16), PAs (66 ± 19), and NPs (62 ± 22) . Qualitative data characterized the difficulties experienced when imaging findings were inconclusive (Table 6-Q4). ## | treatment selection and management ## | applying guidelines in treatment decisions On average, 44% of HCPs "agreed/strongly agreed" that, "Treatment decisions for HCC are difficult because there are too many options" . Interviewed participants reported that existing guidelines do not always optimally support decision-makers in considering multiple treatment options -Q5/Q6). ## | balancing risks and benefits of treatment Quantitative survey data revealed suboptimal knowledge of potential side effects of TKIs (55%), immunotherapeutics (49%), and conventional chemotherapy agents (43%), with variation between profession groups (Table 2-E/F/ G). Interviewees also expressed difficulty finding relevant information on side effect prevalence for newer therapy options -Q7). HEPs (38%), PAs (53%), NPs (54%), and IRs (62%) reported suboptimal skills identifying the safest and most effective treatment option for a specific patient [fig_ref] Table 3 -: B [/fig_ref]. Interviewees expressed the perception that the toxicities of newer agents seem to generally outweigh the benefits of their efficacy . Suboptimal skills managing toxicities to enhance QoL were also reported by HEPs (53%), PAs (37%), and IRs (82%) ( ## | patient profiles and co-morbidities HCPs reported suboptimal knowledge of best practices to treat HCC in patients with comorbid renal failure (59%), lung disease (59%), and cardiovascular disease (57%), with variations between profession groups ( ## | sequencing treatments Knowledge of best practices in treatment sequencing, and skills determining the ideal sequencing of treatments were reported as suboptimal for all professions, ranging from 29% of MOs to 74% of IRs for knowledge [fig_ref] Table 2 -: H/I/J [/fig_ref] # | discussion This study identified multiple challenges experienced by HCPs involved in HCC care with respect to several critical decision points. In relation to screening and diagnosis, these included: (1) adherence to established screening criteria; (2) use of appropriate imaging for diagnosis; (3) making the decision to perform a diagnostic biopsy. In relation to treatment and management, these critical decision points included: (1) applying guidelines in therapeutic decision-making; (2) balancing risks/benefits of treatment regimens; (3) managing various patient profiles and comorbidities; (4) optimal sequencing of modalities; and (5) addressing patients' expectations. The perception that practice guidelines for HCC screening are inadequate has been reported in this study and elsewhere: for example, in a recent evaluation of HCC guidelines by radiation oncologists, the lowest overall score of any of the domains was "applicability," among the 18 guidelines examined. [bib_ref] An evaluation of hepatocellular carcinoma practice guidelines from a radiation oncology perspective, Rim [/bib_ref] This was attributed to the inadequacy of guidelines to adapt to the rapidly-changing treatment landscape and the necessary multi-disciplinary nature of HCC care. [bib_ref] An evaluation of hepatocellular carcinoma practice guidelines from a radiation oncology perspective, Rim [/bib_ref] A general assessment of international HCC guidelines posits that certain challenges (i.e., regional variations in care practices, resources, disease prevalence) make pursuit of universal, applicable, and reliable HCC guidelines unrealistic. [bib_ref] A concise review of updated guidelines regarding the management of hepatocellular carcinoma..., Yu [/bib_ref] Existing guidelines are perceived to be inconsistent in relation to recommendations for identifying high-risk individuals and factoring in the combination of ultrasonography and testing serum AFP levels due to false positives and imprecision. [bib_ref] A concise review of updated guidelines regarding the management of hepatocellular carcinoma..., Yu [/bib_ref] Some of these inconsistencies relate to lack of high-level dependable evidence for the efficacy of the proposed interventions. Our study found that there were significant differences among professions and specialties in their perception of guidelines, with HEP and IR finding them less reliable. This may be explained by the different professional roles (i.e., IRs are not expected to be involved directly in the screening of HCC patients) and by recent controversies regarding both the efficacy and best practices for HCC screening that more directly impact these professionals. [bib_ref] Validation of the updated hepatocellular carcinoma early detection screening algorithm in a..., Tayob [/bib_ref] [bib_ref] The threshold of alphafetoprotein (AFP) for the diagnosis of hepatocellular carcinoma: a..., Zhang [/bib_ref] [bib_ref] Hepatocellular carcinoma: an overview of the changing landscape of treatment options, Koulouris [/bib_ref] The consequences of inadequate or unclear guidelines also impact primary care providers who play an important role in screening for HCC. A lack of knowledge of HCC screening practices, including misconceptions about HCC surveillance, was reported in a 2019 study of primary care providers. [bib_ref] Primary care provider practice patterns and barriers to hepatocellular carcinoma surveillance, Simmons [/bib_ref] Despite these challenges, the scope and scale of the issue is becoming better understood. As the quality of evidence needed to support better guidelines improves, knowledge translation and implementation science will help increase awareness and enhance application of the updated guidelines. [bib_ref] A concise review of updated guidelines regarding the management of hepatocellular carcinoma..., Yu [/bib_ref] Promoting an open exchange of information and experiences between all members of the healthcare team can help deliver a well-informed, evidence-based, cohesive care plan for patients with HCC. In this research, determining an individualized patient care pathway was complicated by challenges specific to interpreting unclear or atypical imaging results. delays (>60 days) in establishing an HCC diagnosis in patients with cirrhosis are associated with guideline nonadherence on the part of the HCP. [bib_ref] Factors associated with delay of diagnosis of hepatocellular carcinoma in patients with..., Choi [/bib_ref] We posit that timeliness and access to rigorous multidisciplinary specialist care could mitigate these diagnostic errors and delays and improve patient outcomes. Patients residing in remote or rural areas often have later clinical presentation and decreased survival rates. [bib_ref] Remoteness of residence predicts tumor stage, receipt of treatment, and mortality in..., Taye [/bib_ref] However, a case review by an interdisciplinary tumor board has the potential to reduce mortality when completed within 30 days of diagnosis. [bib_ref] Association of provider specialty and multidisciplinary care with hepatocellular carcinoma treatment and..., Serper [/bib_ref] In the absence of formal board reviews, creating an environment that promotes and helps to facilitate case discussions between a broad range of members of the healthcare team can be a valuable mechanism to identify diagnostic challenges and pursue appropriate treatment options while minimizing delays. The increased emphasis on multi-disciplinary cancer care is likely to reduce overall discrepancies in the level of knowledge of toxicity found between the professions. Patients often present first with hepatic symptoms and are then treated by a hepatology team. Although patients with early-stage disease being considered for curative intent therapy may not require seeing a medical oncologist, many patients are detected later and should be referred. Many patients present symptoms and imaging data that do not fully reflect HCC diagnostic criteria and thus may not always be referred to an oncologist. This creates siloed care where HCPs are unaware of the need for close collaboration, knowledge of current treatment implications, or the need for effective plans to manage toxicities as a team comprised of oncologists and radiologists, hepatologists, NPs, and PAs. [bib_ref] Clinical characterisation and management of the main treatment-induced toxicities in patients with..., Meriggi [/bib_ref] [bib_ref] Adverse effects of immunecheckpoint inhibitors in hepatocellular carcinoma, Cui [/bib_ref] When a multidisciplinary team is in place, it requires coordinated communication and collaboration which can be complex, especially in light of the fact that quite often four or more providers are concurrently involved in decision-making. In addition to suboptimal skills, such collaboration often is complicated by scheduling issues and limited availability. Interpreting liver imaging results and determining when a biopsy should be done was reported as a challenge. Since biopsy in patients with suspected HCC comes with delays, 38 studies suggest a multidisciplinary and patientspecific approach to these decisions based on a combination of screening approaches and methods of collecting patient data. [bib_ref] Atypical appearance of hepatocellular carcinoma and its mimickers: how to solve challenging..., Kim [/bib_ref] Clinical decision-making algorithms could improve screening and identification of HCC, even with atypical presentation, using radiomics and artificial intelligence. [bib_ref] Radiomics machinelearning signature for diagnosis of hepatocellular carcinoma in cirrhotic patients with..., Mokrane [/bib_ref] These deep learning technologies aid in decision-making for suspected HCC lesions that do not fit the Liver Imaging Reporting and Data System [LI-RADS] criteria. [bib_ref] Deep learningassisted differentiation of pathologically proven atypical and typical hepatocellular carcinoma (HCC)..., Oestmann [/bib_ref] We found that HCPs have suboptimal skills related to improving patients' QoL and suboptimal confidence in making changes to reduce side effects and promote adherence. A study of patient perspectives indicated they want more information throughout their care and would benefit from the services of patient advocates, especially to assist them in navigating decisions resulting from the often-difficult side effects of available treatments. [bib_ref] Insights into the hepatocellular carcinoma patient journey: results of the first global..., Gill [/bib_ref] Improvements in HCC management can be addressed through continuing medical education (CME) and continuing professional development (CPD) activities designed to build skill in establishing trust with the patient and setting realistic expectations in terms of potential treatment side effects. Also, the addition of patient advocate resources and improved patient education may assist in building a base of knowledge within the patient and, in so doing, alleviating the impact of HCC on well-being and autonomy, as well as promoting patient's active participation in the management of the condition. Recent approvals and improvement in systemic and combination treatment options for HCC 1,43 occurred during the data collection phase of this study. Regardless, concerns about treatment safety and management of side effects of antineoplastics remain and the addition of new HCC treatment modalities may introduce additional challenges; for example, some patients experience rapid disease progression following treatment with immune checkpoint inhibitors 12 and decision-making for each type of treatment becomes more challenging due to a lack of evidence for optimal therapy sequencing. [bib_ref] Epidemiology and surveillance for hepatocellular carcinoma: new trends, Singal [/bib_ref] This study highlighted significant differences among professions in their knowledge of guidelines, plus their level of skill interpreting imaging, treating patients, and ensuring QoL, as some competencies are more-directly related to some roles as compared to others. This supports the position of previous studies that HCC, due to its nature and complexity, requires a care team with a wide range of knowledge, experience, and competencies within the context of a multidisciplinary approach. [bib_ref] Clinical characterisation and management of the main treatment-induced toxicities in patients with..., Meriggi [/bib_ref] [bib_ref] Adverse effects of immunecheckpoint inhibitors in hepatocellular carcinoma, Cui [/bib_ref] [bib_ref] Factors and survival implications associated with biopsy of hepatocellular carcinoma, Rho [/bib_ref] [bib_ref] Atypical appearance of hepatocellular carcinoma and its mimickers: how to solve challenging..., Kim [/bib_ref] [bib_ref] Radiomics machinelearning signature for diagnosis of hepatocellular carcinoma in cirrhotic patients with..., Mokrane [/bib_ref] [bib_ref] Deep learningassisted differentiation of pathologically proven atypical and typical hepatocellular carcinoma (HCC)..., Oestmann [/bib_ref] [bib_ref] Insights into the hepatocellular carcinoma patient journey: results of the first global..., Gill [/bib_ref] [bib_ref] Role of multidisciplinary care in the management of hepatocellular carcinoma, Byrd [/bib_ref] It is crucial that members of the healthcare team recognize and respect the role that each one plays in delivering optimal patient care. # | limitations This study's findings stem from self-reported data rather than empirical observations. To minimize self-reporting biases such as social desirability, [bib_ref] Information bias in health research: definition, pitfalls, and adjustment methods, Althubaiti [/bib_ref] our methodology included the use of triangulation (i.e., the combination of different data sources, research methodologies and/or interpretation viewpoints in the study of the same phenomenon) [bib_ref] Bias in research, Smith [/bib_ref] and maximum variation purposive sampling (i.e., where participants are selected to represent a broad spectrum of perspectives). [bib_ref] Purposeful sampling for qualitative data collection and analysis in mixed method implementation..., Palinkas [/bib_ref] Results from the qualitative phase may have been limited by the low sample sizes of some sub-groups (e.g., PAs); however, triangulation with the quantitative findings has ensured the trustworthiness of the overall findings. Caution should be taken when generalizing the findings to other professions/ specialties involved in HCC or to other countries. More studies should be done to inform the development of local educational activities/offerings and region-specific and setting-specific needs assessments should be conducted to ensure the benefits of developing precise activities for the targeted learners and the needs of the patient population. # | conclusion This study illuminated the educational needs of providers involved in the spectrum of care of patients with HCC. It highlighted needs to improve the use, content, and knowledge of guidelines for HCC, as well as to enhance skills needed to appropriately screen and diagnose HCC and to improve decision-making when faced with HCC lesions with atypical presentation. Confidence and skills to enhance QoL were also found to be lacking, despite being critically important for HCC (as a chronic, complex, and severe condition). [bib_ref] Challenges of advanced hepatocellular carcinoma, Colagrande [/bib_ref] Although these challenges could be ameliorated in part by CME/CPD, an emphasis on implementing a multi-disciplinary team approach to HCC care may be equally integral to improving patient outcomes, as the complicated and diverse presentation of HCC, and the number of HCPs involved in care must be better coordinated to optimize the complementary skills of a diverse team. These findings should be taken into consideration by clinicians in their continuous reflection to improve their practice, and by educators when developing educational interventions on early diagnosis and proper management of patients with HCC, especially as the demand for evidence-based CME/CPD increases. [bib_ref] Continuing professional development: progress beyond continuing medical education, Filipe [/bib_ref] AUTHOR CONTRIBUTIONS Ginny Jacobs: Conceptualization (lead); funding acquisition (equal); investigation (equal); methodology (equal); project administration (equal); supervision (lead); validation (equal); writing -original draft (equal); writing -review and editing (equal). Deborah A. Boyle: Methodology (equal); validation (equal); writing -review and editing (equal). Hashem B. El-Serag: Methodology (equal); validation (equal); writing -review and editing (equal). Robert J Lewandowski: Methodology (equal); validation (equal); writing -review and editing (equal). Stacey Stein: Methodology (equal); validation (equal); writing -review and editing (equal). Patrice Lazure: Conceptualization (supporting); data curation (lead); formal analysis (lead); investigation (lead); methodology (equal); project administration (supporting); supervision (equal); validation (equal); visualization (equal); writingoriginal draft (equal); writing -review and editing (equal). Pam McFadden: Conceptualization (equal); funding acquisition (lead); methodology (equal); project administration (equal); supervision (equal); validation (equal); writing -review and editing (equal). ## Acknowledgement We acknowledge the support provided by Morgan Peniuta and Monica Augustyniak (researchers, AXDEV Group), as well as Olivier Jacob (project manager, AXDEV Group) who supported different aspects of the research. We thank all the participants who took part in this study. # Funding information This study was financially supported by Eisai Co., Ltd. ## Conflict of interest GJ and PM are employees of AXDEV Global Inc. DAB has nothing to disclose. HBE is supported by NIH P30DK056338. RJL acted as an advisor to BD, Boston Scientific Corporation, Varian, ABK Medical, and Alhambra Medical. SMS has received compensation for advisory roles with Merck, Genentech, Exelixis, QED, and Astra Zeneca. PL is an employee of AXDEV Group Inc. While AXDEV Global and AXDEV Group are private for-profit companies, they are specialized in educational research with a focus on behavioral and implementation sciences. They do not engage in clinical research or other practice-related business and, therefore, the employment status of the three AXDEV coauthors does not constitute any conflict of interest with the findings of the study. # Data availability statement The data that supports the findings of this study are available upon reasonable request to the corresponding author. Due to privacy and ethical restrictions, the data is not publicly available. ## Orcid Patrice Lazure https://orcid.org/0000-0002-9278-1718 [table] Table 3 -: B). Interviewees shared that initial imaging results do not always meet the HCC criteria such as in the case of patients having atypical features, which renders the diagnosis difficult to establish (Table 6-Q3). [/table] [table] Table 2 -: H/I/J). Suboptimal skills making treatment decisions for patients with multiple comorbidities were also reported by HEPs (47%), PAs (58%), NPs (46%), and IRs (70%) (Table 3-F). Interviewees expressed concerns about the heightened risks for patients with comorbidities, which limits the number of options available (Table 6-Q10). [/table]

Effects of omega-3 fatty acids on patients undergoing surgery for gastrointestinal malignancy: a systematic review and meta-analysis Background: Surgical resection remains the primary treatment for gastrointestinal (GI) malignancy including earlystage cancer. Omega-3 polyunsaturated fatty acids (n-3 PUFAs) have been reported to have beneficial clinical and immune-modulating effects in the prognosis of GI cancer patients undergoing surgery. Methods: We searched PubMed, Embase, EBSCO-Medline, Cochrane Central Register of Controlled Trials (CENTRAL), CNKI and Wanfang to identify primary research reporting the effects of n-3 PUFAs compared with isocaloric nutrition on GI cancer patients who underwent surgery up to the end of June 30, 2016. Two authors independently reviewed and selected eligible randomized controlled trials (RCTs). Results: A total of 9 RCTs (623 participants) were included. The n-3 PUFAs regime resulted in lower levels of Creactive protein (CRP) (P < 0.05), interleukin-6 (IL-6) (P < 0.01), and higher levels of albumin (ALB), CD3 + T cells, CD4 + T cells and CD4 + /CD8 + ratio (P < 0.05) compared with the isocaloric nutrition regime. However, there was no significant difference in the level of tumor necrosis factor-α (TNF-α) between the n-3 PUFAs regime and the isocaloric nutrition regime (P = 0.17). And the level of CD8 + T cells decreased compared with the isocaloric nutrition regime (P < 0.0001). Conclusions: Our meta-analysis revealed that n-3 PUFAs are effective in improving the nutritional status and immune function of GI cancer patients undergoing surgery as they effectively enhance immunity and attenuate the inflammatory response. Background GI cancers are the most common group of malignancies and many types of GI cancer are ranked as the leading cause of cancer death worldwide [bib_ref] Molecular-targeted first-line therapy for advanced gastric cancer, Song [/bib_ref] [bib_ref] PRISMAcompliant article: clinical characteristics and factors influencing prognosis of patients with Hepatoid..., Qu [/bib_ref]. Surgery is the primary treatment for patients with early-stage GI cancer. However, patients undergoing selective GI cancer surgery will face the risk of developing various postoperative complications due to negative impact factors, such as malnutrition, tumor-induced immune suppression, surgical stress, and inflammation. Postoperative complications affect the clinic outcome of patients, resulting in prolonged hospital-stay and increased costs. Of these complications, malnutrition is the most important factor influencing clinical prognosis [bib_ref] Postoperative complications in gastrointestinal cancer patients: the joint role of the nutritional..., Bozzetti [/bib_ref] [bib_ref] ESPEN guidelines on parenteral nutrition: surgery, Braga [/bib_ref]. Current studies indicate that nutritional support can reduce the incidence of adverse events after major GI surgery. Omega-3 polyunsaturated fatty acids (n-3 PUFAs) modulate the level of inflammation and reduce oxidative stress and complications [bib_ref] Nutrition assessment and its relationship with performance and Glasgow prognostic scores in..., Quyen [/bib_ref] [bib_ref] Nutritional management of short bowel syndrome in adults, Sundaram [/bib_ref] [bib_ref] Fish oil supplementation attenuates neuroinflammation and alleviates depressivelike behavior in rats submitted..., Dang [/bib_ref] [bib_ref] Fatty acids and inflammation: the cutting edge between food and pharma, Calder [/bib_ref]. The evidence from these studies indicates that n-3 PUFAs have an anti-inflammatory effect, which promotes wound healing, and enhances the adaptive immune response [bib_ref] Effect of immune-enhanced enteral diet on postoperative immunological function in patients with..., Chen [/bib_ref] [bib_ref] Role of enteral immunonutrition in patients with gastric carcinoma undergoing major surgery, Chen [/bib_ref]. However, interpretation of these studies is problematic due to methodological limitations and small sample sizes. Moreover, the results of several recent RCTs are controversial. Thus, the purpose of this systematic review is to evaluate the potential role of n-3 PUFAs in the outcome of GI cancer patients after surgery. # Methods ## Research design We searched PubMed (January 1, 1976, through April 30, 2016), EMBASE (January 1, 1985, through , the Cochrane Library (January 1, 1987, through , CNKI (January 1, 1986, through April 30, 2016), Wanfang (January 1, 1985, through April 30, 2016) and VIP databases (January 1, 1985, through April 30, 2016) using common keywords related to n-3 PUFAs and GI cancer. The following key words were included: n-3 PUFAs, eicosapentaenoic acid or EPA, docosahexaenoic acid or DHA, gastrointestinal malignancy or cancer surgery. We reviewed the bibliographies of relevant articles for additional publications. ## Selection criteria We included trials that met the following four criteria: (1) the trial enrolled adult patients (male or female aged at least 18 years) undergoing surgery for GI malignancy; (2) the trial design was randomized, double blind, and placebo-controlled; (3) the trial compared n-3 PUFAs support with isocaloric nutrition; (4) the trial reported outcome measures such as CD3 + T cells, CD4 + T cells, CD8 + T cells, CD4 + /CD8 + T cells, ALB, IL-6, TNF-α, and CRP; (5) the study did not include obese patients and there was no difference in body mass index (BMI) between the groups. ## Data extraction Two co-first authors reviewed all the articles independently and discussed the articles until a consensus was reached. Data obtained from the studies included the first author, year of publication, patient source (region), tumor types, and type of study. All data were extracted independently by two investigators. As all the studies were RCTs, we summarized the basic parameters and then assessed the quality of the included studies. ## Quality evaluation We assessed the methodological quality of the included studies using the scale of Risk of bias summary and Risk of bias graph, which is the most widely used assessment tool in meta analyses. The scale measures the following characteristics in RCTs: 1. random sequence generation (selection bias), 2. allocation concealment (selection bias), 3. blinding method used for participants and study personnel (performance bias), 4. blinding method used for outcome assessment (detection bias), 5. incomplete outcome data (attrition bias), 6. selective reporting (reporting bias), 7. other biases. The risk of each included study was rated as "high bias risk", "unclear bias risk" or "low bias risk" according to the information extracted. The graphical results of methodological quality are shown in [fig_ref] Figure 2: Assessment of risk of bias based on the evaluation domains listed in... [/fig_ref]. # Statistical analysis The levels of CRP, IL-6 and TNF-α, ALB, CD3 + T cells, CD4 + T cells, CD8 + T cells, and CD4 + /CD8 + T cells were calculated using the Review Manager 5.0.24 statistical software (Cochrane Collaboration Software). Publication bias was evaluated according to a funnel plot and Begg's and Egger's tests using the Review Manager 5.0.24 package. Heterogeneity was considered statistically significant when P < 0.05. # Results ## Characteristics of the included studies and risk of bias The electronic literature search yielded 672 potential studies for inclusion. And finally, 126 articles had titles and abstracts that appeared to be potentially relevant. Of these studies, 54 studies were excluded because the patients received arginine. 8 studies were reported neither in Chinese nor in English and were thus excluded; 55 studies with full texts were further excluded as the patients received chemotherapy. All procedures were performed by two investigators independently. In total, 9 eligible studies were included in this meta-analysis. The flow chart of retrieval and selection of the studies is shown in [fig_ref] Figure 1: Flowchart of computerized search and the eligible studies included in this systematic... [/fig_ref]. [fig_ref] Table 1: Characteristics of included randomized trials [/fig_ref] summarizes the basic characteristics of the included studies. Of the 8 studies included, 5 trials reported the association between fish oil consumption and the level of CRP [bib_ref] Prospective double-blind randomized study on the efficacy and safety of an n-3..., Ma [/bib_ref] [bib_ref] A prospective, randomized, controlled study of ω-3 fish oil fat emulsion-based parenteral..., Wei [/bib_ref] [bib_ref] Xin fu ze ea: the effect of fish oil on immunologic function..., Xin [/bib_ref] [bib_ref] The role of omega-3 polyunsaturated fatty acid in parenteral nutrion treatment of..., Wei [/bib_ref] , 5 trials described the correlation between PUFAs and the level of IL-6 [bib_ref] Prospective double-blind randomized study on the efficacy and safety of an n-3..., Ma [/bib_ref] [bib_ref] A prospective, randomized, controlled study of ω-3 fish oil fat emulsion-based parenteral..., Wei [/bib_ref] [bib_ref] Impact of fish oil enriched total parenteral nutrition on elderly patients after..., Zhu [/bib_ref] [bib_ref] Enteral nutrition enriched with eicosapentaenoic acid (EPA) preserves lean body mass following..., Ryan [/bib_ref] , 4 trials investigated the association between n-3 PUFAs and the level of TNF-α [bib_ref] Prospective double-blind randomized study on the efficacy and safety of an n-3..., Ma [/bib_ref] [bib_ref] A prospective, randomized, controlled study of ω-3 fish oil fat emulsion-based parenteral..., Wei [/bib_ref] [bib_ref] Impact of fish oil enriched total parenteral nutrition on elderly patients after..., Zhu [/bib_ref] , 4 trials investigated the association between n-3 PUFAs and the level of ALB [bib_ref] A prospective, randomized, controlled study of ω-3 fish oil fat emulsion-based parenteral..., Wei [/bib_ref] [bib_ref] Xin fu ze ea: the effect of fish oil on immunologic function..., Xin [/bib_ref] [bib_ref] Enteral nutrition enriched with eicosapentaenoic acid (EPA) preserves lean body mass following..., Ryan [/bib_ref] , 7 trials investigated the association between n-3 PUFAs and the level of inflammation [bib_ref] Prospective double-blind randomized study on the efficacy and safety of an n-3..., Ma [/bib_ref] [bib_ref] A prospective, randomized, controlled study of ω-3 fish oil fat emulsion-based parenteral..., Wei [/bib_ref] [bib_ref] Xin fu ze ea: the effect of fish oil on immunologic function..., Xin [/bib_ref] [bib_ref] The role of omega-3 polyunsaturated fatty acid in parenteral nutrion treatment of..., Wei [/bib_ref] [bib_ref] Impact of fish oil enriched total parenteral nutrition on elderly patients after..., Zhu [/bib_ref] [bib_ref] Enteral nutrition enriched with eicosapentaenoic acid (EPA) preserves lean body mass following..., Ryan [/bib_ref] , and 7 trials described the correlation between n-3 PUFAs and immune functions . Nutritional status was classified by the Nutritional Risk Index (NRI). If the NRI was >100, the patient was not considered malnourished, 97.5-100 indicated mild malnutrition, 83.5-97.5 indicated moderate malnutrition, and <83.5 indicated severe malnutrition. However, there was no significant difference between the groups in terms of mean weight and BMI in the included studies. Of the included studies, 6 studies were from China [bib_ref] Prospective double-blind randomized study on the efficacy and safety of an n-3..., Ma [/bib_ref] [bib_ref] Xin fu ze ea: the effect of fish oil on immunologic function..., Xin [/bib_ref] [bib_ref] The role of omega-3 polyunsaturated fatty acid in parenteral nutrion treatment of..., Wei [/bib_ref] [bib_ref] Impact of fish oil enriched total parenteral nutrition on elderly patients after..., Zhu [/bib_ref] [bib_ref] The effect of n-3 fatty acid on immunologic function and inflammatory reaction..., Huang Jian [/bib_ref] , one study was from Brazil [bib_ref] A prospective, randomized, controlled study of ω-3 fish oil fat emulsion-based parenteral..., Wei [/bib_ref] , one study was from UK [bib_ref] Randomized clinical trial of omega-3 fatty acid-supplemented enteral nutrition versus standard enteral..., Sultan [/bib_ref] , and one study was from Ireland [bib_ref] Enteral nutrition enriched with eicosapentaenoic acid (EPA) preserves lean body mass following..., Ryan [/bib_ref]. All 9 studies were double-blind and allocation concealment was adequate in all studies. The risk of bias items for each included study are presented in [fig_ref] Figure 2: Assessment of risk of bias based on the evaluation domains listed in... [/fig_ref]. ## Level of inflammation CRP: We identified 5 eligible trials, which included 269 patients, and investigated peripheral blood CRP levels following postoperative n-3 PUFAs supplementation versus isocaloric nutrition. The homogeneous test detected no statistical heterogeneity (P = 0.12), therefore, we adopted a fixed-effects model to perform the analysis. The meta-analysis revealed that n-3 PUFAs effectively decreased the level of CRP (P < 0.05) [fig_ref] Figure 3: Meta-analysis of inflammation level [/fig_ref]. IL-6: We identified 5 eligible trials, which included 329 patients, and investigated IL-6 levels following postoperative n-3 PUFAs supplementation versus isocaloric nutrition. The homogeneous test detected no statistical heterogeneity (P = 0.15), therefore, we adopted a fixedeffects model to perform the analysis. The meta-analysis revealed that n-3 PUFAs effectively decreased the level of IL-6 (P = 0.005) [fig_ref] Figure 3: Meta-analysis of inflammation level [/fig_ref]. TNF-α: We identified 4 eligible trials, which included 276 patients, and investigated TNF-α levels following postoperative n-3 PUFAs supplementation versus isocaloric nutrition. The homogeneous test detected substantial statistical heterogeneity (P = 0.008), therefore, we adopted a random-effects model to perform the analysis. The meta-analysis revealed that TNF-α levels decreased following both n-3 PUFAs supplementation and isocaloric nutrition; however, there was no significant difference in TNF-α level between the two treatment groups (P = 0.17) [fig_ref] Figure 3: Meta-analysis of inflammation level [/fig_ref]. ALB: We identified 4 eligible trials, which included 181 patients, and investigated ALB levels following postoperative n-3 PUFAs supplementation versus isocaloric nutrition. The homogeneous test detected no statistical heterogeneity (P = 0.91), therefore, we adopted a fixedeffects model to perform the analysis. The meta-analysis revealed that n-3 PUFAs effectively increased the level of ALB (P < 0.01) [fig_ref] Figure 3: Meta-analysis of inflammation level [/fig_ref]. Immune status CD3 + T cells: We identified 6 eligible trials, which included 428 patients, and investigated CD3 + T cell levels following postoperative n-3 PUFAs supplementation versus isocaloric nutrition. The homogeneous test detected no statistical heterogeneity (P = 0.25), therefore, we adopted a fixed-effects model to perform the analysis. The meta-analysis revealed that n-3 PUFAs effectively increased the level of CD3 + T cells (P < 0.01) . CD4 + T cells: We identified 7 eligible trials, which included 485 patients, and investigated CD4 + T cell levels following postoperative n-3 PUFAs supplementation versus isocaloric nutrition. The homogeneous test detected substantial statistical heterogeneity (P < 0.00001), therefore, we adopted a random-effects model to perform the analysis. The meta-analysis revealed that n-3 PUFAs effectively increased the level of CD4 + T cells (P = 0.03) . CD8 + T cells: We identified 6 eligible trials, which included 445 patients, and investigated CD8 + T cell levels following postoperative n-3 PUFAs supplementation versus isocaloric nutrition. The homogeneous test detected substantial statistical heterogeneity (P = 0.01), therefore, we adopted a random-effects model to perform the analysis. The meta-analysis revealed that n-3 PUFAs effectively decreased the level of CD8 + T cells (P = 0.03) . CD4 + /CD8 + T cells: We identified 7 eligible trials, which included 485 patients, and investigated CD4 + /CD8 + levels following postoperative n-3 PUFAs supplementation versus isocaloric nutrition. The homogeneous test detected no statistical heterogeneity (P = 0.15); therefore, we adopted a fixed-effects model to perform the analysis. The meta-analysis revealed that n-3 PUFAs effectively increased the level of CD4 + /CD8 + T cells (P < 0.00001) . ## Publication bias There was no evidence of publication bias following assessment by funnel plot, Egger's test (P > 0.05) and Begg's test (P > 0.05). # Discussion The ASPEN guide recommends that for patients with large tumors undergoing surgery, a variety of immune nutrients in the nutritional formulation are conducive for improving prognosis. It is best to start nutritional support 5-7 days before surgery, and it should be continued into the postoperative period [bib_ref] American Society for Parenteral and Enteral Nutrition (ASPEN) Board of Directors. ASPEN..., August [/bib_ref]. N-3 PUFAs have been reported to have a role in enhancing host immunity and attenuating the inflammatory response in GI cancer patients undergoing surgery [bib_ref] Fatty acid patterns and risk of prostate cancer in a case-control study..., Dahm [/bib_ref]. There is evidence to suggest that n-3 PUFAs play an important role in the host immune response and inflammatory reaction in GI cancer, thus n-3 PUFAs are the best option for postoperative management compared with isocaloric nutrition [bib_ref] Interleukin-1 and tumor necrosis factor synthesis by mouse peritoneal macrophages is enhanced..., Lokesh [/bib_ref] [bib_ref] Effects of oils rich in eicosapentaenoic and docosahexaenoic acids on immune cell..., Kew [/bib_ref] [bib_ref] Dietary fatty acids influence the production of Th1-but not Th2-type cytokines, Wallace [/bib_ref] [bib_ref] The effects of dietary lipid manipulation on the production of murine T..., Yaqoob [/bib_ref]. We conducted a systematic review based on eight RCTs involving 583 patients and evaluated the impact of n-3 PUFAs on postoperative inflammation status and immune function. The results of our study showed that n-3 PUFAs significantly decreased the level of inflammation and increased immune function. N-3 PUFAs are beneficial as a dietary supplement for cancer patients as they reduce the level of inflammatory cytokines, including IL-2, IL-6, as well as TNF-α, and promote anti-inflammatory activities. IL-6, an inflammatory cytokine, can down-regulate the stress response, and mainly originates from immune cells (e.g., T cells), endotheliocytes, and macrophages. It can effectively modulate the immune system and fight infection. Serum ALB is a negative acute phase protein and ALB concentration has important roles in the regulation of inflammation [bib_ref] Dietary fatty acids influence the production of Th1-but not Th2-type cytokines, Wallace [/bib_ref] , while CRP is a marker of acute inflammation. Many previously published studies have revealed that n-3 PUFAs can down-regulate the levels of IL-6 and TNF-α in cancer patients postoperatively [bib_ref] Serum albumin: relationship to inflammation and nutrition, Don [/bib_ref] [bib_ref] Oncogenic Ras-induced secretion of IL6 is required for tumorigenesis, Ancrile [/bib_ref] [bib_ref] Role of BNIP3 in TNF-induced cell death-TNF upregulates BNIP3 expression, Ghavami [/bib_ref] [bib_ref] Interleukin-6 stimulates clonogenic growth of primary and metastatic human colon carcinoma cells, Schneider [/bib_ref] [bib_ref] Serum interleukin-6 levels in colorectal cancer patients-a summary of published results, Knüpfer [/bib_ref]. The trial by Turnocket al. revealed that perioperative administration of n-3 PUFAs suppressed the level of CRP in patients undergoing surgery for GI malignancy [bib_ref] Perioperative immunonutrition in well-nourished patients undergoing surgery for head and neck cancer:..., Turnock [/bib_ref]. High EPA and DHA intake, both of which are n-3 PUFAs, was closely related to a reduction in the level of CRP, which indicated a better prognosis. In addition, a nutritional supplement enriched with n-3 PUFAs has shown advantages in serum ALB levels in patients with head and neck cancer [bib_ref] A systematic review and meta-analysis of the n-3 polyunsaturated fatty acids effects..., Mocellin [/bib_ref]. Vasson [bib_ref] Immunonutrition improves functional capacities in head and neck and esophageal cancer patients..., Vasson [/bib_ref] confirmed that immunonutrition improves albuminemia in head and neck and esophageal cancer patients undergoing radiochemotherapy. The results of our meta-analysis are in accordance with these reports, in which n-3 PUFAs reduced host inflammatory response by decreasing the concentration of IL-6, TNF-α, and CRP, and improving hypoalbuminemia. The antiinflammatory response plays an important role in patients with GI cancer [bib_ref] Enteral immuno-enhanced diets with arginine are safe and beneficial for patients early..., Aiko [/bib_ref] [bib_ref] Changes in micronutrient concentrations following anti-inflammatory treatment in patients with gastrointestinal cancer, Mcmillan [/bib_ref]. N-3 PUFAs may be of benefit in down-regulating the strong and discordant inflammatory response which occurs after surgery. N-3 PUFAs are beneficial as a dietary supplement in cancer patients as they enhance immune functions. N-3 PUFAs have been recognized as having immunomodulatory activity, including the activation of T cells and cytokine production [bib_ref] Omega-3 polyunsaturated fatty acids for the treatment and prevention of colorectal cancer, Cockbain [/bib_ref]. CD4 + and CD8 + T cells are important effector cells of cell-mediated immunity. CD8 + T cells are strong effector T cells. All mature T cells express CD3 + ; CD3 + and CD4 + T cells are helper T lymphocytes that promote anti-tumor immunity. CD8 + cells are suppressor T lymphocytes. Presentation of intracellular antigen on MHC class I molecules activates CD8 + T cells, cytotoxic T lymphocytes that will attempt to suppress the intracellular infection. If this does not succeed, the CD8 + T cell will kill the target cell by inducing apoptosis or cell lysis. Elevation of CD4 + /CD8 + ratio, CD3 + and CD4 + lymphocyte percentage were also observed as a result of n-3 PUFAs supplementation. It is essential to understand precisely how specific (n-3) PUFAs modulate immune function. Turbitt [bib_ref] Fish oil enhances T cell function and tumor infiltration and is correlated..., Turbitt [/bib_ref] suggested that it is possible that n-3 PUFAs induced an increase in IL-2 and IFN-g production in T cells, which may drive a Th1 response, enhance antitumor immunity, and contribute to the cancer prevention effect of n-3 PUFAs. Thus, n-3 PUFA supplementation may enhance Th1 cytokine response and may differentially alter the effector function of T cells. Anita [bib_ref] Omega-3 fatty acid is a potential preventive agent for recurrent colon cancer, Vasudevan [/bib_ref] suggested that EPA alone or in combination with 5-FU + Oxaliplatin (FuOx) could be an effective preventive strategy for recurring sporadic colorectal cancer. Cancer stem/stemlike cells (CSCs/CSLCs) are self-renewing undifferentiated cells and are thought to be one of the leading causes of cancer recurrence. EPA acts synergistically with chemotherapy to markedly inhibit the growth of chemo-resistant colon cancer cells which form the bulk of the recurrent tumor. These findings are in accordance with previous evidence that EPA and DHA reduce inflammation in humans and may have anti-neoplastic properties. Kim [bib_ref] Dietary curcumin and limonin suppress CD4+ T-cell proliferation and interleukin-2 production in..., Kim [/bib_ref] confirmed that CD4 + T-cell proliferation Change in IL-6 between n-3 PUFAs and isocaloric nutrition: random-effects model. c Change in TNF-a between n-3 PUFAs and isocaloric nutrition: randomeffects model. d Change in ALB between n-3 PUFAs and isocaloric nutrition: random-effects model was stimulated by a fish oil diet. The level of CD4 + T-cells was higher in the n-3 PUFAs group than in the conventional nutritional support group, indicating that n-3 PUFAs enhanced host immune function. On the other hand, Marano [bib_ref] Clinical and immunological impact of early postoperative enteral immunonutrition after total gastrectomy..., Marano [/bib_ref] suggested that the intake of n-3 PUFAs improved the immune response by increasing peripheral total lymphocytes, including T lymphocytes, and CD4 + Tcells, while several other studies [bib_ref] Comparison between the AA/EPA ratio in depressed and non depressed elderly females:..., Rizzo [/bib_ref] [bib_ref] Leukocyte numbers and function in subjects eating n-3 enriched foods: selective depression..., Mukaro [/bib_ref] [bib_ref] Nutrition intervention using an eicosapentaenoic acid (EPA)-containing supplement in patients with advanced..., Read [/bib_ref] suggested negative or inverse results. Different subsets of mature T cells carry out the functions of cell-mediated immunity, including killing virally infected cells and tumor cells (CD8 + T cells) and providing help for and regulating components of the immune system (CD4 + T cells). Our meta-analysis showed that n-3 PUFAs effectively increased the level of CD3 + T cells, CD4 + T cells and CD4 + /CD8 + T cells in patients undergoing surgery for GI cancer, but could decrease the level of CD8 + T cells, indicating that the immune response was enhanced and rehabilitation was promoted after surgery. Thus modulation of immune responses and reduction of inflammatory responses together lessens postoperative hospital stay for GI cancer patients. And postoperative n-3 PUFAs nutrition for GI cancer is a challenge and need further research. # Conclusions Our study has important limitations. The intake of n-3 PUFAs varies considerably within countries, and this may explain the heterogeneity across studies. The outcome estimates were taken from published data; therefore, systematic biases could not be minimized and the data in some cases were incomplete. However, we confirmed that the addition of n-3 fatty acids improved immune function and reduced the level of inflammation in GI cancer patients postoperatively. Thus, despite these limitations and although further larger trials are needed, these fatty acids should be widely used in the clinic. Abbreviations ALB: Albumin; CRP: C-reaction protein; GI: Gastrointestinal; IL-2: Interleukin-2; IL-6: Interleukin-6; n-3 PUFAs: Omega-3 polyunsaturated fatty acids; TNFα: Tumor necrosis factor-α Meta-analysis of immune indices. a Pooled results of CD3 + Tcells between n-3 PUFAs and isocaloric nutrition: fixed-effects model. b Change in CD4 + T cells between n-3 PUFAs and isocaloric nutrition: random-effects model. c Change in CD8 + T cells between n-3 PUFAs and isocaloric nutrition: random-effects model. d Change in CD4 + /CD8 + T cells between n-3 PUFAs and isocaloric nutrition: fixed-effects model [fig] Figure 1: Flowchart of computerized search and the eligible studies included in this systematic review and meta-analysis [/fig] [fig] Figure 2: Assessment of risk of bias based on the evaluation domains listed in the Cochrane Collaboration Risk of Bias Tool: risk of bias graph (a), risk of bias summary (b) [/fig] [fig] Figure 3: Meta-analysis of inflammation level. a Change in CRP between n-3 PUFAs and isocaloric nutrition: random-effects model. b [/fig] [table] Table 1: Characteristics of included randomized trials [/table]

Genetic diversities of cytochrome B in Xinjiang Uyghur unveiled its origin and migration history Background: Uyghurs are one of the many populations of Central Eurasia that is considered to be genetically related to Eastern and Western Eurasian populations. However, there are some different opinions on the relative importance of the degree of Eastern and Western Eurasian genetic influence. In addition, the genetic diversity of the Uyghur in different geographic locations has not been clearly studied. Results: In this study, we are the first to report on the DNA polymorphism of cytochrome B in the Uyghur population located in Xinjiang in northwest China. We observed a total of 102 mutant sites in the 240 samples that were studied. The average number of mutated nucleotides in the samples was 5.126. A total of 93 different haplotypes were observed. The gene diversity and discrimination power were 0.9480 and 0.9440, respectively. There were founder and bottleneck haplotypes observed in Xinjiang Uyghurs. Xinjiang Uyghurs are more genetically related to Chinese population in genetics than to Caucasians. Moreover, there was genetic diversity between Uyghurs from the southern and northern regions. There was significance in genetic distance between the southern Xinjiang Uyghurs and Chinese population, but not between the northern Xinjiang Uyghurs and Chinese. The European vs. East Asian contribution to the ten regional Uyghur groups varies among the groups and the European contribution to the Uyghur increases from north to south geographically.Conclusion: This study is the first report on DNA polymorphisms of cytochrome B in the Uyghur population. The study also further confirms that there are significant genetic differences among the Uyghurs in different geographical locations. # Background Uyghur is one of the populations living in central Asian where people have undergone unceasing migration and interacted with other populations in prehistoric and historic times. As a consequence, Uyghurs share anthropometric and genetic traits with Eastern and Western Eurasian. They are genetically considered to be an admixture of Eastern and Western Eurasian populations, demonstrated by archaeological, anthropologic and genetic studies [bib_ref] Trading genes along the silk road: mtDNA sequences and the origin of..., Comas [/bib_ref] [bib_ref] Admixture, migrations, and dispersals in Central Asia: evidence from maternal DNA lineages, Comas [/bib_ref] [bib_ref] Different matrilineal contributions to genetic structure of ethnic groups in the silk..., Yao [/bib_ref] [bib_ref] A genome-wide analysis of admixture in Uyghurs and a high-density admixture map..., Xu [/bib_ref] [bib_ref] Analysis of genomic admixture in Uyghur and its implication in mapping strategy, Xu [/bib_ref] [bib_ref] Haplotype-sharing analysis showing Uyghurs are unlikely genetic donors, Xu [/bib_ref]. The Uyghur population has been used as a good human population sample for studying the human migration and gene inflow/drafting [bib_ref] Different matrilineal contributions to genetic structure of ethnic groups in the silk..., Yao [/bib_ref] [bib_ref] Genome-wide association scan for stature in Chinese: evidence for ethnic specific loci, Lei [/bib_ref]. However, the origin of the Uyghur is under discussion. Modern Uyghurs live primarily in the Xinjiang Uyghur Autonomous Region in China. The majority of the Xinjiang Uyghur live in the south of Xinjiang (south of Tian Shan), with minor populations spreading to the north. Due to environmental, war and political reasons, Uyghurs have undergone unceasing migration (from north to southwest) and interacted with other populations in prehistoric and historic times. This might result in genetically differential characteristics of modern Uyghurs living in different regions. However, most genetic studies of the Uyghur have focused on genetic difference between them and other ethnic groups, and utilized relatively small numbers of sample from very limited geographic locations. The differences in genetic characters between different regional Uyghur populations remain largely unclear. Furthermore, previous studies mainly investigated and compared the genetic diversity of mtDNA control region sequences, while DNA polymorphism of mitochondrial Cytochrome B (CYTB) in the Uyghur population in Xinjiang has not been reported. CYTB is located in the coding region of mtDNA (positions 14576-16047), spanning the 1140 bp fragment. CYTB has undergone changes during evolution resulting in the occurrence of multiple single nucleotide polymorphic sites [bib_ref] Polymorphic variants in the human mitochondrial cytochrome b gene, Andreu [/bib_ref] [bib_ref] Polymorphism in the mitochondrial cytochrome B gene in Koreans. An additional marker..., Lee [/bib_ref]. Compared with the control region of mtDNA, CYTB is relatively conservative and has fewer mutations [bib_ref] Polymorphic variants in the human mitochondrial cytochrome b gene, Andreu [/bib_ref] [bib_ref] Mitochondrial cytochrome b: evolution and structure of the protein, Esposti [/bib_ref]. CYTB is widely used in taxonomic research to determine phylogenetic relationships between organisms due to its sequence variability [bib_ref] Multiplex amplification of mitochondrial DNA for human and species identification in forensic..., Bataille [/bib_ref] [bib_ref] Identification of gourmet meat using FINS (forensically informative nucleotide sequencing), Forrest [/bib_ref] [bib_ref] Species identification by means of the cytochrome b gene, Parson [/bib_ref]. It is also considered to be most useful in determining relationships within families and genera. CYTB sequence information drawn from populations reflects their matrilineal ethnohistory and can deduce DNA sample's population based on populationspecific nucleotide mutations existing in the CYTB gene [bib_ref] Polymorphism in the mitochondrial cytochrome B gene in Koreans. An additional marker..., Lee [/bib_ref] [bib_ref] Study of the cytochrome b gene sequence in populations of Taiwan, Hwa [/bib_ref]. Therefore, we hypothesized that the genetic diversity of Uyghur living in different regions of Xinjiang can be identified by the analysis of CYTB sequence. In this study, we for the first time report on the gene diversity of CYTB in the Uyghur of Xinjiang by completely sequencing their CYTB gene. We demonstrate that there was a genetic variance of CYTB occurring in the different regional Uyghurs and CYTB is a very useful genetic marker for the study of genetic differentiation of Uyghurs in Xinjiang. In particular, we showed the founder/ bottleneck event in the Xinjiang Uyghur. # Results High gene diversity of cytochrome B in the Uyghur population 1190 bp PCR fragments were amplified in all studied samples and 1140 bp of DNA fragment encoding CYTB (positions 14747-15886) were sequenced. Compared with the standard reference sequence (rCRS), a total of 102 mutant sites were observed in the studied 240 samples (Additional file 1: [fig_ref] Table 1: The frequency of population-specific polymorphic mutant nucleotides [/fig_ref]. The number of mutation sites in the samples varied from 0 to 10 with an average of 5.126. There were no deletions and/or insertions observed. [fig_ref] Table 1: The frequency of population-specific polymorphic mutant nucleotides [/fig_ref] lists the most common and group-unique mutation positions. Position 15326 A -G transition occurred in almost all samples except 2 samples (one from Sanji and another one from Bortala). The second highest mutation position was nt 14766 C-T with an average frequency of 82.92% observed in 240 studied Uyghur individuals. More than 90% of Uyghur from Korla, Kumul, Sanji and Turpan had a 14766 C-T mutation, while this mutation occurred less than 80% in the individuals from Aksu (77.14%), Gulja (73. 68%), Hotan (78.95%), and Kashgar (79.10%). The third dominant mutation positions were 14783 T-C (36.25%), 15043 G-A (36.25%) and 15301 G-A (35.42%). Position 14783 T/C, 15043G/A, and 15301 G/A mutations simultaneously took place with an average frequency of 34.17% in the overall Uyghurs. The frequency of the triple 14783 T/C-15043G/ A-15301G/A mutant varied from population to population, from region to region . In general, the frequency of the triple 14783 T/C-15043G/A-15301G/A mutant among the ten regional subpopulations decreased in the order of Kumul, Sanji, Turpan, Bortala, Atush, Kashgar, Aksu and Hotan, in line with their geographic location from north to south. The lowest frequency of this triple mutant (15.79%) was observed in the group from Hotan which is located in the south end of Xinjiang, adjacent to Tibet, India and Pakistan. The north end regions, geographically closer to Mongolia, had similar frequencies of the triple 14783 T/C-15043G/ A-15301G/A mutant to the Mongolian people(our unpublished data). Other populations' data are cited from [bib_ref] Study of the cytochrome b gene sequence in populations of Taiwan, Hwa [/bib_ref]. Studying mutations among the regions, we observed the presence of regional group specific mutations in the CYTB sequence [fig_ref] Table 1: The frequency of population-specific polymorphic mutant nucleotides [/fig_ref]. The mutations 14927 A/G and 15440 T/C were only observed in the Bortala with a frequency of 15.38%, 15746 A/G only in Kumul (30.00%), and 14857 T/C only in Turpan (11.11%). Only three samples of the total 240 Uyghurs showed a nucleotide change at position 14798 which occurs more frequently in Caucasians with a frequency of 0.2083, but not in other Asian groups [bib_ref] Study of the cytochrome b gene sequence in populations of Taiwan, Hwa [/bib_ref]. The genetic diversities in the regional population were calculated [fig_ref] Table 2: The gene diversity, nucleotide diversity, and discrimination power noted in these populations [/fig_ref]. The gene diversities varied among groups from 0.8800 in Korla to 0.97449 in Bortala, while nucleotide diversities varied from 0.00269 (in Sanji) to 0.007863 (in Korla), respectively. The discrimination power (DP) was higher than 0.8200, even up to 0.9454 (in Kashgar). The overall genetic diversity and power of discrimination were 0.9480 and 0.9440, respectively, suggesting that MYTC B is one of the highly polymorphic markers useful for maternal lineage identification. ## Founder / bottleneck haplotypes of cytb in the uyghur There were 93 different haplotypes observed in the 240 Uyghur samples (Additional file 2: [fig_ref] Table 2: The gene diversity, nucleotide diversity, and discrimination power noted in these populations [/fig_ref]. Due to absence of a standard for assignment of haplotypes, we called each different sequence as an individual haplotype generated by DnaSP5.0. The frequencies of different haplotypes varied between different regions (Additional file 2: The frequency of triple 14783 T/C-15043G/A-15301G/A mutant. DNA was isolated from 240 samples collected from ten different locations of Xinjiang and sequenced. DNA sequences were compared with the standard reference sequence using MEGA5.0, as described in the methods. The frequency of nucleotide mutation was calculated. ## Different regional uyghur represent different genetic features In order to investigate the genetic diversity of CYTB in ten regional Xinjiang Uyghurs, MDS analysis was first conducted within ten regional Xinjiang Uyghurs and groups were made based on their geography. As shown in [fig_ref] Figure 3: Multiple dimension scale plot by SPSS [/fig_ref] , ten regions were separated from each other by dimension 1 and some subpopulations were gathered closer than others. Kumul, Turpan, Sanji, Atush and Korla were located closely at the left side in the MDS plot forming cluster 1, while Kashgar, Aksu, Hotan and Gulja sit on the right side and formed cluster 2. Bortala was almost in the middle in the MDS plot. However, in dimension 2, Bortala was closely distributed to Kumul (a northern Xinjiang place), but still be separated from cluster 2 consisting of most of southern Xinjiang regions. Tianshan and the ancient Silk road are a geographic dividing line separating Xinjiang into two parts: Southern Xinjiang (Nanjiang) and Northern Xinjiang (Beijiang). Nanjiang consists of Kashgar, Aksu, Hotan, Atush and Korla, while Kumul, Turpan, Sanji, Gulja and Bortala belong to Beijiang. Accordingly we divided ten regional Xinjiang Uyghurs into two subgroups: Southern Uyghur (groups living in Nanjiang) and Northern Uyghur groups (living in Beijiang). We then performed MDS analysis with the Southern and Northern Uyghur and other populations [bib_ref] Study of the cytochrome b gene sequence in populations of Taiwan, Hwa [/bib_ref] [bib_ref] Mitochondrial DNA CA dinucleotide repeats in Koreans: the presence of length heteroplasmy, Chung [/bib_ref] [bib_ref] The polymorphism of CA repeat in Chinese, Zheng [/bib_ref] [bib_ref] Allele frequencies of mitochondrial D-loop (CA)n repeat polymorphism in six Chinese ethnic..., Yan [/bib_ref] in order to reveal the genetic relationships among these populations. [fig_ref] Figure 3: Multiple dimension scale plot by SPSS [/fig_ref] shows that both Northern and Southern Uyghurs and other 4 Asian populations (Thai, Vietnamese, Filipino and Chinese) were distributed closely to the left in MDS plot and formed a cluster, while Caucasians were isolated far away from them and the cluster, suggesting that both Southern and Northern Uyghur are closer to Asian in genetics than to Caucasian. In addition, the Northern Uyghur was separated from the Southern Uyghur. In comparison with the Southern Uyghur, the Northern Uyghur were more closely located to the Chinese. Furthermore, we noticed that Gulja Uyghurs who geographically reside in northern Xinjiang appeared closely related to the southern subgroups (Hotan, Kashgar and Aksu) in the MDS plot, while geographic southern Uyghur subgroups Korla and Atush were close to the northern subgroups (Kumul, Sanji and Turpan) as shown in [fig_ref] Figure 3: Multiple dimension scale plot by SPSS [/fig_ref]. Therefore we regrouped Uyghur samples by subtracting Gulja from northern group, Korla and Atush from the southern group, and conducted the third MDS analysis using new designed groups [fig_ref] Figure 3: Multiple dimension scale plot by SPSS [/fig_ref]. This regrouped MDS plot displayed that the northern group (Turpan, Kumul, Sanji and Bortala) was separated further away from the southern group (Hotan, Aksu and Kashgar). Korla was distributed between Chinese and the new northern group with similarly close distance to both of them. The Gulja and the new southern group formed a cluster, clearly separated from the northern group and Caucasian. In contrast, Atush was distributed relatively more closely to northern Uyghur. ## The patterns of haplotypes of cytb reflected migration history and origin of uyghur To further explore the genetic linkage, the genetic distances between each of the regional groups were determined [fig_ref] Table 3: Genetic distances between ethnic groups [/fig_ref]. Genetic distances between regions shown in [fig_ref] Table 3: Genetic distances between ethnic groups [/fig_ref] nicely reflected their geographic distance. There were significant differences in genetic distance between south region Hotan and north regions Kumul and Turpan. Interestingly, there were significant differences between the south region Korla and Hotan and between Korla and Kashgar. An unrooted NJ tree was constructed based on the ten regional groups' F ST distances as seen in [fig_ref] Table 3: Genetic distances between ethnic groups [/fig_ref] [fig_ref] Figure 4: Unrooted Neighbor-joining tree [/fig_ref]. The tree showed that there were three branches: Hotan and Gulja formed one branch; Turpan, Sanji, Kumul and Korla formed the second branch, Aksu, Kashgar, Bortala and Atush formed the third Branch. The tree also showed that there were genetic diversities within branches with the exception of the first cluster. With the same strategies in MDS analysis, we regrouped samples and calculated genetic distance [fig_ref] Table 4: Genetic distances between ethnic groups [/fig_ref] and constructed phylogenetic trees [fig_ref] Figure 4: Unrooted Neighbor-joining tree [/fig_ref] and 4C) using new assigned groups. [fig_ref] Table 5: Fst based on Tajia Nei [/fig_ref] showed the results of genetic distance generated by regrouping the Uyghur into two major groups--Southern and Northern Uyghur. [fig_ref] Figure 4: Unrooted Neighbor-joining tree [/fig_ref] and 4C are NJ trees constructed with the genetic distances from Tables 4 and 5. We found that both the southern and northern groups showed shorter genetic distance from Chinese than from Caucasian. There were significance in genetic distance between the southern Uyghur and Caucasian, between the southern Uyghur and Chinese, and between the northern Uyghur and Caucasian, while there was no significance between the northern Uyghur and Chinese. The NJ tree [fig_ref] Figure 4: Unrooted Neighbor-joining tree [/fig_ref] intuitively showed the genetic relationship between these populations, supporting [fig_ref] Table 4: Genetic distances between ethnic groups [/fig_ref] results. [fig_ref] Table 5: Fst based on Tajia Nei [/fig_ref] further confirmed the genetic relationship between the southern, northern, Chinese and Caucasian. More importantly, [fig_ref] Table 5: Fst based on Tajia Nei [/fig_ref] and [fig_ref] Figure 4: Unrooted Neighbor-joining tree [/fig_ref] demonstrated that geographic southern cities Atush and Korla showed gene similarity to the northern Uyghur, although northern city Gulja is genetically similar to the Southern Uyghur, but not to the northern Uyghur. The Korla Uyghur was completely different to the southern Uyghur (Hotan, Aksu and Kashgar). # Discussion In this study we first reported the genetic polymorphisms of cytochrome B in Uyghur population of Xinjiang. We demonstrated that cytochrome B is a very useful and informative marker for matrilineal identification and population studies, second to the mtDNA control region. There are population specific and regional population specific nucleotide positions in the cytochrome B sequence, which can help identify DNA sample's region and population [fig_ref] Table 4: Genetic distances between ethnic groups [/fig_ref] (B) and [fig_ref] Table 5: Fst based on Tajia Nei [/fig_ref] (C) an unrooted Neighbor-joint trees were constructed using Phylips 3.6.9 software. Other populations were cited from [bib_ref] Study of the cytochrome b gene sequence in populations of Taiwan, Hwa [/bib_ref]. information. The results of the present study conclude that: 1) there are founder/ bottle neck events in Xinjiang Uyghur; 2) the Uyghur is closer to Chinese, rather than to Caucasian in genetic distance; 3) There are genetic difference between the Southern and Northern Uyghur; 4) Gene influence from Asian (Chinese) is stronger on the Northern Uyghur, than the Southern Uyghur; and 5) CYTB is a good genetic marker for differentiation of subgroups. Xinjiang is located in Central Asia which is an intermediate region of the Eurasian continent. There were 36 nations existing in Xinjiang during prehistorical and historical times. Some nations were defeated by others and destroyed, overrun or forced to migrate to other places. Unceasing wars propelled endless migration and coalition, leading to genetic mixture and gene drafting. Additionally, there were two silk roads extending from east to west through Xinjiang, resulting in extensive exchange of culture and marriage between different populations. Our data confirm that Uyghur is an admixture population with contributions from both Eastern and Western Eurasian ancestries, which is consistent with the results from mtDNA control region sequences [bib_ref] Different matrilineal contributions to genetic structure of ethnic groups in the silk..., Yao [/bib_ref] [bib_ref] A genome-wide analysis of admixture in Uyghurs and a high-density admixture map..., Xu [/bib_ref] [bib_ref] Analysis of genomic admixture in Uyghur and its implication in mapping strategy, Xu [/bib_ref] [bib_ref] Haplotype-sharing analysis showing Uyghurs are unlikely genetic donors, Xu [/bib_ref]. Based on the mutation at position 14798, the CYTB sequences of the Uyghur in Xinjiang contain more inflow of East Asian than European. The proportion of European sequences varies from different geographic regions of the Uyghur (data not shown). The Uyghurs are a Turkic ethnic group living in Eastern and Central Asia. The ancestors of the Uyghur tribe were Turkic pastoralists called Tiele in Northern China, Mongolia, and the Altay Mountains. Due to wars and environmental stress, Uyghur continued migrating from north to south. From the historic perspective, Uyghur originated from Mongolians, rather than Caucasians and inherited Mongolia genetics. The Uyghur population was gradually diluted as they migrated from north to south. Our data on the frequency of triple 14783 T/C-15043G/A-15301G/A mutant and founder/bottle neck haplotypes demonstrated that the Uyghur originated from Mongolia, migrated from north to south, supporting historic reports. In this study, we found that nucleotide positions with most frequent mutation were 14766, 14783, 15043, and 15302, which were most commonly seen in Asians including the Chinese Han population [bib_ref] Study of the cytochrome b gene sequence in populations of Taiwan, Hwa [/bib_ref]. Mutation at position 14766 occurs in more than 93% of Asian samples, but less than 0.72% of Caucasians [bib_ref] Polymorphic variants in the human mitochondrial cytochrome b gene, Andreu [/bib_ref] [bib_ref] Study of the cytochrome b gene sequence in populations of Taiwan, Hwa [/bib_ref]. The frequencies of mutation at the position 14766 in the ten different geographic and 15302 showed the same trend among the ten Uyghur groups as position 14766. Mutations at the above three positions were seldom observed in the Caucasian population. In contrast, mutation at position 14798 that is Caucasian specific was observed in only one Aksu sample and two Hotan samples. Taken together, we found that at the cytochrome B gene, Uyghurs generally have more imprint of East Asian genetic portion than that of Caucasian. These genetic results are in agreement with the history of Uyghur formation. The Uyghurs living in the Nanjiang are different to Beijiang, and the southern Uyghur have relatively less influence from East Asians than the northern. Our data showed that in the same ethnic group, people residing at different places have genetic differences. In this study, we reported that Uyghur from different geographic locations in Xinjiang have differences in the percentage of European/East Asian ancestry component, distinctly by the difference between Aksu/Hotan and Kumul or Korla. There is a significant difference between the south (Hotan, Kashgar, and Aksu ) and the north of Xinjiang (Kumul, Bortala, Turpan and Sanji). East Asian ancestry dominates the Uyghur from the north, while European ancestry imprints more on the southern Uyghur. In addition, we observed that Uyghurs living in two geographically south cities Atush and Korla shared northern Uyghur genetics (Kumul, Turpan and Sanji), in contrast with the southern Uyghur, while Uyghurs from Gulja which is located in the north of Tianshan in geography presented the southern regional Uyghur genetic features. Demographic records show that the dominant ethnic group is Kyrgyz who defeated the Uyghurs in AD 840, despite the fact that Atush is located in the south of Tianshan [Wikipedia, the free encyclopedia, History of the Kyrgyzstan]. Moreover, there were intermarriages in the Uyghur living in the Atush region. This might explain why the gene pattern of cytochrome B in the Atush Uyghurs was between that of the southern and northern Uyghur group, different from the other three southern group samples. As for Gulja, there were two large migrations of the southern Uyghur from south to north Gulja in history. Over time, the migrated southern Uyghurs expanded in Gulja and became the largest minority in Gulja [Wikipedia, the free encyclopedia, History of the Uyghur people.]. As a result, the Gulja Uyghur was and still is a branch of southern Uyghur, having little relationship with the northern Uyghur, which is in line with our data that the Uyghur from Gulja shares similar genetic characters with the Southern Uyghur. Although Korla is located in the south of Tianshan belonging to Nanjiang ( southern of Xinjiang), it is also adjacent to Turpan and Sanji. Korla is located in Bayingolin Mongol Autonomous Prefecture where the current major population is Chinese Han. As early as 94 CE, the Chinese government and military started to administrate this area and forced different populations to exchange and mix. The results from CYTB genotyping are consistent with that from the control region of mtDNA (unpublished data). Our data showed that the Uyghurs from Korla in genetics are different to the south Xinjiang Uyghurs. # Conclusions Cytochrome B is a very useful DNA marker with high discriminate power for matrilineal identification, as well as deduction of the region and population of people. The polymorphisms of CYTB were significantly different between different geographical Uyghur (between south and north). The influences of East Asian and European genetics in Uyghur varies between different geographic locations (particularly south and north) of Uyghur. # Methods ## Sampling A total 240 healthy unrelated samples were collected from ten different Uyghur concentrated locations/communities (10), Sanji [bib_ref] Mitochondrial cytochrome b: evolution and structure of the protein, Esposti [/bib_ref] , and Turpan [bib_ref] Allele frequencies of mitochondrial D-loop (CA)n repeat polymorphism in six Chinese ethnic..., Yan [/bib_ref]. We collected more samples from Kashgar due to more than 36% of Xinjiang Uyghurs live in that area. For the samples it was confirmed that there were no intermarriages between Uyghur and other ethnic groups and no migration history in the latest three generations. The study was approved by the Ethnic Study Committee of Xinjiang University. All blood samples were obtained with informed consent. DNA was extracted from blood using an EasyPure blood genomic DNA extraction kit (TransGen Biotech, Beijing, China). ## Pcr and sequencing The primers used for PCR amplification of the CYTB gene listed in [fig_ref] Table 6: Sequences of primers used for mitochondrial DNA amplification and sequencing L [/fig_ref] were adopted from previous studies [bib_ref] Study of the cytochrome b gene sequence in populations of Taiwan, Hwa [/bib_ref] and synthesized by Shenggong (Shenggong Biotech, Shanghai, China PCR products were visualized on a 1.5% agarose gel and purified using a Biomiga kit (Biomiga, Beijing, China). Sequencing reactions were conducted in a PE-9600 thermocycler (ABI Applied Biosystems, Fostor City, CA, USA), using a BigDye terminator v3.1 Cycle sequencing kit (Applied Biosystems) with the following conditions: 25 cycles of 95°C for 30 s; 50°C for 30 s; and 60°C for 4 min. The primers L14724, H15149, L15283 and H15363 were used for sequencing listed in [fig_ref] Table 6: Sequences of primers used for mitochondrial DNA amplification and sequencing L [/fig_ref]. The DNA sequences were detected with an ABI3730 DNA sequencer (Applied Biosystems) in HuaDa Genome Centre, Beijing. [fig] Figure 3: Multiple dimension scale plot by SPSS. (A) MDS plot of ten different regions' Uyghur people. (B) MDS plot of geographic-based southern and northern Uyghur population. (C) MDS plot of regrouped populations. Other populations were cited from[14]. [/fig] [fig] Figure 4: Unrooted Neighbor-joining tree. Using genetic distance inTable 3 (A), [/fig] [fig] Figure 5: Geographic locations of the samples investigated in this study. [/fig] [table] Table 1: The frequency of population-specific polymorphic mutant nucleotides [/table] [table] Table 2: The gene diversity, nucleotide diversity, and discrimination power noted in these populations [/table] [table] Table S2: third of total haplotypes including H11. The network showed that there are founder/ bottleneck haplotypes existing in the Xinjiang Uyghur which are also observed in the Mongolians (our unpublished data). Median-joining network. A Median-joining network was constructed using NETWORK 4.6.1.0 software with haplotypes of CYTB from the Uyghur. samples into different geographic locations; there were 24 haplotypes observed in 35 Aksu individuals, 16 in 24 Atush, 11 in 13 Bortala, 14 in 19 Gulja, 15 in 19 Hotan, 41 in 67 Kashgar, 13 in 25 Korla, 8 in10 Kumul, 7 in 10 Sanji, and 11 in 18 Turpan individuals. Haplotype 15 (15326G) was predominantly found in the Aksu with a frequency of 20% and in the Kashgar with a frequency of 14.93%. Haplotype 13 (14766 T, 15326G) existed most frequently in the Turpan (27.78%), Atush (16.67%) and Korla (16%), but was not in the Bortala nor Kumul. Haplotype 11 (14766 T, 14783C, 15043A, 15301A, 15326G) appeared frequently in the Korla (32%), Sanji (30%), Turpan (16.67%), and Atush (12.5%). Haplotype 83 (14766 T, 14783C, 15043A, 15301A, 15326G, 15746G) was only observed in the Kumul group with a frequency of 30%, but not in other groups thereby showing region-specificity. Haplotype 89 (14766 T, 14783C, 14857C, 15043A, 15301A, 15326G) was observed only in the Turpan with a frequency of 11.11%, while haplotype 59 (14766 T, 14783C, 15043A, 15301A, 15326G, 15673A) and haplotype 66 (14766 T, 15314A, 15326G, 15452A) were only found in the Kashgar group with a frequency of 2.99%, respectively. [/table] [table] Table 3: Genetic distances between ethnic groups [/table] [table] Table 4: Genetic distances between ethnic groups [/table] [table] Table 5: Fst based on Tajia Nei [/table] [table] Table 6: Sequences of primers used for mitochondrial DNA amplification and sequencing L: light strand; H: heavy strand. *: PCR; ¶ : sequencing. [/table]

An investigation of factors identified at birth in relation to anxiety and depression in old age: the Hordaland Health Study (HUSK) Background: Although life course influences have long been recognised in affective disorder, little is known about the influence of early life factors on late life anxiety and depression. The aim was to investigate the extent to which birth measures, maternal health and family circumstances were associated with symptoms of anxiety and depression in late life. Methods: A retrospective cohort study was constructed from a cross-sectional survey sample of community residents aged 72-74 years, 406 of whom had traceable birth records. Cases and controls for late life anxiety and depression were defined applying standard cut-offs to the Hospital Anxiety and Depression Scale. A range of measures and circumstances were extracted from birth records blind to survey data and compared in age-and gender-adjusted models. Results: There were no differences in any anthropometric measure in either case control comparison. Case-level anxiety and depression were both associated with significantly lower maternal age. Late-life anxiety was additionally associated with smaller maternal pelvic size and the mother's condition being rated as poor at birth/discharge. Late-life depression was associated with a lower status paternal occupation. Conclusions: There was no evidence for a substantial influence of early life size on late life affective disorder. However, there was some evidence in secondary analyses for an enduring influence of the family's socioeconomic environment and maternal health. # Background Anxiety and depression are highly prevalent disorders among older adults [bib_ref] Prevalence and correlates of generalized anxiety disorder among older adults in the..., Gonçalves [/bib_ref] [bib_ref] Risk factors for anxiety and depression in the elderly: a review, Vink [/bib_ref] and, with increasing life expectancy, an area of growing clinical and public health significance [bib_ref] The epidemiology of common late-life mental disorders in the community: themes for..., Gallo [/bib_ref]. A recent review of risk factors for anxiety and depression in old age highlighted personality traits, inadequate coping strategies, previous psychopathology, qualitative aspects of social networks, stressful life events and female gender as important risk factors [bib_ref] Risk factors for anxiety and depression in the elderly: a review, Vink [/bib_ref]. Despite considerable overlap in risk profiles for anxiety and depression, some risk factors appear more relevant for depression specifically, including chronic diseases, poor self-perceived health, functional disability, shrinking social networks and being unmarried [bib_ref] Risk factors for anxiety and depression in the elderly: a review, Vink [/bib_ref]. Facing this combination of proximal and distal risk factors, a life-course perspective including risk factors originating from the early-life environment [bib_ref] Factors associated with symptoms of anxiety and depression in five cohorts of..., Gale [/bib_ref] , is potentially indicated. The developmental origins of adult health and disease (DOHAD), originally known as "the Barker hypothesis", is commonly applied as the framework for understanding empirically observed links between processes at prenatal stages and adult chronic disease [bib_ref] Fetal and infant origins of adult disease, Barker [/bib_ref] [bib_ref] The fetal origins of adult disease: a narrative review of the epidemiological..., Skogen [/bib_ref]. The DOHAD-hypothesis suggests that prenatal risk factors, from intrauterine environmental exposures, affect the foetus irreversibly during specific developmental periods. Thus, according to the hypothesis, the risk of specific diseases such as coronary heart disease and non-insulin dependent diabetes in adult life, will be elevated across the life-span for those who are exposed to the relevant risks in foetal life [bib_ref] The developmental origins of adult disease, Barker [/bib_ref]. A key practical challenge in research on the hypothesis is to attain information about the offspring's development during pregnancy. Anthropometric measures such as birth weight, birth length and head circumference are routinely, and often reliably recorded, and therefore often used as proxy measures [bib_ref] On the importance -and the unimportance -of birthweight, Wilcox [/bib_ref]. The DOHAD-hypothesis has also been applied to investigate the aetiology of mental disorders, proposing that prenatal factors can modify risk [bib_ref] Is there a fetal origin of depression? evidence from the mater university..., Alati [/bib_ref]. Low birth weight has been linked to increased risk of suicide [bib_ref] Low weight gain in infancy and suicide in adult life, Barker [/bib_ref] [bib_ref] Fetal and childhood growth and the risk of violent and non-violent suicide..., Mittendorfer-Rutz [/bib_ref] , schizophrenia [bib_ref] Obsteric complications and schizophrenia: historical and meta-analytic review, Cannon [/bib_ref] , and autism [bib_ref] Prenatal and perinatal risk factors for autism: a review and integration of..., Kolevzon [/bib_ref] , as well as with anxiety and depression [bib_ref] Factors associated with symptoms of anxiety and depression in five cohorts of..., Gale [/bib_ref] [bib_ref] Is there a fetal origin of depression? evidence from the mater university..., Alati [/bib_ref] [bib_ref] Birth weight and the risk of depressive disorder in late life, Thompson [/bib_ref] [bib_ref] Increased susceptibility to stress at a psychological assessment of stress tolerance is..., Nilsson [/bib_ref] [bib_ref] Early origins and adult correlates of psychosomatic distress, Cheung [/bib_ref] [bib_ref] Association between psychological symptoms in adults and growth in early life: longitudinal..., Cheung [/bib_ref] [bib_ref] Birth weight and later risk of depression in a national birth cohort, Gale [/bib_ref] [bib_ref] Prediction from low birth weight to female adolescent depression: a test of..., Costello [/bib_ref] [bib_ref] Birth weight and psychological distress at age 45-51 years, Wiles [/bib_ref] [bib_ref] Andersson S: Depression in young adults with very low birth weight: the..., Raikkonen [/bib_ref] [bib_ref] Depression and its association with diabetes, cardiovascular disease, and birth weight, Paile-Hyvärinen [/bib_ref] [bib_ref] Low birth weight and risk of affective disorders and selected medical illness..., Nomura [/bib_ref] [bib_ref] Increased depressive symptoms in female but not male adolescents born at low..., Van Lieshout [/bib_ref] [bib_ref] Depression in swedish women: relationship to factors at birth, Gudmundsson [/bib_ref] [bib_ref] The association between birth weight and anxiety disorders in young adults, Betts [/bib_ref] [bib_ref] Is birth weight associated with risk of depressive symptoms in young women?..., Inskip [/bib_ref]. In 2007, Alati and colleagues [bib_ref] Is there a fetal origin of depression? evidence from the mater university..., Alati [/bib_ref] reported an inverse linear association between birth weight and depressive symptoms at age 21 in females, but not in males. Wiles and colleagues found the same pattern in relation to risk for psychological distress among middle-aged individuals of both genders [bib_ref] Birth weight and psychological distress at age 45-51 years, Wiles [/bib_ref]. To our knowledge, only three studies have investigated the impact of early life factors and late-life mental health. Thompson and colleagues found that foetal undernutrition predisposed men, but not women, to depression at age 68 [bib_ref] Birth weight and the risk of depressive disorder in late life, Thompson [/bib_ref]. A recent Swedish study found that women born at a lower gestational age and those born with a birth weight less than the median, were more likely to have reported lifetime depression with a follow-up of 92 years of age [bib_ref] Depression in swedish women: relationship to factors at birth, Gudmundsson [/bib_ref]. As the latter study investigated the risk of life-time depression, however, the finding may not reflect an enduring risk of gestational age and birth weight for depression into old age. On the other hand, Gale and colleagues found no association between birth weight and anxiety and depression at age 64-79 as reported on the Hospital Anxiety and Depression Scale [bib_ref] The hospital anxiety and depression scale, Zigmond [/bib_ref] in five different cohorts [bib_ref] Factors associated with symptoms of anxiety and depression in five cohorts of..., Gale [/bib_ref]. Other studies in younger populations have also failed to identify associations between foetal or very early life factors and mental disorders in adolescence, early adulthood and mid-life [bib_ref] Birth dimensions and risk of depression in adulthood: cohort study of danish..., Osler [/bib_ref] [bib_ref] Long-term development for girls and boys at age 16-18 as related to..., Lagerström [/bib_ref] [bib_ref] Low birthweight and subsequent emotional and behavioural outcomes in 12-year-old children in..., Sabet [/bib_ref] [bib_ref] Fetal growth restriction and the development of major depression, Vasiliadis [/bib_ref] [bib_ref] Longitudinal study assessing the joint effects of socio-economic status and birth risks..., Fan [/bib_ref] , leaving the question open for debate [bib_ref] Is birth weight associated with risk of depressive symptoms in young women?..., Inskip [/bib_ref]. A recent systematic review investigating the foetal origins of depression included 18 studies and compared those of low birth weight (<2500g) to those of normal birth weight (>2500g) [bib_ref] Foetal origins of depression? A systematic review and meta-analysis of low birth..., Wojcik [/bib_ref]. The authors found evidence for a weak association (random effects estimate of OR 1.15 (CI95% 1.00-1.32)) between low birth weight and later depression or psychological distress, but also indications of publication bias. Furthermore, studies of associations between perinatal status and later mental disorders have been criticized for relying too much on birth weight as the indicator of the foetal environment [bib_ref] Factors associated with symptoms of anxiety and depression in five cohorts of..., Gale [/bib_ref] and this may contribute to the heterogeneous results, in addition to different lengths of follow-up. It has also been argued that associations may be more prominent in females and less identifiable in male-only cohorts [bib_ref] The impact of gestational stress and prenatal growth on emotional problems in..., Rice [/bib_ref]. In summary, there is some evidence suggesting an association between foetal and early origins of anxiety and depression in adulthood. This is however not consistently supported in the literature, with study design differences and elements of publication bias suggested as possible reasons for heterogeneity. Compared to research on depression in old age, anxiety disorders have received very little attention [bib_ref] Anxiety disorders in the elderly: outdated beliefs and a research agenda, Wetherell [/bib_ref] , despite similar prevalences to depression in old age [bib_ref] Anxiety and depression in later life: Co-occurence and communality of risk factors, Beekman [/bib_ref] and high levels of co-morbidity with depression [bib_ref] Risk factors for anxiety and depression in the elderly: a review, Vink [/bib_ref] [bib_ref] Anxiety and depression in later life: Co-occurence and communality of risk factors, Beekman [/bib_ref]. Taking advantage of a birth record archive in an area covered by a cross-sectional survey of older people's mental health in Bergen (Norway), we constructed retrospective cohort study to investigate the relationship between the early life environment and late life anxiety and depression. This allowed for a follow-up from the 1920s to the 1990s. Primary hypotheses were that smaller anthropometric measures would be associated with case status for both syndromes. Secondary analyses investigated other information reflecting parental health and early life socioeconomic status. Cohort contextcharacteristics of Bergen in the early 20th century and life expectancy During the late 19th century and early 20th century, Bergen city expanded geographically and qualitatively from a semi-rural city to one with more modern characteristics. Primary industry which had dominated, gave ground to expanding secondary and tertiary industry. This change in industry was mostly due to growing production and manufacturing, but also due to an increase in commerce, shipping, transport and service sectors. As a consequence of this, three social classes began to dominate in Bergen during the same period, upper (bourgeoisie), middle and lower, with large differences in income, housing standard and diet. The upper class was characterised by industry proprietors, importers, wholesale dealers and financers. The middle class consisted primarily of merchants, craftsmen and officials, while the lower class comprised regular workers or artisans. During 1925 and 1927 the life expectancy at birth in Norway was approximately 67 years for males, and 74 years for females. # Methods ## Study population The sampling frame for this study comprised all 3,341 (77% of the invited) participants of the old age cohort of the population-based Hordaland Health Study (HUSK) which has been previously described [bib_ref] The hordaland homocysteine study: a community-based study of homocysteine, its determinants, and..., Refsum [/bib_ref]. In summary, all residents of Bergen city or neighbouring areas born between 1925 and 1927 of a previously established study population were invited to participate in a general physical examination and to complete questionnaires on socio-demographic status, general health and healthrelated behaviour. HUSK was conducted from 1997 to 1999 as a collaboration between the National Health Screening Service, the University of Bergen and local health services. In this study, we sought to link as many members as possible of the HUSK sample to their birth records in order to construct a nested retrospective cohort study within the linked subgroup, comparing historically recorded information at birth between participants with/ without a mental disorder in late life. In the Norwegian Population Registry, all inhabitants of Norway are registered with a personal identification number. Using this individual identifier, the names (and maiden name for females), date of birth, place of birth and parents' names (if available) of HUSK participants were retrieved. This information was used to trace the participants born in Bergen to the public maternity ward ("Fødestiftelsen i Bergen") birth records located at the Regional State Archives of Bergen. In the second decade of the 20th century, about one quarter of all births in the Bergen area took place in the official maternity ward (personal communication, State archivist). The portion of deliveries taking place at hospitals increased steeply when the new Women's Clinic ("Kvinneklinikken") was inaugurated in 1926, replacing the old maternity ward. The relevant birth records for the present study were those detailing births between 1st of January 1925 and 31st of December 1927. These records contain detailed information about the pregnancy, the birth and the mother's health recorded by midwives and obstetricians during the hospital stay. The Women's Clinic in question was the main teaching facility for midwifes at the time, and the records were integral to the training, and are therefore considered to be of high quality. ## Early life factorsinformation obtained at birth, 1925-27 The available birth records in the Regional State Archives of Bergen were viewed and coded by researchers blind to all measurements made in HUSK. The following information was abstracted from each record (directly copying original information unless stated otherwise): birth weight (kg), birth length (cm), head circumference (cm) at birth, ponderal index (PI; calculated from weight divided by the third power of length), mother's pelvic size (cm), mother's age (years), and gestation (weeks). The following binary variables were derived from individual free text fields: any recorded disease in the mother, family history of coronary heart disease (yes/no) and tuberculosis (yes/no), the state of mother's teeth (poor/good), complications during birth (yes/no), mother's condition after birth (poor/good; based on the midwives description of the mother (e.g. fatigued, pale, swollen and feverish)), mother's condition at discharge (poor/good; based on the midwives description of the mother (e.g. pale, swollen and feverish), and indicators of socioeconomic status included marital status (married/unmarried), father's occupation (unskilled/manual vs skilled manual/ non-manual), and type of payment for the hospital stay (health insurance/not insurance). Anxiety and depressionfollow-up in HUSK at age 72-74 Cases and controls were defined according to the presence of anxiety and depressive symptoms in late life, ascertained in HUSK participants using the Hospital Anxiety and Depression Scale (HADS) [bib_ref] The hospital anxiety and depression scale, Zigmond [/bib_ref]. HADS assesses anxiety (HADS-A) and depression (HADS-D) on two separate subscales, each consisting of seven questions. The HADS has been used extensively in previous community-based research and is considered a reliable and valid screening tool, with little measurement variance as a function of age in the general population [bib_ref] The validity of the hospital anxiety and depression scale. An updated literature..., Bjelland [/bib_ref] [bib_ref] Hospital anxiety and depression (HAD) scale: factor structure, item analyses and internal..., Mykletun [/bib_ref]. As HADS was developed for a hospital setting, no somatic symptoms of anxiety or depression are included, which renders it especially suitable for assessment in an old age population where significant comorbidity with physical health conditions is anticipated [bib_ref] Factor structure of the hospital anxiety and depression scale in older patients..., Flint [/bib_ref]. In accordance with previous studies, the mean score for all individuals with five or more valid responses on each subscale was computed [bib_ref] Hospital anxiety and depression (HAD) scale: factor structure, item analyses and internal..., Mykletun [/bib_ref]. We employed the recommended cut-offs to define caseness (both ≥8), which have been found to yield an optimal balance between sensitivity and specificity (both 0.8) [bib_ref] The validity of the hospital anxiety and depression scale. An updated literature..., Bjelland [/bib_ref]. Applying these cut-offs, two dichotomous variables for case-level anxiety and depression were created, where the controls for anxiety could include case-level depression and vice versa. Our findings are equivalent when employing a control group restricted to only noncases or HADS as a continuous measure (data not shown). Additional demographic informationfollow-up from HUSK at age 72-74 To assess potential demographic differences between the participants we were able to trace and the remaining HUSK participants, we obtained gender, self-reported level of educational attainment and general health from HUSK. Level of educational attainment was divided into "compulsory only" and "post-compulsory", while general health was divided into "poor" and "good". ## Statistical analyses Of the 3,341 HUSK-participants [bib_ref] The hordaland homocysteine study: a community-based study of homocysteine, its determinants, and..., Refsum [/bib_ref] , we were able to trace 480 to their birth records. Out of these, 406 (84.6%) had five or more valid responses on each HADS subscale (comparable to the rest of the HUSKparticipants) allowing late life case/control status to be ascertained and constituted the total study population. The sample was described, HUSK participants with traceable birth records were compard to the remainder of the sample, and the traced with and without valid HADS-subscales were compared in relation to the exposures. For primary analyses, previous exposure status was compared between cases and controls using ageand gender-adjusted linear regression analyses for continuous exposures and similarly adjusted logistic regression models for categorical exposures. In order to assess adverse conditions in the womb, the association between birth weight and case-level anxiety and depression was adjusted for gestational age in addition to age and gender. In total, 18 potential exposures were investigated in relation to each HADS subscale. Although not entirely uncontroversial, considering the explorative nature of the study and the limited sample size, we chose not to adjust for multiple comparisons. For the significant associations, we also performed a secondary analyses adjusting for educational attainment and self-reported general health at participation in HUSK. Stata version 11.0was employed for all analyses presented in this paper. Using the software G*Power version 3.1.3 [bib_ref] GPOWER: a general power analysis program, Faul [/bib_ref] , a post hoc power analysis indicated that we would be able to detect a small to medium effect size for mean group differences (Cohen's d of 0.39) and binary associations (Cohen's w of 0.18), given a power of 80% for the main HADS-A and HADS-D analyses [bib_ref] Understanding power and rules of thumb for determining sample sizes, Van Voorhis [/bib_ref]. ## Ethics The data in HUSK was collected in accordance with ethical standards required by the regional ethical board of Committees for Medical and Health Research Ethics in Norway (REC). The permission to collect and store the data from HUSK was given by the Norwegian Data Inspectorate. All participation in HUSK was voluntary, and all potential participants received written information about the project before they met for examination. The participants gave their written statements of informed consent, including the specific consents to use information from HUSK in health research and to link this information with other relevant data sources. The participants also gave their written statement that they were informed that no specific time-limit was set for the storage of data. This specific study was reviewed and approved by REC. The current study adheres to the standards of the declaration of Helsinki. # Results No differences between the HUSK-participants with birth journal information and participants without were found regarding gender (p=0.814), self-reported health (p=0.706), case-level anxiety (p=0.531) or case-level depression (p=0.621). However, participants we were able to trace had a significantly higher educational attainment (p=0.025) than the untraced. The proportion of participants with case-level anxiety and case-level depression in the traced sample was 14.3% and 9.4%, respectively. No differences were found on exposures obtained at birth between those with valid HADS subscales (n=406) and those without (n=74), all p-values ranging between 0.087-0.990. Out of the 406 HUSK-participants included in the present study, 54.2% (CI95% 49.3-59.0) were female, and the mean age was 72.3 (SD: 0.9) years. The sample-characteristics obtained at birth are summarized in [fig_ref] Table 1: Sample characteristics at birth obtained from medical records [/fig_ref]. In the main analyses, none of the anthropometric measures at birth, such as weight and length, differed significantly between cases and controls in either comparison [fig_ref] Table 2: Associations between late life anxiety and depression case status and continuous or... [/fig_ref] , and out of the 36 bivariate associations investigated, only 6 (16.7% CI95% 3.9%-29.5%) were statistically significant [fig_ref] Table 2: Associations between late life anxiety and depression case status and continuous or... [/fig_ref]. For continuous and ordinal exposures, both case-level anxiety and depression were associated with lower mother's age (p=0.027 and p=0.036, respectively), and case-level anxiety was associated with a smaller maternal pelvic size (p=0.017). For binary exposures, case-level anxiety was associated with an increased odds of the mother's condition after birth or at discharge being rated as poor (p-values 0.001 and 0.012 respectively). Case-level depression was associated with an increased odds of a lower occupational status for the father (p=0.029). There were no substantial differences between the crude model and age-and gender-adjusted model for any of the associations investigated, and modelling the HADS-subscales as continuous outcomes also yielded similar results as did restriction of controls to those with neither anxiety nor depression (data not shown). For the significant results, adjusting for educational attainment and selfreported general health in HUSK only minimally changed the estimates (data not shown). # Discussion ## Main findings Many previous studies have found an association between early life factors and later mental health problems, but most of these have been limited to follow-up into childhood, early adulthood or middle-age. In this study investigating risk factors present at birth in relation to anxiety and depression in old age, no associations were found for the anthropometric measures of principal interest. However, there was evidence in secondary analyses for associations with factors that might indicate worse maternal health and/or socioeconomic status. # Strengths and limitations This study had several strengths. Having access to birth records from the 1920s and the possibility to link this information to a population-based health survey in the late 1990s enabled examination of an interval between exposure/outcome status in the range of 72-74 years. The exposure source contained detailed, accurate and rich anthropometric measurements (such as birth weight in kilograms with two decimal places), as well as information about maternal health and circumstances, the birth process and the early post-natal period. This information is unlikely to be biased in any particular direction. The birth records were carried out as a key element in the education of midwives under the supervision of the head physician. The HUSK study included a validated and often used screening inventory for symptoms of anxiety and depression. Furthermore, both men and women were represented in our sample, enabling identification of any confounding by gender. Limitations include the relatively small sample size and proportion of the HUSK-participants that could be traced back to their birth records. There are several reasons for this: the birth records were only available for a subgroup as not everyone who participated in HUSK was born in the Bergen area, and some born at home or at other hospitals. Based on a conservative estimate, at least one-third of the HUSK sample would not be within the catchment area of the regional hospital at the time of birth. The small sample size reduces our ability to detect potential small but meaningful effects. A recent meta-analysis investigating the association between low birth weight and depression later on found only small effects [bib_ref] Foetal origins of depression? A systematic review and meta-analysis of low birth..., Wojcik [/bib_ref] and given the lack of precision in our estimates such small effects may not be detectable in our study. In previous DOHAD-studies which have used birth cohorts to trace individuals in their adulthood (i.e. forward tracing), the identified proportions of the original sample are sometimes as low as 2-5% [bib_ref] Review of the evidence on fetal and early childhood antecedents of adult..., Joseph [/bib_ref]. The low traceability and small sample size constitute central limitations to our study, and warrants caution with regards to the interpretation and generalisability of the present study. For birtweight in particular, the small sample size precluded any investigation of low birth weight (<2.5 kg) or high birthweight (>4.5 kg) per se, which is associated with later psychopathology in some studies [bib_ref] A longitudinal typology of symptoms of depression and anxiety over the life..., Colman [/bib_ref] [bib_ref] Fetal growth and the lifetime risk of generalized anxiety disorder, Vasiliadis [/bib_ref]. Due to long follow-up between the exposures and the outcomes, survival and selection effects are likely to influence the nature of the sample overall, including a healthy survivor effect [bib_ref] Characteristics of the healthy survivor effect among male and female hanford workers, Baillargeon [/bib_ref] , and nonparticipation bias [bib_ref] The health status of nonparticipants in a population-based health study: the hordaland..., Knudsen [/bib_ref] rendering the participants in our sample healthier than those from other studies in younger age groups. It is, however, unlikely that such effects would substantially influence the exposure-outcome associations of interest since both cases and controls were survivors and drawn from the same source population. Also we did not find any systematic differences in gender, self- reported health or HADS-scores between the participants we were able to trace, and the remainder of the HUSKparticipants, suggesting reasonable generalisability beyond the analysed sample. The outcome measure, HADS, is a screening instrument for anxiety and depression, primarily designed to measure these syndromes in a general hospital population. In the age group in this study, a measurement which specifically focuses on cognitive rather than somatic symptoms of anxiety and depression may have benefits in avoiding over-ascertainment due to physical health problems, which are relatively common in this age group. The HADS has been found to have good case-finding properties in a GP setting [bib_ref] The hospital anxiety and depression rating scale: a cross-sectional study of psychometrics..., Olsson [/bib_ref] although remains a screening tool rather than a diagnostic instrument. Furthermore, the cutoff ≥8 have been shown to provide the optimal balance between sensitivity and specificity in previous studies, but it is not clear whether this cutoff is optimal in older age groups. Some studies have suggested a cutoff ≥11 in older populations [bib_ref] Factor structure of the hospital anxiety and depression scale in older patients..., Flint [/bib_ref] , but we chose to use 8 due to limitations in sample size. In relation to the number of associations investigated (36 in total), we decided not to adjust for multiple comparison, which is in line with some recommendations [bib_ref] No adjustments Are needed for multiple comparisons, Rothman [/bib_ref]. Applying Bonferroni adjustments would reduce the number of significant associations to one (case-level anxiety associated with mother's condition after birth) but not change our conclusions dramatically. Other limitations were that information on maternal health and family circumstances as from the birth records was relatively crude, and gestational time had relatively little variance (with only 8.9% delivered before 40 weeks). These may have obscured some associations, although systematic bias is unlikely since the information gathering in HUSK was unlikely to have been influenced by birth circumstances and the information from the birth records was historic and transcribed blind to case status. ## Interpretation of our findings Out of the three previously published papers on the relationship between factors present at birth and anxiety and depression in old age [bib_ref] Factors associated with symptoms of anxiety and depression in five cohorts of..., Gale [/bib_ref] [bib_ref] Birth weight and the risk of depressive disorder in late life, Thompson [/bib_ref] [bib_ref] Depression in swedish women: relationship to factors at birth, Gudmundsson [/bib_ref] , only one has included both genders [bib_ref] Factors associated with symptoms of anxiety and depression in five cohorts of..., Gale [/bib_ref]. The authors of the latter also employed HADS as the outcome measure, and found no association between birth weight and anxiety or depression in old age across five different cohorts. They did not investigate other factors present at birth, and speculated that birth weight might be a too crude marker for fetal neurodevelopment [bib_ref] Factors associated with symptoms of anxiety and depression in five cohorts of..., Gale [/bib_ref]. We expanded on this potential limitation, and included several anthropometric measures, as well as factors like mother's condition, maternal familial history of disease and socioeconomic status. Interpretation of our secondary analysis should be done with caution, but the present results are indicative of possible continuation of socioeconomic disadvantage and an adverse impact for the offspring if the health in the mother was judged as poor during the hospital stay. Younger maternal age might lead to less experienced upbringing, or might reflect differences in socioeconomic position. The lack of association with parity at birth, however, does not suggest that the inverse association between case-level anxiety and depression with maternal age is because of participants' positions in their sibships. Associations with smaller pelvic size might be explained by poor nutrition during the mother's childhood, which in turn could negatively affect the development of the placenta and fetus in late pregnancy [bib_ref] Mothers' Pelvic size, fetal growth, and death from stroke and coronary heart..., Martyn [/bib_ref] , although it might also increase the risk of a difficult labour. Worse condition of the mother is of interest as this could reflect a variety of issues, including post-natal disorder, somatic illness or lower socioeconomic position. There was also some suggestion in our findings that socioeconomic markers were more associated with depression, and physical health markers more with anxiety. Our results are, however, not in line with the finding from the Swedish study where both low birth weight and shorter gestational age at birth were independent risk factors for life-time depression among women [bib_ref] Depression in swedish women: relationship to factors at birth, Gudmundsson [/bib_ref] , although this may be related to the previously mentioned low variance in gestational age in our sample. There are other reasons as to why few of the indicators were associated with anxiety and depression in old age. First, it is possible that factors identified at birth are more important in early or mid-life, but become less relevant as the time between exposure and outcome increases. Instead, other more proximal factors become stronger predictors of anxiety and depression in old age, such as social environment [bib_ref] Birthweight and behavioural problems in children: a modifiable effect, Kelly [/bib_ref] stressful life-events and loss of function [bib_ref] Risk factors for anxiety and depression in the elderly: a review, Vink [/bib_ref] [bib_ref] Factors associated with symptoms of anxiety and depression in five cohorts of..., Gale [/bib_ref]. Second, is possible that other measures of the foetal environment than those used here would be better indicators of any true link to anxiety and depression in old age, such as the size and shape of the placental surface [bib_ref] Beyond birthweight: the maternal and placental origins of chronic disease, Barker [/bib_ref]. # Conclusions As the life expectancy in many societies has increased, many individuals will live longer, and the pool of older adults with mental disorders will increase. Despite this development, less attention has been directed towards identifying risk factors for late life mental disorders, especially anxiety, than in younger age groups [bib_ref] Anxiety and depression in later life: Co-occurence and communality of risk factors, Beekman [/bib_ref] [bib_ref] Van Tilburg W: The natural history of late-life depression -a 6-year prospective..., Beekman [/bib_ref]. A better understanding of risk factors across the life-span, also during early life and foetal development, is needed, as this might help to increase identification and treatment of mental disorders in old age [bib_ref] Risk factors for anxiety and depression in the elderly: a review, Vink [/bib_ref]. Our findings indicate that there is only a weak association between status at birth and anxiety and depression in old age, and it might be that the effects of these risk factors are restricted to early or mid-life mental disorders. Although there was no clear evidence of any impact from educational attainment and physical health later in life for the investigated associations, a better perspective might be to adopt a life-course perspective or developmental origins of health and disease perspective, where foetal and early life-factors are considered in conjunction with physical and social exposures across the whole life-span [bib_ref] Rising to the challenges and opportunities of life course epidemiology, Ben-Shlomo [/bib_ref] [bib_ref] Developmental plasticity and human health, Bateson [/bib_ref] [bib_ref] A conceptual framework for the developmental origins of health and disease, Gluckman [/bib_ref]. This will allow for the exploration of early life factors' potential impact on and ability to modify later exposures related to the development of symptoms of anxiety and depression. [table] Table 1: Sample characteristics at birth obtained from medical records [/table] [table] Table 2: Associations between late life anxiety and depression case status and continuous or ordinal exposures recorded at birth [/table] [table] Table 3: Associations between case-level anxiety and depression, and dichotomous exposures in the birth journalsAge-and gender-adjusted OR (95% CI) * p < 0.05, ** p < 0.01, *** p < 0.001. N=394 (Males N=181 and females N=213). N=289 (Males N=135 and females N=154). [/table]

Improved Results in Paediatric Diabetes Care Using a Quality Registry in an Improvement Collaborative: A Case Study in Sweden Background: Several studies show that good metabolic control is important for children and adolescents with type 1 diabetes. In Sweden, there are large differences in mean haemoglobin A1c (HbA1c) in different hospitals and difficulties implementing national guidelines in everyday practice. This study shows how the participation in an improvement collaborative could facilitate improvements in the quality of care by paediatric diabetes teams. The Swedish paediatric diabetes quality registry, SWEDIABKIDS was used as a tool and resource for feedback and outcome measures.Methods: Twelve teams at paediatric diabetes centres, caring for 30% (2302/7660) of patients in Sweden, participated in an 18-month quality improvement program. Each team defined treatment targets, areas needing improvement, and action plans. The main outcome was the centre patients' mean HbA1c levels, but other clinical variables and change concepts were also studied. Data from the previous six months were compared with the first six months after starting the program, and the long-term follow up after another eleven months.Results: All centres reduced mean HbA1c during the second and third periods compared with the first. The mean reduction for all was 3?7 mmol/mol (p,0.001), compared with non-participating centres who improved their mean HbA1c with 1?7 mmol/mol during the same period. Many of the participating centres reduced the frequency of severe hypoglycaemia and/or ketoacidosis, and five centres reached their goal of ensuring that all patients had some sort of physical activity at least once weekly. Change concepts were, for example, improved guidelines, appointment planning, informing the patients, improving teamwork and active use of the registry, and health promotion activities.Conclusions: By involving paediatric diabetes teams in a quality improvement collaborative together with access to a quality register, the quality of paediatric diabetes care can improve, thereby contributing to a reduced risk of late complications for children and adolescents with diabetes. # Introduction Several studies, including the Diabetes Control and Complications Trial (DCCT), have shown that improving metabolic control is important to prevent, delay, or slow the progression of long-term complications from diabetes [bib_ref] Good glycemic control remains crucial in prevention of late diabetic complications-the Linkoping..., Nordwall [/bib_ref] [bib_ref] The significance of the prepubertal diabetes duration for the development of retinopathy..., Olsen [/bib_ref]. In Sweden, children and adolescents with type 1 diabetes are intensively treated following a national management policy according to International Society for Pediatric and Adolescent Diabetes (ISPAD) guidelines [bib_ref] Insulin treatment in children and adolescents with diabetes, Bangstad [/bib_ref]. The population is relatively homogenous, and the centres treat all patients in the catchment area without selection. However, there are substantial differences between the patient's mean haemoglobin A1c (HbA1c) reported at the centres. The latest annual registry report (data from 2012) showed a difference of 13?5 mmol/mol, about 1?2% in National Glycohemoglobin Standardization Program/Diabetes Control and Complications Trial (NGPS/ DCCT), between the centres with the lowest and highest mean HbA1c. These differences are not explained by clinical variables [bib_ref] A1C in children and adolescents with diabetes in relation to certain clinical..., Hanberger [/bib_ref]. Similarly, neither clinical nor treatment variables could explain the persistent differences between centres found by the Hvidoere study group in another large cohort from diabetes centres worldwide [bib_ref] Persistent differences among centers over 3 years in glycemic control and hypoglycemia..., Danne [/bib_ref] [bib_ref] Continuing stability of center differences in pediatric diabetes care: do advances in..., De Beaufort [/bib_ref] [bib_ref] Comparison of metabolic control in a crosssectional study of 2,873 children and..., Mortensen [/bib_ref]. Possible reasons suggested included a centre's effectiveness in implementing treatment regimens [bib_ref] Persistent differences among centers over 3 years in glycemic control and hypoglycemia..., Danne [/bib_ref] [bib_ref] Comparison of metabolic control in a crosssectional study of 2,873 children and..., Mortensen [/bib_ref] and clearly setting glycaemic targets [bib_ref] Target setting in intensive insulin management is associated with metabolic control: the..., Swift [/bib_ref]. Recently, a study within the Swedish paediatric diabetes quality registry, SWEDIABKIDSshowed that team members' policies and approaches affect glycaemic control in children and adolescents. One conclusion was that team members need to be aware of their approach and the importance of effective use of the resources within the team [bib_ref] The influence of structure, process, and policy on HbA(1c) levels in treatment..., Hanberger [/bib_ref]. Quality registries enable us to study clinical variables and outcomes of care. SWEDIABKIDS allows each diabetes centre to follow its results and to benchmark with other centres. Data is continuously registered and can also be followed continuously. Experience within other medical specialties has shown that a systematic quality improvement collaborative in combination with national quality registers can improve clinical results [bib_ref] Improved clinical outcome after acute myocardial infarction in hospitals participating in a..., Carlhed [/bib_ref] [bib_ref] Improving guideline adherence through intensive quality improvement and the use of a..., Peterson [/bib_ref]. With this background, members of the steering committee of SWEDIABKIDS invited all paediatric diabetes teams to participate in a quality improvement collaborative aiming to improve and standardize the quality of paediatric diabetes care. SWE-DIABKIDS should be used as a tool and resource for outcome measures. It was believed that improvement would be reached by changes in work processes and not by an increased work load. # Methods The Swedish paediatric diabetes quality registry Outpatient attendance data from all Swedish paediatric diabetes centres (n = 43) are registered in SWEDIABKIDS, a registry established in 2000. The completeness of centres reporting data increased from 32% to 100% from 2000 to 2007. In Sweden, paediatric departments treat all children and adolescents aged 0 to 18 years with diabetes within their catchment areas. Thus, the registry includes data on almost all (around 99%) of the children and adolescents with diabetes in Sweden. Until the end of 2011, the registry includes data from more than 361,000 outpatient visits. The registry has been web-based since 2008 and is available to all paediatric diabetes centres in Sweden. SWEDIABKIDS is financially supported by the Association of Local Authorities and Regions, SALAR, which represents the interests of Sweden's municipalities, county councils, and regions. SWEDIAB-KIDS has the status of a national quality registry. ## Hba1c analysis and clinical parameters All methods used in Sweden are standardized through the External Quality Assurance in Laboratory Medicine in Sweden (EQUALIS). The data on HbA1c obtained from SWEDIABKIDS was derived from capillary blood samples measured with the Bayer/Siemens DCA-2000 analyser or using local laboratory methods. Because the International Federation of Clinical Chemistry (IFCC) reference method has been adopted in Sweden, HbA1c values will be presented as IFCC (mmol/mol) results. For example, 58 mmol/mol (IFFC) corresponds to 7?5% (NGPS/ DCCT), whereas 10 mmol/mol is about 0?9%. According to the Swedish guidelines, children with diabetes visit the diabetes centre at least 4 times annually until the age of 18 years. At these visits, HbA1c and other clinical parameters such as insulin dose, weight, height, physical activity and blood pressure are measured. Physical activity is divided in 5 levels: never (level 1), less than one time/ week (level 2), one-two times/week (level 3), three-five times/week (level 4) and daily (level 5). Physical activity is defined as activity more the 30 minutes. The program for improvement of quality of diabetes care A quality improvement collaborative was conducted in cooperation between SWEDIABKIDS, Qulturum, the Jönköping County Council, and the Jönköping Academy for Improvement of Health and Welfare, Jönköping University. All 43 paediatric diabetes centres in Sweden were invited to participate in the program. Twelve accepted the invitation. In 2010, about 30% (2302/7660) of the patients in Sweden were cared for at these clinics. The number of patients varied between centres from 53 to 516, and the variance in yearly mean HbA1c was between 58?8 and 68?6 mmol/mol [fig_ref] Figure 1: The participating clinics, respective number of patients, and mean HbA1c levels in... [/fig_ref]. The improvement collaborative was designed with inspiration from the Breakthrough method [bib_ref] A framework for collaborative improvement: lessons from the Institute for Healthcare Improvement's..., Kilo [/bib_ref] [bib_ref] What do collaborative improvement projects do? Experience from seven countries, Wilson [/bib_ref] , included four learningsessions and two follow-up meetings, and had duration of 18 months [fig_ref] Figure 2: The duration of the collaborative was approximately 18 month with four learning... [/fig_ref]. Learning sessions included lectures on improvement methods, teamwork and learning, and sharing experiences between the teams. In the intervals between the learning sessions, the team identified problems and improvement areas at their centres, created action plans, tested changes, and followed up on the results. Most of the improvement work was done at their centre as an integrated part of their own work. The collaborative included learning about and working with systematic improvement methods; for example, the Value Compass, Microsystem analysis, flow charts, fishbone diagrams, and a plan-dostudy-act (PDSA) wheel to test different improvement ideas [bib_ref] A framework for collaborative improvement: lessons from the Institute for Healthcare Improvement's..., Kilo [/bib_ref] [bib_ref] Improving health care, Part 1: The clinical value compass, Nelson [/bib_ref] [bib_ref] A framework for the continual improvement of health care: building and applying..., Batalden [/bib_ref]. In addition to previously used approaches [bib_ref] Improving guideline adherence through intensive quality improvement and the use of a..., Peterson [/bib_ref] , each team also received support from an improvement coach. One member of each team was selected to be the team coach. The coach received extra training and support before the program started in a prophase during the work to effectively support their team at home in the improvement methods, during the action phase, and at the last learning session in a transition phase. The coach should also facilitate the communication with the centres' management to ensure that the work was supported. This coaching model was inspired by the model developed by Godfrey et al [bib_ref] Coaching interprofessional health care improrvement teams: the coachee, the coach and the..., Godfrey [/bib_ref] for coaching improvement teams in collaborative. Outcome variables for the project were clinical, processes, and what kinds of change concepts the team used to improve the work. Clinical variables included HbA1c, severe hypoglycaemia, and ketoacidosis. Process measures were documentation of smoking habits and the degree of physical activity. Each participating team was also allowed to define additional targets and outcome variables. Targets defined by specific centres are exemplified in this report as follows: A) to increase the proportion of patients with HbA1c,55 mmol/mol and to decrease the number of patients with HbA1c.70 mmol/mol; and B) to decrease the patients' mean HbA1c and compare it with the mean value of all clinics in Sweden. To investigate the effect of the quality improvement collaborative, the 6 months prior to the program commencement, November 2010 to April 2011 (period 1), was compared with both the 6-month intensive period (May to October 2011; period 2) and the period thereafter (November 2011 to September 2012; period 3). At the end of the quality improvement collaborative (October 2012), each team presented a final report. These were analysed with a qualitative content analysis [bib_ref] Qualitative content analysis in nursing research: concepts, procedures and measures to achieve..., Graneheim [/bib_ref] to find themes of change concepts used by the teams to improve their work. ## Statistics The statistical methods were mostly descriptive. To determine whether changes were significant, Student's T-test was used. A pvalue ,0.05 was considered significant. # Ethical consideration This study does not treat any data for identifiable individual patients; only aggregate data for different health care organizations. The study concerns improvement efforts undertaken by these organizations; not the actions or performance of individuals. Therefore, the study did not require ethical approval in the Swedish system. # Results ## Project outcome During periods 1, 2 (intensive period), and 3 (follow up period), the centres treated 2,032, 2,004, and 2,119 patients, respectively. At all points of measurement, the mean age of the patients was 1364.1 years, and 53% were males. As seen in [fig_ref] Figure 3: Changes in HbA1c at the different centres between periods 2 and 3... [/fig_ref] , all centres reduced their patients' mean HbA1c in period 2 compared with period 1. The difference was statistically significant for all centres, with a decrease of .1?8 mmol/mol (63?9-62?1 mmol/ mol) for the total population (p,0?01). This difference increased further during period 3 to 3?7 mmol/mol (63?9-60?2 mmol/mol) (p,0?001). The difference between period 2 and period 3 was also significant (p,0?01). The 32 clinics who not participated in the project reduced their mean HbAc1 with 0?9 mmol/mol between period 1 and period 2 (64?0-63?1 mmol/mol) and with 1.7 mmol/mol between period 1 and period 3 (64?0-62?3 mmol/mol, p,0?01). Females, in general, had a higher mean HbA1c during all periods compared with males (64.5, 63?1 and 60?6 mmol/mol compared with 62?8, 61?4, and 59?9 mmol/ mol, respectively), but the decrease was about the same as for males. In line with the decreased mean HbA1c during the three periods, the clinics increased the proportion of children with a mean HbA1c,57 mmol/mol from 31?4% in period 1 to 35?5% in period 2 (p,0?05), and 41?1% in period 3 (p,0?01 compared with period 1) [fig_ref] Figure 4: The proportion of patients with a mean HbA1c,57 mmol/mol [/fig_ref]. As seen in [fig_ref] Figure 4: The proportion of patients with a mean HbA1c,57 mmol/mol [/fig_ref] , the centres with a high mean HbA1c at the start of the project/program were the ones that achieved the best improvement in HbA1c. The clinics that not participated increased the proportion of children with mean HbA1c,57 mmol/mol from 28?8% in period 1 to 33?1% in period 2 and 38?4% in period 3 (p,0?001 compared to period 1). Many of the centres achieved the goal of reducing the frequency of severe hypoglycaemia and/or ketoacidosis (range 0.6%-4.8%). Only two centres increased the frequency (0.9% and 1.4%, respectively). Five of the centres attained the goal of ensuring that all their patients participated in some kind of physical activity of more than 30 minutes duration at least once weekly. ## Process variables The registration of data on smoking, physical activity, and hypoglycaemia/ketoacidosis increased for most of the participating centres. ## Change concepts The final reports show six main themes [fig_ref] Table 1: Theme of the change concepts from the teams' final reports [/fig_ref] according to the content analysis. Eleven out of 12 teams reported that they developed some activities to improve information, including communication and education to both patients and their families, and to staff members at the centre. All teams improved and updated their local guidelines and procedures, including, for example, routines for complication screening or eye examinations, introducing carbohydrate counting, or insulin pump introduction. Seven teams developed special guidelines for newly-diagnosed diabetes patients, and four developed special activities for patients with high HbA1c, with, for example, more frequent outpatient visits, individual care plans, or direct admission at unacceptably high HbA1c levels. Eight teams improved their reception planning and the same number improved how they used the SWEDIAB-KIDS registry in clinical work; for example, continuously reviewing statistics. Eight of the teams improved teamwork and six started or improved health promotion activities for their patients [fig_ref] Table 1: Theme of the change concepts from the teams' final reports [/fig_ref]. # Discussion The mean HbA1c level was reduced during period 2 when the teams had an intensive period of improvement work. Even more important was that the sustainability of the results was confirmed after another year of long-term follow up. This decrease in mean HbA1c at all participating centres shows the positive influence of the current quality improvement collaborative on the quality of paediatric diabetes care. The centres improved in relation to themselves and to the other centres. The decrease of 1?7 mmol/ mol in mean Hba1c at the non-participating clinics could to some degree be secondary to the quality improvement collaborative. This first project was discussed in most of the paediatric diabetes team in Sweden. Many of these teams started on their own to improve their results. This kind of substantial spillover effect on non-enrolled hospitals is known from other studies [bib_ref] National quality campaigns: who benefits?, Hansen [/bib_ref]. A majority of these teams now participate in the second quality improvement collaborative and so far the decrease in HbA1c continuous. The decrease in mean HbA1c is very encouraging. Many children benefit from this improvement and, if the results are sustainable, have less risk of late complications [bib_ref] Good glycemic control remains crucial in prevention of late diabetic complications-the Linkoping..., Nordwall [/bib_ref] [bib_ref] The significance of the prepubertal diabetes duration for the development of retinopathy..., Olsen [/bib_ref]. The results emphasize how important it is for health professionals to work continuously and systematically to improve the treatment, structure and processes of care. Hypoglycaemia can lead to disruptions and practical problems in daily life and have also been found to correlate with lower quality of life [bib_ref] Short-term effects of severe hypoglycaemia in children and adolescents with type 1..., Nordfeldt [/bib_ref]. Fear of hypoglycaemia may have a significant negative impact on diabetes management, metabolic control, and subsequent health outcomes [bib_ref] A critical review of the literature on fear of hypoglycemia in diabetes:..., Wild [/bib_ref]. Many of the centres attained or approached their goals of reduced frequencies of severe hypoglycaemia and/or ketoacidosis, and of ensuring that all their patients engaged in some sort of physical activity at least once weekly. There are clinicians who fear that decreasing HbA1c values increase the risk of severe hypoglycaemia. This project shows the opposite pattern. The benefit of physical activity for children and adolescents with diabetes includes better blood glucose control and enhanced insulin sensitivity [bib_ref] Physical activity, sport, and pediatric diabetes, Riddell [/bib_ref]. Some centres chose their own specific outcome variables, which were easily collected in SWEDIABKIDS; e.g. collecting data on the proportion of patients with low and high HbA1c and comparing the centre's mean HbA1c with the mean HbA1c in Sweden. The result of this study confirm that uniform, simple, and reliable measurements together with a systematic quality improvement stimulate team members and facilitate compliance with the activity plan or changes that the team has set up [bib_ref] Connections between quality measurement and improvement, Berwick [/bib_ref]. The final reports from the teams showed a high level of activity by the team members. Time at the seminars reserved for discussions within the teams provided opportunities to reach agreements on treatment issues, patient education, and treatment targets. In this way, the collaborative combined professional and improvement knowledge to improve the care for the patients, an approach which has been argued by Batalden and Stoltz [bib_ref] A framework for the continual improvement of health care: building and applying..., Batalden [/bib_ref]. Compared with previous collaborative programs [bib_ref] Improving guideline adherence through intensive quality improvement and the use of a..., Peterson [/bib_ref] , the concept of having a coach in each team was developed, a model inspired from the coaching model developed by Godfrey et al [bib_ref] Coaching interprofessional health care improrvement teams: the coachee, the coach and the..., Godfrey [/bib_ref]. The evaluation of the coaching will be presented in a separate publication. Some teams focused on the message to patients and reached agreement on the information to be conveyed to families. These factors have been found to be important for successful treatment and adherence to care plans [bib_ref] Target setting in intensive insulin management is associated with metabolic control: the..., Swift [/bib_ref] [bib_ref] The influence of structure, process, and policy on HbA(1c) levels in treatment..., Hanberger [/bib_ref]. Furthermore, the collaborative contributed to improving the completeness of data reported to the registry. It also contributed to validating the data in the registry by means of the discussions at the seminars. This led to better conditions for auditing and developing paediatric diabetes care. In summary, team members can support to decrease patients' mean HbA1c values for the group they serve. We have in this study shown that the access to a quality register to report data, receiving continuous feedback, and being able to compare the centre's own results over time transparently with other centres are important for successful improvement. Together with systematic work in a quality collaborative with support from a coach improvements can be achieved. Health professionals need to continuously work to improve the quality of paediatric diabetes care to reduce the risk of acute and late complications. Involving paediatric diabetes teams in a quality improvement collaborative can help the teams to improve important clinical results. [fig] Figure 1: The participating clinics, respective number of patients, and mean HbA1c levels in 2010. doi:10.1371/journal.pone.0097875.g001 [/fig] [fig] Figure 2: The duration of the collaborative was approximately 18 month with four learning sessions (LS) and two follow up meetings (FM). The team coaches began with a one-day education session followed thereafter by lunch-meetings (LM) at every learning session and phone-meetings (PM) in between. doi:10.1371/journal.pone.0097875.g002 [/fig] [fig] Figure 3: Changes in HbA1c at the different centres between periods 2 and 3 in relation to period 1. doi:10.1371/journal.pone.0097875.g003 [/fig] [fig] Figure 4: The proportion of patients with a mean HbA1c,57 mmol/mol. doi:10.1371/journal.pone.0097875.g004 [/fig] [table] Table 1: Theme of the change concepts from the teams' final reports. [/table]

It’s Hard to Avoid Avoidance: Uncoupling the Evolutionary Connection between Plant Growth, Productivity and Stress “Tolerance” In the last 100 years, agricultural developments have favoured selection for highly productive crops, a fact that has been commonly associated with loss of key traits for environmental stress tolerance. We argue here that this is not exactly the case. We reason that high yield under near optimal environments came along with hypersensitization of plant stress perception and consequently early activation of stress avoidance mechanisms, such as slow growth, which were originally needed for survival over long evolutionary time periods. Therefore, mechanisms employed by plants to cope with a stressful environment during evolution were overwhelmingly geared to avoid detrimental effects so as to ensure survival and that plant stress "tolerance" is fundamentally and evolutionarily based on "avoidance" of injury and death which may be referred to as evolutionary avoidance (EVOL-Avoidance). As a consequence, slow growth results from being exposed to stress because genes and genetic programs to adjust growth rates to external circumstances have evolved as a survival but not productivity strategy that has allowed extant plants to avoid extinction. To improve productivity under moderate stressful conditions, the evolution-oriented plant stress response circuits must be changed from a survival mode to a continued productivity mode or to avoid the evolutionary avoidance response, as it were. This may be referred to as Agricultural (AGRI-Avoidance). Clearly, highly productive crops have kept the slow, reduced growth response to stress that they evolved to ensure survival. Breeding programs and genetic engineering have not succeeded to genetically remove these responses because they are polygenic and redundantly programmed. From the beginning of modern plant breeding, we have not fully appreciated that our crop plants react overly-cautiously to stress conditions. They over-reduce growth to be able to survive stresses for a period of time much longer than a cropping season. If we are able to remove this polygenic redundant survival safety net we may improve yield in moderately stressful environments, yet we will face the requirement to replace it with either an emergency slow or no growth (dormancy) response to extreme stress or use resource management to rescue crops under extreme stress (or both). ## Preamble How organisms cope with environmental extremes can be traced to the distinctions between life forms explained in basic biology literature. Here the concept of life-cycles is extensively used. The two most important explanations for the universal existence of life cycles is their contributions to first, haploid/diploid switching to generate extreme genetic recombination (variation). Second, this genetic switch, during development of organisms, offers the ability to co-ordinate growth by mitosis and cell expansion with suitability of changing physical aspects of the environment for life. Even in the prokaryotic unicellular world, life cycle alterations switch between grow and no grow cycles when the environment is non-permissive to life. This resulted in the intensive selection for development of dormant spores from growing cells even without the genetic recombination advantage of meiosis. Then, after the advent of meiosis (eukaryotes) switching between spores or spore-like cells and growing cells continued. This developmental pattern was extended to multicellular organisms including higher plants that developed seeds as the dormant stage of life cycles. In addition to plants, animals, despite their mobility, also abundantly use avoidance/dormancy strategies to survive environments that are too extreme. Their developmental switches include hibernation and all the way to the extreme dormancy of tardigrades in their tun stage, which can out-survive any spore or seed (https://www.vox.com/science-and-health [bib_ref] Will the Antarctic tardigrade Acutuncus antarcticus be able to withstand environmental stresses..., Giovannini [/bib_ref]. This use of dormancy as a developmental stage is the clear primary foundation of the success of nearly all life that faces periods of extreme environments (e.g., seasons). We attempt to explain here how this use of dormancy in its many forms of development, not only spores and seeds but also many variations of reduced growth, form the plants stress avoidance strategies that are designed to avoid actual injury and death and eventually extinction. Avoiding extinction is driving the evolution of genes that control growth in threatening stress environments. Agriculturists want to use these genes to another purpose than just survival, namely productivity during stress. This use is in many ways contrary to the evolution of the functions of these genes. This is the underlying concept from which we use the term stress avoidance to explain the myriad biological forms of what we call in the literature, stress tolerance, resistance, adaptation, acclimation and even more derivative terms such as yield stability. Unfortunately, in the stress biology literature the term tolerance has been used by most plant scientists, including us, to describe plants (model systems and/or crop species) that grow exceptionally more than non-tolerant counterparts (species, cultivars and mutants) in a stress environment [bib_ref] Crop salt tolerance-current assessment, Maas [/bib_ref] [bib_ref] Engineering drought tolerance in plants: Discovering and tailoring genes to unlock the..., Umezawa [/bib_ref] [bib_ref] Osmogenetics: Aristotle to Arabidopsis, Maggio [/bib_ref] [bib_ref] Biotechnology of water and salinity stress tolerance, Pardo [/bib_ref]. More growth under stress has been seen as "more tolerance." However, from an evolutionary perspective, exactly the opposite is the case, less growth (avoidance of death) is more actual tolerance or more precisely, more survival [bib_ref] Four Arabidopsis AREB/ABF transcription factors function predominantly in gene expression downstream of..., Yoshida [/bib_ref] [bib_ref] ABF2, ABF3, and ABF4 Promote ABA-Mediated Chlorophyll Degradation and Leaf Senescence by..., Gao [/bib_ref] [bib_ref] Regulation of Leaf Starch Degradation by Abscisic Acid Is Important for Osmotic..., Thalmann [/bib_ref] [bib_ref] Transcription Factor AREB2 Is Involved in Soluble Sugar Accumulation by Activating Sugar..., Ma [/bib_ref]. We may therefore define Evolutionary Avoidance (EVOL-Avoidance) as a survival mode with the ability to avoid injury and death by various means, including dormancy. This is in contrast with Agricultural Avoidance (AGRI-Avoidance). This is a productivity mode, functional to avoid the EVOL-Avoidance response of which is an ancestral response, most of the time un-necessary in agricultural settings. This can be called as in our title: Avoiding Avoidance [fig_ref] Table 1: Relationships of tolerance nomenclature with growth, stress and survival [/fig_ref]. Avoiding avoidance would in nature lead to extinction. This, we suggest, is what would happen to our crops if returned to nature (without us) or, if in Agriculture, not rescued by us or rescued by hard to acquire genetics. We outline here evidence and supporting reasoning that specific genes and alleles that produce the appropriate webbed signal response in plants for abiotic stress "tolerance," as we have been using the word in most of the literature, exist in only few species that evolved in specific environments. Strictly speaking, we have been using the word tolerance from the Agricultural and not Evolutionary perspective. We see the reason for this dominant perspective as over concentrations on Agricultural traits that we consider desirable. So, the genes involved with the interactions of plants and their environments are intrinsically interconnected with those genes that control growth, both as cell number and cell size [bib_ref] Does proline accumulation play an active role in stress-induced growth reduction?, Maggio [/bib_ref]. The sensing and signalling systems involved in growth control are responsible for adjusting cell number and size to be in balance with the prevalent environmental conditions from which natural selection proceeds [bib_ref] Insights into the Origin and Evolution of the Plant Hormone Signaling Machinery, Wang [/bib_ref] [bib_ref] ABA receptor PYL9 promotes drought resistance and leaf senescence, Zhao [/bib_ref] [bib_ref] Combinatorial interaction network of abscisic acid receptors and coreceptors from Arabidopsis thaliana, Tischer [/bib_ref]. Agriculturists, in contrast, have selected in the opposite direction, for increased biomass and yield, with the unintended consequence that crops do not perform well under stressful conditions [bib_ref] Drought resistance, water-use efficiency, and yield potential-Are they compatible, dissonant, or mutually..., Blum [/bib_ref]. Wild plants exist that are less affected by the dichotomy of stress versus growth retardation. Even in crop species, genes exist that can be targets for modifications such that the crops will omit evolutionarily engrained growth retardation. Reduced growth retardation in response to stress, for example, can be bred for and can be engineered. However, this must be combined with management practices that prevent crops from "committing suicide" by continuing growth during progressively increasing stress exposure. Although it seems possible to re-engineer and breed such plants to activate stress protection gene systems besides reduced growth, in the case of drought, the availability of water for irrigation and its prudent management will be needed and not just more genes to reach higher productivity. Conflation of the wording describing tolerance as growing fast and sensitivity as growing slow in agriculture and/or experimental stress settings, while describing fast and slow growth in an evolution context in the opposite way has resulted in confusion. However, the genetic basis of stress tolerance in both Agricultural and Experimental settings and in Evolution has remained the same. That is, genes from evolution evoke tolerance by growing slow and thus avoidance of death. But! to be stress tolerant as it is called in Agricultural/Experimental settings we must find the genetic bases of fast growth under stress conditions but also avoid the eventual death caused by fast growth when stress becomes too extreme. Such plants that grow fast under moderate stress and still avoid death when stress becomes too extreme are rare but they do exist. A brief conceptual synthesis to reframe our search for key mechanisms that allow plant growth under stress follows: (1) Our known genes are almost always from experimental screens for mutants that grow slower and die faster (loss of function screen). (2) From this we conclude that the wildtype allele of this locus is required for full tolerance. (3) We usually conclude that such required genes may convey more tolerance when overexpressed ectopically. This is true but only incremental gains have been achieved. (4) Screening also for mutations that affect stress-induced marker gene expression has been very successful. Most of these mutants do display a stress phenotype also. (5) These required low and altered expression loci have provided information to build models of the signal network that controls plant responses to stress and the genes that the network controls. (6) The genes or alleles of genes that control our desired phenotypes are being used by plant species that display the search target phenotype (grows fast, called tolerant and does not die at high stress). (7) We need opposite screens, ones for more Agriculture/Experimental tolerance (grow faster/dies slower). These are gain of function screens. (8) Single genes can rarely convey a gain of function; therefore, single mutations in single genes are very unlikely to result in faster growth and dying slower (gain of functions). (9) Using molecular genetic tools with such search target natural species with the correct phenotype has not been plausible before. Now our technologies do allow us to carry out gene searches of these species (that have been previously very difficult to use experimentally) by several new and old approaches. ## Avoidance-the basis of plant abiotic stress tolerance Throughout this review, we use the term avoidance not as referred to in most plant stress literature but to indicate a broad genetic program (of which we know only some of its components) that sets plants in a survival mode. This involves the activation of multiple molecular and physiological responses that allow plants to avoid death (EVOL-Avoidance). This is different from the familiar 'drought avoidance,' a response involving adaptive traits that allow plants to maintain (relatively) higher tissue water content despite reduced water content in the soil and thus a relatively sustained growth. If "potentiated" this physiological mechanism would consent "some" growth of the plant relative to non-tolerant cultivars and/or species. This has been the underlying principle on which most genetic engineering for "stress tolerance," as defined in the past decades, has developed [bib_ref] Engineering drought tolerance in plants: Discovering and tailoring genes to unlock the..., Umezawa [/bib_ref]. Some components (traits) of the survival mode genetic program such as rapid life cycling with early flowering and reaching dormancy before resource limitation may turn-out useful in some agricultural contexts [bib_ref] Effects of cold-shock on the growth and flower bud differentiation of tomato..., Li [/bib_ref] [bib_ref] Early Flowering as a Drought Escape Mechanism in Plants: How Can It..., Shavrukov [/bib_ref]. We reason here and in the following sections that plant responses to unfavourable/hostile environments from an evolutionary perspective (EVOL-Avoidance) involve activation of a survival mode which is fundamentally based on deactivation of the growth mode. The most extreme form of the survival mode is dormancy, which indeed allows plants to overcome conditions of extreme stress. Hence, potentiation of all those mechanisms that would allow some or even much growth in plants under stress (i.e., mechanisms that would strengthen tolerance according to standard terminology) is a strategy that goes in the opposite direction to the evolutionary program which is also polygenic and redundant because it conveys survival. Consequently, this usually generates only very marginal growth improvements under stress, often hard to see under field conditions. Stated differently, the strategy to improve plant "tolerance" (AGRO-Avoidance) from an agricultural perspective, one that we have been following experimentally in the last several decades, has been focused on forcing plants to grow against their evolutionary genetic propensity or anthropologically "against their will." This is actually what the domestication process itself has also done. We do not advocate using the AGRO-Avoidance, EVOL-Avoidance terminology in future stress biology reports. It is intended only to help explain traditional terminology. We will discuss here how the survival mode in crop plants is tuned to over-protection from unfavourable growth conditions, which is indeed often needed in natural environments to survive.' Over-protection is mediated by genes that cannot risk failure to enter the survival mode as rapidly as needed in nature. Rapid entry into the survival mode is not usually needed in most agricultural environments where stresses are actually much less in intensity and do not usually persist for time periods long enough to threaten survival. Unravelling the genetic link between reduced growth and stress will open new research avenues to reduce un-necessary over-protection in plants in standard agricultural settings and to potentiate safety mechanisms that are deployed by plants and by us, only under extreme stress that actually threatens plant survival, a condition quite rare in most agricultural contexts [bib_ref] ABA receptor PYL9 promotes drought resistance and leaf senescence, Zhao [/bib_ref] [bib_ref] The unique mode of action of a divergent member of the ABA-receptor..., Zhao [/bib_ref] [bib_ref] Mutations in a subfamily of abscisic acid receptor genes promote rice growth..., Miao [/bib_ref] [bib_ref] Agrochemical control of plant water use using engineered abscisic acid receptors, Park [/bib_ref]. Therefore, this may be a good time to re-think thoughts that have shaped our views of the standard view of plant abiotic stress "tolerance" concept and interpretations of our expanding knowledge. Indeed, much information has been accumulated, much of it interesting and often correct in detail but real genetic based understanding has not yet come from the more than sixty (physiology) or forty (molecular), or twenty (genetics or genomics) years of work on the topic [bib_ref] Abscisic Acid: Emergence of a Core Signaling Network, Cutler [/bib_ref] [bib_ref] ABA perception and signalling, Raghavendra [/bib_ref] [bib_ref] Abiotic Stress Signaling and Responses in Plants, Zhu [/bib_ref]. The "physiology and molecular years" have been spent mostly on a few specific crops and easily cultivated models. With respect to a focus on what is called in standard terminology plant stress 'tolerance' or its opposite, sensitivity, the recent "genetics and genomics revolution" has been centred almost entirely on the model Arabidopsis, which as has been pointed out is not a stress "tolerant" species [bib_ref] Osmogenetics: Aristotle to Arabidopsis, Maggio [/bib_ref] [bib_ref] Improving crop salt tolerance, Flowers [/bib_ref]. Nevertheless, genetics and genomics concepts and several technological innovations applied to Arabidopsis have opened new vistas and paradigms, essentially by studying genes that result in an even more stress-sensitive Arabidopsis plant when altered [bib_ref] In vitro reconstitution of an abscisic acid signalling pathway, Fujii [/bib_ref] [bib_ref] Regulators of PP2C Phosphatase Activity Function as Abscisic Acid Sensors, Ma [/bib_ref]. In turn, it is becoming easier to apply what we learned with Arabidopsis to other plants [bib_ref] Agrochemical control of plant water use using engineered abscisic acid receptors, Park [/bib_ref]. Turning to species adapted to extreme environments can teach us much about what will allow implementation of successful agronomical practices over a broad range of challenging environments [bib_ref] Plant salt tolerance, Zhu [/bib_ref] [bib_ref] Salt Cress. A Halophyte and Cryophyte Arabidopsis Relative Model System and Its..., Inan [/bib_ref] [bib_ref] Salinity stress adaptation competence in the extremophile Thellungiella halophila in comparison with..., Gong [/bib_ref] [bib_ref] Functional gene-mining for salt-tolerance genes with the power of Arabidopsis, Du [/bib_ref] [bib_ref] Learning from Evolution: Thellungiella Generates New Knowledge on Essential and Critical Components..., Amtmann [/bib_ref] [bib_ref] Genome Structures and Halophyte-Specific Gene Expression of the Extremophile Thellungiella parvula in..., Oh [/bib_ref] [bib_ref] Anastatica hierochuntica, an Arabidopsis Desert Relative, Is Tolerant to Multiple Abiotic Stresses..., Eshel [/bib_ref] [bib_ref] Overexpression of AlTMP2 gene from the halophyte grass Aeluropus littoralis in transgenic..., Ben-Romdhane [/bib_ref]. We reason that we might better look for superior characters (e.g., those that uncouple slow growth and stress response) in species that are evolutionarily adapted to thrive sometimes even with high productivity in extreme habitats and challenging climates [bib_ref] A comparative study of salt tolerance parameters in 11 wild relatives of..., Orsini [/bib_ref] [bib_ref] The genome of the extremophile crucifer Thellungiella parvula, Dassanayake [/bib_ref] [bib_ref] Insights into salt tolerance from the genome of Thellungiella salsuginea, Wu [/bib_ref] [bib_ref] Salt Stress in Thellungiella halophila Activates Na + Transport Mechanisms Required for..., Vera-Estrella [/bib_ref] [bib_ref] The Thellungiella salsuginea Tonoplast Aquaporin TsTIP1;2 Functions in Protection Against Multiple Abiotic..., Wang [/bib_ref] [bib_ref] Commercial cultivation and bioremediation potential of sugar kelp, Saccharina latissima, in Danish..., Marinho [/bib_ref] [bib_ref] A Single Amino-Acid Substitution in the Sodium Transporter HKT1 Associated with Plant..., Ali [/bib_ref]. It is there that we will likely be able to find what has eluded detection up to now by using the advanced technologies. Recent work in rice has shown that desensitization of plant response to ABA, through inhibition of specific components of the ABA/PYL receptors system, may enhance plant growth while reducing the negative effects of stress responses [bib_ref] Mutations in a subfamily of abscisic acid receptor genes promote rice growth..., Miao [/bib_ref]. What we will find among these evolutionarily adapted species will be mechanisms that are largely made up of ubiquitous genes that have been used in novel ways to avoid, or work around, problems posed by the environment [bib_ref] A Single Amino-Acid Substitution in the Sodium Transporter HKT1 Associated with Plant..., Ali [/bib_ref] [bib_ref] AtHKT1 drives adaptation of Arabidopsis thaliana to salinity by reducing floral sodium..., An [/bib_ref]. We suggest that plants lack genuine (in the conventional sense) tolerance-conferring genes and proteins but that success in coping with disastrous events comes almost universally from the presence of genes and proteins mediating mechanisms that induce the survival mode (our term-EVOL-Avoidance). These are molecular/physiological traits that allow plants to survive and go to the next generation. These same traits are not necessarily key to improve crop productivity under suboptimal growth conditions. These mechanisms evolved in plants necessitated by the daily struggle to recognize deviations between homeostatic internal optima and continuously changing external environmental conditions that threatened the very survival of a species over long time periods. The true, accurate avoidance concept can be grasped by considering that even a so-called "tolerance" protein must manage not to undergo an environmentally-induced conformation change, denaturation, or inhibitory interaction originating in the environment [bib_ref] Hsp90 as a capacitor of phenotypic variation, Queitsch [/bib_ref]. Likewise, considering the true nature (what has evolved) of "stress tolerance" (our EVOL-Avoidance), the thought has emerged that the same genes conferring protection to high stress in naturally "tolerant," avoiding species that grow slow are also present in sensitive species that grow fast [bib_ref] The Thellungiella salsuginea Tonoplast Aquaporin TsTIP1;2 Functions in Protection Against Multiple Abiotic..., Wang [/bib_ref] [bib_ref] AtHKT1 drives adaptation of Arabidopsis thaliana to salinity by reducing floral sodium..., An [/bib_ref]. The emergence of plant life on land relied on a set of genes present from the beginning, whereas evolution re-used these genes or modified versions (alleles) to produce functions (phenotypes) not needed by fresh water adapted plants (glycophytes). These genes returned to creating species capable of halophytism again. These retro-type species began once again their pre-land occupation functions occupying niches, such as estuaries and salt marshes. A ramification of this view is that, first, all species could be or become either very fast or slow growing species under stress. Second, we have in the past most likely not dealt with really important genes or versions of genes (alleles) since we have only partially succeeded in isolating those components that are introduced and/or duly modify in important crop species could improve their ability to grow fast under stress (ability to avoid avoidance, in the context of agricultural settings, see [fig_ref] Table 1: Relationships of tolerance nomenclature with growth, stress and survival [/fig_ref]. This view, we think, is not in conflict with data on neo-functionalization of genes, for example, starting with alternatively spliced or otherwise processed and increasingly sequence-divergent and so forth isoforms, that can escape from adaptive conflict by further directional selection after duplication [bib_ref] The Thellungiella salsuginea Tonoplast Aquaporin TsTIP1;2 Functions in Protection Against Multiple Abiotic..., Wang [/bib_ref] [bib_ref] A Single Amino-Acid Substitution in the Sodium Transporter HKT1 Associated with Plant..., Ali [/bib_ref] [bib_ref] AtHKT1 drives adaptation of Arabidopsis thaliana to salinity by reducing floral sodium..., An [/bib_ref] [bib_ref] The recruitment of crystallins: New functions precede gene duplication, Piatigorsky [/bib_ref] [bib_ref] Escape from adaptive conflict after duplication in an anthocyanin pathway gene, Des Marais [/bib_ref] [bib_ref] An Arabidopsis PWI and RRM motif-containing protein is critical for pre-mRNA splicing..., Zhan [/bib_ref] [bib_ref] Spliceosomal protein U1A is involved in alternative splicing and salt stress tolerance..., Gu [/bib_ref]. The standard tolerance-avoidance terminology has long ago been introduced [bib_ref] Genetic and agronomic contributions to yield gains: A case study for wheat, Bell [/bib_ref] [bib_ref] Yield Potential: Its Definition, Measurement, and Significance, Evans [/bib_ref] [bib_ref] Recent advances in engineering plant tolerance to abiotic stress: Achievements and limitations, Vinocur [/bib_ref]. Again, we do not advocate supplanting all of it with new terms like avoidance and avoiding avoidance. However, we emphasize that the mechanisms employed by plants to naturally cope with a stressful environment resulting from evolution are overwhelmingly geared to avoid detrimental effects that can lead to injury and death. The new genetic, genomic and bioinformatics approaches and data collected from them have been consistent with the view of avoidance as a unifying concept. ## Survival and avoidance of injury and death are negatively correlated with growth, biomass and productivity The first noticeable casualty of a changing environment is the growth rate. It is often described as being the most sensitive response to stress [bib_ref] Plant Productivity and Environment, Boyer [/bib_ref]. Stress may be viewed as a condition that does not permit optimal growth, measured in comparison to the growth rate of the species or genotype in optimal conditions. Here, it must be clear that "experimental" growth in a controlled, optimal environment, phytotron, growth chamber, or culture plate should not be considered practically comparable with growth in natural settings; many typically stress-induced genes are highly expressed in a seemingly stress-free environment under field conditions [bib_ref] The use of gene expression profiling to dissect the stress sensitivity of..., Zinselmeier [/bib_ref] [bib_ref] Transcript expression profiles of Arabidopsis thaliana grown under controlled conditions and open-air..., Miyazaki [/bib_ref]. In a changing environment, growth modifications occur, either quantitatively (biomass/cell divisions), or qualitatively (switch in developmental phase) or both [bib_ref] Does proline accumulation play an active role in stress-induced growth reduction?, Maggio [/bib_ref] [bib_ref] Analysis of sto1/nced3, an Abscisic Acid-Deficient but Salt Stress-Tolerant Mutant in Arabidopsis, Ruggiero [/bib_ref] [bib_ref] Control of Plant Water Use by ABA Induction of Senescence and Dormancy:..., Zhao [/bib_ref]. Clearly, an altered environment may lead to resource limitation, which then necessarily limits growth but this is also based on a genetic hard-wired program aimed at avoiding a non-sustainable situation. Recent studies showed that the activity of Target-of-Rapamycin (TOR) kinase, the central regulator of energy and biomass production, is inhibited by osmotic stress and abscisic acid (ABA) treatment through SnRK2 mediated phosphorylation of Raptor [bib_ref] Glucose-TOR signalling reprograms the transcriptome and activates meristems, Xiong [/bib_ref] [bib_ref] Reciprocal Regulation of the TOR Kinase and ABA Receptor Balances Plant Growth..., Wang [/bib_ref] , indicating the presence of a pro-active growth control system activated by environmental stresses. Similarly, Zhao et al., [bib_ref] Control of Plant Water Use by ABA Induction of Senescence and Dormancy:..., Zhao [/bib_ref] have shown that the loss of CYCLIN-DEPENDENT KINASE C2 function increases cell division during leaf development and enhances plant growth under drought stress. From an evolutionary perspective, however, survival is what is established in a changing environment by natural genetic responses. Plant growth shows distinct phases of cell divisions, extremely fast or imperceptibly slow cell expansion and many degrees and stages of maturation and reproductive development. The genetic expression guiding those events are pre-programmed, mediated by hormones and fine-tuned in response to the environment by receptors and many layers of signal control networks. It is environmental fine tuning that concerns us because the degree of the effect of external conditions on growth is not hard-wired but follows a malleable program with hard-wired limits [bib_ref] Reciprocal Regulation of the TOR Kinase and ABA Receptor Balances Plant Growth..., Wang [/bib_ref]. In principle, except for situations of extreme resource limitation that can lead to death, there is no functional reason for a plant to initiate slow growth (under moderate and/or transitory stress) other than that it has the genes that direct such a response as an evolutionary survival mode. Arguably, many responses to stress have been interpreted as strictly adaptive, whereas we suggest that most actions are general alarm reactions, or responses aimed at preventing the inevitable resource limitation. However, plant hypersensitivity which activates these alarm reactions before reaching resource limits prevents the achievement of full growth potential under moderately/mild stress. This occurs quite commonly throughout a growth season [bib_ref] Engineering drought tolerance in plants: Discovering and tailoring genes to unlock the..., Umezawa [/bib_ref] [bib_ref] Maize stomatal conductance in the field: Its relationship with soil and plant..., Tardieu [/bib_ref] [bib_ref] How Plants Cope with Water Stress in the Field? Photosynthesis and Growth, Chaves [/bib_ref] [bib_ref] Seasonal and multiannual effects of salinisation on tomato yield and fruit quality, De Pascale [/bib_ref]. If carefully examined, alarm reactions can be revealed as attempts to enter the survival mode (grow slow/low productivity/avoid death) more rapidly. The most observed and obvious alarm is stomatal closing. The most extreme alarm response is senescence and dormancy [bib_ref] Control of Plant Water Use by ABA Induction of Senescence and Dormancy:..., Zhao [/bib_ref]. Plant biologists have used genes involved in the alarm reactions to gain "tolerance" (AGRO-Avoidance) by overexpressing them. This approach has led only to incremental, or marginal, advances toward our ultimate goal. This and the complexity of stress responses, has also resulted in many different, incomparable data appearing in the literature, including measures of plant survival [bib_ref] Stress signaling through Ca 2+ /calmodulin-dependent protein phosphatase calcineurin mediates salt adaptation..., Pardo [/bib_ref] , an important trait under stress but not very helpful in most agricultural contexts. Indeed, increased alarm reactions do usually lead to better survival and subsequent growth after stress dissipates, just as has occurred in evolution. Such genetic variants are good at avoiding death or presenting an appearance (e.g., less injury) that we associate with better "tolerance" or growth (AGRO-Avoidance) [bib_ref] ABA receptor PYL9 promotes drought resistance and leaf senescence, Zhao [/bib_ref] [bib_ref] Control of Plant Water Use by ABA Induction of Senescence and Dormancy:..., Zhao [/bib_ref]. The resurrection plant, which enters a vegetative dormancy before reproductions, represents an extreme alarm reaction that is the result of natural selection in an environment where survival required it. In specific agriculture settings it also could be important as is escape [bib_ref] Biomass Production, Nutritional and Mineral Content of Desiccation-Sensitive and Desiccation-Tolerant Species of..., Yobi [/bib_ref] [bib_ref] The resurrection genome of Boea hygrometrica: A blueprint for survival of dehydration, Xiao [/bib_ref]. A fundamental concept re-emerges here. As dictated by the environment, stress-responsive genes have evolved not for biomass production but as required for survival. Commonly overlooked is that survival genes act in response to conditions that are usually rare in agricultural contexts (production environments). Their functions developed over a long time scale, a scale that by far exceeds the ten millennia of plant culture. We argue that the defect or absence of a gene that is needed only once in many years for the survival of a species-these may include genes controlling dormancy under extreme drought events [bib_ref] Molecular and biochemical mechanisms associated with dormancy and drought tolerance in the..., Pnueli [/bib_ref] will lead as easily to extinction as a gene needed daily (e.g., genes involved in stomatal regulation [bib_ref] Analysis of sto1/nced3, an Abscisic Acid-Deficient but Salt Stress-Tolerant Mutant in Arabidopsis, Ruggiero [/bib_ref]. This concept has a historical foundation. [bib_ref] Molecular and Organismic, Dobzhansky [/bib_ref] has famously stated that "nothing makes sense in biology except in the light of evolution" [bib_ref] Molecular and Organismic, Dobzhansky [/bib_ref]. This has inspired our effort to re-examine what we mean by "tolerance." Earlier, Darwin envisioned biological diversity and plasticity in terms of natural selection and on the basis of heritable diversity upon which natural selection would work, although some rethinking is taking place here in light of a new look at epigenetics. It has taken Mendel's breakthrough and more than 100 years of experimentation to know that there are still "unknown unknowns". The most obvious unknowns are the various members of the RNA world and epigenetic informationthat add until recently unanticipated control layers [bib_ref] Nuclear RNA export and its importance in abiotic stress responses of plants, Chinnusamy [/bib_ref] [bib_ref] Novel long non-protein coding RNAs involved in Arabidopsis differentiation and stress responses, Ben Amor [/bib_ref]. We think it reasonable to modify the often used axiom of 'survival of the fittest' and change it into 'what we see today is what has survived.' However, what has survived can only be compared to a crude paleontological record, revealing important logical gaps in our view of evolution: (1) Fossils are our only way of knowing what has not survived. Records of [CO 2 ] trapped in ice cores, for example, provide valuable but insufficient data points. (2) We cannot reconstruct in any detail the environments in which the non-survivors failed. We barely know essential details of today's environments. Certainly, experiments to test hypotheses about extinctions are not possible. Even in fields of science where direct experimentation is possible, results may lead to hypotheses that give rise to erroneous conclusions. There are two important concepts we would like to put forward: first, survival is the ultimate measure of evolutionary success, not productivity. Agricultural 'success' cannot depend on survival alone but must be concerned with productivity. Seeds represent exquisite survival mechanisms; they represent the ultimate wager. It is however not the number of seeds that counts; rather it is the seeds that produce new plants that are the guarantors of survival. Second then, if increased productivity carries even the slightest disadvantage for survival over a long time period, it will be selected against. Therefore, limits of a productivity phenotype cannot be predicted in a strictly evolutionary context. A corollary of this fact is that productivity ranges in any environment for presently existing species are not predictable as well, because the ranges in productivity are genetically irrelevant for survival. Therefore, they must be empirically used only to reveal the underlying genetic structure and plasticity of the species in question. The crop domestication process provides valid illustrations for these statements. The original wild populations of our domesticated species had nowhere near the productivity of later domesticated cultivars even under optimal conditions. A remarkably small number of genetic changes was necessary to bring about the very large increases in productivity in non-stressful conditions. These few gene alterations coupled with adjustments in agricultural practices designed by the domesticators allowed early farmers to maximize the genetic benefits [bib_ref] The Genetics of Maize Evolution, Doebley [/bib_ref] [bib_ref] A Genomic Scan for Selection Reveals Candidates for Genes Involved in the..., Chapman [/bib_ref] [bib_ref] Genes and Mutations Underlying Domestication Transitions in Grasses, Sang [/bib_ref]. In rice, in fact, the green revolution was brought about by a single mutation in a gene later identified as encoding a GA-synthesis gene [bib_ref] Positional Cloning of Rice Semidwarfing Gene, sd-1: Rice "Green Revolution Gene" Encodes..., Monna [/bib_ref] [bib_ref] A mutant gibberellin-synthesis gene in rice: Green revolution, Sasaki [/bib_ref] [bib_ref] Semidwarf (sd-1), "green revolution" rice, contains a defective gibberellin 20-oxidase gene, Spielmeyer [/bib_ref] [bib_ref] Artificial selection for a green revolution gene during japonica rice domestication, Asano [/bib_ref]. When combined with better fertilization practices, this gene brought about the substantial increase in productivity in the "green revolution" [bib_ref] GIBBERELLIN INSENSITIVE DWARF1 encodes a soluble receptor for gibberellin, Ueguchi-Tanaka [/bib_ref]. Our existing crops already depend on human management to survive and any newly developed crops will most likely be even more dependent on human intervention in their growth. Environmental control of the productivity range or limits of a phenotype are ultimately determined by resource availability. Resource limitations that are potentially associated with environmental stress includes those that have typically been noted, such as reduced amounts of [CO 2 ] when stomata are closed, or water supply when stomata are open, or supply of any essential nutrients. However, we may also add process constraints among the resource limitations. For example, these can be limits to altered membrane transport to accelerate ion homeostasis in saline or water deficit conditions and stress-dependent altered hormone homeostasis, or metabolite disequilibrium (C/N) conditions. Following this perspective, many genes and pathways undoubtedly contributed to the phenotype of abiotic stress avoidance. ## Plant growth versus survival and productivity There is a strong correlation between success, viewed as survival, in extreme environments and natural slow growth, even constitutive slow growth in the absence of stress. We note that this correlation is not absolute and that exceptions have been discussed [bib_ref] ABA receptor PYL9 promotes drought resistance and leaf senescence, Zhao [/bib_ref] [bib_ref] Agrochemical control of plant water use using engineered abscisic acid receptors, Park [/bib_ref] [bib_ref] Stress-Induced Osmotic Adjustment in Growing Regions of Barley Leaves, Matsuda [/bib_ref] [bib_ref] Salt Tolerance and Crop Potential of Halophytes, Glenn [/bib_ref] [bib_ref] Learning from the Arabidopsis Experience. The Next Gene Search Paradigm, Bressan [/bib_ref] [bib_ref] Leveraging abscisic acid receptors for efficient water use in Arabidopsis, Yang [/bib_ref]. We state that a focus on this correlative dichotomy has distorted our views of the true nature of the linkage between stress and productivity. Importantly, the correlation strongly implies that many species have not evolved satisfactory stress EVOL-avoidance mechanisms that allow fast growth in moderately and still persist in more extreme, stressful environments. Corollary to this consideration is that slow growth and success under extreme stressful conditions, although linked in complex fashion, may conceivably be separated genetically [bib_ref] Does proline accumulation play an active role in stress-induced growth reduction?, Maggio [/bib_ref] [bib_ref] ABA receptor PYL9 promotes drought resistance and leaf senescence, Zhao [/bib_ref] [bib_ref] Analysis of sto1/nced3, an Abscisic Acid-Deficient but Salt Stress-Tolerant Mutant in Arabidopsis, Ruggiero [/bib_ref]. We emphasize that this very tight correlation has caused previous considerations to often confuse stress EVOL-avoidance with resource limitation. A good example of the perplexingly strong linkage between compromised growth and stress without resource limitation is provided by describing responses to NaCl relative to water deficit [bib_ref] Stress-Induced Osmotic Adjustment in Growing Regions of Barley Leaves, Matsuda [/bib_ref]. In saline environments water is almost always plentiful but thermodynamically difficult to retrieve and thus requires a genetic response to adjust metabolism to the thermodynamic imbalance, whereas water in a dry environment is limited and may approach zero, thus becoming a limited resource, yet slow growth is tightly linked to both stresses. Similar constraints apply to all micro-and macro-nutrients and to the extreme ranges of temperature or light that restrict growth or induce altered growth and development. Within extreme ranges, however, the altered slower growth phenotype must remain linked to a genetic response program that assures survival even at the expense of growth before the extremes that cause resource limitation and ultimately death, are reached. Exposure to conditions that exceed life chemistry limits, such as taking more water to cool a larger plant, more ions to sustain growth and more energy allocation to fight biotic challenges are avoided to a large degree by small plants. Thus, large plants are less able to utilize micro-niches or to generate sufficient protection from environmental extremes. Productivity is not the issue under such conditions. The correlation between stress and reduced growth emerged from physiological or natural selection paradigms, not from the insights provided by experiments and artificial genetics, an idea that has been difficult to amend let alone discard. At the risk of overstating, in our crops, as in most species, stress reduces growth such that biomass accumulation (overall productivity) and yields and reproductive productivity are almost always compromised. We argue again that the negative correlation between growth in a stressful environment and biomass is mostly based on an EVOL-avoidance strategy in that, importantly, "anticipates" encountering non-survivable environmental conditions [bib_ref] Does proline accumulation play an active role in stress-induced growth reduction?, Maggio [/bib_ref]. The relationship of reduced growth to any experimentally defined version of AGRI-avoidance is presently confusing relative to EVOL-avoidance because these AGRI-avoidance versions almost always neglect the survival advantage or experimental sensitivity [fig_ref] Table 1: Relationships of tolerance nomenclature with growth, stress and survival [/fig_ref]. Equalizing these coincidental but conflicting views of the relationship of growth to the conventional term tolerance (AGRI-avoidance) without considering survival reactions to stress environments (EVOL-avoidance) may have hindered the path to understanding the true importance of different genetic mechanisms. Consequently, we are compelled to again state that slow growth results from being exposed to stress because genes to adjust growth rate to external circumstances have evolved as a survival (EVOL-avoidance), not productivity (AGRI-avoidance), strategy (and not a strategy of altered growth only) that has allowed extant plants to necessarily avoid extinction (EVOL-avoidance). Consistent with this statement, recent research has clearly revealed in unprecedented detail, the reciprocal regulation of ABA and TOR signalling that controls several processes in balancing plant growth and stress responses [bib_ref] Glucose-TOR signalling reprograms the transcriptome and activates meristems, Xiong [/bib_ref] [bib_ref] Reciprocal Regulation of the TOR Kinase and ABA Receptor Balances Plant Growth..., Wang [/bib_ref]. ## The energy limitation myth An analysis calculating the efficiency of photosynthesis for converting solar energy into biomass has been conducted and the differences between achievable maxima and biomass productivity have been discussed [bib_ref] What is the maximum efficiency with which photosynthesis can convert solar energy..., Zhu [/bib_ref] [bib_ref] Photosynthetic Energy Conversion Efficiency: Setting a Baseline for Gauging Future Improvements in..., Slattery [/bib_ref]. Values for C3 and C4 plants emerging from this work were below those that constitute a theoretical maximum. However, the values, based on well-researched bottlenecks and inevitable energy losses, are still higher than what amounts to the average conversion efficiency shown by our major crops. The authors attribute this difference to "unfavourable physical environments"-stresses. The calculations allow for another conclusion. The evolution of land plants over approximately 450 million years has, as far as can be gathered, not resulted in favouring increased conversion efficiency to make use of what is available. Differently stated, it seems that plants did not need to evolve toward improving energy use efficiency in using a resource that is almost always virtually unlimited. This convincingly indicates that energy is not a limiting factor in most ecosystems. That plants do not make use of what is available and often make do with much less, is certainly stress-dependent but we see this as a relationship associated with fewer and smaller cells, or no cells at all, that plants generate by the EVOL-avoidance or survival mode response over long periods of stress. This strategy, again, is geared at avoiding extinction and not to avoid excess energy expenditure. The smallness concept, resulting in lower productivity, has been discussed for a hundred years or more [bib_ref] Osmogenetics: Aristotle to Arabidopsis, Maggio [/bib_ref] [bib_ref] The Wilting Coefficient and Its Indirect Determination, Briggs [/bib_ref] [bib_ref] Soil moisture in relation to the growth of plants, Shantz [/bib_ref]. Others have also deduced that the "energy" problem most likely also reflects the long-term-on an evolutionary scale-avoidance of extinction by resource limitation [bib_ref] The mechanism of background extinction, Wiens [/bib_ref]. We consider this critical strategy as almost never based on energy resource limitations. A striking example is provided by true shade plants that will not grow faster in higher light-a condition that appears to represent the same slow growth phenomenon (EVOL-avoidance) seen in many other abiotic stress conditions, that are not initially resource limiting. Such conditions include water deficit, salinity, temperature extremes, declining nutrient availability and even high light stress that can induce biochemical damage. These stresses generate limits only over time but the decline in plant growth is often much more instantaneous. Accordingly, less energy consumption almost certainly reflects a survival strategy acting over a period of time much longer than agricultural settings when a larger energy allocation-as bigger plants consume more resources-eventually reaches a limited state. Thus any genetic program that continues to guide increased biomass will require even larger resource allocations not to combat stress environments but to avoid in the long run constraints that will likely lead to extinction (see halophytes-discussed in [bib_ref] Osmogenetics: Aristotle to Arabidopsis, Maggio [/bib_ref] [bib_ref] A comparative study of salt tolerance parameters in 11 wild relatives of..., Orsini [/bib_ref]. Again, plants have evolved to respond early and be prepared for challenges by the environment and even often have constitutive genetic programs to meet these potential environmental assaults. Plants embarking on a route that results in relatively increased total biomass, fruit or seed, persevered less well under stress conditions and became extinct not only because of impending resource limitations but also because they became bigger targets for pathogens and predators and supported a much larger expansion of plant species populations even allowing more competition between species. This is also true within a species population that becomes a larger threat to species survival. This is adequately supported by the common observation of within species allelopathic inhibitions in extremely dry environments. Thus, we see the conservation of energy in extant species serving an entirely different purpose (involving long-term survival) that is not related to the notion that energy is often limiting or even close to limiting in agricultural time frames in almost any stressful environment. An important aspect related to the energy consideration is harvestable resource allocation or, rather, the problem of harvest index and yield stability under stress [bib_ref] Plant Productivity and Environment, Boyer [/bib_ref] [bib_ref] Yield potential, yield stability and stress tolerance in maize, Tollenaar [/bib_ref] [bib_ref] Crops for a Salinized World, Rozema [/bib_ref]. In fact, maize breeders unknowingly selected for avoidance of stress avoidance, AGRI-avoidance, by selecting against increased competition caused by higher planting densities, through analysis of yield stability, defined as the overall highest yield under a range of different environments at different planting densities [bib_ref] Yield potential, yield stability and stress tolerance in maize, Tollenaar [/bib_ref]. If actual resource limits can be circumvented by applying appropriate agronomical practices, some species can grow under severe stress conditions at rates that are equal or close to those of a typical crop species under non-stress conditions [bib_ref] Improving crop salt tolerance, Flowers [/bib_ref] [bib_ref] Salt Tolerance and Crop Potential of Halophytes, Glenn [/bib_ref] [bib_ref] Dynamics of Growth and Water Relations of Fodderbeet and Seabeet in Response..., Niazi [/bib_ref]. In other words, extremophile species can exhibit naturally, astoundingly high yield stability [bib_ref] Genetic selection and improvement of hard wood tree species for fuelwood production..., Goel [/bib_ref] [bib_ref] Potential source of ligno-cellulosic biomass for ethanol production, Abideen [/bib_ref]. This fact provides the major justification for paying attention to the growth rates with and without stress of particular plant species that are close relatives of the Arabidopsis model and of extremophile versions of crop species [bib_ref] Anastatica hierochuntica, an Arabidopsis Desert Relative, Is Tolerant to Multiple Abiotic Stresses..., Eshel [/bib_ref] [bib_ref] Overexpression of AlTMP2 gene from the halophyte grass Aeluropus littoralis in transgenic..., Ben-Romdhane [/bib_ref] [bib_ref] Learning from the Arabidopsis Experience. The Next Gene Search Paradigm, Bressan [/bib_ref] [bib_ref] Abiotic Stress and Plant Genome Evolution. Search for New Models, Amtmann [/bib_ref]. ## A new conceptual framework We consider now the plant abiotic stress archetype by incorporating a genetic-genomics viewpoint that has only more recently become possible. Unlike in the past, we have now available genome and transcriptome sequences and collections of tagged or otherwise generated genetically altered lines that reveal by their phenotypes the underlying function of the altered genes. In addition, the gene silencing and editing technologies are further revolutionizing our abilities [bib_ref] Efficient genome editing in plants using a CRISPR/Cas system, Feng [/bib_ref]. Genes that respond to abiotic stresses, based on microarray and RNA sequencing technologies coupled with statistical analyses [bib_ref] Four Arabidopsis AREB/ABF transcription factors function predominantly in gene expression downstream of..., Yoshida [/bib_ref] [bib_ref] ABF2, ABF3, and ABF4 Promote ABA-Mediated Chlorophyll Degradation and Leaf Senescence by..., Gao [/bib_ref] [bib_ref] Transcription Factor AREB2 Is Involved in Soluble Sugar Accumulation by Activating Sugar..., Ma [/bib_ref] [bib_ref] Salinity stress adaptation competence in the extremophile Thellungiella halophila in comparison with..., Gong [/bib_ref] [bib_ref] Transcriptome Changes for Arabidopsis in Response to Salt, Osmotic, and Cold Stress, Kreps [/bib_ref] [bib_ref] Monitoring expression profiles of Arabidopsis gene expression during rehydration process after dehydration..., Oono [/bib_ref] [bib_ref] Integration of Arabidopsis thaliana stress-related transcript profiles, promoter structures, and cell-specific expression, Ma [/bib_ref] [bib_ref] An Arabidopsis gene network based on the graphical Gaussian model, Ma [/bib_ref] [bib_ref] Arabidopsis decuple mutant reveals the importance of SnRK2 kinases in osmotic stress..., Fujii [/bib_ref] [bib_ref] Arabidopsis mutant deficient in 3 abscisic acid-activated protein kinases reveals critical roles..., Fujii [/bib_ref] [bib_ref] Three SnRK2 Protein Kinases are the Main Positive Regulators of Abscisic Acid..., Fujita [/bib_ref] [bib_ref] Arabidopsis PYR/PYL/RCAR Receptors Play a Major Role in Quantitative Regulation of Stomatal..., Gonzalez-Guzman [/bib_ref] [bib_ref] The cbfs triple mutants reveal the essential functions of CBFs in cold..., Jia [/bib_ref] [bib_ref] Three Arabidopsis SnRK2 Protein Kinases, SRK2D/SnRK2.2, SRK2E/SnRK2.6/OST1 and SRK2I/SnRK2.3, Involved in ABA..., Nakashima [/bib_ref] [bib_ref] The Arabidopsis Transcription Factor NAC016 Promotes Drought Stress Responses by Repressing AREB1..., Sakuraba [/bib_ref] [bib_ref] AREB1, AREB2, and ABF3 are master transcription factors that cooperatively regulate ABRE-dependent..., Yoshida [/bib_ref] [bib_ref] Mutational Evidence for the Critical Role of CBF Genes in Cold Acclimation..., Zhao [/bib_ref] , show significant overlap between different stresses and their nature has helped to identify pathways and networks that are affected by different stresses [bib_ref] Four Arabidopsis AREB/ABF transcription factors function predominantly in gene expression downstream of..., Yoshida [/bib_ref] [bib_ref] The Arabidopsis Transcription Factor NAC016 Promotes Drought Stress Responses by Repressing AREB1..., Sakuraba [/bib_ref] [bib_ref] AREB1, AREB2, and ABF3 are master transcription factors that cooperatively regulate ABRE-dependent..., Yoshida [/bib_ref] [bib_ref] Mutational Evidence for the Critical Role of CBF Genes in Cold Acclimation..., Zhao [/bib_ref]. The many examples of past attempts at genetic engineering included modifications of osmotic adjustment (betaine accumulation), ion transport (CAX, HKT1, SOS1, NHX), [bib_ref] A Single Amino-Acid Substitution in the Sodium Transporter HKT1 Associated with Plant..., Ali [/bib_ref] [bib_ref] AtHKT1 drives adaptation of Arabidopsis thaliana to salinity by reducing floral sodium..., An [/bib_ref] [bib_ref] Wheat grain yield on saline soils is improved by an ancestral Na..., Munns [/bib_ref] , membrane potential (ATP synthases), radical oxygen scavenging and redox control (SODs, catalases). Also, morphological traits, hormone homeostasis (ABA, IAA) and signalling (receptors, protein kinases/phosphatases), [bib_ref] ABA receptor PYL9 promotes drought resistance and leaf senescence, Zhao [/bib_ref] [bib_ref] The unique mode of action of a divergent member of the ABA-receptor..., Zhao [/bib_ref] [bib_ref] Leveraging abscisic acid receptors for efficient water use in Arabidopsis, Yang [/bib_ref] [bib_ref] Unique Drought Resistance Functions of the Highly ABA-Induced Clade A Protein Phosphatase..., Bhaskara [/bib_ref] [bib_ref] An ABA-mimicking ligand that reduces water loss and promotes drought resistance in..., Cao [/bib_ref] [bib_ref] Combining chemical and genetic approaches to increase drought resistance in plants, Cao [/bib_ref] [bib_ref] Modulation of Abscisic Acid Signaling in Vivo by an Engineered Receptor-Insensitive Protein..., Dupeux [/bib_ref] [bib_ref] Overexpression of PYL5 in rice enhances drought tolerance, inhibits growth, and modulates..., Kim [/bib_ref] [bib_ref] Potent and selective activation of abscisic acid receptors in vivo by mutational..., Mosquna [/bib_ref] [bib_ref] Activation of dimeric ABA receptors elicits guard cell closure, ABA-regulated gene expression,..., Okamoto [/bib_ref] , or transcription factors involved in the activation of various pathways have been exploited [bib_ref] Recent advances in engineering plant tolerance to abiotic stress: Achievements and limitations, Vinocur [/bib_ref] [bib_ref] AREB1 Is a Transcription Activator of Novel ABRE-Dependent ABA Signaling That Enhances..., Fujita [/bib_ref] [bib_ref] Abscisic acid-dependent multisite phosphorylation regulates the activity of a transcription activator AREB1, Furihata [/bib_ref]. In many cases some avoidance of stress avoidance improvement (AGRI-avoidance) has been observed, which often could be repeated and quantified [bib_ref] Recent advances in engineering plant tolerance to abiotic stress: Achievements and limitations, Vinocur [/bib_ref]. However, with a few exceptions [bib_ref] Engineering drought tolerance in plants: Discovering and tailoring genes to unlock the..., Umezawa [/bib_ref] the engineered lines have not stood up to the test in the field [bib_ref] Genetic engineering to improve plant performance under drought: Physiological evaluation of achievements,..., Lawlor [/bib_ref]. In addition, presumably protective effects disappeared over several generations owing to our unawareness of the importance of epigenetic control circuits that respond to environmental forces. Potentially, practical avoiding of stress avoidance (AGRI-avoidance) has been defeated also by agriculturally negative phenotypes such as undesirable altered flowering times [bib_ref] Double overexpression of DREB and PIF transcription factors improves drought stress tolerance..., Kudo [/bib_ref] linked to AGRI-Avoidance through complexity of signal networks. The next generation of stress AGRI-avoidance engineering, we surmise, will utilize the greatly enhanced knowledge and the gene resources available [bib_ref] Using Information from Arabidopsis to Engineer Salt, Cold, and Drought Tolerance in..., Zhang [/bib_ref]. For example, a condition-specific promoter has been already, long ago, engineered to control induction of a gene leading to altered hormone homeostasis [bib_ref] Isolation and characterization of an ipt gene from the Ti plasmid Bo542, Strabala [/bib_ref] that has also been shown to mediate responses to limited water [bib_ref] ABA receptor PYL9 promotes drought resistance and leaf senescence, Zhao [/bib_ref] [bib_ref] Delayed leaf senescence induces extreme drought tolerance in a flowering plant, Rivero [/bib_ref] Other entities, high in the hierarchy of genes and pathways that may be strengthened are being explored, with transcription factor and receptor genes high on the list [bib_ref] Using Information from Arabidopsis to Engineer Salt, Cold, and Drought Tolerance in..., Zhang [/bib_ref] [bib_ref] Plant nuclear factor Y (NF-Y) B subunits confer drought tolerance and lead..., Nelson [/bib_ref]. Engineering approaches are rapidly being surpassed by gene editing technologies that have addressed the common redundancy of hormone mediated signalling [bib_ref] Mutations in a subfamily of abscisic acid receptor genes promote rice growth..., Miao [/bib_ref] and the very complex regulation of plant growth versus responses to environmental stresses, which may involve additional/unidentified components of osmotic stress signalling [bib_ref] Arabidopsis Duodecuple Mutant of PYL ABA Receptors Reveals PYL Repression of ABA-Independent..., Zhao [/bib_ref]. These technologies may even present allele conversion capabilities because gene editing can produce multiple SNPs (Single Nucleotide Polymorphisms). The most successful strategies in our opinion will include the modification of pathways that govern meristem activities, including flowering, by disconnecting or countermanding the connection between stress sensing and slow growth initiation. The NFYB transcription factor from maize provides a fascinating example of future approaches [bib_ref] Plant nuclear factor Y (NF-Y) B subunits confer drought tolerance and lead..., Nelson [/bib_ref]. Upon over-expression, it suppresses (causes avoidance of avoidance, that is, AGRI-avoidance) the slow growth phenotype that is the typical response to moderate water deficits. The plants no longer 'anticipate' and respond by EVOL-avoidance to an increasingly stressful environment that could lead to death over a longer time period. A short duration and moderate non-resource limiting drought period would result in these engineered plants having increased biomass and yield (AGRI-avoidance) as reported and more recently confirmed under field conditions [bib_ref] Plant nuclear factor Y (NF-Y) B subunits confer drought tolerance and lead..., Nelson [/bib_ref]. We do not know whether these plants would experience some survival problems in the absence of agricultural management (e.g., due to impaired stomatal closure at advanced stress conditions). Similar results have been obtained via overexpression of bacterial RNA chaperones in corn, a strategy that has led to the development of the only commercial corn line for improved grain yield under water-limited conditions [bib_ref] Bacterial RNA Chaperones Confer Abiotic Stress Tolerance in Plants and Improved Grain..., Castiglioni [/bib_ref]. In agricultural environments such stress/slow growth (EVOL-avoidance) decoupled plants will require the addition of precise resource management or deployment only under strict, stable climates to avoid catastrophic crop failures (see Beachy R. [bib_ref] Water: More crop per drop, Marris [/bib_ref]. The increasingly variable weather patterns associated with global climate change is exacerbating such strategies. Four fundamental physiological processes have emerged from decades of abiotic stress research that should help guide future abiotic stress genetic engineering approaches and attempts. These processes and importantly the signal networks that control them are especially important for performance in drought, osmotic and salinity stress conditions since they compose the major responses that allow plants to avoid serious injury and ultimate death. These include (a) osmotic and ionic adjustments [bib_ref] Important roles of drought-and cold-inducible genes for galactinol synthase in stress tolerance..., Taji [/bib_ref] [bib_ref] Regulatory metabolic networks in drought stress responses, Seki [/bib_ref] , (b) radical oxygen induced injury containment [bib_ref] Abiotic Stress Signaling and Responses in Plants, Zhu [/bib_ref] , (c) other aspects of metabolic homeostasis [bib_ref] Abscisic Acid: Emergence of a Core Signaling Network, Cutler [/bib_ref] [bib_ref] Regulation of the Arabidopsis CBF regulon by a complex low-temperature regulatory network, Park [/bib_ref] [bib_ref] Co-evolution of Hormone Metabolism and Signaling Networks Expands Plant Adaptive Plasticity, Weng [/bib_ref] and (d) growth control [bib_ref] Reciprocal Regulation of the TOR Kinase and ABA Receptor Balances Plant Growth..., Wang [/bib_ref]. These mechanisms are controlled by hormone-mediated signalling and therefore this approach must be considered on different levels of "intervention urgency," with the higher-order hormone control circuits and signal networks of superior importance. In this category, two major aspects require attention. First, we need more knowledge about how to disengage damage response aspects of signal networks from the growth arrest and growth maintenance aspects of growth networks. Second, the opposing or combining effects of stress hormone alterations on both EVOL-Avoidance (slow growth) and on developmental thresholds, for example an inappropriate flowering time or pollen silk interval and control of the root/shoot growth ratio in connection to stress [bib_ref] A cis cold memory element and a trans epigenome reader mediate Polycomb..., Yuan [/bib_ref] , must be understood more fully at the genetic level. By moving the community's focus onto the concept AGRI-avoidance as earlier defined rather than premature alarms (EVOL-avoidance) [fig_ref] Table 1: Relationships of tolerance nomenclature with growth, stress and survival [/fig_ref] we hope to stimulate discussion for which we have provided here arguments and a justification for viewing plants as having evolved genes and phenotypes for EVOL-avoidance stress survival. Again, we do not advocate changing existing terminology, only how we think about this terminology. The following are some of the areas of consideration that we advocate to be discussed in efforts to refocus. (1) On an organismal level we need to distinguish the roles in stress of the mechanisms that operate in different tissues and cells and during developmental phase changes that define windows of intervention. (2) Rather than using blinders that exclude whatever is outside our narrow focus, we need to recognize that the sum of effectors that respond to the environment should be visualized as a web or field of vectors which often 'pull' the phenotype into different directions, perhaps in a manner where multiple vectorial components do not act independently and may antagonize or potentiate each other. We point to the example of a spider's web, where the spider unerringly moves to the place the web's contact with prey that may be far from the spider's position and even out of its view. Instead of a "linear thread" that leads directly to the spider, the spider produces a "web" a net or orb shaped structure that can capture prey from many directions and positions. This contact signal initiates a highly specific wave (dependence on the point of contact) that reaches the spider and is perceived and processed (this is poorly understood) to reveal the exact position of contact whereupon the spider moves directly to this position. It appears that the spider has evolved signal network processing perhaps like computer learning or even a crude version of the human brain. We should also keep in mind that both of these processing networks have limitations. We need to know much more about how webbed pathways (interconnected signalling pathways) function and are controlled perhaps by comparisons to neural networks. (3) Such a field of vectors acts not only through ABA or ethylene or auxin, or salicylate or jasmonate but certainly through suppressors and activators mediated by hormone combination. A complete hormonal profiling and signalling molecules in various genetic backgrounds is of highest priority because hormone networks manifest the major phenotypes controlling stress avoidance and survival [bib_ref] A Spatio-Temporal Understanding of Growth Regulation during the Salt Stress Response in..., Geng [/bib_ref] [bib_ref] A small peptide modulates stomatal control via abscisic acid in long-distance signalling, Takahashi [/bib_ref]. Hormone receptors usually have the first contact with an exogenous signal. They exert the most dramatic effects through the webs of developmental and external genetic response programs. Even though some receptors and many intermediates such as kinases, phosphatases and so forth have been discovered, many signal components remain elusive. Especially understanding the roles of interconnections (cross-talking) perhaps will well reveal a mechanism for re-wiring as in computer learning and human memory-which involve likely epigenetic mechanisms. (4) Genes have names that have been given by researchers but the genes may have and many do, have functions that can become obscured by focusing on the labels we contrived. Annotations should never be confused with functions, even with assignments based on phenotypes because they are still most likely incomplete. Genes can have more than one function and they may have different functions depending on the cellular, temporal and conditional context in which they operate [bib_ref] A Single Amino-Acid Substitution in the Sodium Transporter HKT1 Associated with Plant..., Ali [/bib_ref] [bib_ref] The ABA Receptor PYL8 Promotes Lateral Root Growth by Enhancing MYB77-Dependent Transcription..., Zhao [/bib_ref]. A lucid example of the latter has been provided by a study of amyotrophic lateral sclerosis in a mouse model. The mice, over-expressing a SOD1 gene perplexingly both potentiated the disease and were found to respond positively to an inhibitor of SOD1 activity. It was later found in a more rigorous study that these results were explained because the SOD1 enzyme possessed another previously unknown activity that generated other more toxic ROS [bib_ref] Design, power, and interpretation of studies in the standard murine model of..., Scott [/bib_ref]. This and several other cases [bib_ref] Osmotin Is a Homolog of Mammalian Adiponectin and Controls Apoptosis in Yeast..., Narasimhan [/bib_ref] have shown that the DNA sequence does not always reveal its true function. (5) Stress response networks operate on several levels of control, including many that are largely unknown, especially the epigenetic components that have only recently and gradually come into focus [bib_ref] Epigenetic inheritance in plants, Henderson [/bib_ref] [bib_ref] Small RNAs as big players in plant abiotic stress responses and nutrient..., Sunkar [/bib_ref] [bib_ref] Epigenetic control of plant stress response, Boyko [/bib_ref]. Stress adaptation must be also viewed as another epigenetic state such as operates during ontogeny where signal networks may be formed in different patterns [bib_ref] Nuclear RNA export and its importance in abiotic stress responses of plants, Chinnusamy [/bib_ref]. The manifestation of many EVOL-avoidance mechanisms, through classical and epigenetic means, have generated the present biodiversity of species that escaped extinction and established ecosystems in a wide range of climates and conditions. (6) Irrespective of this apparent complexity, we have been encouraged by the surprisingly large phenotypic effects exerted by individual genes emerging from Arabidopsis screens. Yet, it must be remembered that the degree of effect of a single locus on a complex trait depends dramatically on its genetic background (other genes) that can be equated with the evolutionary history of each species. Yield, certainly a complex trait, has been significantly increased by very few domestication genes that were necessary to convert a wild species into a crop that now depends on our care [bib_ref] The Genetics of Maize Evolution, Doebley [/bib_ref] [bib_ref] The nature of selection during plant domestication, Purugganan [/bib_ref]. The trait, flowering time, also illustrates very well the ability of a single locus (or of very few loci) in the appropriate background to have major effects on complex traits. Such phenomena have been recognized as the "Mendelization" of quantitative traits to which one might add that all traits have a quantitative component [bib_ref] Osmogenetics: Aristotle to Arabidopsis, Maggio [/bib_ref] [bib_ref] Quantitative genetics in the age of omics, Keurentjes [/bib_ref] [bib_ref] Recombinant near-isogenic lines: A resource for the mendelization of heterotic QTL in..., Pea [/bib_ref]. (7) Nevertheless, these examples underscore the real possibility that such critical loci affecting positively and negatively EVOL-Avoidance exist. As we have promulgated here, their criticality to survival has almost certainly made them redundant. The formidable late generation sequence technologies and advanced correlative algorithms have allowed them to be exposed to us now. Needed then, are also more inclusive databases (and the curation of misleading annotations therein). Because the information content is getting quite large, we need better statistical methods to sort and categorize and continuously alter and improve the mathematical tools and their applications. The iPlant concept (http://iplantcollaborative.org/) supported by the National Science Foundation was a very early undertaking in the direction of an integrated database of all plant knowledge, which will require additional consideration of this process. Additional examples include international initiatives on data publishing such as FAIR a Findable, Accessible, Interoperable and Reusable data sharing platform [bib_ref] The FAIR Guiding Principles for scientific data management and stewardship, Wilkinson [/bib_ref] ; the PLAZA platform (https://bioinformatics.psb.ugent.be/plaza) a plant-oriented online resource for comparative, evolutionary and functional genomics [bib_ref] PLAZA 4.0: An integrative resource for functional, evolutionary and comparative plant genomics, Van Bel [/bib_ref] ; software to perform genome-wide association studies (GWAS) in a variety of species [bib_ref] Zbrowse: An interactive GWAS results browser, Ziegler [/bib_ref]. The correlation of phenotypic metadata with the genome structure of a large number of Arabidopsis accessions and mutants, for example, has been in the past a high priority [bib_ref] The 1001 Genomes Project for Arabidopsis thaliana, Weigel [/bib_ref]. Also, we need scrutiny of the observational quality provided for several non-crop model species and crops because metadata of older experiments and annotations are often lacking or incorrect. We need integration of transcriptome dynamics, as well as protein amount, turnover, activity control and structure. We also need more information about metabolic pathway interactions and finally the epigenetic status of each stress adaptation state of a crop species. Only large databases can present such information in a cogent manner and preferably in the form of interactive network models which are still ongoing endeavours. Virtual experiments can be generators of ideas, provide clues, predictions and hypotheses that can initiate (well-planned) wet lab experiments, eventually replacing assumptions with reality-tested rules. This may even drive new algorithms to more accurate predictions for future experimental efforts. ## Genomics, bioinformatics and integration Research areas tend to reach stagnation and conceptual deadlock and become in need of rejuvenation. A major problem it seems is the narrow view that results from focusing on a specific, usually easily defensible, hypothesis promising incremental progress in a well-researched field. In fact, funding agencies have insisted on caution and have rewarded timidity. Still focusing on advancing fortunately, Arabidopsis has produced formidable sophistication in molecular and genetic tools for this species that allow us to study stress in an integrative fashion to trace the input factors to their genetic roots and apply them to crop species. As a note of caution, over-emphasis on technical abilities can diminish conceptual progress. Abiotic stress is not to be simply understood by, for example, measuring ABA and the expression of all genes responding to ABA. Improving biomass or yield under stress does not come from simply knowing all components that are involved in the stress response. As we have pointed out, most observed responses are symptoms and indicators of deviations from homeostatic equilibrium with any particular environment. The importance of such responses should be carefully pre-examined (e.g., by edited knock outs) before deciding our level of commitment to them. That Arabidopsis has become the primary stress model is based on superior understanding of the plant's development and the tools that have been developed for analyses. As a stress tolerance (EVOL-avoidance or AGRO-avoidance) model (either or both grows faster under stress or survives higher stress) this plant is arguably inappropriate or at least insufficient. What we have done is make the plant genetically more sensitive (grows slower dies sooner) to a stress and then record the genetic entities that are associated with such engineered sensitivity [fig_ref] Table 1: Relationships of tolerance nomenclature with growth, stress and survival [/fig_ref] [bib_ref] Genetic Analysis of Plant Salt Tolerance Using Arabidopsis, Zhu [/bib_ref]. In some cases of such approaches, more so in development and in the gene-for-gene background of biotic stress responses connected to innate immunity, we have generated finely tuned compromised genetic lines [bib_ref] Plant immunity: Unravelling the complexity of plant responses to biotic stresses, Miller [/bib_ref]. However, this approach also can go only a limited distance before the interconnections of these responses eventually become obviously limiting to our progress (as previously observed in gene for gene disease resistance that cannot be transferred to other species and cold tolerance that is difficult to genetically identify in other species [bib_ref] Regulation of the Arabidopsis CBF regulon by a complex low-temperature regulatory network, Park [/bib_ref] [bib_ref] Plant immunity: Unravelling the complexity of plant responses to biotic stresses, Miller [/bib_ref]. Needed are different models. Models that are true extremophiles and genotypes with either stress EVOL-avoidance or fast growth (AGRI-Avoidance) and especially different species that display either mechanisms that have been provided by evolution will finally yield real world solutions tailored to very specific ecologies and climates. For example, fast growing marine species (e.g., sugar kelp) can provide good sources of some important loci and alleles [bib_ref] Commercial cultivation and bioremediation potential of sugar kelp, Saccharina latissima, in Danish..., Marinho [/bib_ref]. These models should also, for now, if possible exhibit some features that make Arabidopsis so easily studied and manipulated. Such models are found among close relatives of Arabidopsis. We have termed them Arabidopsis Relatives Model Systems, ARMS [bib_ref] Learning from Evolution: Thellungiella Generates New Knowledge on Essential and Critical Components..., Amtmann [/bib_ref] [bib_ref] Learning from the Arabidopsis Experience. The Next Gene Search Paradigm, Bressan [/bib_ref]. Yet we need to go further still, for example, to wild relatives of rice, or cotton [bib_ref] QTL mapping in A-genome diploid Asiatic cotton and their congruence analysis with..., Ma [/bib_ref]. In fact, we believe that such models can be found in each crop family and that we can expect family-specific stress tolerance (EVOL-avoidance) mechanisms that will paraphrase mechanisms that unite plants. The genetic programs supporting stress avoidance of avoidance (AGRI-avoidance) apparent in uncommon wild species can now be reasonably approached with whole-genome forward and reverse genetics, genome-wide association mapping combined with the identification of essential or important loci by whole genome sequencing and re-sequencing [bib_ref] The genome of the extremophile crucifer Thellungiella parvula, Dassanayake [/bib_ref] [bib_ref] Insights into salt tolerance from the genome of Thellungiella salsuginea, Wu [/bib_ref]. This will, we expect, provide targets for more precise gene editing especially targets that are redundant such as PYL receptors [bib_ref] Mutations in a subfamily of abscisic acid receptor genes promote rice growth..., Miao [/bib_ref]. In essence, we consider that one organism's intolerable stress can be another one's normal condition and that comparative studies of extremophiles are an excellent way to finding the ways that disengage productivity from the slow growth survival mode [bib_ref] Mutations in a subfamily of abscisic acid receptor genes promote rice growth..., Miao [/bib_ref] [bib_ref] Functional gene-mining for salt-tolerance genes with the power of Arabidopsis, Du [/bib_ref] [bib_ref] Genome Structures and Halophyte-Specific Gene Expression of the Extremophile Thellungiella parvula in..., Oh [/bib_ref] [bib_ref] A comparative study of salt tolerance parameters in 11 wild relatives of..., Orsini [/bib_ref] [bib_ref] Abiotic Stress and Plant Genome Evolution. Search for New Models, Amtmann [/bib_ref]. ## Thinking around the dogmas-searching for answers where the problems are Crop yields have increased tremendously in the last 100 years, yet the yield potential of crops is still at least some three times higher than what the best lines and best agricultural practices can accomplish today [bib_ref] Yield potential, yield stability and stress tolerance in maize, Tollenaar [/bib_ref] [bib_ref] Improving drought tolerance in maize: A view from industry, Campos [/bib_ref] [bib_ref] Improving photosynthesis and crop productivity by accelerating recovery from photoprotection, Kromdijk [/bib_ref]. Irrespectively, modern lines have steadily increased yields even in less than optimal environments. This "breeding" coupled with appropriate agricultural management largely has been responsible for the increases. An obvious conclusion based on past progress is that our approaches to stresses, especially abiotic stresses, including water deficit and ion imbalance, have let us make only incremental improvements whereas fundamental progress is what farmers, industry and politicians expect and what is desperately needed to feed a growing more affluent global population [bib_ref] Water: More crop per drop, Marris [/bib_ref] [bib_ref] Global food security: Assessing trends in view of guiding future EU policies, Maggio [/bib_ref]. To achieve increased food security calls for unprecedented cooperation between molecular geneticists, breeders, agronomists, geologists, hydrologists, economists and scientists of many other disciplines [bib_ref] Planning for food security in a changing climate, Mckersie [/bib_ref]. Political leaders will have to recognize that food shortages will dwarf the numerous experienced devastations caused by earthquakes, volcanoes, tsunamis, floods, hurricanes and such and will occur before we will be hit by an asteroid and before global climate change exacerbates temperatures, polar ice cap melting and sea levels rise enough to intensify these many problems. When lack of fresh water becomes more prevalent and when hunger and starvation strikes, all other problems, goals and aspirations become irrelevant [bib_ref] The green revolution: Meetings that changed the world, Hardin [/bib_ref]. We have argued here that the evolutionary process converged on growth control as the mechanism most reliable for the survival of a plant species. This is diametrically opposite to the demands we are placing on our crops. Genetic studies that have been directed towards understanding AGRO-avoidance we again emphasize, point towards the lack of actual tolerance mechanisms. How could a plant become tolerant to the absence of water? Plants must avoid the absence of water or remain in some dormant state (e.g., seed, spore, etc.). There is that old adage-plants cannot grow without water [bib_ref] Can the Quest for Drought Tolerant Crops Avoid Arabidopsis Any Longer?, Maggio [/bib_ref]. It should be stated again and clearly understood that the AGRO-Avoidance "tolerance" mechanisms that are usually cited, stomatal closure, enhanced root to shoot growth ratios, epidermis wax layers or trichome growth or altered orientation of the leaves, in our view, are not mechanisms that in themselves make plants AGRO-avoidance "tolerant" [bib_ref] Control of Plant Water Use by ABA Induction of Senescence and Dormancy:..., Zhao [/bib_ref] and grow more under stress. It is clear that evolution favoured survival (EVOL-Avoidance) and almost never AGRO-Avoidance "tolerance" [fig_ref] Table 1: Relationships of tolerance nomenclature with growth, stress and survival [/fig_ref] , which by necessity in evolution, resulted in genetic programs that direct plants to initiate EVOL "avoidance." This response almost universally and importantly reduces growth through activating growth repressing or deactivating growth inducing developmental programs. In evolution, growth (or metabolic) control quickly becomes the key process in any actual or potential resources limited state. The difficulty in finding a mechanism to achieve desired growth control (faster) can be compared to a strategy where plants are "hedging their bets," to be small, grow slowly and to conserve available resources, such as water, that would become limiting earlier in a large plant. This view has inevitably led us to understanding that stressed plants do not initially grow slowly and produce less because of actual resource deficiencies or energy limitations [bib_ref] Osmogenetics: Aristotle to Arabidopsis, Maggio [/bib_ref] [bib_ref] Does proline accumulation play an active role in stress-induced growth reduction?, Maggio [/bib_ref]. Exposing the fallacies surrounding energy and growth limitations we think, can reveal possibilities to more effectively control development that directs energy and growth to the specific parts of the plant that are useful that is, root growth in water deficits or harvest index in all stresses that initiate reduced growth (harvest). Moreover, stress-adapted species, which can be found in many plant families that support this approach as well [bib_ref] Salt Cress. A Halophyte and Cryophyte Arabidopsis Relative Model System and Its..., Inan [/bib_ref] [bib_ref] Learning from the Arabidopsis Experience. The Next Gene Search Paradigm, Bressan [/bib_ref] [bib_ref] Salinity tolerance in halophytes, Flowers [/bib_ref] [bib_ref] Comparative Genomics in Salt Tolerance between Arabidopsis and Arabidopsis-Related Halophyte Salt Cress..., Taji [/bib_ref]. In fact, in the case of saline adapted species, some species are known [bib_ref] Productivity and irrigation requirements of halophytes grown with seawater in the Sonoran..., Glenn [/bib_ref] [bib_ref] Agricultural production of halophytes irrigated with seawater, O&apos;leary [/bib_ref] [bib_ref] An Oilseed Halophyte for Seawater Irrigation, Glenn [/bib_ref] [bib_ref] Bioprospecting and Genetic Engineering of Mangrove Genes to Enhance Salinity Tolerance in..., Das [/bib_ref] , especially those natural to estuaries that have constitutively uncoupled slow growth and stress response. These can grow almost equally fast with saline stress or without it-see fast growing species at river estuaries [bib_ref] Commercial cultivation and bioremediation potential of sugar kelp, Saccharina latissima, in Danish..., Marinho [/bib_ref] and halophytic perennial species already used as forage grass [bib_ref] Sustainable use of salt-degraded and abandoned farms for forage production using halophytic..., Rao [/bib_ref]. It is here that we may find the most important coupling factors. This will be made much easier with new technologies like massive sequences, ortholog searches and CRISPR approaches to test functions [bib_ref] Mutations in a subfamily of abscisic acid receptor genes promote rice growth..., Miao [/bib_ref]. In addition, a reverse strategy is re-emerging to make species that are naturally EVOL-Avoidance tolerant more productive (AGRI-Avoidance tolerant). Examples include Salicornia and Opuntia [bib_ref] Evaluating the growth performance of eleven Salicornia bigelovii populations under full strength..., Lyra [/bib_ref] while new ones like Quinoa are exploiting newer genomics technologies [bib_ref] Prospects for quinoa (Chenopodium Quinoa Willd.) improvement through biotechnology, Jellen [/bib_ref] offering renewed optimism. Various editorials and essays have discussed the discrepancy between the urgency to conserve water, the necessity to generate better, more stress-adapted crops and the perceived failure by plant molecular biology and gene engineering to deliver [bib_ref] Water: More crop per drop, Marris [/bib_ref]. The more-crop-per-drop catchphrase may have inspired but it also may have confused the underlying concepts and diverted from the optimal strategies about how to circumvent inevitable water shortages [fig_ref] Table 1: Relationships of tolerance nomenclature with growth, stress and survival [/fig_ref]. To produce more food with less water, or less suitable water, will depend on adequately supported collaborations between the many disciplines that must become involved. This has consequences for agencies, because the funding of new concepts has been almost invariably declined by programs where thinking out of the box is most often discouraged. Exceptions that charted new ground have been few in the past. These are projects such as the SALK insertion tagging project that has fundamentally altered and advanced our approach to genetics by providing the genomic information to explore the myriads of phenotypes from countless mutant screens [bib_ref] Empirical Analysis of Transcriptional Activity in the Arabidopsis Genome, Yamada [/bib_ref]. The TAIR project has collated information from many different approaches initially for Arabidopsis but now for other species as well [bib_ref] Implementation and Maintenance of a Model Organism Database for Arabidopsis thaliana, Weems [/bib_ref] [bib_ref] Plant Ontology (PO): A Controlled Vocabulary of Plant Structures and Growth Stages, Jaiswal [/bib_ref]. Of similar foresight and impact was the 1001 genome project [bib_ref] The 1001 Genomes Project for Arabidopsis thaliana, Weigel [/bib_ref]. The project continues to provide the genome sequences and phenotypes from ecotypic, evolutionary imprinted variation in the genus Arabidopsis. Importantly, it has revealed both genomic and phenotypic information derived from natural variation and will allow us to expand our knowledge of the genetic details of the solutions that survived evolutionary scrutiny. We will obtain data for many genotype/phenotype comparisons that have been missing, causing a proliferation of hypotheses but perhaps still not having the benefit of sufficient data from diverse and important species. Much more creative effort is needed to provide useful algorithms that can mine these data, especially to identify orthologs and paralogs of genes that have functions known in model species but not in important crops. Hopefully, the evolution oriented stress response genetic programs will be exchanged with an enhanced productivity under stress genetic program. If we are to use genetic manipulation of our crops to produce or yield more in environments that are even moderately stressful, we must finally recognize that plant responses to stress almost always are not, at least at initially, because of any resource limitation. We think that we can succeed in improving yield under moderately stressful conditions by removing the premature alarms that trigger growth reduction and other unproductive responses that are certainly polygenic and redundantly programmed in plants that have assured their survival in evolution. But, we will need to keep a stronger, delayed developmental response like meristem (bud) dormancy that occurs in summer dormancy [bib_ref] ABA receptor PYL9 promotes drought resistance and leaf senescence, Zhao [/bib_ref] , or use resource management to rescue them, or both [bib_ref] ABA receptor PYL9 promotes drought resistance and leaf senescence, Zhao [/bib_ref] [bib_ref] Control of Plant Water Use by ABA Induction of Senescence and Dormancy:..., Zhao [/bib_ref] under extreme stress conditions. [fig] Funding: This work was supported by grants from the National Research Foundation of Korea (NRF) funded by the Korean Government (MSIP No. 2016R1A2A1A05004931) and Global Research Lab (2017K1A1A2013146). [/fig] [table] Table 1: Relationships of tolerance nomenclature with growth, stress and survival. [/table]

Inference of Cross-Level Interaction between Genes and Contextual Factors in a Matched Case-Control Metabolic Syndrome Study: A Bayesian Approach Genes, environment, and the interaction between them are each known to play an important role in the risk for developing complex diseases such as metabolic syndrome. For environmental factors, most studies focused on the measurements observed at the individual level, and therefore can only consider the gene-environment interaction at the same individual scale. Indeed the group-level (called contextual) environmental variables, such as community factors and the degree of local area development, may modify the genetic effect as well. To examine such cross-level interaction between genes and contextual factors, a flexible statistical model quantifying the variability of the genetic effects across different categories of the contextual variable is in need. With a Bayesian generalized linear mixed-effects model with an unconditional likelihood, we investigate whether the individual genetic effect is modified by the group-level residential environment factor in a matched case-control metabolic syndrome study. Such cross-level interaction is evaluated by examining the heterogeneity in allelic effects under various contextual categories, based on posterior samples from Markov chain Monte Carlo methods. The Bayesian analysis indicates that the effect of rs1801282 on metabolic syndrome development is modified by the contextual environmental factor. That is, even among individuals with the same genetic component of PPARG_Pro12Ala, living in a residential area with low availability of exercise facilities may result in higher risk. The modification of the grouplevel environment factors on the individual genetic attributes can be essential, and this Bayesian model is able to provide a quantitative assessment for such cross-level interaction. The Bayesian inference based on the full likelihood is flexible with any phenotype, and easy to implement computationally. This model has a wide applicability and may help unravel the complexity in development of complex diseases. Citation: Wang S-H, Chen WJ, Chuang L-M, Hsiao P-C, Liu P-H, et al. (2013) Inference of Cross-Level Interaction between Genes and Contextual Factors in a Matched Case-Control Metabolic Syndrome Study: A Bayesian Approach. PLoS ONE 8(2): e56693. # Introduction Both genetic and environmental factors play an important role in the development of many diseases, often interacting in ways that elevate disease risk, especially in complex diseases such as obesity, cardiovascular disease, and psychopathology [bib_ref] Gene-environment interaction and obesity, Qi [/bib_ref] [bib_ref] Gene-environment interaction and oxidative stress in cardiovascular disease, Stephens [/bib_ref] [bib_ref] The contribution of gene-environment interaction to psychopathology, Thapar [/bib_ref]. To investigate genetic and environmental (GE) interactions [bib_ref] Gene-environment interactions for complex traits: definitions, methodological requirements and challenges, Dempfle [/bib_ref] [bib_ref] Interactions in epidemiology: relevance, identification, and estimation, Greenland [/bib_ref] [bib_ref] Biological implications of gene-environment interaction, Rutter [/bib_ref] , many studies have focused on environmental factors measured at the individual level, such as individual demographic information or lifestyle. It is possible, however, that group-level environmental factors, usually called contextual variables, also contribute to such risk. For instance, community disparity in medical resources may affect the availability and quality of treatment, and therefore living in a deprived area may result in a somewhat modified effect on the health risk incurred by individuals who engage in unhealthy behaviors such as smoking [bib_ref] Ecological effects in multi-level studies, Blakely [/bib_ref]. This contextual exposure affects the individual health status of each member of the group, and is therefore called the contextual effect [bib_ref] A glossary for multilevel analysis, Roux [/bib_ref] [bib_ref] Ecologic versus individual-level sources of bias in ecologic estimates of contextual health..., Greenland [/bib_ref]. Difficulties arise when it is of interest to model the cross-level interaction between the group and individual variables, especially with data of nested structures. Traditional logistic regression models with health status as the unit of analysis can include both contextual variables and genes as the covariates, but this approach is not appropriate for nested data where individuals within the same group or under the same category may share a certain degree of similarity in risk profile [bib_ref] A glossary for multilevel analysis, Roux [/bib_ref] , not to mention when this similarity is group-or category-specific. Even conditional logistic regression models are not flexible enough to account for the nested structure of heterogeneity in such a study design. Other issues concern the evaluation and computation of the cross-level interaction. These issues differ from what has been considered in most current research where GE interaction is evaluated at the individual level based on a case-control study design. For crosslevel GE interaction in clustered data, one immediate computational complexity involves the existence of multiple sources of variation. For example, homogeneity among individuals in the same group may produce dependence [bib_ref] A glossary for multilevel analysis, Roux [/bib_ref] , whereas the corresponding matched controls for any case may induce further correlation. In addition, the genetic effects may be modified by group-level exposures and may, therefore, correlate with each other. These nested sources of variability hamper the model construction and parameter inference. Several research groups adopted Bayesian or non-Bayesian hierarchical models to alleviate the problem of nested variability [bib_ref] Source partitioning using stable isotopes: coping with too much variation, Parnell [/bib_ref] [bib_ref] Quantifying interand intra-population niche variability using hierarchical bayesian stable isotope mixing models, Semmens [/bib_ref] , but these models are not for matched case-control studies and not for group-level GE interaction [bib_ref] Hierarchical modeling of gene-environment interactions: estimating NAT2 genotype-specific dietary effects on adenomatous..., Aragaki [/bib_ref] [bib_ref] Using hierarchical modeling in genetic association studies with multiple markers: application to..., Hung [/bib_ref] [bib_ref] Bayesian mixture modeling of geneenvironment and gene-gene interactions, Wakefield [/bib_ref] [bib_ref] A flexible Bayesian model for studying gene-environment interaction, Yu [/bib_ref]. To accommodate the nested structure of variation in both the individual-and group-level variables, and to model directly the interaction between individual genetic factors and group-specific variables, Bayesian hierarchical models with random effects can be formulated for inference. By employing hierarchical models with mixed effects, inference of these multiple variance components becomes feasible and may help elucidate whether cross-level interaction does exist, and/or quantify the strength of this interaction from a probabilistic perspective based on posterior distributions. This research was motivated by a matched case-control study of metabolic syndrome [bib_ref] Multilevel analysis of habitual physical activity and metabolic syndrome in Northern Taiwan, Wang [/bib_ref] , a condition which is characterized by a group of metabolic conditions and risk factors. In that metabolic disorder study, the effects of community exercise facility as well as five candidate SNPs (rs1801282, rs7799039, rs12535708, rs822390, and rs182052) were investigated. These factors were selected based on previous publications of association studies linking them to obesity and body mass index (BMI). Many studies have recognized the association between metabolic disorder and both diabetes and coronary heart disease [bib_ref] The metabolic syndrome, Eckel [/bib_ref] [bib_ref] Global and societal implications of the diabetes epidemic, Zimmet [/bib_ref] , while a few studies have shown that both genes and contextual variables, such as community disadvantage, income inequality, and factors affecting community development [bib_ref] Estimating neighborhood health effects: the challenges of causal inference in a complex..., Roux [/bib_ref] [bib_ref] Population effects on individual systolic blood pressure: a multilevel analysis of the..., Merlo [/bib_ref] [bib_ref] A multilevel analysis of race, community disadvantage, and body mass index among..., Robert [/bib_ref] [bib_ref] Body mass index in urban Canada: neighborhood and metropolitan area effects, Ross [/bib_ref] , associate strongly with the risk of metabolic syndrome. In a densely populated country such as Taiwan where the boundary between residential areas and commercial districts is blurred, an individual household may not have enough space for exercise. Even at the community level, there may not be sufficient exercise facilities such as gymnasiums, track fields, parks, or playgrounds for residents to use. This may restrict an individual's accessibility to an exercise facility and influence one's degree of physical inactivity. Although such grouplevel environmental variables can exert an effect on the risk of metabolic syndrome in addition to the effect exerted by individual environmental factors [bib_ref] Estimating neighborhood health effects: the challenges of causal inference in a complex..., Roux [/bib_ref] [bib_ref] Population effects on individual systolic blood pressure: a multilevel analysis of the..., Merlo [/bib_ref] [bib_ref] A multilevel analysis of race, community disadvantage, and body mass index among..., Robert [/bib_ref] [bib_ref] Body mass index in urban Canada: neighborhood and metropolitan area effects, Ross [/bib_ref] [bib_ref] Physical activity is negatively associated with the metabolic syndrome in the elderly, Bianchi [/bib_ref] , it is not certain whether the contextual variable will interact with the genetic effect in the same manner as the individual environmental variable does [bib_ref] Influence of Pro12Ala peroxisome proliferator-activated receptor gamma2 polymorphism on glucose response to..., Adamo [/bib_ref] [bib_ref] Does peroxisome proliferator-activated receptor gamma genotype (Pro12ala) modify the association of physical..., Franks [/bib_ref] [bib_ref] PPARgamma gene polymorphism is associated with exercise-mediated changes of insulin resistance in..., Kahara [/bib_ref] [bib_ref] Endurance training-induced changes in the insulin response to oral glucose are associated..., Weiss [/bib_ref] [bib_ref] Eating, exercise, and ''thrifty'' genotypes: connecting the dots toward an evolutionary understanding..., Chakravarthy [/bib_ref] [bib_ref] Physical activity attenuates the genetic predisposition to obesity in 20,000 men and..., Li [/bib_ref]. In the rest of this article, we reserve the abbreviation GE for the cross-level interaction at the contextual-level, and Ge for the geneenvironment interaction at the individual-level. The purpose of this paper is to explore the cross-level GE interaction in this matched case-control study via a Bayesian hierarchical model with an unconditional likelihood approach. We illustrate our approach with the aforementioned study concerning five candidate SNPs and GE which involves a contextual variable measuring the availability of exercise facilities. The posterior samples of the parameters of interest, the group-specific random effects and the variance among them, will be used to examine the existence of cross-level GE interaction. # Methods ## Motivating example: the metabolic syndrome study This study was conducted by Wang and colleagues in Tao-Yuan County in Taiwan in 2004 [bib_ref] Multilevel analysis of habitual physical activity and metabolic syndrome in Northern Taiwan, Wang [/bib_ref]. Here we briefly summarize the collection procedures of samples, while other details are referred to their original paper. This study recruited 6463 community residents aged greater than 40 years old in an adult health check-up program, and collected a questionnaire and blood sample from each participant. Among the recruited subjects, 1263 were classified with metabolic syndrome based on US National Cholesterol Education Program Adult Treatment Panel III (NCEP ATP III) criteria, with BMI replacing waist measurements. That is, an individual was considered to have metabolic disorder if three or more of the following criteria were met: BMI $ 27 kg/m 2 ; triglycerides $ 150 mg/dl; HDL ,40 mg/dL in men and ,50 mg/dL in women; blood pressure $ 130/85 mmHg or medication for hypertension currently taken; and fasting glucose $ 110 mg/dL or medication for diabetes currently taken. Next, two hundred-and sixty-eight cases were randomly selected and matched with one or two controls by sex, age, educational level, ethnicity, and a contextual variable measuring the availability of exercise facilities. Exercise facilities included swimming pools, parks, bowling alleys, golf courses, fitness centers, and activity centers for the elderly. For each residential area, the exercise facility density was defined as the number of facilities divided by the number of residents, and this density was categorized into four different levels of availability in exercise facility: level I for very low availability, II for low, III for medium, and IV for high. The final data set for analysis contained 268 metabolic syndrome cases and 322 individually matched controls. The terms ''cases'' and ''controls'' here do not intend to indicate that metabolic syndrome is a disease. It is indeed a group of metabolic conditions that associate strongly with increased risk of cardiovascular disease and diabetes. In other words, metabolic syndrome is a clustering of risk factors. Discussions and debates about its definition, inclusion criteria, and usefulness in clinical practice have drawn much attention [bib_ref] Metabolic syndrome: from epidemiology to systems biology, Lusis [/bib_ref] [bib_ref] The metabolic syndrome: useful concept or clinical tool? Report of a WHO..., Simmons [/bib_ref]. Here we call these identified individuals ''cases'' simply for ease of presentation. ## Candidate genes and odds ratios Three candidate genes, the adiponectin gene APM1 on chromosome 3q27, the peroxisome proliferator-activated receptor c gene PPARG on 3p25, and the leptin gene LEP on 7q31, were selected as candidate susceptibility genes for human obesity and Type 2 diabetes [bib_ref] Genetic variation in the gene encoding adiponectin is associated with an increased..., Hara [/bib_ref] [bib_ref] Pro12Ala sequence variant of the PPARG gene is associated with postprandial hypertriglyceridemia..., Cardona [/bib_ref] [bib_ref] Genetic architecture of the APM1 gene and its influence on adiponectin plasma..., Heid [/bib_ref] , and five SNPs (rs1801282 on PPARG, rs7799039 and rs12535708 on LEP, and rs822390 and rs182052 on APM1, respectively) were genotyped for this study. Based on preliminary analysis of genotype-specific odds ratios, we employed an additive model for each of the three SNPs, rs1801282, rs182052 and rs822390, and a recessive model for the other two. [fig_ref] Table 1: The observed genotype counts of metabolic cases [/fig_ref] lists the distributions of the SNP genotypes as well as their genetic effects in terms of conditional odds ratios of GE interaction under each category of the contextual variable. It can be observed that the category-specific genetic effect seems to vary across the four levels, indicating a possible GE interaction. Furthermore, because linkage disequilibrium (LD) was observed between rs12535708 and rs7799039 (D' = 0.89, R-square = 0.46), the SNP-SNP interaction between them is considered in later analysis. In addition, because a protective effect in carriers of the Ala allele of rs1801282 (PPARG_Pro12Ala) against Type 2 diabetes and its association with lower body mass [bib_ref] The common PPARgamma Pro12Ala polymorphism is associated with decreased risk of type..., Altshuler [/bib_ref] [bib_ref] A Pro12Ala substitution in PPARgamma2 associated with decreased receptor activity, lower body..., Deeb [/bib_ref] has not been consistently reported [bib_ref] Association of the Pro12Ala variant in the peroxisome proliferator-activated receptor-gamma2 gene with..., Beamer [/bib_ref] , possible environmental modifiers like exercise may be considered [bib_ref] PPAR gamma and human metabolic disease, Semple [/bib_ref]. Therefore, in the following we focus on the cross-level interaction between rs1801282 and exercise facility availability in the residential environment. ## Bayesian hierarchical model with unconditional likelihood To examine if the effect of the SNP rs1801282 varies among different categories of the contextual covariate, along with the SNP-SNP interaction between rs12535708 and rs7799039 in the same gene LEP on 7q31, we adopt a Bayesian generalized linear mixed-effects model (GLMM). This Bayesian GLMM assumes that each response variable Y ijk , the binary disease status (1 for case and 0 for control) of the k-th individual in the j-th pair of the i-th category, follows a Bernoulli distribution with the parameter p ijk indicating the probability of having metabolic syndrome, [formula] Y ijk Dp ijk *Bernoulli p ijk À Á , [/formula] where the probability of being diseased p ijk is associated in the logit scale with the 5 (SNP) genetic components S (1) ,:::,S (5) (following the same order and coding as in [fig_ref] Table 1: The observed genotype counts of metabolic cases [/fig_ref] and a SNP-SNP interaction S [bib_ref] The contribution of gene-environment interaction to psychopathology, Thapar [/bib_ref] , in addition to other covariates X , [formula] logit p ijk À Á~X 5 g~1 b (g) i |S (g) ijk zc|S (2,3) ijk zh|X ijk zb i zb ij : [/formula] The implication of each parameter is explained in the following paragraphs, and outlined in the supporting information [fig_ref] Table 1: The observed genotype counts of metabolic cases [/fig_ref]. All considered explanatory variables are summarized as well. ## Cross-level ge interaction The covariate S (g) ijk indicates the coding of the SNP g for the k-th individual in the j-th pair of the i-th category. Its associating parameter b (g) i represents the SNP effect under the i-th category for the SNP g. For instance, the evaluation and comparison of the 4 (i~1,:: [formula] :,4) posterior distributions of b (1) 1 , b (1) 2 , b (1) 3 , and b (1) 4 [/formula] can reveal if the effect of the first SNP (g = 1) rs1801282 varies among the 4 different categories of the contextual covariate. Therefore, the variance var(b (g) i ) = s 2 (g) can model directly the heterogeneity in SNP effects among 4 categories for each SNP g. This variance measures the degree to which the genetic effect differs across different categories of the contextual variable. This is thus the parameter of interest when the cross-level GE interaction is to be evaluated. ## Snp-snp interaction and other effects The SNP-SNP interaction between the second SNP rs7799039 and the third SNP rs12535708 in the same gene is measured by c. The parameter h represents the effect of other explanatory variable X . The random intercept b i stands for the categoryspecific effect accounting for the sampling variability among residential areas. The second random intercept b ij is for the pairor cluster-specific random effect. This b ij represents the matching relation within each pair. In other words, subjects in the same pair share the same baseline characteristics and pairs are taken to be independent. The random coefficients are all assumed to follow normal distributions with a large variance representing vague information. A complete model specification including all prior distributions is detailed in the supporting information (Text S1). ## Computation The analysis was conducted based on posterior samples obtained from Markov chain Monte Carlo (MCMC) methods with WinBUGS 1.4.3. The chain contained 50,000 iterations following a burn-in of 5,000 samples to reduce the impact from initial values and the final posterior sample was derived at the thinning rate of 10 to reduce dependence among posterior points. The code in WinBUGS is detailed in the supporting information (Text S2). # Results ## Cross-level gene-environment interaction To investigate the existence of the cross-level GE interaction, we list in the posterior means and standard deviations of the ith category-specific genetic effects b (g) i and variance s 2 (g) for each SNP g, and display the posterior distributions of the b (g) i for SNPs g = 1, …, 5 in [fig_ref] Figure 1: The posterior distributions of b [/fig_ref] -(e). From , it is apparent that for rs1801282 (g = 1) the four category-specific means b (g) i ,i~1,:::,4, differ the most. This implies a relatively large variation of genetic effect across the 4 areas. The same pattern can be observed in the corresponding four posterior distributions of b (g) i in [fig_ref] Figure 1: The posterior distributions of b [/fig_ref]. The curves in [fig_ref] Figure 1: The posterior distributions of b [/fig_ref] are more scattered, as compared with the other four plots [fig_ref] Figure 1: The posterior distributions of b [/fig_ref] , (c), (d) and (e)), indicating substantially larger heterogeneity in the genetic effect of rs1801282 across the four categories. Another supporting evidence lies in the inference of s 2 (g) for g = 1, …, 5. The last column in lists the posterior means of s 2 (g) where the first SNP rs1801282 has the largest mean (1.31), implying again a larger heterogeneity. [fig_ref] Figure 2: The posterior distributions of s 2 [/fig_ref] displays their corresponding distributions. Indeed, s 2 (g) for rs1801282 is the most right skewed, representing an observable difference among areas. Further evidence manifests through the evaluation of the difference in risk across the four categories. For instance, for SNP g = 1, if the posterior probabilities of the risk, P(b (1) 1 .0|y), P [bib_ref] Gene-environment interaction and obesity, Qi [/bib_ref] 2 .0| y), P(b (1) 3 .0| y), and P(b (1) 4 .0| y), deviate from each other, then this SNP has divergent effects among areas. Again, for rs1801282, the corresponding four posterior probabilities contain the largest degree of variation . These findings consistently imply the existence of the cross-level interaction between this SNP and the contextual variable. To be specific, the analyses suggest that, even among those residents with the same genotype C/G for rs1801282, living in an area with low availability of exercise facilities (category I) leads to a higher risk of being metabolically diseased. In other words, the G allele (also called the Ala allele in PPARG_Pro12Ala) is considered protective under categories II-IV, but not under category I. In fact, we have conducted Bayesian model selection using the deviance information criterion (DIC) to compare the model with GE interaction and without (assuming b (1) 1~b (1) 2~b (1) 3~b (1) 4 ). The former model containing the cross-level GE interaction term was selected. This pattern of interaction, however, is not shown for other SNPs. For example, the second SNP (rs7799039) in [fig_ref] Figure 1: The posterior distributions of b [/fig_ref] and the fourth SNP (rs822390) in [fig_ref] Figure 1: The posterior distributions of b [/fig_ref] reveal only mild variation; while the third SNP (rs12535708) in [fig_ref] Figure 1: The posterior distributions of b [/fig_ref] and the ## Snp-snp interaction To evaluate the existence of interaction between SNPs, we examine the posterior distribution of c for the SNP-SNP interaction between rs7799039 and rs12535708. Notice that the posterior distribution of c is left-skewed with a mean of 210.70 and a 95% credible interval (221.8, 23.5), indicating a protective effect when these two SNPs simultaneously carry the G/G and A/ A genotypes, respectively and [fig_ref] Figure 1: The posterior distributions of b [/fig_ref]. The overall aggregation of effects, however, shrinks toward zero when accounting for both single-SNP effects. For instance, in category I, the sum of the average odds ratios for genotype G/G for rs7799039, A/A for rs12535708, and their interaction is about exp(0.25+9.90-10.70) = exp(20.55), resulting in an odds ratio of approximately 0.58( = exp(20.55)). In other words, the interaction diminishes the two strong first-order genetic effects. # Discussion The research reported here was motivated by a matched casecontrol metabolic syndrome study where interest lay in the interaction between genes and contextual variables, as well as SNP-SNP interaction. Such cross-level GE interaction can be as important as the individual level Ge interaction, and thus its influence should not be overlooked. The statistical inference of GE, however, is not straightforward. We adopted here a Bayesian model with random effects to investigate the genetic and environmental interactions by examining the extent to which the allelic effects differ across various categories of the contextual variable. Though many studies have been proposed to test the interaction between genes and environment at the individual level, this Bayesian approach, to the best of our knowledge, is the first Bayesian analysis to infer the cross-level interaction between individual genetic and contextual residential area effects. The advantages of employing a Bayesian unconditional likelihood model are its flexibility in model construction for complex models; the straightforwardness in deriving the posterior distribution of the variance s 2 g to assess cross-level interaction; and the ability to estimate multidimensional parameters simultaneously. ## Biological implications The results indicated that the effect of rs1801282 is likely to be modified by the contextual factor in the sense that individuals with the mutant genotype have higher risk, particularly when living in an area with low availability of exercise facilities. On the basis of the large posterior variability in genetic effects across areas, we conclude that this effect is modulated by the residential environmental factor. Similar observations have been previously reported in other studies where the rs1801282 SNP was shown to exacerbate the negative effect of poor individual exercise habits [bib_ref] Influence of Pro12Ala peroxisome proliferator-activated receptor gamma2 polymorphism on glucose response to..., Adamo [/bib_ref] [bib_ref] Does peroxisome proliferator-activated receptor gamma genotype (Pro12ala) modify the association of physical..., Franks [/bib_ref] [bib_ref] PPARgamma gene polymorphism is associated with exercise-mediated changes of insulin resistance in..., Kahara [/bib_ref] [bib_ref] Endurance training-induced changes in the insulin response to oral glucose are associated..., Weiss [/bib_ref]. Others have observed that physical inactivity and sedentary lifestyle may interfere with optimized expression of the ''thrifty'' genes [bib_ref] Eating, exercise, and ''thrifty'' genotypes: connecting the dots toward an evolutionary understanding..., Chakravarthy [/bib_ref] and that the availability of recreational resources can relate directly to an individual's physical activity level [bib_ref] Availability of recreational resources and physical activity in adults, Diez Roux [/bib_ref]. These findings are in agreement with our finding that the availability of recreational resources will interact with the effect of rs1801282 on metabolic syndrome in such a way that living in areas with lower exercise facility availability will increase the risk of disease, particularly for individuals carrying the mutant type of rs1801282. The model we have introduced here for the matched design can accommodate SNP-SNP interaction as well. For instance, SNPs rs12535708 and rs7799039 locate closely in the gene LEP on chromosome 7q31, where rs7799039 is in the 59 regulatory promoter region and rs12535708 is at the transcription-factor binding site. The strong LD between these two SNPs has been documented earlier [bib_ref] Common variants in the 59 region of the leptin gene are associated..., Jiang [/bib_ref] , and the existence of their statistical interaction indicates that these two SNPs may be involved in similar functional pathways. An existing study reported that rs7799039 cannot add further information when other multimarker haplotypes containing rs12535708 have been included for analysis [bib_ref] Common variants in the 59 region of the leptin gene are associated..., Jiang [/bib_ref]. Our results replicate the finding that, when both SNPs are present in the model, rs12535708 exhibits strong association (the posterior modes of b (3) 1 *b (3) 4 are all around 10), while the influence of rs7799039 is mild (posterior modes of . In the upper half of the table, numbers in each row are the posterior means and standard deviations of the areaspecific genetic effects (b (g) i ) and variance (Var(b (g) i ) = s 2 (g) ) for each candidate SNP g. Posterior mean (se) . Numbers are P(b (g) i w0Dy), the posterior probability of (b (g) i w0), for the g-th SNP in the i-th category (area) under the unconditional model. are between 0 and 1). When accounting for both markers along with their interaction, the estimated odds ratios of being metabolic, for persons carrying genotype G/G for rs7799039 and A/A for rs12535708, are 0.57, 0.83, 1.02, and 1.54 for the four categories, respectively. Each of the resulting values implies the existence of SNP-SNP interaction, and the heterogeneity among the four odds ratios provides evidence for GE interaction. Further research about the functional polymorphism of these two SNPs, or haplotype analysis, would be worth pursuing to unravel their biological interplay. [formula] SNP Covariates b (g) 1 , I b (g) 2 , II b (g) 3 , III b (g) 4 , IV s 2 (g) S (1) , rs1801282(PPARG_Pro12Ala [/formula] Other Statistical Models for GE interaction 1. Bayesian conditional logistic regression model. If only some of the previously mentioned parameters are of interest, then their inference can be made under the Bayesian conditional logistic regression (BCLR) model assuming p ijk = e ijk = P k e ijk , and a log link formulation on e ijk log e ijk~X [formula] 5 g~1 b (g) i |S (g) ijk zc|S (2,3) ijk zh|X ijk , [/formula] where for each SNP g, b (g) iÑ Normal m (g) ,s 2 (g) ,i~1,:::,4. [fig_ref] Table S2: The formulation and parameter interpretation of the Bayesian conditional logistic regression model [/fig_ref] displays the model and parameters for comparison. [fig_ref] Table 4: The variance parameters represent the variability among areas for each SNP [/fig_ref] lists the posterior means and standard errors of s 2 (g) for g~1,:::,5. The posterior distributions of s 2 (g) under the Bayesian conditional logistic regression do not vary much from that under the Bayesian unconditional likelihood model. In addition, we have investigated the pattern based on a limited simulation study with only 10 replications, each containing five SNPs for 100 cases and 100 matched controls in every one of the 4 areas. The consistency in the inference of variance between the Bayesian unconditional and conditional models stays clearly. However, as compared with the previous unconditional model, the disadvantages are that one is unable to infer the sampling variability among areas or within each pair under this BCLR, and the computational difficulty is greatly elevated under this setting if the number of genetic markers is large. Our experience shows that even when the computation under BCLR reaches convergence, the computation time for BCLR is about 4.5 times that of the Bayesian full likelihood model, when computed with an Intel core i7-2620M (2.7/3.4GHz) dualcore processor. Although the BCLR model could be viewed as a workable alternative model in this case, its computational burden and lack of information on sampling variability within clusters hinder its use in such a cross-level GE study. The unconditional likelihood approach uses matched-pair specific random effects to account for the selection bias, while the conditional approach adopts matched-pair specific intercepts [bib_ref] Equivalence between conditional and random-effects likelihoods for pair-matched case-control studies, Rice [/bib_ref]. Another reason for favoring the Bayesian unconditional likelihood approach is that the full likelihood model can accommodate complex nested structures easily and intuitively, such as those involving matched cases and controls in the same category. These various sources of dependence can be modeled straightforwardly. 2. Models with cross-product terms. Other alternative for handling the cross-level GE interaction in a matched design is through the use of cross-product terms of dummy variables for areas and SNP covariates. Some possibilities are outlined in the supporting information (Text S3). We do not recommend however the use of these other models with cross-product terms, because the large number of coefficients of GE interaction often leads to failure in the MCMC computations. Other non-Bayesian multi-level models with the MQL option in MLwiN and glmmPQL in R are theoretically possible, but the computation for this matched case-control metabolic syndrome study collapses when it runs into a nonpositive definite matrix during the iteration procedure. ## Further notes Some further notes are worth mentioning. First, for the sensitivity analysis of the posterior inference, we have tried other prior distributions. For example, we have used different gamma distributions for the precision parameters (the inverse of the variance components), and obtained similar conclusions. Second, we have adopted a Bayesian hierarchical model with no GE interaction for rs1801282, and performed model selection with DIC. The model with the cross-level GE was selected, and therefore the posterior probability under the GE model was evaluated for inference. Last but not least, although Bayesian models may at first seem complicated with respect to their formulation and computation, the advancement of MCMC methods has greatly enhanced the feasibility of using Bayesian inference in daily practice. Several software applications, such as WinBUGS, R, and MLWin, are handy for carrying out Bayesian analysis. The model presented here for analyzing interaction between genes and contextual variables could achieve broad applicability with the aid of these statistical analysis tools. [fig_ref] Table 1: The observed genotype counts of metabolic cases [/fig_ref] The formulation and parameter interpretation of the unconditional likelihood Bayesian model. Text S1 Complete specification of the Bayesian model. Text S3 Model formulation and parameter interpretations with cross-product terms for cross-level interaction. ## Supporting information (DOCX) [fig] Figure 1: The posterior distributions of b (g) i (i = 1,…,4) for four categories are displayed in (a)-(e) for SNP g = 1, g = 2,…, g = 5, respectively, under the Bayesian unconditional likelihood model. doi:10.1371/journal.pone.0056693.g001 [/fig] [fig] Figure 2: The posterior distributions of s 2 (g) for g = 1, …, 5 SNP, respectively, under the Bayesian unconditional likelihood model. doi:10.1371/journal.pone.0056693.g002fifth SNP (rs182052) inFigure 1(e) show almost no difference in the 95% credible intervals across the four categories. In addition, the density functions of s 2 (1) , s 2 (2) , …, and s 2 (5) differ from each other, which implies different patterns in heterogeneity among the 5 SNPs. [/fig] [table] Table 1: The observed genotype counts of metabolic cases (cs) and controls (cn) under each category of exercise facility availability. [/table] [table] Table S2: The formulation and parameter interpretation of the Bayesian conditional logistic regression model. (DOCX) [/table] [table] Table 4: The variance parameters represent the variability among areas for each SNP. [/table]

Homocysteine Levels in Chronic Gastritis and Other Conditions: Relations to Incident Cardiovascular Disease and Dementia Background Homocysteine levels in circulation are determined by several factors and hyperhomocysteinemia is reportedly associated with cardiovascular diseases and dementia. The aim of this study is to determine the relation of chronic gastritis and other conditions to homocysteine levels and their relation to incident cardiovascular diseases and dementia. Methods An adult population-based cohort (N = 488) was screened for H. pylori infection, gastro-duodenitis (endoscopic biopsies), disease history, and lifestyle factors. Blood samples were analyzed for pepsinogen I and II (gastric function), vitamin B12, folate, homocysteine, and cystatin C (renal function). The methylenetetrahydrofolate reductase C677T polymorphism reportedly associated with hyperhomocysteinemia was analyzed by pyrosequencing. Incident cardiovascular diseases and dementia were monitored during a median follow-up interval of 10 years. Results At baseline, there was a positive relation of S-homocysteine to male gender, age, S-cystatin C, methylenetetrahydrofolate reductase 677TT genotype and atrophic gastritis. During follow-up, cardiovascular diseases occurred in 101/438 and dementia in 25/488 participants, respectively. Logistic regression analysis (adjusting for gender, age at baseline, follow-up interval, BMI, smoking, alcohol consumption, NSAID use, P-cholesterol, and P-triglycerides) showed an association of S-homocysteine higher than 14.5 lmol/l to cardiovascular diseases (OR 2.05 [95% c.i. 1.14-3.70]), but not to dementia overall. Conclusions Gender, age, vitamin B12, folate, renal function, atrophic gastritis and the methylenetetrahydrofolate 677TT genotype were significant determinants of homocysteine levels, which were positively related to incident cardiovascular diseases. # Introduction Homocysteine (Hcy) metabolism is dependent on several factors, such as S-adenosylmethionine (SAM), vitamin B12 as a cofactor, and methyltetrahydrofolate as the substrate. Deficiency of these or of other substances, as well as enzymatic defects, disturb the metabolism and may lead to hyperhomocysteinemia [bib_ref] Homocysteine metabolism, Selhub [/bib_ref]. Methylenetetrahydrofolate reductase (MTHFR) catalyzes the reduction of 5, 10-methylenetetrahydrofolate to 5-methyltetrahydrofolate, acting as methyl donor in the conversion of Hcy to methionine [bib_ref] Homocysteine metabolism, Selhub [/bib_ref]. The MTHFR single nucleotide polymorphism (SNP) C677T (SNPdb ID: rs1801133) is known to yield an enzyme with reduced activity, which leads to mild hyperhomocysteimia [bib_ref] A candidate genetic risk factor for vascular disease: a common mutation in..., Frosst [/bib_ref]. Hyperhomocysteinemia may cause damage to the intima of the vascular wall [bib_ref] Homocysteine, a risk factor for cardiovascular disease, Prasad [/bib_ref]. It has been linked to cardiovascular diseases, dementia, diabetes mellitus (DM), hypothyreosis, and renal disease [bib_ref] Homocysteine, a risk factor for cardiovascular disease, Prasad [/bib_ref] [bib_ref] Hyperhomocysteinaemiaa common finding in a psychogeriatric population, Nilsson [/bib_ref] [bib_ref] Elevated plasma homocysteine level is an independent predictor of coronary heart disease..., Soinio [/bib_ref] [bib_ref] Interaction of plasma homocysteine and thyroid hormone concentrations in the pathogenesis of..., Evrengul [/bib_ref] [bib_ref] Why is homocysteine elevated in renal failure and what can be expected..., Van Guldener [/bib_ref]. The levels of Hcy in circulation are inversely related to renal function [bib_ref] Serum cystatin C as an endogenous parameter of the renal function in..., Randers [/bib_ref]. The association between Hcy and cardiovascular diseases [bib_ref] Homocysteinemia as a risk factor for atherosclerosis: a review, Nehler [/bib_ref] [bib_ref] Homocysteine and cardiovascular disease: evidence on causality from a meta-analysis, Wald [/bib_ref] and dementia [bib_ref] Plasma homocysteine as a risk factor for dementia and Alzheimer's disease, Seshadri [/bib_ref] [bib_ref] Folate deficiency is a common finding in psychogeriatric patients, Hultberg [/bib_ref] has been shown in several studies, but is controversial [bib_ref] The relation between Helicobacter pylori and atherosclerosis cannot be explained by a..., Bloemenkamp [/bib_ref] [bib_ref] Homocyst(e)ine and cardiovascular disease: a systematic review of the evidence with special..., Ford [/bib_ref]. The cut-off value for the Hcy concentration used to define hyperhomocysteinemia is poorly defined. Levels around 14-15 lmol/l have been used [bib_ref] Homocysteine, a risk factor for cardiovascular disease, Prasad [/bib_ref] [bib_ref] Elevated plasma homocysteine level is an independent predictor of coronary heart disease..., Soinio [/bib_ref] [bib_ref] Why is homocysteine elevated in renal failure and what can be expected..., Van Guldener [/bib_ref] [bib_ref] Homocysteine and cardiovascular disease: evidence on causality from a meta-analysis, Wald [/bib_ref] [bib_ref] Plasma homocysteine as a risk factor for dementia and Alzheimer's disease, Seshadri [/bib_ref] [bib_ref] Folate deficiency is a common finding in psychogeriatric patients, Hultberg [/bib_ref] [bib_ref] Facts and recommendations about total homocysteine determinations: an expert opinion, Refsum [/bib_ref]. Deficiencies of vitamin B12 and folate are common etiologies of hyperhomocysteinemia. Untreated celiac disease is frequently associated with folate deficiency and, to a lower rate, with vitamin B12 deficiency [bib_ref] Prevalence of hyperhomocysteinemia in adult gluten-sensitive enteropathy at diagnosis: role of B12,..., Saibeni [/bib_ref] [bib_ref] Homocysteine and related Bvitamin status in coeliac disease: effects of gluten exclusion..., Dickey [/bib_ref]. A frequent cause of vitamin B12 deficiency is atrophy of the gastric corpus mucosa, which leads to reduced intrinsic factor secretion, i.e., latent or overt pernicious anemia. Helicobacter pylori (H. pylori) infection per se, irrespective of atrophy of the gastric mucosa, was found to be associated with reduced vitamin B12 in circulation [bib_ref] An association between Helicobacter pylori infection and serum vitamin B12 levels in..., Shuval-Sudai [/bib_ref] [bib_ref] Relation of Helicobacter pylori infection to plasma vitamin B12, folic acid, and..., Tamura [/bib_ref]. However, a meta-analysis showed no association of H. pylori infection to coronary heart disease [bib_ref] Risk factors for coronary heart disease and infection with Helicobacter pylori: meta-analysis..., Danesh [/bib_ref]. The objective of this 10-year follow-up study of a population-based cohort was to investigate what impact H. pylori infection, chronic gastritis, MTHFR C677T SNP, renal dysfunction, and some other conditions have on variations in Hcy levels in circulation and, furthermore, to explore the relations of Hcy levels to subsequent development of cardiovascular diseases and dementia. # Methods ## Study population and follow-up The study was performed in accordance with the Helsinki Declaration and was approved by the local ethics committee. Informed written consent was obtained from all participants. The prevalence of gastritis, duodenitis, and H. pylori infection in this cohort have been published [bib_ref] Prevalence of gastroduodenitis and Helicobacter pylori infection in a general population sample:..., Borch [/bib_ref] [bib_ref] Prevalence of coeliac disease and relations to Helicobacter pylori infection and duodenitits..., Borch [/bib_ref]. Fivehundred-and-one randomly selected subjects (age 35-85 years and equal number of women and men at selection) in the general population underwent esophago-gastroduodenoscopy (EGD) with biopsy. A complete histomorphological examination of biopsies from both the gastric corpus and antrum was obtained in 488 participants. These 488 participants make up the cohort studied here and which comprises 225 women and 263 men with median ages of 62 (min 38-max 85) years and 60 (37-82) years, respectively, at baseline. Histomorphological examination of biopsies from the second part of the duodenum was achieved in 474 participants. At baseline, the participants received a self-administered questionnaire exploring details about previous and current diseases, body weight, and height, current smoking, use of alcohol, and current medications including the use of aspirin (including low-dose) or other potentially ulcerogenic NSAIDs. Eight years (minimum) after baseline, the same questionnaire was distributed. In addition, in-hospital diagnoses recorded before and after baseline including causes of death among deceased participants were extracted from local patient files (including those of the Memory Reception, Department of Geriatric Medicine) as well as from the records of Statistics Sweden (SCB) and the Swedish National Board of Health and Welfare. Nationwide reporting to the latter goes back to 1964 (full cover by 1987). Files were screened up to 2005. In case of incident cardiovascular diseases, one diagnosis was assigned to each participant. Myocardial infarction was given the highest priority. The priority of other diagnoses in descending order was: ischemic heart disease (angina pectoris), aortic or other arterial aneurysm, transient ischemic attack (TIA) or stroke, and cardiac arrhythmia and/or heart failure; all without a registered diagnosis of myocardial infarction. Hypertension as the sole diagnosis was not included since primarily this diagnosis is set in primary health care, and is probably underreported and latent/subclinical in the general population. In the total cohort of 488 participants, 50 had a history of a cardiovascular disease at baseline (myocardial infarction ten, angina pectoris 17, TIA/stroke four, and cardiac arrhythmia and/or heart failure 19). None had known dementia, since that precluded participation. Ninety-nine of 485 responders were current smokers, 78 of 476 responders used spirits and/or wine every week, and 81 of 476 responders used aspirin or other NSAID every week. There were 438 participants (207 women and 231 men, age 59.4 (36.8-84.9) years) with no history of a cardiovascular disease or dementia at baseline (smokers 93 of 435 responders; use of alcohol every week 71 of 429 responders; use of aspirin or other NSAID every week 60 of 427 responders). At baseline, 14 participants had DM (one type 1 and 13 type 2). During the follow-up period, four developed type 2 DM. Twenty-six participants had thyroid disease, classified as hypothyroid disease in six. During follow-up, one participant underwent surgery for non-specified thyroid disease. Three participants had a history of renal disease (one operated and one non-operated for cystic disease, and one operated for renal carcinoma). During follow-up, nephritis was diagnosed in one and renal carcinoma in two participants. Histological Examinations and H. pylori Status Routinely prepared biopsy sections (three biopsies from the corpus, antrum and second part of the duodenum) were stained with hematoxylin-eosin, Alcian blue-periodic acid-Schiff, and Giemsa stain as previously described [bib_ref] Prevalence of gastroduodenitis and Helicobacter pylori infection in a general population sample:..., Borch [/bib_ref]. Gastritis was classified according to the Sydney system, i.e., severity, topographic distribution and the occurrence of H. pylori [bib_ref] Classification and grading of gastritis. The updated Sydney system, Dixon [/bib_ref]. Considering glandular atrophy of the gastric corpus mucosa, moderate to severe atrophy was taken into account in the present study. Atrophy of the duodenal mucosa was histologically classified according to Alexander, and in the present study it was defined as grade III or grade IV. An experienced pathologist (FÁP) who had no information about other data performed the microscopic examinations. H. pylori status was classified as previously described, i.e., positive if at least one of serology, histology and urease-test on fresh biopsy specimens [bib_ref] Prevalence of gastroduodenitis and Helicobacter pylori infection in a general population sample:..., Borch [/bib_ref] indicated infection. The study was non-interventional regarding H. pylori infection. However, the infection was treated at baseline in 15 participants who at screening were shown to have H. pylori associated ulcer disease [bib_ref] Prevalence of gastroduodenitis and Helicobacter pylori infection in a general population sample:..., Borch [/bib_ref]. None of these had atrophic gastritis. ## Blood analyses At baseline, blood samples were collected in a fasting state immediately preceding EGD. The samples were stored at -80°C. Sera were analyzed (Biohit Diagnostics, Helsinki, Finland) for vitamin B12 (pmol/l), folate (nmol/l) and Hcy (lmol/l) concentrations with the Immulite Ò 1000 method (DPC, Los Angeles, CA, USA) according to the manufacturer's instructions. Samples were run at the same time to minimize assay variability. Pepsinogens in serum were measured with a sandwich enzyme immunoassay (ELISA) utilizing a pepsinogen I (PGI) and pepsinogen II (PGII) specific capture antibody, respectively, and a horseradish peroxidase (HRP) detection antibody (from GastroPanel Ò , Biohit Diagnostics, Helsinki, Finland). There is no cross reactivity between the two assays. The reference interval is 30.0-120.0 lg/l for PGI, 3.0-10.0 lg/l for PGII and 3.0-20.0 for the ratio PGI/PGII. Values of PGI/PGII lower than 3.0 were considered as indicative of significant atrophy of the gastric corpus mucosa. P-total cholesterol (mmol/l) and P-triglycerides (mmol/l) were measured with the Hitachi 717 (Boehringer Mannheim Scandinavia AB, Stockholm, Sweden) using specific reagent 1. Cystatin C in serum was analyzed with Advia 1650 (Bayer AB, Gothenburg, Sweden) using reagent DakoCystatin C at the Dept. of Clinical Chemistry, University Hospital, Linköping, Sweden (reference interval for ages 1-50 years : 0.7-1.21 mg/l and for ages[50 years: 0.84-1.55 mg/l). ## Mthfr snp 677 genotype analysis DNA from blood samples (whole blood, plasma, or serum) was extracted, amplified and quality assessed as previously described [bib_ref] Multiple strand displacement amplification of DNA isolated from human archival plasma/serum: identification..., Sun [/bib_ref]. Primers for MTHFR C677T (rs1801133) pyrosequencing analysis were designed based on the available MTHFR genomic sequence (GenBank accession number NT_021937, region: 6388104-6400540) using the PSQ Assay design software (Biotage, Uppsala, Sweden). The primers Biotin-MTHFR677.se (ccgaagcagggagctttga), MTHFR677.as (caagtgatgcccatgtcg), and sequencing primer Seq-MTHFR677 (cgtgatgatgaaatcg) were used. PCR products were obtained using the HotStarTaq Plus Mastermix kit (Qiagen, Hilden, Germany) following the manufacturer's instructions, using 5 pmol of each primer, and the PCR program: 95°C 5 min; 35 cycles of 95°C 30 s, 60°C 30 s, 72°C 1 min; final extension 72°C 10 min. The PCR products were analyzed by pyrosequencing as previously described [bib_ref] Multiple strand displacement amplification of DNA isolated from human archival plasma/serum: identification..., Sun [/bib_ref]. ## Statistical analyses Numerical data were summarized as median (min-max) unit. The Mann-Whitney U-test was used for comparisons of continuous data between groups while Fisher's exact test was used for dichotomous data. Correlations were tested with Spearman's rank correlation test. A general linear model (GLM) was used to identify factors with impact on S-Hcy concentrations. Logistic regression analysis was performed in which incident cardiovascular diseases overall or dementia overall were included as the dependent variable. Goodness-of-fit was tested with the Hosmer-Lemeshow's method. Coefficients in GLM and odds ratios (OR) in logistic regression are given with [95% confidence intervals]. In every analysis, a two-sided P-value of \0.05 was regarded as significant. # Results At baseline, 246 participants had histologically normal gastric mucosa with negative H. pylori status, 209 had chronic gastritis with positive H. pylori status, and 33 had chronic gastritis with negative H. pylori status. Atrophy of the corpus mucosa was diagnosed histomorphologically in 21 of the 488 participants (15 with positive H. pylori status). Twenty of 474 examined participants had a subnormal PGI/PGII value (\3.0) and 28 of 475 participants had atrophy either according to histological examination and/or the PGI/PGII value. Villous atrophy of the duodenal mucosa was diagnosed in nine participants. The distribution of the MTHFR 677 genotype among 483 successfully analyzed participants was 50.1% CC, 38.5% TC, and 11.4% TT while an S-Cystatin C concentration above the reference range was present in 18 of 473 participants examined. Spearman analysis showed a weak although significant (P \ 0.001) correlation between age and S-Hcy (rho = 0.26), age and S-cystatin C (rho = 0.40), S-Hcy and S-cystatin C (rho = 0.22), S-vitamin B12 and S-Hcy (rho = -0.28) and S-folate and S-Hcy (rho = -0.24). S-Hcy concentrations were lower in women than in men (12.0 (2.9-35.7) lmol/l vs 13.9 (5.8-50.0) lmol/l; 222 women and 257 men were examined; P \ 0.001) and Svitamin B12 concentrations were higher in women than in men (288 (118-789) pmol/l vs. 259 (98-742) pmol/l; 208 women and 233 men were examined; P = 0.005). S-folate and S-cystatin C concentrations did not differ between the sexes. S-Hcy concentrations were elevated in participants with DM and in those with elevated S-cystatin C [fig_ref] Table 1: S-Hcy concentrations in participants with condition reportedly associated with hyperhomocysteinemia P-value from... [/fig_ref]. shows that S-Hcy concentrations were increased among participants with atrophic gastritis, but not in those with H. pylori-associated gastritis without atrophy. Considering participants with subnormal PGI/PGII, indicative of atrophy of the gastric corpus mucosa, results were similar with median S-Hcy 17.5 (9.8-50.0) lmol/l among participants with subnormal PGI/PGII and 12.5 (2.9-33.1) lmol/l among those with PGI/PGII within normal range (P \ 0.001). S-Hcy levels did not differ significantly between participants with the MTHFR 677 TT genotype (13.0showed that S-Hcy concentrations were related to gender, age, S-cystatin C, S-vitamin B12, S-folate, MTHFR 677 TT genotype, and the occurrence of atrophic gastritis. Substituting the morphological diagnosis of atrophic gastritis infor subnormal PGI/PGII yielded similar results overall and the coefficient for subnormal PGI/PGII was 5.96 (95% c.i. 3.35-8.56, P \ 0.001). When H. pylori-associated chronic gastritis without significant atrophy of the corpus mucosa was included in the model, no significant relation to S-Hcy was shown. At baseline, 438 participants had no history of any cardiovascular disease. They were followed-up for 129 (9-178) months (4,796 person years), during which 101 were diagnosed with cardiovascular disease. These were myocardial infarction [bib_ref] Prevalence and mechanisms of hyperhomocysteinemia in elderly hospitalized patients, Salles-Montaudon [/bib_ref] , angina pectoris [bib_ref] Homocysteine and related Bvitamin status in coeliac disease: effects of gluten exclusion..., Dickey [/bib_ref] , abdominal aortic aneurysm [bib_ref] Interaction of plasma homocysteine and thyroid hormone concentrations in the pathogenesis of..., Evrengul [/bib_ref] , TIA/stroke [bib_ref] An association between Helicobacter pylori infection and serum vitamin B12 levels in..., Shuval-Sudai [/bib_ref] , and cardiac arrhythmia/heart failure [bib_ref] Meta-analysis of MTHFR 677C-&T polymorphism and coronary heart disease: does totality of..., Lewis [/bib_ref]. The median follow-up time for all 488 participants was 128 (1-180) months (5,262 person years). Among all 488 participants, 25 developed dementia. In six, the form of dementia was classified as Alzheimer's disease, in three as vascular dementia, in six as mixed dementia, and in ten it was not classified. Mild cognitive impairment (diagnosed in five participants) was not defined as dementia. The prevalence of atrophic gastritis was 9/101 among participants with incident cardiovascular diseases and 7/337 among participants not diagnosed with any cardiovascular disease during follow-up (P = 0.004). Corresponding figures for dementia were 3/25 and 18/463, respectively (P = 0.09). The prevalence of positive H. pylori status did not differ significantly between participants with and without incident cardiovascular diseases or dementia. S-Hcy was found to be elevated among participants developing cardiovascular diseases lmol/ l) compared to those without such diseases (12.1 (2.9-48.3) lmol/l) (P = 0.001). S-Hcy levels in subgroups are shown in . The median for S-Hcy was 15.7 lmol/l among participants developing dementia and 12.7 (2.9- The difference of their S-Hcy levels in relation to participants without dementia was close to significant for Alzheimer's disease (P = 0.05) and significant for non-specified dementia (P = 0.004). [fig_ref] Table 5: Associations of S-Hcy levels with risk of cardiovascular diseases grouped together [/fig_ref] shows the results of the logistic regression analyses with cardiovascular diseases considered together as the dependent variable and S-Hcy as a continuous or dichotomous independent variable with different cut-off levels. Exchanging S-Hcy in this model for gastric mucosal atrophy (as determined morphologically or with PGI/PGII), MTHFR 677 TT genotype and elevated cystatin C taken together as an independent variable, showed no significant relation to cardiovascular diseases (OR 1.70 [0.92-3.13] with morphological determination). As shown in [fig_ref] Table 5: Associations of S-Hcy levels with risk of cardiovascular diseases grouped together [/fig_ref] , there was no significant association between S-Hcy and subgroups of dementia considered together. # Discussion As is the case with most population-based cohorts, one may question whether the present one is representative. The incidence of myocardial infarction found was at approximately the same level as reported by the Swedish National Board of Health and Welfare (www.socialstyrelsen.se). Also, the incidence of dementia was at the same level as previously reported in Sweden [bib_ref] Incidence of dementia and major subtypes in Europe: a collaborative study of..., Fratiglioni [/bib_ref]. However, considering the invasive character of the EGD, which the participants were asked to undergo, the participation rate became relatively low and there may have been a selection bias based on the occurrence of digestive symptoms [bib_ref] Prevalence of gastroduodenitis and Helicobacter pylori infection in a general population sample:..., Borch [/bib_ref]. A strength of the study is its long follow-up interval of 10 years overall (4,796 person years for cardiovascular disease and 5,262 person years for dementia) and that, in addition to relations of Hcy to disease development, it includes main determinants of Hcy levels. The evaluation of these relations was enhanced by the access to data on known relevant confounders. Another strength of the study is the reliability of the files of the Swedish National Board of Health and Welfare in which personal identification numbers are linked to in-hospital diagnoses (www.social styrelsen.se) and, furthermore, that the information obtained from these files was verified or supplemented by information from the local hospital files and by the participants themselves. Regarding determinants of Hcy levels in circulation, the GLM analysis showed a positive relation of S-Hcy to age, male gender, MTHFR 677 TT genotype and atrophic gastritis (both as determined morphologically and functionally with PGI/PGII). Not unexpectedly, there was a negative relation of S-Hcy to renal function, S-vitamin B12, and S-folate. These findings are in accord with those in other studies [bib_ref] A candidate genetic risk factor for vascular disease: a common mutation in..., Frosst [/bib_ref] [bib_ref] Prevalence of hyperhomocysteinemia in adult gluten-sensitive enteropathy at diagnosis: role of B12,..., Saibeni [/bib_ref] [bib_ref] Homocysteine and related Bvitamin status in coeliac disease: effects of gluten exclusion..., Dickey [/bib_ref] [bib_ref] Helicobacter pylori eradication lowers serum homocysteine level in patients without gastric atrophy, Ozer [/bib_ref] [bib_ref] Prevalence and mechanisms of hyperhomocysteinemia in elderly hospitalized patients, Salles-Montaudon [/bib_ref] [bib_ref] Association of plasma homocysteine with coronary artery calcification in different categories of..., Kullo [/bib_ref] [bib_ref] Prevalence of low vitamin B12 and high homocysteine in serum in an..., Sipponen [/bib_ref] [bib_ref] The 677 C/T MTHFR polymorphism is associated with essential hypertension, coronary artery..., Ilhan [/bib_ref] , although no single study evaluated all these parameters. We found no significant association between the MTHFR 677 TT genotype and Hcy levels in the univariate analysis. This is not in agreement with a previous study showing an association with mild hyperhomocysteinemia [bib_ref] Homocysteine metabolism, Selhub [/bib_ref]. However, recent studies have shown a reduced effect of the MTHFR 677 TT genotype in populations with high folate intake [bib_ref] Meta-analysis of MTHFR 677C-&T polymorphism and coronary heart disease: does totality of..., Lewis [/bib_ref] [bib_ref] The Hordaland homocysteine study: a community-based study of homocysteine, its determinants, and..., Refsum [/bib_ref]. In the GLM analysis in the present study, this SNP had a significant impact on Hcy levels. The appearance of this significance may be explained by the inclusion of S-folate in the model. Inclusion in the GLM analysis of H. pylori-associated chronic gastritis without atrophy of the corpus mucosa did not show any significant relation to S-Hcy. This is in agreement with the finding of others, namely that H. pylori per se has no significant impact on Hcy levels [bib_ref] Atrophic gastritis as a cause of hyperhomocysteinaemia, Santarelli [/bib_ref] [bib_ref] Correlation between Helicobacter pylori infection, gastric inflammation and serum homocysteine concentration, Leung [/bib_ref]. There was no significant relation of S-Hcy to the occurrence of atrophy of the duodenal mucosa, DM, or hypothyreosis in the GLM analysis. Others have found increased S-Hcy levels in these diseases [bib_ref] Elevated plasma homocysteine level is an independent predictor of coronary heart disease..., Soinio [/bib_ref] [bib_ref] Interaction of plasma homocysteine and thyroid hormone concentrations in the pathogenesis of..., Evrengul [/bib_ref] [bib_ref] Association of plasma homocysteine with coronary artery calcification in different categories of..., Kullo [/bib_ref]. Since the number of cases in these groups in the present study was small, we do not draw any conclusions regarding their impact on Hcy levels. When using dichotomized values of S-Hcy with a cutoff level of 14.5 lmol/l we found a positive correlation between S-Hcy and the incidence of cardiovascular diseases; in fact, the risk was doubled (OR 2.05 [1.14-3.70]). With S-Hcy as continuous variable the relation was weaker (OR 1.06 [1.00-1.12]). Exchanging S-Hcy for its explanatory variables, atrophic gastritis, MTHFR 677 TT genotype and elevated cystatin C taken together as an independent variable in the analysis yielded a weak association, if any (OR 1.70 [0.92-3.13]). Since there are other Results are from logistic regression analysis adjusting for gender, age at baseline (years), follow-up interval (months), body mass index (kg/m 2 ), current smoking (yes/no), alcohol use (weekly use yes/no), use of NSAID (weekly use yes/no), P-total cholesterol (mmol/l) and P-triglycerides (mmol/l) causes of variations in Hcy levels than those studied here, this lack of an association is not unexpected. Considering the relation of Hcy to cardiovascular diseases, our results are in accordance with findings in other prospective studies, although the cut-off levels for Hcy concentrations used vary between the studies [bib_ref] Homocysteine and ischemic heart disease: results of a prospective study with implications..., Wald [/bib_ref] [bib_ref] Nordrehaug JE. Serum total homocysteine and coronary heart disease, Arnesen [/bib_ref] [bib_ref] Serum homocysteine and risk of coronary heart disease and cerebrovascular disease in..., Stehouwer [/bib_ref] [bib_ref] Homocysteine and shortterm risk of myocardial infarction and stroke in the elderly:..., Bots [/bib_ref] [bib_ref] Serum total homocysteine and coronary heart disease: prospective study in middleaged men, Whincup [/bib_ref]. In one of these studies [bib_ref] Homocysteine and ischemic heart disease: results of a prospective study with implications..., Wald [/bib_ref] , which included 229 cases and 1,126 controls (all men, mean follow-up time 8.7 years), there was a continuous dose-response relationship with a 41% (20%-65%) increased risk of death from ischemic heart disease for each 5-lmol/l increase in the S-Hcy level. In an other study [bib_ref] Nordrehaug JE. Serum total homocysteine and coronary heart disease, Arnesen [/bib_ref] of 122 cases and 478 controls, the relative risk of ischemic heart disease hospital discharges and deaths resulting from a 4 lmol/l increase in S-Hcy was 1.32 (1.05-1.65). Whincup et al. [bib_ref] Serum total homocysteine and coronary heart disease: prospective study in middleaged men, Whincup [/bib_ref] , who followed a population for 12.8 years (386 cases and 454 controls, men aged 40-59 years at entry), found that a 47% increase in S-Hcy was associated with a slight increase in the OR for myocardial infarction of 1.15 (1.00-1.32). For values of S-Hcy higher than 16.5 lmol/l, the OR for myocardial infarction was 1.77 (1.28-2.42). As shown in the comprehensive meta-analysis conducted by the Homocysteine Studies Collaboration, there are also prospective studies that show no correlation of Hcy levels to cardiovascular diseases. This was also the case in the study by Fallon et al. [bib_ref] Homocysteine and coronary heart disease in the Caerphilly cohort: a 10 year..., Fallon [/bib_ref] , which included 312 men with coronary heart disease and 1,248 matched controls. Although according to the results of univariate analysis in the present study, S-Hcy levels seemed to be elevated in subgroups with dementia, including dementia overall as dependent variable in the logistic regression analysis yielded no significant association to Hcy levels. In a prospective study of 1,092 participants (667 women and 425 men) with a median follow-up interval of 8 years, Seshadri et al. [bib_ref] Plasma homocysteine as a risk factor for dementia and Alzheimer's disease, Seshadri [/bib_ref] found that increased P-Hcy is an independent risk factor for dementia and Alzheimer's disease. The relative risk of dementia was 1.4 (1.1-1.9) for each increase of one SD in logarithmically transformed Hcy values. The corresponding figure for Alzheimer's disease was 1.8 (1.3-2.5). In an Italian dementia-free cohort of 816 participants, Ravaglia et al. [bib_ref] Homocysteine and folate as risk factors for dementia and Alzheimer disease, Ravaglia [/bib_ref] related baseline P-Hcy to incident dementia, including Alzheimer's disease, during a mean follow-up interval of 4 years. They concluded that increased Hcy and low folate concentrations are independent risk factors. Hyperhomocysteinemia was considered to be present with a P-Hcy higher than 15.0 lmol/l. With this cut-off level the hazard ratio for dementia was 2.08. To summarize, in this population-based cohort Hcy concentrations in circulation were dependent on gender, age, the levels of vitamin B12 and folate, as well as on renal function, the occurrence of atrophic gastritis and the MTHFR 677 TT genotype. There was a positive relation of Hcy levels to incident cardiovascular diseases. [fig] 7. 6 -: 50.0) lmol/l) and those with the CT (12.7 (4.8-33.1) lmol/l) or CC genotype (12.7 (2.9-48.3) lmol/l), nor was there any relation between S-Hcy levels and the presence of the T allele (12.8 (4.8-50.0) lmol/l).GLM analysis [/fig] [table] Table 1: S-Hcy concentrations in participants with condition reportedly associated with hyperhomocysteinemia P-value from the Mann-Whitney U-test when compared to participants with normal gastric mucosa b P-value from the Mann-Whitney U-test when compared to participants with mild or no villous atrophy of the duodenal mucosa [/table] [table] Table 5: Associations of S-Hcy levels with risk of cardiovascular diseases grouped together (A) and dementia subgroups considered together (B) as the dependent variableCut-off level for S-HcyOdds ratio [95% confidence interval] [/table]

Trends in high deductible health plan enrolment and spending among commercially insured members with and without chronic conditions: a Natural Experiment for Translation in Diabetes (NEXT-D2) Study ## Major concerns 1) The most important issue that needs to be addressed in this study is the interpretation of enrollment patterns given by and results in "Rates of HDHP Enrollment Over Time" section. The authors note that the "rate of enrollment of HDHPs increased markedly over the study period for all disease categories, increasing by approximately 6 percentage points per year…." The figure and the percentage point result both show a constant rate of increase, not an increasing rate. An increasing rate would have a convex shape, whereas the figure appears roughly linear. Further, you note that you used GEE models to model trends and annual rates, but never give results that suggest differences year-overyear, or test whether these differences are increasing at a statistically significant rate. It also does not appear that you tested year-to-year differences in prevalence of healthier vs chronic disease enrollees-only differences in average rates. The methodology notes you test the average difference in rates between disease groups and healthier members (p. 5; Statistical Analysis). Your research question implies a test of whether the composition of enrollees is changing over time, which implies a formal analysis of not just average differences but also year-toyear rates of enrollment. 2) Related to the first concern, much of the paper will need to be rewritten. The abstract Conclusion and manuscript Discussion sections are currently not supported by the results of this study. 3) Odds ratios have limitations not noted in the paper. Notably, they lack external validity because odds ratio are quite dependent on the number of explanatory variables and sample. . Additionally, odds ratios can be hard to interpret, and are often mistaken as probabilities. Perhaps consider a more intuitive measurement in , such as predicted prevalence. 4) I am confused about exactly how the sample is drawn-what percent of enrollees in your sample actively select into plans vs are subject to rollover by employers? Some of my confusion is likely semanticin the Intro you describe the rigorous methodology of the NEXT-D1 study, and deem this the NEXT-D2 study-but do not appear to use same methodology here. I think this confusion could be cleared up by deleting description of methodology for NEXT-D1 in Intro. However, the issue of selection into these plans remains and is never explicitly addressed. Chronic disease patient selecting into these plans is very different behaviorally from being 'forced' in by an employer decision. I think your enrollment results will be much more meaningful if you can tease that out. 5) Citations to other work need to be included. These authors primarily cite their own work. While they have contributed greatly to the literature on this topic, there are a number of researchers who have also contributed to enrollment in HDHPs (Kaiser Family Foundation); selection into HDHPs (e.g. and spending in HDHPs (e.g. Melinda Buntin; Paul Fronstin (EBRI)) Minor Issues 1) Put years of sample into Abstract 2) How do you handle enrollees who turn 65 during plan year? 3) The finding that OOP costs decreased over time seems counter to the idea of rising health care costs and rising share of cost shifting to enrollees. Any idea what's going on? 4) The finding that members with chronic disease paid a lower share of total costs is also interesting. Why is that? Is it because of higher costs in the first place? Are they more likely to hit their deductible? If so, what's their incentive for spending less (in the standardized spending algorithm) implying lower utilization? Some of this may be beyond the scope of this article, but any insights you can put into Discussion along these lines would be welcome. 5) You note in the Discussion that HSA-eligible members spend more OOP. This finding is interesting and should be highlighted more, perhaps with context of recent finding (Kullgren et al, JAMA Open, 2020) that >50% of people with accounts do not put any money into HSAs. 6) Tables in Appendix 4 are hard to interpret. I think they could use better labeling or more exposition in notes. ## Version 1 -author response ## Reviewer: 1 Comments to the Author Thank you for the opportunity to review this very interesting manuscript. I have a few questions related to statistical methodology. 1. What proportion of the subjects in each of the three groups had a higher deductible? Our original analysis focused on the percent of members with deductibles >$1000 (and we called this a HDHP). For the resubmission, we split our HDHP cohort into two groups -those with deductibles $1000-2499 and those with deductibles greater or equal to $2500for the analyses that examines the percentage of members enrolled in a HDHP. These results appear in Appendix 5 and are mentioned in the results section (under "Percentage of Members Enrolled in a HDHP Over Time"): "The percentage of members enrolled in a very high deductible health plan (≥$2500) increased over the study period for all disease categories, from less than 1% in 2005 to 14-15% in 2015 (Appendix 5). In the last three years of the study period, the percentage of members in a HDHP with a deductible between $1000 and $2499 remained relatively flat, and the increase observed in HDHP plan enrollment overall was driven by enrollment in very high deductible health plans." 2. Was the association between income and size of deductible explored, overall and in the three groups? This is an insightful question, but answering it is beyond the scope of the current paper. We are currently working on another project that examines socio-economic disparities in HDHP uptake and spending and we will add this suggested analysis to that paper. 3. Were subjects with a deductible between 500 and 1000 dollars just excluded from the analysis? Members with deductibles of $501-$999 were excluded from the analyses that examine OOP and total costs since we are comparing members in HDHP and low deductible health plans head to head. However, they are included in the models that assess percentage of members enrolled in a HDHP (they fall in the denominator). We updated the methods section to clarify this, as follows: "In the models to assess percentage of members enrolled in a HDHP and predictors of HDHP enrollment, the denominator was all members in that disease category. The analyses that examined OOP and total costs compared members in HDHP (≥ $1000) to members in low deductible plans (≤$500) and therefore excluded members with deductibles of $501-$999." 4. I couldn't find much mention of missing data. How many of the variables were missing and where? GEE requires complete data, so I'm wondering how many subjects were removed from the analysis due to this? It's also described how the deductible is imputed, but it appears that single imputation was performed. Is that correct? If that's true, please use multiple imputation instead, as single imputation underestimates the variance in any models. A very small percent of the members had missing values for the neighborhood level variables; we now include the n and % missing in the demographic table (Appendix 3). We excluded members with missing variables from the GEE models. We added a statement about missingness and the GEE variables to the limitations section of the paper. The question about imputation of the deductible level is one that we have thought about at length and have addressed in other peer-reviewed publications that use the same deductible imputation. We did not use multiple imputation for 3 reasons: 1. A key requirement for multiple imputationthat missingness be randomis not met. Missingness in deductible level is in fact highly systematic given that it is almost entirely among larger employers. The reason for this systematic missingness is that we have a benefits table (list of exact deductible levels per employer ID) that includes mostly small employers and fewer large employers. 2. The uncertainty about the deductible we inferred from claims is low. Moreover, as we now demonstrate below, sensitivity and specificity improve substantially at larger employers (where missingness in deductible level predominates) because such employers generate more claims, providing better evidence about actual deductible levels at their firms. 3. Carrying out formal multiple imputation would be a practical challenge. The programming steps that make up and follow our imputation algorithm are enormously time consuming due to our large sample size (the process incorporates analysis of every single claim from every one of the many millions of members in the entire dataset) and the complexity of SAS programs. We previously validated our imputation algorithm in a single category of employer size that has substantial overlap with gold standard data (i.e., 75 to 100 enrollees per employer). However, we do have gold standard data on a smaller number of large employers. Therefore, we created 3 additional categories of employer size (101-400, 401-700, and 701-1000) and tested sensitivity and specificity, finding a range from 96% to 100% that increased across employer size category. We now include these results as Appendix .a-1.d. Furthermore, in the limitations section of the Discussion, we now include the following sentences: "Although we knew the exact deductible level of most smaller employers, we had to infer it from claims at large employers. However, the sensitivity and specificity of our algorithm was high and increased across employer size category, ranging 96% to 100% (Appendix 1). We expect adjustment for the uncertainty of the imputation process would have a negligible effect on confidence bounds." Reviewer: 2 Comments to the Author This study describes trends in enrollment and costs for those in high-deductible health plans (HDHPs), comparing healthier enrollees to those with two common chronic conditions, diabetes and cardiovascular disease. It uses data from a large national insurer from 2005-2013. While some of the findings, particularly those on out-of-pocket costs-are well-established in the literature, enrollment trends among people with different health risks are not, making this paper a potentially important addition to the literature. I have outlined several concerns with this project below, notably with the presentation and interpretation of results. I appreciated the chance to read this paper and all comments are intended to help the authors improve upon their work. Major Concerns 1. The most important issue that needs to be addressed in this study is the interpretation of enrollment patterns given by and results in "Rates of HDHP Enrollment Over Time" section. The authors note that the "rate of enrollment of HDHPs increased markedly over the study period for all disease categories, increasing by approximately 6 percentage points per year…." The figure and the percentage point result both show a constant rate of increase, not an increasing rate. An increasing rate would have a convex shape, whereas the figure appears roughly linear. Further, you note that you used GEE models to model trends and annual rates, but never give results that suggest differences year-over-year, or test whether these differences are increasing at a statistically significant rate. It also does not appear that you tested year-to-year differences in prevalence of healthier vs chronic disease enrollees-only differences in average rates. The methodology notes you test the average difference in rates between disease groups and healthier members (p. 5; Statistical Analysis). Your research question implies a test of whether the composition of enrollees is changing over time, which implies a formal analysis of not just average differences but also year-to-year rates of enrollment. We apologize that the terminology in our original submission was confusing. Throughout the manuscript, we replaced "rate" with "percentage of members in a HDHP," to describe our HDHP enrollment analysis. In addition to examining the percentage of members in a HDHP each year, we also measured the slope of the enrollment (i.e., year-over-year differences)we now refer to these analyses as "trend" (e.g., the trend in HDHP enrollment increased by 5 percentage points a year). In the revised manuscript, we tested whether these trends were statistically significant (they all were) and added the 95% confidence intervals of the slopes. We conducted these two analyses (i.e., examined the percentage of members in a HDHP and the trend of HDHP enrollment) separately for each disease category. Since HDHP enrollment in each disease category had similar year-to-year trends (i.e., slopes), we did not test the difference in trends, as you noted (although the reader can examine the trends head to head by looking at the confidence intervals for each disease area provided in the results section). Rather, we focused on the average difference in the percentage of members in a HDHP for diabetes vs. healthy members and CVD vs. healthy members. For the OOP and total cost analyses, we provide the following analyses to compare HDHP and lowdeductible health plans within each disease category. -Average absolute difference in cost between HDHP and low-deductible health plans over the study period -Average relative difference in cost between HDHP and low-deductible health plans over the study period -Slope (i.e., year-over-year change, aka "trend") of average costs for HDHP and lowdeductible health plan members (Note that we do not explicitly compare costs across the different disease categories since the key analysis is HDHP vs. low deductible plans, although the differences between disease categories are apparent in the Related to the first concern, much of the paper will need to be rewritten. The abstract Conclusion and manuscript Discussion sections are currently not supported by the results of this study. With the changes made above, we now more clearly describe the results and the conclusion and discussion sections are supported by the results. 3) Odds ratios have limitations not noted in the paper. Notably, they lack external validity because odds ratio are quite dependent on the number of explanatory variables and sample. (EC Norton et al. JAMA. July 2018). Additionally, odds ratios can be hard to interpret, and are often mistaken as probabilities. Perhaps consider a more intuitive measurement in , such as predicted prevalence. Per your suggestion, we have replaced the odds ratios in with average adjusted predictions, which we refer to as "predicted probability of HDHP enrollment." We also now use this method to calculate the percentage of members in a HDHP. We have updated the methods section and all of the results in the manuscript and and . ## 4) i am confused about exactly how the sample is drawn-what percent of enrollees in your sample actively select into plans vs are subject to rollover by employers? Some of my confusion is likely semanticin the Intro you describe the rigorous methodology of the NEXT-D1 study, and deem this the NEXT-D2 study-but do not appear to use same methodology here. I think this confusion could be cleared up by deleting description of methodology for NEXT-D1 in Intro. However, the issue of selection into these plans remains and is never explicitly addressed. Chronic disease patient selecting into these plans is very different behaviorally from being 'forced' in by an employer decision. I think your enrollment results will be much more meaningful if you can tease that out. You are correct. The study population is the entire population of members in the study categories (i.e., diabetes, cardiovascular disease, healthy) in the database; it is not limited to the NEXTD-1 study population (i.e., members forced into HDHPs by their employer). We see how this led to confusion and removed the details of the NEXTD-1 study design from the intro. The issue of members' choice of HDHPs is a very interesting one and a topic of another paper on which we are working. We alluded to this limitation and future work in the discussion section of the original submission ("Finally, our data combined HDHP members whose employers offered only a HDHP with members who were offered a choice by their employer to enroll in a HDHP or a lowerdeductible plan. Future research should examine HDHP enrollment among members who have plan choice to better understand factors associated with selecting HDHPs.") However, in the revision, we have decided to include a few of the choice calculations: (1) the percentage of members who are offered a choice of a HDHP and a non-HDHP from their employer and (2) the percentage of members with employer-level choice who choose to enroll in a HDHP. These now appear in the methods and results sections and in Appendix 6 and we allude the results in the discussion section: "Among the subset of members who were offered a choice of a HDHP or lower deductible plan from their employer, most members opted for a lower deductible health plan and members with chronic diseases were less likely to choose a HDHP than healthier members." 5) Citations to other work need to be included. These authors primarily cite their own work. While they have contributed greatly to the literature on this topic, there are a number of researchers who have also contributed to enrollment in HDHPs (Kaiser Family Foundation); selection into HDHPs (e.g. Lave, Men et al., and spending in HDHPs (e.g. Melinda Buntin; Paul Fronstin (EBRI)) Thank you for these suggestions. We added references from Lave, Buntin and Fronstin to the introduction and discussion sections of the paper. We had already included references to the Kaiser Family Foundation in the original submission. ## Minor issues 1) put years (2005-2013) of sample into abstract The years are in the abstract (under participants). 2) How do you handle enrollees who turn 65 during plan year? We assigned age at the annual level using member age at the anniversary month, which is the last month of each member's 12 month cycle. Anyone with age >64 at the anniversary month would have been excluded so there was no opportunity for anyone to turn 65 during the study period. 3) The finding that OOP costs decreased over time seems counter to the idea of rising health care costs and rising share of cost shifting to enrollees. Any idea what's going on? OOP costs remained relatively flat over the time period for the HDHP members and decreased slightly for the low deductible plan members. There are multiple explanations for this. One, our cost figures are adjusted for medical cost inflation (unlike the figures in reports like this one from Kaiser Family Foundation: https://www.healthsystemtracker.org/brief/tracking-the-rise-in-premiumcontributions-and-cost-sharing-for-families-with-large-employer-coverage/). Also, unlike the KFF report, our OOP figures are adjusted for changes in demographic and health characteristics and employer characteristics and stratified by disease category (meaning that increases in the prevalence of diabetes and cardiovascular disease over time cannot impact the OOP costs in our study). 4) The finding that members with chronic disease paid a lower share of total costs is also interesting. Why is that? Is it because of higher costs in the first place? Are they more likely to hit their deductible? If so, what's their incentive for spending less (in the standardized spending algorithm) implying These are all very interesting and important questions. Yes, members with chronic diseases paid OOP for a lower share of total costs than healthier members because members with chronic diseases have much higher total costs and these members either hit the deductible or OOP max limits. Fronstin et al. showed that half of high cost members (which they define as members with the top 10% of spending) hit their deductible and OOP max. We added this to the discussion section of the article. Also, you are correct -there is evidence that once members hit a deductible they have little incentive to spend less. HDHP members who exceed the deductible increase utilization of low value care and screenings (e.g., and adding cost-containing measures (copayments/co-insurance) after the deductible is met decreases spending (Fronstin EBRI Issue Brief November 19, 2020, No 519). However, analyzing spending before and after a member hits their deductible is beyond the scope of this study. There is evidence from multiple studies (mentioned in the intro) that deductibles provide a financial disincentive to seek needed carethis is likely the driver of lower costs we observe in HDHP members compared to members with lower deductibles. lower utilization? Some of this may be beyond the scope of this article, but any insights you can put into Discussion along these lines would be welcome. 5) You note in the Discussion that HSA-eligible members spend more OOP. This finding is interesting and should be highlighted more, perhaps with context of recent finding (Kullgren et al, JAMA Open, 2020) that >50% of people with accounts do not put any money into HSAs. Thank you for this suggestion. We added the Kullgren et al. reference to the discussion section. 6) Tables in Appendix 4 are hard to interpret. I think they could use better labeling or more exposition in notes. Thank you for this comment. We split Appendix 4 into two appendices (Appendix 4a and 4b), renamed the Appendix title and figure titles within the appendix, and added a note to explain the denominator. We also added to the results section of the manuscript that "non-account HDHPs were the most common HDHP type for HDHP members in all three disease areas" (this refers to data in Appendix 4a) and then made it clear that the next result mentioned ("the higher percentage of enrollment in HDHPs among healthy members, compared to members with chronic diseases, was associated with higher enrollment in HSA-eligible HDHPs") specifically refers to the figures in Appendix 4b. article is ready to be accepted. In re-reading the manuscript, I had 2 minor changes suggested to increase clarity: 1) In "Results:Percentage of Members Enrolled in an HDHP Over Time", sentence reading "Members with chronic diseases were less likely than healthier members to be in an HDHP throughout the entire study period ." I interpreted this statement as implying as longer spells of enrollment by healthier HDHP members. Based on , I think what you mean is that, throughout the study period, healthier members had consistently higher levels of HDHP enrollment. Change the wording to reflect that interpretation. Same issue in second sentence of the Discussion. Make it clear that you are talking about the level of enrollment not the length of spells. Perhaps something like "In all study period years, members with diabetes and CVD had lower levels of enrollment." 2) In notes for , add more details about how the percentage is adjusted, i.e. the exact regression model used. ## Version 2 -review ## Version 2 -author response We responded to Reviewer 1's comments and made the following minor changes: 1. Revised two sentences (one in the results section and one in the discussion section) to make it clear that members with chronic diseases had lower levels of HDHP enrollment than healthier members (and not shorter spells of HDHP enrollment). 2. Updated the notes in all figures (not just as the reviewer requested) to provide more details about how we calculated the adjusted estimates. Review 2 did not have any suggested revisions. In response to the editor's comments, we removed one of the "strengths and limitations" bullets and now have 5 bullets. We also added an ethics statement just before the references list. We look forward to having our work published in BMJ Open! Please let me know if you need me to make any further changes.

Hospital Characteristics Associated With Heterogeneity in Institutional Postacute Care Spending Reductions Under the Comprehensive Care for Joint Replacement Model ## Eappendix 1. additional methodology Our studied used causal forests using the `grf` R package version 2.0.2 to assess the heterogeneous treatment effect of CJR model on institutional post-acute care spending. Causal forests are type of generalized random forest described in . Generalized random forests in turn are extension of the common random forest machine learning model . ## Theoretical framework Random forests are a type of ensemble machine learning that seek to predict an outcome y using a set of predictor features x. Random forests average over a large number of classification and regression trees, which are very simple non-parametric prediction models that attempt to split the data into recursive partitions in such a way to maximize differences in the outcome between subgroups. The final subgroups in the tree are referred to as the leaves, and hopefully identify distinct subpopulations predictive of the outcome. This type of model is often referred to as "greedy" in the sense that they will leverage random variance unique to the sample to predict in a way that may not be generalizable to other samples. To avoid this, it is common to split the data into two samples for each tree, the training sample that is used to build the tree and the test sample which is used to predict. In our sample we also apply a "honest" correction which further splits the training sample into one subsample that is used to identify the splitting rules for the tree and another that estimates the predictions for each leaf. Each tree in the forest also each only consider a subset of the covariates so that they do not all identify the same set of subgroups. The final prediction for each observation is the average of predictions for all trees where the observation was not part of the training sample, also referred to as "out of bag predictions" because the observation was not part of the "bag" used to create the trees. Unlike random forest, generalized random forests are optimized to split based on something other than simply maximizing the differences in the outcome. In the case of the causal forests we use, the forests split instead to maximize differences in the treatment effect. This is accomplished using what is called the R Learner objective function described in , which is an extension of the doublyrobust estimator for treatment effect described in . The method uses predicted values (or rather the residuals from those predictions) for both the outcome and the treatment to estimate the treatment effect, and is often called doubly robust because the method is robust to misspecifications in either (but not both) of the prediction models. The full conceptual framework for the approach is better described in . ## Conditional average treatment effects (cate) For each hospital in our dataset, the causal forest model will estimate a treatment effect conditional on the hospital's characteristics, also called a Conditional Average Treatment Effect (CATE). This can be interpreted as the estimated difference in outcomes between hospitals with similar features that fall into the same leaves based on the splitting rules in the causal forest. The precision of these estimates will depend both on the strength of the underlying heterogeneity, the association between the provided covariates and that heterogeneity, and the amount of observations that share a covariate space. ## Running the causal forest In practice, our application of the causal forest method followed a number of distinct steps borrowed from Athey & Wager (2019). First, we predict our outcome (per-episode institutional post-acute care spending) using a random forest fit with our full list of covariates, minus treatment. Second, we estimate the probability of treatment, which in the case of the CJR program is the selection probability used by CMS as part of the stratified random sampling for participation in the CJR program. Next we feed both of those predicted values into a naïve causal forest using the default settings and including all covariates. For each covariate in this model, we measure its importance, or how often it was used as a splitting rule in the forest. We use this measure of importance to drop the covariates in the bottom half of importance. This culling allows the final forest to focus on splitting only the most important covariates. Finally, we run a causal forest using the remaining covariates. This final run also runs a series of smaller forests to tune some of the hyperparameters that guide the fit of the forest, specifically the minimize sample size for leaves and the number of covariates considered in each tree. All of the models described above are also specified to account for clustering of hospitals within MSAs, both because our treatment was assigned at the MSA level and also because hospitals in the same geographic area often share post-acute care resources. found that failing to account for clustering will produce CATE estimations with greater precision that overestimate the heterogeneity in the model. We found similar results and have chosen to use the more conservative clustering approach. The full code used to run this analysis is presented in the eSection 2 for both transparency and to facilitate replication. ## Model assumptions The causal forest approach requires two major assumptions and a number of smaller assumptions. First, similar to other estimators of treatment effect it assumes the treatment assignment is exogenous, and thus that the treatment effect is not confounded by unmeasured variables that contribute to both the treatment assignment and the outcome. This assumption is well met in our study because CJR participation was randomly assigned by CMS and is thus independent of any potential confounders. The selection probabilities for each strata of the random sample were provided by the Lewin Group, who were the official evaluators of the CJR model. Second the model assumes that the covariates provided adequately predict heterogeneity. The model is unable to estimate heterogeneity that is independent of the covariates provided. This assumption is not testable, however we have attempted to provide the model with a rich set of patient, hospital and MSA-level covariates. The model also assumes that the treatment and control groups have sufficient overlap in their probability of receiving treatment (met by the random assignment), and that the probability of receiving treatment is bound within 0,1 (met by the random assignment). ## Best linear predictor test The best linear predictor test serves as both a test of model fit and an omnibus test of whether our model is predicting heterogeneity . This test attempts to fit a linear model predicting the treatment effect estimator using two predictors, the mean prediction from the causal forest and the amount each hospital's CATE differs from that mean prediction. A significant coefficient near one for the mean prediction coefficient indicates that the model is accurately capturing the average treatment effect. A significant coefficient near one for the differential coefficient indicates that the model is accurately capturing heterogeneity in the model, while a value of zero implies that the model did not capture heterogeneity. Thus, the p-value for the differential coefficient can be used as a test of treatment heterogeneity. The results of the best linear projection test for our model are presented below. This test suggests that our model is accurately capturing the average effect, but that it does not detect significant heterogeneity (though the estimate is relatively close to 1). ## Estimate Std . ## Median test of treatment heterogeneity Another test we used to check for treatment heterogeneity was to split the dataset into two groups, those with CATE estimates above the median CATE and those below the median, and then calculate the Average Treatment Effect for each group. We can then calculate the difference in ATEs and test its significance using a two-sample t-test. This method is suggested in Athey & Wager (2019). ## Best linear project test The best linear projection test examines whether a single predictor variable is linearly associated with treatment effect heterogeneity. It simply predicts the treatment estimator using a linear model including the predictor of interest. This method may not be able to identify non-linear associations, which is why we also examined the ATE and CATE distribution stratified into quintiles of each hospital characteristic of interest. ## Eappendix 2. analysis code and documentation ## Preliminary work Switches strata_vars <-c("pct_complex_pt", "pct_dual", "lejr_n", "pac_spending ") ## Set seed set.seed (8321091) Set paths inpath <-"/home/groups/chse/WorkingData/ARG_30_NIHJointReplacement/D ata/analytic_datasets/DiffTx" codepath <-"/home/groups/chse/WorkingData2/User_Folders/Thomas/active _projects/CJR_Disparities/Manuscripts/DiffTx/analysis/" plot_path <file.path("/home/groups/chse/WorkingData/ARG_30_NIHJointReplacement/ Output/Plots/p6_DiffTx", format(Sys.Date(), "%Y%m%d")) tab_path <file.path("/home/groups/chse/WorkingData/ARG_30_NIHJointReplacement/ Output/Tables/p6_DiffTx", format(Sys.Date(), "%Y%m%d")) ## If(!dir.exists(plot_path)){dir.create(plot_path)} if(!dir.exists(tab_path)){dir.create(tab_path)} Load Required Packages packs <-c("data.table", #Best package for managing large datasets "ggplot2", #Makes pretty graphics "stringr", #String manipulation "ggridges", #Ridgeline Plots "patchwork", #Combine Plots "grf", #GRF functions "dplyr", #Tidyverse functions "scales", #Scaling functions "rpart", #CART models "tableone", #Table One creation "magrittr" #Enables usf function piping ) ## ## Attaching package: 'dplyr' ## The following objects are masked from 'package:data. plot <temp %>% #Factor group mutate(group = factor(group, levels = c(1, 0), labels = c("Treatment", "Control"))) %>% #Create bins mutate(ps_binned = cut(psvalue, breaks = seq(-0.025, 1, 0.025), include.lowest = T, labels = seq(0, 1, 0.025))) %>% #Calculate distribution of bins for each group group_by(group) %>% mutate(n = n()) %>% group_by(group, ps_binned) %>% summarize(p = n()/max(n)) %>% mutate( p = case_when(group == "Treatment" ~ p, group == "Control" ~ -p) ) %>% #Plot ggplot(aes(x = ps_binned, y = p, fill = group)) + geom_bar(stat = 'identity', width = 0.75, position = position_stack()) + geom_hline(yintercept = 0, color = "black") + geom_text(aes(y = max(p)*1.1, x = "0.5", label = "Treatment"), color = "dodgerblue", size = 12) + geom_text(aes(y = min(p)*1.1, x = "0.5", label = "Control"), color = "gray60", size = 12) + scale_fill_manual(values = c("dodgerblue", "gray60")) + scale_x_discrete(name = "Propensity Score", breaks = seq(0.0, 1, 0.1), limits = seq(0.0, 1, 0.025) %>% as.character, label = seq(0, 1, 0.1) %>% percent) + scale_y_continuous("Density", breaks = 0) + theme_minimal() + theme(legend.position = "none") ## Fancy quantile Creates a fancy quantile label with length equal to the input string. Saves some coding later. fancy_quantile <-function(x, quantiles, label_func = identity){ #calcualte quantiles and cut q_cuts <-quantile(x, quantiles) q_pred = cut(x, breaks = q_cuts, include.lowest = T) #Calculate Quantile number q_lab_pre = cut(x, breaks = q_cuts, include.lowest = T, labels = paste0("Q", 1:(length(q_cuts)-1))) #Choose whether to close range with inclusion or exclusion # (Brackets or Parens) open_type <-substr(as.character(q_pred), 1,1) close_type <-substr(as.character(q_pred), nchar(as.character(q_pred)), nchar(as.character (q_pred)) vars <-c("owner_type", "maj_teach", "op_margin_pat", "bed_cnt", "in_system", "sys_multistate", "sys_acutehosp", "pac_affil", "pac_spending", "lejr_n", "pct_hip", "pct_frac", "pct_male", "pct_black", "pct_hisp", "pct_oth", "pct_white", "pct_dual", "pct_poverty", "mean_age", "pct_complex_pt", "pct_ext_surg", "elix_score_m" , "elix_score_r", "msa_pop_2010", "hhi", "msa_mcr_adv_pen", "msa_hha_rate", "m sa_snf_rate", "msa_bpci_pre") vars_cat <-c("owner_type", "maj_teach", "in_system", "sys_multistate", "pac_affil") tab1 <-CreateTableOne(vars = vars, strata = "treat", data = analytic2, factorVars = vars_cat) ps_vars <-c("msa_group", "in_system", "pac_spending", "hhi", "msa_mcr_adv_pen", "msa_hha_rate", "msa_snf_rate", "msa_pop_2010", "msa_bpci_pre", "owner_type", "pac_affil", "bed_cnt", "maj_teach", "op_ma rgin_pat", "lejr_n", "pct_hip", "pct_frac", "pct_comp", "pct_male", "pct_black", "pct_hisp", "pct_oth", "pct_whit e", "pct_dual", "pct_poverty", "mean_age", "pct_complex_pt", "pct_ext_surg", "elix_score_m", "elix_score_r") x_frame <analytic x <-model.matrix(~., data = x_frame) x <-x[, -1] ## Clusters We cluster our sampling on MSA (since MSAs were CMS's sampling unit, and there is no treatment variation within MSA) msa <-analytic$msa %>% as.numeric() ## Weights Weights are the combination of sample weights and LEJR counts (to match the patient level ATE) ## Weight <-analytic$lejr_n_post Plot CATE ate <-average_treatment_effect(pac_spending_cf) ate_est <-ate[["estimate"]] ate_se <-ate[["std.err"]] ate_low <-ate_est -1.96*ate_se ate_high <-ate_est + 1.96*ate_se ate_label <sprintf("ATE = %s (%s, %s)", dollar_format(1)(ate_est), round(ate_low), round(ate_high)) plot_cate <analytic %>% ggplot(aes(x = pac_cate)) + geom_histogram(binwidth = 100) + geom_vline(xintercept = ate_est, color = "red", linetype = 2, size = 1) + annotate("text", x = ate_est, y = 60, label = ate_label, size = 4, color = "white", angle = 90, vjust = -1) + scale_x_continuous("CATE ($)", labels = dollar, limits = c(-2500,250 0)) + scale_y_continuous("Hospitals (N)", labels = comma) + theme_minimal() + theme(axis.title = element_text(size = 14), axis.text = element_text(size = 12)) plot_cate ## Warning: Removed 2 rows containing missing values (geom_bar). ## ---## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 ## Test individual predictors stored plot settings Savings some plot settings here for easier swapping later, First, theme: my_theme <-theme_minimal() + theme(panel.grid.major.y = element_blank(), legend.position = "none", plot.title = element_text(size = 16), plot.title.position = "plot", axis.title.y = element_text(margin = margin(r = 10), size = 14), axis.title.x = element_text(margin = margin(t = 10), size = 14), axis.text.x = element_text(size = 12), axis.text.y = element_text(size = 12), axis.ticks = element_line(), axis.ticks.length.y = unit(5, "mm") ) Stored values cate_breaks <-seq , 1000) #CATE breakpoints ate_breaks <-seq Predictor: Complex Patient Percent Proportion of patients in the top 25% of the Elixhauser readmission risk score. ## Best linear projection Test if predictor is linearly associated with DR scores (i.e. Treatment effect) Next, the ridgeline plot complex_pat_ridge <analytic %>% ggplot(aes(x = pac_cate, y = complex_pt_q)) + geom_density_ridges2(fill = "white") + geom_vline(xintercept = med_cate, color = "red", linetype = 2, size = 1) + ggtitle("A. Medical Complexity") + scale_y_discrete("Percent Medically Complex\nPatients (Quintiles)", expand = expansion(mult = c(0, 0.2))) + scale_x_continuous("CATE Estimates", labels = dollar, breaks = cate_breaks) + my_theme Stratified ATE by quintile complex_pat_q <-quantiles_ate(pac_spending_cf, predictor = analytic$pct_complex_pt, quantiles = seq(0,1,0.2)) Plot complex_pat_ate <ggplot(complex_pat_q, aes(x = ate_est, xmin = ate_low, xmax = ate_high, y = q_label )) + geom_point() + geom_errorbarh() + geom_vline(xintercept = ate[[1]], color = "red", linetype = 2, size = 1) + ggtitle("A. Medical Complexity") + scale_y_discrete("Pre-period Medically Complex\nPatient Percentage ( Quintiles)", expand = expansion(mult = c(0, 0.2))) + scale_x_continuous("ATE Estimates", labels = dollar, breaks = ate_breaks) + my_theme complex_pat_ate CATE Ridgeline by Quintile analytic[, dual_q := fancy_quantile(pct_dual, quantiles = seq(0, 1, 0.2), label_func = percent_format(0.1))] Next, the ridgeline plot dual_ridge <analytic %>% ggplot(aes(x = pac_cate, y = dual_q)) + geom_density_ridges2(fill = "white") + geom_vline(xintercept = med_cate, color = "red", linetype = 2, size = 1) + ggtitle("B. Social Complexity") + scale_y_discrete("Percent Socially Complex\nPatients (Quintiles)", expand = expansion(mult = c(0, 0.2))) + scale_x_continuous("CATE Estimates", labels = dollar, breaks = cate_breaks) + my_theme Stratified ATE by quintile dual_quint <-quantiles_ate(pac_spending_cf, predictor = analytic$pct_dual, quantiles = seq(0,1,0.2)) Plot dual_ate <ggplot(dual_quint, aes(x = ate_est, xmin = ate_low, xmax = ate_high, y = q_label )) + geom_point() + geom_errorbarh() + geom_vline(xintercept = ate ], color = "red", linetype = 2, size = 1) + ggtitle("B. Social Complexity") + scale_y_discrete("Pre-period Socially Complex\nPatient Percentage (Q uintiles)", expand = expansion(mult = c(0, 0.2))) + scale_x_continuous("ATE Estimates", labels = dollar, breaks = ate_breaks) + my_theme dual_ate Predictor: LEJR Volume Number of joint replacements provided in the pre-intervention period. ## Best linear projection Test if predictor is linearly associated with DR scores (i.e. Treatment effect) Next, the ridgeline plot volume_ridge <analytic %>% ggplot(aes(x = pac_cate, y = volume_q)) + geom_density_ridges2(fill = "white") + geom_vline(xintercept = med_cate, color = "red", linetype = 2, size = 1) + ggtitle("C. Surgical Volume") + scale_y_discrete("Number of Pre-period\nLEJR Encounters (Quintiles)" , expand = expansion(mult = c(0, 0.2))) + scale_x_continuous("CATE Estimates", labels = dollar, breaks = cate_breaks) + my_theme (1))] Next, the ridgeline plot pac_ridge <analytic %>% ggplot(aes(x = pac_cate, y = pac_q)) + geom_density_ridges2(fill = "white") + geom_vline(xintercept = med_cate, color = "red", linetype = 2, size = 1) + ggtitle("D. Pre-period Institutional PAC Spending") + scale_y_discrete("Pre-Period Institutional Post-Acute\nCare Spending (Quintiles)", expand = expansion(mult = c(0, 0.2))) + scale_x_continuous("CATE Estimates", labels = dollar, breaks = cate_breaks) + my_theme pac_ridge ## Picking joint bandwidth of 94 Stratified ATE by quintile pac_quint <-quantiles_ate(pac_spending_cf, predictor = analytic$pac_spending, quantiles = seq(0,1,0.2), label_func = dollar_format(1)) Plot pac_ate <ggplot(pac_quint, aes(x = ate_est, xmin = ate_low, xmax = ate_high, y = q_label )) + geom_point() + geom_errorbarh() + geom_vline(xintercept = ate[[1]], color = "red", linetype = 2, size = 1) + ggtitle("D. Pre-period Institutional PAC Spending") + scale_y_discrete("Pre-Period Institutional Post-Acute\nCare Spending (Quintiles)", expand = expansion(mult = c(0, 0.2))) + scale_x_continuous("ATE Estimates", labels = dollar, breaks = ate_breaks) + my_theme pac_ate [table] Table of Contents: Context ......................................................................................................................................... 9 Preliminary Work ........................................................................................................................ 10 Switches .................................................................................................................................. 10 Set Seed .................................................................................................................................. 10 Set paths ................................................................................................................................. 10 Load Required Packages .......................................................................................................... 10 Custom Functions .................................................................................................................... 11 Check Overlap ..................................................................................................................... 13 split_ate .............................................................................................................................. 14 Fancy Quantile .................................................................................................................... 15 Quantile ATE ....................................................................................................................... 16 Table One ................................................................................................................................ 17 Structure Data ......................................................................................................................... 19 Outcomes ............................................................................................................................ 19 Treatment Variable ............................................................................................................. 20 X covariates ......................................................................................................................... 20 Clusters ............................................................................................................................... 20 Weights ............................................................................................................................... 20 Analysis ....................................................................................................................................... 21 Run Model ............................................................................................................................... 21 Plot CATE ................................................................................................................................. 22 Test Heterogeneity .................................................................................................................. 23 Median Test ........................................................................................................................ 23 BLP Test ............................................................................................................................... 23 Test for treatment heterogeneity using median and BLP test 10. Plot stratified density curves for CATE across deciles of variables of interest [/table]

Comparative inhibitory profile and distribution of bacterial PARPs, using Clostridioides difficile CD160 PARP as a model Poly-ADP-ribose polymerases (PARPs) are involved in the regulation of important cellular processes, such as DNA repair, aging and apoptosis, among others. They have been considered as promising therapeutic targets, since human cancer cells carrying BRCA1 and BRCA2 mutations are highly sensitive to human PARP-1 inhibitors. Although extensive work has been carried out with the latter enzyme, little is known on bacterial PARPs, of which only one has been demonstrated to be active. To extend this limited knowledge, we demonstrate that the Gram-positive bacterium Clostridioides difficile CD160 PARP is a highly active enzyme with a high production yield. Its phylogenetic analysis also pointed to a singular domain organization in contrast to other clostridiales, which could be due to the long-term divergence of C. difficile CD160. Surprisingly, its PARP becomes the first enzyme to be characterized from this strain, which has a genotype never before described based on its sequenced genome. Finally, the inhibition study carried out after a high-throughput in silico screening and an in vitro testing with hPARP1 and bacterial PARPs identified a different inhibitory profile, a new highly inhibitory compound never before described for hPARP1, and a specificity of bacterial PARPs for a compound that mimics NAD + (EB-47).Post-translational modifications (TMPs), which are widespread throughout the phylogenetic scale, consist of chemical modifications that occur in proteins catalysed by specific enzymes 1 . TMPs allow cells to produce rapid responses to changes in the environment. Among the different types described in both prokaryotic and eukaryotic cells is the so-called ADP-ribosylation 2,3 , which introduces units of ADP-ribose (ADPr) at the expense of NAD + . This reaction is catalysed by a special class of glycosyltransferases, named ADP-ribosyltransferases (ARTs). They were first described in the diphtheria toxin and then in the choleric toxin as a form of interference with important proteins (e.g. elongation factor 2, G proteins, and Rho GTPases), thereby disrupting host cell biosynthetic, regulatory and metabolic pathways as a way of gaining advantage during the infection process 4 . ARTs can be divided into two main groups based on active site amino acids: the so-called ADP-ribosyl transferases cholera toxin-like (ARTCs) and ADP-ribosyl transferases diphtheria toxin-like (ARTDs). The first group includes GPI-anchored extracellular or secreted enzymes containing an R-S-E (Arg-Ser-Glu) motif, which catalyse the mono-ADP-ribosylation (MARylation) of their substrates 5 . The remaining group comprises intracellular ADP-ribosyl transferases able to transfer either a single ADP-ribose residue (H-Y-I/L motif) or several ADP-ribose residues (H-Y-E motif), resulting in linear or branched chains of ADP-ribose (poly-ADP-ribosylation or PARylation) 6 . In the latter group, the invariant Glu (E) is the key catalytic residue that coordinates the transfer of ADP-ribose to the acceptor site, the His (H) forms a hydrogen bond with the N-ribose, and the tyrosine (Y) side chain stacks with the N-ribose and the nicotinamide moiety, thus facilitating the binding of NAD + 7 . However, when the catalytic glutamate residue is replaced by a small hydrophobic residue in enzymes of the mono-ARTD group (mARTD), a glutamate residue of the substrate is used as the catalytic glutamate, giving rise to a substrate-assisted catalysis to transfer the ADP-ribose moiety. This produces a modified glutamate residue, which is then no longer available for the addition of new ADPr molecules [bib_ref] Substrate-assisted catalysis by PARP10 limits its activity to mono-ADP-ribosylation, Kleine [/bib_ref]. PARylation in mammal cells plays a crucial role in cellular functions, including mitosis, DNA repair and cell death [bib_ref] Matic, I. & Ahel, I. Specificity of reversible ADP-ribosylation and regulation of..., Crawford [/bib_ref]. Among the seventeen PARP enzymes identified in the human genome [bib_ref] New directions in poly(ADP-ribose) polymerase biology, Bock [/bib_ref] , only Poly(ADP-ribose) polymerase-1 (PARP1 or ARTD1), PARP2, PARP3, PARP4, Tankyrase1 (TNKS1, also known as ARTD5 or PARP5a) and Tankyrase2 (TNKS2, also known as ARTD6 or PARP5b) are capable of catalysing poly-(ADP-ribosyl)ation, whereas PARP10, PARP12, PARP14 and PARP15 are mono-(ADP-ribosyl)transferases [bib_ref] New directions in poly(ADP-ribose) polymerase biology, Bock [/bib_ref]. The remaining members of the family, PARP9 and PARP13, appear to be enzymatically inactive [bib_ref] The recognition and removal of cellular poly(ADPribose) signals, Barkauskaite [/bib_ref]. Among them, human PARP-1 (hPARP1) is the most abundant and most active protein in the PARP family, being a nuclear chromatin-associated protein [bib_ref] The recognition and removal of cellular poly(ADPribose) signals, Barkauskaite [/bib_ref]. It is also the best-studied protein in the PARP family since monotherapy with PARP-1 inhibitors selectively kills tumours harbouring deficiencies in BRCA1 and BRCA2 genes, which are involved in homologous recombination DNA repair pathway [bib_ref] Laying a trap to kill cancer cells: PARP inhibitors and their mechanisms..., Pommier [/bib_ref]. This 'synthetic lethality' has attracted clinical attention over the years as more potent and selective inhibitors have been identified. Several clinical trials are currently being conducted with them as a form of 'personalized' cancer therapy hPARP1 has a modular architecture comprising six domains [bib_ref] Structural basis for DNA damage-dependent poly(ADP-ribosyl)ation by human PARP-1, Langelier [/bib_ref]. The N-ter site consists of two zinc finger domains (Zn1 and Zn2) that recognize the damaged DNA ends, and a third zinc finger domain (Zn3) that intervenes in DNA-dependent activation. There is also a central BRCA C-terminal-like domain (BRCT) that modulates protein-protein interactions and accomplishes PAR self-modification, and a tryptophan-glycine-arginine (WGR) domain that is important for DNA-dependent activation after interaction with DNA. The last portion of the protein is the catalytic domain, which has an α-helix domain serving in the allosteric regulation (PARP_reg) followed by an ART domain (PARP_cat), which contains the conserved catalytic glutamate [bib_ref] Structural basis for DNA damage-dependent poly(ADP-ribosyl)ation by human PARP-1, Langelier [/bib_ref]. The last three domains (WGR-PARP_reg-PARP_cat) are also found in hPARP2 and hPARP3 but fused with a variable N-ter tail, as well as in most eukaryotes except for yeasts [bib_ref] Structural Implications for Selective Targeting of PARPs, Steffen [/bib_ref]. Nevertheless, the number of sequences in prokaryotes is reduced to only 28 PARP homologue sequences in 27 bacterial species [bib_ref] Distribution of protein poly(ADP-ribosyl)ation systems across all domains of life, Perina [/bib_ref]. Curiously, its activity has only been experimentally tested by western blot with anti-PAR antibodies with a recombinant enzyme cloned from the filamentous predatory gram-negative bacterium Herpetosiphon aurantiacus [bib_ref] The structure and catalytic mechanism of a poly(ADP-ribose) glycohydrolase, Slade [/bib_ref]. The above enzyme (HaPARP) was active in the presence of activated DNA and inhibited with the hPARP1 inhibitor KU-0058948 [bib_ref] The structure and catalytic mechanism of a poly(ADP-ribose) glycohydrolase, Slade [/bib_ref]. In addition, H. aurantiacus also has a DUF2263 protein (UniProt code: T3D766) that is capable of effectively removing PAR [bib_ref] The structure and catalytic mechanism of a poly(ADP-ribose) glycohydrolase, Slade [/bib_ref] , and whose sequence contains a poly (ADP-Ribose) glycohydrolase (PARG) signature (GGG-X 6-8 QEE) [bib_ref] Identification of three critical acidic residues of poly(ADP-ribose) glycohydrolase involved in catalysis:..., Patel [/bib_ref]. Thus, the presence of both enzymes in the same microorganism suggests that certain bacteria may have a functional PAR metabolism [bib_ref] Distribution of protein poly(ADP-ribosyl)ation systems across all domains of life, Perina [/bib_ref]. Another example of a microorganism with both putative PARP and PARG homologues is the rod-shaped, Gram-positive spore-forming anaerobic bacillus Clostridium difficile CD160, now reclassified as Clostridioides difficile CD160 [bib_ref] Clostridium difficile infection: a comprehensive review, Kachrimanidou [/bib_ref] [bib_ref] Reclassification of Clostridium difficile as Clostridioides difficile (Hall and O'Toole 1935) Prevot..., Lawson [/bib_ref]. C. difficile is a nosocomial opportunistic antibiotic-associated pathogen of humans responsible for a spectrum of diseases known collectively as C. difficile infections (CDI), which range from a mild self-limiting diarrhoea to pseudomembranous colitis and toxic megacolon. These pathologies often result in death, causing 29,000 deaths and costs in excess of US 6.0 billion dollars per annum in the USA alone [bib_ref] Cost of hospital management of Clostridium difficile infection in United States-a meta-analysis..., Zhang [/bib_ref]. The major virulence factors of this microorganism are the potent A-type monoglycosyltransferases toxins A (TcdA) and B (TcdB) that attach glucose to Rho proteins using UDP-glucose as a cosubstrate [bib_ref] Glucosylation of Rho proteins by Clostridium difficile toxin B, Just [/bib_ref] , and the C. difficile transferase (CDT) toxin, frequently produced by so-called hypervirulent strains. CDT is a two component toxin, CDTa being involved in the ADP-ribosylating activity and the CDTb in binding 21 . All the above-mentioned toxins are located within two defined loci, the PaLoc (Pathogenic Locus, with tdcA, tdcB, tdcR, tdcE and tdcC genes) and the CdtLoc (with cdtA and cdtB genes). Interestingly, non-toxigenic C. difficile strains (which lack the genes for toxins A and B and the binary toxin) are also relatively common, although little is known about their biology [bib_ref] Draft genome sequence of the nontoxigenic Clostridium difficile strain CD37, Brouwer [/bib_ref]. The same is the case with C. difficile CD160, which was isolated from a clinical stool at the University of Michigan Medical Centre, to test for its toxigenicity. However, the evaluation of this isolate by PCR with two different primers revealed that it was negative for C. difficile toxin genes A and B (tcdA − and tcdB − ). Nevertheless, C. difficile CD160 was selected for sequencing because CE-PCR analysis showed it to be a unique ribotype (Prof. Seth Walk, Montana State University, USA, personal communication). Apart from this uniqueness, only one protein has been studied from this microorganism, a type IV pilin region (PilA1) with an unusual conformation compared with PilA1 from C. difficile R20291 and NAP08 strains [bib_ref] Structural and evolutionary analyses show unique stabilization strategies in the type IV..., Piepenbrink [/bib_ref]. The present work makes a comprehensive phylogenetic analysis of bacterial PARPs to better understand the domain organization of these enzymes. Surprisingly, the domain organization in C. difficile CD160 PARP (CdPARP) differed from that of other clostridial PARPs. A detailed study of this anomalous distribution suggested the fact that C. difficile CD160 long ago diverged from other C. difficile strains and also defined a new toxigenotype. In addition, the presence of an active DUF2263 protein in C. difficile CD160 (CdPARG) revealed the existence of a functional PAR metabolism. Kinetic characterization also revealed CdPARP to be a highly active enzyme, with an activity that was 3-fold higher than that observed previously in HaPARP. Finally, the inhibition profile of the bacterial PARPs studied was also different from that of hPARP1, providing new information for the development of novel bacterial inhibitors with well-defined selectivity. # Results ## Cdparp shows an atypical parp domain organization compared with other clostridial parps. A phylogenetic analysis of bacterial PARPs was carried out to compare the sequence and domains of CdPARP, using the data available in the UniProt and NCBI databases. Seventy-two sequences were found containing at least the PARP domain (PARP_cat) (Supplementary Table S1) compared to the 28 previously described in the bibliography [bib_ref] Distribution of protein poly(ADP-ribosyl)ation systems across all domains of life, Perina [/bib_ref]. These sequences are distributed in six phyla (Actinobacteria, Bacteriodetes, Choloflexi, Cyanobacteria, Firmicutes, and Proteobacteria) and twelve orders [fig_ref] Figure 1: Phylogenetic analysis of bacterial PARPs [/fig_ref]. The phylogenetic tree shows that these sequences are divided into two large clades. Clade 1 groups all the sequences of two orders, Cytophagales (Clade SCIentIFIC RepoRts | (2018) 8:8056 | DOI:10.1038/s41598-018-26450-0 1.1) and Clostridiales (Clade 1.2), with the exception for the sequence of CdPARP, which is in Clade 2 (see below). Almost all members of Clade 1 are characterized by the presence of WGR and PARP_cat domains without a defined PARP_reg domain. However, when these sequences are modelled, their corresponding three-dimensional structures show the presence of a helical domain, which could fulfil the same regulatory role as the canonical PARP_reg domain does. Clade 1.1 also contained the longest bacterial PARP used in this study (764 amino acids), which corresponded to one of the two PARPs reported for the bacterium Microscilla maritima, and whose origin seems to be linked to a fusion of a tail of 350 residues in the N-ter. In addition, Bacteroidetes bacterium ADurb.BinA174 PARP (UniProt code: A0A1V5GYX3) is a sister clade of Clade 1.1. On the other hand, Clade 2 is characterized by the balanced presence of both WGR-PARP_reg-PARP_cat and WGR-PARP_cat domain configurations. In the latter case, modelling of the sequences again shows the presence of a helical domain such as that found in the proteins of Clade 1. An example of this dichotomy can be found in Clade 2.1, which consists of four Deltaproteobacteria sequences with the two domain organizations described above. The rest of the Deltaproteobacteria found in the databases were encountered in Clade 2.2, where they formed a sister group with different bacilli. All members of this Clade 2.2, including CdPARP, have the canonical WGR-PARP_reg-PARP_cat organization, an architecture also found in Clades 2.3 and 2.4, formed by proteins belonging to the orders Herpetosiphonales (Clade 2.3), Pleurocapsales, Pseudomonadales and Vibronales (Clade 2.4). By contrast, in Clade 2.5, constituted by two Burkholderiales and one Firmicutes bacterium, the predominant organization was the WGR-PARP_cat. Finally, Clade 2.6 contains all the actinobacterial sequences found in this study, divided into two large groups, belonging to the Micrococcales and Corynebacteriales orders, the latter being the most abundant. The only member of this group with the three domains is the PARP of Leifsonia sp. Leaf264, a microorganism of the Arabidopsis leaf microbiota. The MISTIC server 27 (mutual information server to infer coevolution) was used to analyse and visualize the extent of the coevolutionary relationship between two positions in the bacterial PARP protein family by using the information contained within the above MSA (multiple sequence alignment) [bib_ref] Mutual information in protein multiple sequence alignments reveals two classes of coevolving..., Gloor [/bib_ref]. The mutual information (MI) obtained is usually used to find structurally or functionally important positions in a given protein fold family [bib_ref] Integrative view ofalpha2,3-sialyltransferases (ST3Gal) molecular and functional evolution in deuterostomes: significance of..., Petit [/bib_ref] [bib_ref] Characterization and mutational analysis of a nicotinamide mononucleotide deamidase from Agrobacterium tumefaciens..., Martinez-Monino [/bib_ref]. The residue-based Kullback-Leibler conservation score , coloured rectangles) and cumulative mutual information score (cMI) for each residue , black histograms) were fairly similar, both revealing that highly coevolving residues were primarily localized in the donor site of the catalytic domain (N316-Y432 in CdPARP; , and in the DNA binding site of the WGR domain (V20-N100, . In addition, some residues of PARP_reg domain are also highly coevolving (K163-Y239, . The three main regions of the donor site corresponding to a nicotinamide (NI site; G323, Y359, S367, Y370 and E428; , green circles), a phosphate (PH site; D234 and D237; , black circles) and an adenine-ribose binding (AD site; D241, H322, S324, Y336 and K338; , blue circles) sites are in zones characterized by high cMI scores . Among these residues, especially conserved are those involved in the catalytic triad (H322, Y359 and E428; , red stars). The circos representation also shows three areas of high cMI scores in the WGR domain (N34-Y39; Y57-G61 and K88-Y93), which are related with the DNA-dependent activity of PARPs, the highest being those corresponding to G58 and R59. Of note is that, in the bacterial sequences used, the first position of WGR zone is mostly occupied by a tyrosine (Y), followed by a tryptophan (W) or a phenylalanine (F). In the PARP_reg domain, two leucines (L181 and L236) involved in the stabilization of PARP_reg hydrophobic core are also highly conserved. These two leucines correspond to L269/L233 and L321/L286 in hPARP2/hPARP3, respectively, whose alanine mutants showed an increase in DNA-independent activity, mimicking of distortion of the helical domain produced after DNA binding [bib_ref] PARP-2 and PARP-3 are selectively activated by 5′ phosphorylated DNA breaks through..., Langelier [/bib_ref]. ## Figure 2. Circos representation of the bacterial PARPs. The outer ring shows the amino acid code corresponding to CdPARP (Uniprot code; T3DQ72). Coloured rectangular boxes of the second circle indicate the KL (Kullback-Leibler) conservation score (from red to cyan, red: highest; cyan: lowest) [bib_ref] MISTIC: Mutual information server to infer coevolution, Simonetti [/bib_ref]. The third circle shows the cMI (cumulative Mutual Information score) scores as histograms. Lines in the centre of the circle connect pairs of positions with MI (Mutational Information) score > 6.5. Red lines represent the top 5%, the black lines between 70 and 95%, and the grey lines account for the last 70%. Sequence distribution of WGR, PARP regulatory and PARP catalytic domains are shown. Amino acids belonging to the catalytic triad are marked with red stars, the amino acids from the NI site, PH site and AD site are marked with green, black, and blue circles, respectively. SCIentIFIC RepoRts | (2018) 8:8056 | DOI:10.1038/s41598-018-26450-0 The results described above, along with those found in [fig_ref] Figure 1: Phylogenetic analysis of bacterial PARPs [/fig_ref] , which shows that the 12 bacterial sequences corresponding to each of the 12 different orders found are grouped together with DNA-repairing eukaryotic PARPs in Clade 1 [bib_ref] Distribution of protein poly(ADP-ribosyl)ation systems across all domains of life, Perina [/bib_ref] [bib_ref] Evolutionary history of the poly(ADP-ribose) polymerase gene family in eukaryotes, Citarelli [/bib_ref] , may indicate that the proposed organization into three structural domains for the last common ancestor of all eukaryotes may occasionally have been acquired by bacteria through horizontal gene transfer, in some cases remaining intact and in others changing their canonical sequence but not theirs helical structure. This conclusion corroborates what Ahel's group previously proposed Genetic organization of C. difficile CD160 reveals a new toxinotype. In order to explain the above anomalous localization of CdPARP in a clade other than that of Clostridiales, a maximum likehood tree from the cdu1 genes of at least three C. difficile strains for each of the six clade already described 24 was generated, including for the first time the corresponding gene from C. difficile CD160 (Uniprot code: QEW_0946) . The result showed that C. difficile CD160 forms a sister group within Clade C-I, a highly divergent lineage containing toxigenic and non-toxigenic strains [bib_ref] Evolutionary history of the Clostridium difficile pathogenicity locus, Dingle [/bib_ref] , including SA10-0505 and CD10-165 strains. These latter strains, along with toxinotypes XXX and XXXI, are a group of C. difficile that lack a complete tcdA gene while preserving tcdB and CDT genes (toxinotype A − B + CDT + ); however, CD160 strain is a completely new toxinotype (A − B + CDT − ) with the same PaLoc organization as toxigenic SA10-0505 and CD10-165 strains (cdtR, tcdR, tcdB, tcdE) but without cdtB-cdtA genes . CD160 strain also contains the endolysin cwlH gene corresponding to a N-acetylmuramoyl-L-alanine amidase, which corroborates the phage origin of the PaLoc. In addition, the homology of the TcdR (UniProt code: T3DH32) and CdtR (UniProt code: T3DFV6) regulators in CD160 and those of the reference strain CD630 (Uniprot codes: Q189K4 and Q182U3 respectively) is low (73% and 52%, respectively), but in the range of values described for SA10-0505 and CD10-165 strains (70% and 62%, respectively). These low values suggest a long-term divergence of CD160 strain compared not only with those of the C. difficile strains already studied but perhaps also with other clostridiales; indeed, a different origin for the C. difficile CD160 PARP gene is plausible. C. difficile CD160 has functional PAR metabolism. In order to test whether or not CdPARP is a bona fide DNA-dependent PARP as predicted by in silico analysis, it was cloned into pET28a vector and transformed into E. coli Rosetta 2(DE3). In addition, due to the great variability that the bibliography mentions for kinetic parameters depending on the assay used, two additional PARPs with demonstrated DNA-dependent PARP activity were also cloned: the full-length hPARP1 (1-1014) into pET24b vector and the bacterial HaPARP into pET28a vector [bib_ref] The structure and catalytic mechanism of a poly(ADP-ribose) glycohydrolase, Slade [/bib_ref] [bib_ref] Large-scale preparation and characterization of poly(ADP-ribose) and defined length polymers, Tan [/bib_ref]. The soluble recombinant proteins obtained at 20 °C after induction with IPTG were isolated in three simple steps, as described in Materials and Methods. SDS-PAGE pure enzymes were obtained with their corresponding molecular masses of 113.5 kDa, 52 kDa and 47 kDa for hPARPl, CdPARP and HaPARP, respectively . CdPARP rendered the highest yield of purified protein (4.5 mg/L culture), followed by HaPARP (3.9 mg/L culture), whereas hPARPl had the lowest (1.2 mg/L). HaPARP and hPARP1 DNA-dependent automodification has previously been demonstrated by western blot and PAR antibodies [bib_ref] The structure and catalytic mechanism of a poly(ADP-ribose) glycohydrolase, Slade [/bib_ref]. When this process was assayed with CdPARP, using hPARP1 as a control, both PARPs showed a noticeable shift in mobility as seen by western blot using both NAD + and activated DNA as substrates [fig_ref] Figure 4: PARylation assay of CdPARP and hPARP1 [/fig_ref] , thus indicating poly(ADP-ribosyl)ation of the corresponding PARP. This process was also sensitive to rucaparib, a well-known hPARP1 inhibitor [fig_ref] Figure 4: PARylation assay of CdPARP and hPARP1 [/fig_ref]. In addition, the PARylation shown by hPARP1 and CdPARP was abolished by recombinant protein corresponding to the Clostridioides difficile CD160 QEW_4455 gene (Uniprot code: T3D766). This uncharacterized protein has a DUF2263 domain similar to that of hPARG and HaPARG, which also reverses PARylation [fig_ref] Figure 4: PARylation assay of CdPARP and hPARP1 [/fig_ref]. These results demonstrate that C. difficile CD160 has a functional PAR metabolism with both functional CdPARP and PdPARG, as has also been described for H. aurantiacus Biochemical characterization of recombinant CdPARP. The kinetic parameters of hPARP1 were determined using different assay methods with PARP alone or combined with histones, together with NAD + (radioactive, biotinylated or as ADP-ribose-pNP) and with activated DNA or double-stranded DNA oligomers [bib_ref] A colorimetric substrate for poly(ADP-ribose) polymerase-1, VPARP, and tankyrase-1, Nottbohm [/bib_ref] [bib_ref] The Zn3 domain of human poly(ADP-ribose) polymerase-1 (PARP-1) functions in both DNA-dependent..., Langelier [/bib_ref]. This resulted in different K M (0.06-0.27 mM), k cat (0.41 s −1 ) and k cat /K M (1.47-6.95 mM −1 s −1 ) values [bib_ref] Molecular mechanism of poly(ADP-ribosyl)ation by PARP1 and identification of lysine residues as..., Altmeyer [/bib_ref]. Since no kinetic parameters have been described for any bacterial PARP, recombinant hPARP1 was used as a reference to validate the fluorescent method used in this work, which measures the consumption of substrate by the chemical conversion of the remaining NAD + into a stable fluorescent condensation product upon treatment by acetophenone in ethanol at basic pH, followed by heating at acidic conditions [bib_ref] An enzymatic assay for poly(ADP-ribose) polymerase-1 (PARP-1) via the chemical quantitation of..., Putt [/bib_ref] [bib_ref] Development and validation of high-throughput screening assays for poly(ADP-ribose) polymerase-2 inhibitors, Zhu [/bib_ref]. Among all the enzymes used, hPARP1 was the most active with a k cat /K M at 3.06 × 10 mM −1 s −1 , value 5.2-and 17-fold higher than those of CdPARP and HaPARP, respectively . These results highlighted the important contribution of the three zinc fingers and BRCT domain for the activity, which results in a lower K M and a higher k cat . As regards bacterial PARPs, both showed a similar K M but the k cat was higher in CdPARP, resulting in a 3-fold higher catalytic efficiency compared with HaPARP . These data, together with the purification yield of CdPARP, point to this enzyme as a promising biocatalyst for PAR production. The thermal stability of the above PARPs was also studied under different conditions by monitoring Sypro Orange fluorescence while heating the samples, and determining the midpoints of the transitions (melting temperature or Tm) [bib_ref] Characterization and mutational analysis of a nicotinamide mononucleotide deamidase from Agrobacterium tumefaciens..., Martinez-Monino [/bib_ref]. Previous investigations showed that this assay provides a good estimate of binding affinity, optimal storage pH and the suitable protein stabilizers to be used [bib_ref] Family-wide chemical profiling and structural analysis of PARP and tankyrase inhibitors, Wahlberg [/bib_ref] [bib_ref] Structural and functional analysis of Oceanobacillus iheyensis macrodomain reveals a network of..., Zapata-Perez [/bib_ref]. To carry out such study, the melting temperature of each enzyme in MilliQ ® water was taken as a reference, and thus the increment in Tm (ΔTm) was being the most stable. When the effect of pH on enzyme stability was studied [fig_ref] Table 2: IC 50 values of well-known and new compounds against hPARP1, PdPARP and... [/fig_ref] , both hPARP1 and CdPARP showed higher increments in Tm at pH 7.5, while, in the case of HaPARP the values were higher at pH 8.0-8.5. However, pH values below pH 7.0 or above pH 8.5 decreased the stability of these enzymes. The protein-stabilizing compounds used (ammonium sulphate and hydroxyectoine), while producing an increase in their Tm with all the enzyme, presented a differential effect between them [fig_ref] Table 2: IC 50 values of well-known and new compounds against hPARP1, PdPARP and... [/fig_ref]. Thus, ammonium sulphate was the best stabilizer for CdPARP, while hydroxyectoine was the best stabilizer for HaPARP and hPARP1. This information may be relevant for the case of hPARP1 as it is a commercial enzyme. In fact, when comparing the stability of hPARP1 in the presence of glycerol (10%) and hydroxyectoine (200 mM) at 4 °C, the latter compound preserved 100% of the activity after 24 h, while in the presence of glycerol it is reduced to 75% [fig_ref] Figure 5: PARP inhibitor EB-47 binding site [/fig_ref]. This proportion is maintained even at 72 hours, where the activity in the presence of glycerol is only 8% compared with 32% in the presence of hydroxyectoine [fig_ref] Figure 5: PARP inhibitor EB-47 binding site [/fig_ref]. Finally, NAD + and ADP-ribose had a destabilizing effect at 1 mM, but nicotinamide and the inhibitor 3-aminobenzamide at the same concentration increased the Tm of all enzymes, especially in the case of HaPARP, where the Tm increase was 7 °C [fig_ref] Table 2: IC 50 values of well-known and new compounds against hPARP1, PdPARP and... [/fig_ref]. This increase in Tm due to the presence of 3-aminobenzamide has previously been observed with an hPARP1 construct comprising the domains PARP_reg-PARP_cat (K654-L1013), but not with the entire protein 40 . Bacterial PARPs are specifically inhibited by a compound that mimics NAD + . The inhibitory profile of bacterial PARPs was assessed using a subset of small molecules that were originally selected following an in silico high-performance screening campaign on two complementary libraries of 50,000 compounds each (DIVERSet-EXP and DIVERSet-CL, ChemBridge), designed to provide the greatest coverage of pharmacophore while maintaining structural diversity. In addition, nine well-known hPARP1 inhibitors were also tested using a more sensitive fluorescent method due to the high potency of some of them. Thus, the remaining NAD + was amplified by coupling the enzymes alcohol dehydrogenase and diaphorase. Each time the NAD + was recycled through these enzymes, a highly fluorescent resorufin molecule was generated. This assay was first used to determine the enzyme activity of NMN-adenyltransferase and ADP-ribosyl cyclase [bib_ref] A novel cycling assay for cellular cADP-ribose with nanomolar sensitivity, Graeff [/bib_ref] [bib_ref] A novel cycling assay for nicotinic acid-adenine dinucleotide phosphate with nanomolar sensitivity, Graeff [/bib_ref]. The selected PARP inhibitors and analogues were initially screened at 10 µM and 50 µM with both hPARP1 and bacterial PARPs (CdPARP and HaPARP), respectively (Supplementary . Of the 44 compounds analysed, twenty-three showed more than 50% inhibition with hPARP1 . Half-inhibitory concentration (IC 50 ) values were calculated for the first 17 hPARP1 inhibitors, obtaining very similar values for the 8 inhibitors already described in the bibliography [fig_ref] Table 2: IC 50 values of well-known and new compounds against hPARP1, PdPARP and... [/fig_ref]. After the 5 highly selective known inhibitors (rucaparib, ABT-888, PJ-34, EB-47 and DPQ) surprisingly, four compounds derived from 4-substituted 3-nitrophenyl-1(2 H)-phthalazinone appeared [fig_ref] Table 2: IC 50 values of well-known and new compounds against hPARP1, PdPARP and... [/fig_ref] ; . Among them, standing out for its selectivity, was compound 7655698 with an IC 50 of 81 nM, which, to the best of our knowledge, has never been used as hPARP1 inhibitor, and which could be optimized to increase its potency and selectivity. The others members of this family of compounds with IC 50 values between 136-481 nM have been tested as modulators of cytokine activity 44 or maternal embryonic leucine zipper kinase [bib_ref] Novel inhibitor discovery through virtual screening against multiple protein conformations generated via..., Mahasenan [/bib_ref] , and only one (7660328) has been used to calculate its Tm increment with a hPARP1 construct (K654-L1013) [bib_ref] Exploring the effect of PARP-1 flexibility in docking studies, Antolin [/bib_ref]. This was also the case with compound 7650155 (1,3-dioxo-2,3-dihydro-1H-benzo[de]isoquinoline-5-sulfonamide) [bib_ref] Exploring the effect of PARP-1 flexibility in docking studies, Antolin [/bib_ref] , which together with another two [fig_ref] Table 2: IC 50 values of well-known and new compounds against hPARP1, PdPARP and... [/fig_ref]. The latter inhibitor was designed to mimic the NAD + within the hPARP1 substrate-binding site [bib_ref] The discovery and synthesis of novel adenosine substituted 2,3-dihydro-1H-isoindol-1-ones: potent inhibitors of..., Jagtap [/bib_ref] and has been co-crystallized with tankyrase-2 with an IC 50 of 45 nM 48 . When a structural alignment of this tankyrase-2 with EB-47 (PDB: 4BJ9) was carried out, both with hPARP1 (PDB: 4opx) and with the modelled CdPARP and HsPARP, it was observed that at the nicotinamide site the residues that form the hydrogen bonds (G1032 and S1068; TKNS2 numbering) and produce π-π stacking (Y1071) with isoindolinone moiety are also conserved in the rest of the studied PARPs [fig_ref] Figure 5: PARP inhibitor EB-47 binding site [/fig_ref]. In addition, the residues involved in binding the ribose hydroxyls (H1031 and S1033) and the two H-bonds with the adenosine (G1043 and D1045) were also structurally conserved [fig_ref] Figure 5: PARP inhibitor EB-47 binding site [/fig_ref]. The different IC 50 of the above PARPs and TKNS2 could be related with the specific interaction of EB-47 with the D-loop of each enzyme, as previously described for TKNS2 48 . # Discussion Poly-ADP-rybosilation is a remarkably, well-conserved post-transcriptional modification in eukaryotes, but it is less common in bacteria. This was also observed in the phylogenetic studies carried out, mainly focusing on the evolutionary distribution of the poly(ADP-ribose) polymerases in eukaryotes. In this study, we looked into bacterial PARPs, not only as regard their phylogenetic distribution, by representing their first phylogenetic three [fig_ref] Figure 1: Phylogenetic analysis of bacterial PARPs [/fig_ref] , but also the key amino acids responsible for catalysis, and DNA-dependent activation, using the coevolutionary relationship between two positions provided by the MISTIC server. With this aim, we focused on the C. difficile CD160 putative protein corresponding to QWE_4625 gene, since its protein in the PFAM database indicates the presence of a WGR-PARP_reg-PARP_cat canonical organization. The updated phylogenetic analysis was carried out with 72 sequences rather than the 27 previously used. Unexpectedly, the result showed that even when PFAM did not found in most of the protein sequences the PARP_reg domain, the corresponding modelled sequences displayed a similar helical domain to the canonical PARP_reg domain, which is necessary for the destabilization of PARP_cat domain after DNA-binding at the WGR domain. The study of this last domain produced two important findings: an YGR domain exits in most of bacterial PARPs along with other highly coevolved amino acid, which have never before been studied, such as G92 and Y93 . In addition, the organization of the CdPARP anomalous domain compared with that of other clostridial PARPs suggests a long-term divergence of CD160 strain, based on its cdu1 gene phylogeny and homology in TcdR and CdtR. Of note too, is that the latter, poorly studied strain (just one paper mentions it), presents a new toxinotype never before described (A − B + CDT − ) with an atypical "Bi-Toxin Paloc" comprising only the CdtR, tcdR, tcdB, tcdE, CwlH genes. Surprisingly, this PaLoc organization differs from that of the closely related SA10-50 and CD10-165 strains. Furthermore, this microorganism has also a functional bona fide PAR metabolism, not only by the existence of PARP [fig_ref] Figure 4: PARylation assay of CdPARP and hPARP1 [/fig_ref] , but also by the presence of a putative MacroD (T3DIR9), which is the enzyme predictably responsible for the removal of the terminal ADP-ribose unit after PARG has acted, giving rise to a completely unmodified protein [bib_ref] Macrodomains: Structure, Function, Evolution, and Catalytic Activities, Rack [/bib_ref]. Surprisingly, no ARH3-like protein was found in UniProt for C. difficile CD160, whereas three genes were found in H. auranticus ATCC 23779 (Haur_3413, Haur_2362 and Haur_0057), in addition to the existing three MacroD genes (Haur_0008, Haur_4555 and Haur_2355). The presence of different genes involved in the PAR metabolism in Bacteria, together with the presence of sirtuins-like genes would seem to be rather more than a random horizontal gene transfer phenomenon, and more likely an adaptive advantage that has been preserved over time. Such an advantage has been recently described in a new toxin-antitoxin (TA) system, which possibly contributes to bacterial persistence 50 . The latter system is based on the combined action of a toxin protein (DarT) that specifically leads to a sequence-specific ADP-ribosylation on single-stranded DNA and an antitoxin macrodomain protein (DarG) that reverse such DNA modification [bib_ref] The Toxin-Antitoxin System DarTG Catalyzes Reversible ADP-Ribosylation of DNA, Jankevicius [/bib_ref]. Finally, our data revealed different inhibitory profiles between hPARP1 and bacterial PARPs, the latter enzymes being specifically inhibited by EB-47, a compound that mimics NAD + [fig_ref] Table 2: IC 50 values of well-known and new compounds against hPARP1, PdPARP and... [/fig_ref]. However, the major discovery in this study was a compound, never before used with PARP, that selectively inhibited hPARP1 (compound 7655698) but not the bacterial PARPs. The absence of Pan-PARP inhibition for bacterial PARPs, as produced by rucaparib [fig_ref] Table 2: IC 50 values of well-known and new compounds against hPARP1, PdPARP and... [/fig_ref] , makes this compound as an ideal selective inhibitor for seeking further improvements in its potency. Taken together, the above described results will provide the foundation for new future bacterial PAR metabolism studies and the development of new specific PARP inhibitors. # Materials and methods Protein expression and purification. Bacterial and human PARP and PARG proteins were expressed from the pET28a vector (EMD Millipore, Madrid, Spain) bearing an N-terminal His-tag, with the exception of the hPARP1 [fig_ref] Figure 1: Phylogenetic analysis of bacterial PARPs [/fig_ref] , which has a C-terminal His-tag and was kindly provided by J. M. Pascal (Université de Montréal, Montréal, QC) in a pET24b vector. PARP (QEW_4625) and PARG (QEW_4455) genes from Clostridioides difficile CD160 (BioProject and SRA Accession numbers PRJNA85757 and SRR593185, respectively) were obtained from GenScript (Piscataway, USA). Genomic DNA from Herpetosiphon aurantiacus DSM785, purchased from DSMZ (Braunschweig, Germany), was used to clone PARG (Haur_1618) and PARP (Haur_4763) genes. hPARG-pCMV6-XL5 construct was purchased from Origene (Rockville, USA). Genes were amplified by PCR using KAPA HiFi DNA polymerase and the corresponding primers, including NheI and XhoI restriction site extensions. The cloning and transformation techniques used were essentially those previously described [bib_ref] Molecular characterization of a novel N-acetylneuraminate lyase from Lactobacillus plantarum WCFS1, Sanchez-Carron [/bib_ref]. All PARP constructs were expressed in Escherichia coli strain BL21(DE3) Rosetta2. The clones containing the constructions were grown in 1 litre of Terrific Broth (TB) medium supplemented with kanamycin (50 µg/mL) and chloramphenicol (30 µg/mL). When the culture reached an optical density of 2.5 at 600 nm, it was induced with 0.2 mM isopropyl-β-D-thiogalactoside (IPTG) for hPARP1 or with 0.1 mM IPTG for both HaPARP and CdPARP for 16 h at 20 °C with constant shaking. Pelleted cells were resuspended in lysis buffer (50 mM HEPES pH 7.5, 250 mM NaCl, 1 mM β-mercaptoethanol, 1 mM PMSF, 1 mM benzamidine and cOmplete EDTA-free protease inhibitor cocktail) before being disrupted by sonication (450-D Sonifier, Branson). After ultracentrifugation (40000 g, 40 min), the supernatant was incubated with Ni-NTA beads (Macherey-Nagel, Germany) at 4 °C for 1 h. Then, they were washed with lysis buffer, and the protein was eluted with the same buffer containing 250 mM imidazole. The PARP-containing fractions were loaded onto a HP heparin column coupled to a FPLC chromatography system (ÄKTA Prime Plus, GE Lifesciences) [bib_ref] Purification of human PARP-1 and PARP-1 domains from Escherichia coli for structural..., Langelier [/bib_ref] , and eluted with 50 mM Tris-HCl pH 7.5, 1 M NaCl, 1 mM EDTA and 1 mM β-mercaptoethanol. Finally, the protein was loaded onto a Superdex 200 HiLoad 16/600 column (GE Lifesciences), obtaining an electrophoretically pure enzyme. PARP aliquots were stored at −20 °C with 10% glycerol. All PARG constructs were expressed as above described but induced with 0.2 mM IPTG when the culture reached an optical density of 4.0 at 600 nm in TB medium. After ultracentrifugation (40000 g, 40 min), the resulting supernatant was purified in two steps comprising a Ni 2+ -chelating affinity chromatography (HiPrep IMAC 16/10 FF column) followed by gel filtration onto a Superdex 200 HiLoad 16/600 column (GE Life Sciences). PARG aliquots were also stored at −20 °C with 10% glycerol. The protein concentration was determined using Bradford reagent (Bio-Rad) and BSA as standard. PARP activity assays. PARP activity was determined by a fluorescence-based assay based on the reaction of N-alkylpyridinium compounds, such as NAD + , with acetophenone to give rise a fluorescent compound after heating in acid [bib_ref] An enzymatic assay for poly(ADP-ribose) polymerase-1 (PARP-1) via the chemical quantitation of..., Putt [/bib_ref]. The hPARP1 automodification reaction was allowed to proceed in a black 96-well fluorescence plate (Greiner Bio-One, USA) for 20 min at 25 °C in a reaction medium containing 90 nM of hPARP1, 1 mM NAD + (Trevigen), 75 µg/mL of activated DNA (Trevigen), and PARP assay buffer (100 mM Tris-HCl pH 8.0, 10 mM MgCl and 1 mM DTT) in a final volume of 60 µL. The NAD + present after the reaction was then determined by the addition of 20 µL of an aqueous 2 M KOH solution and 20 µL of a 20% acetophenone solution in ethanol. The plate was incubated at 4 °C for 10 min. Then, 80 µL of formic acid 88% was added, incubated at 110 °C for 5 min and allowed to cool for 30 min. Fluorescence was measured on Synergy HT equipment (Biotek Instruments) at an excitation wavelength of 360 nm and emission wavelength of 445 nm. HaPARP and CdPARP activity assays were carried out as above but at 30 °C for 1 h and at enzyme concentrations of 850 nM and 500 nM, respectively. K M values were estimated using plots of initial rates vs. NAD + concentrations. Due to the potency of some of them, the inhibitors were tested using a more sensitive cycling assay involving alcohol dehydrogenase (ADH) and diaphorase (DP) [bib_ref] A novel cycling assay for cellular cADP-ribose with nanomolar sensitivity, Graeff [/bib_ref] [bib_ref] A novel cycling assay for nicotinic acid-adenine dinucleotide phosphate with nanomolar sensitivity, Graeff [/bib_ref]. ADH reduces the NAD + to NADH and oxidizes ethanol to acetaldehyde, while DP turns NADH back into NAD + with the reduction of resazurin, providing resorufin, which has an excitation maximum at 544 nm and an emission maximum at 590 nm. The reaction mixture contains enzyme (90 nM hPARP1 or 0.5 µM CdPARP or 0.85 µM HsPARP), 100 nM NAD and 75 µg/mL of activated DNA in PARP assay buffer. The inhibitors were added at different concentrations at a constant volume of 0.3 µL of DMSO. The PARP automodification reaction was allowed to proceed for 20 min at 25 °C (hPARP1) or 60 min at 30 °C (CdPARP and HsPARP). Then, 30 µL of cycling reagent, containing 2% ethanol, 50 µg/mL ADH, 50 µM Resazurin, 5 µg/mL DP, 10 µM FMN in PARP assay buffer was added. The enzyme-coupled reaction was measured over 15 min in a Synergy HT (Biotek Instruments). The compounds were screened at 10 μM (hPARP1) or at 50 μM (CdPARP and HsPARP) in triplicate, and the best compounds were selected for measuring their corresponding IC [bib_ref] The Toxin-Antitoxin System DarTG Catalyzes Reversible ADP-Ribosylation of DNA, Jankevicius [/bib_ref] Western blot. PARP automodification was also assayed by western blot in a reaction containing different concentrations of enzyme (250 nM HaPARP, 200 nM CdPARP, or 85 nM hPARPl) in assay buffer, with or without NAD + (1 mM), activated DNA or rucaparib (1 µM), respectively. The reactions were allowed to proceed for 1 h at 25 °C. PAR-mediated PARG hydrolysis was analysed using the above-mentioned PARP automodification reaction after stopping it with rucaparib, followed by the addition of PARG enzymes (500 nM) and incubation for 1 hour at 30 °C. The above reactions were then run on 7-10% SDS-PAGE gels and transferred to a nitrocellulose membrane. The presence of poly ADP-ribose (PAR) was detected by the use of rabbit polyclonal anti-PAR antibodies (1:1000; Trevigen), and goat anti-Rabbit IgG antibody conjugated to horseradish peroxidase (1:5000; Bio-Rad), and Opti-4CN as optimized colorimetric substrate (Bio-Rad). Thermal stability assay. Protein melting curves to determine the thermal stability of PARPs were obtained using the fluorescent dye SYPRO Orange (Molecular Probes), as previous described [bib_ref] Characterization and mutational analysis of a nicotinamide mononucleotide deamidase from Agrobacterium tumefaciens..., Martinez-Monino [/bib_ref] [bib_ref] Structural and functional analysis of Oceanobacillus iheyensis macrodomain reveals a network of..., Zapata-Perez [/bib_ref]. Proteins (10 µg) were preincubated with 10X Sypro Orange (excitation at 490 nm, emission at 530 nm) in the presence of different buffers (100 mM), compounds (NAD + , ADP-ribose, nicotinamide, ammonium sulphate and hydroxyectoine) or inhibitors (3-aminobenzamidine). The assay, performed in 7500 RT-PCR equipment (Applied Biosystems), monitors Sypro Orange fluorescence while heating the samples from 5 to 98 °C. Each experiment was carried out in triplicate. In silico analysis. Protein sequences were obtained from NCBI non-redundant (NR) and UniProt databases using Clostridiales difficile CD 160 PARP protein sequence as a query. Incomplete sequences and duplicates were removed, rendering the sequences described in . A Neighbour-Joining (NJ) tree with 1000 replicates was constructed using the MAFF server (https://mafft.cbrc.jp/alignment/server/). Domain architectures of retrieved sequences were obtained from the Pfam (http://www.sanger.ac.uk/Software/Pfam) and GenomeNet (http://www.genome.jp/tools/motif/) databases. Mutation correlation analysis was made with the retrieved bacterial PARP sequences, using Mistic (Mutual Information Server to Infer Coevolution) web server (http://mistic.leloir.org.ar) [bib_ref] MISTIC: Mutual information server to infer coevolution, Simonetti [/bib_ref]. The x-ray structure of hPARP1 (PDB: 2DR6) was used as model for the in silico high-performance screening campaign on two complementary libraries of 50,000 compounds each (DIVERSet-EXP and DIVERSet-CL, ChemBridge) using LeadIT and SeeSAR (https://www.biosolveit.de/). Protein sequences were modelled with SwissModel 53 . Protein structure images and structural alignments were obtained with Chimera 54 . Cdu1 gene phylogenetic three of C. difficile CD160 was constructed with MEGA 7.0 55 using maximum likelihood method and the data provided by Prof. Marc Monot (Institute Pasteur, FR). PaLoc and CdTLoc representations of C. difficile strains were obtained from Patric web (https://www.patricbrc.org/). [fig] Figure 1: Phylogenetic analysis of bacterial PARPs. The Neighbour-Joining (NJ) tree was obtained from 1000 replicates. Protein domain architecture is shown behind each protein code: WGR domain (Red), PARP regulatory domain (Blue), and PARP catalytic domain (Green). Protein codes are summarized in Supplementary Table S1. SCIentIFIC RepoRts | (2018) 8:8056 | DOI:10.1038/s41598-018-26450-0 [/fig] [fig] Figure 3 0: Phylogenetic analysis of the C. difficile cdu1 genes. Maximum likelihood tree with 1000 replicates was constructed using representative clade strains 24 . SCIentIFIC RepoRts | (2018) 8:8056 | DOI:10.1038/s41598-018-26450-Supplementary [/fig] [fig] Figure 4: PARylation assay of CdPARP and hPARP1. It was carried out with CdPARP (A) and hPARP1 (B) as control in the presence of absence of NAD + , activated DNA and the PARP inhibitor rucaparib. Numbers on the right margin indicate protein markers (kDa). K M (mM) k cat (s −1 ) k cat /K M (mM −1 s −1 ) [/fig] [fig] Figure 5: PARP inhibitor EB-47 binding site. Structural alignment of TNKS2 (PDB: 4BJ9, orange), hPARP1 (PDB: 4OPX, cyan) and modelled CdPARP (salmon) and HaPARP (green) was carried out using chimera54 . hPARP1 AD site is represented by R878, G876, H862 and S864, whereas Ni site by G863, S904 and Y907, respectively.SCIentIFIC RepoRts | (2018) 8:8056 | DOI:10.1038/s41598-018-26450-0 [/fig] [table] Table 2: IC 50 values of well-known and new compounds against hPARP1, PdPARP and HaPARP. Literature values are from different authors 47,56-62 . PARPs, only five compounds showed inhibition in excess of 50% with CdPARP, and three compounds with HsPARP (Supplementary Table S3). However, only rucaparib and EB-47 were able to inhibit them completely. When IC 50 values were determined, rucaparib showed values of 5.3 and 7.2 µM, while EB-47 showed values of 0.8 and 1.0 µM, respectively ( [/table]

Research on outpatient capacity planning combining lean thinking and integer linear programming Background The size and cost of outpatient capacity directly affect the operational efficiency of a whole hospital. Many scholars have faced the study of outpatient capacity planning from an operations management perspective.Objective The outpatient service is refined, and the quantity allocation problem of each type of outpatient service is modeled as an integer linear programming problem. Thus, doctors' work efficiency can be improved, patients' waiting time can be effectively reduced, and patients can be provided with more satisfactory medical services.Methods Outpatient service is divided into examination and diagnosis service according to lean thinking. CPLEX is used to solve the integer linear programming problem of outpatient service allocation, and the maximum working time is minimized by constraint solution.Results A variety of values are taken for the relevant parameters of the outpatient service, using CPLEX to obtain the minimum and maximum working time corresponding to each situation. Compared with no refinement stratification, the work efficiency of senior doctors has increased by an average of 25%. In comparison, the patient flow of associate senior doctors has increased by an average of 50%.ConclusionIn this paper, the method of outpatient capacity planning improves the work efficiency of senior doctors and provides outpatient services for more patients in need; At the same time, it indirectly reduces the waiting time of patients receiving outpatient services from senior doctors. And the patient flow of the associate senior doctors is improved, which helps to improve doctors' technical level and solve the problem of shortage of medical resources. # Introduction Outpatient service is one of the key service resources of the medical service system, and it is the earliest place for patients to contact for medical treatment. The efficiency of outpatient resource scheduling directly affects the operation efficiency of follow-up departments and the whole hospital [bib_ref] Enhancing implementation of a standardized initial assessment for demand management in outpatient..., Breckner [/bib_ref]. However, most medical institutions currently face the problems of a large demand for outpatient services but low operation efficiency, long waiting times, and short treatment times for patients [bib_ref] The medical arms race and its impact in Chinese hospitals: implications for..., Qian [/bib_ref]. In order to alleviate the contradiction between the supply and demand of outpatient resources, it is urgent to improve the utilization rate of medical resources [bib_ref] Questionnaire survey about use of an online appointment booking system in one..., Zhang [/bib_ref]. How to efficiently schedule and operate limited medical resources and guide patients to seek medical treatment orderly have become the primary task of outpatient scheduling research. Long waiting time in the outpatient department greatly impacts timely access to medical services and medical experience. Many scholars have studied the problem of long waiting times for patients. Young patients may arrive earlier or later than the appointment time in many cases, and patients' unpunctuality have a negative impact on the normal operation of the appointment arrangement system as planned [bib_ref] Association between age and outpatient clinic arrival time: Myth or reality? BMC, Faiz [/bib_ref]. It affects the utilization of medical resources and the efficiency of healthcare staff, which increases healthcare costs and reduces the ability of clinics to serve patient populations. Laan et al. proposed a solution of stochastic mixed integer programming based on the fluctuation of patient arrival and the unavailability of doctors to optimize appointment scheduling related to access time [bib_ref] Static and dynamic appointment scheduling to improve patient access time, Laan [/bib_ref]. Discrete event simulations are used to evaluate the effectiveness and limitations of the approach, which flexibly allocated 2% of patient flow, resulting in improved clinic performance. Zhu et al. proposed a simulation framework combined with heuristic strategies to improve the performance of the reservation scheduling system given patients' unpunctuality [bib_ref] Outpatient appointment scheduling with unpunctual patients, Zhu [/bib_ref]. This method has dramatic advantages in current practice, thus improving the patient capacity of the clinic. In recent years, the study of outpatient flow scheduling has attracted more and more attention from business and academic circles. Najmuddin et al. developed an optimized scheduling method combined with a discrete event simulation model to reduce further waiting time and increase flow in obstetric outpatients and improved the optimization effect of this method through multiple simulations run [bib_ref] A simulation approach: improving patient waiting time for multiphase patient flow of..., Najmuddin [/bib_ref]. Lenin et al. optimized the appointment model of obstetrics and gynecology clinics and distinguished three different patient types according to the possibilities of each type of patient attending appointments [bib_ref] Optimizing appointment template and number of staff of an OB/GYN clinic-micro and..., Lenin [/bib_ref]. The best solution resulted in adding a medical assistant and modifying the appointment system to eliminate bottlenecks without sacrificing resource utilization by reducing patient waiting times. Viana et al. proposed a hybrid discrete event simulation optimization model for pregnancy clinics to better plan medical resources and improve patient flow in the outpatient department [bib_ref] Capacity and patient flow planning in post-term pregnancy outpatient clinics: a computer..., Viana [/bib_ref]. This model can effectively deal with uncertain events such as the expansion of catchment areas, the modification of overdue pregnancy guidelines, and women being able to deliver before an appointment. Doctors' time is precious in medical care, and medical resources are minimal. In order to better play the role of limited resources, many scholars began to use integer programming methods to optimize resource allocation. During the COVID-19 emergency, social distancing was identified as one of the most effective measures to limit the spread of the virus. proposed an integer linear programming model to determine the optimal layout of outpatient services to reduce the congestion in the waiting room [bib_ref] A mathematical formulation for reducing overcrowding in hospitals' waiting rooms, Caselli [/bib_ref]. The experimental results on real hospital data show that this method can reduce congestion by 80% on average. In view of the uncertainty of patients' reservation demand, Aslani et al. proposed the outpatient capacity planning method using cardinality constraints [bib_ref] A robust optimization model for tactical capacity planning in an outpatient setting, Aslani [/bib_ref]. This method considers the first-visit and the second-visit patients and provides a feasible capacity allocation scheme for all reservation requirements within the allowable range of uncertain appointments. Izady et al. proposed a compact planning scheme for different reservation needs. They proposed a model of reserving free space for more reservation services, thereby reducing patients waiting time [bib_ref] An integrated approach to demand and capacity planning in outpatient clinics, Izady [/bib_ref]. By adjusting the working hours to improve patient flow, overtime costs are minimized under the premise of ensuring the constraints of waiting and appointment access. Santibanez et al. put forward a mixed integer programming model for operating room management, which arranges each specialty's operation blocks into the operating room [bib_ref] Surgical block scheduling in a system of hospitals: an application to resource..., Santibáñez [/bib_ref]. At the same time, considering the time availability of the operating room and the limitation of postoperative resources, the experimental results show that the hospital can handle more cases through different specialty arrangements without increasing postoperative resources. Lean thought originated from the lean production mode invented by Toyota in Japan. The lean production mode brought advantages in quality and cost to Japanese cars, which then made the automobile industry in other countries unable to rise. The core of lean thinking is to create as much value as possible with less manpower, equipment, time, and space [bib_ref] Implementation of lean thinking: one health system's journey, Kim [/bib_ref]. Healthcare organizations have begun to apply this approach to hospital management. The University of Michigan Health System adopts lean thinking as its unified approach to quality improvement and strives to become a completely lean organization. Many scholars also use lean thinking to improve the management efficiency of the hospital [bib_ref] Hospital innovation management under the lean thinking, Liang [/bib_ref]. Improta et al. apply lean thinking to the hospital's emergency room to increase the flow of patients, improve the process that helps promote the flow of patients in all stages of medical treatment, and eliminate all bottlenecks and activities that produce waste [bib_ref] Lean thinking to improve emergency department throughput at AORN Cardarelli hospital, Improta [/bib_ref]. Mu et al. combined lean thinking to optimize the flow of the obstetric ward, and divided the obstetric ward into an observation ward, cesarean section ward, and natural delivery ward according to lean thinking [bib_ref] Research on obstetric ward planning combining lean thinking and mixed-integer programming, Mu [/bib_ref]. The problem of how to allocate the number of wards of each type is modeled as a mixed integer programming problem, which maximizes the patient flow of pregnant women in obstetric hospitals. It increases the patient flow by 19-25% compared to before improvement. The above methods have been studied on outpatient patient flow and waiting time issues. However, there has yet to be a report on methods of planning outpatient capacity from a more refined perspective. Based on lean thinking, this paper proposes dividing outpatient service into two types: examination service and diagnosis service. Further refinement and classification of outpatient services will make it easier for doctors to manage patients with more centralized distribution, thus indirectly improving their work efficiency. Patients can enjoy more detailed medical services at different stages, thus obtaining more professional medical services in unit time and effectively reducing patient waiting time. It is a matter of planning how the amount of time allocated for each type of clinic is divided into more detailed sections. Suppose the service time of a certain type of outpatient service is allocated relatively less. In that case, there will be a bottleneck and more patients cannot make an appointment, wasting medical resources. Suppose one type of clinic has a relatively high allocation of service time and the other type has a relatively low allocation of service time. In that case, the patient will be hindered from moving to the next process. In this paper, the outpatient capacity planning problem is modeled as an integer linear programming model where the maximum outpatient working time is minimized. The model is solved by means of an integer linear programming solver. This method has an important guiding role in the planning of outpatient capacity, especially how to plan the working time of different types of outpatient service when the number of patients or doctors on duty changes greatly. # Methods ## Lean outpatient capacity planning Famous or senior doctors are scarce resources that are seriously insufficient. Patients often fail to see doctors on time, and doctors work overtime. Because of the natural advantages of their platforms, top hospitals attract patients and service leaders who should have seen doctors in other hospitals or grass-roots hospitals, which is called the siphoning effect in the industry. The siphon effect not only attracts patients from primary hospitals, but also simultaneously "hollows out" senior talents from primary hospitals. Also, due to the siphonage effect, more patients come to the top hospitals. This leads to long patient waiting times, which seriously affects the medical experience. The worse situation is that there is no appointment with a senior doctor, which also leads to the failure of some incurable diseases to receive the diagnosis and treatment services of senior doctors, seriously affecting the fairness of medical resource allocation. Senior doctors have been engaged in clinical and academic research for many years in the hospital, and have a high level of professional technology and academics. Associate senior doctors generally have been engaged in clinical and academic research for a short time, and need further study in professional technology and academia. They can only be employed as senior doctors after passing the examination. For associate senior doctors, relatively few patients choose them for treatment. Doctors often wait for patients, which leads to a waste of medical resources. More importantly, this situation severely limits the associate senior doctors' learning and accumulation of diagnosis and treatment experience for difficult diseases, which is not conducive to the growth of associate senior doctors. The general clinical pathway of patients receiving outpatient services from senior doctors is shown in [fig_ref] Figure 1 a: Clinical pathway of outpatient services [/fig_ref]. Patients line up outside senior doctors' offices during their appointments. When it is the patient's turn to receive outpatient service, the patient presents the chief complaint in front of the senior doctor, who gives medical advice after understanding the situation. The patient then pays the prescribed fee and has the corresponding medical examination as ordered. After receiving the results of medical examinations, patients return to the doctor's office and wait in line. When the patient receives outpatient service, the doctor will diagnose the disease according to the examination results and give the corresponding treatment plan. Patients should communicate with senior doctors at least twice when they receive outpatient services. In some cases, they will have a third or more communication with senior doctors to change their medical orders or ask for treatment plans. Suppose a patient has an appointment with a senior doctor. In that case, the patient needs to wait in line every time he or she receives the services of the senior doctor because the senior doctor is serving other patients. Patients have to wait in line several times to receive outpatient services from senior doctors. For patients, time is relatively tight, with the vast majority of examinations to be completed in the morning or before work. Otherwise, they have to make another doctor's appointment the next day. Usually, patients' waiting time is more than the communication time with senior doctors, and the long waiting time seriously affects patients' rest and emergency treatment, and seriously affects their medical experience. During this period, patients often occupied the time of senior doctors, wasting medical resources and failing to receive more accurate medical services. For senior doctors, when they provide fixed medical services, the time period for their services is not concentrated. Multiple visits to each patient, and senior doctors have to re-master each patient's condition, seriously consuming the energy and time of senior doctors. This has led to more effort and time for senior doctors to provide the same service. What is more, senior doctors are not serving more patients with more time and energy. Instead, fewer patients may be served because of energy consumption and time delays. The outpatient clinical pathway of the senior doctors mentioned above is the mainstream clinical pathway adopted by the hospital at present. This clinical pathway is mostly a guide in time, without planning for the doctors and space to implement the clinical pathway. On the premise of not changing the clinical pathway, this paper combined lean thinking to plan the implementation of the clinical pathway from the perspective of space and doctors. Based on lean thinking, outpatient services are divided into two types of services in more detail: examination services and diagnosis services. The examination service includes the chief complaint and the first doctor's order. Beyond that, all phases are divided into diagnosis services. In most hospitals, the classification of departments is very fine, and the senior associate doctor is fully competent for patient examination services. According to lean thinking, the examination services of patients who make appointments with senior doctors are assigned to associate senior doctors. In contrast, the diagnosis services are still assigned to senior doctors. The post-planning outpatient service flow is shown in [fig_ref] Figure 1 a: Clinical pathway of outpatient services [/fig_ref]. Patients wish to receive examination services from non-senior doctors and diagnosis services from senior doctors, see 1b. Patients will receive examination services from the senior associate doctor first. After completing the appropriate examination services, the patient is returned to the senior doctors for diagnosis services. The procedure for patients who want to receive examination and diagnosis services from senior doctors is represented in [fig_ref] Figure 1 a: Clinical pathway of outpatient services [/fig_ref]. Such lean planning has the following advantages: Patients can reduce the waiting time of patients without affecting their medical treatment. Because the number of patients treated by associate senior doctors is small, the medical experience of patients is improved; For senior doctors, energy consumption and working time are reduced so that more energy and time are available to provide services to other patients in need. For the associate senior doctors, the number of patients received has significantly increased, which helps improve the professional level but also helps to solve the problem of difficult medical treatment. This model is proposed based on many years of research on outpatient service of a class III class A hospital in China. This model has a guiding significance for managing door diagnostic services in all hospitals. This model is still in the theoretical research stage, and further application in practice is the next work to be carried out. ## Integer programming model This paper's outpatient scheduling problem is modeled as an integer linear programming model. The constraints of outpatient doctors, examination service time, diagnosis service time, and a number of patients are given below, and then the mathematical programming model of outpatient scheduling is constructed. Assume that the total number of doctors is td, the number of senior doctors is tds. Assume that the total number of patients is tp, and examination and diagnosis service times are marked as exam_time and diagnosis_ time respectively. Two-dimensional data dpat[][] is used to represent the appointment relationship between all patients and each doctor. npp[] indicates that each doctor can be transferred in or out the number of patients. Transferring in a patient refers to letting the current doctor serve the patient who has made an appointment with the other doctor. Transfer out a patient refers to letting other doctors serve the patient who has made an appointment with the current doctor. The maximum working time of each doctor is set as work_time[], and the maximum working time after scheduling optimization is set as LWT. [fig_ref] Table 1: Description of parameters and decision variables of the problem [/fig_ref] gives the description of the decision variables and parameter variables related to the mathematical model of this planning. Based on the above constraints, the integer linear programming solver is used to solve, and then the maximum working time of the outpatient department is minimized. The working time of each doctor i after scheduling given by the sum of diagnosis, examination, and transfer times should not exceed their limit on working time work_time[i]: For each senior doctor, the working time after scheduling should not exceed LWT: For associate senior doctors, the working time after scheduling should not exceed LWT: [formula] tp−1 j=0 (et + dt) * dpat[i][j] + et * npp[i] ≤ work_time[i], 0 ≤ i < td. [/formula] For senior doctors, if they have more patients booked than the average number of outpatient visits, they should call out some patients. Otherwise, some patients should be enrolled. The corresponding constraints are as follows: For associate senior doctors, if the number of patients they have booked is greater than the average number of outpatient reception, patients will not be transferred. Otherwise, some patients should be enrolled. The corresponding constraints are as follows: For the number of patients transferred in and out by all doctors, its sum should remain unchanged. The corresponding constraints are as follows: This planning problem is to solve the minimization of the maximum working time of the outpaent department, and its constraints are as follows. The objective of the considered planning problem is to minimize the maximum working time of the outpatient department, which is formulated as follows: # Results Today, more than 1000 universities and more than 100 leading software companies worldwide choose to use the IBM CPLEX optimizer to help solve planning problems in various industries . The CPLEX optimizer provides flexible, high-performance mathematical programming solutions for integer programming problems and [formula] tp−1 j=0 (et + dt) * dpat[i][j] + et * npp[i] ≤ LWT , 0 ≤ i < tds. tp−1 j=0 (et + dt) * dpat[i][j] + et * npp[i] ≤ LWT , tds ≤ i < td. ∀i, 0 ≤ i ≤ tds − 1, npp[i]        ≤ 0, tds−1 � i=0 dpat[i][j] > tp/td ≥ 0, tds−1 � i=0 dpat[i][j] ≤ tp/td . ∀i, td ≤ i ≤ td − 1, npp[i]          ≤ 0, td−1 � i=tds dpat[i][j] > tp/td ≥ 0, td−1 � i=tds dpat[i][j] ≤ tp/td . td i=0 npp[i] = 0. minimize LWT. [/formula] is available in free versions for educational and scientific research. In this paper, the above outpatient planning problems are coded in CPLEX, and CPLEX is used as a mixed integer programming solver. Our method is proposed based on many years of research and practice on the outpatient services of a class III A Chinese hospital. The data used in our experiments are generated based on the real activity of this hospital. All experiments are run on the hardware of the Windows 10 64-bit operating system. Due to the small scale of the problem and the high efficiency of CPLEX in solving integer programming problems, all test cases are solved within 1 min. [fig_ref] Table 2: Results of planned outpatient capacity [/fig_ref] shows the minimum of the maximum working time obtained by using CPLEX and the patient information adjusted by senior doctors and associate senior doctors. Detailed parameters corresponding to each test are shown in [fig_ref] Table 3: Parameters for experiments [/fig_ref]. RTS represents the improvement of the work efficiency of senior doctors receiving the same number of patients after planning and before planning; RTDS represents the improvement of the number of patients received by associate senior doctors after planning and before planning. Their calculation formula is as follows. SNPP and ANPP represent the number of patients transferred in or out by senior doctors and associate senior doctors respectively. According to the data obtained from the experimental solution after the planning, the work efficiency of senior doctors after the planning increased by 42% at the highest, 8% at the lowest, and 24% on average. The number of associate senior doctors increased by 83% at the highest level, 16% at the lowest level, and 51% on average. At present, the ratio of senior and associate senior doctors in most class III class A hospitals is 1:1. In this proportion, after using the method proposed in this paper, the work efficiency of senior doctors is up to 25%, and the number of patients of associate senior doctors is up to 50%. According to the experimental results, senior doctor's work efficiency has improved, and the number of patients treated by associate senior doctors has increased significantly. This is due to the fact that the associate senior doctors are assigned to take care of the examination service of patients who make appointments with senior doctors. Senior doctors can save more time when serving the same number of patients. At the same time, it also effectively reduces patients' waiting time in the senior doctor. For associate senior doctors, the opportunity to contact more patients every day is increased, which is conducive to the cultivation of professional skills. Below, we analyze the minimum of the maximum working time corresponding to the change of values of relevant factors from an experimental perspective. In order to be more representative, the following cases are analyzed respectively: ET value increased from 2 to 8, DT value decreased from 8 to 2, SD value increased from 2 to 18, and ASD value decreased from 18 to 2. The test results are shown in Figs. 2 and 3. shows the corresponding changes of ANPP when ET and DT change. As ET increases from 2 to 8, i.e. DT decreases from 8 to 2, SNPP decreases from 36 to 11. As ET time increases from 2 to 8, the time of examination service per patient gradually increases. While the free time of associate senior doctors is fixed, ANPP is also reduced from 36 to 11. shows the corresponding changes of ANPP and SNPP when SD and ASD change. As SD increased from 2 to 18, ASD decreased from 18 to 2, SNPP decreased from 30 to 6, and ANPP increased from 6 to 30. The free time of each senior associate doctor is fixed, and as ASD decreases from 18 to 2, the totally free time of all associate senior doctors gradually decreases. As a result, the number of patients referred by senior doctors gradually decreased. That the SNPP decreased from 30 to 6. [formula] RTS = 1 − LWT * SD tds i=0 doctor_pat[i][j] * (et + dt) RTDS = LWT * ASD �   td � i=tds+1 doctor_pat[i][j] * (et + dt)   − 1 [/formula] In [fig_ref] Table 4: Six detailed experimental results before and after planning [/fig_ref] , we give detailed experimental results before and after planning under the condition of whether senior doctors and associate senior doctors are busy or not. The parameters in this experiment are as follows: There are ten senior doctors and ten associate senior doctors respectively. Columns 3 through 12 in [fig_ref] Table 4: Six detailed experimental results before and after planning [/fig_ref] represent the number of patients corresponding to senior doctors. The 13 to the 22nd cases are the number of patients corresponding to associate senior doctors; ET and DT time are 5 for each patient. Column 1 in [fig_ref] Table 4: Six detailed experimental results before and after planning [/fig_ref] is the name, where 1 through 6 represent the six pre-planning scenarios.1-p to 6-p are the number of patients after the corresponding planning. In order to show the comparison before and after planning in more detail, the comparison of data trends before and after planning in 6 cases is shown in [fig_ref] Figure 4: Minimum [/fig_ref]. In [fig_ref] Figure 4: Minimum [/fig_ref] , the red line represents the working time of each doctor before the plan, and the blue line shows the working time of each doctor after the plan. The abscissa 1 to 10 represents 10 senior doctors, and 11 to 20 represents 10 associate senior doctors. Before planning, the working time of senior doctors are mostly higher than those of associate senior doctors. Senior doctors work longer time, while associate senior doctors work shorter time. After planning, the working time of senior doctors are significantly reduced, while those of associate senior doctors are increased, resulting in the working time of senior doctors and associate senior doctors being equal. According to lean thinking, the patient outpatient service time is divided into examination service time, and diagnosis service time, and the problem is modeled as an integer programming problem to solve. As seen from the above experimental results, assigning patient examination services to associate senior doctors can effectively improve the work efficiency of senior doctors by 24% on average. At the same time, it also increased the number of patients served by associate senior doctors by an average of 50%. By using integer programming, the minimum maximum working time is obtained. # Discussion Outpatient service is the first step for hospital to provide high-quality and safe medical service for patients, and it is a key link to hospital quality control. The quality of outpatient procedures directly affects the image of the hospital, which reflects the overall service level of the hospital and relates to the overall interests of the hospital. Recent studies have shown that major problems in outpatient clinics include long patient waiting times, burnout among doctors, and the resulting inefficient use of resources. In order to further improve patients' medical experience and medical resource utilization, outpatient scheduling research has become a hot issue in medical operation management. This paper will discuss the latest progress in outpatient scheduling research from two aspects patient waiting time and patient flow. Then, the advantages and limitations of the mathematical programming model presented in this paper are discussed. ## Reduce waiting time With the transformation of the medical service model and the development of the market economy, patients' demand for medical services is getting higher and higher, and patient experience is getting more and more attention from medical circles. Good patient experience can improve patient treatment compliance, reduce medical costs, achieve better medical effects, and improve patient satisfaction. Long waiting time in the outpatient department has a great influence on timely access to medical services and medical experience. Many scholars have studied the problem of long waiting times. Li et al. studied the satisfaction of outpatient service in community health centers and secondary and tertiary hospitals [bib_ref] Patient satisfaction between primary care providers and hospitals: a cross-sectional survey in..., Li [/bib_ref]. Patients at county and tertiary hospitals complained of long waiting times and low overall satisfaction with outpatient services. Compared with secondary hospitals, tertiary hospitals face greater challenges in outpatient satisfaction [bib_ref] Patient satisfaction with outpatient care in China: a comparison of public secondary..., Yan [/bib_ref]. Several studies have shown that reducing uncanceled missed appointments can have a huge impact and improve resource utilization in physician productivity. Almuhaideb et al. predicted the situation of non-payment of outpatient appointments [bib_ref] Prediction of hospital no-show appointments through artificial intelligence algorithms, Almuhaideb [/bib_ref]. The Hoeffding tree algorithm is used to predict appointments with high absence risk in real-time. Then appropriate active interventions are set up to reduce the negative effects of absence. Barghash et al. studied the effect of predetermined overload percentage and patient interval on waiting for time, overtime and utilization rate [bib_ref] Enhancing outpatient appointment scheduling system performance when patient no-show percent and lateness..., Barghash [/bib_ref]. The study found that over time increase excessively when both the basic interval and the overload percentage increase. If the basic interval must be reduced to achieve low overtime time, the utilization of doctors will also increase. Hribar et al. used discrete event simulation to model the outpatient eye clinic workflow to test a new scheduling template that could reduce patient waiting time and improve clinic efficiency [bib_ref] Evaluating and improving an outpatient clinic scheduling template using secondary electronic health..., Hribar [/bib_ref]. Analyze EHR data and its compliance with templates to gain insight into new policies to better balance time to meet patient needs and minimize waiting time. Creps et al. used decision trees to analyze influencing factors significantly related to patients' visiting behaviors, and then evaluated the possibility of patients not appearing [bib_ref] A dynamic approach for outpatient scheduling, Creps [/bib_ref]. On this basis, a dynamic appointment scheduler is developed, which uses different overbooking strategies for different appointment patients. Dynamic scheduler improves scheduling efficiency through overbooking. Munavalli et al. proposed an overall patient scheduling model combining the status and information of all departments in the outpatient department for the long waiting time of patients [bib_ref] Integral patient scheduling in outpatient clinics under demand uncertainty to minimize patient..., Munavalli [/bib_ref]. This model guides the patient to the optimal path upon arrival and solves the problem of scheduling patients without appointments in real-time. Experimental results show that full patient scheduling significantly reduces waiting time, and realtime path optimization makes scheduling more efficient because it can immediately capture changes in the outpatient clinic. Berg et al. studied the use of constraint optimization modeling to balance doctors' work schedules in specialized outpatient clinics, minimizing the variability of timetables, improving work efficiency, and thus reducing the maximum number of doctors making visits [bib_ref] A method for balancing provider schedules in outpatient specialty clinics, Berg [/bib_ref]. Nguyen et al. proposed precise and relaxing appointment scheduling rules to arrange the appointment of each new patient according to the uncertainty of future arrival [bib_ref] Scheduling rules to achieve leadtime targets in outpatient appointment systems, Nguyen [/bib_ref]. Its scheduling rules are designed to maximize the use of program resources (i.e., physician staff ) or, equivalently, to maximize the number of patients admitted. Numerical experiments test the proposed scheduling rules, and the experimental results show that the scheduling rules are efficient and effective in terms of service objective satisfaction and resource utilization. Silver et al. studied the reasons for the increased waiting time of cancer outpatient patients with large volumes [bib_ref] Reducing patient wait times in a head and neck cancer outpatient clinic:..., Silver [/bib_ref]. Arranging too many patients in a short interval of clinic start time and arranging patients with more than the doctor's hourly patient capacity will increase the waiting time of patients. Doctors with the shortest waiting times for patients are interviewed and the best outpatient paths are identified. It proposes to reduce patient waiting time by benchmarking the optimal outpatient pathway. Which records and analyzes patient workflow and scheduling processes. The above methods can improve patients' medical examination and patient satisfaction by reducing patients' waiting time. In this paper, a mathematical programming model combined with lean thinking is proposed to effectively reduce the waiting time of patients through a twostage outpatient process. In addition, on the premise of not changing the outpatient process, this method plans the implementation space and doctors of the outpatient process from the perspective of space, and increases the patient flow while improving the patient attendance rate of associate senior doctors. Hospital management can further improve the service management model of the hospital by using lean thinking for reference. Eliminating redundant links and patients' unnecessary waiting in hospital management, and developing a standardized optimization flow chart of the hospital treatment process. Healthcare reform needs to do more to reduce patient waiting time, such as improving the quality of primary care diagnosis and treatment, and creating more targeted referral systems. ## The patient flow Many scholars have studied the refined management method of patient flow to increase the capacity of outpatient reception and the number of inpatients by effectively increasing mobility. Lin et al. proposed a novel heuristic algorithm based on two-stage simulation to evaluate and optimize various tactical and operational decisions for multiple objectives [bib_ref] Resource allocation and outpatient appointment scheduling using simulation optimization, Lin [/bib_ref]. Resource allocation plans are sought in the first phase, and overall outpatient appointment schedules are determined based on patient categories and daily operational level service discipline. Hahn et al. proposed outpatient dynamic template scheduling, a new technology combining active optimization and online optimization, and applied it to the scheduling problem of chemotherapy outpatient clinics [bib_ref] Dynamic optimization of chemotherapy outpatient scheduling with uncertainty, Hahn-Goldberg [/bib_ref]. A deterministic optimization model and appointment samples are used to create an active template for the expected dates of chemotherapeutic centers, which solves the dynamic uncertainty caused by appointment requests arriving in real-time and the uncertainty caused by last-minute schedule changes. Scheduling an outpatient appointment involves a complex set of factors and different stakeholders. Families, administrators, and clinicians may have different experiences in scheduling clinic appointments. Maira et al. studied the views and experiences of scheduling pediatric outpatient appointments from the perspective of stakeholders [bib_ref] Pediatric ambulatory appointment scheduling: a qualitative study of stakeholders' perceptions and experiences, Quintanilha [/bib_ref]. Qualitative content analysis is used to analyze the three stakeholder groups. It is found that the treatment process, skills, and services have the greatest impact on the pediatric outpatient scheduling system. DeWaters et al. reviewed the commissioning methods for internal medicine residents [bib_ref] The impact of block ambulatory scheduling on internal medicine residencies: a systematic..., Dewaters [/bib_ref]. The effects of block and traditional outpatient scheduling on inpatientrelated outcomes are systematically examined. In the lastest decade, nearly half of internal medicine residents have implemented block-based outpatient scheduling. While block-based scheduling has improved residents' satisfaction, the conflict between inpatient and outpatient responsibilities, and outpatient training time, there may be important trade-offs with poor continuity of care. Klassen et al. investigated the circumstances under which additional mid-level service providers (MLSP) could be added to a single-physician clinic and identified scheduling strategies in a single-phase environment [bib_ref] Appointment scheduling in multi-stage outpatient clinics, Klassen [/bib_ref]. Compared with a single-phase system in which the physician performs all parts of the service, the results showed that the addition of MLSP reduced patient waiting time, patient flow time, and physician service time. Steen et al. adjusted the location of workload generation and activities in working days, improving clinic utilization rate and a total number of visits in the clinical pharmacy expert compound group [bib_ref] Improving clinic utilization and workload capture for clinical pharmacy specialists, Steen [/bib_ref]. The implementation of the intervention shows that pharmaceutical administrators can improve workload and obtain quality care by cooperating with clinical pharmaceutical experts. Munavalli et al. modeled the outpatient clinic as a multi-agent system, and proposed an intelligent real-time scheduler that could schedule patients and resources according to the actual state of the department [bib_ref] An intelligent realtime scheduler for out-patient clinics: a multi-agent system model, Munavalli [/bib_ref]. The results show that the intelligent real-time scheduler significantly improves the performance indexes such as waiting time, cycle time, and utilization rate. The outpatient system becomes a pull system by scheduling resources and patients according to system status and demand. The online reservation scheduling system is designed to solve the problem of the traditional reservation scheduling systems. In Iran, most outpatient clinics and our study population are facing high rates of unattended patients and long waiting times due to the non-use of an online appointment systems. Habibi et al. studied the effect of online appointment scheduling systems by comparing the evaluation indicators of appointment scheduling before and after intervention [bib_ref] Effect of an online appointment scheduling system on evaluation metrics of outpatient..., Habibi [/bib_ref]. This pre-and post-pilot study is conducted in 10 outpatient clinics in different specialties. An online appointment scheduling system effectively reduced patient wait time, and patient absence rate and improved physician punctuality. The above methods reduce patient waiting time and improve the utilization of doctors and medical resources by increasing mobility. The mathematical programming model presented in this paper improves the work efficiency of senior doctors, and then gives senior doctors more time to provide services to patients in greater need. In addition, it also improves the number of associate senior doctors' patients. On the one hand, it effectively alleviates the waiting time of patients and accumulates diagnosis and treatment experience in difficult and miscellaneous diseases, which is conducive to the rapid growth of associate senior doctors. Medical care has become one of the largest industries in the world. In recent years, the cost of medical care in developed countries has risen sharply, of which about one-third is from hospital expenses. Health and government departments of various countries have issued relevant policies, forcing hospitals to improve operating efficiency, control costs, and increase income. The improvement of patient flow is in essence to improve the utilization rate of medical resources, reduce the average hospitalization cost per patient, and at the same time, it is more conducive to the hospital to control costs and increase income. # Advantages and limitations We actively involve hospital management and medical staff constructing of the mathematical model, to more comprehensively consider the complexity of the outpatient scheduling model, which increases the possibility of model planning results to guide decision-making, and effectively solves the challenging problems in analysis, statistics, and modeling methods. The mathematical model is based on constraint relations, and the minimum and maximum working time is solved automatically by CPLEX. The mathematical model is easy to model and solve. By providing a set of relevant input parameters to the model, relevant information about senior physicians, senior associate physicians, and patient distribution can be solved. Based on the existing constraint relations of the mathematical model proposed in this paper, other constraint relations can be added to the model. For example, the restrictions on the working time of individual doctors, doctors' and patients' preferences, and the capacity of relevant medical resources. This mathematical model is customized according to outpatient procedures, but it can also be applied to other spatial planning problems. This mathematical model is constructed after discussion with outpatient management and can only be applied to guide the allocation of doctors and patients when patient flow, doctors, or related medical resources change. # Conclusion Aiming at the key problems in outpatient service, this paper considers the work efficiency of senior doctors and the waiting time of patients, and combines lean thinking and integer programming method to optimize outpatient service. This study enriches and improves the theory and method system of medical operation management, especially outpatient service scheduling. This method has strong practicability and can provide a theoretical basis and decision support for hospitals to the perfect outpatient service systems and reasonably schedule outpatient resources. It can help the hospital reduce the cost of outpatient operations, improve the utilization rate of medical resources and patient satisfaction, and comprehensively improve the quality of medical service. In this paper, based on lean thinking, the outpatient service is subdivided into examination service and diagnosis service, and the examination service provided by the senior doctor is assigned to the senior associate doctor. For senior doctors, energy consumption and working time are reduced, so that more energy and time are available to provide services to other patients in need; For patients, the waiting time of patients is reduced without affecting their medical treatment; For the associate senior doctors, the number of patients received has significantly increased, which not only helps to improve the professional level but also helps to solve the problem of difficult medical treatment. Using CPLEX to solve the outpatient planning problem, the minimum and maximum working time are obtained, and the examination service and diagnosis service allocation are planned from a fine perspective. Compared with those without outpatient service planning, the work efficiency of senior doctors increased by 25% on average, and the patient flow of associate senior doctors increased by 50% on average. [fig] Figure 1 a: Clinical pathway of outpatient services (Normal). b Clinical pathway of outpatient services (Refined) [/fig] [fig] Figure 2, Figure 3: Patients of ET and DT Patients [/fig] [fig] Figure 4: Minimum [/fig] [table] Table 1: Description of parameters and decision variables of the problem [/table] [table] Table 2: Results of planned outpatient capacity [/table] [table] Table 3: Parameters for experiments [/table] [table] Table 4: Six detailed experimental results before and after planning [/table]

Challenging response latencies in faking detection: The case of few items and no warnings Faking detection is an ongoing challenge in psychological assessment. A notable approach for detecting fakers involves the inspection of response latencies and is based on the congruence model of faking. According to this model, respondents who fake good will provide favorable responses (i.e., congruent answers) faster than they provide unfavorable (i.e., incongruent) responses. Although the model has been validated in various experimental faking studies, to date, research supporting the congruence model has focused on scales with large numbers of items. Furthermore, in this previous research, fakers have usually been warned that faking could be detected. In view of the trend to use increasingly shorter scales in assessment, it becomes important to investigate whether the congruence model also applies to self-report measures with small numbers of items. In addition, it is unclear whether warning participants about faking detection is necessary for a successful application of the congruence model. To address these issues, we reanalyzed data sets of two studies that investigated faking good and faking bad on extraversion (n = 255) and need for cognition (n = 146) scales. Reanalyses demonstrated that having only a few items per scale and not warning participants represent a challenge for the congruence model. The congruence model of faking was only partly confirmed under such conditions. Although faking good on extraversion was associated with the expected longer latencies for incongruent answers, all other conditions remained nonsignificant. Thus, properties of the measurement and properties of the procedure affect the successful application of the congruence model. In recent years, researchers have paid increasing attention to the problem of faking in psychological measurement. Much of the research has highlighted that faking on psychological measures can be a serious issue, because people do fake (e.g., [bib_ref] A meta-analytic investigation of job applicant faking on personality measures, Birkeland [/bib_ref] [bib_ref] Meta-analyses of fakability estimates: Implications for personality measurement, Viswesvaran [/bib_ref] , and faking has an influence on scale means (e.g., [bib_ref] The impact of response distortion on preemployment personality testing and hiring decisions, Rosse [/bib_ref] [bib_ref] Effects of the testing situation on item responding: Cause for concern, Stark [/bib_ref] [bib_ref] Meta-analyses of fakability estimates: Implications for personality measurement, Viswesvaran [/bib_ref] , rank orders (e.g., [bib_ref] Correcting the 16PF for faking: Effects on criterion-related validity and individual hiring..., Christiansen [/bib_ref] [bib_ref] The impact of response distortion on preemployment personality testing and hiring decisions, Rosse [/bib_ref] , and the validity of test scores (e.g., [bib_ref] Five-factor inventories have a major general factor related to social desirability which..., Bäckström [/bib_ref]. Furthermore, measures that were considered to be supposedly immune to faking have turned out to be fakeable (e.g., [bib_ref] Exaggeration is harder than understatement, but practice makes perfect! Faking success in..., Röhner [/bib_ref]. Given the potential detrimental consequences of faking, it is not surprising that researchers have spent considerable time investigating methods for detecting such behavior. ## Response latencies as a method of faking detection The rationale behind this approach is that fakers, in contrast to non-fakers, differ in the amount of time they take to respond to psychological measures. A popular, albeit simplified, assumption can be summarized as: "Lying takes time" (e.g., [bib_ref] Lying takes time: A meta-analysis on reaction time measures of deception, Suchotzki [/bib_ref]. Being considered as an effortful process that requires extra cognitive processing and editing, it is quite intuitive that faking, in contrast to non-faking, results in longer response times. However, research has produced contradictory findings. On one hand, findings have indicated that faking is associated with longer response times (e.g.,, while on the other hand, research has also demonstrated that faking is related to shorter response times (e.g.,. Derived from schema theory, have articulated a complex model of faking indicating that whether fakers respond faster or slower depends on the congruence between the generated response and the faking schema (see also [bib_ref] Response latency detection of fakers on personnel tests, Holden [/bib_ref]. According to this model, people who fake good will provide favorable responses (i.e., congruent answers) faster than they provide unfavorable (i.e., incongruent) responses. Complementarily, people who fake bad will provide unfavorable (i.e., congruent) responses faster than they will provide favorable (i.e., incongruent) responses. Thus, schema-incongruent responding is slower than schemacongruent responding. One might question why fakers provide answers that are incongruent with their faking schema. [bib_ref] Response latencies are alive and well for identifying fakers on a self-report..., Holden [/bib_ref] argue that faking can be a nuanced and sophisticated process rather than simple, naïve answering, especially if fakers seek to avoid presenting an extremely obvious and detectable faking pattern. ## Challenging the congruence model of faking In general, can the latencies that individuals take to respond to psychological measures indicate whether or not they faked? According to recent meta-analyses [bib_ref] The relationship between faking and response latencies: A meta-analysis, Maricuțoiu [/bib_ref] [bib_ref] Lying takes time: A meta-analysis on reaction time measures of deception, Suchotzki [/bib_ref] , response times can be informative, but how informative depends on their measurement properties. The results of the meta-analysis by [bib_ref] The relationship between faking and response latencies: A meta-analysis, Maricuțoiu [/bib_ref] clearly revealed that there are differences in response times between faking good and honest responding and between faking bad and honest responding. However, they point to properties of the measurement that impact response latencies by demonstrating moderator effects of item type (i.e., larger effects computed on response latencies of positively keyed items as compared to response latencies of negatively keyed items). In the meta-analysis by [bib_ref] Lying takes time: A meta-analysis on reaction time measures of deception, Suchotzki [/bib_ref] , results also clearly revealed the expected and large differences in response times between faking and honest responding. However, they also point to properties of the measurement that impact response latencies. For example, the type of faking instruction played a role (i.e., whether or not participants received instructions to avoid being detected as fakers impacted their response times). Further, the to-befaked test played a role (i.e., autobiographical Implicit Association Test produced smaller effects than the Concealed Information Test, the Sheffield Lie Test, and the Differentiation of Deception paradigm). In addition, the authors point to the fact that, like other deception measures, response time-based deception measures may be susceptible to countermeasures. Thus, the authors conclude that response time-based measures of deception have potential information, but still more research is needed. Summing up the results of both meta-analyses, the properties of the measurement (e.g., faking direction, number of items, construct to be faked, warnings) are relevant factors that need to be taken into consideration. Empirical support for the congruence model has been found in a variety of experimental studies ranging from induced faking with university students [bib_ref] Instructed faking and MMPI-2 response latencies: The potential for assessing response validity, Brunetti [/bib_ref] [bib_ref] Differentielle Reaktionslatenzzeiten beim Bearbeiten von Persönlichkeitsfragebogen als möglicher Indikator für Verfälschungstendenzen [Differential..., Esser [/bib_ref] , to incarcerated offenders , and unemployed persons who are actively seeking employment [bib_ref] Detecting fakers on a personnel test: Response latencies versus a standard validity..., Holden [/bib_ref]. However, in virtually all these investigations, the scales examined included 60 or more items. Considering the ongoing trend in big data analyses to use increasingly shorter scales in research, it becomes important to investigate whether the congruence model can be applied to self-report measures with smaller numbers of items. For economic reasons and also for time efficiency, research is steadily migrating toward using as few items as possible (e.g., [bib_ref] Can't we make it any shorter? The limits of personality assessment and..., Rammstedt [/bib_ref]. Whether the congruence model successfully applies in such instances, however, has not yet been evaluated. Furthermore, in addition to using a substantial number of items, most previous research involving the congruence model has warned participants that faking can be detected and that they should avoid activating any detection systems. Research has already indicated that warning participants that faking can be identified impacts participants' responding. On one hand, previous research (e.g., [bib_ref] The validity of psychophysiological detection of information with the Guilty Knowledge Test:..., Ben-Shakhar [/bib_ref] [bib_ref] Memory detection with the Concealed Information Test: A meta-analysis of skin conductance,..., Meijer [/bib_ref] has revealed larger effects on response latencies when participants received an extra incentive to avoid detection-following from this, the effects of warnings on response latencies could be a result of the motivation to avoid detection, something that is absent under nonwarned conditions. On the other hand, warnings also could lead to more and better attempts to counteract faking detection (e.g., [bib_ref] A socialcognitive framework for understanding serious lies: Activationdecision-construction-action theory, Walczyk [/bib_ref] -following from this, the effects of warnings on response latencies may be even greater in nonwarned conditions. A recent meta-analysis by [bib_ref] Lying takes time: A meta-analysis on reaction time measures of deception, Suchotzki [/bib_ref] points to a small but nonsignificant effect of warnings on response latencies. However, whether the inclusion of a warning is a necessary aspect of the congruence model is currently unclear. This is not trivial, because in applied settings it may not be best practice to inform test-takers that their faking can be detected. This could be both because valid faking detection methods are still missing for most psychological tests, and because respondents may not believe in the quality of detection methods. Further, warnings could have negative side effectsthat may keep test users from utilizing them. For example, warnings may decrease the perceived test quality or increase test anxiety in test-takers (e.g., [bib_ref] Impact of electronic warnings on online personality scores and testtaker reactions in..., Burns [/bib_ref] ; see also [bib_ref] Faking behavior, Röhner [/bib_ref]. In this regard, recent research by [bib_ref] The role of emotions as mechanisms of mid-test warning messages during personality..., Li [/bib_ref] has demonstrated that warnings can induce different emotions in participants (i.e., guilt, fear, anger). These emotions are related to different outcomes. While guilt was related to a desired outcome of warnings (i.e., guilt boosted the personality scores of fakers to accuracy), fear and anger were related to negative side effects of warnings (i.e., fear led to overcorrection of non-fakers, anger reduced the perceived test fairness by fakers and non-fakers). Thus, there may be at least three reasons to avoid the use of warnings in applied settings. ## Present study Particularly because of the use of short scales without faking detection warnings in applied settings, this research seeks to challenge the congruence model of faking by using scales that include only a few items and by not warning participants about faking detection measures. Toward this goal, we reanalyze two data sets involving faked and non-faked self-description measures of extraversion and need for cognition. Extraversion was chosen because of its use in previous faking research and because, for this construct, both faking directions (i.e., good and bad) are plausible (e.g., [bib_ref] An assessment of the fakeability of self-report and implicit personality measures, Mcdaniel [/bib_ref] [bib_ref] What do fakers actually do to fake the IAT? An investigation of..., Röhner [/bib_ref] [bib_ref] Is the Implicit Association Test immune to faking?, Steffens [/bib_ref]. Further, consistent with the congruence model, previous research has demonstrated that faking involving an extraversion scale can be detected using response latencies, at least when 60 items are used and when participants are warned about faking detection indicators [bib_ref] Response latencies are alive and well for identifying fakers on a self-report..., Holden [/bib_ref]. In addition, we chose need for cognition because this construct has not been previously examined in such research. However, given the research that has demonstrated small but positive associations between need for cognition and socially desirable responding (e.g., [bib_ref] Dispositional differences in cognitive motivation: The life and times of individuals varying..., Cacioppo [/bib_ref] [bib_ref] The very efficient assessment of need for cognition: Developing a six-item version, De Holanda Coelho [/bib_ref] , it is a construct that may encourage respondents to answer in a manner that makes them seem more interested in thinking in order to impress others. Overall, therefore, our study's research focus is summarized as an investigation of whether the congruence model of faking is applicable to scales having only a small number of items and where test-takers are not warned regarding the presence of faking indicators. # Method ## Data sets To examine whether response latencies differ between fakers and non-fakers under the abovementioned challenging conditions, we reanalyzed data sets that were previously collected under the supervision of the lead author in an investigation of faking on scales measuring extraversionand both extraversion and need for cognition. We chose these data sets for several reasons: First, data sets were from studies that included both faking good and faking bad instructions. Because our interest was on the impact of faking good and of faking bad, it was necessary that both faking directions were contained in the same data set. Second, the scales used in these studies met our precondition of including comparably few items (i.e., extraversion scale: 12 items; need for cognition scale: 16 items). Third, in these studies, participants were not warned that faking could be detected, something that was a precondition for our reanalyses. Finally, because 255 (i.e., extraversion) and 146 (i.e., need for cognition) participants were included in these studies, power analyses using G*Power 3.1.7 [bib_ref] G* Power3: A flexible statistical power analysis program for the social, behavioral,..., Faul [/bib_ref] revealed a power > .99 for ANOVAs concerning the manipulation check analyses to detect a moderate effect size at an alpha level of .05 and revealed a power of > .99 (i.e., extraversion; N = 130) and >. 91 (i.e., need for cognition; N = 86) for ANOVAs concerning the response latency analyses to detect the expected large effect size at an alpha level of .05. Participants took part in the studies in exchange for personal feedback and/or partial university course credit. In both studies, individuals completed the extraversion scale (Study 1 and 2;and/or the need for cognition scale (Study 2 only; [bib_ref] Need for cognition: Eine Skala zur Erfassung von Engagement und Freude bei..., Bless [/bib_ref] twice. On the first occasion (i.e., baseline), participants completed the measures under standard instructions. On the second occasion, individuals were randomly assigned to one of three conditions (i.e., control, faking good, or faking bad). Participants in the control condition again responded under standard instruction. Fakers were asked to fake either high scores or low scores on the measures according to a personnel selection scenario. To assess the faking behavior of participants as would normally occur within a personnel context, fakers were not provided with any strategies on how to fake (i.e., naïve faking; see, e.g., [bib_ref] What do fakers actually do to fake the IAT? An investigation of..., Röhner [/bib_ref] , for further information), nor were they informed of the presence of any faking detection measures. In the instructions for faking good, participants were asked to imagine they had been unemployed for one year and had now received a very attractive job offer. They were asked to fake high on extraversion (and/or need for cognition) in order to maximize the chances of being offered the job. The instructions for faking bad included the description of a very unattractive job offer. To avoid being offered the job, participants were asked to fake low extraversion (and/or need for cognition). For our analyses, we combined the data on the extraversion scale from both studies. Thus, the final sample for the extraversion scale consisted of 255 participants (93 faking bad, 86 control, 76 faking good; 186 women, 68 men, 1 no response; 251 students) with an average age of 22.34 years (SD = 4.59). The overall gender ratio was 72.9% women, 26.7% men, and 0.4% no response, which was very comparable to the gender ratios in subgroups (faking low: 75.3% women, 24.7% men; control group: 76.7% women, 23.3% men; faking high: 65.8% women, 32.9% men, and 1.3% no response). The final sample for the need for cognition scale consisted of 146 participants (51 faking bad, 47 control, 48 faking good; 110 women, 36 men; 145 students) with an average age of 21.89 years (SD = 4.34). The overall gender ratio was 75.3% women, 24.0% men, and 0.7% no response, which was very comparable to the gender ratios in subgroups (faking low: 74.5% women, 23.5% men, and 2.0% no response; control group: 80.9% women, 19.1% men; faking high: 70.8% women, 29.2% men). These data were used for the manipulation check analyses. Concerning the response latency analyses, these data had to be checked for whether the inclusion criterion of having both high and low responses was met. After the exclusion of participants who did not fulfill this criterion, the data on the extraversion scale consisted of 130 participants (41 faking bad, 69 control, 20 faking good). The overall gender ratio was 73.8% women, 26.2% men, which, with the exception of faking high, was comparable to the gender ratios in subgroups (faking low: 73.2% women, 26.8% men; control group: 81.2% women, 18.8% men; faking high: 50.0% women, 50.0% men). The data on the need for cognition scale consisted of 86 participants (26 faking bad, 40 control, 20 faking good). The overall gender ratio was 76.7% women, 22.1% men, and 1.2% no response, which was comparable to the gender ratios in subgroups (faking low: 76.7% women, 23.1% men, and 3.8% no response; control group: 85.0% women, 15.0% men; faking high: 65.0% women, 35.0% men). ## Measures ## Extraversion scale Participants completed the extraversion scale of the German adaptation of the NEO-Five Factor Inventory . It consists of 12 items that are answered on 5-point ratings ranging from 0 ("strongly disagree") to 4 ("strongly agree"). Scale characteristics in the present study were: M = 27.74, SD = 43.08; coefficient α = .83 at baseline assessment and M = 25.76, SD = 204.83; coefficient α = .96 under faking/ retest. ## Need for cognition scale Participants completed the German adaptation of the 16-item need for cognition scale [bib_ref] Need for cognition: Eine Skala zur Erfassung von Engagement und Freude bei..., Bless [/bib_ref]. It consists of 16 items that are answered on 7-point ratings ranging from −3 ("strongly disagree") to +3 ("strongly agree"). Scale characteristics in the present study were: M = 15.73, SD = 11.31; coefficient α = .85 at baseline assessment and M = 5.64, SD = 986.19; coefficient α = .98 under faking/retest. ## Analytical strategy Data, syntaxes, and the outputs of our analyses are available at the OSF (https://osf.io/bh98z/). ## Manipulation check Prior to analyzing the response latencies of fakers and comparing them to non-fakers, we verified whether the participants in the faking conditions had followed the faking instructions. To do so, we used repeated-measures ANOVAs on the individual participants' mean scale scores of the extraversion and need for cognition scales (see e.g., [bib_ref] Exaggeration is harder than understatement, but practice makes perfect! Faking success in..., Röhner [/bib_ref]. ## Preparation and analysis of response latencies Separately for each scale, and based on previous procedures [bib_ref] Detecting fakers on a personnel test: Response latencies versus a standard validity..., Holden [/bib_ref] as explicitly outlined in [bib_ref] Measuring self-enhancement: From self-report to concrete behavior, Paulhus [/bib_ref] , raw item response times for the second administration of items were adjusted to control for the effects of statistical outliers and were twice standardized, once to adjust for confounding respondent factors (e.g., reading speed, sex) and a second time to adjust for confounding item variables (e.g., length, vocabulary level). This involved the following: First, to mitigate the influence of statistical outliers, response latencies of less than 0.5 s or greater than 40 s were set to 0.5 or 40 s, respectively. Second, response latencies were standardized across items, within each participant, to control for irrelevant differences between individual respondents. Third, response times were standardized across participants within each item to adjust for differences between items. Of note, the means and standard deviations used for this second standardization were from the corresponding items for all participants during the baseline administration (i.e., prior to the experimental assignment to different instructional conditions). Fourth, response times were readjusted for outliers, such that latencies of less than −3.00 or greater than 3.00 were set to −3.00 or 3.00, respectively. Overall, this preparation procedure yields response latencies that are calculated both relative to the respondent and relative to the scale item, and are unconfounded by main effect influences of specific persons, specific items, and statistical outliers. Following this preparation, these adjusted latencies were aggregated for each respondent to produce means for high scores on items (scores of 3 or 4 on extraversion items; scores of 1, 2, or 3 on need for cognition items) and low scores on items (scores of 0 or 1 on extraversion items; scores of −3, −2, or −1 on need for cognition items.). For each participant, these two mean adjusted response latency scores were the units of analysis for each scale. Item latencies for "neutral" responses (i.e., scored 2 on extraversion items; scored 0 on need for cognition items) were excluded from the analyses. ANOVAs then were used to compare mean adjusted response latencies for the control, faking good, and faking bad groups. # Results ## Manipulation check As expected, ANOVAs revealed that participants were motivated and well able to fake high and low scores on the extraversion scale and on the need for cognition scale. A 2 (measurement occasion) × 3 (experimental group) ANOVA with repeated measures on the first factor and the scores on the extraversion scale as the dependent variable indicated that participants in the faking groups faked according to their faking instructions on the extraversion scale (see [fig_ref] Table 1: Descriptive Variables and Post Hoc Comparisons Regarding the Means of the Extraversion... [/fig_ref]. The significant main effect of group, F(2, 252) = 165.96, p < .001, ω 2 = .56, was qualified by the expected large and significant interaction effect, F(2, 252) = 428.88, p < .001, ω 2 = .77. The main effect of measurement occasion, F(1, 252) = 3.77, p = .053, ω 2 = .01, was nonsignificant. A similar 2 (measurement occasion) × 3 (experimental group) ANOVA for scores on the need for cognition scale also confirmed that participants faked according to their instructions for the need for cognition scale (see [fig_ref] Table 1: Descriptive Variables and Post Hoc Comparisons Regarding the Means of the Extraversion... [/fig_ref]. The significant main effects of measurement occasion, F(1, 143) = 71.92, p < .001, ω 2 = .33, and group, F(2, 143) = 179.34, p < .001, ω 2 = .71, were qualified by the expected large and significant interaction effect, F(2, 143) = 356.48, p < .001, ω 2 = .83. ## Response latency analyses According to the congruence model of faking, participants were included in the response latency analyses only if they had a latency for high responses and for low responses. If a participant provided no low responses or no high responses, that person was excluded from analyses. Applying this inclusion criterion led to samples of 130 (i.e., extraversion) and 86 participants (i.e., need for cognition) for the response latency analyses. Mean adjusted item response latencies as a function of response and faking condition are presented for extraversion and for need for cognition scales in [fig_ref] Table 2: Mean [/fig_ref]. For extraversion, using response type (i.e., high vs. low responses) as a within-subject factor and faking instructional group (i.e., control, faking good, faking bad) as a between-subjects factor, a significant type of response by group interaction (see , F(2, 127) = 4.30, p < .05, partial η 2 = .06, approximating a medium effect size, partially supported the congruence model. However, based on the Games-Howell procedure (not requiring homogeneous variances), post hoc comparisons among faking groups within each response type indicated only differences for the faking good group taking longer to provide low extraversion answers than the control group. For need for cognition, no significant response type by group interaction emerged (see , F(2, 83) = 0.95, p > .05, partial η 2 = .02. Note. N = 255 for the extraversion scale. N = 146 for the need for cognition scale. Different lettered subscripts indicate significant differences between experimental groups (i.e., columns); different numbered subscripts identify significant differences between measurement occasions (i.e., rows) at p < .05 # Discussion We reanalyzed two data sets in order to investigate whether the congruence model of faking could be applied successfully on scales that have comparably few items (i.e., 12 and 16) and in contexts where warnings that faking can be detected are not used. In view of the growing tendency to use shorter scales and the various approaches regarding warnings, this is a crucial question. Our results revealed that the congruence model of faking was only partly supported under these conditions (i.e., for incongruent responses to faking good on extraversion). Thus, the applicability of the model depends on the measurement procedure. Can response latencies indicate fakers on scales with few items in unwarned situations? With regard to faking on the extraversion scale, our analyses revealed that, under these circumstances and in line with the congruence model of faking, the response latencies of fakers differed significantly from those of non-fakers when respondents were instructed to fake good. The differences in response latencies were limited to faking-good-incongruent answering. When faking bad was instructed, any differences in response latencies were nonsignificant. For faking on the need for cognition scale, the same pattern was found, but only at a statistically nonsignificant descriptive level. The differences between the two scales could be attributed either to the tobe-faked measure, to the number of items that were included (i.e., the need for cognition scale included four more items than the extraversion scale), or to both. The present results align with findings from the metaanalysis by Maricotoiu and Sârbescu (2016) that revealed that differences in response latencies are more pronounced when faking good than when faking bad is instructed. The findings also support the results of recent research by [bib_ref] The nature of faking: A homogeneous and predictable construct?, Bensch [/bib_ref] , who explored differences between faking good and faking bad. [bib_ref] The nature of faking: A homogeneous and predictable construct?, Bensch [/bib_ref] note that faking good and faking bad are related to different psychological processes. They indicate that faking good is easier than faking bad and suggest that, when faking good, it might be obvious which items are relevant to be faked and how one should respond to portray oneself in a positive way. Alternatively, when faking bad, it might also be obvious which items are relevant, but the best way to portray oneself in a negative way may be a less clear process. Their findings may assist in explaining our results. The congruence model fit the "easier" faking good direction, because respondents had to process less on how to fake, and thus differences in response times became more apparent. It could be speculated that, in children, being nice and presenting positively is something that becomes thoroughly ingrained during upbringing and is automatic in processing. The model did not fit "the more difficult" faking bad direction, because it entailed more cognitive processing of something that is not automatic, and response latencies were, thus, more comparable to what respondents do in determining an honest response that represents their personality. Earlier research has also often warned participants that faking can be detected. In unwarned conditions, such as in the current investigation, the effects of faking on response latencies were smaller than what has been found under warned conditions (e.g., [bib_ref] The validity of psychophysiological detection of information with the Guilty Knowledge Test:..., Ben-Shakhar [/bib_ref] [bib_ref] Memory detection with the Concealed Information Test: A meta-analysis of skin conductance,..., Meijer [/bib_ref]. Thus, warnings seem to be a necessary precondition. From an applied point of view, this is relevant because warnings cannot always be included, particularly when the original development of a scale has not included warnings, and there is a desire to use testing conditions that are identical to those associated with established test normative scores. Another consideration may be the degree to which respondents believe or fail to believe in any warnings that are given. This is a direction for future research. The issue arises as to why short scales with no warnings about faking produce a weakened effect for the congruence model. In fact, the applicability of the model depends on a respondent producing incongruent answers, responses that are discrepant from the test-taker's goal. However, with a small number of items per scale, there was less opportunity for faking-discrepant responses to occur. Although using incongruent answers represents a nuanced faking strategy (e.g., [bib_ref] Response latencies are alive and well for identifying fakers on a self-report..., Holden [/bib_ref] designed to avoid the detection of obvious faking, not including warnings might decrease the use of such nuanced faking. Thus, in addition to less opportunity to use incongruent answers, it might also be possible that with no warnings present, more elaborate strategies of faking are not invoked because respondents are less concerned about being caught faking. Of note are the relatively few degrees of freedom associated with the F-ratios of the mean response latencies reported for the interaction of type of response with group congruence. Using the extraversion scale as an example, although there were 255 participants, 125 of them did not provide both low and high responses for the 12 items of that scale and, thus, because mean adjusted latencies could not be calculated, could not be included in the analysis of response times. With more items and a warning about faking detection, a larger number of faking-discrepant responses would have resulted in stronger support for the congruence model that contrasts congruent and incongruent responding. # Limitations Our study has potential limitations regarding the applicability of the congruence model. First, we examined only two constructs (i.e., extraversion and need for cognition). Future research should investigate the applicability of the congruence model of faking using other to-be-faked constructs. Second, with regard to the number of items of a measure, we used scales that involved 12 and 16 items. Previous research has focused on measures of 60 or more items. Thus, an important issue is the minimum number of items necessary to produce a meaningful result. Third, our samples primarily consisted of female (> 70%) students who were instructed to fake. Considering research that has demonstrated differences in response times between female and male respondents (e.g., [bib_ref] Sex and handedness differences in eye-hand visual reaction times in handball players, Dane [/bib_ref] [bib_ref] Sex differences in reaction time mean and intraindividual variability across the life..., Dykiert [/bib_ref] , the overrepresentation of women might raise a concern about the generalizability of our results. However, the doublestandardization method as used in the congruence model of faking adjusts for potential confounding respondent factors (e.g., reading speed, sex) and for potential confounding item variables (e.g., length, vocabulary level). Nevertheless, whether the results are generalizable to samples from other populations (e.g., job applicants, gender-balanced) and to naturally occurring faking are avenues for future research. Fourth, in examining the congruence model of faking, our focus was on response type rather than on other potential influences such as item keying. Additional research could serve to articulate the influence of other factors that could impact on faking. Fifth, the inclusion criterion for the congruence model of faking led to the exclusion of numerous participants for the response latency analyses. Thus, the power was sufficient in each data set (above .91) to detect large effect sizes as expected from the large and significant overall average effect size found in a recent meta-analysis by [bib_ref] Lying takes time: A meta-analysis on reaction time measures of deception, Suchotzki [/bib_ref]. However, the power of our study was not sufficient to detect moderate or small effect sizes, as indicated by the meta-analysis by [bib_ref] The relationship between faking and response latencies: A meta-analysis, Maricuțoiu [/bib_ref]. We calculated our power estimation based on the results of the more recent meta-analysis by [bib_ref] Lying takes time: A meta-analysis on reaction time measures of deception, Suchotzki [/bib_ref] , because we believed that their effect size estimation is more reliable, for the following reasons: [bib_ref] Lying takes time: A meta-analysis on reaction time measures of deception, Suchotzki [/bib_ref] included 114 studies, whereas [bib_ref] The relationship between faking and response latencies: A meta-analysis, Maricuțoiu [/bib_ref] included only 16 studies; and [bib_ref] The relationship between faking and response latencies: A meta-analysis, Maricuțoiu [/bib_ref] included mainly studies on faking good, which is typically less pronounced than faking bad (e.g., [bib_ref] Exaggeration is harder than understatement, but practice makes perfect! Faking success in..., Röhner [/bib_ref] , and which in turn might have caused an underestimation of the effect size. Nevertheless, researchers interested in investigating the congruence model may thus want to ensure they have a very large number of participants and items, particularly if wanting to use the information at the individual respondent level. # Summary and conclusion Our results indicate that having a small number of items on a measure and not warning participants about faking detection has a deleterious effect on successfully applying the congruence model of faking. Under these conditions, the model was only partly supported. Although faking good on the extraversion scale was associated with the expected longer latencies for incongruent answers, all other conditions failed to support the model. Thus, properties of the measurement and aspects of warnings impact whether the congruence model can be applied successfully. As such, measurement conditions are an important consideration when applying response times for the detection of faking. Funding Open Access funding enabled and organized by Projekt DEAL. This research was partly funded by a grant from the equal opportunities office at the University of Bamberg. The funding source had no involvement in the study design or analyses. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. [table] Table 1: Descriptive Variables and Post Hoc Comparisons Regarding the Means of the Extraversion Scale and the Need for Cognition Scale [/table] [table] Table 2: Mean (SD) Adjusted Response Latency by Response and Faking Condition [/table]

b-arrestins play critical roles in chemotaxis and cytoskeletal reorganization downstream of several receptor types, including G protein-coupled receptors (GPCRs), which are targets for greater than 50% of all pharmaceuticals. Among them, receptors for lysophosphatidic acid (LPA), namely LPA 1 are overexpressed in breast cancer and promote metastatic spread. We have recently reported that b-arrestin2 regulates LPA 1 -mediated breast cancer cell migration and invasion, although the underlying molecular mechanisms are not clearly understood. We show here that LPA induces activity of the small G protein, Rap1 in breast cancer cells in a b-arrestin2-dependent manner, but fails to activate Rap1 in non-malignant mammary epithelial cells. We found that Rap1A mRNA levels are higher in human breast tumors compared to healthy patient samples and Rap1A is robustly expressed in human ductal carcinoma in situ and invasive tumors, in contrast to the normal mammary ducts. Rap1A protein expression is also higher in aggressive breast cancer cells (MDA-MB-231 and Hs578t) relative to the weakly invasive MCF-7 cells or non-malignant MCF10A mammary cells. Depletion of Rap1A expression significantly impaired LPA-stimulated migration of breast cancer cells and invasiveness in three-dimensional Matrigel cultures. Furthermore, we found that b-arrestin2 associates with the actin binding protein IQGAP1 in breast cancer cells, and is necessary for the recruitment of IQGAP1 to the leading edge of migratory cells. Depletion of IQGAP1 blocked LPAstimulated breast cancer cell invasion. Finally, we have identified that LPA enhances the binding of endogenous Rap1A to b-arrestin2, and also stimulates Rap1A and IQGAP1 to associate with LPA 1 . Thus our data establish novel roles for Rap1A and IQGAP1 as critical regulators of LPA-induced breast cancer cell migration and invasion. Citation: Alemayehu M, Dragan M, Pape C, Siddiqui I, Sacks DB, et al. (2013) b-Arrestin2 Regulates Lysophosphatidic Acid-Induced Human Breast Tumor Cell Migration and Invasion via Rap1 and IQGAP1. PLoS ONE 8(2): e56174. # Introduction Breast cancer is the leading cause of cancer-related deaths in women world-wide and metastasis accounts for the majority of these deaths [bib_ref] Breast cancer metastasis: markers and models, Weigelt [/bib_ref]. Thus, characterization of the signaling mechanisms involved in breast cancer cell migration and invasion, processes that are critically required for the metastatic spread of cancer is crucial for the identification of new therapeutic targets. The b-arrestins (b-arrestin1 and 2) are ubiquitously expressed proteins that are instrumental in attenuating G protein-coupled receptor (GPCR) signaling [bib_ref] Transduction of receptor signals by betaarrestins, Lefkowitz [/bib_ref] [bib_ref] Emerging paradigms of beta-arrestindependent seven transmembrane receptor signaling, Shukla [/bib_ref]. b-Arrestins can also function as molecular scaffolds for the organization of signaling complexes and thereby regulate cell migration [bib_ref] Transduction of receptor signals by betaarrestins, Lefkowitz [/bib_ref] [bib_ref] Beta-arrestins as regulators of signal termination and transduction: how do they determine..., Defea [/bib_ref] [bib_ref] beta-arrestin-dependent actin reorganization: bringing the right players together at the leading edge, Min [/bib_ref] downstream of various receptors including GPCRs [bib_ref] beta-Arrestin 1 and Galphaq/11 coordinately activate RhoA and stress fiber formation following..., Barnes [/bib_ref] [bib_ref] Beta-arrestin2 is critically involved in CXCR4-mediated chemotaxis, and this is mediated by..., Sun [/bib_ref] [bib_ref] Beta-arrestindependent regulation of the cofilin pathway downstream of protease-activated receptor-2, Zoudilova [/bib_ref] [bib_ref] G protein-coupled receptors stimulation and the control of cell migration, Cotton [/bib_ref] , receptor kinases such as transforming growth factor b receptor-III and insulin-like growth factor-1 receptor [bib_ref] Beta-arrestins as regulators of signal termination and transduction: how do they determine..., Defea [/bib_ref] [bib_ref] The type III TGF-beta receptor regulates epithelial and cancer cell migration through..., Mythreye [/bib_ref] [bib_ref] Betaarrestin and Mdm2 mediate IGF-1 receptor-stimulated ERK activation and cell cycle progression, Girnita [/bib_ref]. Emerging roles of b-arrestins in tumorigenesis have been demonstrated using in vitro [bib_ref] Expression of G protein-coupled receptor kinase 4 is associated with breast cancer..., Matsubayashi [/bib_ref] [bib_ref] Betaarrestin links endothelin A receptor to beta-catenin signaling to induce ovarian cancer..., Rosano [/bib_ref] and in vivo model systems [bib_ref] Role of beta-arrestin 1 in the metastatic progression of colorectal cancer, Buchanan [/bib_ref] [bib_ref] Depletion of beta-arrestin-2 promotes tumor growth and angiogenesis in a murine model..., Raghuwanshi [/bib_ref] [bib_ref] Rapid xenograft tumor progression in beta-arrestin1 transgenic mice due to enhanced tumor..., Zou [/bib_ref]. b-arrestins can associate with and regulate the oncoprotein Mdm2, a negative regulator of the tumor suppressor p53 [bib_ref] Beta-arrestin signaling and regulation of transcription, Ma [/bib_ref] [bib_ref] Identification of betaArrestin2 as a corepressor of androgen receptor signaling in prostate..., Lakshmikanthan [/bib_ref]. In breast cancer cells, b-arrestins regulate stress fiber formation via Rho GTPases [bib_ref] beta-Arrestin 1 and Galphaq/11 coordinately activate RhoA and stress fiber formation following..., Barnes [/bib_ref] , or by activating the actin filament-severing protein cofilin [bib_ref] Beta-arrestindependent regulation of the cofilin pathway downstream of protease-activated receptor-2, Zoudilova [/bib_ref]. Recently, a direct role for barrestins in regulating breast cancer metastasis has been demonstrated using a xenograft model with MDA-MB-231 cells [bib_ref] beta-arrestin1 mediates metastatic growth of breast cancer cells by facilitating HIF-1-dependent VEGF..., Shenoy [/bib_ref]. We have previously reported that b-arrestin2 mRNA levels are elevated in patient breast tumors samples at advanced stages of the disease [bib_ref] Betaarrestin/Ral signaling regulates lysophosphatidic acid-mediated migration and invasion of human breast tumor..., Li [/bib_ref]. Consistent with these observations, a recent study has shown that b-arrestin2 protein expression increased with the progression of breast cancer invasiveness [bib_ref] Differential expression of arrestins is a predictor of breast cancer progression and..., Michal [/bib_ref]. Furthermore, we found that b-arrestins regulate breast cancer invasion by regulating the activity of matrix metalloproteases [bib_ref] Betaarrestin/Ral signaling regulates lysophosphatidic acid-mediated migration and invasion of human breast tumor..., Li [/bib_ref]. Although b-arrestin2 has been suggested to regulate actin cytoskeleton organization by acting as scaffolds for actin binding proteins [bib_ref] beta-arrestin-dependent actin reorganization: bringing the right players together at the leading edge, Min [/bib_ref] [bib_ref] Beyond desensitization: physiological relevance of arrestin-dependent signaling, Luttrell [/bib_ref] and thereby regulate cell motility and cancer invasion, the underlying molecular mechanism by which barrestins regulate cancer cell migration and invasion remain largely unknown. b-Arrestins regulate internalization of LPA 1 , a member of the endothelial differentiation gene (EDG2) [bib_ref] A requirement for membrane cholesterol in the beta-arrestin-and clathrin-dependent endocytosis of LPA1..., Urs [/bib_ref]. LPA 1 binds lysophosphatidic acid (LPA), a blood-borne chemoattractant mitogenic lipid molecule that regulates actin cytoskeletal dynamics necessary for cell migration and thus plays important roles in processes such as wound healing [bib_ref] Regulation of fibroblast functions by lysophospholipid mediators: potential roles in wound healing, Watterson [/bib_ref]. An LPA 1 antagonist is currently in clinical trials for fibrotic disease (http://www.amirapharm.com). LPA has emerging roles in cancer and can mediate the 'hallmarks of cancer'. LPA induces the transformation of benign cells into invasive carcinoma cells, stimulates tumor growth, angiogenesis, invasion, and metastasis [bib_ref] The emerging role of lysophosphatidic acid in cancer, Mills [/bib_ref]. Autotaxin, a key enzyme in LPA production in blood, is overexpressed in various human malignancies, including breast cancer [bib_ref] The emerging role of lysophosphatidic acid in cancer, Mills [/bib_ref] [bib_ref] The ins and outs of lysophosphatidic acid signaling, Moolenaar [/bib_ref]. LPA signaling via LPA 1 , stimulates proliferation and metastasis of MDA-MB-231 breast cancer cells in a xenograft mouse model [bib_ref] The type 1 lysophosphatidic acid receptor is a target for therapy in..., Boucharaba [/bib_ref]. Although LPA 1 is present in normal mammary epithelial cells, it is aberrantly expressed in breast cancer [bib_ref] Betaarrestin/Ral signaling regulates lysophosphatidic acid-mediated migration and invasion of human breast tumor..., Li [/bib_ref] [bib_ref] Expression of autotaxin and lysophosphatidic acid receptors increases mammary tumorigenesis, invasion, and..., Liu [/bib_ref]. Additionally, expression of LPA 1 in mammary epithelial cells of transgenic mice promotes the development of metastatic mammary tumors [bib_ref] Expression of autotaxin and lysophosphatidic acid receptors increases mammary tumorigenesis, invasion, and..., Liu [/bib_ref]. A few studies have shed light on the mechanisms by which LPA 1 regulates breast cancer cell migration, implicating a role for phosphatidylinositol 3-kinases/PAK1/ERK signaling [bib_ref] Lysophosphatidic acid induces MDA-MB-231 breast cancer cells migration through activation of PI3K/PAK1/..., Du [/bib_ref] and Rho kinases [bib_ref] LPA2 (EDG4) mediates Rhodependent chemotaxis with lower efficacy than LPA1 (EDG2) in..., Chen [/bib_ref]. However, we have shown a major role for G i/o signaling pathways in LPA 1 -stimulated breast cancer cell migration and invasion signaling via b-arrestins and Ral GTPases [bib_ref] Betaarrestin/Ral signaling regulates lysophosphatidic acid-mediated migration and invasion of human breast tumor..., Li [/bib_ref]. We found that LPA 1 is overexpressed in patient breast tumors [bib_ref] Betaarrestin/Ral signaling regulates lysophosphatidic acid-mediated migration and invasion of human breast tumor..., Li [/bib_ref] and high levels of LPA 1 transcripts have been found in numerous breast cancer cell lines [bib_ref] LPA2 (EDG4) mediates Rhodependent chemotaxis with lower efficacy than LPA1 (EDG2) in..., Chen [/bib_ref]. We showed that LPA 1 activation induces migration and invasion of breast cancer cells via b-arrestin2 and the small GTPase Ral [bib_ref] Betaarrestin/Ral signaling regulates lysophosphatidic acid-mediated migration and invasion of human breast tumor..., Li [/bib_ref]. Ral activity can be regulated by its guanine nucleotide exchange factors (GEFs) such as RalGDS as well as b-arrestins in a Ras-independent manner [bib_ref] Betaarrestin/Ral signaling regulates lysophosphatidic acid-mediated migration and invasion of human breast tumor..., Li [/bib_ref] [bib_ref] Beta-arrestins regulate a Ral-GDS Ral effector pathway that mediates cytoskeletal reorganization, Bhattacharya [/bib_ref]. RalGDS can also associate with Rap1A GTPases with 100-times more affinity than Ras [bib_ref] Differential interaction of the ras family GTP-binding proteins H-Ras, Rap1A, and R-Ras..., Herrmann [/bib_ref]. Although Rap1 (composed of A and B subtypes) was originally discovered as a revertant of k-Ras-mediated cell transformation, several studies have since identified Ras-independent Rap1 signaling involved in various cellular processes including gene transcription, proliferation and growth, adhesion, polarity and migration [bib_ref] Rap1, a mercenary among the Ras-like GTPases, Frische [/bib_ref] [bib_ref] Rap1 stabilizes betacatenin and enhances beta-catenin-dependent transcription and invasion in squamous cell..., Goto [/bib_ref] [bib_ref] Rap1 GTPase: functions, regulation, and malignancy, Hattori [/bib_ref]. In the past decade, a role for Rap1 and its regulators has emerged in various cancers including breast cancer [bib_ref] Rap1, a mercenary among the Ras-like GTPases, Frische [/bib_ref] [bib_ref] Rasrelated protein 1 and the insulin-like growth factor type I receptor are..., Furstenau [/bib_ref] [bib_ref] Rap1 integrates tissue polarity, lumen formation, and tumorigenic potential in human breast..., Itoh [/bib_ref]. However, whether or not LPA signals via Rap1 in breast cancer cells is unknown. Here we show that Rap1A is aberrantly expressed in human breast tumors. We identify Rap1A as a novel regulator of LPAinduced breast cancer chemotaxis and invasion, and show that LPA stimulates Rap1 activity in breast cancer cells downstream of b-arrestin2, but fails to activate Rap1 in non-malignant mammary epithelial cells. We identify Rap1A and the actin-binding protein IQGAP1 as novel binding partners for b-arrestin2 as well as LPA 1 . IQGAP1 regulates directional cell migration and the establishment of polarized cell morphology [bib_ref] IQGAP1 in cellular signaling: bridging the GAP, Brown [/bib_ref] [bib_ref] IQGAP1 promotes cell motility and invasion, Mataraza [/bib_ref] [bib_ref] IQGAP1: a key regulator of adhesion and migration, Noritake [/bib_ref]. We also found that b-arrestin2 is co-localized with IQGAP1 in lamellipodia of migratory breast cancer cells in response to LPA, and appears to critically regulate the recruitment of IQGAP1 to the leading edge. Thus, our data establish novel roles for Rap1A and IQGAP1, downstream of b-arrestin2 as critical regulators of LPA-induced breast cancer cell migration and invasion. # Results ## Rap1a is aberrantly expressed in human breast tumor tissue We have previously shown that the mRNA levels of b-arrestin2 and Ral GTPases are elevated in human breast tumors compared to non-malignant tissue and that these molecules regulate breast cancer migration and invasion [bib_ref] Betaarrestin/Ral signaling regulates lysophosphatidic acid-mediated migration and invasion of human breast tumor..., Li [/bib_ref]. Since the Ral guanine exchange factor, RalGDS, strongly associates with b-arrestins [bib_ref] Beta-arrestins regulate a Ral-GDS Ral effector pathway that mediates cytoskeletal reorganization, Bhattacharya [/bib_ref] and Rap1A GTPases [bib_ref] Differential interaction of the ras family GTP-binding proteins H-Ras, Rap1A, and R-Ras..., Herrmann [/bib_ref] , we determined if the expression of Rap1A was also altered in patient breast tumors. In order to quantify Rap1A expression in vivo, Rap1A mRNA levels were measured in an array of 48 cDNA samples derived from breast cancer patients at Stages 0-IV of the disease. We found significantly higher Rap1A mRNA levels in Stage I-IV samples relative to Stage 0 (pathology-confirmed normal) breast tissue [fig_ref] Figure 1: Rap1A is abundantly expressed in human breast tumors [/fig_ref]. We also performed immunohistochemical analyses in a panel of normal, ductal carcinoma in situ (DCIS), and invasive human mammary tissue samples. We observed that in normal ductal cells, which strongly displayed the epithelial marker cytokeratin CAM5.2, contained weak Rap1A expression. The relatively little Rap1A expression was detected in the epithelial tissue and absent from the stromal mammary tissue [fig_ref] Figure 1: Rap1A is abundantly expressed in human breast tumors [/fig_ref]. However, Rap1A was robustly expressed in DCIS lesions and invasive cancer. We have previously shown that the mRNA levels of b-arrestin2 and Ral GTPases are high in human breast tumors that these molecules regulate breast cancer migration and invasion [bib_ref] Betaarrestin/Ral signaling regulates lysophosphatidic acid-mediated migration and invasion of human breast tumor..., Li [/bib_ref]. We also performed immunohistochemistical analysis to look at the localization of b-arrestin2 in normal and tumor tissue. We observed that normal ductal cells, which strongly displayed the epithelial marker cytokeratin CAM5.2 also expressed b-arrestin2 [fig_ref] Figure 1: Rap1A is abundantly expressed in human breast tumors [/fig_ref]. We found that similar to Rap1A, b-arrestin2 was also present in DCIS and invasive infiltrating ductal carcinoma [fig_ref] Figure 1: Rap1A is abundantly expressed in human breast tumors [/fig_ref]. These results therefore show that b-arrestins, but not Rap1 GTPases, are strongly expressed in normal mammary duct, and that both Rap1 and b-arrestins are intensely expressed in DCIS and invasive tumors. ## Rap1a is aberrantly expressed in breast cancer cell lines and regulates invasiveness We next determined whether or not Rap1A proteins were altered in three well-established human breast cancer cell lines (MCF-7, MDA-MB-231, and Hs578T) compared to non-malignant mammary epithelial (MCF-10A) cells. Western blot analysis revealed significantly higher levels of Rap1A in the invasive breast cancer Hs578T and MDA-MB-231, compared to the non-invasive MCF-7 cells or the non-tumorigenic MCF-10A cells [fig_ref] Figure 2: Rap1A is abundantly expressed in invasive human breast cancer cells and regulates... [/fig_ref] , *, p,0.05), thus revealing that Rap1A expression is dysregulated in the aggressive breast cancer cells. Although Rap1 is a key regulator of E-cadherin [bib_ref] Rap1 GTPase: functions, regulation, and malignancy, Hattori [/bib_ref] and has well-established roles in cell adhesion, its role in breast cancer invasiveness is unknown. In order to determine if Rap1A signaling is involved in LPA-mediated breast cancer cell invasion, we depleted Rap1A from the highly aggressive breast cancer MDA-MB-231 cells by stably expressing two individual Rap1A shRNA constructs (shRNA-1 and shRNA-2). The knockdown of Rap1A protein was verified by Western blot analysis [fig_ref] Figure 2: Rap1A is abundantly expressed in invasive human breast cancer cells and regulates... [/fig_ref]. Transwell chamber Matrigel-invasion assays showed that depletion of Rap1A significantly reduced cell invasion towards LPA [fig_ref] Figure 2: Rap1A is abundantly expressed in invasive human breast cancer cells and regulates... [/fig_ref]. The effects of Rap1A knockdown on breast cancer cell invasion was also examined using rigid three-dimensional (3D) cell invasion assays using a reconstituted extracellular matrix (Matrigel) that mimics the in vivo micro-environment [bib_ref] Betaarrestin/Ral signaling regulates lysophosphatidic acid-mediated migration and invasion of human breast tumor..., Li [/bib_ref]. Although cell viability was not affected by the depletion of Rap1A [fig_ref] Figure 1: Rap1A is abundantly expressed in human breast tumors [/fig_ref] , we observed a marked reduction in the number of invasive stellate colonies [fig_ref] Figure 2: Rap1A is abundantly expressed in invasive human breast cancer cells and regulates... [/fig_ref]. These cells formed spherical organoids (spheroids) that were composed of multiple intact nuclei [fig_ref] Figure 2: Rap1A is abundantly expressed in invasive human breast cancer cells and regulates... [/fig_ref] , last column of panel), similar to those displayed by non-malignant MCF-10A cells [bib_ref] Betaarrestin/Ral signaling regulates lysophosphatidic acid-mediated migration and invasion of human breast tumor..., Li [/bib_ref]. These results indicate that Rap1A regulates breast cancer cell invasiveness. Rap1A interacts with LPA 1 and regulates LPA-induced breast cancer cell migration Having observed that Rap1A depletion inhibited invasive properties of tumor cells, we next assessed if the depletion of endogenous Rap1A also blocked LPA-stimulated MDA-MB-231 cell migration. Since LPA is a major constituent of serum, all experiments in this study were done with cells serum-starved for at least 4h. We have previously shown that LPA signaling via pertussis-toxin sensitive G proteins stimulates MDA-MB-231 cell migration and invasion [bib_ref] Betaarrestin/Ral signaling regulates lysophosphatidic acid-mediated migration and invasion of human breast tumor..., Li [/bib_ref]. Additionally, dose-response studies indicated that LPA-mediated migration of MDA-MB-231 began to plateau at the 10 mM LPA concentration and thus all experiments were conducted using this concentration of LPA, as reported in other studies [bib_ref] Lysophosphatidic acid induces MDA-MB-231 breast cancer cells migration through activation of PI3K/PAK1/..., Du [/bib_ref] [bib_ref] Mechanisms for lysophosphatidic acid-induced cytokine production in ovarian cancer cells, Fang [/bib_ref] [bib_ref] Mechanisms in LPA-induced tumor cell migration: critical role of phosphorylated ERK, Stahle [/bib_ref]. Using a scratch assay, we found that Rap1A is necessary for LPA-induced cell motility since loss of Rap1A inhibited scratch closure by with LPA [fig_ref] Figure 3: Rap1A associates with LPA 1 and regulates breast cancer cell migration [/fig_ref] , and movie S1). Because we found that Rap1A regulated LPA-stimulated directional cell migration, we examined if Rap1A expression was required for LPAstimulated lamellipodia formation, using a scratch assay to observe polarized cell migration. LPA stimulated lamellipodia formation that were positive for F-actin [fig_ref] Figure 3: Rap1A associates with LPA 1 and regulates breast cancer cell migration [/fig_ref] , upper panel). Breast cancer cells stably expressing Rap1A shRNA showed a pronounced reduction of LPA-stimulated lamellipodia formation in cells at the leading edge [fig_ref] Figure 3: Rap1A associates with LPA 1 and regulates breast cancer cell migration [/fig_ref] , lower panel) compared to cells expressing scrambled shRNA [fig_ref] Figure 3: Rap1A associates with LPA 1 and regulates breast cancer cell migration [/fig_ref] , upper panel). Furthermore, depletion of Rap1A also inhibited LPA-induced cell migration assessed using Transwell chamber assays [fig_ref] Figure 3: Rap1A associates with LPA 1 and regulates breast cancer cell migration [/fig_ref]. However, we did not observe any effect of Rap1A depletion on cell adhesion [fig_ref] Figure 2: Rap1A is abundantly expressed in invasive human breast cancer cells and regulates... [/fig_ref]. These results reveal that Rap1A is required for LPA stimulated directional migration. We also examined the spatial localization of LPA 1 in migratory cells. For this purpose, we generated MDA-MB-231 cell lines stably expressing FLAG-LPA 1 as we have previously described [bib_ref] Betaarrestin/Ral signaling regulates lysophosphatidic acid-mediated migration and invasion of human breast tumor..., Li [/bib_ref] , and receptors were detected by immunostaining using an anti-FLAG antibody. We found that in unstimulated cells, FLAG-LPA 1 is primarily localized intracellularly, as we have previously reported [bib_ref] Betaarrestin/Ral signaling regulates lysophosphatidic acid-mediated migration and invasion of human breast tumor..., Li [/bib_ref] [fig_ref] Figure 3: Rap1A associates with LPA 1 and regulates breast cancer cell migration [/fig_ref] , left panel). However, upon LPA stimulation, we observed the in addition to being localized intracellularly, LPA 1 re-localized to the plasma membrane, and also at the leading edge of migratory MDA-MB-231 cells [fig_ref] Figure 3: Rap1A associates with LPA 1 and regulates breast cancer cell migration [/fig_ref] , right panel). We next sought to determine whether Rap1 associates with LPA 1 receptor, to mediate LPA signaling. Using co-immunoprecipitation studies, we found that endogenous Rap1 weakly associated with LPA 1 under basal conditions. Upon 5 min of LPA stimulation, the interaction was seen to more than double and then returned to background levels by 15 min after stimulation [fig_ref] Figure 3: Rap1A associates with LPA 1 and regulates breast cancer cell migration [/fig_ref]. Taken together, these results suggest that Rap1 associates with LPA 1 to regulate LPA-induced breast cancer cell migration. ## Lpa activates rap1 in breast cancer cells but not in nonmalignant mammary cells Although LPA has been shown to activate Rap1 in fibroblasts [bib_ref] Extracellular signal-regulated activation of Rap1 fails to interfere in Ras effector signalling, Zwartkruis [/bib_ref] and endothelial cells [bib_ref] ) attenuates LPA(1)-mediated VE-cadherin translocation and cell-cell dissociation through G(12/13) protein-Src-Rap1, Kimura [/bib_ref] , whether or not this occurs in any Immunofluorescence and confocal microscopy on formalin-fixed and paraffin-embedded human tissue samples using (B) a rabbit anti-Rap1A polyclonal antibody followed by Alexa-Fluor 488 secondary antibody (green) and (C) a goat anti-b-arrestin2 polyclonal antibody followed by Alexa-Fluor 555 secondary antibody. In order to distinguish cells of epithelial origin, staining for cytokeratin CAM 5.2 was performed using a mouse monoclonal anti-CAM 5.2 antibody (red). Hematoxylin and eosin staining of normal and tumor sections from each group were examined (representative red boxes). Images are representative of four independent experiments. Scale bar, 20 mm. doi:10.1371/journal.pone.0056174.g001 subjected to a scratch assay, as described above. Immunostaining for FLAG (red) with a rabbit anti-FLAG antibody; nuclei were stained using Hoechst 33258. (E) LPA 1 interacts with endogenous Rap1A. Co-immunoprecipitation studies were performed cancer cell types has not been shown. We thus examined if LPA could stimulate Rap1 activity in breast cancer cells. MDA-MB-231 cells were treated with 10 mM LPA, as we previously determined from dose-response studies [bib_ref] Betaarrestin/Ral signaling regulates lysophosphatidic acid-mediated migration and invasion of human breast tumor..., Li [/bib_ref] ; this concentration of LPA has also been used by other investigators [bib_ref] Mechanisms for lysophosphatidic acid-induced cytokine production in ovarian cancer cells, Fang [/bib_ref] [bib_ref] Mechanisms in LPA-induced tumor cell migration: critical role of phosphorylated ERK, Stahle [/bib_ref]. All assays were done using serum starved cells to eliminate the effects of LPA present in serum. The content of active (GTP-bound) Rap1 was measured through its ability to strongly bind to its effector protein RalGDS, specifically the Rap1-binding domain of RalGDS (RalGDS-RBD). LPA treatment of MDA-MB-231 cells led to increased Rap1 activity within 5 min of stimulation, and diminished within 30 min [fig_ref] Figure 4: LPA stimulates Rap1 activity in breast cancer cells but not in non-malignant... [/fig_ref]. However, LPA failed to activate Rap1 in the nontumorigenic MCF-10A cells [fig_ref] Figure 4: LPA stimulates Rap1 activity in breast cancer cells but not in non-malignant... [/fig_ref] , suggesting that LPA may only induce Rap1 activity selectively in breast cancer cells. We have previously shown that MDA-MB-231 cells express significantly higher levels of LPA 1 than MCF-10A cells [bib_ref] Betaarrestin/Ral signaling regulates lysophosphatidic acid-mediated migration and invasion of human breast tumor..., Li [/bib_ref]. To determine if the lack of LPA 1 in MCF-10A cells is responsible for the inability of LPA to stimulate Rap1, we examined Rap1 activity in LPA-treated MCF-10A cells stably expressing LPA 1 . We found that LPA did indeed induce Rap1 activity in MCF-10A cells expressing FLAG-LPA 1 [fig_ref] Figure 4: LPA stimulates Rap1 activity in breast cancer cells but not in non-malignant... [/fig_ref] ; we have previously shown that expression of LPA 1 in the non-malignant MCF-10A cells stimulates these cells to invade in 3D cultures [bib_ref] Betaarrestin/Ral signaling regulates lysophosphatidic acid-mediated migration and invasion of human breast tumor..., Li [/bib_ref]. Knockdown of b-arrestin2 blocks LPA-stimulated Rap1 activity in breast cancer cells b-arrestin signals via Ral GTPases to mediate breast cancer cell invasiveness [bib_ref] Betaarrestin/Ral signaling regulates lysophosphatidic acid-mediated migration and invasion of human breast tumor..., Li [/bib_ref]. Since RalGDS, a RalGEF and a direct binding partner for b-arrestin2 [bib_ref] Beta-arrestins regulate a Ral-GDS Ral effector pathway that mediates cytoskeletal reorganization, Bhattacharya [/bib_ref] , strongly binds to active Rap1 [bib_ref] Differential interaction of the ras family GTP-binding proteins H-Ras, Rap1A, and R-Ras..., Herrmann [/bib_ref] , we examined if b-arrestin regulates LPA-induced Rap1 activation. For these studies, we employed a b-arrestin2 shRNA MDA-MB-231 cell line exhibiting reduced b-arrestin2 levels that results in impaired LPA-induced migration and invasion [bib_ref] Betaarrestin/Ral signaling regulates lysophosphatidic acid-mediated migration and invasion of human breast tumor..., Li [/bib_ref]. Here we observed that while LPA significantly induced Rap1 activation in control cells expressing scrambled shRNA sequences, LPA failed to alter Rap1 activity in cells where b-arrestin2 levels were reduced [fig_ref] Figure 5: b-arrestin2 regulates Rap1 activity and associates with Rap1A and IQGAP1 in breast... [/fig_ref]. These results indicate that b-arrestin2 is required for LPA-mediated Rap1 activation in breast cancer cells. ## Rap1a and b-arrestin2 associate with iqgap1 in breast cancer cells We next examined if b-arrestin2 associates with Rap1 to regulate its activity. Using co-immunoprecipitation studies, we found that LPA stimulated the endogenous association of Rap1 with b-arrestin2 after 5-15 min which diminished to basal levels within 30 min of stimulation [fig_ref] Figure 5: b-arrestin2 regulates Rap1 activity and associates with Rap1A and IQGAP1 in breast... [/fig_ref]. Rap1A has been shown to interact directly with the actin-binding protein, IQGAP1, which has established roles in breast cancer cell migration and invasion [bib_ref] IQGAP1 promotes cell motility and invasion, Mataraza [/bib_ref] [bib_ref] IQGAP1 binds Rap1 and modulates its activity, Jeong [/bib_ref]. We therefore tested if IQGAP1 was part of the b-arrestin2/Rap1A complex and found that b-arrestin2 did in fact associate with endogenous Rap1A and IQGAP1 in MDA-MB-231 cells [fig_ref] Figure 5: b-arrestin2 regulates Rap1 activity and associates with Rap1A and IQGAP1 in breast... [/fig_ref]. Addition of LPA modulated only the interaction between b-arrestin2 and Rap1A and did not significantly alter the level of association between b-arrestin2 and IQGAP1. Furthermore, we observed that b-arrestin2 associates with endogenous Rap1A upon LPA stimulation in Hs578T breast cancer cells [fig_ref] Figure 5: b-arrestin2 regulates Rap1 activity and associates with Rap1A and IQGAP1 in breast... [/fig_ref]. We also observed that endogenous b-arrestin2 binds IQGAP1 under basal conditions in Hs578T cells. In contrast to MDA-MB-231 cells, treatment with LPA further increased the interaction between b-arrestin2 and IQGAP1 and this association persisted even 30 min post stimulation [fig_ref] Figure 5: b-arrestin2 regulates Rap1 activity and associates with Rap1A and IQGAP1 in breast... [/fig_ref]. The differences in the time frame of association between these signaling molecules might be due to the differences in their expression levels in these two cell lines. Hs578T cells express less b-arrestin2 [bib_ref] Betaarrestin/Ral signaling regulates lysophosphatidic acid-mediated migration and invasion of human breast tumor..., Li [/bib_ref] compared to MDA-MB-231 cells. Overall, these data suggest that the interaction among these endogenous proteins, b-arrestin2, Rap1A and IQGAP1 is not breast cancer cell-type specific IQGAP1 associates with LPA-1 and regulates LPAstimulated breast cancer cell invasion Although it has been demonstrated that IQGAP1 directly interacts with Rap1A in vitro, using purified proteins [bib_ref] IQGAP1 binds Rap1 and modulates its activity, Jeong [/bib_ref] , whether or not this association occurs in breast cancer cells is not known. Thus we next sought to determine whether or not Rap1A associates with IQGAP1. We found that endogenous IQGAP1 associated constitutively with YFP-Rap1A in MDA-MB-231 cells, and this interaction was not modulated by LPA [fig_ref] Figure 6: IQGAP1 associates with Rap1A and LPA 1 and regulates LPA-induced breast cancer... [/fig_ref]. Since IQGAP1 has been shown to associate with various small GTPases including cdc42 and Rac1 [bib_ref] Rac1 and Cdc42 capture microtubules through IQGAP1 and CLIP-170, Fukata [/bib_ref] , we examined whether or not RalA, previously shown to signal downstream of b-arrestin in cancer cells [bib_ref] Betaarrestin/Ral signaling regulates lysophosphatidic acid-mediated migration and invasion of human breast tumor..., Li [/bib_ref] , can associate with IQGAP1. We found that unlike Rap1A, RalA did not co-immunoprecipitate with endogenous IQGAP1 from MDA-MB-231 cells [fig_ref] Figure 6: IQGAP1 associates with Rap1A and LPA 1 and regulates LPA-induced breast cancer... [/fig_ref] , although RalA was present in immunoprecipitates (data not shown), indicating that the interaction of IQGAP1 is specific to a subset of small GTPases. We next sought to determine whether or not IQGAP1 binds LPA 1 and plays a role in transducing its signals to the cytoskeleton to thereby stimulate invasion. We found that endogenous IQGAP1 weakly associated with FLAG-LPA 1 in MDA-MB-231 cells and this interaction increased significantly upon treatment of cells with LPA for 5 min [fig_ref] Figure 6: IQGAP1 associates with Rap1A and LPA 1 and regulates LPA-induced breast cancer... [/fig_ref]. The interaction between IQGAP1 and the receptor decreased within 15 min of stimulation, approaching basal levels after 30 min of LPA stimulation. To determine a role for IQGAP1 in LPA-induced breast cancer invasiveness, IQGAP1 expression was stably depleted in MDA-MB-231 cells which express the highest level of this protein [bib_ref] IQGAP1 stimulates proliferation and enhances tumorigenesis of human breast epithelial cells, Jadeski [/bib_ref] using IQGAP1 siRNA constructs [bib_ref] IQGAP1 promotes cell motility and invasion, Mataraza [/bib_ref] [fig_ref] Figure 6: IQGAP1 associates with Rap1A and LPA 1 and regulates LPA-induced breast cancer... [/fig_ref]. We found that depletion of IQGAP1 from MDA-MB-231 cells significantly blocked serum-induced invasion compared to cells expressing scrambled controls [fig_ref] Figure 6: IQGAP1 associates with Rap1A and LPA 1 and regulates LPA-induced breast cancer... [/fig_ref]. Furthermore, diminished levels of IQGAP1 in MDA-MB-231 cells abrogated invasion and migration of breast cancer cells towards LPA, in the absence of serum, compared to cells expressing scrambled controls [fig_ref] Figure 6: IQGAP1 associates with Rap1A and LPA 1 and regulates LPA-induced breast cancer... [/fig_ref]. These results implicate IQGAP1 as a novel regulator of LPAmediated breast cancer cell invasion. b-Arrestin2 depletion disrupts LPA-stimulated lamellipodia formation and IQGAP1 localization at the leading edge Overall, our data indicate that b-arrestin2 scaffolds IQGAP1 and Rap1A in breast cancer cells and that the b-arrestin2/Rap1A interaction is LPA-dependent. The leading edge of motile cells compartmentalizes different protein complexes that orchestrate oriented cell motility [bib_ref] Life at the leading edge, Ridley [/bib_ref]. Studies have revealed that the actin scaffolding protein IQGAP1 localizes to the leading edge of motile using serum-starved (4h) cells stably expressing FLAG-LPA 1 and using a mouse anti-FLAG monoclonal antibody. Endogenous Rap1A was detected by a rabbit polyclonal anti-Rap1A antibody. A representative blot from three independent experiments is shown. Graph illustrates densitometric analysis of blots; data are expressed relative to unstimulated cells. Error bar, SEM *, p,0.05. doi:10.1371/journal.pone.0056174.g003 cells and interacts with actin filaments to cross-link them and thereby regulate cell migration [bib_ref] IQGAP1 in cellular signaling: bridging the GAP, Brown [/bib_ref] [bib_ref] IQGAP1 promotes cell motility and invasion, Mataraza [/bib_ref] [bib_ref] IQGAP1: a key regulator of adhesion and migration, Noritake [/bib_ref]. b-Arrestins have also been shown to facilitate cell migration by sequestering regulators of actin assembly at the leading edge [bib_ref] Beta-arrestins as regulators of signal termination and transduction: how do they determine..., Defea [/bib_ref]. We therefore investigated whether the b-arrestin2 was necessary for the recruitment of IQGAP1 to the leading edge of migratory cells. We first characterized the sub-cellular localization of b-arrestin2 and IQGAP1 in motile MDA-MB-231 breast cancer cells that were scratched and allowed to migrate in the presence or absence of LPA. In the absence of LPA, we found that b-arrestin2 colocalized with IQGAP1 in cells ; this is consistent with our finding that b-arrestin2 and IQGAP1 are found in a complex under basal condition [fig_ref] Figure 5: b-arrestin2 regulates Rap1 activity and associates with Rap1A and IQGAP1 in breast... [/fig_ref]. Upon LPA treatment, b-arrestin2 co-localized with IQGAP1 in lamellipodia and membrane ruffles that were formed at the leading edge of migrating cells . Breast cancer cells stably expressing b-arrestin2 shRNA showed a pronounced reduction of LPA-stimulated lamellipodia formation in cells at the leading edge [fig_ref] Figure 8: b-arrestin2 regulates LPA-stimulated localization of IQGAP1 to the leading edge [/fig_ref] , lower panel) compared to cells expressing scrambled shRNA [fig_ref] Figure 8: b-arrestin2 regulates LPA-stimulated localization of IQGAP1 to the leading edge [/fig_ref]. Furthermore, we observed less accumulation of endogenous IQGAP1 at the leading edge of cells expressing b-arrestin2 shRNA, compared to scrambled controls [fig_ref] Figure 8: b-arrestin2 regulates LPA-stimulated localization of IQGAP1 to the leading edge [/fig_ref]. These results reveal that b-arrestin2 is required for the localization of IQGAP1 at the leading edge of motile breast cancer cells and for directional migration. This far our data indicate that b-arrestin2 scaffolds Rap1A and IQGAP1 under basal conditions, and LPA stimulation further enhances these interactions [fig_ref] Figure 5: b-arrestin2 regulates Rap1 activity and associates with Rap1A and IQGAP1 in breast... [/fig_ref]. Since IQGAP1 binds Rap1A and b-arrestin2 constitutively [fig_ref] Figure 5: b-arrestin2 regulates Rap1 activity and associates with Rap1A and IQGAP1 in breast... [/fig_ref] , respectively), we therefore sought to determine whether depletion of IQGAP1 from MDA-MB-231 cells impairs Rap1A/b-arrestin2 interactions. Indeed we found a significant loss of binding between these two proteins in cells stably expressing IQGAP1 siRNA, compared to scrambled controls [fig_ref] Figure 8: b-arrestin2 regulates LPA-stimulated localization of IQGAP1 to the leading edge [/fig_ref]. Taken together, our results suggest that a complex of b-arrestin2, Rap1A and IQGAP1 regulates LPA induced lamellipodia formation and breast cancer cell migration. # Discussion In this study, we are the first to describe a link between LPA 1 stimulation, b-arrestin2-containing complexes and the activation of Rap1 GTPases that is necessary for the acquisition of a migratory and invasive phenotype. We have identified that Rap1A and the scaffolding protein IQGAP1 are novel binding partners for b-arrestin2 as well as LPA 1 receptors in breast cancer cells. We found that Rap1A, as well as b-arrestin2 and LPA 1 [bib_ref] Betaarrestin/Ral signaling regulates lysophosphatidic acid-mediated migration and invasion of human breast tumor..., Li [/bib_ref] are aberrantly expressed in breast cancer cells and patient tumors, and previous studies have shown that IQGAP1 levels are elevated in breast carcinoma [bib_ref] IQGAP1 stimulates proliferation and enhances tumorigenesis of human breast epithelial cells, Jadeski [/bib_ref]. We thus postulate that in breast tumors, LPA 1 activation recruits b-arrestins and leads to the activation of Rap1, and signaling via IQGAP1. This modulates the actin cytoskeleton, a process necessary for cell migration and invasion. Based on our data, we propose a model for chemoattractantinduced breast cancer cell cytoskeletal reorganziation. Under basal conditions, IQGAP1 scaffolds b-arrestin2 and Rap1A in cells. In response to LPA 1 activation, b-arrestin2 is recruited to LPA 1 , to mediate LPA 1 endocytosis [bib_ref] A requirement for membrane cholesterol in the beta-arrestin-and clathrin-dependent endocytosis of LPA1..., Urs [/bib_ref] , LPA enhances the binding of b-arrestin2 to Rap1A which then activates this GTPase. b-Arrestin2 also recruits IQGAP1 to the leading edge of migratory cells. Whether LPA 1 activation further regulates the level of association between b-arrestin2/IQGAP1 appears to be cell-specific. Although Rap1 has been implicated in growth factor signaling pathways [bib_ref] Rap1, a mercenary among the Ras-like GTPases, Frische [/bib_ref] [bib_ref] Rap1 GTPase: functions, regulation, and malignancy, Hattori [/bib_ref] [bib_ref] Linking Rap to cell adhesion, Bos [/bib_ref] , the direct involvement of Rap1A in LPA b-Arrestins function as signaling scaffolds facilitate cell migration by sequestering regulators of actin assembly at the leading edge [bib_ref] beta-arrestin-dependent actin reorganization: bringing the right players together at the leading edge, Min [/bib_ref] [bib_ref] G protein-coupled receptors stimulation and the control of cell migration, Cotton [/bib_ref]. b-Arrestins associate with cofilin, an actin filament severing protein, localizing it to the leading edge of cells to promote barbed end formation and membrane protrusions [bib_ref] Beta-arrestins as regulators of signal termination and transduction: how do they determine..., Defea [/bib_ref]. b-Arrestins also interact directly with Arp2/3 complex and WASP family of proteins and can thereby affect actin nucleation [bib_ref] Functional specialization of beta-arrestin interactions revealed by proteomic analysis, Xiao [/bib_ref]. IQGAP1 also serves as a scaffold to integrate multiple signaling molecules to facilitate cell migration [bib_ref] Sacks DB (2011) IQGAP1 and its binding proteins control diverse biological functions, White [/bib_ref]. b-Arrestins and IQGAP1 can bind microtubules that control polarized cell motility [bib_ref] IQGAP1: a key regulator of adhesion and migration, Noritake [/bib_ref] [bib_ref] Arrestin mobilizes signaling proteins to the cytoskeleton and redirects their activity, Hanson [/bib_ref]. We establish that the molecular scaffolds b-arrestins and IQGAP1 associate with each other in breast cancer cells, to perhaps fine tune their signaling capacity. Indeed other reports have described complex formation between scaffolds [bib_ref] Scaffold proteins: hubs for controlling the flow of cellular information, Good [/bib_ref] [bib_ref] Organizing signal transduction through A-kinase anchoring proteins (AKAPs), Logue [/bib_ref]. We identified a novel interaction between LPA 1 and IQGAP1 in MDA-MB-231 cells. IQGAP1 has been shown to be enriched at invadopodia in MDA-MB-231 cells, structures that form at contact sites between invasive tumor cells and the extracellular matrix [bib_ref] The interaction of IQGAP1 with the exocyst complex is required for tumor..., Sakurai-Yageta [/bib_ref]. Although it is not surprising that knock-down of IQGAP1 disrupted general cell invasion, we found that LPAinduced breast cancer cell invasion specifically required the presence of IQGAP1, thus establishing it as a key player in this process. A recent study reported that b-arrestin2-mediated scaffolding of the RAF-MEK-ERK1/2 signaling components is required for the expression and secretion of MMP-1 by bronchial epithelial cells [bib_ref] Diesel exhaust particles activate the matrix-metalloproteinase-1 gene in human bronchial epithelia in..., Li [/bib_ref]. Thus, both IQGAP1 and b-arrestins act to scaffold components of the MEK/ERK pathways [bib_ref] IQGAP1 in cellular signaling: bridging the GAP, Brown [/bib_ref] , suggesting that IQGAP1 and b-arrestin2 may utilize the MAPK signaling cascade to regulate LPA-induced breast cancer cell invasiveness. Rap1 has been shown to elicit both pro-and anti-invasive phenotypes depending on the cancer type [bib_ref] Rap1 GTPase: functions, regulation, and malignancy, Hattori [/bib_ref] [bib_ref] Rap1 integrates tissue polarity, lumen formation, and tumorigenic potential in human breast..., Itoh [/bib_ref] [bib_ref] Activation of Rap1 promotes prostate cancer metastasis, Bailey [/bib_ref] [bib_ref] Rap1GAP promotes invasion via induction of matrix metalloproteinase 9 secretion, which is..., Mitra [/bib_ref]. However, few studies have examined a role for Rap1 in breast cancer [bib_ref] Rap1 integrates tissue polarity, lumen formation, and tumorigenic potential in human breast..., Itoh [/bib_ref] [bib_ref] Polymorphisms of the SIPA1 gene and sporadic breast cancer susceptibility, Hsieh [/bib_ref]. Rap1 activity was found to be higher in malignant human T4-2 breast cancer cells compared to the non-malignant S1 cells [bib_ref] Rap1 integrates tissue polarity, lumen formation, and tumorigenic potential in human breast..., Itoh [/bib_ref] and expression of constitutively active Rap1 in T4-2 cells yields larger tumors with higher grade of malignancy in vivo [bib_ref] Rap1 integrates tissue polarity, lumen formation, and tumorigenic potential in human breast..., Itoh [/bib_ref]. Recently, Rap1 has been shown regulate breast cancer cell migration via the adhesion molecule JAM-A [bib_ref] Breast cancer cell migration is regulated through junctional adhesion molecule-Amediated activation of..., Mcsherry [/bib_ref]. We found higher levels of Rap1A mRNA in Stage I-IV breast cancer tumors relative to Stage 0, normal non-diseased tissue samples. This implicates Rap1A in tumorigenesis before the cancer has spread to the nodes (Stage I) as well as during tumor metastasis (Stages II-IV). Correspondingly, we observed that Rap1A protein expression increased in DCIS and invasive tumors in comparison to normal ducts. These findings are in line with recent findings by Furstenau et al. that show that Rap1 is weakly expressed in normal and benign breast tissue while DCIS lesions and areas with invasive tumors show enhanced Rap1 expression [bib_ref] Rasrelated protein 1 and the insulin-like growth factor type I receptor are..., Furstenau [/bib_ref]. We also observed increased levels of Rap1A protein in invasive breast cancer cells (MDA-MB-231 and Hs578T) compared to non-invasive MCF-7 cells or the non-malignant MCF-10A cells. Overall, these results indicate that the activity or function of Rap1A may be altered in aggressive cancer cells. Indeed, LPA only induced Rap1 activity in the highly metastatic MDA-MB-231 cells and not in the nonmalignant MCF-10A cells. These results suggest that chemoattractants such as LPA may selectively stimulate Rap1 in tumor cells. We identified a novel interaction between LPA 1 and Rap1A in breast cancer cells. MDA-MB-231 cells primarily express LPA 1 mRNA [bib_ref] Betaarrestin/Ral signaling regulates lysophosphatidic acid-mediated migration and invasion of human breast tumor..., Li [/bib_ref]. Studies looking at endogenous LPA receptors could not be conducted due to lack of effective antibodies against LPA receptors. The binding of Rap1A to LPA 1 might stimulate signaling pathways involved in cell motility and invasion. Reduction of Rap1A expression using shRNA sequences inhibited LPA-stimulated breast cancer cell migration and formation of invasive stellate structures in 3D cultures. These 3D tumor invasion models provide important insights into breast cancer biology including cellular differentiation and tissue organization [bib_ref] Betaarrestin/Ral signaling regulates lysophosphatidic acid-mediated migration and invasion of human breast tumor..., Li [/bib_ref]. Although the specific mechanisms by which Rap1A mediates cell invasion are unclear, Rap1A has been shown to stimulate MMP-7 gene transcription in squamous cell carcinoma cells via bcatenin [bib_ref] Rap1 stabilizes betacatenin and enhances beta-catenin-dependent transcription and invasion in squamous cell..., Goto [/bib_ref]. The possible role of Rap1 in the expression of genes involved in breast cancer invasion is currently under investigation. In summary, our data indicate that b-arrestin2 regulates multiple and distinct, yet convergent, signaling pathways to modulate the cytoskeletal reorganization necessary for breast cancer cell migration and invasion. Whether the b-arrestin2/ Rap1A/IQGAP1 pathway regulates cell migration downstream of other GPCRs remains to be tested. Since b-arrestins regulate signaling of several types of receptor systems, a better understanding of the molecular pathways by which b-arrestins regulate breast cancer invasiveness will aid in the identification of new targets for the design of novel therapeutics against metastatic disease. # Materials and methods ## Cell culture Cell lines (MCF-10A, MCF-7, MDA-MB-231 and Hs578T) were from American Type Culture Collection (ATCC) and grown as described [bib_ref] Betaarrestin/Ral signaling regulates lysophosphatidic acid-mediated migration and invasion of human breast tumor..., Li [/bib_ref]. Cells were directly obtained from and characterized by ATCC as per Cell Line Verification Test Recommendations (ATCC Technical Bulletin No. 8). All cells were passaged in our laboratory for a maximum of two months. ## Reagents and constructs LPA (18:1) is from Avanti Polar Lipids. YFP-Rap1A and myc-Rap1A are generous gifts from Dr. Mina Bissell (Berkeley, CA) and Dr. Johannes Bos (Utrecht, Netherlands), respectively. FLAG-LPA 1 has been described [bib_ref] Betaarrestin/Ral signaling regulates lysophosphatidic acid-mediated migration and invasion of human breast tumor..., Li [/bib_ref]. pSuper-IQGAP1 siRNA were generously provided by Dr. David Sacks (Bethesda, MD). Rap1A-pRS-shRNA is from Origene Technologies. ## Quantitative real-time pcr Experiments were done as described [bib_ref] Betaarrestin/Ral signaling regulates lysophosphatidic acid-mediated migration and invasion of human breast tumor..., Li [/bib_ref]. A cDNA array generated from breast cancer tissue (Stages 0-IV) was purchased from Origene Technologies. Stage 0 samples were derived from . b-Arrestin2 co-localizes with IQGAP1 at the leading edge of migratory breast cancer cells. MDA-MB-231 cells were serum starved for 4h and scratched with a sterile pipette tip. Cells were then placed in serum-free medium containing 10 mM LPA and allowed to migrate into the scratch. Cells were fixed, permeabilized, immunostained and visualized using confocal microscopy. Time course of endogenous IQGAP1 immunostaining performed with rabbit anti-IQGAP1 antibody followed by Alexa-Fluor 488 (green). Endogenous b-arrestin2 immunostaining was performed with a polyclonal goat anti-b-arrestin2 antibody followed by Alexa-Fluor 555 (red). Areas of colocalization in lamellipodia at the leading edge of motile (arrows, overlay) is shown in yellow. Scale bar, 20 mm. doi:10.1371/journal.pone.0056174.g007 tissues that were determined to be normal, as determined by a pathologist (www.origene.com). The sequences for Rap1A primers are 5`-ACAGGACCTGAGGGAACAGA-3`and 5`-CCCTGCTCTTTGCCAACTAC-3`. ## Tissue staining and confocal microscopy Formalin-fixed and paraffin-embedded human normal breast tissue, DCIS, and invasive cancer tissue sections (5 mm thick) were obtained from London Health Sciences Center and characterized by a pathologist on our team (Dr. Siddiqui). These studies were conducted in consultation with a pathologist on our team (Dr. Siddiqui). Serial sections were de-paraffinized, cleared, rehydrated through an ethanol series, and subjected to either hematoxylin and eosin staining or immunofluorescence (IF) using polyclonal anti-Rap1A rabbit antibody (1:50, Santa Cruz), polyclonal anti-b-arrestin2 goat antibody (1:50, C-18, Santa Cruz) and monoclonal anti-CAM5.2 mouse antibody (undiluted, BD Biosciences). Tissue sections were imaged using Zeiss LSM-510 META laser scanning microscope (Zeiss). ## Stable transfections and mrna silencing MCF-10A and MDA-MB-231 cells were electroporated with 25 mg of DNA using Gene Pulser Xcell (Bio-Rad) as described [bib_ref] Betaarrestin/Ral signaling regulates lysophosphatidic acid-mediated migration and invasion of human breast tumor..., Li [/bib_ref]. Heterogeneous populations of cells transfected with YFP-Rap1A or FLAG-LPA 1 were selected using G418. Generation of MDA-MB-231 cells stably expressing b-arrestin 2 shRNA constructs have been previously described [bib_ref] Betaarrestin/Ral signaling regulates lysophosphatidic acid-mediated migration and invasion of human breast tumor..., Li [/bib_ref]. To knockdown Rap1A, four different Rap1A shRNA were individually electroporated into MDA-MB-231 cells. The two constructs that gave the best knockdown were selected and a heterogeneous population of cells expressing each Rap1A shRNA construct was used for subsequent studies (shRNA-1, GCCAACAGTGTATGCTC-GAAATCCTGGAT; shRNA-2, CTATTACAGCTCA-GTCCACGTTTAACGAC).SiIQGAP1-pSuperRenilla [bib_ref] IQGAP1 promotes cell motility and invasion, Mataraza [/bib_ref] was microporated (Bio-Rad) according to the manufacturer's instructions. ## Co-immunoprecipitation and immunoblots Rap1A expression in cell lines was examined using a rabbit anti-Rap1A antibody (1:500, Santa Cruz). Equal loading of the samples was verified by checking b-actin expression (Rabbit antibody, 1:2000, Sigma Aldrich). Co-immunoprecipitation experiments between endogenous proteins were conducted as previously described [bib_ref] Betaarrestin/Ral signaling regulates lysophosphatidic acid-mediated migration and invasion of human breast tumor..., Li [/bib_ref]. To investigate the interaction between LPA 1 receptor and YFP-Rap1A, MDA-MB-231 FLAG-LPA 1 stables [bib_ref] Betaarrestin/Ral signaling regulates lysophosphatidic acid-mediated migration and invasion of human breast tumor..., Li [/bib_ref] were stably transfected with YFP-Rap1A using Lipofectamine 2000 (Invitrogen) and cell lysates subjected to co-immunoprecipitation assays. ## Rap1 activation assay A Rap1-activation assay was performed according to the manufacturer's instructions (Upstate Millipore). Cells were serum-starved for 4h and active Rap1 was pulled down from lysates using agarose beads conjugated to RalGDS-RBD and detected by Western blot analysis using a rabbit anti-Rap1 antibody (Upstate). ## Immunostaining Assays were conducted as previously described [bib_ref] Betaarrestin/Ral signaling regulates lysophosphatidic acid-mediated migration and invasion of human breast tumor..., Li [/bib_ref]. We employed rabbit anti-IQGAP1 (1:500, Santa Cruz) and goat anti-b-arrestin2 (C-18, 1:50, Santa Cruz) primary antibodies followed by anti-rabbit AlexaFluor-488 (1:250, Molecular Probes) and anti-goat AlexaFluor555 (1:1000, Molecular Probes) secondary antibodies. Nuclei were labeled using Hoechst 33258 (1:10000, Invitrogen). All images were taken using Zeiss LSM-510 META laser scanning microscope (Zeiss). ## 3d invasion assays 3D invasion assays were conducted as described [bib_ref] Betaarrestin/Ral signaling regulates lysophosphatidic acid-mediated migration and invasion of human breast tumor..., Li [/bib_ref]. Cells were plated in a 1:1 dilution of phenol red-free Matrigel and culture medium containing 2.5610 6 cells/mL on Matrigel-coated 35 mm glass-bottomed culture dishes (MatTek). Cell colonies were scored in a blinded manner as either stellate or spheroidal. Stellate colonies were characterized by one or more projections from the central sphere of cells. Images were taken on an IX-81 microscope (Olympus) using InVivo Analyzer Suite (Media Cybernetics). ## Cell migration and invasion assays Transwell chamber migration and Matrigel-invasion assays (Millipore filters, 8 mm pores) were performed as detailed [bib_ref] Betaarrestin/Ral signaling regulates lysophosphatidic acid-mediated migration and invasion of human breast tumor..., Li [/bib_ref]. Serum-starved (4h) cells [bib_ref] IQGAP1 promotes cell motility and invasion, Mataraza [/bib_ref] were placed in the upper chamber and allowed to migrate for 18 hours. Ten fields of nuclei images were acquired and quantified (20x magnification). Migration was expressed as a percentage of control cells. We typically observed approximately 200 cells migrated per field. ## Cell adhesion assays Quantitative cell adhesion assays were carried out as described [bib_ref] AFAP-110 is required for actin stress fiber formation and cell adhesion in..., Dorfleutner [/bib_ref] , using non-tissue culture-treated 96-well plates coated with 10 mg/ml fibrnectin for 1h, at room temperature. Briefly, fibronectin coated and uncoated control wells were blocked for 1 h with 5% bovine serum albumin (BSA). Cells were split one day prior to the experiment to achieve a subconfluent culture. Briefly, serum-starved cells (2X10 4 ) were added to each well and incubated for 60 min at 37uC. Each experiment was done in triplicate. Unattached cells were removed and stained with trypan blue to exclude the possibility of lack of adhesion, due to cell death. Wells were washed three times with PBS and to exclude nonspecific adhesion. Attached cells were fixed in 3.7% formaldehyde, rinsed with PBS, and stained with 1% crystal violet in 2% ethanol, lysed in 10% acetic acid and OD 600 was determined in a plate reader. ## Scratch assays for cell motility Serum-starved (4h) cells were scratched and allowed to migrate for 16 hours. Time-lapse images and movies of scratch closure were collected every 10 minutes using an IX-81 inverted microscope (Olympus). Distance travelled (in mm) was measured using ImagePro software (Media Cybernetics). Experiments were done in duplicate. The quantitation of the number of cells containing proteins at the leading edge of the cell was carried out using 100 cells/condition from three experiments. Briefly, a cell containing a positive immunofluorescence signal only along the edge of the plasma membrane that was directly adjacent to the scratch was scored as positive for leading edge localization. If a cell contained signal along the complete cell periphery, it was scored as negative for leading edge staining. # Statistical analysis Student's t test or one-way ANOVA was used to assess statistical significance with GraphPad Prism 5.0 software (GraphPad Software, Inc.). Differences with p , 0.05 were considered significant. ## Supporting information [fig] Figure 1: Rap1A is abundantly expressed in human breast tumors.(A) Patient breast tumors from Stages I-IV of the disease express higher levels of Rap1A mRNA compared to non-diseased Stage 0 samples, as determined by quantitative real-time PCR. Data points shown are log-transformed values of relative mRNA expressions compared to average expression in Stage 0 samples. Levels of b-actin were used by the manufacturer to pre-normalize samples. One-way ANOVA statistical analysis was performed on the log-transformed relative expression values and a Dunnett's multiple comparison posthoc test was conducted. a, p,0.01, compared to Stage 0. (B) [/fig] [fig] Figure 2: Rap1A is abundantly expressed in invasive human breast cancer cells and regulates invasiveness.(A) The aggressive breast cancer cell lines MDA-MB-231 and Hs578T express elevated levels of Rap1A protein relative to the non-invasive MCF-7 cells or the non-malignant MCF-10A mammary epithelial cells. Endogenous expression of Rap1A in each cell line was determined by Western blot analysis. A representative blot from five separate experiments is shown. In the graphs, bars represent relative expression to MCF-10A cells. Error bar, SEM *, p,0.05. (B) Knockdown of Rap1A protein expression in MDA-MB-231 cells stably expressing two separate Rap1A shRNA constructs was verified by Western blot analysis. Densitometric analysis shows significantly lower levels of Rap1A expression in Rap1A shRNA expressing cells. *, p,0.05. (C) Knock-down of endogenous Rap1A in serum-starved MDA-MB-231 cells inhibits LPA (10 mM) stimulated cell invasion as assessed using Transwell chamber (Matrigel) invasion assays. (D) MDA-MB-231 cells expressing Rap1A shRNA and suspendend in Matrigel for 5 days display reduced growth of invasive stellate structures relative to cells expressing scrambled control. Nuclei were stained using Hoechst 33258 and visualized by confocal microcopy (blue, last column of panel). Images are representative of five independent experiments, scale bar = 40 mm. Shape of the colonies was scored as either stellate or spheroidal and the number of stellate colonies is expressed as percentage of the total number of colonies. *, p,0.05 compared to formation of stellate structures by MDA-MB-231 cells expressing scrambled shRNA. doi:10.1371/journal.pone.0056174.g002 [/fig] [fig] Figure 3: Rap1A associates with LPA 1 and regulates breast cancer cell migration. (A) Depletion of Rap1A expression significantly blocks scratch closure of MDA-MB-231 cells in response to LPA treatment. Cells were grown to confluence, serum starved for four hours and scratched with a sterile pipette tip. Serum-free medium containing 10 mM LPA was added to the cells which allowed to migrate into the scratch for 16 hours and visualized using an IX-81 Olympus microscope. Cell migration was quantitated using ImagePro software and bars represent the mean distance travelled (in mm) after 16 hours of culturing in LPA. Data from four experments.*, p,0.05, scale bar = 100 mm. (B) LPA (10 mM) stimulates lamellipodia formation at the leading edge of motile MDA-MB-231cells expressing scrambled shRNA (control, upper panels) in a scratch assay. The effect of depleting endogenous Rap1A from LPA-stimulated cells is shown (lower panel). Cells, grown to confluency were serum starved for 4h and scratched with a sterile pipette tip. Cells were then placed in media containing LPA and allowed to migrate into the scratch for 4 hours, fixed, permeabilized, immunostained and visualized using confocal microscopy. Alexa-Fluor 555 conjugated phalloidin (red) was used to stain F-actin. Loss of color (grayscale image) shows the lamellipodia at the leading edge (white arrowheads). Direction of cell migration shown by blue arrow. Scale bar, 20 mm.Quantitation of cells with lamellipodia was carried out as described under ''Materials and Methods'' and graphed. Error bar, SEM *, p,0.05. Data from three independent experiments. (C) Migration of serum-starved, Rap1A-depleted MDA-MB-231 cells towards 10 mM LPA is significantly inhibited, as determined by Transwell chamber assays. *, p,0.05. Data from four experments. (D) Confocal micrographs showing the localization of LPA 1 in MDA-MB-231 cells stably expressing FLAG-LPA 1 [/fig] [fig] Figure 4: LPA stimulates Rap1 activity in breast cancer cells but not in non-malignant cells.MDA-MB-231 breast cancer cells were serum-starved for 4 hours and stimulated with 10 mM LPA for the indicated times. Cells were then lysed and activated Rap1 was pulled down using RalGDS-RBD-conjugated beads and immunoblotted using a rabbit anti-Rap1 antibody. A representative blot from four independent experiments is shown. (B) MCF-10A cells were stimulated with 10 mM LPA for the indicated times and Rap1 activation was determined. Blot is representative of three independent experiments. (C) MCF-10A cells stably expressing FLAG-LPA 1 were treated for 15 minutes with 10 mM LPA and subjected to Rap1 activation assay. Blot is representative of three independent experiments. doi:10.1371/journal.pone.0056174.g004induction of lamellipodia formation and forward movement is a novel function for this GTPase. [/fig] [fig] Figure 5: b-arrestin2 regulates Rap1 activity and associates with Rap1A and IQGAP1 in breast cancer cells. (A) Depletion of b-arrestin2 in breast cancer cells blocks LPA-induced Rap1 activation. MDA-MB-231 cells stably expressing either scrambled shRNA or b-arrestin2-targeting shRNA were serum starved for 4 hours and stimulated with 10 mM LPA. Cells were lysed and activated Rap1 was pulled down. Knockdown of b-arrestin2 in MDA-MB-231 cells stably expressing shRNA against b-arrestin2 was verified by Western blot analysis. Representative blots from four independent experiments are shown. (B) The association of endogenous proteins, b-arrestin2, Rap1A and IQGAP1, in breast cancer cells. MDA-MB-231 cells were serum-starved for 4 hours and stimulated with LPA in serum-free media (10 mM). Isolated lysates were subjected to coimmunoprecipitation assays using a rabbit anti-Rap1A antibody or rabbit anti-IQGAP1 antibody and immunoblotted using a monoclonal anti-b-arrestin2 antibody. A representative blot from four independent experiments is shown. Graph illustrates densitometric analysis of blots; data are expressed relative to unstimulated cells. Error bar, SEM *, p,0.05. (C) Endogenous proteins, b-arrestin2, Rap1A and IQGAP1 association in Hs578T cells as described in (B). A representative blot from three independent experiments is shown. doi:10.1371/journal.pone.0056174.g005 [/fig] [fig] Figure 6: IQGAP1 associates with Rap1A and LPA 1 and regulates LPA-induced breast cancer cell invasion. (A) Endogenous IQGAP1 binds Rap1A in MDA-MB-231 breast cancer cells. Serum-starved (4h) cells stably expressing YFP-Rap1A or GFP vector control were stimulated with LPA (10 mM) and lysates were subjected to immunoprecipitation assays with an anti-GFP monoclonal antibody and immunoblotted with a rabbit anti-IQGAP1 polyclonal antibody. (B) IQGAP1 does not associate with RalA in MDA-MB-231. Serum starved (4h) cells were treated with 10 mM LPA and coimmunoprecipitation studies done with an anti-GFP monoclonal antibody. Endogenous IQGAP1 was detected using a rabbit anti-IQGAP1 antibody. (C) LPA 1 interacts with endogenous IQGAP1. Co-immunoprecipitation studies were performed using serum-starved (4h) cells stably expressing FLAG-LPA 1 and using a mouse anti-FLAG monoclonal antibody. Endogenous IQGAP1 was detected by a rabbit polyclonal anti-IQGAP1 antibody. (D) Depletion of IQGAP1 blocks LPA-induced cell migration and invasion. Serum starved (4h) MDA-MB-231 cells expressing IQGAP1 siRNA were subjected to Transwell migration or Matrigel-invasion assays towards 10 mM LPA (in serum-free media). Results represent data from three independent experiments. *, p,0.05. IQGAP1 expression in MDA-MB-231 cells stably expressing siRNA against IQGAP1 was verified by Western blot analysis. A representative blot from three independent experiments is shown. doi:10.1371/journal.pone.0056174.g006 [/fig] [fig] Figure 8: b-arrestin2 regulates LPA-stimulated localization of IQGAP1 to the leading edge. (A) MDA-MB-231 cells were serum starved for 4h and scratched with a sterile pipette tip. Cells were then placed in serum-free medium containing 10 mM LPA and allowed to migrate into the scratch. Cells were fixed, permeabilized, immunostained and visualized using confocal microscopy. LPA (10 mM) stimulates lamellipodia formation and the accumulation of IQGAP1 at the leading edge of motile MDA-MB-231cells (white arrowheads) expressing scrambled shRNA (control, upper panel) in a scratch assay. The effect of depleting endogenous b-arrestin2 from LPA-stimulated cells is shown (lower panel). Endogenous IQGAP1 immunostaining was performed with rabbit anti-IQGAP1 antibody followed by Alexa-Fluor 488 (green). Scale bar, 20 mm. (B) b-arrestin2 expression in MDA-MB-231 cells stably expressing shRNA against b-arrestin2 was verified by Western blot analysis. A representative blot from three independent experiments is shown. Graph showing blind quantification of cells containing IQGAP1 at the leading edge of migrating cells from (A) carried out after stimulation of cells with LPA for 4 hours, as described under ''Experimental Procedures''. Error bar, SEM *, p,0.05. Data from three independent experiments. (C) Depletion of IQGAP1 blocks the endogenous association of Rap1A and b-arrestin2 in MDA-MB-231 cells. Co-immunoprecipitation studies were performed using a rabbit polyclonal anti-Rap1A antibody. Endogenous b-arrestin2 was detected with a monoclonal anti-b-arrestin2 antibody. A representative blot from three independent experiments. doi:10.1371/journal.pone.0056174.g008 [/fig] [fig] Figure S1: Knockdown of Rap1A by shRNA does not affect cell viability.MDA-MB-231 cells stably expressing scrambled shRNA or Rap1A shRNA were subjected to MTT cell viability assays. Absorbance was measured at 575 nm and a reference reading at 750 nm. Columns represent mean absorbance from three independent experiments. (TIF) Figure S2 Effect of Rap1A depletion on MDA-MB-231 cell adhesion. MDA-MB-231 cells stably expressing scrambled shRNA or Rap1A shRNA were subjected to adhesion assays to fibronectin. Columns show % adhesion compared to scrambled control; data from four independent experiments. (TIF) Movie S1 Knockdown of Rap1A by shRNA inhibits motility of MDA-MB-231 cells by scratch assay. The top panel of the video shows live imaging of cells expressing a scrambled control while the bottom portion shows cells knocked down for Rap1A. Serum-starved cells were stimulated with 10 mM LPA and live-cell microscopy conducted using an IX-81 microscope once every 10 min over a course of 16 hours and compiled into a time-lapse video. Top panel, cells expressing scrambled shRNA control; bottom panel, cells expressing Rap1A shRNA. (MOV) [/fig]

Kounis Syndrome Associated With Takotsubo Syndrome in an Adolescent With Peutz-Jeghers Syndrome We describe the case of a 15-year-old female patient with Peutz-Jeghers syndrome who presented with vomiting and abdominal pain secondary to ileoileal invagination. Initial analgesic treatment was not effective, and subsequent tramadol infusion resulted in clinical manifestations compatible with Kounis and Takotsubo syndromes. However, the patient had an excellent recovery. (Level of Difficulty: Advanced.) (J Am Coll Cardiol Case Rep 2021;3:1602-1606) The authors attest they are in compliance with human studies committees and animal welfare regulations of the authors' institutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information, visit the Author Center. ment of another hospital because of vomiting and abdominal pain. Alizapride and paracetamol, followed by ketorolac, were administered. Vital signs were within limits, and she was afebrile. An abdominal ultrasound scan revealed ileoileal invagination. Because of persisting abdominal pain, a tramadol infusion was started. During the tramadol infusion, the patient had a reaction characterized by generalized skin rash. Moreover, high blood pressure (blood pressure, 140/105 mm Hg) was recorded, and headache developed. A few minutes later, orthopnea and dyspnea occurred in association with widespread pulmonary rales, suggestive of acute pulmonary edema. Heart sound auscultation showed a systolic murmur grade 2/6 on the Levine scale. After 2 hours, the patient's blood pressure dropped to 80/ 50 mm Hg, her peripheral blood oxygen saturation level was 89% without oxygen support, heart rate was 100 beats/min, respiratory rate was 30 breaths/min, body temperature was 36.2 C, and Glasgow Coma Scale score was 15. ## Learning objectives To be able to make a differential diagnosis among common causes of acute heart failure in pediatric patients. To be aware of the potential cardiac complications of "pharmacological cocktails" and hypersensitivity reaction. To consider that KS and TTS may coexist in the same patient. The patient was transferred to our center (Meyer Children's University Hospital, Florence, Italy), which is the regional pediatric referral center. The electrocardiogram (ECG) showed ST-segment elevation in leads I and aVL, as well as ST-segment depression in the inferior leads [fig_ref] FIGURE 1: Electrocardiographic Findings [/fig_ref]. The echocardiogram ## Management In the acute phase, pulmonary edema was treated with a furosemide infusion and continuous positive airway pressure support, with rapid clinical and radiologic improvement. The patient's hemodynamic instability required the administration of levosimendan and epinephrine. After partial recovery, treatment with a low dose of bisoprolol was introduced because of residual systolic dysfunction, and the patient was discharged on therapy with this drug. # Discussion KS is a coronary hypersensitivity disorder characterized by an anaphylactoid, anaphylactic, allergic, or To the best of our knowledge, we report such a potential characteristic for the first time. Although it is rare, KS can occur in pediatric patients, and it should be suspected, for example, in patients with chest pain developing in the context of a potential reaction, associated with ST-segment abnormalities and transient wall motion dysfunction (4). A possible role of cardiac magnetic resonance in KS has been suggested, to identify an earlyphase subendocardial contrast defect and edema without late gadolinium enhancement as typical features (5). An increase of tryptase and the demonstration of a specific allergic sensitization against the potential trigger of the reaction could support the diagnosis of KS. Unfortunately, serum tryptase was not analyzed in our patient, and it was not possible to identify a specific sensitization through skin testing or sIgE. However, these negative results cannot rule out KS [bib_ref] Kounis syndrome: a review article on epidemiology, diagnostic findings, management and complications..., Abdelghany [/bib_ref]. Indeed, the patient was advised to strictly avoid drugs potentially associated with the event in the future. Conversely, TTS is a transient form of LV dysfunction that mimics an acute coronary syndrome, usually triggered by an emotional or physical stressor [bib_ref] International expert consensus document on takotsubo syndrome (part I): clinical characteristics, diagnostic..., Ghadri [/bib_ref]. Most commonly described in postmenopausal women, TTS has also been reported in patients younger than 18 years of age [bib_ref] Takotsubo syndrome in the paediatric population: a case report and a systematic..., Urbinati [/bib_ref]. Compared with adults, who usually present chest pain, pediatric patients more commonly have heart failure, usually with more severe LV impairment and nonapical variants. Cardiac magnetic resonance is required for the diagnosis, especially in patients with nonapical TTS; a typical finding in TTS is myocardial edema without late gadolinium enhancement [bib_ref] International expert consensus document on takotsubo syndrome (part II): diagnostic workup, outcome,..., Ghadri [/bib_ref]. [fig] FIGURE 1: Electrocardiographic Findings (A) Electrocardiogram on admission showing ST-segment elevation in leads I and AVL as well as ST-segment depression in the inferior leads. (B) Electrocardiogram recorded in the intensive care unit showing sinus tachycardia, Q waves in leads I and aVL, poor R-wave progression in leads V 1 to V 4 and ST-segment elevation with inverted T waves in the same leads. after the admission, the patient's clinical condition significantly improved. She remained afebrile, without signs or symptoms, and hemodynamically stable. Her blood hs-cTnT and NT-proBNP values progressively decreased (Figure 3), as did her procalcitonin, C-reactive protein, and lactate values. Chest radiography showed resolution of the lung congestion. The electrocardiographic abnormalities resolved. On day 11, the patient underwent cardiac magnetic resonance with and without a gadolinium infusion; the imaging showed findings suggestive of edema with the absence of late gadolinium enhancement (Figures 4A to 4D). PAST MEDICAL HISTORY In the past medical history of this patient with Peutz-Jeghers syndrome, there were several intestinal polypectomies, as well as a recent hospitalization for ileoileal invagination that was treated conservatively. She had rhinoconjunctivitis from 10 years of age, with sensitization to inhalant allergens (dust mite, animal dander, and pollens). However, the girl did not have cardiovascular disease in her past clinical history or acute illnesses in the previous month. Furthermore, the familiar history was negative for cardiovascular disease or sudden death, although it was positive for atopic diseases (mother with rhinoconjunctivitis and asthma, with sensitization to inhalant allergen). DIFFERENTIAL DIAGNOSIS Takotsubo syndrome (TTS), myocarditis, serotonin syndrome, and Kounis syndrome (KS) were considered. The ECG abnormalities, the transient LV dysfunction with diffuse middle and basal hypokinesia, the BNP elevation, and a potential physical trigger (invagination) first suggested the diagnosis of TTS. However, headache and hypertension within a likely time frame with potentially associated drugs (tramadol and alizapride) suggested possible serotonin syndrome, but that was ruled out on the basis of the diagnostic criteria (1). The results of cardiac magnetic resonance with and without a gadolinium infusion let us exclude myocarditis. This decision was also supported by the clinical course of the patient. Conversely, the close temporal association between the drug treatment and the clinical manifestations (probable association according to the Naranjo scale), including a skin rash and the onset of transient heart failure with diffuse middle and basal hypokinesia, raised our suspicion of KS. Moreover, electrocardiographic and echocardiographic findings additionally supported this hypothesis, and tramadol has already been described as a potential trigger for KS (2). INVESTIGATIONS The patient underwent skin prick testing with tramadol (50 mg/mL), with a negative result. According to current standards, histamine, 10 mg/mL (Lofarma), and normal saline were used as positive and negative control substances, respectively. Furthermore, research into serum-specific immunoglobulin E (sIgE) against tramadol was performed by using radioimmunoassay with the solid phase obtained by coupling the tramadol to epoxy-activated sepharose (Sepharose 6B, Cytiva Life Sciences), and the result was negative (radioactive uptake, 0.69%; nonspecific binding, 0.53%; values >3 times the nonspecific binding are considered positive). [/fig] [fig] FIGURE 2: Echocardiographic associated with an acute coronary syndrome (3). Possible causes of KS include several substances, such as tramadol in our patient, and environmental exposures. Mast cells are well represented in the cardiac tissues, by locating preferentially inside the coronary arteries, where they can degranulate. Thus, we suppose that the echocardiographic finding of transient LAD artery dilatation [/fig] [fig] FIGURE 3: High-Sensitivity Troponin T and N-Terminal Pro-B-Type Natriuretic Peptide HS ¼ high sensitivity; Nt-proBNP ¼ N-terminal pro-B-type natriuretic peptide. [/fig] [fig] FIGURE 4: Cardiac Magnetic Resonance Findings (A) T 2 -weighted sequence showing a hyperintense signal in the interventricular septum and the posterior and anterior wall of the left ventricle, findings suggestive of edema. (B) T 1 mapping sequence suggestive of diffuse edema. (C and D) Absence of late gadolinium enhancement. the KS hypothesis in our case. [/fig]

Challenges amid COVID-19 times - Review of the changing practices in a clinical chemistry laboratory from a developing country pandemic is the defining global health crisis of our time. Compared with its neighbors China and Iran, which were rated as epi-centers of the outbreak, Pakistan has lower standards of health care, unstable economy and dearth of financial resources to tackle the outbreak. Like other institutes and industries in the country, clinical laboratories were succumbed to a variety of challenges. This article is based on the experience and adapted workflow measures from the Clinical Chemistry laboratory at the Aga Khan University Hospital (AKUH), Karachi, which serves as a national referral center with its widespread network of satellite laboratories and phlebotomy centers across the country. It highlights the challenges faced and the appropriate responses to ensure the provision of diagnostic facilities during the COVID-19 outbreak. Furthermore, the lessons acquired and necessary preparations for the post crisis situation are also incorporated. # Background The marks of the Coronavirus Disease 2019 (COVID-19) outbreak trace back to 31 st December 2019, when a pneumonia of unknown etiology was informed to the world health organization (WHO) in Wuhan, China. Since then the outbreak termed as a 'pandemic', declared by WHO on 11 th March 2020, has continued to cause severe morbidity and mortality in most countries globally with devastating implications. Pakistan reported its first confirmed case in the metropolis of Karachi on 26 th February 2020, in an Iran returned traveler. A point of concern for the country, was frequent exchange of travelers; with Iran and China, main reason being religious pilgrimages and business activities respectively. Consequently, by mid-March the pandemic, initially seen only in travelers coming from abroad was confirmed to have spread throughout the country with cases of local transmission and the first death on 20 th March 2020. Compared with its bordering neighbors China and Iran, which were rated as the initial epi-centers of the outbreak, Pakistan has lower standards of health care, unstable economy and dearth of resources to tackle the outbreak. Most of the cities including the major economical hubs i.e. Karachi, Lahore and Islamabad were put on a lock down in fourth week of March to prevent the local spread with negative impact on the financial outcomes and revenue generation of most entities. Health care facilities, especially tertiary care setup like the Aga Khan University Hospital (AKUH), Karachi, Pakistan, catering to COVID-19 cases, were the ones which sustained more serious blows. Routine patient inflow was significantly reduced owing to the imposed lock down and psychological insecurities of visiting a facility testing and housing COVID-19 patients. Laboratories have a critical role to play during the pandemic right form diagnosis till surveillance, with the laboratory professionals as the front-line warriors. However, unlike most laboratories in the developed countries with the essential infrastructure and financial resources, the current situation poses enormous challenges for the laboratories based in the developing world. The Clinical Chemistry laboratory at the AKUH, serves as a referral center, and caters to on an average 16000 routine and specialized tests requests per day from all the provinces owing to its network of more than 250 phlebotomy stations spread across Pakistan. The laboratory was the first to be accredited by Joint Commission International Accreditation (JCIA) and the only with College of American Pathologist (CAP) accreditation in Pakistan. From an organizational perspective, laboratory is considered a critical service in a tertiary care hospital and is expected to continue 24/7 operations even during an infectious pandemic. This is also vital for the individual patient, who expects the provision of diagnostic services not to be disrupted. To ensure such continuity of services, the aim of this review is to highlight the challenges faced by a Clinical https://doi.org/10.1016/j.amsu.2020.06.004 Received 6 May 2020; Received in revised form 1 June 2020; Accepted 1 June 2020 Chemistry laboratory in a developing country amid the COVID-19 crisis, the challenges, implementing changes and the lessons learnt. 1.1. Day to day challenges faced ## Declining test volumes The financial ramifications of the COVID crisis were significant. Owing to the lock down across the country and social distancing campaigns, during the initial phase of the outbreak the testing volumes decreased significantly by approximately 52% of the baseline compared to pre-COVID months. This led to substantial financial decline and the revenue generation targets suffered enormously. The major reasons identified were a smaller number of subjects visiting outreach phlebotomy centers across the country, cessation of outpatient clinics and postponement of all non-emergency surgeries causing limited flow of patients to the tertiary care hospital, with which the laboratory is associated. To cope up with the situation, free of cost home sample collection was started with the aid of outreach phlebotomy centers. The hospital initiated tele clinics for outpatients and door step sample collection for required diagnostic work up was propagated. Following these measures and due to the ease in lockdown, a steady increment in volume of approximately 5% was noted at the time of publication. ## Work from home movement and remote meetings The Clinical Chemistry faculties were divided into teams (e.g. Teams 1 and 2). This is so that, if one team needs to be quarantined, the other team could provide continuity of service. Each team lead by a team leader consisted of three specialist pathologists and one post graduate trainee. The second team was made to work from home via Microsoft Teams and was connected with the team present in the section via sectional WhatsApp group. Within the workplace, the teams had dedicated workspaces and a distance of at least 2 m was practiced. All the essential meetings were converted to virtual meetings via Microsoft Teams and ZOOM to ensure continuity of services and plans as laid down at the beginning of the year by the management. In the beginning few issues pertaining to low bandwidth connection at home and lack of working knowledge of online platforms posed challenges but gradually improved. The technical staff working at the bench side were also limited to minimize exposure. Furthermore, the staff was offered to avail their paid leaves during the outbreak. However, this does not lead to any ease for the section as majority of the staff were reluctant to go on annual earned leaves. The major reasons communicated were an apparent fear of losing the job amid the financial crisis and hindrances in execution of vacation plans which were normally considered to be undertaken during the leaves. Moreover, Pakistan, being a developing country has substantially compromised living conditions. Not unusual to find large households with an average of six to seven persons living together. Many considered staying at work as a better and safer alternate than being lock downed at home. However, the staff was counselled to abide by the situation and physical presence of the staff at the laboratory was reduced by approximately 30% compared to normal by aid of modifications in duty rosters. ## Preserving jobs by cutting expenditure, salaries and compensation As a measure to ease fixed costs and stave off job losses salaries of mid-level and senior level staff were revised with no annual increases, and pay cuts ranging from 10 to 30% were implemented. However, salaries of approximately 73% of the total staff including most junior staff and trainees was preserved to save them from excess economic burden. Furthermore, all travel, new projects and hiring were put on hold. Additionally, it was announced that the staff will further have no appraisals as scheduled for the fiscal year. ## Ensuring adequate inventory and supplies Adequate supply of Personal Protective Equipment (PPE) as an essential requirement was ensured with the assistance of the purchase department. Staff was counselled to practice judicious use of PPE and sanitizers. Bench in charges were reinforced to re-visit their inventory requirements owing to the expected delays in shipment due to the air space closure, to maintain demand-supply chain. Despite the reduction in overall test volumes, a few tests e.g. Albumin, Troponin-I (Trop-I) and C -reactive protein (CRP) exhibited more than usual demands as they were being used for COVID-19 cases in supplementation to the molecular diagnosis and later for prognosis. Furthermore, due to rapid influx of inpatients with COVID-19 to the hospital to which the laboratory serves, there was a surge in demand of point of care tests (POCT) including arterial blood gases and glucose. To meet the increasing POCT workload, additional instruments, were procured and installed on an urgent basis. Alongside the POCT operator trainings were conducted on ad hoc basis and validation requirements were also fulfilled in a timely manner by the POCT team. On the other hand, the laboratory also had to cope with the shortage of Trop-I kit, a vital biomarker, due to shipment delay and a few kits were borrowed from other sources until the receipt of the next shipment. Furthermore, the lab had to discontinue provision of another special biochemistry test, Immunoglobulin G4 (IgG4), due to the shortage because of delays in production and shipment of supplies from the manufacturer in United Kingdom, one of the countries hardest hit by the COVID-19 crisis. 1.1.5. Impact on external quality assurance and accreditation requirements due to airspace closure Due to logistic challenges Proficiency testing surveys From CAP were also not able to meet the timelines and the CAP directed the participating labs to perform and document alternate assessment testing by either split sample analysis, clinical correlation studies or direct observation of technique dependent tests as appropriate. The quality assurance group in liaison with the sectional chemical pathologists came up with a plan swiftly and it was implemented to ensure the lab continues to deliver the highest standards of quality. Furthermore, the European Research Network for evaluation and improvement of screening, Diagnosis and treatment of inherited disorders of metabolism (ERNDIM), an external quality assurance scheme for the biochemical genetics laboratory also extended their dates of result submission, amid the airspace closure, which further eased the path for maintaining compliance with this scheme. At this instant, again having the privilege of having a virtual quality control platform in shape of Bio-Rad Unity Real Time, which allows for external peer group analysis in a networkable quality control program proved to be beneficial for the laboratory and was utilized efficiently. Due to limited availability of testing kits and consumables, scheduled method validation/verification activities were also disrupted. The laboratory was preparing for an upcoming accreditation compliance audit by CAP in the later part of 2020, however as most audits, most of which requires physical presence of inspectors were called off by CAP due to the COVID-19 crisis, the status of the upcoming event also became uncertain. ## New tests introduction process halted In order to ensure continuous growth, our lab core group plans introduction of new tests into the system keeping in sight the productivity, demand and regulatory approvals after appropriate protocols. For the current year, the laboratory had completed the required task associated with the new tests it was bound to offer including Anti-Phospholipase A2 Receptor antibodies (PLA2R) and acylcarnitine on its newly acquired tandem mass spectrometry platform. However, due to logistic issues the supply of reagents and consumables was either slowed down or halted. This further negatively impacted the anticipated revenue growth. ## Processes re-defined ## Management of patient influx at the main laboratory As the laboratory caters to both outpatients and inpatients, a noteworthy threat was the local spread of infection from un-screened patients and attendant visiting the facility for diagnostic workup. A special counter was set up with appropriate safety protocol outside the main lab premises for screening using a standard questionnaire. Suspected COVID-19 cases were refrained from entering the laboratory premises and were diverted to the hospital's specially designated area for screening and management of such cases with appropriate protocol. As COVID-19 is transmitted by droplets and close contacts, patients, staffs and visitors to the laboratory had to wear mandatory masks and were provided with hand sanitizers at the counter and movement in common corridors were restricted. ## Internal surveillance measures All healthcare professionals serving in the laboratory were strictly made to wear surgical masks and gloves as per the safety guidelines and safety officers were held responsible to ensure compliance. Staff who developed symptoms were asked to report immediately and were given medical consultation within the hospital staff clinic. This enabled symptomatic staff to be identified promptly. Furthermore, the organization announced that all staff and their dependents, in case, if tested positive, their complete care will be fully covered financially by the hospital. This was re-assuring and taken positively by all the staff. Additionally, refresher training on the use of PPE was also conducted. ## Updates to the existing policies Following good laboratory practices, all samples received at the section of clinical chemistry were regarded as infectious whether or not coming from suspected or confirmed cases and hazardous risk was minimized by reducing aerosolization, spill prevention and reducing unnecessary sample handling and movement. A fume hood, on war footing, was added to the processing bench to avoid aerosol and droplet spread specially for fecal samples. For provocative tests e.g. sweat chloride test and hydrogen breath test the technical staff was required to wear N95 mask, face shields, gowns (with sleeves) and gloves. Following sample collection, the procedure room was disinfected with hypochlorite and left vacant for at least 60 min for air changes before the next procedure. ## Housekeeping activities re-scheduled Housekeeping cleaning schedules were re-defined and high exposure areas like the toilets, lifts, desks, tables were cleaned and wiped down several times a day with 5% Hypochlorite and alcohol-based sanitizers as appropriate. The premises were mopped three times a day with disinfectant. Discarded PPE and other clinical waste were collected in biohazard bags and disposed in a proper manner. The sectional safety officer was assigned the task of compliance monitoring. ## Continuing medical education (cme) via virtual platforms An integral component of our daily work-based learning is dissemination of new knowledge via one to one lecture by faculty and senior managers for technologists, post graduate trainees and trainee technologists, case-based discussion, instrumentation module for resident and journal clubs. To foster teaching and learning environment and continue the routine teaching-learning activities the Section transitioned from face to face teaching sessions to online teaching activities. Virtual Learning Environments were explored by the faculty and all live lectures and Journal Clubs were being conducted via Microsoft Teams for synchronous teaching with the provision of recording for those who miss it. After getting familiar with Microsoft Teams. The Section has been conducting all administrative and educational meetings on Microsoft Teams or even ZOOM since the crisis started. The faculty was formally trained to operate these software's by the information technology department and further utilized as trainers for residents and other staff. Feedback from the residents and technologists on the use of virtual podiums was positive as they can efficiently participate from their desk via their cell phones alongside the performance of other tasks assigned rather than scheduled events requiring their physical presence. ## Staff mental wellbeing measures Most of the staff were demoralized by the pay cut and abandonment of appraisals, however they were encouraged by the leadership that the positive aspect is their jobs are maintained with provision of free health care amid the times of the crisis. Cookies with a thank you note were distributed to the staff to create recognition. Furthermore, the management was reinforced to ensure easily accessible channels for staff feedback and concerns. This COVID pandemic is generating significant stress and anxiety in our laboratory personnel particularly in those with existing mental health problems, in those with elderly or infants at home and in those with chronic disease. It was essential for all laboratory personnel, seniors and junior staff, to remain connected throughout this crisis. One of the main reasons for stress came from a feeling of uncertainty. As fear and uncertainty continued, the only key was to spread positivity by staying connected. An important stress releaser activity at workplace is social networking with peers especially during intervals and time offs. Initially our dedicated dining room's seating arrangements were not in compliance as adequate distance was not maintained. However, to keep the staff connected and ensure relaxation time temporary arrangements were made to split the seating and staff's interval schedules were redefined to ensure compliance with social distancing. ## Preparations for the post covid-19 scenario Most organization worldwide, including the Pakistani masses were not prepared to face a looming threat of a COVID-19 pandemic. The preparations for such a crisis never existed in the books of laboratories especially for resource constrained setups in developing world where most focus is on cost savings and revenue enhancement. While one cannot predict when the next major pandemic will occur. However, the results of the timely measures highlighted above will be more accurately evident in the times to come, a substantial success was achieved in terms of infection control as of today, and none of our staff was tested positive during the outbreak, despite handling COVID positive samples. A few staff members were quarantined as part of the internal surveillance measures due to clinical suspicion or contact but their tests also turned negative and they resumed work after appropriate interval. A few salient lessons learnt from the crisis, which will enlighten the pathway for adequate workflow of the laboratory in the post COVID-19 world are summarized as follows. - In times of such crisis, follow a proactive approach in anticipating and planning for the impact. Crisis response teams should be part of the regular laboratory groups with appropriate training. If need arises, prior standard operating procedures addressing the challenges the lab faced during the pandemic should be in place for immediate referral. - The management, faculty and key stake holder should engage and coordinate formulating policies to tackle such emergency situations. The sectional leadership should pay special attention in identifying the "bottlenecks" in their respective domains for appropriate alternate plans. - Decrease workload where possible in anticipation of depletion of resources. Built infrastructure for home phlebotomy and sample collection to facilitate the patients in lock downs and prevent unnecessary movement and reduce probability of exposure in the laboratory phlebotomy stations. - Establish staff segregation either spatially or temporally, work from S. Ahmed, et al. Annals of Medicine and Surgery 55 (2020) 300-304 teams to be encouraged in order to have a few back up personnel if the front liners are exposed and need to be isolated. - Staff training on the use of virtual connectivity podiums e.g. Microsoft Teams and ZOOM should be part of the regular training regimen at the entry level. - It may be wise for educational institutes to encourage faculty to spend some of their time teaching online to gain experience and prepare themselves. - As lack of human interaction will be the new norm in the post pandemic scenario, work culture will need modification on the grounds of social distancing. All meetings and consultations are to be conducted online with maximum participation. Social activities at the workplace including celebration parties, leisure breaks are to be halted. - Rescheduling the tea and meal breaks for the staff in shifts, segregating the employs meal area with provision of limited seating and facility of prepacked meals ensured safety precautions are compliant. - Salary cuts, postponed appraisals and the looming threat of losing their jobs created a negative impact on already disturbed mental status for most employs faced with painstaking objectors potentially risking the lives of their families when undertaking duties in a contagious outbreak. The lack of social activities further overburdened the scenario. Managers as effective counsellors should play a role and keep the morale high for effective functioning. Furthermore, the budget allocation should be diverted as such to ensure better incentives and remuneration; and may be fairer to give preference to groups that are more vulnerable. - Post pandemic world with the anticipated overwhelming financial crisis, will require major resource allocation decisions for inventory management and prioritizing needs for new developments and projects. The post COVID lab functioning was evidently different from the normal scenario. Even though, the lab continued to serve 24/7, in all three shifts as prior; but the number of staff was reduced to minimize exposure. A significant time of the daily work place activities of Clinical Chemistry faculty and managerial staff is spent in quality control & quality assurance activities, administrative meetings, internal and external audits, teaching sessions, grand rounds, journal clubs etc. requiring physical interaction and interventions. However, amid the social distancing measures all such activities were transformed to virtual sessions, including a few temporary postponements. Additionally, most of the staff were trained on using virtual platforms using self-directed learning, which is the new norm and essential skill for the post COVID era. Additionally, a literature review of approaches adopted by clinical laboratories worldwide during the pandemic, revealed a report from a United States of America based cytology laboratory, which implemented likewise measures specifically pertaining to social distancing, revision of certain policies, safety measures, re-evaluation of staffing needs, online meetings and virtual academic activities. Taken together, many questions remain for this emerging infection, the take home message derived is that a few of yesterday's practices and policies may be too old or irrelevant in the post COVID-19 laboratory. As social distancing will be the new normal, a potential impact may be seen in an intensification of digital infrastructure of the laboratories, to allow few of the stakeholders to effectively function remotely in compliance with the regulatory and accreditation requirements. New ways of ensuring cost cuttings and reaching a break-even point at minimum has to be formulated, keeping the interests and security of your frontline warriors in sight. # Conclusion With the current state of local spread in the country and amid the high chances of contracting COVID-19, it is inevitable that the clinical laboratories take drastic measures and succumb to acceptable alternate plans for ensuring the safety and interests of its valuable employs alongside continuousness of provision of diagnostic services for better health outcomes, in times of the pandemic. We must be prepared to adapt to the constantly evolving phases and emergent scenarios during the pandemic with a proactive approach. We hope that other clinical chemistry laboratories globally will benefit from our experience in dealing with the COVID-19 crisis. In 2020, we woke up in a different world, with different set of rules; for work environment, financial priorities and social norms, both at home and workplace. The earlier we adapt and embrace the change, the earlier we will be able to create a safe, workable, financial, social, environment for ourselves, our colleagues and families. ## Provenance and peer review Not commissioned, externally peer reviewed. # Ethical approval Not required as it is a narrative review and does not include any human data or intervention. # Sources of funding None. # Author contribution SA performed the majority of the writing work in the first draft. LJ helped in designing the outline and the writing work. HM helped with references search and retrieval and gave suggestions to improve the first draft. AHK and FG did the critically revisions of the manuscript. IS conceived the idea and coordinated the writing of the paper and reviewed the final draft. All authors have reviewed the final draft and agreed upon. ## Consent Not required. ## Registration of research studies Name of the registry: Not requiredUnique Identifying number or registration ID: Not requiredHyperlink to your specific registration (must be publicly accessible and will be checked):

Indomethacin Polymorph δ Revealed To Be Two Plastically Bendable Crystal Forms by 3D Electron Diffraction: Correcting a 47‐Year‐Old Misunderstanding [fig_ref] Table S2: Melting point [/fig_ref] [fig_ref] Table S3: Summary of the crystallographic information used for the structure solution of melt... [/fig_ref] [fig_ref] Table S5: Merging statistics of seven cRED datasets obtained from melt δ-IDM [/fig_ref] [fig_ref] Table S8: Summary of the crystallographic information used for the structure solution of solution... [/fig_ref] [fig_ref] Table S9: Solution δ-IDM [/fig_ref] # Supplementary figures ## References ## Fourier transform infrared spectroscopy (ftir) The FTIR spectra of IDM polymorphs were collected in attenuated total reflectance (ATR) mode using a Spectrum Two spectrometer (Perkin Elmer, Waltham, USA) with a PerkinElmer horizontal ATR accessory. The samples were scanned from 400 to 4000 cm −1 at a resolution of 4 cm −1 . ## Differential scanning calorimetry (dsc) DSC thermograms were recorded using a Netzsch DSC 200-F3 (200 F3 Maia, Selb, Germany). The crystallographically pure samples of melt δ-IDM (θ-IDM) and solution δ-IDM (δ-IDM) (3-5 mg) were accurately weighed into alumina pans and heated at 10 °C/min under nitrogen purging conditions (40 ml/min). Data were analyzed using Proteus analysis software (NETZSCH, Version 5.2.1, Germany). All experiments were triplicated. ## 3d ed sample preparation and data collection Single crystals were first separated from the supercooled melt and gently crushed between two glass slides. Cu TEM grids were then glow discharged for 40s using a PELCO easiGLOW with a current of 20 mA. The glow-discharged Cu TEM grid was gently placed over the crushed sample to pick up the broken fragments and then lightly tapped to remove any larger fragments, leaving only microcrystals on the grid. 3D ED (MicroED) data were collected on a JEOL JEM-2100 LaB6 TEM equipped with a fast Timepix hybrid pixel detector (512 x 512 pixels, Amsterdam Scientific Instruments) operating at 200 kV in selected area electron diffraction (SAED) mode. The software Instamaticwas used for electron diffraction data collection. Diffraction data were collected by continuously rotating the crystal at a rate of 1.13 ° s -1 . The exposure time was 0.3 s, meaning the individual diffraction images were integrated over 0.34° of reciprocal space. Several data sets from different crystals were acquired from starting angles between -60° to + 60°. The spot size and camera length were 3 and 25 cm, respectively. ## 3d ed data processing and structure solution Rotation ED processing software (REDp)was used to determine the unit cell and space group based on the symmetry and reflection conditions of the diffraction patterns [fig_ref] Figure S1: PXRD patterns of IDM polymorphs [/fig_ref]. Data were processed using crystallography software (XDS).Data sets were merged together based on their cross-correlation, in order to improve completeness and I/σ(I) and obtain a single data set suitable for structure solution and refinement (Tables S2, S3, S4 and S5). The structures were solved using the simulated annealing method as implemented in the program Sir2014.The atom connectivity of γ-indomethacin (as reported in the CSD, reference code INDMETwas used to create the rigid-body starting fragment in the form of a mol file. The general conditions can be found in . The structures with the lowest cost functions were refined by least-squares against the ED data using SHELXLthrough the OLEX2 interface.The unit cell parameters were refined against the PXRD data using the Pawley methodwith the TOPAS-Academic v6 program package.Results and Discussion : Crystallographic information of IDM polymorphs. [1-(4-chlorobenzoyl)-5-methoxy-2-methyl-1H-indole-3-acetic acid], C19H16ClNO4, Mr = 357.8 [formula] Polymorph γ α Melt (θ-IDM) Solution (δ-IDM) [/formula] Solved yr. 19722002(1) 18.301 (5) b (Å).969 (1) 25.310 (9) 56.999 (9) 5.123 (1) c (Å) 9.742 (1) 18.152 (7) 12.908 (2) 18.564 (6) : Pair-wise cross-correlation of individual data sets of melt δ-IDM (θ-IDM). ## Data sets number of common reflections Correlation between i, j ## Ratio of common intensities (i,j) b-factor between i,j #i #j ## General conditions Cost function R structure factor Resolution (Å) 2 Random seed 1 ## Number of runs 10 Starting temperature 10 (automatic) ## Number of moves 490 (automatic) Temperature reduction factor 0.9 : Melt δ-IDM (θ-IDM) experimental crystallographic data and refinement data. The unit cell parameters were refined against PXRD data. ## Melt δ-idm (θ-idm) Crystal [bib_ref] Author Contributions, Andrusenko [/bib_ref] (blue). Viewed along the a-axis. The root-meansquare deviation (RMSD) is 0.236 Å (non-hydrogen atoms). The small differences between the two structures can arise from our differing refinement procedures. Solution δ-IDM was refined against 3D ED data while the model from Andrusenko et al. was refined against powder X-ray diffraction data. Both models have reasonable bond lengths and angles. [fig] Figure S1: PXRD patterns of IDM polymorphs. [/fig] [fig] Figure S2: FT-Raman spectra of IDM polymorphs. [/fig] [fig] Figure S3: FTIR spectra of IDM polymorphs. [/fig] [fig] Figure S4: Low magnification TEM images of melt δ-IDM (θ-IDM) (left) and solution δ-IDM (δ-IDM) (right). [/fig] [fig] Figure S5: The reciprocal lattice of melt δ-IDM (θ-IDM). [/fig] [fig] Figure S6: The reciprocal lattice of solution δ-IDM (δ-IDM). [/fig] [fig] Figure S7: Profile fit after refinement of the melt δ-IDM (θ-IDM) crystal structure model against the powder XRD data. [/fig] [fig] Figure S8: Profile fit after refinement of the solution δ-IDM (δ-IDM) crystal structure model against the powder XRD data. [/fig] [fig] Figure S9: Overlay of solution δ-IDM (δ-IDM) with the δ-IDM structure solved by Andrusenko et. al. [/fig] [table] Table S10: Table of torsion angle θ(C4-C7-N1'-C16) (°) of IDM polymorphs. [/table] [table] Table S2: Melting point (Tm) and melting enthalpy (ΔHm) of melt δ-IDM (θ-IDM) and solution δ-IDM (δ-IDM) (n=3). [/table] [table] Table S3: Summary of the crystallographic information used for the structure solution of melt δ-IDM (θ-IDM). [/table] [table] Table S5: Merging statistics of seven cRED datasets obtained from melt δ-IDM (θ-IDM).Table S6: General conditions used for the simulated annealing calculations. [/table] [table] Table S8: Summary of the crystallographic information used for the structure solution of solution δ-IDM (δ-IDM). [/table] [table] Table S9: Solution δ-IDM (δ-IDM) experimental crystallographic data and refinement data. The unit cell parameters were refined against PXRD data. [/table]

Taking Advantage of the Morpheein Behavior of Peroxiredoxin in Bionanotechnology Morpheeins are proteins that reversibly assemble into different oligomers, whose architectures are governed by conformational changes of the subunits. This property could be utilized in bionanotechnology where the building of nanometric and new high-ordered structures is required. By capitalizing on the adaptability of morpheeins to create patterned structures and exploiting their inborn affinity toward inorganic and living matter, "bottom-up" creation of nanostructures could be achieved using a single protein building block, which may be useful as such or as scaffolds for more complex materials. Peroxiredoxins represent the paradigm of a morpheein that can be applied to bionanotechnology. This review describes the structural and functional transitions that peroxiredoxins undergo to form high-order oligomers, e.g., rings, tubes, particles, and catenanes, and reports on the chemical and genetic engineering approaches to employ them in the generation of responsive nanostructures and nanodevices. The usefulness of the morpheeins' behavior is emphasized, supporting their use in future applications. # Introduction Advances in biochemistry have shed light on unusual proteins that defy the dogma "one sequence, one structure, one function", i.e., chameleonic proteins, intrinsically disordered proteins, metamorphic proteins, and morpheeins.These biological polymers can fold into multiple secondary and tertiary structures and as morpheeins self-assemble into several quaternary complexes starting from a single protomer.Due to different conformational states of the protomer, often related to the motions associated with the protein primary functions, morpheeins naturally undergo self-assembly into different homo-oligomers, which may have distinct activities directly correlated to the exposure of hidden surfaces and associated with recognition of new ligands. Thus, morpheein behavior can be seen a stratagem adopted by nature to generate new protein functions or tune pre-existing ones by changing the oligomeric state.Furthermore, the equilibrium between morpheein oligomers can be shifted by changes in temperature, pH, ionic strength, and redox potential as well as single point mutations, 2 encouraging the use of the resulting supramolecular architectures as templates and scaffolds in bionanotechnology. Bionanotechnology relies on the application of knowledge of the structural and functional properties of biomolecules obtained by in vivo and/or in vitro studies to create new architectures at the nanoscale showing chemical and/or physical properties useful for practical purposes: the possibility of building different supramolecular structures relies on the natural occurrence of biological building blocks characterized by different shapes. Among these, proteins are the most adaptable due to their unique collection of chemical functionalities and their variability in shape with respect to other, simpler biomolecules such as RNA, DNA, and peptides. Proteins can establish a plethora of specific covalent and noncovalent interactions exploiting their intrinsic and unique chemical versatility with affinity toward nanomaterials including zero-dimensional (0D) inorganicand organic nanoparticles (NPs),two-dimensional (2D) lattices such as graphene, graphene oxide (GO) and reduced GO (rGO),as well as one-dimensional (1D) carbon nanotubes.Moreover, gold (Au) and silver (Ag) nanomaterials are very suitable for conjugation with proteins due to their affinity for binding to the −SH and −S−CH 3 groups of cysteine and methionine.In addition, such chemical versatility can be improved by genetic engineering or chemical cross-linking, where ad hoc insertion of functionalities offers endless possibilities of protein engineering. This inherent capability is further refined upon folding of the protein polymer into the tertiary and quaternary structure: during folding some amino acidic functional groups are exposed at the protein surface, creating distinct patches of chemical functionalities, and the subsequent oligomerization process patterns them in a regular manner onto the resulting 3D tertiary and quaternary assembly. The oligomers usually possess higher structural stability and enhanced biological functions with respect to single subunits as well as increased binding affinity for enzyme substrates and receptor ligands.Further, proteins are prone to anisotropic hierarchical self-assembly under mild conditions, forming regular 1−100 nm large supramolecular homo-and heterooligomers,including filaments,rings,tubes,cages, 17−20 catenanes,and knots.In some cases, protein superstructures have been used as scaffolds to direct the polymerization of more durable materials, or to synthesize mixed polymers endowed with specific properties, e.g., electrical conductivity.The success of such protein assemblies starts to impact even at commercial level, for instance, next-generation sequencing of nucleic acids using protein nanopore-based devices by the Oxford Nanopore Technologies.However, even though many natural proteins have evolved on their own to self-associate, in general the lack of control over the assembly and the need of structural stability in vitro limits their applications. To overcome this, different approaches, including the matching rotational symmetry method, interface design, and directed evolution, have been explored to create several oligomeric assemblies; these approaches are engineeringintensive for protein surface and dependent on the precision of the design.However, using a single building block to create different high-order oligomeric protein assemblies remains a challenge.Looking for such "ready-to-use" biological "tools" for bionanotechnology, morpheeins are worth considering, as they can avoid the afore-mentioned problems if the structural shift of the building block that gives rise to different nanometric architectures can be controlled. Members of the typical 2-Cys peroxiredoxin subfamily (Prx) are naturally able to assemble in various high-order homo-oligomers such as rings, tubes of stacked rings, cage-like particles, and even catenanes 27−30. Notably, these oligomers can be easily accessed in vitro under nonphysiological conditions making Prxs unique within the building blocks trialed so far for practical purposes. The use of Prx in bionanotechnology has expanded, and this morpheein has emerged as an adaptable platform where "bottom-up" tailored building of responsive nanostructures and nanodevices can be easily achieved. This review collects and discusses the best characterized examples of bioinspired "bottom-up" strategies for the construction of responsive nanostructures and nanodevices using the Prx morpheeins. The data are presented according to (i) the biochemical and structural features which underlie the morpheein behavior of Prxs, (ii) the chemical and/or genetic engineering expedients applied to stabilize their patterned oligomers, and (iii) the strategies exploited to create nanostructures and nanodevices using the oligomers as templates and scaffolds. ## Typical 2-cys peroxiredoxins: structure−function relationship of a morpheein Prxs are enzymes with cysteine-dependent peroxidase activity widely found in prokaryotic and eukaryotic organisms.Their high abundance in cells suggests a remarkable role, a condition common to other ring-shaped proteins formed in vivo.The family of Prxs encompasses six subfamilies grouped according to structural and peptide sequences at the active site containing the reactive cysteine, namely, AhpC/Prx1, Prx6, BCP/PrxQ, Tpx, Prx5, and AhpE, as reported in the peroxiredoxin classification database PREX.The AhpC/Prx1 subfamily is the largest and includes the so-called typical 2-Cys Prxs, a class of proteins with a pair of reactive cysteines forming intersubunit disulfide bonds during the peroxidatic cycle (see subsection 2.1). Some members of this subfamily show the (a) Secondary, tertiary, and quaternary structure of SmPrxI. The β-strands and α-helices are in yellow and red cartoons, respectively. The catalytic Cys48 P and Cys169 R are shown as solvent-accessible surfaces. The elements involved in the monomer−monomer B-interface (β7-strand/antiparallel β7-strand, N-terminal region/β-hairpin/α5-helix, β7-α6 loop/α6-helix, α2helix/C-terminal region), dimer−dimer A-interface (α4-helix/β3-α2 loop at the external side of the ring, β5-α5 loop/β-hairpin on the internal side), and ring−ring R-interface (α2-and α6-helices of the upper ring/α2-and α6-helices of the lower one, β2-strands of the upper ring/β2-strands of the lower one) are indicated by arrowheads. (b) Step 1: under physiological conditions, the LMW ring peroxidase scavenges peroxides (ROOH) into reduced species (ROH) undergoing oxidation at the catalytic Cys P residues (Cys P −SH) into cysteine sulfenic acids (Cys P −SOH). Step 2: local unfolding at the active sites and at the C-terminal arms allow Cys P −SOH to move toward the resolving Cys R (Cys R −SH) forming intersubunit disulfide bridges (−S−S−) which break the ring into oxidized homodimers. Step 3: Trx reduces the disulfide bridges to enable the homodimers to reassemble into a fully folded active ring. Step 4: under oxidative stress conditions, the oxidized Cys P −SOH residues undergo further oxidation to form cysteine sulfenic acids (Cys P −SO 2 H) triggering structural changes that induce the rings to stack into HMW complexes with holdase and intracellular signaling activity. Step 5: Srx reduces the overoxidized HMW species at the catalytic sites to form Cys P −SOH, which, in turn, can undergo further reduction into LMW peroxidase. moonlighting behavior as they can act as a peroxidase enzyme to detoxify cells from hydrogen or alkali peroxides as well act as an ATP-independent molecular chaperones (holdases) able to prevent precipitation of unfolded proteins.This shift in function is accompanied by a change in shape from homodimers and single rings to high-order oligomers, such as tubes of stacked rings and cage-like particles, due to conformational changes as described at atomic details for the Prx (isoform I) from Schistosoma mansoni (SmPrxI)and two mitochondrial Prx (isoforms III) from Bos taurus (BtPrxIII, also called SP-22)and Homo sapiens (hPrxIII). However, other typical 2-Cys Prxs can form particles and tubes, such as human PrxI and PrxII,Thiol Specific Antioxidant I (TsaI) from Saccharomyces cerevisiae,and Prx from Pseudomonas aeruginosa, 45 but in these cases, the 3D structures of their high-order oligomers and the molecular mechanism underlying the structural shift have not yet been clarified (see. In the next two subsections, the structural−functional switch of SmPrxI will be considered due to its unique morpheein and moonlighting behavior representing the first one investigated by high-resolution X-ray crystallographyand exploited in bionanotechnology.Notably, at least for SmPrxI, the moonlighting behavior strictly depends on the oligomerization state.This gain of function is due to the exposure of hydrophobic surfaces, likely at the rim of the ring, coupled to an increased disorder of its polypeptide making SmPrxI more "sticky" than its orthologuesand thus more prone to potentially interact with different nanomaterials, as similarly observed for other ring-shaped molecular chaperones.Differences in the oligomerization mechanism of human and bovine mitochondrial PrxIII proteins with respect to SmPrxI will be highlighted and discussed. 2.1. Low Molecular Weight (LMW) Homo-Oligomers of Typical 2-Cys Peroxiredoxin: Dimers, Rings, and Peroxidase Activity. Typical 2-Cys Prxs can form supramolecular oligomers made by identical subunits with molecular mass ∼22 kDa and possessing a thioredoxin (Trx)-like fold composed of a seven-stranded β-sheet (β1−β7) wrapped between six α-helices (α1−α6). Within each subunit, the first turn of the α2-helix hosts the so-called peroxidatic Cys P , while the flexible C-terminus contains the resolving Cys R involved in the formation of the intersubunit (typical 2-Cys Prxs) or intrasubunit (atypical 2-Cys Prxs) disulfide bonds during the catalytic mechanism (Figure 2a) and absent in 1-Cys Prxs.Typical 2-Cys Prxs exist as obligate homodimers (∼44 kDa) where the two subunits interact via an isologous interface, called the B-interface. One subunit is rotated 180°with respect to the other one, thus conferring to the homodimer an internal twofold symmetry and placing the catalytic cysteines close to each other, that is, Cys P of one subunit facing Cys R from the other subunit. Therefore, the resulting homodimer contains two identical catalytic sites with twofold symmetry. The Binterface relies on the interaction between (i) the β7-strand from one subunit with the corresponding antiparallel β7-strand belonging to the other subunit resulting in a large 14-stranded β-sheet, (ii) the N-terminal region of one subunit with the βhairpin and part of the α5-helix of the other subunit on one side of the homodimer, (iii) the β7-α6 loop of one subunit and the α6-helix of the other subunit, and (iv) the α2-helix from one subunit and the C-terminal region belonging to the other one.Under reducing physiological conditions, typical 2-Cys Prx homodimers self-assemble into a ring-shaped complex (∼220 kDa), that is usually referred to as low molecular weight (LMW) species. For instance, SmPrxI under reducing conditions at pH = 7.4 is a LMW ring complex made by five obligate homodimers (pentamer of homodimers) resulting in a regular decameric complex (PDB: 3ZTL) 32,37 and resembling a pentagon with 52-point group symmetry and superimposable with homologous LMW species.The SmPrxI LMW ring has a thickness of ∼4 nm, a diameter of ∼13 and a pore of ∼6 nm. The seven-stranded β-sheet of each monomer is sandwiched by helices α1 and α5 and one short β-hairpin at the internal face of the ring, and by helices α2, α3, α4, and α6 at the external face. The subunits of each homodimer are correlated by a pseudo twofold axis perpendicular to the plane containing the extended 14-stranded β-sheet that forms an angle of ∼27°w ith respect to the fivefold symmetry axis of the ring complex. The monomer−monomer B-interface within each homodimer and the dimer−dimer interface formed upon assembly into the LMW ring, so-called A-interface, resemble those observed in other typical 2-Cys Prx rings.The dimer−dimer Ainterface responsible for the LMW ring assembly is also symmetrical and relies on contacts between (i) the α4-helix of one homodimer and the β3-α2 loop of the adjacent one on the external side of the ring and (ii) the β5-α5 loop of one homodimer and the short β-hairpin of the adjacent one on the internal side. Similar interactions at the A-and Binterface have also been observed in the LMW species of mitochondrial Prxs, which, however, are dodecameric ring complexes, i.e., hexamers of homodimers, with pore and diameters of ∼7 and ∼15 nm, respectively. Typical 2-Cys Prxs alternate between reduced LMW rings and oxidized homodimers as a consequence of the protein's peroxidase activity and its reduction by cytosolic Trx, 54 the latter being pivotal in parasites such as S. mansoni for fulfillment of redox pathwaysand their survival as observed by drug discovery studies. The active sites of the reduced SmPrxI LMW ring are in the so-called Fully Folded (FF) conformation,i.e., residues 47−50 of each subunit, including Cys48 P , contribute in forming the first turn of the α2-helix, while residues 165−185 of the C-terminal tail, containing Cys169 R , are folded in a hairpin-like structure. In this conformation, the sulfur atom of each Cys48 P is ∼13 Å distant from the sulfur atom of Cys169 R . This implies that during the catalytic cycle, a remarkable structural rearrangement occurs within each homodimer allowing the formation of both the intersubunit disulfide bridges 50 and rings to disassemble into homodimers (see below).Prxs share a similar 2-phase enzymatic mechanism and the same conserved reactive cysteine residue, i.e., Cys P , being first involved in the peroxidase reaction. The peroxidase catalytic cycle of the LMW ring starts with the reduction of the peroxide, e.g., H 2 O 2 , formation of sulfenic acid residues (Cys P −SOH), and the concomitant release of water. The first step of the reaction requires a thiolate nucleophile to break the O−O bond of the peroxide. In each subunit, Cys P is surrounded by three conserved residues: Pro49, Thr45, and Arg124 (numbering is according to SmPrxI). Pro49 limits the solvent accessibility of Cys48 P in its reduced form, while when Cys48 P is oxidized, its presence facilitates the α2-helix unfolding allowing interaction with Cys169 R . Thr45 assists in the binding of H 2 O 2 . Arg124 lowers the pK a of Cys P by stabilizing its deprotonated form using the positively charged guanidine group.In fact, Cys P in typical 2-Cys Prxs has a pK a around 6 instead of 8.3, the value of free Cys in solution. The Bioconjugate Chemistry pubs.acs.org/bc Review second phase of the peroxidase reaction leads to resolution of the cysteine sulfenic acid residue and distinguishes the three Prx classes.In the typical 2-Cys Prx class, including SmPrxI and the mitochondrial PrxIII, the α2-helices containing the oxidized the Cys P −SOH residues undergo a conformational change from an FF to a Locally Unfolded (LU) state moving apart the pocket formed by the Pro, Thr, and Arg and exposing the oxidized sulfur to the solvent. This conformational change is coupled to a movement of the flexible C-terminal arms of the symmetrical subunit of the homodimer placing the resolving cysteine Cys R close to the sulfenic acid for the formation of an intersubunit disulfide bonds. The LU conformation and the unfolding of the C-terminal arm in each subunit result in weakening of the decameric LMW ring complex until a critical state is reached where it breaks down into oxidized inactive homodimers. However, the cytosolic reductant Trx can process the intersubunit disulfide bridges allowing the oxidized homodimers to self-assemble into the reduced peroxidase LMW ring characterized by the FF conformation at the active sitepurple panel).2.2. High Molecular Weight (HMW) Homo-Oligomers of Typical 2-Cys Peroxiredoxin: Tubes of Stacked Rings, Cage-Like Particles, and Molecular Chaperone Activity. The morpheein behavior of typical 2-Cys Prxs allows them to undergo further assembly into high molecular weight (HMW) species where the LMW rings interact to form hollow tubular structures or particles.For most of the Prx morpheeins, the ability to generate HMW species is correlated with the acquisition of holdase activity and exposure of hydrophobic regions on the ring surface,while in the case of human mitochondrial PrxIII, the HMW species seems to be correlated with the protection of its peroxidase activity under stress conditions; i.e., stacked rings are still able to reduce peroxides without gaining the holdase activity. Stacking of rings into HMW tubes occurs upon chemical or physical stimuli such as high concentration of peroxides, pH changes, high temperature, and phosphorylation.For instance, SmPrxI under acidic conditions at pH 4.2 is an HMW complex of two stacked LMW decameric rings as observed by X-ray crystallography at 3.0 Å resolution (PDB: 3ZVJ).The interface between the rings is called R-interface and comprises two regions, namely, (i) the helices α2 and α6 of each subunit belonging to the upper ring in contact with the corresponding helices (α2 and α6) of the lower ring and (ii) the β2-strand of each subunit belonging to the upper ring in contact with the corresponding strand (β2) of the lower one. Upon formation of the HMW complex, all the α2-helices and α6helices are oriented in such a way that the carbonyl groups, hence the helix dipoles, point toward the R-interface. This creates a negatively charged area centered on the Lys164 and His165 (numbering is according to SmPrxI) located at the end of the α6-helices of both the upper and lower rings. Therefore, the positively charged side chains of Lys and His residues (both are expected to be protonated at pH 4.2) stabilize the negative charge of the helix dipoles, thus contributing to the stabilization of the R-interface.Because of the 52-point group symmetry, the rings can stack one on top of the other indefinitely, at least in theory. However, in such a complex each ring resembles a "cogwheel" where the secondary structure elements facing the ring−ring R-interface are "pawls" through which two rings are interlocked due to a rotation of ∼10°around the fivefold symmetry axes of a ring with respect to the next one.It is worth noting that in the crystal structure of the stacked SmPrxI HMW complex the sequence segment 47−49 of each subunit containing Cys48 P is disordered with Cys48 P being solvent exposed; moreover, the C-terminal tail including Cys169 R is unstructured as well.A similar way of stacking has been found in the 3D structures of the hPrxIII HMW complexes at acidic pH obtained by cryoelectron microscopy (Cryo-EM). In this case, the dodecamer interlocking is possible by a rotation around the sixfold axis of one ring with respect to the other of ∼8°. The contact at the β2 seems lost in the hPrxIII HMW assembly, but, again, at acidic pH both the first round of α2-helices and the Cterminus tails are unstructured exposing Cys P to the solvent. However, it seems that hPrxIII can stack also at pH 8.5, without requiring the unfolding of the active site, but the resulting stacks are found in the crystal lattice and it is not clear how stable they are in solution. The protonation of Cys P represents one of the triggers of the morpheein behavior of Prxs, because it disrupts the salt-bridges involving a conserved Arg residue (Arg147 in SmPrxI), known to lower the cysteine's pK a , making Cys P a better nucleophile toward peroxides (see subsection 2.1). The positively charged residue comes from a position distant in sequence with respect to Cys P and is kept in place by an interaction with a residue of the second turn of the α2-helix that can be either a Glu, a Gln, or a His in the subfamily members.The breakage of the saltbridge between Cys P and R124, induced by acidic pH, unwinds the first turn of α2. This event triggers conformational changes at (i) the B-interface leading to C-terminus unfolding of the parent subunit, (ii) the A-interface, where a slight reorientation of the dimers constituting the ring occurs, and (iii) the Rinterface, where the proper alignment of the α2 and α6 due to the resulting dimer orientations allows the proper electrostatic interaction between the same structural elements belonging to another stacking ring. Indeed, the mutation of Cys P to Ser, that mimics cysteine protonation, allows all these Prx to stack constitutively even at physiological pH.So, there is a strong coupling between Cys P protonation, active site unfolding, and HMW formation. Another proof about this remarkable linking between the unfolding of the active site and the HMW formation at acidic pH comes from small molecule binding: by means of in-solution studies and crystal structure analyses, it has been demonstrated that sulfate ions bound at the active site force the afore-mentioned Cys48Ser mutant of SmPrxI mutant to adopt a FF conformation bridging Ser48 and Arg147 therefore stabilizing the LMW decameric ring.Similarly, the effect of sulfate, i.e., the capability to shift from high-order to low-order oligomers, has been observed also in hPrxIII.Another important trigger for the formation of HMW oligomers is the overoxidation of the Cys P side chain to sulfinic (−SO 2 H) or sulfonic (−SO 3 H) acid, fostered by high peroxide concentrations under oxidative stress conditions. In this case, it is not clear if the structural shift is coupled to active site unfolding, as the crystal structure of the human erythrocyte Prx, where Cys P is in the "sulfinic acid" state, is still characterized by folded α2-helices thanks to a salt-bridge with the conserved Arg; 65 however, upon overoxidation several Prxs are able to form both particles and small tubes,which are implicated in both cell signaling and holdase activity in intact cells at the expanse of the peroxidatic onepink panel).However, sulfiredoxin (Srx), a small protein of 14 kDa, specifically reduces the sulfinic form of Cys P restoring the LMW forms and the Prx's peroxidatic activity. ## Peroxiredoxin-based bionanotechnology A survey of the literature indicates that the structural shift of morpheein oligomers has been best characterized in three typical 2-Cys Prxs, i.e., SmPrxI, hPrxIII, and BtPrxIII. Based on published results, this section is intended to provide information about (i) the genetic engineering/chemical expedients used to stabilize the SmPrxI, hPrxIII, and BtPrxIII oligomers in vitro, i.e., rings, tubes, particles, and catenanes and (ii) the strategies utilized to create nanostructures and nanodevices using such oligomers as templates and scaffolds. 3.1. Rings, 2D Arrays, Tubes, Particles, and Catenanes: Controlling the Prx Oligomerization. Protein rings, 0D cages, 1D strings, and tubes as well as 2D crystal sheets have generated a great deal of interest in applications including artificial enzyme mimics, light harvesting antenna, drug and imaging nanocarriers, and nanoreactors.In this view, typical 2-Cys Prx morpheeins show noticeable versatility, as they can naturally self-assemble into rings, tubes, cage-like particles, and catenanes. It is apparent that each Prx assembly described in the following paragraphs depends on specific conformations of the redox-active site lying at the B-interface, which can be induced by particular experimental conditions or site-directed mutagenesis, and on the engineering of the A-, Rinterface and of the N-termini. 3.1.1. Prx Rings. Ring proteins are closed oligomers, in which the rotational symmetry between subunits causes saturation of all the subunit−subunit interaction surfaces leading to structures with defined stoichiometry.Therefore, the sites available for binding nanomaterials are periodically and geometrically arranged with nanometric precision on the surface. Ring proteins exhibit four surfaces available for conjugation and for further derivatization: two surfaces above and below the ring plane, the third one around the pore, and the fourth outside with respect to the pore; 32,72 the bottom and top surfaces are structurally identical as well as the inner and outer ones, as typical 2-Cys Prx rings are made by homodimers with twofold symmetry (see subsection 2.1). These structural properties, the inherent ability to bind nanomaterials and the possibility to engineering the surface (see section 1) make ring proteins ideal tectons in bionanotechnology.Bioconjugate Chemistry pubs.acs.org/bc Review Stable, monodispersed Prxs LMW rings are obtained in vitro under reducing conditions at basic pH, i.e., between 7.2 and 8.0. For instance, hPrxIII observed by transmission electron microscopy (TEM) forms rings at pH 8.0 in the presence of reducing agent tris(2-carboxyethyl)phosphine (TCEP). Projection class averages of the rings show the hexagonal dodecameric assembly (hexamers of homodimers) with diameter of ∼16 nm, pore of ∼7 nm, and thickness of ∼4 nm when observed face-on, i.e., with the sixfold symmetry axis perpendicular to the substrate; alternatively, the rings appear as rods when observed side-on,in agreement with the crystal structure (PDB: 5JCG). Alternatively, Prx rings are obtained by genetic engineering at basic pH but without relying on the redox conditions. For instance, hPrxIII engineered with a N-terminal polyHis tag forms rings resembling the untagged protein at pH 8.0, 29 as confirmed by analytical ultracentrifugation (AUC).Alike hPrxIII, rings can be obtained using the N-terminal polyHistagged SmPrxI at pH 7.6,46 characterized by the pentagonal decameric assembly (pentamer of homodimers; PDB: 3ZTL).The rings appear as spherical globules with height profile of ∼12 or ∼4 nm when imaged by atomic force microscopy (AFM) and, notably, are stable after heating to 75°C , therefore demonstrating remarkable thermostability.An unusual "mini-ring" with square-like octameric assembly (tetramer of homodimers) is obtained using the typical 2-Cys Prx type 1 from Plasmodium yoelii (PyPrxI) when the first seven N-terminal residues are replaced with the polyHis tag, as observed by X-ray crystallography (PDB: 4L0W).A different approach relies on ad hoc amino acid mutations that prevent stacking of the rings into HMW species by destabilizing the ring−ring R-interface without hampering the ring assembly. For instance, the N-terminal polyHis-tagged mutants His164Glu, His164Ala, and Thr163Val of hPrxIII form rings at pH 7.2 in the presence of the metal chelator ethylenediamine tetraacetic acid (EDTA); conversely, the tagged wild-type protein can still stack,in contrast to other results (see above).To explain such unusual behavior, the authors considered the involvement of π−π stacking between the imidazole groups of the polyHis tags from adjoining hPrxIII rings and the possibility that the protein, as an ATPindependent holdase activity, may recognize the 33-aminoacid-long polyHis tag as unfolded peptide thus triggering HMW tube self-assembly, aided by electrostatic interactions between the positively charged tags and the negatively charged α2-helices at the R-interface. EDTA prevents the assembly by sequestering any residual divalent ions present in solution, thus hindering the strong histidine ions coordination bonds between rings which would form HMW complexes (see section 3.1.2). The stabilization of the single rings by pointmutations reflects the importance of the interactions at the Rinterface in Prx stacking; the mutation of His164, belonging to the α6-helix of hPrxIII and involved in the HMW assembly 40 as also observed in SmPrxI,into the negatively charged Glu or uncharged Ala hinders the ionic and polar contacts necessary to form the HMW stacks (see subsection 2.2). Similarly, mutation of Thr163, also belonging to the α6-helix, into the hydrophobic nonpolar Val prevents the stacking likely abolishing the polar contacts through the hydroxyl group.Interestingly, oligomers depending on redox conditions are obtained using the untagged hPrxIII carrying the substitution Ser78Ala. In this case, the mutant forms single rings in the presence of TCEP at pH 8.0, while assembling HMW cage-like particles under nonreducing conditions 29 (see section 3.1.2). Ser78Ala localizes at the A-interface and likely contributes to ring stabilization by increasing hydrophobicity between dimers of hPrxIII. This view is also supported by recent findings on peroxiredoxin from Aeropyrum pernix, where cross-linking with small aromatic compounds at residues constituting the Ainterface, effectively shifts the dimer-ring equilibrium versus the high order oligomer.3.1.2. Prx 2D Arrays, Tubes, Particles, and Catenanes. In general, high-order oligomers self-assemble by two main mechanisms: (i) a noncommutative or hierarchical mechanism, where closed structures characterized by point group symmetry self-assemble and must be formed first, thus creating the new surfaces to interact with other identical closed structures leading to the final oligomer; (ii) a commutative mechanism, where single subunits may give rise to infinite polymerization as the interaction surfaces always exposed and free to interact with incoming building blocks. Prx morpheeins employ both mechanisms by changing shape of the single subunits: for instance, Prx HMW cage-like particles and tubes follow a noncommutative mechanism, i.e., LMW rings must form first before assembling into HMW species, while HMW catenanes are likely formed by a commutative mechanism.Complex nanostructures such as 2D crystalline arrays can be obtained using PrxII from human erythrocyte Prx rings adsorbed on flat, unfunctionalized mica surfaces, in the presence of polyethylene glycol (PEG) and ammonium molybdate. Under these conditions, all the rings adsorb faceon over the substrate thus avoiding side-on adsorption that would disrupt the flat 2D arrangement. 41 Prx can form stable HMW tubes from reversible stacking of rings through pH changes. For instance, the N-terminal polyHis-tagged hPrxIII rings (see section 3.1.1), undergo stacking into long, regular tubes after tag cleavage and decreasing the pH from 8.0 to 4.0.To note, the tubes reversibly and quickly (within tens of seconds) disassemble into rings upon increasing the pH from 4.0 to 8.0 as observed by MS and AUC.Interestingly, the size of the tubes can be scaled by altering the ionic strength with ammonium sulfate as the average length progressively decreases by adding 100, 200, and 400 mM (NH 4 ) 2 SO 4 , 29 likely due to a sulfate-induced refolding of the active site as observed for SmPrxI (see section 2.2).These results reflect the physiological behavior of Prx HMW stacks which selfassemble upon exposure to acidic stress and the importance of the interactions established by key residues such as Glu, Lys, and His (Glu20, Lys22, and His164 in hPrxIII) at the α6helices that stabilize the R-interface between the rings. The polyHis tags if ad hoc engineered can provide an additional site that can be exploited to drive the assembly of tubes. An example is the N-terminal polyHis-tagged SmPrxI ring (see section 3.1.1) that forms hybrid ring-Ni 2+ complexes able to stack into tubes of 20−90 nm length. The tubes appear as cylindrical structures with height profile of ∼12 nm and, to note, undergo reversible disassembly upon addition of imidazole likely due to the competition against the polyHis tags for the binding to the metal.These results account for the high affinity of Ni 2+ ions for the polyHis sequences (K d = 15 nM)located at the ring pore that become driving forces to achieve stacking into tubes, acting as polymerizing agents increasing the local concentration of the rings. In fact, both the top and bottom surfaces of the polyHis-tagged rings are suitable environments to bind divalent metal ions within a binding surface of ∼130 nm 2 .Similarly, the N-terminal polyHis-tagged hPrxIII constitutively forms short stacks at pH 7.2, even in the presence of EDTA.Interestingly, changing the length of the tag affects the tube elongation with the number of rings incorporated within the stacks being 2−10, 2− 66, and 4−100 for tags containing 2, 4, and 6 histidine residues, respectively.This effect can be ascribable to two factors modulated by the elongation of the N-termini: (1) π−π stacking of the histidines' imidazole groups between adjoining rings and/or (2) coordination binding of monovalent and/or divalent metal ions, not efficiently sequestered by EDTA, to the polyHis tags resulting in bridging bonds between rings. Site-directed mutagenesis of residues belonging to the secondary structure elements involved in the formation of the stacks, as Cys P and Cys R , can enable constitutive assembly of tubes without relying on the experimental conditions. Indeed, the mutations Cys48Ser, Cys48Asp, and Cys48Pro of the N-terminal polyHis-tagged SmPrxI and its deletion from residue 166 to 182 at the C-terminal region (ΔC-ter), lead irreversibly to tubes at pH 7.4 whose length depends on the change introduced: the Cys48Asp, Cys48Pro, and ΔC-ter mutants form short stacks of 2 to 6 rings (∼8−24 nm in length), while the Cys48Ser mutant forms tubes composed of 20−30 stacked rings (∼80−120 nm in length).These results relate with the physiological behavior of SmPrxI as the Cys48Ser and Cys48Asp mutants mimic the Cys−SH and Cys−SO 2 H residues of Cys48 P that form upon acid or oxidative stress, respectively, while the Cys48Pro mutant elicits constitutive unfolding of the first turns in the α2-helices favoring the UF conformational state, both required during the formation of HMW stacks; on the other hand, the ΔC-ter mutant removes the steric hindrance caused by the flexible Cterminal tails which hamper the stacking process (see subsection 2.2).A strong structure−function relationship is thus present in Prxs, as observed for other proteins as well.Amino acid mutations also allow the assembly of patterned, nontubular Prx HMW oligomers. For instance, complexes of interlocked rings are obtained using the Cys47Ser−Ser78Ala double mutant of the N-terminal polyHis-tagged hPrxIII at pH 7.2,resembling catenanes formed by the Cys168Ser and Phe190Leu mutants of BtPrxIII.These structures seem to rely again on particular structural rearrangements of the Prx's active site and not on the polyHis tag presence or on experimental conditions. For instance, this HMW state seems to rely on the destabilization of the Cterminus, and thus on conformational changes at the Binterface. The F190L site-directed mutant of BtPrxIII, a residue belonging to the YF motif known to modulate Cys P overoxidation and orientations during the catalytic cycle of typical 2-Cys Prxs, destabilizes the first turns of the α2-helix. The crystal structure of this mutant in the reduced state ## Ikhlsvn-derived ribbons BtPrxIII Nanoribbons 89 Colloidal Hydrogel 89 Self-assembling (SP-22) Nanofiber-like optoelectronic compoundsColloidal 1D conductive nanofilaments in field-effect transistors 90 Self-assembling Semiconductive LMW (do)decameric rings ## Smprxi Ring-trapped gold NPs 46 Colloidal Nanoelectrodes Self-assembling Electrically responsive Ring-shaped gold NPs arraysColloidal SERS nanoprobes GO composites doped with ring-trapped metal NPsColloidal 3D metal-doped graphene nanoscaffolds for gas sensing Self-assembling Low weight Microporous Ring-shaped silver rings on graphene-coated solidstate membranesColloidal Nanopores for single-molecule detection and sequencing Self-assembling Plasmonic Fluorescent Ring-linked gold NPs arraysColloidal Nanometric probes for intracellular Raman imaging Self-assembling Plasmonic SERS Ring-doped GO compositesBiocompatibile Scaffolds for cell differentiation and growth 93 Pro-differentiating hPrxIII Ring-trapped iron NPs arraysColloidal Nanoelectrodes Electrically responsive Ring-tethered gold layersPreferential ring orientation Biosensors Ring-tethered gold layers coated with SAMPreferential ring orientation Biosensors HMW tubes of stacked rings hPrxIII Tube-trapped iron NPs arraysColloidal 1D conductive nanofilaments Self-assembling Electrically responsive Tube-tethered gold layersPreferential tube orientation Biosensors Tube-tethered gold layers coated with SAMPreferential tube orientation Biosensors SmPrxI Tube-trapped gold NPs arraysColloidal 1D conductive nanofilaments Self-assembling Electrically responsive Tube-containing coatingsBiocompatibile Scaffolds for cell differentiation and growth 95 Pro-differentiating BtPrxIII Tube-containing coatingsBiocompatibile Scaffolds for cell differentiation and growth 95 (SP-22) Pro-differentiating a It includes all known (referenced) and putative potential applications (unreferenced). Bioconjugate Chemistry pubs.acs.org/bc Review highlights the presence of an unusual position of the conserved Arg residue, now far from Cys P , pointing with its guanidinium group toward the conserved glutamate residue at the second turn of the α2-helix 69 (see subsection 2.2). Furthermore, the mutation Ser78Ala, located at the dimer−dimer A-interface and far from the R-interface, in the untagged hPrxIII produces HMW cage-like particles of linked rings at pH 8.0.3.2. Prx-Based Nanofabrication and Biomaterials. Broadly speaking, when used to control the assembly or synthesis of nanostructures, biomolecules are referred to as "soft templates". Nanofabrication through the "bottom-up" strategy based on soft templates is efficient to create nanostructures with various morphologies.However, obtaining protein templates with defined, patterned shapes is extremely difficult using computational methods as these macromolecular architectures require complex and elaborate protein−protein interactions and specific control of protein orientation.This section illustrates how the Prx morpheeins, especially rings and tubes, and Prx-derived peptides can be used to create nanostructures and nanodevices with specific chemical, physical, and biological properties useful for a vast array of applications. 3.2.1. 3D Hydrogels and Bio-Organic Field-Effect Transistors. The B-interface in typical 2-Cys Prx homodimers relies on contacts between the seven-stranded β-sheet from one monomer and the same secondary structure of the other monomer, resulting in an extended 14-stranded β-sheet (see subsection 2.1). It is observed that this structure is found in about 15% of native homodimer interfaces of different Prx families and frequently includes the amino acidic sequence Ile-Lys-His-Leu-Ser-Val-Asn (IKHLSVN).The IKHLSVN sequence derived from BtPrxIII can be synthesized as a water-soluble, stand-alone peptide with both N-acetylated and C-amidated regions to maintain the net charge of the corresponding interface. This peptide forms 3D birefringent hydrogels at concentrations between 20 and 200 mg mL −1 in pure water exhibiting liquid crystalline texture and resulting from spontaneous self-assembly of elongated nanoribbons.The IKHLSVN peptide can be employed to obtain perylene imide-based organic semiconductor assemblies for transistor devices. Namely, the peptide derivatives glycine-IRHLSVNglutamate-glutamate-glutamate (GIRHLSVNEEE) and Nacetylated EEEIRHLSVN-ethylamine (Ac-EEEIRHLSVN-ethylamine) can be conjugated with perylene diimide (PDI) or perylene imide bisesters (PIBEs) to obtain hybrid bio-organic compounds with luminescence efficiency and optoelectronic properties. While the glycine and ethylamine groups act as lowsteric linkers between the peptides and the organic compounds, the three glutamate ionizable residues improve the solubility of the final hybrids and provide a means to invoke pH-triggered self-assembly. The compounds form diffusive, interlaced networks of nanofibers with roughly 200−300 nm up to micrometer length, likely due to the synergistic effect of the IRHLSVN self-assembly behavior and the π−π stacking between PDI or PIBEs. By taking advantage of the nanofibers as semiconductor layers between Ag electrodes, bio-organic field-effect transistors (biOFET) can be built showing output and transfer characteristics upon exposure to voltages.3.2.2. Colloidal 0D NPs. The inborn affinity of amino acids toward metalsand the possibility to insert polyHis sequences to further expand this affinity allow Prx to bind metals and build hybrid Prx-metal nanostructures.For instance, imidazole-mediated binding of Ni 2+ -coated 1.8 nm AuNPs to N-terminal polyHis-tagged SmPrxI rings can be achieved at pH 7.6. The nanoparticles are partially trapped within the ring scaffolds thus giving to the hybrid SmPrxI-AuNPs complexes electrical responsiveness if exposed to a 6 V bias voltage, as demonstrated by repulsive-mode electrostatic force microscopy (EFM) performed in air on samples adsorbed on silicon dioxide substrates.Furthermore, the binding affinity for gold of the ring without dependence on the polyHis tag is observed using bare 2 nm AuNPs. In this case, the nanoparticles appear adsorbed all over the ring surface yielding arrays with circular arrangement and size of ∼21.2 nm,thus fitting with size scale of the crystal structure (PDB: 3ZTL).On the other hand, templated synthesis of nanoparticles can be obtained taking advantage of the polyHis tags as nucleation sites. For instance, binding of Fe 2+ ions to Nterminal polyHis tagged hPrxIII rings is achieved by addition of Fe 2+ at pH 8.0 and 4°C in the presence of citrate, which stabilizes the ions and prevents oxidation by molecular oxygen. Bioconjugate Chemistry pubs.acs.org/bc Review Under these conditions, the Fe 2+ ions are captured in the ring pore to allow seeding growth of electron-dense ∼4 nm iron oxyhydroxide NPs (FeNPs), which confer electrical responsiveness to the complex adsorbed on mica substrates upon application of a 5 V bias voltage in air.These examples show how to generate stable, responsive metal−protein complexes by tidily regulating the strong metal−ligand coordination bonds and the weak ones occurring in the protein quaternary assembly to avoid precipitation as well as disassembly of the complex. For instance, imidazole used for assembling the SmPrxI-AuNP complexes is efficient in balancing the strong nickel-polyHis interaction (K d = 15 nM)and the weaker molecular contacts holding the SmPrxI LMW ring (K d ≈ 1 μM for the dimer/decamer equilibrium).Moreover, hybrid protein−NP conjugates represent tools for addressing many of the difficulties in biomedicine, e.g., as biocompatible drug delivery systems: these hybrid nanostructures, indeed, improve interactions with biological material such as cells because the protein coating can increase penetration of cell membranes by NPs.3.2.3. NPs-Graphene 3d Composites and 2d Plasmonic Nanopores Arrays. Graphene-based materials can be obtained taking advantage of the inherent affinity of proteins toward the 2D lattice of graphene.As an example, the N-terminal polyHis-tagged SmPrxI rings can be utilized to drive aggregation of 0.3−1 μm GO layers into a microporous 3D composite at pH 7.4, which can be freeze-dried and observed by scanning electron microscopy (SEM). Furthermore, scanning TEM (STEM) and energy dispersive X-ray spectrometry (EDS) demonstrate that the rings can act as scaffolds or templates to dope the composite with Ni 2+ -coated 1.8 AuNPs or ∼3.3 nm palladium NPs (PdNPs) obtained by Bioconjugate Chemistry pubs.acs.org/bc Review chemical reduction of Pd 2+.To note, the rings adsorb face-on over the lattice and perform chemical reduction of GO into rGO at the expanse of the native cysteine residues.This example demonstrates how to master the assembly of complex materials by exploiting the biochemical and structural features of the Prx morpheeins: stacking is achieved taking advantage of the perfect symmetry of SmPrxI and its identical top and bottom surfaces (see section 2.1) as well as the protein-catalyzed reduction into rGO, which is known to yield of 3D graphene-based materials.The graphene-bound Prx rings can be exploited to precisely place metal nanorings on regular, porous 2D arrays in order to assemble devices with improved properties. For instance, Nterminal polyHis-tagged SmPrxI rings can be used as templates to synthesize Ag nanorings with diameter and pore of ∼28 and ∼3 nm, respectively, on graphene layers by chemical reduction of Ag + . The nanorings can be labeled with fluorescent dyes and selectively placed on 2D nanopores arrays on solid-state silicon nitride (Si 3 N 4 ) membranes as observed by confocal fluorescence microscopy (CM), 92 taking advantage of electrophoresis as reported for other 2D lattices.The hybrid nanopores show improved fluorescence lifetimes of 1.0 ± 0.8 ns, corresponding to half of the values obtained using silver-free labeled SmPrx1 rings thus suggesting plasmonic behavior of the device. Notably, drilling of 2 nm holes on the graphene surface at the nanoring pore can be achieved by focused electron beam resulting.This strategy accounts on a synthesis at pH 5.5 in citrate buffer that stabilizes the LMW ring oligomer and acts as adjuvant for Ag + binding by providing a net negative charge to the citrate-coated protein. In addition, at pH 5.5 the polyHis sequences are mostly protonated (pK a of histidine is 6) thus discouraging their binding to Ag + to the advantage of the native surface amino acids of SmPrxI.Self-assembly of graphene-based materials represents a feasible route to tailor and enhance the properties of graphene while making more manageable 3D structures and devices for practical applications, e.g., energy conversion/storage and environmental decontamination.For instance, the device reported herein represents one of the first examples of hybrid plasmonic nanopores integrated on solid-state membranes that would be useful for next-generation sequencing and singlemolecule detection.3.2.4. Gold-Tethered 2D Protein Rings and Colloidal Plasmonic 1D AuNPs Arrays. Gold surfaces are useful platforms to create protein-reactive assemblies that can be exploited to build nanodevices for bioelectronic and biosensing applications.An example of Prxs assemblies on flat Au is the N-terminal polyHis-tagged hPrxIII rings that adsorb at low concentration of 5 μg mL −1 and pH 7.2 over mica substrates coated with 300 nm Au layer. Scanning tunneling microscopy (STM) show that that the rings preferentially bind face-on to the gold appearing as discrete globules with 15 nm diameter and 0.5 nm thickness, thus creating a flat protein coating.To note, the stability of the interaction is increased by treating the gold surface with a self-assembling monolayer (SAM) of SH-nitrilotriacetic acid (NTA-thiol).These data represent the first example of building Prx supramolecular structures with preferential orientation tethered to gold surfaces, which would be particularly useful in the preparation of ordered functional 2D or 3D nanoscale assemblies for different applications, including cell, virus, and bacterial adhesion, as well as biomaterial and biodevice engineering.The affinity of Prxs toward gold can be exploited to assemble colloidal 1D metal NPs arrays labeled with molecular tags, which can find use in applications based on surface enhanced raman scattering (SERS). Namely, 20 nm AuNPs can be labeled with alkyne-based compounds through sulfur−gold interactions, as previously reported for silver nanostructures,and induced to self-assemble upon binding to Nterminal polyHis-tagged SmPrxI rings. The hybrid alkynelabeled SmPrxI-AuNP complexes are arranged as short 1D arrays with the interparticle gaps indicating the position of the SmPrxI rings. Notably, the arrays exhibit Raman scattering enabling detection of the amino acids and alkyne molecules, the latter being distinguishable for their CC bonds in a biomolecule-silent spectral region.Alike the SmPrxI-GO hybrid composites where the patterned ring shape drives stacking of GO (see section 3.2.3), the 1D assembly of AuNPs is likely to occur due to the geometrical protein complex and the interaction with the polyHis-tagged identical bottom and top surfaces; 9,96 moreover, even the native amino acids, including methionine and cysteine, at the SmPrxI ring surface would play a role in the assembly process for being sensitive to changes of pH and ionic strength.By comparison, 1D assemblies of nanoparticles have not been as thoroughly explored as their 2D and 3D counterparts because of the difficulties in their preparation and isolation for analysis. However, 1D assemblies have potential applications in a variety of optoelectronic, electronic, photonic, and magnetic purposes.For instance, these nanostructures are currently exploited for nondestructive intracellular imaging by Raman spectroscopy, where they are used as nanoprobes in theranostics, e.g., controlled NPs delivery and imagingand plasmonic photothermal treatment.3.2.5. Graphene-Based 3D Scaffolds for Cell Growth and Differentiation. As 3D graphene-based composites doped with inorganic nanomaterials are being exploited for electronic and environmental purposes,protein-graphene materials show interesting properties in biological applications, taking advantage of the inherent biocompatibility of GO.For instance, polyHis-tagged SmPrxI rings adsorbed on GO can be used as substate for adhesion, growth, and differentiation of human neuroblastoma cells SH-SY5Y into neuronal-like cells without relying on the pro-differentiating agent N2 as observed by immunofluorescence microscopy (IFM). The effect fits with the concurrent expression of the marker of neuronal NF200 and leads to morphological changes of the neuronal-like phenotype and formation of a cell network over the hybrid SmPrxI-GO substrate as confirmed by an increasing neurite length and neurite/neuron number.Furthermore, the YAP and TAZ transcriptional coregulators are found to move across the cell, from nucleus to cytoplasm, likely acting as molecular triggers for mechano-transduction process of differentiation.Accordingly, down-regulation of proliferating and pluripotency factors such as RhoA and Sox2 is observed in cells cultured over the SmPrxI-rGO composite.The differentiation is attributed to increased stiffness gained by GO 112 likely due to SmPrxI-induced reduction to rGO (see section 3.2.3). Graphene-based scaffolds for tissue engineering are now at the forefront in medicine as attractive materials for their biocompatibility, versatile chemical states, suitable flexibility, and physicochemical properties, the latter necessary for stabilizing the growth and differentiation of cells. In this context, neural cells are currently considered as the main model for tissue engineering to be assessed in regenerative therapies for various diseases and disorders, e.g., spinal cord injuries, strokes, and Alzheimer's as well as Parkinson's disease.However, even though the surface chemistry of graphene is known to play a significant role in influencing the cell behavior and fate, there is little known about molecular mechanisms underlying the differentiation by mechanical stimuli of the intracellular response and, therefore, this topic remains a hot issue to be fully explored. 3.2.6. Colloidal 1D Polarizable NPs Arrays. Nanometric assemblies with 1D architecture or nanotubes are popular structures and probably the best characterized. In this regard, protein tubes of self-assembling Prx rings can be obtained with special features for wide applications. For instance, N-terminal polyHis-tagged SmPrxI rings can bind Ni 2+ -coated 1.8 nm AuNPs, taking advantage of the Ni 2+ -polyHis interaction, slowly undergoing imidazole-mediated self-assembly into SmPrxI-AuNPs hybrid 1D arrays where the nanoparticles locate between adjoining rings, thus acting as gold metal linkers between adjacent protein molecules. The arrays appear as ∼12 nm cylinders and exhibit EFM electrical responsiveness along the whole structure if adsorbed on silicon dioxide and exposed to a 6 V bias voltage, likely due to the presence metal embedded within the tube cavity.Though the polyHis tag-Ni 2+ interaction is likely the main driving force of the assembly process (see section 3.2.2), even the ring−ring Rinterface can play a role in the ring stacking as AuNPs are small enough to partially penetrate the ∼6 nm ring pore; thus, the nanoparticles act as driving forces to increase the local protein concentration thus inducing the conformational changes required to stabilize the electrostatic and polar contacts at the R-interface (see sections 2.1 and 2.2). As an alternative strategy to achieve 1D assemblies, metal NPs can be obtained by direct chemical synthesis on the Prx ring scaffold and triggering of the stacking through pH changes. For instance, the N-terminal polyHis-tagged hPrxIII rings can be used as templates at pH 8.0 to synthesis at the ring pore ∼4 nm FeNPs O 2 -mediated to oxidation of citratestabilized Fe 2+ ions (see section 3.2.2). Taking advantage of the pH-sensitivity of hPrxIII (see section 3.1.2), on decreasing the pH to 6.0, the hPrxIII-FeNPs complexes can be induced to stack, forming long regular 1D arrays that, according to their hybrid composition, exhibit electrical response along the whole structure when applying voltages of 5 V when adsorbed on mica.To note, oxyhydroxide NPs forming the 1D arrays appear embedded within the cavity but also lying at the outer surface of the hPrxIII stacks likely due to partial deprotonation of histidines of the polyHis sequence at pH 6.0, with loss of affinity for the mineralized iron.In this process, the stacking is likely due to the protonation of the catalytic Cys P (Cys47 P in hPrxIII) inducing the conformational changes necessary to stabilize the ring−ring R-interface as reported.Nanotubes are found widely as structures made by carbon, metals, silicon, and DNA as well as peptides and proteins, the latter occurring naturally in cells as self-assembling structures, e.g., microtubulesand the bacterial flagella.Due to their highly desirable properties such as biodegradability, biocompatibility, and ease of tailored surface functionalization, protein nanotubes currently found several applications in food, pharmaceutical, and cosmetics sectors. These reasons make them a promising alternative to carbon nanotubes, which, even though possessing excellent tensile strength, large surface area and appropriate electronic, thermal and chemical properties, still evoke concerns about their toxicity and compatibility with biological matter.Though small nanoparticles do not exhibit coupling phenomena, such as surface plasmonic resonance or scattering to be used for instance as biosensors, their regular arrangement into 1D arrays confer the final nanostructure with other properties useful in applications such as conductive nanowires 117 and nanostructured catalysts for oxygen reduction reaction in fuel cells.3.2.7. Gold-Tethered 2D Protein Tubes and Scaffolds for Cell Growth and Differentiation. Proteins supramolecular structures with preferential orientation on flat 2D gold are promising candidates to build ordered functional 2D or 3D assemblies for applications ranging from cell adhesion to biodevice engineering.In this regard, N-terminal polyHistagged hPrxIII tubes can be assembled from stacking of rings on flat gold surfaces with preferential orientation at pH 7.2. The protein adsorbs as single discrete rings at low concentration (see section 3.2.4) that, however, stack into tubes upon increasing at concentration 50 μg mL −1 in side-on (laterally) or face-on mode.Furthermore, face-on adsorption is favored when coating the gold with SAM of 4mercaptobenzoic acid (MBA) and attaching the protein by chemical cross-linking and, notably, by pretreating the protein at pH 4.0 resulting in long tubes, in agreement with other data (see section 3.1.2).The use of Prx tubes is not limited to the interaction with nonliving matter such as inorganic surfaces or NPs. Examples are the N-terminal polyHis-tagged SmPrxI and BtPrxIII whose mutants Cys48Ser and Cys47Ser are known to constitutively self-assemble into long tubes of stacked ringsto be used as biocompatible substrates for cells. Namely, human neuroblastoma SH-SY5Y can be seeded and differentiated into neuronal-like phenotypes sprouting neurites and forming a cell network without the addition of N2, as shown by phase contrast microscopy (PCM) and confirmed by expression of neural markers β-tubulin III and NF-200. The differentiation is observed on both Prx-based substrates while being almost absent when seeding the cells over the wild-type proteins forming single rings.To note, very similar effects are observed if using neural cancer stem cells (NCSCs) from human glioblastoma which are prone to adhesion of the neurospheres before spreading and differentiation on both the Prx tubes.Though the mechanism resulting is still under investigation, these data suggest that the sole tube-like architecture is mandatory to trigger the differentiation pathway whatever the Prx used, likely supporting these materials as universal tissue-free proteinaceous scaffolds for tissue engineering. Protein-based self-assembling scaffolds are increasingly gaining interest as favorable and efficient scaffolds in tissue engineering applications as they can self-assemble under mild conditions into supramolecular structures mimicking the native extracellular matrix. ## Conclusions and future perspectives The expanding universe of protein-based nanotechnology is currently flowing through the growing field of applied nanoscience, with broad impacts on nearly economic sectors, from electronics to energy, biomedicine, cosmetics, defense, automotive, and agriculture. Indeed, the global nanotechnology market is expected to exceed US $125 billion by 2024,thus advancing technology through increasing funding resources for R&D activities, miniaturization of devices, and strategic alliances between countries. Analysis of expanding market trends show that electronics, energy, and biomedicine account for over 70% of the growth, with nanoparticles, nanolithography, and nanodevices being the most significant. This review aimed at highlighting that the structural and biochemical plasticity of Prxs can be also exploited across different multidisciplinary areas of applied science from electronics to biomedicine. The reported examples clearly demonstrate that control over the assembly/disassembly of Prxs into supramolecular complexes and their interactions with inorganic and organic materials can be easily achieved under mild experimental conditions in vitro. This is augmented using genetic engineering and/or chemical strategies: (1) presence/ absence and length of N-terminal polyHis tags, (2) presence/ absence of divalent metal cations and metal chelators, (3) pH changes, (4) reducing/oxidizing conditions, (5) amino acid mutations, (6) ionic strength, (7) protein concentration, (8) SAM-coated surfaces, and (9) design of Prx-derived peptides. Interestingly, in most cases changing these factors facilitates tuning the morpheein behavior and the interaction with nanomaterials. Evidence is presented for how to harness the oligomeric transitions of Prxs in solution from rings to hollow colloidal and surface-tethered tubesas well as cage-like particles and catenanes.Studies also show methods to improve such complexes making them suitable tectons for building nanostructures such as 2D protein arrays 41 and nanoribbonbased 3D hydrogels.Their derivatives can be coupled to optoelectronic compounds for building transistor devices, 90 0D and 1D arrays of colloidal metal and mineralized NPs with electrical polarizabilityand 1D SERS-active metal NPs clusters.The "sticky" features of the Prx rings have been shown to be useful for assembling metal NP-doped 3D GO materialsand improving the stiffness and biological effects of rGO to induce growth and differentiation of tumor-derived cellsas well as to obtain 2D plasmonic nanopores over graphene-coated solid-state membranes.Finally, bare tubes of stacked Prx rings have been demonstrated to be efficient biological scaffolds for adhesion and differentiation of stem and tumor-derived cells with no need of differentiating supplements.The use of Prx for bionanotechnology purposes is just initiated and more efforts should be made in the future to completely master its morpheein behavior in order to easily access all the oligomeric structures available for Prx. For example, the interfaces involved in the Prx's nanocage structure are not determined at the atomic level, even though some indications on how to stabilize this array already exist; the employment of this supramolecular oligomer for practical purposes is, at present, limited, but still very promising if one considers the many applications of ferritin-like or virus-like nanocages in biomedicine and nanotechnology reported so far. The valuable adaptability of such patterned protein complexes expands even more broadly if looking at other ringshaped protein oligomers that do not belong to the Prx family. Taken together, all these features make protein morpheeins with supramolecular structures suitable for hierarchical nanofabrication of structures and devices, 86,88 a concept that is currently emerging to overcome the limitations and necessities inherent to the "top-down" and "bottom-up" strategies, i.e., the scaling down of the size limit of nanofabrication and improving the control over self-assembling structures to create nanometric objects with collective behavior and coupling phenomena with augmented abilities to interact with light and biological matter.

Only minimal regions of tomato yellow leaf curl virus (TYLCV) are required for replication, expression and movement The IL-60 platform, consisting of a disarmed form of tomato yellow leaf curl virus (TYLCV) and auxiliary components, was previously developed as a nontransgenic universal vector system for gene expression and silencing that can express an entire operon in plants. IL-60 does not allow rolling-circle replication; hence, production of viral single-stranded (ss) DNA progeny is prevented. We used this double-stranded (ds) DNA-restricted platform (uncoupled from the dsDNA?ssDNA replication phase of progeny viral DNA) for functional genomics studies of TYLCV. We report that the noncoding 314-bp intergenic region (IR) is the only viral element required for viral dsDNA replication. None of the viral genes are required, suggesting recruitment of host factors that recognize the IR. We further show that IR-carrying reporter genes are also capable of replication but remain confined to the cells into which they were introduced. Only two sense-oriented viral genes (V1 and V2) need to be added to the IR-carrying construct for expression and movement. Hence, any IR-dsDNA construct supplemented with V1 and V2 becomes a replication-competent, mobile and expressing plant plasmid. All viral functions (replication, expression and movement) are determined by the IR and the sense-oriented genes. The complementary-oriented viral genes have auxiliary roles in the late phase of the virus ''life cycle''. The previously reported involvement of some viral genes in expression and movement is therefore revised.O. Gover and Y. Peretz contributed equally to this paper.Electronic supplementary material The online version of this article ( # Introduction Tomato yellow leaf curl virus (TYLCV) is a monopartite begomovirus with a characterized genome organization consisting of six overlapping transcribed open reading frames (ORFs) [bib_ref] The molecular biology of geminiviruses, Stanley [/bib_ref] [bib_ref] Family geminiviridae, Stanley [/bib_ref] [bib_ref] Geminivirus replication origins have a group-specific organization of iterative elements: a model..., Argüello-Astorga [/bib_ref] [bib_ref] Tomato yellow leaf curl virus-a whitefly-transmitted geminivirus with a single genomic component, Navot [/bib_ref]. Briefly, TYLCV consists of six overlapping open reading frames (ORFs) transcribed in opposite directions from two promoters situated at either end of an intergenic region (IR). The 314-bp IR carries the universal motif TAATATT/AC. In addition to carrying promoters, it also serves as the viral origin of replication. It carries motifs (iterons) for binding the replicase-associated protein (REP-the product of ORF C1 [bib_ref] The molecular biology of geminiviruses, Stanley [/bib_ref]. Two ORFs are expressed in viral (sense) orientation-V2 and V1 [pre-coat and coat protein (CP), respectively]-and four ORFs in the complementary orientation (C1 to C4). Geminiviral DNA replication involves a number of steps [bib_ref] Family geminiviridae, Stanley [/bib_ref] [bib_ref] Geminiviruses: models for plant DNA replication, transcription, and cell cycle regulation, Hanley-Bowdoin [/bib_ref]. The uncoated open circular single-stranded (ss) DNA converts into a covalently closed circular double-stranded (ds) DNA. In begomoviruses, this stage requires priming by RNA, with the primosome complex provided by the host [bib_ref] Geminivirus replication origins have a group-specific organization of iterative elements: a model..., Argüello-Astorga [/bib_ref] [bib_ref] Geminiviruses: models for plant DNA replication, transcription, and cell cycle regulation, Hanley-Bowdoin [/bib_ref]. The conversion of covalently closed circular DNA into supercoiled forms is carried out solely by the host enzymatic machinery [bib_ref] Geminivirus DNA replication, Gutierrez [/bib_ref] [bib_ref] MVMIV replication initiator protein (Rep): roles at the initiation and elongation steps..., Singh [/bib_ref]. Closed circular dsDNA forms are bound by host histones to form minichromosomes, leaving sufficient gaps at the origin of replication, the promoter sites and the iterons to allow interactions with plant proteins and REP [bib_ref] The molecular biology of geminiviruses, Stanley [/bib_ref]. The dsDNA within the minichromosomes replicates and is used for transcription of the viral genes. When a sufficient amount of REP becomes available, it binds to the IR iterons, initiates a specific nick for rolling-circle replication, and produces the ssDNA-molecule progeny. The viral genes V1, V2, C4 [bib_ref] Family geminiviridae, Stanley [/bib_ref] [bib_ref] Geminivirus replication origins have a group-specific organization of iterative elements: a model..., Argüello-Astorga [/bib_ref] [bib_ref] Geminiviruses: models for plant DNA replication, transcription, and cell cycle regulation, Hanley-Bowdoin [/bib_ref] and (indirectly) C2 [bib_ref] The molecular biology of geminiviruses, Stanley [/bib_ref] are involved in symptom development and expression, disease severity and movement [bib_ref] Family geminiviridae, Stanley [/bib_ref] [bib_ref] Geminivirus replication origins have a group-specific organization of iterative elements: a model..., Argüello-Astorga [/bib_ref] [bib_ref] Geminiviruses: models for plant DNA replication, transcription, and cell cycle regulation, Hanley-Bowdoin [/bib_ref] [bib_ref] Geminivirus DNA replication, Gutierrez [/bib_ref]. Trafficking of the geminiviral genome within a cell, between cells and over long distances within the host has been attributed in some cases to the viral ssDNA [bib_ref] The molecular biology of geminiviruses, Stanley [/bib_ref] [bib_ref] Family geminiviridae, Stanley [/bib_ref] [bib_ref] Geminivirus replication origins have a group-specific organization of iterative elements: a model..., Argüello-Astorga [/bib_ref] [bib_ref] Geminiviruses: models for plant DNA replication, transcription, and cell cycle regulation, Hanley-Bowdoin [/bib_ref]. In other cases, both ssDNA and dsDNA have been implicated as moving entities in a non-sequence-specific manner [bib_ref] Bean dwarf mosaic geminivirus movement proteins recognize DNA in a form-and size-specific..., Rojas [/bib_ref]. The CP carries nuclear-localization and nuclear-export signals [bib_ref] The role of host and viral proteins in intra-and inter-cellular trafficking of..., Gafni [/bib_ref] [bib_ref] Genetic requirements for local and systemic movement of Tomato golden mosaic virus..., Jeffrey [/bib_ref] [bib_ref] Maize streak virus coat protein is karyophyllic and facilitates nuclear transport of..., Liu [/bib_ref] [bib_ref] Two proteins of a plant DNA virus coordinate nuclear and plasmodesmal transport, Noueiry [/bib_ref] [bib_ref] The geminivirus BR1 movement protein binds single-stranded DNA and localizes to the..., Pascal [/bib_ref] [bib_ref] Characterization of signals that dictate nuclear/nucleolar and cytoplasmic shuttling of the capsid..., Sharma [/bib_ref] and is thus involved in facilitating ssDNA invasion of the nucleus and possibly directing viral nuclear components back to the cytosol. C3 and C2 code for factors involved in replication and transcription enhancement, respectively. Thus, five of the viral ORFs have no direct role (but some auxiliary roles) in viral dsDNA replication. The sixth ORF (C1) is only essential for rolling-circle replication, a late-phase function which may, in turn, provide a template for dsDNA replication and play some auxiliary roles in recruiting plant proteins for optimization of viral replication and expression. In a previous study, the IL-60 platform of TYLCVderived plasmids (see further on for construct definitions) was found to be capable of expressing foreign genes or silencing native plant genes in a nontransgenic manner in every plant tested [bib_ref] A universal expression/silencing vector in plants, Peretz [/bib_ref]. Moreover, it can harbor long inserts and express an entire operon [bib_ref] Expression of an entire bacterial operon in plants, Mozes-Koch [/bib_ref]. In the IL-60 system, the activity of the ORF C1 (REP) product is compromised, thus eliminating rolling-circle replication and the production of progeny ssDNA. Therefore, the inserted IL-60 DNA components replicate in plant cells only as dsDNA. In our previous study [bib_ref] A universal expression/silencing vector in plants, Peretz [/bib_ref] , we further demonstrated that a foreign gene could be placed within the IL-60-pBlueScript (BS) construct or under IR control on a different construct for expression. In the latter case, IL-60-BS served as the ''driver'', providing needed elements for replication and expression [bib_ref] A universal expression/silencing vector in plants, Peretz [/bib_ref]. In the present report, we show that p1470, carrying the IR and only two sense-oriented TYLCV genes (without any of the four complementary-oriented genes), can replace the larger construct, IL-60-BS, in driving expression of a foreign gene in cis as well as in trans. Here, the IL-60 system (including its various components) was used to study aspects of TYLCV functional genomics. A major discovery was that the short noncoding viral DNA segment (IR) is sufficient for basal levels of viral replication and that the addition of only two viral genes (IR-V2-V1 = p1470) allows expression and movement. It was also found that a reporter gene is expressed in cis and in trans when fused to p1470 or placed under IR control with p1470 serving as the driver (the limited-size construct p1470 could replace IL-60-BS as a driving entity in transactivation of IR-X). This paper demonstrates that the IR is fundamental to viral replication, expression and movement and, at a later phase, is regulated to function (along with other viral and host factors) in enhancing progeny production. The association of TYLCV CP with IR is also demonstrated. # Materials and methods ## Clones and constructs In this report, shorter components of the previously reported IL-60 system [bib_ref] A universal expression/silencing vector in plants, Peretz [/bib_ref] were explored, as described in the section ''IR-carrying constructs'' below. The term ''IL-60 system'' is used when we are not referring to a particular component of the system. ''IL-60-BS'' refers to a manipulated construct of TYLCV fused to the plasmid Blue-Script, 5,682 bp in length, as described by Peretz et al. [bib_ref] A universal expression/silencing vector in plants, Peretz [/bib_ref]. All constructs not carrying IL-60-BS were built for the resent study or their use was largely elaborated in the present paper. In some of the constructs described here and in our previous report [bib_ref] A universal expression/silencing vector in plants, Peretz [/bib_ref] , the BS plasmid could be replaced by another plasmid (pDRIVE). The IR segment of TYLCV was PCR-amplified and cloned into the T/A cloning vector pDRIVE. This construct presumably carries the ribosomal binding site of V2. A bglucuronidase (GUS) gene (uidA) was force-cloned with SalI and SacI into pDRIVE-IR downstream of the IR. Similarly, the gene for green fluorescent protein (GFP) was force-cloned with SalI and HindIII. C2 was cloned on the opposite side of GUS using PstI and KpnI [fig_ref] Figure 1: Illustrations [/fig_ref]. In all constructs, IRs were positioned so that the transcripts would be regulated by the promoter controlling V2-V1 transcription. 35S:GFP served as an IR-less control. ## Propagation of il-60 constructs and their administration into plants Escherichia coli cells were transformed with the pertinent IL-60 construct and propagated under ampicillin selection, and the construct was extracted by standard procedures. IL-60 constructs were administered to plants by injection [bib_ref] A universal expression/silencing vector in plants, Peretz [/bib_ref] or root uptake-root tips of tomato plantlets at the first leaf stage were trimmed and immersed for 1-2 days at room temperature in an aqueous solution containing 1 lg DNA per plant until solution imbibition was complete. TYLCV inoculation TYLCV-viruliferous Bemisia tabaci was maintained on TYLCV-infected Datura stramonium or tomato plants. For inoculation, viruliferous insects (50 insects per plant) were placed in cages containing uninfected, IL-60-carrying tomato plants at the second-true-leaf stage. A week later, the insects were killed using Confidor and the plants were transferred to an insect-free cage. A week after that, all plants were inspected for residual insect infestation, and all insect-free plants were transferred to the greenhouse for phenotypic observation and molecular analysis. # Molecular analysis Southern, northern, western, PCR, quantitative (q) PCR and RT-PCR analyses were carried out according to standard procedures . Unless stated otherwise, PCR assays consisted of 40 cycles. Probes for Southern and northern analyses were labeled by the PCR-DIG procedure (Roche Molecular Biochemicals). GUS expression was monitored by staining according to Jefferson et al. [bib_ref] GUS fusions: beta-glucuronidase (gus) as a sensitive and versatile gene fusion marker..., Jefferson [/bib_ref]. Western blot analysis of plant material was performed on total tomato protein extracts after removal of the masking Rubisco protein according to Xi et al. [bib_ref] Polyethylene glycol fractionation improved detection of lowabundant proteins by two-dimensional electrophoresis analysis..., Xi [/bib_ref]. Total DNA from BY-2 protoplasts was subjected to overnight digestion by DpnI (Fermentas) in the original buffer at 20 units per 5 ll of genomic DNA. Digested DNA was PCR-amplified using primers for the entire length of the GUS gene (uidA) (ESM . ## Protoplast extraction and transformation BY-2 cells were grown in Linsmaier and Skoog medium in which KH 2 PO 4 and thiamine-HCl were increased to 370 and 1 mg/l, respectively, and sucrose and 2,4-D were supplemented to 3 % and 0.2 mg/l, respectively. Cells were harvested 3 days after culture dilution to reach peak mitotic stage [bib_ref] Tobacco BY-2 cell line as the ''HeLa'' cell in the cell biology..., Nagata [/bib_ref]. BY-2 cells were centrifuged at 500 rpm and the supernatant was gently removed. Cell walls were then digested in an enzymatic solution ## Dna extraction from by-2 protoplasts At each sampling, cells were removed and frozen at -80°C, then placed in liquid nitrogen and homogenized in lysis buffer (100 lg/ml proteinase K, 0.4 % SDS) at 55°C for 1 h using a mortar and pestle. DNA was extracted from the cell suspension with phenol:choloroform:isoamyl alcohol (25:24:1, v/v). Following centrifugation, the upper phase was decanted into a clean tube and supplemented with 1 volume of isopropanol and 20 ll 5 M NaCl. After another centrifugation the pellet was washed twice with 70 % EtOH and suspended in molecular-grade water. ## Confocal microscopy Post-transfection, protoplast suspensions were studied over time under a confocal microscope (Zeiss 100M). Excitation was at 488 nm, and GFP emission was detected at 505-550 nm. Chlorophyll autofluorescence was detected at wavelengths longer than 560 nm. Data were processed by the built-in program LSM 51. ## Immunocapture pcr Immunocapture PCR was carried out, with modifications, as described by Wetzel et al. [bib_ref] A highly sensitive immunocapture polymerase chainreaction method for plum pox potyvirus detection, Wetzel [/bib_ref]. Plant extracts (1 g leaves in 3 ml extraction buffer [20 mM Tris-HCl, pH 8.0, 138 mM NaCl, 3 mM KCl, 1 mM PVP, 0.05 % Tween 20]) were treated with DNase. Polystyrene microtiter plates or polypropylene tubes were coated with 100 ll anti-TYLCV antibodies (diluted 1:500 in carbonate buffer, pH 9.6) and incubated overnight at 4°C. The wells were washed five times with PBS-Tween. Then, 100 ll of DNase-treated plant extract (diluted 1:100 in extraction buffer) was added and subjected to UV irradiation (12,000 J/cm 2 ) for 4 h. PCR amplification was then carried out in the same wells with primers for TYLCV CP or phytoene desaturase (PDS) genes. ## Real-time qpcr Each DNA sample was diluted 1:50 before amplification. Approximately 10 ng of DNA was taken for each PCR amplification. Real-time qPCR in a LightCycler 480 (Roche) was performed with the primers and probes listed in ESM , and data analysis was performed using the built-in software. The real-time PCR program was as follows: 95°C for 10 min, followed by 45 cycles of 95°C for 10 s and 60°C for 30 s, ending with a step at 40°C for 30 s. DNA levels were standardized using the 18S rRNA gene as an internal control. Plasmid DNA harboring the GUS gene was quantified using the standard calibration curve of serial 10-fold dilutions of IR-GUS plasmid. A tobacco 18S quantification curve was obtained using serial 10-fold dilutions of Nicotiana tabacum genomic DNA. Light Cycler software was used for actual quantification of each DNA species, and tomato actin was used as a standard. Primers and probes for GUS amplification and 18S DNA amplification are presented in ESM . # Results ## Ir-carrying constructs For foreign gene expression, we cloned the reporter genes encoding GUS (bases 1466 to 3274, GenBank accession no. M14641) or GFP (excised from the plasmid 30b-GFP3, courtesy of William O. Dawson, University of Florida, Lake Alfred) downstream of the IR (IR-GUS and IR-GFP). IR-GUS was described previously [bib_ref] A universal expression/silencing vector in plants, Peretz [/bib_ref] and is presented here by another illustration. IR-GFP is a similar construct, in which GFP replaces GUS. To assay the functions of the sense-oriented genes, we fused V2-V1 to the IR, forming a TYLCV-derived 1,467-bp insert in pDRIVE, and named it ''p1470''. Another construct named ''p1470-GFP'' consisted of the entire gene encoding GFP downstream of p1470. The bacterial plasmid component for all IR-carrying constructs was pDRIVE, but other plasmids, such as pBS, have also been employed successfully. In some cases, a nopalin terminator (termed nos or ter) was added to the reporter gene. In most cases, the tested genes were placed downstream to the ''right'' promoter of the IR (driving the expression of sense-oriented genes). In a certain limited number of constructs, genes were placed under the control of the ''left'' promoter of the IR. For example, in IR-C2-GUS, the C2 and GUS genes were placed under the ''right'' promoter, and in C2-IR-GUS, C2 was placed under the ''left'' promoter and GUS was placed under the ''right'' promoter. The constructs are diagrammatically represented in [fig_ref] Figure 1: Illustrations [/fig_ref]. ## Foreign dna longevity in plant cells is extended by the ir We previously demonstrated foreign gene expression by plasmids consisting of a reporter gene downstream of the TYLCV IR when co-inoculated into plants with IL-60 [bib_ref] A universal expression/silencing vector in plants, Peretz [/bib_ref]. Here, we tested the effect of the IR alone on the longevity (stability) of foreign DNA in situ in the absence of any detectable viral gene. To ensure ribosome binding to the RNA transcribed from the IR, the 5' terminal sequence of V2 (presumably carrying a ribosomal binding site) was fused between the IR and the reporter gene. IR-GFP [fig_ref] Figure 1: Illustrations [/fig_ref] was injected into the petioles of 4-to 5-week-old tomato plants. In parallel, a GFP-carrying plasmid (pDRIVE) lacking the IR, as well as pDRIVE harboring GFP-coding sequences downstream of the 35S promoter, were simultaneously and similarly injected into control tomato plants. DNA was then extracted from the tissue excised from the treated site 1, 5, 10 and 20 days post-injection. The constructs were administered into only a few cells, and therefore the presence of GFP sequences in the tissue samples could only be determined by PCR. [fig_ref] Figure 2: Stability of IR-carrying constructs in intact plants [/fig_ref] shows that GFP sequences from all three constructs were present in planta for 1 day but only persisted for 5 days or more only in plants treated with IR-carrying constructs. We concluded that DNA stability is conferred by the IR but not by 35S, which is a strong virus-derived promoter for expression in plants. The IR is the only viral element required for basal dsDNA replication DNA constructs fused to the IR are stabilized in plant cells. This may result from, among other possibilities discussed further on, the ability of the construct to replicate in the cells. To test the potential for IR-directed replication, protoplasts were transfected with IR-GFP or IR-GUS. Slot-blot hybridization assay showed IR-GFP accumulation, indicating replication. This was corroborated by realtime qPCR, which showed replication of IR-GUS [fig_ref] Figure 3: Left frame [/fig_ref]. pDrive-GFP and 1470-BS-GUS served as IR-less controls. These results clearly show that GUS and GFP constructs carrying only the IR as the viral element accumulate with time in protoplasts. The level of DNA at zero time is considered to be the level of input DNA. Since the sensitivity of PCR is greater than that of hybridization, the level of input DNA is detected by PCR, while in the hybridization assay it is below the detection limit. Geminiviral DNA evades methylation in plant cells [24, and references within]. Therefore any geminiviral DNA made in plants is DpnI resistant. Therefore, non-methylated geminiviral DNA indicates that it has been propagated in planta and not in bacteria. DNA replicated in protoplasts was resistant to DpnI, indicating its plant (and not bacterial) origin [fig_ref] Figure 4: Evidence that the IR-GUS extracted from protoplasts is not methylated, indicating that... [/fig_ref]. Viral sense-oriented genes support IR-directed expression GUS protein was previously shown to be expressed from IR-GUS driven by IL-60 [bib_ref] A universal expression/silencing vector in plants, Peretz [/bib_ref]. The western blot shown in [fig_ref] Figure 5: Western blot analysis for GUS [/fig_ref] illustrates that IR is necessary for GUS expression and cannot be replaced by the 35S promoter. The constructs were introduced into the plants by root uptake; proteins were extracted from plant leaves 2 weeks later, and the Rubisco protein was removed by polyethylene fractionation as described by Xi et al. [bib_ref] Polyethylene glycol fractionation improved detection of lowabundant proteins by two-dimensional electrophoresis analysis..., Xi [/bib_ref]. The proteins were subjected to western blot analysis with anti-GUS antibodies [fig_ref] Figure 5: Western blot analysis for GUS [/fig_ref]. Expression was also followed in vivo by GFP fluorescence. IR-GFP alone, IR-GFP together with IL-60, and the fused construct p1470-GFP (carrying only sense-oriented TYLCV genes, as depicted in [fig_ref] Figure 1: Illustrations [/fig_ref] were introduced into protoplasts by transfection. IR alone did not support expression of the IR-fused reporter gene . However, IR-directed expression did occur when IL-60 was added in trans or when the reporter gene was fused to the senseoriented viral genes (V2, V1; exemplified by p1470-GFP). Adding p1470 to IR-GFP in trans also promoted expression (data not shown). ## Tylcv provides in-trans factors that facilitate mobilization of genes fused to the ir in plants In our previous paper [bib_ref] A universal expression/silencing vector in plants, Peretz [/bib_ref] , TYLCV infection was shown to provide factors enabling the mobilization of IL-60-BS. Here, we demonstrate that the target of said factors is IR. GUS or GFP under the transcriptional control of the IR was injected into tomato leaf petioles. PCR amplification did not reveal any traces of the inserts in leaves remote from the point of injection. Seven days post-administration, the IR-GUS-and IR-GFP-treated plants were inoculated with TYLCV. Successful infection was confirmed by PCR for TYLCV CP [fig_ref] Figure 7: The effect of TYLCV infection on the expression and spread of IR-GFP [/fig_ref]. GFP distribution and expression in IR-GFPtreated plants were observed by fluorescence and PCR, respectively [fig_ref] Figure 7: The effect of TYLCV infection on the expression and spread of IR-GFP [/fig_ref]. Mobilization of IR-GUS in TYLCV-infected plants was detected by staining [fig_ref] Figure 7: The effect of TYLCV infection on the expression and spread of IR-GFP [/fig_ref]. Note that addition of the sense-oriented V1 and V2 genes (in the form of p1470-X) supported mobilization in planta, which was comparable to what was observed in tissues infected with TYLCV (as reported further on). The IR and products of the sense-oriented viral genes V1 and V2 are sufficient for dsDNA mobilization in plants IR-carrying constructs are capable of replication, but, as already noted, the products remain localized within the We defined V1 and/or V2 as the movement determinants by showing that, in addition to their role in directing expression, these sense-transcribed TYLCV genes under IR regulation can promote dsDNA mobilization in TYLCV-free plants. P1470-GFP was injected into the petiole of the second true leaf of tomato plants with 4-6 leaves. Four days later, the treated leaves and the younger leaves positioned above them were analyzed by confocal microscopy and PCR [fig_ref] Figure 8: Replication and movement of p1470-GFP in tomato plants [/fig_ref]. GFP fluorescence was evident in the cytoplasm of cells remote from the point of administration [fig_ref] Figure 8: Replication and movement of p1470-GFP in tomato plants [/fig_ref]. GFP fluorescence was also observed in parallel locations in neighboring cells on both sides of the cell wall, suggesting cell-to-cell movement [fig_ref] Figure 8: Replication and movement of p1470-GFP in tomato plants [/fig_ref]. In newly emerged leaves, expression and transport of GFP in the presence of V2 and V1 under IR control were confirmed by PCR and RT-PCR analyses [fig_ref] Figure 8: Replication and movement of p1470-GFP in tomato plants [/fig_ref]. [fig_ref] Figure 8: Replication and movement of p1470-GFP in tomato plants [/fig_ref] demonstrates the absence of TYLCV contamination, as indicated by the lack of C2. We propose that expression of V1 and V2 facilitated the mobilization of IR-GFP and its trafficking through the plant. Evidence of IR association with CP As already stated, IR-carrying constructs are induced to move systemically and express inserted sequences following challenge inoculation with TYLCV. CP has been indicated (along with V2) as a factor involved in virus mobility. Immunocapture-PCR assays were performed with IR-PDS as the template (PDS is easily distinguished from CP). The PCR tubes were coated with anti-TYLCV-CP antibodies. Extracts of IR-PDS-treated, TYLCV-infected plants were incubated in the coated tubes, and unbound material was washed away. The coated tubes were subjected to DNase treatment and UV irradiation to remove any DNA bound to the outside of the trapped entities, and PCR was carried out with primers for the PDS gene. Aside from the expected trapped TYLCV, the immunocaptured material also included PDS DNA [fig_ref] Figure 9: Immunocapture assays of plants treated with IR-PDS [/fig_ref]. Neither PDS nor TYLCV was amplified in the absence of antibodies, ruling out the possibility that amplification resulted from plant material not associated with CP. Thus, TYLCV infection appears to provide the capsid part for the IR-PDS association with CP. The association with in-trans-provided CP suggests that at least some of the roles of the IR are carried out by IR-CP complexes. # Discussion It was previously impossible to determine the functional phase of the TYLCV ''life cycle'' at which a certain event occurs. Using the IL-60 system [bib_ref] A universal expression/silencing vector in plants, Peretz [/bib_ref] , however, dsDNA replication and expression can be uncoupled from the normal ''life cycle'' of TYLCV [ssDNA?(dsDNA? dsDNA?dsDNA) n ?ssDNA]. This enables studying events occurring in the middle phase (dsDNA and its various functions) and their separation from early-phase events (entry, uncoating, transport to the nucleus, etc.) and late-phase functions (replication of progeny ssDNA, virus assembly, transmission to other plants). Many pieces of data provided in this paper are restricted to the middle phase of the TYLCV ''life cycle''. Employing the IL-60 system, we reveal new features of the noncoding (IR) component of the TYLCV genome and revise the role of some of the other TYLCV genes. The small-size TYLCV exerts its activities mostly by recruiting host factors through protein-protein and protein-DNA interactions [bib_ref] DNA replication and cell cycle in plants: learning from geminiviruses, Gutierrez [/bib_ref] [bib_ref] Geminivirus DNA replication and cell cycle interactions, Gutierrez [/bib_ref] [bib_ref] Reprogramming plant gene expression: a prerequisite to geminivirus DNA replication, Hanley-Bowdoin [/bib_ref] [bib_ref] RKP, a RING finger E3 ligase induced by BSCTV C4 protein, affects..., Lai [/bib_ref] [bib_ref] Dual interaction of geminivirus replication accessory factor with a viral replication protein..., Settlage [/bib_ref] [bib_ref] Geminivirus C3 protein: replication enhancement and protein interactions, Settlage [/bib_ref] [bib_ref] Arabidopsis protein kinases GRIK1 and GRIK2 specifically activate SnRK1 by phosphorylating its..., Shen [/bib_ref]. These interactions are sometimes regulated by interactions between the viral products themselves. In several cases, virus-host complexes increase due to induced expression of the host proteins participating in complex formation or of their regulatory elements [bib_ref] The interaction between geminivirus pathogenicity proteins and adenosine kinase leads to increased..., Baliji [/bib_ref] [bib_ref] C2 from Beet curly top virus promotes a cell environment suitable for..., Caracuel [/bib_ref] [bib_ref] Geminivirus C2 protein might be the key player for geminiviral co-option of..., Lozano-Duran [/bib_ref] [bib_ref] Suppression of RNA silencing by a geminivirus nuclear protein, AC2, correlates with..., Trinks [/bib_ref]. Some of the roles of viral genes, discussed further on, involve complexes of viral gene products or viral sequences with host elements. ## Function of complementary-oriented tylcv gene products Only the IR and the sense-oriented genes V1 and V2 are required for replication, expression and movement. However, the complementary-oriented genes (C1-C4) are known to have a role in replication and expression as well. This seeming contradiction can be explained by separating the various roles along a time line. During the early events, basal levels of the viral genes are expressed. Some of the gene products interact among themselves or with host factors, producing complexes that will serve in a later phase when progeny viral particles are produced. As already noted, a viral gene product (CP) interacts with the IR [fig_ref] Figure 9: Immunocapture assays of plants treated with IR-PDS [/fig_ref]. The IR-CP complex may enable TYLCV infection to proceed to the next stage (i.e., late events). Antisense-oriented viral genes have been reported to be associated with DNA replication [30 and citations therein]. However, as shown, none of the complementary-oriented TYLCV gene products are required for dsDNA replication. Nevertheless, C3, the product of ORF C3 of TYLCV, has been characterized as a DNA-replication enhancer. These seemingly contradictory data might be explained by taking into account that these activities can occur at different phases of TYLCV replication. C3 affects the rolling-circle phase and replication of progeny viral ssDNA, probably by interaction with C1 and by self-oligomerization [bib_ref] Geminivirus C3 protein: replication enhancement and protein interactions, Settlage [/bib_ref]. Despite being termed ''DNA-replication enhancer'', it is not involved in, or required for, any aspect of dsDNA replication in the middle phase. Complementary-oriented viral gene products (C1, C3, C4) are involved in recruiting host factors that restore cellcycle programming in resting plant cells [bib_ref] DNA replication and cell cycle in plants: learning from geminiviruses, Gutierrez [/bib_ref] [bib_ref] Geminivirus DNA replication and cell cycle interactions, Gutierrez [/bib_ref] [bib_ref] Reprogramming plant gene expression: a prerequisite to geminivirus DNA replication, Hanley-Bowdoin [/bib_ref] [bib_ref] RKP, a RING finger E3 ligase induced by BSCTV C4 protein, affects..., Lai [/bib_ref] [bib_ref] Dual interaction of geminivirus replication accessory factor with a viral replication protein..., Settlage [/bib_ref] [bib_ref] Geminivirus C3 protein: replication enhancement and protein interactions, Settlage [/bib_ref] [bib_ref] Arabidopsis protein kinases GRIK1 and GRIK2 specifically activate SnRK1 by phosphorylating its..., Shen [/bib_ref] and stimulate the expression or activation of proteins of the transduction pathway [bib_ref] Arabidopsis protein kinases GRIK1 and GRIK2 specifically activate SnRK1 by phosphorylating its..., Shen [/bib_ref] [bib_ref] Global analysis of Arabidopsis gene expression uncovers a complex array of changes..., Ascencio-Ibáñez [/bib_ref]. C3 interacts with the host proliferating cell nuclear antigen (PCNA), which is involved in normal cell-cycle processes and might therefore affect dsDNA replication at a later phase. Virusinduced reprogramming of the expression of host genes and the resultant interactions between viral and host-cell components might be required for the acceleration of dsDNA replication in the late phase, when many progeny ssDNA are produced from dsDNA and the production of template dsDNA is accelerated as well. C2 has been identified as a transcriptional activator of late viral gene expression and a silencing suppressor. C2 analogs in other geminiviruses have also been reported to be transcriptional activators [bib_ref] The two nonstructural proteins from wheat dwarf virus involved in viral gene..., Collin [/bib_ref] [bib_ref] The Tomato golden mosaic virus transactivator (TrAP) is a single-stranded DNA and..., Hartitz [/bib_ref] [bib_ref] Regulation of a geminivirus coat protein promoter by AL2 protein (TrAP). Evidence..., Sunter [/bib_ref]. However, with C2 attached [fig_ref] Figure 1: Illustrations [/fig_ref] or in the absence of C2, the construct IR-V2-V1 drives expression to similar levels (data not shown). Hence, C2 is not required for activation of the basal transcription of sense-oriented genes. These data may not be contradictory: the presently studied middle-phase activities of TYLCV require only factors for dsDNA replication, expression and movement. In the late phase, on the other hand, many CP molecules are required for progeny virus encapsidation, and the rate of CP production must be considerably increased. Thus, C2 produced in the middle phase has no role in middle-phase activities but is required for the late phase of TYLCV infection; with its binding capacity to ssDNA [bib_ref] The Tomato golden mosaic virus transactivator (TrAP) is a single-stranded DNA and..., Hartitz [/bib_ref] , C2 may be regulated by the accumulated ssDNA and/or by the IR-CP complex, thus elevating transcription rates to accommodate the increased need for CP in the late phase. ## Role of ir in dna longevity and replication The IR was shown to be a stabilizing element for dsDNA. One possible reason might be that it attracts DNA-binding host proteins to construct a protective shield. Indeed, geminiviral dsDNA has been reported to associate with histones to form minichromosomes [bib_ref] Abutilon mosaic geminivirus doublestranded DNA is packed into minichromosomes, Pilartz [/bib_ref] [bib_ref] Mapping of Abutilon mosaic geminivirus minichromosomes, Pilartz [/bib_ref] , and the IR may mediate this process. Another possibility is that IR-X (X being any DNA) stability is gained by its autonomous replication. Our findings support the latter notion that dsDNA ''stabilization'' (increased longevity) by IR is due to perpetual replication. We demonstrated that IR is the only viral component required for dsDNA replication. Since replication does not require any viral gene product, it must be entirely dependent on the recruitment of host factors. Presumably, the IR attracts the host factors, initiating the replication of IRcarrying DNA, which then becomes replication-competent. ## Expression None of the geminivirus (including TYLCV) genes code for RNA polymerases. Hence, viral transcription depends entirely on host polymerases and transcription factors that recognize the promoters located in the IR. Here, we demonstrate that the minimal requirement for expression (transcription and translation) is the two IR-directed senseoriented TYLCV genes V1 and V2. Foreign genes fused to IR-V2-V1, such as with p1470-GFP [fig_ref] Figure 1: Illustrations [/fig_ref] , are also expressed. Foreign genes directed by the IR would also be expressed in trans, i.e., gene X would replicate as an IR-X construct but would be expressed only if supplemented in trans by p1470. None of the complementary-sense TYLCV genes seem to be essential for expression, replication, or movement. Nevertheless, they might affect the levels of replication and expression due to interactions with host factors or direct association with IR promoters. The ORFs V1 and V2 were found to be necessary for both expression and movement. These two aspects may be linked together. It is possible that the initial expression took place in the primary cells and the entities spreading to distant cells were the transcripts. In such a case, movement would be an automatic consequence of expression. ## Movement It is well established that the karyophilic CP facilitates TYLCV ssDNA import into the nucleus [bib_ref] The role of host and viral proteins in intra-and inter-cellular trafficking of..., Gafni [/bib_ref] [bib_ref] Characterization of signals that dictate nuclear/nucleolar and cytoplasmic shuttling of the capsid..., Sharma [/bib_ref] [bib_ref] Nuclear import of the capsid protein of tomato yellow leaf curl virus..., Kunik [/bib_ref] [bib_ref] Maize streak virus coat protein is karyophyllic and facilitates nuclear transport of..., Liu [/bib_ref]. Its spread, however, depends on export of the viral DNA to the cytosol, movement to the cell periphery and traversing to neighboring cells. CP has indeed been reported to carry nuclear export signals as well, and to mediate the nuclear exit of viral DNAs [bib_ref] Functional analysis of proteins involved in movement of the monopartite begomovirus, Tomato..., Rojas [/bib_ref] [bib_ref] Nuclear import of the capsid protein of tomato yellow leaf curl virus..., Kunik [/bib_ref] [bib_ref] Maize streak virus coat protein is karyophyllic and facilitates nuclear transport of..., Liu [/bib_ref] , which move within the plant in both their ss and ds forms [bib_ref] Bean dwarf mosaic geminivirus movement proteins recognize DNA in a form-and size-specific..., Rojas [/bib_ref]. Transport from the nucleus to the cytoplasm and cell-to-cell trafficking through plasmodesmata have been suggested to be mediated by various CP-protein complexes [bib_ref] Characterization of signals that dictate nuclear/nucleolar and cytoplasmic shuttling of the capsid..., Sharma [/bib_ref] [bib_ref] Interaction of DNA with movement proteins of geminiviruses revised, Hehnle [/bib_ref]. The level of expression of CP from IR-V1-V2 is sufficient for viral DNA trafficking and possibly, when combined with host factors and/or viral DNA sequences, the initiation of late viral functions as well. The products of C2 and C4, as well as some host proteins, have also been reported to be involved in virus trafficking [bib_ref] Movement of Tomato yellow leaf curl geminivirus (TYLCV): involvement of the protein..., Jupin [/bib_ref] [bib_ref] Involvment of C4 protein of Beet severe curly top virus (familiy Geminiviridae)..., Teng [/bib_ref]. However, our results show that the sense-oriented viral genes (V1, V2) are sufficient for movement of viral dsDNA within the plant. It has been suggested that both ssDNA and dsDNA are shuttled within the infected plant. C2 and C4 may play a role in ssDNA movement and be redundant for dsDNA shuttling. As discussed above, the moving entities could be the transcripts made following the enabling of expression. Thus, the initiation of expression in one cell allows the spread of the expressing entity to other cells. If this is the case, the ''expression'' and ''movement'' are cause and consequence. Other features of the IR As shown in this report, the IR, even when fused to a sequence foreign to the TYLCV system, binds CP. All IR-X constructs in this paper are dsDNA. Hence, the conclusion is that the IR in its dsDNA form is associated with CP. dsDNA is not encapsidated, making it unlikely that such an interaction is required for virion assembly. As reported here and elsewhere, CP is probably involved in dsDNA movement. It is therefore likely that CP binds to the IR in its dsDNA form. We thus propose that CP has a regulatory role in one or more IR functions: for instance, enhancement of IR promoter activity, better driving of V2-V1 expression following CP binding, and the formation of a positive loop in which the CP produced at an early stage enhances its own production when more CP is needed at a later stage of the viral ''life cycle''. A similar role was assigned to the gene product of C2, which, in complex with host proteins, exerts late promoter activity, probably, as indicated in this report, by interacting with IR-CP. From a satellite to a replication-competent plant plasmid and virus Many naturally occurring geminivirus-associated satellites have been isolated and characterized (e.g., see Briddon and Stanley [bib_ref] Subviral agents associated with plant single-stranded DNA viruses, Briddon [/bib_ref]. Most of them are about half the size of the helper virus genome, except for the very short (682 bp) satellite described by Dry et al. [bib_ref] A novel subviral agent associated with a geminivirus: the first report of..., Dry [/bib_ref]. All hitherto-described geminivirus-related satellites carry motifs that are also in the helper virus IR. They are encapsidated and insecttransmissible. Here, we further developed IL-60-BS [bib_ref] A universal expression/silencing vector in plants, Peretz [/bib_ref] into a shorter construct (p1470), carrying only the IR and two viral genes, that replicates, expresses and moves independently of all other viral components. P1470 can therefore be considered a plant plasmid. For expression, p1470 relies on a helper virus, or parts thereof, and hence can be considered a satellite. It appears that the two parts of the TYLCV genome assume different sets of functions. The sense-oriented genes are involved in all matters concerning the intraplant dsDNA phase of infection as expected of a plasmid, i.e., stability, replication, expression and movement. Thus, independent of the complementary-oriented part of the genome, the sense-oriented segment can develop into an autonomous entity capable of moving within the plant. However, it cannot produce ssDNA progeny or assemble into virions. Hence, it is unable to be naturally transmitted from one plant to another. The complementary-oriented genes are required for conversion of the dsDNA to its progeny ssDNA (by rolling-circle replication) and also have roles in recruiting plant factors that aid in replication and expression. The TYLCV progeny ssDNA, however, can be encapsidated, and the resultant viral particles are transmissible from one plant to another by insects. Hence, the complementary-oriented part of the TYLCV genome enables conversion of a plasmid to a virus. [fig] Figure 1: Illustrations (not to scale) of the various IR-carrying constructs. pT7 and SP6, sites of the T7 and SP6 promoters; *V2, the first 282 bp of ORF V2; C4*, the last 279 bp of the interrupted C1 ORF (the truncated N terminus of C4); C2, the entire ORF of TYLCV C2Requirements for TYLCV replication, expression and movement 2265 [/fig] [fig] Figure 2: Stability of IR-carrying constructs in intact plants. All DNA samples were taken from the point of injection and were analyzed by PCR for the presence of GFP sequences. A. Stability of a GFPharboring plasmid (pDRIVE-GFP) versus the same plasmid carrying the IR. Lanes 1 and 3, IR-carrying plasmids; lanes 2 and 4, IR-lacking plasmids; lane 6, negative control-DNA extract from untreated plants. B. Stability of 35S-vs. IR-carrying plasmids. PCR was performed at various times postinjection as specified [/fig] [fig] Figure 3: Left frame: Replication of IR-GUS in protoplasts as indicated by real-time qPCR. The time after application of the construct is shown on the x-axis. Real-time qPCR (average of three repetitions) was performed on DNA extracted from 10 5 protoplasts, and results were normalized to those for the 18S rRNA gene. DNA extracted from untransfected protoplasts served as a control for real-time PCR. Right frame: Slot-blot assay of IR-GFP. The slot-blot was carried out by applying 10 5 lysed protoplasts to each slot and hybridizing to a GFP probe [/fig] [fig] Figure 4: Evidence that the IR-GUS extracted from protoplasts is not methylated, indicating that it was replicated in the protoplasts and does not represent the input DNA. DNA extracted from protoplasts at various times following IR-GUS transfection was digested with DpnI, and the entire GUS sequence was PCR-amplified. Failure of DpnI digestion indicated that the DNA was not methylated. Lane 1, size markers; lane 2, control-DNA extracts from untransfected protoplasts; lane 3, digestion of bacterial-extracted IR-GUS; lanes 4-6, digestion of PCR-amplified DNA extracts from IR-GUS-transfected protoplasts at 16, 24 and 48 h posttransfection, respectively [/fig] [fig] Figure 5: Western blot analysis for GUS. Lane 1, extracts of untreated plants; lane 2, extracts of plants treated with IL60?C2-IR-GUS-ter; lane 3: extracts of plants treated with IL60?35S:GUS-ter (no IR); lane 4, extract of plants treated with IL60?IR-GUS-ter; lane 5, extract of transgenic GUS-expressing plants Fig. 6 GFP expression in protoplasts transfected with various IL-60 constructs. GFP (left-hand column) and chlorophyll (center column) fluorescence was detected using a confocal microscope supplemented with suitable filters. Superpositions of the left-hand column on the center column are presented in the right-hand column [/fig] [fig] Figure 7: The effect of TYLCV infection on the expression and spread of IR-GFP. A. Detection of TYLCV in infected plant tissues by PCR, using TYLCV-CP primers. Lane 1, size markers; lane 2, DNA extract from untreated, uninoculated plants as template; lane 3, DNA extract from IR-GFP-treated, uninoculated plants as template; lanes 4, 6 and 7, DNA extracts from IR-GFP-treated, TYLCV-inoculated plants as template; lane 5, positive control-a plasmid carrying TYLCV CP served as the template. B. Expression and movement of IR-GFP and IR-GUS in TYLCV-infected plants. Confocal images were taken from leaves of IR-GFP-treated plants. I, TYLCV-inoculated, plasmid-devoid plant; II, plant treated with IR-GFP, 14 days after TYLCV inoculation; III and IV, plants treated with IR-GFP, 28 days after TYLCV inoculation. Bars represent 100 lm. V, image of a stained leaf section of an IR-GUS-treated plant, 14 days after TYLCV inoculation. VI, image of a stained leaf section of an untreated plant. C. Detection of GFP sequences (by PCR with GFP primers) in systemic leaves of IR-GFP-treated plants 28 days post-TYLCV inoculation. Lane 1, size markers; lane 2, DNA extract from a plasmid-devoid, TYLCV-infected plant as template; lane 3, positive control-a GFP-harboring plasmid served as a template; lane 4, DNA extract from IR-GFP-treated plant not inoculated with TYLCV as template; lanes 5 and 6, DNA extracts from IR-GFP-treated, TYLCV-inoculated plants as template; pD-IR-GFP, IR-GFP inserted in the plasmid pDRIVE Requirements for TYLCV replication, expression and movement 2269 treated cells. The presence of TYLCV-mobilizing factors, however, enables in vivo movement through the treated plants. [/fig] [fig] Figure 8: Replication and movement of p1470-GFP in tomato plants. A. GFP detection in systemic leaves of a plant treated with p1470-GFP. B. GFP expression in parallel locations across the cell wall in two adjacent cells. C. PCR with primers for TYLCV CP confirms the spread of p1470. (?) and (-) indicate p1470-GFP-treated and untreated plants, respectively. D. RT-PCR with primers for TYLCV CP, confirming the expression of V1 (CP) from p1470. (?) and (-) indicate p1470-GFP-treated and untreated plants, respectively. Lane 8, negative control-PCR performed without template. E. PCR with GFP primers, corroborating the presence of p1470-GFP. (?) and (-) indicate p1470-GFP-treated and untreated plants, respectively. Lane 6, negative control-no template. F. PCR with primers for TYLCV C2. Two p1470-GFP-treated plants were analyzed for the presence of CP and C2. Lane 6, negative control-no template. The absence of C2 indicates a lack of TYLCV contamination. Lane 1 in frames C-F, size markers [/fig] [fig] Figure 9: Immunocapture assays of plants treated with IR-PDS. Lane 1, PCR of the antibody-bound fraction from a TYLCV-infected, untreated plant; lane 2, PCR of the antibody-bound fraction from IR-PDS-treated plants with primers for TYLCV CP; lanes 3-6, same as in lane 2 but with primers for PDS amplification; lanes 7 and 8, negative controls-PCR with plant extracts in the absence of anti-TYLCV CP antibodies; lane 9, empty; lane 10, positive control-a PDS-carrying plasmid served as the template Requirements for TYLCV replication, expression and movement 2271 [/fig]

Study on Horizon Scanning with a Focus on the Development of AI-Based Medical Products: Citation Network Analysis Horizon scanning for innovative technologies that might be applied to medical products and requires new assessment approaches to prepare regulators, allowing earlier access to the product for patients and an improved benefit/risk ratio. The purpose of this study is to confirm that citation network analysis and text mining for bibliographic information analysis can be used for horizon scanning of the rapidly developing field of AI-based medical technologies and extract the latest research trend information from the field. We classified 119,553 publications obtained from SCI constructed with the keywords "conventional," "machine-learning," or "deep-learning" and grouped them into 36 clusters, which demonstrated the academic landscape of AI applications. We also confirmed that one or two close clusters included the key articles on AI-based medical image analysis, suggesting that clusters specific to the technology were appropriately formed. Significant research progress could be detected as a quick increase in constituent papers and the number of citations of hub papers in the cluster. Then we tracked recent research trends by re-analyzing "young" clusters based on the average publication year of the constituent papers of each cluster. The latest topics in AI-based medical technologies include electrocardiograms and electroencephalograms (ECG/EEG), human activity recognition, natural language processing of clinical records, and drug discovery. We could detect rapid increase in research activity of AI-based ECG/EEG a few years prior to the issuance of the draft guidance by US-FDA. Our study showed that a citation network analysis and text mining of scientific papers can be a useful objective tool for horizon scanning of rapidly developing AI-based medical technologies. # Introduction The application of innovative technologies to the development of medical products is expected as a potential new treatment or diagnostic tool for various diseases. Conversely, in some cases, the application of conventional development and evaluation concepts and/or regulatory frameworks to innovative technologies is inappropriate. In some cases in the past, a guidance document was issued when the clinical development is about to begin, the development of companion diagnostics in Japan, the U.S., and the EU. However, the guidance must be provided earlier, such as before starting clinical development planning. Therefore, the early identification of innovative technologies with a potential application to medical products through horizon scanning would encourage regulatory authorities to establish new approaches to assess their quality, efficacy, and safety to advice developers and revise their regulations if necessary. Doing so contributes to timely patient access and improve the benefit/risk ratio of the product [1]. The International Coalition of Medicines Regulatory Authorities (ICMRA), consisting of regulatory authorities, has recognized the need to respond quickly to innovative technologies and promotes the use of "horizon scanning" to identify such technologies. The ICMRA Innovation concept notedescribes horizon scanning as a broad-reaching information-gathering monitoring activity to anticipate emerging products and technologies and potentially disruptive research avenues. Two major methods exist for acquiring the data needed to create high-quality horizon scanning [bib_ref] Science and technology roadmaps, Kostoff [/bib_ref]. The expert-based approach mainly uses the tacit knowledge of domain experts, such as the Delphi method. Traditionally, horizon scanning has been conducted predominantly in Europe for policy making, scientific research funding, and health-care budgeting purposes, by surveying a variety of sources such as the Internet, government, international organizations and companies, databases, and journals [bib_ref] Scanning the horizon: a systematic literature review of methodologies, Hines [/bib_ref]. This type of expert-based approach is very difficult to implement in the current information explosion. Moreover, individual experts must subdivide their domain of expertise to keep up with the growth of their respective domains, which makes their perception of the big picture extremely subjective [bib_ref] Creating an academic landscape of sustainability science: an analysis of the citation..., Kajikawa [/bib_ref]. Computer-based approaches collect and analyze vast amounts of formal knowledge, such as articles, patents, and newspapers. Recently, the European Commission (EC) published reports, such as "Weak signals in Science and Technologies 2019 Report" based on Tools for Innovation Monitoring (TIM)that use text mining and keywords in the scientific literature. The Japanese National Institute of Science and Technology Policy (NISTEP) also uses a digital tool to analyze academic papers; the top 1% of citations contributes to science and innovation policy planning. These cover the medical field as a sub-survey of the overall science survey and are used in efforts to identify and evaluate advanced technologies. Hines et al. reported that, in the medical and health-care field, most horizon-scanning methods used manual or semiautomated, with relatively few automated aspects, which may be resolved in the not-too-distant future via the rapidly evolving fields of machine learning and artificial intelligence [bib_ref] Scanning the horizon: a systematic literature review of methodologies, Hines [/bib_ref]. To solve this challenge, a computer-based approach can complement the expert-based approach as it fits the scale of the information [bib_ref] Visualizing knowledge domains, Börner [/bib_ref] [bib_ref] Mapping the backbone of science, Boyack [/bib_ref] because they are compatible with the scale of the information. The two types of computer-based approaches are citation mining and text mining. The citation-based approach assumes that the cited papers and their research topics are similar. Analyzing this citation network allows us to understand the structure of the research areas constituting the large volume of papers that we can read. These methods have been widely used as powerful tools to visualize and understand the structure of a research field and to identify new trends and research directions; they also have been proven effective in various studies [bib_ref] Visualising semantic spaces and author co-citation networks in digital libraries, Chen [/bib_ref] [bib_ref] Visualizing and tracking the growth of competing paradigms: two case studies, Chen [/bib_ref] [bib_ref] Visualizing science by citation mapping, Small [/bib_ref]. For example, Kajikawa et al. [bib_ref] Creating an academic landscape of sustainability science: an analysis of the citation..., Kajikawa [/bib_ref] used citation network analysis to track emerging research areas in the field of sustainable science effectively and efficiently. Many fields have applied similar approaches, including energy research [bib_ref] Tracking emerging technologies in energy research: toward a roadmap for sustainable energy, Kajikawa [/bib_ref] , regenerative medicine, robotics, and gerontology [bib_ref] Finding linkage between technology and social issue: a literature based discovery approach, Ittipanuvat [/bib_ref]. Sakata et al. [bib_ref] Identifying knowledge structure of patent and innovation research, Sakata [/bib_ref] proposed a meta-structure of academic knowledge on patent and innovation research to effectively assist policy discussions on intellectual property system reform. They have shown that network analysis and machine learning methods are useful for understanding and predicting the development of technologies such as solar cells [bib_ref] Identifying emerging research related to solar cells field using a machine learning..., Sasaki [/bib_ref] and nanocarbons [bib_ref] Emerging scientific field detection using citation networks and topic models: a case..., Sasaki [/bib_ref]. Many fields have used also text mining to analyze technology trends; analyzed multi-word phrase frequencies and phrase proximity to extract energyrelated taxonomic structures. Another study discussed the trend in the field of information security by creating a network of co-occurring words and focusing on clusters with network centralities. [bib_ref] Trends in research foci in life science fields over the last 30..., Ohniwa [/bib_ref] focused on the MeSH terms included in the top 5% of the increase rate in a given year in the field of life science [bib_ref] Trends in research foci in life science fields over the last 30..., Ohniwa [/bib_ref]. A study to discuss a community's the future prospects by calculating the cosine similarity of terms in the session content from the data of conference proceedings focused on the field related to the World Wide Web [bib_ref] Identifying the evolutionary process of emerging technologies: a chronological network analysis of..., Furukawa [/bib_ref]. R&D strategists and policymakers in many fields find citation network analysis and text mining useful to understand the broad scope of scientific and technological research. It is difficult to understand the semantics of clusters based on citation relations alone. Text mining can reveal subject relationships across citations and provide insights into the diffusion of knowledge into interdisciplinary research and development. The addition of text mining to citation-based bibliometrics makes accessible the large-scale multigenerational citation studies necessary to display the full impact of research [bib_ref] Citation mining: integrating text mining and bibliometrics for research user profiling, Kostoff [/bib_ref]. Text mining is extremely sensitive to certain terms. When only text mining is used, the problem of terminological distortions cannot be ignored. In addition, it is difficult to separate homonyms that are used in different fields with different meanings. [bib_ref] A bibliometric analysis of text mining in medical research, Hao [/bib_ref] attempted to identify research fronts using only text mining in the medical field [bib_ref] A bibliometric analysis of text mining in medical research, Hao [/bib_ref]. They highlighted the challenges of clustering by text similarity, which makes the results vulnerable to the method selection. At the same time, they observed that citation relationships are highly valuable in explaining relationships in scientific knowledge. Therefore, there are challenges in analyzing trends using only one citation network and text mining. The associations between papers in citation networks reflect authors' background knowledge which cannot be extracted by simple text mining. Our study proposes an objective methodology for horizon scanning that identifies innovative technologies to be applied to medical products from entire research papers in the target field using citation network analysis methods and text mining. The three types of citation network analysis are direct citation, bibliographic merging, and co-citation. Existing studies have shown that direct citation is the most appropriate for obtaining leading-edge information on trends [bib_ref] Co-citation analysis, bibliographic coupling, and direct citation: which citation approach represents the..., Boyack [/bib_ref]. Other fields have widely used the approach of clustering the subject area into subcategories by direct citation networks and interpreting the contents of the clusters by text mining [bib_ref] Creating an academic landscape of sustainability science: an analysis of the citation..., Kajikawa [/bib_ref] [bib_ref] Tracking emerging technologies in energy research: toward a roadmap for sustainable energy, Kajikawa [/bib_ref] [bib_ref] Finding linkage between technology and social issue: a literature based discovery approach, Ittipanuvat [/bib_ref] [bib_ref] Identifying knowledge structure of patent and innovation research, Sakata [/bib_ref] , but insufficient examples exist of the application of advanced technologies in medical-related fields. We focus on AI-based medical image analysis as a retrospective example of AI-based medical devices that have been developed in recent years, applied in many fields, and selected for consideration in ICMRA [1]. # Methods ## Extraction of paper data for analysis We used "convolutional" OR "deep learning" in the review article of medical image analysis [bib_ref] Gradient-based learning applied to document recognition, Lecun [/bib_ref] ; we used "machinelearning" to include a wide range of conventional studies. As a result, we obtained 140,794 papers that contain "convolutional*" OR "machine-learning" OR "deep-learning" from the SCI (Science Citation Index) and SSCI (Social Sciences Citation Index) indexed by Web of Science Core Collection (WoS, Clarivate analytics), between January 1, 1900, and . This database has the longest history of containing bibliographic information from academic papers. It is also used for many bibliometric analyses because of its excellent searchability and comprehensiveness as a database platform [bib_ref] Creating an academic landscape of sustainability science: an analysis of the citation..., Kajikawa [/bib_ref] [bib_ref] Tracking emerging technologies in energy research: toward a roadmap for sustainable energy, Kajikawa [/bib_ref] [bib_ref] Finding linkage between technology and social issue: a literature based discovery approach, Ittipanuvat [/bib_ref] [bib_ref] Identifying knowledge structure of patent and innovation research, Sakata [/bib_ref]. In addition to the data in 1900-2020, we created datasets for , and 1900-2019 and identified the cluster that contains key articles for each year. To track the development history of AI-based medical image analysis and to select keywords for the extraction of the papers for citation network analysis, we selected 13 key articles [bib_ref] A survey on deep learning in medical image analysis, Litjens [/bib_ref] [bib_ref] Learning hierarchical features for scene labeling, Farabet [/bib_ref] [bib_ref] ImageNet classification with deep convolutional neural networks, Krizhevsky [/bib_ref] [bib_ref] Automated analysis of retinal images for detection of referable diabetic retinopathy, Abramoff [/bib_ref] [bib_ref] U-net: convolutional networks for biomedical image segmentation, Ronneberger [/bib_ref] [bib_ref] Deep learning, Lecun [/bib_ref] [bib_ref] Deep residual learning for image recognition, He [/bib_ref] [bib_ref] Pulmonary nodule detection in CT images: false positive reduction using multi-view convolutional..., Setio [/bib_ref] [bib_ref] Improved automated detection of diabetic retinopathy on a publicly available dataset through..., Abramoff [/bib_ref] [bib_ref] Dermatologist-level classification of skin cancer with deep neural networks, Esteva [/bib_ref] [bib_ref] Validation of automated screening for referable diabetic retinopathy with the IDx-DR device..., Van Der Heijden [/bib_ref] presents eight articles included in the analysis data), including several papers cited in the review article [bib_ref] A survey on deep learning in medical image analysis, Litjens [/bib_ref] on the application of deep learning in medical image analysis and a study [bib_ref] Validation of automated screening for referable diabetic retinopathy with the IDx-DR device..., Van Der Heijden [/bib_ref] that led to the clinical development of IDx-DR, a retinal imaging software approved as a medical device by the US Food and Drug Administration (FDA) in 2018. # Citation network analysis In this study, we converted the citation network into an unweighted network with papers as nodes and citation relationships as links. Papers with no citations as the largest component were considered digressional and were ignored in this study (Step 2 in . The core paper with the highest number of citations appears at the center of the citation relations. Papers with no citation relationships with other papers were considered deviant and ignored in this study. The network was then divided into several clusters using the topological clustering method. Topological clustering is a clustering method based on the graph structure of a network; here, we use modularity Key articles and the clusters in which they are contained The key articles that have contributed to the development of AI-based medical image analysis were selected based on a review article on AIbased medical image analysis. The clusters obtained from the citation network analysis of these articles are indicated. The clusters are numbered in descending order of the number of constituent papers included. The cells for papers not included in the analysis were shadowed. 8 articles are listed, excluding the 5 articles maximization. A cluster module in a citation network is a group of papers in which the citation relations are divided by using a modularity (Q value) maximization method and are densely aggregated (Louvain method) [bib_ref] Emerging scientific field detection using citation networks and topic models: a case..., Sasaki [/bib_ref] [bib_ref] Fast unfolding of communities in large networks, Blondel [/bib_ref]. The modularity maximization method appreciates network partitioning so that the intracluster is dense and the intercluster is sparse. The modularity maximization method determines an optimal partitioning pattern by extracting the partitioning pattern that maximizes the modularity using a greedy algorithm. Q is an evaluation function of the degree of coupling within a cluster and between clusters, as follows: where A ij represents the weight of the edge between i and j , k i = ∑ j A ij is the sum of the weights of the edges attached to vertex i, c i is the community to which vertex i is assigned, δ-function δ(u, v) is 1 if u = v and 0 otherwise, and m = 1 2 ∑ ij A ij . The clusters are assigned labels corresponding to the size of the number of papers included. The characteristics of each cluster were confirmed by extracting a summary of frequently cited academic papers in the cluster and the characteristic keywords in the cluster. Moreover, we computed the term frequency-inverse cluster frequency (TF-ICF) to extract the characteristic keywords of each cluster. The TF gives a measure of the importance of a term in a particular sentence, whereas the ICF provides [formula] Q = 1 2m ∑ i,j A ij − k i k j 2m c i , c j , [/formula] a measure of the general importance of a term. The TF-ICF of a given term i in a given cluster j is given by where N is the total number of sentences. Each cluster was labeled based on the resulting keywords and sentences. To confirm the trends in the research field, we extracted the mean or median year of publication of papers in each cluster, as well as information on journals, authors, and affiliated institutions. After clustering the network, visualization is converted to intuitively infer relationships among these clusters. We used a large graph layout (LGL) based on a force-direct layout algorithm [bib_ref] LGL: creating a map of protein function with an algorithm for visualizing..., Adai [/bib_ref] [bib_ref] Academic landscape of hydropower: citation-analysis-based method and its application, Sasaki [/bib_ref]. This layout can display the largest connected component of the network to generate coordinates for nodes in two dimensions. We visualize the citation network by expressing inter-cluster links with the same color (Step 4 in . However, the position of the clusters and the distance between clusters did not indicate an approximation of the content. shows an overview of this process. For the extracted dataset, we converted the citation network into an unweighted network with papers as nodes and citation relationships as links (Step 2). The network was then divided into several clusters using the topological clustering method (Step 3). Moreover, a LGL, based on a force-direct layout algorithm, displayed the largest connected component of the network to generate coordinates for the nodes in two dimensions, visualizing the citation network by expressing inter-cluster links with the same color (Step 4). Step4: Visualization Those groups of papers (Clusters) are mapped into an Academic Landscape, which helps visualize the relationship of technologies. Steps of clustering and making Academic Landscape based on citation network. This figure has been published in reference [bib_ref] Mapping the backbone of science, Boyack [/bib_ref]. The procedure of the citation network is as follows: (1) Extract the dataset of academic papers for analysis. (2) To extract the data, convert the citation network into an unweighted network with papers as nodes and citation relationships as links. (3) Divide the network into several clusters by using the topological clustering method. (4) Use a large graph layout (LGL), based on a force-direct layout algorithm, to display the largest connected component of the network to generate coordinates for the nodes in two dimensions and to visualize the citation network by expressing inter-cluster links with the same color. [formula] TF-ICF = tf i,j ⋅ icf i = tf i,j ⋅ log(N∕cf i ), [/formula] # Results ## Results of citation network analysis We analyzed 140,794 papers and found that 119,553 (85%) formed a citation network. We divided this network into 36 clusters by extracting the largest linkage component from all linkage components via direct citation of papers (excluding the gray linkage not involved in cluster formation shown in. The contents of the top 10 clusters, which contain approximately 75% of the papers in a citation network, were estimated from the characteristic keywords appearing in each cluster and the titles and abstracts of the papers with the highest number of citations. The cluster numbers (number of papers) and their contents are as follows: Cluster 1 (14,033): Basic studies on deep learning and convolutional neural networks (CNNs), including geographic information system (GIS) image analysis using remote sensing. , with the circle sizes representing the approximate number of citations in the cluster for each paper Cluster 10 (4333): Classification of individuals based on the analysis of text information from social media, such as emotions and behavior. presents the clusters in which key articles were included. Three papers (labeled A, B, and C) based on image recognition were found in clusters 1 and 5 (labeled D, E, F, G, and H) on image diagnosis in cluster 3, including the review article "Deep Learning" [bib_ref] Deep learning, Lecun [/bib_ref] (labeled D), which is often cited in medical field papers. This indicates that we appropriately formed clusters related to medical imaging in cluster 3. ## Tracking the time series of key articles We analyzed papers published each year and identified the cluster containing the key papers in and the number of citations within the cluster to assess the position of the research on medical imaging in the past. As shown in, all the papers were included in the same cluster until 2015 and the rank of cluster number increased by one until 2014. In 2015, the number of papers in this field increased rapidly and the rank of cluster numbers rose from 13th in 2014 to 6th, suggesting that great scientific attention has increased. In 2016, a key paper on the imaging diagnosis of diabetic retinopathy (F in was in cluster 7, which comprised papers on medical image analysis, and the other seven key articles were in cluster 3. Subsequently, in 2017, cluster 1 contained all the key articles, but from 2018 onward, a new separate cluster containing papers on image analysis using deep learning was formed. It should be noted that the number of citations of key articles also increased. Thus, most key articles were in one or two clusters, suggesting that we properly formed the clusters related to the targeted AI-based medical image analysis. The research status of the clusters can also be confirmed by the cluster numbers, which reflect the number of papers comprising the cluster and the number of citations of the key articles. ## Recent research trends in ai-based medical products To detect the latest research trends in AI-based medical products, we focused on "younger" clusters with an average publication year later than 2017 as research progress could be observed over three years for AI-based medical image analysis [fig_ref] Figure 3: Tracking clusters related to ECG and EEG [/fig_ref]. We re-analyzed clusters 3, 15, 12, 5, 13, and 2, which we considered to be closely related to AIbased medical technologies. We listed these clusters in order of average publication year. [fig_ref] Table 2: Sub-clustering results for clusters of AI-based medical technologies [/fig_ref] lists the sub-clusters formed by re-analysis of the most cited articles (hub-paper) [bib_ref] Deep learning, Lecun [/bib_ref] in each subcluster, suggesting recent research trends in this field as follows: Cluster 3 Applied research in medical image analysis. Cluster 15 Electrocardiogram, electroencephalogram, and other electrical biosignals of human activity. Cluster 12 Human activity recognition. Cluster 5 Natural language processing of clinical records. Cluster 13: Neuroimaging analysis. The clusters of AI-medical technologies were re-analyzed and the characteristics of the top five sub-clusters, that is, the number and average of publications of constituent papers, specific keywords, and the title of hub paper are shown Predicting protein structural classes for low-similarity sequences by evaluating different featuresCluster 2 Drug discovery with machine learning related to proteins, peptides, etc. Among these AI-based medical technologies, EEG analysis was identified for applications in epileptic seizure prediction, emotional analysis, and brain-computer interfaces, for which the FDA issued draft guidance on non-clinical and clinical trials in 2019. Electrocardiograms (ECGs) and electroencephalograms (EEGs) in cluster 15 are most likely to be applied to new medical devices; therefore, we tried to follow the cluster containing a key paper on the application of deep learning to EEG analysis [bib_ref] Convolutional neural networks for P300 detection with application to brain-computer interfaces, Cecotti [/bib_ref] , which was one of the triggers for the development of this field. During 2015-2016, the article was included in the same cluster as other neuroimaging techniques, such as MRI (MEG, fNIRS, etc.). In 2017, the key article was found in a separate cluster numbered 20 from other neuroimaging techniques, suggesting that a new cluster specific to the application of deep learning to EEG was formed. Then, in 2018, we included the article in cluster 1 of the applications of deep learning in various fields but was included in specific clusters re-formed, numbered 14 and 15 in 2019 and 2020, respectively; the number of citations of the article increased. This suggests that research in this field has developed rapidly since 2017. # Discussion In this study, we examined the possibility of using this analysis method for horizon-scanning targeting AI-based medical image analysis. IDx-DR, an image-analysis software for the automatic diagnosis of diabetic retinopathy, received FDA certification in 2018. The AI characteristics are self-learning, the algorithm for learning data during the development of a medical product is in a black box, and performance changes as the product continues learning during clinical use. This has become an interesting dilemma for regulators [bib_ref] FDA backs clinician-free AI imaging diagnostic tools, Ratner [/bib_ref]. We assessed the feasibility of using citation network analysis and text mining to identify trend history in AI-based medical image analysis research and development as follows: Research on convolutional neural networks (CNNs), the current leading technology in deep learning that arose in the 1970s, renewed interest in neural networks was Werbos's multi-layer networks. LeNet [bib_ref] Deep convolutional and LSTM recurrent neural networks for multimodal wearable activity recognition, Ordonez [/bib_ref] , a CNN-based handwritten number recognition system-was developed and succeeded by a CNN called AlexNet [bib_ref] ImageNet classification with deep convolutional neural networks, Krizhevsky [/bib_ref] , which is a key trigger for renewed interest in neural networks. Later, the U-net [bib_ref] U-net: convolutional networks for biomedical image segmentation, Ronneberger [/bib_ref] architecture was proposed, which consists of an upsampling section that uses "up" convolution to increase the image size. Furthermore, the combination of CNNs and recurrent neural networks (RNNs), represented by long short-term memory (LSTM), has been applied to analysis involving time-series data [bib_ref] A survey on deep learning in medical image analysis, Litjens [/bib_ref] [bib_ref] Deep convolutional and LSTM recurrent neural networks for multimodal wearable activity recognition, Ordonez [/bib_ref]. We evaluated 13 key articles, including these milestones in the development of AI-based medical image analysis, to determine how citation network analysis can capture key articles. We identified eight articles in one or two clusters , with a concentration of the characteristic keywords of the clusters, and the titles and abstracts of the articles with the highest number of citations confirmed that the clusters were related to AI-based medical image analysis and that identifying actual research trends was possible. Moreover, we analyzed the papers reported each year and found that the number of constituent papers of the cluster containing the key articles increased dramatically after 2014, with the rank rising from 13 to 6th, suggesting that the technology related to diagnostic imaging has progressed dramatically. This might have led to a major clinical trial of IDx-DR in 2017. Since then, research activity has increased in this field, as can be seen from the rank of cluster numbers and number of citations in the key articles. We did not include five of the 13 selected articles in the analysis: three papers were not included in the WoS and the other two [bib_ref] Automated analysis of retinal images for detection of referable diabetic retinopathy, Abramoff [/bib_ref] [bib_ref] Validation of automated screening for referable diabetic retinopathy with the IDx-DR device..., Van Der Heijden [/bib_ref] on clinical evaluation were not found with the set query, because there was no mention of the underlying technology in the abstract or title, and the methods were mainly described as product names or computer detection in either paper. Next, we explored trends in the development of new medical products using AI by re-analyzing "young" clusters with a late average of the publication year of constituent papers to identify more specific topics by sub-clustering [fig_ref] Table 2: Sub-clustering results for clusters of AI-based medical technologies [/fig_ref]. This allowed us to objectively look at the landscape of AIbased medical technology. We focused on EEG and ECG, which have the potential to lead to the development of new medical devices, and followed the cluster containing the key article on this topic. As shown in [fig_ref] Figure 3: Tracking clusters related to ECG and EEG [/fig_ref] , the increase in constituent papers and citations of key articles suggested that this topic developed significantly between 2017 and 2018, a couple of years before the FDA issued a guidance draft on brain-computer interfaces in 2019, which was finalized in 2021. Regarding the FDA's activity, a public workshop was held on , to promote open discussion of scientific and clinical considerations related to the development of BCI devices, suggesting that the FDA might consult public on product development. The ECG is already at the stage of realization in smartwatches and other devices and was judged to be of low novelty. The FDA has already approved the app for the Apple Watch®. This study also showed that analysis every several months might allow us to identify the candidate topics for further investigation through the rapid rise of the rank of cluster number, i.e., a sharp increase in constituent papers (2014-2015 inand 2017-2018 in [fig_ref] Figure 3: Tracking clusters related to ECG and EEG [/fig_ref] , or the emergence of a new cluster spun out of the original one (2017-2018 inand 2016-2017 in [fig_ref] Figure 3: Tracking clusters related to ECG and EEG [/fig_ref] , which may be a signal of significant research progress. This analysis has the following limitations. Which would be detected by this method as well. Therefore, it is necessary to determine whether the candidate topic is a good idea or We included papers in major journals in WoS relatively quickly after publication, but there might be a delay of approximately six months for almost all journals and some research areas may not be reflected in WoS sufficiently quickly, which may delay the identification of research trends. Until the birth of Scopus and Google Scholar in 2004, WoS was the only tool for citation analysis [bib_ref] The journal coverage of web of science and scopus: a comparative analysis, Mongeon [/bib_ref]. Even today, WoS is known to have a longer record period than Scopus and is one of the most effective databases in the field of history. In addition to WoS, Scopus and PubMed have also become powerful databases, and future studies are needed to evaluate the robustness of those databases. Although this paper does not show these data, we also analyzed the papers obtained from PubMed; however, approximately 30% of the papers formed a citation network and only five of the 13 key articles were included. One possible reason for not being able to extract appropriate research papers from PubMed was that many papers did not use terminology related to AI-based technologies. This suggests that the choice of the literature database according to the target technology is also critical. Furthermore, research results in the field of machine learning, which covers basic technologies in the field of AI and other informatics fields, tends to be published as proceedings of international conference or arXiv.com as preprints than peer-reviewed journals, where researchers can directly exchange papers with each other via the Internet; therefore, the latest results cannot be covered by databases of academic papers, such as WoS or PubMed. A comparison of peer-reviewed journal-based analysis and proceeding-or preprints-based analysis-needs to be conducted in the future. Experts who have a deep understanding of innovative technologies would be able to predict the development of medical products based on the technology. However, it might sometimes be inappropriate to narrow the scope of consideration based solely on experts' opinions [bib_ref] A critical review of the Delphi technique, Beretta [/bib_ref]. Extracting a limited number of novel topics that may affect pharmaceutical regulations from a vast amount of information on a human basis is difficult and using a computer-based method (such as this study) is reasonable and appropriate. This study assumes that the ultimate users are regulators who evaluate technologies in the mid to long term. Because policymakers and decision makers are not always experts in their fields, providing the status of the academic field in a systematic method supports decision making that can be reproduced by anyone. In our study, we used citation network analysis and text mining to classify the entire papers in the target field in terms of research topic. Furthermore, we identified the topics of the clusters based on the characteristic cluster keywords and titles of the most cited papers. We objectively evaluated the popularity and novelty of a topic based on the number of papers and the median year of publication. We consider that the feature of the method is suitable for a primary screening by regulators to pick up candidate topics from wide range of scientific fields, and the topics would be further evaluated based on the opinion of experts of the topic and other sources such as patents. We considered that limiting the search to papers in the clinical development stage was rather inappropriate because the purpose of horizon scanning is to detect technologies that have the potential to reach clinical development in the pre-clinical stage. When searching for papers on clinical development, papers on related technologies in the earlier stages are less likely to not include in the analysis, which involves the risks that do not reflect the overall picture of the field. The overall landscape of R&D can be grasped more objectively by analyzing a wide range of papers, for example and then target cluster, the cluster on clinical development is obtained by clustering and re-clustering. Information on the clinical development stage can be directly and timely obtained from clinical trial registries such as ClinicalTrials. gov. The information provided by these other tools from analysis such as "Tools for Innovation Monitoring (TIM)" is useful for determining the query for the papers data in our method. Another possible bias, as mentioned [bib_ref] Trends and typology of emerging antenna propagation technologies: citation network analysis, Takano [/bib_ref] , is that researchers mainly check and cite papers written in their native language or journals they contribute to, or that they tend to search and cite papers using the same terminology and not others, even when the technological meaning is the same. Considering the opinions of experts in the field regarding candidate topics to be investigated will help in overcoming the aforementioned limitations. Our method provides information about the median and average year of publication of the papers in the cluster and the newness of the Hub paper, but prioritization requires the perspective of an expert in the field. Academic size and speed of discussion do not necessarily determine prioritization; however, depending on the social demands and feasibility of the technologies included in the individual topics. Hence, a content-based evaluation is necessary. # Conclusion This study showed that citation network analysis and text mining for bibliographic information analysis of the rapidly developing field of AI-based medicine can be used for horizon scanning for medical products that require new assessment approaches. We detected recent research 1 3 developments, including AI-based ECG/EEG. We suggest that this method be used as a primary screening tool for horizon scanning, and that the analysis results be used more effectively and appropriately by incorporating the opinions of experts. [fig] Cluster 2: (13,309): Drug discovery technologies related to proteins, peptides, etc., using machine learning.Cluster 3(10,992): Applied research in medical image analysis.Cluster 4 (9867): Feature classification using ensemble methods to increase accuracy by combination.Cluster 5 (7829): Natural language processing of clinical records.Cluster 6 (7412): Application of deep learning to fault diagnosis, for example, motor condition monitoring for machines running on electric motors.Cluster 7 (6571): Machine learning (ML) and data mining (DM) methods for cyber analysis.Cluster 8 (5815): Application to traffic flow information analysis for the implementation of intelligent transport systems.Cluster 9 (4371): Single-image super-resolution (SR) to reconstruct high-quality data. [/fig] [fig] Figure 2: Tracking clusters containing key articles. We analyzed papers obtained from WoS published up to the indicated years. We plotted the cluster numbers that contained the eight key articles shown in [/fig] [fig] Figure 3: Tracking clusters related to ECG and EEG. We analyzed papers obtained from WoS published up to the indicated years. A cluster number indicates the cluster on ECG and EEG. The circle sizes indicate the approximate citation frequency of the key article,[73] and the number in each circle represents the number of citations in the cluster. Clusters on ECG and EGG were first detected in 2015 as cluster number 10 and were classified into cluster numbers 11, 21, 1, 15, and 15 for 2016, 2017, 2018, 2019, and 2020, respectively [/fig] [table] Table 2: Sub-clustering results for clusters of AI-based medical technologies [/table]

Risk Modifying Factors of Anxiety and Depressive Disorders, Using the Example of a Population Study in the Żywiec District Citation: Lubecka, B.; Lubecki, M.; Kasperczyk, J.; Jośko-Ochojska, J.; Pudlo, R. Risk Modifying Factors of Anxiety and Depressive Disorders, Using the Example of a Population Study in theŻywiec District. Int. J. # Introduction Mental disorders' prevalence rates have increased significantly in recent years. The most common ones are anxiety disorders and depression. According to WHO (World Mental Health) data, depression accounts for over 4% of the global burden of all diseases and is the second most common cause of disability after cardiovascular disease. Anxiety disorders can be considered as a separate nosological unit or component of the depressive episode. These disorders are often comorbid, as anxiety is a risk factor for depression and vice versa. The most up-to-date data on the prevalence of these disorders in Poland are described in the EZOP Poland (Epidemiology of Psychiatric Disorders and Availability of Psychiatric Healthcare) research conducted in 2012. It showed that 23.4% of the respondents had been diagnosed with at least one mental disorder during their lifetime, of which 9.6% had symptoms of anxiety disorders and 3.4% met the criteria for an episode of major depression. These data only include the adult population below 65 years of age. The European-wide ESEMeD study (The European Study of the Epidemiology of Mental Disorders), published in 2004, showed that major depression and specific phobias were among the most common mental disorders, which occurred with a frequency of 13% and 8%, respectively, during the lifetime of the respondents. Nowadays, more and more importance is committed to the correct diagnosis and treatment of the described disorders. The current standards assume that thanks to properly conducted preventive measures, many new cases can be avoided. Prevention programs carried out in Poland so far only cover people belonging to high-risk groups, such as women after childbirth, adolescents and people over 65, which from the point of view of the increasing frequency of anxiety disorders and depression in the general population seems to be insufficient. To create effective prevention programs, it is necessary to conduct many population studies, the effect of which will be to identify new and verify already studied risk factors for the occurrence of anxiety disorders and depression in individual populations. Apart from the aforementioned EZOP study, the population studies conducted in Poland so far are based on the analyses of less numerous groups selected for specific risk factors, such as school students, students, patients with another mentalor somatic diseaseand the elderly. This publication presents the analysis of the correlation between individual risk modifying factors and the occurrence of anxiety and depressive disorders in the adult population of theŻywiec district. The modifying factors were grouped in terms of increasing the risk of developing the disease or having a protective effect, which made it possible to distinguish other groups requiring special preventive measures. # Materials and methods The study population consisted of the inhabitants of theŻywiec district. This region is inhabited by 124,616 people, of which 64,149 are women and 60,467 men. Data were collected in the city ofŻywiec, as well as in all regions of the district. Due to its location, compact geographic structure, organization of healthcare and demographic characteristics, theŻywiec district was considered optimal for obtaining reliable data for epidemiological research. In terms of socio-demographic factors, the studied group corresponded to the district population. Therefore, the collected data can be extrapolated to the Polish population. The obtained data were also used to assess the disproportion between the occurrence and treatment of the studied disorders, which was described in another publication. Data for the study were collected with anonymous questionnaires. The study group consisted of randomly selected adult inhabitants of theŻywiec district who voluntarily and independently completed the questionnaires. The inclusion criteria were: living in the district, being 18 years of age and giving informed consent to participate in the study. The exclusion criteria were as follows: uncorrected visual disturbances making it impossible to fill in the questionnaire, cognitive disorders that make filling in the questionnaire impossible and submitting an incomplete questionnaire or questionnaire containing interpretation doubts. In order to obtain a representative study group, activities were carried out in many different places (Poviat Starosty, City Hall, secondary vocational and high schools, family doctor clinics and multi-specialist clinics, among employees of hospitals and local institutions and companies). They included several ways of reaching residents: direct contact and handing over a questionnaire, making it possible to take questionnaires from prepared and marked places in public buildings, and handing over the questionnaires to people ready to distribute them among their families, friends and colleagues. The distributed questionnaires were collected in person from designated places. A research questionnaire consisting of two parts was used for the study. The first part included questions about age, marital status, place of residence, professional activity, physical activity, the use of psychoactive substances, previously diagnosed somatic and mental illnesses and medications taken. In the second part, the Hospital Scale of Anxiety and Depression was used to diagnose studied disorders, which is an inventory of self-esteem. Due to the high sensitivity and specificity in detecting generalized anxiety and depressive disorders, this scale was also used in the diagnosis of healthy adults in over 700 scale world studies, showing satisfactory psychometric properties for both the Anxiety (HADS-A) and Depression (HADS-D) subscales. In this research, the Polish version of the scale was used. The use of this scale in research is supported by the short time needed for self-completion, clearly formulated questions (regardless of gender, age and level of education) and the possibility of simultaneous testing of both disorders. The Hospital Scale of Anxiety and Depression consists of two subscales with 7 questions each-HADS-D (Hospital Scale of Anxiety and Depression-Depression Subscale) and HADS-A (Hospital Scale of Anxiety and Depression-Anxiety Subscale)-sequentially for the study of depression and anxiety disorders. The total number of points obtained corresponds to the severity of the symptoms of anxiety and depression. Values from 0 to 7 points correspond to the norm, values from 8 to 10 points correspond to the limit value and values greater than or equal to 11 points represent an episode of depression (HADS-D) or an anxiety disorder (HADS-A). The correlation of the results obtained in the scales with the risk factors listed in the questionnaire was examined. Statistical analysis was performed using the Statistica 13.0 (Statsoft, Kraków, Poland) package and included: descriptive statistics (group sizes, percentage fractions for qualitative variables, mean for quantitative variables, standard deviation) and intergroup comparisons for quantitative variables, which were made using the t-test or analysis of variance-in the event of failure to meet the test assumptions, their non-parametric equivalents were used. For qualitative variables, the following tests were used: Pearson's Chi2 and Maximum Likelihood. In the analyses, the p-value < 0.05 was considered significant. This study is part of a larger project analyzing the epidemiology of depression, anxiety and sleep disorders in the mentioned region. The first part includes the assessment of the disproportion between the incidence of anxiety and depressive disorders and the use of psychiatric healthcare by the inhabitants of the district. It has been published already in Polish science magazine "Psychiatry" vol. 17, no. 1, 1-8 in 2020. The research project was presented to the Bioethics Committee of the Medical University of Silesia, which concluded that the study did not require its consent. All respondents gave their informed consent to participate in the study. # Results A total of 5000 questionnaires were distributed, 1659 of which were fully completed. Incomplete or incorrectly completed questionnaires were rejected. There were 934 women (56.3%) and 725 men (43.7%) in the group. The age of the respondents ranged from 18 to 92 years old. The mean age was 43.7 ± 14.61 years, and the median was 44 years. Three age groups were adopted: 18-39 years old, 40-59 years old and from 60 years old, representing 38.5%, 47% and 14.5% of the entire sample, respectively. The largest group in the study were people under 60 (85.5%). The subjects who were in a stable relationship prevailed (69.2%). Unmarried persons accounted for 30.8%. Due to the district's demographic structure, inhabitants of communes predominated in the study (69.4%). Most respondents had secondary or vocational education (62%), and the most common form of employment was full-time work (62.9%). The characteristics of the study group in terms of selected risk modifying factors are presented in. Physical activity was confirmed by 75.1% of the respondents, of which 6.9% performed it daily. Most of the respondents used alcohol sporadically or not at all. The use of stimulants was declared by: 19.8% who smoked cigarettes, 3.7% marijuana and derivatives and 1.3% psychostimulants. A total of 33.9% of the respondents confirmed the coexistence of somatic diseases, of which 6.5% were without regular treatment. Among the respondents, as many as 23.4% confirmed the use of over-the-counter calming supplements/medications and hypnotics, and 8.7% were treated for insomnia. In the study population, on the basis of the HADS results, depressive disorders were found in 14.4% of the respondents, and the criteria for anxiety disorders were met by 11.2% of the participants. In the study population, on the basis of the HADS results, depressive disorders were found in 14.4% of the respondents, and the criteria for anxiety disorders were met by 11.2% of the participants. The examined risk factors were divided into groups. The first group included sociodemographic variables such as sex, age, marital status and place of residence. Detailed results obtained in the scale study are presented below. To present the results, the study group was divided into subgroups depending on the result on the HADS scale. On the depression subscale (HADS-D), the sum of the points obtained by respondents from 0 to 7 points corresponds to the norm (HADS-D1), from 8 to 10 is the limit value of depression (HADS-D2) and a sum greater than or equal to 11 points represents an episode of depression (HADS-D3). On the anxiety subscales (HADS-A), the sum of points obtained during the study from 0 to 7 testifies to the norm (HADS-A1), a sum from 8 to 10 points is the limit value of an anxiety disorder (HADS-A2) and a sum greater than or equal to 11 points represents an anxiety disorder (HADS-A3). The examined risk factors were divided into groups. The first group included sociodemographic variables such as sex, age, marital status and place of residence. Detailed results obtained in the scale study are presented below. To present the results, the study group was divided into subgroups depending on the result on the HADS scale. On the depression subscale (HADS-D), the sum of the points obtained by respondents from 0 to 7 points corresponds to the norm (HADS-D1), from 8 to 10 is the limit value of depression (HADS-D2) and a sum greater than or equal to 11 points represents an episode of depression (HADS-D3). On the anxiety subscales (HADS-A), the sum of points obtained during the study from 0 to 7 testifies to the norm (HADS-A1), a sum from 8 to 10 points is the limit value of an anxiety disorder (HADS-A2) and a sum greater than or equal to 11 points represents an anxiety disorder (HADS-A3). Modifying factors, the relationship of which with the occurrence of anxiety and depressive disorders was statistically significant, are presented in the diagrams. The prevalence of anxiety disorders in women and men was similar. Women met the criteria for depressive disorders significantly more often than men. In the group of respondents over 60 years of age, depressive disorders occurred much more often and were found in 34.9% of people compared to those in the age group 18-39 years old (5.7%) and 40-59 years old (3.3%). Anxiety disorders also slightly prevailed in this age group, but this was not a statistically significant difference. Marital status and place of residence did not significantly affect the occurrence of anxiety and depression. Data sorted by the percentage of the studied disorders depending on the individual subgroups of modifying factors, the relationship of which was statistically significant, are presented in. The prevalence of anxiety disorders in women and men was similar. Women met the criteria for depressive disorders significantly more often than men. In the group of respondents over 60 years of age, depressive disorders occurred much more often and were found in 34.9% of people compared to those in the age group 18-39 years old (5.7%) and The prevalence of anxiety disorders in women and men was similar. Women met the criteria for depressive disorders significantly more often than men. In the group of respondents over 60 years of age, depressive disorders occurred much more often and were found in 34.9% of people compared to those in the age group 18-39 years old (5.7%) and . Anxiety disorders also slightly prevailed in this age group, but this was not a statistically significant difference. Marital status and place of residence did not significantly affect the occurrence of anxiety and depression. Data sorted by the percentage of the studied disorders depending on the individual subgroups of modifying factors, the relationship of which was statistically significant, are presented in. Most often, depressive disorders were presented by people with primary education (15.9%) and vocational education (12.8%). Compared to them, the incidence of depression was almost three times lower in the group of respondents with a master's degree (5.4%). The level of education did not have a significant influence on the incidence of anxiety disorders. Most often, depressive disorders were presented by people with primary education (15.9%) and vocational education (12.8%). Compared to them, the incidence of depression was almost three times lower in the group of respondents with a master's degree (5.4%). The second group of analyzed factors concerned variables related to professional activity, type of work and physical activity. In the group of retired people, depressive disorders concerned over half of the respondents (51%). Compared to them, the incidence of depression in the unemployed, students or part-time workers was 4%, while in those working full-time or who are self-employed, it was 2%. The unemployed, retired and self-employed people were in the highest risk group of anxiety disorders. They occurred in this group twice as often as among full-time employees. A dozen times higher incidence of depression was observed in the respondents engaged in mental work than in the group of physical workers. The type of work was not significant for the occurrence of anxiety disorders. In the studied population, the prevalence of anxiety and depression was significantly influenced by physical activity. In people who did not take up physical activity at all, anxiety disorders and depression occurred much more often than in other groups. However, in the group of people practicing sports every day, depressive disorders occurred slightly more often than among respondents who took care of physical activity less regularly. The third group of the analyzed risk factors involved the use of psychoactive substances. The relationship between the frequency of alcohol consumption and the occurrence of anxiety disorders was confirmed, which was most often found in the group of people who drink alcohol three or more times a week (26.5%). It was observed that in the group of people who maintain abstinence, the frequency of anxiety disorders was slightly higher (18.2%) than among the sporadic drinkers. There was no relationship between the frequency of alcohol consumption and the occurrence of depressive disorders. In the group of smokers, both anxiety disorders and depression were significantly more frequent. The use of marijuana was confirmed by a small group of people, a total of 62 respondents. A total of 75% of people in this group were under the age of 29. Among them, no statistically significant difference was found in the assessment of the prevalence of anxiety and depression. The use of psychostimulants was confirmed by 22 people. Respondents using stimulants almost twice as often presented symptoms of anxiety disorders. However, their use was not associated with a higher incidence of depression. The fourth group of analyzed risk factors concerned variables related to the presence and the treatment of somatic diseases, the use of hypnotics, and thus the occurrence of insomnia, and the use of over-the-counter supplements or sedative agents. Among the subjects burdened with somatic diseases, anxiety disorders and depression were significantly more frequent than in healthy respondents. Depression was slightly more frequent (17.8%) in the group of patients receiving regular medical treatment for somatic diseases than in the group not taking regular medications (11.1%). Contrary to this observation, it was shown that somatically burdened people who did not receive regular treatment confirmed the occurrence of anxiety disorders significantly more often (34.3%) compared to people who regularly take medications (20.7%). Depression was much more common in the group of people using sleeping pills than in respondents who refused to use them. This difference, probably due to the small size of the group, was not statistically significant. Anxiety disorders occurred significantly more often in subjects who used sleeping pills than in those untreated because of insomnia. Anxiety disorders and depression were significantly more frequent among respondents using herbal preparations or dietary supplements with a sedative effect compared to the rest of the studied population. In the group of people who confirmed the daily use of the above-mentioned substances, depressive disorders occurred twice as often and anxiety disorders were almost four times more frequent than in the population of people who refused to take them. # Discussion Due to the significant social costs of anxiety and depression, it is important to identify risk factors and intervene early. Because of a small number of new national population studies on anxiety and depression in recent years, the authors planned and conducted a study of inhabitants of theŻywiec district, which was considered optimal for obtaining data for epidemiological research. Due to the planned large size of the study group, the only possible method was a questionnaire study. The scale included in the survey that was used for the analysis was the Hospital Scale of Anxiety and Depression. The use of this scale is supported by the short time needed to complete it independently, clearly formulated questions-regardless of gender, age and level of education-and the possibility of the simultaneous examination of both disorders. In the studied population, an episode of depression was found in 14.4% of the respondents, and the point criteria for anxiety disorders were met by 11.2% of the individuals. The results are higher than in the nationwide EZOP study from 2012, which revealed that 9.6% of people had symptoms of anxiety disorders, 0.4% met the criteria of minor depression and 3.4% had an episode of major depression. The HADS-D subscale does not differentiate episode severity. On the other hand, the results are consistent with the European ESEMeD studies, which showed that during the lifetime of the respondents depressive disorders occurred with a frequency of 14.0%, and anxiety disorders in 13.6%. The first risk modifying factor tested was gender. The results obtained by the authors of this study show a similar frequency of anxiety disorders in both women and men. This phenomenon may result from the progressive process of standardizing or reversing social roles. There is an increase in the importance of women in the public and professional spheres, greater consent to showing weakness by men and more frequent use of medical help by men. Similar observations can be found in some internationaland Polish studies. The study showed that depressive disorders occur twice as often in the group of women than in men, which is confirmed by other publications. In the EZOP study cited earlier, minor depression was equally common in both sexes, while major depression and anxiety disorders occurred twice as often in women as in men, and it was found that the differences in prevalence were greater in older age groups. The reason for these differences can be explained among others through hormonal changes in the group of women related to the menstrual cycle, pregnancy, postpartum and menopause. In a population study conducted in 2005-2007 in Poland, it was shown that postmenopausal women frequently deteriorate in their mental state-19.8% of them have symptoms of depressed mood, and 27% of increased irritability. The analysis of the responses of women and men did not show any significant differences between them. This result proves that the HADS scale can be used without fear that the gender of the examined person will influence the obtained answers. Similar conclusions were the result of other Polish and foreign authorsAge was the second researched risk modifying factor. Depressive disorders occurred much more often (34.9%) in the population of people over 60 than in other age groups. People between 18 and 39 years of age were in the intermediate risk range, and here depression was present at the level of 5.7%. The most numerous group and at the same time burdened with the lowest risk of these disorders were people between 40 and 59 years of age (3.3%). The most frequent occurrence of anxiety disorders was recorded among respondents over 60 years of age, but this difference was not statistically significant. Interpretations of the influence of age on the risk of depression may include both biological and psychological factors. Early and middle adulthood is an extremely intense period in terms of the necessity to make key decisions regarding the development of an individual. Increasing social pressure, constant volatility and instability in terms of place of residence, work, career and interpersonal relations are conducive to the feeling of strong stress. The simultaneous need to enter the labor market, gain financial independence, continue personal development and start a family can be overwhelming and may cause anxiety and depression to occur more frequently in this age group. The stage after the age of 40 seems to be a period of relative financial stabilization, subsidence in life roles and the stability of somatic health. Depressive disorders appear more and more frequently in the elderly with the emergence of many new psychological factors and somatic diseases. The factors contributing to depression include loneliness, a feeling of uselessness after retirement, suffering from somatic diseases and a reduction in psychophysical fitness and thus independence. It should be noted that the oldest respondent in this study was 92 years old. Several analyses of the prevalence of depressive disorders in the elderly were carried out in the country. In the PolSenior study, the prevalence of depression increased with age and ranged from 20 to 33%, and in a study of inmates of care and treatment facilities in the Bielsko district conducted in the years 2008-2010, severe depression was found in 10.9% of respondents. The presence of depressive symptoms impairs the quality of life of the elderly and often worsens the prognosis of comorbidities. It is known from the literature that the prevalence of anxiety disorders in the elderly is lower than among young people and ranges from 3.2% to 21.6% of the elderly population. It is widely accepted that anxiety disorders develop most frequently in adolescence or early adulthood. Less than 1% of anxiety disorders develop in people over 65, and 90% in people under 40. In this study, no significant correlation was observed between marital status and place of residence, understood as living in a city or rural area by the respondents, and the presence of anxiety and depression. The EZOP study shows that an important risk factor for anxiety and depressive disorders is the termination of a relationship with a close person, e.g., due to divorce or the death of a partner. However, this does not mean that staying in a stable relationship is a guarantee of mental health. The quality of the relationship, which includes the scope of mutual support, is of great importance. Some studies show that people in long-term marriages or partnerships reported anxiety about their mental health slightly more often than others. The phenomenon of the disappearance of obvious differences also applies to the lifestyle related to the place of residence. The stereotypical approach is that the inhabitants of rural areas far from the city noise, living closer to nature, eat healthier and spend more time in the fresh air. In the studies that analyzed the incidence of mental disorders among inhabitants of rural areas and cities, it was found that people living in cities obtained higher incidence rates due to anxiety disorders and depression. On the other hand, it would seem that the rural community is characterized by lower health awareness, stronger fears of stigmatization due to psychiatric treatment and less access to specialized healthcare. In one of the Polish studies, in a population of 90 patients of the hospital inŚwiecie, it was proved that a lower level of education and living in rural areas were strong risk factors for a depressive episode. The results of the research conducted by the authors of this publication support the ongoing process of rural urbanization. This consists of abandoning the rural lifestyle, taking up employment within cities and increasing health awareness and the availability of medical care. One should not forget about the more and more common phenomenon of willingly inhabiting rural areas by inhabitants of large cities. These situations more and more often lead to the similarity of rural and urban environments, also in terms of the spread of health phenomena. In this study, people with primary and vocational education presented depressive disorders three times more often than in the group with higher education. In the group of respondents with secondary education, a slightly higher incidence of depression was observed than in people with higher education. There was no significant relationship between the occurrence of anxiety disorders and the level of education. A lower level of education may be associated with insufficient knowledge about a healthy lifestyle, including the proper mechanisms of stress reduction, a healthy diet, attention to physical activity and adherence to the principles of sleep hygiene. The relationship between the higher incidence of depression in people with lower education is also confirmed by the results of studies conducted by other authors. It was observed that the protective factor for the occurrence of depressive disorders is a permanent form of employment, such as full-time work and own activity, where the incidence of depressive symptoms was found to be 2.8% and 2.5%, respectively. A slightly higher percentage of these disorders was presented by people performing odd jobs, people studying and the unemployed-between 4.0 and 4.4%. The greatest morbidity due to depression (51%) was found in the group of retirees and disability pensioners. Among the factors affecting mental health analyzed under the EZOP project and studied by the Public Opinion Research Center in 2012, one of the most important factors was employment, understood as a source of income, but also social support. People who lose their jobs have a much higher risk of developing depression. On the other hand, the appearance of a mental illness in professionally active people negatively affects their functioning at work. Due to the obtained results, one should consider such a significant difference in the incidence of depression among the unemployed and retirees and pensioners. It seems that a significant difference between the two forms of economic inactivity is the sense of irreversibility accompanying the transition to a disability or retirement pension. The retirement period, often identified with the beginning of old age, a sense of uselessness, a lack of motivation to act, inability to use free time and the deterioration of somatic health, is a strong depressant factor. In the study, the most important risk factor for anxiety disorders was unemployment (28.9%). Retirement/disability pension (25.2%) and own activity (22.5%) are in second and third place. A slightly lower percentage of anxiety symptoms affects people in education (19%), and the lowest among respondents employed full-time (13.1%) and part-time workers (12.3%). One of the elements of social security is social insurance. Research shows that unemployed people are characterized by worse health than professionally active people. Psychosomatic symptoms of anxiety, free-flowing anxiety and insomnia are more common in them. The analysis revealed a significant protective effect of physical work on the prevalence of depressive disorders. Mentally working people were several times more likely to develop symptoms of depression. The type of work did not affect the occurrence of anxiety disorders. The impact of physical work can be compared to the effect of physical activity, the significant impact of which on the occurrence of depression and anxiety disorders was noted in the analyses. The subjects who did not engage in any activity had a much higher risk of both disorders. The lowest risk was for people who were active several times a week, but not every day. The studies conducted so far show that systematic physical activity has a protective effect against the occurrence of anxiety and depressive disorders. Interestingly, this study showed that people practicing sports every day were characterized by a slightly higher percentage of depression and anxiety disorders than people who showed less regular physical activity. In the literature, you can come across the term addiction to physical exercise. It consists of regular, daily physical exercise regardless of the circumstances, also during illness, after an injury or when there are other health contraindications. Physical activity begins to dominate the addict's life, which leads to the neglect of other areas of functioning, such as work, family life or social contact. In studies of a group of people practicing sports every day, after a break in exercise lasting more than a day, symptoms of withdrawal syndrome were found, the symptoms of which were anxiety, nervousness, irritability and depressed mood, which was dictated by changes in neurotransmitters. In addition, too intense and too frequent training can be interpreted by the body as an excessive burden, generating chronic inflammation and initiating neurodegeneration. Another group of analyzed factors was related to the use of psychoactive substances. According to the results of the study, the use of alcohol over three or more times a week was associated with an increased risk of anxiety disorders (26.5%). In this study, the influence of frequent alcohol use on the occurrence of depression has not been proven. The literature shows that alcohol dependence increases the incidence of depression almost fourfold, and anxiety disorders more than twofold. Among alcohol-dependent patients, 37% suffer from other mental disorders at the same time. This relationship is two-way, namely primary mental disorders may favor excessive drinking and the development of alcohol dependence, and alcohol abuse may induce or aggravate the occurrence of mental disorders. Drinking alcohol can also be an attempt to "self-medicate" anxiety or depression symptoms. In the study group, anxiety and depressive disorders occurred significantly more often among smokers. The negative effects of cigarette smoking have been documented in many scientific studies. Additionally, the coexistence of depressive symptoms in a smoker supports the habit of reaching for a cigarette. One cohort study found that as exposure to higher doses of nicotine increased, the risk of depressive disorders increased, with women being more exposed at lower doses than men. In addition, it is emphasized that people addicted to nicotine use marijuana and derivatives more often and with greater freedom. With regard to stimulants other than tobacco and alcohol, which were included in this study, it should be emphasized that the percentage of people who confirmed their use was small and for marijuana it was 3.7% of the respondents, for stimulants 1.3%. In addition, the researchers in the questionnaire did not ask about the amount of substances and frequency of use, while being unable to assess the risk of addiction in the respondents. The use of marijuana has been confirmed mainly by young people, up to 29 years of age. There was no statistically significant difference in the assessment of the prevalence of anxiety and depression in this group. However, among the 22 people who confirmed the use of psychostimulants, the symptoms of anxiety disorders were significantly more frequent-36.4% vs. 16.3%. According to the results of the EZOP survey, 1.3% of respondents admitted to using psychoactive substances in Poland. It is known from the literature that regular marijuana use is associated with an increased risk of depression. The dependence of the occurrence of general sadness was demonstrated in 20-30% of people who smoked it. In the Polish study of users of psychostimulants, higher values of the Beck Depression Scale were found in their group. The total value of points on the scale was higher, the more psychoactive substances the respondents used. Many publications show that the use of amphetamine and methamphetamine and its derivatives may cause anxiety disorders of considerable intensity and uncontrolled impulsive behavior. The described symptoms may also occur in users of these substances during the period of abstinence. It is estimated that the frequency of anxiety disorders among users of psychostimulants may be as high as 30.2%. Among the subjects with somatic disease, anxiety disorders (23.3% vs. 13.2%) and depression (16.5% vs. 5.1%) were significantly more frequent than in healthy subjects. In this case, the relationship is bidirectional. In long-term studies, it was found that the presence of a depressive episode is associated with a much higher risk of developing somatic diseases, and vice versa, the presence of a somatic disease is a risk factor for depressive disorders, which significantly worsen the course of treatment and rehabilitation. Pain, which is a symptom of many diseases, especially at severe stages, may be associated with the onset of a depressive or anxiety disorder. In addition, it has been shown that the coexistence of anxiety disorders in patients with somatic disease increases the perceived somatic symptoms, while reducing the effectiveness of treatment. This research showed that in the group of somatic patients, depression was slightly more frequent among people receiving regular treatment than in those who did not take medications permanently (17.8% vs. 11.1%). There are many reports that the inadequate treatment of somatic diseases carries a greater risk of depressive episodes. However, the observations made by the authors of this study may suggest that patients who do not take medications regularly may not feel any discomfort due to the coexistence of comorbidities. A typical example of this phenomenon is metabolic diseases such as arterial hypertension, diabetes, hyperlipidemia or gout, which at the beginning may not cause significant suffering to patients, and thus the incidence of depressive disorders in this group will also be lower. The feeling of being healthy strengthens the tendency to displace or downplay the disease, and it is known from the available analyses that compliance by patients leaves much to be desired. Somatically burdened people who did not receive regular treatment confirmed the occurrence of anxiety disorders significantly more often compared to those who regularly took medication-34.3% vs. 20.7%. In the analysis of patients' attitudes towards the diagnosis and treatment of the disease, there is a view that a significant reduction in the fear of finding oneself in the role of a sick person is achieved by an active attitude consisting of searching for information and actually relating it to one's own situation, as well as focusing on controlling the state of affairs and overcoming the disease. In the group of people using sleeping pills every day, depression was more than twice as common as in those untreated for insomnia. This difference, probably due to the small size of the group of people taking hypnotics, was not statistically significant. Analyzing the frequency of anxiety disorders, it was much higher in people taking hypnotics in general (43.8-55%) compared to the group of people not receiving treatment (13.8%). The literature shows that the most common form of sleep disturbance is insomnia. Its prevalence, depending on the research method adopted, ranges from 4.4% to 48%. It is known from Polish studies that the prevalence of insomnia is estimated at 50.5% . Insomnia is three times more common in people diagnosed with depressive disorders. As a rule, it is a symptom of the current episode. Additionally, sleep disorders increase the risk of the occurrence, recurrence and more severe course of depressive disordersand may be a risk factor for anxiety disorders. The frequency of use of over-the-counter medications or sedative supplements in the study population was directly proportional to the frequency of anxiety and depressive disorders. In the group of people using the above-mentioned preparations every day, anxiety disorders occurred almost four times as often and depressive disorders more than twice as often as in people who refused to use them. Every year, more and more new over-the-counter preparations are created. They are often the first link in the selfhealing process of the studied disorders. For many people, supplements, including herbal drugs, seem to be more natural and less harmful than synthetic psychotropic drugs. Usually, they are available without a prescription and do not require a visit to a specialist doctor, which theoretically is to protect users from stigma. Research conducted among people with anxiety and mood disorders has shown that nearly half of them use so-called complementary medicine. Their use can often prolong the poor functioning of patients and delay the initiation of adequate treatment based on the principles of EBM (Evidence-Based Medicine). This publication is an example of a population study, the results of which enable the identification of new risk factors for the described disorders and the planning of changes in the scope of health policy, including targeted interventions. # Limitation Analyses of factors modifying the risk of anxiety disorders and depressive disorders were carried out thanks to a population study in theŻywiec district. Due to its demographic characteristics and compact geographic location, this area was selected as optimal for this type of research. A limitation, however, may be the structure of this region, the vast majority of which consists of rural areas. Therefore, much less data were obtained from people living in the city, which could result in a lack of significant differences in the incidence of anxiety disorders and depression depending on the place of residence. People over 60 were the least numerous age group in the study. At the same time, in this group, the greatest number of respondents met the depression criteria. There is a chance that if the number of respondents in this age group was increased, the frequency of the examined disorders would be even greater. However, it can be assumed that elderly people who suffered from the tested symptoms were more likely to participate in the study, which could significantly inflate the results. In the group of people who denied alcohol use, anxiety disorders were more frequently observed. The question about alcohol use was about the last month. It is difficult to estimate whether abstinence was forced by previous alcohol addiction, taking medications due to comorbidities or fear of addiction. The too small size of the group of people using marijuana was probably the reason why smoking was not significantly influenced by anxiety and depression, which could be due to the reluctance of the respondents to disclose this information. The HADS scale, which is a short and easy-to-fill diagnostic tool, detects depressive symptoms while not differentiating the severity of a depressive episode, which may lead to the overestimation of the percentage of respondents with depressive disorders. On the other hand, there is a risk that people with the greatest severity of symptoms were unable to complete the questionnaire on their own, which could lead to underestimating the scale of depression. Due to the presence of more than one risk factor in one person, it is difficult to assess their independent influence on the occurrence of the analyzed disorders. Institutional Review Board Statement: The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Ethics Committee of the Medical University of Silesia (protocol code KNW 0022/KB/156/15), which concluded that the study did not require its consent. # Informed consent statement: Informed consent was obtained from all subjects involved in the study. # Data availability statement: The data presented in this study are available on request from the last author.

Salivary Outer Membrane Vesicles and DNA Methylation of Small Extracellular Vesicles as Biomarkers for Periodontal Status: A Pilot Study Citation: Han, P.; Bartold, P.M.; Salomon, C.; Ivanovski, S. Salivary Outer Membrane Vesicles and DNA Methylation of Small Extracellular Vesicles as Biomarkers for Periodontal Status: A Pilot Study. Int. # Introduction Periodontitis is a highly prevalent chronic inflammatory disease, affecting up to 50% of the global population. It is associated with microbial dysbiosis and characterised by the destruction of periodontal tissues that may result in tooth loss if left untreated. Current clinical diagnosis relies on a manual instrument (probe) and radiography to determine periodontal status by measuring various parameters, such asbleeding on probing (BOP), plaque index (PI), periodontal pocket depth (PPD) and periodontal bone loss. Early detection and diagnosis of periodontitis would allow timely interventions and appropriate treatments. A non-invasive, biologically based diagnostic technique is yet to be developed and it may improve clinical diagnosis for routine periodontal screening and monitoring of periodontitis patients. Saliva analysis has been investigated widely as a promising vehicle for periodontitis diagnosis , due to being non-invasive, easy to access and harboring a wide range of systemic and local components (i.e., cytokines, extracellular vesicles (EVs), bacteria or bacterial by-products, redox enzymes). Besides the common cytokines, salivary redox, such as stress oxidants and antioxidants, has emerged as a potential biomarker for periodontitis, but does not have the ability to differentiate between the stages of periodontal disease. Salivary small extracellular vesicles (sEVs, also named exosomes), a group of biological nanoparticles produced by virtually every species, are emerging as potential biomarkers for periodontitis, owing to their cargos of nucleic acids (microRNAs, DNAs, etc.), lipids and protein. Most of the current literature has only investigated human-derived CD9/CD63/CD81+ sEVs from biofluids, with little attention given to Gram-negative bacteria-derived outer membrane vesicles (OMVs). Recent studies have explored the potential of sEVs as diagnostic tools in periodontitis. It has been shown that CD9/CD81-positive salivary sEVs were decreased in periodontitis patients compared to healthy controls. Salivary sEV (exosomal) programmed deathligand 1 (PD-L1) mRNA was increased (p < 0.01) in periodontitis versus non-periodontitis subjects. Moreover, 26 salivary sEV (exosomal) proteins were only detected in severe periodontitis patients and 58 proteins were identified only in the healthy group. Three salivary sEV miRNAs (hsa-miR-140-5p, hsa-miR-146a-5p and hsa-miR-628-5p) were significantly increased in periodontitis patients compared to healthy controls. Furthermore, salivary exosomal hsa-miR-125a-3p (AUC = 1) is a potential biomarker for chronic periodontitis compared to healthy controls. However, it remains unknown whether salivary OMVs and human sEV-associated DNA methylation landscapes can be used as potential biomarkers for periodontitis diagnosis. There is strong evidence that Gram-negative periodontal pathogens, such as Tannerella forsythia, Porphyromonas gingivalis, Treponema denticola, Prevotella intermedia, Fusobacterium nucleatum, Campylobacter rectus, Peptostreptococcus anaerobius and Eikenella corrodens, are associated with a microbial dysbiosis that results in periodontal attachment loss and disease progression. All of these species secrete outer membrane vesicles (OMVs), nanosized proteoliposomes enriched with an outer layer of lipopolysaccharide (LPS), which play a vital role in intracellular communication, microbial virulence and host immune response. It has been shown that P. gingivalis OMVs facilitate bacterial co-aggregation and influence the bacterial composition of periodontal plaque, which may enhance subgingival biofilm growth. Detecting saliva LPS+ OMVs and specific bacterium-derived OMVs is important to understand microbiome-host interaction in periodontal disease. DNA methylation is a heritable epigenetic change involving the addition of a methyl group to either cytosine or adenine, which is capable of modulating gene expression. The most common form of DNA methylation is 5 methylcytosine (5mC) and 5mC can be demethylated to 5-hydroxymethylcytosine (5hmC). Beyond 5mC and 5hmC, N6-Methyladenosine modification in DNA (m6dA) is the most abundant DNA modification. Recent research suggests that global cytosine methylation (5mC) patterns of genomic DNA (gDNA) from tumor tissues and cell-free DNA (cfDNA) from serum can be used as universal biomarkers for breast, prostate and colorectal cancers. However, the global methylation pattern of 5mC, 5hmC and m6dA in salivary sEVs and whole saliva has not been assessed in relation to periodontal status. The purpose of this pilot study was to investigate the global DNA epigenetic patterns, LPS-positive OMVs population and specific periodontal pathogen-derived OMVs in salivary sEVs from healthy, gingivitis and periodontitis patients and to determine the diagnostic power of aberrant DNA epigenetics or specific bacterial OMVs as potential biomarkers in periodontitis. # Results ## Participant characteristics As shown in, the group of 22 participants (n = 22) was mixed gender (15 males, 7 females) and mostly non-smokers from various ethnic backgrounds (9 Caucasians and 13 Asians), with an age range from 24 to 66 years. It is noted that a statistically significant age difference was found between periodontitis and non-periodontitis patients. ## Salivary sev characterisation and quantification of cd9+ sevs and lps+ omvs We isolated the salivary sEVs using the SEC method and verified them by transmission electron microscopy (TEM) and nanoparticle tracking analysis (NTA). Salivary sEVs were confirmed as having a cup-shaped morphology, with sEV-associated protein (TSG101 and CD9) expression. The NTA results demonstrated that the average modeand particle concentrationof sEVs were comparable between the healthy, gingivitis and periodontitis groups. difference in saliva LPS in gingivitis and periodontitis groups compared to the healthy group, while LPS levels in sEVs were significantly increased in the periodontitis group when compared to healthy patients. We next investigated CD9 and LPS expression levels in saliva and salivary sEVs. Using our in-house CD9 enzyme-linked immunosorbent assay (ELISA,, we found that CD9 in saliva and CD9+ sEVs were comparable between the healthy, gingivitis and periodontitis groups. Furthermore, an endotoxin quantification kit was used for LPS+ sEVsand the results showed that there was no statistically significant difference in saliva LPS in gingivitis and periodontitis groups compared to the healthy group, while LPS levels in sEVs were significantly increased in the periodontitis group when compared to healthy patients. ## Global methylation profile in saliva and salivary sevs The gDNA from 200 µL saliva and sEV-associated gDNA (from 10 9 particles) were comparable in quantityand qualitybetween healthy, gingivitis and periodontitis groups. Commercial Abcam global 5mC, 5hmC and m6dA ELISA kits were used to measure the global methylation for both saliva and sEV samples, as illustrated in. The results demonstrated that in saliva, there was no significant difference in global 5mC ## Bacterium and omv detection in saliva and salivary sevs Our pilot study identified the presence of periodontal pathogen-secreted outer membrane vesicles (OMVs) in isolated salivary sEVs using genomic DNA qPCR. The results demonstrated that, except for salivary P. gingivalis, all the other bacterial strainsshowed no difference in saliva between healthy, gingivitis and periodontitis groups. Meanwhile, only periodontitis salivary sEVs contained higher levels of periodontal pathogens, including T. denticola, E. corrodens, P. gingivalisand F. nucleatum. ## Discrimination power of up-regulated parameters in periodontitis Our research showed that periodontitis sEV LPS+ OMVs, global 5mC methylation and OMVs secreted by four periodontal pathogens (T. denticola, E. corrodens, P. gingivalis and F. nucleatum) were significantly increased compared to the healthy group. Consequently, receiver operating characteristic (ROC) curves and area under the curve (AUC) values were used to examine the discrimination power of these differentially regulated parameters as potential biomarkers for periodontal disease compared to healthy patients. The data showed that most of the parameters in salivary sEVs performed better in discriminating between gingivitis, periodontitis and healthy subjects compared to saliva. Peridontitis sEVs had a far better discriminatory performance in the selected parameters than gingivitis sEVs when both were compared to healthy patients. Of note, LPS+ OMVs used to measure the global methylation for both saliva and sEV samples, as illustrated in. The results demonstrated that in saliva, there was no significant difference in global 5mC, 5hmC (Figure 2e) and m6dAmethylation between the healthy, gingivitis and periodontitis groups. However, periodontitis sEVs exhibited significantly increased global 5mCand m6dAmethylation compared to those from the healthy groups, while there was no change regarding sEV h5mC methylationbetween the three groups. The discrimination power of tested parameters between gingivitis and periodontitis was also carried out. Except salivary E. corroden (AUC = 0.86), salivary EVs displayed a better performance in distinguishing gingivitis from periodontitis. It is noted that global 5mC methylation (AUC = 1) and F. nucleatum OMVs (AUC = 0.94) in salivary sEVs can distinguish gingivitis from periodontitis. Our pilot study identified the presence of periodontal pathogen-secreted outer membrane vesicles (OMVs) in isolated salivary sEVs using genomic DNA qPCR. The results demonstrated that, except for salivary P. gingivalis, all the other bacterial strainsshowed no difference in saliva between healthy, gingivitis and periodontitis groups. Meanwhile, only periodontitis salivary sEVs contained higher levels of periodontal pathogens, including T. denticola, E. corrodens, P. gingivalisand F. nucleatum. # Discussion This pilot study has provided fundamental insight into the human global DNA methylation profiles of sEVs and Gram-negative bacterial OMVs in different periodontal conditions (healthy, gingivitis and periodontitis). Our research showed that LPS + OMVs, global 5mC methylation and four periodontal pathogen (T. denticola, E. corrodens, P. gingivalis and F. nucleatum) OMVs were significantly increased in periodontitis sEVs compared to sEVs those from healthy groups. Notably, salivary sEV global DNA methylation (5mC) appears to be a highly sensitive marker (with an AUC of 1) for differentiating periodontitis and healthy patients. Conversely, in the whole saliva samples, only P. gingivalis showed a statistically significant increase (p < 0.05) in periodontitis compared to healthy patients, and the discriminatory power of this marker was low (AUC = 0.82). Generally, sEVs appear to have superior potential compared to whole saliva samples as diagnostic biomarkers of periodontitis. This may be because sEVs are more likely to reflect the local oral environment compared to the whole saliva, which reflects an individual's overall systemic status. This is consistent with our findings that sEV miRNAs are better biomarkers for periodontitis than whole saliva. Salivary sEV isolation methods are still being developed. Four published studies reporting on salivary sEV diagnosis research in periodontitis used either a precipitationbased ExoQuick kitor ultracentrifuge method, with potential nuclear acid and protein contamination. Our very recent research used the SEC method to isolate salivary sEVswith enriched salivary sEV particles compared to the ultracentrifuge method, and this approach was applied in the current study. Our current study was consistent with our previous study, showing that salivary sEV mode and particle concentration were comparable between healthy, gingivitis and periodontitis groups. Extracellular vesicles (EVs) are heterogeneous populations secreted by all species and play vital roles in physiologic and pathological cellular processes. The role of various subpopulations of sEVs from different tetraspanin proteins (CD63+, CD9+, CD81+ subtypes) in periodontal disease has not been widely investigated and requires further elucidation. A recent report demonstrated salivary CD9+ and CD 81+ sEVs were decreased in periodontitis patients, while another report showed that salivary CD63+ sEVs are comparable in periodontitis and non-periodontitis patients. In our study, we developed an in-house CD9 ELISA assay to quantify CD9+ sEVs, and the results demonstrated that salivary CD9 expression and salivary CD9+ sEVs were comparable between healthy, gingivitis and periodontitis groups, which was in line with Chaparro Padilla et al.. It is widely recognised that certain Gram-negative periodontal pathogens are associated with periodontal diseases, such as T. forsythia, P. gingivalis, T. denticola, P. intermedia, F. nucleatum, C. rectus, P. anaerobius and E. corrodens. Thus, detection of these bacteria in saliva has been used for periodontal disease diagnosis. In line with Damgaard et al., our results demonstrated that salivary P. gingivalis is significantly increased in periodontitis patients compared to healthy patients. Outer membrane vesicles (OMVs) are Gram-negative bacterial by-products, enriched with lipopolysaccharide (LPS), which can be potentially used to quantify OMV subpopulations among isolated salivary sEVs. The data showed that salivary LPS was not statistically significantly different in the healthy, gingivitis and periodontitis groups, while LPS+ OMV populations were significantly enhanced in periodontitis patients compared to healthy patients. The diagnostic power of salivary LPS and LPS+ OMVs was also evaluated and LPS+ OMVs showed good biomarker discrimination (AUC = 0.89) for periodontitis, compared to the healthy group. To further confirm which bacterial strains contributed to the LPS+ OMVs, gDNA qPCR was performed for purified salivary sEVs. The data demonstrated that P. gingivalis-, T. denticola-, E. corrodensand F. nucleatum-secreted OMVs were significantly enhanced in periodontitis patients compared to healthy patients. It was noted that P. gingivalis and T. denticola OMVs are good discriminators (AUC at 0.9 and 0.82, respectively), suggesting that they are strongly associated with periodontitis. Notwith-standing the limitation of the small sample size of this pilot study, our data provide initial evidence that LPS+ OMVs, P. gingivalis OMVs and T. denticola OMVs could be potential diagnostic biomarkers for periodontitis. However, this concept can only be fully explored once pure microbial OMVs or OMVs from a specific bacterial strain can be isolated from saliva and validated with a larger patient cohort. As periodontal diseases are associated with genetic and epigenetic factors, it is important to understand salivary DNA epigenetics in different periodontal states. DNA methylation is an epigenetic mechanism involving the addition of a methyl group to cytosine or adenosine. 5-methylcytosine (5mC), 5-hydroxymethylcytosine (5hmC) and N6methyladenosine modifications (m6dA) are the most critical DNA methylation epigenetics in regulating physiological and pathological processes. It has been reported that increased global DNA methylation (5-methylcytosine, 5mC) is associated with diabetes and breast/prostate/colorectal cancers, however, there have been no attempts at investigating global 5mC, 5hmC and m6dA methylation levels in both saliva and salivary sEVs in periodontal disease. Moreover, in the periodontology field, most studies focus on gene-specific DNA methylation in gingival tissues or blood samples (reviewed in . Our current pilot study is the first to utilise simple ELISA kits (turnaround time of approximately 2 h) to investigate global DNA epigenetics. While no differences in global methylation could be detected between the whole salivary samples from the three groups, significantly increased global 5mC and m6dA were detected in periodontitis sEVs compared to the healthy controls, with good discriminatory power (AUC = 1, 0.87, respectively). These differences in global methylation require further investigation in larger patient cohorts. Further, whole methylome investigations, such as using bisulphite sequencing or methylated DNA immunoprecipitation sequencing, could be used to determine individual salivary sEV methylation candidate biomarkers associated with different periodontal states. Aside from the small sample size, a limitation of this pilot study was the age difference between the control/gingivitis and periodontitis groups, which should be addressed in future studies by recruiting older healthy and gingivitis patients. Another limitation of this study was that smoking status is different among the groups (two smokers in the periodontitis group; others are non-smokers). Thus, future studies should recruit a large cohort with age, gender and smoking status-matched non-smoker participants to confirm these findings. In summary, notwithstanding the limitations of this pilot study, the findings suggest that sEVs may represent superior biosamples compared to the whole saliva for assessing periodontal status via global methylation and OMV assessment. # Materials and methods ## Participant recruitment This pilot study recruited 22 participants, with approved human ethics and written informed consent from the University of Queensland Human Ethics Committee (approval number 2018001225; approval date: 12 November 2018). Written informed consent was obtained from all subjects involved in the study with the following inclusion and exclusion criteria: (a) inclusion criteria: ≥18 years old; ≥20 teeth (excluding third molars); no periodontal treatment or antibiotic therapy three months prior to investigation, no longterm use of anti-inflammatory drugs; (b) exclusion criteria: metabolic bone diseases, autoimmune disease, unstable diabetes or post-menopausal osteoporosis, pregnancy. The Florida probe periodontal charting system (Florida Probe Corporation Gainesville, FL, USA) was calibrated and used to collect periodontal parameters using a standardised probing force. Comprehensive periodontal charting was performed by two independent experienced periodontists for each participant to determine bleeding on probing (BOP) and periodontal pocket depths (PPDs). Healthy (n = 7), gingivitis (n = 7) and periodontitis (n = 8) subjects were defined as described previouslyand the new classification of periodontitis guidelines: (a) healthy: no periodontal disease history; PPD < 3 mm; BOP < 15 % sites; (b) gingivitis: no periodontal pocket, PPD < 3 mm; BOP > 30 % sites; (c) stage III/IV periodontitis: > 30% of the sites with PPD ≥ 3 mm and BOP, at least five sites with PPD ≥ 5mm on at least three non-adjacent teeth. All participants had no underlying systemic diseases. The clinical characteristics of the participants are shown in. Unstimulated whole saliva samples were collected before the full-mouth periodontal charting, as described previously, by asking the participants to spit the saliva samples into a sterile tube. The saliva samples were collected between 9 a.m. and 12 p.m. from the participants who refrained from eating and drinking for at least 1 h (up to 2 h). Before the saliva collection, the participants were asked to rinse their mouth to remove any food debris using 10 mL of water. The whole saliva was collected by spitting into a sterile Falcon tube, and the samples were kept on ice. Then, fresh saliva was aliquoted and frozen in a −80 - C freezer. ## Salivary sev isolation and characterisation Salivary sEVs were isolated using size exclusion chromatography (SEC) columns (miniPURE-EVs, HansaBioMed, Lonza, Tallinn, Italy) according to the manufacturer's instructions and as described previously. Briefly, 250 µL of saliva were diluted in 250 µL of 1x phosphate-buffered saline (PBS, without calcium, magnesium, phenol red; In Vitro Technologies Pty Ltd., Australia). The samples were subjected to differential centrifugation at 4 - C: 300 g for 15 min, 2600 g for 15 min, 16,000 g for 20 min, prior to fractionating the supernatant on an SEC column. Each 100 µL fraction was collected; fractions 7 to 11 were obtained and concentrated to 100 µL using an Amicon Ultra 0.5 Centrifugal Filter Unit (10 kDa, Merck Millipore, QLD, Australia) by centrifugation at 14,000 g for 5 min at 4 - C. Following the recommendation from the International Society of Extracellular Vesicles, sEVs were characterised via morphology by transmission electron microscopy (TEM), EV-associated protein markers by Western blot and sEV particle numbers by nanoparticle tracking analysis (NTA). The TEM analysis was described previously. Briefly, sEV samples were fixed in 3% glutaraldehyde and adsorbed on Formvar carbon-coated electron microscopy grids. After washing with PBS, the grids were placed in uranyl-oxalate solution (pH 7) for 3 min and then imaged using an FEI Tecan 12 transmission electron microscope (FEI, Hillsboro, OR, USA). Western blotting was performed to detect sEV-associated proteins markers: CD 9 and tumour susceptibility gene 101 (TSG 101), as previously described. Briefly, the sEV protein samples were separated by SDS-PAGE and transferred to a polyvinylidene difluoride membrane. The membrane was blocked with Odyssey ® Blocking Buffer at room temperature for 1 h, and primary antibodies (CD9, 1:1000, Santa Cruz Biotechnology; TSG101, 1:1000, Santa Cruz Biotechnology, Dallas, TX, USA) were incubated overnight at 4 - C. Then, the membrane was incubated with anti-rabbit DyLight 800 secondary antibody (1:10,000 in Odyssey Blocking Buffer) and anti-mouse DyLight 700 secondary antibodies, prior to being visualised on an Odyssey ® Infrared Imaging System (LI-COR Biotechnology, Inc., Lincoln, NE, USA). The NTA analysis was carried out to determine the sEV particle size and concentration, as described previously. Polystyrene latex beads (100 nm, Malvern NTA 4088) were used as a positive control and 1x PBS was used as a negative control. A NanoSight NS500 instrument (NanoSight, Salisbury, UK) with a 488 nm laser module and NTA 3.1 software was used to obtain 5 videos of 30 s for each sample with a camera level of 14 and a detection threshold set at 5. The video file was processed and analysed to determine the mode of particle size and particle concentration. ## Determination of cd9 and lipopolysaccharide (lps) in saliva and sevs To quantify saliva CD9 and CD9+ sEVs, an in-house enzyme-linked immunosorbent assay (ELISA) was developed following PeproTech's TMB ELISA Development Kit protocol (illustrated in. For CD9 quantification, 12.5 µL of whole saliva and 10 8 sEVs particles from each group were used. Briefly, the ELISA plates were coated with 4 µg/mL of monoclonal mouse anti-human CD9 antibody (HansaBioMed, Lonza) overnight at 4 - C. After 1 h of blocking, saliva or salivary sEVs samples were added and incubated for 1 h at 37 - C. After washing 3 times with 1% Tween 20 in PBS, monoclonal biotin-conjugated mouse anti-human CD9 at 4 µg/mL (HansaBioMed, Lonza) was added to the wells to bind with sEVs or saliva. After 3 further washes with PBST, diluted streptavidin-HRP conjugated HRP (1:2000) was added and incubated for 30 min at 37 - C. Then, 3,3 ,5,5 -Tetramethylbenzidine (TMB) substrate was used for colour development and stopped with 1 M HCl stop solution. The OD value was be measured at 450 nm. According to the manufacturer's protocol, a Pierce TM Chromogenic Endotoxin Quantification Kit (ThermoFisher Scientific) was used to determine endotoxin LPS levels by measuring the interaction of endotoxin (LPS) with proenzyme Factor C. For LPS quantification, 12.5 µL of whole saliva and 10 8 sEV particles from each group were used after measuring the yellow colour at an absorbance of 405 nm (illustrated in. ## Genomic dna isolation and dna methylation landscape quantification Genomic DNA (gDNA) was isolated from 200 µL of saliva and 10 8 sEV particles, using Trizol TM reagent (Invitrogen, ThermoFisher Scientific, Melbourne, Australia) following the manufacturer's instructions and as described previously. The DNA quality and quantity were measured using a NanoDrop™ One spectrophotometer (Thermo Fisher Scientific, Australia). Global methylation analysis of 5-methylcytosine (5mC), 5-hydroxymethylcytosine (5hmC) and N6-methyladenosine (m6dA) for DNA was performed by using a Global DNA Methylation Assay Kit (5mC, ab233486, Abcam), a Global DNA Hydroxymethylation Assay Kit (5hmc, ab233487, Abcam) and an m6dA DNA Methylation Assay Kit (m6dA, ab233488, Abcam), respectively, as per the manufacturer's instructions. Briefly, sample DNA, positive controls at 6 different concentrations (to generate standard curves) and the negative control were mixed with DNA-binding solution and incubated at 37 - C for 60 min. After washing three times with 150 µL washing buffer, 5mC/5hmC/m6dA antibodies along with signal indicator and enhancer solution were added and incubated at room temperature for 1 h. Following thorough washing (five times) with wash buffer, 50µL developer solution were added and incubated for 3 min at room temperature until the positive control with the highest concentration turned blue. Subsequently, 50 µL of stop solution were added to each well for 2 min to stop the enzyme reaction. The absorbance was measured at 450 nm for 2 min on an Infinite Pro spectrometer. The global methylation level of all DNAs was calculated using the following equations: 5mc/5hmC % = * 100% m6dA % = * 100% where the slope (OD/1%) was determined from the standard curve using linear regression; S = amount of input sample DNA in ng; P = amount of input positive control in ng; OD = optical density. ## Gdna real-time quantitative pcr (qpcr) Salivary bacterial outer membrane vesicles (OMVs) and bacteria were detected using gDNA qPCR, as described previously. The primers for 8 periodontal pathogens (Tannerella forsythia, Porphyromonas gingivalis, Treponema denticola, Prevotella intermedia, Fusobacterium nucleatum, Campylobacter rectus, Peptostreptococcus anaerobius and Eikenella corrodens) are listed in. The gDNA qPCR was performed using SYBR Green reagent in StepOnePlus equipment (Applied Biosystems). The relative expression was normalised with 16S rRNA and calculated as 2 -(normalised average Cts) . # Statistical analysis All the data are presented as mean ± standard deviation (SD) and ordinary one-way ANOVA with Tukey's multiple comparison test (a single pooled variance) was used to compare the data between healthy, gingivitis and periodontitis patients with GraphPad Prism 8.3.1 software (San Diego, SA, USA). The normality of data distribution was calculated in Prism using a QQ plot. p < 0.05 was considered a statistically significant difference. Receiver operating characteristic (ROC, using the Wilson/Brown method) curves and the area under the curve (AUC) were generated using data from healthy controls and patients (gingivitis and periodontitis, respectively) with GraphPad Prism 8.3.1 software (San Diego, SA, USA). The area under the curve (AUC, indicating the discriminatory power of the biomarkers) and p values were calculated by the software.

Impact of low-frequency ultrasound technology on physical, chemical and technological properties of cereals and pseudocereals A B S T R A C T Cereals (CE) and pseudocereals (PSCE) play a pivotal role in nourishing the human population. Low-frequency ultrasound (LFUS) modifies the structure of CE and PSCE macromolecules such as starch and proteins, often improving their technological, functional and bioactive properties. Hence, it is employed for enhancing the traditional processes utilized for the preparation of CE-and PSCE-based foods as well as for the upcycling of their by-products. We report recent advances in LFUS treatments for hydration, germination, extraction of bioactive compounds from by-products, and fortification of CEs and PSCE, including kinetic modelling and underlying action mechanisms. Meta-analyses of LFUS influence on compounds extraction and starch gelatinization are also presented. LFUS enhances hydration rate and time lag phase of CE and PSCE, essential for germination, extraction, fermentation and cooking. The germination is improved by increasing hydration, releasing promoters and eliminating inhibitors. Furthermore, LFUS boosts the extraction of phenolic compounds, polysaccharides and other food components; modifies starch structure, affecting pasting properties; causes partial denaturation of proteins, improving their interfacial properties and their peptides availability. Overall, LFUS has an outstanding potential to improve transformation processes and functionalities of CE and PSCE. # Introduction Ultrasound (US) is a mechanical wave with frequency higher than 20 kHz, i.e., beyond the audible frequency range of humans. It can be differentiated in three frequency ranges: 1) 20 to 100 kHz, conventional high-power or low-frequency ultrasound (LFUS); 2) 100 kHz to 2 MHz, sonochemistry range and 3) 5 to 10 MHz, medical or analytical range. Over the past 20 years, the application of LFUS has gained significant potential interest in many industrial sectors because of the versatility in modifying as well as generating microstructures. Thanks to its adaptability, relative simplicity, low energy requirements and limited impact on the environment, LFUS has a very promising future in food technology and has already been applied to solve technical issues such as increasing yield, reducing treatment time [bib_ref] Advances in application of ultrasound in food processing: A review, Bhargava [/bib_ref] [bib_ref] Ultrasound-assisted synthesis of heterocyclic compounds, Ziarani [/bib_ref] , extracting with nontoxic solvents [bib_ref] Ultrasound: A clean, green extraction technology, Tiwari [/bib_ref] and limiting energy needs [bib_ref] The uses of ultrasound in food technology, Mason [/bib_ref]. LFUS can be employed to make changes in liquids, dispersions, solid and gaseous media, but they are particularly efficient when applied through a liquid [bib_ref] Ultrasound processing, Feng [/bib_ref]. LFUS effects in liquid systems are mainly related to the cavitation phenomenon: during the negative pressure half-wave (rarefaction) the medium is stressed by tensile forces that increase the distance among molecules; once the cavitation threshold is reached, the liquid breaks down and empty bubbles form. When a cavity cannot tolerate the surrounding liquid pressure, it implodes violently, and energy is immediately released. Cavities collapse is far more rapid than heat diffusion, creating a localized hot spot while the liquid bulk remains cold. Inside the cavities, temperatures of 1700-4700 - C and pressures of 180-300 MPa are reached. As temperature rises, cavitation threshold goes down due to drops in surface tension and viscosity, thus a liquid will start to cavitate at lower sound pressure. However, water vapor pressure will be higher and vapor-filled bubbles will form, reducing pressure difference and cushioning bubbles implosions. Very high tensile forces are needed for cavitation. Energy and intensity, along with medium viscosity, surface tension, vapor pressure, nature and concentration of dissolved gas, presence of solid particles, temperature, and treatment pressure, determine the extent of cavitation [bib_ref] Effect of ultrasound on the technological properties and bioactivity of food: a..., Soria [/bib_ref]. Gas-filled bubbles or particulate produce weak spots in the liquid that allow the process to begin; alternatively, pre-generated cavitation bubbles can be starting points. The most common frequency range applied is 20-40 KHz, because cavitation is hardly achieved at high frequencies unless very great intensities are adopted. The true mechanical power in the liquid is usually far lower than the nominal one; in addition, devices with different efficiencies can be used and the volume may change as well. Therefore, to compare results it is necessary to experimentally evaluate the specific amount of energy per volume unit, an often-overlooked aspect [bib_ref] Ultrasoundassisted hydration of wheat grains at different temperatures and power applied: Effect..., Guimarães [/bib_ref]. The volumetric power can be estimated by the following formula [bib_ref] Ultrasoundassisted hydration of wheat grains at different temperatures and power applied: Effect..., Guimarães [/bib_ref] [bib_ref] Calorimetric method for measurement of acoustic power absorbed in a volume of..., Margulis [/bib_ref] : [formula] P = mC p dT dt [/formula] where m is the water mass (kg), C p its specific heat (4.186 kJ/(kg - C)) and dT/dt ( - C/s) the temperature increase during the first 90 s of sonication. Besides the mechanical effect, LFUS generates free radicals through water homolytic cleavage: H 2 O → H- + OH - . However, free radicals are produced at the fastest rate from water at around 200 kHz (sonochemistry range), while at 20 kHz their development is minimal [bib_ref] Modification of food ingredients by ultrasound to improve functionality: A preliminary study..., Ashokkumar [/bib_ref]. Food processing procedures using LFUS have been applied to improve product quality, including bioactivity, and process efficiency in cereal and pseudocereals. However, to the best of our knowledge, a review focused on the effect of LFUS on key cereal and pseudocereal processes like hydration, germination and fortification is not available. Additionally, we summarize as forest plots the effect of LFUS on the extraction of bioactive compounds from cereal and pseudocereal byproducts. Finally, we report and discuss the influence of LFUS on structural changes of starch, protein, and dietary fibre. # Methods ## Research and selection criteria Literature research from 1994 to 2022 was carried out on Scopus and Web of Science databases. The keywords used were Ultrasound OR Ultrasonic followed by common or Latin name of each cereal and pseudocereal (i.e., Wheat, Rice, Corn or Maize, Barley, Rye, Oat, Sorghum, Teff, Triticale, Millet, Quinoa, Amaranth, Buckwheat, and Chia). All articles concerning the application of low-frequency ultrasound to cereals and pseudocereals, or to derived ingredients, aimed at improving processing, increasing yield, or determining a modification in their chemical and technological behaviour were considered. Sources which presented LFUS as an ameliorating technique but without providing control trials were considered biased and were excluded. ## Statistical elaboration On two occasions the results were summarized through meta-analysis by creating forest plots with the software RevMan v5.4.1 (https://www. cochrane.org). For [fig_ref] Figure 1: Forest plot for random-effects meta-analysis of extraction yield [/fig_ref] , the effect size was standardized by computing log response ratios and standard errors as illustrated by Borenstein et al.: [formula] ln(R) = ln(X 1 ) − ln(X 2 ) SE ln(R) = ̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅ ̅ S 2 pooled ( 1 n 1 (X 1 ) 2 + 1 n 2 (X 2 ) 2 ) √ √ √ √ [/formula] When the studies provided multiple non-independent outcomes the effect size was computed as their mean log response ratio, while the variance was determined as follows: [formula] var ( 1 m ∑ m i=1 Y i ) = ( 1 m ) 2 ( ∑ m i=1 V i + ∑ i∕ =j ( r ij ̅̅̅̅ ̅ V i √ ̅̅̅̅ ̅ V j √ ) ) [/formula] where m is the number of correlated variables and r is the correlation coefficient; when not calculated, r was cautiously assumed to be 1, resulting in a lower weight of the study. The data were plotted with the generic inverse variance method by RevMan, which provides automatic antilogarithm. For , the effect size (i.e., the raw mean difference) and standard error were computed as described by Borenstein et al.: [formula] D = X 1 − X 2 SE D = ̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅ ̅ S 2 1 n 1 + S 2 2 n 2 √ [/formula] ## Low-frequency ultrasound application in cereals and pseudocereals processing Because of ease of utilization, energy-saving, and relatively low costs, high-power LFUS has been proposed for processes as different as analysis, extraction, chemical and technological properties modification, emulsification, drying, freezing [bib_ref] Applications of ultrasound in analysis, processing and quality control of food: A..., Awad [/bib_ref] and even microbial decontamination [bib_ref] Screening of post-harvest decontamination methods for cereal grains and their impact on..., Schmidt [/bib_ref]. In the cereals area, the research is focused mainly on the effects of LFUS on physical and chemical changes of proteins [bib_ref] Sensory and quality evaluation of traditional compared with power ultrasound processed corn..., Janve [/bib_ref] [bib_ref] Effect of low-frequency ultrasonic-assisted enzymolysis on the physicochemical and antioxidant properties of..., Liang [/bib_ref] [bib_ref] Ultrasonic modification of starch -Impact on granules porosity, Sujka [/bib_ref] [bib_ref] Effects of ultrasound pretreatment on the enzymolysis and structural characterization of wheat..., Zhang [/bib_ref] and starch [bib_ref] Effect of high-intensity ultrasound treatment on nutritional, rheological and structural properties of..., Kaur [/bib_ref] , enzymatic activity [bib_ref] Does ultrasound improve the activity of alpha amylase? A comparative study towards..., Oliveira [/bib_ref] , antioxidant compounds retrieval [bib_ref] Changes in the GABA and polyphenols contents of foxtail millet on germination..., Sharma [/bib_ref] and biofortification [bib_ref] Increased vitamin B 5 uptake capacity of ultrasonic treated milled rice: A..., Bonto [/bib_ref] [bib_ref] The ultrasound assisted hydration as an opportunity to incorporate nutrients into grains, Miano [/bib_ref] [bib_ref] Efficient fortification of folic acid in rice through ultrasonic treatment and absorption, Tiozon [/bib_ref] , hydration and germination [bib_ref] Ultrasoundassisted hydration of wheat grains at different temperatures and power applied: Effect..., Guimarães [/bib_ref] [bib_ref] Sono-hydro priming process (ultrasound modulated hydration): Modelling hydration kinetic during paddy germination, Kalita [/bib_ref] [bib_ref] Ultrasound technology enhances the hydration of corn kernels without affecting their starch..., Miano [/bib_ref] [bib_ref] Characterizing physicochemical, nutritional and quality attributes of wholegrain Oryza sativa L. subjected..., Xia [/bib_ref] as well as food technological and sensorial quality improvement [bib_ref] Sensory and quality evaluation of traditional compared with power ultrasound processed corn..., Janve [/bib_ref] [bib_ref] Power ultrasound assisted mixing effects on bread physical properties, Pa [/bib_ref] [bib_ref] Effect of ultrasonic treatment of brown rice at different temperatures on cooking..., Cui [/bib_ref] [bib_ref] Determination of the effect of high energy ultrasound application in tempering on..., Yüksel [/bib_ref] [bib_ref] The effect of ultrasound on the functional properties of wheat gluten, Zhang [/bib_ref]. ## Hydration Hydration is essential for many processes, such as cooking, extraction, fermentation, germination and malting [bib_ref] The hydration of grains: A critical review from description of phenomena to..., Miano [/bib_ref]. During kernels hydration, the water uptake follows preferential paths, according to grain shape and structure. For instance, in barley caryopses the water penetrates through the hilar fissure and the micropyle, while in maize kernels ## Fig. 2. Forest plot for random-effects meta-analysis of starch gelatinization enthalpy (ΔH; J/g): control group vs. the LFUS-treated group (under the most severe conditions). The raw mean difference between the control and the LFUS-treated group was used as the effect size. enters by the tip cap, filling the empty space between germ and endosperm [bib_ref] Ultrasound technology enhances the hydration of corn kernels without affecting their starch..., Miano [/bib_ref] [bib_ref] Ultrasound assisted hydration improves the quality of the malt barley, Borsato [/bib_ref]. Several studies suggest that LFUS treatments enhance hydration. Miano et al. [bib_ref] Mechanisms for improving mass transfer in food with ultrasound technology: Describing the..., Miano [/bib_ref] proposed two different mechanisms: a) direct, related to inertial flow and sponge effect, where the ultrasound-induced cavitation creates a compression-rarefaction turnover that squeezes and releases the matrix, pumping water through pre-existing pores; and b) indirect, leading to the formation of new micro-channels because of physical damage to the tissues. However, low water activity does not allow cavitation, hence dry caryopses (a w ~ 0.65) undergo a slow hydration [bib_ref] Mechanisms for improving mass transfer in food with ultrasound technology: Describing the..., Miano [/bib_ref]. In fact, new cavities begin to form in the matrix only when an adequate a w is reached, accelerating the hydration process. For the above-mentioned reason, LFUS treatments are more effective when applied on pre-soaked kernels [bib_ref] Effect of soaking and sprouting treatment on germination rate of paddy, Chatchavanthatri [/bib_ref]. The relative extent of direct and indirect effects on the kernels varies according to the species [bib_ref] Ultrasound technology enhances the hydration of corn kernels without affecting their starch..., Miano [/bib_ref]. Generally, cereals hydration kinetics exhibit a downward concave curve with one or two steps [bib_ref] The hydration of grains: A critical review from description of phenomena to..., Miano [/bib_ref] that can be modelled by the Peleg model [bib_ref] An empirical model for the description of moisture sorption curves, Peleg [/bib_ref] : [formula] M t = M 0 + t k 1 + k 2 t [/formula] where M t is the moisture (% dry weight) at t time, M 0 the initial moisture, k 1 and k 2 are Peleg's constants. The k 1 is equal to the inverse of the initial moisture uptake rate and the k 2 is inversely related with equilibrium moisture. As an alternative the Weibull model, an exponential empirical model useful for its simplicity and good accuracy in describing complex processes with high variability, may be used. [bib_ref] Ultrasoundassisted hydration of wheat grains at different temperatures and power applied: Effect..., Guimarães [/bib_ref] : [formula] M t − M 0 M eq − M 0 = 1 − exp [ − ( t β ) ] α [/formula] where M eq is the equilibrium moisture, β the scale parameter and α the shape parameter; α describes the initial water uptake behavior and β represents the time needed to achieve 63% moisture; both parameters are inversely related to the process rate [bib_ref] Effect of milk fat and total solids concentration on the kinetics of..., Machado [/bib_ref]. The parameter α is related to moisture migration during hydration: when α > 1 the process is governed by internal moisture diffusion as well as external mass transfer, while when α < 1 the external mass transfer is negligible [bib_ref] Ultrasoundassisted hydration of wheat grains at different temperatures and power applied: Effect..., Guimarães [/bib_ref]. Both Peleg and Weibull models have been used to model the hydration kinetic of wheat treatment with LFUS [bib_ref] Ultrasoundassisted hydration of wheat grains at different temperatures and power applied: Effect..., Guimarães [/bib_ref]. The LFUS technology accelerates the hydration process by increasing the hydration rate and reducing the lag phase, contributing to reach a higher final moisture. Li et al. [bib_ref] Effect of ultrasonic treatment on the hydration and physicochemical properties of brewing..., Li [/bib_ref] examined the initial 120 min of rice water uptake and observed that ultrasound hastened the maximum absorption rate in a power-dependent manner and that the moisture absorption rate increased according to a Peleg-like curve. Miano et al. [bib_ref] Ultrasound technology enhances the hydration of corn kernels without affecting their starch..., Miano [/bib_ref] reported a 300 min drop in hydration time for flint maize kernels, albeit no new cavities were formed because of their extremely vitreous endosperm, while Patero and Augusto [bib_ref] Ultrasound (US) enhances the hydration of sorghum (Sorghum bicolor) grains, Patero [/bib_ref] observed a decrease from 320 to 190 min in sorghum kernels. The combination of LFUS and temperature may further improve these results. Kalita et al. [bib_ref] Sono-hydro priming process (ultrasound modulated hydration): Modelling hydration kinetic during paddy germination, Kalita [/bib_ref] compared paddy rice hydration by soaking and LFUS-assisted soaking at different temperatures and noticed that the process was abridged from ≥ 24 h (traditional) to 3 h (LFUS). Similarly, Guimarães et al. [bib_ref] Ultrasoundassisted hydration of wheat grains at different temperatures and power applied: Effect..., Guimarães [/bib_ref] observed that LFUS per se shortened wheat hydration time by 28-42%, while the combination of LFUS and temperature (25 - C) cut it by 72%. Patero and Augusto [bib_ref] Ultrasound (US) enhances the hydration of sorghum (Sorghum bicolor) grains, Patero [/bib_ref] determined that the temperature (53 - C) had a greater effect than LFUS on sorghum hydration, and a similar result was obtained by Shafaei et al. [bib_ref] Neural computing efforts for integrated simulation of ultrasoundassisted hydration kinetics of wheat, Shafaei [/bib_ref] in wheat at different temperatures (from 30 - C to 70 - C). Nevertheless, high temperatures may have relevant drawbacks, including protein denaturation, starch gelatinization or industrial plant insulation costs. Moreover, the increase of processing temperature above 60 - C reduces cavitation, because the amount of water vapor inside the bubble increases and a collapse occurs [bib_ref] Engineering principles of ultrasound technology, Kentish [/bib_ref]. Hence, a 35-40 - C range is generally appropriate to accelerate the process without damaging seed properties [bib_ref] Ultrasound assisted hydration improves the quality of the malt barley, Borsato [/bib_ref]. In barley malt production, the combination of LFUS and mild temperatures (20-25 - C) accelerates the steeping phase; the LFUS intensity affects water absorption mainly in the early phase (Peleg's k 1 ), while the temperature has a stronger effect on higher final moisture. Overall, their combination reduces the hydration time by half [bib_ref] Effect of intermittent high-intensity sonication and temperature on barley steeping for malt..., De Carvalho [/bib_ref] or even more [bib_ref] Ultrasound assisted hydration improves the quality of the malt barley, Borsato [/bib_ref]. Temperatures above 40 - C worsen malt quality, therefore LFUS represents a promising approach to reduce steeping time; additionally, the LFUS mechanical power is dissipated as heat, reducing the energetic costs. ## Germination Germination is the sum of events that break the seed quiescence status and allow the embryonic axis to appear. As a first step, the seed hydrates so the pre-existent enzymes are activated and initiate the Krebs cycle or, in rice, even anaerobic pathways, the endosperm reserves are mobilized, and the protein synthesis becomes intensive. The germination ends with cells elongation, that leads to root and shoot emission [bib_ref] Seed germination, mobilization of food reserves, and seed dormancy, Srivastava [/bib_ref]. From the point of view of food sciences, the activation of dormant enzymes during germination leads to significant changes in biochemical and nutritional characteristics. The αand β-amylases rapidly degrade the carbohydrates, with a consequent surge in reducing sugars, while other enzymes decompose cell walls and enhance the accessibility of internal nutrients [bib_ref] Impact of preharvest sprouting on endogenous hydrolases and technological quality of wheat..., Olaerts [/bib_ref]. The proteins are hydrolyzed, increasing peptides and amino acids availability [bib_ref] Impact of preharvest sprouting on endogenous hydrolases and technological quality of wheat..., Olaerts [/bib_ref]. Different bioactive compounds, such as tocols, thiamine, riboflavin, folic acid, vitamin C and carotenoids, are biosynthesized to generate the nutrients necessary for seedling growth [bib_ref] Impact of preharvest sprouting on endogenous hydrolases and technological quality of wheat..., Olaerts [/bib_ref] [bib_ref] Sprouted and freeze-dried wheat and oat seedsphytochemical profile and in vitro biological..., Aborus [/bib_ref]. The metabolic activities fuel phenolic compounds biosynthesis [bib_ref] Phenolic compounds in grains, sprouts and wheatgrass of hulled and non-hulled wheat..., Benincasa [/bib_ref] [bib_ref] Antioxidant properties and heat damage of water biscuits enriched with sprouted wheat..., Hidalgo [/bib_ref] , while the degradation of cell walls leads to an increase of readily available free phenolic compounds [bib_ref] Phenolic acid composition of sprouted wheats by ultra-performance liquid chromatography (UPLC) and..., Van Hung [/bib_ref]. The abundant presence of bioactive molecules boosts the antioxidant capacity [bib_ref] Polyphenol composition and in vitro antioxidant activity of amaranth, quinoa buckwheat and..., Alvarez-Jubete [/bib_ref]. Furthermore, sprouting increases phytase activity, leading to phytic acid degradation and better micronutrients absorption in the gastrointestinal tract [bib_ref] Impact of preharvest sprouting on endogenous hydrolases and technological quality of wheat..., Olaerts [/bib_ref] [bib_ref] Reduction of phytic acid and enhancement of bioavailable micronutrients in food grains, Gupta [/bib_ref]. Overall, controlled germination improves the nutritional composition of the kernels, but the enzymatic activity has a negative impact on some technological properties of the flours (e.g., leavening), stressing the importance of a carefully controlled process to avoid excessive hydrolysis [bib_ref] Impact of preharvest sprouting on endogenous hydrolases and technological quality of wheat..., Olaerts [/bib_ref]. US, together with high pressure, cold plasma, and pulsed electric fields, is an emerging technology used to regulate seeds germination [bib_ref] Characterizing physicochemical, nutritional and quality attributes of wholegrain Oryza sativa L. subjected..., Xia [/bib_ref]. The LFUS treatment enhances germination primarily by its positive effect on hydration. In fact, cavitation appears to cause fissures in the pericarp, augmenting the availability of water and oxygen [bib_ref] Physicochemical properties of germinated dehulled rice flour and energy requirement in germination..., Ding [/bib_ref] [bib_ref] Application of ultrasonic waves as a priming technique for accelerating and enhancing..., Yaldagard [/bib_ref]. The absorbed water stimulates the embryo to release gibberellic acid, a chemical promoter of germination [bib_ref] Application of ultrasonic waves as a priming technique for accelerating and enhancing..., Yaldagard [/bib_ref] ; at the same time, the hydration promotes the leakage of germination inhibitors, such as the abscisic acid [bib_ref] Seed germination, mobilization of food reserves, and seed dormancy, Srivastava [/bib_ref]. With regards to other phytohormones, Wei et al. [bib_ref] Enhancement of the differentiation of protocormlike bodies of Dendrobium officinale to shoots..., Wei [/bib_ref] observed that LFUS decreased the indole-3-acetic acid (IAA)/cytokinin ratio in Dendrobium officinale protocorms. IAA is a root growth promoter while cytokinin stimulates shoot growth [bib_ref] Seed germination, mobilization of food reserves, and seed dormancy, Srivastava [/bib_ref] , therefore a higher IAA presence may lead to earlier germination. Moreover, LFUS probably acts facilitating the mobilization of endosperm nutrients by disrupting cell membranes [bib_ref] Application of ultrasonic waves as a priming technique for accelerating and enhancing..., Yaldagard [/bib_ref] , as corroborated by the sonication-induced erosion at the joint between cells observed by Ananthakrishnan et al. [bib_ref] Ultrasonic treatment stimulates multiple shoot regeneration and explant enlargement in recalcitrant squash..., Ananthakrishnan [/bib_ref] , which would stimulate a faster flow of water and nutrients. Finally, Chen et al. [bib_ref] Ultrasonic vibration seeds showed improved resistance to cadmium and lead in wheat..., Chen [/bib_ref] documented that LFUS-treated wheat seeds exhibited greater antioxidant activities of catalase, superoxide dismutase and glutathione reductase. Apparently, reactive oxygen species (ROS) generated by cavitation elicit an antioxidant response that enhances plant vigor. Other improvements were found in number of germinated seeds, protein and total chlorophyll content, shoot length, fresh and dry biomass, and levels of cell damage indicators [bib_ref] Ultrasonic vibration seeds showed improved resistance to cadmium and lead in wheat..., Chen [/bib_ref]. Although the mechanism that induces faster germination is not fully understood, numerous observations have been collected [fig_ref] Table 1: Conditions tested in ultrasound processing for improving seed hydration and germination [/fig_ref]. Ding et al. [bib_ref] Physicochemical properties of germinated dehulled rice flour and energy requirement in germination..., Ding [/bib_ref] found that LFUS improves rice sprout growth speed by 22.3-26.9%, Xia et al. [bib_ref] Characterizing physicochemical, nutritional and quality attributes of wholegrain Oryza sativa L. subjected..., Xia [/bib_ref] reported that an additional 10-15% of rice seeds sprouted in the initial 32 h, while Goussous et al. [bib_ref] Enhancing seed germination of four crop species using an ultrasonic technique, Goussous [/bib_ref] described an enhanced wheat germination rate of about 60% compared to plain water soaking; in addition, a positive interaction between LFUS and temperature was found. Similarly, Yaldagard et al. [bib_ref] Application of ultrasonic waves as a priming technique for accelerating and enhancing..., Yaldagard [/bib_ref] observed that LFUS reduced barley steeping time from the usual 46 h to 25-30 h, depending on the applied intensity, and that the percentage of germinated seeds increased from 93.3% (control) up to 99.4% in the most intensely treated group. Furthermore, the hair roots were luxuriant and, at the highest intensity, the germination period was shortened from 7 to 4 days. A slight increase in barley germinated seeds was observed also by Miano et al. [bib_ref] Effect of ultrasound technology on barley seed germination and vigourEffect of ultrasound..., Miano [/bib_ref] and was attributed to porosity increment, since plastic bags insulated the seeds from the water during the sonication. Similarly, Hassan et al. [bib_ref] Phytochemical characterization of ultrasound-processed sorghum sprouts for the use in functional foods, Hassan [/bib_ref] observed that 5 min LFUS treatment greatly improved the number of germinated sorghum grains from 78% (control) to 94%; a longer process decreased the germinated seeds, therefore they inferred that at high amplitude the treatment damaged cells and/or enzymes. Kratovalieva et al. [bib_ref] Ultrasound influence on coleoptile length at Poaceae seedlings as valuable criteria in..., Kratovalieva [/bib_ref] studied LFUS application to different cereals and noticed significant effects on coleoptile and mesocotyl elongation. Oat and rye caryopses were very sensitive to LFUS stimulation (+119% and +65% growth, respectively), while triticale and wheat showed smaller increases. In a comparison among different pre-treatment, LFUS proved to be better than microwaves or heat in enhancing buckwheat germination by 4-4.5% [bib_ref] Effects of ultrasonic waves, microwaves, and thermal stress treatment on the germination..., Wang [/bib_ref]. Beyond the described effect, sonication before or during germination in rice and wheat appeared to increase the synthesis of γ-aminobutyric acid, a health-enhancing compound, by 1.7 and 3-fold respectively [bib_ref] Enhancement of gamma-aminobutyric acid (GABA) and other health-related metabolites in germinated red..., Ding [/bib_ref] [bib_ref] Stimulation of antioxidant activity and -aminobutyric acid synthesis in germinated wheat grain..., Naumenko [/bib_ref]. ## Extraction The most popular LFUS application is improving the extraction of different compounds. [fig_ref] Table 2: Optimized experimental conditions and yields [/fig_ref] summarizes a variety of LFUS-assisted extraction trials with their respective optimized experimental conditions. [fig_ref] Figure 1: Forest plot for random-effects meta-analysis of extraction yield [/fig_ref] reports a forest plot for random-effects performed considering twenty studies comparing LFUS-assisted vs. traditional or non-US extractions of different fractions (phenolics compounds, oil, polysaccharides, protein and tocopherols) from diverse substrates; the extraction yields of LFUS-assisted processes were compared with a control group of non-LFUS or traditional processes. The projection of each point on the abscissa axis represents, if higher than 1, the n of an nfold increase in the extraction yield. An overall pooled increase of 54% in yield is achieved by implementing LFUS. Extraction of phenolic compounds, especially from byproducts and waste materials, was the most studied process. Although in some cases yield improvements are negligible, other advantages must be considered. In fact, LFUS-assisted extraction generally allows to shorten the extraction period, to employ food-grade non-hazardous solvents and to use lower temperatures. For instance, in Balasubramaniam et al. [bib_ref] Effect of enzyme pretreatment in the ultrasound assisted extraction of finger millet..., Balasubramaniam [/bib_ref] experiments, although yields were similar, the phenolic compounds extraction from millet was performed in half the time (30 vs. 60 min) and at lower temperature (50 vs. 60 - C). Izadifar [bib_ref] Ultrasound pretreatment of wheat dried distiller's grain (DDG) for extraction of phenolic..., Izadifar [/bib_ref] found that a 30 s sonication increased the extracted phenolics content from wheat by 14.3%. Further reductions in time were reported by Wang et al. [bib_ref] Optimisation of ultrasound-assisted extraction of phenolic compounds from wheat bran, Wang [/bib_ref] , from 15 h of Soxhlet extraction to 25 min; Giopato Viell et al. [bib_ref] Comparison between ultra-homogenisation and ultrasound for extraction of phenolic compounds from teff..., Viell [/bib_ref] , 5 min instead of 10; Teslić et al. [bib_ref] Defatted wheat germ as source of polyphenols-Optimization of microwave-assisted extraction by RSM..., Teslić [/bib_ref] , 30 min instead of 24 h, with 11% higher yield; Chen et al. [bib_ref] Ultrasound-assisted extraction and characterization of anthocyanins from purple corn bran, Chen [/bib_ref] , 32.5% higher yields in half the time; Melini and Acquistucci [bib_ref] Extraction of free and insoluble-bound phenolic compounds from pigmented rice by commonly..., Melini [/bib_ref] , 15 min instead of 1 h. As reported by Chen et al. [bib_ref] Ultrasound-assisted extraction from defatted oat (Avena sativa L.) bran to simultaneously enhance..., Chen [/bib_ref] , even though the highest yields were achieved at high temperature (70 - C), at 20 - C LFUS extracted in 5 min 54% more phenolics than a 65 min control trial. Therefore, LFUS represents a strategy to perform shorter low-temperature extractions and deserves to be introduced in analytical protocols to ensure correct quantification of phenolics [bib_ref] Modelling and optimization of ultrasound-assisted extraction of phenolic compounds from black quinoa..., Melini [/bib_ref]. LFUS-assisted extraction in some cases can efficiently use ethanol in place of chlorinated methanol [bib_ref] Effect of enzyme pretreatment in the ultrasound assisted extraction of finger millet..., Balasubramaniam [/bib_ref] [bib_ref] Ultrasound-assisted extraction and characterization of anthocyanins from purple corn bran, Chen [/bib_ref]. On the other hand, a certain amount of methanol contributes to lower the average viscosity of solvent mixture, thus facilitating cavitation [bib_ref] Comparison between ultra-homogenisation and ultrasound for extraction of phenolic compounds from teff..., Viell [/bib_ref]. In addition, with LFUS, green new developed deep eutectic solvents (DESs) may substitute or outperform the traditional ones. DES was 13% more effective than 30% ethanol in extracting ferulic and sinapic acid from corn silk [bib_ref] Optimizing ultrasound-assisted deep eutectic solvent extraction of bioactive compounds from Chinese wild..., Zeng [/bib_ref] and extracted phenolic compounds from wheat bran better than 60% EtOH (+33%) and alkaline hydrolysis (+71%) [bib_ref] Highefficiency extraction of phenolics from wheat waste biomass (bran) by combining deep..., Cherif [/bib_ref]. Besides, properly formulated DESs can be compatible with HPLC as dilution solvent [bib_ref] Optimizing the ultrasound-assisted deep eutectic solvent extraction of flavonoids in common buckwheat..., Mansur [/bib_ref]. The LFUS ability at improving extraction is probably due to cavitation-driven cellular damage (especially cell walls), particle size reduction, strong mixing, increase in porosity and in specific surface area. These power-and time-dependent events allow the solvent to easily access tissues and dissolve phenolic compounds [bib_ref] Ultrasound pretreatment of wheat dried distiller's grain (DDG) for extraction of phenolic..., Izadifar [/bib_ref] [bib_ref] Extraction of free and insoluble-bound phenolic compounds from pigmented rice by commonly..., Melini [/bib_ref] [bib_ref] Highefficiency extraction of phenolics from wheat waste biomass (bran) by combining deep..., Cherif [/bib_ref] [bib_ref] Pre-treatment of ferulic acid esterases immobilized on MNPs to enhance the extraction..., Gadalkar [/bib_ref]. In enzyme-assisted extraction, cavitation leads to higher available surface for enzyme digestion [bib_ref] Pre-treatment of ferulic acid esterases immobilized on MNPs to enhance the extraction..., Gadalkar [/bib_ref]. Giopato Viell et al. [bib_ref] Comparison between ultra-homogenisation and ultrasound for extraction of phenolic compounds from teff..., Viell [/bib_ref] suggested that part of extracted flavonoids from teff could come from release of bound fraction because of cavitation, while Chen et al. [bib_ref] Ultrasound-assisted extraction from defatted oat (Avena sativa L.) bran to simultaneously enhance..., Chen [/bib_ref] inferred that at (continued on next page) 70 - C heat-led cleavage of glycosidic bonds could have freed phenolics from bound forms. Nevertheless, a LFUS overtreatment may worsen the extraction because of phenolics oxidation or destruction [bib_ref] Ultrasound pretreatment of wheat dried distiller's grain (DDG) for extraction of phenolic..., Izadifar [/bib_ref] [bib_ref] Ultrasound assisted extraction of anthocyanins and total phenolic compounds from dried cob..., Muangrat [/bib_ref] due to temperature increase [bib_ref] Ultrasound pretreatment of wheat dried distiller's grain (DDG) for extraction of phenolic..., Izadifar [/bib_ref] and radicals generation [bib_ref] Pre-treatment of ferulic acid esterases immobilized on MNPs to enhance the extraction..., Gadalkar [/bib_ref] [bib_ref] From 'green' technologies to 'red' antioxidant compounds extraction of purple corn: a..., Jiang [/bib_ref]. According to Dzah et al. [bib_ref] Ultrasound-induced lipid peroxidation: Effects on phenol content and extraction kinetics and antioxidant..., Dzah [/bib_ref] , although above 65 - C the phenolic molecules are degraded by hydrogen peroxide and their antioxidant activity decreases, the sonication time has not a direct impact on them. Improving oil extraction is another LFUS application. In a study by Cravotto et al. [bib_ref] Improving solvent-free extraction of policosanol from rice bran by high-intensity ultrasound treatment, Cravotto [/bib_ref] , alkaline aqueous extraction with LFUS support outperformed hexane, resulting in 22% higher yield. The extraction time was also cut by 8-fold (4 h to 30 min). Additionally, in the presence of LFUS rice wax was hydrolyzed to policosanols mixture under milder temperature and alkali concentration [bib_ref] Improving solvent-free extraction of policosanol from rice bran by high-intensity ultrasound treatment, Cravotto [/bib_ref] [bib_ref] Extraction of policosanols from hydrolysed rice bran wax by high-intensity ultrasound, Liu [/bib_ref]. and Chekin [bib_ref] The ultrasound-assisted aqueous extraction of rice bran oil, Khoei [/bib_ref] tested LFUS as pre-treatment in aqueous extraction carried out at 45 - C and pH 12 for 15 min, and achieved an 20% yield, similar to a Soxhlet extraction (23.4%) which, nevertheless, used hexane. In addition, the small Soxhlet extraction surplus in yield appeared linked to free fatty acids content, 3.2% in hexane-extracted oil vs. almost zero in aqueousextracted oil: this could suggest that the aqueous extraction is selective against free fatty acids. Han et al. [bib_ref] Ultrasound-assisted aqueous enzymatic extraction of corn germ oil: analysis of quality and..., Han [/bib_ref] confirmed that LFUS, being more efficient than steam and heat, could be a viable pre-treatment also in enzyme-led aqueous extraction. In Krishnan et al. [bib_ref] Ultrasound assisted extraction of oil from rice bran: a response surface methodology..., Krishnan [/bib_ref] experiments, LFUS-assisted extraction of oil with ethanol achieved a yield 29% higher compared to the traditional process with hexane, and 78.5% higher compared to non-US ethanol extraction. Additionally, no changes in oil composition or modification in peroxide value were found [bib_ref] Ultrasound assisted extraction of oil from rice bran: a response surface methodology..., Krishnan [/bib_ref] [bib_ref] Optimization of rice lipid production from ultrasound-assisted extraction by response surface methodology, Xu [/bib_ref]. Therefore, a green solvent like ethanol could be conveniently used in oil extraction. By implementing a LFUS phase in their supercritical CO 2 extraction, Soares et al. [bib_ref] Extraction of rice bran oil using supercritical CO2 combined with ultrasound, Brazilian, Soares [/bib_ref] improved the yield by 27% and cut by 60% the time, while extracting four oryzanol precursors, campesterol, β-sitosterol, stigmasterol and 4-methylenecycloartanol. As above mentioned, sometimes LFUS over-treatment may lead to poor performances. According to Xu et al. [bib_ref] Optimization of rice lipid production from ultrasound-assisted extraction by response surface methodology, Xu [/bib_ref] , a power excess lowers yields since cavitation bubbles act as hindrance for wave propagation, while too long a treatment increases the quantity of suspended impurities and worsens solvent penetration. An over-treatment could also result in lipids emulsification, hence lower extraction yield [bib_ref] The ultrasound-assisted aqueous extraction of rice bran oil, Khoei [/bib_ref]. Lastly, Khoei and Chekin [bib_ref] The ultrasound-assisted aqueous extraction of rice bran oil, Khoei [/bib_ref] , Han et al. [bib_ref] Ultrasound-assisted aqueous enzymatic extraction of corn germ oil: analysis of quality and..., Han [/bib_ref] and Xu et al. [bib_ref] Optimization of rice lipid production from ultrasound-assisted extraction by response surface methodology, Xu [/bib_ref] pointed out the cushioning effect as one of the major factors preventing yield enhancement by raising temperature. Polysaccharides are a third class of compounds whose LFUS-assisted extraction has been studied. Ðordević and Antov [bib_ref] Ultrasound assisted extraction in aqueous two-phase system for the integrated extraction and..., Ðordević [/bib_ref] developed a protocol to extract at the same time xylo-oligosaccharides and phenolic compounds from wheat chaff, and the LFUS treatment improved by 2 and 1.3-fold the respective yields. In xylan extraction from corn cob, a 10 min LFUS pre-treatment at 70 - C could substitute the traditional one at 95 - C for 1 h [bib_ref] Study of the classical and ultrasound-assisted extraction of the corn cob xylan, Hromádková [/bib_ref]. In a more recent work, Hromádková et al. [bib_ref] Comparison of conventional and ultrasound-assisted extraction of phenolics-rich heteroxylans from wheat bran, Hromádková [/bib_ref] confirmed that, thanks to LFUS implementation, heteroxylans extraction from wheat bran could be hastened by 55 min and with minor yield loss. Successively, Benito-Román et al. [bib_ref] Ultrasound-assisted extraction of β-glucans from barley, Benito-Román [/bib_ref] found that LFUS-assisted extraction of β-glucans from barley outperformed the stirred tank technique by improving yield up to 58% and hastening the processing time from 3 h to 16 min. Conversely, Sun and Tomkinson [bib_ref] Characterization of hemicelluloses obtained by classical and ultrasonically assisted extractions from wheat..., Sun [/bib_ref] found a negligible increase in hemicellulose extraction yield from wheat straw. Wang et al. [bib_ref] Combination of ultrasound and heat in the extraction of chia seed (Salvia..., Wang [/bib_ref] optimized the extraction of mucilage (i.e., heteropolysaccharides) from chia seeds and observed that the adoption of LFUS increased 10-fold the extraction yield. Besides yields, LFUS proved to be energy-efficient and timesaving. Reis et al. [bib_ref] Improved efficiency of brewer's spent grain arabinoxylans by ultrasound-assisted extraction, Reis [/bib_ref] designed a 25 min-long LFUS-assisted extraction from brewer's spent grain that yielded the same amount of arabinoxylan they got with a 7 h optimized alkaline extraction. In another case the LFUSassisted extraction of β-glucan from hull-less barley gave a 6% lower yield, but was still convenient because lasted 5 min instead of the standard 90 min [bib_ref] Ultrasound-assisted extraction of β-D-glucan from hull-less barley: Assessment of physicochemical and functional..., Sourki [/bib_ref]. Shorter times often allow to save energy because heating is needed for a briefer period. According to Benito-Román et al. [bib_ref] Ultrasound-assisted extraction of β-glucans from barley, Benito-Román [/bib_ref] , LFUS could cut energy demand by at least 52%. The yield loss, in case of over-treatment, seems to be the main LFUS drawback [bib_ref] Optimization of ultrasonicmicrowave assisted alkali extraction of arabinoxylan from the corn bran..., Jiang [/bib_ref] [bib_ref] Box-Behnken design based statistical modeling for ultrasound-assisted extraction of corn silk polysaccharide, Maran [/bib_ref] [bib_ref] Optimisation of ultrasound-assisted enzymatic extraction of arabinoxylan from wheat bran, Wang [/bib_ref]. Although the most drastic ultrasound conditions resulted in lower polysaccharides molecular weight, due to cavitationdriven chain-breaking, experimental parameters could be adjusted to minimize unwanted depolymerization [bib_ref] Study of the classical and ultrasound-assisted extraction of the corn cob xylan, Hromádková [/bib_ref] [bib_ref] Ultrasound-assisted extraction of β-glucans from barley, Benito-Román [/bib_ref]. In addition, at the highest amplitude, unspecific disruption of cell compartments results in different polysaccharides release, thus threatening the extract purity [bib_ref] Ultrasound-assisted extraction of β-D-glucan from hull-less barley: Assessment of physicochemical and functional..., Sourki [/bib_ref]. As detailed in the last rows of [fig_ref] Table 2: Optimized experimental conditions and yields [/fig_ref] , LFUS-assisted extraction was tested for a variety of other compounds as well. Protein can be extracted efficiently with LFUS: Li et al. [bib_ref] Effect of ultrasound on alkali extraction protein from rice dreg flour, Li [/bib_ref] reported a 2-fold increase in yield, thanks to better solvent penetration. Similarly, Quintero-Quiroz et al. [bib_ref] Physicochemical properties and functional characteristics of ultrasoundassisted legume-protein isolates: a comparative study, Quintero-Quiroz [/bib_ref] saw a 2.4-fold higher yield, while Zhu et al. [bib_ref] Optimization of ultrasound-assisted extraction of defatted wheat germ proteins by reverse micelles, Zhu [/bib_ref] stated that LFUS improved forward mass transfer in reverse micelles extraction. Laqui-Vilca et al. [bib_ref] Ultrasound-assisted optimal extraction and thermal stability of betalains from colored quinoa (Chenopodium..., Laqui-Vilca [/bib_ref] extracted betalains from quinoa seed hulls in 10-40 s instead of 10 min. Loypimai et al. [bib_ref] Optimization of tocols and γ-oryzanol extraction from rice bran using ultrasound and..., Loypimai [/bib_ref] developed a LFUS-assisted soybean oil enriching process: the oil was directly employed as green solvent to simultaneously extract tocopherols and γ-oryzanol from rice bran; the yields were slightly lower than with conventional extraction, but this technique simplified the process and avoided the use of hazardous solvents. Wizi et al. [bib_ref] Ultrasound-microwave assisted extraction of natural colorants from sorghum husk with different solvents, Wizi [/bib_ref] recovered 3.6-fold higher red colorants, mainly apigeninidin and luteolinidin, from sorghum husk by coupling ultrasound and microwave technologies. Ye et al. [bib_ref] Ultrasound-assisted extraction of corn carotenoids in ethanol, Ye [/bib_ref] extracted carotenoids from corn meal and observed that LFUS were better than magnetic stirring, inasmuch that after 60 min they achieved 0.08 mg/ml vs. 0.004 mg/ml. However, the experiment did not last enough to find out the optimum conditions. ## Fortification Because of its capacity to produce cracks and pores, LFUS has also been tested as a method to fortify grains. Bonto et al. [bib_ref] Increased vitamin B 5 uptake capacity of ultrasonic treated milled rice: A..., Bonto [/bib_ref] treated polished white rice in a 53 Hz ultrasonic bath: a time-dependent fragmentation of the cells external layer was observed, and the starchy endosperm arrangement was disrupted, but no starch granules were damaged. Although at first sonication led to complete B-vitamins loss, after soaking in pantothenic acid the LFUS-treated grains absorbed and retained 140% more vitamin B 5 than non-sonicated rice. Similarly, Tiozon et al. [bib_ref] Efficient fortification of folic acid in rice through ultrasonic treatment and absorption, Tiozon [/bib_ref] fortified rice by soaking it in 800 ppm folic acid solution after 5 min-long sonication. Brown rice absorbed about 1982-fold its natural folic acid content, while a 4054-fold increase was seen in white rice; in addition, after washing and cooking, they retained 93.5% and 86.5% folic acid, respectively. Yanova et al. [bib_ref] Fortification of cereal groats with iron and zinc using ultrasound, Yanova [/bib_ref] attempted mineral fortification of barley and oat groats by LFUS treatment with 65 mg/ L solutions of Fe 2+ and Zn 2+ . A positive linear relationship between amount of absorbed minerals, temperature and treatment duration was found; furthermore, at higher ultrasound frequency an additional amount of iron and zinc was soaked up, possibly because of the higher number of impingement cycles. In must be stressed that higher frequencies are less likely to produces grains fractures because cavitation collapses are less violent. ## Other applications Cui et al. [bib_ref] Effect of ultrasonic treatment of brown rice at different temperatures on cooking..., Cui [/bib_ref] and Dang et al. [bib_ref] Effects of ultrasonic and enzymatic treatment on physical and chemical properties of..., Dang [/bib_ref] suggested LFUS as a viable alternative to reduce brown rice cooking time by about 16%, with the additional advantage that less vitamins and other solids are lost; the LFUS pre-treatment resulted also in higher grain expansion volume, water uptake and softness. The rice cooking time reduction was confirmed by Yang et al. [bib_ref] Effects of ultrasonic treatment on the cooking and fermentation properties of Shanlan..., Yang [/bib_ref]. In addition, the glycemic index was subjected to negligible changes with respect to untreated brown rice [bib_ref] Effects of ultrasonic and enzymatic treatment on physical and chemical properties of..., Dang [/bib_ref] [bib_ref] The effects of ultrasound -assisted recrystallization followed by chilling to produce the..., Kunyanee [/bib_ref]. LFUS washing at acidic pH improved rice bran quality by reducing lipase and lipoxygenase activities, heavy metals contaminants, phytic acid and coliforms [bib_ref] Effect of washing, soaking and pH in combination with ultrasound on enzymatic..., Mohammadi [/bib_ref]. Habuš et al. [bib_ref] Highintensity ultrasound treatment for prolongation of wheat bran oxidative stability, Habuš [/bib_ref] [bib_ref] Influence of particle size reduction and high-intensity ultrasound on polyphenol oxidase, phenolics,..., Habuš [/bib_ref] confirmed that the shelf life of wheat bran can be extended with an ultrasound treatment thanks to the reduction of the enzymatic activities of lipase, lipoxygenase and polyphenol oxidase. However, the effect is partly due to thermal denaturation from the heat generated, because when the sample was cooled the LFUS treatment appeared less effective than the conventional heating. According to Yüksel and Elgün [bib_ref] Determination of the effect of high energy ultrasound application in tempering on..., Yüksel [/bib_ref] submitting wheat to 30 s sonication during tempering should improve its bread-making quality, although the wet gluten amount was slightly diminished, its quality was better as higher gluten index, dough stability, energy, loaf volume and specific volume were observed. In addition, LFUS-assisted mixing may result in better air incorporation and thus higher loaf development [bib_ref] Power ultrasound assisted mixing effects on bread physical properties, Pa [/bib_ref]. Similarly, sonication during pre-mixing of corn bread dough resulted in better softness, porosity, general sensory acceptability and yellower crust and crumb [bib_ref] The effect of the ultrasound process and pre-gelatinization of the corn flour..., Jalali [/bib_ref]. Hence, LFUS treatment of wheat and quinoa flour suspension could produce tailored-made flours with defined rheological properties, such as decrease in gelatinization enthalpy, lower viscosities and lower gel hardness and cohesiveness [bib_ref] Effect of ultrasound on structural and physicochemical properties of sweetpotato and wheat..., Cui [/bib_ref] [bib_ref] Modification of quinoa flour functionality using ultrasound, Zhu [/bib_ref]. The use of LFUS during corn nixtamalization cooking, reduced the steeping time needed to reach the correct kernel softness from 20 to 1 h; a slight inferior mass leaching was also observed [bib_ref] Enhancement of corn nixtamalization by power ultrasound, Janve [/bib_ref]. A further study confirmed that LFUS reduces steeping time (from 18 to 1.5 h) and does not affect nixtamal quality as well as sensory acceptability of derived products like tortilla chips [bib_ref] Sensory and quality evaluation of traditional compared with power ultrasound processed corn..., Janve [/bib_ref]. reports the broad range of different conditions tested for modifying cereal starches. Ultrasound treatments induce pores, depressions and cracks in starch granules because of the rapid water jets produced by the cavitation bubbles collapse. Yang et al. [bib_ref] Controlled ultrasound treatments modify the morphology and physical properties of rice starch..., Yang [/bib_ref] even reported that rice starch granules were peeled off as power increased. However, LFUS augmented the total pores volume but not their average diameter [bib_ref] Ultrasonic modification of starch -Impact on granules porosity, Sujka [/bib_ref] and affect mainly the largest granules [bib_ref] Impact of ultrasonication on functional and structural properties of Indian wheat (Triticum..., Karwasra [/bib_ref] [bib_ref] Ultrasound-treated starch: SEM and TEM imaging, and functional behaviour, Sujka [/bib_ref]. Additionally, in the range 25-55 - C the temperature effect is inversely correlated to the damage [bib_ref] Morphological, physicochemical, and viscoelastic properties of sonicated corn starch, Amini [/bib_ref] because the collapse of the bubbles favors cavitation but is less violent at higher temperature. Conversely, the application of multiple frequencies is more effective due to the sum of cavitation events [bib_ref] Ultrasonic frequency effect on corn starch and its cavitation, Hu [/bib_ref] [bib_ref] Dual-frequency ultrasonic effect on the structure and properties of starch with different..., Hu [/bib_ref]. ## Effect of low-frequency ultrasound on cereals and pseudocereals components ## Starch Boufi et al. [bib_ref] Ultrasonic assisted production of starch nanoparticles: Structural characterization and mechanism of disintegration, Boufi [/bib_ref] reported that LFUS ruptured starch particles into smaller ones: a 15 min treatment reduced both 1200 nm and 950 nm granules to 600 nm, while a 90 min process trimmed them down to 40 nm, a size reduction easily visualized by beam light scattering, because the starch suspension becomes clearer over the time. Similar observations were reported by Kang et al.for waxy starches, but the amylose-rich samples underwent a slight particles diameter increase due to amylose aggregation, as observed also by Li et al.. The correlation between granule size and degree of polymerization is not clear. For instance, Yang et al. [bib_ref] Controlled ultrasound treatments modify the morphology and physical properties of rice starch..., Yang [/bib_ref] reported a decrease in particle size but no changes in chain-length. However, many studies describe a depolymerization process. According Kang et al.and Li et al. [bib_ref] Effect of ultrasound pretreatment on enzymolysis and physicochemical properties of corn starch, Li [/bib_ref] , LFUS first disrupts weak interactions and then cleaves covalent C-O-C α-1,6 glycosidic bonds (apparently less stable than α-1,4 bonds) thus leading to amylopectin debranching [bib_ref] Fine structure, crystalline and physicochemical properties of waxy corn starch treated by..., Yang [/bib_ref]. Coherently, Li et al.saw a roughly halved weight average molar mass (M w ) and Huang et al.reported a drastic increase in hydrolysis degree. Furthermore, the number average molecular weight (M n ) and polydispersity (M w ∕M n ) diminished with sonication time and their decreasing rate was related to temperature and power [bib_ref] Ultrasonic Degradation of Waxy Rice Starch, Isono [/bib_ref]. Due to these events, Zhou et al. [bib_ref] Study on the structure and digestibility of high amylose Tartary buckwheat (Fagopyrum..., Zhou [/bib_ref] indicated LFUS as the best alternative to produce flavonoids-enriched starch complex from Tartary buckwheat. In addition, they saw an increase in resistant starch (+20%) and a slower starch digestion. Swelling power and solubility were generally increased by LFUS [bib_ref] Impact of ultrasonication on functional and structural properties of Indian wheat (Triticum..., Karwasra [/bib_ref] [bib_ref] Ultrasound-treated starch: SEM and TEM imaging, and functional behaviour, Sujka [/bib_ref] [bib_ref] Morphological, physicochemical, and viscoelastic properties of sonicated corn starch, Amini [/bib_ref] [bib_ref] Ultrasound effect on physical properties of corn starch, Jambrak [/bib_ref] [bib_ref] Sonication of sugary-2 corn: A potential pretreatment to enhance sugar release, Montalbo-Lomboy [/bib_ref] and the temperature-time interaction was positively correlated to these parameters [bib_ref] Morphological, physicochemical, and viscoelastic properties of sonicated corn starch, Amini [/bib_ref]. A slight increase in suspension transparency was noticed by Sujka and Jamroz [bib_ref] Ultrasound-treated starch: SEM and TEM imaging, and functional behaviour, Sujka [/bib_ref] , Amini et al. [bib_ref] Morphological, physicochemical, and viscoelastic properties of sonicated corn starch, Amini [/bib_ref] and Li et al. [bib_ref] Comparative studies on structure and physiochemical changes of millet starch under microwave..., Li [/bib_ref]. Jambrak et al. [bib_ref] Ultrasound effect on physical properties of corn starch, Jambrak [/bib_ref] observed that the sonicated suspension became clear after one-day storage. However, Dey and Sit [bib_ref] Modification of foxtail millet starch by combining physical, chemical and enzymatic methods, Dey [/bib_ref] stated that LFUS turned starch 21% darker. reports the forest plot for random effects performed considering fourteen different studies related to LFUS effect on starch gelatinization enthalpy (ΔH); the raw mean difference between the control and the LFUS-treated group was used as the effect size. The overall pooled LFUS treatment causes a slight decrease (0.97 J/ g) in ΔH, a tendency also observed by Amini et al. [bib_ref] Morphological, physicochemical, and viscoelastic properties of sonicated corn starch, Amini [/bib_ref] , Boufi et al. [bib_ref] Ultrasonic assisted production of starch nanoparticles: Structural characterization and mechanism of disintegration, Boufi [/bib_ref] , Li et al., Huang et al., Zhou et al. [bib_ref] Study on the structure and digestibility of high amylose Tartary buckwheat (Fagopyrum..., Zhou [/bib_ref] and Yu et al. ## Table 3 Conditions tested in ultrasound processing for modifying cereal starches. [bib_ref] Effects of ultrasound processing on the thermal and retrogradation properties of nonwaxy..., Yu [/bib_ref]. The decrease in ΔH was related to the loss of amylopectin double helices, essential structures for granule integrity preservation [bib_ref] Loss of crystalline and molecular order during starch gelatinisation: origin of the..., Cooke [/bib_ref]. Controlled-temperature LFUS treatments allow to obtain starches that demand lower energy when residual helical structures unfold [bib_ref] Ultrasound effect on physical properties of corn starch, Jambrak [/bib_ref] [bib_ref] Different-frequency ultrasonic effects on properties and structure of corn starch, Hu [/bib_ref]. Since amorphous lamellae are disrupted first, their disappearance could determine a temporary higher crystallinity, till the inner lamellae are attacked [bib_ref] Dual-frequency ultrasonic effect on the structure and properties of starch with different..., Hu [/bib_ref]. In fact, the onset temperature (T o ) increased slightly, hence when weak structures are damaged, the remaining stronger crystals will require higher energy to melt [bib_ref] Different-frequency ultrasonic effects on properties and structure of corn starch, Hu [/bib_ref]. Similarly, a rise-and-fall pattern due to short-lived increased interactions between crystals and amorphous chains was reported by Li et al. [bib_ref] Effect of ultrasonic treatment on the hydration and physicochemical properties of brewing..., Li [/bib_ref]. Ultrasound power is inversely related to ΔH [bib_ref] Fine structure, crystalline and physicochemical properties of waxy corn starch treated by..., Yang [/bib_ref] [bib_ref] Ultrasound effect on physical properties of corn starch, Jambrak [/bib_ref] [bib_ref] Effects of ultrasound processing on the thermal and retrogradation properties of nonwaxy..., Yu [/bib_ref] , although an opposite behavior has been observed at very high power (600-1000 W) probably because small amounts of amylose leaked out from the granules and gelled, insulating the surface from external water [bib_ref] Controlled ultrasound treatments modify the morphology and physical properties of rice starch..., Yang [/bib_ref] [bib_ref] Effects of ultrasound processing on the thermal and retrogradation properties of nonwaxy..., Yu [/bib_ref]. These behavioral discrepancies may be explained by the different nature of starch granules: A-type granules are characterized by a tight monoclinic lattice, while B-type granules have a hexagonal arrangement enclosing a cavity with freezable water. Consequently, Atype granules are more resistant, while B-types are more susceptible to LFUS-induced disruption. Yang et al. [bib_ref] Controlled ultrasound treatments modify the morphology and physical properties of rice starch..., Yang [/bib_ref] , Yang et al. [bib_ref] Fine structure, crystalline and physicochemical properties of waxy corn starch treated by..., Yang [/bib_ref] and Hu et al. [bib_ref] Different-frequency ultrasonic effects on properties and structure of corn starch, Hu [/bib_ref] provided observations that support this crystallinity changes theory, Karwasra et al. [bib_ref] Impact of ultrasonication on functional and structural properties of Indian wheat (Triticum..., Karwasra [/bib_ref] , Hu et al. [bib_ref] Dual-frequency ultrasonic effect on the structure and properties of starch with different..., Hu [/bib_ref] and Luo et al.did not observe any significant modification: probably, different experimental conditions (e.g., ultrasound power, time, volume treated, higher efficiency of the probe system with respect to bath) led to higher total energy per volume unit. In fact, conclusive evidence from Boufi et al. [bib_ref] Ultrasonic assisted production of starch nanoparticles: Structural characterization and mechanism of disintegration, Boufi [/bib_ref] showed that LFUS flattened thermogram and x-ray diffraction spectrum, thus supporting the formation of amorphous particles. Because of LFUS-induced starch modification, the pasting properties of cereal starches and flours are also modified. LFUS treatment decrease starch paste peak viscosity, breakdown, and setback, but did not affect pasting temperature [bib_ref] Ultrasonic frequency effect on corn starch and its cavitation, Hu [/bib_ref] [bib_ref] Effect of ultrasound pretreatment on enzymolysis and physicochemical properties of corn starch, Li [/bib_ref] [bib_ref] Fine structure, crystalline and physicochemical properties of waxy corn starch treated by..., Yang [/bib_ref] [bib_ref] Ultrasonic Degradation of Waxy Rice Starch, Isono [/bib_ref] [bib_ref] Comparative studies on structure and physiochemical changes of millet starch under microwave..., Li [/bib_ref] [bib_ref] The pasting properties of sonicated waxy rice starch suspensions, Zuo [/bib_ref]. However, Cui et al. [bib_ref] Effect of ultrasonic treatment of brown rice at different temperatures on cooking..., Cui [/bib_ref] , Yang et al. [bib_ref] Controlled ultrasound treatments modify the morphology and physical properties of rice starch..., Yang [/bib_ref] and Park and Han [bib_ref] Quality controlling of brown rice by ultrasound treatment and its effect on..., Park [/bib_ref] reported lower pasting temperature, higher peak viscosity, and breakdown. According to Li et al. [bib_ref] Effect of ultrasound pretreatment on enzymolysis and physicochemical properties of corn starch, Li [/bib_ref] and Zuo et al. [bib_ref] The pasting properties of sonicated waxy rice starch suspensions, Zuo [/bib_ref] the viscosity drop is due to amylose and long linear amylopectin depolymerization, a finding consistent with the results by Luo et al., who saw no modification in waxy maize starch, but a viscosity fall in amylose-rich Amylomaize V. Amylose is mainly responsible for the viscosity increase during setback [bib_ref] Starch use in foods, Mason [/bib_ref] , therefore its hydrolysis will result in a weaker network. The magnitude of changes in pasting behavior varies from almost negligible to complete starch liquefaction, as in thermal assisted (60 - C) sonication [bib_ref] Ultrasonic Degradation of Waxy Rice Starch, Isono [/bib_ref]. The LFUS and temperature interaction effect on pasting properties was confirmed by Zuo et al. [bib_ref] The pasting properties of sonicated waxy rice starch suspensions, Zuo [/bib_ref] , who reported a peak viscosity decrease down to a quarter. These findings encourage the use of LFUS to further improve starch saccharification, as described by Montalbo-Lomboy et al. [bib_ref] Ultrasonic pretreatment of corn slurry for saccharification: A comparison of batch and..., Montalbo-Lomboy [/bib_ref] and Shewale and Pandit [bib_ref] Enzymatic production of glucose from different qualities of grain sorghum and application..., Shewale [/bib_ref]. ## Proteins Ultrasound-assisted modification of cereal and pseudocereal protein isolates has been extensively studied. [fig_ref] Table 4: Conditions tested in ultrasound processing for modifying cereal proteins [/fig_ref] reports the broad range of conditions tested for modifying cereals and pseudocereals proteins. The results demonstrate that cavitation disrupts weak interactions (electrostatic bonds, hydrogen bonds, hydrophobic effect), thus leading to conformational changes in the secondary and tertiary structures. The proteins unravel and the hydrophobic cores, rich in phenylalanine, tyrosine, and tryptophan, are exposed, thus augmenting the surface hydrophobicity [bib_ref] Modifications of physicochemical and functional properties of amaranth protein isolate (Amaranthus cruentus..., Constantino [/bib_ref] [bib_ref] The use of ultrasound for enzymatic preparation of ACE-inhibitory peptides from wheat..., Jia [/bib_ref] [bib_ref] Effects of sonication on the in vitro digestibility and structural properties of..., Jin [/bib_ref] [bib_ref] Effects of ultrasound and ultrasound assisted alkaline pretreatments on the enzymolysis and..., Li [/bib_ref] [bib_ref] Enzymolysis kinetics and structural characteristics of rice protein with energygathered ultrasound and..., Li [/bib_ref] [bib_ref] Effects of high-intensity ultrasound pretreatment with different levels of power output on..., Li [/bib_ref] [bib_ref] Effect of ultrasound treatments on functional properties and structure of millet protein..., Nazari [/bib_ref] [bib_ref] The effect of ultrasound upon the physicochemical and emulsifying properties of wheat..., O&apos;sullivan [/bib_ref] [bib_ref] Effects of sweeping frequency ultrasound pretreatment on the hydrolysis of zein: angiotensin-converting..., Ren [/bib_ref] [bib_ref] Effects of different working modes of ultrasound on structural characteristics of zein..., Ren [/bib_ref] [bib_ref] Effects of ultrasound pretreatment with different frequencies and working modes on the..., Yang [/bib_ref] [bib_ref] Effects of multi-mode S-type ultrasound pretreatment on the preparation of ACE inhibitory..., Yang [/bib_ref] [bib_ref] Effects of multi-frequency ultrasound on physicochemical properties, structural characteristics of gluten protein..., Zhang [/bib_ref] [bib_ref] Mechanism study of multimode ultrasound pretreatment on the enzymolysis of wheat gluten, Zhang [/bib_ref] [bib_ref] Relationships between the structure of wheat gluten and ACE inhibitory activity of..., Zhang [/bib_ref] [bib_ref] Pretreatment of defatted wheat germ proteins (by-products of flour mill industry) using..., Zhou [/bib_ref] [bib_ref] Comparative study of the effects of ultrasonic power on the structure and..., Zhang [/bib_ref]. Covalent disulphide bonds are also affected by LFUS, and the amount of free-sulfhydryls and disulphides depends on the power intensity used185, 186]. In fact, far higher energy is required to reduce S-S bonds (226 kJ/mol) than to break weak interactions (13 kJ/mol) [bib_ref] The effect of ultrasound upon the physicochemical and emulsifying properties of wheat..., O&apos;sullivan [/bib_ref]. An increase in SH groups, accompanied by a less evident decrease in S-S, was reported [bib_ref] Modifications of physicochemical and functional properties of amaranth protein isolate (Amaranthus cruentus..., Constantino [/bib_ref] [bib_ref] Effects of multi-mode S-type ultrasound pretreatment on the preparation of ACE inhibitory..., Yang [/bib_ref] [bib_ref] Effects of multi-frequency ultrasound on physicochemical properties, structural characteristics of gluten protein..., Zhang [/bib_ref] [bib_ref] Relationships between the structure of wheat gluten and ACE inhibitory activity of..., Zhang [/bib_ref] [bib_ref] Pretreatment of defatted wheat germ proteins (by-products of flour mill industry) using..., Zhou [/bib_ref] [bib_ref] Comparative study of the effects of ultrasonic power on the structure and..., Zhang [/bib_ref] [bib_ref] Characteristics of enzymatic hydrolysis of the wheat gluten proteins treated by ultrasound..., Jin [/bib_ref] , leading to conclude that the breaking of disulphide bonds LFUS led to a much looser gluten matrix [bib_ref] Effects of ultrasound pretreatment on the enzymolysis and structural characterization of wheat..., Zhang [/bib_ref] [bib_ref] Effects of multi-frequency ultrasound on physicochemical properties, structural characteristics of gluten protein..., Zhang [/bib_ref] [bib_ref] Mechanism study of multimode ultrasound pretreatment on the enzymolysis of wheat gluten, Zhang [/bib_ref]. With regards to the secondary structure, the relative number of α-helices decreases in favor of β-sheets and random coils [bib_ref] Effects of ultrasound pretreatment on the enzymolysis and structural characterization of wheat..., Zhang [/bib_ref] [bib_ref] Effects of ultrasound and ultrasound assisted alkaline pretreatments on the enzymolysis and..., Li [/bib_ref] [bib_ref] Effects of high-intensity ultrasound pretreatment with different levels of power output on..., Li [/bib_ref] [bib_ref] Effects of ultrasound pretreatment with different frequencies and working modes on the..., Yang [/bib_ref] [bib_ref] Effects of multi-frequency ultrasound on physicochemical properties, structural characteristics of gluten protein..., Zhang [/bib_ref] [bib_ref] Relationships between the structure of wheat gluten and ACE inhibitory activity of..., Zhang [/bib_ref] [bib_ref] Comparative study of the effects of ultrasonic power on the structure and..., Zhang [/bib_ref] , while intramolecular β-sheets are converted to intermolecular β-sheets [bib_ref] Effects of ultrasound pretreatment on the enzymolysis and structural characterization of wheat..., Zhang [/bib_ref]. Ultrasoundinduced denaturation may result in better digestibility due to increased accessibility for the digestive enzymes: in fact, Jin et al. [bib_ref] Effects of sonication on the in vitro digestibility and structural properties of..., Jin [/bib_ref] observed an improvement from 41.4 (control) to 58.2% (LFUS-treated) with buckwheat protein isolate. These changes do not affect protein molecular weight but disrupt aggregates [bib_ref] Comparative study of the effects of ultrasonic power on the structure and..., Zhang [/bib_ref]. Nevertheless, overtreatment may result in new intermolecular disulphide and hydrophobic bonds that increase the dimensions of the aggregates [bib_ref] Physicochemical properties and functional characteristics of ultrasoundassisted legume-protein isolates: a comparative study, Quintero-Quiroz [/bib_ref] [bib_ref] Modifications of physicochemical and functional properties of amaranth protein isolate (Amaranthus cruentus..., Constantino [/bib_ref] [bib_ref] The effect of ultrasound upon the physicochemical and emulsifying properties of wheat..., O&apos;sullivan [/bib_ref] [bib_ref] Pretreatment of defatted wheat germ proteins (by-products of flour mill industry) using..., Zhou [/bib_ref] [bib_ref] Comparative study of the effects of ultrasonic power on the structure and..., Zhang [/bib_ref] [bib_ref] Conformational and physicochemical properties of quinoa proteins affected by different conditions of..., Vera [/bib_ref]. Constantino and Garcia-Rojas [bib_ref] Modifications of physicochemical and functional properties of amaranth protein isolate (Amaranthus cruentus..., Constantino [/bib_ref] suggested that hydrogen peroxide generated by hydroxy radical addition could act as an oxidizing agent, thus converting free sulfhydryl into disulphide bridge. Unexpectedly, LFUS denaturation increases protein solubility in water [bib_ref] Modifications of physicochemical and functional properties of amaranth protein isolate (Amaranthus cruentus..., Constantino [/bib_ref] [bib_ref] Effects of high-intensity ultrasound pretreatment with different levels of power output on..., Li [/bib_ref] [bib_ref] Effect of ultrasound treatments on functional properties and structure of millet protein..., Nazari [/bib_ref] [bib_ref] Effects of sweeping frequency ultrasound pretreatment on the hydrolysis of zein: angiotensin-converting..., Ren [/bib_ref] [bib_ref] Effects of multi-frequency ultrasound on physicochemical properties, structural characteristics of gluten protein..., Zhang [/bib_ref] [bib_ref] Conformational and physicochemical properties of quinoa proteins affected by different conditions of..., Vera [/bib_ref]. Nazari et al. [bib_ref] Effect of ultrasound treatments on functional properties and structure of millet protein..., Nazari [/bib_ref] suggested that sonication leads to the emergence of hidden residues with a negativelycharged side chain, thus explaining the more negative zeta potential (25% and 14%) observed by Vera et al. [bib_ref] Conformational and physicochemical properties of quinoa proteins affected by different conditions of..., Vera [/bib_ref] and Zhang et al., respectively. Alterations in spatial arrangement improve the interfacial properties of the protein because expose residues in accordance with polarity, thus acting like surfactants [bib_ref] Modifications of physicochemical and functional properties of amaranth protein isolate (Amaranthus cruentus..., Constantino [/bib_ref]. LFUS-treated wheat gluten increased foam capacity (+138%), foam stability (+42-118%), emulsion activity index (2-fold) and emulsion stability (>3-fold) [bib_ref] Mechanism study of multimode ultrasound pretreatment on the enzymolysis of wheat gluten, Zhang [/bib_ref]. Similarly, LFUS improved millet protein foaming capacity (2.75-fold), foam stability (22-fold), emulsion activity index (almost 2-fold) and emulsion stability (4.4-fold) [bib_ref] Effect of ultrasound treatments on functional properties and structure of millet protein..., Nazari [/bib_ref]. By emulsifying finer oil droplets in water, LFUS enhanced wheat proteins foaming even better than Tween 80 [bib_ref] The effect of ultrasound upon the physicochemical and emulsifying properties of wheat..., O&apos;sullivan [/bib_ref]. Recently, Zhang et al.produced a high internal phase emulsion, which mimics a solid fat, using quinoa protein nanoparticles as an emulsifier whose interfacial properties were adjusted by varying the ultrasonic density. The low-power high-frequency ultrasound has often been utilized as pre-treatment to produce bioactive peptides from cereals. The LFUSassisted enzymolysis was very effective at enhancing the ACE inhibitory activity of peptides by 8-99% . Protein denaturation, as mentioned above, leads to enhanced exposure of the enzymatic cleavage sites, improving the quality of the digested products, because hydrophobic-ending peptides have higher ACE inhibitory activity [bib_ref] The use of ultrasound for enzymatic preparation of ACE-inhibitory peptides from wheat..., Jia [/bib_ref] [bib_ref] Effects of ultrasound pretreatment with different frequencies and working modes on the..., Yang [/bib_ref] [bib_ref] In situ and real-time monitoring of an ultrasonic-assisted enzymatic hydrolysis process of..., Luo [/bib_ref]. Furthermore, cavitation disaggregates protein-starch complexes, thus exposing larger attack surfaces [bib_ref] Immobilized enzymolysis of corn gluten meal under triple-frequency ultrasound, Qu [/bib_ref]. Enzymatic affinity is increased, because of the reduction (13-42%) in Michaelis-Menten constant [bib_ref] Effects of ultrasound pretreatment on the enzymolysis and structural characterization of wheat..., Zhang [/bib_ref] [bib_ref] The use of ultrasound for enzymatic preparation of ACE-inhibitory peptides from wheat..., Jia [/bib_ref] [bib_ref] Immobilized enzymolysis of corn gluten meal under triple-frequency ultrasound, Qu [/bib_ref] [bib_ref] Ultrasonic treatment effect on enzymolysis kinetics and activities of ACE-inhibitory peptides from..., Wang [/bib_ref] causes higher initial hydrolysis rates [bib_ref] Effects of ultrasound pretreatment on the enzymolysis and structural characterization of wheat..., Zhang [/bib_ref] [bib_ref] The use of ultrasound for enzymatic preparation of ACE-inhibitory peptides from wheat..., Jia [/bib_ref] [bib_ref] Ultrasonic treatment effect on enzymolysis kinetics and activities of ACE-inhibitory peptides from..., Wang [/bib_ref] [bib_ref] Effect of multifrequency counter-current S type ultrasound pretreatment on the enzymatic hydrolysis..., Musa [/bib_ref]. The affinity towards the substrate may also be raised by conformational changes in the enzymes [bib_ref] Immobilized enzymolysis of corn gluten meal under triple-frequency ultrasound, Qu [/bib_ref]. ## Dietary fibre Ultrasound modification of dietary fibre was experimented by Hassan et al. [bib_ref] Ultrasound-assisted modification of insoluble dietary fiber from chia (Salvia hispanica L.) seeds, Hassan [/bib_ref] on the insoluble fraction pre-extracted from chia seeds, and by Wei et al. [bib_ref] Effects of high-temperature, high-pressure, and ultrasonic treatment on the physicochemical properties and..., Wei [/bib_ref] on purified soluble fraction from millet bran: the LFUS treatments produced about two-fold increases of both water and oil holding capacity. Those effects were attributed to the shear stress leading to disruption of structure and fragmentation of particles, with the consequent increase of hydrophilic functional groups, but also to a greater access to hydrophobic cavities where oil can be retained. LFUS capability to shatter dietary fibre was evidenced by Vaitkeviciene et al. [bib_ref] Functionalisation of rice bran assisted by ultrasonication and fermentation for the production..., Vaitkeviciene [/bib_ref] : they ultrasonicated rice bran and observed an average 10% decrease of insoluble fibre, and an increase of the soluble fraction; formation of resistant starch, likely due to chains rearrangement, was also detected. Similarly, Zadeike et al. [bib_ref] A comparative study on the structural and functional properties of water-soluble and..., Zadeike [/bib_ref] , documented a coarse surface with multiple fractures in LFUS-treated rice bran and an increase in its absorption of water and oil. Li et al. [bib_ref] The influence of ultrasonic modification on arabinoxylans properties obtained from wheat bran, Li [/bib_ref] observed a reduction in the number-average molecular weight of wheat bran arabinoxylan (AX) treated with LFUS (120-160 W for 15-45 min) and hypothesized a power and time-dependent chainbreaking effect because of cavitation, while Fan et al. [bib_ref] Improvement of Chinese noodle quality by supplementation with arabinoxylans from wheat bran, Fan [/bib_ref] confirmed the LFUS role in the reduction of AX molecular weight; both group of authors suggested that this could represent an alternative way to regulate the rheological properties of AX fraction. In addition, AX covalently binds phenolic acids [bib_ref] The influence of ultrasonic modification on arabinoxylans properties obtained from wheat bran, Li [/bib_ref] , thus its heightened depolymerization could imply a transition of phenolic compounds from insoluble to soluble forms. In fact, Fan et al. [bib_ref] Improvement of Chinese noodle quality by supplementation with arabinoxylans from wheat bran, Fan [/bib_ref] observed an increase in ferulic acid after sonication, but being AX a major component of wheat bran [bib_ref] Contents of dietary fibre components and their relation to associated bioactive components..., Andersson [/bib_ref] this phenomenon deserves further, more detailed examination. # Conclusions Low-frequency ultrasound (LFUS) is a promising green technology for improving key processing steps in cereals and pseudocereals as well as for modifying structural, physical, chemical, technological, functional, and biological properties of macromolecules such as carbohydrates and proteins. Moreover, in addition to saving energy and reducing processing time in agreement with its environment-friendly nature, LFUS allows to improve many extraction processes and to obtain value-added ingredients from cereal and pseudocereal byproducts in the frame of a circular economy. LFUS enhances the hydration rate and the time lag phase during pretreatments essential for cooking, extraction, fermentation, and germination of cereals and pseudocereals. Additionally, it improves and accelerates sprouting by increasing hydration, which in turn releases promoters and eliminates inhibitors of germination. Therefore, LFUS could be easily employed to speed-up the synthesis of bioactive compound (e.g., phenolics) in cereals and pseudocereals during the germination under stress-inducing condition that promote the production of antioxidants. Additionally, LFUS boosts the extraction rate of bioactive compounds, such as phenolic compounds and polysaccharides, is compatible with the use of some green solvents and improves the fortification with vitamin and minerals due to its ability to produce cracks and pores. However, the mechanism underlying such improvements still need be investigated in detail to increase scientific knowledge and optimize LFUS treatments. An important property of LFUS is that can improve and regulate the technological properties of fundamental food macromolecules like carbohydrates (starch and dietary fibre) and proteins. Peptides produced from the enzymatic hydrolysis of proteins denatured by LFUS present enhanced bioavailability and bioactivity. Therefore, LFUS may represent an efficient and viable alternative to produce nanoparticles of proteins or carbohydrates and bioactive compounds with improved bioavailability and bioactivity. Nevertheless, to scale-up similar processes to future industrial applications, LFUS conditions should be carefully assessed and measured by calorimetric and chemical dosimetry, rarely studied in cereals and pseudocereals. Although LFUS has demonstrated numerous possible applications in cereals and pseudocereals, its use in manufacturing has been scarcely developed. The challenge currently is to go beyond basic research and to transfer the promising lab results to pilot and industrial scale. ## Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper [fig] Figure 1: Forest plot for random-effects meta-analysis of extraction yield: LFUS-assisted extraction vs. a similar control group of non-US-assisted extraction or vs. traditional extraction. The projection of each point on the abscissa axis represents, if higher than 1, the n of an n-fold increase in the extraction yield. [/fig] [table] Table 1: Conditions tested in ultrasound processing for improving seed hydration and germination. [/table] [table] Table 2: Optimized experimental conditions and yields (mean ± SD) of LFUS-assisted extraction. Experiments marked with asterisk employed LFUS along with another major technique or lasted longer than the net LFUS treatment time. [/table] [table] Table 4: Conditions tested in ultrasound processing for modifying cereal proteins. [/table]

Prevalence of chronic kidney disease stages 3–5 in low- and middle-income countries in Asia: A systematic review and meta-analysis Abbreviations: eGFR, estimated glomerular filtration rate; CKD, chronic kidney disease; EPI, epidemiology collaboration equation; MDRD186, Modification of Diet in Renal Disease Study with constant factor of 186; MDRD175, Modification of Diet in Renal Disease Study with constant factor of 175; N/A, not applicable from studies. . Characteristic of studies included in the meta-analysis. ## Author Year Country [table] Table S4: Newcastle-Ottawa quality assessment scale (NOS) of included studies in the Systematic Review: A cross-sectional studies. Comparability of subjects in different outcome groups on the basis of design or analysis. Confounding factors controlled. [/table] [table] Table S5: Newcastle-Ottawa quality assessment scale (NOS) of included studies in the Systematic Review: A cohort studies Study quality Scoring: High = 7-9 points, Moderate= 4-6 points, and Poor = 0-3 points [/table]

Can Agricultural Productive Services Promote Agricultural Environmental Efficiency in China? # Introduction Along with the increasing attention paid to environmental issues, the agricultural environment has gradually become one of the most attractive research topics in the agricultural field in China. Agricultural environment efficiency, which brings environmental factors into traditional agricultural technical efficiency, is an effective index to balance agricultural development and agricultural environment. Especially since China has set forth the aims of carbon peak in 2030 and carbon neutrality in 2060, agricultural environmental efficiency based on carbon emissions and its determinants have caused increasing concern. Existing fruitful studies on the determinants of agricultural environmental efficiency are developed from two aspects of the external agricultural environment and the internal productive factors. In terms of the external environment, agricultural development level [bib_ref] Carbon emissions and driving forces of China's power sector: Input-output model based..., Luo [/bib_ref] [bib_ref] Ecoefficiency of China's agricultural sector: What are the spatiotemporal characteristics and how..., Zeng [/bib_ref] , urbanization, natural disasters, production risk, innovation [bib_ref] The influencing factors of CO 2 emission intensity of Chinese agriculture from..., Long [/bib_ref] , and energy price [bib_ref] Does the popularization of agricultural mechanization improve energy-environment performance in China's agricultural..., Jiang [/bib_ref] are taken into account. In terms of internal factors, agricultural structure [bib_ref] Ecoefficiency of China's agricultural sector: What are the spatiotemporal characteristics and how..., Zeng [/bib_ref] , agricultural labor force, 2 of 18 reservoir infrastructure construction, agricultural mechanization service organizations [bib_ref] Does the popularization of agricultural mechanization improve energy-environment performance in China's agricultural..., Jiang [/bib_ref] , as well as mechanization [bib_ref] Does the popularization of agricultural mechanization improve energy-environment performance in China's agricultural..., Jiang [/bib_ref] [bib_ref] Infante-Amate, J. From animals to machines. The impact of mechanization on the..., Aguilera [/bib_ref] [bib_ref] Does agricultural mechanization improve agricultural environment efficiency? Evidence from China's planting industry, Zhu [/bib_ref] are frequently mentioned. At the same time, due to the ubiquity of small-scale and fragmented farming, agricultural productive services adapted to China's national conditions have spread and popularized rapidly. Agricultural productive services are an essential driving force for improving agricultural productivity and promoting agricultural modernization [bib_ref] The paradox of developing agricultural mechanization services in China: Supporting or kicking..., Qiu [/bib_ref] , for its benefits of realizing economies of scale and reducing agricultural production costs without changing the social security functions of cropland [bib_ref] Determinants and impacts of outsourcing pest and disease management: Evidence from China's..., Sun [/bib_ref] [bib_ref] The impact of mechanization on crop production in China, Qiao [/bib_ref]. Nevertheless, agricultural productive services mainly rely on large machinery that consumes diesel and other fuels, emitting large amounts of carbon dioxide. It is worth noting that the fuel consumption of energy activity has been one of the most fundamental origins of agricultural carbon emissions [bib_ref] Relationship and integrated development of low-carbon economy, food safety, and agricultural mechanization, Li [/bib_ref]. However, few studies concentrate on whether the promotion of agricultural productive services exerts a negative impact on the agricultural environment and causes the variation of agricultural environmental efficiency. To date, it is still unclear whether agricultural productive services will lead to the improvement of the agricultural environment. Therefore, more imperative exploration of whether agricultural productive services improve agricultural environmental efficiency from a holistic evaluation perspective can not only conduce to understanding the formation and evolution of agricultural environmental efficiency, but also help to promote the relevant policies for the promotion and application of agricultural productive services. Agricultural environmental efficiency is determined by agricultural inputs, outputs, and carbon emissions [bib_ref] Does agricultural mechanization improve agricultural environment efficiency? Evidence from China's planting industry, Zhu [/bib_ref] , which are influenced by agricultural productive services. In re-ality, agricultural productive services, also known as agricultural production outsourcing, refer to outsourcing some or all agricultural production stages to service providers or other farmers [bib_ref] How to improve the welfare of smallholders through agricultural production outsourcing: Evidence..., Mi [/bib_ref]. Due to its incremental, overlapping, complementary advantages [bib_ref] The rapid rise of agricultural mechanization in Myanmar, Belton [/bib_ref] , agri-cultural productive services can help optimize agricultural factor inputs while keeping agricultural output increased or unchanged. This not only reduces agricultural production costs [bib_ref] Mechanization outsourcing clusters and division of labor in Chinese agriculture, Zhang [/bib_ref] , but also reduces carbon emissions by optimizing agricultural chemical inputs. In addition, more generally, the cost of agricultural productive services to a service provider is lower than the cost of labor to complete the same task because of specialization. Accordingly, through the adoption of productive services, farmers can not only save production costs, but also save labor time to engage in non-agricultural work to obtain more income [bib_ref] Outsourcing Agricultural Production: Evidence from Rice Farmers in Zhejiang Province, Ji [/bib_ref]. Moreover, agricultural production services embed green production techniques and green production materials in farmers' production processes, change their customs and experience with fertilizer application, and thus achieve a reduction in agricultural chemical materials. In addition, agricultural productive services have a mix of benefits, such as increasing the speed of operations, enhancing the timeliness of crucial production stages, improving the ability to cope with weather-related risks, and reducing losses in the harvest process, which exert positive effects on agricultural output [bib_ref] The rapid rise of agricultural mechanization in Myanmar, Belton [/bib_ref]. revealed that agricultural outsourcing had a significant impact on farmers' crop income and net crop income in Zimbabwe [bib_ref] Assessment of an outsourced agricultural extension service in the Mutasa district of..., Machila [/bib_ref]. Agricultural outsourcing contributes significantly and substantially to household crop income and the net income of farmers who participated in the program of outsourced extension service [bib_ref] Impact of outsourced agricultural extension program on smallholder farmers' net farm income..., Baiyegunhi [/bib_ref]. From the perspective of economics, farmers share land operation rights with service providers by purchasing socialized agricultural services, which benefit from specialization arising from the division of labor [bib_ref] Mechanization outsourcing clusters and division of labor in Chinese agriculture, Zhang [/bib_ref] and will inevitably affect agricultural production efficiency and agricultural environmental efficiency. Agricultural productivity increased by 25.61% for China's farmer households who chose agricultural productive services, and the productivity would increase by 10.86% if non-outsourcing farmer households chose to outsource [bib_ref] Does outsourcing affect agricultural productivity of farmer households? Evidence from China, Deng [/bib_ref]. In summary, the literature on the impact of agricultural productive services on agricultural inputs and outputs are growing. However, the research on agricultural environmental impact is still scarce, which needs further theoretical analysis and empirical verification. In particular, carbon emissions are the inevitable product of modern agricultural production for null-jointness and weak disposability imposed on agricultural production. It means that the impact of agricultural productive services on agricultural environmental efficiency requires not only further theoretical analysis but also more micro-empirical results for verification. To fill this observed deficiency, this paper empirically examines the effects of productive agricultural services on agricultural environmental efficiency. Furthermore, comprehensive theoretical analyses and empirical tests are needed to explain the environmental effects of agricultural productive services, i.e., the mechanisms by which agricultural productive services contribute to agricultural environmental efficiency. To make up for this theoretical gap, this paper presents a comprehensive overview of the possible influencing pathways, and empirically validates some of them. In these contexts, based on the panel data of 30 provinces in mainland China from 2004 to 2019, this paper investigates the role of agricultural productive services in improving agricultural environmental efficiency. The findings indicate that agricultural productive services significantly improve agricultural environmental efficiency, which holds steady after endogeneity treatment and a series of robust analyses. Meanwhile, the impact of agricultural productive services on agricultural environmental efficiency shows a marginal decreasing trend. Further research conclusions suggest that the spread of agricultural productive services affects both agricultural inputs and outputs through promoting technology progress, changing cropping structure, and optimizing factor input structure. The above pathways have significant spatial spillover effects. This paper contributes to the existing research from the following three aspects. First, we examine the heterogeneity of agricultural productive services affecting agricultural environmental efficiency based on accurately measuring agricultural environmental efficiency, expanding not only the study of the effects of agricultural productive services, but also the analysis of the factors influencing agricultural environmental efficiency. Second, we develop an analytical framework for the impact of agricultural productive services on agricultural environmental efficiency, namely that agricultural productive services affect agricultural environmental efficiency through three aspects, i.e., inputs, outputs, and environmental factors. Third, we propose the main mechanisms through which agricultural productive services affect agricultural environmental efficiency at a theoretical level and test empirically through the causal steps approach for the mediating effect test. The remainder of this paper is structured as follows: the "Methodology and Data" section describes analysis framework, empirical methods, and the nature of data. Econometric results and discussion are presented in the "Results" and "Discussion" sections, and the "Conclusions" section sets out the main conclusions and some policy implications. # Methodology and data ## Analysis framework of agricultural productive services affecting agricultural environmental efficiency In order to reveal the relationship between agricultural productive services and agricultural environmental efficiency in depth, we attempt to explore the influencing pathways of agricultural productive services on agricultural environmental efficiency from four aspects. Firstly, agricultural environmental efficiency is influenced by agricultural productive services through agricultural technology progress. Most farmers in China face high labor costs and credit constraints, making it challenging to purchase advanced and costly agricultural machinery and choose the reduction scheme for agricultural chemical materials. Agricultural productive services are the best solutions to these problems by reducing the cost of purchasing costly agricultural equipment and lowering the threshold for adopting advanced agricultural technology. On the one hand, the agricultural technology progress, which contains advanced farming technology and advanced energy technology, conduces to improve energy efficiency and directly reduces carbon emissions from energy consumption. On the other hand, agricultural technology progress can indirectly reduce carbon emissions by optimizing the structure of agricultural energy and fertilizer consumption. For example, the growing use of renewable energy can reduce the carbon emission intensity per unit of energy, and the increasing use of new biological fertilizers can increase the carbon sequestration capacity of soil. Secondly, agricultural environmental efficiency is influenced by agricultural productive services through agricultural planting structure adjustment. Agricultural productive services help to accelerate the large-scale development of agriculture, promote the proportion of grain sown area [bib_ref] Why are Chinese small farmers planting more grain crops?-Review of the Logic..., Zhang [/bib_ref] , and achieve the agglomeration of planting varieties. On the one hand, research shows that, compared with horticultural crops and cash crops, grain crops require the least input, including labor and fertilizer [bib_ref] Why are Chinese small farmers planting more grain crops?-Review of the Logic..., Zhang [/bib_ref]. Consequently, the increase in the proportion of grain crops sown has led to a decrease in the demand for agricultural chemical materials (i.e., pesticides, agricultural films, and fertilizers) per unit area. On the other hand, grain crops are inclined to be produced by large machinery, resulting in an increased demand for diesel fuel consumption. Furthermore, grain crops generally grow more organic matter (i.e., fruit and straw) than cash crops such as vegetables and flowers, and thus have a more substantial carbon sink effect. Therefore, the bigger the cultivated area of grain crops, the higher agricultural environmental efficiency. Thirdly, agricultural environmental efficiency is influenced by agricultural productive services through factor allocation optimization. Agricultural productive services are accompanied by the spatial flow of mechanical resources and the transmission of beneficial agricultural information, and thus realize the effective allocation of regional resources [bib_ref] How to improve the welfare of smallholders through agricultural production outsourcing: Evidence..., Mi [/bib_ref]. The popularization of agricultural productive services is accompanied by a shift of agricultural labor to non-agricultural sectors, so the outflow of agricultural labor will strengthen the scarcity of labor as a primary production factor, and then improve the willingness of farmers to use other capitals as a substitute for labor in agricultural production. The decreasing price of machinery relative to labor has led to a decrease in labor input intensity in farm production and an increase in fertilizers, pesticides, and other agricultural chemicals per unit area. Taking fertilizers for example, the outflow of agricultural labor may reduce the frequency of fertilization, but increase the amount of a single fertilization. In addition, the increased use of machinery requires more fuel consumption and brings more carbon emissions, leading to changes in agricultural environmental efficiency. Agricultural productive services contribute to increased specialization in all segments of the agricultural industry chain. The gains from specialization that arise from the division of labor promote the efficiency of agricultural production and realize economies of scale. At the regional level, because of the formation and expansion of productive service markets, agricultural production factors can fully flow and be exchanged between different agriculture operators with the help of agricultural productive services. In other words, agricultural productive services have realized the change of varying factor combinations. The improvement of resource utilization rate eases the excessive use of agricultural chemical resources, and thus leads to the change in carbon emissions. Fourthly, agricultural environmental efficiency is influenced by agricultural productive services through spatial spillover. From a regional perspective, the higher the level of agricultural productive services in a region (i.e., the higher proportion of persons engaged in productive services and the larger market for agricultural productive services), the greater the empowering effect on agricultural producers in the neighborhood. Relying on technological innovation, technology diffusion, specialized division of labor, collaboration, and the similarity of agricultural production resource endowment conditions in adjacent regions, agricultural machinery, which has apparent diffusion and spillover, forms the agricultural agglomeration effect and enhances the network connection effect in neighboring areas. Based on the above analysis, we establish an analysis framework (shown in [fig_ref] Figure 1: Analysis framework. [/fig_ref] as follows. Based on the above analysis, we establish an analysis framework (shown in [fig_ref] Figure 1: Analysis framework. [/fig_ref] as follows. ## Empirical models ## Benchmark model To test the relationship between agricultural productive services and agricultural environmental efficiency, the ordinary least squares regression with fixed-effect panel model is employed: [formula] 01 it it j itj i t it j AEE APS Z       = + + + + + (1) [/formula] where it AEE represents the agricultural environmental efficiency of province i at year t, it APS denotes the level of agricultural productive services, it Z refers to control variables; 0  is the intercept term, 1 , j  are the estimation coefficients of explanatory variable and control variables, i  and t  represent the fixed effects of province and year, and it  is a stochastic error term. ## Causal steps approach for mediating effect test To explore the pathways of agricultural productive services affecting agricultural environmental efficiency, based on the test method of [bib_ref] The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical..., Baron [/bib_ref] [bib_ref] The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical..., Baron [/bib_ref] , the causal steps approach for mediating effect test is set as follows: [formula] 01 it it j itj i t it j M APS Z       = + + + + +  (2) 0 1 2 it it it j itj i t it j AEE APS M Z        = + + + + + + (3) [/formula] where it M represents the mediating variables, 00 ,  are the intercept terms, and [formula] 1 1 2 [/formula] , , , j     are the estimation coefficients of corresponding variables. Equation (2) is used to test the effect of the independent variable on the mediating variables, and Equation (3) is used to test the effect of the independent variable on the dependent variable after introducing mediating variables. If the regression coefficient on agricultural productive services decreases or becomes insignificant, it indicates that the impact of agricultural productive services on agricultural environmental efficiency comes partly or entirely through the pathway of mediating variable. ## Empirical models ## Benchmark model To test the relationship between agricultural productive services and agricultural environmental efficiency, the ordinary least squares regression with fixed-effect panel model is employed: [formula] AEE it = α 0 + α 1 APS it + ∑ j λ j Z itj + µ i + η t + ξ it (1) [/formula] where AEE it represents the agricultural environmental efficiency of province i at year t, APS it denotes the level of agricultural productive services, Z it refers to control variables; α 0 is the intercept term, α 1 , λ j are the estimation coefficients of explanatory variable and control variables, µ i and η t represent the fixed effects of province and year, and ξ it is a stochastic error term. ## Causal steps approach for mediating effect test To explore the pathways of agricultural productive services affecting agricultural environmental efficiency, based on the test method of [bib_ref] The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical..., Baron [/bib_ref] [bib_ref] The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical..., Baron [/bib_ref] , the causal steps approach for mediating effect test is set as follows: [formula] M it = β 0 + β 1 APS it + ∑ j λ j Z itj + µ i + η t + ξ it (2) AEE it = γ 0 + γ 1 APS it + γ 2 M it + ∑ j λ j Z itj + µ i + η t + ξ it(3) [/formula] where M it represents the mediating variables, β 0 , γ 0 are the intercept terms, and β 1 , γ 1 , γ 2 , λ j are the estimation coefficients of corresponding variables. Equation (2) is used to test the effect of the independent variable on the mediating variables, and Equation (3) is used to test the effect of the independent variable on the dependent variable after introducing mediating variables. If the regression coefficient on agricultural productive services decreases or becomes insignificant, it indicates that the impact of agricultural productive services on agricultural environmental efficiency comes partly or entirely through the pathway of mediating variable. ## Spatial econometric model We employ the Spatial Dubin Model with fixed effects to verify the spatial spillover effect of agricultural productive services on agricultural environmental efficiency: [formula] AEE it = α 0 + τ AEE i,t−1 + ρWAEE it + α 1 APS it + α 2 WAPS it + ∑ j λ j Z itj + µ i + η t + ξ it (4) [/formula] where W represents the spatial weight matrix, and geographical distance matrix is adopted in the paper. τ is the first-order lag coefficient of dependent variable, ρ is the spatial correlation coefficient, and α 1 , α 2 , λ j are the estimated coefficients of each explanatory variable. ## Endogeneity and two-stage least squares Biases might remain in empirical models because of two types of endogenous problems. The first is the omitted variable problem, i.e., some variables that affect both agricultural productive services and agricultural environmental efficiency are omitted from the regression model. The second is reverse causality, i.e., agricultural environmental efficiency might affect agricultural productive services. For example, provinces with high agricultural environmental efficiency level are usually regions with the high level of factor endowments. These regions, due to the high level of regional development and high price relative to labor, will adapt to local agricultural development requirements through large-scale adoption of agricultural productive services, which in turn contribute to the improvement of their agricultural productive services. This means that the core independent variable and the dependent variable suffer from reverse causality. To solve the problem of possible omitted variables, we employ a fixed-effect model with panel data to control the unobservable effects at provincial and time level, while adding as many control variables as possible in the regression analysis based on existing studies, such as rural human capital and production risk. To solve the problem of reverse causality, we introduce a suitable instrumental variable and employ two-stage least squares method to alleviate endogenous problems. The first stage is used to obtain prediction to replace endogenous variables, and the second stage is used to draw the final estimation result. ## Variables 2.3.1. dependent variable: agricultural environmental efficiency Following Zhu et al. [bib_ref] Does agricultural mechanization improve agricultural environment efficiency? Evidence from China's planting industry, Zhu [/bib_ref] , we adopt the multi-output stochastic frontier analysis method, which is based on the output-oriented distance function, to obtain agricultural environmental efficiency. In terms of the production function form, we introduce a time-varying parameter model to estimate elastic changes across time accurately. Drawing on the traditional literature [bib_ref] A non-parametric bootstrap-data envelopment analysis approach for environmental policy planning and management..., Toma [/bib_ref] [bib_ref] Agricultural reforms and production in China: Changes in provincial production function and..., Gong [/bib_ref] , input variables and output variables are selected to calculate agricultural environmental efficiency. The main input variables of agricultural production are labor, machinery, fertilizer, land and fuel, which are measured by the number of employees (in millions), the total power of planting machinery (in million kilowatts), the sum of the gross weight of various fertilizers (in million tons), the sown area (in million hectares) reflecting the actual utilization of the cultivated land, and diesel oil (in million tons) in the planting industry, respectively. We employ the gross value of output (in million CNY) and net carbon sinks (in million tons of CO 2 -equivalent) in the planting industry as the output variables because agriculture has the attribute of net carbon sinks. The calculation formula, coefficient of agricultural carbon sinks, and carbon emissions are based on the research of [bib_ref] Does agricultural mechanization improve agricultural environment efficiency? Evidence from China's planting industry, Zhu [/bib_ref]. The descriptive statistics on agricultural input and output variables on provincial level are shown in [fig_ref] Table 1: Descriptive statistics on agricultural input and output variables [/fig_ref] , and the changing trends of agricultural inputs and outputs on national level in China from 2004 to 2019 are shown in [fig_ref] Figure 2: The changing trend of agricultural inputs and outputs in China from 2004... [/fig_ref]. ## Independent variable: agricultural productive services From the point of view of the demand side (i.e., the objects of agricultural productive services), the sown area and the number of farmers adopting productive services are the ideal indexes to measure the level of productive services; from the point of view of the supply side, the supplier (i.e., the number of agricultural machinery service organizations and agricultural machinery professional service households or organizations) and the power of agricultural machinery for productive services are reasonable indicators. Limited by data availability, we employ the number of people engaged in agricultural productive services per unit area to measure the level of agricultural productive services. ## Independent variable: agricultural productive services From the point of view of the demand side (i.e., the objects of agricultural productive services), the sown area and the number of farmers adopting productive services are the ideal indexes to measure the level of productive services; from the point of view of the supply side, the supplier (i.e., the number of agricultural machinery service organizations and agricultural machinery professional service households or organizations) and the power of agricultural machinery for productive services are reasonable indicators. Limited by data availability, we employ the number of people engaged in agricultural productive services per unit area to measure the level of agricultural productive services. ## Instrumental variable To solve the problem of endogeneity, the usual approach is to introduce a suitable instrumental variable. Following Zhu et al. [bib_ref] Does agricultural mechanization improve agricultural environment efficiency? Evidence from China's planting industry, Zhu [/bib_ref] , this paper adopts the ratio of road mileage to farmland area in the region as an instrumental variable for agricultural productive services. As the transportation infrastructure, roads do not directly impact agri-cultural output value and net carbon sinks, but they can improve the degree of agricultural productive services via improving road conditions and reducing transportation costs for agricultural machinery. ## Mediating variables Based on the analysis in the "Literature Review" section, technology progress, planting structure, and input structure are selected to study the influencing mechanism of agricultural productive services affecting agricultural environmental efficiency. Agricultural machinery, which contains advanced planting and harvesting techniques, is accompanied by the transfer and diffusion of management techniques in its service process. Consequently, technology progress is characterized by the ratio of the total power of machinery to farmland area. There are significant differences between cash crops and grain crops in mechanical demand and agricultural chemical materials input [bib_ref] Does the popularization of agricultural mechanization improve energy-environment performance in China's agricultural..., Jiang [/bib_ref] , so the planting structure is represented by the ratio of grain sowing area to crop sowing area. The main agricultural inputs include machinery, farmland, labor, and chemical fertilizer. Considering the rele-vance of machinery and the stability of farmland, the ratio of fertilizer to labor per unit area is used to characterize the input structure. ## Control variables Agricultural operation scale, planting industry development level, rural human capital, production risk, regional economic development level, part-time employment of labor, and urban-rural income gap are controlled in our models. Agricultural operation scale (in hectare per household), which is an essential factor affecting the consumption of farm machinery and chemical materials [bib_ref] Relationship and integrated development of low-carbon economy, food safety, and agricultural mechanization, Li [/bib_ref] , is represented by the ratio of farmland to rural households, and its square term is introduced to examine the possible threshold of land. Based on the viewpoint of comparative advantage, the development level of the planting industry is measured by the share of the planting industry in agriculture. As a factor with strong positive externalities, rural human capital (in years) is calculated as the average length of schooling of the rural population. As farmers will increase fertilizer and pesticide inputs to avoid risks [bib_ref] Factors influencing Chinese farmers' proper pesticide application in agricultural products-A review, Pan [/bib_ref] , production risk is selected as the control variable, which refers to the ratio of the affected area to the total sown area of crops. Regional economic development level (in CNY per person), which affects carbon emissions intensity [bib_ref] Carbon emissions and driving forces of China's power sector: Input-output model based..., Luo [/bib_ref] , is expressed by the ratio of a region's GDP to its population. As an essential factor affecting the capital inputs (i.e., machine, chemical fertilizer, and pesticides) in agricultural production [bib_ref] Relationship and integrated development of low-carbon economy, food safety, and agricultural mechanization, Li [/bib_ref] , the part-time employment of labor is measured as the proportion of wage income of rural residents. The urban-rural income gap, which leads to the flow of the labor to cities [bib_ref] Agricultural inputs, urbanization, and urban-rural income disparity: Evidence from China, Wang [/bib_ref] , is represented by the ratio of disposable income of urban residents to rural residents. In addition, due to the complexity of agricultural production, agricultural environmental efficiency will be affected by many unobserved factors, such as social culture and natural conditions [bib_ref] A critical assessment of provincial level variation in agricultural GHG emissions in..., Han [/bib_ref] , so provincial and year fixed effects are controlled too. The descriptive statistics on variables are shown in [fig_ref] Table 2: Descriptive statistics on variables [/fig_ref]. The differences of agricultural environmental efficiency and agricultural productive services among different provinces are presented in [fig_ref] Figure 3: Kernel density of agricultural environmental efficiency [/fig_ref] , showing that there are great differences between samples, which are suitable for regression analysis. ## Data # Results # Benchmark regression results The regression results in columns (1) and (2) of present the parameter estimates for the benchmark regression based on Equation (1). Among them, column (1) does not include control variables, while column (2) adds control variables, with both columns controlling province and time differences. As shown in column (1), the coefficient on agricultural productive services is positive and statistically significant at the 1% level, indicating that the development of agricultural productive services promotes the level of agricultural environmental efficiency. However, the analysis is likely to be confounded by omitted variables. After introducing control variables, the results in column [bib_ref] Ecoefficiency of China's agricultural sector: What are the spatiotemporal characteristics and how..., Zeng [/bib_ref] show that the coefficient on agricultural productive services is still positive and statistically signifi- # Results # Benchmark regression results The regression results in columns (1) and (2) of present the parameter estimates for the benchmark regression based on Equation (1). Among them, column (1) does not include control variables, while column (2) adds control variables, with both columns controlling province and time differences. As shown in column (1), the coefficient on agricultural productive services is positive and statistically significant at the 1% level, indicating that the development of agricultural productive services promotes the level of agricultural environmental efficiency. However, the analysis is likely to be confounded by omitted variables. After introducing control variables, the results in column [bib_ref] Ecoefficiency of China's agricultural sector: What are the spatiotemporal characteristics and how..., Zeng [/bib_ref] show that the coefficient on agricultural productive services is still positive and statistically significant at the 1% level. To be explicit, the augmentation of agricultural productive services by 1% leads to an increase of 7.16% in agricultural environmental efficiency. This means that, despite the increasing effect in carbon emissions caused by energy consumption, the increasing carbon sink effect, which is brought about by the increase of output and the reduction of agricultural chemical inputs, is more potent. In fact, the carbon emissions of China's planting industry began to decline after 2016, but the carbon sinks have been increasing, showing an obvious net carbon sink effect. . Impact of agricultural productive services on agricultural environmental efficiency. ## Variables ols 2sls (1) Note: **, ***: statistically significant at 5% and 1%, respectively; Standard error in parentheses. also identifies some statistically significant control variables. An inverted U-shaped relationship exists between farmland operation scale and agricultural environmental efficiency. With the expansion of farmland operation scale, agricultural environmental efficiency gradually improves, which is unsurprising as the appropriate expansion of farmland operation scale is conducive to the rational allocation of production factors by agricultural producers and the reduction of material waste. When farmland operation scale exceeds the threshold value, agricultural environmental efficiency declines. The reason lies in the difficulty of managing and realizing the maximum potential if the farmland per capita is too large. There is a significant positive relationship between rural human capital and agricultural environmental efficiency at a 1% level. This is due to the considerable positive externality of human capital, which has a significant incremental improvement on environmental quality. Production risk is significantly and negatively related to agricultural environmental efficiency. The greater the frequency and degree of natural disasters, the greater the production risk perceived by farmers, and the more likely it is to use agricultural chemical materials to avoid risks and reduce losses, resulting in the decline of environmental efficiency. The high degree of part-time labor employment and the low degree of urban-rural income gap improve agricultural environmental efficiency through higher demand for productive services. ## Two-stage least squares results Columns (3) and (4) of present the parameter estimates for the two-stage least squares method. The results of the first stage show a significant positive correlation between the instrumental variable and agricultural productive services, whether with or without control variables, which is consistent with the hypothesis of correlation of instrumental variables. The results of the second stage regression showed that agricultural productive services still had a significant contribution to agricultural environmental efficiency, which confirmed the robustness of the benchmark regression results. # Robust analysis results To further verify the robustness of the impact of productive agricultural services on agricultural environmental efficiency, we conduct the robustness analysis from three aspects: replacing the dependent variable, replacing the core independent variables, and subdividing the sample into main and non-main grain-producing areas. Firstly, we recalculate agricultural environmental efficiency using the translog production function and re-estimate the impact of agricultural productive services. The results in column (1) of [fig_ref] Table 4: Results of robust analysis [/fig_ref] presents that the effect of agricultural productive services on agricultural environmental efficiency remains positive and significant at a 1% level. Note: **, ***: statistically significant at 5% and 1%, respectively; Standard error in parentheses. Secondly, we adopt the number of productive service institutions per unit area to express the level of agricultural productive services. The results in column (2) of [fig_ref] Table 4: Results of robust analysis [/fig_ref] show that the independent variable has a positive correlation with agricultural environmental efficiency with significance at a 5% level, which further confirms the contribution of agricultural productive services to agricultural environmental efficiency. Thirdly, we divide the samples into main and non-main grain-producing areas. There are 13 provinces in main grain-producing areas, such as Heilongjiang and Henan, with a high level of agricultural productive services. The coefficients on agricultural productive services are positive and statistically significant at least at a 5% level in columns (3) and (4) of [fig_ref] Table 4: Results of robust analysis [/fig_ref] , but differ a lot in the main grain-producing areas and non-main grain-producing areas. The coefficient on agricultural productive services is 0.1394 in non-main grainproducing areas, which is more significant than that of main grain-producing areas. This means that the marginal effect of agricultural productive services is somewhat larger in non-main grain-producing areas, i.e., agricultural productive services in non-main grainproducing areas are still in the early development stage and have ample space for growth. # Mechanism analysis results To examine the transmission mechanism of agricultural productive services to agricultural environmental efficiency, we employ Equations (2) and (3) for mediating effect tests from three perspectives of technology progress, planting structure, and input structure, and adopt the spatial econometric model for the spatial spillover effect test. [fig_ref] Table 5: Results of the causal steps approach for mediating effect test [/fig_ref] reports the results of the mechanism analysis. The regression results of the first step are presented in columns (1), (2), and (3). The coefficients are positive and significant at a 1% level, indicating that the increase in agricultural productive services has significantly improved technology progress, and optimized the structure of planting crops and factor inputs. Column (4) exhibits the impact of agricultural productive services on agricultural environmental efficiency without mediating variables, and columns (5), [bib_ref] Infante-Amate, J. From animals to machines. The impact of mechanization on the..., Aguilera [/bib_ref] , and (7) demonstrate the regression results of the second step. As shown in column (5), after the inclusion of technology progress, the coefficient on agricultural productive services and its significance become lower, and the coefficient of technology progress is significant at a 5% level, indicating that the mediating mechanism of technology progress is verified. This is because technology progress is a fundamental driver of agricultural productivity and is a critical factor in reducing carbon emissions in agriculture [bib_ref] The influencing factors of CO 2 emission intensity of Chinese agriculture from..., Long [/bib_ref]. Similarly, the mediating pathways of planting structure and input structure are manifested in columns (6) and (7). Grain crops have a more substantial carbon sink effect and their demand for chemical materials is less than cash crops. Hence, the increase in the share of grain crops sown area has a positive effect on agricultural environmental efficiency. The optimization of inputs structure can effectively reduce the cost of agricultural production and contribute to the improvement of agricultural environmental efficiency. [fig_ref] Table 4: Results of robust analysis [/fig_ref] Note: *, **, ***: statistically significant at 10%, 5% and 1%, respectively; Standard error in parentheses. [fig_ref] Table 6: Results of the spatial econometric model [/fig_ref] reports the results of Equation (4) under the geographical distance matrix. The spatial coefficient and spatial spillover effects are significantly positive, showing that agricultural productive services have a significant spatial spillover effect on agricultural environmental efficiency, i.e., agricultural productive services in neighboring regions have a positive effect on agricultural environmental efficiency in this region. This reflects that agricultural productive services not only serve the region where they are located, but also radiate to the surrounding regions based on the similar crop structure and the different crop maturity time. Therefore, agricultural productive services not only contribute to local agricultural environmental efficiency, but also promote the efficiency of surrounding regions due to the spatial spillover effect caused by their cross-regional operation. Note: **, ***: statistically significant at 5% and 1%, respectively; Standard error in parentheses. # Discussion ## The relationship between agricultural productive services and agricultural environmental efficiency Under the context of sustainable development, it is valuable to investigate agricultural productivity in combination with environmental factors and analyze its critical influencing factors. Our empirical results show that agricultural productive services have a positive impact on agricultural environmental efficiency. Although the existing literature does not directly explore the relationship between them, there are two related types of literature for comparison, both of which have the common basis-mechanization. One is about agricultural mechanization and agricultural environmental efficiency, but the conclusions of their relationship are not consistent. Jiang et al. believed that mechanization led to the decline of agricultural environmental performance based on the provincial data from 2000 to 2017 in China [bib_ref] Does the popularization of agricultural mechanization improve energy-environment performance in China's agricultural..., Jiang [/bib_ref]. In the study of Zhu et al., the argument that mechanization and agricultural environmental efficiency presented an inverted U-shaped relationship was verified based on the study sample of China's 30 provinces during 2001-2019 [bib_ref] Does agricultural mechanization improve agricultural environment efficiency? Evidence from China's planting industry, Zhu [/bib_ref]. The reason for this inconsistency may lie in the inconsistency of the measurement methods and indicators of agricultural environmental efficiency adopted by them. Jiang et al. employed the DEA method and selected carbon emissions and solid residues as environmental outputs [bib_ref] Does the popularization of agricultural mechanization improve energy-environment performance in China's agricultural..., Jiang [/bib_ref] , while Zhu et al. adopted the SFA method and selected net carbon sinks as environmental outputs [bib_ref] Does agricultural mechanization improve agricultural environment efficiency? Evidence from China's planting industry, Zhu [/bib_ref]. The other is about agricultural green production technologies and agricultural environmental efficiency. The majority of green production technologies in China rely on machinery and energy, which can cause excessive carbon emissions. In the research of He et al., the impact of different green production technologies on low-carbon efficiency is not uniform, and the results vary across regions [bib_ref] The role of agricultural green production technologies in improving low-carbon efficiency in..., He [/bib_ref]. For example, mechanized straw returning improves the low-carbon efficiency in the north, while mechanized deep ploughing and scarification improve low-carbon efficiency in the south. Agricultural environmental efficiency is determined by agricultural productivity and agricultural carbon emissions [bib_ref] The role of agricultural green production technologies in improving low-carbon efficiency in..., He [/bib_ref]. To be specific, agricultural inputs and outputs determine agricultural productivity, while agricultural carbon emissions affect agricultural environment [bib_ref] Does agricultural mechanization improve agricultural environment efficiency? Evidence from China's planting industry, Zhu [/bib_ref]. Agricultural productive services, which mainly rely on large machinery that consumes diesel and other fuels, not only affect agricultural performance through inputs and outputs, but also emit large amounts of carbon dioxide from energy consumption. From the perspective of input, as a capital input replacing labor, productive services significantly reduce labor costs and optimize agricultural chemical materials, but increase diesel consumption. From the standpoint of output, most studies confirm the positive effect of productive services on agricultural output [bib_ref] Determinants and impacts of outsourcing pest and disease management: Evidence from China's..., Sun [/bib_ref] [bib_ref] Why are Chinese small farmers planting more grain crops?-Review of the Logic..., Zhang [/bib_ref]. From the perspective of carbon emissions, agricultural chemical materials and energy consumption are the two primary sources of carbon emissions in agriculture. The optimization of the former leads to reducing carbon emissions, but the rising diesel consumption results in an increase in carbon emissions. Moreover, the increase effect of carbon emissions is greater than the reduction effect. The path of agricultural productive services impacting agricultural environmental efficiency is determined by two aspects. The first aspect is to increase agricultural productivity. Studies have confirmed that agricultural productive services are applied to maximize total yield and benefit potential production. At the macro level, Sheng and Chancellor found that the hiring of capital services may lift the productivity of small farms in the Australian grains industry [bib_ref] Exploring the relationship between farm size and productivity: Evidence from the Australian..., Sheng [/bib_ref]. At the peasant household level, Bravo-Ureta et al. analyzed the impact of a canal irrigation project for smallholders in the Philippines and found that irrigation technology promoted beneficiaries' production potential. Taking drought-tolerant maize as a research object in the Northern Region of Ghana, Martey et al. revealed that the climate-smart agricultural technologies positively impacted yield and commercialization intensity. The second aspect is to minimize the damage to the agricultural environment. For the planting industry, its attribute of net carbon sink dictates that we should maximize the net carbon sinks rather than minimize carbon emissions. From the perspective of carbon emissions, the adoption of agricultural productive services can reduce carbon emissions by lowering chemical inputs, but agricultural machinery and its fuel consumption, which support agricultural productive services, directly and indirectly cause carbon emissions. Moreover, in terms of outcome, the carbon emission effect of energy is greater than that of chemical input reduction. However, from the perspective of net carbon sinks, the planting industry showed a prominent positive net carbon sink characteristics and gradually increased. ## The mechanism of agricultural productive services affecting agricultural environmental efficiency Technology progress is an essential driving force for productivity growth, and the production frontier will continue to expand outward with technology progress. Taking paddy smallholders in the Philippines, for example, the frontier output was significantly promoted by improved irrigation technology. If there is no change in the productivity of the areas below the production frontier, the environmental efficiency values in these areas will become lower. However, our empirical results show that the environmental efficiency values are improved in almost all regions, suggesting that technology progress plays a role in all regions. Furthermore, the efficiency of the regions with low environmental efficiency rises faster than that of the regions with high environmental efficiency, suggesting that agricultural productive services embodying technology progress play a more significant role in the regions with low environmental efficiency. Agricultural productive services help to achieve the concentration of crop varieties, especially the increase in the proportion of the sown area of grain crops. In general, compared to cash crops, grain crops have higher productivity, relatively fewer chemical inputs, and relatively lower environmental impact. In particular, it is worth mentioning that the carbon absorption rate of grain crops is relatively high, which can easily form more carbon sinks. For example, the carbon absorption rates of wheat and maize are 0.485 and 0.471, respectively, which are higher than that of cash crops (the carbon absorption rate is 0.45). As a product and cost of agricultural production are dual problems, the optimization of inputs structure means that the cost can be minimized with constant output. We should seek measures to improve agricultural productivity from a cost perspective in the case that the grain output per unit area cannot be significantly increased. Hence, the improvement of agricultural productivity lies in optimizing cost, which depends on the optimal allocation of different input factors. The spread of agricultural productive services has effectively solved the limitation of labor scarcity and realized the rational allocation of resources through the gains from specialization that arise from the division of labor. On a regional scale, the neighboring regions often have similar crop varieties, and the difference in the distribution along the latitude makes the crop maturity have time differences, which provides a broad market space for agricultural productive services. Agricultural productive services have obvious diffusion and spillover, for they not only serve the local agricultural production, but also provide support for neighboring regions. From the viewpoint of micro-entities, farmers can learn advanced knowledge, technology, and experience from other farmers, agricultural cooperatives, and agricultural production service organizations in adopting agricultural productive services, and thus promote productivity improvement through positive externalities of learning. Employing panel data of 13 prefecture-level cities in Jiangsu province of China from 2000 to 2016, Wu et al. revealed that the spatial spillover effect of mechanization was significant on grain yield due to the cross-regional operation of mechanization [bib_ref] Threshold effect or spatial spillover? The impact of agricultural mechanization on grain..., Wu [/bib_ref]. There are several reasons for this paper does not explore the pathway of scale operation, which is analyzed in some other articles [bib_ref] Does agricultural mechanization improve agricultural environment efficiency? Evidence from China's planting industry, Zhu [/bib_ref]. First, the effect of scale operation is not the expansion of land operation scale, but the optimal combination of land and other endowment inputs of farmers. Scale operation at farmer's level is ultimately reflected in factors input structure, which has been analyzed above. Second, scale operation at the region's level is reflected in the change of planting structure. The more similar the adjustment of planting structure, the easier it to form a scale agglomeration effect in the region. Third, this paper considers the influence of land operation area among the control variables. To the best of our knowledge, this is the first attempt to reveal the effect of agricultural productive services on agricultural environmental efficiency and its influencing mechanism at the provincial level in China. ## Heterogeneity impact of different levels of productive agricultural services By comparing the effects of agricultural productive services on agricultural environmental efficiency in different regions in , we find that the marginal effect of agricultural productive services is more prominent in non-main grain-producing areas. Then, is this because the level of agricultural productive services is lower in non-main grain-producing areas? As a factor of production, do agricultural productive services satisfy the law of diminishing marginal returns? To test these assumptions, we further divide the level of agricultural productive services into three groups, conduct an econometric test (the results are shown in [fig_ref] Table 7: Heterogeneity impact of different levels of productive agricultural services [/fig_ref] , and find that the above inferences are valid. This means that, as an input factor, the input of agricultural productive services also conforms to the law of diminishing marginal returns, which is a fundamental law of short-term production. There is an optimal combination of productive services and other input factors. When the ratio is exceeded, the marginal benefit of increasing productive services decreases. Note: ***: statistically significant at 1%; Standard error in parentheses. # Conclusions Based on the panel data of 30 provinces in China from 2004 to 2019, we adopt a multi-output stochastic frontier analysis method to measure agricultural environmental efficiency based on net carbon sink, then explore the effects and mechanisms of agricultural productive services on agricultural environmental efficiency using OLS with fixed-effect panel data, two-stage least squares with instrumental variable, causal steps approach, and a spatial econometric method. The main conclusions are as follows: Firstly, agricultural productive services have a significant contribution to agricultural environmental efficiency. Agricultural productive services enhance agricultural productivity as well as the agricultural environment through inputs and outputs, and then improve the efficiency of the agricultural environment. This result still holds after using instrumental variables to deal with endogeneity, changing the measurement of dependent and independent variables, and subdividing the sample. Secondly, the mechanism pathways of agricultural productive services affecting agricultural environmental efficiency are mainly reflected in technology progress, planting structure adjustment, factor allocation optimization, and spatial spillover. Advanced agricultural technology and management measures can improve agricultural productivity and energy efficiency, and optimize the energy con-sumption structure. A reasonable planting structure is conducive to reducing agricultural chemical inputs, improving factor utilization efficiency, and realizing economies of scale. The rational allocation of factors is the fundamental guarantee for the role of agricultural input resources such as labor, farmland, and chemical fertilizer. Agricultural agglomeration and different crop maturity periods make cross-regional services of agricultural productive services possible, thus generating spatial spillover effects. Thirdly, the impact of agricultural productive services on agricultural environmental efficiency is more significant in the regions with low agricultural productive services level because agricultural productive services, as a factor input in agricultural production, conform to the law of diminishing marginal returns. Given the above evidence and arguments, we draw some policy implications. Firstly, policy attention should be paid to improving the level of agricultural productive services continuously according to the actual situation in different regions. For regions with low level of agricultural producer services due to terrain factors or crop structure factors, the marginal contribution of agricultural producer services is more pronounced. It is valuable to promote appropriate agricultural productive services in these areas to promote the quality of agricultural development and environmental efficiency. For regions with a high agricultural producer services level, such as main grain-producing areas, the marginal contribution of agricultural producer services is relatively small. The critical measure is to update or replace the existing productive service equipment, and promote the use of agricultural productive services with more advanced technology and higher efficiency. Secondly, efforts should be made to strengthen information diffusion and regional cooperation to realize the positive spatial spillover effect of agricultural productive services. For regions with similar resource endowments and similar crop types, information cooperation is used to guide and promote the cross-regional operation of agricultural productive services, to achieve rational flow and scientific allocation of agricultural machinery. Thirdly, it is necessary to improve the operating conditions of agricultural productive services, such as appropriately increasing the investment in transportation infrastructure, guiding the centralized farmland transfer, and realizing regional agglomeration of crop planting. Improving transportation infrastructure can provide primary conditions and convenience for agricultural machinery operations. Farmland transfer and crop agglomeration are conducive to realizing economies of land scale and service scale in agricultural production. The scale of expansion is beneficial to cultivating the agricultural productive services market and introducing small farmers to modern agriculture. There are also some potential limitations in this paper. First, in terms of the research scope, this study was conducted at the provincial level. Future research can be conducted at the county level, which potentially has richer data and more accurate results. Second, in terms of the research depth, as agricultural productive services exist in all aspects of agricultural production, such as tillage, sowing, irrigation, fertilization, and harvesting, a possible research direction is to profoundly investigate the impact of different stages of productive services on agricultural environmental efficiency. ## Institutional review board statement: not applicable. Informed Consent Statement: Not applicable. # Data availability statement: The datasets analyzed during the current study are available from the corresponding author on reasonable request. [fig] Figure 1: Analysis framework. [/fig] [fig] Figure 2: The changing trend of agricultural inputs and outputs in China from 2004 to 2019. [/fig] [fig] Figure 3: Kernel density of agricultural environmental efficiency. [/fig] [fig] Figure 4: Kernel density of agricultural productive services. [/fig] [fig] Author: Contributions: Conceptualization, Y.Z. (Yingyu Zhu); methodology, Y.Z. (Yingyu Zhu); validation, Y.Z. (Yingyu Zhu); formal analysis, Y.Z. (Yingyu Zhu); resources, M.W. and Y.T.; data curation, W.Y.; writing-original draft preparation, Y.Z. (Yingyu Zhu), M.W. and Y.T.; writing-review and editing, J.D.; visualization, Y.Z. (Yingyu Zhu); supervision, Y.Z. (Yan Zhang). All authors have read and agreed to the published version of the manuscript. [/fig] [fig] Funding: This research was funded by the Humanities and Social Sciences of the Ministry of Education of China (Grant No. 19YJA880083), the National Social Science Foundation of China (Grant No. 18BGL170), the National Natural Science Foundation of China (Grant No. 71703106), and the China Postdoctoral Science Foundation (Grant No. 2018M631823). [/fig] [table] Table 1: Descriptive statistics on agricultural input and output variables. [/table] [table] Table 2: Descriptive statistics on variables. [/table] [table] Table 4: Results of robust analysis. [/table] [table] Table 5: Results of the causal steps approach for mediating effect test. [/table] [table] Table 6: Results of the spatial econometric model. [/table] [table] Table 7: Heterogeneity impact of different levels of productive agricultural services. [/table]

Text mining to support abstract screening for knowledge syntheses: a semi-automated workflow Background: Current text mining tools supporting abstract screening in systematic reviews are not widely used, in part because they lack sensitivity and precision. We set out to develop an accessible, semi-automated "workflow" to conduct abstract screening for systematic reviews and other knowledge synthesis methods. Methods: We adopt widely recommended text-mining and machine-learning methods to (1) process title-abstracts into numerical training data; and (2) train a classification model to predict eligible abstracts. The predicted abstracts are screened by human reviewers for ("true") eligibility, and the newly eligible abstracts are used to identify similar abstracts, using near-neighbor methods, which are also screened. These abstracts, as well as their eligibility results, are used to update the classification model, and the above steps are iterated until no new eligible abstracts are identified. The workflow was implemented in R and evaluated using a systematic review of insulin formulations for type-1 diabetes (14,314 abstracts) and a scoping review of knowledge-synthesis methods (17,200 abstracts). Workflow performance was evaluated against the recommended practice of screening abstracts by 2 reviewers, independently. Standard measures were examined: sensitivity (inclusion of all truly eligible abstracts), specificity (exclusion of all truly ineligible abstracts), precision (inclusion of all truly eligible abstracts among all abstracts screened as eligible), F1-score (harmonic average of sensitivity and precision), and accuracy (correctly predicted eligible or ineligible abstracts). Workload reduction was measured as the hours the workflow saved, given only a subset of abstracts needed human screening.Results: With respect to the systematic and scoping reviews respectively, the workflow attained 88%/89% sensitivity, 99%/99% specificity, 71%/72% precision, an F1-score of 79%/79%, 98%/97% accuracy, 63%/55% workload reduction, with 12%/11% fewer abstracts for full-text retrieval and screening, and 0%/1.5% missed studies in the completed reviews. # Conclusion: The workflow was a sensitive, precise, and efficient alternative to the recommended practice of screening abstracts with 2 reviewers. All eligible studies were identified in the first case, while 6 studies (1.5%) were missed in the second that would likely not impact the review's conclusions. We have described the workflow in language accessible to reviewers with limited exposure to natural language processing and machine learning, and have made the code available to reviewers. Keywords: Systematic review, Scoping review, Text mining, Natural language processing, Machine learning, Classification model, Abstract screening, Automation Background Well-conducted knowledge syntheses such as systematic reviews (SRs) provide valid evidence to inform decisionmaking. However, SRs in healthcare can be timeconsuming (e.g., 1 year),labor-intensive (e.g., 1,139 person-hours, 5 reviewers), and expensive (e.g., > $100,000). Automation that aims to minimize timelines, person-time, and cost expenditures is of interest to producers and users of knowledge synthesis internationally. Study selection is one of the most important steps in the SR process. The recommended practice is to have two reviewers screen titles and abstracts and to resolve discrepancies between reviewers in order to maximize the chance of identifying all eligible studies. Experienced reviewers are trained on eligibility criteria to reduce discrepancies and then conduct abstract screening independently. Abstract screening consumes about 25% of the total person-time per review, estimated to range between 1000 and 2000 person-hours. Despite the rigorously recommended methods, sources of between-screener variation remain, including lack of information and varying interpretation for eligibility determination due to abstracts that report limited details on methods and results, and errors associated with distraction and fatigue, to name a few. Recent advances in text mining (the science of extracting information from text), and machine learning (the study of methods that learn patterns from data and make decisions with minimal human intervention)have enabled solutions to previously intractable problems. Since 2006, these advances have been adopted to support the automation of title and abstract screening. Substantial progress has been made to partially automate the process, with tools deployed for real-world use , and their use described in review protocols. With the increasing use of SR methods for different types of knowledge syntheses, continuing efforts have been expended to improve the performance of the automation tools. However, few tools for title and abstract screening attain the level of sensitivity and precision consistent with published benchmarks for pairs of human reviewers. We set out to develop a workflowconsecutive steps starting from importing titles and abstracts to a computing platform, to ending with a set of all eligible abstracts for the automation of title and abstract screening that is comparable to the recommended practice of screening by two reviewers, independently. Our motivation was guided by making the workflow accessible to review teams conducting different types of knowledge syntheses, requiring minimal technical expertise and training. The current paper describes the design, development, and evaluation of the workflow using two case studies. # Methods ## Workflow structure The workflow is proposed to handle title and abstract screening for knowledge syntheses addressing clinical or non-clinical research questions. We adopted widely recommended text-mining and machine-learning methods to phase 1) process title-abstract texts into computing data; and phase 2) identify all eligible abstracts through repeated interactions between human reviewers and software, using a classification model and a nearestneighbor search procedure. The two phases of the proposed workflow are outlined below, with terminologies related to text mining and machine learning described in a glossary (Additional File 1, Appendix A). ## Phase 1: preparation of abstracts for machine learning, and creation of the training dataset In this phase, title-abstract texts (henceforth referred to as "abstracts") are converted into numerical training data to train the machine learning model that classifies abstracts as eligible or ineligible. Using a simple example in , we illustrate the initial steps of transforming the content of a collection of abstracts into a "document-feature matrix," where the rows denote the abstracts, the columns represent the features (e.g., in the form of words), and the values in the matrix are weights of the features in the corresponding abstracts. In this example, the weights are frequency of the feature in the abstract. Features with relatively high frequencies are selected and retained in a reduced document-feature matrix for further analyses.outlines the 9 steps of this phase. Step 1. Import abstracts -Titles and abstracts from the literature search are imported into a computing platform for text mining and machine learning, excluding citations with title only and no abstract. Step 2. Pre-process text -Texts of each title and abstract are combined. Text is pre-processed through tokenization, lemmatization, parts of speech tagging, and as needed, semantic annotation. Tokenization is a step which splits longer strings of text into smaller pieces (tokens), such as words separated by white space. Lemmatization is a step which replaces a word by its normalized form (lemma). For example, "runs", "ran", and "running", all have a common lemma, "run." These steps ensure that words of the same meaning but expressed in slightly different forms are processed uniformly. Parts of speech tagging is done by marking up a word in a text as a noun, verb, pronoun, or adjective (among others), based upon the definition of the word and the context of its use (e.g., the context is defined by surrounding words). Illustrated workflow steps to process a simple collection of abstracts for machine learning classification DFM document-feature matrix. IDDM insulin-dependent diabetes mellitus. *The features in this example are words. ** The DFM contains frequency counts of the features. Highlighted columns in the DFM denote relatively high-frequency features that are selected and retained for further analysis Semantic annotation is a natural language processing task specifically designed for detecting and disambiguating terms mentioned in the texts into machineunderstandable terminology, such as concepts in the Unified Medical Language System (UMLS). The UMLS Metathesaurus is a large biomedical thesaurus, organized by concept or meaning, and different expressions of the same concept are linked. For example, "heart attack", "coronary thrombosis", and "cardiac arrest" are different expressions of the same concept -"Myocardial infarction"registered in the UMLS Metathesaurus. To perform semantic annotation, we use RysannMD, a general-purpose biomedical semantic annotator that balances accuracy and computing time. Step 3. Construct 3 sets of features to represent the content of abstracts -Once pre-processing of the abstracts is complete, the contents are summarized in a document-feature matrix where the columns represent each of 3 sets of features: (1) short phrases of 1, 2, or 3 contiguous words, (2) nouns and verbs, and (3) words for knowledge synthesis addressing the non-clinical review questions, or, in the case of clinical reviews, UMLS concepts. As we will show in the evaluation, multiple features are used to improve the sensitivity of the workflow, which aims to identify all eligible abstracts. Step 4. Select features and perform dimension reduction of the 3 document-feature matrices -For the document-feature matrix with short phrases, only phrases common across abstracts are retained. A mathematical method known as Singular Value Decomposition (SVD) is then applied to the matrix to further reduce its dimension (e.g., from > 100,000 features to about 300 derived features). SVD results in a matrix transformation that obtains a more compact, computationally efficient representation of the abstracts, while at the same time preserving their semantic relations. For the document-feature matrix of nouns and verbs, a topic-modeling method known as Latent Dirichlet Allocation is applied to identify common topics across abstracts. For, example, words like "cortisone", "anti-viral", and "rituximab", if seen relatively frequently, might be grouped under the topic "Covid-19 treatments", while "case", "hospitalization", "ventilator" might fall under "Covid-19 outcomes." Only nouns and verbs are used since their use tends to generate consistent and meaningful topics. The content of all abstracts is then summarized in a reduced document-feature matrix, with each row representing the probability distribution of the topics within an abstract (e.g., 300 common topics identified from all abstracts). For the document-feature matrix of words for knowledge synthesis, only words with representations in GloVe are retained in the matrix. GloVe is an opensource project that has derived global vectors for word representations (commonly known as word embeddings). GloVe characterizes a word by other words that tend to appear with it, assuming that the words near a given word encode a large amount of information regarding that word's meaning. Word embeddings model this contextual information by creating a lowerdimensional space such that words that appear in similar contexts are nearby in this new space (e.g., a 300dimensional space). For example, semantically close words such as "effect" and "consequence" are mapped to close points in the low-dimensional space where representations of semantically unrelated words such as "effect" and "reject" are more distant. Each abstract is then replaced by a weighted average vector of the word embeddings that represent the words mentioned in the abstract, weighting on the frequency count of the words. A similar approach is applied to the document-feature matrix of UMLS concepts, with concept embeddings obtained from an open-source project that uses massive sources of multimodal medical data to derive concept embeddings. Step 5. Generate 3 distance matrices representing pairwise distances between the 3 numerical vectors that represent the abstracts -Any two abstracts can be compared by taking the cosine of the angle between their vector representations (which are rows from a document feature matrix), with values approaching 1 denoting semantically close abstracts, and values approaching 0 denoting distant abstracts. The distance between two abstracts is defined to be inversely proportional to the cosine of the angle. For the document-feature matrices derived with word embeddings, the word mover distance (WMD) between two abstracts represents the minimum total cosine distance that the embedded words of one abstract need to "travel" to reach the embedded words of the other abstract. Compared to other distance measures, the use of WMD reduces error rates in document classification. Step 6. Identify abstracts similar to eligible seed abstracts -This step requires a few seed abstracts (e.g., 3-5 abstracts) that are known to be eligible and are often identified in the preparation of the review protocol. For each of these seed abstracts, a fixed number, k, of nearest-neighbor abstracts are identified based upon the distances defined in step 5. For example, if we have 3 known eligible abstracts, with k= 8 and three distance matrices, we would identify a batch of approximately 3 × 8 × 3 = 72 abstracts (with duplications removed). Information from a study examining characteristics of a representative sample of the population of SRs is used to guide the selection of k relative to the median number of 15 included studies (interquartile range: 8-25). Step 7. Screen abstracts by pairs of reviewers -The batch of abstracts identified in step 6 is screened by two reviewers independently to identify eligible abstracts. We expect a high proportion to be identified as eligible because their content is "near" those of known eligible seed abstracts. These screening results include results for both eligible and ineligible abstracts and will form part of the training dataset for the classification model (discussed below). Step 8. Assess the cumulative number of screened abstracts relative to a pre-set sample size for the training dataset -If the cumulative number of screened abstracts falls below a pre-set sample size for the training dataset (e.g., 300-600 screened abstracts), additional eligible abstracts need to be identified. Step 9. Identify new eligible abstracts -The screening results from step 7 identified eligible abstracts in addition to the original seeds in step 6. These newly identified abstracts are then input to step 6 as "seeds" so that yet additional and potentially eligible abstracts are screened for addition to the training dataset. Steps 6-9 are iteratively applied to the eligible abstracts newly identified in each iteration until the number of screened abstracts exceeds the pre-set sample size requirement for the training dataset. As we will show in the evaluation, phase 1 requires the screening of approximately 300 to 600 potentially eligible abstracts. We will also show that these abstracts are approximately 5 times more likely to be eligible than a randomly selected abstract from the literature search results. ## Phase 2: screening of abstracts through human-guided machine-learning This phase aims to identify all abstracts that are eligible for full-text screening. It involves fitting a classification model to the training dataset, predicting eligible abstracts using the fitted model, screening the predicted eligible abstracts by 2 human reviewers to identify eligible abstracts, identifying abstracts similar to the eligible abstracts (using step 6), screening the similar abstracts by 2 human reviewers, updating the training dataset to include the screening results, updating the predictive model (which is then fit to the training dataset), and repeating the described steps until no newly eligible abstracts can be identified.outlines the 7 steps of this phase. Step 1. Assemble training dataset -The screening results from step 7, phase 1 are merged with the document-feature matrix with the SVD-based features (from step 4, phase 1) to generate a training dataset, with columns of the document-feature matrix being treated as predictors and the screening results as the binary response, for the development of the classification models. We will show in the evaluation that the classification model with SVD-based features attains a higher F1-score (the average of sensitivity and precision) than corresponding models based upon word-embedding and topic-modeling features (Additional File 1, Appendix B). Step 2. Train a random-forest model to classify eligible abstracts -Among recently developed classification models for binary responses, a "random forest" is a combination of several decision trees (Glossary)that attains relatively high precision, high sensitivity, and fast computing time. A random-forest model is fit to the training dataset using the recommended method of multifold cross-validation to maximize the sensitivity of modeled prediction. To deal with imbalanced distribution of eligible versus ineligible abstracts in each crossvalidation fold (e.g., 5% versus 95%, respectively), the SMOTE algorithm is used to rebalance the distribution by generating synthetic abstracts with high eligibility probability (glossary, Additional File 1, Appendix A). Step 3. Predict eligible abstracts -The fitted model is used to predict eligible abstracts among all abstracts. Step 4. Screen predicted eligible abstracts -The set of predicted eligible abstracts is screened by two human reviewers to identify eligible abstracts. Step 5. Assess for newly identified eligible abstracts in step 4 -The eligible abstracts identified in step 4 are Step 6. Identify abstracts similar to eligible abstracts -For each of the newly identified abstracts in step 4 and using the distance measures from step 5 of phase 1, a fixed number k of nearest-neighbor abstracts are identified as abstracts similar to the eligible abstracts (see also step 6 of phase 1). Step 7. Screen the set of similar abstracts by pairs of reviewers -The batch of similar abstracts identified in step 6 is screened by 2 human reviewers, independently. Steps 1-7 are then iteratively applied, using an updated training set including the newly screened abstracts and their eligibility assessments, until no newly eligible abstracts can be identified in step 5. As we will show in the evaluation, phase 2 involves approximately 5 iterations, requires the screening of 37% to 45% of all abstracts, and identifies approximately 90% of the eligible abstracts among all abstracts screened by pairs of reviewers. We will show that compared to the reference standard of screening by pairs of reviewers, the workflow attains literature saturation, in the sense that additional eligible abstracts, if any and if identified by the reviewer pairs, would not change inferences in the knowledge synthesis. in Additional file 1 summarizes the implementation of the workflow in R (a programing language for statistical computing), using publicly available R packages for text mining and machine learning, notably the "caret" package that streamlines model training and evaluation. also displays values of key parameters in the workflow, including values for the main analysis and alternative values for sensitivity analysis. Values for the main analysis were selected through trial and error to optimize the performance of the workflow with respect to maximizing both the sensitivity of modeled classification and workload reduction (below). The selection necessarily involved uncertain judgment, and alternative parameter values likely to affect the workflow's performance were identified for sensitivity analyses by varying the parameters one at a time. Human and computing resources required for the implementation are reported in Additional File 1, Appendix A. The R code is reproduced in Additional File 1, Appendix C, and the case study original screening results can be replicated using the material available here (https://knowledgetranslation.net/text-mining-to-supportabstract-screening-for-knowledge-syntheses-a-workflowapproach/). ## Evaluating the performance of the workflow to identify eligible abstracts From the database of our knowledge synthesis team, two case studies were selected based upon the following criteria: i) different types of knowledge synthesis, ii) availability of a review protocol, iii) broad eligibility criteria, and iv) the review results published in peer-reviewed journals. We selected a SR of the efficacy and safety of insulin formulations for patients with type-1-diabetes, and a scoping review on knowledge synthesis methods (Additional File 1, Appendix A). The protocol and planned search strategies for each SR are accessible at https://osf.io/xgfud, and https://bmcmedresmethodol. biomedcentral.com/articles/10.1186/1471-2288-12-114, respectively. The search strategies may also be found within each result's publication. The screening results by the recommended practice of screening abstracts with 2 reviewers were considered the reference standard in the evaluation of the proposed workflow. As such, for each of the two case studies, the results from the original review were used as the reference standard. The proposed workflow was evaluated with respect to the following performance measures: N P : the number of predicted eligible abstracts identified by the workflow at the end of phase 2, N WF : the number of eligible abstracts (as determined by the reference standard) identified by the workflow at the end of phase 2, N S : the number of eligible abstracts identified by human reviewers (the reference standard) after screening all abstracts (N), ΔN = N S − N WF : The number of eligible abstracts missed by the workflow, The number of missed studies due to the fulltext screening of the N WF eligible abstracts instead of full-text screening the N S eligible abstracts, Precisionthe percentage of eligible abstracts predicted by the workflow that are confirmed via the reference standard (correctly predicted eligible abstracts) among all predicted eligible abstracts (N WF /N P ) *100, Sensitivityor recall, the percentage of correctly predicted eligible abstracts among the eligible abstracts identified via the reference standard (N WF / N S ) *100, F1-scorethe harmonic average of sensitivity and precision, Specificitythe percentage of correctly predicted ineligible abstracts among abstracts identified as ineligible via the reference standard ((N − N WF )/(N − N S )) *100, Accuracythe percentage of correctly predicted (based on the reference standard) eligible or ineligible abstracts, Workload reductionthe difference between the total number of abstracts and the number of abstracts screened by the workflow, assuming each abstract is screened by two reviewersPerson-hours savedthe reduction in personhours associated with the workload reduction, assuming that on average, a reviewer screens about 200 abstracts per hour. To provide benchmarking measures for the workflow's performance, we conducted a literature review of studies reporting data on the performance of human reviewers conducting abstract screening for SRs (Additional File 1, Appendix A). Studies were identified from a recent SR of methods for study selection, including forward searching of citations of studies identified by the SR to identify eligible studies published after the SR.displays the results of the evaluation of the workflow's performance; step-specific results of the workflow performance are included in Additional File 1, Appendix B.displays the workflow's performance on various performance measures relative to the recommended practice for abstracts screening. For the main analysis of the SR of type 1 diabetes, the workflow attained an 88% sensitivity, 71% precision, F1-score of 79%, 99.3% specificity, 98% accuracy, 63% workload reduction, or equivalently, a saving of 91 person-hours. Pairs of reviewers identified 743 eligible abstracts, while the workflow identified 655 eligible abstracts, or 88 fewer eligible abstracts. Compared to screening by pairs of reviewers, the workflow recommended 88 fewer eligible abstracts for full-text retrieval and screening, and this did not lead to any missed studies among the 80 studies included in the SR, which were identified via full-text screening of the 743 eligible abstracts. # Results For the main analysis of the scoping review of knowledge synthesis methods, the workflow attained an 89% sensitivity, 72% precision, F1-score of 79%, 99.3% specificity, 97% accuracy, 55% workload reduction, or equivalently, a saving of 95 person-hours. Pairs of reviewers identified 957 eligible abstracts, while the workflow identified 852. Compared to screening by pairs of reviewers, the workflow recommended 105 fewer eligible abstracts for full-text retrieval and screening. This reduction led to 6 missed studies among the 409 studies included in the scoping review, which were identified via full-text screening of the 957 eligible abstracts (an error rate of 1.5%).also displays results of the sensitivity analysis. The workflow's sensitivity increased to approximately 95% (from 88%/89%) with larger value k for the k-nearest-neighbors in phase 2 (k = 25, base value k = 15), but the workload reduction decreased by approximately 15%. Compared to the use of two feature representations in step 3 of phase 1 of the workflow, the use of the three feature representations led to higher sensitivity, which was reduced by 4%/7% by excluding the topic-modeling features. The results were robust against other parameter values, especially the threshold used to select common features in the derivation of the SVD-based features. The workflow performance was consistent regardless of the clinical and non-clinical topics, and across the SR and scoping review. Given the computing resources (Additional File 1, Appendix A) and the implementation details of the workflow , Additional File 1, Appendix B), it took approximately 6 computing hours to derive the document-feature matrix with SVD-based features, approximately 18 hours to derive the matrix with topicmodeling-based features, and approximately 60 hours to derive the matrix with word or concept embeddings. These initial steps to set up the workflow were timeconsuming. Phase 1 required approximately 3 iterations, with no computing delays between iterations that were noticeable by reviewers. The saturation of newly identified eligible abstracts was attained after approximately 5 iterations of the steps in phase 2, with approximately 45minute delays between iterations to update the classification model. Results of the literature review evaluating the performance of human reviewers are given in Additional File 1, Appendix A. We identified three studies reporting benchmarking data, with varying review topics (postal survey methods in study 1, diet research in study 2, and brain injury in study 3, and varying reviewers' experience (4 experienced reviewers, 12 reviewers with 6 experienced and 6 student reviewers, and 58 student reviewers, respectively). When reported, the sensitivity of reviewers ranged from 47% to 90% (the workflow's sensitivity was 88-89% as reported above), specificity from 73% to 100% (workflow: >99%), precision from 55 to 90% (workflow: 71-72%), F1-score from 56 to 77% (workflow: 79%); and pairs of reviewers did not miss any eligible studies identified via full-text screening, with an estimated range of 0% to 1% of missed eligible studies (workflow: 0-1.5%). # Discussion Currently, the recommended practice is to screen titles and abstracts for knowledge syntheses with two reviewers, independently, and to err on the side of over-inclusion during screening. Until now, the use of pairs of reviewers seemed to be the only known approach to reduce errors and subjectivity in study selection. Compared to the recommended practice, the proposed workflow was sensitive, as it identified all eligible studies in the first case, while missing 6 studies (1.5%) in the second that would likely not impact the review's conclusions. Thus, most importantly, we infer that in these cases, the results of the systematic reviews would not have changed had our semi-automated workflow originally been implemented. The workflow was also reasonably precise, with approximately 7 truly eligible abstracts out of 10 predicted eligible. It was efficient, as it substantially reduced the workload of abstract screening by approximately 60%. Also, it referred 10% fewer abstracts for full-text retrieval and full-text screening, while ensuring literature saturation. Using two case studies, we showed that the workflow was generalizable to two different review topics and two different types of knowledge synthesis -a SR of a clinical review topic, and a scoping review of a methodology topic. To overcome skepticism towards automation, both within the scientific community and among recognized SR, regulatory, and health technology assessment bodies, we designed the workflow with a strong emphasis on close interactions between human reviewers and the text-mining and machine-learning application. We endeavored to describe the proposed workflow in a way that is accessible to reviewers with limited exposure to text mining and machine learning, including a glossary of common terms (Additional File 1, Appendix A). We outlined the step-specific implementation in detail and implemented the workflow with publicly available software tools. Our study results can be replicated using materials available online, and the R codes are reproduced in Additional File 1, Appendix C. We hope all this would serve to facilitate the application, adoption and diffusion of the workflow into routine practice for review teams with interest in SR automation. We recommend our workflow be considered in a de novo implementation when the number of abstracts to be screened is at least 5000. As well, users may want to first consider piloting the workflow on a systematic review they completed, and comparing the results before using the workflow routinely. We compared the workflow with other automation tools for abstract screening that are currently in use. Abstrackr, a commonly used tool, has been recently evaluated in four abstract-screening projects. Across the projects, sensitivity ranged from 79 to 97%, and precision ranged from 15 to 65%. The text-mining algorithm implemented in the online SR system EPPI-Reviewer has been evaluated recently using a case study. According to the results, sensitivity could be very *Sensitivity or recall; results of the sensitivity analyses are displayed in decreasing sensitivity of the workflow's performance. ♣ Person-hours that were tallied across reviewers. a The main analysis was conducted with distance definitions from three feature representations (SVD-based, LDA-based and word-embedding features), a threshold ϕ = 70%, k-nearest-neighbor (k-NN 1 ) for phase 1 of 8, k-NN for phase 2 (k-NN 2 ) of 15, and initial sample size r = 600 high (e.g., 99%) at reasonable precision (e.g., 50%). The text-mining algorithm implemented in the online SR system Rayyan was evaluated using a sample of 15 SRs. According to the results, sensitivity ranged from 62 to 100%, and workload reduction ranged from 3 to 55%. The algorithm implemented in the Distiller SR platform was evaluated using a sample of 15 SRs, with workload reduction ranging from 9 to 62%. In comparison, our proposed workflow was associated with high sensitivity (approaching 90%), high precision (approximately 70%), and high workload reduction (approximately 60%). The workflow as described requires two systematic reviewers as part of the iterative human-machine-learning process. Certainly, one experienced reviewer could be considered instead, but we would expect inferior results. One reviewer could be used for the first few iterations, or for establishment of the training dataset, while the potential abstracts are "nearest" the seeds, and two reviewers could be used for the remaining iterations, without inferior results. This would be an item for further research. There are limitations to our study. First, the evaluation of the workflow was retrospective, with potential bias associated with the fact that the reference standard was known prior to the evaluation. We only used two case studies in our evaluation because we wanted to provide a detailed analysis of the proposed automation approach. The step-specific methods we used in the workflow are simple; they work well together, but they might not be optimal for individual steps. The values we used for the parameters governing the steps of the workflow in might not be the best values to optimize the overall performance of the workflow. In this regard, we did not try to establish optimal step-specific methods or parameter values, since we believe optimality would depend on the specific SRs or types of knowledge syntheses. In the second case study (scoping review), the workflow missed 6 eligible studies that were identified by pairs of reviewers. We however believe that the workflow identified a saturated set of eligible studies, in the sense that the 6 missed studies would not change the inferences in the scoping review. These studies were eligible for inclusion for the scoping review but were not a major focus of the scoping review. Another major limitation is that the workflow did not perform well with title-only citations, with low F-1 score (e.g., 7%, data not shown). Reviewers who are interested in using this workflow may have to apply the proposed automation approach to the set of titles only and no abstracts (separately from the handling of titles and abstracts) or manually screen citations with titles only. # Conclusions The workflow was a sensitive, precise and efficient alternative to the recommended practice of abstract screening with 2 reviewers, independently. All eligible studies were identified in the first case, while 6 studies (1.5%) were missed in the second that would likely not impact the review's conclusions. We have described the workflow in language that is accessible to reviewers with limited exposure to natural language processing and machine learning as well as making the codes accessible to reviewers.

Three-dimensional reconstruction for coherent diffraction patterns obtained by XFEL The three-dimensional (3D) structural analysis of single particles using an X-ray free-electron laser (XFEL) is a new structural biology technique that enables observations of molecules that are difficult to crystallize, such as flexible biomolecular complexes and living tissue in the state close to physiological conditions. In order to restore the 3D structure from the diffraction patterns obtained by the XFEL, computational algorithms are necessary as the orientation of the incident beam with respect to the sample needs to be estimated. A program package for XFEL single-particle analysis based on the Xmipp software package, that is commonly used for image processing in 3D cryo-electron microscopy, has been developed. The reconstruction program has been tested using diffraction patterns of an aerosol nanoparticle obtained by tomographic coherent X-ray diffraction microscopy. # Introduction Biological molecules, such as proteins, nucleic acids and their complexes, are responsible for many vital cellular functions, including gene transcription and protein synthesis, and their dysfunctions result in severe diseases. Information on the structure and dynamics of biomolecules and biomolecular complexes is helpful for understanding their functional mechanisms, and X-ray crystallography is currently the most widely used technique to determine the three-dimensional (3D) structure of biomolecules. However, this technique requires molecules to be crystallized, and it is difficult to apply to insoluble molecules or intrinsically disordered proteins. Although cryo-electron microscopy (cryo-EM) does not require crystallization and can observe heterogeneous samples, applications to observe the inner structure of thick objects (more than 500 nm) can be challenging due to multiple-scattering events. Single-particle imaging using femtosecond X-ray pulses from X-ray free-electron lasers (XFELs) could allow the inner structure of biological molecules to be observed in a state close to nature without crystallization. Radiation damage is a serious problem in high-reso- lution microscopy as it reduces the spatial resolution of the molecular structure. XFELs can significantly relax the resolution barrier imposed by radiation damage by recording the diffraction pattern before specimen destruction, due to femtosecond-short pulse duration. XFEL experimental data are becoming increasingly available and several low-resolution structures from single-particle approaches have been reported. It has also been shown, theoretically, that high-resolution 3D structures could be obtained using millions of diffraction patternsand that dynamics could be directly extracted from two-dimensional (2D) data. However, there are still challenging problems in obtaining high-resolution 3D structures of biomolecules from XFEL experimental data. Because the diffraction intensity from biological molecules is too weak, an insufficient photon count is a serious problem, especially at high-wavenumber pixels which determine the resolution in real space. On the other hand, the diffraction intensity at low-wavenumber pixels is too strong and saturates the detection range, which hinders the determination of the overall shape of the system by phaseretrieval procedures. In addition, 3D imaging requires the assembly of diffraction patterns from many identical copies of a reproducible object. Therefore, 2D diffraction images should be obtained from structurally homogeneous samples, but it is generally difficult for sub-micrometer systems, which is the typical target size for XFEL single-particle analysis. Along with these challenges, one of the critical algorithmic problems to be solved in order to reconstruct 3D structures from diffraction patterns obtained in XFEL experiments is the estimation of the orientation of single particles with respect to the laser beam (three Euler angles). This is also a common problem in cryo-EM single-particle analysis. Several opensource software packages have been developed to reconstruct 3D molecular structures from cryo-EM projection images. There are currently two main strategies for estimating the 3D orientation of particles. One is called projection matching, where the best angular parameters are estimated by finding an image which has a maximum correlation coefficient in the reference library and is used to construct a new volume. Another widely used strategy is the maximum-likelihoodbased algorithm. In these approaches, a number of angular assignments are considered for each target image, and are concurrently used for reconstruction with relative weights based on the similarities between the target image and reference images. While a cryo-EM electron density image contains structural information in real space, an XFEL diffraction pattern contains structural information in Fourier space and is related to a slice (Ewald sphere) of the 3D diffraction intensity. Thus, 3D reconstruction from XFEL data can be performed using similar procedures to those used in cryo-EM. For instance, 'slice matching' can be performed to determine the orientation of the diffraction patterns. One of the algorithms successfully applied for 3D reconstruction from XFEL data is EMC , which uses a 'maximum likelihood' approach. A 'maximum cross correlation' algorithm for 3D reconstruction was previously tested with synthetic data using a large number (8000-100000) of diffraction patterns. Here, we aim to reconstruct a 3D model employing a 'maximum cross correlation' algorithm with experimental data and to examine necessary parameter calibrations in detail. Our maximum cross-correlation algorithm is similar to that of Tegze & Bortel, but there are also some differences between the two algorithms as follows: we reconstructed the 3D structure factor amplitude instead of the 3D diffraction intensity by using a weight function based on the Kaiser-Bessel window. A phase recovery procedure is needed to obtain a 3D model in real space from the assembled 3D model in Fourier space (assembled from the diffraction patterns at the determined orientations). However, this task is also not trivial and its result may depend on the choice of the phase recovery algorithm; therefore we here focus on the angular assignment process only. Our program was implemented on top of Xmipp, which is an image-processing software package primarily aimed at single-particle 3D cryo-EM (de la. The program suite contains many useful tools for image analysis that could be used for analyzing XFEL data. We tested our reconstruction program using experimental diffraction data of an aerosol nanoparticle obtained by tomographic coherent X-ray diffraction microscopy (CXDM)instead of data from XFEL. These data are very similar to those from XFEL experiments, the main difference being that XFEL is 'single-shot' while X-ray diffraction microscopy allows (weaker but) repeating exposure. The speckle patterns from CXDM were employed to simulate XFEL single-particle data, i.e. a set of diffraction patterns obtained from a non-crystalline nanoparticle with different sample orientations. Sample orientations were estimated using the approach demonstrated here, and the assigned angles were later compared with the angles actually used for the data collection. We discuss the choice of parameters and protocols for a successful estimation of the incident beam angles. ## Theory: reconstruction of 3d structure in fourier space from diffraction patterns We performed the reconstruction of the structure factor amplitude distribution (hereafter called 'volume'), from diffraction patterns, based on the 'slice matching' protocol research papers 728 Miki Nakano et al. 3D reconstruction for coherent diffraction patterns [fig_ref] a: Schematic view of our program for volume reconstruction from experimental diffraction patterns [/fig_ref]. While others reconstruct 3D diffraction patterns by assembling 2D diffraction patterns in 3D space, we here reconstruct 3D amplitudes, because the 3D amplitude can be directly used for phase retrieval. More precisely, we convert 2D diffraction patterns to structure factor amplitude and then assemble into 3D space (comparison results between 3D diffraction amplitude distribution and 3D diffraction intensity are shown in x4). The protocol is as follows: (i) Create an initial volume, V initial Fourier . In general, we create V initial Fourier using a set of 2D diffraction patterns with incident beam angles that are randomly assigned. Each 2D diffraction pattern has been centro-symmetrized, and the square roots of the intensities are calculated to obtain structure factor amplitude before calculation. In the first iteration, V initial Fourier is used as the reference volume, V reference Fourier . (ii) Create a reference library of N ref diffraction patterns by extracting, using cubic spline interpolation, slices from V reference Fourier at orientations (angles ' and ) distributed over a sphere evenly, using a given angular sampling (discretization) step. Then, square slice pixel values to obtain the corresponding diffraction patterns. (iii) Calculate the zero-mean normalized cross correlation, CC, between each experimental diffraction pattern and all reference patterns rotated by angles in plane (with a given angular discretization step) using the following equation, [formula] CC pq ¼ 1=N pix À ÁP N pix i I exp;p ðiÞ À I I exp;p  à I ref;q ðiÞ À I I ref;q Â Ã È É I exp;p I ref;q ;ð1Þ [/formula] where I exp,p (i) and I ref,q (i) are the diffraction intensity at pixel i of the pth experimental and qth reference diffraction patterns, respectively ( p = 1 to N exp , q = 1 to N ref ). N pix is the number of pixels in each diffraction pattern. I I exp;p and I I ref;q are the average intensities of the pth experimental and qth reference diffraction patterns, and I exp;p and I ref;q are their standard deviations, respectively. I ref;q ðiÞ is the diffraction intensity of the qth reference pattern rotated by angle in the plane. The reference pattern and resulting in the maximum CC coefficient (CC max ) are denoted by M opt ref and opt , respectively. The incident beam angles ' opt and opt used to create M opt ref and the in-plane angle opt assigned to the experimental image are the Euler angles set which determines the particle orientation. For the calculation, we apply masks as described below to enhance the sensitivity of the angular assignment. (iv) Reconstruct a volume, V Fourier , using the experimental images with the Euler angles assigned as described above. The diffraction amplitude at voxel k in the reconstructed volume, A(k), is obtained as the weighted average of the square roots of the diffraction intensities, ½I exp;p ðiÞ 1=2 , in the experimental diffraction images [equation (2)]. To calculate A(k), a weight function based on the Kaiser-Bessel window is used, which depends on the distance d kj between the position k and j within V Fourier : k is the center position of the voxel k and j is the position of the pixel i within the 2D diffraction image, p, in a 3D volume [equation (3)], [formula] AðkÞ ¼ X N exp p ¼ 1 X N pix i ¼ 1 w d kj À Á I exp;p ðiÞ Â Ã 1=2 0 X N exp p ¼ 1 X N pix i ¼ 1 w d kj À Á ; ð2Þ w d kj À Á ¼ ð; Þ I 0 ðÞ I 0 1 À 2d kj À À 1 À 1 2 " # 1=2 8 < : 9 = ; [/formula] ; [formula] 0 d kj :ð3Þ [/formula] I 0 is the zeroth-order modified Bessel function, is the maximum radius for interpolation, and is a variable which determines the decreasing rate of w(d kj ). (,) is the normalization factor determined by and . With a large , the diffraction intensity would be interpolated using the pixels farther in the mapped position in the reconstructed volume. With a large , the weight for the interpolation would be decreased quickly as d kj increases. These parameters need careful calibration for successful reconstructions and are discussed later in detail. Finally, we centro-symmetrized the reconstructed volume in Fourier space. (v) Update V reference Fourier using the volume reconstructed in the previous step. The reference library sampling steps will be made smaller for the refinement of angle assignment in the next iteration. The method allows the correlation coefficient between the experimental diffraction image and the reference diffraction images to be calculated in the vicinity of the currently assigned angles. This is an option of the method that should not be used in the beginning of the iterative assignment process. (vi) Iterate from (ii) to (v) until convergence of V reference Fourier . The final 3D structure is denoted as V final Fourier . In the calculation of CC, we excluded the center and outer regions of the diffraction patterns to improve the sensitivity of the matching. The center region is excluded in the calculation since these very large intensity values reduce the sensitivity of the CC calculation. Also, in practice, these diffraction intensities are often too strong to be measured and are thus missing in the data, especially in XFEL applications. In the outer region of diffraction patterns, intensities are usually too weak to be detected and are uncertain because of noise. Therefore, we only calculated CC for the annular regions defined by the inner and outer radii, q L and q H , as shown in [formula] (b). [/formula] The 'slice matching' approach described above [steps (i)-(vi)] was implemented based on a projection matching protocol included in Xmipp, which uses a Fourier-space representation of the reference volume for library creation as well as 3D interpolation in Fourier space for 3D reconstruction. # Results ## Pre-processing of experimental diffraction patterns To test our 'slice matching' protocol, we used tomographic CXDM data of aerosol nanoparticles. The similarity between CXDM and single-particle XFEL resides in the absence of sample crystallization and in the sample irradiation from different incident beam orientations. However, different diffraction patterns collected in a CXDM experiment correspond to different orientations of the same sample while each single-particle XFEL diffraction pattern corresponds to an orientation of a different sample. Additionally, in tomographic CXDM, the sample is only rotated around one axis (one angle, , is known) with respect to the incident beam, whereas all three Euler angles (', and ) are unknown in single-particle analysis using XFELs. Our final goal is to reconstruct the 3D structure using XFEL experimental diffraction images. The aim of this study is to validate the incident beam angle estimation in our reconstruction program using experimental diffraction images including noise and obtain insight into the calibration of the necessary parameters. Diffraction images obtained by tomographic CXDM are ideal data for our purpose, since incident beam angles with respect to the sample are known (the tilt angles were set experimentally). Pretending that we do not know the incident beam angles, we used our program to estimate the orientation of each diffraction pattern, and compared against the actual orientation. A total of 53 diffraction patterns at various sample orientations (tilting angles from À69 to +69 in 1 to 4 increments in an equal slope scheme) were measured using a CCD camera. The specimen size used in this study was about 1.5 mm. The data were all 3  3 binned and resized to 430 pixel  430 pixel. All diffraction patterns were centro-symmetrized.shows a cross-section view of the arrangement of the full-size experimental diffraction patterns in 3D Fourier space using the ground-truth incident beam angles, V true Fourier , with the interpolation parameters = 1 pixel and = 15. The tilt axis is perpendicular to the cross section shown in. The empty regions in this 3D space are related to the limitations in the maximum tilt angle achievable with the given CXDM apparatus, which is identical to the missing wedge problem in cryoelectron tomography. Also, there is one common line shared by all experimental diffraction patterns. Before applying the 3D reconstruction algorithm, we cropped the outer region of the diffraction patterns to remove the pixels which do not practically carry diffraction information (the values of pixels much farther from the central pixel are usually zero or too small to be distinguished from noise). The resolution in reciprocal space at the edge of the cropped image is approximately 0.011 nm À1 (= q max ), which corresponds to 91 nm resolution in real space. Furthermore, we reduced the image size from 100 pixel  100 pixel to 50 pixel  50 pixel by binning . These reductions were necessary in order to cover a sufficient number of voxels to ensure that nearby 2D images in the assembled 3D volume are close enough to detect the consistency (and inconsistency) of the 3D reconstruction. Cropping the outer region of a diffraction pattern corresponds to reducing the resolution in real space, while diffraction pattern binning (reducing the oversampling in Fourier space) corresponds to denoising in Fourier space (without binning, the diffraction intensity fluctuates more due to noise, including Poisson noise). The curvature of the Ewald sphere can be ignored, taking into account the current resolution limit. Regarding searching orientations of diffraction patterns by 2D slice matching of a 3D structure in Fourier space, diffraction pattern cropping is important because the search is focused on the central region that contains information about the object's shape in real space. Cross section of the volume V true Fourier that was assembled from experimental diffraction patterns arranged in 3D Fourier space using the ground-truth angles. The tilt axis is perpendicular to the cross section. ## Adjustments of beam intensity variations To reconstruct 3D structures, variations of beam intensity embedded in the diffraction patterns need to be normalized. [fig_ref] a: Schematic view of our program for volume reconstruction from experimental diffraction patterns [/fig_ref] shows the diffraction intensities averaged over the cropped region reduced to 50 pixel  50 pixel. In [fig_ref] a: Schematic view of our program for volume reconstruction from experimental diffraction patterns [/fig_ref] , the average diffraction intensity over all ground-truth tilt angles was 32.59 AE 4.15. It is expected that the average diffraction intensity is not the same for different ground-truth tilt angles and that a continuous change is possible. In this study, we normalized (scaled) the cropped diffraction patterns (50 pixel  50 pixel) so that the average intensity is the same for all ground-truth tilt angles [Figs. 4(a) and 4(d)]. Note that, in all scaled diffraction patterns, the sum of diffraction intensities at low wavenumbers (q L < 5 pixel) is 95% of the total sum. This scaling smoothens the variations of the intensity averaged over the annular regions defined by q L and q H , as shown in [fig_ref] Figure 4: Rad [/fig_ref] , 4(c), 4(e) and 4(f ). ## Adjustment of the interpolation parameters To find optimal interpolation parameters for 3D reconstruction, we reconstructed volumes using experimental images (pre-processed as described in the previous sections) and their ground-truth orientations (ground-truth tilt angles), with various interpolation parameters for 3D reconstruction [Figs. 5(a)-5(e)]. The interpolation radius, , needs to be sufficiently large to fill the reciprocal space where experimental data are missing. The parameter controls the relative weight for the interpolation; with a smaller , data points are more equally counted, regardless of the distance, and the reconstructed volume becomes blurred. The maximum Intensity of each diffraction pattern for non-scaled data [(a)-(c)] and for scaled data [(d)-( f )] averaged over the cropped image regions of size reduced to 50 pixels  50 pixels (a, d), within annular region q L = 5 pixels and q H = 10 pixels (b, e), and within annular region q L = 5 pixels and q H = 20 pixels (c, f ). The horizontal-axis tilt(true) represents the ground-truth incident beam angles. ## Figure 3 Pre-processing of the experimental diffraction patterns. diffraction pattern frequency used for the 3D reconstruction was set to 0.8 q max (= 20 pixels) to avoid the protrusion of the intensity outside the volume box when using large values of . The volumes reconstructed with the interpolation parameter, = 1, show discontinuity inside and 5(c)]. In the 3D reconstruction algorithm, angles are assigned to maximize the consistency between neighboring slices. With such a small , each voxel is determined merely by the nearest slice (diffraction pattern); in other words, neighboring slices have no influence on each other, and thus there is no inconsistency to be reduced. Indeed, when angular assignment is started from a set of random angles, convergence is often reached within a few iteration steps without any improvement (results not shown). Thus, we need to use a sufficiently large . With = 10, the interpolated 3D volume shows continuity and 5(e)], indicating that diffraction patterns sufficiently influence each other in an assembled 3D space. We then optimized other interpolation parameters for angular assignment. We performed iterations starting from the initial reference volumes obtained from images (preprocessed as explained above) with ground-truth orientations with the different interpolation parameters, and examined the resulting tilt angles. A total of 139 reference diffraction patterns were created from À69 to +69 (which was the incident beam angle range used in the tomography experiment), with 1 tilt angle intervals with cubic spline interpolation. It should be noted that the angular range for reference diffraction patterns would not be limited in the case of singleparticle XFEL data because these reference patterns would need to be computed in all orientations in such case. Two different matching regions were defined as earlier, i.e. one having q L = 5 and q H = 10 pixels and the other having q L = 5 and q H = 20 pixels.shows the assigned tilt angles as a result of this test. With the combination of = 10 and = 15, the tilt angles are not well estimated [Figs. 6(a) and 6(b)], presumably because the reconstructed volume becomes too blurred as shown in. The use of large values of compensates for this effect by emphasizing the nearby images for interpolation. By increasing to 100 with = 10, the agreement between the estimated and ground-truth angles greatly improved [Figs. 6(c) and 6(d)], and angles similar to groundtruth angles are obtained. As we described in x2, we excluded the center region of the experimental diffraction patterns (q < 5 pixel) for CC calculations. The diffraction intensities in this region are very high and have large fluctuations such that the CC measure is much less discriminative if the pixels in this region are included in the calculations. The results with q H = 10 were slightly better than those with q H = 20 Estimated tilt angles of experimental diffraction patterns using the alignment initiated with ground-truth angles and the 3D reconstruction with various interpolation parameters. large matching regions would not be beneficial since the pixels farther from the image center are more prone to noise. Therefore, all further experiments were performed using scaled cropped experimental diffraction patterns of size reduced to 50 pixel  50 pixel, = 10 pixel and = 100 for volume reconstruction, and q L = 5 pixel and q H = 10 pixel for orientation estimation (these q L and q H correspond to 0.004 nm À1 and 0.002 nm À1 in reciprocal space, respectively). ## Estimation of the incident beam angle from random initial reference volumes Using the parameters calibrated in the preliminary analysis presented above, we performed the tilt angle estimation using an initial reference volume computed from experimental images by assigning random orientations. We created five reference initial volumes using five sets of random angles (in the range [À69 , 69 ]) for each experimental image. It should be again noted that there would not be such an angular range limit when using single-particle XFEL data (the initial volumes would be generated from experimental images at any random orientation in the case of single-particle XFEL data). [fig_ref] Figure 7: Rad [/fig_ref] shows the results of the tilt angle estimation. The parameters converged within 20 to 50 iterations. We obtained good results in two of five trials (random#2 and random#5), meaning that the estimated angles were similar to the groundtruth angles in these two trials. For the other three trials, although the estimated tilt angles largely deviated from the correct angles, the relative orientations of nearby diffraction patterns (the diffraction patterns with similar orientations) were partially recovered. For example, in the trial random#3, it appears as if should be wrapped to [À100, À70], but this is because this reconstructed volume is made from two disconnected groups of images; in one group, the images from [true À35, 70] are correctly aligned and in another the images from [true À65, À35] are aligned, but the connection around the image with [true À35] was not recovered during this iteration. Thus, two groups exchanged their positions in this iteration, and appear 'swapped'. It should be noted, however, that such optimization issues may be less severe for data from a typical XFEL single-particle experiment. In such experiments, different pairs of diffraction patterns may share different common lines, unlike the single-tilt-axis tomographic CXDM data in which all patterns share just one common line. In the above particular example, if we had a few diffraction patterns that are (near) perpendicular to the tilt-axis, i.e. transverse to the other images, such remaining inconsistencies could be identified and possibly resolved. These results show that our 3D reconstruction method can successfully identify relative orientations of similarly oriented images although the results depend on their initial volume. shows the cross sections of the five initial random volumes and the corresponding final reconstructed 3D volumes. For all the trials, the cross sections of the initial and final volumes are significantly different. For the random#2 and random#5 trials, which showed good matching results in [fig_ref] Figure 7: Rad [/fig_ref] , final cross sections were similar to the cross sections of the volume reconstructed using the ground-truth angles and the same interpolation parameters ( = 10 pixel and = 100). In the random#3 trial, slices were not assigned from 20 to 45 in both [fig_ref] Figure 7: Rad [/fig_ref] and . ## Cross-correlation map among diffraction patterns To further assess the results of the five trial runs, we examined the correlation coefficients between experimental images and all reference images which were sliced from the final volume. We created the cross-correlation map (CC map) as shown in Figs. 9(a)-9(e). In Figs. 9(a)-9(e), horizontal axes represent the image number, which were sorted by the assigned tilt angle at the end of the refinement to check the CC coefficient between assigned and all reference tilt angles. For the random#2 and random#5 trials, which showed good matching results, the maximum CC coefficient values mostly follow diagonal lines. This is more the case for random#2 than for random#5, which is consistent with the matching results as shown in [fig_ref] Figure 7: Rad [/fig_ref]. This means that the diffraction patterns are more smoothly arranged in the final volume for random#2 and random#5 than for other trials. On the other hand, for unsuccessful reconstructions such as random#1, correlations between the experimental image and the reference slices are not continuously distributed. This is more evident in Figs. 9( f )-9( j), which show the maximum CC coefficient (CC max ) between each experimental diffraction pattern and their optimal reference images from the final reconstruction. There are multiple gaps in random#1, Tilt angles assigned to the experimental images by the proposed method initiated with five random initial volumes. The horizontal and vertical axes represent the ground-truth incident tilt angles and the tilt angles assigned by the proposed method, respectively. Iter is the number of iterations until convergence. random#3 and random#4, for example from À45 to À30 , from À20 to À10 , from 0 to 20 and from 25 to 35 in random#1, i.e. these tilt angles were not assigned in their final reconstruction. In some cases, multiple images are assigned to the same angles. For random#2, a set of images with groundtruth angles from À69 to À45 are assigned to the same angles . A similar result was obtained for random#5 (from À45 to À20 ) as shown in . This is likely to be due to similarities in the diffraction patterns at these angles. In addition, CC max values are discontinuous before and after the gaps, indicating the fragmented matching. These results show that angular assignments are initial value dependent. Geometrical relations between similar images are identified by the algorithm and nearby images are sorted correctly; however, discontinuities in the image alignment occasionally occur, resulting in a set of independent blocks of correctly ordered images. However, these artifacts should be less significant for typical single-particle experiments as the rotation angles are expected to be distributed evenly. # Discussion Our slice matching and 3D reconstruction protocol is similar to the cross-correlation maximization method previously developed by, in which a 3D reconstruction of NapAB protein molecule using a large number of Cross section of (a-e) random initial volumes and ( f-j) finally reconstructed volumes. The tilt axis is perpendicular to the cross section. The same crosssection view is shown for all volumes. The r-values shown below panels ( f )-( j) are the correlation coefficients between finally reconstructed volumes and ground-truth volume. ## Figure 9 (a-e) CC map between the experimental images and the references created from the final volume. CC coefficients were calculated only for the matching regions (q L = 5 pixels, q H = 10 pixels). The horizontal axis represents the image numbers which were sorted by the assigned tilt angle. ( f-j) Distribution of the assigned angle for each experimental image and the maximum CC coefficient (CC max ) against the reference images created from the final model. simulated diffraction patterns (100000 images) was performed. In this study, we calculate the 3D structure factor amplitude distribution instead of the intensity distribution. Our tests have shown that matching results obtained using the 3D amplitude [fig_ref] Figure 7: Rad [/fig_ref] were slightly better than those obtained using the 3D intensity on the system studied here . This may be attributed to the fact that the structure factor amplitude is smoother than the intensity because the square roots are taken. Also, we consider that it is more useful to reconstruct the 3D amplitude than the 3D intensity because the 3D amplitude can later be directly used to obtain the structure in real space. The results of our study show that such protocols can estimate the orientation of the incident beam angles and the 3D structure factor amplitude with experimental diffraction patterns. Key parameters of our protocol are the parameters that determine the matching region (q L and q H ) and the interpolation parameters ( and ). Sometimes, these parameters are best adjusted by trial and error. We found that large q L and small q H values give optimal results since the experimental diffraction pattern pixels much further from the image center are unreliable due to noise and the central pixels have too strong intensities which reduce the discriminative power of the CC measure. It is also possible to use a low-wavenumber region first to determine the orientation of diffraction patterns to reconstruct the 3D structure factor amplitude coarsely, and later use a high-wavenumber region to determine the structure with finer details. In addition, we carefully examined the choice of parameters for the interpolation with the Kaiser-Bessel window. In our test, we assembled the diffraction images of 50 pixel  50 pixel (the wavenumber at the edge of the diffraction patterns is approximately 0.011 nm À1 ). Assuming that we evenly distribute 53 images over an angular range of 140 , the distance between the edge pixels on two adjacent images in 3D space is about 1 voxel. Using the combination of = 10 and = 100, the weight would decrease to a half value at the voxel about 1.5 units away from the position where the pixel on the 2D diffraction image is mapped within the 3D volume [fig_ref] a: Schematic view of our program for volume reconstruction from experimental diffraction patterns [/fig_ref]. Thus, for each voxel, pixels from three to four images are weight-averaged. These parameter values are appropriate for 3D interpolation from a limited number of 2D diffraction patterns. We have found that the parameters and can be selected in a general case based on the rule that pixels from three to four images should be weight-averaged in 3D space. We also showed that the reliability of the reconstructed volume could be examined by evaluating the continuity of CC max with respect to the assigned beam angle [Figs. 9( f)-9( j)].proposed an approach that evaluates the distribution of CC values between the diffraction patterns and the reference slices [like in Figs. 9(a)-9(e) of our study] assuming that just a few particular diffraction patterns should have high CC values. Our approach is complementary to this approach, focusing on the distribution of assigned angles and CC max continuity. In single-particle XFEL experiments, the signal-to-noise ratio could be improved by using a large number of diffraction patterns (by averaging many diffraction patterns with similar 3D orientations, as done in single-particle electron microscopy). However, this has not been demonstrated with experimental tomographic CXDM data because a small number of diffraction patterns is usually obtained. Also in our test study (with tomographic CDXM data), only 53 patterns were mapped onto a 140 angular space. Thus, the strategy to improve the signal-to-noise ratio with actual experimental data requires further study. In addition, in this study, we simply assigned initial orientations randomly, because both the number of images and possibility of assignment angles were small. To create a reliable initial volume could be a challenging problem in singleparticle analysis. However,demonstrated encouraging results that the projection matching of NapAB protein starting from random orientations using 100000 diffraction patterns could converged after 15 iterations. Some optimization methods have been proposed in cryo-EM analysis. These studies could help us to create an appropriate initial volume using XFEL diffraction images. By combining with a 3D classification for many final volumes obtained by a large number of trials, we can select the most populated classes as the most plausible volumes. # Conclusion We have developed a protocol for the 3D reconstruction of the absolute value of the structure factor from XFEL diffraction Tilt angles assigned to experimental images by reconstructing the diffraction intensity distribution (using slice matching and starting from the same random angles as in [fig_ref] Figure 7: Rad [/fig_ref]. patterns and software based on Xmipp libraries to run it. In this article, we presented this protocol and the results of its tests with experimental diffraction patterns of an aerosol nanoparticle obtained by CXDM. Although the angle estimation for these data is difficult, as only one common line exists between diffraction patterns, encouraging results were obtained. Reconstruction was performed starting from randomly created reference 3D structures, and nearly correct volumes were obtained. We also showed that the plausibility of reconstructed volumes could be evaluated by examining the continuity of CC max with respect to the assigned beam angles. Finally, we showed the sensitivity of the protocol to different parameter values. Additional testing with more experimental data would be necessary to establish more general guidelines to the parameter adjustment. # Funding information Funding for this research was provided by: FOCUS for Establishing Supercomputing Center of Excellence; JSPS KAKENHI (award No. 16K05527). [fig] 729, Figure 1: Rad. (2017). 24, 727-737 Miki Nakano et al. 3D reconstruction for coherent diffraction patterns [/fig] [fig] a: Schematic view of our program for volume reconstruction from experimental diffraction patterns. (b) Matching region in the diffraction pattern used for calculation of cross correlation. Center and outer regions of the diffraction pattern are masked (filled with pink stripes). [/fig] [fig] Figure 4: Rad. (2017). 24, 727-737 Miki Nakano et al. 3D reconstruction for coherent diffraction patterns 731 [/fig] [fig] Figure 7: Rad. (2017). 24, 727-737 Miki Nakano et al. 3D reconstruction for coherent diffraction patterns 733 [/fig] [fig] 737, Figure 8: Nakano et al. 3D reconstruction for coherent diffraction patterns J. Synchrotron Rad. (2017). 24, 727- [/fig] [fig] 735, Figure 10: Rad. (2017). 24, 727-737 Miki Nakano et al. 3D reconstruction for coherent diffraction patterns [/fig]

Association between organizational climate and perceptions and use of an innovation in Swedish primary health care: a prospective study of an implementation Background: There is a need for new knowledge regarding determinants of a successful implementation of new methods in health care. The role of a receptive context for change to support effective diffusion has been underlined, and could be studied by assessing the organizational climate. The aim of this study was to assess the association between organizational climate when a computer-based lifestyle intervention tool (CLT) was introduced in primary health care (PHC) and the implementation outcome in terms of how the tool was perceived and used after 2 years. Methods: The CLT was offered to 32 PHC units in Sweden, of which 22 units agreed to participate in the study. Before the introduction of the CLT, the creative climate at each participating unit was assessed. After 24 months, a follow-up questionnaire was distributed to the staff to assess how the CLT was perceived and how it was used. A question on the perceived need for the CLT was also included.Results:The units were divided into three groups according to the creative climate: high, medium and low. The main finding was that the units identified as having a positive creative climate demonstrated more frequent use and more positive perceptions regarding the new tool than those with the least positive creative climate. More positive perceptions were seen at both individual and unit levels.Conclusions: According to the results from this study there is an association between organizational climate at baseline and implementation outcome after 2 years when a tool for lifestyle intervention is introduced in PHC in Sweden. Further studies are needed before measurement of organizational climate at baseline can be recommended in order to predict implementation outcome. # Background Factors influencing the implementation of new methods in health care have been studied, but there is still a lack of knowledge regarding determinants of a successful implementation. The role of context has been highlighted, and is one essential part in the Promoting Action on Research Implementation in Health Services (PARIHS) framework; implementation success is considered to be a function of evidence, context and facilitation [bib_ref] Ingredients for change: revisiting a conceptual framework, Rycroft-Malone [/bib_ref]. Cummings et al. [bib_ref] Influence of organizational characteristics and context on research utilization, Cummings [/bib_ref] have described how contextual factors in terms of culture, leadership and evaluation have an impact on research uptake among nurses. Another contextual factor that has been suggested to influence the uptake of research findings is organizational climate, defined by Greenhalgh et al. [bib_ref] Diffusion of Innovations in Health Service Organisations: A Systematic Literature Review, Greenhalgh [/bib_ref] as the extent to which staff in an organization "feel that it's OK to experiment with new ideas". Organizational climate can be studied explicitly, or as one factor in what is often labelled as a receptive context [bib_ref] Diffusion of Innovations in Health Service Organisations: A Systematic Literature Review, Greenhalgh [/bib_ref]. The role of a receptive context for change to support effective diffusion of research evidence has also been underlined. Another important factor associated with the potential adopters is whether they see a need for the innovation, meaning that it is perceived to have advantages compared with current practice. However, effective implementation needs not only a receptive climate but also a good fit between what is implemented and the potential adopters' needs and values [bib_ref] The challenge of innovation implementation, Klein [/bib_ref]. This was considered when a computerbased tool for lifestyle intervention was developed and introduced in Swedish primary health care (PHC) [bib_ref] Computerized lifestyle intervention in routine primary health care: evaluation of usage on..., Carlfjord [/bib_ref]. In Sweden, health care has an obligation not only to provide medical care but also to address lifestyle issues in order to prevent illness and promote health. Preventive services, however, are not provided to the fullest extent due to lack of time, knowledge and skills, a problem that has also been recognized in other parts of the world [bib_ref] One minute for prevention: the power of leveraging to fulfill the promise..., Stange [/bib_ref] [bib_ref] Attitudes of Swedish general practitioners and nurses to working with lifestyle change,..., Geirsson [/bib_ref] [bib_ref] Under what circumstances are nurses willing to engage in brief alcohol interventions?..., Johansson [/bib_ref]. To meet the needs expressed by practitioners, a computer-based lifestyle intervention tool (CLT) was developed and offered to nine PHC units operating in Östergötland County in south east Sweden in 2006. The CLT was evaluated and found to be feasible for both staff and patients [bib_ref] Computerized lifestyle intervention in routine primary health care: evaluation of usage on..., Carlfjord [/bib_ref]. Adoption of the CLT was found to vary substantially between the PHC units and questions were raised about how contextual factors at baseline might have influenced the implementation and about sustainability over time. This resulted in a decision to perform another study, offering the tool to the remaining 32 PHC units in Östergötland County. If there is an association between organizational climate and willingness to adopt a new way of working, climate could be seen as an implementation facilitator, and a climate assessment could help in developing a tailored implementation strategy. Organizational climate has earlier been found to be associated with youth outcomes in child welfare systems [bib_ref] Organizational climate, services, and outcomes in child welfare systems, Glisson [/bib_ref] , but to our knowledge, no studies have compared organizational climate at baseline with a long-time perspective outcome of an implementation. The aim of this study was to assess the association between organizational climate when a tool for lifestyle intervention was introduced in PHC and implementation outcome in terms of how the tool was perceived and used after 2 years. # Methods ## Setting and study population In 2008-2010, the CLT was offered to all 32 PHC units (i.e. health care centres with general practitioners (GPs) and other health care professionals) in Östergötland County, Sweden, that had not yet tried it. Of these, 22 units agreed to participate in the study. ## The clt The CLT consists of a lifestyle assessment performed on a touch-screen computer, and provides tailored advice to the patient, based on their individual answers. Staff members at the PHC unit are encouraged to refer their patients (aged ≥18 years) to the computer, and to give them the opportunity to discuss the results and provided advice. The CLT has been described in detail previously [bib_ref] Computerized lifestyle intervention in routine primary health care: evaluation of usage on..., Carlfjord [/bib_ref]. ## Implementation strategy The implementation strategy used was based on Rogers' innovation-decision process including knowledge, persuasion, decision and implementation. The implementation process began with an information session where a change agent from the research team visited the unit to provide information about the CLT. The information session was followed by a trial for 1 month, during which all staff members were encouraged to perform the lifestyle assessment provided by the CLT and give their opinions about it. The trial was part of the persuasion stage in Rogers' model and the aim was to make the staff aware of the trialability and observability of the CLT. After the 1-month trial, the change agent visited the unit again. There was a discussion about how the CLT could be used within the daily routine, and mutual agreement to incorporate it as a working method or not was made. The purpose of this was to facilitate the decision stage. After the second meeting, the CLT was made available to patients. ## Data collection Before the first visit at each unit, a questionnaire was sent to all staff members who meet patients in their daily work (administrative staff excluded) to assess the organizational climate. The Creative Climate Questionnaire (CCQ) developed by Ekvall was found to be feasible for the study [bib_ref] Organizational climate for creativity and innovation, Ekvall [/bib_ref]. The instrument measures individual perceptions of the organizational environment and has often been used in health care settings [bib_ref] Organizational climate for creativity and innovation, Ekvall [/bib_ref] [bib_ref] Evidence-based practice and determinants of research use in elderly care in Sweden, Bostrom [/bib_ref] [bib_ref] The relationship between organizational climate and the content of daily life for..., Norbergh [/bib_ref] [bib_ref] Leadership behaviour of nurse managers in relation to job satisfaction and work..., Sellgren [/bib_ref]. The CCQ instrument covers ten dimensions of organizational climate and consists of 50 statements that are answered on a 4-point scale (0-3) showing extent of agreement. The statements are put as: "There are many new ideas floating around here". The ten dimensions are: Challenge, Freedom, Idea support, Trust/Openness, Dynamism/Liveliness, Playfulness/Humour, Debate, Conflicts, Risk taking and Idea time. The higher the value, the more creative the climate, except for the dimension Conflicts for which the reverse applies. Ekvall also presents reference values for innovative and stagnated organizations for each of the ten dimensions [bib_ref] Organizational climate for creativity and innovation, Ekvall [/bib_ref]. When the CLT had been in operation for 24 months, another questionnaire was sent to the same staff categories. Respondents were regarded as a closed cohort, so that only those who had received the first questionnaire were included in the study population for the second questionnaire. As perceived need for the innovation has been found to be an important predictor for implementation, there was one question about this: "I see no need for the CLT", to which the responder could agree or disagree. The other questions were formulated to capture how well the CLT had been adopted by the staff members. Some of the questions concerned how the CLT was perceived by the staff group, according to the responder, and are considered an evaluation of perceptions at the unit level; other questions concerned the individual's personal perceptions of the CLT. One question concerned frequency of referral to the CLT, and is considered a measurement of use at the individual level. The questions and the response alternatives are given in the Results section. # Data analysis The CCQ assessment at baseline resulted in a mean value for each CCQ dimension, as suggested by Ekvall et al.. Based on these results, the units were categorized into three groups, high CCQ (Hi), medium CCQ (Med) and low CCQ (Lo). The categorization was based on how many dimensions that, statistically significant, exceeded or fell short to the reference values presented by Ekvall et al. [bib_ref] Organizational climate for creativity and innovation, Ekvall [/bib_ref] using the Student's t-test. Units with a higher number of dimensions exceeding than falling short to the reference values were defined as Hi, units with less than six more dimensions falling short to than exceeding the reference values were defined as Med and units with at least six more dimensions falling short to than exceeding the reference values were defined as Lo. This resulted in 4 units in group Hi, 15 units in group Med,and 3 units in group low. See also [fig_ref] Table 1: CCQ values in units defined as Hi or Lo and reference values... [/fig_ref]. When the numbers were calculated the reverse nature of the dimension Conflicts was taken into account. The cut-off values were chosen to assure that units presenting top scores were compared to units presenting bottom scores. Associations between perceived need for the CLT and the CCQ dimensions were analysed with Spearman's rank correlation coefficient. The Kruskall-Wallis test was used to calculate mean ranks regarding perceived need. Generalized Estimating Equations (GEE), which allows for analysis of clustered data, was used to calculate the Wald Chi-Square test in order to compare the distribution of responses to the follow-up questionnaire among the individual staff members from the three CCQ groups. Statistical analyses were performed using the computer-based analysis program, Statistical Package for the Social Sciences (SPSS) version 22.0. Statistical significance was set at p ≤ 0.05. ## Ethics The study was carried out in compliance with the Helsinki Declaration and was approved by the Regional Ethical Review Board in Linköping, Sweden (Dnr Ö 16-08). As the participants were staff members, written consent was not required according to Swedish regulations (SFS 2003:460). # Results The average response rate was 53 % (range 30-81 %) at baseline and 52 % (range 20-86 %) at follow-up. The total number of respondents was 322 at baseline and 239 at follow-up. To be included at baseline, the responder had to have completed the CCQ instrument, leaving 275 (45 %) respondents for the analysis. An analysis of the drop-outs showed no differences between responders and the entire sample in terms of profession or gender, except that at baseline, the proportion of GPs was lower among the responders (12 %) than in the entire sample (20 %). CCQ scores for the units in group Hi and group Lo are shown in [fig_ref] Table 1: CCQ values in units defined as Hi or Lo and reference values... [/fig_ref]. At follow-up, internal non-response occurred for some of the questions; the number of responders for each question is shown in Tables 2 and 3. The question on perceived need for the CLT revealed considerable differences between the units (p = 0.004). The mean ranks for each unit are presented in [fig_ref] Table 4: CCQ group and perceived need for the CLT in terms of mean... [/fig_ref]. When the item was analysed on a group level, groups Med and Lo showed homogeneity, with no significant differences within groups (p = 0.060, p = 0.808). Group Hi showed withingroup differences (p = 0.003). It was found that the unit with the lowest perceived need was one of the units in the Hi group; this unit was excluded from further analysis, leaving three units in group Hi, which thus was homogeneous regarding perceived need (p = 0.481). The unit with lowest perceived need was also found to score low on perceptions and use. The question capturing use in terms of frequency of referral to the CLT showed significantly higher levels of use in group Hi compared with group Lo (p < 0.001) as shown in [fig_ref] Table 2: Perceptions and use of the CLT related to the individual [/fig_ref]. For the follow-up questions assessing individual perceptions of the CLT, there was a significant difference between respondents in two of the three items when group Hi was compared with group Lo. Compared with staff in group Lo, staff in group Hi stated to a higher degree that the CLT facilitated their own work (p = 0.002) and that they believed the CLT could influence patients' lifestyles (p < 0.001). For questions regarding perceptions at the unit level, group Hi was significantly more positive than group Lo for all four items assessed [fig_ref] Table 3: Perceptions of the CLT related to the unit [/fig_ref]. Among staff at the units in group Hi, the CLT was discussed more often (p = 0.006), use of the CLT was more supported (p < 0.001), was perceived to a higher degree to facilitate the work with lifestyle issues (p < 0.001), and was seen as an important part of the work with lifestyle issues (p = 0.001), compared with group Lo. # Discussion The aim of the study was to assess the association between the creative climate at baseline and perceptions and use of a tool for lifestyle intervention in PHC 2 years after its introduction. The main finding was that the units having a positive creative climate (group Hi) adopted the new tool to a higher degree than those with the least positive creative climate (group Lo). Differences were found regarding use at the individual level, and regarding perceptions at both the individual and unit levels. This result was based on self-reported data, and, at the individual level, concerned use in terms of frequency of referral, and perceptions of the CLT such as facilitating work with lifestyle issues and being an effective way to influence lifestyle. At the unit level, staff in group Hi were found to discuss the CLT, support its use and find that it facilitated work with lifestyle issues at the unit. Staff in group Hi did refer patients to the CLT to a higher degree than staff in group Lo, but only one third of the respondents in group Hi referred patients more than once a month. This could not be considered a frequent use of the CLT. The fact that staff members express positive perceptions and seem to have adopted the new tool, but do not use it on a regular basis, could be explained in terms of conceptual or instrumental use, terms often applied in the context of research use. Conceptual use refers to a change in the way practitioners think, their knowledge, understanding and attitudes, whereas instrumental use refers to the direct impact on practice decisions. In the present study, staff at the units with a positive creative climate appreciate the CLT, which they find useful in their practice, but they still, after 2 years, have not incorporated it in their daily practice. In a former qualitative evaluation of the CLT at six PHC units, staff claimed that they often forget to refer their patients to the CLT, despite their good intentions [bib_ref] Key factors influencing adoption of an innovation in primary health care: a..., Carlfjord [/bib_ref]. This could also be due to the role of habit as a barrier for the implementation of change in clinical practice, which sometimes seems to be a stronger factor than knowledge and intentions [bib_ref] Creatures of habit: accounting for the role of habit in implementation research..., Nilsen [/bib_ref]. The positive perceptions of the CLT found in group Hi, might result in a higher level of use in a longer perspective, but probably additional efforts from a change agent or the local management is still needed. Measuring creative climate at baseline could be one way to predict implementation outcome. It could be helpful in tailoring the interventions when an implementation process is planned, so that the implementation activities can be designed to suit the specific setting. A simplified way of describing tailored implementation is that it is similar to the tailoring of a clinical treatment to a diagnosed health problem [bib_ref] Developing and selecting interventions for translating knowledge to action, Wensing [/bib_ref]. Using that metaphor assessment of creative climate could help to diagnose the problem, and a tailored approach would be the treatment. In our study the unit with the lowest level of perceived need showed high scores on creative climate, but low scores on perceptions and use of the innovation. Previous research has shown that a certain level of perceived need is important for an innovation to be successfully implemented [bib_ref] The challenge of innovation implementation, Klein [/bib_ref]. If the potential adopters see no need for a certain new practice, it probably should not be implemented at all until staff have gained the required knowledge or curiosity about the innovation so that there is a so-called user pull as described by Lavis et al. [bib_ref] Assessing country-level efforts to link research to action, Lavis [/bib_ref]. When the implementation of the CLT was evaluated in another setting, in terms of sustainability after 2 years, staff expectations, perceptions of the innovation's relative advantage and potential compatibility with existing routines were factors found to be positively associated with implementation outcome [bib_ref] Sustained use of a tool for lifestyle intervention implemented in primary health..., Carlfjord [/bib_ref]. # Methodological considerations Implementation is known to be a complex process, and there might have been other factors influencing the outcome, which could be seen as a limitation of the study. For example, adopter characteristics, outer context or the presence of opinion leaders, all potentially influencing factors according to Greenhalgh et al. [bib_ref] Diffusion of Innovations in Health Service Organisations: A Systematic Literature Review, Greenhalgh [/bib_ref] , were not taken into account. Another limitation of the study is the use of self-reported data. It is known that selfreporting might be biased by social desirability and it cannot be ruled out that the respondents overestimated their perceptions and use of the CLT. This, however, could be assumed to be equal for all the participants. The sample size in this study is rather small, which may have influenced the GEE test so that the results are not as conservative as they should be, which is also a limitation. Measuring creative climate using the CCQ is one way of assessing context. An advantage with the CCQ instrument is that it was originally developed in Sweden, and the disadvantages associated with translating, such as misinterpretation of questions, are avoided. The instrument has also been used and found feasible in other studies in Swedish health care settings [bib_ref] Evidence-based practice and determinants of research use in elderly care in Sweden, Bostrom [/bib_ref] [bib_ref] The relationship between organizational climate and the content of daily life for..., Norbergh [/bib_ref] [bib_ref] Leadership behaviour of nurse managers in relation to job satisfaction and work..., Sellgren [/bib_ref]. Creative climate is known to be stable over time if no efforts are made to alter it [bib_ref] Organizational climate for creativity and innovation, Ekvall [/bib_ref] , which is why no data on creative climate were collected at follow-up. Unexpected changes in creative climate might have affected the results. The unit with the lowest perceived need and the highest creative climate was excluded from further analysis. A lack of perceived need at this stage, 2 years after the introduction of the CLT, was interpreted to mean that other ways of handling lifestyle issues were available, which made the CLT unnecessary and thus not used at all. Thus, the unit differed in such an important way from the other units in the CCQ Hi group that it was considered most appropriate to exclude it. # Conclusions According to the results from this study there is an association between organizational climate at baseline and implementation outcome after 2 years when a tool for lifestyle intervention is introduced in PHC in Sweden. Since the findings are based on self-reports and the use of the CLT was relatively low, further studies are needed before measurement of organizational climate at baseline can be recommended in order to predict implementation outcome. [table] Table 1: CCQ values in units defined as Hi or Lo and reference values for innovative organizations [/table] [table] Table 2: Perceptions and use of the CLT related to the individual [/table] [table] Table 3: Perceptions of the CLT related to the unit [/table] [table] Table 4: CCQ group and perceived need for the CLT in terms of mean rank I see no need for the CLT (agree/partly agree/ partly disagree/ disagree) Mean rank according to the Kruskall-Wallis test for the units. Lower mean rank indicates less perceived need Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color figure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit [/table]

A rare case of spontaneous abdominal aorta thrombosis Spontaneous aortic thrombosis is rare, and prompt diagnosis is needed whenever encountering a case of unexplained abdominal pain. The cause of the thrombosis needs to be evaluated thoroughly to rule out any underlying thrombophilias. # | introduction Spontaneous abdominal aorta thrombosis is extremely rare with an unclear incidence due to its infrequent occurrence. Symptoms vary widely, ranging from being asymptomatic to a more catastrophic presentation with organ ischemia. We report the case of a 40-year-old lady who presented with 1-week history of abdominal pain and was found to have spontaneous abdominal aorta thrombosis. A 40-year-old lady was presented to the hospital with 1-week history of generalized abdominal pain, moderate to severe in intensity, associated with nausea and four episodes of vomiting. An initial CT scan of abdomen with contrast showed features of abdominal aorta thrombosis with ischemic colitis. Screening for an autoimmune etiology or hereditary thrombophilia was negative. The patient was treated with anticoagulation and she showed evidence of improvement in clinical symptoms and partial resolution of thrombus on follow-up radiology. Spontaneous aortic thrombosis is rare in occurrence. Evaluation includes abdominal imaging with either CT scan or MRI, and the cause of the thrombosis needs to be evaluated thoroughly. Treatment includes anticoagulation and if needed re-vascularization by endovascular methods. Spontaneous abdominal aorta thrombosis is extremely rare with an unclear incidence due to its infrequent occurrence. To this date, there are very limited data on spontaneous abdominal aorta thrombosis in literature. [bib_ref] Mobile thrombus in the ascending aorta associated with acute myocardial infarction, Lemaire [/bib_ref] It is of clinical significance because it can lead to devastating manifestations if it is not detected and treated in a timely manner. Symptoms vary widely, ranging from being asymptomatic to a more catastrophic presentation with organ ischemia. [bib_ref] A loose cannon: free-floating thrombus in ascending aorta, Madershahian [/bib_ref] We present a case of a patient with severe abdominal pain who was found to have ischemic colitis secondary to spontaneous abdominal aorta thrombosis. ## | case presentation A 40-year-old lady was presented to the hospital with 1-week history of generalized abdominal pain, moderate to severe in intensity, associated with nausea and four episodes of vomiting, with no aggravating or relieving factors. There was no history of fever, hematemesis, bleeding per rectum, chest pain, or palpitations. The patient was a known case of hypothyroidism, vitiligo, and type 1 diabetes mellitus. She was a non-smoker. The patient had no personal or family history of thrombophilia and there was no history of recurrent abortions. Physical examination revealed an afebrile patient, with a heart rate of 92 beats per minute, blood pressure 124/76 and respiratory rate 17 breaths per minute. Abdominal examination showed diffuse tenderness, no organomegaly, normal percussion note with normal bowel sounds. Cardiac, respiratory, and nervous system examination were normal. Labs investigations showed mild neutrophilic leukocytosis, high C-reactive proteins, microcytic anemia, normal urea, creatinine, electrolytes, and liver function tests . Coagulation testing showed normal PT, aPTT, and INR . A computed tomography (CT) scan of the abdomen with contrast showed diffuse thickening of the splenic flexure, descending colon, and sigmoid colon with surrounding mesenteric edema and fat stranding as well as a thrombus in the abdominal aorta at the level of L3 vertebra . However, all the thrombophilia screen and autoimmune work up were negative . The patient was managed as a case of ischemic colitis due to spontaneous abdominal aorta thrombosis. She was started on anticoagulation with warfarin bridged by therapeutic dose enoxaparin with a target INR of 2-3. A follow-up CT scan of the abdomen that was done 2 months afterward showed resolution of thrombus and improvement of the inflammatory changes in the bowel . Three months afterward, she presented with left-sided abdominal pain, nausea, and vomiting. CT scan of the abdomen was done, and it showed focal segmental circumferential mural thickening of the distal descending colon as a complication of the ischemic colitis the patient was admitted and underwent left hemicolectomy. She was then discharged home after improvement. # | discussion Spontaneous arterial thrombosis is rare in occurrence, but it has a significant clinical importance as it leads to several complications secondary to blood flow impediment. [bib_ref] Arterial thrombosis in unusual sites: a practical review, O&apos;donnell [/bib_ref] It can involve any organ system with the resulting ischemia and/ or infarction. Prompt diagnosis and management are always needed as any delay can cause acute complications which can include renal infarction, stroke, abdominal, and peripheral ischemia with gangrene and might eventually lead to death. Presentation depends on the system or the organ involved. [bib_ref] Arterial thrombosis in unusual sites: a practical review, O&apos;donnell [/bib_ref] Clinically significant thrombosis can be because of a primary hypercoagulable state or as secondary to trauma, immobilization, or other coagulopathic abnormalities. Suspicion should arise toward an underlying abnormality whenever a patient has history of unprovoked arterial or venous occlusion. Evaluation for causes such as factor V Leiden, antithrombin III deficiency, or protein C or S deficiency is warranted in unprovoked cases.Spontaneous abdominal aorta thrombosis is rare as the blood flowing in the aorta has a high velocity which, along with a high caliber of the artery, makes it less prone to thrombosis and it is almost always secondary to an underlying hypercoagulable status. [bib_ref] Complicated course consequences of a floating thrombus in ascending aorta, Mandegar [/bib_ref] It can present with manifestations that could range from asymptomatic in cases of chronic thrombosis to a more severe one in patients with acute thrombosis which can lead to devastating results with hemodynamic instability and even peripheral or abdominal organ loss. [bib_ref] Floating, non-occlusive, mobile aortic thrombus and splenic infarction associated with protein C..., Sanon [/bib_ref] As the abdominal aorta is the major source of arterial supply, its involvement can lead to renal artery ischemia with renal infarction, colonic ischemia, and limb ischemia. Diagnosis should always be prompt. The initial evaluation should start with basic laboratories such as CBC, which can show increased white cell count, increased lactate denoting ischemia, acute kidney injury, or liver injury. Urgent imaging evaluation should be done next. CT scan with contrast is usually the imaging diagnostic of choice as it can reveal the thrombus CT scan can also help in excluding other diagnoses. [bib_ref] Ascending aorta thrombus: a diagnostic and treatment dilemma, Soo [/bib_ref] A full and thorough evaluation for possible underlying hypercoagulability must be sought. Treatment should be pursued as soon as possible, and it usually depends on the presentation and hemodynamic stability. Usually, the approach in stable patients with an established thrombus is to start on anticoagulation with warfarin and bridging enoxaparin. [bib_ref] Anticoagulation is an effective treatment for aortic mural thrombi, Bowdish [/bib_ref] There are currently limited data at present on the novel anticoagulants use in hereditary thrombophilia's and in spontaneous aorta thrombosis. The duration of anticoagulation is usually guided by the underlying cause, as some patients require life-long anticoagulation for prophylaxis. 9 # | conclusion Spontaneous aortic thrombosis is rare in occurrence, and prompt diagnosis is needed whenever encountering a case of unexplained abdominal pain or embolic phenomena. Evaluation includes abdominal imaging with either CT scan or MRI, and the cause of the thrombosis needs to be evaluated thoroughly to rule out any secondary hypercoagulable states as well as evaluating for primary hypercoagulability. Treatment includes anticoagulation and if needed revascularization by endovascular methods.

Anterior cruciate ligament reconstruction complicated by pyoderma gangrenosum We report a case of pyoderma gangrenosum as a complication of an anterior cruciate ligament reconstruction in a patient with inflammatory bowel disease, which was misdiagnosed initially as a post-operative wound infection. An early dermatology opinion and skin biopsy should be considered in cases of suspected infection where thorough surgical debridement and antimicrobial therapy has failed to improve the clinical picture. Pyoderma gangrenosum (PG) is a rare, immunological, ulcerative condition of the skin that may occur spontaneously or secondarily to injury or surgery. This condition is known to be associated with inflammatory bowel disease and, in the context of surgery, may be misdiagnosed as a post-operative wound infection. We present a case of PG as a complication of an anterior cruciate ligament (ACL) reconstruction, mimicking a surgical site infection. ## Case history A 33-year-old male engineer presented with symptoms of right knee instability following a severe valgus twisting injury when he fell downstairs. He was known to have a history of ulcerative colitis. Following magnetic resonance imaging, he went on to have an uneventful, arthroscopically assisted ACL reconstruction using an ipsilateral hamstrings graft, with concomitant medial and lateral meniscal repairs. Four days following surgery, the patient was admitted under the gastroenterologists with abdominal pain secondary to an acute exacerbation of his ulcerative colitis. His temperature on admission was 39°C and inflammatory markers showed a C-reactive protein (CRP) level of 329mg/l (105mg/l pre-operatively) and a white cell count of 16 x 10 9 /l. Our physicians commenced him on intravenous (IV) co-amoxiclav. At this stage, the right knee was quiescent and the post-operative wounds looked healthy. On the tenth day after surgery, the patient had persistent fever and rigours, and the CRP level remained high (340mg/l). The IV co-amoxiclav was switched to vancomycin. The following day, the graft harvest site appeared clini- cally infected with significant erythema and pus-like discharge [fig_ref] Figure 1: Initial appearance of wound on post-operative day [/fig_ref]. Given these findings, he was taken to theatre where he underwent wound debridement and lavage with aspiration of the knee joint using a superolateral approach . A VAC ® dressing (Kinetic Concepts, San Antonio, TX, US) was applied to the wound and IV antibiotics were continued. Over the next 12 days, the wound failed to settle and a further 5 surgical debridements were performed. A large area of skin ulceration was evident over the anteromedial aspect of the right proximal tibia, measuring approximately 15cm x 15cm . Despite sending multiple microbiology specimens at each stage, we failed to grow any bacterial organism. A dermatology opinion was gained, and a diagnosis of PG was made on the basis of the clinical findings and the histology results taken from a skin biopsy. The patient was commenced immediately on high dose oral steroid therapy. Sixteen weeks after his primary ACL reconstruction, the wound had fully healed with no requirement for plastic surgical intervention or skin grafting. The ACL graft remains intact. # Discussion PG is one of the most dramatic and rapidly progressing skin conditions. It was first reported by Brunsting et al in 1930 when they published a series of five cases presenting with inflammatory sterile neutrophilic dermatosis. [bib_ref] Pyoderma (ecthyma) gangrenosum: clinical and experimental observations in five cases occurring in..., Brunsting [/bib_ref] The cases were misdiagnosed initially as infection. Since then, PG has been a diagnostic dilemma for clinicians. Due to the unknown aetiology, there is no clear diagnostic test; microbiology cultures are negative and histopathology is usually not definitive. PG can be idiopathic or associated with other conditions such as inflammatory bowel disease, myeloproliferative disease, Wegener's granulomatosis and, more worryingly, it can be paraneoplastic, 2 requiring rapid diagnosis and appropriate treatment. It can also present after surgery, espe-cially abdominal and breast surgery. To our knowledge, only six orthopaedic cases have been reported with PG following surgery. These cases occurred after hip and knee arthroplasty, knee arthroscopy, tarsal tunnel release and an operation on a rheumatoid hand. [bib_ref] Pyoderma gangraenosum as a complication to knee arthroscopy, Madsen Jt [/bib_ref] [bib_ref] Successful treatment of wound breakdown caused by pyoderma gangrenosum after total knee..., Nakajima [/bib_ref] The patients present with a rapidly developing pustule surrounded by an area of erythema and oedema. It then ulcerates, developing undermined borders with bluish colouration. These ulcers can develop a purulent cover. They can become painful and the patient develops systemic symptoms. Blood tests show raised inflammatory markers. Histopathology is not specific; it shows leucocytoclastic vasculitis and sometimes necrotising granulomatous inflammatory changes. [bib_ref] Pyoderma gangrenosum: a review, Crowson [/bib_ref] These features are often misdiagnosed as infection. While PG is treated by immunosuppressive therapy, post-operative wound infections are managed with high dose IV antibiotics and surgical debridement as these can be limb or even life threatening. All the reported cases, including our patient, had high doses of IV antibiotics and underwent multiple surgical debridement and lavage procedures. This often resulted in the development of skin and soft tissue defects that required a prolonged hospital stay and plastic surgical intervention. In some cases, the decision to amputate or perform an arthrodesis of the limb was considered. [bib_ref] Pyoderma gangraenosum as a complication to knee arthroscopy, Madsen Jt [/bib_ref] [bib_ref] Successful treatment of wound breakdown caused by pyoderma gangrenosum after total knee..., Nakajima [/bib_ref] In the present case, the surgical wound did not respond to standard surgical management or IV antibiotic therapy. A dermatology opinion was therefore sought and the diagnosis of PG was made. The patient responded to systemic corticosteroid treatment and did not require any plastic surgical input. # Conclusions We believe PG must be considered in all cases of suspected wound infection that are refractory to standard treatment options. PG is often a clinical diagnosis of exclusion and an early opinion from an experienced dermatologist should be Wound following initial surgical debridement Wound following extensive surgical debridement over a 12-day period BAgOURI SmITH gEUTjENS ANTERIOR CRuCIATE LIgAmENT RECONSTRuCTION COmPLICATEd By PyOdERmA gANgRENOSum sought in these cases with samples sent for histology along with the microbiology samples, especially in cases at risk or those not responding to appropriate antibiotics and surgical treatment. [fig] Figure 1: Initial appearance of wound on post-operative day [/fig]

Baseline Amino Acid Substitutions in the NS5A ISDR and PKR Binding Domain of Hepatitis C and Different Fibrosis Levels and Levels of Development of Hepatocellular Carcinoma in Patients Treated with DAAs S.P.); [email protected] (A.P.);Abstract: Variations in the interferon sensitivity-determining region (ISDR) within the NS5A region were related to the development of hepatocellular carcinoma (HCC) in patients infected with hepatitis C virus (HCV). The aim of the study was to investigate a relationship between ISDR/PKR substitutions and their association with liver fibrosis or HCC development. A total of 316 patients infected with HCV and treated with DAAs were evaluated. HCV RNA was quantified and sequenced before treatment. The liver fibrosis stage was assessed by transient elastography and equalized to METAVIR scores. Multivariate analysis showed that ≥3 substitutions in ISDR and ≥6 in PKR-bd were significantly associated with advanced fibrosis. Advanced fibrosis was observed in patients with higher substitutions in ISDR and PKR-bd. A higher correlation between advanced fibrosis and a high frequency of ≥3 substitutions in ISDR and ≥6 in PKR-bd was observed in patients infected with genotype 2c. In addition, in a higher proportion of HCC patients, advanced fibrosis (40.4% vs. 88.2%; p < 0.001) and ≥6 substitutions in PKR-bd (15.4% vs. 41.2%; p = 0.01) was observed. In conclusion, a higher number of substitutions in ISDR and PKR-bd were associated with advanced liver fibrosis, suggesting a use of like predictors for progression in the liver damage. A significantly higher number of PKR-bd substitutions was observed in HCC patients; in particular, in patients infected with HCV genotype 2c. # Introduction Hepatitis C virus (HCV) infection is one of the main causes of hepatocellular carcinoma (HCC). Therapy to eradicate HCV prevents the progression of liver fibrosisand the development of HCC. However, HCC can develop even in patients with chronic hepatitis C who achieve sustained virologic response (SVR). Previous studies have shown that liver fibrosis is strongly associated with the development of HCC after SVR. Thus, patients at risk of HCC should continue to have close surveillance after SVR. In addition, it is important to determine which patients might have higher risk factors for HCC after SVR, and for how long they should be monitored. Due to the recent introduction of direct-acting antiviral drugs (DAAs) for HCV, an increased number of patients with chronic hepatitis C have achieved SVR. However, the high degree of HCV genetic diversity may play an important role in the virus's ability to evade the immune response and antiviral therapy's selective pressure. This variability could be focused on changes occurring at different positions in the NS3 protease, NS5B polymerase, and NS5A protein. Several studies have highlighted the influence of sequence heterogeneity within a particular region, especially in the NS5A, on the outcome of IFN-based therapy due to its relationship with IFN responsiveness. NS5A modulates HCV replication through interactions with other viral proteins and host proteins to form the HCV replication complex and regulate host cell functions, including viral pathogenesis. In particular, amino acid substitutions in the interferon sensitivity-determining region (ISDR) within the NS5A region have been related to advanced fibrosis and the development of HCC in patients infected with HCV. In addition, it is known that ISDR is located within the protein kinase R-binding domain (PKR-bd) involved in a tumor suppressor function. Other studies reported that NS5A resistance-associated substitutions (RASs) were associated with advanced liver fibrosis. However, the relationship between NS5A substitutions and liver disease progression remains unclear. The aim of the present study is to investigate a potential relationship between the number of baseline ISDR/PKR-bd substitutions and the liver fibrosis stage and/or the development of HCC in SVR after treatment with DAAs. # Materials and methods ## Patient and clinical characteristics This study was conducted in accordance with the principles of the Declaration of Helsinki and approved by the Institutional Review Board of the Fondazione IRCCS Policlinico San Matteo (protocol no. 20080009620, approval date: 11 March 2008). Written informed consent was obtained from all participants in the study. A total of 316 patients infected with HCV with genotypes 1a, 1b, 2c, 3a, and 4d referred to the Fondazione IRCCS Policlinico San Matteo, Pavia between January 2017 and April 2019 were evaluated in the study. All patients were treated with DAAs. In addition, 70/316 patients had previously experienced peg-IFN treatment. The liver fibrosis stage was assessed by transient elastography and equalized to METAVIR scores, and F0-F2 was classified as low fibrosis, while F3-F4 as the advanced fibrosis stage. Patients were monitored for at least 3 months after the completion of treatment to evaluate SVR. The follow-up period was defined as the period from the confirmation of SVR to HCC development or the last visit to assess HCC. Currently, patients with advanced fibrosis before treatment are monitored after stopping therapy for the duration of their lives to see the possible progression in HCC. ## Hcv-rna quantification and sequencing Serum samples were collected at the baseline in all 316 patients treated with DAAs included in the study, and the HCV load was quantified by using the Abbott HCV-RNA assay (Abbott Park, IL, USA). HCV genotyping was performed using the Abbott RealTime HCV Genotype II assay. In addition, in order to resolve possible ambiguities during the genotyping, the NS3/NS5B region was sequenced to further subtype the HCV strains. Data were analyzed using the Blast program (http://blast.ncbi.nlm.nih.gov). Viral RNA was extracted from serum samples using the automatic Easy Mag extractor (Biomerieux, Lyon, France). HCV NS5A domain I (aa 1-406), NS3-protease (aa 1-181), and NS5B (aa 1-591) genes were amplified using a nested RT-PCR. The amplified NS5A region Viruses 2020, 12, 255 3 of 10 included PKR-bd (from aa 237 to 302) and ISDR (from aa 237 to 302). Direct sequencing was performed by using an automatic sequencer (ABI PRISM 3130xl genetic analyzer DNA Sequencer, Applied Biosystems, Foster City, CA, USA) and the BigDye Terminator v1.1 Cycle Sequencing kit (Applied Biosystems) to evaluate the effect of the most represented viral quasispecies as already reported. The RASs were defined according to the geno2pheno algorithm (http://hcv.geno2pheno.org/index.php) and other clinical and in vitro data for RAS interpretation. Nucleotide sequences were aligned with MEGA 7.0 software and compared with the confirmed references: AF009606 for subtype 1a, D90208 for subtype 1b, JX227965 for subtype 2c, HM042073 for subtype 3a, and FJ462437 for subtype 4d. The reference strains have been chosen to be as homologous as possible to the HCV viral strains of the infected patients. # Statistical analysis Continuous variables (i.e., viral load) were compared using the Mann-Whitney U test for independent nonparametric data. Categorical variables were compared by the chi-square or Fisher's exact test, as appropriate, and p-values of ≤0.05 were considered statistically significant. All these statistical analyses were performed using Graph Pad Prism software (version 5.00.288). The cut-off values for the number of ISDR and PKR-bd amino acid substitutions were defined as three and six, respectively, by using ROC curve analysis. Univariate and multivariate analyses were performed under logistic regression models fitted for predicting high-grade fibrosis or HCC. These statistical analyses were performed using StataCorp USA 2017 (v15.1). # Results Overall, only 9/316 (2.8%) patients failed to achieve SVR after treatment with DAAs. Development in HCC was observed in 17/316 (5.4%) patients within a year after treatment with DAAs. The evaluation of variants of the HCV NS5A-ISDR and PKR-bd genes was performed at pretreatment in all 316 patients. Overall, comparing ISDR and PKR-bd substitutions between SVR and failing-treatment patients, no significant differences were observed. In detail, the mean number of mutations observed in SVR and failing-treatment patients was comparable (1.85 ± 1.80 and 1.67 ± 1.32 in ISDR and 4.51 ± 2.42 and 4.67 ± 2.55 in PKR-bd, respectively p > 0.05). On the other hand, among all patients at baseline, a greater number of substitutions in both regions were observed in patients with advanced fibrosis levels. The mean number of mutations observed in the two regions of HCV patients with low and high fibrosis levels was comparable in ISDR (1.56 ± 1.28 and 2.20 ± 2.31; p = 0.13), while it was significantly different in PKR-bd (4.04 ± 1.74 and 5.07 ± 2.94; p = 0.006). The results of univariate and multivariate analysis of factors associated with high fibrosis levels are shown in. In detail, 180/316 (56.9%) patients had low fibrosis values (F0-F2), and 136/316 (43.0%) patients had high fibrosis values (F3-F4). Univariate analysis showed that higher HCV RNA levels (p = 0.003), aspartate aminotransferase (AST) level (p < 0.001), γ-glutamyl transpeptidase (γ-GTP) levels (p < 0.001), male gender (p = 0.019), lower platelet (PLT) counts (p < 0.001), ≥3 substitutions in ISDR (p < 0.001), ≥6 substitutions in PKR-bd (p < 0.001), and PKR-bd with INS/DEL were associated with advanced fibrosis. According to multivariate analysis, higher HCV RNA level (p = 0.002), γGTP level (p = 0.002), lower platelet count (p < 0.001), ≥3 substitutions in ISDR (p = 0.007), and ≥6 in PKR-bd (p < 0.001) were identified as independent factors significantly associated with the development of advanced fibrosis. The number of substitutions observed in PKR-bd were significantly higher than those observed in ISDR (p < 0.001 vs. p = 0.007). Clinical characteristics analyzed according to substitutions in the ISDR and PKR-bd are shown in. Among patients with ≥3 substitutions in ISDR and ≥6 in PKR-bd, older age (p = 0.006, p < 0.001) higher AST levels (p = 0.03, p < 0.001), γGTP levels (p = 0.01, p = 0.003), ALT levels (p = 0.01) in PKR-bd, and advanced fibrosis (p < 0.001, p < 0.001) were more frequently observed. In particular, a higher proportion of patients with ≥3 substitutions in ISDR and ≥6 in PKR-bd was observed in patients infected with genotype 2c (p < 0.001 in both genes). In addition, according to genotypes, a higher correlation between advanced fibrosis and ≥3 substitutions in ISDR and ≥6 substitutions in PKR-bd was observed in patients infected with genotype 2c (p < 0.001, p < 0.001), followed by genotype 1b (p = 0.007, p = 0.04) and genotype 1a (p = 0.013, p = 0.002). To note, HCV with insertions/deletions (INS/DEL) in PKR-bd was observed in 16/316 (5.0%) patients: 10 infected with genotype 2c, 3 with genotype 1b, and 3 with genotype 3a. Among them, a significantly greater proportion of patients had advanced fibrosis stage (11/16, 67.8%; p = 0.0037). In particular, significant lower HCV RNA levels were observed in patients carrying HCV with INS/DEL in ISDR (amino acid 237 to 270), compared to patients with HCV carrying INS/DEL in the remaining part of PKR-bd (amino acid 271 to 302) (p = 0.021). The comparison of factors according to the development of HCC showed that older age (p = 0.05), lower HCV RNA levels (p < 0.001), higher alanine aminotransferase (ALT) level (p < 0.001), AST level (p < 0.001), γGTP level (p = 0.02), and lower platelet count (p < 0.001) were observed at the baseline of patients who developed HCC after treatment with DAAs. In particular, in a higher proportion of patients who developed HCC advanced fibrosis (40.4% vs. 88.2%; p < 0.001) and ≥6 substitutions in PKR-bd (15.4% vs. 41.2%; p = 0.01) was observed. In addition, the proportion of patients who developed HCC was higher among patients infected with genotype 1b (47%) than with genotype 1a (0%), genotype 2c (29.4%), genotype 3a (23.5%), or genotype 4d (0%). higher correlation between advanced fibrosis and ≥3 substitutions in ISDR and ≥6 substitutions in PKR-bd was observed in patients infected with genotype 2c (p < 0.001, p < 0.001), followed by genotype 1b (p = 0.007, p = 0.04) and genotype 1a (p = 0.013, p = 0.002). To note, HCV with insertions/deletions (INS/DEL) in PKR-bd was observed in 16/316 (5.0%) patients: 10 infected with genotype 2c, 3 with genotype 1b, and 3 with genotype 3aThe comparison of factors according to the development of HCC showed that older age (p = 0.05), lower HCV RNA levels (p < 0.001), higher alanine aminotransferase (ALT) level (p < 0.001), AST level (p < 0.001), γGTP level (p = 0.02), and lower platelet count (p < 0.001) were observed at the baseline of patients who developed HCC after treatment with DAAs. In particular, in a higher proportion of patients who developed HCC advanced fibrosis (40.4% vs. 88.2%; p < 0.001) and ≥6 substitutions in PKR-bd (15.4% vs. 41.2%; p = 0.01) was observed. In addition, the proportion of patients who developed HCC was higher among patients infected with genotype 1b (47%) than with genotype 1a (0%), genotype 2c (29.4%), genotype 3a (23.5%), or genotype 4d (0%). Nevertheless, according to genotype, a significantly higher proportion of virus-carrying ≥3 substitutions in ISDR and ≥6 in PKR-bd was observed in patients who developed HCC infected with genotype 2c (5/5, 100%) than with patients infected with genotype 1b (2/8, 25%) (p < 0.001). No virus carrying high substitution levels was observed among HCC patients infected with genotypes 1a, 3a, and 4d. Nevertheless, according to genotype, a significantly higher proportion of virus-carrying ≥3 substitutions in ISDR and ≥6 in PKR-bd was observed in patients who developed HCC infected with genotype 2c (5/5, 100%) than with patients infected with genotype 1b (2/8, 25%) (p < 0.001). No virus carrying high substitution levels was observed among HCC patients infected with genotypes 1a, 3a, and 4d. ## Sequence analysis of other viral regions Sequence analysis of the NS3 and NS5B regions was performed in all available samples to evaluate HCV variability in different genes. As expected, the proportion of amino acid changes detected in each region was variable, but no differences between samples from patients with low and high fibrosis levels were observed comparing NS3 and NS5B genes (data not shown). # Discussion HCV eradication by DAA therapy reduces the progression of liver disease among patients with SVR. However, in a small proportion of them, a progression on HCC was observed. The mechanisms and risk factors are still being investigated. The optimal strategy for monitoring the incidence of advanced fibrosis and HCC remains unknown. Thus, it is important to determine the risk factors for liver damage progression after SVR. The association between substitutions in ISDR and development of HCC in patients who achieved SVR by IFN-based therapy has been reported in patients infected with HCV genotype 1b, which suggests that substitutions in ISDR might be a useful predictor for HCC together with higher γ-GTP level and older age. Here, the genetic variations in HCV NS5A-ISDR and PKR-bd of five different subtypes were evaluated. No differences were observed comparing SVR and treatment-failing patients. However, higher γ-GTP levels, lower PLT counts and plasma HCV RNA levels, as well as presence of higher level of substitutions in ISDR and in PKR-bd of HCV were associated with advanced liver fibrosis stage among all baseline treatment patients. In addition, a greater proportion of advanced fibrosis stage was observed among patients infected with HCV carrying INS/DEL in PKR-bd, as well as lower plasma HCV RNA levels being observed among patients infected with HCV carrying INS/DEL in ISDR compared to INS/DEL found in the rest of PKR-bd. As has already been hypothesized for the duplication of the V3 domain in NS5A, Viruses 2020, 12, 255 8 of 10 a higher number of substitutions in PKR-bd might alter protein functionality. In fact, it is known that PKR can affect the pathogenesis of malignancies leading to cell growthand may induce HCC cell proliferation with HCV infection. Since the NS5A region suppresses PKR function, amino acid changes in ISDR and PKR-bd affect the binding and repression of PKR, thus determining a progression in the liver damage and hepatocarcinogenesis. NS5A of HCV genotype 1b carrying an ISDR with amino acid substitutions was associated with a reduced efficiency of infectious virus production in an in vitro human hepatocyte model determining a lower propagation ability. Since lower plasma HCV RNA levels were observed among patients infected with HCV carrying higher substitutions in PKR-bd and carrying INS/DEL in ISDR, our hypothesis is that the mutated PKR-bd virus trapped in hepatocytes might indirectly determine a role in hepatocyte damage that could trigger a process of degeneration in HCC. Thus, higher levels of substitutions in the ISDR and even more in PKR-bd might be a useful predictor for advanced liver fibrosis and possibly for progression in HCC. Among all patients infected with HCV carrying ≥3 substitutions in ISDR and ≥6 in PKR-bd, advanced fibrosis was observed in a higher proportion of patients infected with genotype 2c followed by patients infected with genotype 1b and genotype 1a. Overall, a higher proportion of patients who developed HCC had advanced fibrosis and ≥6 substitutions in PKR-bd compared to patients who did not develop HCC. Among pegIFN/RBV-treated patients, the highest percentage of subjects developing HCC was infected with HCV genotype 1b, and still now, among patients treated with DAAs, the higher proportion is infected with genotype 1b. However, among patients who developed HCC, a significant higher proportion infected with HCV genotype 2c had (i) higher substitutions in ISDR and PKR-bd, compared to patients infected with other genotypes, (ii) greater variability associated with advanced fibrosis, and (iii) a higher number of patients who developed HCC carried high numbers of PKR-bd substitutions at the baseline samples. Thus, it would be possible that over time this will lead to a change on the incidence of a greater frequency of development in HCC among patients infected with this subtype. However, further investigations are needed to assess (i) how many other recently treated patients will still develop HCC in a longer follow-up, (ii) the correlation between ISDR or PKR-bd substitutions and development in HCC, (iii) the correlation between advanced fibrosis after SVR and HCC development, and (iv) which genotype will be the most prone to develop HCC. With this perspective, it would be necessary to consider in the patient's follow-up greater controls in assessing the state of the liver for longer periods and in particular considering, before starting DAA treatment, the presence of a greater number of ISDR/PKR-bd substitutions as a prognostic marker for advanced fibrosis and HCC.

Percutaneous osteoplasty for painful bony lesions: a technical survey Percutaneous osteoplasty (POP) is defined as the injection of bone cement into various painful bony lesions, refractory to conventional therapy, as an extended technique of percutaneous vertebroplasty (PVP). POP can be applied to benign osteochondral lesions and malignant metastatic lesions throughout the whole skeleton, whereas PVP is restricted to the vertebral body. Common spinal metastases occur in the thoracic (70%), lumbosacral (20%), and cervical (10%) vertebrae, in order of frequency. Extraspinal metastases into the ribs, scapulae, sternum, and humeral head commonly originate from lung and breast cancers; extraspinal metastases into the pelvis and femoral head come from prostate, urinary bladder, colon, and uterine cervical cancers. Pain is aggravated in the dependent (or weight bearing) position, or during movement (or respiration). The tenderness and imaging diagnosis should match. The supposed mechanism of pain relief in POP is the augmentation of damaged bones, thermal and chemical ablation of the nociceptive nerves, and local inhibition of tumor invasion. Adjacent (facet) joint injections may be needed prior to POP (PVP). The length and thickness of the applied needle should be chosen according to the targeted bone. Bone cement is also selected by its osteoconduction, osteoinduction, and osteogenesis. Needle route should be chosen as a shortcut to reach the target bony lesions, without damage to the nerves and vessels. POP is a promising minimally invasive procedure for immediate pain relief. This review provides a technical survey for POPs in painful bony lesions. # Introduction The first percutaneous vertebroplasty (PVP) was performed in a case of hemangioma of the cervical vertebral body by Harve Deramond in France in 1984. Applications of PVP have increased from benign primary tumors and osteoporotic compression fractures to malignant osteolytic/osteoblastic/mixed painful vertebral lesions. Percutaneous osteoplasty (POP) has been developed as an extended technique of PVP for the treatment of painful bony disorders in the whole body. POPs include PVPs and extraspinal POPs. The human adult skeleton (206 bones) can be divided into axial (80) and appendicular (126) bones. The axial bones consist of the bones of the skull, ossicles, hyoid bone, ribs, sternum, and vertebrae. The appendicular bones are composed of the shoulder girdle, upper extremities, pelvic girdle, and lower extremities . Various bones can also be classified by their shapes: long, short, flat, sesamoid, and irregular bones. In POP, the classification by the shape of bones is important when inserting the needle and placing the bone cement, as well as in preserving the effectiveness of the cement augmentation against weight bearing in a dependent position. For example, POP in the humeral head is enough to improve its strength, however, POP in the femoral head needs additional intramedullary nailing in the femoral shaft to bear the weight of the body. Bone is the third most frequent site of metastasis, after the lung and liver. Prostate and breast cancers are responsible in the majority of bone metastases (up to 70%), followed by thyroid, lung, urinary bladder, and renal cancers, and finally melanoma. Median survival rate after diagnosis of bone metastases is known to be 4 years in thyroid cancer, 2 years in breast cancer, 1-4.5 years in prostate cancer, 1 year in renal cancer, 6-9 months in urinary bladder cancer, and 6 months in lung cancer and melanoma. Fortunately, the 3 most common cancers which develop bone metastases (prostate, breast, and thyroid cancers) show relatively high median survival rates. (2), scapula (2) Cranial bones: parietal, temporal, frontal (1), occipital (1), ethmoid (1), sphenoid (1) Facial bones: maxilla, mandible (1), zygomatic, nasal, palatine, inferior nasal concha, lacrimal, vomer (1) Auditory ossicles: malleus, incus, stapesUpper extremities (60): humerus, radius, ulna, carpals, metacarpals, phalangesHyoid (1) Pelvic girdle: coxal (hip) bones (2) Vertebral column: cervical spine (7), thoracic spine, lumbar spine (5), sacrum (1), coccyxLower extremities (60): femur, tibia, fibula, patella, tarsals, metatarsals, phalanges (28) Thoracic cage: sternum (1), ribsThe numbers in parentheses indicate the number of bones. ## 377 The primary tumors in the neck and thorax, such as those of the thyroid gland, breast, and lung, frequently metastasize to the cervicothoracic vertebrae, ribs, scapulae, sternum, and humeral head; the primary tumors of the abdomen and pelvis, such as those of the stomach, colon, prostate, urinary system, and uterine cervix, invade the lumbosacral vertebrae, pelvic bones (the ilium, ischium, and pubis), and femoral head. This technical survey introduces various POPs in axial/ appendicular or malignant/benign painful bony lesions. ## Main body ## General concepts ## 1) mechanisms of pain relief in pop, including pvp The exact mechanisms of pain relief in PVP are not fully understood. POP creates 1) mechanical stabilization of microfractures, resulting in a reduction of irritation to free nerve endings and restoring strength and stiffness in the vertebral body, 2) thermal ablation, as high as 70°C, to the free nerve endings, 3) decompression of intraosseous pressure following compression fractures in the vertebral body, and 4) an antitumor effect due to the local toxicity of polymethylmethacrylate (PMMA). The mechanisms of pain relief in POP for various shapes of bones do not seem different from those of PVP. ## 2) basic principle for needle insertion and selecting needle gauge sizes The most important concern for inserting the needle in POP is choosing the shortcut from the skin to targeted bone lesions, in order to reduce procedural time and effort while avoiding damage to the adjacent or traversing nerves and vessels of the targeted bone. Vulnerable vessels and nerves of the targeted bones will be discussed in each POP. The best method for reducing the risk of piercing important nerves and vessels is not only having knowledge about anatomy but also reducing the speed of needle insertion near these structures in order to give them enough time to escape from the approaching needle. According to the thickness of the targeted bone or the approaching route of the needle, selection of the needle gauge size and length should be different. An 11 gauge, 10 cm long needle is chosen for the vertebral bodies from the typical cervical to the sacral vertebrae, humeral head, scapula, femoral head, greater or lesser trochanter of the femur, as well as the pelvic bones, including the ilium, is-chium, and pubic bones. A 13 gauge, 10 cm long needle is selected for the lateral masses of the atlas and the body of the axis. A 13 gauge, 5 cm long needle is recommended for the ribs and sternal body. ## 3) bone cement It should be considered whether the inserted cement helps improve strength and stability in a certain bone, such as long, short, flat, or irregular bones. It is better to discuss this in each POP case, according to the shape of the bones. The result for insertion of the bone cement into the head or shaft (with a medullary cavity) in long bones (humerus and femur) or flat bones (ribs, sternum, scapulae, or ilium) may be quite different in regards to improving strength and stability. The flat bones are composed of 2 outer layers of compact (trabecular) bone and intervening cancellous (spongy) bone. The long bones consist of proximal and distal epiphyses connected to the diaphysis via the metaphysis. Commonly available injectable bone cements (IBCs) include (1) PMMA or acrylic bone cements (ABCs), (2) calcium phosphate cements (CPCs), calcium sulfate cements (CSCs), or hydroxyapatite cements (HACs: tetra-calcium phosphate and di-calcium phosphate anhydrous powders), and (3) glass polyalkenoate (ionomer) cements (GPCs or GICs, alumino-silicate glass). Bone cements are selected according to their osteoconduction, osteoinduction, and osteogenesis. CPCs or HACs, instead of PMMA, have been used for the treatment of benign osteochondral lesions (OCLs) of the talus or tibia for osteoconduction, osteoinduction, and osteogenesis. ## 4) necessity of adjacent (facet) joint injections prior to pop (pvp) Facet joint injections or medial branch blocks are required when the vertebral body (anterior component of the spine) cannot bear the weight due to benign osteoporotic compression fracture or malignant metastatic osteolytic compression fracture, resulting in overloaded facet joints (posterior components of the spine). Unilateral or bilateral facet joint injections are needed prior to the PVP because: (1) they help physicians confirm the exact currently painful vertebrae among mixed old and new compressed vertebrae, (2) patients can cooperate in maintaining a prone position during the PVP under local anesthesia or monitored anesthetic care, and (3) it can reduce patients' dissatisfaction from incomplete pain relief from PVP while facet joint pain remains. Glenohumeral joint injection is commonly required in a painful humeral head metastasis, resulting in a long-term immobilization of the glenohumeral or acromioclavicular joint due to the painful bony lesion. Sacroiliac joint injections are also needed in painful sacral or iliac metastatic fractures due to the joint being overloaded, which produces Honda sign in bone scans. ## 5) deciding whether to perform pop according to life expectancy of patients with bone metastases and evaluation for effectiveness of pop The Mirels score is helpful in assessing the risk of pathologic fracture due to lytic lesions of the long bones, based on these variables with 3 grades, including the site of the pathologic fracture, its size, pain intensity, and the presence of a lytic lesion. If the total score is greater than 9, the need for prophylactic fixation, due to a high risk of fracture, should be considered. The revised Tokuhashi score is useful for assessing the life expectancy of patients with bone metastases. The score includes 6 variables, including general condition based on performance status, the number of extraspinal bone metastases foci, number of metastases in the vertebral body, metastases to the major internal organs, primary site of the cancer, and palsy, according to the Frankel classification. Life expectancy for proper treatment is supposed to be less than 6 months for conservative treatment, 6-12 months for palliative surgery, or 12 months or more for excisional surgery, these periods being determined by a total score of less than 8, 9-11, or 12-15, respectively. Even though POPs are considered palliative surgery, they can be performed on any patient who can tolerate a bone biopsy. The Karnofsky performance scale (KPS) is useful for comparing the effectiveness of therapies, including POPs, and for assessing the prognosis of patients. The lower the KPS score, the worse the chance of survival. The Eastern Cooperative Oncology Group performance status (ECOG PS) is also commonly used for evaluation for physical performance status. The higher the ECOG PS score, the worse the chance of survival. These evaluation systems use a markedly different score for the same performance status. These PS evaluation systems can be used for evaluating the effectiveness of POP before and after the procedure. ## 6) diagnostic procedures for painful metastatic bony lesions It is common to find painful metastatic bony lesions after the diagnosis of primary malignant lesions. However, it is not rare to find painful metastatic bony lesions before the diagnosis of the primary malignant lesions or lesions of an unknown origin. Most patients complain of aggravation from incidental pain in a certain position or during positional changes. Tenderness on the bony lesions can suggest the presence of fracture. For a confirmation of the presence of metastatic bony lesions through history and physical examination, imaging diagnosis is required, such as X-ray, computed tomography or positron emission tomography-computed tomography/magnetic resonance imaging, and bone scans. ## Percutaneous vertebroplasty (spinal pop) Metastases to the axial skeleton, rather than the appendicular skeleton, are common due to its hematopoietically active red bone marrow, paravertebral network, and micro-environment support, based on the seed-and-soil hypothesis. ## 1) cervical vertebrae (1) atypical cervical vertebrae ① c1 vertebra The atlas has no vertebral body. Instead of a vertebral body, the lateral masses, which are the bulkiest and hardest parts of the atlas, support the weight of the head. Each mass has the superior and inferior articular processes. The superior articular process has a concave articular surface corresponding with the occipital condyle, making the atlanto-occipital (AO) joint, which allows the nodding (yes) movement of the head. The inferior articular process has a flat or convex articular surface corresponding with the superior articular process of the axis, making the atlantoaxial (AA) joint, which permits rotational (no) movement of the head. The lateral masses, the major weight bearing units of the atlas, become the targeted metastatic bony structures for osteoplasty. Candidates for POP at the lateral mass of the atlas are not able to bear the weight of the head due to the osteolytic metastatic lesion of the lateral mass. They usually complain of intractable pain in the suboccipital area with a limited range of neck motion. They can only sleep on the side opposite the lesion with a pillow. On physical examination, there is tenderness in the one side of the suboccipital area and severe limitation of ipsilateral rotation of the neck. A destructive lesion should be found in the lateral mass of the atlas from the computed tomography/magnetic resonance imaging and bone scans. For performing POP at the lateral mass, the patients are ## Kim and kim placed in the supine position with a pillow under their shoulders to create neck extension. After sterile draping, local anesthetic with 1% lidocaine is given at the point of the anticipated needle entry. A 13 gauge vertebral needle is introduced after determining the location of the carotid artery under ultrasound guidance. While advancing the needle, great care should be given to avoid piercing the vertebral artery and the spinal nerves by observing the depth of penetration using lateral fluoroscopy. After confirming the needle position in the targeted lateral mass, contrast medium is injected to observe any possible leakage of bone cement into the vessel or epidural space. The amount of bone cement for this procedure is less than 1 mL. Immediately after POP at the lateral mass of the axis, suboccipital headache disappears and the limitation of the neck motion is decreased. The anterolateral approach for POP, under the guidance of both ultrasound and fluoroscope, is safe under local anesthesia. The transoral approach has also been introduced; however, it can be performed under general anesthesia with nasotracheal intubation. AO and/or AA joint injections may be needed before/ after POPs in the C1 and C2 vertebrae, as with other facet or glenohumeral joint injections in spinal or extraspinal POPs. ## ② c2 vertebra The axis is composed of the body, dens, transverse processes, pedicles, pars interarticularis, and superior/inferior processes. The anteriorly located dens (odontoid process) is a characteristic structure which is connected to the body. The superior articular facets and dens of the axis (C2) articulate with the inferior articular facets and the facet for dens of the atlas (C1), respectively. The inferior articular facets of the axis articulate with the superior articular facets of the C3 vertebra. All transverse processes in the first 6 cervical vertebrae have a pair of distinct transverse foramina, each of which contains a vertebral artery and vein. The 7th vertebra also has a pair of transverse foramina, each of which contains only the vertebral vein, not the artery. Patients with a C2 body and/or dens fracture complain of suboccipital headache and cannot sleep lying down on their back or dependent side, and aggravate their pain in an upright position. Therefore, patients can usually be found in a sitting position supporting their head with neck flexed. Plain X-ray, computed tomography/magnetic resonance imaging, and bone scans should match with the above symptoms and signs. Patients are placed in the supine position with neck slightly extended. The entry point of the puncture for the anterolateral approach is the edge of the mandibular angle. Ultrasound is also helpful to avoid the major vessels/nerves and tracheoesophageal complex (TEC), as with performing POP at C1. Along with a 10 cm, 23 gauge spinal needle for local anesthesia from the skin to the vertebral body, a 10 cm, 13 gauge vertebral needle is advanced and anchored to the targeted vertebral body under alternating anteroposterior and lateral fluoroscopic guidance. After removal of the stylet of the needle, contrast medium is injected to confirm any leakage into the vessels or epidural space. Bone cement, PMMA, with barium powder for increasing opacity, less than 1 mL, is injected under continuous fluoroscopic control. Postoperative computed tomography should show proper filling with bone cement in the destructive lesion of the body and/or dens. Patients can get immediate pain relief and recovery of the range of motion of the neck. AA and AO joint injections may be needed, before or after PVP at C2. ## (2) typical cervical vertebrae (c3-c7) Characteristic preoperative symptoms and signs are weight-bearing pain and limitation of the range of motion in the neck. The preoperative imaging diagnostic tools, including plain X-ray, bone scans, and magnetic resonance imaging/computed tomography, should be correlated with the painful metastatic lesion. Radiation therapy is somewhat helpful in relieving pain, however, the role of augmentation by PVP may be necessary in most cases. The key principle using an anterolateral approach to the vertebral body, from the C3 to C7, is to avoid the TEC, internal carotid artery, and internal jugular vein. The method for pushing the TEC away from the approaching needle is the same as in performing percutaneous endoscopic cervical discectomy. Right-handed operators should stand to the left side of the patient. After the left middle and index fingers split the TEC medially and carotid artery laterally, infiltration of local anesthetic, using a 10 cm, 25 gauge spinal needle into the skin, subcutaneous tissue, and anteromedial vertebral body, is performed. While leaving the spinal needle in place, and placing the fingers with the middle finger against the TEC and the index finger against the carotid artery, a 10 cm, 11 gauge vertebral needle is inserted alongside the spinal needle. After anchoring the vertebral needle to the left anteromedial vertebral body by hammering against the slippery anterior longitudinal ligament, the position of the needle can be directed to the targeted bony lesion. The anterior segment of the cervical spine, the vertebral body, is a weight-bearing structure supporting the head. If the targeted bony lesion affects the anterior vertebral body, the needle can be placed into the anterior 1/3 to 1/2 of the vertebral body. Even if there is a disruption of the posterior wall of the vertebral body, it is better to place the needle in the anterior 1/3 of the vertebral body. The next step is to check for leakage of the contrast medium into the blood vessels or anterior epidural space. If the contrast medium is contained inside the vertebral body for a moment before leaking into the blood vessels or anterior epidural space, rather than using the usual toothpaste consistency, thicker bone cement is required. It is better, in cases using a reduced volume of injected bone cement, to rely on pain relief from thermal and chemical ablation of the nerves rather than augmentation by the injected volume. Normally, the injected cement volume is less than 2 mL. ## A b c ## Kim and kim Facet joint injections may be also needed before and after PVP for the classic cervical vertebrae from C3 to C7. ## 2) thoracic and lumbar vertebrae (1) transpedicular versus extrapedicular approach Depending on the route of the vertebral needle passing the pedicle, the method of approach to the vertebral body can be divided into a transpedicular and extrapedicular approach. The transpedicular approach is a conventional, safe method, even though it is difficult for the needle to penetrate the hard consistency of the pedicle. It has a risk for spinal cord or dorsal root ganglion injury. The transpedicular approach can be divided into unipedicular and bipedicular methods. The unipedicular approach can be performed in the thoracolumbar vertebrae. It can reduce time, risk, and effort in performing PVP . ## ① unipedicular approach Patients are placed in a prone position. Monitored anesthetic care, using 30 mg of ketorolac and 50 μg of fentanyl, is performed. Facet joint injections are performed prior to PVP, if necessary. The targeted vertebral body, the midline passing through the spinous processes, the pedicle with its quadrants, especially the outer, upper paadrant, aswell as lines marking 1, 2, and 3 pedicular distances parallel to the midline should be drawn in the anteroposterior fluoroscopic view. A Kirschner (K) wire is placed as it passes from the upper pedicle to the anterior 1/3 to 1/4 of the targeted body in the lateral fluoroscopic view. Local anesthesia is performed using a 10 cm, 23 gauge spinal needle. The needle is inserted from 2-3 pedicular distances into the outer, upper pedicle in the anteroposterior fluoroscopic view after local infiltration of 1% lidocaine into the skin. A 10 cm, 11 gauge vertebral needle is inserted along with the spinal needle after skin incision using a No. 15 blade scalpel along with kin crease. A vertebral needle is anchored to the outer, upper quadrant of the pedicle by hammering. While passing the pedicle, hammering is not recommended to avoid spinal cord or nerve root injury, and aiming of the vertebral needle upward and outward is also helpful to reduce the risk of nerve injury. After penetrating the posterior wall of the vertebral body, rotation of the bevel of the vertebral needle toward the medial, lateral, superior, or inferior direction is helpful to reach to targeted anterior 1/3 or 1/4 of the vertebral body at in the midline. When the vertebral needle is placed in the vertebral body, hammering is allowed while the depth of the needle is fixed with curved Kelly forceps to prevent passing the targeted anterior 1/3 to 1/4 of the vertebral body. A leakage test into the epidural space or blood vessel, using contrast medium, is essential. If found, a slice of gelatin sponge is helpful and the test should be performed again. The injected bone cement volume is less than 3 or 5 mL in the thoracic or lumbar vertebrae, respectively. ## ② extrapedicular approach The extrapedicular approach is selected when a pedicle , and lumbar [L1-L4] arteries) may be increased. The posterior intercostal arteries arise directly from the thoracic aorta (T3-T11) or subclavian artery (T1-T2). The posterior intercostal artery has a dorsal and a collateral branch. The dorsal branch is divided into the spinal branch (to the spinal cord) and medial and lateral cutaneous branch (to the posterior thoracic wall). The collateral branch supplies blood to the lateral thoracic wall. The 1st to 6th anterior intercostal arteries arise directly from the internal thoracic artery, and the 7th to 9th anterior intercostal arteries indirectly from the musculophrenic artery (a branch of the internal thoracic artery), and connects to the posterior intercostal artery. The anterior intercostal artery is not present between the 10th and 11th intercostal spaces. Therefore, the posterior intercostal arteries supply each space solely. The 12th subcostal artery, below the 12th rib, arises directly from the thoracic aorta. The 4 pairs of lumbar arteries are located parallel to the intercostal arteries and subcostal artery. The segmental arteries usually run along thoracolumbar vertebral bodies at the infra-pedicular level. Therefore, it is recommended to anchor the vertebral needle at the pedicular level and the midline of the lateral vertebral body with a reduced inserting speed. As with the transpedicular approach, the vertebral needle should target the anterior 1/3 to 1/4 of the vertebral body. ## (2) benign versus malignant painful thoracolumbar lesions It is common to predict metastatic vertebral fractures in patients with known malignancies. However, intractable back pain in a weight bearing position may be the first symptom to suggest metastatic lesions. Attention should be given in patients who are young, in a low risk group for osteoporosis, and men. Palpation after adjacent facet joint injections among multiple metastatic vertebral fractures makes the current painful fracture levels clear. After removal of facet joint pain, it is easy to recognize precise supraspinous tenderness among multiple fractures. In addition, the effects of facet joint injections help patients to lie on their abdomen with less pain and increases cooperation during the PVP procedure under local anesthesia. Spinal metastases, in order of frequency, are to the thoracic (70%), lumbosacral (20%), and cervical (10%) vertebrae. Multiple spinal metastases range from 17% to 30%. The first metastatic involvement within the vertebra is the posterior portion of the vertebral body before the pedicles. Absence of the pedicle due to metastatic destruction, in plain X-ray film, is a characteristic feature comparable to an osteoporotic compression fracture . . Imaginary lines for the unipedicular approach in percutaneous vertebroplasties (PVPs) of the lumbar and thoracic vertebrae. If a vertebral needle passes through the structure of the pedicle, the approach method should be called "a transpedicular approach" for PVP. If not, it should be called an "extrapedicular approach". The vertebral needle should not penetrate the inner wall of the pedicle, so as not to pierce the spinal cord or dorsal root ganglion. The target point of the vertebral needle is the anterior 1/3 to 1/4 of the vertebral body. (A) In the lumbar vertebrae, the shortest distance from the midline is 2 pedicular distances, and the longest distance from the midline is 3 pedicular distances. (B) On the contrary, in the thoracic vertebrae, the shortest distance from the midline is 2 pedicular distances and the longest distance from the midline is 2.5 pedicular distances. Osteoporotic vertebral compression fractures are also frequent in the thoracolumbar vertebrae. ## 3) sacrum The sacrum, 5 fused sacral vertebrae, is composed of the vertebral body (from the promontory to the apex of the sacrum) and alae (wings). Honda sign is a characteristic Hshaped (wider upper and narrow lower H letter) increase in radioactive uptake in the sacral body and both alae, parallel to the sacroiliac joints, observed on bone scans, indicating a bilateral sacral insufficiency fracture. ## (1) sacral body Sacroplasty in the sacral body needs bipedicular approaches because it is dangerous for the vertebral needle to penetrate to the midline of the sacrum . An epidurogram, usually through the S1 foramen, is helpful in avoiding neural passage prior to vertebral needle insertion into the sacral body. The amount of injected bone cement is less than 5 mL for each approach. ## (2) sacral wings (alae) Sacroplasty, in each sacral wing, from the S1 to S5, can be performed by insertion of the vertebral needle once from the S1 to S5 alae, and the cement can be inserted while withdrawing the needle from the S5 to S1 alae . The sacroplasty in the ala is quite similar to performing POP in the ilium, in that caution should be given not to pierce the cauda equina or spinal nerves, or inject bone cement into the sacroiliac joints. The amount of injected bone cement is less than 10 mL for each sacral sacral wing. Sacroiliac joint injections with contrast medium or a steroid may be needed in performing sacroplasty of the sacral wings for both preventing leakage into the sacroiliac joint and relieving joint pain, respectively. ## Extraspinal pop in malignant metastases There are 33 irregular bones in the human body: the spine The vertebral needle is inserted into the left sacral wing from the S1 to the S5, and then bone cement is inserted from the S5 to the S1 while withdrawing the needle in the anteroposterior and lateral fluoroscopic views, respectively. ## A b ## Percutaneous osteoplasty Korean J Pain 2021;34(4):375-393 www.epain.org 385 (from the cervical vertebrae to the coccyx) and the skull (temporal, sphenoid, ethmoid, zygomatic, maxilla, mandible, palatine, inferior nasal concha, and hyoid bone). Flat bones are commonly found in the human body, such as the skull (occipital, parietal, frontal, nasal, lacrimal, and vomer bone), the thoracic cage (sternum and ribs), and pelvis (ilium, ischium, and pubis), with the function of protecting their internal organs. Long bones include the humerus, radius, ulna, metacarpals, and phalanges in the upper extremities, and the femur, tibia, metatarsals, and phalanges in the lower extremities. ## 1) bones in the thoracic cavity, scapula, and humerus The origin of the cancers metastasized into the bones in the thoracic cavity and humerus are, commonly, the breast, thyroid glands, and lungs. ## (1) costoplasty The ribs are the second most common metastatic bones from lung cancers, after the spine. Most metastatic bone cases are found in both the thoracic spine and ribs. The ribs can be divided into true ribs (R1-R7, anteriorly articulated with the sternum by the costal cartilages), false ribs (R8-R10, anteriorly articulated with the rib above), and floating ribs (R11 and R12, without articulation anteriorly). A rib is composed of the head, neck, and body. The head and neck of the first rib are attached to their own superior costal facet of the vertebral body and costal facet of the transverse process. From the 2nd to 10th ribs, the head of the rib is attached to both the rib above and their own vertebral bodies. The 11th and 12th ribs have their costal facets on the vertebral body alone, without the costal facets of the transverse process. Patients with rib metastasis complain of pain, depending on their position. Pain is aggravated in the supine, lateral, or prone position when the rib metastasis involves the posterior, lateral, or anterior body of the rib. In addition to movement-related weight-bearing pain, incidental breakthrough pain can develop in coughing and sneezing. Tenderness is apparent on multiple ribs or 2 or 3 points on the same ipsilateral rib. It is quite different from anterior or lateral chest pain from referred pain, originating from thoracic facet joint syndrome, which can be recognized by tenderness on the supra-facetal area. Corresponding lesion imaging, plain X-ray film, computed tomography/positron emission tomography-computed tomography, and bone scans, should match with palpation of the tenderness. A 5 cm, 13 gauge, short-bevel vertebral needle is suitable for costoplasty, in order not to pierce the pleura. The flat bone is composed of spongy bone between 2 compact bones. A perpendicular approach against bone osteolytic lesions is more convenient than approaching from above into the narrow upper margin or from below into the complicated complex of the intercostal costal nerve, artery, ## Kim and kim and vein. Anchoring to the slippery outer compact bone of the targeted rib requires hammering, while preventing penetration of the vertebral needle through the inner compact bone, using Kelly forceps. After the arrival of the vertebral needle inside the spongy bone, the stylet of the vertebral needle is removed and the cannula, without the stylet, should be fixed. Contrast medium should be injected for checking leakage into the vessels or pleura. Bone cement should be injected under lateral fluoroscopic monitoring. The amount of injected cement volume per each costoplasty is less than 2 mL. ## (2) sternoplasty The sternum is a blade-like flat bone, consisting of the manubrium, body and xiphoid process. The 1st costal notches of the manubrium articulate with the pair of the 1st costal cartilages. The 2nd costal notches at the sternal angle, between the manubrium and body, articulates with a pair of the 2nd costal cartilages. The body articulates with a pair of the 3rd to 7th costal cartilages at the 3rd to 7th costal notches. The true ribs, from the 1st to 7th, articulate with the sternum. Patients with sternal metastases complain of continuous pain or incidental pain during respiration. Tenderness should be proven to be osteolytic lesions in computed tomography or active lesions in bone scans. A 13 gauge, 5 cm vertebral needle is inserted perpendicular to the body of the sternum. The amount of injected cement volume per each sternoplasty is less than 2 mL. ## (3) scapuloplasty The scapula, or shoulder blade, is a triangular flat bone of the thoracic cage. It has 3 borders (medial [vertebral], lateral [axillary], and superior borders), and superior, inferior, and glenoid angles (or glenoid cavity). It also has costal and dorsal surfaces. When inserting the vertebral needle into the scapula, it is better to choose the medial and lateral borders, because the superior border contains the suprascapular notch, part of the route of the suprascapular nerve and artery. The suprascapular nerve originates from the ventral rami of the C5 and C6 spinal nerves. The suprascapular nerve innervates the supraspinatus and infraspinatus muscles, and receives sensory innervation from the acromioclavicular and glenohumeral joints. Patients with scapula metastases complain of pain when they lie on their back, with/without persistent bone pain. Therefore, they usually sleep on their side. They sometimes have limited shoulder motion. Tenderness on the involved scapula is apparent. Imaging, including bone scans and positron emission tomography-computed tomography/computed tomography, should match the tenderness. After drawing the 3 borders and scapular spine, tender points over the osteolytic lesions are also marked on the skin of the dorsal scapular surface. More than one vertebral needle may be needed if multiple osteolytic lesions exist. The superior border, especially the scapular notch, should be not be chosen for the entry of the vertebral needle, due to the passage of the suprascapular nerve and artery. An 11 gauge, 10 cm vertebral needle is required. After advancing the needle to the targeted lesion, a leakage test should be performed. Bone cement insertion is performed while withdrawing the needle. The injected volume of the cement is variable according to osteolytic lesions. Pain relief can be immediate after POP in the scapula, when lying on the back or in a dependent position. ## (4) humeroplasty The humerus is a long bone. It consists of 2 epiphyses (proximal end [head and anatomical/surgical neck] and distal end) and 1 diaphysis (shaft). POP, using bone cement, is effective for a metastasis to the proximal end rather than the shaft, which has an intramedullary canal. The greater tubercle is the common insertion site of the supraspinatus, infraspinatus, and teres minor muscles of the rotator cuff. The lesser tubercle is the insertion site of the subscapular muscle of the rotator cuff. The bicipital groove is the passage of the long head of the biceps brachii. The deltoid and pectoral branches of the thoracoacromial artery and anterior and posterior circumflex humeral arteries supply blood to the humeral head and neck. Therefore, the greater tubercle, lesser tubercle, bicipital groove, as well as the thoracoacromial and circumflex humeral arteries should be avoided during entry of the vertebral needle. Patients with humeral head metastases may have pain and limited shoulder motion. Tenderness on the humeral head should match imaging, including computed tomography/magnetic resonance imaging and bone scans. Vertebral needle insertion, a leakage test, and bone cement insertion are performed under fluoroscopy, and proper cement insertion is confirmed by postoperative computed tomography. The amount of injected bone cement is less than 10 mL. ## 2) pelvic bones and femur The pelvic girdle is composed of the 2 coxal bones and the sacrum. The 2 coxal bones are connected with each other at the cartilaginous symphysis pubis, and to the sacrum via the sacroiliac joints. The orientation of the ilium, ischium, and pubis of the coxal bones is quite easy to understand from the tri-radiate cartilage of a child's acetabulum: (1) the ilium: superiorly, (2) the ischium: inferiorlyposteriorly, and (3) the pubis: inferiorly-anteriorly. Metastases to the pelvic bones and femoral head commonly originate from prostate cancer, renal cancer, colorectal cancer, as well as breast and thyroid cancer. ## Kim and kim ## (1) ilioplasty The iliac crest between the anterior superior iliac spine and posterior superior iliac spine is the most common site for bone marrow biopsy. The entry point of the vertebral needle for POP of the ilium is the iliac crest, just as with bone marrow biopsy. The needle should be advanced along with the curvature of the ilium from the iliac crest to the superior acetabulum, if necessary. Caution should be given not to insert bone cement into the sacroiliac joint. The injected cement volume per each ilioplasty is less than 10 mL. ## (2) ischioplasty The ischium consists of the body, superior ramus, and inferior ramus. The entry point of the vertebral needle for ischioplasty is the ischial tuberosity, in a prone position. If necessary, the needle should be advanced into the inferoposterior acetabulum. The injected cement volume per each ischioplasty is less than 5 mL. ## (3) puboplasty The pubic bone is located inferiorly and anteriorly among the pelvic bones. It consists of a body, superior ramus, and inferior ramus. The left and right bodies of the pubic bones join together at the symphysis pubis. An 11 gauge, 5 cm vertebral needle is inserted perpendicular to the pubic tubercle. The injected cement volume per each puboplasty is less than 5 mL. ## (4) femoroplasty Unlikely POP at the humeral head, POP at the femoral head alone without intramedullary fixation of the femoral shaft is unstable, because the femur and acetabulum of the pelvic bones are weight-bearing units. In case of greater or lesser trochanter metastases, after intramedullary fixation of the femoral shaft, POP at the greater or lesser trochanter may be needed. The injected cement volume per each femoral head osteoplasty is less than 10 mL. ## Extraspinal pop in ocls (osteochondritis dissecans) An OCL is an injury or small fracture of the cartilage surface and subchondral bone, commonly in the talus or tibia ## A b c Kim and Kim. The talus articulates with the tibia and fibula superiorly, forming a mortise, with the navicular anteriorly, and calcaneus inferiorly. Instead of using PMMA cement in metastatic bone lesions, calcium phosphate or HAC is used in order to reduce thermal ablation and increase osteoconduction. Monitoring under arthroscopy and fluoroscopy decreases the risk of cement leakage into the ankle joint. The injected cement volume for OCL of talus or tibia is less than 2 or 5 mL, respectively. # Conclusions For the decision regarding and evaluation of POPs, the preoperative revised Tokuhashi scoring system for the evaluation of patient life expectancy with bone metastases, preoperative and postoperative comparison of KPS and ECOG PS for comparing the effectiveness of therapies and accessing the prognosis of patients, and the preoperative Mirels score, for investigating the risk of pathologic fracture, are also required. It should be emphasized that adjacent facet or shoulder/ sacroiliac joint/hip joint injections are essential prior to POPs. They help to determine whether to perform POPs after removal of adjacent joint pain, to permit patients be more cooperative in assuming specific operating positions, and in determining the precise current levels of pain among mixed old/new and painful/non-painful multiple vertebral fractures. For reducing the risk of leakage into the blood vessels or epidural space, augmentation by injecting cement volume or recovery of original bone shape after POP is less important because other mechanisms for pain relief by POP, such as chemical or thermal ablation of the nociceptive nerves, exist. In conclusion, even though POP is a simple and fast pain relief method for painful bony lesions, basic understanding of anatomy, related to the shape of the various bones and adjacent blood vessels and nerves, is required. Decreased analgesic requirements and increased physical performance will be shown immediately after removal of deep somatic pain by POP. POPs are helpful for both painful malignant osteolytic/osteoblastic metastases and benign osteoporotic vertebral compression fracture/OCLs.

Undetectable ultrasensitive PSA after radical prostatectomy for prostate cancer predicts relapse-free survival Radical retropubic prostatectomy (RRP) is a well-established treatment option for localized prostate cancer in patients less than 70 years of age. Unfortunately, in a proportion of these patients postoperative prostate-specific antigen (PSA) does not reach undetectable levels and biochemical relapse occurs(Pound et al, 1997;Graefen et al, 1999). Patients with organ-confined disease Summary Radical retropubic prostatectomy is considered by many centres to be the treatment of choice for men aged less than 70 years with localized prostate cancer. A rise in serum prostate-specific antigen after radical prostatectomy occurs in 10-40% of cases. This study evaluates the usefulness of novel ultrasensitive PSA assays in the early detection of biochemical relapse. 200 patients of mean age 61.2 years underwent radical retropubic prostatectomy. Levels ≤ 0.01 ng ml-1 were considered undetectable. Mean pre-operative prostatespecific antigen was 13.3 ng ml-1. Biochemical relapse was defined as 3 consecutive rises. The 2-year biochemical disease-free survival for the 134 patients with evaluable prostate-specific antigen nadir data was 61.1% (95% CI: 51.6-70.6%). Only 2 patients with an undetectable prostate-specific antigen after radical retropubic prostatectomy biochemically relapsed (3%), compared to 47 relapses out of 61 patients (75%) who did not reach this level. Cox multivariate analysis confirms prostate-specific antigen nadir ≤ 0.01 ng ml-1 to be a superb independent variable predicting a favourable biochemical disease-free survival (P < 0.0001). Early diagnosis of biochemical relapse is feasible with sensitive prostate-specific antigen assays. These assays more accurately measure the prostate-specific antigen nadir, which is an excellent predictor of biochemical disease-free survival. Thus, sensitive prostate-specific antigen assays offer accurate prognostic information and expedite decision-making regarding the use of salvage prostate-bed radiotherapy or hormone therapy. are more likely to have prolonged biochemical disease-free survival (BDFS). Consequently, presence of extracapsular disease (pT3) or positive margins have been used not only to predict outcome, but also to decide on adjuvant therapy [bib_ref] Serum PSA after anatomic radical prostatectomy, Partin [/bib_ref] [bib_ref] The combination of pre-operative prostate-specific antigen and post-operative pathological findings to predict..., D&apos;amico [/bib_ref]. However, this criterion is not reliable in predicting biochemical relapse, and large numbers of patients (up to 70%) with positive margins do not progress [bib_ref] The combination of pre-operative prostate-specific antigen and post-operative pathological findings to predict..., D&apos;amico [/bib_ref]. Moreover, patients with negative margins are not guaranteed to be free from progression, approximately 10% progress in some series [bib_ref] 5-year tumour recurrence rates after anatomical radical retropubic prostatectomy for prostate cancer, Catalona [/bib_ref] [bib_ref] The combination of pre-operative prostate-specific antigen and post-operative pathological findings to predict..., D&apos;amico [/bib_ref]. In an attempt to improve outcome from RRP, some have attempted to select patients with a good prognosis. Probability nomograms combining PSA levels, clinical stage and Gleason score (so-called have been advocated to predict pathological stage, and by inference, risk of progression . Although use of these tables has had the effect of improving BDFS figures, they are not accurate enough to confidently predict clinical or biochemical outcome for an individual patient. Consequently, some patients with unfavourable risk factors may be denied potentially curative treatment. Pre-operative criteria other than Partin's tables and postoperative criteria other than surgical margins are needed to reliably predict biochemical or clinical outcome. PSA nadir measured with a sensitive assay offers the possibility of confidently predicting the removal of all significant PSA producing tissue within a few weeks of the operation. In contrast, histological evidence of relapse is usually not attained with more than 60% accuracy until the PSA rises above 2 ng ml -1 [bib_ref] Local recurrence after radical prostatectomy: characteristics in size, location, and relationship to..., Connolly [/bib_ref] which may not occur for many years after RRP. Thus, the earliest indicator of prostate cancer relapse after radical prostatectomy is a rising PSA. Most contemporary published series use a PSA level > 0.2 ng ml -1 in defining undetectable levels of PSA [bib_ref] Natural history of progression after PSA elevation following radical prostatectomy, Pound [/bib_ref] , but levels ranging from 0.1-4 ng ml -1 have also been used [bib_ref] Is prostate-specific antigen of clinical importance in evaluating outcome after radical prostatectomy?, Frazier [/bib_ref] [bib_ref] Hazard rates for progression determined by PSA, after radical prostatectomy for T1-T2..., Dillioglugil [/bib_ref] [bib_ref] A review of radical prostatectomy from three centres in the UK: clinical..., Feneley [/bib_ref] [bib_ref] Clinical and pathological significance of the level and extent of capsular invasion..., Scardino [/bib_ref]. The advent and availability of a sensitive PSA assay [bib_ref] Ultrasensitive time-resolved immunofluorometric assay of prostate-specific antigen in serum and preliminary clinical..., Yu [/bib_ref] raises questions regarding the appropriateness of these unvalidated levels. Sensitive PSA assays have been reported as being clinically useful but their role in early detection and prediction of biochemical relapses needs further assessment. Moreover, the benefits of early detection of relapse are still unknown; early detection of prostate cancer relapse might offer therapeutic advantages. Indeed, a recent study suggests that salvage external beam radiotherapy (EBRT) to the prostate bed may be more effective if given at low PSA levels [bib_ref] Limited suppression of prostate-specific antigen after salvage radiotherapy for its isolated elevation..., Egawa [/bib_ref]. A reasonable hypothesis is that postoperative PSA nadir levels measured by a sensitive PSA assay may provide useful prognostic information and expedite decision-making regarding the use of early salvage therapy. Thus, the objective of this study was to evaluate the role of sensitive PSA assays following radical prostatectomy. # Patients and methods A consecutive series of 200 patients of mean age 61.2 years (range 43-73) were selected for RRP and simultaneous pelvic lymph node dissection according to the following criteria. Patients had to have clinically organ-confined disease with histological confirmation of prostate cancer by transrectal ultrasound-guided needle biopsy or transurethral resection of prostate (TURP). Isotope bone scans were performed if the PSA was over 20 ng ml -1 or the patient had musculoskeletal pain. All patients had a presumed life expectancy in excess of 10 years, and no significant comorbidity. Patients over the age of 70 years were only considered for RRP in exceptional circumstances. Patients were not refused surgery because of high biopsy Gleason scores or high pre-operative serum PSA, providing they fulfilled the other criteria, and had no evidence of extracapsular disease on CT scan or transrectal ultrasound. 9 patients had pre-operative radiotherapy and were excluded from subsequent analysis of biochemical disease-free survival (BDFS). 12 patients had pre-operative hormone therapy. Serum PSA levels were measured with the Roche COBAS ® CORE assay in the first 80 patients (lower detecting limit = 0.1 ng ml -1 ). From 1997 onwards, 120 patients had measurements using the IMMULITE ® 'Third Generation' sPSA assay (Diagnostic Products Corporation, DPC, Gwynedd, UK). This assay detects PSA down to below levels of 0.01 ng ml -1 . Interassay variation is negligible at ≥ 0.01 ng ml -1 . This low detection level is a consequence of the efficient centrifugal wash, which results in a low non-specific signal accompanied by a large specific signal afforded by the chemiluminescent label [bib_ref] The Immulite assay tube: a new approach to heterogeneous ligand assay, Babson [/bib_ref]. Mean pre-op PSA for the whole group was 13.3 ng ml -1 (range 0.18-59 ng ml -1 , median 10 ng ml -1 ). (The patient with a pre-operative PSA of 0.18 ng/mL -1 had been previously treated by EBRT.) Around half (48%) of patients had pre-op PSA levels of >10 ng ml -1 [fig_ref] Table 1: Pre and post-operative clinicopathological characteristics of 200 men with clinically localized prostate... [/fig_ref]. One surgeon between 1992-1999 performed all the operations via the retropubic approach. When possible, a nerve-sparing technique was used [bib_ref] Radical prostatectomy and cystoprostatectomy with preservation of potency. Results using a new..., Walsh [/bib_ref]. Two pathologists graded the specimens using the Gleason classification system [bib_ref] Histologic grading and clinical staging of carcinoma of the prostate, Gleason [/bib_ref]. The new TNM staging system was used for clinical and pathological staging. At initial biopsy, 31 (15.5%) patients had Gleason grade 7 or above. Clinicopathological data is shown in [fig_ref] Table 1: Pre and post-operative clinicopathological characteristics of 200 men with clinically localized prostate... [/fig_ref]. Patients were followed-up by PSA measurements every 3 months until a year and then every 6 months. Only 12 patients who are still alive have not had a sensitive PSA measurement. Biochemical relapse was defined as 3 consecutive rising PSA values. PSA readings had to be at least 3 months apart. This is consistent with the definition proposed by ASTRO [bib_ref] Consensus Statement: Guidelines for PSA following radiation therapy, Cox [/bib_ref]. An undetectable PSA nadir is defined as ≤0.01 ng ml -1 . The log rank test was used to analyse differences in survival duration. Multivariate analysis of prognostic variables was performed by the Cox proportional hazards model. In addition, survival duration from the date of surgery to death from any cause and survival free of biochemical relapse were estimated by the Kaplan-Meier method. # Results Clinicopathological details are outlined in [fig_ref] Table 1: Pre and post-operative clinicopathological characteristics of 200 men with clinically localized prostate... [/fig_ref]. 4 patients had histopathologically proven nodal involvement, despite CT scans showing normal pelvic lymph node morphology. 82 (41%) had pT3 disease despite being considered clinically organ-confined. 49 (25%) had pathological specimen Gleason scores ≥7. Median follow-up was 1.7 years (maximum 6.5 years). Overall 2-year survival was 98.3% (95% CI: 96.4-100%). 4 patients have died, one of metastatic prostate cancer (5 years after RRP) and 3 from other causes. Overall prostate cancer specific survival was 100% at 2 years. Patients who did not have evaluable nadirs. These included those who had had prior radiotherapy (n = 9), early deaths (n = 3) and adjuvant postoperative EBRT for positive margins (n = 4). Another 44 patients had nadirs below the lower detection limit for the Roche COBAS ® CORE system. 13 patients have not yet reached their nadir. PSA nadir data was therefore analysed in 134 patients. 73 patients achieved an undetectable PSA nadir (≤ 0.01 ng ml -1 ). The median time to reach PSA nadir ≤ 0.01 ng ml -1 was 10.4 weeks (range 2.6-214 weeks). Of the 200 patients, those who had pre-operative (n = 9) or adjuvant EBRT (n = 4) or early deaths (n = 3) were excluded from analysis of BDFS. Of the remaining 184 patients biochemical relapse occurred in 72 (39%). Two-year BDFS was 68.2% (61.5-75.9%). Of the 134 patients with evaluable PSA nadir data recorded, 49 (36.5%) have failed biochemically. This gives a 2-year BDFS of 61.1% (95% CI: 51.6-70.6%). Out of the 73 patients who have reached an undetectable nadir, only 2 have failed (3%); this compares with 47 relapses out of the 61 patients (76%) who did not reach undetectable levels. Patients who achieve undetectable PSA nadir have significantly greater BDFS (Log Rank Test P < 0.0001). The 2-year BDFS for PSA nadir >0.01 ng ml -1 was 26.4% (95% CI: 14.4-38.4%) compared to those with a PSA nadir ≤0.01 ng ml -1 of 94.8% (95% CI: 87.5-100%). shows BDFS for nadir evaluable patients by comparing achievement of undetectable PSA nadir. Univariate analysis of variables predictive of BDFS showed no significant association with age or biopsy Gleason score [fig_ref] Table 2: Predictors of BDFS [/fig_ref]. However, seminal vesicle invasion (P=0.0013), surgical margins (P=0.0004), specimen Gleason score (P=0.0002), extracapsular spread (P ≤ 0.0001), pre-operative PSA (P < 0.0001) and undetectable PSA nadir (P < 0.0001) were prognostic variables. All these variables were entered into a Cox multivariate model to determine which variables were of independent prognostic significance. [fig_ref] Table 3: Predictors of BDFS [/fig_ref] shows that only undetectable PSA nadir (P ≤ 0.0001), positive surgical margins (P = 0.03) and preoperative PSA (P = 0.04) were independent variables predicting a favourable BDFS. On the basis of the finding that PSA nadir is the most significant predictor of BDFS, a model to predict PSA nadir and hence BDFS was constructed using pre-operative variables to aid patient selection. The nomograms derived from the Charing Cross dataset is shown in . Patients with low-grade tumours had a better outcome. Moreover, the deleterious effect of increasing levels of pre-op PSA level was less marked in well-differentiated tumours. # Discussion Radical retropubic prostatectomy is a popular treatment for men less than 70 years with clinically localized prostate cancer. After RRP, a proportion of patients will have pathological non-organconfined disease, often with positive margins. This has led to a quest for accurate pre-operative markers with the aim of increasing the identification of patients with organ-confined ## Relapse-free survival after prostatectomy 1435 British Journal of Cancer , disease. Patients with organ-confined disease are more likely to have complete excision of the tumour, which should result in cure. Thus, in previous large-scale studies of outcome after radical prostatectomy, attention has largely focused upon histology of the RRP specimen to define success of surgery. The overall incidence of positive surgical margins after radical prostatectomy has been reported to range from 7-45% in patients with clinically localized cancer prior to surgery [bib_ref] Radical prostatectomy with preservation of sexual function: pathological findings in the first..., Eggleston [/bib_ref] [bib_ref] Nerve-sparing radical prostatectomy: evaluation of results after 250 patients, Catalona [/bib_ref]. The accuracy with which surgical margins can predict outcome is reasonably good, especially if combined with grade. Pound et al found that in men with capsular penetration with Gleason score 2-6, the status of the surgical margin had a significant effect on outcome at 10 years. However, the outcome in both groups was good with the likelihood of an undetectable PSA at 10 years with a negative margin being 89% and 72% for men with a positive margin [bib_ref] PSA following anatomical radical prostatectomy: patterns of recurrence and cancer control, Pound [/bib_ref]. Thus, biochemical relapse is not always a consequence of positive surgical margins [bib_ref] The combination of pre-operative prostate-specific antigen and post-operative pathological findings to predict..., D&apos;amico [/bib_ref]. Opting for adjuvant therapy on the basis of a positive surgical margin status alone may result in unnecessary treatment of the patient. This lack of reliability may be because of the heterogeneity of a positive margin [bib_ref] Frequency and location of extracapsular extension and positive surgical margins in radical..., Rosen [/bib_ref]. Positive apical margins are apparently less clinically significant than positive posterior-lateral margins [bib_ref] Prognostic implications of a positive apical margin in radical prostatectomy specimens, Fesseha [/bib_ref]. The pathological determination of margin positive disease also requires its discrimination from anatomical artefact, remembering that the prostate is in direct contact with the rectum and pelvic sidewall with little or no surrounding connective tissue elsewhere. Despite these limitations, we identified positive margins as an independent predictor of BDFS in our series. Postoperative PSA readings are regarded by many to be the best criteria for determining tumour-free status. However, levels less than 0.2 ng ml -1 are generally considered of little clinical value, despite some large clinical studies demonstrating early detection of relapse with sensitive PSA assays [bib_ref] Detection of prostate cancer relapse with prostate specific antigen monitoring at levels..., Yu [/bib_ref]. Indeed, the most common level below which PSA is considered 'undetectable' is 0.2 ng mL -1 . In a recent study of the natural history of PSA elevation following RRP, 23% of patients with 'undetectable' PSA (defined as less than 0.2 ng ml -1 ) biochemically relapsed after 5 years [bib_ref] Natural history of progression after PSA elevation following radical prostatectomy, Pound [/bib_ref]. This suggests that PSA levels <0.2 ng ml -1 are often associated with the presence of residual prostate cancer. In this study, sensitive PSA was used to measure nadir post RRP. A level of greater than 0.01 ng ml -1 was found to be more accurate than any other criteria for prediction of subsequent biochemical relapse. Measuring the PSA to low levels minimizes the risk of falsely reassuring patients of surgical cure. The most sensitive PSA assay ever reported was by [bib_ref] Ultrasensitive detection of prostate-specific antigen by a time-resolved immunofluorometric assay and the..., Ferguson [/bib_ref]. Such ultrasensitive assays which detect PSA to levels less than 0.01 ng ml -1 may subsequently be shown to be even more accurate in predicting biochemical outcome. It is not possible to exclude some of the difference on BDFS accounted for by prolonged lag time in patients who achieve undetectable nadirs. However, the contour of the BDFS Kaplan-Meier Curve for patients with undetectable nadir suggests that a plateau is achieved after 2 years. It is anticipated that prolonged follow-up of this series of patients will confirm this. In addition sensitive PSA assays can detect disease relapse by consecutive PSA rises from very low levels [bib_ref] Early detection of cancer relapse after prostatectomy using very sensitive prostate-specific antigen..., Witherspoon [/bib_ref]. Concern regarding the use of sensitive assays in this way is the potential lack of specificity in diagnosing prostate cancer recur-rence. This anxiety relates to the fact that extraprostatic sources of PSA production are known to exist [bib_ref] Extraprostatic localisation of prostatic acid phosphatase and prostate-specific antigen: distribution in cloacogenic..., Kamoshida [/bib_ref]. Nevertheless, this is a very small contribution to serum PSA [bib_ref] The periurethral glands do not significantly influence the serum prostate specific antigen..., Oesterling [/bib_ref] and is believed not to complicate the interpretation of monitoring since these sources contribute a stable amount of PSA in the serum, which does not change with time. Similarly, although PSA values rise as part of the natural history in benign as well as malignant prostatic disease, this has only been reported in patients with the prostate in-situ. For example, patients with benign prostatic hyperplasia typically have a PSA rise of 0.2 ng ml -1 year -1 [bib_ref] Relationship between prostate-specific antigen, prostate volume and age in the benign prostate, Collins [/bib_ref]. It is not necessarily the case that residual benign prostatic tissue after RRP will have the similar PSA kinetics. Recent papers have suggested the use of PSA doubling time to predict malignant potential [bib_ref] Recurrence patterns after radical retropubic prostatectomy: Clinical usefulness of prostate-specific antigen doubling..., Patel [/bib_ref] , but even this parameter is of questionable value. Our data suggest that all patients who relapse biochemically continue to have an unremitting rise, and this occurs whether biochemical relapse is detected early (with sensitive assays) or late. Another reason for the limited enthusiasm for sensitive assays is the lack of proven benefit from the use of early salvage external beam radiotherapy (EBRT). Critics claim that salvage EBRT to the prostate bed risks over-treating patients with rising PSA due to either undiagnosed metastatic disease or benign disease. However, potential benefits do exist for salvage EBRT. In a recent study investigating the outcomes of patients who received salvage EBRT after radical retropubic prostatectomy, 40% of the 32 patients enrolled had 'undetectable' post-radiotherapy PSA values at an average follow-up of 12 months. Moreover, the best outcomes were associated with patients who had the lowest pretreatment PSA levels [bib_ref] Limited suppression of prostate-specific antigen after salvage radiotherapy for its isolated elevation..., Egawa [/bib_ref]. Probability tables as shown in this paper provide information on the likelihood of achieving an undetectable PSA nadir based on pre-op PSA and biopsy Gleason score. Although this is useful, the ability to predict accurately biochemical outcome in an individual patient prior to surgery is still unattainable. Reliably selecting patients for RRP who will not relapse biochemically after surgery continues to be a problem. However, ultrasensitive PSA assays are of benefit in the postoperative detection of relapse which is now known to be advantageous in planning adjuvant therapy. # Conclusions The role of RRP has been controversial in the UK partly because of the scepticism regarding the effectiveness of the procedure. Undoubtedly, not all patients undergoing RRP are cured of prostate cancer and patient selection is an important factor. Patients are keen to know if they have been cured of cancer after RRP. Until recently, to predict this we have been reliant upon histology alone. However, ultrasensitive PSA assays now enable us more accurately to advise patients of their chance of PSA relapse within 2 months of surgery and appears to be a reliable predictor of cure of prostate cancer. [table] Table 1: Pre and post-operative clinicopathological characteristics of 200 men with clinically localized prostate cancer Number of patients (%) Figure 1 Kaplan-Meier biochemical disease-free survival curves for patients with PSA nadir ≤0.01 ng ml -1 and PSA nadir >0.01 ng ml -1 . Chi-square = 71.67, P < 0.0001 [/table] [table] Table 2: Predictors of BDFS. Univariate analysis [/table] [table] Table 3: Predictors of BDFS. Multivariate analysis. [/table]

Calcific Aortic Valve Disease-Natural History and Future Therapeutic Strategies # Introduction Calcific aortic valve disease (CAVD) is the most common valve disease worldwide. Epidemiological studies show that 2.8% of adults over 75 years old have some CAVD degree, and as many as 25% of adults over 65 years old have at least valvular sclerosis. CAVD is a chronic process characterized by progressive fibrotic tissue remodeling and mineralization. Over the years, there is a disease continuum from sclerosis to chronic inflammation and finally leaflet calcification, culminating with severe stenosis. Human pathologic samples have shown that key features in the CAVD development include pathological concentrations of inflammatory cells and lipid species. Several risk factors have been identified as relevant in the CAVD progression. Among others, male gender, high triglyceride levels, and smoking have been independently associated with early aortic valve replacement in the presence of CAVD. Levels of oxidative stress, higher in patients that are more exposed to certain risk factors than others, could be responsible for the starting of the sclerotic CAVD phase. Calcification is recognized as an active disease process driven by the native valvular interstitial cells (VICs). These cells acquire an osteogenic and pro-calcific profile in response to different pathological stimuli, such as inflammatory mediators, endothelial damage, low-density lipoprotein (LDL) accumulation, reactive oxygen species (ROS), increase calcium/phosphate (Ca/Pi) levels, modified lipids, and cyclic stretch. The underlying process that ends with the ectopic mineralization of the aortic valve is still not entirely understood. Treatment options include operative valve replacement and percutaneous implantation of valve prosthesis. To date, there is no medical treatment available to prevent or reverse calcium deposition within the valve leaflets. Conventional cardiovascular drugs studied in clinical trials failed to influence the disease progression or reduce adverse outcomes, therefore additional research is required to understand the mechanisms of disease progression and, identify novel therapeutic targets. ## Cavd biology histological structure of the aortic valve The aortic heart valve consists of three leaflets that allow blood flow from the left ventricle to the aorta without regurgitation. Each leaflet has a trilaminar structure that is vital for the biomechanical properties of the aortic valveas shown in . Histologically, the leaflets are composed of three distinct layers: fibrosa, spongiosa, and ventricularis. Fibrosa and ventricularis are the external layers, facing the aorta and the left ventricle respectively. Fibrosa consists largely of collagen fibers with dispersed VICs, that are thought to be responsible for reinforcing the valvular structure. The central layer, spongiosa, is rich in glycosaminoglycans (GAGs) and is responsible for absorbing some of the mechanical stress generated during the cardiac cycle. The ventricularis is localized on the ventricular side of the leaflets and consists of collagen and elastin fibers. The tissue composition of ventricularis provides more compliance and grants the apposition of free edge leaflet regions, thus preventing the backward blood flow into the left ventricle during diastole. The normal valve leaflet is avascular and free of infiltrating lymphocytes or monocytes. ## Pathobiology of cavd The CAVD pathophysiology is schematically depicted in . The process of progressive fibrocalcific remodeling of the aortic valve is multifactorial, involving genetic predisposition, endothelial shear stress, chronic inflammation, lipid deposition, and valve calcification. In its early stages, CAVD resembles atherosclerosis. The existence of shared risk factors for the disease development and the correlation between the severity of CAVD and that of coronary calcification suggest a shared disease pathway, at least in the initial phases of both diseases. Alike atherosclerosis, the triggering event in CAVD is endothelial damage resulting from increased mechanical stress and reduced shear stress. Physiologic fluid shear stress (FSS) contributes to valve homeostasis, whereas altered shear stress, on the contrary, stimulates the endothelial upregulation of vascular cell adhesion molecule-1 (VCAM-1) and intracellular adhesion molecule-1 (ICAM-1). The endothelial damage following the altered shear stress allows the infiltration of lipids, especially LDL and lipoprotein (a) [Lp(a)], starting the recruitment of inflammatory cells into the leaflets. Later appears the formation of focal subendothelial plaque-like lesions constituted from LDL, Lp(a) and infiltrates of inflammatory cells which interact in processes that release ROS causing LDL oxidation. In time, the expansion of this inflammatory process triggers the release of factors that cause VICs to acquire an osteogenic profile, with the formation of microcalcifications initiating leaflet mineralization. The two disease processes differ in the later disease phases: while in atherosclerosis, smooth muscle cells are the active players in the chronic inflammation within the plaque, in CAVD fibroblasts are involved in the prominent mineralization process and only with a small number of smooth muscle cells is present. The hallmark of the early CAVD stages is inflammation, characterized by the activation of the leaflet endothelium via enhanced expression of cell adhesion molecules (VCAM-1, ICAM-1). Disease initiation involves the activation of VICs, recruitment of immune cells, and subsequent sclerosis of the valve leaflets owing to fibrosis and formation of calcific nodules. All these phenomena start and interest more the fibrosa layer within the valve. The first macroscopic change in the leaflets, seen as microcalcifications, or focal thickening with preserved valve function, is nominated aortic valve sclerosis, nonetheless, the initiating events likely occur much earlier. The activation degree of the immune system seems to be different on different examined valve areas. Interestingly, the thickening process and the formation of calcium nodules accompanied by neoangiogenesis, are both localized near the aortic surface of the leaflets. The final disease, namely calcific aortic stenosis, is characterized by large calcified noduli on the leaflet surface, that protrude into the sinuses of Valsalva, hindering the leaflet mobility. In this final phase the leaflets are infiltrated by immune cells and concomitantly angiogenesis occurs, along with deposition of lipids, proteoglycans, and cell debris. Finally, the calcification of the valvular matrix leads to an increased stiffness and obstruction of the blood flow . With an orifice under 1 cm 2 , versus 2.5-4.5 cm 2 observed in a normal valve, the stenotic valve generates a pressure gradient of over 40 mmHg, categorized as severe stenosis with an indication to valve replacement. ## Impact of inflammation in cavd remodeling Inflammation is the primary response of innate immunity and occurs after endothelial damage with its activation and lipid deposition. The innate immune response, with both its components, cellular and humoral responses are implicated in this process. As response to an injury induced by foreign organisms, dead cells or physical irritants, the innate immune system represents the first response to external or internal triggers and initiates the process of tissue regeneration. Once the inflammation process has been triggered, it will proceed along a certain course of events until the inflammatory stimulus is eradicated and the healing mechanism can begin. However, if the inflammatory source cannot be eliminated, inflammation will progress, varying in intensity over time. Histological samples of human CAVD valves are characterized by calcified areas rich in lymphocytes, macrophages, and osteoblast-like cells. The inflammatory process can be acute or chronic. Cote´et al. showed that chronic inflammatory infiltrates, composed of the CD45 + leukocytes, CD68 + macrophages, and a few scattered CD3 + T cells, were present near the calcified areas. Moreover, the chronic inflammatory infiltrates in the aortic valve were independently associated with several indices of remodeling, suggesting that inflammation may participate in mineralization and the fibrotic process. A significant association between the degree of aortic valve inflammation and the development of calcification has been previously reported (and recently confirmed) using 18Ffluorodeoxyglucose positron emission tomography. In these studies, a high degree of inflammation and calcification was documented in patients with severe CAVD, with the latter being the predominant pathogenic process. This could in part explain the failure of statin therapy in slowing the CAVD progression. Previously thought to be a passive disorder, CAVD is now recognized as an active process, whereby endothelial progenitor cells (EPCs) and inflammatory cells promote tissue remodeling. In summary, mechanical shear stress and atherogenic factors activate VICs and initiate the recruitment of inflammatory cells. The following remodeling of extracellular matrix with leaflet stiffening and valvular dysfunction causes further mechanical stress that maintains the self-perpetuating cycle of endothelial dysfunction. As suggested by studies with molecular imaging, the continuous maintenance of this shear stress-inflammation cycle could result irreversibly to calcium deposition and finally to severe CAVD. ## Role of lipids Over recent years, several studies have highlighted the central role of ox-LDL as an activator of inflammation, both in atherosclerosis and in CAVD . In atherosclerotic plaques, ox-LDL activates the inflammatory cells and the production of cytokines promoting tissue remodeling and disease progression. Existing evidence of valve infiltration by ox-LDL in cases of CAVD, supports the concept that the development of valve calcification may be, at least in part, influenced by ox-LDL and, that an association exists between high plasma levels of ox-LDL and CAVD . Although having some similarities with atherosclerosis, CAVD differs from it since the aortic valve has some inherent properties that differ from the vascular wall. Indeed, the cellular organization, as well as the hemodynamic stress imposed upon the aortic valve, vary from that of the arteries. Furthermore, the phases that characterize the pathological process occurring within the valve, from early inflammation stages to the calcification ones, are more numerous and qualitatively different from atherosclerosis, as shown in . Most importantly, statins, although efficient in reducing adverse events in patients with atherosclerosis, are inefficient in CAVD. ## Role of the immune response in cavd The immune system, with innate and adaptive responses, plays a central role in the development of different chronic disorders, including atherosclerosis and CAVD. Evidence suggests that inflammation, as the primary response of the innate immunity, promotes the mineralization of VICs in response to several factors. Adaptive immunity, on the other hand, could play a role in instrumenting the immune response. In this regard, experimental studies have shown that the increased leukocyte density in mineralized aortic valves correlates with faster disease progression. ## Innate response in cavd The innate immune response in CAVD |+| }_be initially triggered by several oxidized lipid species, through both the toll-like receptors (TLRs) and the nuclear factor-k B (NF-kB) pathway. VICs express TLRs, known to play a key role in inflammation and initiation of antigen-specific adaptive immune responses. TLRs can be activated by several lipid species, especially ox-LDL, with the signal passing through the recruitment of specific adaptor molecules to the activation of the transcription factor NF-kB. In this way, TLRs signal inflammation through the NF-kB pathway and are vital for maintaining tissue homeostasis. NF-Kappa B Pathway. The NF-kB plays an important role in signal integration by responding to mediators of endothelial injury. VICs actively participate in the regulation of inflammation by producing a high level of cytokines, and the NF-kB pathway contributes to a stimulus-dependent and cell-typespecific manner. The NF-kB activation (canonical and noncanonical) pathway is mediated by different extracellular signals including angiotensin II, ox-LDL, CD40 ligand, advanced glycation end-products, and inflammatory cytokines. The canonic pathway is activated, among others, by tumor necrosis factor alpha (TNF-a) and interleukin-1b (IL-1b). Angiotensin II-induced ROS generated also on mineralized aortic valves, are involved in the activation of the NF-kB canonic pathway. The non-canonical pathway is activated, among others, by vasoactive peptides, ox-LDL, activated CD40 receptor, B cell-activating factor (BAFF), lymphotoxin b (LTb) monocyte released cytokines, or advanced glycation endproducts. In summary, activated NF-kB controls the initiation of vascular inflammation, through "leucocyte adherence" and "chemotaxis" and is a crucial signaling integrator of vascular injury. Cytokines. One important target of the NF-kB pathway is IL-6. Highly expressed in valves with severe CAVD, IL-6 is a pleiotropic cytokine secreted by multiple vascular cell types, such as macrophages, lymphocytes, fibroblasts, endothelial cells, and smooth muscle cells. It mediates local vascular monocyte activation and protection from ROS-induced cellular stress via the downstream transcription effector signal transducer and activator of transcription 3 (STAT3). By controlling the monocyte activation vial the IL-6 pathway, NF-kB mediates the systemic acute phase response and plays a central role in the initiation and maintaining of the vascular inflammation. IL-1b is another intriguing cytokine involved in the valve calcification process. Its levels are indeed increased in the stenotic aortic valve, making IL-1b subject to numerous research studies in the last decade. The IL-1b enhances the expression of matrix metalloproteinases (MMPs), a group of enzymes responsible for the degradation of extracellular matrix proteins. The MMPs exacerbate the process of valvular stenosis and activate the NF-kB pathway, leading to the production of several cytokines, such as IL-6, IL-8, and MCP-1 (monocyte chemoattractant protein-1) involved in both pathological processes, namely atherosclerosis and CAVD. Finally, the IL-1 receptor agonist (IL-1Ra) plays a protective role in valvular disease, as its deficiency is closely associated with inflammatory cell infiltration and valve thickening. Using in vitro and in vivo models, Zeng et al. elucidated the role of IL-37, an anti-inflammatory member of the IL-1 family. IL-37 attenuates the expression of bone morphogenetic protein 2 (BMP2) and alkaline phosphatase (ALP) enzyme, inhibiting the osteogenic process. In CAVD, the expression levels of IL-37 are very low, consequently, BMP2 is free to promote VICs calcification by ALP expression leading subsequently to aortic valvular thickening. Confirming the protective role of IL-37, in vivo experiments showed that mice expressing this human cytokine, displays significantly lower BMP-2 levels and a lower degree of aortic valve thickening. ## Adaptive response in cavd The key role of lymphocytes in CAVD and the presence of Tlymphocyte clonal expansion in stenotic valves has been highlighted by various studies. In a study involving patients with severe CAVD, Wu et al. observed the presence of CD8 + /CD28 − T cells near the mineralized nodules of the aortic valve and a higher prevalence of circulating CD3 + T cells, namely the subset of CD8 + and CD57 + T cells expressing HLA-DR, in subjects with CAVD. CD8 + /CD28 − T cells play a crucial role in CAVD, as demonstrated by the correlation between the degree of clonal expansion and the severity of valve calcification. Furthermore, these cells play a crucial role in both activation and differentiation of memory effector status among circulating T cells. Winchester et al., investigated the composition of the lymphocytic infiltration in patients with (bi-and tri-cuspid CAVD). They showed an aggregate of infiltrating lymphocytes containing CD4 + and CD8 + T cells, with a preponderance of CD4 T cells. The clones shared between blood and the valve were found in the memory-effector CD8 + CD28 − T cell subset, demonstrating the trafficking of members of the same T cell clone between the peripheral circulation and the valve. Importantly, the proportion of activation of peripheral blood T-lymphocytes strongly correlated with the degree of valve calcification, indicating that the extent of these events is somehow related to CAVD severity. This data suggests that in patients with CAVD, an adaptive systemic immune response is occurring, coupled with the valvular lymphocytic infiltration, probably triggered by recognition of antigens expressed in the valve. The valvular events that lead to an immune response, such as intracellular pathogens, or self-antigens, remain unclear. One hypothesis is that in response to shear stress the VICs express stress-induced molecules which could be perceived as antigens and generate an immune response. ## Mineralization At some point during the CAVD progression, VICs start to produce a calcified matrix and enter an osteogenic phase. The mechanisms of this switch are still poorly understood. One hypothetical pathway involves IL-6 and the tumor necrosis factor ligand superfamily member 11 (also called receptor activator of NF-kB ligand-RANKL). IL-6 has been shown to induce the expression of RANKL in bone cells, which activates its cognate receptor RANK that finally activates VICs to produce extracellular matrix and might therefore promote matrix calcification. This could be a hypothetical pathway, which through IL-6-RANKL overexpression promotes the osteogenic reprogramming of VICs. Another pathway involving the expression of BMP2 via inflammatory cytokines and oxidized lipid derivatives was shown to induce osteogenic reprogramming in several cell types including VICs. ## Genetic factors implicated in cavd Numerous genetic factors are implicated in the CAVD pathogenesis. Two recent genome-wide association (GWAS) studies showed that a genetic variation in the LPA locus, mediated by levels of Lp(a), is involved in both atherosclerotic disease and CAVD and that several pathways are shared by both conditions. Lp(a) and non-high-density lipoprotein cholesterol as shared risk factors contributed to the frequent co-existence of these disorders. A recent transcriptome-wide association study (TWAS) identified PALMD (palmdelphin), a gene that promotes myoblast differentiation and muscle regeneration, as an additional key gene in CAVD. The study showed that lowered expression levels of PALMD mRNA in valve tissues are associated with risk alleles for CAVD and higher disease severity. Recently, Theriault et al. identified some additional susceptibility genes in patients with CAVD. Using both techniques, GWAS and TWAS, IL-6, ALPL (alkaline phosphatase), and NAV1 (neuron navigator 1) emerged as important contributors involved not only in pulse and blood pressure modulation-two important factors associated with blood flow turbulence, already identified as risk factors for CAVD development-but also in the valve mineralization process. Moreover, the ALPL gene codes for tissue-nonspecific alkaline phosphatase, a crucial enzyme involved in mineralization. This gene was present in calcified aortic valves contrary to non-calcified valves. All these findings taken together constitute a new step toward the further elucidation of the CAVD pathogenesis and may open new paths for the development of innovative therapeutic approaches. ## Gender as a risk factor for cavd The role of gender as a biological variable, especially that of the male sex as a risk factor for cardiovascular diseases, including CAVD, has been long established; Institute of Medicine Board on Population H and Public Health P, 2012). However, data regarding the significance of male sex in CAVD is lacking. Indeed, most studies addressing sex-specific differences in CAVD focus on conditions caused by CAVD, such as ventricular dysfunction, rather than CAVD itself. It is mostly unclear whether the onset and progression of CAVD are inherently different in men and women, due to genetic and cellular-scale sex differences, or whether the initial disease process is common in both sexes but undergoes subsequent differentiation after. Aggarwal et al. reported that male gender correlates with an higher incidence of CAVD, as men displayed higher levels of calcification for the same degree of aortic stenosis when compared to women. Another study showed significant differences between stenotic valves in males and females, where the latter were more prone to fibrosis.Evidence points towards genetic and epigenetic differences, specifically the differential production of miRNAs, being the origin of these sex-specific differences. ## Epigenetic regulation of cavd The remarkable developments in the field of epigenetics could offer answers regarding the link between environmental and proatherogenic factors, induction and perpetuation of inflammation, and CAVD development. The four major epigenetic mechanisms, collectively known as "epigenome", include non-coding regulatory RNAs (ncRNA), DNA methylation, histone modifications, and chromatin remodeling. Studying the link between epigenome and CAVD could be clinically relevant since epigenetic markers could represent non-invasive biomarkers for monitoring CAVD initiation and progression and predicting prognosis. ncRNAs, a group of single-stranded cellular RNAs made up of 18-25 nucleotides, represent a new area of interest in the field of human biology. These molecules, normally not coding for proteins, were previously regarded as "dark matter" since their function was largely unknown. Increasingly discovered to play a central role in the regulation of various molecular pathways and cell differentiation, ncRNAs are categorized in short, medium, or long (lncRNAs) based on their transcript size. MicroRNAs (miRNAs) are the most well-studied ncRNAs, involved in cell development, proliferation, lipid metabolism, cancer metabolism, and angiogenesis. They regulate gene expression after transcription, by translational repression or transcript degradation. The role of miRNAs and especially lncRNAs in regulating atherosclerotic processes has been investigated in several studies suggesting that some miRNAs may be important inducers of proinflammatory pathways and others stimulators or inhibitors of calcification in the human aortic VICs. Among others, miRNA-30b, miRNA-138, and miRNA-204 can suppress the differentiation of mesenchymal cells into osteoblasts whereas miRNA-29b and miRNA-214 promote calcification. miRNA-138 was indicated as a suppressor of osteoblastic differentiation of human aortic VICs by targeting FOXC1, representing, therefore, an inhibitor of VIC osteogenic differentiation during the development of CAVD. In 2017, the role of miRNA-29b as a promoter of valvular calcification, through the TGF-ß3/Smad3 and wnt3/ßcatenin pathways was reportedwhereas in 2019, miRNA-214 was identified as a promoter of the human aortic VICs calcification by accelerating inflammatory chain reactions through MyD88/NF-kB signaling . Another epigenetic mark, the alteration of DNA methylation, could have a role in the osteogenic differentiation of aortic VICs by downregulating Notch/drosophila/homolog 1/translocationassociated (NOTCH1), a regulator of calcification-related gene networks in human valvular endothelium. Markers of histone modification have also been reported to have a pro-inflammatory and osteogenic role in CAVD development. They participate in the shear-stress induced proinflammatory pathways via alteration of the silent information regulator-two (SIRT) gene expression. In summary, increasing evidence supports the role of epigenetic factors as regulators of key processes underlying valvular tissue remodeling and participators in the landscape of phenotypical changes occurring in CAVD. Understanding the epigenetic mechanisms involved in the initiation and progression of CAVD will not only help to enlighten the pathology, but since these markers are potentially reversible, it could offer important targets for diagnostic and therapeutic interventions. ## The putative role of the microbiome Herein we hypothesize a putative role of the microbiome in the development and progression of CAVD, either via stimulation of immune response and valve calcification or via direct valvular damage caused by specific bacterial taxa. Different hints point to a role microbiome in the promotion of CAVD and its deterioration to severe calcified aortic stenosis. Some known risk factors for CAVD, such as age, male sex, hypertension, and diabetes mellitus, have a known relationship to the microbiome. In fact, patients with type II diabetes mellitus (T2DM), which is associated with the gut microbiota dysbiosis, are at higher risk for developing CAVD. Tissue histopathological studies demonstrated more calcification in tissue samples from patients with T2DM compared to nondiabetic patients. Furthermore, the presence of metabolic syndrome is associated with faster disease progression and worse outcome in patients with aortic stenosis. Here we discuss in detail the involvement of oral and gut microbiome (GM) in pathophysiological processes that promote cardiovascular diseases, including CAVD. ## Oral microbiome The oral microbiome represents, after the GM, the secondlargest microbial community in humans and plays a fundamental role in maintaining the oral cavity homeostasis and preventing the development of several diseases. The involvement of the oral microbiome in pathologies such as obesity, diabetes, cancer, and cardiovascular diseases is well documented. Different studies suggest that oral pathologies with oral microbiome alteration, have a tight link with heart diseases; for example, the gingival ulceration in periodontitis causes bacteremia and could induce the formation of atherosclerotic plaques. Increasing evidence suggests a close association between oral microbiome alteration and CAVD, demonstrating the presence of oral bacteria in the valvular tissue by using a polymerase chain reaction (PCR). Nomura et al., investigated the mechanisms through which Streptococcus mutans (S. mutants), the major pathogen responsible for dental caries, colonizes heart valves constituting a potential virulence factor for the development of infective endocarditis. They demonstrated that specific strains of S. mutants have selective virulence for infectious endocarditis. There is an increasingly accepted hypothesis that bacterial endocarditis is one of the causes of aortic valve calcification, supported by in vivo studies showing larger calcifications in animals inoculated with endocarditisrelated pathogens.Gut Microbiome The composition of the human microbiome is specific to each individual and the study of the GM is increasingly focusing on its role in the physiology of the host organism in during phases of health and disease. Changes in the composition of GM are involved in several diseases including atherosclerosis, hypertension, heart failure, chronic kidney disease, obesity, diabetes, inflammatory bowel disease, and colon cancer. Growing evidence suggests that intestinal bacteria play an essential role in cardiovascular diseases. GM metabolites such as choline, betaine, and trimethylamine N-oxide (TMAO), are the subject of research studies investigating the correlation between the intestinal microbiome and the calcification of the aortic valve. Diets based on red meat, eggs, and dairy products are rich in choline, lecithin, and carnitine, constituting, therefore, a potential source of TMAO. Recent metabolomic studies indicate that fasting TMAO, choline, and phosphatidylcholine plasma levels are related to high risk of cardiovascular diseases. Furthermore, carnitine has been indicated as an independent predictor of major adverse cardiovascular events, with TMAO as the main driver behind the risk association. The idea that gut microbiota could contribute to CAVD is new. Very recently Kocyigit et al. reported a significant association between gut microbiota metabolites and CAVDwhereas showed that patients with CAVD and those with coronary artery disease suffer from different gut microbial dysbiosis . The next section discusses the putative link between the GM and CAVD development. ## The role of the gm in adaptive immunity Several studies indicate a very close connection between the GM and adaptive immunity. Bacterial metabolites function as a link between the commensal microbiome and the immune system, by affecting the shifting of the immune response between pro-and anti-inflammatory pathways. Hypothetically, the gap between the microbiome and immune system could be filled by shortchain fatty acids (SCFAs), generated by bacterial fermentation of dietary fiber in the intestinal lumen. SCFAs such as acetate, propionate, and butyrate act as important inductors of regulatory T (Treg) cell differentiation and reinforce the gut barrier function by expressing the transcription factor Foxp3 and consequently suppressing inflammatory responses in the intestine. A diet rich in SCFAs might antagonize several immunological defects, therefore GM-derived metabolites could offer a therapeutic advantage for patients suffering from various immunological conditions. Significant links between fungal-and bacterial-induced cytokine responses and specific gut bacterial species were identified through the Human Functional Genomics Project. For instance, the production of interferon-g (IFN-g) and TNFa is strongly associated with the GM. The fermentation of bacterial SCFAs acts through activation of Gprotein coupled receptors by modulating the activity of intestinal epithelial cells and leukocyte life cycle. Furthermore, due to their direct effect on lymphocytes and by activating macrophages and dendric cells, SCFAs can induce a T-lymphocyte tolerogenic profile, acting as a link between the microbiome and the immune system (Correa-. In summary, the role of adaptive immunity in CAVD has been explored, but the valvular events that lead to an immune response, such as intracellular pathogens or self-antigens, remain largely undetermined. Alike atherosclerosis, one could hypothesize that the presence of a certain type of gut or mouth microbiome could lead to the presence of bacterial DNA inducing valvular specific antigens that could trigger an immune response. The putative interaction between the microbiome and immune response leading to CAVD development is schematized in . ## Novel potential therapeutic strategies Currently, no drug strategies can prevent or reverse CAVD and valve replacement remains the only treatment option. Patients with severe CAVD undergo valve replacement either with surgery or via percutaneous transcatheter procedures. Drugs already used to treat vascular complications might also improve CAVD outcomes, but the mechanisms of valve calcification differ from other vascular conditions. A randomized trial of intensive lipid-lowering therapy failed to stop ore reverse the progression of CAVD. However, the possibility that this specific pharmacological treatment induces a small reduction in disease progression or in the main clinical endpoints cannot be excluded. Antiinflammatory agents represent a promising therapeutic strategy that requires additional research in CAVD patients. Due to the role of angiotensin II in CAVD development, inhibitors of angiotensin-converting enzyme (ACE-I) have also been explored. Notably, no significant difference in disease progression was observed in a retrospective cohort of more than 200 patients. The benefits of ACE-I are mostly related to their effect on left ventricular remodeling. Human monoclonal antibodies have been used successfully in many conditions. In 2015 Lerman et al. found that denosumab, a human monoclonal antibody targeting the receptor activator of nuclear factor-kB ligand, may reduce valvular calcium deposition to basal levels. Furthermore, denosumab reduced calcium deposition in the aorta, although the mechanisms by which it affected ectopic calcification, must be further examined.Nonetheless, this field remains attractive and promising. After GWAS and TWAS studies discovering several genes implicated in the CAVD pathogenesis, genetic therapy could also be a possible therapeutic development. A growing body of experimental evidence supports the role of epigenetic factors such as miRNAs and markers of DNA methylation, as regulators of key processes underlying valvular tissue remodeling and participators in the landscape of phenotypical changes FIGURE 3 | The putative interaction between the microbiome and immune response in CAVD development. Increased shear stress, genetic factors, and oxygen reactive species cause the endothelial damage that allows ox-LDLs, immune and inflammatory cells to infiltrate the valve. Macrophages and ox-LDLs stimulate VICs inflammation through NF-kB/IL-6 signaling pathway. The microbiome and microbial metabolites can modulate the immune system through the intense effect of SFCAs on T cells' differentiation. Furthermore, the bacterial metabolite TMAO can induce ox-LDL and maintain the inflammatory loop. IL-6, interleukin 6; NF-kB, nuclear factor k-light-chain enhancer of activated B cells; ox-LDL, oxidized low-density lipoprotein; SCFA, short-chain fatty acid; TMAO-Trimethylamine-N oxide; VECs, valve endothelial cells; VICs, vascular interstitial cells. occurring in CAVD. Since these markers are potentially reversible, they could offer important targets for diagnostic and therapeutic interventions. The recently introduced, tissue-engineered heart valves [tissue engineered heart valves (TEHVs)] may constitute a valid alternative to traditional strategies of valve replacement. In TEHVs, a decellularized scaffold is seeded with patient's cells subjected to an appropriate series of stimuli able to promote cellular activity. TEHVs could offer the advantages of biocompatibility and longevity, while preventing pathological tissue responses and regurgitation. Additionally, TEHVs could allow VICs to produce their own extracellular matrix, thus displaying the capacity for growth and remodeling, which is critical for pediatric patients. TEHVs can be produced from synthetic or natural materials. # Conclusions Despite its high prevalence and associated mortality, there are no drugs to prevent or cure CAVD. Therefore, the development of new therapeutic strategies capable of counteracting the CAVD development is crucial. Being a multifactorial disease, CAVD is characterized by a complex pathobiology that involves environmental and genetic factors, immune-molecular pathways, hemodynamic factors, and shear stress, all intertwined in the systemic district. Several of the many steps of disease progression have been clarified but some others, like the critical switch from inflammatory-fibrotic to osteogenic program, or the putative role of microbiota, remain still unclear. The quickly evolving field of epigenetic regulation of CAVD, involving especially miRNAs and lncRNAs, could offer novel potential biomarkers for the development of new diagnostic and therapeutic interventions. Furthermore, the microbiome could play a role in promoting and perpetuating CAVD by inducing the production of endogenous or exogenous antigens via the activation of inflammatory pathways. The elucidation of these conundrums will help identify new disease targets, host or microbial, and support the design of new potential therapeutic strategies to prevent or reverse the effects of CAVD. # Author contributions BA, LC, AL, and AA were involved in the study concept and design. BA, LC, and MC drafted the manuscript. AL, UL, HG, and AA revised the manuscript for important intellectual contents. All authors had access to all the data, including figures, and approved the manuscript for final submission.

Introducing patient and public involvement practices to healthcare research in Austria: strategies to promote change at multiple levels ## Reviewer Barker, Jacqueline University of the West of England Bristol, Faculty of Business and Law REVIEW RETURNED 08-Feb-2021 ## General comments This is an interesting paper. I think the following needs to be addressed: 1. There is no clear research question, therefore it is hard to judge the methods. 2. The methods are not completely described (eg how was the survey administered, how were the questions developed etc). 3. In research about PPI, it is accepted practice to have involved the public. It looks as if the public were not involved in this paper so it is important to say why. 4. The UK organisation INVOLVE has been replaced -see this link https://www.nihr.ac.uk/news/nihr-launches-new-centre-forengagement-and-dissemination/24576 so some of your wording may need to be updated. 5. You seem to use the term 'levels' in 2 different ways: 1) to describe levels of PPI from consultation to control and 2) in your abstract and elsewhere in terms of levels of organisation e.g. the individual, the organisation, the governance structures. You may need to clarify and flesh out these two different concepts and how they interact in your paper. 6. Proofing point: there is one use of 'und' instead of 'and' 7. In the background section, you mention the reasons for PPI but only list 2 items. I think you are missing arguments about citizens' rights 8. Towards the end of your paper you use the term 'co production' which has specific meaning and for many people is not interchangeable with 'involvement'. I think you need to either stick to involvement or reflect the debate around involvement and co production in your paper. ## Version 1 -author response Reviewer: 1 Dr. Claudio Di Lorito, University of Nottingham Comments to the Author: ## Dear authors Thank you for your manuscript, which I read with great interest, given my previous experience with PPI. I commend your commitment and the initiative you have undertaken to further PPI practice in your country. However, I feel that at this stage, your manuscript requires further work before it is ready for publication. I have provided extensive points for improvement in the attached PDF, which I would invite you to carefully address, before I recommend publication to the Editor. Other than several other issues, as it stands, the main problem with the manuscript is how it is structured. To me, you have two distinct elements to your work : 1. The surveys; 2. The anecdotal evidence based on personal experience (currently entitled "points for attention"). At present, you have structured the paper as if the surveys are the only and core element of your data collection, but it does not work, because the current discussion is mostly unrelated to the survey findings. In fact, the discussion presents new data/findings and introduces new concepts based on your personal experience. I would re-structure the manuscript so that the results section includes the surveys and anecdotal personal experience. The discussion can then focus on the initiative you have been working on to advance PPI in Austria. This should be also reflected in the abstract and Aims and Objectives section which you need to include at the end of the introduction. Looking forward to seeing these comments addressed. ## Best wishes and good luck Thank you for your suggestions and the specific feedback provided in the annotated PDF file. We have addressed all comments, which we have listed below referring to the comment sequence and page number in the annotated PDF file: ## Comment 1, page 4: We have re-written the abstract, which is now in a structured format and follows the order in which we have presented information in the main text: survey, anecdotal evidence, strategies to promote PPI. Comment 2, page 4: We fully acknowledge the reviewer's point that survey findings and observations from our personal experience need to be explicitly distinguished in the abstract and main text. We have re-structured and re-worded accordingly, and we have also explicitly described our personal observations as anecdotal evidence. Comment 1, page 5: Thank you for noticing this, and we changed "citizens" to "members of the public" throughout the manuscript. . The checklist could serve as best practice example and standard operating procedure for Austrian ethics committees in dealing with PPI. Applying the checklist to their own work, applicants may be asked to, e.g., describe the role of patients and members of the public in their project, distinguish between study participation and involvement and highlighting possible ethical issues. This could support quality assurance and implementation of standards for PPI and give researchers an opportunity to self-evaluate their ethical considerations around PPI." Comment 2, page 6: We deleted "formative" throughout the manuscript, as it does not add information to the manuscript. ## Comment 3, page 6 We added further detail on the survey design, sampling, and distribution (page 5, lines 18-38) Comment 4, page 6: We clarified that dissemination does not actively involve members of the public and therefore is not considered PPI (page 5, line 42, to page 6, line 2): "Respondents were generally active in disseminating research findings to patients or the public, via traditional media, social media, popular science events, and other channels, which represents engagement/dissemination and not involvement." Comment 5, page 6: We have added describing the results of the second online survey to research ethics committees. Comment 6, page 6: We thank the reviewer for this comment, and we admit that there is some irony in not including any PPI in a manuscript about PPI. We hope that the manuscript does convey our good faith intentions to realise a genuine PPI approach in our research community, and that it provides some useful real-world insights into the types of difficulties we have encountered. Input and feedback from members of the public would undoubtedly be valuable; however, in this first scoping step we did not include members of the public in the study design or procedure. Our perception was that the immediate barriers we encountered in our work related to lack of awareness and understanding among researchers, and we therefore intended to scope current PPI practices among this group. For the reviewer's information, members of the public have been involved in the design and concept of the PPIE programme and funding model in 2019. We rephrased the sentence (page 6, lines 37-39): "Members of the public were not involved in the design and conduct of the surveys, because the immediate barriers to PPI we encountered in our work seemed to relate to awareness, knowledge, and perceptions among researchers." Comment 1, page 7: We removed the sentence here, and we have clarified the distinction between active involvement (PPI) and dissemination earlier in the manuscript (see above, your comment 4, page 6). Comment 2, page 7: Thank you for the suggestion, we changed "gradient" to "trend" (page 7, line 1): Comment 3, page 7: Thank you for the comment. We expanded the statement (page 7, lines 4-6): "This snapshot encourages us to further promote PPI on individual level by offering training and facilitating exchange among researchers; and to introduce support structures on institutional and national level." Comment 4, page 7: Thank you for this comment. We rewrote the sentence in more general terms not focusing on co-researchers (page 7, lines 6-7): "In our second survey, most representatives from research ethics committees were unfamiliar with the PPI concept, but interested in discussing its ethical aspects." Comment 5, page 7: Thank you for this comment. We moved the "anecdotal evidence based on personal experience" to the Scoping section (page 6, lines 18-35) and the "organisational, individual, and structural strategies" to the Background section (page 4, line 34, to page 5, line 2). We further discuss the initiative and working group with Austrian ethic committees in the Discussion section (page 8, line 24, to page 9, line 2). ## Comment 6, page 7: We rephrased the sentence (page 7, lines 20-24): "Signposting researchers to international scientific communities in which PPI is an established and valued practice (also considering other descriptors which are used internationally to describe an approach that is similar to PPI in spirit, such as 'community engagement or 'patient-focused drug development' could stress the importance of PPI and increase motivation for researchers to acquire adequate knowledge and skills for PPI." Comment 7, page 7: We added detail to this statement (page 6, lines 21-22): ## "we have encountered scepticism towards the usefulness and impact of ppi among lbg researchers and austrian ethics committees." Comment 8, page 7: Thank you for the suggestion. Indeed, these are very relevant considerations also in our institutes, whereby the discussions we are having are differentiated, e.g., some PPI contributors value donating their time and would prefer not to receive financial remuneration. For the reviewer's information, at one of our institutes we have advertised a salaried PPI contributor ("patient researcher") position and found that the purpose and background to this role is challenging to convey, partly, undoubtedly, because such a position is rather unusual in the Austrian research landscape; but also because the particular contribution and added value of a researcher with lived experience to an institute such as ours takes some explaining of background, history and epistemology of PPI. The latter could be provided in training events, to create some equity in power by sharing knowledge about the value and purpose of PPI contributors' work. We It is very encouraging to see the development of PPI in research in Austria and for this to be undertaken in a strategic way informed by surveying researchers and ethical committee members. The points for attention in the discussion state clearly the priorities for PPI informed by the survey findings and research evidence on PPI. Thank you for this encouraging comment. The manuscript reporting is difficult to follow. This compromises the robust reporting required for publication of a communication. Thank you for the comment. We restructured the manuscript as follows: first, reporting the surveys, second, the anecdotal evidence (see Scoping section), and last, the strategies for promoting of PPI in Austria (see Discussion section). Although this is a communication, greater clarity is needed on what you did, why and how. We thank the reviewer and appreciate this comment, which is also mirrored by comments from Reviewers #1 and #3. We restructured the manuscript: first, reporting the surveys, second, the anecdotal evidence (see Scoping section, page 5ff), and last, the strategies for promoting PPI in Austria (see Discussion section, page 6ff). We have added an explicit statement of the problem and the survey aims at the end of the Background section (manuscript page 5, lines 3-14): We have provided more detail on the survey distribution and content of questionnaires (page 5, lines 18-38). The abstract is difficult to follow to understand the why PPI is important, the methods used with two surveys and the results should report the participants and the findings. BMJ Open recommend a structured abstract style in communication pieces. Using the structured abstract format would enable clearer presentation of the salient points of the importance of PPI in research, the aim of the study, methods and results, and the conclusions informed by the results detailing the key messages such as state the useful examples to introduce PPI -what are the strategic priorities. We apologise for providing an unstructured abstract and have re-written the abstract in a structured format, also taking into account comments from Reviewer #1. This PPI strategic work is undertaken with the Ludwig Boltzmann Institute (LBG) e.g. sampling their researchers. What is unclear is how this survey work of researchers and ethical committees links with the LBI stakeholder consultation in 2019 to develop PPI 'How to Guide'. We appreciate the reviewer's comment and realise that in our first manuscript version information about the 'How to Guide' was introduced after the surveys, when in fact this work had preceded the surveys. We have now moved this passage to the Background and hope that this provides a better description of the sequence of events; and how the activities conducted by the LBG OIS Center to foster PPI in Austria dovetailed with the practical experiences of researchers at the two Ludwig Boltzmann Institutes for Digital Health, which in turn sparked the surveys and led to further work with ethics committees. This laid foundation for investment in research on PPI for researchers to apply for grants to support implementation. This initiative is detailed towards the end of the discussion. To set the context of PPI this detail would be better reported in the background, then aim of this work. In the discussion helpful to consider the organizational/national level initiatives needed to drive change. We thank the reviewer and have moved the description of the PPI grant scheme to the Background. We have added a new statement of the aims of this work at the end of the Background section. We have kept the discussion of initiatives at organisational/national levels in the Discussion section. The background could be reduced to salient points re what is PPI, what is the importance for research and the situation in Austria to set context for the importance and relevance of the communication aim and findings reported. We have restructured the Background to present some of the details which were previously given in the Discussion section. We feel that the passages on PPI governance in the UK and on international developments in PPI (FDA and EUPATI) provide important context, as they situate PPI on an international landscape of research and signal to (Austrian) readers that PPI is in line with the international medical scientific community's direction of travel. In our national discourse this offers a helpful, sometimes very impactful argument. We therefore hope that the reviewer will agree that these points offer meaningful content to the Background. Greater detail is needed on how you scoped PPI practices for example the survey design -from the findings series of closed questions with a ranked scale, but key words state qualitative research. Use of open questions in the survey is not detailed. Also important to reference the underpinning evidence that informed the focus of the questions e.g. national guidance INVOLVE guidance on standards for PPI, systematic reviews on impact of PPI. Sampling and participants -detail is very limited. Thank you for this comment. We added details on the study design, sampling, and participants in the surveys (page 5, lines 18-38) and underpinning evidence that informed the focus of the questions. We apologise for including the keyword "qualitative research", which we realise could be misleading. There were some qualitative findings in our surveys from open questions, and we highlight the issue of researchers confusing/conflating PPI conversations with qualitative research, so the keyword was meant to refer to these two aspects; but we realise that it is better removed. Limitations of the work -Response from the ethical committee members very low 8/32, clearly this weakens the findings especially to understand that ethical committee members indicating limited experience, yet researchers indicating seek advise from ethical committee members. Clearly a priority is to engage ethics committee members to develop the strategic priorities e.g. as members of the national network on PPI Thank you for the comment. We agree on the argument and the limitation of the second survey. This was the reason why we asked the ethics committees to continue as a working group exploring ethical aspects in PPI in research. We expanded our previous acknowledgement of limitations in the Discussion section (page 7, lines 8-9) and highlighted the importance of engaging ethics committee members in developing strategic priorities for PPI (page 9, lines 7-10 and lines 15-18). You make important conclusions about how this work mapping PPI practices in research in the LBG community in Austria has led to wider discussions. Important to state clearly in the abstract and background that this work intended to map (or scope both terms are used in the manuscript) PPI in research in Austria to inform what ? We appreciate the reviewer's comment (which is also echoed by the other reviewers) regarding the lack of a statement of aims in the abstract and background, which we have now added: "We therefore undertook a scoping exercise and conducted surveys among researchers and research ethics committees in Austria. The aim was to explore current PPI practices, experiences, and ethical and operational challenges with PPI, to gauge in how far our personal experiences might be reflected within our wider research community; and to draw insights which may inform strategies for supporting PPI in research in Austria going forward." For consistency in wording, we have removed "map"/"mapping" from the manuscript and only use "scope". In the conclusion -state the main strategic priorities that detailed in the discussion. What are the main messages from the priorities detailed. This would strength the message in the conclusion about the priorities, rather than detailing the aim of the study. Thank you for this comment. We rewrote the conclusion specifically highlighting strategic priorities for PPI in Austria and the view going forward, rather than re-stating the aim of our surveys (page 9, lines 5-18). Reviewer: 3 Dr. Jacqueline Barker, University of the West of England Bristol Comments to the Author: This is an interesting paper. I think the following needs to be addressed: 1. There is no clear research question, therefore it is hard to judge the methods. We thank the reviewer for raising this point, which we have addressed by formulating a statement of the problem as it presented itself to us, and the aims of our online surveys. This passage is provided at the end of the Background section (page 5, lines 5-14): ## "…initial experiences made by researchers at these two institutes (stk, ek, mkp, es and ash among others), however, have surfaced challenges in implementing ppi practices, including lack of awareness and knowledge about the ppi concept in the local scientific communities; lack of appreciation of the value of involving patients as 'experts by experience'; and fear of violating research ethics if ppi activities are carried out without formal ethical approval. we therefore undertook a scoping exercise and conducted surveys among researchers and research ethics committees in austria. the aim was to explore current ppi practices, experiences, and ethical and operational challenges with ppi, to gauge in how far our personal experiences might be reflected within our wider research community; and to draw insights which may inform strategies for supporting ppi in research in austria going forward." We hope that through restructuring the manuscript as advised by all 3 reviewers, this passage contributes to a more apparent and coherent thread through the manuscript, linking problem statement, aims, methods, results, discussion, and conclusion. 2. The methods are not completely described (eg how was the survey administered, how were the questions developed etc). Thank you for this comment. We added further detail on the survey design, sampling, and distribution (page 5, lines 18-38). 3. In research about PPI, it is accepted practice to have involved the public. It looks as if the public were not involved in this paper so it is important to say why. Thank you for this comment, which has also been raised by Reviewer #1. We admit that there is some irony in not including any PPI in a manuscript about PPI. We hope that the manuscript does convey our good faith intentions to realise a genuine PPI approach in our research community, and that it provides some useful real-world insights into the types of difficulties we have encountered. Input and feedback from members of the public would undoubtedly be valuable; however, in this first scoping step we did not include members of the public in the study design or procedure. Our perception was that the immediate barriers we encountered in our work related to lack of awareness and understanding among researchers, and we therefore intended to scope current PPI practices among this group. For the reviewer's information, members of the public have been involved in the design and concept of the PPIE programme and funding model in 2019. We rephrased the sentence (page 6, lines 37-39): "Members of the public were not involved in the design and conduct of the surveys, because the immediate barriers to PPI we encountered in our work seemed to relate to awareness, knowledge, and perceptions among researchers." ## The uk organisation involve has been replaced -see this link: https://www.nihr.ac.uk/news/nihr-launches-new-centre-for-engagement-and-dissemination/24576 so some of your wording may need to be updated. Thank you for this information. We updated the wording in this passage (keeping the reference to INVOLVE, since the INVOLVE webpage remains active and appears to have been updated recently in 2021), and we added a reference to the NIHR website "Involve patients" which includes a link to NIHR Evidence (page 4, line 18). 5. You seem to use the term 'levels' in 2 different ways: 1) to describe levels of PPI from consultation to control and 2) in your abstract and elsewhere in terms of levels of organisation e.g. the individual, the organisation, the governance structures. You may need to clarify and flesh out these two different concepts and how they interact in your paper. We appreciate this point and thank the reviewer. For clarity, we have revised the use of 'levels' in the manuscript, so that 'levels' is now only used to indicate levels of organisation; and we use the term 'degree' to describe the concept of 'how much' involvement is afforded to patients and members of the public (no involvementconsultationcollaborationcontrol, as outlined in the discussion and with reference to Oliver et al. 2008). ## Proofing point: there is one use of 'und' instead of 'and' Thank you for noticing this mistake, we corrected it. 7. In the background section, you mention the reasons for PPI but only list 2 items. I think you are missing arguments about citizens' rights Thank you for highlighting this, and we have added this point to the opening paragraph (page 4, lines 9-10): ## "…;and a political dimension, based on citizens' rights and proposed advantages of alliances between researchers, patients and the public [2]." 8. Towards the end of your paper you use the term 'co production' which has specific meaning and for many people is not interchangeable with 'involvement'. I think you need to either stick to involvement or reflect the debate around involvement and co-production in your paper. Thank you for this comment. We are aware and appreciate the debate around involvement and coproduction, and we realise that our use of "co-production" did not consider this discourse. Upon reflection, we feel that this discussion is somewhat outside the remit of our manuscript, and to avoid misinterpretation we decided to use "involvement" throughout the manuscript. ## Version 2 -review ## Reviewer Di Lorito, Claudio University of Nottingham, Division of Rehabilitation, Ageing and Wellbeing ## Review returned 19-Mar-2021 ## General comments Thank you for addressing the comments previously provided and for improving sensibly the quality of the manuscript. I only have the following two points that I would like to see addressed before I am happy to recommend publication of your manuscript: 1. The body of the manuscript should follow the structure of the Abstract. At the moment, it is difficult to understand in the text which section corresponds to which section of the abstract. Please use the same heading of the abstract (i.e. introduction, methods, results, discussion, conclusion) in the text; 2. I am sorry but I do not quite understand your current justification as to why you did not use PPI to design the survey: "Members of the public were not involved in the design and conduct of the surveys, because the immediate barriers to PPI we encountered in our work seemed to relate to awareness, knowledge, and perceptions among researchers." Could you please rephrase to clarify? It would be worth mentioning in the paper that "members of the public have been involved in the design and concept of the PPIE programme and funding model in 2019" as this is evidence of good practice. Looking forward to seeing these comments addressed. Good Luck. ## Reviewer ## Evans, catherine King's College London REVIEW RETURNED 06-Apr-2021 ## General comments Thank you for revising this paper following detailed comments from the three peer-reviewers. The paper is improved, but could be strengthened to convey the methods of data analysis, detail on the checklist in the discussion and greater clarity in the abstract to report the aim and the results. Abstract: This is clearer with the structured format. The aim stated in the main paper is very clear. Please state in the abstract we aimed toand give detail as stated in the main paper. Background and aim -Background two sentences. We aimed to ….. Detail on response to difficulties could be removed or reduced, to increase word count to report the methods. This detail can be given in the main paper background to set the context for the work. Detail is needed in the methods about data analysis of the survey findings, and how you 'summarise five key challenges' from personal experiences. What is the relationship between the survey findings and your personal experiences? The objective is stated as to gauge the extent your personal experiences were echoed (or not) by other researchers in Austria. The abstract results need to report against thiswere your experiences echoed in the survey? Or not as the surveyed researchers rarely involved PPI. The detail on the results in the main paper could be better reflected in the abstract. For example, researchers experience of PPI was limited to dissemination of research. You want to present key results in the abstract that speak to the conclusions drawn. The conclusions need to relate to the results reported. The conclusions state development of checklist for ethical committees. Which findings inform this? The main paper reporting on this is clearera checklist for ethical aspects of PPI. In the abstract better to state the findings inform checklist for ethical aspects of PPI. ## Main manuscript Methodsdetail is needed on methods of data analysis. Stating for example, descriptive statistical analysis and collation of textual responses to open questions. Throughout: Please carefully proof read the manuscript to correct typos and grammatical errors. Best to avoid use of semi-colons or use minimally. Abstract results line 20 clearer to use a full stop and new sentence starting However, Throughout the manuscript please review use of semi-colon. In most instances a comma, or full stop and new sentence would be clearer. ## Dear authors Thank you for addressing the comments previously provided and for improving sensibly the quality of the manuscript. We thank the reviewer for the positive feedback on our revision. I only have the following two points that I would like to see addressed before I am happy to recommend publication of your manuscript: 1. The body of the manuscript should follow the structure of the Abstract. At the moment, it is difficult to understand in the text which section corresponds to which section of the abstract. Please use the same heading of the abstract (i.e. introduction, methods, results, discussion, conclusion) in the text; We thank the reviewer and have amended the headings in the abstract and main text, so they correspond. (We also took into account the further comments regarding the abstract by Reviewer #2.) The headings used in both abstract and main text are now: Background, Methods, Results, Discussion, Conclusion. 2. I am sorry but I do not quite understand your current justification as to why you did not use PPI to design the survey: "Members of the public were not involved in the design and conduct of the surveys, because the immediate barriers to PPI we encountered in our work seemed to relate to awareness, knowledge, and perceptions among researchers." Could you please rephrase to clarify? Apologies for not being more explicit in the description of the reason why we did not use PPI in the design of the surveys. We would like to request that the text in the manuscript may remain unchanged, as we thought very carefully about this passage to provide a wording that is truthful (we did not use PPI to design the surveys) but also appropriately sensitive and diplomatic to the readership in our community. (For the reviewer's information, we might re-iterate that in our attempts to implement PPI we encountered barriers related to awareness, knowledge, and perceptions among researchers; and one might consider a scenario in which senior medical scientists who are unfamiliar with PPI forbid researchers to engage with patients unless this has been formally approved by a research ethics committee.) It would be worth mentioning in the paper that "members of the public have been involved in the design and concept of the PPIE programme and funding model in 2019" as this is evidence of good practice. ## Amended Looking forward to seeing these comments addressed. Good Luck. We thank the reviewer. Reviewer: 3 Dr. Jacqueline Barker, University of the West of England Bristol Comments to the Author: Congratulations on this major rewrite. Your paper is now very clear on all the points raised last time. We thank the reviewer for the positive feedback. I think there is only 1 very small revision now required: please can you briefly describe the ethical permission you received for your research (sorry if it's there and I missed it) and/or what you did about informed consent and withdrawal from the study. We thank the reviewer for raising this point. We have completed the section on research ethics approval in the ScholarOne online system, indicating that "This study involves human participants but was not approved by an Ethics Committee(s) or Institutional Board(s)" and specifying the reason why formal ethics approval was not obtained: "In Austria, online surveys which are considered low-risk are commonly conducted without formal review and approval by a research ethics committee. Our surveys were considered low-risk, due to the target group (scientists and research ethics committee representatives, who may be assumed to be familiar with processes and governance in research), and we therefore did not formally submit these survey projects for ethical review. We followed ethical research practice as outlined in the Declaration of Helsinki (i.e. voluntary participation; assurance of anonymity, data protection and confidentiality; advance information of purpose and content; provision of contact details of the research team; and full disclosure of involved organisations). This information was summarised in the survey invitation email and described in the opening pages of the online questionnaires." and: "In line with common practice in online survey research, completion of the online questionnaire was understood as implied consent for participation. We followed standard practice for ethical research. The online surveys included opening pages which provided participants with the relevant information about the project (content and purpose of the questionnaires, voluntary participation, assurance of no disadvantage in case of non-participation, assurance of anonymity and data protection, dissemination plans, contact details of researchers and involved organisations). Participants' answers were registered only after the final questionnaire page had been completed. If a participant exited the questionnaire before completing the final page, the entries they had made up to that point were not stored by the online survey platform. Participants were informed that, should they wish to withdraw their data after having completed the questionnaire, they should contact us and we would do our best to accommodate this, but that this may be difficult due to the anonymised dataset." We have added one sentence to the Methods section: "In the design and conduct of the surveys we followed standard ethical research guidelines." We would ask the Editor for guidance whether additional detail should be provided within the main manuscript. Reviewer: 2 ## Dr. catherine evans, king's college london Comments to the Author: Thank you for revising this paper following detailed comments from the three peer-reviewers. The paper is improved, but could be strengthened to convey the methods of data analysis, detail on the checklist in the discussion and greater clarity in the abstract to report the aim and the results. We thank the reviewer for the positive feedback on our revision and for their further suggestions to improve the manuscript. We think the reviewer is referring to the 3 keywords listed on the cover page of the PDF proof ("ETHICS (see Medical Ethics), STATISTICS & RESEARCH METHODS, MEDICAL ETHICS"). These are keywords required by the ScholarOne online system, and a minimum of 3 need to be selected from a given list which is not equivalent to the MeSH catalogue. It is not possible to enter other keywords and most of these terms specify medical conditions or interventions. There is no keyword "Survey" or Questionnaires", but we have replaced "Medical Ethics" with "Qualitative Research", which could fit the aspect of qualitative (free text) responses in our survey. Please note that in the main manuscript, we have also defined our own 6 keywords: "Citizen Science, Community Engagement, Open Innovation in Science, Participation, Patient and Public Involvement, Service User Involvement", and we have double-checked these terms against the PubMed MeSH catalogue. We have subsequently amended this list of keywords to: - We thank the reviewer for this prompt. Abstract: This is clearer with the structured format. The aim stated in the main paper is very clear. Please state in the abstract we aimed toand give detail as stated in the main paper. Background and aim -Background two sentences. We aimed to ….. Detail on response to difficulties could be removed or reduced, to increase word count to report the methods. This detail can be given in the main paper background to set the context for the work. ## Amended Detail is needed in the methods about data analysis of the survey findings, and how you 'summarise five key challenges' from personal experiences. What is the relationship between the survey findings and your personal experiences? ## Amended The objective is stated as to gauge the extent your personal experiences were echoed (or not) by other researchers in Austria. The abstract results need to report against thiswere your experiences echoed in the survey? Or not as the surveyed researchers rarely involved PPI. The detail on the results in the main paper could be better reflected in the abstract. For example, researchers experience of PPI was limited to dissemination of research. You want to present key results in the abstract that speak to the conclusions drawn. We have revisited the results section of the abstract to address these points: With the revised aim/objective statement, the aspect of personal experiences echoed in the survey is now removed from the abstract. The key findings have been re-worded, to highlight the five current key challenges. (In the previous round of review, the point has been made that dissemination activities are not involvement, so we would prefer to avoid the wording that "PPI was limited to dissemination of research", as this could be understood to imply that dissemination activities do constitute PPI.) The conclusions need to relate to the results reported. The conclusions state development of checklist for ethical committees. Which findings inform this? The main paper reporting on this is clearera checklist for ethical aspects of PPI. In the abstract better to state the findings inform checklist for ethical aspects of PPI. ## Amended We thank the reviewer for these comments and hope that we have successfully incorporated all suggestions in the revised abstract. ## Main manuscript Methodsdetail is needed on methods of data analysis. Stating for example, descriptive statistical analysis and collation of textual responses to open questions. We have added the following passage to the Methods section: "We conducted descriptive statistical analyses for quantitative survey data and collated textual responses to open questions. Using survey data to contextualise our personal experiences, we articulated five current challenges to implementing PPI practices in Austria." Throughout: Please carefully proof read the manuscript to correct typos and grammatical errors. We have proof-read and corrected throughout the manuscript. Best to avoid use of semi-colons or use minimally. Abstract results line 20 clearer to use a full stop and new sentence starting However, ## Amended Throughout the manuscript please review use of semi-colon. In most instances a comma, or full stop and new sentence would be clearer. We thank the reviewer and have reduced the use of semi-colon throughout the manuscript. Research ethical committees, members of research ethical committees is correct not ethics committee, which implies belonging to. Apologies, but we are unsure whether the reviewer is suggesting that we should use the wording of "research ethic*al* committee" rather than "research ethic*s* committee"? To our knowledge the common wording is "research ethic*s* committee", e.g., on the HRA website, in the WHO International Ethical Guidelines for Health-Related Research Involving Humans, and on the BMJ Open ScholarOne online system. We would be more than happy to follow editorial guidance. Line 36should state vision ## Amended We thank all reviewers for their further helpful comments. ## Version 3 -review ## Reviewer Di Lorito, Claudio University of Nottingham, Division of Rehabilitation, Ageing and Wellbeing REVIEW RETURNED 29-Jun-2021 ## General comments Thank you for revising the manuscript and for addressing the remaining comments I provided. In relation to your statement as to why you did not have PPI input, I think that the explanation provided for me is much more understandable. Could you please reword into: "In our attempts to implement PPI, we encountered barriers related to awareness, knowledge, and perceptions among researchers". I would suggest adding the very illustrative example; "one might consider a scenario in which senior medical scientists who are unfamiliar with PPI forbid researchers to engage with patients unless this has been formally approved by a research ethics committee". But I would leave the decision to this last sentence to you.

The AFF4 scaffold binds human P-TEFb adjacent to HIV Tat Human positive transcription elongation factor b (P-TEFb) phosphorylates RNA polymerase II and regulatory proteins to trigger elongation of many gene transcripts. The HIV-1 Tat protein selectively recruits P-TEFb as part of a super elongation complex (SEC) organized on a flexible AFF1 or AFF4 scaffold. To understand this specificity and determine if scaffold binding alters P-TEFb conformation, we determined the structure of a tripartite complex containing the recognition regions of P-TEFb and AFF4. AFF4 meanders over the surface of the P-TEFb cyclin T1 (CycT1) subunit but makes no stable contacts with the CDK9 kinase subunit. Interface mutations reduced CycT1 binding and AFF4-dependent transcription. AFF4 is positioned to make unexpected direct contacts with HIV Tat, and Tat enhances P-TEFb affinity for AFF4. These studies define the mechanism of scaffold recognition by P-TEFb and reveal an unanticipated intersubunit pocket on the AFF4 SEC that potentially represents a target for therapeutic intervention against HIV/AIDS. # Introduction At many genes in humans-including the integrated HIV genome as well as loci that regulate development and mediate responses to stress-RNA polymerase II initiates transcription but forms a stable paused complex after the synthesis of 30-50 nucleotides [bib_ref] Dynamic transcriptional events in embryonic stem cells mediated by the super elongation..., Lin [/bib_ref] [bib_ref] Paused RNA polymerase II as a developmental checkpoint, Levine [/bib_ref] [bib_ref] The super elongation complex family of RNA polymerase II elongation factors: gene..., Luo [/bib_ref] [bib_ref] RNA polymerase II elongation control, Zhou [/bib_ref]. These paused polymerases are poised for rapid, synchronous efficient transcription. For these genes, escape of the paused polymerase from the promoter-proximal region and elongation of the mRNA are rate-limiting regulated processes. Promoter escape requires positive transcription elongation factor b (P-TEFb), a heterodimeric protein kinase composed of CDK9 and cyclin T1 (CycT1) subunits. P-TEFb triggers promoter escape by directly or indirectly stimulating phosphorylation of the RNA polymerase II C-terminal domain and the associated factors, NELF (negative elongation factor) and DSIF (DRB sensitivity inducing factor). Consequently, recruitment of active P-TEFb to the paused polymerase complex serves as an important checkpoint for gene expression [bib_ref] Paused RNA polymerase II as a developmental checkpoint, Levine [/bib_ref] [bib_ref] The super elongation complex (SEC) family in transcriptional control, Luo [/bib_ref] [bib_ref] RNA polymerase II elongation control, Zhou [/bib_ref]. P-TEFb cycles between inactive and active complexes [bib_ref] The Yin and Yang of P-TEFb regulation: implications for human immunodeficiency virus..., Zhou [/bib_ref]. Recent studies of gene fusions in myeloid leukemias [bib_ref] A higher-order complex containing AF4 and ENL family proteins with P-TEFb facilitates..., Yokoyama [/bib_ref] , as well as complexes recruited to the HIV promoter by the HIV Tat protein [bib_ref] HIV-1 Tat and host AFF4 recruit two transcription elongation factors into a..., He [/bib_ref] [bib_ref] HIV-1 Tat assembles a multifunctional transcription elongation complex and stably associates with..., Sobhian [/bib_ref] [bib_ref] Global landscape of HIV-human protein complexes, Jager [/bib_ref] , uncovered a family of Super Elongation Complexes (SECs) that bring together active P-TEFb and other transcription elongation factors. The SECs act at many normal human genes to stimulate mRNA elongation not only by triggering promoter escape but also by limiting proteolytic degradation of transcription elongation factors and increasing the processivity of RNAP II [bib_ref] HIV-1 Tat and host AFF4 recruit two transcription elongation factors into a..., He [/bib_ref] [bib_ref] AFF4, a component of the ELL/P-TEFb elongation complex and a shared subunit..., Lin [/bib_ref] [bib_ref] Function of leukemogenic mixed lineage leukemia 1 (MLL) fusion proteins through distinct..., Biswas [/bib_ref] [bib_ref] The ubiquitin ligase Siah1 controls ELL2 stability and formation of super elongation..., Liu [/bib_ref]. The SECs also couple to the PAF complex, which stimulates efficient transcript elongation [bib_ref] Human polymerase-associated factor complex (PAFc) connects the super elongation complex (SEC) to..., He [/bib_ref]. In addition to P-TEFb, the SECs contain subunits in the AF4, ELL, and ENL/AF9 protein families. Despite the major roles of SECs in metazoan gene expression and human disease, little is known about the architecture of these complexes. To define the interactions that mediate SEC assembly and to understand how HIV Tat recruits P-TEFb [bib_ref] A novel CDK9-associated C-type cyclin interacts directly with HIV-1 Tat and mediates..., Wei [/bib_ref] in the context of these large protein complexes, we mapped contacts among SEC subunits [fig_ref] Figure supplement 1: Expression levels of AFF4 variants [/fig_ref] [bib_ref] HIV-1 Tat recruits transcription elongation factors dispersed along a flexible AFF4 scaffold, Chou [/bib_ref]. Here, we report the structural and functional analysis of P-TEFb in complex with the cognate binding site on the SEC scaffold protein, AFF4. These studies reveal that AFF4 recognizes P-TEFb by binding the CycT1 subunit on the side opposite from CDK9. AFF4 is positioned to make direct contacts with HIV-1 Tat. Tat increases the affinity of P-TEFb for AFF4 by over an order of magnitude in vitro and rescues P-TEFb binding of AFF4 interface mutants in vivo. These results suggest that the SEC scaffold is an unanticipated direct partner of HIV-1 Tat, and an intersubunit Tat-binding pocket in the AFF4-P-TEFb complex may afford an unexpected site to target with selective inhibitors of HIV transcription. # Results The AF4 proteins, AFF1-4, form intrinsically disordered scaffolds that bind other transcription elongation factors not through folded domains but rather through dispersed short binding sites in the first 750 amino acids [bib_ref] HIV-1 Tat recruits transcription elongation factors dispersed along a flexible AFF4 scaffold, Chou [/bib_ref] [bib_ref] Leukemia fusion target AF9 is an intrinsically disordered transcriptional regulator that recruits..., Leach [/bib_ref] [fig_ref] Figure supplement 1: Expression levels of AFF4 variants [/fig_ref]. Residues 2-73 of AFF4, for example, are sufficient to bind P-TEFb through the CycT1 subunit, and a peptide encompassing AFF4 2-73 folds upon binding CycT1. Using this binding site, we determined the 2.9-Å-resolution crystal structure of AFF4 2-73 -P-TEFb-AMPPNP (R/R free = 0.207/0.245; [fig_ref] Figure supplement 1: Expression levels of AFF4 variants [/fig_ref] , [fig_ref] Figure supplement 1: Expression levels of AFF4 variants [/fig_ref] -figure supplements 1 and 2, and . In all three independent copies of the complex in the crystals, AFF4 residues Leu34-Ile66 are ordered, binding to the second cyclin domain of CycT1 opposite CDK9 [fig_ref] Figure 2 -: figure supplement 1B [/fig_ref]. In one complex, an isolated helix containing AFF4 residues 3-21 bridges symmetry-related CDK9 molecules in the crystals [fig_ref] Figure 2 -: figure supplement 1B [/fig_ref]. We focus on the shared features of the three independent complexes. Consistent with the coupling of folding to binding, AFF4 lacks intramolecular tertiary contacts as it snakes across the CycT1 surface. An extended segment in AFF4 tracks eLife digest The rates at which many genes are expressed as proteins are limited by the efficiency of a process called transcriptional elongation. This process takes place as the stretch of DNA that defines the gene is transcribed into an RNA molecule and it is catalyzed by an enzyme called RNA polymerase II. However, this enzyme can become trapped, and another enzyme called P-TEFb (positive transcription elongation factor b) is needed to release it. P-TEFb and other elongation factors therefore have an important role in gene expression. The human immunodeficiency virus (HIV) is a retrovirus that hijacks the gene expression processes in human immune cells to replicate the RNA genome of the virus. To do this, the virus produces a protein called Tat that recruits P-TEFb as part of a multi-protein machine called the super elongation complex. This ensures that the process of transcriptional elongation, and hence the overall replication process, is highly efficient. There are gaps, however, in our knowledge of the architecture of the super elongation complex, which is known to be organized on a flexible scaffold. In turn, the molecular basis for the interaction between HIV-1 Tat and P-TEFb within the super elongation complex is not well understood. Now Schulze-Gahmen et al. show that only one of the two subunits in P-TEFb-a cyclin known as CycT1-binds to the AFF4 scaffold protein in the super elongation complex. In addition to assisting with the expression of hundreds of human genes, super elongation complexes containing P-TEFb-AFF4 are hijacked in various forms of cancer and viral infections, including HIV/AIDS. Schulze-Gahmen et al. show that AFF4 can directly contact HIV-1 Tat, which binds to the P-TEFb-AFF4 complex much more strongly than it binds to P-TEFb alone. This suggests that HIV-1 Tat evolved to work within the super elongation complex. Moreover, Schulze-Gahmen et al. reveal that HIV-1 Tat binds to a cleft between the P-TEFb enzyme and the AFF4 protein, which raises the possibility that this cleft could be used as a target for anti-HIV/AIDS drugs. antiparallel to the H3′-H4′ loop in CycT1, and two short AFF4 helices fit into shallow grooves in CycT1 between H3′ and H5′ and the surface formed by H2′ and H3′ [fig_ref] Figure 2 -: figure supplement 1B [/fig_ref]. Compared to the separated models, 1457 Å 2 of AFF4 and 1251 Å 2 of CycT1 accessible surface are buried in the complex. Twentysix of the 33 ordered residues in AFF4 contact CycT1, emphasizing the extensive recognition determinants in the scaffold. Many hydrophobic and aromatic residues in CycT1 mediate contacts with AFF4 [fig_ref] Figure 2 -: figure supplement 1B [/fig_ref] -C and [fig_ref] Figure 2 -: figure supplement 1B [/fig_ref]. A pocket between CycT1 Leu163-Val164-Arg165, Trp221, and Tyr224, for example, anchors AFF4 Phe35, which forms a classic edge-to-face interaction [bib_ref] Aromatic-aromatic interaction: a mechanism of protein structure stabilization, Burley [/bib_ref] with the Tyr [fig_ref] Figure 2 -: figure supplement 1B [/fig_ref]. CycT1 Trp210 adjusts to create a site that buries AFF4 Pro38 and allows a short antiparallel β-sheet to form between AFF4 Tyr39-Val41 and CycT1 Asn 209-Glu211 [fig_ref] Figure 2 -: figure supplement 1B [/fig_ref]. CycT1 Trp207 shifts to form a hydrogen bond with the Gly57 carbonyl in the loop between the two AFF4 helices, additional van der Waals contacts with the loop backbone and nonpolar contacts with AFF4 residues Leu56 and Tyr59 In comparison to the structure of P-TEFb alone (CDK9 1-330 /CycT1 1-259 ; PDB ID 3BLQ; [bib_ref] The structure of P-TEFb (CDK9/cyclin T1), its complex with flavopiridol and regulation..., Baumli [/bib_ref] , the last 20 residues present in the CycT1 subunit undergo a major rearrangement upon AFF4 binding [fig_ref] Figure 2 -: figure supplement 1B [/fig_ref]. AFF4 overlaps with the position of residues 243-259 in the superimposed P-TEFb cyclin subunit. To accommodate AFF4, helix H5′ straightens and residues 251-256 form a helical segment that packs against CycT1 helices H1 and H2' and forms part of the AFF4 binding surface. In addition to burying hydrophobic residues in the first AFF4 helix, the adjustments in CycT1 helix H5′ mediate formation of a hydrogen-bonded ion pair between AFF4 Arg51 and CycT1 Glu246. CycT1 Trp256, the last residue ordered in the AFF4 complex, moves over 13 Å upon AFF4 binding. To probe the basis for CycT1 binding, we measured the effects of interface mutations on the stability of the AFF4 complex in vitro. The AFF4 2-363 and 33-67 fragments bound to CycT1 with similar affinities [fig_ref] Table 2: Binding affinities of AFF4 segments [/fig_ref] , suggesting that the ordered AFF4 segment in the crystal structure captures the major CycT1 binding determinants. Mutations throughout the CycT1 interaction surface reduced AFF4 binding [fig_ref] Figure 3: AFF4 interface mediates P-TEFb recognition [/fig_ref] and [fig_ref] Table 3: Dissociation constants of AFF4 32-67 for Cyclin T1 mutants [/fig_ref]. The Trp210Ala and Trp207Ala substitutions in CycT1 had the largest effects, respectively, reducing AFF4 affinity by 21-and 58-fold. These results point to these CycT1 Trp residues as interaction hotspots and suggest that contacts all along the interface observed in the crystal structure mediate AFF4 recognition. To examine the functional roles of the N-terminal 72 residues of AFF4, we measured the effects of tandem alanine mutations on the stimulation of expression of a luciferase reporter gene driven by the HIV-1 promoter in HeLa cells [bib_ref] Human polymerase-associated factor complex (PAFc) connects the super elongation complex (SEC) to..., He [/bib_ref] [fig_ref] Figure 3: AFF4 interface mediates P-TEFb recognition [/fig_ref] and [fig_ref] Figure 3: AFF4 interface mediates P-TEFb recognition [/fig_ref] -figure supplement 1). Consistent with the structure, mutations of AFF4 residues that contact P-TEFb reduced Tat-independent transcriptional stimulation. Three of the most deleterious variations-alanine substitutions at Pro33/ Leu34, Phe35 (Ala36 was maintained), and Met55/Leu56-shorten hydrophobic side chains that are buried in the interface. Tandem alanine substitutions (Glu45/Asp46 and Gly57/Asn58) that remove side chains that cap and stabilize the AFF4 helices also reduced luciferase expression. In contrast, residues such as Glu61/Met62 and Ile71/Pro72, which are more tolerant of di-alanine substitutions, are more solvent exposed or flexible. Tandem alanine replacements in residues 3-32, which flank the ordered CycT1 contacts in the crystal structure, showed significant but generally smaller effects on AFF4 transcriptional stimulation activity [fig_ref] Figure 3: AFF4 interface mediates P-TEFb recognition [/fig_ref]. These results implicate the P-TEFb binding site, as well as the flanking flexible sequences, in the function of the AFF4 N-terminal segment. The HIV-1 Tat protein was shown nearly 15 years ago to bridge P-TEFb and the TAR RNA element near the 5′ end of nascent HIV transcripts to recruit the active CDK9 kinase to the HIV promoter [bib_ref] A novel CDK9-associated C-type cyclin interacts directly with HIV-1 Tat and mediates..., Wei [/bib_ref]. The discovery that Tat recruits P-TEFb as part of a larger SEC [bib_ref] HIV-1 Tat and host AFF4 recruit two transcription elongation factors into a..., He [/bib_ref] [bib_ref] HIV-1 Tat assembles a multifunctional transcription elongation complex and stably associates with..., Sobhian [/bib_ref] raised the question of how Tat distinguishes the SEC from free P-TEFb, particularly since . X-ray data collection and refinement statistics for AFF4-P-TEFb-AMPPNP Tat binds P-TEFb in isolation. Moreover, Tat shows specificity for SECs containing AFF4 and AFF1 [bib_ref] HIV-1 Tat and host AFF4 recruit two transcription elongation factors into a..., He [/bib_ref] [bib_ref] HIV-1 Tat assembles a multifunctional transcription elongation complex and stably associates with..., Sobhian [/bib_ref]. What accounts for this specificity? The AFF4-P-TEFb crystal structure provides a simple and unsuspected explanation-Tat binds in a position to make direct contacts with the scaffold [fig_ref] Figure 4 -: figure supplement 1 [/fig_ref]. Superposition of the CycT1 subunits of the structures of Tat-P-TEFb [bib_ref] Crystal structure of HIV-1 Tat complexed with human P-TEFb, Tahirov [/bib_ref] and AFF4-P-TEFb shows that Tat is positioned to pack against helix 2 of AFF4. Tat Met1, Lys28 (which is reversibly acetylated in vivo; [bib_ref] HIV-1 tat transcriptional activity is regulated by acetylation, Kiernan [/bib_ref] [bib_ref] The control of HIV transcription: keeping RNA polymerase II on track, Ott [/bib_ref] and Phe32, in particular, are predicted to interact with AFF4 Glu61, Met62, Phe65, and Ile66. The disordered C-terminus of the AFF4 peptide also neighbors Glu2 and His13-Gly15 of Tat. These seven residues of Tat are crucial for transcriptional activation, even though they are exposed to solvent in the Tat-P-TEFb complex [bib_ref] Transition step during assembly of HIV Tat:P-TEFb transcription complexes and transfer to..., D&apos;orso [/bib_ref]. The juxtaposition of AFF4 and Tat on the P-TEFb surface predicts that the scaffold enhances Tat binding. Direct measurements of Tat affinity are problematic, however, because Tat is unstable in vitro in the absence of partners and difficult to maintain in an active form. To overcome this problem, we took advantage of a thermodynamic cycle that illustrates that AFF4 and Tat mutually influence the P-TEFb affinity of each other by the same amount [fig_ref] Figure 4 -: figure supplement 1 [/fig_ref]. In quantitative in vitro assays [fig_ref] Figure 4 -: figure supplement 1 [/fig_ref] and [fig_ref] Table 2: Binding affinities of AFF4 segments [/fig_ref] , the purified AFF4 2-73 peptide bound Tat-P-TEFb (K d = 0.85 ± 0.15 nM) ∼11 times more tightly than P-TEFb (K d = 10 ± 1 nM). The AFF4 2-363 segment also bound Tat-P-TEFb (K d = 0.6 ± 0.1 nM) ∼11 times more tightly than P-TEFb (K d = 7 ± 1 nM). These results support the model of direct Tat interactions with AFF4 and suggest that the AFF4 2-73 peptide includes the principal residues that contact Tat [fig_ref] Figure 4 -: figure supplement 1 [/fig_ref] and To test the importance of the AFF4 interactions for P-TEFb and Tat recognition in vivo, we measured the effects of mutations in AFF4 on the binding of SEC subunits and Tat in HeLa cells. Cells were transfected with wild-type or mutant AFF4 containing a C-terminal 3× FLAG tag, and immunoprecipitations using an anti-FLAG antibody were probed for the presence of associated factors. Alanine substitutions in P-TEFb contacts such as AFF4 Phe35, Tyr59/Asp60, and Lys63/Asp64 reduced binding of P-TEFb but not the ELL2, AF9, or ENL subunits of the SEC [fig_ref] Figure 4 -: figure supplement 1 [/fig_ref]. These defects were rescued by overexpressing a stably integrated gene for Tat, which strengthened the AFF4-P-TEFb association. In contrast, tandem alanine substitutions for AFF4 Glu61/Met62, which are more exposed in the P-TEFb interface and predicted to contact Tat, caused a small reduction (∼30%) in the binding of P-TEFb that was not rescued by overexpressing Tat [fig_ref] Figure 4 -: figure supplement 1 [/fig_ref]. The specificity of these mutational effects supports the model that Tat interacts directly with AFF4. In keeping with the tenuous contacts of the AFF4 3-21 helix with CDK9, AFF4 2-73 (K d = 10 ± 1 nM) binds only 3.6 times more tightly than AFF4 32-67 (K d = 36 ± 4 nM) to P-TEFb [fig_ref] Table 2: Binding affinities of AFF4 segments [/fig_ref]. In addition, the CDK9 kinase subunit structure is little changed upon AFF4 binding to P-TEFb (CDK9 15-360 RMSD = 0.54 Å vs 1BLQ). In vivo, the Glu13Ala/Arg14Ala AFF4 mutant in the heart of the CDK9 interface in the crystals is associated with a <25% decrease in transcription stimulation [fig_ref] Figure 3: AFF4 interface mediates P-TEFb recognition [/fig_ref] and little change in P-TEFb binding [fig_ref] Figure 4 -: figure supplement 1 [/fig_ref]. To measure if recruitment of P-TEFb to the SEC scaffold regulates the kinase, we assayed the effects of AFF4 on the in vitro phosphorylation of an RNA polymerase II CTD substrate by purified P-TEFb and Tat-P-TEFb [bib_ref] The mixed-lineage leukemia fusion partner AF4 stimulates RNA polymerase II transcriptional elongation..., Bitoun [/bib_ref]. AFF4 2-73 inhibited CTD phosphorylation by approximately twofold in a purified system and [fig_ref] Figure 2 -: figure supplement 1B [/fig_ref]. By comparison, Tat stimulated P-TEFb by sevenfold, and addition of AFF 2-73 did not further stimulate the kinase activity of the Tat-P-TEFb complex. Taken together, these results point away from CDK9 as the primary physiological partner of AFF4 3-21 and suggest that AFF4 functions as a SEC scaffold but not an allosteric activator of CDK9. # Discussion The AFF4-P-TEFb crystal structure reveals that a high density of contacts in residues 34-66 of the AFF4 scaffold mediates binding to the CycT1 subunit of P-TEFb. These contacts are largely conserved in AF4 family members and can be perturbed physically and functionally by mutations [fig_ref] Figure 3: AFF4 interface mediates P-TEFb recognition [/fig_ref] and [fig_ref] Figure 2 -: figure supplement 1B [/fig_ref]. Consistent with these results, tandem alanine substitutions of Pro33/Leu34, Val41/Thr42, Arg51/Ile52, and Met55/Leu56 in the CycT1 binding site (but not Arg3/Glu4 or Glu25/ Asp26 in the preceding segment) also reduce P-TEFb binding in vivo [bib_ref] HIV-1 Tat recruits transcription elongation factors dispersed along a flexible AFF4 scaffold, Chou [/bib_ref]. The impacts of alanine substitutions on transcriptional stimulation by AFF4 [fig_ref] Figure 3: AFF4 interface mediates P-TEFb recognition [/fig_ref] show that residues in the P-TEFb interface, as well as helix stabilizing residues, play crucial roles. The additional sensitivity of transcriptional stimulation to alanine substitutions of disordered residues flanking the CycT1 binding site suggests that the flexibility and potential interactions with other ligands are important for function. HIV-1 Tat binds adjacent to AFF4, increases the affinity of AFF4 for P-TEFb by over an order of magnitude, and rescues P-TEFb binding to AFF4 interface mutants in vivo. The AFF4 Glu61/Met62 double alanine substitution in the proposed Tat interface blocks this rescue. These results suggest that direct contacts between Tat and the AFF4 scaffold in complex with P-TEFb mediate the selective recruitment of the SEC to the HIV promoter. The functional AFF4-Tat interface, including the acetylated Lys28 in Tat as a potential regulator of affinity [bib_ref] HIV-1 tat transcriptional activity is regulated by acetylation, Kiernan [/bib_ref] , supports the idea that Tat evolved to function within the SEC. In the crystal structure of the (low-affinity) Tat-P-TEFb subcomplex [bib_ref] Crystal structure of HIV-1 Tat complexed with human P-TEFb, Tahirov [/bib_ref] , Tat binds to a relatively open groove. In the context of the AFF4-P-TEFb structure, however, AFF4 creates an unanticipated pocket for Tat [fig_ref] Figure 4 -: figure supplement 1 [/fig_ref]. This pocket may ultimately provide a suitable therapeutic target for the development of small-molecule inhibitors of Tat binding that selectively block HIV transcription. # Materials and methods Expression of P-TEFb, P-TEFb-Tat, and AFF4 CDK9 (1-330) and cyclin T1 (1-264) were cloned into a modified pFastBac Dual donor plasmid, and HIV-1 Tat (1-86) was cloned into the pFastBac1 donor plasmid using the Bac-To-Bac system from Life Technologies (Carlsbad, CA). CDK9 was cloned with a Tobacco Etch Virus (TEV) protease cleavable N-terminal His-tag while CycT1 and Tat remained un-tagged. Each virus genome was transfected into Sf9 cells to generate the baculovirus according to manufacturer's protocol. Each baculovirus was amplified (2×), plaque purified, and amplified (3×) to obtain a stock with 10 8 plaque forming units (PFU)/ml. Test infections were screened for expression levels of the target proteins by Western blot analysis, and the highest expressing virus stock was used. For large-scale production of P-TEFb, 4 l of High5 cells at 1 × 10 6 cells/ml in ESF921 medium (Expression Systems, Reno, NV) were infected with 20 ml of virus stock per liter of culture. The flasks were incubated for 52 hr at 27°C on a rotary shaker. Cells were harvested by centrifugation for 30 min at 350xg in a Beckmann JLA-8.1 rotor, washed quickly in 50 mM Tris pH 7.5, 150 mM NaCl, and centrifuged at 350×g for 10 min. The supernatant was removed, and the cell pellets were frozen in liquid nitrogen. For large-scale production of P-TEFb-Tat, insect cells were coinfected with 20 ml each of CDK9, CycT1, and Tat virus stocks. The remaining steps were the same as for production of P-TEFb. AFF4 2-73 was cloned into a modified pET28 plasmid. The recombinant protein includes a N-terminal TEV-protease-cleavable His-tag. The plasmid was transformed into Rosetta(DE3)pLysS. 3 l of transformed Escherichia coli were grown at 37°C to OD 600 = 0.6, and expression was induced with 0.3 mM IPTG at 18°C for 16 hr. Cells were harvested, and pellets were frozen in liquid nitrogen. ## Purification of the aff4-p-tefb complex Pellets from 4 l infected High5 cells were resuspended in 75 ml lysis buffer (20 mM Na-HEPES pH 7.4, 10 mM NaCl, 1 mM DTT) with 1 × Roche Complete Protease Inhibitor and 0.5 mM AEBSF. Cells were lysed with a Dounce homogenizer. The lysate was brought up to 0.2 M NaCl by adding 3.0 ml 5 M NaCl, incubated on ice for 10 min, and centrifuged at 5800xg in a SS34 rotor. The supernatant was saved, and the pellet extracted again with 30 ml lysis buffer in the homogenizer. The supernatants of the two centrifugations were combined, cleared by centrifugation at 210,000×g in an ultracentrifuge, and filtered through a 0.8-μm syringe filter. The cleared lysate was loaded onto a 5 ml His-Trap HP column (GE Healthcare, Piscataway, NJ) equilibrated in buffer A (20 mM Na-HEPES pH 7.4, 0.3 M NaCl, 10% glycerol, 1 mM DTT, 20 mM imidazole). After washing for 10 column volumes with buffer A + 1 M NaCl, followed by 10 column volumes of buffer A, P-TEFb was eluted with a gradient from 0% A to 100% B (20 mM Na-HEPES pH 7.4, 0.3 M NaCl, 10% glycerol, 0.5 M imidazole, 1 mM DTT). The eluted P-TEFb was dialyzed for 3 hr against 2 L buffer A. TEV-protease was added to the protein at a 1:25 (wt/wt) ratio, and the digest was incubated for 1 hr at room temperature and 4°C overnight. The digest was loaded on a His-Trap HP column, and P-TEFb lacking the His tag eluted in the flow-through of the column, while undigested protein and TEV-protease eluted later in the imidazole gradient. The yield was ∼0.8 mg P-TEFb/L High5 cell culture. The Tat-P-TEFb complex was purified in a similar way. The cell lysate was purified over a 5 ml His-Trap HP column, dialyzed, and digested with TEV protease as described above. The digested complex was diluted with 1.1 volumes of 20 mM Na-HEPES pH 7.3, 1 mM DTT, 1% β-octyl-glucoside to a final concentration of 0.14 M NaCl, and 0.5% β-octyl-glucoside and applied to a Source S anion exchange column equilibrated in 20 mM Na-HEPES pH 7.3, 10% glycerol, and 1 mM DTT. The column was developed with a linear gradient to 20 mM Na-HEPES pH 7.3, 0.75 M NaCl, 10% glycerol, and 1 mM DTT. P-TEFb-Tat eluted as single peak at about 0.25 M NaCl. To purify AFF4 2-73 , 30 g of E. coli cell pellet was resuspended in 125 ml lysis buffer (25 mM Tris/HCl pH 7.5, 0.2 M NaCl, and 1 mM DTT). Lysozyme was added to 0.1 mg/ml final concentration and incubated for 30 min. After adding Roche Complete Protease Inhibitor without EDTA, 0.5 mM AEBSF, and -P-TEFb complex using the cyclin subunit (yellow) shows the close proximity of AFF4 (blue) and Tat (red). Tat Lys28 (pink), where acetylation stimulates function, as well as other residues essential for Tat transcriptional activation [bib_ref] Transition step during assembly of HIV Tat:P-TEFb transcription complexes and transfer to..., D&apos;orso [/bib_ref] that are exposed to solvent in the Tat-P-TEFb complex (bright red) are positioned adjacent to AFF4. (B) Tat enhances AFF4 binding in vitro. Fluorescence polarization of fluorescein-labeled AFF4 32-67 (5 nM) is plotted as a function of the concentration of CycT1 (blue circles), P-TEFb (red squares), and Tat-P-TEFb (green triangles). (C) Alanine substitutions in the P-TEFb binding site of AFF4 reduce CycT1 binding but not associations of other SEC subunits in HeLa cells. Western blots show associations of each indicated factor with different FLAGtagged AFF4 variants (top). Lysates were immunoprecipitated with an anti-FLAG antibody. Expression of Tat (right) rescues defects in CycT1 binding, except for the E61/M62 double alanine mutant. This mutant in the predicted AFF4-Tat interface shows equal small defects in P-TEFb binding in the absence (left) and presence (right) of Tat. (D) AFF4 (blue) and CycT1 (yellow) create an intersubunit pocket where Tat (red) can bind with minor structural adjustments. The program DoGSiteScorer [bib_ref] DoGSiteScorer: a web server for automatic binding site prediction, analysis and druggability..., Volkamer [/bib_ref] assigns this cleft a high druggability score (0.83 out of 0-1.0) and shows that it contains the most nonpolar surface of any pocket in the AFF4-P-TEFb structure. DOI: 10.7554/eLife.00327.014 The following figure supplements are available for figure 4: DNaseI (5 units/ml), the cells were lysed by sonication. The lysate was centrifuged for 1 hr at 17,000 rpm in a SS34 rotor. The supernatant was filtered through a 0.8-μm syringe filter and applied to a 5 ml His-Trap column. The protein was purified as described for P-TEFb. The separately purified P-TEFb and AFF4 2-73 were combined at a 1:1.4 (mol/mol) ratio, concentrated to 0.6 ml, and injected onto an analytical Superdex S200 gel exclusion column equilibrated with 25 mM Na-HEPES pH 7.4, 0.2 M NaCl, and 1 mM DTT. The center fractions of the eluted three-protein peak were used for crystallization. ## Crystallization and structure determination The purified AFF4-P-TEFb complex was concentrated to 10 mg/ml using Millipore Ultrafree centrifugal devices. Crystals were grown from 1.0 µl protein combined with 0.5 µl silver bullet condition 30 (Hampton Research, Aliso Viejo, CA; 0.33% wt/vol Gly-Phe, 0.33% wt/vol Gly-Tyr, and 0.33% wt/vol Leu-Gly-Gly in 20 mM Na-HEPES pH 6.8) and 0.5 µl reservoir solution equilibrated against 16-18% PEG 3350, 0.1 M Na-HEPES pH 7.0. After equilibrating for 24 hr, diluted microseeds from previous crystallization experiments were added with a hair. Seeding produced large single crystals (0.3 × 0.2 × 0.2 mm) with and without 1 mM AMPPNP, 5 mM MgCl 2 . The presence of all three proteins in the crystals was confirmed by gel electrophoresis and mass spectrometry of dissolved crystals. Crystals were soaked in cryoprotectant (25% PEG 3350, 0.1 M Na-HEPES pH 7.0, 10% glycerol, 1:4 silver bullet 30) and flash frozen in liquid nitrogen. X-ray data were collected at Beamline 8.3.1 at the Advanced Light Source at the Lawrence Berkeley National Laboratory [bib_ref] Suite of three protein crystallography beamlines with single superconducting bend magnet as..., Macdowell [/bib_ref]. The best data were collected from a crystal that was grown in the presence of 1 mM AMPPNP and 5 mM MgCl 2 . The reflections were processed using HKL2000 [bib_ref] Processing of X-ray diffraction data collected in oscillation mode, Otwinowski [/bib_ref]. The structure was determined by molecular replacement with PHENIX [bib_ref] PHENIX: a comprehensive Python-based system for macromolecular structure solution, Adams [/bib_ref] using P-TEFb from the P-TEFb-Tat complex (PDB ID 3MI9) as the search model. The asymmetric unit contains three complexes. Initial refinement using PHENIX was performed with model restraints, as well as noncrystallographic symmetry restraints. Model restraints were removed in later stages of refinement. AFF4 was built manually and the model was adjusted using Coot [bib_ref] Coot: model-building tools for molecular graphics, Emsley [/bib_ref]. The model was refined using gradient minimization with weight optimization and maximumlikelihood targets, TLS refinement, and individual atomic B-factor refinement. The model was checked against composite omit maps. Density was missing for residues 1-7 and 88-95 in CDK9 mol1 and mol3, and residues 1-7 and 89-97 in CDK9 mol2. Density also was absent for residues 1-7 and 253-264 in CycT1 mol1, residues 1-7 and 257-264 in CycT1 mol2 and mol3, and residues 2-33 and 67-73 in AFF4, mol1 and mol3. For AFF4 in molecule 2 residues 2, 22-33, and 67-73 are missing. The register of the AFF4 sequence in the electron density was confirmed by locating the Se atoms in crystals grown with SeMet-labeled AFF4. AFF4 2−73 was labeled with SeMet [bib_ref] Atomic structures of the human immunophilin FKBP-12 complexes with FK506 and rapamycin, Van Duyne [/bib_ref] and purified as described above. Optimized crystallization conditions were seeded with unlabeled microcrystals. Anomalous differences were calculated from data collected at 12,657 eV, the peak of Se fluorescence measured from the crystal. # Structure analysis Buried surface areas were calculated with PISA [bib_ref] Inference of macromolecular assemblies from crystalline state, Krissinel [/bib_ref]. Surface pockets were identified with the program DoGSiteScorer [bib_ref] DoGSiteScorer: a web server for automatic binding site prediction, analysis and druggability..., Volkamer [/bib_ref]. Figures of molecular structures were prepared with PyMOL Version 1.5 (Schroedinger LLC, New York, NY). ## Aff4 affinity for cyct1 Protein binding was measured using fluorescence anisotropy of a 36-residue segment of AFF4 (residues 32-67) encompassing the protein-protein contacts in the crystal structure. The AFF4 peptide was synthesized at the University of Utah DNA/Peptide Facility using the following sequence: C-FAM-GABA-SPLFAEPYKVTSKEDKLSSRIQSMLGNYDEMKDFIG-amide where FAM indicates 5-carboxyfluoroscein and GABA indicates a γ-amino-butyric acid spacer. Varying amounts of purified CycT1, P-TEFb, and Tat-P-TEFb were incubated for 30 min with 5 nM labeled peptide at room temperature in the dark in 25 mM HEPES pH 7.5, 100 mM NaCl, 10% glycerol, 0.1% NP40, and 0.5 mM Tris(2-carboxyethyl)phosphine (TCEP). Competition titration experiments of unlabeled peptides were performed using 75 nM CycT1, 45 nM P-TEFb, and 6 nM Tat-P-TEFb in 25 mM HEPES pH 7.5, 100 mM NaCl, 10% glycerol, 0.1% NP40, 0.5 mM TCEP, and 5 nM fluorescent peptide. Fluorescence anisotropy was measured using a Victor 3V (Perkin Elmer) multi-label plate reader. Data points represent the average of six independent measurements. Binding curves were fit to the single-site binding equation using Prism version 5.0c (Graphpad Software). ## Aff4 stimulation of tat-independent transcription HeLa cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% FBS at 37°C in a humidified atmosphere with 5% CO 2 . Cells were seeded at 5 × 10 5 cells/ml in TC-treated 96-well plates one day prior to plasmid transfection using the 25-kDa linear polyethyleneimine reagent (Sigma-Aldrich, St. Louis, MO). Cells were cotransfected with 100 ng of an HIV-LTR firefly luciferase reporter construct [bib_ref] HIV-1 Tat and host AFF4 recruit two transcription elongation factors into a..., He [/bib_ref] and 350 ng of pCDNA3.1 containing AFF4 variants with a C-terminal 3× FLAG tag. Following stimulation for 48 hr with the indicated ligands, the cells were lysed in passive lysis buffer (Promega, Fitchburg, WI) for 5 min at 25°C. The cell lysates were incubated with firefly luciferase substrate, and luminescence was measured on a SpectraMax L microplate reader (Molecular Devices, Sunnyvale, CA). The relative luminescence was normalized to the concentration of AFF4 in the cell determined by Western blotting using an anti-FLAG primary antibody. ## Co-immunoprecipitation and detection of proteins bound to aff4 HeLa cells were seeded in 10-cm TC-treated plates at 2.2 × 10 5 cells/ml, Incubated for 24 hr and transfected with pCDNA3.1 (1 μg) encoding an AFF4 variant with a C-terminal 3× FLAG tag. After incubating for 2 days, cells were collected in PBS, washed, and resuspended in hypotonic buffer (20 mM Tris-HCl pH 7.4, 10 mM NaCl, and 3 mM MgCl 2 ). After 15 min on ice, 0.4% Triton X-100 was added, and the cell suspension was mixed and centrifuged at 3000×g for 10 min. The pellet was resuspended in nuclear extraction buffer (100 mM Tris-HCl pH 7.4, 100 mM NaCl, 1% Triton X-100, 1 mM EDTA, 10% glycerol, 0.7% Tween 20, and protease inhibitors [AEBSF, leupeptin and E64]) for 30 min on ice followed by centrifugation at 14,000×g for 30 min to yield nuclear extract. Anti-FLAG agarose beads (Sigma-Aldrich) were incubated in the nuclear extract at 4°C for 2 hr and washed with nuclear extraction buffer. The beads were eluted with 0.1 M glycine-HCl pH 3.5, and the neutralized eluate was analyzed by Western blotting with the indicated antibodies. ## P-tefb kinase assays Kinase assays were performed in LoBind tubes (Eppendorf) in 20 μl reactions containing 50 mM HEPES pH 7.3, 50 mM NaCl, 1 mM DTT, 10 mM MgCl 2 , and 0.05 mM or 0.5 mM ATP. For assays with 0.05 mM ATP, 0.15 μCi [ 32 P]-γ-ATP was used, and for reactions with 0.5 mM ATP, 1.5 μCi [ 32 P]-γ-ATP was used. Reactions in conventional tubes gave distinct less reproducible results. Purified recombinant P-TEFb or P-TEFb-Tat was pre-incubated with 500 ng purified recombinant GST-CTD (52 C-terminal domain repeats from human RNA polymerase) in the absence or presence of 0.3 µM purified AFF4 2-73 for 15 min at 20°C. After addition of ATP, the kinase reactions were stopped at different times by addition of 5 μl of 5× SDS sample buffer. The samples were analyzed by SDS-polyacrylamide gel electrophoresis, followed by measurement of the radioactive protein bands on a Typhoon phosphorimager (GE Healthcare). For Western blots of kinase reactions performed with non-radioactive ATP, 0.2 µg of GST-CTD from each reaction was loaded per lane of a 4-20% SDS-PAGE gel, transferred onto a PVDF membrane (Immobilon-FL; Millipore, Billerica, MA) and processed by standard Western Blot procedures. The primary antibodies were ab5095 (αpS2; Abcam) and ab5131 (αpS5; Abcam) at 1:1000 dilution for both. The secondary antibody was a fluorescently labeled goat anti-rabbit antibody (Odyssey; LI-COR Biosciences, Lincoln, NE) at 1:20,000 dilution. ## Additional files ## Major datasets The following dataset was generated: [fig] Figure 2 -: figure supplement 1B). Eight intermolecular hydrogen bonds help establish the chemical complementarity between AFF4 and CycT1. Twenty-one of the 26 interacting residues in AFF4 are conserved in AFF1-3 (Figure 2-figure supplement 2). [/fig] [fig] Figure 1, Figure supplement 1: AFF4 binds CycT1 distal to CDK9. (A) Schematic model of the SEC. AFF4 is an intrinsically disordered scaffold that binds partners via 20-50 residue segments. (B) Ribbon diagram showing the strand-helix-helix arrangement of AFF4 (blue) bound to CycT1 (yellow) remote from CDK9 (teal). AFF4 adopts an extended conformation with no intramolecular tertiary contacts. AMPPNP (spheres) is bound to CDK9. DOI: 10.7554/eLife.00327.003 The following figure supplements are available for figure 1: Electron density for AFF4 2-73 . DOI: 10.7554/eLife.00327.004 Figure supplement 2. A crystal contact formed by AFF4 2-21 . DOI: 10.7554/eLife.00327.005 [/fig] [fig] Figure 2 008, Figure supplement 2: Basis for AFF4 scaffold recognition by P-TEFb. (A) AFF4 residues 34-66 (blue spheres) fill grooves on CycT1 (yellow surface). (B) Chemical complementarity mediates AFF4 binding. Exposed hydrophobic residues of CycT1 (yellow surface) are buried in the AFF4 complex. Hydrogen bonds (black dotted lines) also mediate binding. (C) AFF4 Phe35 is buried in a hydrophobic pocket formed by aromatic and nonpolar residues on the surface of CycT1. (D) The C-terminus of the CycT1 cyclin domain (gray in the ellipse) adjusts to make contacts with AFF4 (blue). DOI: 10.7554/eLife.00327.007 The following figure supplements are available for figure 2: Figure supplement 1. Example interactions between AFF4 and P-TEFb. DOI: 10.7554/eLife.00327.Conserved AFF4 sequences mediate P-TEFb recognition. DOI: 10.7554/eLife.00327.009 [/fig] [fig] Figure 4 -: figure supplement 1). [/fig] [fig] Figure 3: AFF4 interface mediates P-TEFb recognition. (A) Mutations of CycT1 contact residues reduce AFF4 affinity. Fluorescence polarization of fluorescein-labeled AFF4 32-67 (5 nM) is plotted as a function of the concentration of the indicated CycT1 variant. (B) Transcriptional effects of AFF4 tandem Ala mutants. Stimulation of Tatindependent transcription from the HIV LTR was measured in extracts of cells cotransfected with a luciferase reporter construct and an expression vector for the indicated di-Ala AFF4 variant. Activity was normalized to the level of AFF4 expression. Values represent the mean of three independent assays. Tandem alanine substitutions cover the first 72 residues of AFF4. Horizontal lines correspond to the mean stimulation of di-Ala substitutions in residues 3-32 (left; 113.9 ± 5.1) and 33-66 (right; 73.6 ± 4.9). DOI: 10.7554/eLife.00327.011The following figure supplements are available for figure 3: [/fig] [fig] Figure supplement 1: Expression levels of AFF4 variants. DOI: 10.7554/eLife.00327.012 [/fig] [fig] Figure 4: AFF4 binds in position to make direct contacts with HIV-1 Tat. (A) Superposition of the AFF4-P-TEFb complex and the [/fig] [fig] Figure 5 017, Figure supplement 2: Kinase activity of P-TEFb and P-TEFb-Tat complexes with AFF4. (A) Autoradiogram showing phosphorylation of GST-CTD (500 ng) by P-TEFb and P-TEFb-Tat with and without excess (0.28 μM) AFF4 2-73 in the presence of low (50 μM) ATP. (B) Phosphorylation of 500 ng GST-CTD by P-TEFb and P-TEFb-Tat with and without excess (0.28 μM) AFF4 2-73 in the presence of saturating (500 μM) ATP. AFF4 reduces the activity of P-TEFb twofold and has little influence on the kinase activity of Tat-P-TEFb. Tat-P-TEFb is, however, sevenfold to tenfold more active than P-TEFb. Lane 3 in panels (A and B) is a control without GST-CTD. (C) Quantitation of the radioactive GST-CTD in (A and B). DOI: 10.7554/eLife.00327.016 The following figure supplements are available for figure 5: Figure supplement 1. SDS polyacrylamide gel of P-TEFb and P-TEFb-Tat at the same ratio as they were used in the kinase assay. DOI: 10.7554/eLife.00327.Western blots of kinase reaction products from panel B. Phosphorylated CTD was detected with anti-phoshoSer2 and anti-phoshoSer5 antibodies. DOI: 10.7554/eLife.00327.018 [/fig] [table] Table 2: Binding affinities of AFF4 segments [/table] [table] Table 3: Dissociation constants of AFF4 32-67 for Cyclin T1 mutants [/table]

Early and late HIV-1 membrane fusion events are impaired by sphinganine lipidated peptides that target the fusion site Figure S1PBDK-sphing RP-HPLC profile post-purification PBDK-sphing was purified using RP-HPLC as described in the Materials and methods section of the main text. To validate purity, 20 μg of the purified peptide stock was run at an acetonitrile gradient of 10-90 % for 40 min on a C 4 column. All peaks were validated as the same compound by ESI analysis. The resulting peaks are most probably oligomers of the peptide since collecting and then re-injecting each peak results in the above spectrum. The reading was done at 215 nm with a reference of 650 nm. mAU, milli-absorption units.Figure S2 PKH67 staining on TZM-bl cells reveals a high staining efficacyIn order to assess PKH67 (excitation, 490 nm and emission, 504 nm) general staining efficacy stained (green) and unstained (red) TZM-bl cells were subjected to FACS analysis. A close to 100 % staining efficacy was observed for the PKH67 evident by the lack of unstained cells in the population that underwent staining.Figure S3 PKH67 staining does not affect HIV-1 infectivityHIV-1 virions were stained with PKH67 at the noted doses and subjected to a virus-cell infectivity assay with TZM-bl cells. At the evaluated doses no significant change of infectivity was observed. The 2 % (v/v) dose was eventually used for virus labelling. The results are normalized to the infectivity of the unstained virions. Results are means + − S.E.M. (n = 2). P > 0.05 for all columns in relation to unstained HIV. PBDK-sphing RP-HPLC profile post-purification PBDK-sphing was purified using RP-HPLC as described in the Materials and methods section of the main text. To validate purity, 20 μg of the purified peptide stock was run at an acetonitrile gradient of 10-90 % for 40 min on a C 4 column. All peaks were validated as the same compound by ESI analysis. The resulting peaks are most probably oligomers of the peptide since collecting and then re-injecting each peak results in the above spectrum. The reading was done at 215 nm with a reference of 650 nm. mAU, milli-absorption units. ## Figure s2 pkh67 staining on tzm-bl cells reveals a high staining efficacy In order to assess PKH67 (excitation, 490 nm and emission, 504 nm) general staining efficacy stained (green) and unstained (red) TZM-bl cells were subjected to FACS analysis. A close to 100 % staining efficacy was observed for the PKH67 evident by the lack of unstained cells in the population that underwent staining. [fig] Figure: S3 PKH67 staining does not affect HIV-1 infectivity HIV-1 virions were stained with PKH67 at the noted doses and subjected to a virus-cell infectivity assay with TZM-bl cells. At the evaluated doses no significant change of infectivity was observed. The 2 % (v/v) dose was eventually used for virus labelling. The results are normalized to the infectivity of the unstained virions. Results are means + − S.E.M. (n = 2). P > 0.05 for all columns in relation to unstained HIV. [/fig]

The tobacco genome sequence and its comparison with those of tomato and potato The allotetraploid plant Nicotiana tabacum (common tobacco) is a major crop species and a model organism, for which only very fragmented genomic sequences are currently available.Here we report high-quality draft genomes for three main tobacco varieties. These genomes show both the low divergence of tobacco from its ancestors and microsynteny with other Solanaceae species. We identify over 90,000 gene models and determine the ancestral origin of tobacco mosaic virus and potyvirus disease resistance in tobacco. We anticipate that the draft genomes will strengthen the use of N. tabacum as a versatile model organism for functional genomics and biotechnology applications. C ommon tobacco (Nicotiana tabacum) is one of the most widely cultivated non-food crops worldwide and is grown in B120 countries [bib_ref] Tobacco cultivation, Peedin [/bib_ref]. It belongs to the Nicotiana genus, which is named after Jean Nicot de Villemain who, in 1560, became the first person to import these plants from the Americas to Europe. The term Nicotiana was originally used by Adam Lonitzer to describe tobacco plants in and in 1788 by Carl von Linné (Linnaeus) to designate the entire genus. Over 75 naturally occurring Nicotiana species, including 49 native to America and 25 native to Australia 4 , have been classified by and Knapp [bib_ref] Nomenclatural changes and a new sectional classification in Nicotiana (Solanaceae), Knapp [/bib_ref]. Most commercial tobaccos cultivated today belong to the species Nicotiana tabacum L., for which 41,600 N. tabacum cultivated varieties (cultivars) are listed in the National Plant Germplasm System 7 . The three most commonly used tobacco types are Flue-Cured (or Virginia), Burley and Oriental, which are traditionally grown and harvested under different agricultural practices. Tobacco is a model plant organism for studying fundamental biological processes [bib_ref] Dynamic metabonomic responses of tobacco (Nicotiana tabacum) plants to salt stress, Zhang [/bib_ref] , and is the source of the BY-2 plant cell line, which is a key tool for plant molecular research. It is also used as a model for plant disease susceptibility, which it shares with other Solanaceae plants including potato, tomato and pepper. Diseases affecting tobacco include the tobacco mosaic virus (TMV), the tobacco vein mottling virus (TVMV), the tobacco etch virus (TEV), and the potato virus Y (PVY); the TN90 variety of tobacco, which we sequenced here, is notable in that is has been bred to resist these viral infections. Considerable interest has centred on understanding the origin, organization and evolution of the N. tabacum genome. Tobacco stands out as a complex allotetraploid with a large 4.5 Gb genome with significant proportion (470%) of repeats [bib_ref] DNA sequence organization in the genome of Nicotiana tabacum, Zimmerman [/bib_ref] [bib_ref] Next generation sequencing reveals genome downsizing in allotetraploid Nicotiana tabacum, predominantly through..., Renny-Byfield [/bib_ref]. As a species, N. tabacum (2n ¼ 4x ¼ 48) evolved through the interspecific hybridization of the ancestors of Nicotiana sylvestris (2n ¼ 24, maternal donor) and Nicotiana tomentosiformis (2n ¼ 24, paternal donor) about 200,000 years ago [bib_ref] The ups and downs of genome size evolution in polyploid species of..., Leitch [/bib_ref]. Because of its complexity and importance, the tobacco genome is a target for the SOL-100 sequencing project, which aims to decipher the genomes of the most important Solanaceae species. The genome sequences of modern varieties of ancestral species were recently reported [bib_ref] Reference genomes and transcriptomes of Nicotiana sylvestris and Nicotiana tomentosiformis, Sierro [/bib_ref] , and limited evidence suggests that Nicotiana otophora is an alternative paternal donor [bib_ref] Characterization of the Nicotiana tabacum L. genome by molecular cytogenetics, Kenton [/bib_ref] [bib_ref] Structure and expression of the gene family encoding putrescine N-methyltransferase in Nicotiana..., Riechers [/bib_ref]. In this report, however, we demonstrate that this is unlikely because of the higher sequence identity of the N. tabacum T-genome with that of N. tomentosiformis. We show the chromosomal rearrangements between the ancestral and tobacco chromosomes, and provide an explanation for an apparent genome reduction following the hybridization. In addition, we present a genomic comparison of tobacco to two other solanaceous species, tomato and potato. Significant chromosomal reshuffling is clearly observed for all chromosomes despite the conservation of their overall count, confirming previous reports [bib_ref] COSII genetic maps of two diploid Nicotiana species provide a detailed picture..., Wu [/bib_ref]. Tobacco's rich metabolism (involving 44,000 chemical components) and exceptional ability to express proteins (440% of its dry weight) have prompted numerous initiatives to harness its potential for the production of biologically active substances [bib_ref] Plant molecular farming: systems and products, Horn [/bib_ref]. Here, we describe the major alkaloid biosynthesis pathway in Nicotiana species, as well as glutamate/aspartate pathways in the three main tobacco types. In this work, we sequence the genomes of key representatives of the three major types of tobacco and combine them with genetic and physical maps of tobacco [bib_ref] A high density genetic map of tobacco (Nicotiana tabacum L.) obtained from..., Bindler [/bib_ref] [bib_ref] Whole genome profiling physical map and ancestral annotation of tobacco Hicks Broadleaf, Sierro [/bib_ref]. We verify genome assembly accuracy by mapping transcriptomics and Exon Array 22 data of corresponding varieties, we assess the consistency of assemblies and published physical and genetic maps, and we compare N. tabacum S-and T-genomes with those of N. sylvestris and N. tomentosiformis. # Results Sequencing and assembly. Genome assembly of polyploid species, such as coffee (Coffea arabica), potato (Solanum tuberosum) and wheat (Triticum aestivum) is challenging. Even the assembly of the relatively small Brassica napus genome (1.2 Gb), which has well-annotated ancestral reference sequences, is still ongoing [bib_ref] Dissecting the genome of the polyploid crop oilseed rape by transcriptome sequencing, Bancroft [/bib_ref]. We sequenced the genomes of three inbred varieties of the allotetraploid N. tabacum, K326 (Flue-cured), TN90 (Burley) and Basma Xanthi (BX, Oriental), using a whole-genome shotgun sequencing approach with 100 bp Illumina HiSeq-2000 pairedend and mate-pair reads. The size of the tobacco genome has been estimated to be 4.46 Gb by flow cytometry using the fluorochrome propidium iodide [bib_ref] First nuclear DNA amounts in more than 300 angiosperms, Zonneveld [/bib_ref] , each amount to 3.7 Gb, representing a coverage of 480% of the tobacco genome. The remaining B20% is likely to consist of repetitive regions that could not be resolved using the short read de novo shotgun approach. This represents a great improvement from the previous publicly available tobacco genome assembly (ftp://ftp.solgenomics.net/ tobacco_genome/assembly/), which contains 294,750 sequences (8.5% of the tobacco genome). We also sequenced libraries from N. otophora to investigate its contribution to N. tabacum. We evaluated the quality of the scaffolding by mapping the assembly scaffolds to long BAC sequences (altogether 1.8 Mb) obtained from Nicotiana tabacum Hicks Broadleaf. We did not observe any sequence inversions due to scaffolding, and identified only three cases where a contig was not incorporated in a scaffold and remained as singleton (Supplementary Data 1). We also identified 13 cases of insertions or deletions. Altogether this represents 8.6 structural differences per megabase, if we assume that the Hicks Broadleaf sequence is the same as the TN90, K326 and BX sequence for the selected BACs. The sequence-based Whole Genome Profiling (WGP) physical map was used to super-scaffold each genome assembly, and simple sequence repeats (SSRs) were mapped to super-scaffolded assemblies to anchor them to the 24 tobacco linkage groups. This resulted in 84% of the de novo assembly of TN90 being anchored to the WGP physical map (83 and 84% for K326 and BX, respectively), and 19% to the genetic map (17 and 16% for K326 and BX, respectively). [fig_ref] Figure 1 |: map linkage group b Sequence-based linkage group origin c WGP map-based origin... [/fig_ref] shows the composition of the N. tabacum TN90 genome anchored to the linkage groups of the tobacco genetic map. The three methods for assigning an ancestral origin to the linkage groups, SSR amplification in N. sylvestris and N. tomentosiformis 20 , sequence identity with donor species and using the WGP physical map 21 , are concordant. The latter two confirmed the previously reported colour inversion in linkage group , and the central part of [fig_ref] Figure 1 |: map linkage group b Sequence-based linkage group origin c WGP map-based origin... [/fig_ref] clearly shows the correspondence between regions of different ancestral origins within regions of the N. tabacum genome linked by sequence homology. By mapping SSR markers to the sequenced genomes of N. tabacum K326, TN90 and BX, we were able to predict SSR length differences between the three varieties. When comparing TN90 to K326, 57% of the predictions obtained exactly matched experimental measurements, and 10% had only 2-bp differences. The larger differences observed between experimental and in silico SSR measurements are likely to be caused by difficulties in resolving the SSR using short read assembly. We used sequence identity to assign an ancestral origin to each 2-Mb region. The only region for which N. otophora appears as the most likely ancestor is at the end of linkage group 14 [fig_ref] Figure 1 |: map linkage group b Sequence-based linkage group origin c WGP map-based origin... [/fig_ref] , indicating that if it contributed to the N. tabacum genome, then only marginally. This observation favours the hypothesis that the predominant paternal donor was N. tomentosiformis [bib_ref] The origin of tobacco's T genome is traced to a particular lineage..., Murad [/bib_ref]. We evaluated the completeness of our assemblies by mapping reference gene sequences to each genome using BLAT [bib_ref] BLAT-the BLAST-like alignment tool, Kent [/bib_ref] (Supplementary . For this we used NCBI and SGN tobacco Unigene constructs, as well as the SGN tobacco transcriptome [bib_ref] Deciphering the complex leaf transcriptome of the allotetraploid species Nicotiana tabacum: a..., Bombarely [/bib_ref]. We also used the coding sequences of tomato (ITAG v2.3) and potato (PGSC v3.4). We mapped 82-86% of the tobacco transcriptome; the remainder is likely to consist of genes spanning different genome sequences that have not yet been scaffolded. Between 50-59% of the Unigene constructs from NCBI and SGN could be mapped, presumably reflecting sequence diversity in the Unigene set [bib_ref] A draft genome sequence of Nicotiana benthamiana to enhance molecular plant-microbe biology..., Bombarely [/bib_ref]. Approximately 60% of the tomato and potato coding sequences could be mapped to the genomes, which compares favourably with the rate of transcript mapping to the Nicotiana benthamiana genome 31 (56.5 and 58.5% for tomato and potato, respectively). In 92.8% of the cases, the best possible alignment of a tomato protein overlaps with a tobacco gene model from our TN90 transcriptome (nine tissues). The genetic and physical map anchoring results as well as alignment to the N. sylvestris and N. tomentosiformis genomes confirmed the quality of the three N. tabacum genome assemblies. Synteny with other Solanaceae. The linkage groups from the genetic maps of both N. tomentosiformis and N. acuminata 18 , and the results of SSR amplification in N. tomentosiformis and N. sylvestris 20 , were used to show rearrangements that are found in tobacco (Supplementary Data 2 and 3). [fig_ref] Figure 2 |: Synteny between Nicotiana tabacum, Nicotiana tomentosiformis and Nicotiana acuminata genetic linkage groups... [/fig_ref] illustrates one such rearrangement, where part of linkage groups 4 and 8 of N. tomentosiformis are fused to give N. tabacum linkage groups 12 and 23, whereas N. acuminata linkage groups 4 and 8 did not fuse, and gave N. tabacum linkage groups 16 and 1, respectively. The synteny between the genomes of N. tabacum TN90, K326 and BX and those of tomato and potato was evaluated at the protein level by mapping tomato and potato proteins to tobacco sequences anchored to the linkage groups of the genetic map 20 to detect homologous genes (Supplementary Data 2 and 3), and by detecting further homologous DNA blocks in genomes masked for repeats (Supplementary Data 4-7). Similar results were obtained using MCScanX 32 , which is a toolkit specifically designed for the detection and analysis of gene synteny and collinearity (Supplementary Data 8 and 9). Not all the regions identified by either of these methods should be considered as truly syntenic, as some of them rely only on a limited number of anchors. Likewise, additional syntenic regions are likely to exist, which cannot be detected by the methods used here. Nevertheless, these approaches confirm results reported earlier based on COSII (ref. [bib_ref] COSII genetic maps of two diploid Nicotiana species provide a detailed picture..., Wu [/bib_ref] and SSR markers [bib_ref] A high density genetic map of tobacco (Nicotiana tabacum L.) obtained from..., Bindler [/bib_ref]. Repetitive elements in the genomes. The repeat content of the N. tabacum K326, TN90 and BX genomes is summarized in Supplementary . Between 72 and 79% of the sequenced genomes are reported as repeat elements by RepeatMasker. This estimation is lower than that reported for barley (84%) and close to the original estimate (B80%) by Zimmerman [bib_ref] DNA sequence organization in the genome of Nicotiana tabacum, Zimmerman [/bib_ref]. Based on the amount of the sequenced genome covered by repeats, we evaluated the DNA portion of the tobacco genome containing non-repeat coding regions to about 1 Gb. This is equivalent to the sum of the same DNA portion from the descendants of both ancestral genomes. The observed 4-8% reduction in genome size is thus likely to have occurred in the repetitive region of the genome. The sum of the genome sizes from the descendants of both ancestors is of 5.04 Gb, N. sylvestris accounting for 53% of it, and N. tomentosiformis for 47%. The tobacco scaffolds to which an origin could be assigned show 55-57% of S origin and 43-45% of T origin. The genome reduction thus is likely to have been more important in the T part of the genome than in the S part, which corresponds to what was reported by Renny-Byfield et al. [bib_ref] Next generation sequencing reveals genome downsizing in allotetraploid Nicotiana tabacum, predominantly through..., Renny-Byfield [/bib_ref] The reported assemblies of three tobacco varieties cover 480% of the genome, which is comparable to that for smaller diploid genomes (76-90%) [bib_ref] Genome sequence and analysis of the tuber crop potato, Xu [/bib_ref] [bib_ref] The tomato genome sequence provides insights into fleshy fruit evolution, Sato [/bib_ref] and for the smaller (3 Gb) allotetraploid N. benthamiana (81-87%) [bib_ref] Deciphering the complex leaf transcriptome of the allotetraploid species Nicotiana tabacum: a..., Bombarely [/bib_ref] [bib_ref] Advanced engineering of lipid metabolism in Nicotiana benthamiana using a draft genome..., Naim [/bib_ref]. They represent some of the largest assembled plant genomes together with barley (5.1 Gb) [bib_ref] A physical, genetic and functional sequence assembly of the barley genome, Mayer [/bib_ref] , Norway spruce (20 Gb) [bib_ref] The Norway spruce genome sequence and conifer genome evolution, Nystedt [/bib_ref] and the partial wheat genome [bib_ref] Analysis of the bread wheat genome using whole-genome shotgun sequencing, Brenchley [/bib_ref]. Contrary to other allotetraploid genomes, for which ancestral information is either not available (N. benthamiana) or limited (wheat), the ancestral origin of the tobacco sequences was identified and confirms previously reported assignments based on genetic markers 20 and the physical map [bib_ref] Whole genome profiling physical map and ancestral annotation of tobacco Hicks Broadleaf, Sierro [/bib_ref]. Tobacco root and leaf transcriptome analysis. For each variety, three biological replicates were obtained from roots and leaves, which are two metabolically highly active tobacco tissues. In addition, nine tissues were sampled for TN90, so as to get good coverage of the gene regions. For each RNA-Seq sample, 86.5-94.7% of reads were mapped to the genome of the corresponding variety (Supplementary . Using RNA-seq data from two tissue types, we generated gene models for each of the three varieties, the gene number estimate being 81,000 for TN90 (Supplementary . These numbers do not represent the whole transcriptome of tobacco, as we sampled RNA from only nine tissues, and hence will be missing transcripts not expressed in any of them. For the root and leaf transcriptomes, gene ontology (GO) terms could be assigned by InterProScan 38 to B40,000 proteins (from B28,000 genes) (Supplementary ; for the nine tissue samples, this number increases to 450,000, although there is notable increase in the number of unique GO terms assigned. A recent analysis of 454 N. sylvestris, N. tomentosiformis and N. tabacum next-generation sequencing transcriptomes showed neither differences in gene expression nor the creation of a new function for homoeologous genes between N. tabacum S-and T genomes [bib_ref] Deciphering the complex leaf transcriptome of the allotetraploid species Nicotiana tabacum: a..., Bombarely [/bib_ref]. Similarly, we found no new genes or new functionality of existing genes in the three varieties. We analysed the GO term enrichment for differentially expressed genes in each tissue. Photosynthesis-and biosynthetic-related genes were highly expressed and enriched in leaf tissue, as were genes involved in oxidation-reduction processes [fig_ref] Figure 2 |: Synteny between Nicotiana tabacum, Nicotiana tomentosiformis and Nicotiana acuminata genetic linkage groups... [/fig_ref]. Root tissue has a very distinctive profile of upregulated genes . Besides lipid transport and regulatory gene overexpression, lignin and cell wall metabolism genes were prominent. Moreover, defence and oxidative stress response genes were heavily upregulated, which may be related to cell death processes apparent from the expression profile. The overexpression of 'pollination' genes reflects the electronic annotation of several upregulated proteins within the PFAM domain PF00954, the socalled 'S locus glycoprotein-like' domain. While this protein family is best known for being involved in the pollination process, some members are involved in defence response regulation [bib_ref] S-glycoproteinlike protein regulates defense responses in Nicotiana plants against Ralstonia solanacearum, Maimbo [/bib_ref]. It is thus more likely that the proteins identified here are involved in defence, rather than in pollination. We used OrthoMCL to analyse the functional overlap between tobacco and three other, increasingly divergent plant species: N. benthamiana as a further representative of the Nicotiana genus, Solanum lycopersicum (tomato) as a further species in the Solanaceae family and Arabidopsis thaliana, another species in the eudicotyledons . The bulk of the protein clusters are shared between all dicotyledons (10,362), we observed 2,024 and 4,044 clusters specific to N. benthamiana and N. tabacum, respectively, and 2,206 Nicotiana-specific clusters shared between both species. We also observed 3,706 clusters shared between all Solanaceae but not with Arabidopsis. Classical tobacco pathways. We extracted the sequences of biochemical pathway enzymes and observed copy numbers and expression in roots and leaves under different growth conditions (Supplementary Note 1). No major differences in gene expression were observed, but one new putrescine N-methyltransferase gene was identified in the alkaloid pathway in addition to the four already reported [bib_ref] Putrescine N-methyltransferase-the start for alkaloids, Biastoff [/bib_ref]. Furthermore, the two quinolinate phosphoribosyltransferase (QPT) genes of S origin were not found in N. tabacum K326 (Supplementary . Most described alkaloid genes are expressed in roots where most alkaloids are synthesized; however, some transcripts from AO, QPT, QS and MPO genes were also detected in leaf, suggesting that they have different functions (Supplementary and Supplementary Note 1). In addition, we have mapped the putative steroidal alkaloid biosynthesis genes from Nicotiana genomes to the syntenic regions recently discovered on chromosomes 7 and 12 of S. lycopersicum 41 . In N. tabacum we identified two copies of each region, corresponding to their ancestral origin, indicating that these regions are conserved within Nicotiana species . For Burley tobacco, which is known for its high nitrogen requirement, nitrogen assimilation is stronger than for Flue-cured tobacco . We therefore compared the number and expression levels of genes related with the glutamate/ aspartate pathway and Supplementary Note 2). With the exception of one AAT5 isoform missing in K326 but present in BX and TN90, we observed neither CNVs nor major transcriptomic variations, thereby suggesting that the nitrogen assimilation at the level of the glutamate/aspartate pathway is close between Burley tobacco and Flue-cured tobacco. However, it is not enough to conclude that the efficiency of ammonium assimilation is similar in both tobaccos, other downstream gene products from root to leaf being involved in the pathway as well as adaptations to environmental conditions. The analyses of the complete set of genes related to amino-acid assimilation will certainly help to understand the effect of human artificial selection on the high nitrogen requirement of Burley tobacco. Disease resistance. TMV resistance was introduced in tobacco in the 1930s as a single dominant locus from an interspecific hybrid with Nicotiana glutinosa [bib_ref] Molecular and genetic characterization of Nicotiana glutinosa L. chromosome segments in tobacco..., Lewis [/bib_ref] [bib_ref] Inheritance of resistance to tobacco-mosaic disease in tobacco, Holmes [/bib_ref] [bib_ref] Analysis of the N gene of Nicotiana, Dunigan [/bib_ref]. This locus was shown to harbour the N gene (NGU15605) encoding a (TIR)-NBS-LRR protein [bib_ref] The product of the tobacco mosaic virus resistance gene N: similarity to..., Whitham [/bib_ref] that triggers a hypersensitive response following recognition of the viral helicase [bib_ref] The helicase domain of the TMV replicase proteins induces the N-mediated defence..., Erickson [/bib_ref]. Among the three varieties sequenced here, only TN90 is TMVresistant. TN90 has inherited the N gene from the variety Burley 21, which in turn inherited it from the variety Kentucky 56, which in turn inherited it from N. glutinosa hybrids [bib_ref] Development of Burley 21, the first wildfire-resistant tobacco variety, including results of..., Heggestad [/bib_ref] showed weak identity in K326 and BX genomes (B90% identity on o35% of N. glutinosa genomic DNA), whereas a nearly perfect match was found for TN90 (99.9% identity over 7,158 bp). PVY is a potyvirus, which is transmitted by aphids, and has a broad host range including potato and tomato. N. tabacum is naturally susceptible to PVY and other potyviruses such as TVMV and TEV; interestingly, the ancestors N. tomentosiformis and N. sylvestris are resistant and susceptible, respectively, to PVY. Most modern tobacco varieties bred for PVY resistance carry alleles of the Va locus. The recessive va allele, which was first obtained by deletionin the line TI 1406 (Virgin A Mutant) (Supplementary Note 3), confers resistance by preventing virus cell-to-cell movement in a similar way to the recessive resistances of pepper, potato and tomato, suggesting the involvement of a eukaryotic initiation factor [bib_ref] Complementary functions of two recessive R-genes determine resistance durability of tobacco 'Virgin..., Acosta-Leal [/bib_ref]. The characterization of the PVY resistance gene using a recombinant inbred line population was recently presented by Julio et al.The eukaryotic translation initiation factor eIF4E was shown to be strongly expressed in susceptible plants, but not in resistant plants. We carried out a sequence comparison of the genome of TN90, which is resistant to the potyviruses TVMV, TEV and PVY, and the genomes of K326 and BX, which are susceptible. We identified eIF4E1, eIF4E2 and eIF(iso)4E genes in the TN90, K326 and BX genome assemblies by first mapping corresponding tomato genes to N. sylvestris and N. tomentosiformis genomes, then mapping the respective genes to the tobacco genomes. One copy of each gene was found in the N. sylvestris genome, whereas two copies of eIF4E1 and one copy of the other genes were found in N. tomentosiformis. With the exception of the N. sylvestris eIF4E1 gene in TN90, all identified N. sylvestris and N. tomentosiformis genes were found in TN90, K326 and BX genomes. Gene-specific PCR verification showed that the absence of the N. sylvestris eIF4E1 in TN90 is not an artefact of genome assembly, but rather that it is missing from this variety. This confirms the observation that TVMV, TEV and PVY resistance in TN90 is caused by genomic deletion of the S-form eIF4E1 locus. This is in line with the pattern of resistance observed in tobacco ancestors, suggesting that the dominant sensitivity of N. tabacum to potyviruses via the eIF4E1 mechanism was conferred by N. sylvestris. The fact that both S and T copies of N. tabacum eIF4E1 are expressed concomitantly (Supplementary allows us to speculate that either the interaction between the N. tomentosiformis eIF4E1 copies and viral proteins are compromised, or that the viral-host protein complex cannot provide the cell-to-cell movement required for systemic infection. # Discussion Next-generation sequencing data were used to construct the genomes of three tobacco varieties, representing the three main market classes of commercially grown tobacco. Assembly accuracy was verified by mapping tobacco transcriptomics data, by assessing the consistency with tobacco physical and genetic maps, and through comparison of N. tabacum S-and T genomes to ancestral (N. sylvestris and N. tomentosiformis) genomes. The assemblies cover 480% of the genome, 75% being annotated as repeats. Sequencing and mapping of N. otophora in addition to mapping N. sylvestris and N. tomentosiformis genome sequences to the tobacco genomes confirmed that N. sylvestris and N. tomentosiformis are the most likely progenitors of N. tabacum. We observed a 4-8% genome reduction in the allotetraploid N. tabacum compared with its ancestral species. Between 81,000 and 94,000 gene models were identified for the three varieties. The alkaloid biosynthesis pathway characteristic of Nicotiana species and the glutamate/aspartate pathway did not show marked CNVs or preferential expression of genes from one ancestor within the investigated genes; indeed, only four of 32 identified homoeologous pairs were differentially expressed. This suggests that there was no adaptation affecting these two pathways at the genome or transcriptome level. The comparison of the genomes of the three varieties might shed some light on the effect of human selection, in addition to natural selection, resulting in resistance to TMV and potyviruses. However, more analyses about the evolution of R genes after the formation of allotetraploid will be necessary to understand this phenomenon. The draft genomes are sufficiently complete to both make and test hypotheses at the biological level, as exemplified by the analysis of virus resistance and alkaloid pathway genes. Together with the genome of N. otophora, they represent an important contribution to the SOL-100 genome project. Alongside genomes of the ancestral species N. sylvestris and N. tomentosiformis, the De novo genome assembly. Raw DNA reads were preprocessed with FASTX toolkit utilities 52 by first trimming 3 0 bases with qualities lower than 30, and then discarding reads shorter than 50 bases or with o90% of the bases with qualities lower than 30. The paired-end libraries with insert sizes shorter than 200 bases were further preprocessed using FLASH 53 to merge the paired-end reads into extended single reads. The paired and single reads from the paired-end libraries were then assembled into contigs using SOAPdenovo 27 with a k-mer of 63. Paired reads from paired-end and mate-pair libraries were used for scaffolding by increasing library size. To improve scaffolding, mate-pair libraries from the closely related Nicotiana species N. sylvestris and N. tomentosiformis [bib_ref] Reference genomes and transcriptomes of Nicotiana sylvestris and Nicotiana tomentosiformis, Sierro [/bib_ref] were also used (Supplementary . SOAPdenovo was instructed to use these libraries only during its scaffolding step, and not during the building of contigs or during gap closing. Gaps resulting from the scaffolding were closed using GapCloser 27 and all sequences shorter than 200 bases were discarded from the final assemblies. After closing the gaps, singletons were blasted against the scaffolds and removed if matching was higher than 97% to avoid artificial duplication of short sequences. [fig_ref] Figure 1 |: map linkage group b Sequence-based linkage group origin c WGP map-based origin... [/fig_ref] shows details of the N. tabacum TN90 genome. Synteny with other Solanaceae. The tobacco assemblies were superscaffolded using the sequence-based WGP tobacco physical map. The obtained superscaffolds were then anchored to the tobacco genetic map by mapping of SSR markers. Synteny with the tomato and potato chromosomes (Supplementary Data 2 and 3) was determined by mapping of tomato and potato reference proteins (ITAG2.3 and PGSC_DM_4.03) to the genetic map-anchored superscaffolds using BLAT [bib_ref] BLAT-the BLAST-like alignment tool, Kent [/bib_ref]. The mapping was filtered to retain hits with at least 50% coverage and 80% identity. Supplementary lists the number of proteins from each tomato and potato chromosome that are mapped to each linkage group. In addition to using the SSR markers shared between the N. tabacum, N. tomentosiformis and N. acuminata genetic maps, TAIR10 proteins for COSII markers present on two latter maps were also mapped with BLAT and filtered to retain hit with at least 50% coverage and 50% identity. To determine synteny between tomato or potato and tobacco using whole genomes , we masked the genomes using tantan [bib_ref] A new repeat-masking method enables specific detection of homologous sequences, Frith [/bib_ref] , tandem repeat finder [bib_ref] Tandem repeats finder: a program to analyze DNA sequences, Benson [/bib_ref] and RepeatMasker and used nucmer [bib_ref] Versatile and open software for comparing large genomes, Kurtz [/bib_ref] to detect syntenic DNA blocks. Analysis using MCScanX 32 was performed using the tomato and potato reference proteins and set of predicted proteins derived from the N. tabacum TN90 RNA-seq data . Estimation of the quality of the assembly. N. tabacum Hicks Broadleaf BAC sequences were sequenced by Roche 454 and assembled with newbler. In total 1,864,003 bases were obtained in sequences longer than 80 kb. The TN90, K326 and BX assemblies were mapped to these BACs sequences using blastn and a threshold of 98% identity, and the identified scaffolds remapped to the BAC sequences using LAST [bib_ref] Parameters for accurate genome alignment, Frith [/bib_ref] [bib_ref] Adaptive seeds tame genomic sequence comparison, Kie&quot;basa [/bib_ref]. Repeat content estimation. The repeat content of the genome assemblies was estimated using RepeatMasker with the eudicot repeat library available from the Sol Genomics Network, the TIGR Solanaceae repeat library and RepeatScout 59 libraries created using sequences of at least 200 kb from the draft genome assemblies. Classification of the repeat types was performed using blastn 60 hits to known repeat elements. Transcriptome assembly and quantitative analysis. The transcriptomes were derived by mapping RNA-Seq reads from each variety to the corresponding reference genome. For each variety, three biological replicates were obtained for each tissue. Mapping was performed using the 'tuxedo' suite of short read mapping tools (Bowtie v2.0.0 beta 61 , Tophat v2.0.4 (ref. [bib_ref] TopHat2: accurate alignment of transcriptomes in the presence of insertions, deletions and..., Kim [/bib_ref] and Cufflinks v2.0.2 (ref. [bib_ref] Transcript assembly and quantification by RNA-Seq reveals unannotated transcripts and isoform switching..., Trapnell [/bib_ref]. The resulting fragments were used as input for the Trinity software suite 64 open reading frame finder; the resulting predicted protein sequences were filtered for uniqueness and used for further analysis. GO terms were assigned to the unique subset of predicted protein sequences using InterProScan [bib_ref] InterProScan-an integration platform for the signature-recognition methods in InterPro, Zdobnov [/bib_ref]. Protein clusters were determined using OrthoMCL 65 with standard parameters. Besides the predicted proteins from the TN90 transcriptome, we used the following sources to obtain proteins from other species: SGN for N. benthamiana, ITAG (version 2.3) for S. lycopersicum and TAIR (version 10) for A. thaliana. All protein sets were filtered for a minimum protein size of 100 amino-acid residues. For the quantitative analysis, the mapped RNA-seq reads were counted using HTSeq, and DESeq 66 was used to calculate differences in expression between the two tissues. Genes with a 10-fold upregulation and an associated adjusted P-value cutoff of 0.001 were deemed to be significantly differentially regulated. The set of overexpressed genes in leaves and roots was analysed for GO term enrichment. The software tool BiNGO 67 was used to calculate and visualize the enrichment. For determining enriched terms, the hypergeometric test was used with subsequent Benjamini and Hochberg false discovery rate corrections; a P-value threshold of 0.001 was used as a cutoff for inferring enrichment. Pathway gene identification. Genes of interest were identified by mapping TAIR10 or UniProt proteins using BLAT 29 to the genomes of N. tabacum ancestors N. sylvestris and N. tomentosiformis 15 , extracting the identified genomic regions and 5 kb flanking regions, remapping the query protein using Exonerate 68 and predicting the target gene using SNAP [bib_ref] Gene finding in novel genomes, Korf [/bib_ref] with the Exonerate hints. The thuspredicted N. sylvestris and N. tomentosiformis coding DNA sequences were then mapped to N. tabacum assemblies, and the N. tabacum coding DNA sequences and proteins were extracted using the above method. PCR eIF4 gene family detection. Sequences related to eIF4E were identified by mapping tomato homologues as described above. In addition to the previously described PVY test primers [bib_ref] Development of SCAR markers linked to three disease resistances based on AFLP..., Julio [/bib_ref] , specific primer pairs were designed for each eIF4E member . Genomic DNA of different varieties was isolated as described in the plant section. PCR of genomic DNA was conducted using a 2-ng ml À 1 template, 0.25 mM of each primer and GoTaq Hot Start polymerase (Promega AG, Switzerland). PCR was performed at 95°C for 8 min, then 40 cycles of 30 s at 95°C, 55°C and 72°C, sequentially, with a final extension at 5 min at 72°C. PCR products were separated on a 1.2% agarose gel in 0.5 Â TAE and visualized using ethidium bromide. [fig] Figure 1 |: map linkage group b Sequence-based linkage group origin c WGP map-based origin d N. sylvestris coverage (%) e N. tomentosiformis coverage (%) f N. otophora coverage (%) g Tomato protein coverage (Tobacco genome. Blue and red represent features of S and T origins in the Nicotiana tabacum TN90 genome, respectively. Track a indicates the origin of the linkage group previously determined by SSR amplification. Track b shows the assignment of the linkage group origin based on sequence identity with N. sylvestris, N. tomentosiformis and N. otophora. Track c shows the origin of the WGP physical map contig used for super-scaffolding. The blue, red and green histograms (tracks d, e and f, respectively) indicate the percentage of each 2 Mb region covered by an N. sylvestris, N. tomentosiformis or N. otophora sequence of at least 1,000 bp and 98% identity. Track g shows the density of coding regions identified by mapping of reference tomato proteins. Track h gives the position of SSR markers mapped to the genome sequences. The centre links regions of the N. tabacum genome based on sequence homology, as determined by reference tomato proteins mapping at two locations. [/fig] [fig] Figure 2 |: Synteny between Nicotiana tabacum, Nicotiana tomentosiformis and Nicotiana acuminata genetic linkage groups and tomato chromosomes. Links between tomato and tobacco are based on the mapping of tomato proteins to linkage group-anchored tobacco sequences. Note that not all regions identified by this method should necessarily be considered as syntenic. Links between tobacco and N. tomentosiformis (red) and N. acuminata (blue) are based on shared SSR (full lines) and COSII (dotted lines) markers. Triangles indicate the amplification of the corresponding SSR in N. tomentosiformis (red) or N. sylvestris (blue). [/fig]

Validating the validation: reanalyzing a large-scale comparison of deep learning and machine learning models for bioactivity prediction ## Cv simulations Adapting the simulations from [bib_ref] Cross-validation failure: small sample sizes lead to large error bars, Varoquaux [/bib_ref] , we performed the simulations described in the text for N train = 30, 100, 300, 1000. The results are shown in [fig_ref] Figure 5: The error in estimating generalization performance from 3-fold cross validation, using the... [/fig_ref]. As is expected, the errors in estimating generalization performance from a cross-validation procedure decrease as the sample size increases. ## Correlation of model mean auc performance In we provide the correlation between mean AUC-ROC performances for all models. This differs from the plot in the main text, which only includes the performance on large assays. As is expected, the results are much more variable when including the many assays with small sample sizes. Additionally, we provide the correlation matrix corresponding to figure 7. [fig] 2019, Figure 1: Email address: [email protected] (Alpha A. Lee) Preprint submitted to (Journal Name)May 24, Distributions of classification scores for a theoretical classifier of positive and negative examples. The negative scores follow a B(a = 1, b = 1) distribution, while the positive scores follow a B(a = 4, b = 1) distribution. Adapted from the simulations in[1] [/fig] [fig] Figure 3: Distributions of classification scores for a theoretical classifier of positive and negative examples. The negative scores follow a B(a = 1, b = 4) distribution, while the positive scores follow a B(a = 1, b = 1) distribution. Adapted from the simulations in [1] [/fig] [fig] Figure 4: ROC, PRC, and enrichment factor curves for predictions from the theoretical classifier shown in 3. The curves result from 10 runs of a simulation with large class imbalance (∼ 1% actives) [/fig] [fig] Figure 5: The error in estimating generalization performance from 3-fold cross validation, using the procedure described in the main text. [/fig] [fig] Figure 6 7, Figure 7: The correlation (Pearson) matrix of all model mean AUC-ROC performances for all assays. This table summarizes the information shown graphically in figure The correlation of all model mean AUC-ROC performances for all assays. The density plots on the diagonal represent the distribution of mean AUC-ROC prediction scores for a given classifier. [/fig] [fig] Figure 8: The comparison of AUC-ROC performance on all test folds of FNN against Naive Bayes (NB), k-nearest neighbors (KNN), and random forests (RF) [/fig]

Noradrenergic Dysfunction in Alzheimer's and Parkinson's Diseases—An Overview of Imaging Studies Frontiers in Aging Neuroscience | www.frontiersin.org Noradrenergic dysfunction contributes to cognitive impairment in Alzheimer's Disease (AD) and Parkinson's Disease (PD). Conventional therapeutic strategies seek to enhance cholinergic and dopaminergic neurotransmission in AD and PD, respectively, and few studies have examined noradrenergic dysfunction as a target for medication development. We review the literature of noradrenergic dysfunction in AD and PD with a focus on human imaging studies that implicate the locus coeruleus (LC) circuit. The LC sends noradrenergic projections diffusely throughout the cerebral cortex and plays a critical role in attention, learning, working memory, and cognitive control. The LC undergoes considerable degeneration in both AD and PD. Advances in magnetic resonance imaging have facilitated greater understanding of how structural and functional alteration of the LC may contribute to cognitive decline in AD and PD. We discuss the potential roles of the noradrenergic system in the pathogenesis of AD and PD with an emphasis on postmortem anatomical studies, structural MRI studies, and functional MRI studies, where we highlight changes in LC connectivity with the default mode network (DMN). LC degeneration may accompany deficient capacity in suppressing DMN activity and increasing saliency and task control network activities to meet behavioral challenges. We finish by proposing potential and new directions of research to address noradrenergic dysfunction in AD and PD. Keywords: norepinephrine, dopamine, neurodegeneration, neurodegenerative, locus coeruleus, ventral tegmental area, midbrain, MRI ANATOMICAL AND NEUROBIOLOGICAL CONSIDERATIONS ## Alzheimer's disease Alzheimer's disease (AD) is a well-known cause of dementia that is associated with the accumulation of intraneuronal neurofibrillary tangles (NFTs; [bib_ref] A sequence of cytoskeleton changes related to the formation of neurofibrillary tangles..., Braak [/bib_ref] and extraneuronal neuropil threads [bib_ref] Neuropil threads of Alzheimer's disease show a marked alteration of the normal..., Perry [/bib_ref]. NFTs are composed of abnormally phosphorylated tau, a protein supporting cytoskeletal structure in neurons [bib_ref] Alzheimer's disease: Tau proteins, the promoting factors of microtubule assembly, are major..., Delacourte [/bib_ref] [bib_ref] Accumulation of abnormally phosphorylated τ precedes the formation of neurofibrillary tangles in..., Bancher [/bib_ref] [bib_ref] A sequence of cytoskeleton changes related to the formation of neurofibrillary tangles..., Braak [/bib_ref]. Neuropil threads (NTs) are composed of tau and ubiquitin and located typically at distal dendrites [bib_ref] Neuropil threads of Alzheimer's disease show a marked alteration of the normal..., Perry [/bib_ref]. The progression of AD, first detailed by in six stages that correlate with accumulation of NFTs and NTs. Stage I and II are regarded as the entorhinal stage and associated with accumulation of NFTs and NTs in the transentorhinal regions [bib_ref] Alzheimer lesions in the entorhinal region and isocortex in Parkinson's and Alzheimer's..., Jellinger [/bib_ref] [bib_ref] Neuropathological staging of Alzheimer lesions and intellectual status in Alzheimer's and Parkinson's..., Bancher [/bib_ref] [bib_ref] A sequence of cytoskeleton changes related to the formation of neurofibrillary tangles..., Braak [/bib_ref]. Stage III and IV are known as the limbic stage and progress to involve considerable portions of the entorhinal cortex. The limbic stage is also associated with minor hippocampal change. Clinically, stage III/IV are associated with impaired cognitive functioning and minor personality changes [bib_ref] Alzheimer lesions in the entorhinal region and isocortex in Parkinson's and Alzheimer's..., Jellinger [/bib_ref] [bib_ref] Neuropathological staging of Alzheimer lesions and intellectual status in Alzheimer's and Parkinson's..., Bancher [/bib_ref] [bib_ref] A sequence of cytoskeleton changes related to the formation of neurofibrillary tangles..., Braak [/bib_ref]. The last two stages, stage V and VI, represent the isocortical stages of AD. The feature that distinguishes the isocortical stages is significant involvement of the cerebral cortex and hippocampus. Stage V and VI manifest in clinically diagnosable AD. It is important to note that AD does not follow a uniform pattern of neuroanatomical changes. A more recent study of AD pathology revealed three NFT distribution patterns described as typical AD, hippocampal-sparing AD and limbic predominant AD [bib_ref] Neuroimaging correlates of pathologically defined subtypes of Alzheimer's disease: a case-control study, Whitwell [/bib_ref]. Typical AD presents with diffuse NFT deposition in the hippocampus and cortex. Hippocampal-sparing AD demonstrates less NFT deposition in the hippocampus but more in the cortex, when compared to typical AD. Lastly, limbic predominant AD involves significantly more NFT deposition in the hippocampus than either of the other two groups. Thus, AD appears to be a relatively heterogeneous disease in terms of its etiological processes as characterized by chemical neuroanatomy. Whereas clinical staging focuses on changes in the cerebral cortex and hippocampus, other studies have implicated changes in the locus coeruleus (LC) and nucleus basalis of Meynert (NbM). The loss of cholinergic neurons in the NbM was demonstrated as early as 1981 by Nissl-stained histological sections and cell counts in post-mortem studies of AD [bib_ref] Alzheimer disease: evidence for selective loss of cholinergic neurons in the nucleus..., Whitehouse [/bib_ref]. Patients with AD demonstrated degeneration of >75% of the neurons within the NbM [bib_ref] Alzheimer's disease and senile dementia: loss of neurons in the basal forebrain, Whitehouse [/bib_ref]. The NbM cholinergic neurons project throughout the brain and loss of the cholinergic inputs may account for behavioral changes such as short-term memory loss, disorientation and problems with language in AD [bib_ref] Alzheimer's disease, Burns [/bib_ref]. Other studies demonstrated significant loss of LC neurons in AD [bib_ref] Noradrenergic changes, aggressive behavior, and cognition in patients with dementia, Matthews [/bib_ref] [bib_ref] Compensatory changes in the noradrenergic nervous system in the locus ceruleus and..., Szot [/bib_ref] [bib_ref] Dogs with canine counterpart of Alzheimer's disease lose noradrenergic neurons, Insua [/bib_ref]. In fact, LC degeneration has been documented in AD too as early as 1981 [bib_ref] Cell loss in the locus coeruleus in senile dementia of Alzheimer type, Tomlinson [/bib_ref]. In comparative studies of disease burden in patients with advanced-stage AD the NbM demonstrated the most significant cell loss followed by the LC and transentorhinal cortex [bib_ref] Early neurone loss in Alzheimer's disease: cortical or subcortical?, Arendt [/bib_ref]. LC degeneration also occurs during healthy aging, which could amount to 50% of cell loss in the tenth decade of life [bib_ref] Locus coeruleus cell loss in the aging human brain: a non-random process, Manaye [/bib_ref]. In a longitudinal clinical-pathologic cohort study, 165 individuals completed a battery of cognitive tests and underwent brain autopsy upon death to examine neuronal density of the LC and other aminergic nuclei in the midbrain/brain stem. Modeled together, only LC neuronal density was related to cognitive decline, suggesting that the LC may be a structural component of neural reserve. Parkinson's Disease Structural brain changes in Parkinson's Disease (PD) similarly follow a time course. PD is associated with the formation of intraneuronal aggregates of the protein alpha synuclein, known as Lewy bodies. The accumulation of Lewy bodies is closely associated with the degeneration of the ventrolateral substantia nigra including specifically the pars compacta (SNc; [bib_ref] Ageing and Parkinson's disease: substantia nigra regional selectivity, Fearnley [/bib_ref] [bib_ref] The substantia nigra of the human brain: II. Patterns of loss of..., Damier [/bib_ref]. The loss of SNc neurons results in diminution of extranigral dopaminergic projections throughout the striatum, allocortex, and neocortex [bib_ref] Staging of brain pathology related to sporadic Parkinson's disease, Braak [/bib_ref]. As with AD, post-mortem pathology has been described in six different stages which were determined using alphasynuclein-immunopositive Lewy neurites and Lewy bodies to track disease progression in PD [bib_ref] Staging of brain pathology related to sporadic Parkinson's disease, Braak [/bib_ref]. Stage 1 is associated with lesions in the dorsal IX/X motor nucleus. Stage 2 includes the findings of stage 1 with additional involvement of the caudal raphe nuclei, gigantocellular reticular nucleus, and coeruleus-subcoeruleus complex. Stage 3 is characterized by significant lesions in the midbrain specifically in the area of the SNc. Progression to stage 4 represents first cortical involvement specifically in the transentorhinal region and allocortex. Stage 5 involves the neocortex with lesions present in sensory association areas and prefrontal regions. Finally, progression to stage six is characterized by additional lesions in sensory association areas and new lesions in the premotor, primary sensory, and primary motor areas. In summary, lesions in PD start at the dorsal motor nucleus and progress to involve the LC, SNc, transentorhinal and cortical regions, in that order. It is worth noting that Braak staging of PD implies an ascending propagation of pathology similar to that of prion diseases and offers to explain the neural processes underlying symptom progression. However, the exact mechanisms of stage progression remain unclear. In particular, studies have reported that only about half of postmortem brains in PD demonstrated Lewy body pathology consistent with the Braak staging [bib_ref] The dorsal motor nucleus of the vagus is not an obligatory trigger..., Kalaitzakis [/bib_ref] [bib_ref] A critical evaluation of current staging of α-synuclein pathology in Lewy body..., Jellinger [/bib_ref] [bib_ref] Evaluation of the Braak hypothesis: how far can it explain the pathogenesis..., Halliday [/bib_ref] , challenging the notion that PD progresses in a sequence similar to prion diseases [bib_ref] Parkinson's Disease Is Not Simply a Prion Disorder, Surmeier [/bib_ref]. Thus, whereas degeneration of the SNc is the most classic neuropathological finding in PD, postmortem studies also demonstrate Lewy body burden and associated loss of noradrenergic neurons in the LC [bib_ref] Alterations in catecholamine neurons of the locus coeruleus in senile dementia of..., Chan-Palay [/bib_ref] [bib_ref] Cognitive decline correlates with neuropathological stage in Parkinson's disease, Braak [/bib_ref] [bib_ref] Common factors among Alzheimer's disease, Parkinson's disease, and epilepsy: possible role of..., Szot [/bib_ref]. Interestingly, the loss of LC neurons in PD has been documented to occur earlier and in greater magnitude than that of the SNc [bib_ref] Norepinephrine: the redheaded stepchild of Parkinson's disease, Rommelfanger [/bib_ref]. For instance, in Braak staging of PD involvement of the LC starts in stage 2 while involvement of the SNc is not present until stage 3. Loss of LC neurons in PD is associated with depletion of noradrenergic inputs in the frontal cortex, cerebellum, striatum, thalamus, and hypothalamus [bib_ref] Cerebellar norepinephrine in patients with Parkinson's disease and control subjects, Kish [/bib_ref] [bib_ref] Noradrenaline, dopamine and serotonin levels and metabolism in the human hypothalamus: observations..., Shannak [/bib_ref] [bib_ref] Progression of monoaminergic dysfunction in Parkinson's disease: a longitudinal 18 F-dopa PET..., Pavese [/bib_ref] , all of which receive projections from the LC. Together, although postmortem anatomical studies have focused on changes in the SNc, there is significant evidence in support of changes of the noradrenergic system as a critical etiological process of PD. [fig_ref] TABLE 1 |: Summary of anatomical changes in Alzheimer's Disease and Parkinson's Disease [/fig_ref] highlights the anatomical and neurobiological changes in AD and PD. I/II-NFTs and NTs accumulate in transentorhinal regions III/IV-pathology in entorhinal cortex with hippocampal change V/VI-significant accumulation in cerebral cortex and hippocampus -Early loss of neurons is also seen in NbM and the LC [bib_ref] Compensatory changes in the noradrenergic nervous system in the locus ceruleus and..., Szot [/bib_ref] PD -Accumulation of LBs is closely associated with degeneration of the SNc [bib_ref] Ageing and Parkinson's disease: substantia nigra regional selectivity, Fearnley [/bib_ref] [bib_ref] The substantia nigra of the human brain: II. Patterns of loss of..., Damier [/bib_ref] -Loss of SNc neurons is associated with diminution of extranigral dopaminergic projections through striatum, allocortex, and neocortex [bib_ref] Staging of brain pathology related to sporadic Parkinson's disease, Braak [/bib_ref] -Loss of LC neurons occurs earlier and in greater magnitude than that of the SNc [bib_ref] Norepinephrine: the redheaded stepchild of Parkinson's disease, Rommelfanger [/bib_ref] -Loss of LC neurons is associated with depletion of noradrenergic inputs in the frontal cortex, cerebellum, striatum, thalamus, and hypothalamus [bib_ref] Cerebellar norepinephrine in patients with Parkinson's disease and control subjects, Kish [/bib_ref] [bib_ref] Noradrenaline, dopamine and serotonin levels and metabolism in the human hypothalamus: observations..., Shannak [/bib_ref] [bib_ref] Progression of monoaminergic dysfunction in Parkinson's disease: a longitudinal 18 F-dopa PET..., Pavese [/bib_ref] ## Anatomical changes in ad and pd ## Ad ## Structural brain imaging in ad and pd ## Alzheimer's disease Many studies have correlated Braak staging with atrophic changes on magnetic resonance imaging (MRI) using voxelbased morphometry (VBM) [bib_ref] Changes in brain morphology in Alzheimer disease and normal aging: is Alzheimer..., Ohnishi [/bib_ref] [bib_ref] Longitudinal evaluation of both morphologic and functional changes in the same individuals..., Matsuda [/bib_ref] [bib_ref] Dissociating atrophy and hypometabolism impact on episodic memory in mild cognitive impairment, Chetelat [/bib_ref] [bib_ref] Verbal episodic memory impairment in Alzheimer's disease: a combined structural and functional..., Rémy [/bib_ref] , focusing largely on hippocampal atrophy as a biomarker of AD. The patterns of cerebral atrophy as detected on MRI vary across pathological subtypes of AD [bib_ref] Neuroimaging correlates of pathologically defined subtypes of Alzheimer's disease: a case-control study, Whitwell [/bib_ref] [bib_ref] Anatomical heterogeneity of Alzheimer disease: based on cortical thickness on MRIs, Noh [/bib_ref]. Medial temporal cortical atrophy is seen in limbic-predominant while severe cortical atrophy is seen more in hippocampal-sparing AD [bib_ref] Neuroimaging correlates of pathologically defined subtypes of Alzheimer's disease: a case-control study, Whitwell [/bib_ref]. Distinct anatomical changes have also been identified on MRI when early and late onset AD are contrasted [bib_ref] The topography of grey matter involvement in early and late onset Alzheimer's..., Frisoni [/bib_ref]. Individuals with early onset AD tend to demonstrate more extensive occipital gray matter atrophy whereas those with late onset AD demonstrate more extensive hippocampal atrophy. The notion of prioritizing anatomical integrity of an isolated region as a biomarker of AD may be convenient; however, the anatomical heterogeneity brings to question the sensitivity of structural changes of individual brain regions as a diagnostic tool of AD. Although the LC is implicated by postmortem anatomical studies, only recently have studies begun to employ highresolution fast spin-echo T1-weighted imaging to reveal signal attenuation of the LC in individuals with mild cognitive impairment (MCI) and AD [bib_ref] Detection of changes in the locus coeruleus in patients with mild cognitive..., Takahashi [/bib_ref]. This finding corresponds well to pathological findings of a decrease in neuromelanin contents as a result of LC neuronal loss in these patients. It is important to note that conventional MRI has not been successful in delineating the LC. However, fast spinecho T1-weighted imaging, which relies on the paramagnetic properties of neuromelanin, has been demonstrated as a reliable method to quantify signal attenuation or volume loss in neuromelanin-containing tissues such as the LC and SNc [bib_ref] Effects of the interaction between ferric iron and L-dopa melanin on T1..., Tosk [/bib_ref] [bib_ref] Paramagnetic metal scavenging by melanin: MR imaging, Enochs [/bib_ref]. Neuromelanin is a pigment found in both SNc and LC [bib_ref] Magnetic resonance imaging of the substantia nigra in Parkinson's disease, Lehéricy [/bib_ref]. Neuromelanin is identified in monoamines-containing neurons and formed by polymerization of 4,5-dihydroxyindole monomers [bib_ref] Fe (III)-coordination properties of neuromelanin components: 5, 6-dihydroxyindole and 5, 6-dihydroxyindole-2-carboxylic acid, Charkoudian [/bib_ref] via enzymatic processes that involve monoamine oxidase [bib_ref] Neuromelanin synthesis in rat and human substantia nigra, Rabey [/bib_ref] , and its concentrations increase with age. Neuromelanin can chelate metals and protect against oxidative stress [bib_ref] Identification of L-ferritin in neuromelanin granules of the human substantia nigra: a..., Tribl [/bib_ref]. On the other hand, excessive neuromelanin accumulation compromises neuronal integrity and causes dissolution of cell mass and decrease in neuromelanin signals, as may occur in neurodegenerative conditions. Neuromelanin has paramagnetic-T1 shortening effects and, when combined with metals like iron and copper, would make the SNc and LC appear hyper-intense on MRI. On neuromelanin-MRI there is signal attenuation in MCI participants who do or do not to progress to AD, suggesting that attenuated LC signal alone may not be diagnostic of AD. In fact, the LC exhibited signal changes during healthy aging [bib_ref] In vivo MRI assessment of the human locus coeruleus along its rostrocaudal..., Betts [/bib_ref]. Most recently, a stereological study of postmortem human brains reported that as the Braak staging increases by 1 unit, LC volume decreases by 8.4%. Together, although not necessarily pathognomonic of AD, LC volume measurements offer a promising biomarker to track the progression to AD from presymptomatic stages. ## Parkinson's disease Findings from VBM studies of PD demonstrate varying results with the most common changes identified in the frontal and parieto-occipital regions [bib_ref] Brain volume changes in Parkinson's disease and their relationship with cognitive and..., Biundo [/bib_ref] [bib_ref] Voxel-wise meta-analysis of gray matter abnormalities in idiopathic Parkinson's disease, Pan [/bib_ref] [bib_ref] Early cortical gray matter loss and cognitive correlates in nondemented Parkinson's patients, Lee [/bib_ref]. A meta-analysis of 498 patients with idiopathic PD revealed reductions in gray matter (GM) volume in the left frontal temporal cortices encompassing inferior frontal and superior temporal gyri [bib_ref] Voxel-wise meta-analysis of gray matter abnormalities in idiopathic Parkinson's disease, Pan [/bib_ref]. Other brain areas demonstrating GM reductions in PD were the left insular cortex, with insular GM density potentially related to memory scores [bib_ref] Voxel-wise meta-analysis of gray matter abnormalities in idiopathic Parkinson's disease, Pan [/bib_ref] , and the middle and superior frontal gyrus [bib_ref] Brain volume changes in Parkinson's disease and their relationship with cognitive and..., Biundo [/bib_ref]. In a study where patients were scanned twice 2 years apart, PD but not controls demonstrated significant GM loss in the putamen and parietal cortex. On the other hand, some studies did not reveal any significant differences in GM volume between PD and healthy controls [bib_ref] Comprehensive morphometry of subcortical grey matter structures in early-stage Parkinson's disease. Hum, Menke [/bib_ref] [bib_ref] Structural and functional imaging study in dementia with Lewy bodies and Parkinson's..., Borroni [/bib_ref]. In one of these studies PD patients were split into those with and without dementia and the PD group with dementia but not the group without demonstrated reductions in frontal regional GM volume [bib_ref] Structural and functional imaging study in dementia with Lewy bodies and Parkinson's..., Borroni [/bib_ref]. However, compared to health controls, the two groups combined did not show frontal GM reduction. Together, these findings suggest that variability across studies may relate to the cognitive status of study participants. Other issues including sample size, duration of illness, medication status may all impact the imaging findings. Neuromelanin Imaging has also been widely used to examine the integrity of SNc function in PD [bib_ref] Differentiation of early-stage parkinsonisms using neuromelaninsensitive magnetic resonance imaging, Ohtsuka [/bib_ref] [bib_ref] Automated neuromelanin imaging as a diagnostic biomarker for Parkinson's disease, Castellanos [/bib_ref] [bib_ref] Substantia nigra neuromelanin-MR imaging differentiates essential tremor from Parkinson's disease, Reimão [/bib_ref]. In melanin-sensitive MR sequences both the SNc and the LC demonstrate reduction in signal intensity or contrast ratio (CR) in PD . The finding of decreased neuromelanin CR in the SNc has been reported in multiple studies of PD [bib_ref] Neuromelanin magnetic resonance imaging in Parkinson's disease and multiple system atrophy, Matsuura [/bib_ref] [bib_ref] Evaluation of Parkinson disease and Alzheimer disease with the use of neuromelanin..., Miyoshi [/bib_ref] [bib_ref] Neuromelanin MRI in a family with mitochondrial parkinsonism harboring a Y955C mutation..., Mukai [/bib_ref] [bib_ref] Differentiation of early-stage parkinsonisms using neuromelaninsensitive magnetic resonance imaging, Ohtsuka [/bib_ref] [bib_ref] Neuromelanin imaging and dopaminergic loss in Parkinson's disease. Front, Isaias [/bib_ref]. A recent work of neuromelanin-sensitive MRI employed an automated segmentation algorithm to quantify the volumes of the SNc and LC [bib_ref] Automated neuromelanin imaging as a diagnostic biomarker for Parkinson's disease, Castellanos [/bib_ref]. Receiver operating characteristic analysis demonstrated better accuracy using the SNc than LC volumes in the diagnosis of PD. Neuromelaninsensitive MRI has also been used to differentiate essential tremor from PD [bib_ref] Substantia nigra neuromelanin-MR imaging differentiates essential tremor from Parkinson's disease, Reimão [/bib_ref] and early-stage PD from healthy controls [bib_ref] Differentiation of early-stage parkinsonisms using neuromelaninsensitive magnetic resonance imaging, Ohtsuka [/bib_ref]. Thus, neuromelanin MRI offers a promising technique for characterizing structural and potentially functional changes of the SNc and LC in PD. ## Summary Neuromelanin-sensitive MRI has become a popular tool to investigate LC function and dysfunction. Neuromelanin CRs enable better visualization and assessment of the anatomical and functional integrity of the LC. In healthy adults neuromelanin CR of the LC increased until the sixth decade of life at which point it started to decrease . This as well as another study [bib_ref] Neuromelanin marks the spot: identifying a locus coeruleus biomarker of cognitive reserve..., Clewett [/bib_ref] suggests that LC neuromelanin accumulates in an inverted U pattern with age and peaks around 60 years. In older adults, greater LC signal intensity was associated with higher composite scores of cognitive reserve [bib_ref] Neuromelanin marks the spot: identifying a locus coeruleus biomarker of cognitive reserve..., Clewett [/bib_ref]. Neuromelanin CRs were greater in healthy adults as compared to patients with major depression, schizophrenia, MCI, PD, and AD . Together, these findings demonstrate the importance of LC imaging in investigating cognitive changes in neurodegenerative conditions and other psychiatric illnesses that implicate monoaminergic dysfunction. ## Functional brain imaging in ad and pd Functional MRI (fMRI) utilizes blood-oxygenation level dependent (BOLD) signals as a proxy for neural activity. The premise of using BOLD signals to reflect neural activities lies in the fact that oxyhemoglobin and deoxyhemoglobin exhibit different magnetic properties. As neurons are activated, more oxygen is consumed, resulting in a change in the ratio of oxyhemoglobin and deoxyhemoglobin, which can be picked up by BOLD imaging. Many studies have combined fMRI and a behavioral task to examine cerebral responses to cognitive and affective challenges. However, a decade of work has suggested that functional organization of the brain can be delineated by how BOLD signals of brain regions are correlated during a resting state. Low-frequency BOLD signal fluctuations reflect connectivity between functionally related brain regions [bib_ref] Functional connectivity in the motor cortex of resting human brain using echo-planar..., Biswal [/bib_ref] [bib_ref] A method for using blocked and event-related fMRI data to study "resting..., Fair [/bib_ref] [bib_ref] Spontaneous fluctuations in brain activity observed with functional magnetic resonance imaging, Fox [/bib_ref]. Studies of this "spontaneous" activity have provided insights into the intrinsic functional architecture of the brain and shown that coordinated spontaneous fluctuations are present in motor, visual, auditory, default mode, memory, language, dorsal attention, and ventral attention systems [bib_ref] Spontaneous fluctuations in brain activity observed with functional magnetic resonance imaging, Fox [/bib_ref]. Other studies have suggested connectivity analysis of resting-state fMRI data as a useful alternative to characterizing functional subdivisions of a brain region [bib_ref] Functional connectivity mapping of the human precuneus by resting state fMRI, Zhang [/bib_ref]. In a study of resting state functional connectivity (rsFC), the LC demonstrated positive connections to bilateral superior frontal gyrus, primary motor cortex, inferior parietal cortex, inferior temporal cortex, anterior parahippocampal gyrus, posterior insula, putamen, pallidum, ventrolateral thalamus, midbrain, and large areas of the cerebellum . The LC showed negative connectivities too to a wide swath of brain regions, including bilateral visual cortex, middle/superior temporal cortex, precuneus, retrosplenial cortex, posterior parahippocampal gyrus, frontopolar cortex, caudate nucleus, and the dorsal and medial thalamus. Findings of positive connectivity support a role for norepinephrine (NE) in orienting and sensorimotor responses to external stimuli. The findings of negative connectivity to the precuneus support a role for the LC in regulating activity of the default-mode network (DMN). Together, through connectivity to the cerebral cortex the LC may suppress DMN activity in response to external stimuli and facilitate engagement of the saliency ("alert and orient") and executive control systems [bib_ref] Greater activation of the "default" brain regions predicts stop signal errors, Li [/bib_ref]. On the other hand, one should be cautioned against over-interpreting rsFC in terms of functional implications. A positive rsFC does not necessarily mean that the two brain regions are functionally congruent. Long range cortical-cortical/subcortical projections may target inhibitory interneurons within the recipient regions [bib_ref] Functional maps of neocortical local circuitry, Thomson [/bib_ref] [bib_ref] Laminarly orthogonal excitation of fast-spiking and low-thresholdspiking interneurons in mouse motor cortex, Apicella [/bib_ref] [bib_ref] Connectivity of mouse somatosensory and prefrontal cortex examined with trans-synaptic tracing, Denardo [/bib_ref]. For instance, the amygdala and ventromedial prefrontal cortex (vmPFC) show positive rsFC [bib_ref] Whole brain resting-state analysis reveals decreased functional connectivity in major depression, Veer [/bib_ref] [bib_ref] Resting-state connectivity of the amygdala is altered following Pavlovian fear conditioning, Schultz [/bib_ref] but may respond in isolation or in opposite directions to behavioral challenges such as affective regulation [bib_ref] Amygdala and ventromedial prefrontal cortex are inversely coupled during regulation of negative..., Urry [/bib_ref] and extinction learning of fear [bib_ref] Extinction learning in humans: role of the amygdala and vmPFC, Phelps [/bib_ref]. The proposition that the LC circuit responds to behavioral tasks is supported by task-based studies demonstrating LC activation to stimulus change. In an oddball task LC activity increased to the detection of novel, oddball stimuli [bib_ref] Modulation of locus coeruleus activity by novel oddball stimuli, Krebs [/bib_ref]. In a picture-word interference task the LC responded to novel as compared to familiar stimuli [bib_ref] Picture novelty attenuates semantic interference and modulates concomitant neural activity in the..., Krebs [/bib_ref]. In an attention task, exposure to the alerting condition was associated with increased activation of the LC [bib_ref] Modulation of prefrontal functioning in attention systems by NPSR1 gene variation, Neufang [/bib_ref]. The LC is also consistently activated in response to stimuli that trigger arousal responses via fear, pain, and anger [bib_ref] A direct brainstem-amygdala-cortical 'alarm'system for subliminal signals of fear, Liddell [/bib_ref] [bib_ref] Resolving the Brainstem Contributions to Attentional Analgesia, Brooks [/bib_ref] [bib_ref] Neural indicators of interpersonal anger as cause and consequence of combat training..., Gilam [/bib_ref]. The role of the LC in arousal is examined by a recent study using dexmedetomodine to elicit a state of reduced arousal [bib_ref] Pharmacological modulation of noradrenergic arousal circuitry disrupts functional connectivity of the locus..., Song [/bib_ref]. Dexmedetomidine, a potent sedative, is an α2 agonist that reduces LC neuronal firing and NE release. Along with decreasing arousal, dexmedetomidine diminished connectivity between the LC and posterior cingulate cortex (PCC) as compared to the awake state. Further, dexmedetomidine disrupted PCC connectivity with subcortical and cortical structures central to executive control. These findings together support the role of the LC in facilitating attention shifts in response to stimulus change and in situations that demand a higher level of alertness (see 2017 for a review of task-related LC studies). Finally, an overview of electrophysiological studies of LC neuronal functions would place human imaging findings in a clearer perspective (Aston-Jones and Waterhouse, 2016). First, cortical projections of the LC are predominantly ipsilateral whereas subcortical projections are more likely to be bilateral. Further, projections of individual LC neurons collateralize to cover functionally related brain regions. Neurons projecting to different circuits are segregated in the LC and may release NE asynchronously in response to task demands. Second, studies employing iontophoresis of NE in target regions and electrical stimulation of the LC showed that LC suppresses spontaneous neuronal activity in the cerebellum and cerebral cortex, establishing NE as an inhibitory neurotransmitter at central synapses. Other studies showed that the spontaneous activities are suppressed to a greater extent than stimulusevoked activity, resulting in an increase of the signal noise ratio in response to stimulus. Third, LC neurons discharge both tonically and phasically. LC neurons discharge tonically at a moderate rate but respond to task-related stimuli phasically during focused attention. The Adaptive Gain Theory proposes that LC phasic responses facilitate behavioral adaptation to changing environment whereas tonic activities suppress behavior of low utility (Aston-Jones and Waterhouse, 2016). ## Functional imaging studies of ad The default mode network (DMN) comprises a set of brain regions, including parts of the precuneus and posterior cingulate cortex (PCC; [bib_ref] The role of the posterior cingulate cortex in cognition and disease, Leech [/bib_ref] , that are more active during mind wandering, retrieval of autobiographical memories, and monitoring of the arousal state [bib_ref] The role of the posterior cingulate cortex in cognition and disease, Leech [/bib_ref] , than during exposure to environmental stimuli. A host of studies have noted changes in DMN activity and connectivity in AD during rest [bib_ref] Distinctive resting state network disruptions among Alzheimer's disease, subcortical vascular dementia, and..., Kim [/bib_ref]. Reduced DMN connectivities can be observed from early stages of MCI to late stage AD [bib_ref] Widespread disruption of functional brain organization in early-onset Alzheimer's disease, Adriaanse [/bib_ref] [bib_ref] Functional connectivity changes differ in early and late-onset alzheimer's disease. Hum, Gour [/bib_ref] , as well as in asymptomatic patients genetically at risk of AD [bib_ref] Impaired default network functional connectivity in autosomal dominant Alzheimer disease, Chhatwal [/bib_ref] [bib_ref] Evolving brain functional abnormalities in PSEN1 mutation carriers: a resting and visual..., Sala-Llonch [/bib_ref] [bib_ref] Brain imaging and blood biomarker abnormalities in children with autosomal dominant Alzheimer..., Quiroz [/bib_ref]. Reduced DMN connectivity in AD also correlates with disease severity, as measured by the Clinical Dementia Rating scale [bib_ref] Divergent network connectivity changes in behavioural variant frontotemporal dementia and Alzheimer's disease, Zhou [/bib_ref] [bib_ref] Default mode network connectivity in stable vs progressive mild cognitive impairment, Petrella [/bib_ref] [bib_ref] Loss of intranetwork and internetwork resting state functional connections with Alzheimer's disease..., Brier [/bib_ref]. Of all the brain networks studied, the DMN demonstrates the most consistent connectivity changes in AD. The salience network (SAN) is a network of structures including most prominently the dACC and anterior insula that processes a constant stream of external stimuli to identify salient inputs for goal-directed behavior. In contrast to the DMN, which is an inwardly driven network, the SAN integrate sensory, visceral, and autonomic information to facilitate decision making [bib_ref] Dissociable intrinsic connectivity networks for salience processing and executive control, Seeley [/bib_ref]. Changes in the SAN activity and connectivity are also implicated in AD. Network based analyses frequently demonstrate hyperconnectivity within the SAN in patients with AD [bib_ref] Functional connectivity in autosomal dominant and late-onset Alzheimer disease, Thomas [/bib_ref] [bib_ref] Differentially disrupted functional connectivity of the subregions of the inferior parietal lobule..., Wang [/bib_ref]. Abnormal SAN connectivity has been documented across the disease spectrum from early to late onset AD [bib_ref] Differentially disrupted functional connectivity of the subregions of the inferior parietal lobule..., Wang [/bib_ref] , in patients with autosomal dominant AD [bib_ref] Functional connectivity in autosomal dominant and late-onset Alzheimer disease, Thomas [/bib_ref] , and in those who carry the APOEe4 gene, which is known to increase the risk for AD [bib_ref] Effect of APOE ε4 status on intrinsic network connectivity in cognitively normal..., Machulda [/bib_ref] [bib_ref] Functional network endophenotypes unravel the effects of apolipoprotein E epsilon 4 in..., Goveas [/bib_ref]. One of the key hubs of the SAN is the anterior insular cortex [bib_ref] Dissociable intrinsic connectivity networks for salience processing and executive control, Seeley [/bib_ref]. In a large, voxelbased meta-analysis, the anterior insula demonstrated significant hyperconnectivity in AD. The SAN receives limbic and autonomic inputs and facilitates activity shifts between the DMN and frontoparietal network to help guide behavior [bib_ref] Dissociable intrinsic connectivity networks for salience processing and executive control, Seeley [/bib_ref]. Moreover, changes in SAN connectivity may perturb DMN function and vice versa [bib_ref] Salience network integrity predicts default mode network function after traumatic brain injury, Bonnelle [/bib_ref] [bib_ref] Abnormal resting-state functional connectivity of insular subregions and disrupted correlation with working..., Zhao [/bib_ref]. Thus, the SAN plays a critical role in regulating DMN activity and failure to suppress DMN activity during task challenges correlates with worse cognitive performance. The implications of these changes for AD will be discussed in more depth later. Other studies demonstrated hyperconnectivity of the limbic network [bib_ref] Basal functional connectivity within the anterior temporal network is associated with performance..., Gour [/bib_ref] [bib_ref] Functional connectivity changes differ in early and late-onset alzheimer's disease. Hum, Gour [/bib_ref] , mostly in the hippocampus and entorhinal cortex, in AD. Intriguingly, anterior temporal network hyperconnectivity was correlated positively with memory performance in patients with early onset AD [bib_ref] Functional connectivity changes differ in early and late-onset alzheimer's disease. Hum, Gour [/bib_ref] , perhaps reflecting a transient compensation before progression to late-stage dementia. Together with findings on the DMN, network connectivity changes appear to be a prominent feature of AD. The significance of these findings needs to be evaluated with longitudinal records of disease progression. ## Functional imaging studies of pd In patients with PD slower processing speeds were associated with decreased DMN connectivity at rest, specifically between the posterior cingulate, medial prefrontal and inferior parietal nodes. Off medication, PD patients were unable to suppress DMN activity during a facial emotion recognition task [bib_ref] Dopaminergic modulation of the default mode network in Parkinson's disease, Delaveau [/bib_ref]. In the Montreal card-sorting task, a task that involves manipulation of short-term memory, patients with PD showed less deactivation of the DMN as compared to healthy controls [bib_ref] Dysfunction of the default mode network in Parkinson disease: a functional magnetic..., Van Eimeren [/bib_ref]. Another study reported worse performance on executive functioning, psychomotor speed and verbal memory in association with increased positive connectivity between the SAN and DMN [bib_ref] Salience and default mode network coupling predicts cognition in aging and Parkinson's..., Putcha [/bib_ref]. In other words, reduced anti-correlation between these two networks was associated with impaired cognitive performance in PD. The latter study implicates dysregulation of both SAN and DMN in PD. The right inferior parietal cortex (IPC), a region associated with the DMN [bib_ref] On the relationship between the "default mode network" and the "social brain, Mars [/bib_ref] [bib_ref] Mapping the self in the brain's default mode network, Davey [/bib_ref] , demonstrated aberrant connectivity in PD patients [bib_ref] Resting-state functional reorganization in Parkinson's disease: an activation likelihood estimation meta-analysis, Tahmasian [/bib_ref] with and without depression [bib_ref] Abnormal baseline brain activity in non-depressed Parkinson's disease and depressed Parkinson's disease:..., Wen [/bib_ref] , particularly in those who were cognitively impaired [bib_ref] Resting-state functional connectivity associated with mild cognitive impairment in Parkinson's disease, Amboni [/bib_ref] [bib_ref] To rise and to fall: functional connectivity in cognitively normal and cognitively..., Gorges [/bib_ref]. Altered IPL activity was demonstrated consistently across different neural metrics, including the amplitude of low frequency fluctuations and regional homogeneity [bib_ref] Resting-state functional reorganization in Parkinson's disease: an activation likelihood estimation meta-analysis, Tahmasian [/bib_ref]. In a meta-analysis of rs-fMRI studies, connectivity changes in the right posterior IPC represent the most consistent finding in PD [bib_ref] Resting-state functional reorganization in Parkinson's disease: an activation likelihood estimation meta-analysis, Tahmasian [/bib_ref] and may disrupt activity of the precuneus, PCC, mid-cingulate cortex, middle and superior frontal gyrus, and orbitofrontal gyrus. It is worth noting again that, the posterior IPC is part of the DMN [bib_ref] Mapping the self in the brain's default mode network, Davey [/bib_ref]. Sub-cluster analysis of functional connectivity identified four functionally dissociable subregions in the IPC [bib_ref] Functional clustering of the human inferior parietal lobule by whole-brain connectivity mapping..., Zhang [/bib_ref] [bib_ref] Subspecialization in the human posterior medial cortex, Bzdok [/bib_ref] and the posterior subregion, in the area of angular gyrus and with the strongest connectivity to the DMN, is the exact region that demonstrated aberrant connectivity in PD [bib_ref] Resting-state functional reorganization in Parkinson's disease: an activation likelihood estimation meta-analysis, Tahmasian [/bib_ref]. PD patients demonstrated increased parietal cortical activation when off as compared to on medication [bib_ref] Functional neuroimaging of motor control in parkinson's disease: a meta-analysis, Herz [/bib_ref]. Mutation carriers of leucine-rich repeat kinase 2 (LRR2), a population with an elevated risk for PD, also demonstrated reduced connectivity between the IPL and posterior putamen, a task network region [bib_ref] Reorganization of corticostriatal circuits in healthy G2019S LRRK2 carriers, Helmich [/bib_ref]. Altered DMN activity was associated with decreased SNc activity and with disease severity in PD [bib_ref] Basal ganglia circuits changes in Parkinson's disease patients, Wu [/bib_ref]. In summary, as with AD, dysconnectivity of the DMN appears a core neural feature and may contribute to cognitive decline in PD. ## Other imaging studies Conventional T1/T2 MRI has not been successful at identifying changes in the SNc or other structures in PD. More advanced techniques including magnetization transfer [bib_ref] Magnetization transfer contrast (MTC) and tissue water proton relaxation in vivo, Wolff [/bib_ref] [bib_ref] Measuring in vivo myelination of human white matter fiber tracts with magnetization..., Rademacher [/bib_ref] [bib_ref] Axonal loss in multiple sclerosis lesions: magnetic resonance imaging insights into substrates..., Van Waesberghe [/bib_ref] [bib_ref] Improved segmentation of deep brain grey matter structures using magnetization transfer (MT)..., Helms [/bib_ref] , adiabatic and MR microscopy [bib_ref] Magnetic resonance imaging of the substantia nigra in Parkinson's disease, Lehéricy [/bib_ref] and relaxometry have revealed changes in the SNc in individuals with PD. In addition to neuromelanin imaging, iron imaging has been studied in PD. Brain iron content is increased in association with dopamine loss in the SNc in PD patients [bib_ref] Alterations in the levels of iron, ferritin and other trace metals in..., Dexter [/bib_ref] [bib_ref] Midbrain iron content in early Parkinson disease: a potential biomarker of disease..., Martin [/bib_ref] [bib_ref] Quantitative estimation of regional brain iron with magnetic resonance imaging, Martin [/bib_ref] [bib_ref] Movement disorders: role of imaging in diagnosis, Mascalchi [/bib_ref]. Thus, imaging iron with nonionizing MRI offers a strategy to detect neuroanatomical changes in PD. In a two-year follow-up study iron-related relaxation increased in the anterior globus pallidus, caudate nucleus and medial SNc and the changes in the globus pallidus and SNc were related to the development of MCI [bib_ref] Imaging brain iron and diffusion patterns: a follow-up study of Parkinson's disease..., Rossi [/bib_ref]. More recently an iron imaging study employing quantitative susceptibility mapping (QSM) reported increased QSM magnetic values in PD patients as compared to healthy controls. The difference between SNc iron deposition between healthy controls and patients with advanced PD was the most prominent, with SNc iron content correlated with symptom severity in PD . Positron emission tomography (PET) imaging of PD focused primarily on changes in dopamine transporter (DAT) density. In a recent meta-analysis DAT and vesicular monoamine transporter in early to moderate PD is decreased most significantly in the posterior putamen followed by the anterior putamen and caudate. Moreover, disease severity was linearly correlated with dopamine loss [bib_ref] Striatal dopamine in Parkinson's disease: a meta-analysis of imaging studies, Kaasinen [/bib_ref]. No studies to our knowledge investigated norepinephrine transporter (NET) density in PD. [(11)C]MENNET is a novel PET radiotracer with high affinity and selectivity for NET and may provide insights into noradrenergic dysfunction in PD [bib_ref] Compartmental modeling of [11C] MENET binding to the norepinephrine transporter in the..., Adhikarla [/bib_ref]. PET imaging has also demonstrated utility in tracking AD progression. Fluorine-18 fluorodeoxyglucose (FDG) PET imaging assessed regional cerebral glucose metabolism and showed decreased metabolic rates in the medial temporal lobes, lateral temporoparietal cortex, posterior cingulate cortex, and precuneus in patients with AD in comparison to normal aging [bib_ref] PET imaging in neurology: Alzheimer's and Parkinson's diseases, Sarikaya [/bib_ref]. In a post-mortem autoradiographic study of (S,S)-[(18)F]FMeNER-D(2), a selective ligand for NET, AD patients showed a reduction in NET density at the LC and thalamus in correlation with disease progression by Braak staging [bib_ref] The norepinephrine transporter (NET) radioligand (S, S)-[18F] FMeNER-D2 shows significant decreases in..., Gulyás [/bib_ref]. The size of the LC is below the spatial resolution of PET imaging but in vivo imaging of NET density in the thalamus may be useful to highlight early noradrenergic dysfunction in AD [bib_ref] The norepinephrine transporter (NET) radioligand (S, S)-[18F] FMeNER-D2 shows significant decreases in..., Gulyás [/bib_ref]. [fig_ref] TABLE 2 |: Summary of imaging findings in Alzheimer's Disease and Parkinson's Disease [/fig_ref] highlights VBM, neuromelanin and other imaging findings of AD and PD. ## Noradrenergic dysfunction in ad and pd ## An overview The LC sends noradrenergic projections to the hippocampus [bib_ref] Efferent projections of nucleus locus coeruleus: morphologic subpopulations have different efferent targets, Loughlin [/bib_ref] , amygdala [bib_ref] Catecholamine innervation of the basal forebrain II. Amygdala, suprarhinal cortex and entorhinal..., Fallon [/bib_ref] , and prefrontal cortex (PFC; [bib_ref] Locus coeruleus projections to cortex: topography, morphology and collateralization, Loughlin [/bib_ref]. Phasic LC activation in response to target stimuli facilitates anticipation [bib_ref] Impulse conduction properties of noradrenergic locus coeruleus axons projecting to monkey cerebrocortex, Aston-Jones [/bib_ref] and release of norepinephrine (NE) in the cortex [bib_ref] Detection thresholds for stimuli in humans and monkeys: comparison with threshold events..., Mountcastle [/bib_ref] [bib_ref] An integrative theory of locus coeruleusnorepinephrine function: adaptive gain and optimal performance, Aston-Jones [/bib_ref] prior to a motivated action. NE signals in the PFC regulate attention, learning and working memory [bib_ref] From arousal to cognition: the integrative position of the prefrontal cortex, Robbins [/bib_ref]. On the other hand, NE interacts with other catecholamines like dopamine (DA) to support these functions, with NE often playing a regulatory role in DA signaling. For example, chemical modulation or electrical stimulation of the LC increases the extracellular concentrations of both NE and DA [bib_ref] CNS dopamine transmission mediated by noradrenergic innervation, Smith [/bib_ref]. The ability of the LC to effect direct and indirect control of catecholamines has important implications on the arousal state as well as autonomic, motor, sensory, and cognitive functions. Both AD and PD involve noradrenergic dysfunction. For example, orthostatic and postprandial hypotension in PD can be related to autonomic dysfunction and NE deficiency [bib_ref] Autonomic dysfunction in Parkinson disease, Kaufmann [/bib_ref]. Disruption of the circadian rhythm and arousal/wakefulness cycles manifest in both diseases and are associated with noradrenergic dysfunction [bib_ref] The locus coeruleus-noradrenergic system: modulation of behavioral state and state-dependent cognitive processes, Berridge [/bib_ref] [bib_ref] The locus ceruleus norepinephrine system: functional organization and potential clinical significance, Benarroch [/bib_ref]. In PD, decreased CSF concentration of NE is associated with freezing of gait and administration of a NE precursor improves gait . Animal studies involving degeneration of the LC also support noradrenergic dysfunction in AD. In transgenic mice carrying homozygous forms of amyloid precursor protein/presenilin 1 genes, induction of LC degeneration with N-(2-chloroethyl)-N-ethyl-bromo-benzylamine resulted in an exacerbation of olfactory memory deficits and weakening of ## Non-functional imaging in ad and pd ## Ad Voxel-based morphometry Hippocampal atrophyMedial temporal atrophy in limbic-predominant AD [bib_ref] Neuroimaging correlates of pathologically defined subtypes of Alzheimer's disease: a case-control study, Whitwell [/bib_ref] Severe cortical atrophy in hippocampal-sparing AD [bib_ref] Neuroimaging correlates of pathologically defined subtypes of Alzheimer's disease: a case-control study, Whitwell [/bib_ref] More extensive occipital GM atrophy in early-vs. late-onset AD [bib_ref] The topography of grey matter involvement in early and late onset Alzheimer's..., Frisoni [/bib_ref] More extensive hippocampal atrophy in late-vs. early-onset AD [bib_ref] The topography of grey matter involvement in early and late onset Alzheimer's..., Frisoni [/bib_ref] Neuromelanin Imaging LC demonstrates neuromelanin signal attenuation in MCI [bib_ref] In vivo MRI assessment of the human locus coeruleus along its rostrocaudal..., Betts [/bib_ref] LC volume decreases by 8.4% with progression to each consecutive Braak stage, as measured by neuromelanin signalsPET Imaging PET imaging with F 18 -FDG radioligand reveals decreased cerebral metabolic rates in the medial temporal lobes, lateral temporoparietal cortex, posterior cingulate cortex and precuneus [bib_ref] PET imaging in neurology: Alzheimer's and Parkinson's diseases, Sarikaya [/bib_ref] (S,S)-[(18)F]FMeNER-D(2), a radioligand specific for norepinephrine transporter (NET), demonstrates reduced NET density in the LC and thalamus on postmortem brains [bib_ref] The norepinephrine transporter (NET) radioligand (S, S)-[18F] FMeNER-D2 shows significant decreases in..., Gulyás [/bib_ref] PD Voxel-based morphometry GM volume reductions in the left frontal temporal cortices encompassing inferior frontal and superior temporal gyri [bib_ref] Voxel-wise meta-analysis of gray matter abnormalities in idiopathic Parkinson's disease, Pan [/bib_ref] GM reductions in left insular cortex [bib_ref] Voxel-wise meta-analysis of gray matter abnormalities in idiopathic Parkinson's disease, Pan [/bib_ref] PD patients with dementia have more prominent reductions in frontal regional GM [bib_ref] Structural and functional imaging study in dementia with Lewy bodies and Parkinson's..., Borroni [/bib_ref] Iron Imaging Brain iron content in the SNc is increased in PD patients, in association with loss of DA neurons [bib_ref] Alterations in the levels of iron, ferritin and other trace metals in..., Dexter [/bib_ref] [bib_ref] Midbrain iron content in early Parkinson disease: a potential biomarker of disease..., Martin [/bib_ref] [bib_ref] Quantitative estimation of regional brain iron with magnetic resonance imaging, Martin [/bib_ref] [bib_ref] Movement disorders: role of imaging in diagnosis, Mascalchi [/bib_ref] Increased iron-content in the globus pallidus and anterior and medial SNc, in correlation with MCI in PD [bib_ref] Imaging brain iron and diffusion patterns: a follow-up study of Parkinson's disease..., Rossi [/bib_ref] Iron content in the SNc as measured by quantitative susceptibility mapping correlates with the symptom severity of PD Neuromelanin Imaging SNc and LC demonstrate reduction in signal intensity in PD [bib_ref] Spontaneous fluctuations in brain activity observed with functional magnetic resonance imaging, Fox [/bib_ref] [bib_ref] Functional connectivity mapping of the human precuneus by resting state fMRI, Zhang [/bib_ref] [bib_ref] Resting-state functional connectivity of the locus coeruleus in humans: in comparison with..., Zhang [/bib_ref] Differences on neuromelanin-sensitive MRI distinguish essential tremor from PD and early-stage PD from healthy-controls [bib_ref] A method for using blocked and event-related fMRI data to study "resting..., Fair [/bib_ref] [bib_ref] Spontaneous fluctuations in brain activity observed with functional magnetic resonance imaging, Fox [/bib_ref] PET Imaging Decline in dopamine transporter occurs most significantly in the posterior putamen followed by anterior putamen and caudate and there is a correlation between dopamine loss and disease severity [bib_ref] Striatal dopamine in Parkinson's disease: a meta-analysis of imaging studies, Kaasinen [/bib_ref] olfactory discrimination [bib_ref] Locus coeruleus degeneration exacerbates olfactory deficits in APP/PS1 transgenic mice, Rey [/bib_ref]. As 90% of patients with early AD experience olfactory dysfunction [bib_ref] Olfaction in neurodegenerative disorder, Hawkes [/bib_ref] , the latter study suggests a potential role of the LC in the etiological processes of early AD. In another study of transgenic mice with a low amyloid load, LC degeneration elicited with the same chemical triggered massive glial inflammation and augmented the deposition of amyloid plaques [bib_ref] Locus ceruleus degeneration promotes Alzheimer pathogenesis in amyloid precursor protein 23 transgenic..., Heneka [/bib_ref]. Together, these studies provide substantial evidence for noradrenergic dysfunction in AD and PD. ## Noradrenergic dysfunction and the hippocampus in ad and pd AD and PD are known to have distinct pathology, yet they share many cognitive and behavioral manifestations as the illness progresses, including dementia, motor dysfunction, and behavioral disinhibition. Both AD and PD are associated with significant degeneration of the LC [bib_ref] Loss of neurons of origin of the adrenergic projection to cerebral cortex..., Bondareff [/bib_ref] [bib_ref] Loss of nerve cells from locus coeruleus in Alzheimer's disease is topographically..., Marcyniuk [/bib_ref] [bib_ref] Norepinephrine: the redheaded stepchild of Parkinson's disease, Rommelfanger [/bib_ref]. In PD the degeneration of the LC occurs earlier and in greater magnitude in comparison to the SNc [bib_ref] Norepinephrine: the redheaded stepchild of Parkinson's disease, Rommelfanger [/bib_ref]. AD and PD also share extensive pathology in the hippocampus. In AD, memory deficits tend to manifest early and correlate with hippocampal pathology [bib_ref] A sequence of cytoskeleton changes related to the formation of neurofibrillary tangles..., Braak [/bib_ref]. In PD, loss of NE-containing neurons is associated with cognitive decline [bib_ref] Parkinson's disease and dementia: norepinephrine and dopamine in locus ceruleus, Cash [/bib_ref]. As the LC projects heavily to the hippocampus and parahippocampal formation , loss of noradrenergic signaling may lead to hippocampal dysfunction and memory deficits. In patients with MCI and a Clinical Dementia Rating score of 0.5, decreased LC connectivity to the left parahippocampal gyrus was associated with lower memory performance [bib_ref] Relevance of parahippocampal-locus coeruleus connectivity to memory in early dementia, Jacobs [/bib_ref]. These findings support the proposition that noradrenergic dysfunction in AD and PD contributes to memory impairment. In rodent studies, the LC has extensive anatomical connectivity (likely stronger than the ventral tegmental area) to the hippocampus [bib_ref] Locus coeruleus and dopaminergic consolidation of everyday memory, Takeuchi [/bib_ref]. Optical activation of LC terminals in the dorsal hippocampus enhances object location memory [bib_ref] Dopamine release from the locus coeruleus to the dorsal hippocampus promotes spatial..., Kempadoo [/bib_ref]. Further, depletion of noradrenergic terminals in the hippocampus by neurotoxics or ablation of LC neurons impaired location memory in mice [bib_ref] Selective noradrenaline depletion impairs working memory and hippocampal neurogenesis, Coradazzi [/bib_ref] [bib_ref] Hippocampal release of dopamine and norepinephrine encodes novel contextual information, Moreno-Castilla [/bib_ref]. Numerous other studies have implicated noradrenergic control of other aspects of cognitive functioning [bib_ref] Noradrenergic control of error perseveration in medial prefrontal cortex, Caetano [/bib_ref]. Reduced cortical noradrenergic neurotransmission was associated with increased neophobia and impaired spatial memory in aged rats [bib_ref] Reduced cortical noradrenergic neurotransmission is associated with increased neophobia and impaired spatial..., Collier [/bib_ref]. Work on transgenetics associated reduced tissue levels of NE and LC degeneration with behavioral phenotypes of mouse models of AD [bib_ref] Reduced tissue levels of noradrenaline are associated with behavioral phenotypes of the..., Francis [/bib_ref] [bib_ref] Locus coeruleus degeneration exacerbates olfactory deficits in APP/PS1 transgenic mice, Rey [/bib_ref]. In transgenic mice that accumulated amyloid burden at early ages, treatment with a NE precursor increased CNS NE levels and improved learning in the Morris water maze [bib_ref] The noradrenaline precursor L-DOPS reduces pathology in a mouse model of Alzheimer's..., Kalinin [/bib_ref]. Noradrenergic innervations from the LC are needed to maintain beta amyloid clearance, and LC degeneration could contribute to the pathogenesis of AD [bib_ref] Noradrenaline deficiency in brain increases β-amyloid plaque burden in an animal model..., Kalinin [/bib_ref]. The potential roles of NE in the etiological processes of AD and PD and memory impairment can also be examined with pharmacologic manipulations. Administration of methylphenidate, a stimulant medication known to increase NE and DA levels and commonly used in the treatment of ADHD [bib_ref] Differential effects of adjunctive methylphenidate and citalopram on extracellular levels of serotonin,..., Weikop [/bib_ref] , improved word recall when taken before learning [bib_ref] Methylphenidate significantly improves declarative memory functioning of adults with ADHD, Verster [/bib_ref] [bib_ref] Methylphenidate produces selective enhancement of declarative memory consolidation in healthy volunteers, Linssen [/bib_ref]. Administration of methylphenidate after learning also increased 1-week retention of intentionally and casually learned information [bib_ref] Age-dependent and age-independent human memory persistence is enhanced by delayed posttraining methylphenidate..., Izquierdo [/bib_ref]. In a fMRI study, methylphenidate administration increased connectivity between the LC and the hippocampus [bib_ref] The effects of methylphenidate on resting-state functional connectivity of the basal nucleus..., Kline [/bib_ref]. The latter study implicates NE in producing a measurable change in connectivity between the LC and hippocampus, both of which suffer significant neuronal loss and functional dysconnectivity in AD and PD. Although methylphenidate increases the extracellular levels of DA in addition to NE, the finding of increased connectivity between the LC and hippocampus and its association with improved memory function helps establish a specific link to the noradrenergic system. Together, considered in the context of LC degeneration and the resulting denervation of the hippocampus, these findings support noradrenergic dysfunction as a neural mechanism of memory impairment in AD and PD. ## Noradrenergic connectivity dysfunction in ad and pd The LC undergoes early and significant deterioration and recent imaging studies suggest LC functional dysconnectivity that involves brain regions other than the hippocampus in AD. A recent work employed Granger causality mapping to create a model of directed connectivity and characterized regional functional coupling in AD and healthy controls [bib_ref] Investigating focal connectivity deficits in Alzheimer's disease using directional brain networks derived..., Zhao [/bib_ref]. With group differences determined with a generative model of pathology, the authors identified the LC and right orbitofrontal cortex as the two foci of disruption in AD. These findings provide further support of a critical role of the LC in functional impairment in AD. The saliency network (SAN) facilitates network activity transition from the DMN to the frontoparietal network during task challenges [bib_ref] Dissociable intrinsic connectivity networks for salience processing and executive control, Seeley [/bib_ref]. For instance, the right anterior insula temporally precedes activation of the central executive network and de-activation of the DMN in task based paradigms [bib_ref] A critical role for the right fronto-insular cortex in switching between central-executive..., Sridharan [/bib_ref]. Temporal control over these network activities is central to an intact cognition, and impaired suppression of the DMN is associated with impaired performance. As discussed earlier, abnormal DMN connectivity represents one of the most consistent finding in AD. The SAN, specifically the anterior insula, also demonstrates abnormal connectivity in AD. DMN and SAN dysfunction may result in large part from noradrenergic dysfunction in AD. The insula responds strongly to deviances in a stream of continuous stimuli and does so across auditory, visual, and tactile modalities [bib_ref] The functional neuroanatomy of target detection, an fMRI study of visual and..., Linden [/bib_ref] [bib_ref] A multimodal cortical network for the detection of changes in the sensory..., Downar [/bib_ref] [bib_ref] The effect of task relevance on the cortical response to changes in..., Downar [/bib_ref] [bib_ref] Where and when the anterior cingulate cortex modulates attentional response: combined fMRI..., Crottaz-Herbette [/bib_ref]. Regions of the SAN coactivate in response to faces of loved ones [bib_ref] The neural correlates of maternal and romantic love, Bartels [/bib_ref] , pleasurable touch [bib_ref] How do you feel? Interoception: the sense of the physiological condition of..., Craig [/bib_ref] , emotional dimensions of pain [bib_ref] Functional imaging of brain responses to pain. A review and meta-analysis, Peyron [/bib_ref] , "chills" to music [bib_ref] Intensely pleasurable responses to music correlate with activity in brain regions implicated..., Blood [/bib_ref] and other forms of saliency [bib_ref] Does rejection hurt? An FMRI study of social exclusion, Eisenberger [/bib_ref] [bib_ref] Empathy for pain involves the affective but not sensory components of pain, Singer [/bib_ref] , all supported by LC activity and an intact system of physiological arousal. As the LC shores up arousal and responses to salient stimuli, noradrenergic signaling is in a position to influence SAN and DMN network activities. It is highly likely that deficient noradrenergic signaling may compromise functional coupling between the SAN and DMN and, as a result, deficits in the suppression of DMN activity to meet task demands. In support, atomoxetine, a selective NE reuptake inhibitor commonly used in the treatment of ADHD, increased connectivity between the right anterior insula and dorsolateral prefrontal cortex [bib_ref] Noradrenaline transporter blockade increases fronto-parietal functional connectivity relevant for working memory, Hernaus [/bib_ref]. Increased connectivity between these regions correlated with decreased reaction time variability in an N-back working memory test and predicted auditory digit span of the Wechsler Adult Intelligence Scale. In a predictive learning task atomoxetine facilitated extinction learning along with increased activity in the insula [bib_ref] Noradrenergic stimulation modulates activation of extinction-related brain regions and enhances contextual extinction..., Lissek [/bib_ref]. In a study of inhibitory control, atomoxetine improved performance in a stop-signal task in correlation with increased insula activity [bib_ref] Atomoxetine modulates right inferior frontal activation during inhibitory control: a pharmacological functional..., Chamberlain [/bib_ref]. These findings suggest that NE influences insula activity and improves a broad spectrum of cognitive functions, providing a pathway whereby noradrenergic dysfunction may contribute to aberrant SAN and DMN activations in AD. As discussed earlier, the IPL is part of the DMN [bib_ref] Subspecialization in the human posterior medial cortex, Bzdok [/bib_ref] and consistently demonstrates abnormal connectivity in PD. Engagement in a cognitive task is associated with significant posterior IPL deactivation [bib_ref] Functional connectivity in the resting brain: a network analysis of the default..., Greicius [/bib_ref] [bib_ref] The brain's default network: anatomy, function, and relevance to disease, Buckner [/bib_ref]. Altered IPL and DMN connectivity in PD during task engagement [bib_ref] Dysfunction of the default mode network in Parkinson disease: a functional magnetic..., Van Eimeren [/bib_ref] [bib_ref] Dopaminergic modulation of the default mode network in Parkinson's disease, Delaveau [/bib_ref] [bib_ref] Salience and default mode network coupling predicts cognition in aging and Parkinson's..., Putcha [/bib_ref] may represents a failure in activity transition between functional networks [bib_ref] Salience and default mode network coupling predicts cognition in aging and Parkinson's..., Putcha [/bib_ref]. Successful task engagement requires suppression of internally driven processes such as mind-wondering, and the salience network monitors external stimuli and facilitates the suppression of DMN activity and these internally driven processes. One study discussed earlier demonstrated reduced anti-correlation between the SAN and DMN in link with diminished executive functioning, psychomotor speed and verbal memory [bib_ref] Dysfunction of the default mode network in Parkinson disease: a functional magnetic..., Van Eimeren [/bib_ref]. The LC responds to saliency and functionally connects to key regions within the DMN including the IPL, PCC, and precuneus . It is also one of the first brain regions affected in PD [bib_ref] Norepinephrine: the redheaded stepchild of Parkinson's disease, Rommelfanger [/bib_ref]. Thus, early deterioration and dysconnectivity of the LC is positioned to effect cognitive decline in PD. Pharmacological investigations provide more evidence relating cognitive decline to LC dysfunction in PD. Enhancing noradrenergic signaling improves cognitive performance in PD [bib_ref] Attentional deficits in Parkinson's disease: partial reversibility with naphtoxazine (SDZ NVI-085), a..., Bédard [/bib_ref] [bib_ref] Atomoxetine for the treatment of executive dysfunction in Parkinson's disease: a pilot..., Marsh [/bib_ref] [bib_ref] Atomoxetine for depression and other neuropsychiatric symptoms in Parkinson disease, Weintraub [/bib_ref] [bib_ref] Deficits in inhibitory control and conflict resolution on cognitive and motor tasks..., Obeso [/bib_ref] [bib_ref] Targeting impulsivity in Parkinson's disease using atomoxetine, Kehagia [/bib_ref] [bib_ref] Multiple modes of impulsivity in Parkinson's disease, Nombela [/bib_ref] [bib_ref] Improving response inhibition in Parkinson's disease with atomoxetine, Ye [/bib_ref]. For example, naphtoxazine, a selective α1 agonist, as compared to placebo, ameliorated performance deficits in a reaction time task and the Stroop task in PD patients [bib_ref] Attentional deficits in Parkinson's disease: partial reversibility with naphtoxazine (SDZ NVI-085), a..., Bédard [/bib_ref]. A recent study of healthy adults reported that LC functional connectivity increased in the Stroop task, implicating the noradrenergic system during interference control [bib_ref] Differential involvement of brainstem noradrenergic and midbrain dopaminergic nuclei in cognitive control, Köhler [/bib_ref]. In other studies, atomoxetine improved performance on measures of global cognition (mini-mental status exam) and executive functions, as evaluated by the Frontal Systems Behavior Scale and Connors Adult ADHD Rating Scale [bib_ref] Atomoxetine for the treatment of executive dysfunction in Parkinson's disease: a pilot..., Marsh [/bib_ref] [bib_ref] Atomoxetine for depression and other neuropsychiatric symptoms in Parkinson disease, Weintraub [/bib_ref] in patients with PD. In a stop signal task, atomoxetine improved stop signal reaction times in a subset of PD patients [bib_ref] Deficits in inhibitory control and conflict resolution on cognitive and motor tasks..., Obeso [/bib_ref] [bib_ref] Targeting impulsivity in Parkinson's disease using atomoxetine, Kehagia [/bib_ref] [bib_ref] Multiple modes of impulsivity in Parkinson's disease, Nombela [/bib_ref] [bib_ref] Improving response inhibition in Parkinson's disease with atomoxetine, Ye [/bib_ref]. Atomoxetine also restored functional connectivity between the presupplementary motor area, inferior frontal gyrus, and subthalamic nucleus [bib_ref] Atomoxetine restores the response inhibition network in Parkinson's disease, Rae [/bib_ref] of the cortical subcortical circuit to support response inhibition in PD patients. As discussed earlier, atomoxetine improved performance on stop signal tasks along with increased insular activity and engagement of the saliency network in healthy adults [bib_ref] Atomoxetine modulates right inferior frontal activation during inhibitory control: a pharmacological functional..., Chamberlain [/bib_ref]. This is important because control of the DMN relies critically on SAN function [bib_ref] Salience network integrity predicts default mode network function after traumatic brain injury, Bonnelle [/bib_ref] [bib_ref] Abnormal resting-state functional connectivity of insular subregions and disrupted correlation with working..., Zhao [/bib_ref]. In a study of medication-naïve adults with ADHD, atomoxetine facilitated suppression of the DMN in correlation with improved clinical symptoms [bib_ref] Atomoxetine treatment strengthens an anticorrelated relationship between functional brain networks in medication-naïve..., Lin [/bib_ref]. Together, these findings support noradrenergic dysfunction and the efficacy of noradrenergic agents in improving cognition in PD. [fig_ref] TABLE 3 |: Summary of functional connectivity changes of the default mode and saliency networks... [/fig_ref] highlights key findings of DMN dysfunction in AD and PD. [fig_ref] FIGURE 1 |: Hypothetical models of the influence of LC on the activity of the... [/fig_ref] illustrate a hypothetical model of noradrenergic circuit regulation of functional neural networks. ## Conclusions and future research Dysfunction of the noradrenergic system is a key feature of both AD and PD. The deficit is manifested in post-mortem studies which reveal loss of NE neurons in the LC in both diseases. It is also manifested in structural brain imaging as atrophic changes in structures connected with the LC and in functional imaging as changes in network activity and connectivity. Key networks implicated in noradrenergic dysfunction include the hippocampus, DMN and the SAN. Specifically, noradrenergic dysfunction may disrupt the ability to monitor external stimuli and shift attentional demands accordingly. These shifts in attentional demand rely on the SAN to regulate and suppress the DMN during task-oriented processes. Pharmacologic studies with NE reuptake inhibitors highlight the role of NE in remediating aberrations in functional connectivity and deficits ## Functional connectivity changes in ad and pd ## Ad dmn Reduced DMN connectivity during resting state in early stage MCI to late stage AD and in asymptomatic patients at genetic risk for AD [bib_ref] Dissociable intrinsic connectivity networks for salience processing and executive control, Seeley [/bib_ref] [bib_ref] Divergent network connectivity changes in behavioural variant frontotemporal dementia and Alzheimer's disease, Zhou [/bib_ref] [bib_ref] Effect of APOE ε4 status on intrinsic network connectivity in cognitively normal..., Machulda [/bib_ref] [bib_ref] Functional connectivity in autosomal dominant and late-onset Alzheimer disease, Thomas [/bib_ref] [bib_ref] Differentially disrupted functional connectivity of the subregions of the inferior parietal lobule..., Wang [/bib_ref] Reduced DMN connectivity in AD correlates with disease severity as measured by the Clinical Dementia Rating [bib_ref] Basal functional connectivity within the anterior temporal network is associated with performance..., Gour [/bib_ref] [bib_ref] Abnormal resting-state functional connectivity of insular subregions and disrupted correlation with working..., Zhao [/bib_ref] SAN Abnormal SAN connectivity across disease spectrum and those at genetic risk of AD [bib_ref] Dysfunction of the default mode network in Parkinson disease: a functional magnetic..., Van Eimeren [/bib_ref] [bib_ref] Divergent network connectivity changes in behavioural variant frontotemporal dementia and Alzheimer's disease, Zhou [/bib_ref] Anterior insular cortex, a key hub of the SAN, demonstrates hyperconnectivity to the SAN in AD [bib_ref] Dopaminergic modulation of the default mode network in Parkinson's disease, Delaveau [/bib_ref] PD DMN The most consistent connectivity change in PD involves a region of the R posterior IPC, which is part of the DMN [bib_ref] Measuring in vivo myelination of human white matter fiber tracts with magnetization..., Rademacher [/bib_ref] and altered IPC connectivity is more prominent in PD patients with cognitive impairment [bib_ref] Midbrain iron content in early Parkinson disease: a potential biomarker of disease..., Martin [/bib_ref] [bib_ref] Quantitative estimation of regional brain iron with magnetic resonance imaging, Martin [/bib_ref] Slower processing speeds in PD are associated with decreased DMN connectivity at rest, specifically between the posterior cingulate, medial prefrontal and inferior parietal nodes [bib_ref] Subspecialization in the human posterior medial cortex, Bzdok [/bib_ref] SAN Reduced anti-correlation between the SAN and DMN is associated with worse performance on tests of executive functioning, psychomotor speed and verbal memory [bib_ref] Basal ganglia circuits changes in Parkinson's disease patients, Wu [/bib_ref] AD, Alzheimer's disease; PD, Parkinson's disease; DMN, default mode network; MCI, mild cognitive impairment; SAN, salience network; R posterior IPC, right posterior inferior parietal cortex. in cognitive performance. Together, the studies may facilitate research of the etiology of AD and PD as well as development of pharmacotherapy for these debilitating illnesses [bib_ref] Novel GLP-1 (glucagon-like peptide-1) analogues and insulin in the treatment for Alzheimer's..., Calsolaro [/bib_ref] [bib_ref] Memantine ER/donepezil: a review in Alzheimer's disease, Greig [/bib_ref] [bib_ref] Aβ-degrading proteases: therapeutic potential in Alzheimer disease, Leissring [/bib_ref] [bib_ref] Dopaminergic therapies for nonmotor symptoms in Parkinson's disease, Schaeffer [/bib_ref]. Despite the advances in methodology, there are limitations in current imaging techniques and pharmacological approaches to precisely localize the LC and quantify LC functions. For instance, neuromelanin imaging provides data with low spatial resolution with respect to the size of the LC and with signal nonuniformity in a given plane [bib_ref] Neuromelanin-sensitive MRI, Sasaki [/bib_ref]. These issues may compromise the utility of neuromelanin imaging in quantifying signal alteration in the LC. Pharmacological manipulations have been used to investigate noradrenergic dysfunction in AD and PD. However, many "noradrenergic" agents are not specific to the noradrenergic system. Further, because of co-localization of synaptic NE and DA, the findings from pharmacological studies would require careful interpretation. A few potential research directions can be explored to further our knowledge of LC function in health and illness. First, despite electrophysiological studies characterizing the roles of LC neurons in cognitive performance in behaving primates, imaging work in humans has been hampered by the small size of LC. Thus, studies combining neuromelanin imaging to more accurately localize the LC and fMRI to explore taskrelated activations are needed to fully understand how LC and LC connectivity with SAN, DMN, and frontoparietal network partake in task challenges. As VTA/SNc and LC are both involved in saliency response, it would be of great interest to evaluate how the two systems are differentially involved in cognitive performance and whether reward plays a specific role in differentiating noradrenergic and dopaminergic circuit functions. Likewise, studies of pharmacological manipulations can take advantage of anatomical localization of the LC and more precisely pinpoint how noradrenergic agents influence regional activity and connectivity during cognitive performance. Second, there have been neuromelanin imaging studies of the VTA/SNc in PD but less work has focused on elucidating signal changes in the LC in either PD or AD. In particular, longitudinal studies to characterize how LC signals evolve during healthy aging and in individuals with MCI or at risk of developing PD and AD would be instrumental in determining the significance of neuromelanin imaging as a tool in predicting onset and progression of the illnesses. In individuals at risk for PD, it would also be of interest to contrast findings on VTA/SNc and LC to determine how the noradrenergic and dopaminergic circuits are differentially involved in the pathogenesis of PD. Likewise, longitudinal studies of PD and AD patients undergoing treatment may benefit from knowledge whether LC neuromelanin signal intensity may serve as a biomarker of treatment outcomes. Third, the LC projects to multiple brain regions and the interaction of LC with these neural networks are likely complex and defy simple correlation analyses. More sophisticated analytical tools such as Granger causality analysis [bib_ref] Functional connectivity delineates distinct roles of the inferior frontal cortex and presupplementary..., Duann [/bib_ref] [bib_ref] Anticipating conflict: neural correlates of a Bayesian belief and its motor consequence, Hu [/bib_ref] and detrended partial cross correlation [bib_ref] Detrended partial cross correlation for brain connectivity analysis, Ide [/bib_ref] [bib_ref] Time scale properties of task and resting-state functional connectivity: Detrended partial cross-correlation..., Ide [/bib_ref] will be useful in delineating the direction of influence and distinguish direct functional interaction from influences via a common "third party." These analyses would be tremendously useful in confirming the hypothesis that the LC response to saliency and its projection to the SAN facilitates activity transition from the DMN to frontoparietal network for goal-directed behavior. Together, these new studies will address many unanswered questions in cognitive and clinical neuroscience, and the findings would not only advance knowledge but also better patient care. # Author contributions All authors listed have made a substantial, direct and intellectual contribution to the work, and approved it for publication. # Acknowledgments This review was supported by NIH grants MH113134, DA023248, and AA021449. We declare no conflicts of interest in the current work. The NIH was otherwise not involved in the preparation of or in the decision to publish this review. [fig] FIGURE 1 |: Hypothetical models of the influence of LC on the activity of the saliency network (SAN), frontoparietal task network (FPN), and default model network (DMN). Box position with respect to the dashed line represents changes in activity level from a resting to task state, and arrows indicate the location of LC action in response to a task challenge. (A) LC increases activity of the SAN and FPN and suppresses activity of the DMN; (B) LC increases activity of the SAN, which in turn increases activity of the FPN and suppresses activity of the DMN. [/fig] [table] TABLE 1 |: Summary of anatomical changes in Alzheimer's Disease and Parkinson's Disease. [/table] [table] TABLE 2 |: Summary of imaging findings in Alzheimer's Disease and Parkinson's Disease. [/table] [table] TABLE 3 |: Summary of functional connectivity changes of the default mode and saliency networks in Alzheimer's disease and Parkinson's disease. [/table]

Using Surveillance Data to Inform Community Action: The Effect of Alcohol Sale Restrictions on Intentional Injury-related Ambulance Pickups Youth violence disproportionately affects inner city, urban minority communities in the USA. This article illustrates the use of surveillance data to inform and evaluate community action directed at this serious problem. Community efforts in response to surveillance data indicating high rates of violence surrounding convenience stores with unrestricted alcohol beverage licenses provided a natural experiment to examine the impact of imposing licensing restrictions on intentional injury rates. Rates of ambulance pickups for intentional injuries in the 15-to 24-year-old population in five census tracts where alcoholic beverage sales were restricted were compared to five census tracts with similar demographic characteristics near stores where restrictions were not instituted. Time periods included an 18-month baseline period, a 6-month period during which restrictions were in effect in the intervention communities, and an 18-month period following lifting of this restriction resulting from legal action by store owners. The monthly average rate of ambulance pickups for violent injuries showed a significantly greater baseline-to-intervention phase decrease in the intervention communities (i.e., from 19.6 to 0 per 1,000) than in the control communities (i.e., 7.4 to 3.3 per 1,000). This rate subsequently increased to 11.4 in the intervention communities after the restriction was removed. This study illustrates the potential value of surveillance data for guiding community mobilization efforts and for evaluating the impact of such efforts. It also demonstrates the potential impact of restricting inexpensive, single-serve alcoholic beverages on rates of violence. # Introduction Intentional injury among youth is a serious public health problem. Homicide is the second leading cause of death among 15-to 24-year olds in the USA (Centers for Disease Control and Prevention 2010). Although mortality represents the most severe outcome of violence, non-fatal injuries are much more common and may have serious consequences. Disability, disruption of social services, health disparities, increased health care cost, and reduction in property values are all associated with violence. The frequency of non-fatal injuries also predicts more severe forms of violence such as fatal injuries and disabilities. For each fatal injury, there are approximately 20 to 40 non-fatal violence victims who receive treatment in hospitals. The repercussions of youth violence are far reaching and impact many facets of society. Its consequences include increased burden on health and welfare services, a loss of productivity, depreciation of property values, and neighborhood decay. The impact of death, injury, and disability among youth affects not only the victims, but extends to their communities, families, and friends; US Department of Health and Human Services 2001). Public health providers and communities often rely on surveillance data to monitor fatal and non-fatal outcomes. Youth violence surveillance entails the analysis of primary and secondary datasets garnered from emergency rooms, hospital discharges, law enforcement, schools, juvenile justice systems, public health surveys, medical examiners, and vital statistics. Reports based on these data may be used to guide, implement, and monitor prevention efforts by community-university partnerships in areas where youth violence is pervasive. This article illustrates this process, based on a study that used local surveillance data to examine the impact of a community-based youth violence prevention effort that restricted the sale of inexpensive, single-serve alcohol beverages at convenience grocery stores. The relation between alcohol outlet density and violence is well established in the literature. Numerous studies examining this spatial relation have reported significant associations between alcohol outlet density and violence [bib_ref] Immigrants and violence: The importance of neighborhood context, Alaniz [/bib_ref] [bib_ref] Alcohol consumption, alcohol outlets, and the risk of being assaulted with a..., Branas [/bib_ref] [bib_ref] Neighborhood level spatial analysis of the relationship between alcohol outlet density and..., Britt [/bib_ref] [bib_ref] The relationship between liquor outlet density and injury and violence in New..., Escobedo [/bib_ref] [bib_ref] Alcohol outlets and violent crime in, Franklin [/bib_ref] [bib_ref] Spatial dynamics of alcohol availability, neighborhood structure and violent crime, Gorman [/bib_ref] [bib_ref] Reassessing the alcoholviolence linkage: Results from a multiethnic city, Nielsen [/bib_ref] [bib_ref] The risk of assaultive violence and alcohol availability in Los Angeles county, Scribner [/bib_ref] [bib_ref] Alcohol availability and homicide in New Orleans: Conceptual considerations for small area..., Scribner [/bib_ref] [bib_ref] Alcohol outlet density and violence: A geospatial analysis, Zhu [/bib_ref]. Researchers have also observed that alcohol outlets are often located within communities that are already at increased risk for violence and other health-related problems. These communities are generally located in inner-city areas with high rates of poverty where the majority of residents represent ethnic minorities (LaVeist and Wallace 2000; [bib_ref] Neighbourhood deprivation and alcohol consumption: Does the availability of alcohol play a..., Pollack [/bib_ref]. These communities are often characterized by social disorganization such that crime and violence are highly prevalent and considered normative. This is supported by a recent study that reported a strong positive association between risk for violence and the number and density of alcohol outlets in areas where social disorganization exists . Areas with high densities of alcohol outlets may also serve as an "attractor" to youth and uninhibited behavior, contributing to a variety of undesirable activities such as illegal drug sales, prostitution, and violence [bib_ref] Immigrants and violence: The importance of neighborhood context, Alaniz [/bib_ref]. Alcohol outlets in urban areas that are licensed for offpremise sales (i.e., licensed to sell alcohol to be consumed outside of the store) represent a particular problem based on studies showing their relation to elevated rates of violence [bib_ref] The spatial dynamics of violence and alcohol outlets, Lipton [/bib_ref] [bib_ref] The risk of assaultive violence and alcohol availability in Los Angeles county, Scribner [/bib_ref] [bib_ref] Alcohol availability and homicide in New Orleans: Conceptual considerations for small area..., Scribner [/bib_ref]. Although the same rationale regarding alcohol outlet density as a causal factor for violence can be applied to single-serve container sales, outlets that primarily sell single-serve containers may have a compounded impact on the prevalence of violence by increasing the affordability of alcoholic beverages. Although such a relation appears quite plausible, few studies have examined the impact of single-serve alcoholic beverages or malt liquor beverages on violence [bib_ref] Type of alcohol drink and exposure to violence: An emergency department study, Chavira [/bib_ref] [bib_ref] Alcohol and malt liquor availability and promotion and homicide in inner cities, Jones-Webb [/bib_ref]. These studies related to single-serve alcoholic beverages had focused on promotion and availability of these beverages as factors for racial disparities in homicide rates [bib_ref] Alcohol and malt liquor availability and promotion and homicide in inner cities, Jones-Webb [/bib_ref] and analyzing the consumption of such beverages as a predictor for exposure to violence [bib_ref] Type of alcohol drink and exposure to violence: An emergency department study, Chavira [/bib_ref]. Unlike the above studies, a more specific evaluation, conducted by, reported a significant association between the availability of refrigerated shelf space dedicated to the sale of single-serve beverages and neighborhood-level violence. However, a recent quasiexperimental study examining the impact of malt liquor sale restrictions on crimes reported a mixed finding. Although the study showed a reduction in disorderly conduct in the restricted sales areas, there was an increase in larceny, burglary, and aggravated assault. Evidently, there is a scant and inconsistent body of scientific literature that has specifically examined the impact of single-serve alcohol sales or alcohol licensing restrictions on violence-related injuries. The limited number of studies addressing this issue may largely reflect the difficulty of assessing the impact of community-wide prevention efforts. The Centers for Disease Control, in its publication, Guide to Community Preventive Services [bib_ref] The effectiveness of limiting alcohol outlet density as a means of reducing..., Campbell [/bib_ref] , identified a gap in the literature regarding how local decisions are made and emphasized the need to examine policies affecting alcohol outlet density. The Guide acknowledges that, although this type of study may be difficult to conduct, the results could provide a basis for guiding policy decisions concerning alcohol outlets. The collection of local surveillance data on youth violence-related incidents has been a central focus of the Virginia Commonwealth University Clark-Hill Institute for Positive Youth Development. The Institute operates as a CDC-funded Academic Center of Excellence for the prevention of youth violence. Through its community surveillance system, the Institute has been continuously gathering and analyzing youth violence data from hospital emergency rooms, vital registry, ambulance pickups, schools, police, and juvenile justice systems since 2000. These data are routinely summarized and disseminated to guide prevention interventions and policy actions. In 2003, through the Institute's community outreach efforts, residents were made aware of the increased prevalence of violence surrounding convenience stores that sold inexpensive, single-serve alcoholic beverages. Alarmed by these data, residents held several community forums to discuss issues surrounding the problems associated with these "Mom and Pop" (i.e., nonchain, locally owned) convenience stores. Community members recognized that the availability of these beverages significantly contributed to undesirable activities such as violence, loitering, littering, and overall neighborhood decay. They collaborated with the local civic leagues to push for restrictions on any new Alcoholic Beverage Control (ABC) licenses and to prevent renewal of existing licenses that allowed convenience grocery stores in their communities to sell these beverages. More specifically, they requested that the Virginia ABC board issue restricted licenses to allow only the sale of 6 packs, 12 packs, or cases of beer and not the inexpensive and popular 40-or 22-oz bottles. Although these community efforts were maintained for 6 months, enforcement subsequently declined in response to strong opposition to these restrictions by some convenience grocery store owners and their legal counsel. This rather unique situation provided a natural experiment to examine the impact of restricting licenses for alcohol beverage sales on intentional injury-related ambulance pickups. It also provides an illustration of methods used to translate surveillance data into community action. # Methods An ecological panel study was used to examine the impact of restricted licenses for alcohol beverage sales on intentional injury-related ambulance pickups. Rates of ambulance pickups for violent intentional injuries in census tracts where convenience grocery stores were restricted from selling inexpensive, single-serve alcoholic beverages were compared to rates in control census tracts across three phases. The baseline phase represented the 18-month period prior to imposing restrictions in the intervention community. The intervention phase represented the 6 months during which the restriction was in effect. The reversal phase represented the 18 months following removal of the restriction. ## Setting and design This study was conducted in Richmond, Virginia. Approximately 52 % of the city's overall population was African American, and 60 % of youth aged 10-24 were African American (US Census Bureau 2009). The 2008 homicide rate in Richmond was 15.5 homicides per 100,000, nearly three times the national average of 5.4 per 100,000 (US Department of Justice and Federal Bureau of Investigation 2009). The majority of homicides were among youth 15 to 24 years of age. The city harbors a disproportionately high density of public housings and convenience grocery stores that sell alcoholic beverages and goods. Convenience grocery stores where community action had led to license restrictions between January and June of 2003 were identified from the ABC licensure database, and their locations were geocoded. These 18 stores were located within five separate census tracts. The five tracts containing these stores were selected to represent the intervention communities. We then selected five demographically similar census tracts (see [fig_ref] Table 1: Comparison of population characteristics between intervention and control communities [/fig_ref]. These comparison communities contained 18 ABC-licensed stores where restrictions had not been imposed. To prevent bias due to diffusion effects of the intervention, the comparison communities were selected from a separate part of the city where youth violence activity is also historically high. Whereas the intervention tracts were located in the central and northern areas of the city, the selected controls were located in the city's south side [fig_ref] Figure 1: Intervention and control areas and distribution of convenience grocery stores across the... [/fig_ref]. ## Data source and analysis Demographic data for each census tract were obtained from the US Census Bureau (US Census Bureau 2002). Indicators examined included proportions of residents between 15 and 24 years, educational attainment of residents 25 years and older, female-headed households with children 18 years of age or younger, race and ethnic groups, and poverty ratio (ratio of income to poverty level). The proportions were computed for each census tract, and means were compared across conditions (i.e., intervention versus control). Data on ambulance pickups for intentional injury between July 2001 and December 2004 were obtained from the Richmond Ambulance Authority (RAA). The RAA has the franchise to provide emergency medical services to the city. Although there are other transport companies that are volunteer rescue squads or "concierge" services, the RAA is the main service provider in the city. The type of intentional injury was coded based on the paramedics' assessments and patients' reports. Events where an assault occurred, including rape, fight/brawl, shooting, or stabbing, were categorized as a violent injury-related event and included in the analysis. Because violent injuries disproportionately affect the youth population, this analysis was conducted for youth 15 to 24 years of age. The RAA data also included information on longitude and latitude to indicate the location of the incidents for geocoding. Spatial analysis using Geographic Information System software (ArcGIS Desktop version 9.1) was conducted to examine data by census tracts. The rate of intentional injury-related ambulance pickups was calculated using the average number of RAA pickups per month for each census tract per 1,000 populations in the specified time period. Rates were calculated for the 18 months preceding restrictions (baseline phase), the 6 months during which restrictions were in place (intervention phase), and the 18 months following removal of the restrictions (reversal phase). Analyses of changes in ambulance pickups for intentional injuries across the three phases were conducted using SAS Proc Mixed. Analyses were based on a multilevel modeling approach in which observations during each of the three phases of the study (i.e., baseline, intervention, and reversal) were considered nested within census tracts. Phase was treated as a random effect using dummy coding of the intervention and reversal phases with baseline phase as the reference. This allowed us to model changes in the rate of ambulance pickups within census tracts across the three phases. We considered this preferable to a more typical growth curve approach to modeling changes across phases because we anticipated non-linear change across the baseline, intervention, and reversal phases in census tracts where ABC license restrictions were imposed and then withdrawn. We concluded that modeling intervention effects on changes relative to baseline would provide a more easily interpretable and direct test than examining intervention effects on growth parameters within a non-linear change model. These effects were readily incorporated into a mixed effects analysis as reflected in the following equations for the level 1 and level 2 models. [formula] Level 1 equation : Y ti ¼ b 0i þ b 1i P I þ b 2i P R þ " tið1Þ [/formula] This level 1 equation (see Eq. 1) represents Y ti , the rate of ambulance pickups during phase t for census tract i, as a function of the rate during the baseline phase (β 0i ), the change from the baseline to the intervention phase (β 1i , where P I is coded 1 for the intervention phase and 0 otherwise), the change from the baseline to the reversal phase (β 2i , where P R is coded 1 for the reversal phase and 0 otherwise), and a residual term (ε ti ). [formula] Level 2 equation : b 0i ¼ g 00 þ g 01 INT þ μ 0i ; ð2Þ b 1i ¼ g 10 þ g 11 INT þ μ 1ið3Þb 2i ¼ g 20 þ g 21 INT þ μ 2ið4Þ [/formula] The level 2 equations (see Eqs. 2, 3, and 4 above) included intercepts γ 00 , γ 10 , and γ 20 that represent the grand means for census tracts in the control condition for the parameters in Eq. 1, and g 01 , g 11 , and g 21 that represent the intervention effects on these parameters, respectively. INT is the dummycoded intervention condition (1 for intervention condition and 0 for control). This provided a basis for testing differences between tracts in the intervention and control conditions (a) in rates at baseline, g 01 ; (b) in change in rates from the baseline to intervention phase, g 11 ; and (c) in change in rates from the baseline to reversal phase, g 21 . Follow-up tests were also conducted to examine effects within each condition using procedures in Proc Mixed that allow for focused tests on parameter estimates. For example, testing the significance of g 10 +g 11 indicates whether the change across the baseline and intervention phases is significant for tracts in the intervention condition. An unstructured model of the covariance structure was used as it fit the data significantly better than models with more restrictive assumptions (e.g., symmetric, compound symmetry) based on differences in the deviance (−2 log likelihood) for the various models. # Results Demographic characteristics of the intervention and control communities during the study period (3.5 years) are reported in [fig_ref] Table 1: Comparison of population characteristics between intervention and control communities [/fig_ref]. Intervention and comparison communities did not significantly differ at p<0.05 in the proportion of residents who were African Americans, were married, were living in female-headed households with children, were between 15 and 24 years old, were at or below the poverty level, or who had less than a high school education. The average rate of ambulance pickups per month for intentional injuries during the baseline phase was higher in the intervention community than in the control community (19.6 versus 7.4 per 1,000 15- to 24-year-old population, respectively), but this difference was not statistically significant at p<0.05. Parameter estimates for the mixed effects models are reported in Tables 2 and 3. Monthly rates of ambulance pickups for intentional injury per 1,000 in the 15-to 24year-old population across the three phases by condition are reported in along with 95 % confidence intervals. Ambulance pickup rates for intentional injuries were not significantly different for tracts in the intervention and control conditions at baseline (see g 01 in. There was, however, a significant difference between tracts in the intervention and control conditions in their degree of change. For tracts within the control condition, rates per 1,000 15 to 24-year olds during the intervention and reversal phases did not significantly differ from rates during the baseline phase (see and g 10 and g 20 in. In contrast, the average rate of ambulance pickups in the intervention tracts sharply declined between the baseline and intervention phases, decreasing from 19.6 per 1,000 15-to 24-year olds at baseline to 0 during the intervention phase (g 10 +g 11 =−19.55, t(16)=−5.16, p<0.001), and subsequently increased to 11.4 during the reversal phase (i.e., [g 20 +g 21] −[g 10 +g 11 ]=11.44, t(16)=4.52, p=0.022). This pattern of findings was consistent with our hypothesis that rates would decline following imposition of restricted ABC licensing in tracts in the experimental condition, but would increase when this policy was reversed. # Discussion This study revealed an association between the restriction of single-serve alcoholic beverages at convenience stores within the intervention communities and a decline in the rate of intentional injury-related ambulance pickups among youth 15 to 24 years old. The study also showed a gradual increase in injury-related ambulance pickup rates after the restriction was reversed. Although numerous studies have examined the association between alcohol outlet density and violence [bib_ref] Immigrants and violence: The importance of neighborhood context, Alaniz [/bib_ref] [bib_ref] Alcohol consumption, alcohol outlets, and the risk of being assaulted with a..., Branas [/bib_ref] [bib_ref] Neighborhood level spatial analysis of the relationship between alcohol outlet density and..., Britt [/bib_ref] [bib_ref] The relationship between liquor outlet density and injury and violence in New..., Escobedo [/bib_ref] [bib_ref] Alcohol outlets and violent crime in, Franklin [/bib_ref] [bib_ref] Spatial dynamics of alcohol availability, neighborhood structure and violent crime, Gorman [/bib_ref] [bib_ref] The spatial dynamics of violence and alcohol outlets, Lipton [/bib_ref] [bib_ref] Reassessing the alcoholviolence linkage: Results from a multiethnic city, Nielsen [/bib_ref] [bib_ref] The risk of assaultive violence and alcohol availability in Los Angeles county, Scribner [/bib_ref] [bib_ref] Alcohol availability and homicide in New Orleans: Conceptual considerations for small area..., Scribner [/bib_ref] [bib_ref] Alcohol outlet density and violence: A geospatial analysis, Zhu [/bib_ref] , we are not aware of any studies that have evaluated the impact of restricting sales of these inexpensive, single-serve alcoholic beverages on the rate of intentional injury ambulance pickups or communitylevel violence-related outcomes. The results of our study are, however, consistent with few related studies. A recent quasi-experimental study that utilized a pre-post design examined the impact of malt liquor sales restrictions on crimes, at three stores, in a large midwestern city in the USA. The study reported a reduction in disorderly conduct in the restricted sales areas, although an increase in other forms of crimes was observed. Another study, by, that examined refrigerated shelf space dedicated to single-serve containers as a proxy for single-serve alcohol sale, also reported a positive association between single-serve sales availability (i.e., shelf space) and rates of violence. This pattern of findings is consistent with a recent review by [bib_ref] The effectiveness of limiting alcohol outlet density as a means of reducing..., Campbell [/bib_ref] who noted a relation between violent crime and off-premise outlets. Their review also calculated mean elasticities. They reported a high mean elasticity of 0.48; which meant that for every 1 % increase in outlet density, there was a 0.48 % increase in violent crime. A related cross-sectional study by [bib_ref] Alcohol consumption, alcohol outlets, and the risk of being assaulted with a..., Branas [/bib_ref] that examined alcohol outlets and assaults involving firearms also reported a twofold increased risk of firearmrelated assaults in high-density off-site alcohol venues compared to areas with low density area (OR=2.00, 95 % CI= 1.05-3.75). Few longitudinal studies have examined the relation between violence and off-site alcohol venues. In 2006, the sentinel longitudinal study conducted by reported a 1.7 % increase in violence, for every 10 % increase in off-site alcohol outlets. Another longitudinal study, conducted in Australia, also found a positive relation between three types of alcohol premises (general, on premise, and packaged) and violence; however, these types of outlets may not be comparable to those found in the USA [bib_ref] A longitudinal analysis of alcohol outlet density and assault. Alcoholism, Livingston [/bib_ref]. More generally, this study illustrated the utility of using youth violence surveillance data to promote community mobilization and community-led action [bib_ref] Developing a comprehensive approach to youth violence prevention in a small city, Meyer [/bib_ref]. It also highlighted the use of surveillance data for community-based participatory research to promote community-academic partnerships. Routine collection and geo-mapping of surveillance data suggested elevated violence-related incidents in areas served by convenience stores with unrestricted licenses. Outreach workers presented these data to community groups to document the magnitude of the problem. Unlike many academic-initiated studies where researchers impose predetermined objectives on the community, the presentation of the surveillance data mobilized the communities to convene and pursue policy actions. This approach allowed the implementation of community-based participatory research, by "equitably involving all partners in the research process" with the "aim of combining knowledge with action" (Community Health Scholars Program website, as cited in [bib_ref] Representing your community in community-based participatory research: Differences made and measured. Preventing..., Katz [/bib_ref]. These types of strategies not only serve to inform and educate stakeholders, but also assist in engaging the community in their own health [bib_ref] Addressing perinatal disparities using community based participatory research, Masho [/bib_ref]. Such efforts are consistent with emerging environmental efforts to promote and evaluate change at the community level [bib_ref] Translational research in action: Implementation of the communities that care prevention system..., Fagan [/bib_ref]. Although this community effort allowed us to take advantage of a "natural" experiment to examine the effect of restricted sale of inexpensive, single-serve alcoholic beverages on youth violence, the quasi-experimental nature of this study, covering a relatively small geographic area, has several clear limitations. The use of surveillance data led to community mobilization to pursue policy changes to address this problem. Once initiated, these efforts could not be easily subverted by suggesting that the participating communities be subjected to random assignment to intervention and control conditions. This hampers our ability to draw clear causal inferences regarding the impact of policy changes versus other changes that may have occurred within the intervention and comparison communities. The unfortunate and unexpected reversal of policy resulted in a fairly short intervention phase. More generally, the limited number of communities in each condition and fairly low base rate of violence-related ambulance pickups resulting from aggregating data within fairly short windows of time within communities provided limited statistical power to detect differences. A longer intervention phase and replication across communities would have provided a clearer basis for drawing a connection between the intervention and observed outcomes [bib_ref] The value of interrupted time-series experiments for community intervention research, Biglan [/bib_ref]. Another limitation was the small number of tracts included in this analysis (N=10). Additionally, this study was unable to examine other moderating factors such as illegal drug sales and loitering. Finally, these data did not examine other less serious intentional injuries that did not require emergency services, which may have been more common than the serious injuries. Although this study lacked the rigor of a large-scale experimental study, we believe it illustrates the potential value of using surveillance data to inform community mobilization efforts and to evaluate the impact of such efforts. In particular, it demonstrates the potential impact of policies restricting sale of inexpensive, single-serve alcoholic beverages on the rates of intentional injury-related ambulance pickups. It also provides relevant pilot information to emerging large-scale youth violence prevention efforts. Furthermore, this study demonstrated the utility of surveillance data to assess policy changes and community efforts in making significant impacts in communities. It highlights the use of surveillance data to promote and enhance community-academia partnerships in conducting community-based participatory research. Public health professional and community advocates should be aware of the potential impacts of inexpensive, single-serve alcoholic beverages on violence. Future studies should be conducted using large-scale community-wide data to confirm the finding of this study. Open Access This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited. [fig] Figure 1: Intervention and control areas and distribution of convenience grocery stores across the baseline and intervention phases (see g 11 in [/fig] [table] Table 1: Comparison of population characteristics between intervention and control communities [/table]

Participatory Assessment and Selection of Workforce Health Intervention Priorities for Correctional Supervisors Objective: A team of academics and unionized correctional supervisors collaborated to assess workforce health and determine intervention priorities using participatory methods and tools. Methods: Correctional supervisors took a web-based survey. Univariate and bivariate tests examined attitudes/ behaviors, exposures, and outcomes most strongly associated with health; risk based on rank within chain-of-command; and health behaviors amenable to change. We used a voting process tool to prioritize intervention topics. Results: Some health behaviors and outcomes were poor (89% overweight/ obese, 41% poor-quality sleep). We also found favorable health behaviors (annual check-ups) and psychosocial conditions (meaningful work). Some health risks (excessive overtime) were not amenable to change or resisted acknowledgment (poor mental health). The team voted to develop interventions on sleep, mental health, and obesity. Conclusions: Comprehensive health assessment informed the prioritization process, enabling the team to quickly reach consensus on intervention priorities. C orrectional employees demonstrate a number of occupational health disparities that distinguish them from other occupational groups. Correctional employees have a shorter life span than the national average (58 years vs 75 years, respectively). [bib_ref] The experience of stress for correction officers: a double-bind theory of correctional..., Cheek [/bib_ref] Compared with other occupational groups, correctional employees have higher rates of injuries on the job, 1,5 musculoskeletal disorders, [bib_ref] Musculoskeletal disorder symptoms in correction officers: why do they increase rapidly with..., Warren [/bib_ref] cardiovascular disease, [bib_ref] Understanding police stress research, Abdollahi [/bib_ref] [bib_ref] Talking about health: correction employees' assessments of obstacles to healthy living, Morse [/bib_ref] suicide, [bib_ref] Suicide risk among correctional officers: a logistic regression analysis, Stack [/bib_ref] [bib_ref] Eleven years of occupational mortality in law enforcement: the census of fatal..., Tiesman [/bib_ref] anxiety and depression, [bib_ref] Eleven years of occupational mortality in law enforcement: the census of fatal..., Tiesman [/bib_ref] and alcohol and drug abuse. [bib_ref] Correctional officer stress: a review of the literature 1977-2007, Morgan [/bib_ref] They are also more prone to job stress, work-family conflict, and burnout. [bib_ref] Psychosocial work environment, interpersonal violence at work and mental health among correctional..., Bourbonnais [/bib_ref] [bib_ref] Burnout among prison caseworkers and corrections officers, Carlson [/bib_ref] [bib_ref] Leave your job at work: the possible antecedents of work-family conflict among..., Lambert [/bib_ref] [bib_ref] Depression and work family conflict among corrections officers, Obidoa [/bib_ref] As public safety workers, correctional employees' working conditions in prison or jail are unique to their line of work. Their work environment is characterized by high psychological demands (eg, interpersonal conflict, inmate violence, hypervigilance, stress) and physical demands (eg, static sitting and standing, heavy lifting, emergency responses). [bib_ref] Musculoskeletal disorder symptoms in correction officers: why do they increase rapidly with..., Warren [/bib_ref] [bib_ref] Correctional officer stress: a review of the literature 1977-2007, Morgan [/bib_ref] Correctional employees also work extended and irregular shifts, and may be required to work mandatory overtime on short notice. [bib_ref] Shift work and correctional officers: effects and strategies for adjustment, Swenson [/bib_ref] Eliminating the health and safety disparities of correctional employees has received limited research and policy attention, and few programs designed to improve correctional worker health have been evaluated for effectiveness. [bib_ref] Does the job matter? Comparing correlates of stress among treatment and correctional..., Armstrong [/bib_ref] [bib_ref] Progress in corrections worker health: the National Corrections Collaborative utilizing a Total..., El Ghaziri [/bib_ref] While research on correctional employees has increased somewhat in the last two decades, most has focused on front-line correctional officers (COs). [bib_ref] Occupational stress among correctional supervisors, Owen [/bib_ref] [bib_ref] Job stress and burnout among correctional officers: a literature review, Schaufeli [/bib_ref] Health-related findings from this research may not be wholly generalizable to other subgroups of workers in corrections such as supervisors and prison health care personnel, whose work function and exposures considerably differ from that of COs. Correctional supervisors are the middle managers in correctional facilities, whose job titles reflect quasi-military assignment (eg, lieutenant, captain). They are primarily in charge of administrative tasks and maintaining both security and policy within a correctional facility. [bib_ref] The positive effects of participative decision making for midlevel correctional management, Reeves [/bib_ref] [bib_ref] A comparison of line and supervisory officers and the impact of support..., Vickovic [/bib_ref] Compared with COs, correctional supervisors have a higher level of responsibility and fewer organizational peers, which may place them at increased health risk. Although separated from COs due to their higher rank within the chain of command, supervisors are usually promoted from the CO ranks which confers common background experience. There are also differences among supervisors themselves. In the state correctional system discussed in this study, supervisors include lieutenants, captains, and counselor supervisors. Captains and counselor supervisors (equivalently ranked) have a higher rank in the hierarchy compared with lieutenants, and they typically only work on weekdays, mostly on the first shift, have shorter commuting distances (due to more discretion over their job site), and are not obliged to fill posts. The result is a more predictable schedule with less overtime. Lieutenants are the largest group of supervisors, and they work on the front line with COs, which exposes them to stressors that are different from those of captains and counselor supervisors, and more similar to those of COs. Therefore, identification of stressors unique to correctional supervisors (particularly lieutenants) that account for differences in rank and responsibilities is an important research area for development, and a foundational step to interventions that aim to improve supervisor health. ## Health improvement through employee control (hitec) Health Improvement Through Employee Control (HITEC) is a research project initiated by investigators from the Center for the Promotion of Health in the New EnglandWorkplace (CPH-NEW), a National Institute for Occupational Safety and Health (NIOSH) Center of Excellence for Total Worker Health.Total Worker Health (TWH) is an approach whereby workplace interventions simultaneously protect and promote worker health, by jointly addressing individual-and organizational-level factors that affect worker health. This approach is gathering increasing attention in corrections research and practice. [bib_ref] Progress in corrections worker health: the National Corrections Collaborative utilizing a Total..., El Ghaziri [/bib_ref] HITEC is a collaboration of public university researchers and a state Department of Correction (DOC) in the northeast US. It has effectively used participatory techniques to study workplace factors and to design interventions that impact the health and well-being of the correctional workforce. Over 15 years of HITEC research has demonstrated that in corrections, workplace interventions designed with grassroots involvement of front-line workers using a participatory approach are more effective in improving worker health than traditional top-down, administratively-driven health interventions. [bib_ref] Process evaluation of two participatory approaches: implementing Total Worker Health1 interventions in..., Dugan [/bib_ref] [bib_ref] Total Worker Health1 needs assessment to identify workplace mental health interventions in..., Jaegers [/bib_ref] In addition to having higher participation rates and greater acceptability, [bib_ref] Participatory action research in corrections: the HITEC 2 program, Cherniack [/bib_ref] participatory interventions are perceived by endusers as having more relevance, appropriateness, credibility, and compatibility with organizational culture. [bib_ref] Dissemination and implementation research for occupational safety and health, Dugan [/bib_ref] Organizational barriers such as bureaucracy and hierarchy, poor leadership, and power imbalances, organizational climates characterized by a lack of flexibility and support, and poor employee motivation, can pose barriers to participatory methods, yet participatory programs have been more successful in the public sector (compared with the private sector) because of opportunities for worker empowerment and voice [bib_ref] A comparative study of workplace bullying among public and private employees in..., Ariza-Montes [/bib_ref] [bib_ref] Workplace bullying and mobbing in the public service sector and the role..., Sorozan [/bib_ref] [bib_ref] Do gender differences matter to workplace bullying?, Wang [/bib_ref] In 2014, within the HITEC project, a new partnership began between seven members of a correctional supervisors' union in the state (eg, captains, lieutenants, counselor supervisors) and two CPH-NEW academic researchers who formed a ''Design Team'' (a team that is tasked with developing interventions) to improve the health of correctional supervisors using community-based participatory research (CBPR) methods. In CBPR, members of a community (ie, people connected by shared norms and values, common language and customs, similar goals and needs, and collective interest in community well-being) [bib_ref] Development and implementation of principles for community-based research in public health, Schulz [/bib_ref] are actively and equitability involved in all aspects of the research process, alongside scientific investigators. CBPR partnerships with the highest degree of equity are those in which community partners help determine research priorities, define research questions, contribute to study and intervention design, and make joint decisions about putting study findings into action to create change. [bib_ref] Development and implementation of principles for community-based research in public health, Schulz [/bib_ref] A review of occupational and environmental health CBPR research found this level of participation in only a quarter of studies. [bib_ref] Integrating research and action: a systematic review of community-based participatory research to..., Cook [/bib_ref] Our goal in this study was to form and assess a partnership with the highest levels of involvement from the Design Team, as representatives of the community of correctional supervisors. We did this by following the process and using tools from CPH-NEW's Healthy Workplace Participatory Program 33 (HWPP) on-line toolkit. Per the HWPP, the Design Team's initial task was to create a comprehensive and contextually-relevant survey to assess the health of their workforce of correctional supervisors. [bib_ref] Community-based participation in survey design and implementation: The Healthy Environments Partnership Survey, Schulz [/bib_ref] They first utilized the HWPP Focus Group Guide for Workplace Safety, Health and Well-beingto generate key themes regarding relevant health-related experiences and challenges faced by correctional supervisors. Then they used those themes to adapt the HWPP All Employee Survey (AES) to create their own customized survey, entitled the Correctional Supervisors' Healthy Workplace Survey. The AES is a generic workforce health assessment questionnaire.Details of survey adaptation and the participatory survey design process are provided in a prior publication by Dugan et al. [bib_ref] Participatory survey design of a workforce health needs assessment for correctional supervisors, Dugan [/bib_ref] In addition to providing a point-in-time assessment of supervisor health, the survey was intended to confirm whether the experiences described in focus groups held true for the entire supervisor workforce, thus enabling the Design Team to identify priority health concerns and develop targeted interventions to address them. In focus groups, supervisors identified 12 themes as organizational-or individual-level factors relevant to supervisor health and for possible intervention opportunities. These included unhealthy organizational culture, masculinity, work-family conflict, family support, stress and trauma, positive aspects of the job, health literacy and efficacy, health and risk behaviors, sleep, obesity, income and retirement, and the need make health a personal priority. [bib_ref] Participatory survey design of a workforce health needs assessment for correctional supervisors, Dugan [/bib_ref] In this paper, we present the results of the survey developed by a Design Team using participatory methods, explain how survey results confirm or contradict previous work (including focus groups), and demonstrate how survey findings were used to inform decision-making about health intervention priorities. To guide the current study, the Design Team posed two research questions to examine the health-related attitudes and behaviors, exposures, and health and work outcomes of correctional supervisors. They also posed one research hypothesis to examine comparative differences between supervisors of different ranks (lieutenants vs captains and counselor supervisors). This hypothesis was posed in order to evaluate whether interventions should be targeted to different groups. We propose the following: Research Question 1: What health-related attitudes and behaviors, exposures, and outcomes are most strongly associated with health in general? Hypothesis 1: Supervisors with the rank of lieutenant will be at greater risk for poor health (worse health-related attitudes and behaviors, more exposures, worse outcomes) than those with the higher rank of captain or counselor supervisor. Research Question 2: What specific health behaviors are supervisors most interested in changing? # Methods In this study, the Design Team who created the Correctional Supervisors' Healthy Workplace Survey, 37 continued the HWPP process by using the survey and another key tool: the HWPP Prioritizing and Selecting Safety and Well-being Concerns for Intervention Group Activity.Details regarding the selection of Design Team members (ie, seven union members, two university researchers), as well as the 25-minute web-based survey assessing health-related attitudes and behaviors, attitudes and behaviors related to income and retirement, psychosocial exposures at work and home, health outcomes, work outcomes, and job and personal information [fig_ref] TABLE 1: The Correctional Supervisors' Healthy Workplace Survey [/fig_ref] are detailed in a prior study. [bib_ref] Participatory survey design of a workforce health needs assessment for correctional supervisors, Dugan [/bib_ref] Protocols for the current study were approved by the University's Institutional Review Board. ## Survey participants and administration Correctional supervisors (ie, lieutenants, captains, counselor supervisors) working in the public sector (Department of Correction) in a northeastern US state across 19 correctional facilities (ie, JOEM - Volume 64, Number 7, July 2022 Workforce Health Priorities prisons, jails) were invited to participate in the cross-sectional survey. The 19 facilities housed approximately 13,400 incarcerated people and had 5210 workers staffing the facilities, including 423 supervisors. All participants were members of the correctional supervisors' union who supervised correctional officers in the state's Department of Correction facilities. The survey was distributed by the union vice president to all 423 union members via an email that explained the survey purpose (ie, to conduct a health needs assessment for intervention planning) and invited supervisors to participate in the online survey by clicking a web link. The survey was open for one month and there were no incentives for study participation. Two weeks after the survey launch, the Design Team held a meeting to discuss and implement ways to promote study participation to increase the response rate (eg, sending a reminder email, making announcements at union meetings). ## Prioritizing and selecting safety and well-being concerns for intervention group activity The HWPP Prioritizing and Selecting Safety and Well-being Concerns for Intervention Group Activity starts off with the Design Team first reviewing and discussing the results of the workforce health assessment; this is intended to inform opinions in preparation for deciding upon health intervention priorities. Descriptive analyses of survey data, as well as univariate and bivariate analyses, were conducted by the academic team members and presented as tables and figures in a PowerPoint presentation (key figures are presented in Appendix A, http://links.lww.com/JOM/B75) for the Design Team's review. Next, each individual member creates a list of issues that they would like the Design Team to prioritize for developing interventions. Each Design Team member shares their top three issues to generate a larger group list, and if topics are similar in theme, they may be grouped together. Each Design Team member votes for three topics on the group list. The topics that receive the top three votes are designated as priorities. However, the topic that receives the most votes is not necessarily the choice for developing an intervention. Rather, the team discusses which priority topic among the three is the best choice for the first intervention initiative. Mitigating considerations include feasibility and potential for positive impact on the workforce. If consensus on a first topic is not reached by discussion, each person may again vote and the topic that receives the most votes is selected. Corrections employees and union leaders are typically cautious about worker privacy, and since this was a CBPR study where the Design Team provided input into study procedures, they opted not to audio-record and transcribe the group activity. Rather, two Design Team members took detailed notes, recording what was discussed throughout the group activity's consensus-building and voting processes. Health-related behaviors and attitudes Psychosocial family exposures Healthy eating1 Family health climate 4 Physical activity1 Health outcomes Cigarette smoking1 General health [bib_ref] Individual differences in two emotion regulation processes: Implications for affect, relationships, and..., Gross [/bib_ref] 4 Work outcomes Interpersonal caring and dominance [bib_ref] The measurement of psychological androgyny, Bem [/bib_ref] [bib_ref] Exploratory and confirmatory studies of the structure of the Bem Sex Role..., Choi [/bib_ref] 8 S t r e s s2 Fatalistic thinking about health [bib_ref] Fatalism reconceptualized: a concept to predict health screening behavior, Straughan [/bib_ref] 4 W o r k -f a m i l y c o n f l i c t 61 4 Readiness for change [bib_ref] The transtheoretical model of health behavior change, Prochaska [/bib_ref] 8 Behavior-based work-family conflict [bib_ref] Construction and initial validation of a multidimensional measure of work-family conflict, Carlson [/bib_ref] [bib_ref] Workplace cohort studies in times of economic instability, Cherniack [/bib_ref] 1 Work information Retirement-financial situation [bib_ref] Workplace cohort studies in times of economic instability, Cherniack [/bib_ref] 1 Overtime hours* 1 Psychosocial work exposures 1 Commuting time1 Civility norms [bib_ref] Assessing workgroup norms for civility: the development of the civility norms questionnaire-brief, Walsh [/bib_ref] Work history4 Job content: social support2 Facility/Location* 1 Job security4 O t h e r Emotional job demands [bib_ref] Dual processes at work in a call centre: an application of the..., Bakker [/bib_ref] 1 Informed consent 1 Org. support for health and safety [bib_ref] Perceptions of safety at work: a framework for linking safety climate to..., Griffin [/bib_ref] 1 Use of social media* 1 Work health climate [bib_ref] A practical scale for multi-faceted organizational health climate assessment, Zweber [/bib_ref] 2 Advice to new recruits* 1 Supervisor communication [bib_ref] The team climate inventory: development of the TCI and its applications in..., Anderson [/bib_ref] 3 Other comments1 Masculine org. culture4 Effect of traumatic events at work* 6 Meaningful work [bib_ref] Psychological empowerment in the workplace: dimensions, measurement, and validation, Spreitzer [/bib_ref] 18 DT, Design Team; total No. survey items = 170. *Indicates an original item created by the Design Team. # Data analysis All analyses were conducted by the two university researchers, using IBM SPSS Statistics for Windows, version 26.0 (IBM Corp., Armonk, NY). Using principles of participatory research, the Design Team provided input into which analyses would be conducted. We used descriptive statistics, univariate analyses (t tests), and bivariate analyses (correlations) because the data were collected primarily for use by the Design Team, rather than for research (its secondary purpose). To answer the first research question, we examined the bivariate Pearson correlations of various health-related attitudes and behaviors, psychosocial exposures, and outcomes with a ''General Health'' variable that assessed self-reported health (''In general would you say your health is. . .'') on a scale from 1 (poor) to 5 (excellent). The study's hypothesis was tested using independentsamples t tests to assess differences between two groups of supervisors (lieutenants vs captains/counselor supervisors). The second research question was answered by examining the eight individual items that made up the ''Readiness for Change'' measure, in which participants indicated interest in adopting eight health behaviors through responses of 0 (not interested in changing) or 1 (interested in changing). Of those who completed the survey, missing data were excluded from the analyses. Although unplanned, during the review and discussion of results, the Design Team also opted to compare some of the healthrelated survey findings from correctional supervisors to those of the general US adult population. They also requested some targeted post hoc analyses (t tests, Pearson correlations) in order to further understand relationships among study variables (results described but not shown).We present this information along with the results of the planned analyses as it was instrumental in helping the Design Team make decisions about intervention priorities. # Results Of the 423 supervisors invited, 157 participated in the online survey, representing a response rate of 37%. Demographic administrative data provided by the DOC Human Resources office confirmed that the 37% who participated were representative of the larger DOC supervisor population. The mean (SD) age of participants was 42 years (6.05) (see [fig_ref] TABLE 2: Demographics and Work Information of Overall Sample by Job Classification [/fig_ref]. Fifty percent of the sample was between 41 and 50 years old. The sample was 78% men, 69% White/non-Hispanic, 73% married/partnered, and 46% had bachelor's degree or higher. Most were lieutenants (59%) and the rest were captains and counselor supervisors. Most participants (64%) worked the first (daytime) shift and worked a mean (SD) of 12.8 (11.6) overtime hours per week. Mean (SD) DOC tenure was 15.4 years (4.72). There were statistically significant differences between the two supervisor groups. Compared with captains/counselor supervisors, lieutenants had a lower education level and fewer years of job tenure, were more likely to work second or third shift (rather than first), and worked more overtime hours. ## Correctional supervisors' healthy workplace survey The Design Team reviewed and discussed the survey results as detailed below. ## Health-related behaviors and attitudes Most respondents reported moderately meeting expert recommendations for healthy eating and physical activity (see [fig_ref] TABLE 3: Alphas, Means With Standard Deviations, and Correlations for Health Behaviors/Attitudes by Job... [/fig_ref]. Specifically, 37% of respondents never or rarely met expert recommendations for physical activity, and 43% never or rarely met expert recommendations for fruit and vegetable consumption in Appendix A, http://links.lww.com/JOM/B75). Consistent with the established anti-smoking policies of DOC, a small number of respondents smoked cigarettes daily (10%), but a larger number chewed tobacco daily (11%). The Design Team compared these findings to the 14% cigarette smoking rate of US adults 66 and the 2.4% of adults aged 18+ who report use of smokeless tobacco.A small number of respondents reported regularly smoking a cigar/ pipe or gambling. Regarding alcohol, most respondents (59%) reported having 1 to 14 drinks per week, with 8% consuming 15+ drinks per week. This was much higher than the 5.1% of US adults aged 22 to 44 years old who consume 14+ drinks per week [bib_ref] Heavy drinking among U.S. adults, Boersma [/bib_ref] Caffeine consumption was high, with 30% consuming 4+ beverages daily. This was much higher than the 10% of general population who consume a mean daily caffeine intake of ∼4 beverages (380 mg) per day, [bib_ref] Beverage caffeine intakes in the US, Mitchell [/bib_ref] the limit recommended by the US Food and Drug Administration.Post hoc analyses requested by the Design Team (not shown) indicated that caffeine consumption was positively correlated with overtime hours, sleep difficulty, work and home stress, and psychological symptoms (anxiety, depression, hostility). Most respondents had a primary care doctor/provider (96%) and received annual checkups (97%) with recommended screenings (mean ratings on a 3-point scale were both ∼3). The Design Team noted this was much higher than the United States adult population in which 75% have an identified source of primary care 71 and onethird receive annual physicals (preventive examinations). [bib_ref] Improving value in health care-against the annual physical, Mehrotra [/bib_ref] Considerably fewer (22%) supervisors had an identified mental health professional to receive psychological care from (∼1.5 on a threepoint scale). Most reported a strong sense of personal responsibility for their health (∼4 on a five-point scale). Respondents reported moderate levels of emotional suppression (∼3 on a five-point scale) and post hoc tests requested by the Design Team (not shown) indicated emotional suppression was positively correlated with being affected by work trauma (inmate-directed), behavior-based work-to-family conflict, and psychological symptoms (depression, anxiety). Respondents reported moderate levels of both interpersonal caring and dominance (both 3.4 on a five-point scale). Post hoc independent-samples t tests, which were requested by the Design Team, showed that female supervisors reported being more caring than males, but there were no differences in reported social dominance when comparing women and men in Appendix A, http://links.lww.com/JOM/B75). Other post hoc tests showed that some poorer health behaviors (excess caffeine consumption, emotional suppression, more overtime hours, prioritizing earnings over health) were correlated with social dominance and one health behavior (fewer overtime hours) was correlated with caring. Several attitudes and behaviors were positively correlated with general health including healthy eating, physical activity, and personal responsibility for one's health. General health was negatively correlated with smoking cigars/pipes, chewing tobacco, consuming caffeine, and less overall readiness to improve health behaviors. Means and standard deviations listed were calculated by excluding missing cases. Healthy eating, physical activity (1 = Never, 5 = Always); cigarette smoking (1 = Never smoked, 5 = 10+ cig weekly); smoking cigars/pipe, chewing tobacco, gambling (1 = Never, 5 = Daily); consuming alcohol (1 = 0 drinks weekly, 5 = 21+ drinks weekly); consuming caffeine (1 = 0 beverages daily, 5 = 10+ beverages daily); access care (1 = no, 2 = unsure, 3 = yes); suppressing emotions, fatalism (1 = strongly disagree, 5 = strongly agree); interpersonal caring and dominance (1 = almost never true; 5 = almost always true); readiness for change (0 = not interested in change any health behaviors, 8 = interested in changing 8 health behaviors). c Independent samples t test. *P < 0.05. **P < 0.01. ***P < 0.001. There were statistically significant differences between the two supervisor groups on some variables. Compared with captains/ counselor supervisors, lieutenants reported smoking more (P = 0.003), drinking more alcohol (P = 0.048), being less likely to have a primary care physician (P = 0.04), and showing less interpersonal caring (P = 0.02). They also reported marginally more interpersonal dominance. In terms of participants' readiness to make changes or improvements to health, respondents on average were moderately interested in improving health behaviors overall (mean response showed interest in changing ∼4 out of 8.0 possible health behaviors). in Appendix A, http://links.lww.com/JOM/B75, shows detail of specific health behaviors supervisors were most interested in changing.) The highest percentage of respondents (80%) reported an interest in reducing stress, followed by improving sleep (71%), practicing healthy eating habits (66%), getting physical greater activity (62%), and losing weight (59%). Fewer respondents were interested in reducing use of caffeine, cigarettes/tobacco, or alcohol. With regard to reported attitudes on health, longevity, income, and overtime hours (see [fig_ref] TABLE 4: Means With Standard Deviations and Correlations for Attitudes on Longevity, Income, Overtime,... [/fig_ref] , most supervisors thought often about retirement (∼4 on a five-point agreement scale). Most reported that they would retire after 20 years (as early as possible) and that they would be healthier and happier after they retire. Most were aware that the lifespan of correctional workers is short and that earlier death is a job risk (2.9 and 2.6, respectively, on a three-point scale). Most thought that health should not have to be delayed until one retires (2.7 on a three-point scale). Many reported that they currently felt they had to prioritize earnings over their health (∼2 on a three-point scale), but acknowledged they would probably be healthier if they worked less overtime (∼2 on a three-point scale). Two attitudes were positively correlated with supervisors' general health, including feeling confident in their ability to meet financial needs in retirement and thinking that they would be financially comfortable after they retire. General health was negatively correlated with frequent thoughts of retirement, thinking that health and happiness would increase after retirement, thinking that working less overtime would improve health, awareness that an earlier death is a job risk, and prioritizing earnings over health. There were statistically significant differences between the two supervisor groups on several variables. Captains/counselor supervisors were closer to retirement (in years) than lieutenants (Mean [SD] of 5.87 (3.90) vs 7.39 respectively; P = 0.039). However, compared with captains/counselor supervisors, lieutenants thought more often about retirement (P = 0.01), planned to retire after 20 years (as early as possible) (P < 0.001), prioritized earnings over health (P = 0.02), and thought that they would be healthier with less overtime (P < 0.001) ## Psychosocial exposures Regarding psychosocial exposures at work and home (see [fig_ref] TABLE 5: Alphas, Means With Standard Deviations, and Correlations for Psychosocial Work Exposures by... [/fig_ref] , job security and meaningful work were high (mean ratings on a five-point agreement scale were 4.3 and 4.0, respectively). Emotional job demands were high (3.9 on a five-point scale), and post hoc tests requested by the Design Team (not shown) indicated that emotional job demands were positively correlated with feeling affected by work trauma (inmate-directed, self/peer-directed), stress (work, home), burnout, work-to-family conflict and family-to-work conflict (all forms), and psychological symptoms (depression, anxiety, hostility). Supervisor and coworker social support, organizational support for health and safety, and supervisor communication were moderately high (∼3.5 on a five-point scale). Family health climate and work health climate were fairly high (∼3.5 on a five-point scale). Ratings of civility norms and masculine organizational culture were moderate (∼3 on a five-point scale), and post hoc tests (not shown) indicated a positive correlation between perceiving a masculine organizational culture and being a female supervisor. Regarding the effect of traumatic events at work, ratings were low, and respondents reported being more greatly affected by trauma directed to self/peers (∼2 on a five-point scale) than by trauma directed to inmates (∼1 on a five-point scale). Post hoc analyses requested by the Design Team (not shown) indicated Notes. Means and standard deviations listed were calculated by excluding missing cases. Retirement thoughts often (1 = strongly disagree, 5 = strongly agree); less overtime healthier, early death job risk, aware of short life span, prioritize earnings over health, health should not wait, plan to retire at 20 years (1 = no, 2 = unsure, 3 = yes); proximity to retirement (no. of years until retirement); retirement financial confidence (1 = not at all, 5 = very); retirement financial situation (1 = won't have enough to meet basic needs, 4 = able to live comfortably). Exposures that were positively correlated with general health included coworker social support, organizational support for health and safety, work health climate, supervisor communication, meaningful work, and family health climate. General health was negatively correlated with emotional job demands. Several variables differed based on supervisor group. Compared with captains/counselor supervisors, lieutenants reported lower civility norms (P = 0.049), lower coworker support (P = 0.02), lower organizational support for health and safety (P = 0.01), and less supervisor communication (P = 0.02). ## Health outcomes Regarding health outcomes (see [fig_ref] TABLE 6: Alphas, Means With Standard Deviations, and Correlations for Health Outcomes by Job... [/fig_ref] , respondents reported relatively good general health (3.4 on a five-point scale) and an average of one chronic health condition (mean response was presence of ∼1 out of 5.0 possible chronic health conditions). The most common chronic conditions (see in Appendix A, http://links.lww.com/JOM/B75) were high cholesterol (24%) and hypertension/high blood pressure (23%). In a post hoc analyses requested by the Design Team (not shown) lieutenants were more likely to report having anxiety/depression compared with captains/ counselor supervisors. In terms of body mass index (BMI), 51% fell in the obese category (compared with the 42% age-adjusted prevalence of obesity in US adults).73 (See in Appendix A, http:// links.lww.com/JOM/B75) An additional 38% fell in the BMI category of overweight. Average ratings of musculoskeletal pain and health interference with work were low (∼2 on a five-point scale). Supervisors reported low psychological symptoms of anxiety, depression, and hostility (mean ratings ranged from 1.6 to 2.0 on a five-point scale). (See in Appendix A, http://links.lww. com/JOM/B75) Hostility (2.0) was the most severe self-reported psychological symptom experienced, followed by anxiety (1.9) and depression (1.6). Supervisors also reported their perceptions of the symptoms experienced by other supervisors they work with, which showed that supervisors perceived others' symptoms as more severe than their own. Hostility (2.6) was perceived to be the most severe psychological symptom attributed to other supervisors, followed by anxiety (2.4) and depression (2.0). Sleep-related findings showed that 41% of supervisors reported (very or fairly) poor sleep quality on a typical night. 57% of respondents reported 6 or fewer hours of sleep each day during a regular work week, while only 11% reported 8 or more hours, the amount recommended by sleep experts. (See in Appendix A, http://links.lww.com/JOM/B75) Comparing actual sleep hours to needed sleep hours, respondents had much less sleep than they needed to have good functioning. About 32% reported (moderate to severe) difficulty sleeping due to physical and emotional problems. General health was positively correlated with sleep quality. Health outcomes that were negatively correlated with general health included number of chronic health conditions, BMI, self-reported psychological symptoms, health interference with work, musculoskeletal pain, sleep difficulty, and needed sleep hours. In terms of difference between supervisor groups, lieutenants reported poorer sleep quality compared with captains/counselor supervisors (P = 0.04); this may be related to other differences described earlier, including lieutenants being less likely to work the dayshift and working more overtime hours. Post hoc tests requested by the Design Team (not shown) indicated that some poorer health outcomes (difficulty sleeping, lower sleep quality) were correlated with social dominance, and some better health outcomes (higher sleep quality, fewer reports of back disease/spine problems as a chronic condition, lower cholesterol) were correlated with caring. They also showed that poorer mental health (greater effect of inmate-directed trauma, depressive and hostility symptoms, work stress, and burnout) was correlated with social dominance, and better mental health (less hostility, burnout, anxiety, and home stress) and less work-family conflict (behavior-based work-to-family conflict, behavior-based family-to-work conflict) were correlated with caring. ## Work outcomes Regarding work outcomes (see [fig_ref] TABLE 7: Alphas, Means With Standard Deviations, and Correlations for Work Outcomes by Job... [/fig_ref] , findings showed moderate levels of overall stress (∼3 on a five-point scale). Ratings on general work-to-family conflict (WFC) and familyto-work conflict (FWC) were low-to-moderate (2.8 and 2.2, respectively, on a five-point scale). (See in Appendix A, http://links.lww.com/JOM/B75) Ratings of behavior-based WFC and FWC (which occurs when behaviors that are effective in one life role are inappropriate in another role) were higher (3.2 and 3.1, respectively, on a five-point scale). Ratings on job satisfaction were high (∼4 on a five-point scale), burnout was moderate (∼3 on a five-point scale), and turnover intentions were low (∼2 on a five-point scale). General health was positively correlated with job satisfaction. Health outcomes that were negatively correlated with general health included stress, general WFC and FWC, behavior-based WFC and FWC, burnout, and turnover intent. In terms of difference between supervisor groups, lieutenants reported greater behaviorbased FWC compared with captains/counselor supervisors ( P = 0.02). ## Prioritization and selection of safety and well-being concerns After reviewing and discussing the workforce health assessment survey results, Design Team members each generated a list of issues that they would like the team to prioritize for developing interventions, and then shared their top three issues with team members. Similar topics were grouped together, resulting in a group list that included 10 possible topics for intervention (see [fig_ref] TABLE 8: List of 10 Health Priority Topics for Intervention [/fig_ref] : improving sleep quality/quantity, weight loss, reducing stress, reducing alcohol/substance use, improving healthy eating, increasing physical activity, reducing caffeine use, improving healthy emotional expression, reducing high blood pressure/cholesterol, and reducing musculoskeletal pain. Each Design Team member then voted for three priority topics from the group list. The Design Team tallied votes, noted the most-voted-for topics, and then discussed what topic would be most feasible, impactful, and well-accepted as a first intervention initiative. The Design Team determined that the best approach would be to focus on health concerns that affected a large number of supervisors and that they were motivated to address. To that end, much of the discussion centered on survey responses to the question regarding supervisor readiness to change various health behaviors. For example, during the review of survey results, sleep and obesity were seen as areas of concern affecting many supervisors, and were topics that both received the highest number of votes (4) as a priority topic, but because more survey respondents indicated that they were ready to improve their sleep (71%) than lose weight (59%), the Design Team chose to develop a sleep intervention as their first priority. The mental health-related issues of reducing stress and alcohol/substance use both received the second highest number of Design Team votes (3) as a priority topic. Although 80% of supervisors respondents were ready to reduce stress (it was the health behavior they were most interested in changing) and the Design Team noted that most survey variables measuring stress and poor mental health were associated with poorer overall health, survey responses showed low self-reports of mental health problems, suggesting that supervisors may not be aware of or recognize their experiences as being of concern for mental health. Notably, although alcohol/substance use was voted for as a priority topic c Independent samples t test. *P < 0.05. **P < 0.01. ***P < 0.001. because of survey results showing heavy alcohol use, alcohol use was the behavior that survey respondents were least ready to change (18%), and the Design Team felt that there would be a lack of interested participants for alcohol interventions. For these reasons, the Design Team chose stress reduction as their second priority. Although weight loss received the same level of Design Team prioritization votes as sleep, fewer survey respondents indicated that they were ready to lose weight (59%) compared with making other health behavior changes. Therefore, weight loss was decided as the third priority for the Design Team. Finally, the Design Team had decided on an intervention program to address the following three topics, in turn: sleep, stress, and weight loss. However, because some topics on the list of 10 voted-for health priorities were related or similar in theme, the Design Team began to group priorities together to be as inclusive as possible in addressing these priorities and the problems raised by the survey. The Design Team expanded the first intervention topic on sleep to include other sleep-related topics such as overtime and shiftwork, stress, caffeine use, physical activity, healthy eating, and obesity. They broadened the second intervention to not only address stress management, but also to incorporate the mental health-related topics of emotional well-being and alcohol/substance use. They expanded the third intervention on weight loss to also address the behavioral topics of healthy eating and physical activity, as well as cardiovascular health. The Design Team reached final consensus that they would develop interventions to address the following three broad topics, in order: sleep quality/quantity, stress management/ mental health, and weight loss/healthy eating/physical activity. # Discussion This study presents the findings from a workforce health assessment survey, which were utilized for prioritizing intervention topics. The survey was developed using participatory design methods and HWPP tools. The primary purpose was to assess the health and well-being of a unionized, public sector workforce of correctional supervisors, and then to identify health topics for intervention. HWPP has been used in previous studies and can be easily adapted and implemented by organizations to improve worker health and well-being. Analysis of survey data collected using the thorough and contextually-relevant workforce assessment also served a secondary purpose, which was to contribute to research by providing deeper knowledge about supervisors' lived experiences of work, wellbeing, and lifestyle. The presumption was that participatory methods would elucidate an under-researched subject area more effectively than conventional methods. To inform the selection of three priority health concerns for intervention planning, the Design Team posed two research questions and a hypothesis. They wanted to know which health-related attitudes and behaviors, psychosocial occupational exposures, and outcomes were most strongly associated with general health; whether more junior lieutenants had greater health risk compared with higher-ranked supervisors; and which health behaviors supervisors were most interested in changing. During the process of reviewing survey results, the Design Team employed two other unplanned methods for evaluation: comparing their cohort results to national averages and established norms (when possible), and running post-hoc analyses to examine associations among some variables of particular interest. ## Assessment of workforce health ## Health outcomes and health behavior/attitudes Survey results confirmed several health problems raised in our initial focus group study [bib_ref] Participatory survey design of a workforce health needs assessment for correctional supervisors, Dugan [/bib_ref] showing a high prevalence of overweight/obesity (high BMI) that exceeded national averages 73 Means and standard deviations listed were calculated by excluding missing cases. Stress (1 = no stress, 5 = extreme stress); work-to-family conflict, family-to-work conflict (1 = never, 5 = always); behavior-based work-to-family conflict, behavior-based family-to-work conflict, burnout, satisfaction, intent to turnover (1 = strongly disagree, 5 = strongly agree). c Independent samples t test. *P < 0.05. **P < 0.01. ***P < 0.001. as well as a reported lack of behavioral adherence to healthy eating and physical activity guidelines. All of these were correlated with poorer overall health. Linked to obesity, poor cardiovascular health affected about a quarter of all supervisors (ie, they reported having high cholesterol and/or hypertension as a chronic health condition). In terms of BMI, although 51% of supervisors fell in the obese category, these results may be inflated given the body structure of many male correctional workers in the population, which is largeframed and muscular. This is because although people with a high BMI are likely to have a body composition with high fat body mass, high BMI can also result from high lean body mass (muscle and bone).Findings regarding obesity and sleep align with observations from previous studies of obesity [bib_ref] Talking about health: correction employees' assessments of obstacles to healthy living, Morse [/bib_ref] [bib_ref] Work characteristics as predictors of correctional supervisors' health outcomes, Buden [/bib_ref] [bib_ref] Associations among work and family health climate, health behaviors, work schedule and..., Buden [/bib_ref] [bib_ref] Health behavior knowledge and self-efficacy as predictors of body weight, Faghri [/bib_ref] and sleep deficiencies 16,78 among corrections workers. Survey findings confirmed concerns about sleep raised in focus groups, [bib_ref] Participatory survey design of a workforce health needs assessment for correctional supervisors, Dugan [/bib_ref] showing that most supervisors did not get sufficient hours of sleep, and many reported poor quality sleep and difficulty sleeping, all of which were correlated with worse overall health. Lieutenants had poorer sleep quality than higher-ranked supervisors. Also corroborating focus group conclusions 37 were survey findings showing heavy alcohol and caffeine consumption that exceeded national averages. The Design Team noted that caffeine use was correlated with poorer overall health and that alcohol use was higher among lieutenants than higher-ranked supervisors. Our findings support existing research showing heavy alcohol consumption among corrections workers 8 ; however further insight could be gained from future research examining factors associated with alcohol use. For example, some research suggests that for correctional workers, drinking alcohol may be used as a form of coping, [bib_ref] Emotional demands and alcohol use in corrections: a moderated mediation model, Shepherd [/bib_ref] which is supported by our post hoc analyses showing links between alcohol consumption and trauma at work. This is also aligned with the experiences reported by Design Team members that after-work drinks with colleagues at local bars often serve as an opportunity to debrief about work stresses and receive social support. More research on alcohol and social support is needed, and could possibly inform future psychoeducational interventions focused on making supervisors aware of healthier forms of coping and social support. We could not identify other research focused on caffeine use as a health behavior in correctional workers. We found an association between caffeine use and poorer overall health, which may contradict research suggesting that caffeine, a stimulant, has possible benefits (ie, better performance and safety at work).80,81 Findings suggested that a third of correctional supervisors in our sample consume excessive amounts of caffeine, as defined by the FDA70 Excessive caffeine intake is associated with increased risk of hypertension and cardiovascular disease, as well as nervousness/ anxiety, headaches, insomnia, nausea, rapid heartbeat, frequent urination, and muscle tremors. [bib_ref] Caffeine (1, 3, 7-trimethylxanthine) in foods: a comprehensive review on consumption, functionality,..., Heckman [/bib_ref] [bib_ref] Effects of caffeine on human health, Nawrot [/bib_ref] [bib_ref] caffeine and blood pressure: a critical review, Nurminen [/bib_ref] [bib_ref] The effect of coffee on blood pressure and cardiovascular disease in hypertensive..., Mesas [/bib_ref] This aligns with our post hoc analyses showing caffeine consumption was related to difficulty sleeping, stress, and psychological symptoms; it was also associated with working more overtime hours. Links between heavy caffeine use and sleep impairment are especially pronounced in extendedshift workers, like those in corrections, who rely on caffeine at work to manage fatigue, stay awake, and improve alertness. [bib_ref] The safety of ingested caffeine: a comprehensive review, Jennifer [/bib_ref] More research is needed on the potentially harmful effects of excessive caffeine consumption, and of new trends in caffeine consumption (eg, energy drinks, synthetic caffeine products) which are popular among corrections workers. Tobacco use was a health risk raised in our focus group study, 37 but survey findings indicated that a relatively small portion of supervisors used tobacco in some form. Chewing tobacco was associated with poorer overall health among supervisors, and over four times higher than the national rate, 67 while cigarette smoking was slightly lower 66 than the national average. The Design Team attributed this discrepancy to the DOC policy explicitly prohibiting cigarette smoking among workers in its facilities-but not smokeless tobacco-and some corrections workers do chew tobacco on the job. As research on tobacco use in correctional settings has almost exclusively focused on health effects of the incarcerated population, [bib_ref] Smoking characteristics of community corrections clients, Cropsey [/bib_ref] [bib_ref] Smoking among female prisoners: an ignored public health epidemic, Cropsey [/bib_ref] [bib_ref] Prison tobacco control policies and deaths from smoking in United States prisons:..., Binswanger Ingrid [/bib_ref] these findings contribute to the very limited research available on tobacco use among corrections workers. [bib_ref] Talking about health: correction employees' assessments of obstacles to healthy living, Morse [/bib_ref] An unexpected positive finding was that supervisors held strong attitudes of personal responsibility for one's health, and the overwhelming majority of supervisors reported having a primary care provider and receiving annual checkups with recommended screenings; much higher rates than in the United States population. [bib_ref] Characteristics of Americans with primary care and changes over time, Levine [/bib_ref] [bib_ref] Improving value in health care-against the annual physical, Mehrotra [/bib_ref] The Design Team attributed these findings to high supervisor enrollment in the recently-implemented State Employees' Health Enhancement Program, which offered cost-sharing reductions to enrollees who committed to having yearly physicals/wellness visits, recommended screenings and preventive care, two dental cleanings, and disease management programs for conditions including diabetes, high blood pressure, heart disease, asthma, and chronic obstructive pulmonary disorder (COPD). ## Attitudes related to overtime and health Survey findings regarding overtime, and related attitudes about prioritizing income and retirement, provided key insights on health effects. Supervisors worked ∼1.5 overtime shifts weekly, and although many thought working less overtime would allow them to be healthier, they felt the need to prioritize earnings over health. Lieutenants reported these attitudes more than higher-ranking supervisors, likely because they worked more overtime. The effect of overtime on well-being is supported by research with corrections workers showing that long hours (>48 hours weekly) are associated with musculoskeletal symptoms, [bib_ref] Investigating the relationship between working time characteristics on musculoskeletal symptoms: a cross..., Garza [/bib_ref] and that work intensity (working 6+ days in a row) is associated with work burnout, likely due to prolonged physical and psychosocial work exposures and insufficient recovery during non-work time. [bib_ref] Working time characteristics and mental health among corrections and transportation workers, Cavallari [/bib_ref] Further, extended and irregular work hours pose barriers to physical and emotional health, social well-being (ie, family, leisure, community participation), and health behaviors (ie, sleep, nutrition, exercise, weight management). [bib_ref] Worker perspectives on the impact of non-standard workdays on worker and family..., Suleiman [/bib_ref] Opportunities to work overtime are well compensated and therefore competitively pursued; however there remains a deficient understanding of the root causes of sacrificing health to maximize earnings (particularly among lieutenants). These root causes include financial and gender role pressures in the family (eg, pressure to fulfill the traditional masculine family breadwinner role) and society (ie, economic volatility). The Design Team noted that although turnover intent was low, work withdrawal was salient. Supervisors frequently thought about retiring (∼7 years in the future) and most planned to retire as soon as eligible. Most knew that their job was associated with decreased longevity and thought they would be happier and healthier after retirement. Compared with higher-raking supervisors, lieutenants more frequently thought about retirement and planned to retire as soon as possible, even though they had a longer wait until retirement. Survey results echoed focus group findings 37 in which the decision by supervisors to remain in a job that poorly affected health and mortality was explained as a constrained choice. Working for a limited career (20 years), having an early retirement (often in their 40s), and receiving a lifetime of retirement income and fully-paid-for comprehensive health insurance, was perceived as an acceptable tradeoff in exchange for the occupational risks. This illustrates the influence of the social ecology on health behavior (eg, policy at the organizational, state, and national levels). The DOC wage structure, which allows workers to substantially increase current and retirement income with extensive overtime motivates them to work the maximum possible overtime in their late career because retirement payouts are based on the highest three earning years. Also, living in the United States, where financial security and full medical insurance are not guaranteed for citizens prior to Medicare eligibility, makes it difficult for workers to turn down a job with post-retirement security in early or late middle age. ## Psychosocial work exposures and outcomes As the focus group study [bib_ref] Participatory survey design of a workforce health needs assessment for correctional supervisors, Dugan [/bib_ref] suggested, there are various positive aspects of correctional work (eg, high job security, satisfaction) and this was confirmed by survey results. In general, supervisors reported a positive psychosocial work environment (eg, meaningful work, supervisor communication, coworker social support) which correlated positively with general health. This new evidence that positive aspects of corrections work do exist, contrasts with past research (and popular media portrayals) that characterizes it as a stigmatized profession with low occupational prestige. [bib_ref] Congruence work in stigmatized occupations: a managerial lens on employee fit with..., Ashforth [/bib_ref] [bib_ref] Good job or dirty work-public perceptions of correctional employment, Sundt [/bib_ref] [bib_ref] Sexuality, masculinity, and taint management among firefighters and correctional officers: Getting down..., Tracy [/bib_ref] Findings also raise the possibility of publicizing and promoting positive aspects of corrections work (identified by the Design Team 37 ) as a strategy to improve personal and occupational selfesteem, possibly enhancing worker mental well-being. Survey results related to culture/climate measures (ie, moderate civility climate and masculine culture, relatively high work health climate ratings) are somewhat incongruent with focus group 37 concerns about an unhealthy DOC culture. In particular, pressures in the DOC environment to adhere to masculine norms were concerning due to their perceived link with certain health risk behaviors (ie, caffeine consumption, tobacco chewing), traditional gender role pressures (ie, family breadwinner), emotional labor (ie, unwavering displays of strength/control, suppression of emotions), and mental health stigma (discouraging acknowledgement of mental health challenges and help-seeking behavior).Yet survey data provided a subtler picture with moderate ratings of masculine culture and emotional suppression, and moderate-highly levels of both social dominance (ie, norms for masculinity) and caring (ie, sociocultural norms for femininity) among both women and men, showing more androgynous (than exclusively masculine) behavior. Survey reporting on gender-related measures may have been influenced by social disability bias, underreporting, or unawareness. The muted depiction could be attributable to the phenomena that in masculine cultures, masculine practices are ubiquitous and grant privilege to people (usually men), who embody and engage in such practices, rendering masculinity invisible to those privileged by it (and highly visible to those without such privilege). [bib_ref] Invisible masculinity, Kimmel [/bib_ref] Given the limited research on the effect of a masculine culture on corrections workers (which to date has mainly pertained to inmate experiences [bib_ref] A prison within a prison? The masculinity narratives of male prisoners, Evans [/bib_ref] [bib_ref] Doing" masculinity as an adaptation to imprisonment, Jewkes [/bib_ref] , focusing on gender is a promising area for further research, especially given post hoc tests suggesting that social dominance may be a health risk factor, while caring seems to be a protective factor. As we found these associations between norms for gender (femininity, masculinity) and health, rather than sex (female, male) and health, it is noteworthy that socially-constructed gender practices can have health implications for people regardless of their sex. [bib_ref] Sex and gender role differences in occupational exposures and work outcomes among..., El Ghaziri [/bib_ref] This supports research recommendations to assess both sex and gender in health research, given that social and behavioral differences may influence the risk of developing certain illness more than biological differences. 101 ## Mental health Compared with focus group findings which brought up concerns around work stress, trauma exposure, PTSD, and suicide, survey results provided a subtler picture of mental health. [bib_ref] Participatory survey design of a workforce health needs assessment for correctional supervisors, Dugan [/bib_ref] Stress and burnout were reportedly moderate, but reports of all psychological symptoms were low. Supervisors attributed greater psychological symptoms to their colleagues than to themselves, with hostility being the most severe symptom reported for both self and others (consistent with masculine emotional expression). Ratings of the effect of traumatic events at work were also low, with respondents reporting being more affected by trauma directed to self/peers than trauma directed to inmates. These low levels could be due to a lack of mental health literacy, a lack of willingness to admit to mental health problems or to caring for inmates (which is contrary to self-presentations of masculine strength), or underreporting resulting from social disability bias within the DOC context, with its high degree of mental health stigma. Despite muted self-reports in survey measures, other findings point to mental health as a concern. Reducing stress was the health behavior that supervisors were most interested in changing, and our findings suggest that work is a particular source of stress. WFC, which is related to mental well-being (depression, anxiety, alcohol/ substance use) [bib_ref] Depression and work family conflict among corrections officers, Obidoa [/bib_ref] [bib_ref] Work-family conflict and employee psychiatric disorders: the national comorbidity survey, Frone [/bib_ref] was moderate, with work interfering with family more than family interfered with work. Post hoc analyses showed positive correlations between experienced trauma and alcohol consumption. Other post hoc analyses showed that lieutenants were more likely to have chronic anxiety/depression compared to high-ranking supervisors. These findings align with prior research showing that correctional workers have higher rates of anxiety and depression, 10 alcohol and drug abuse, [bib_ref] Correctional officer stress: a review of the literature 1977-2007, Morgan [/bib_ref] and suicide 9,10 than other work groups, but additional research is needed to identify the root causes of poor mental health and develop acceptable interventions. This may be a challenge, given worker reticence to acknowledge personal mental health difficulties, much less report about them in surveys (even anonymously) [bib_ref] Process evaluation of two participatory approaches: implementing Total Worker Health1 interventions in..., Dugan [/bib_ref] Focusing on behavior and emotions could provide promising new directions for intervention. Behavior appears to be a driver of conflict because behavior-based WFC and FWC, which occur when behaviors that are effective in one life role are inappropriate in another (eg, when a corrections worker behaves at home in the same rough or aloof manner as at work), were higher than general WFC and FWC. This confirms a focus group concern about the challenge of living in two different worlds (in/out of prison) and managing two personas (work, family). [bib_ref] Participatory survey design of a workforce health needs assessment for correctional supervisors, Dugan [/bib_ref] More research is needed to identify the worker behaviors that are perceived as problematic (or helpful) at home (eg, we found negative correlations between caring behavior and behavior-based WFC). Post hoc analyses also suggest that emotions play a role, given positive correlations between behavior-based WFC and both emotional suppression and emotional job demands. ## Differences in rank We found partial support for the hypothesis that supervisors with the rank of lieutenant are at greater risk for poor health than those with the higher rank of captain or counselor supervisor, as the two groups differed on 15 study variables. Differences pertained mainly to lieutenants having poorer health-related attitudes and health behaviors, and more psychosocial occupational exposures. Fewer differences were found in work and health outcomes. Differences were interpreted by the Design Team as resulting from the distinct social location of lieutenants compared with captains and counselor supervisors. In addition to a lower rank in the organizational hierarchy, lieutenants had a lower education level, had fewer years of tenure, were less likely to work the dayshift, and worked substantially more overtime. Moreover, they work on the front line with COs, exposing them interpersonal conflicts and other stressors that may arise among COs or inmates. Based on these results, the Design Team plans to make special efforts to recruit lieutenants for participation in future health interventions. ## Challenges and lessons learned The participatory survey design process was essential in yielding a highly-detailed, comprehensive picture of supervisors' health. The Design Team perceived it to be time-consuming, yet in the end they were especially satisfied with depth and breadth of the survey results. Another challenge was having numerous survey items representing various domains, which required us to find the optimal way to summarize survey results for the Design Team's review. This is an inherent problem with survey-based research that aligns only with the preferences of the researchers. It was especially helpful for the Design Team to generate specific research questions/ hypotheses that they wanted to examine using the data (ie, flagging variables with especially low or high ratings, identifying variables associated with general health, and comparing two supervisors groups of different rank). They also drew comparisons with comparable US population data when possible, and decided to run post hoc analyses when they wanted further detail about interrelations among study variables to inform health priorities. The use of CBPR and HWPP was particularly helpful in providing a systematic process and evidence-based tools to structure and guide the Design Team's collaborative work, provide momentum, and keep it on track until completion. The CBPR method of involving supervisors in the research process improved our usual survey response rates, promoted diffusion of information about the project within the workforce being surveyed, and generated early interest and buy-in for the interventions that eventually resulted from the health assessment survey. Using HWPP tools was essential to the Design Team's success. Data from their survey (a customized version of the HWPP All Employee Survey) enabled the investigation of specific research questions and a hypothesis which, combinedwith discussion generated by the HWPP Prioritizing and Selecting Safety and Well-being Concerns for Intervention Group Activity, helped the team to see patterns among variables that were helpful in deciding priority health topics. Four particular categories were noted, enabling the team to assess whether health intervention topics would be perceived by supervisors as acceptable and appropriate, two key intervention characteristics that can determine whether future implementation is successful. [bib_ref] Dissemination and implementation research for occupational safety and health, Dugan [/bib_ref] Specifically, in their review of the survey findings, the Design Team first noted that some health problems (eg, poor health outcomes and behaviors such as high BMI, poor sleep, use of alcohol, caffeine, tobacco) were recognized as problematic, with supervisors willing to personally report the problems in a survey, and showing some desire to change them. Second, the Design Team found that some perceived problem areas were unexpectedly revealed as strengths (eg, good health behaviors such as preventative healthcare visits, positive psychosocial work exposures such as high coworker support, and positive outcomes such as high job satisfaction) that supervisors acknowledged; these can now be considered resources and assets within the DOC community, and may be built upon to improve health (identifying community strengths is a key CBPR principle [bib_ref] Development and implementation of principles for community-based research in public health, Schulz [/bib_ref]. Third, some health problems were personally acknowledged by supervisors as problematic (eg, working overtime to maximize income), but they also showed a deliberate desire not to change their situation due to the rewards associated with the behavior (eg, financial security). Fourth, other problems (eg, links between masculine culture, mental health, alcohol use, work-family conflict) were seen by the Design Team as blind spots among supervisors, not recognized as problematic despite a preponderance of evidence, with supervisors unwilling to personally acknowledge the problems and resistant to changing them, in order to uphold a particular self-presentation (ie, strength, control). In the end, health issues that affected a large portion of supervisors and fell into the first category-recognized as problematic, admitted to by supervisors, and amenable to change-were seen as the most promising intervention priorities for obtaining workforce buy in. The Design Team thought it would be futile to begin the health intervention program by addressing blind spots or highly intractable attitudes and behaviors such as working excessive overtime for higher incomes; in fact, team members felt that such change efforts would likely result in backlash against the program. They felt that these more challenging health concerns would require additional time for developing innovative solutions, likely involving educational initiatives to heighten awareness and improve health literacy. Concerns remain about biased reporting using certain survey measures to convey a socially-desirable self-image, and the Design Team is interested in finding innovative solutions to solve this problem. For example, due to underreporting concerns, the Design Team used two versions of the Brief Symptom Inventory to assess psychological symptoms, an original version that asked supervisors to self-report psychological symptoms, and an adapted version that asked supervisors to report perceptions of other supervisors' experiences of the same symptoms. Supervisors perceived peers as having more severe symptomology, in the same order of magnitude as themselves (hostile, anxious, depressive). This provided insight about the psychological health status of a work group that is reluctant to acknowledge personal or emotional difficulties on surveys, but more work is needed to address underreporting among workers in corrections and similar occupations with masculine cultures and mental health stigma. # Strengths and limitations The participatory approach in a unionized public sector workforce is a study strength, although caution should be taken in generalizing findings to private or non-unionized corrections systems. Having a survey designed with input from a large group of supervisors increases the internal validity and applicability of results, ensuring that measures and findings are relevant to our specific population. It also provided a comprehensive picture of the workforce's health. Determining health priorities using a discussion and voting process allowed for informed decision-making and enabled all voices to be heard in an equal way, which is especially beneficial. However, limitations exist from an epidemiological perspective. This study was cross-sectional in nature and limited to only 37% participation, although we were able to confirm that the survey respondents were representative of the larger population. Healthy worker survivor bias may have influenced the findings if workers who were sensitive to the exposures of the job left the workforce and were not included in the cross-sectional survey. If this were the case, the correlation between poor working conditions and general health may actually be stronger than what was observed in the study. Also, rather than using multivariate regression models, simple correlations were used to examine relationships. While this limits the ability to take into account correlations between potential exposures, it was important for us to maintain CBPR methods. The Design Team found correlations to be more accessible than sophisticated models and thus they were used to examine relationships. Limitations regarding the survey include measures being adapted from other studies, or having been originally created by the Design Team. This means that some measures require further development and validation, and prevent us from comparing our study findings with those of other studies. ## Next steps The next step will be to design interventions for the three health priorities in turn, using an HWPP intervention planning tool to brainstorm, evaluate, rank, and select interventions. Design Team discussions thus far favor approaches that are educational, to improve health literacy and self-efficacy, build upon strengths and positive aspects of corrections work to improve self-worth and esteem by others, and capitalize on supervisors' strong sense of personal responsibility for their own health. Another option discussed is obtaining input from DOC supervisors exhibiting ''positive deviance,'' 103 who engage in healthy attitudes and behaviors, value a healthy and fit appearance, and prioritize health and longevity over earnings. Finally, the Design Team is considering policies for negotiation during collective bargaining, as many of their discussions focused on policies as an important determinant of health (eg, the DOC tobacco policy that prohibits smoking but not chewing, the retirement policy with payouts based on highest-year earnings, a scheduling policy that enables excessive overtime, the Health Enhancement Program with incentives for preventative healthcare.) # Conclusion The Design Team successfully used CBPR methods and HWPP evidence-based tools to systematically assess the health of a unionized correctional supervisor workforce in the public sector and select intervention priorities. The findings from the health assessment were instrumental in identifying and ultimately prioritizing health interventions that were important and relevant to the correctional supervisors. These intervention priorities were: improving sleep quality/quantity, stress management/mental health, and weight loss/healthy eating/physical activity. Given the success of this participatory process, several other CPH-NEW projects have adopted the approach in assessing worker health, safety and wellbeing, with promising results. [table] TABLE 1: The Correctional Supervisors' Healthy Workplace Survey: Constructs, Measures, and Number of Items [/table] [table] TABLE 2: Demographics and Work Information of Overall Sample by Job Classification (n = 157) [/table] [table] TABLE 3: Alphas, Means With Standard Deviations, and Correlations for Health Behaviors/Attitudes by Job Classification (n = 157) [/table] [table] TABLE 4: Means With Standard Deviations and Correlations for Attitudes on Longevity, Income, Overtime, and Retirement by Job Classification (n = 157) [/table] [table] TABLE 5: Alphas, Means With Standard Deviations, and Correlations for Psychosocial Work Exposures by Job Classification (n = 157)Bold values indicate statistical significance. Means and standard deviations listed were calculated by excluding missing cases. Civility norms, coworker and supervisor social support, job security, emotional job demands, organizational support for health and safety, work health climate, supervisor communication, masculine organizational culture, effect of trauma, meaningful work, family health climate (1 = strongly disagree, 5 = strongly agree). [/table] [table] TABLE 6: Alphas, Means With Standard Deviations, and Correlations for Health Outcomes by Job Classification (n = 157) [/table] [table] TABLE 8: List of 10 Health Priority Topics for Intervention [/table] [table] TABLE 7: Alphas, Means With Standard Deviations, and Correlations for Work Outcomes by Job Classification (n = 157) [/table]

National School-Based Health Lifestyles Intervention in Chinese Children and Adolescents on Obesity and Hypertension Introduction: This study aimed to examine the effectiveness of the national school-based intervention on both obesity and high blood pressure in Chinese children and adolescents aged 6-18 years.Methods: The national school-based cluster non-randomized controlled trial was done in seven provinces from September 2013 to February 2014. A total of 23,175 children and adolescents in the control group and 25,702 in the intervention group were included in this trial with a mean follow-up of 6.7 ± 0.9 months. Mixed-effects regression models were used to evaluate the effect of the interventions on body weight and blood pressure (BP).Results: A significant upward in the body mass index (BMI) levels but downward in systolic BP (SBP), diastolic BP (DBP), BMI Z-scores, SBP Z-scores, and DBP Z-scores were witnessed in the intervention group compared to those in the control group (<0.001). Subgroup analyses presented significant intervention effects in children aged 6-12 years for BMI, SBP, DBP, and their standardized values Z-scores, but no effective results were found in adolescents aged 13-18 years. Stratification analyses based on the dynamic weight changes presented non-differential HBP, SHBP, and DHBP prevalence gaps between the control and intervention groups. Children aged 6-12 years with higher BMI percentiles at baseline presented obvious declines in SBP and DBP standardized values Z-scores.Conclusion: A mean 6-month multi-centered school-based comprehensive obesity intervention in China yields a small to null effect on obesity and hypertension with increasing age; the early age before 12 years may be the key period for interventions, and the younger, the better. Precise and high-intensity interventions targeting the population at different stages of childhood and adolescence are urgently needed to be developed.Clinical Trial Registration: https://www.clinicaltrials.gov/, identifier: NCT02343588 # Introduction Obesity continues to rise rapidly in many countries. In China, the prevalence of overweight and obesity increased from 4.4% in 1995 to.4% in 2014. Overweight and obesity commonly begin in childhood or adolescence and predict not only continuing obesity but also increased cardiovascular mortality. High blood pressure (HBP) is a further early life indicator of risks of later cardiovascular disease and one that is linked to obesity in childhood. Thus, theoretically, interventions targeting obesity in childhood may bring double benefits from weight control to healthy blood pressure. Schools have been regarded as one promising setting for scalable obesity interventions. Since the 1990s, many local school-based intervention studies for obesity prevention have been conducted among Chinese children and adolescents. Most trials have targeted the lifestyle of school-aged children including dietary behaviors, physical activity, and sleep duration. Some did produce temporary improvements, but without powerful advocating because of the complex and powerful factors that drive the obesity epidemic at the individual, family, school, and societal levels. In addition, a few have attempted to shift the school and family environments, key settings for children and adolescents. Additionally, a few have evaluated the effects of health lifestyles on blood pressure. The Health Lifestyles Intervention in Chinese Children and Adolescents (HLI-CCA) was a multicenter cluster nonrandomized controlled school-based intervention aiming to prevent obesity in children and adolescents. It included elements of education around nutrition and physical activity as well as modifications to school environments and engagement of families. We hypothesized that obesity intervention would have effects not only on obesity but also on the blood pressure in children and adolescents because obesity is becoming one of the leading risk factors for HBP. For this reason, we examined whether the school and family interventions for obesity were effective in controlling and preventing HBP in children and adolescents aged 6-18 years. # Methods ## Study design This national trial with a school-based cluster non-randomized controlled design took place in seven provinces (centers) including Liaoning, Ningxia, Shanghai, Chongqing, Hunan, Tianjin, and Guangdong. This study aimed to determine the effectiveness of HLI-CCA in preventing obesity with 6-month intervention duration from September 2013 to February 2014. The flow of the participants and the trial design are presented in. The full trial protocol has been published previously in detail. ## Sampling and participants A multistage cluster non-randomized sampling was used to obtain the representative sample of children and adolescents in each center, where schools meeting the survey requirements, willing to participate, and fulfilling the inclusion criteria were selected according to the trial protocol. In the end, 12-17 primary and secondary schools were recruited in each center. Altogether, 92 schools were included in our study and were assigned to the control or the intervention group. In each school, all participants selected were extracted from n classes from each grade (n depended on the average size of classes and was no <200 students per school). According to the protocol, the schools and the participants were adjusted slightly to meet the matched standards of balanced schools in the same stratification in each center and the equal distribution of participants in the control or the intervention group (i.e., boy/girl = 1:1, primary school/secondary school = 1:1, urban/rural = 1:1, control/intervention = 1:1). All participants underwent a regular physical examination every year, and those who had one or more of the following conditions were excluded based on the previous medical history and physical examination data: (1) serious organ disease (e.g., heart, lung, liver, and kidney); (2) abnormal physical development (e.g., pygmyism or gigantism); (3) physical impairment or deformity (e.g., severe scoliosis, pectus carinatum, limp, genu valgum, and genu varum); or (4) acute disease symptoms (e.g., diarrhea and high fever) during the past month and not yet recovered. Furthermore, all principals of the selected schools were assessed comprehensively through a face-to-face interview to ensure the smooth execution of the project. According to the trial protocol, a planned sample size of 7,000 has 90% power to detect a 10% difference in the obesity intervention group at 5% statistical significance (two-sided). A sample size of 5,000 had 85% power to detect a 10% difference. Ultimately, a total of 5,1685 participants were included in the present analysis with a mean follow-up of 6.7 ± 0.9 months. ## Intervention The HLI-CCA aimed to deliver a general healthy lifestyle message encouraging a healthy energy balance. In schools assigned to the intervention, the HLI-CCA was delivered to children and adolescents with integrated intervention strategies focused on changing specific behaviors related to energy intake and expenditure, such as decreasing the consumption of sweetened fizzy drinks, increasing the consumption of vegetables, ensuring proper protein intake, reducing sedentary behaviors including screen time, and maintaining at least 1 h of moderate to vigorous physical activity. Integrating the above intervention elements, the comprehensive intervention strategy of the HLI-CCA includes the following four parts during the whole intervention period, which were described in the previously published protocol in detail. As shown in, the intervention model had four strategy designs. Firstly, we targeted the school and family environments through a few strategies with creating a supportive environment of both physical activity and healthier dietary choices. Elements included the provision of the necessary facilities for physical activity in schools, mandating 1-h physical activity time each school day, providing participants and parents with information on obesity prevention through posters, school broadcasts and school website, and restricting sweetened fizzy drinks on the school grounds. Secondly, we developed specialized curriculum and class activities (or campus activities). Intensive health lifestyle educational strategies were established in schools through increasing related knowledge in the health education curriculum, theme class meetings and activities, and health education lectures to parents to enhance the participants' social support in dimensions of peers, schools, and families. Thirdly, strategies were used to instruct and promote school physical education by designing and revising physical education and activities to establish a standardized and rational physical education, such as ensuring enough time for physical education (three times a week) and maintaining moderate to vigorous physical activity 1 h every day under supervision and recording. Fourthly, obesity-related behaviors were monitored and instructed through increasing awareness of self-behaviors using physical activity and dietary behavioral logs, strengthening health lifestyle knowledge with regular measurements of weight and height, and increasing self-efficacy under parental supervision. Project members, school managers, class teachers, school doctors, parents, and students themselves were all involved in the intervention project. The investigators arranged to supervise the intervention in schools throughout the program. All schools completing the trial would be offered all intervention materials from the project team to reduce loss of follow-up. In the control schools, no specific intervention strategies or activities were carried out throughout the study period, but only records of their daily activities according to ordinary monitoring were conducted by project members. Detailed quality control measures and process evaluation were implemented according to the protocol, including the specialized intervention manual, professional trainings for all the intervention school doctors, health education teachers, and physical education teachers, routine supervision by project managers, and appropriate incentives to avoid loss of follow-up. ## Data collection and measurement At baseline and follow-up examinations, information on the demographic characteristics, parental education level, and occupation were collected by questionnaire. All the anthropometric measurements were taken in a sensitive manner in private rooms in both the control and intervention schools. All the measurements followed a standardized procedure by professionals who had passed the training course. Participants underwent a complete anthropometric measurement at baseline and outcome including height, weight, waist circumference, hip circumference, and blood pressure. About 5% of students would be rechecked, and if the error exceeds 10%, all the students should be measured again. All the participants wore only light clothing and stood erect, barefoot, and at ease while being measured. Height was measured to the nearest 0.1 cm with a portable stadiometer (model TZG, China), and weight was measured to the nearest 0.1 kg with a standardized scale (model RGT-140, China). According to the WHO anthropometric standardization and comparison of different standards, waist circumference and hip circumference were measured to the nearest 0.1 cm, and their measurements were located at 1 cm above the umbilicus and the maximal protrusion of the buttocks, respectively. Blood pressure was measured using an auscultation mercury sphygmomanometer (model XJ1ID, China) with an appropriate cuff for children. Three cuff sizes (7, 9, and 12 cm width) were selected according to the age and upper arm circumference of the children, which stipulated that the cuff bladder width should cover 50-75% of the mid-arm circumference. The midupper arm circumference determined the cuff size. The cuff was placed ∼2 cm above the crease of the elbow. The child was seated comfortably for at least 10 min prior to the first reading. The feet of children were placed on a platform during BP measurement. Blood pressure was measured twice, with a 1-min break between each measurement. The participants were asked to remain quiet and to sit still while each reading was being taken. Systolic blood pressure was determined by the onset of the first Korotkoff sound (K1) and diastolic blood pressure determined by the fifth Korotkoff sound (K5). The stadiometers, scales, steel tape, and auscultation mercury sphygmomanometer were calibrated before use, and similar instruments were used in all measurements at the investigated schools. ## Outcomes The two primary outcomes were body mass index (BMI) and blood pressure (BP), including continuous variables [BMI, systolic BP (SBP), diastolic BP (DBP), and their standardized values Z-scores] and categorical variables (overweight and obesity, SHBP, DHBP, and HBP). BMI was calculated as body weight (in kilograms) divided by height (in meters) squared. Overweight and obesity were classified using the sex-and agespecific BMI reference values developed by the International Obesity Task Force (IOTF). BMI Z-scores were calculated according to the U.S. Centers for Disease Control Growth Charts (www.cdc.gov/growthcharts). BMI values were also grouped according to percentiles for age and sex as follows: less than fifth percentile, fifth to 24th, 25th−49th, 50th−74th, 75th−84th, 85th−94th, and 95th percentile or higher (3). BMI Z-scores were also grouped according to their changes during the intervention duration to determine the effects of body weight dynamic changes on the blood pressure as follows: Decline group (BMI Z-score changes of −0.5 or less), Stable group (BMI Zscore changes between −0.5 and +0.5), and Increase group (BMI Z-score changes +0.5 or greater). Systolic HBP (SHBP) and diastolic HBP (DHBP) were defined as SBP and DBP greater than or equal to the reference age-, sex-, and height-specific 95th centile, respectively, according to the National High Blood Pressure Education Program reference in the Fourth Report. HBP was defined as the SHBP or DHBP of children. The SBP and DBP Z-scores were calculated according to reference in the Fourth Report. # Statistical analysis Baseline characteristics were described as the mean (SD) for continuous variables or number (percentage) for categorical variables. The analysis of mixed-effects regression models was used to evaluate the effect of intervention on outcomes after adjusting for age, sex, provinces, urban/rural areas, and the baseline disequilibrium for multiple variables. Two-level (individual level and school level) mixed-effects regression models were used in our analyses due to the grouping structure of the data consisting of multiple levels of nested groups. Linear mixed-effects regression models were used to evaluate the intervention on continuous variables (BMI, SBP, DBP, and their standardized values) with β coefficients, whose positive values meant an increase in the variables after intervention, while negative values meant an effective decline with significant p-values. Similarly, mixed-effects models for binomial responses were used to evaluate the intervention on categorical variables (overweight and obesity, SHBP, DHBP, and HBP) with odds ratios (ORs), whose values more than 1 meant an increased effect in the variables after intervention, while values <1 meant an effective declined effect with significant p-values. Random intercepts with assumed Gaussian distribution were also used in the mixed-effects regression models. Subgroup analyses were used in our study by age groups as follows: 6-9, 10-12, 13-15, and 16-18 years, which were used to identify the effective or sensitive subgroups for intervention. Allowing for the different intervention effects between the age groups in this study and the age composition of the stages of primary and high schools in China, the study further evaluated the differences in the intervention effects between the two age groups of 6-12 and 13-18 years based on the analysis of the four age groups above. Stratification analyses based on the dynamic change of BMI during the intervention periods were used to assess whether the body weight changes had mediating effects on the intervention effects by age groups. Stratification analyses based on BMI percentiles at baseline were used to explore which groups with different initial body weights were sensitive to intervention. Due to the blood pressure sustained increment along with age by genders, we used the blood pressure standardized values to assess the highly sensitive population with weight at baseline. In each BMI percentile group, changed values of variables = (Standardized value Z-scores in post-intervention in intervention group-Standardized value Z-scores at baseline in the intervention group)-(Standardized value Z-scores in post-intervention in the control group-Standardized value Zscores at baseline in the control group). The same analyses of the mixed-effects regression models were conducted for the different BMI percentile groups with β coefficients and ORs to further quantify the intervention effects of BMI on blood pressure. Statistical analyses were conducted using Stata version 14.0 (StataCorp). Statistical significance was defined using a two-sided test with p-values of 0.05. # Results ## Participants The baseline characteristics of the study participants are shown in. The control and intervention groups consisted of 23,175 and 25,702 participants, respectively, with approximately half being males. The mean age at the baseline medical evaluation was 10.9 ± 3.2 and 11.3 ± 3.3 years in the control and intervention groups, respectively. A significant difference between groups was found for parental educational level, parental occupation, and monthly household income, but not in being from a onechild family. The control group had lower height, weight, BMI, hip circumference, waist circumference, and lower prevalence of HBP and SHBP than those in the intervention group. There is no difference in the prevalence of overweight and obesity and DHBP or the levels of SBP and DBP between the two groups. ## Overall intervention effect on weight and blood pressure After an average 6 months' intervention with participation rates of 94.3% in the intervention group and 94.8% in the control group, the intervention group had an increase of BMI values of 0.14 kg/m 2 compared to 0.11 kg/m 2 in the control group. The β coefficient of 0.04 (p < 0.001) indicated a greater increase in BMI in the intervention group. Both control and intervention groups had declines in the prevalence of overweight and obesity, but with a non-significant trend for lower levels in the intervention effects (OR = 0.85, 95% CI = 0.70-1.05, P = 0.134) after adjusting for age, sex, provinces, urban/rural areas, and the baseline disequilibrium for multiple variables. As for the BMI Z-scores, the negative significant β coefficient of −0.0035 (p = 0.002) indicated that there was a significant decrease of the BMI Zscores in the intervention group compared to the control group. An overall larger decrease of the SBP (−0.46 mmHg) and DBP (−0.88 mmHg) levels was found in the intervention compared with the control group (0.33 and −0.41 mmHg). The negative significant β coefficients, after adjusting for age, sex, provinces, urban/rural areas, and the baseline disequilibrium for multiple variables, predicted that the intervention would lead to a 0.77 and a 0.46-mmHg absolute decrease in SBP and DBP, as well as to a 0.047 and a 0.033 decrease in the SBP and DBP Z-scores for children and adolescents aged 6-18 years, respectively (p < 0.001). Similar results were found in both sexes, with 0.78 mmHg (SBP) and 0.40 mmHg (DBP) absolute decreases in boys and 0.75 and 0.53 mmHg absolute decreases in girls, respectively. However, overall, no significant difference for the intervention effects existed for HBP, SHBP, and DHBP, except for DHBP in girls, although we saw a larger decline in the prevalence of HBP, SHBP, and DHBP in the intervention group than that in the control group. For example, the prevalence of HBP decreased from 8.6 to 5.8% with −2.8 percentage points, which was higher than that in the control group (−2.6 percentage points) with nonstatistical effects of the OR value (0.92, 95% CI = 0.83-1.02, p = 0.107). ## Subgroup analysis by age We further divided the participants into four age groups to evaluate the intervention effect on weight and blood pressure, as follows: 6-9, 10-12, 13-15, and 16-18 years. Overall significant intervention effects were found in children aged 6-12 years for BMI, SBP, DBP, and their standardized value Z-scores, with statistically significant β coefficients, particularly in those aged 6-9 years with relatively better intervention effects. However, there seemed no obvious intervention effects for adolescents aged 13-18 years. Similar results were found for the prevalence of overweight and obesity, HBP, SHBP, and DHBP, with significant ORs in children aged 6-12 years, but null ORs in adolescents aged 13-18 years . For example, the intervention resulted in 0.03 kg/m 2 in BMI, 1.31 mmHg in SBP, and 0.98 mmHg absolute decreases and in 19% in overweight and obesity, 22% in HBP, 22% in SHBP, and 26% in DHBP reduced risks in children aged 6-9 years after adjusting for several confounders and the baseline disequilibrium for several indicators. ## Influence of weight change on blood pressure Even though the intervention group in both age groups presented larger decrements in the HBP, SHBP, and DHBP prevalence than those in the control group in each body weight change group, a non-statistically significant difference existed between the control group and the intervention group for prevalence gap with overlapped 95% confidence interval . When further stratified by the BMI percentile categories at baseline, children aged 6-12 years presented more obvious declines in the SBP and DBP standardized value Z-scores in all BMI percentile groups compared to adolescents aged 13-18 years. In group of 6-12-year-olds, there were reductions in the prevalence of HBP, SHBP, and DHBP, even in subjects who increased their BMI Zscores, while in the group of 13-18-year-olds, the increases in the BMI scores were associated with increases in the prevalence of HBP, SHBP, and DHBP. Furthermore, children aged 6-12 years with higher BMI percentiles had larger decrements in the SBP and DBP Z-scores than their peers with lower BMI percentiles. ## Adverse events reports There were no adverse events reported by concerned parents, teachers, or project coordinators. # Discussion To our knowledge, this is the first and the largest trial assessing the intervention effectiveness of obesity prevention on blood pressure in children and adolescents. This average 6-month school-based, family-involved, comprehensive intervention seems not to affect both weight and blood pressure in all the children and adolescents aged 6-18 years being studied. However, subgroup analyses revealed that children aged 6-12 years were more sensitive and the intervention on preventing overweight and obesity and HBP more effective than in those aged 13-18 years. Given the significant downward trend of blood pressure in younger children during the intervention period, it is plausible that the comprehensive obesity-targeted intervention in China has an obvious effect on decreasing the HBP risks in children aged 6-12 years, especially in those with higher BMI levels. Previous studies and pediatric hypertension guidelines had emphasized the importance of comprehensive prevention and intervention strategies on blood pressure, including lifestyle intervention and the control of obesity so as to maintain a healthy blood pressure level in children. However, evidence has been lacking in the pediatrics clinical practice or epidemiological studies. Our findings in the multicenter non-randomized controlled trials (RCTs) demonstrated that the comprehensive schoolbased obesity intervention, which started in early childhood and particularly in those with overweight status, might be effectively used to decrease both obesity and HBP risks. Previous reviews of childhood obesity prevention focused largely on schools were numerous, and the findings have been inconsistent. One meta-analysis found that, of the 147 RCTs conducted to prevent or treat childhood obesity in high-income countries, few have demonstrated successful BMI change, but with small effect sizes. One latest RCT conducted in United States also found that a 36-month behavioral intervention did not change the BMI trajectory among underserved preschoolaged children. The RCT conducted in the UK with an intervention of 24 months also reached an insistent conclusion . β values indicate the effects of interventions for BMI, systolic BP, and diastolic BP using the mixed-effects regression models in each age group adjusting for age, sex (total), and the baseline disequilibrium for socio-demographic indicators. The subfigure (A) reflect the intervention effect in BMI, (B) for BMI Z-Score, (C) for systolic blood pressure, (D) for systolic blood pressure Z-Score, (E) for diastolic blood pressure, and (F) for diastolic blood pressure Z-Score. FIGURE 4 | Analysis by age group of the intervention effects on overweight and obesity (OWOB), systolic high blood pressure (SHBP), diastolic high blood pressure (DHBP), and high blood pressure (HBP). Odds ratios (ORs) indicate the effect of interventions for BMI, systolic BP, and diastolic BP using the mixed-effects regression models in each age group adjusting for age, sex (total), and the baseline disequilibrium for socio-demographic indicators. The subfigure (A) reflect the intervention effect in overweight and obesity, (B) for high blood pressure, (C) for systolic high blood pressure, (D) for diastolic high blood pressure. that no effect was found for the intervention on preventing overweight or obesity. However, some other studies provided evidence on the effectiveness of obesity intervention. One RCT conducted in Sweden found that comprehensive obesity interventions based on parental support program to promote healthy dietary and physical activity habits showed FIGURE 5 | Effect of interventions on high blood pressure (HBP, A for 6-12 years, B for 13-18 years), systolic high blood pressure (SHBP, C for 6-12 years, D for 13-18 years), and diastolic high blood pressure (DHBP, E for 6-12 years, F for 13-18 years) in different BMI Z-score change groups. Three BMI Z-score change groups were defined for the BMI Z-score changes between baseline and post-intervention: BMI Z-score changes between −0.5 and 0.5 as the Stable group, < −0.5 as the Decline group, and more than 0.5 as the Increase group. Red circles represent the prevalence of HBP, SHBP, and DHBP in the control group at baseline and blue circles represent that after the intervention. The prevalence of HBP, SHBP, and DHBP in the intervention group at baseline is represented by red triangles and that after the intervention is represented by blue triangles. Green bars represent the prevalence gaps of HBP, SHBP, and DHBP between the baseline and post-intervention. significant intervention effects. Consistent conclusions were also reported in other studies. However, different settings, a single intervention measure, or the sample sizes may limit scaling up the research. Schools have been identified as a recommended setting for public health strategies to preventing childhood obesity. Our studies added the evidence that a comprehensive intervention in the school setting with family involvement is more effective in preventing overweight and obesity in younger children who are the priority of intervention strategies. The development of precise intervention measures and non-standardized or general intervention programs is warranted for policy-making in the future. Theory and evidence are abundant in the literature about HBP being positively associated with overweight and obesity. In addition, lifestyle factors, including sedentary behaviors, physical activity, unhealthy dietary habits, and adverse family environment, are noted as the major contributing factors to HBP in children and adolescents. In the specific group, our findings confirmed the necessity of a school-based comprehensive obesity intervention implementation combined with lifestyle factors for improvements in children's blood pressure levels and HBP risks. One previous meta-analysis of 25 trials showed reductions in SBP and DBP levels due to weight loss during the intervention period in adults. Feasible biologic pathways may address the effects of weight intervention on the decreased risk of HBP in children. The reninangiotensin-aldosterone system is overactivated early in obese children, whose renin activity and aldosterone concentrations are higher than those in lean peers . But the sympathetic nervous system is more active in hypertensive and obese subjects. Furthermore, weight interventions may inhibit the natriuretic peptide system, of which the functional effects are vasodilatation and natriuresis. In addition, decreased insulin sensitivity and hyperinsulinemia due to weight loss and lifestyle factor reduction might also provide evidence for the decline of hypertension risks. More obvious effects in children with higher BMI percentiles, also in the present study, confirmed the biological pathways above, a shift from an epidemiological perspective. An interesting finding of the present study is that there was a significant age difference in the effective evaluation of the obesity intervention in that children aged 6-12 years showed more sensitivity to the intervention, but not those in middle and late adolescence (aged 13-18 years). Childhood and adolescence are two different but closely linked periods before adulthood. Adolescence is a critical transition period between childhood and adulthood, which brings rapid physical growth and behavioral alterations. However, girls begin reaching puberty with an average age of menarche at about 12 years, although with a 2-year lag in boys. After the onset of puberty, secretions of different hormones begin to accelerate, including folliclestimulating hormone, luteinizing hormone, testosterone, and estradiol levels. It has been proven that an obvious increase in sex hormone secretion has an important influence on blood pressure regulation, whose effects might be more pronounced in middle and late adolescence. In addition, studies have shown that obesity is associated with earlier puberty onset. In addition, it is hard to change some behaviors in adolescents compared to their younger peers. Thus, obesity intervention in childhood, before 12 years, targeting some key factors, such as lifestyle factors and environmental changes, could produce more effective effects on the blood pressure levels and HBP risks. One small pilot trial conducted in Germany also found that early preventive measurements in overweight and obese primary school children were effective in decreasing blood pressure levels. Although no significant effects were found in adolescence, more declines in the BP levels and HBP risks, and even in the significant DBP Z-scores, in adolescents aged 16-18 years after the intervention in the intervention group compared to the control group reflected that the significance of the comprehensive intervention in adolescents may not be apparent temporarily due to a few crucial factors such as the intensity and duration of the intervention measures. Thus, intensive or precise interventions in adolescents could have a positive effect. One study reported the effective effects of a schoolbased obesity intervention on BP levels only in adolescents. They were different from the findings of our study, but provided evidence of the meaning of continued intervention for adolescents. Recent studies confirmed that weight increments in early childhood had a higher risk of cardiovascular disease than those overweight or obese individuals in adolescence. Thus, our study not only confirmed the importance of early obesity intervention for HBP prevention in childhood but also provided directions for future interventions: that preventing HBP in adolescence may require more efforts and larger changes in weight. Given the importance and difficulty of interventions in adolescence, precise interventions, even costly, need to be developed further. Previous studies reported that the risk of HBP would increase for individuals whose body status changed from normal weight to overweight and obesity during a long period, but they failed to examine the effects of weight loss on preventing HBP in adults. Weight changes were important for the control of hypertension in children and adolescents during and after the intervention period. Our analyses found that the significant intervention effects in children aged 6-9 years disappeared after stratification according to the weight changes before and after intervention, with undifferentiated HBP prevalence between the control group and the intervention group. This finding provided evidence that weight changes during the intervention period might mediate the effects of intervention on the blood pressure levels and HBP risks. We also found inconsistent trends of the HBP difference in the two age groups, with declined trends in children aged 6-12 but increased trends in adolescents aged 13-18 years in the increased weight group pre-and post-intervention period, while without statistically significant prevalence gaps. However, the decrement for HBP prevalence in children aged 6-12 years was larger and the increment in adolescents aged 13-18 years was smaller for individuals whose weight had increased in the intervention group than for those in the control group, which also occurred in both children and adolescents whose weights declined or remained stable. Compared to children with lower BMI percentiles at baseline, those with higher BMI percentiles had more obvious intervention effects for HBP prevention, which reflected that children with higher body weight may be more responsive to comprehensive interventions in the short term. Previous studies found that the baseline BMI is associated with a future incident risk of hypertension, even after accounting for weight change during the follow-up period; thus, the impact of obesity on incident hypertension was independent of weight gain among adults. The average BMI used in many other studies was not a complete reflection in predicting the BP changes because it counteracted the difference in the different percentiles, might be in the bottom two, but the changes in BP with BMI percentiles were not even. Therefore, this finding reinforced evidence that sustained weight intervention and management as early as possible in childhood, and throughout adolescence, is effective and necessary in preventing HBP risks. The present study is the first and largest multicenter school-based and family-involved obesity intervention and BP effectiveness assessment study in Chinese children and adolescents. Although the selection of schools in each center did not realize the randomization, willingly participating schools will increase the degree of cooperation for intervention and the effectiveness of the implementation program. Thus, the design of the multicenter cluster non-randomized controlled trial provided a strong reference for obesity intervention on HBP prevention in Chinese children and adolescents with a broad age band and a large sample size, especially valuable for the present weak to null results among adolescents. However, the undesirable intervention results produced in those older might suggest that we should not waste resources in a general intervention among all adolescents in the future. The more precise, the more helpful; in other words, adolescents could not obtain more benefits from a general intervention, but more need of a precise intervention because the target population among them is small. In addition, the primary outcomes of the present study were fully assessed using a comprehensive set of variables around the theme of weight and blood pressure in order to capture potential intervention effects on later cardiovascular risks, including weight and blood pressure levels and their standardized (including BMI, SBP, and DBP Z-scores) and risk (including overweight and obesity, SHBP, DHBP, and HBP) outcomes. However, three limitations should be mentioned. Firstly, because the study was conducted among Chinese children and adolescents, the findings might not be generalized to other populations. However, a lot of previous literatures were conducted in high-income countries, so our study should provide a good reference for middle-and low-income countries. Secondly, the present results adopted the classical BP criteria developed in American children in 2004, and we did not use the updated US BP criteria in children, but the 2004 criteria were widely applied by researchers, which makes it easy to directly compare our results to those of other studies. The latest studies have demonstrated that the associations between HBP with BMI and other medical and behavioral factors remained unchanged using the updated US BP criteria, so our results would not be affected by the different BP definitions. Thirdly, data on lifestyle factor changes, such as dietary behaviors, physical activity, health literacy, and family support, during the followup period and their independent effects on obesity and HBP risks were not analyzed, which need future research. # Conclusion In summary, a 6-month multicenter school-based and familyinvolved obesity intervention in China did not observe overall risk reductions of overweight and obesity and HBP in children and adolescents aged 6-18 years. However, children aged 6-12 years could be more sensitive to the intervention, with a more significant drop in overweight and obesity and in HBP risk than those aged 13-18 years. The intervention in children aged 6-12 years with overweight status can yield a more effective blood pressure control. Therefore, a comprehensive health lifestyle intervention was needed as early as possible in school settings in order to reduce the risks of obesity, HBP, and other cardiovascular diseases in the future. Future research with longerperiod precise interventions should assess the effectiveness of interventions in adolescents. # Data availability statement The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation. Proposals should be directed to [email protected] and [email protected]. # Ethics statement This project was approved by the Medical Research Ethics Committee of Peking University Health Science Center (IRB00001052-13034). Written informed consent to participate in this study was provided by the participants' legal guardian/next of kin. # Author contributions YD conceptualized and designed the study, completed the statistical analyses, drafted the initial manuscript, and reviewed and revised the manuscript. YS, YM, and JM contributed to the conceptualization and design of the study, supervised the data collection, statistical analyses, initial drafting of the manuscript, and reviewed and revised the manuscript. ZZ and HW assisted with the statistical analyses and critically reviewed and revised the manuscript. PH and BD assisted with the data processing, statistical analyses, and the interpretation of the data. All authors approved the final manuscript as submitted and agreed to be accountable for all aspects of the work.

Analysis of Transformation Plasticity in Steel Using a Finite Element Method Coupled with a Phase Field Model An implicit finite element model was developed to analyze the deformation behavior of low carbon steel during phase transformation. The finite element model was coupled hierarchically with a phase field model that could simulate the kinetics and micro-structural evolution during the austenite-to-ferrite transformation of low carbon steel. Thermo-elasticplastic constitutive equations for each phase were adopted to confirm the transformation plasticity due to the weaker phase yielding that was proposed by Greenwood and Johnson. From the simulations under various possible plastic properties of each phase, a more quantitative understanding of the origin of transformation plasticity was attempted by a comparison with the experimental observation. # Introduction The transformation plasticity is believed to be a deformation mechanism that causes permanent deformation during the phase transformation of allotropic polycrystalline materials, even under an extremely small applied stress. For ideally plastic materials, Greenwood and Johnson [bib_ref] The deformation of metals under small stresses during phase transformations, Greenwood [/bib_ref] derived an analytical solution for permanent strain due to the transformation plasticity assuming that plastic deformation occurs in a weaker phase to accommodate the external and internal stresses caused by volume mismatch between two allotropic phases. The transformation plastic strain increment De tp under an uniaxial stress state was derived as follows: [formula] De tp~5 3 DV V s s yð1Þ [/formula] where DV/V is the absolute value of the volume mismatch, and s and s y are the externally applied stress and uniaxial yield stress of the weaker phase, respectively. Although their description is a widely accepted in the diffusional transformation plasticity, a later study by Zwigl and Dunand [bib_ref] A non-linear model for internal stress superplasticity, Zwigl [/bib_ref] showed that Greenwood and Johnson's derivation was valid only for small applied stresses compared to the yield stress. In their work [bib_ref] A non-linear model for internal stress superplasticity, Zwigl [/bib_ref] , Greenwood and Johnson's theory was extended to relatively higher applied stress. However, the extended analytical solution could not provide any information on the internal or macroscopic strains that are dependent on time during a phase transformation. A couple of years later, they proposed a numerical model [bib_ref] A numerical model of transformation superplasticity for iron, Zwigl [/bib_ref] that can generate time dependent information as well considering the temperature dependent properties of a material. The proposed model of transformation plasticity for an elastic, ideally plastic material was established through a two dimensional plane strain formulation considering both the temperature and displacement. Recently, Greenwood and Johnson's model, so called the internal stress model, was elaborated theoretically as an explicit expression of the transformation plastic strain rate from the effort of Taleb and Sidoroff [bib_ref] A micromechanical modeling of the Greenwood-Johnson mechanism in transformation induced plasticity, Taleb [/bib_ref]. They improved the micro-mechanical model originally suggested by Leblond et al. [bib_ref] Mathematical modeling of transformation plasticity in steels I: Case of ideal-plastic phases, Leblond [/bib_ref] by removing some assumptions: elastic behavior of the product phase, and rigid plastic behavior of the parent phase. All these studies were based on conventional plasticity theory or the continuum mechanics. Transformation plasticity is closely related to a phase transformation including interfacial movement, a morphologic construction, and other kinematical phenomena, of which combination eventually produces actual microstructure. However, the continuum-based theories have an obvious limitation in understanding the transformation plasticity because information on the microstructural evolution during phase transformation is absent. For this reason, in previous studies based on macroscopic conventional plasticity, the transformation plastic strain rate was adopted as an additional strain rate to reflect the additional plastic deformation caused by transformation plasticity. Computer simulation methods have been effectively used to better understand phase transformations. The cellular automata (CA) method, for example, was used for the austenite-to-ferrite transformation in steels [bib_ref] Modeling austenite decomposition into ferrite at different cooling rate in low-carbon steel..., Lan [/bib_ref] [bib_ref] Prediction of phase transformation and microstructure in steel using cellular automaton technique, Kundu [/bib_ref]. The CA method simulates the impingement between newly formed grains well, but cannot take into account grain coarsening. A phase field model (PFM) has many advantages for the analysis of phase transformations comparing to the CA method. The PFM can handle grain coarsening and the impingement phenomenon, as well as consider diffusion, interface mobility, and the effect of interface energy [bib_ref] The phase-field approach and solute drag modeling of the transition to massive..., Loginova [/bib_ref]. Therefore, for the clearer understanding of transformation plasticity, it will be beneficial to incorporate microstructural information obtained from PFM into conventional continuum based theories, which have been suggested to interpret the transformation plasticity. Many PFMs have been reported for the austenite-to-ferrite transformation. Yeon et al. [bib_ref] A phase field study for ferrite-austenite transitions under para-equilibrium, Yeon [/bib_ref] modified their phase field model for multicomponent alloy solidification [bib_ref] A phase field model for isothermal solidification of multicomponent alloys, Cha [/bib_ref] to describe the austeniteto-ferrite transformation of Fe-C-Mn ternary alloy under paraequilibrium. Mecozzi et al. [bib_ref] Analysis of the cRa transformation in a C-Mn Steel by phase-field modeling, Mecozzi [/bib_ref] modified the model proposed by Steinbach et al. [bib_ref] A phase field concept for multiphase systems, Steinbach [/bib_ref] and analyzed the microstructure evolution in the austenite-ferrite transformation of Fe-C-Mn alloy. Huang et al. [bib_ref] A phase-field simulation of austenite to ferrite transformation kinetics in low carbon..., Huang [/bib_ref] combined the solute diffusion model by and the model proposed by Warren et al. [bib_ref] Extending phase field models of solidification to polycrystalline materials, Warren [/bib_ref] for multi-phase field and analyzed the microstructure evolution in the austenite-ferrite transformation of Fe-C binary alloy. Recently Cha et al. [bib_ref] Phase field study on the austenite/ ferrite transition in low carbon steel, Cha [/bib_ref] proposed the PFM for the ferrite growth in the austenite polycrystal including the effect of transformation stress. In this study, we adopted the PFM proposed by Cha et al. [bib_ref] Phase field study on the austenite/ ferrite transition in low carbon steel, Cha [/bib_ref] to describe the ferrite growth in the polycrystalline austenite microstructure. The effect of transformation stress, which requires expensive cost in the calculation, was not considered because our focus is not to develop rigorous PFM, rather than to investigate transformation plasticity associated with realistic evolution of microstructure. An implicit numerical solution procedure to calculate the deformation during the phase transformation of low carbon steel was implemented into the general purpose implicit finite element (FE) program. The procedure was coupled hierarchically with a PFM that could simulate the kinetics and microstructural evolution of the austenite-to-ferrite transformation, to evaluate the internal stress from volume mismatch between each phase involving the transformation. Therefore, any additional strain rate term for the transformation plastic deformation is not necessary to analyze the transformation plasticity. Only the thermo-elastic-conventional plastic constitutive equations for each phase were adopted to confirm the transformation plasticity due to the weaker phase yielding that was proposed by Greenwood and Johnson [bib_ref] The deformation of metals under small stresses during phase transformations, Greenwood [/bib_ref]. From the simulation, the origin of the transformation plasticity was discussed quantitatively in the context of Greenwood and Johnson's model [bib_ref] The deformation of metals under small stresses during phase transformations, Greenwood [/bib_ref] and compared with the other possible mechanisms [bib_ref] Diffusion-controlled transformation plasticity of steel under externally applied stress, Han [/bib_ref] [bib_ref] A model for transformation plasticity during bainite transformation of steel under external..., Han [/bib_ref]. # Methods ## Constitutive formulations The Cauchy stress increment, ds, is [formula] ds~C e : de eð2Þ [/formula] where C e and de e are the elastic stiffness tensor and elastic strain increment, respectively. In the matter of elastic stiffness, isotropic elastic modulidepending on the temperature were used, as listed in [fig_ref] Table 1: Elastic modulus of steel at various temperatures[20] [/fig_ref]. Poisson's ratio of the steel was assumed to be constant, 0.3 [bib_ref] A micromechanical modeling of the Greenwood-Johnson mechanism in transformation induced plasticity, Taleb [/bib_ref]. The total strain increment, de T is [formula] de T~d e e zde v zde pð3Þ [/formula] where de v is the volumetric strain increment due to the phase transformation and temperature variation, and de p is the conventional plastic strain increment. For the volumetric strain calculation, the linear mixture of strain increments of existing phases is assumed as follows: [formula] de v~X a de v a zX c de v c~X a : 1 3 r a dr a IzX c : 1 3 r c dr c ! Ið4Þ [/formula] where X and r are the phase fraction and density of each phase. The subscript a and c mean the ferrite (a-iron) and austenite (ciron), respectively. I is identity tensor. The densities of austenite and ferrite were defined separately as a function of the temperature and chemical composition from Miettinen's data, as listed in [fig_ref] Table 2: Densities of austenite and ferrite phase as a function of temperature and... [/fig_ref]. For the plastic strain increment, the flow rule can be written as follows under the von-Mises criterion. [formula] de p~d e p 3 2 S s Yð5Þ [/formula] where de p is the equivalent plastic strain increment, S is deviatoric stress, and s Y is the yield stress. The yield condition is [formula] s Y~s0 e p ,T ð Þð6Þ [/formula] where s 0 (e p , T) is a function of equivalent plastic strain (e p ) and temperature (T), which is defined for each phase. However, the reported data on the independent plastic behavior for each phase at a given temperature are found very rarely in literature, because the austenite and ferrite in low carbon steel generally coexist during the transformation. Here, based on the previous experimental data, it was assumed that the yield stresses of austenite and weaker ferrite phase are 100,300 MPa and 80,240 MPa with a linear work hardening rate of 8 MPa, respectively. The stress increment then becomes [formula] ds~C e : de T {de v {de p À Áð7Þ [/formula] The constitutive formulations were incorporated into the user material subroutine UMAT of ABAQUS/Standard, a commercial FE program. ## Phase field model If we consider polycrystalline system where ferrite phase coexists with austenite phase, the governing equation of phase field model for austenite-to-ferrite transformation is [bib_ref] Phase field study on the austenite/ ferrite transition in low carbon steel, Cha [/bib_ref] : [formula] Lw q Lt{ 2 S X n r=q s q s r M qr dF dw q { dF dw r zf q (u q ){f r (u r ){(u q {u r )m m !ð8Þ [/formula] where a and c represent ferrite (a-iron) and austenite (c-iron), respectively. The order parameter w q (q = 1, 2, 3, …, n) gives the orientation state of a point in a polycrystalline system containing n grains, and the sum of all phase-field values in a point (i, j, k) is conserved as: [formula] X n q~1 w q i,j,k ð Þ~1ð9Þ [/formula] If q.n/2, we define w q as the orientation state of a point in the polycrystalline ferrite phase. Likewise, q,n/2 corresponds to austenite phase. A step function s q = 1, if w q .0 and s q = 0 otherwise. The number of grains coexisting in a given point is S i,j,k ð Þ~P n q~1 s q i,j,k ð Þ. M qr is the phase field mobility. u q and f q (u q ) are carbon concentration and the free energy density of q grain.m m is the chemical potential of carbon, and we used the following constitutive equation [bib_ref] Phase-field model for binary alloys, Kim [/bib_ref] [bib_ref] Phase field study on the austenite/ ferrite transition in low carbon steel, Cha [/bib_ref] : [formula] df a du a~d f c du c :m mð10Þ [/formula] The total free energy functional F includes the grain boundary (GB) or interphase boundary (IB) energy density F GB [bib_ref] Phase field study on the austenite/ ferrite transition in low carbon steel, Cha [/bib_ref] : [formula] F~ð V F GB dV~ð V X n r,q~1 e 2 rq 2 +w r : +w q zv rq w r w q dVð11Þ [/formula] The parameters v rq , e rq , and M qr in Eqs. [bib_ref] The phase-field approach and solute drag modeling of the transition to massive..., Loginova [/bib_ref] and (11) have the definite relationship with the GB or IB energy s qr with its width 2j qr as follows: [formula] v qr~2 s qr j qr , e rq~4 p ffiffiffiffiffiffiffiffiffiffiffiffi j qr s qr qð12Þ [/formula] and the mobility M qr is determined by the GB or IB mobility. The governing equation of carbon diffusion is given as follows [bib_ref] Phase field study on the austenite/ ferrite transition in low carbon steel, Cha [/bib_ref] : [formula] Lu Lt~+ : D a h(w a )+u a zD c ½1{h(w a )+u c À Áð13Þ [/formula] where D a and D c are the carbon diffusivities in the ferrite and austenite phases, respectively, and w a is defined as the sum of the ferrite phase fields existing in a point, w a : P iwn=2 w i . All points in the system are considered as the mixture of the ferrite and austenite phases and their fractions are given by h(w a ) and (12h(w a )), respectively. A function, h(w a ), is a monotonically increasing function for w a between h(0) = 0 and h(1) = 1. The concentration, u, in the left-hand side of Eq. (13) is defined as [formula] u~h(w a )u a z½1{h(w a )u cð14Þ [/formula] Although various functions for h(w a ) can be used, this study employed the following function to minimize the solute trapping [bib_ref] Large-scale three-dimensional simulation of Ostwald ripening, Kim [/bib_ref] h(w a )~0 [formula] a t w a v 1 2 {D 1 2 z 2w a {1 4w a at 1 2 {Dƒw a ƒ 1 2 zD 0 a t w a w 1 2 zD 8 > > > > > > < > > > > > > :ð15Þ [/formula] The phase field mobility, M qr , is obtained from the interface mobility, M ij exp , which is measured experimentally. The following relationship between M ij and M ij exp is obtained at a thin interface limit [bib_ref] Phase-field model with a reduced interface diffuseness, Kim [/bib_ref] : for the ferrite/austenite interface: [formula] E ij M ij e 2 ij~1 M exp ij z e ij (u e c {u e a ) 2 ffiffiffiffiffiffiffiffiffi 2$ ij p ð 1 0 Y w a ½1{h(w a ) ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi w a (1{w a ) p dwð16Þ [/formula] where [formula] E ij M ij e 2 ij~1 M exp ijð17Þ [/formula] Hierarchical multi-scale modeling: PFM into FEM The PFM and finite element model (FEM) were established within a two-dimensional square shape domain, which has 256 mm on one side, as shown in [fig_ref] Figure 1: Two-dimensional square domain in PFM and FEM calculation [/fig_ref]. The domain is divided into 5116511 elements (or grids) in both calculations. This fine mesh system is needed because localized micro-stress field, which is caused by the volume change due to nucleation and growth of a newly formed phase, should be evaluated precisely to analyze the transformation plasticity. According to Greenwood and Johnson's model [bib_ref] The deformation of metals under small stresses during phase transformations, Greenwood [/bib_ref] and other previous studies [bib_ref] A non-linear model for internal stress superplasticity, Zwigl [/bib_ref] [bib_ref] A numerical model of transformation superplasticity for iron, Zwigl [/bib_ref] [bib_ref] A micromechanical modeling of the Greenwood-Johnson mechanism in transformation induced plasticity, Taleb [/bib_ref] [bib_ref] Mathematical modeling of transformation plasticity in steels I: Case of ideal-plastic phases, Leblond [/bib_ref] , the localized microstress field can generate transformation plasticity, which originates from the micro-plasticity of the weaker phase. [fig_ref] Figure 2: Flow of PFM-FEM hierarchical multi-scale simulation [/fig_ref] shows the flow of a hierarchical multi-scale simulation that makes a connection between the PFM and FEM. Phase information for each element and each time step obtained from the PFM calculation is transferred into the FEM. However, for the temperature information, the temperature history in both cases is quite similar for the small calculation domain due to the relatively high conductivity of the steel. Therefore it is reasonable to assume a homogeneous temperature distribution within the entire domain. In the boundary conditions, the periodic boundary conditions (PBCs) should be adopted into the FE calculation to correspond to the PFM calculation. The unit cell size in the PFM is not changed during the calculation due to its finite difference scheme. However, the unit cell size in the FEM would be expanded or contracted according to the result of stress/displacement analysis. To be adequate requirements of PBCs in FEM, a linear multi-points constraint was applied to the boundaries of the FE domain. For example, let the lower edge of the domain in [fig_ref] Figure 1: Two-dimensional square domain in PFM and FEM calculation [/fig_ref] be composed of (n-m) nodes (node number: m, m+1, …, n-1, n), as shown in [fig_ref] Figure 3: Schematic diagram of lower edge of FE calculation domain [/fig_ref]. The linear multi-points constraint for the lower edge of the domain is as follows: where Z represents a nodal variable, the subscript means i-th degree of freedom, and the superscript means node number. In the above case, the i-th degree of freedom should be the displacement along the second axis. The phase information for each position at each time step, which was calculated from PFM, was transferred to the corresponding element in the FEM that has exactly equivalent domain with PFM. In the PFM, the phase information is obtained as binary numbers, 0 or 1, for each node. Number 0 represents the austenite phase, and number 1 represents the transformed ferrite phase. Since the FE calculation of volumetric strain increment is performed on the Gaussian integration point of each element, not the nodal points, an interpolation technique for determining phase fraction was adopted in the FE calculation. The volumetric strain increment due to both phase transformation and temperature change was calculated in the FEM by using Eq. (4) and the densities as a function of temperature and chemical composition, which are listed in [fig_ref] Table 2: Densities of austenite and ferrite phase as a function of temperature and... [/fig_ref]. # Results and discussion Hierarchical simulation for austenite-to-ferrite transformation The evolution of austenite-to-ferrite transformation of low carbon steel was simulated using PFM assuming the chemical composition to be 0.003C-1.1 Mn (in wt.%). Steel with an initial temperature of 865uC was cooled continuously to 801uC with a cooling rate of 1uC/sec. The interface mobility between austenite and ferrite and the grain boundary mobilities of austenite and ferrite were assumed to be isotropic and to have the value obtained from the austenite/ferrite interface mobility proposed by Wits et al. [bib_ref] A study on the austenite-to-ferrite phase transformation in binary substitutional iron alloys, Wits [/bib_ref]. The austenite/ferrite interface energy (c ca ) and grain boundary energies of austenite and ferrite are as follows: grain boundary energies of austenite and ferrite are 1.95 and 1 J/m 2 , respectively, and c ca = 1 J/m 2 . This combination of the interface and grain boundary energies produces ,26 degrees of wetting angle. Due to the wetting angle, the ferrite phase nucleated on the grain boundary of austenite phase grows along austenite grain boundary like allotriomorph. It was assumed that the diffusion of Mn is negligible. shows the calculated microstructures, which consist of initially full austenite and finally 80% ferrite with 20% austenite remaining. For validation of the calculated microstructures, the evolution of transformed microstructure was obtained experimentally. The quenching process, which means very rapid cooling, causes a diffusionless, martensitic transformation in the austenite remaining in the steel. This allows a measurement of the morphological features of ferrite and austenite during the transformation. The specimen for the experimental validation was 0.15C-1.4 Mn-0.25Si steels. The carbon content was designed to be slightly higher comparing to the PFM calculation to secure sufficient hardenability of the steel. Quenching after the small austenite-to-ferrite transformation was performed using a hot deformation simulator (THERMEC MASTER Z). The specimen was prepared as a cylindrical shape with 8 mm (W, diameter)612 mm (length). The route of thermal processing is listed as follows: (a) Heating to 1150uC at a heating rate of 5uC/sec (b) Holding for 3 minutes (c) Cooling to 700uC with a cooling rate of 2uC/sec (d) Holding for 10, 20, 30 and 40 seconds (e) Quenching with He gas [fig_ref] Figure 6: Calculated distribution of von-Mises stress during austenite-to-ferrite transformation [/fig_ref] shows the calculated distribution of von-Mises stress during the austenite-to-ferrite transformation of steel. The microstress field developed by the evolution of ferrite was calculated from FE analysis. [fig_ref] Figure 7: Calculated distribution of equivalent plastic strain during austenite-to-ferrite transformation [/fig_ref] presents the distribution of equivalent plastic strain during the austenite-to-ferrite transformation. The evolution of micro-plasticity due to an externally and internally developed stress is concentrated mainly in the weaker ferrite phase. It is also found that the front position (see arrows in [fig_ref] Figure 7: Calculated distribution of equivalent plastic strain during austenite-to-ferrite transformation [/fig_ref] of the growing ferrite grains receives relatively larger plastic deformation. This result appears to be in agreement with the internal stress model by Greenwood and Johnson [bib_ref] The deformation of metals under small stresses during phase transformations, Greenwood [/bib_ref] , which suggests weaker phase yielding. However, micro-plastic deformation was also observed in the stronger austenite region, even though the amount of deformation is relatively small. It is possible that this result could be affected by the material properties used in the calculation, such as the hardening curves or yield stress ratio of ferrite to austenite. These effects will be discussed in the next section. From the above procedure, transformation plasticity, which might be caused by micro-plasticity in the weaker phase, can be evaluated without any additional terms for the transformation plastic strain with assistance of (a) accurate morphological data of the microstructure from the PFM calculation and (b) a very fine mesh system that is sufficient for application to the micro-scale phenomenon. However, the quantitative amount of transformation plasticity could not be obtained directly from the calculated deformation, since the amounts of deformation along each side of the domain are not accurately equal to each other, even without external stress. In other words, the deformation of a domain could not be isotropic, even though there is no transformation plasticity. This means that the amount of non-isotropic deformation in this calculation does not match the amount of transformation plasticity directly. Moreover, the total amount of deformation contains not only the transformation plastic deformation but also the conventional plastic, thermal, and elastic deformation not relevant to the transformation plasticity. Here, the amount of transformation plasticity was evaluated as follows. (a) Calculate the deformation in a reference state that represents the deformation without externally applied stress. The elastic spring-back was also considered in the determination of transformation plasticity. ## Analysis of transformation plasticity The principal factors that possibly affect the transformation plastic deformation in the present calculations can be classified into (a) an externally applied stress, (b) yield stress of each phase, and (c) yield stress ratio of ferrite to austenite. Initially, variations in the transformation plastic deformation with externally applied stresses along horizontal direction of the domain were evaluated. The applied stress was varied as 2, 3, 5 and 8 MPa, and the other calculation conditions, such as the yield stress of austenite and ferrite, and the cooling rate were the same as the calculation in the previous section. For a comparison, the experimental data in the previous study was adopted. The experimentally measured transformation plasticity has been reported in many studies [bib_ref] The deformation of metals under small stresses during phase transformations, Greenwood [/bib_ref] [bib_ref] A model for transformation plasticity during bainite transformation of steel under external..., Han [/bib_ref] [bib_ref] A model for deformation behavior and mechanically induced martensitic transformation of metastable..., Han [/bib_ref] [bib_ref] TRIP-assisted steels?, Bhadeshia [/bib_ref] [bib_ref] A constitutive model for transformation superplasticity under external stress during phase transformation..., Han [/bib_ref] [bib_ref] Experimental analysis of transformation plasticity, Taleb [/bib_ref] [bib_ref] Transformation induced plasticity in maraging steel: An experimental study, Nagayama [/bib_ref] [bib_ref] Experimental study of the phase transformation plasticity of 16MND5 low carbon steel..., Coret [/bib_ref] [bib_ref] A new view on transformation induced plasticity (TRIP), Fischer [/bib_ref] [bib_ref] High-rate tensile properties of Si-reduced TRIP sheet steels, Choi [/bib_ref] [bib_ref] Modeling of cementite precipitation kinetics on solute carbon content in extra and..., Choi [/bib_ref]. A few years ago, some of the present authors [bib_ref] Diffusion-controlled transformation plasticity of steel under externally applied stress, Han [/bib_ref] suggested a model for transformation plasticity in diffusional transformation with the experimental results during the austeniteto-ferrite transformation in plain low carbon steel. These experimental results are compared with the calculated results in the present study. [fig_ref] Figure 8: Comparison between measured [18] and calculated transformation plastic strains according to externally... [/fig_ref] shows the measured [bib_ref] Diffusion-controlled transformation plasticity of steel under externally applied stress, Han [/bib_ref] and calculated transformation plastic strains corresponding to externally applied stresses of 2, 3, 5 and 8 MPa. As expected from the internal stress model [bib_ref] The deformation of metals under small stresses during phase transformations, Greenwood [/bib_ref] , the transformation plastic deformation was linearly proportional to the applied stress in both the measured and calculated cases. However, the measured amounts of transformation plastic deformation were significantly higher than the calculated ones. Since this disagreement might be caused by uncertainty in the mechanical properties of ferrite and austenite, several case studies were carried out under the conditions of the possible yield stress ranges of ferrite and austenite based on the previous experimental data [bib_ref] Large-scale three-dimensional simulation of Ostwald ripening, Kim [/bib_ref]. The amount of transformation plastic deformation varying with the yield stress of austenite and ferrite was evaluated in a consistent manner. The externally applied stress was fixed to 2 MPa. The yield stress of ferrite, which is assumed to be 80% of austenite strength, was varied as 80, 120, 160, 200 and 240 MPa. The calculated results are represented in [fig_ref] Figure 9: Measured [18] and calculated transformation plastic strain according to yield stress of... [/fig_ref]. As shown in the figure, the amount of deformation due to transformation plasticity is in inverse proportion to the yield stress of ferrite. This tendency is also consistent with the internal stress model of Eq. (1), which suggests that the strain rate due to transformation plasticity is inversely proportional to the yield stress of the weaker phase. Another interesting point is that the transformation plasticity obtained under the condition of the extremely lowest yield stress of ferrite was still about over 4 times smaller than the measured one. Lastly, the amount of transformation plastic deformation varying with the yield stress ratio of ferrite to austenite was obtained from FE calculations. In this case, the externally applied stress was fixed to 2 MPa, and the yield stress of austenite was fixed to 200 MPa as well. The yield stress of ferrite was changed as 140, 160, 180, and 200 MPa, which correspond to the yield stress ratio of 0.7, 0.8, 0.9, and 1.0, respectively. [fig_ref] Figure 1: Two-dimensional square domain in PFM and FEM calculation [/fig_ref] shows the calculated transformation plasticity according to the yield stress ratio of ferrite to austenite. As the yield stress of ferrite approaches that of austenite, the transformation plastic deformation decreases because (a) the micro-plastic deformation by the evolution of phase transformation would be distributed more uniformly into both phases and (b) the yield stress of ferrite is increased. [fig_ref] Figure 1: Two-dimensional square domain in PFM and FEM calculation [/fig_ref] shows the difference in the equivalent plastic strain distribution for the two extreme cases in these calculations, 0.7 and 1.0 yield stress ratio. The figure also shows that plastic deformation occurs in both stronger austenite phase and weaker ferrite phase. Indeed, the plastic deformation in the stronger austenite phase always occurs in all the calculations in this study. Therefore, the micro-plastic deformation during the phase transformation is not confined to the weaker phase, which is in contrast to that suggested by Greenwood and Johnson [bib_ref] The deformation of metals under small stresses during phase transformations, Greenwood [/bib_ref]. From Figs. 8, 9 and 10, it might be confirmed that only the conventional plasticity, which is caused by the volume mismatch between the prior and resultant phase, might not be sufficient to explain the transformation plastic deformation experimentally observed. In other words, the transformation plasticity might require the other mechanisms in addition to the Greenwood and Johnson's one. Indeed, Han et al. [bib_ref] An observation of permanent strain during recrystallization and growth of steel under..., Han [/bib_ref] [bib_ref] Diffusion-controlled recrystallization and grain growth-induced plasticity of steel under externally applied stress, Han [/bib_ref] reported that considerable permanent strain was observed in extra low carbon steel during the recrystallization and growth under a small applied stress, even much lower than the yield stress. Since there is little volume mismatch in recrystallization and growth, the model based on the internal stress model by Greenwood and Johnson [bib_ref] The deformation of metals under small stresses during phase transformations, Greenwood [/bib_ref] is difficult to explain the above permanent deformation. Therefore, they suggested a model based on a migrating interface diffusion mechanism [bib_ref] Diffusion-controlled transformation plasticity of steel under externally applied stress, Han [/bib_ref] [bib_ref] Diffusion-controlled recrystallization and grain growth-induced plasticity of steel under externally applied stress, Han [/bib_ref] , which was formulated as an accelerated Coble creep, to explain the permanent deformation due to both the phase transformation and recrystallization/growth. This model was confirmed recently by Cho et al. [bib_ref] A finite element modeling for dilatometric nonisotropy in steel, Cho [/bib_ref] for the case of various heat treatments on low and ultra low carbon steels. They incorporated this model into an implicit FEM, then showed the simulated results were in good agreement with the observed deformation behaviors. As another possible mechanism for transformation plasticity, the selection of the specific variant during the phase transformation under external stress could be considered. In this mechanism, the preferentially selected variant with the specific orientation relationship between the variant and the applied stress field could lead to the occurrence of transformation plasticity [bib_ref] A model for transformation plasticity during bainite transformation of steel under external..., Han [/bib_ref] [bib_ref] A model for deformation behavior and mechanically induced martensitic transformation of metastable..., Han [/bib_ref]. # Conclusion To calculate the deformation behavior during the phase transformation of low carbon steel, the numerical procedure was coupled hierarchically with PFM and FEM. PFM could simulate the kinetics and micro-structural evolution during the austenite-toferrite transformation. Phase information for each element and each time step obtained from the PFM calculation was transferred into the thermo-elastic-plastic FEM. From the developed method, it was confirmed that the transformation plasticity could be caused by a conventional plastic deformation of the weaker phase, which was suggested as the internal stress model by Greenwood and Johnson [bib_ref] The deformation of metals under small stresses during phase transformations, Greenwood [/bib_ref]. Through the case studies, the quantitative amounts of deformation due to the transformation plasticity versus (a) the externally applied stress, (b) the yield stress of weaker ferrite, (c) and the yield stress ratio of ferrite to austenite were presented, and the tendency of the transformation plastic deformation was confirmed with the one predicted by the internal stress model. However, the calculation results suggest that plastic flow is not confined to the weaker ferrite phase, which is unlike that suggested by the internal stress model. Rather, it occurs in both the weaker and stronger phases. The important result is that the calculated amount of transformation plastic deformation was much less than the measured transformation plastic deformation. This difference suggests the possibility that the transformation plasticity cannot be fully understood only with the micro-plastic deformation of the weaker phase and the other mechanisms such as the accelerated Coble creep [bib_ref] Diffusion-controlled transformation plasticity of steel under externally applied stress, Han [/bib_ref] or the specific variant selection [bib_ref] A model for transformation plasticity during bainite transformation of steel under external..., Han [/bib_ref] [bib_ref] A model for deformation behavior and mechanically induced martensitic transformation of metastable..., Han [/bib_ref] during the phase transformation may work simultaneously. A limit of this study is related mainly with a fundamental restriction of continuum mechanics. Although we obtain several important results that overcome the conventional continuum approaches with an assistance of PFM, the atomistic phenomena such as the accelerated creep or the variant selection cannot be considered essentially in this FE based scheme. We believe that lower scale simulations such as molecular dynamics or ab-initio calculation, which describe the atomistic motion directly, would be helpful for more fundamental understanding to the transformation plasticity. [fig] Figure 1: Two-dimensional square domain in PFM and FEM calculation. doi:10.1371/journal.pone.0035987.g001 [/fig] [fig] Figure 2: Flow of PFM-FEM hierarchical multi-scale simulation. doi:10.1371/journal.pone.0035987.g002 [/fig] [fig] Figure 3: Schematic diagram of lower edge of FE calculation domain. doi:10.1371/journal.pone.0035987.g003 [/fig] [fig] Figure 5: shows the microstructures observed by optical microscopy. As shown in the figure, it could be considered that the microstructures calculated by PFM represent a comparable result to the experimentally observed one.The calculated morphological evolution with PFM was transferred hierarchically into the FE calculation, which has an equivalent domain at the initial state. Thermo-elastic-plastic analyses were performed for the phase transformation of low carbon steel through the FE calculation. The following lists one of the calculation conditions:N Yield stress of austenite: 200 MPa N Slope of linear work hardening of austenite: 8 MPa N Yield stress of ferrite: 160 MPa N Slope of linear work hardening of ferrite: 8 MPa N Externally uniaxial applied stress: 8 MPa along horizontal direction [/fig] [fig] Figure 6: Calculated distribution of von-Mises stress during austenite-to-ferrite transformation. doi:10.1371/journal.pone.0035987.g006 [/fig] [fig] Figure 7: Calculated distribution of equivalent plastic strain during austenite-to-ferrite transformation. (arrow: front position of growing ferrite). doi:10.1371/journal.pone.0035987.g007 [/fig] [fig] Figure 9: Measured [18] and calculated transformation plastic strain according to yield stress of ferrite. (Yield stress ratio of ferrite to austenite: 0.8, Externally applied stress: 2 MPa). doi:10.1371/journal.pone.0035987.g009 [/fig] [fig] Figure 8: Comparison between measured [18] and calculated transformation plastic strains according to externally applied stress. (Yield stress of ferrite: 160 MPa, Yield stress of austenite: 200 MPa). doi:10.1371/journal.pone.0035987.g008 [/fig] [table] Table 1: Elastic modulus of steel at various temperatures[20]. [/table] [table] Table 2: Densities of austenite and ferrite phase as a function of temperature and chemical composition[21]. [/table]

Time series analysis of age related cataract hospitalizations and phacoemulsification Background: Cataract surgery remains a commonly performed elective surgical procedure in the aging and the elderly. The purpose of this study was to utilize time series methodology to determine the temporal and seasonal variations and the strength of the seasonality in age-related (senile) cataract hospitalizations and phacoemulsification surgeries.Methods: A retrospective, cross-sectional time series analysis was used to assess the presence and strength of seasonal and temporal patterns of age-related cataract hospitalizations and phacoemulsification surgeries from April 1, 1991 to March 31, 2002. Hospital admission rates for senile cataract (n = 70,281) and phacoemulsification (n = 556,431) were examined to determine monthly rates of hospitalization per 100,000 population. Time series methodology was then applied to the monthly aggregates.Results: During the study period, age-related cataract hospitalizations in Ontario have declined from approximately 40 per 100,000 to only one per 100,000. Meanwhile, the use of phacoemulsification procedures has risen dramatically. The study found evidence of biannual peaks in both procedures during the spring and autumn months, and summer and winter troughs. Statistical analysis revealed significant overall seasonal patterns for both age-related cataract hospitalizations and phacoemulsifications (p < 0.01).Conclusion:This study illustrates the decline in age-related cataract hospitalizations in Ontario resulting from the shift to outpatient phacoemulsification surgery, and demonstrates the presence of biannual peaks (a characteristic indicative of seasonality), in hospitalization and phacoemulsification during the spring and autumn throughout the study period. # Background Cataract surgery represents one of the most common forms of elective surgery worldwide in persons over age 65, and one of the leading causes of hospital admissions in the province of Ontario [bib_ref] Popular beliefs regarding the treatment of senile cataract, Temporini [/bib_ref]. In North America, cataracts afflict nearly one in six North Americans over the age of 40, approaching one in two by age 80. Cataracts are a major cause of blindness, although its etiology is still not fully understood [bib_ref] Agerelated cataract, Asbell [/bib_ref]. Previous studies have demonstrated an empirical link between age-related cataracts and personal and environmental factors such as, UV radiation, heavy alcohol consumption, and smoking [bib_ref] Agerelated cataract, Asbell [/bib_ref]. Predisposing conditions such as diabetes, glaucoma, myopia in early age, in addition to prior eye injuries, have also been linked to the development of cataracts [bib_ref] Bassett K: Cataract surgery at the Vancouver Eye Care Centre: Do patient..., Mildon [/bib_ref]. Although nonlife threatening, cataracts affect a person's ability to perform routine activities, lowering his or her overall quality of life. Cataract surgery in Ontario is paid for by the Ontario Health Insurance Program (OHIP) at no cost to the patient. In 2003-2004, the median wait time for cataract surgery in Ontario was found to be 15 weeks [bib_ref] Cataract Surgery, Bell [/bib_ref]. As a result, the efficacy of current treatments, coupled with the cost of health care, support the need for a thorough understanding of the patterns of cataract surgery in order to improve delivery of eye care services. Studies examining the prevalence of cataracts have shown that the number of cataract operations have been growing since 1980 [bib_ref] Trends in cataract surgery and postoperative endophthalmitis in Western Australia (1980-1998): the..., Semmens [/bib_ref]. Rates of cataracts are linked to age, and with a burgeoning elderly population, the demand for surgery is likely to greatly increase [bib_ref] Bassett K: Cataract surgery at the Vancouver Eye Care Centre: Do patient..., Mildon [/bib_ref]. Current literature on cataract hospitalizations is limited to analysis of single communities and has been performed only at annual levels. In order to address these limitations, this study employs a large population distributed over a vast geographical area, encompassing both urban and rural settings. Furthermore, this study examines monthly rates of cataract hospitalizations using population-based data for the province of Ontario over an 11-year period. Specifically, the objectives are: (1) to examine the overall trends in inpatient surgery versus outpatient surgery; (2) to determine if there are seasonal patterns of hospitalizations by gender; and (3) to assess the strength of these patterns. # Methods We conducted a retrospective, population-based study to assess temporal patterns in hospitalizations for agerelated cataracts, and phacoemulsification and aspiration of cataract from April 1, 1991 to March 31, 2002. Approximately 14 million residents of Ontario eligible for universal health care coverage during this time were included for analysis. The Canadian Institute for Health Information (CIHI) Discharge Abstract Database was used to obtain information on hospitalizations for senile cataract (ICD-9 code: 366.1) and phacoemulsification (ICD-9 code: 13.41) as the principal diagnoses. This database records discharges from all Ontario acute care hospitals, documenting inpatient hospital stays with a scrambled patient identifier, date of admission and discharge, up to 16 diagnoses as coded by the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM), and up to 10 procedures. Researchers have found that diagnoses within this database are coded with a high degree of accuracy, with less than 1% of the basic information on patients missing [bib_ref] Appendix I: A summary of studies on the quality of health care..., Williams [/bib_ref] [bib_ref] Relation between physician characteristics and prescribing for elderly people in New Brunswick, Davidson [/bib_ref] [bib_ref] Increased prescribing of antidepressants subsequent to betablocker therapy, Thiessen [/bib_ref]. All records with a principal discharge diagnosis of senile cataract (n = 70,281; male: 35%, female: 65%) and phacoemulsification (n = 556,431; male: 39%, female: 61%) were selected. The total number of discharges by gender were assessed for each month. Annual census data for each age group for residents of Ontario were provided by Statistics Canada. Monthly population estimates were derived through linear interpolation. Using this data, monthly hospitalization rates per 100,000 population were calculated and normalized for length of month. All transfers from one acute care hospital to another within this study group were excluded from the analysis. Analysis of the data involved several statistical techniques in order to assess the statistical significance of seasonal patterns and the consistency and magnitude of seasonal effects. Spectral analysis was conducted to test for seasonality, detecting periodicity in time series, by plotting spectral density against period. The data was detrended using moving averages of order 13 prior to conducting spectral analysis. Seasonality was tested using Fisher's κ (FK) and Bartlett Kolmogorov Smirnov (BKS). The autocorrelation function (ACF) was then used to measure the correlation between observations at different time lags. A strong correlation between the observations at 12 time lags indicates a strong seasonality of the period 12. Finally, R-squared autoregression coefficients (R 2 Autoreg ) were calculated. Autoregression uses the coefficient of determination (R 2 ) of the autoregressive regression model fitted to the data, and can be used to measure the strength of seasonality within a set of serially correlated observations that occurs with time series data [bib_ref] Autoregression as a means of assessing the strength of seasonality in a..., Moineddin [/bib_ref]. R 2 Autoreg is interpreted in the same way as the coefficient of determination in classical regression: 0-0.4 non-existent to weak seasonality, 0.4-0.7 moderate to strong seasonality, and 0.7-1.0 strong to perfect seasonality. All statistical analyses were performed using SAS 8.2 (SAS Institute Inc., Cary, NC, USA). Ethical approval for this study was obtained from the Sunnybrook & Women's College Health Sciences Centre Research Ethics Board. shows decreasing monthly hospitalization rates over the study period. The overall trend can be divided into two distinct periods. From 1991 through 1995, hospitalization rates dropped from 40 per 100,000 to 15 per 100,000. From 1995 to 2002, rates in autumn dropped from 15 per 100,000 to one per 100,000. From 1991 through 1995, rates routinely peak immediately after April and October with a relative drop in hospitalizations over the summer. In contrast, phacoemulsification procedures increased quickly from under 50 per 100,000 in 1991 to over 200 per 100,000 by the end of 2002. Gender analysis illustrates that hospitalization rates were higher among females than males throughout the study, and over time both genders fluctuate in near synchrony (data not shown). # Results The seasonal distributions of age-related cataract hospitalizations and phaocemulsification surgeries for both genders are illustrated in . Biannual peaks are present in both cataract hospitalizations and phacoemulsifications. Peak phases are more prominent during the spring and less prominent in the autumn throughout the study. Spring hospitalization phases (21 per 100,000) were slightly higher than autumn peaks (19 per 100,000). Peak phacoemulsification phases differed greatly between the spring (140 per 100,000) and autumn (158 per 100,000). However, both hospitalizations and phacoemulsifications have troughs between peak phases. Results of the time series analysis are seen in [fig_ref] Table 1: Seasonality of age-related cataract hospitalizations and phacoemulsification by gender [/fig_ref]. Fisher's κ demonstrates significant seasonality (p < 0.01) in both genders overall for hospitalizations and phacoemulsifications, as well as by gender. Similarly, BKS showed significant seasonality (p < 0.01). The R 2 Autoreg value for both genders combined was 0.64 for hospitalizations, suggesting moderately strong seasonality, and 0.78 for phacoemulsification, indicating strong seasonality. # Discussion This study revealed several important findings. Agerelated cataract hospitalizations were found to peak in the spring and autumn, and drop during the summer and winter. Further analysis revealed that overall seasonal patterns were statistically significant for both hospitalizations and phacoemulsifications. The most striking finding is the downward trend in age-related cataract hospitalizations concurrent with a significant rise in phacoemulsification over the study, a finding consistent for both genders. Similar patterns of hospitalization are seen in otolaryngological day surgery, where, for example, nasal polyp extractions (a common elective procedure) were found to peak in April-June, drop significantly during midsummer, and peak again in September-October [bib_ref] Day case surgery in otolaryngology: a 10-year analysis, Lesser [/bib_ref]. The troughs in hospitalizations coincide with the months when physicians and surgeons request vacation time, and are therefore unlikely to schedule elective surgical procedures (i.e. July, August and December). This is seen in Scotland, where the lowest elective surgical activity occurs during summer months (July through September) [bib_ref] Lazy days of summer, Pettinger [/bib_ref]. Over the course of this study, hospitalization rates for agerelated cataracts decreased from approximately 50-55 per 100,000 to roughly one to two per 100,000, while phacoemulsification grew more than 100-fold. Previous studies set precedence for such an occurrence [bib_ref] Trends in cataract surgery and postoperative endophthalmitis in Western Australia (1980-1998): the..., Semmens [/bib_ref] [bib_ref] Changing trends in cataract surgery in Singapore, Lee [/bib_ref] [bib_ref] Trends in health service delivery for cataract surgery at a large Australian..., Yi [/bib_ref] [bib_ref] Practice styles and preferences of ASCRS members -2003 survey, Leaming [/bib_ref] [bib_ref] Current Trends in Cataract and Refractive Surgery in Japan: 1999 Survey, Oshika [/bib_ref]. As an elective surgical procedure, the reduced surgical times, in addition to other benefits, such as lower costs for Age related cataract hospitalizations and phacoemulsification, aggregated by month Age related cataract hospitalizations and phacoemulsification, aggregated by month. Time plot of age related cataract hospitalizations and phacoe-mulsification, aggregated by gender Time plot of age related cataract hospitalizations and phacoemulsification, aggregated by gender. the procedure and improved surgical outcome, may serve to further motivate patients to undergo surgery [bib_ref] Trends in cataract surgery and postoperative endophthalmitis in Western Australia (1980-1998): the..., Semmens [/bib_ref]. A rise in phacoemulsification procedures is the likely explanation for the downward trend in hospitalizations . A few limitations of our study should be noted. The current study is descriptive in nature and thereby does not fully address potential explanatory factors for the hosptialization patterns. In addition, we are unable to measure temporal trends or seasonal variation in the use of other health services such as emergency department or physician visits. We have counted each procedure in the numerator, although one person could have two procedures. However, it is exceedingly unlikely for an individual to have two cataract procedures at the same time, or within the same calendar month so this does not bias the measures. The strengths of this study lie in its longitudinal base and large population size coupled with the use of a comprehensive time series analysis approach applied to gender. These results improve our understanding of gender differences as well as overall trends in age-related cataract hospitalizations in Ontario. # Conclusion This study serves to illustrate the decline of age-related cataract hospitalizations and rise of phacoemulsification surgery, plus the existence of a significant seasonal pattern in age-related cataract hospitalizations. The evidence collected supports the notion that elective surgical procedures are not seasonally driven by disease prevalence, but in fact driven by extrinsic factors, such as physician vacation periods, hospital administrations which regulate surgical schedules, and patients evaluating treatment options. Further time series analysis of patterns of elective procedures can illustrate the distribution of elective surgeries throughout a calendar year. Such data can be used by health care providers for future cost analyses and allocation of services effectively based on seasonal patterns of system utilization. [table] Table 1: Seasonality of age-related cataract hospitalizations and phacoemulsification by gender.* Tests the null hypothesis that the series is Gaussian white noise against the alternative hypothesis that the series contains an added deterministic periodic component of unspecified frequency. † Tests the null hypothesis that the series is white noise. ‡ p < 0.01. FK, Fisher's κ; BKS, Bartlett's Kolmogorov-Smirnov [/table]

Quality of care and prescription patterns among patients with diabetic kidney disease—a large-scale cohort study from Taiwanese clinics Aims.To investigate the quality of care and prescription patterns of patients with diabetic kidney disease (DKD) receiving primary care at local clinics in Taiwan. Methods. A retrospective chart review was conducted in 43 primary care clinics in Taiwan. The patients' baseline characteristics, laboratory tests, presence of complications and antidiabetic agents prescribed were analyzed. Results. 7,200 patients with type 2 diabetes mellitus were enrolled. Percentage of HbA1c, blood pressure (BP), and low density lipoprotein cholesterol (LDL-C) goals reached were 52.5% in HbA1c < 7%, 40.9% in BP < 130/80 mmHg and 79.7% in LDL-C < 2.59 mmol/L. 18.3% achieved all three ABC goals. However, patients with DKD had a lower rate of ABC goal attainment and higher rate of complications. Among DKD patients with eGFR ≥ 30 ml/min/1.73 m 2 and on monotherapy, metformin was most frequently prescribed. As for dual therapy, the most common combinations were metformin with sulfonylurea and metformin with DPP-4 inhibitors. Conclusions. Diabetes patients in Taiwan receiving primary diabetes care at local clinics had generally satisfactory management performance. However, more aggressive HbA1c, BP, and LDL-C management among DKD patients should be emphasized. Contrary to current recommendations, SGLT-2 inhibitors and GLP-1 receptor agonists as frontline therapy were under-prescribed. PC, Group SS. 2009b. Effects of structured versus usual care on renal endpoint in type 2 diabetes: the SURE study: a randomized multicenter translational study. Diabetes Care 32:977-982 . 2015. Blood pressure lowering in type 2 diabetes: a systematic review and meta-analysis. JAMA 313:603-615 AT. 2010. Effect of intensive treatment of hyperglycaemia on microvascular outcomes in type 2 diabetes: an analysis of the AC-CORD randomised trial. Raz I. 2001. Renoprotective effect of the angiotensin-receptor antagonist irbesartan in patients with nephropathy due to type 2 diabetes. The New England Journal of Medicine # Introduction Type 2 diabetes is a major challenge for the healthcare system globally and locally; the prevalence of diabetes mellitus (DM) has increased from 4.7% in the 1980s to 8.5% in 2014 and is still rising . The age-adjusted comparative prevalence of diabetes in Taiwan has grown from 9.1% in 2000 to 9.7% in 2021, and is expected to be 11.5% in 2030 . Patients with diabetes are at higher risk for microvascular and macrovascular complications, which result in heavy economic burden for the society. It was estimated that diabetes-related health expenditure was 1,314.0 USD per person . The aggregated expenditure of diabetes patients accounted for 29.7% of NHI expenditure. Diabetic kidney disease (DKD) is one of the most common complications among type 2 diabetes. In Taiwan, DKD is the most common cause for ESRD. From 2005 to 2014, the prevalence of DKD patients increased from 10.5% to 17.9%, and were expected to grow as the diabetic population increases . Statistically, Taiwan has the highest incidence and prevalence of dialysis in the world . Target attainment and routine assessment for HbA1c, blood pressure (BP) and low density lipoprotein cholesterol (LDL-C) among diabetes has long been advocated by the International Diabetes Federation (IDF) and American Diabetes Association (ADA) Guidelines . Multifactorial controls has been widely recognized in reducing complications and improving outcomes of diabetes [bib_ref] Prevention of diabetic nephropathy by tight target control in an asian population..., Tu [/bib_ref]. However, past reports showed general low HbA1c, BP, and LDL-C (ABC) target attainment. The Joint Asia Diabetes Evaluation Program found that among seven Asian countries only 5.4% diabetes achieving the ABC goals [bib_ref] Comprehensive risk assessments of diabetic patients from seven Asian countries: the Joint..., So [/bib_ref]. And The International Diabetes Management Practices Study (IDMPS) found only 3.6% diabetes from 17 countries in Europe, Asia, Latin America and Africa achieving all ABC targets . In Taiwan, survey data from the National Diabetes Health Promotion Centers reported only 30-35% achieving individual goals and 4.5% achieving all ABC goals. To date, most studies regarding diabetes care in Taiwan were conducted at either medical centers or regional hospitals. Yet, 40% of diabetes population received care at local clinics; the report of quality of care at the primary care has been limited. Hence, this report sought to conduct a retrospective, cross-sectional study investigating the characteristics, goal attainment, complications, and prescription patterns among DKD patients at the primary medical care level in Taiwan. # Materials & methods ## Study design and study population This study aimed to investigate the quality of care and prescription patterns of patients with DKD receiving primary care at local clinics in Taiwan. A retrospective chart review was conducted among 43 primary care clinics located in northern, middle, and southern Taiwan. These primary care clinics were all registered members of the Diabetes Share Care Network, a government plan featuring a teamoriented and pay-for-performance (P4P) model operated by the Health Promotional Administration of Taiwan. To be registered as a member, the clinic is required to (1) provide multidisciplinary professional care consisting of doctors, nurses, and dietitians, and (2) follow-up on DM patients' routine checkups. In its service patient pool of the P4P program, each clinic selected and enrolled 50-350 patients according to their service capacity with following methods. In the study period (July to October 2019), each DM patient who visited one of the network's clinics for diabetes management and fulfilled the inclusion criteria was enrolled in this study until each clinic achieved its maximum enrollment capacity. The inclusion criteria included: 1. Diagnosis of type 2 diabetes at least one year prior to the index day (index day: first DM clinic visit within the study period). 2. Age of 20 years and older. 3. Has made at least one DM visit within 3-6 months prior to the index day. Pregnant patients were excluded. This study complied with the ethical principles of the Declaration of Helsinki and was reviewed and approved by the Institutional Review Board of Antai Medical Care Cooperation Antai Tian-Sheng Memorial Hospital. ## Data collection At each DM checkup (including index day), the patient's baseline characteristics, physical examination, laboratory examinations, presence of complications and antidiabetic agents prescribed were recorded as claim data for P4P program. It was later analyzed retrospectively. History of diabetic microvascular complications (retinopathy, neuropathy, and nephropathy) and macrovascular complication (coronary artery disease (CAD), stroke, and peripheral arterial disease (PAD)) were obtained. Definitions of the complications are listed in [fig_ref] Table 1: Definition of complications [/fig_ref]. HbA1c level in the past 6 months; LDL-C level in the past year. # Statistical analysis The proportion of patients attaining HbA1c < 7%, BP < 130/80 mmHg, LDL-C < 2.59 mmol/L or attainment of all three ABC targets constitutes quality of care. The baseline characteristics, attainment of management goals, presence of complications and prescription patterns of antidiabetic agents were compared to patients with and without The odds ratio (OR) of macrovascular complications, attainment of individual's goals, and attainment of ABC goals in patients with and without DKD was calculated by logistic regression. Patients' complications were unadjusted and patients' goal attainment was adjusted by age, sex, body mass index (BMI) and diabetes duration. The OR of antidiabetic medications among DKD patients were adjusted based on baseline characteristics (age, sex, BMI, diabetes duration) and HbA1c. All statistical analyses were conducted with R version 4.0.0. P < 0.05 was considered statistically significant. # Results ## Baseline characteristics A total of 7,200 type 2 DM patients were enrolled in this study, baseline characteristics shown in [fig_ref] Table 2: Baseline characteristics of the study population [/fig_ref] : 49.6% male, mean age 62.56 ± 11.99 years old, diabetes duration of 8.67 ± 6.79 years and generally overweight (mean BMI 26.49 ± 4.41 kg/m 2 ). DM complications among enrolled patients included 7.1% with CAD and 2.3% with documented cerebrovascular events. In addition, 20.96% were with retinopathy, 13.53% with neuropathy, and 44% with DKD. Mean eGFR is 79.46 ± 26.33 ml/min/1.73 m 2 . Patients with DKD were generally older and has had longer diabetes duration. Their HbA1c level was significantly higher and were at higher comorbidity risk for hypertension yet not dyslipidemia. The average number in classes of antidiabetic agents prescribed and percentage of patients on insulin therapy were both higher among DKD patients, predictable among advanced diabetes patients. ## Attainment of treatment goals and complications Overall, enrolled DM patients had a mean HbA1c of 7.22 ± 1.36%, a mean LDL-C 2.09 mmol/L, a mean systolic blood pressure [fig_ref] Figure 1: The percentages of patients attaining ABC goals [/fig_ref]. Retinopathy and neuropathy occurred in 28.00% and 18.14% of these DKD patients. The prevalence of PAD, stroke and CAD were 0.38%, 3.35% and 10.01%, respectively. After adjustment for multivariable regression, DKD patients were less likely to attain HbA1c, BP, and LDL-C goals and were more likely to experience microvascular and macrovascular diseases [fig_ref] Figure 2: Adjusted odds ratio of complications among DM patients with and without DKD [/fig_ref]. ## Gender difference of dkd patients The prevalence of CKD in male and female were 43.2% and 44.5%, respectively. Among DKD patients, male patients are more likely to have CAD (11.7% vs. 8.4%, p = 0.003) and proteinuria (82.7% vs. 77.1%, p < 0.001), but overall eGFR was higher (67.4 vs. 64.8 ml/min/1.73 m 2 ). The prevalence of PAD, stroke, retinopathy, and neuropathy were similar between different genders. ## Prescription pattern After adjustment, DKD patients were more likely to receive DPP-4 inhibitors (OR [fig_ref] Figure 3: Adjusted odds ratio of individual classes of prescribed antidiabetic agents among patients... [/fig_ref]. Stratified by different treatment regimens and eGFR status, [fig_ref] Figure 4: The combinations and proportions of antidiabetic agents prescribed in different treatment regimens... [/fig_ref] shows the percentage of each class of antidiabetic agent prescribed. Among patients with eGFR ≥ 30 ml/min/1.73 m 2 on monotherapy, metformin was most frequently prescribed (73.8%) followed by sulfonylurea (6.6%) and then insulin (6.6%). Among patients with eGFR <30 ml/min/1.73 m 2 on monotherapy, nearly half were prescribed insulin followed by DPP-4 inhibitors. Patients on dual therapy with eGFR ≥ 30ml/min/1.73m 2 , metformin-based combination was the most common regimen. The most common combinations were metformin with sulfonylurea (26.4%), metformin with DPP-4 inhibitors (23.6%) and then metformin with SGLT-2 inhibitors (19.3%). 30% dual therapy patients received other combinations in the order of DPP-4 inhibitors with sulfonylurea (3.5%), metformin with thiazolidinedione (3.0%) and SGLT-2 inhibitors with sulfonylurea (3.0%). Lastly, among dual therapy patients with eGFR < 30 ml/min/1.73 m 2 , their regimen was characterized by DPP-4 inhibitor-based medications. Insulins is frequently prescribed among eGFR < 30 ml/min/1.73 m 2 patients. # Discussion To our knowledge, this was the first large-scale primary diabetes care study initiated by local clinics on diabetes goal attainment, complications, and prescription patterns of DKD patients in Taiwan. The results not only revealed real-world practice outcomes, it but also revealed the unmet needs of DKD care at the primary care level. DKD is the major cause for ESRD globally and attributes to multiple cardiovascular related morbidities. Numerous studies in the past had been conducted to determine the effectiveness of intensive intervention to improve outcomes of DM patients with DKD. First, intensive glycemic control to attain goals of HbA1c < 6.5% or fasting glucose of < 6 mmol/L was shown in multiple clinical trials to slow the decline of eGFR and progression of proteinuria [bib_ref] Intensive blood glucose control and vascular outcomes in patients with type 2..., Patel [/bib_ref] [bib_ref] Follow-up of blood-pressure lowering and glucose control in type 2 diabetes, Zoungas [/bib_ref]. Second, strict control of SBP was shown to reduce mortality and proteinuria in type 2 diabetes (Accord Study . Treatment with angiotensin converting enzyme inhibitors (ACEIs) or angiotensin II receptor blockers (ARBs) decreases progression to . As for LDL-C management, 160 type 2 DM patients with persistent microalbuminuria were enrolled in the STENO-2 Study, they were randomly assigned to receive either intensive (experiment group) or conventional therapy (control group). The experimental group had targets of attaining HbA1c < 6.5%, fasting total cholesterol < 4.53 mmol/L, fasting serum triglyceride (TG) < 1.69 mg/dL and BP < 130/80 mmHg. Compared to the conventional therapy group, the intensive therapy group had a lower risk of cardiovascular related deaths, cerebrovascular events, and progression to ESRD; confirming the cardio-and reno-protective effects of multifactorial control . In our study, DKD patients represented the population with advanced DM complications compared to DM non-DKD patients; DKD patients has had longer diabetes duration since diagnosis (10.3 years vs. 7.4 years), higher HbA1c levels (7.4% vs. 7.1%), and were prescribed more categories of antidiabetic agents than non-DKD DM patients. A significantly higher rate of micro-and macrovascular complications was observed among DKD patients. The percentage of attaining either individual HbA1c, BP, LDL-C goal or attainment of all three ABC goals were all lower among DKD patients. Although insulin resistance and beta cell exhaustion may be of challenge for glycemic management in the later years of diabetes progression, we do not find poor LDL-C control directly associated with DM etiology. Thus, stricter LDL-C management should be emphasized for better DKD management in the future. The importance of attaining ABC goals is widely advocated across different guidelines in diabetes care. In the USA, the National Health and Nutrition Examination Survey (NHANES) shows better diabetes management in attaining all three ABC goals from 2.7% in . Yet, compliance dropped to 20.7% between 2011-2014 possibly due to the adaptation of a less stringent goal set for frail DM patients [bib_ref] Poor control of risk factors for vascular disease among adults with previously..., Saydah [/bib_ref]. In Taiwan, comparing our study to a prior study conducted in 2006; we, too, find diabetes management improved over the years. In the 2006 study, 4.5% DM patients achieved all ABC goals and 30-35% attained individual goals. In our study, 18.3% DM patients achieved all ABC goals, 52.5% attained HbA1c, 40.9% attained BP and 79.7% attained their LDL-C goal. The improvements may be due to the effective primary care at local clinics. In our study, all primary care clinics were registered under the Diabetes Share Care Network in Taiwan, a P4P program shown to provide better outcomes for diabetes patients. Lee et al. points out that participants of the P4P program had a significantly lower all-cause, diabetes-, and cancer-related mortality in each analytic year between 2005 to 2014. Meanwhile, frequency of diabetes-related hospitalization and emergency department visits were also lower. Over the years, due to our National Health Insurance (NHI) service fee reimbursement policy for lab and eye examinations, utilization of nephro-and retinopathy screening rose and detection of nephro-and retinopathy among diabetes grew . In our cohort study, most DKD patients were detected at either routine examination or by urinary examination. For the past 20 years, the percentage of diabetes goal attainment kept improving. As routine diabetes health education is reimbursed with additional bonus to patients attaining the goal of HbA1c and LDL-C, diabetes health education attendance increased and dietitian-led interventions proved to provide better glycemic control under P4P [bib_ref] Effectiveness of dietitianled diabetes management program on glycemic and diet control after..., Sung [/bib_ref]. As for medication treatment, diabetes patients in our study utilized more classes of antidiabetic medications compared to the 2005-2014 cohort study, indicating a more aggressive medication intervention . Overall, the Share Care Network provided better self-management support, complication detection, and enabled goal-oriented treatments that contributed to better diabetes management and control. Our data revealed that male DKD patients are associated with higher prevalence of CAD and proteinuria, but also higher eGFR. This is compatible with the observation with other studies, which pointed out that men DM men are at higher risk for albuminuric phenotype, whereas women are at higher risk for eGFR impairment. The observations are not conclusive and require more study to confirm . Because of the cardio-and reno-protective effects, SGLT-2 inhibitors and GLP-1 RAs have been recommended to diabetes patients with ASCVD, heart failure or CKD (American Diabetes Association Professional . Several large-scale randomized controlled trials have proved SGLT-2 inhibitors improve renal composite outcomes, slow eGFR decline and decrease albuminuria . The reno-protective effect is now considered independent of glucose lowering effect. GLP-1 RAs reduce albuminuria, but the efficacy on hard renal endpoint remained equivocal . The Taiwan Society of Cardiology published a consensus on the pharmacological management of patients with type 2 diabetes and cardiovascular diseases in 2020 that diabetes patients with stage 3 CKD, an SGLT-2 inhibitor in combination with metformin is the only recommended regimen in dual therapy . However, in our survey, only 19% (152/798) of diabetes patients received the combination of an SGLT-2 inhibitor with metformin, in contrast to 23.3% (186/798) on DPP-4 inhibitor with metformin and 26% (208/798) on sulfonylurea with metformin. This finding shows a relatively low prescription rate of SGLT-2 inhibitors and GLP-1 RAs among DKD patients. Similar to our results, the multinational study on cardiovascular disease pointed out that an SGLT2i and GLP-1 RA were prescribed in only 16.0% and 10.1% of type 2 DM patients [bib_ref] CAPTURE: a multinational, cross-sectional study of cardiovascular disease prevalence in adults with..., Mosenzon [/bib_ref]. In our study, DKD patients were more frequently prescribed with insulin and DPP-4 inhibitors, but less likely to be prescribed with metformin. As DKD patients are often older, with lower eGFR, and higher HbA1c, DPP-4 shows a safer profile among aged and renal impaired patients. In addition, insulin was more frequently prescribed due to poor glycemic control and limited medication choices among diabetes patients with lower eGFR. Diabetes patients with or without DKD had similar chances of receiving sulfonylureas. In our study, maybe due to a relatively loose NHI reimbursement criteria and clinical inertia, sulfonylurea is the most prescribed second line medication in combination with metformin among DKD patients. despite its increased risk for hypoglycemia in patients with impaired renal function. Although being prohibited, it is observed that 5.7% DKD patients with eGFR < 30 ml/min/1.73 m 2 received off-label metformin. Finally, our analysis finds that among DKD patients with HbA1c > 7%, 18.1% were on monotherapy, 38.3% on dual therapy, and 58.1% on triple therapy. More aggressive glycemic control should be followed by DKD patients not attaining recommended goals. Our study has several limitations. First, in this observational, cross-sectional study, causal inferences cannot be performed due to limited temporal data. Second, as the dosage data of each oral antidiabetic agent (OAD) cannot be obtained, we used the categories of OAD as an indicator of the level of difficulty in controlling diabetes; we are aware, different stages of medication combinations and dosages may confound our outcome. Third, as the data were obtained from 43 primary clinics without a well-established electronic health record system or care management system, comorbidities and complications requiring advanced detection equipment, prevalence of complications (such as PAD), retinopathy, and neuropathy may be underestimated. Sampling bias may be minimized because the patients were selected randomly according to the visiting days in the study interval. # Conclusions In conclusion, diabetes patients in Taiwan receive a generally satisfactory diabetes primary care at local clinics. However, we suggest more aggressive HbA1c, blood pressure, and LDL-C management among DKD patients to prevent micro-and macrovascular complications. Prescription of SGLT-2 inhibitors and GLP-1 RAs are frontline therapy medications under-prescribed probably due to reimbursement criteria or patients' and physicians' choice. Further investigation and analysis are required. [fig] Figure 1: The percentages of patients attaining ABC goals. Full-size DOI: 10.7717/peerj.13636/fig-1 [/fig] [fig] Figure 2: Adjusted odds ratio of complications among DM patients with and without DKD. Full-size DOI: 10.7717/peerj.13636/fig-2 [/fig] [fig] Figure 3: Adjusted odds ratio of individual classes of prescribed antidiabetic agents among patients with and without DKD. Full-size DOI: 10.7717/peerj.13636/fig-3 [/fig] [fig] Figure 4: The combinations and proportions of antidiabetic agents prescribed in different treatment regimens and eGFR stages. (A) Monotherapy in eGFR ≥30 ml/min/1.73 m 2 , (B) monotherapy in eGFR <30 ml/min/1.73 m 2 , (C) dual therapy in eGFR ≥30 ml/min/1.73 m 2 and (D) dual therapy in eGFR <30 ml/min/1.73 m 2 . Full-size DOI: 10.7717/peerj.13636/fig-4 [/fig] [table] Table 1: Definition of complications. [/table] [table] Table 2: Baseline characteristics of the study population. DKD by Students' T -test for continuous variables and Fisher's exact test for categorical variables. [/table]

Combination targeted pulmonary hypertension therapy in the resolution of Dasatinib-associated pulmonary arterial hypertension Dasatinib is a small-molecule tyrosine kinase inhibitor used in the treatment of hematological malignancies. Pulmonary arterial hypertension (PAH) is a rare but known complication. The mainstay of treatment is cessation of Dasatinib, and while clinical improvement is rapid, complete hemodynamic resolution of pulmonary hypertension (PH) still remains exceedingly uncommon. We present a case of Dasatinib-induced PAH in a woman with chronic myeloid leukemia, who demonstrated rapid and complete clinical and hemodynamic resolution following treatment with combination pulmonary vasodilator therapy using an endothelin receptor antagonist and a phosphodiesterase-5 inhibitor. This case suggests there may be an association between the use of targeted PH medication in combination and the complete resolution of dasatinib-associated PAH, but further investigation is required. A 61-year-old woman with no past history of cardiopulmonary disease was diagnosed with chronic myeloid leukemia (CML) and initially started on Imatinib therapy. Four years later, a bone marrow biopsy showed progression of CML to precursor B-cell acute lymphoblastic leukemia (ALL) and the patient was switched to therapy with Dasatinib at a dose of 140 mg daily due to disease progression on Imatinib. At this time, the patient also received two cycles of chemotherapy with Cyclophosphamide, Vincristine, Adriamycin, and Dexamethasone. Following this, a repeat bone marrow aspirate showed complete hematologic, cytogenetic, and molecular remission and the patient was continued on Dasatinib therapy. However, 26 months later, the patient developed a persistent dry cough and severe dyspnea on exertion. Initial laboratory testing showed a white blood cell count of 3700, with 46% neutrophils and 45% lymphocytes. Blood cultures were negative for infection, and the patient did not produce adequate sputum for culture. A chest X-ray showed bilateral pleural effusions with cardiomegaly, suggestive of right ventricular failure. A CT scan was performed, which noted an enlarged central pulmonary artery and bilateral pleural effusions; no other abnormalities concerning the lung parenchyma, pulmonary veins, or interstitium were reported. No pulmonary embolism was seen. An echocardiogram confirmed the presence of a dilated and hypokinetic right ventricle and a dilated right atrium. Other potential causes of pulmonary hypertension (PH) were excluded by a negative V/Q scan and a review of past records showing normal echocardiographic testing one year prior. A right heart catheterization (RHC) was performed with a mean pulmonary artery pressure (mPAP) of 54 mmHg, pulmonary capillary wedge pressure (PCWP) of 11 mmHg, and pulmonary vascular resistance (PVR) of 18.22 Woods Units, yielding a diagnosis of pulmonary arterial hypertension (PAH). The right atrial (RA) pressure was 19 mmHg, and the cardiac output (CO) by the Fick method was 2.36 L/min (Cardiac index of 1.46 L/min/m 2 ). Vasoreactivity testing was not performed. In clinic followup one month after diagnosis, a D-dimer test was done and found to be elevated at 1.73 mg/L (reference ¼ 0-0.5 mg/L); however, a polymerase chain reaction (PCR) test for the BCR-ABL gene was < 0.001% at that time, consistent with her levels while in remission. A complete autoantibody screen was also done and found to be negative (antinuclear antibody, anti-double strand DNA, anti-Smith, ANCA, ribonucleoprotein, anti-histone, SCL-70, chromatin, centromere, SSA/SSB). Her BNP was noted to be 560 pg/mL (reference ¼ 0-100 pg/mL), her 6-minute walk test (6MWT) was 250 m using 2 L oxygen via nasal cannula with a nadir pulse oximetry saturation of 96%, and her functional class was WHO-FC III. Dasatinib was suspected to be the precipitating factor for the patient's new PAH given the temporal association with the patient's symptoms and exclusion of other etiologies, and was promptly replaced with Nilotinib (a related second-generation tyrosine kinase inhibitor with a lower rate of drug-induced PAH compared with Dasatinib) to continue therapy for CML. Given the severity of the patient's symptoms and RHC data, the patient was also started on combination therapy with a phosphodiesterase-5 inhibitor, Tadalafil, at 20 mg daily and an endothelin receptor antagonist, Ambrisentan, at 5 mg daily. The patient was also started on furosemide at 10 mg daily and was not started on systemic anticoagulation. The Tadalafil was uptitrated over a period of four weeks to 40 mg daily, followed by an up-titration of Ambrisentan to 10 mg daily over the following four weeks. The patient experienced marked symptomatic improvement, and repeat chest X-rayand echocardiogram four months after beginning therapy showed complete resolution of the PH with a normal right ventricular size and function (Figs 5 and 6). A repeat RHC performed at the same time showed full resolution of PAH, with mPAP of 18 mmHg, PCWP of 9 mmHg, and PVR of 1.96 Woods Units. The RA pressure was 5 mmHg and the CO by the Fick method was 4.59 L/min (cardiac index of 2.51 L/min/m 2 ). Her BNP was measured in clinic at this time and had decreased to 18 pg/mL, her 6MWT had improved to 425 m on room air with a nadir pulse oximetry saturation of 96%, and her functional class was WHO-FC I. Given the marked improvement both symptomatically and hemodynamically, Ambrisentan and furosemide were discontinued and Tadalafil was weaned sequentially over the following four months. Despite the cessation of her pulmonary vasodilator therapy, the patient did not have a return of her dyspnea or exercise intolerance. She continues to remain symptom-free after five months without targeted PH therapy. # Discussion Dasatinib is a potent oral small-molecule tyrosine kinase inhibitor of the BCR-ABL oncogene used in the treatment of diseases such as CML and ALL. 1,2 PAH has been reported as a rare complication of Dasatinib therapy that can occur at any time during the treatment course, ranging from months to over one year after initiation.Drug-induced PH secondary to Dasatinib is rare, and the exact mechanism is unknown but thought to be related to the platelet-derived growth factor pathway (PDGF) and the Src family of kinases.While its main therapeutic effect in CML patients is high-affinity inhibition of the BCR-Abl oncoprotein, Dasatinib is a non-specific tyrosine kinase inhibitor and also inhibits a number of different kinases, including Src and the receptor for PDGF.In animal models of PH, dysfunction of the PDGF pathway has been involved in the development of PH, and the Src family of kinases is involved in smooth muscle cell proliferation and vasoconstriction of the pulmonary vasculature. More recent work has shown a direct link between Dasatinib exposure and pulmonary endothelial cell damage, dysfunction, and ultimately apoptosis in both a rat PAH model and cultured human pulmonary endothelial cells. Interestingly, treatment with Imatinib, a related tyrosine kinase inhibitor, did not have the same effect.Indeed, evidence suggests related inhibitors such as Nilotinib and Imatinib do not have the same frequency of drug-induced pulmonary vascular dysfunction as Dasatinib, and the pulmonary vascular toxicity for Dasatinib is molecule-specific.Some related tyrosine kinase inhibitors such as Imatinib have also been used as treatment for PH, highlighting how much is still unclear regarding the exact mechanism of druginduced PH from Dasatinib.While drug-induced PAH secondary to Dasatinib is accompanied by serious clinical complications, evidence suggests it may be reversible to some degree.To date, the cornerstone of treatment is cessation of the offending drug in question; targeted PH medication is frequently used to address the underlying hemodynamic and functional abnormalities present. While this leads to rapid improvement in both symptoms of right heart failure and hemodynamic measures of PH severity, complete resolution of PH is fairly uncommon. In an analysis of a French Registry of nine Dasatinib-induced PAH cases, the initial hemodynamic abnormalities seen on diagnostic RHC (median mPAP of 46 mmHg, median PVR of 5.9 Woods Units) improved significantly with cessation of Dasatinib. However, none of the patients had complete resolution of PH at 15-month follow-up. This was despite two patients receiving treatment with endothelin receptor antagonists and one patient receiving therapy with calcium channel blockers.Similarly, in an analysis of the Bristol-Meyers Squibb database of adverse drug events due to Dasatinib, 41 cases of PAH confirmed by RHC due to Dasatinib were identified, of which there was symptomatic or echocardiographic resolution of PAH in 21 cases. Only five patients had a subsequent RHC following drug cessation, of which none had complete resolution of PAH as defined by mPAP < 25 mmHg at rest. While there was no relationship between the duration of Dasatinib exposure and improvement of PAH (ranging from one week to 75 months of Dasatinib therapy), the majority of improved cases (13 of 21, 62%) involved treatment with pulmonary vasodilator medications such as endothelin receptor antagonists (ERA) or phosphodiesterase-5 inhibitors (PDE-5i). Of note, two of the improved cases used combination targeted PH therapy with a PDE-5i and an ERA simultaneously, showing rapid clinical resolution in less than one month. Unfortunately, following clinical resolution neither case had a subsequent RHC to demonstrate hemodynamic resolution of PAH.A PubMed/Medline search revealed several case reports of Dasatinib-induced PAH, some of which resolved with cessation of the drug alone,and others that resolved with the addition of PAH-specific medications.These reports differ in both the duration of Dasatinib treatment prior to the development of PAH, the type of PAH-specific therapy used, as well as the duration of time required for resolution of PAH; however, none of these published cases involved the use of combination therapy. Notably only two cases confirmed complete hemodynamic resolution of PAH with RHC, and both involved therapy with a PDE-5i (Sildenafil). The time to hemodynamic resolution was seven months and six months, respectively.Currently there are no data on the use of PAH-specific medication to increase the extent and/or rapidity of Dasatinib-induced PAH resolution, either alone or in combination. As noted above, our patient was not anticoagulated, and currently there is conflicting evidence regarding systemic anticoagulation in PAH. The publication of the Registry to Evaluate Early and Long-term PAH disease Management (REVEAL) showed there was no survival benefit to systemic anticoagulation in idiopathic PAH (IPAH) patients, and a trend towards increased mortality in patients with systemic sclerosis-associated PAH (SSc-PAH). This is contrary to the Comparative, Prospective Registry of Newly Initiated Therapies for Pulmonary Hypertension (COMPERA) registry, which showed a statistically significant survival benefit at three years in IPAH patients on systemic anticoagulation. Given the uncertainty regarding the survival benefit in IPAH patients, as well as the concern for increased mortality in SSc-PAH exposed to systemic anticoagulation, the decision is currently left to the provider and patient on an individualized basis and the optimal strategy in drug-induced PAH patients is unknown. The use of combination targeted PH therapy in PAH patients is supported by the AMBITION trial, where initial combination therapy resulted in a 50% reduction in the rate of clinical failure when compared with monotherapy in previously untreated patients with PAH.As noted above, complete resolution of Dasatinib-induced PH is exceedingly uncommon, but registries and case reports suggest there is an association between the use of targeted PH medications and complete resolution. We believe this case suggests a potential association between the use of combination targeted PH therapy and the complete resolution of Dasatinib-associated PAH. It is unclear if combination therapy may be useful in increasing the chances of complete resolution after drug cessation, and additional research is warranted to investigate any potential association that may exist. Therapy with Dasatinib, a small-molecule tyrosine kinase inhibitor used in the treatment of CML and ALL, is known to cause drug-induced PAH. The mainstay of treatment is cessation of Dasatinib, which can lead to improvement in the severity of PAH but does not typically result in complete resolution. As seen in this case, there may be an association between the use of targeted PH medications in combination and the complete resolution of PAH, but further investigation is required.

Palmitoylethanolamide Reduces Proinflammatory Markers in Unvaccinated Adults Recently Diagnosed with COVID-19: A Randomized Controlled Trial Background: Inflammation is at the core of many chronic conditions and exacerbates infectious conditions, including the severity of coronavirus disease 2019 infections.Objectives:This study aimed to examine the effects of a novel food supplement, palmitoylethanolamide (PEA), specifically Levagen+, as compared with a placebo on proinflammatory biomarkers in adults recently diagnosed with COVID-19 who were unvaccinated and nonhospitalized.Methods:This study was a double-blind randomized placebo-controlled trial conducted October 2020-March 2021 (clinicaltrials.gov: NCT04912921). Participants aged 19-53 y were unvaccinated and recently infected with COVID-19 as indicated by a positive test result per RT-PCR or antigen test, and they reported to the test site following diagnosis as allowed by the CDC's return-to-work policy. Participants were stratified by age, sex, and BMI and randomly assigned by coin toss to receive 600 mg Levagen+ twice daily (LEV) or placebo tablets twice daily (CON) for 4 wk. At baseline and week 4, participants completed health histories, 24-h dietary recalls, anthropometrics, and nonfasting blood sampling.The primary outcomes were the 4-wk change between groups for IL-6, C-reactive protein, ferritin, intercellular adhesion molecule 1, soluble P-selectin (sP-selectin), and neutrophil/lymphocyte ratio. Multiple linear regression models were utilized to assess treatment effects on outcomes, adjusting for covariates.Results: A total of 60 participants completed the study (LEV: n = 30; CON: n = 30). After 4 wk of supplementation, sP-selectin (β = −11.5; 95% CI: −19.8, −3.15; P = 0.0078), IL-1β (β = −22.9; 95% CI: −42.4, −3.40; P = 0.0222), and IL-2 (β = −1.73; 95% CI: −3.45, −0.065; P = 0.0492) concentrations were significantly reduced in the LEV group compared with the CON group. Conclusions: Inflammatory mechanisms are crucial to optimal resolution of infectious conditions, yet unchecked secretion of inflammatory mediators can promote the dysregulated immune response implicated in COVID-19 complications. Overall, PEA supplementation produced anti-inflammatory effects in individuals recently diagnosed with COVID-19 who were nonhospitalized. J Nutr 2022;152:2218-2226. # Introduction Inflammation is at the core of many chronic and infectious conditions, including the severity of coronavirus disease 2019 (COVID- [bib_ref] Emerging roles of PPARs in inflammation and immunity, Daynes [/bib_ref] infection [bib_ref] Could natural products modulate early inflammatory responses, preventing acute respiratory distress syndrome..., Amaral-Machado [/bib_ref]. The SARS-CoV-2 virus, responsible for COVID-19, is classified in the β genus of the Coronaviridae family [bib_ref] Molecular immune pathogenesis and diagnosis of COVID-19, Li [/bib_ref]. This family of enveloped viruses has a positive-strand RNA genome capable of encoding critical structural proteins, including the spike surface glycoprotein and the membrane, envelope, and nucleocapsid proteins, which play a pivotal role in the pathogenic outcomes of the virus [bib_ref] COVID-19: pathogenesis, cytokine storm and therapeutic potential of interferons, Nilea [/bib_ref]. Upon exposure to SARS-CoV-2, largely through transmission by respiratory droplets, the viral spike protein interacts with angiotensin-converting enzyme 2 receptors on host cells, triggering subsequent changes in spike protein conformation and endosomal formation, events that facilitate viral entry into host cells [bib_ref] Molecular mechanism of interaction between SARS-CoV-2 and host cells and interventional therapy, Zhang [/bib_ref]. Once inside a cell, the virus releases its RNA and replicates, and high numbers of formed virus leave the cell to invade other cells, furthering infection [bib_ref] Mechanisms of SARS-CoV-2 entry into cells, Jackson [/bib_ref]. Derangements in angiotensin-converting enzyme 2 activity caused by viral attachment in individuals infected with SARS-CoV-2 are supported by observed increases in plasma angiotensin II [bib_ref] Clinical and biochemical indexes from 2019-nCoV infected patients linked to viral loads..., Liu [/bib_ref]. Overactive immune responses may be tied to this loss in angiotensin-converting enzyme 2 activity and therefore inhibition of its protective mechanisms related to inflammation, such as its conversion of angiotensin II to angiotensin 1-7 and its inhibitory effects on macrophage cytokine expression [bib_ref] Protective role of ACE2 and its downregulation in SARS-CoV-2 infection leading to..., Banu [/bib_ref]. Furthermore, signaling mechanisms associated with the innate immune response against viral infection lead to the activation of transcription factors, such as NF-κB, which induce production of a diverse array of proinflammatory cytokines and chemokines [bib_ref] Overview of COVID-19 inflammatory pathogenesis from the therapeutic perspective, Lee [/bib_ref]. COVID-19 has also been reported to induce lymphopenia among dendritic cells, T cells, and NK cells and to upregulate other immune cells, such as neutrophils and monocytes [bib_ref] Innate immune signaling and proteolytic pathways in the resolution or exacerbation of..., Sallenave [/bib_ref] [bib_ref] Clinical features of patients infected with 2019 novel coronavirus in Wuhan, China, Huang [/bib_ref]. This overproduction of cytokines leads to the enhanced expression of adhesion molecules on the surface of endothelial cells, further expanding the inflammatory sequelae. For example, elevated circulating P-selectin has been demonstrated to predict thrombosis in patients with COVID-19 [bib_ref] Plasma P-selectin is an early marker of thromboembolism in COVID-19, Fenyves [/bib_ref]. Excessive cytokine production, termed "cytokine storm," can potentiate negative downstream consequences, including a variety of pulmonary and extrapulmonary complications, which are associated with severity of COVID-19 outcomes [bib_ref] Extrapulmonary manifestations of COVID-19, Gupta [/bib_ref]. Thus, investigations have utilized proinflammatory biomarkers to characterize disease severity and risk for death. Consequently, strategies to control inflammation are considered a key approach for slowing the progression of disease and tissue pathology. Whereas the heightened release of inflammatory mediators is typically observed in the acute stage of COVID-19 [bib_ref] The first 12 months of COVID-19: a timeline of immunological insights, Carvalho [/bib_ref] , some patients develop symptoms lasting beyond 12 wk, often occurring in those with mild-moderate infection severity. Recently, in a study of patients with mild-moderate SARS-CoV-2 infection, those experiencing persistent symptoms had increased immune cell activation and exhibited elevations in proinflammatory mediators 8 mo after infection [bib_ref] Immunological dysfunction persists for 8 months following initial mild-to-moderate SARS-CoV-2 infection, Phetsouphanh [/bib_ref]. Hence, approaches to reduce inflammatory mediators in those with even mild-moderate symptoms could be of key importance to mitigating the acute and long-term impacts of COVID-19. Palmitoylethanolamide (PEA)-a fatty acid amide first isolated from lipid fractions of egg yolks, peanuts, and soybeanshas been documented as being beneficial for reducing inflammation [bib_ref] Palmitoylethanolamide: a natural body-own anti-inflammatory agent, effective and safe against influenza and..., Hesselink [/bib_ref]. PEA is also endogenously produced in human tissues and functions as a lipid mediator targeting ion channels, nuclear hormone receptors, and G protein-coupled receptors, with wide-reaching effects on metabolism [bib_ref] The pharmacology of palmitoylethanolamide and first data on the therapeutic efficacy of..., Petrosino [/bib_ref]. One of the direct targets for PEA has been identified as nuclear receptor peroxisome proliferator-activated receptor α (PPAR-α), which is expressed in several cells, including immune cells, and supports PEA's ability for use in modulating inflammatory responses [bib_ref] The nuclear receptor peroxisome proliferator-activated receptor-α mediates the anti-inflammatory actions of palmitoylethanolamide, Verme [/bib_ref] [bib_ref] Emerging roles of PPARs in inflammation and immunity, Daynes [/bib_ref]. PEA's interaction with PPAR-α has been shown to mediate its anti-inflammatory effects, and PEA's binding to PPAR-α on immune cells leads to reductions in proinflammatory and pain signals [bib_ref] The nuclear receptor peroxisome proliferator-activated receptor-α mediates the anti-inflammatory actions of palmitoylethanolamide, Lo Verme [/bib_ref]. Furthermore, PEA exerts its action on G protein-coupled receptor 55, a cannabinoid-like receptor, and indirectly influences activity of cannabinoid receptors 1 This research was funded by Gencor Pacific Ltd. and 2 via inhibition of cannabinoid degradation, termed the "entourage effect" [bib_ref] The basal pharmacology of palmitoylethanolamide, Rankin [/bib_ref]. Importantly, research has demonstrated that PEA inhibits the migration and degranulation of mast cells by direct and indirect mechanisms [bib_ref] The pharmacology of palmitoylethanolamide and first data on the therapeutic efficacy of..., Petrosino [/bib_ref]. Given the emerging evidence that COVID-19 infection activates mast cells via tolllike receptor 4, the ability of PEA to lessen immune cellinduced inflammation via toll-like receptor 4-dependent PPARα activation may have important physiologic relevance [bib_ref] Sodium chromo-glycate and palmitoylethanolamide: a possible strategy to treat mast cell-induced lung..., Gigante [/bib_ref]. PEA has been utilized in several trials to treat influenza and the common cold and was shown to be effective in treating upper respiratory infections [bib_ref] Palmitoylethanolamide: a natural body-own anti-inflammatory agent, effective and safe against influenza and..., Hesselink [/bib_ref]. Given the parallels between COVID-19 and the mechanisms by which PEA has been successful as an immunomodulator in conditions such as influenza, it may be a viable adjunctive treatment for COVID-19. Notably, PEA is endogenously produced and has not been associated with adverse side effects. Although PEA has been put forth as a possible adjunctive treatment strategy for COVID-19, there is a dearth of evidence on its use in patients with COVID-19 [bib_ref] Ultramicronized palmitoylethanolamide (um-PEA): a new possible adjuvant treatment in COVID-19 patients, Noce [/bib_ref]. In this context, the aim of this research is to expand this literature related to PEA in a novel manner and to evaluate the efficacy of 4-wk PEA supplementation to reduce the mediators of inflammation in adults who recently tested positive for COVID-19 but were not hospitalized, a group likely to experience elevations in inflammatory mediators following infection. We hypothesized that 4 wk of PEA supplementation would result in a favorable modulation in inflammatory mediators as compared with placebo supplementation. Given that inflammatory mediators would be expected to return toward homeostatic levels following infection, changes in inflammatory mediators are expected to improve to a larger degree in the treatment group as compared with the placebo group, all else being equal. # Methods ## Participants Healthy adults between the ages of 18-65 y who recently tested positive for COVID-19 and were not hospitalized for their illness were recruited via online advertisements, local news outlets, e-mail lists, and word of mouth to participate in this study. Inclusion in the study required a recent positive COVID-19 test result per PCR in asymptomatic/symptomatic individuals, although positive antigen test results were also accepted upon symptomatic infection consistent with COVID-19 per the symptoms outlined by the CDC. The exclusion criteria included the following: any unstable or serious illness; serious mood disorders; neurologic disorders, such as multiple sclerosis or cognitive damage; active smoking and/or nicotine or drug abuse; active regular marijuana or other cannabinoid use; other street/recreational drug use; chronic past and/or current alcohol use (>14 alcoholic drinks/wk); allergies to any of the ingredients in the active or placebo formula, including peanuts and eggs; pregnancy or lactation; medical prescription of drugs that affect immune and/or inflammatory responses; malignancy treatment in the last 5 y; chronic use of steroids; and a BMI > 40 kg/m 2 . Eligible respondents were sent an electronic consent form and scheduled to visit the test site based on CDC return-to-work guidance. All CDC guidance for workplace safety was followed at the test site, including hand hygiene practices, environmental infection control, and personal protective equipment. Study recruitment was conducted October 2020-March 2021, and all participants were unvaccinated for COVID-19. The Arizona State University Institutional Review Board approved this study (No. 00,012,406), and all participants provided written consent. The study is registered at clinicaltrials.gov as NCT04912921. ## Study protocol The study followed a placebo-controlled randomized parallel-arm study design. Participants were stratified by age, sex, and BMI and randomly assigned by coin toss to receive the active ingredient or placebo treatment for 4 wk. Randomization was performed by a research team member not involved in data collection or analysis. Participants were screened for eligibility using an online questionnaire. Those who met the inclusion criteria were interviewed by phone for a secondary screening to confirm eligibility and to reduce physical contact with investigators during infectious periods. Eligible individuals were scheduled to visit the test site (Arizona Biomedical Collaborative Laboratory Building) based on CDC return-to-work guidance: 1) at least 10 d since symptoms first appeared or from positive test result in asymptomatic individuals, 2) at least 24 h since last fever without the use of fever-reducing medications, and 3) improvement of symptoms (e.g., cough, shortness of breath). Participants attended two in-person study visits and completed a health history questionnaire and 24-h dietary recall. Anthropometrics were collected during study visits, including height, weight, and BMI. At study visits, venous blood samples were collected by a certified staff phlebotomist into appropriate collection tubes for subsequent processing and analyses of inflammatory mediators. At study baseline, participants were provided with the assigned supplements; a calendar to track supplement intake; and directions on supplement ingestion, symptom reporting, and over-the-counter medication tracking. The final posttrial visit was scheduled after 4 wk, and the baseline assessments and blood draw protocol were repeated. Nutrient intakes obtained from 24-h dietary recalls at baseline and posttrial were assessed using Food Processor Nutrient Analysis Software (ESHA Research). ## Supplementation Participants in the active treatment group ingested 600 mg Levagen+ twice daily (LEV) for 4 wk. A 2013 review [bib_ref] Palmitoylethanolamide: a natural body-own anti-inflammatory agent, effective and safe against influenza and..., Hesselink [/bib_ref] reported on the role of PEA as an anti-inflammatory and therapeutic agent for influenza and the common cold. In six clinical trials in nearly 4000 volunteers, PEA demonstrated effectiveness and safety for the treatment of these indications. The effective dose range has consistently been 10-30 mg PEA/kg bodyweight. Levagen + also contains LipiSperse: a novel delivery system designed to increase the dispersion of lipophilic agents in aqueous environments and indicated to enhance PEA bioavailability by nearly 75% [bib_ref] Increased absorption of palmitoylethanolamide using a novel dispersion technology system (LipiSperse) study..., Briskey [/bib_ref]. As such, each pill contains 350 mg Levagen + with the label claim "Not less than 300 mg palmitoylethanolamide." Ten percent of the 350 mg is LipiSperse; therefore, each pill contains 315 mg PEA. This is per USP regulations for finished products such that Levagen + contains 270-330 mg PEA (i.e., 90%-110% of the label's claimed ingredients). Those in the placebo group (CON) ingested placebo capsules (maltodextrin) twice daily for 4 wk. LEV and CON supplements were identical in appearance, and supplement blinding was completed by an investigator not involved in data collection or analyses. For adherence checking, participants were asked to record pill ingestion daily and to return unemptied pill bottles. Participants recorded any physical symptoms and/or over-the-counter medication use daily during the trial. Participants also received weekly prompts via e-mail from investigators, which contained reminders related to the study protocol. ## Blood processing and analyses During baseline and posttrial visits, participant blood samples were collected from the antecubital vein by standard phlebotomy techniques, with blood collection sets in 4-mL dipotassium EDTA tubes and 8.5-mL serum separator tubes. After centrifugation of serum separator tubes, serum samples were treated with 1% Triton X-100 to inactivate SARS-CoV-2 virus (25). Whole blood collected in EDTA tubes was immediately lysed in Trizol RNA extraction buffer (TRIzol LS; Thermo Fisher Scientific) and prepared for qRT-PCR for gene expression analysis. Whole blood collected in EDTA tubes was transported and analyzed for complete blood count with differentials, specifically utilized for calculation of neutrophil/lymphocyte ratios by Sonora Quest Laboratories. Serum samples were analyzed for high-sensitivity Creactive protein and ferritin by Sonora Quest Laboratories. Samples were stored at −80 - C until time of analyses. Serum cytokines and chemokines were analyzed by human focused 15-plex discovery assay (Eve Technologies Corporation): IL-6, IL-1β, IL-1Ra, IL-2, IL-4, IL-5, IL-8, IL-10, IL-12p40, IL-12p70, IL-13, monocyte chemoattractant protein 1, IFN-γ , granulocyte-macrophage colony-stimulating factor, and TNF-α. Serum soluble P-selectin (sP-selectin) and intercellular adhesion molecule 1 were analyzed via commercially available ELISA methods (Invitrogen; Thermo Fisher Scientific). ## Rna isolation and qrt-pcr The expression of NF-κB, IL-6, and CD177 was quantified from whole blood samples. Trizol-treated whole blood was sonicated for 1 min and then centrifuged to remove cell debris. RNA was purified using the Direct-zol RNA MicroPrep Kit (R2062; Zymo Research), and cDNA was synthesized from total RNA using the cDNA Reverse Transcriptase Kit (4,374,966; Thermo Fisher Scientific) according to the manufacturers' instructions. Amplification of NF-κB, IL-6, CD177, and GAPDH genes was accomplished using SYBR Green Master Mix (A46012; Thermo Fisher Scientific) according to the following primers: [formula] NF-κB2 (F- GAAGACAAGGAAGAGGTGCAG, R-CACCTCCTCCAGCTCCT), IL-6 (F-TGAGAAAGGAGACATGTAACAAGAG, R-CAAGTCTCC TCATTGAATCCAGA), CD177 (F-GAGGAGGCATCTTCTCCAATC, R-CCCGATGACAGGTCAGAAA), [/formula] and GAPDH (F-GAA GGTGAAGGTCGGAGTC, R-GAAGATGGTGATGGGATTTC). The RNA levels of each gene were normalized with GAPDH and expressed as Ct. # Statistical analysis The sample size (N = 60) was informed by a meta-analysis reporting 11 randomized controlled trials that investigated the impact of coenzyme Q10 on reducing IL-6 in chronic inflammatory diseases. In these trials, the sample size averaged 44 (range: 26-60), and power analyses based on several of these trials indicated that a sample size of 30-63 was adequate for 80% power at an alpha of 0.05 to detect a significant change in IL-6 (26). We examined histograms of input variables and performed log transformations as needed for analyses to reduce skewness. The primary outcomes included the change in circulating levels of IL-6, Creactive protein, ferritin, intercellular adhesion molecule 1, sP-selectin, and neutrophil/lymphocyte ratio between groups. Secondary analyses were conducted on the changes in a panel of serum cytokines and chemokines and expression of NF-κB, IL-6, and CD177 in whole blood between groups. A series of linear regression models was built to test if the postchange as compared with the prechange in each inflammatory marker was associated with the treatment. In these models, the response variable was the postchange as compared with the prechange; the independent variable was the treatment group; and the covariates were age, sex, BMI, and existing conditions (binary). An overall P value <0.05 indicated statistical significance. The Benjamini-Hochberg procedure [bib_ref] Controlling the false discovery rate: a practical and powerful approach to multiple..., Benjamini [/bib_ref] was employed to account for multiple comparisons in analyses of primary outcomes with a false discovery rate of 10%. Data were analyzed using open source R software. # Results From the 323 study respondents, 61 participants were enrolled in the trial. One LEV participant withdrew due to scheduling conflicts, and 60 participants (n = 30/group) completed the trial [fig_ref] FIGURE 1: Study participant flow [/fig_ref]. Average compliance to supplement ingestion was high such that 87% and 91% of the tablets were taken during the study in the CON and LEV groups, respectively. Baseline characteristics of participants are shown in [fig_ref] TABLE 1: Baseline characteristics of study participants [/fig_ref]. Baseline characteristics of participants were comparable between study arms. Additionally, of the 60 participants who completed the trial, 2 in the CON group and 1 in the LEV group reported asymptomatic infection. Types of COVID-19 symptoms reported by participants were also comparable between groups [fig_ref] FIGURE 1: Study participant flow [/fig_ref]. Based on the 24-h dietary records obtained throughout the trial, no significant baseline or posttrial differences were seen between the LEV and CON groups in dietary intakes of energy, carbohydrates, proteins, fats, saturated fatty acids, sugar, total dietary fiber, vitamin C, vitamin D, calcium, iron, sodium, and zinc (Supplemental [fig_ref] TABLE 1: Baseline characteristics of study participants [/fig_ref]. There were no significant differences between groups in the reductions of IL-6, C-reactive protein, ferritin, intercellular adhesion molecule 1, or neutrophil/lymphocyte ratio [fig_ref] TABLE 2: Status of primary outcome serum inflammatory biomarkers at baseline and after the... [/fig_ref]. However, the reduction in sP-selectin from baseline was significantly associated with LEV treatment compared with CON, which saw increases in these markers after adjustment for covariates and remained significant with a Benjamini-Hochberg false discovery rate ≤ 0.10 (β = −11.5; 95% CI: −19.8, −3.15; P = 0.0078). sP-selectin fell 8% in the experimental group (mean ± SD: -6.8 ± 17.3) and rose 5% in the control group (3.4 ± 15.3). The changes in circulating cytokines and chemokines and whole blood expression of inflammatory markers after 4 wk of supplementation are presented in . Our analyses did not reveal significant associations between the treatment group and the change from baseline in most serum inflammatory factors or whole blood expression of NF-κB, IL-6, or CD177. However, significant reductions from baseline in IL-1β (β = −22.9; 95% CI: −42.4, −3.40; P = 0.0222) and IL-2 (β = −1.73; 95% CI: −3.45, -0.065; P = 0.0492) were associated with LEV treatment compared with CON . # Discussion These data suggest that daily PEA supplementation reduced sP-selectin concentrations in adults recently diagnosed with COVID-19. Specifically, sP-selectin fell 8% in the LEV experimental group and rose 5% in the CON group, a difference that was statistically significant after false discovery rate adjustment using the Benjamini-Hochberg procedure (P = 0.0078). Furthermore, LEV treatment was associated with significant reductions from baseline in IL-2 (P = 0.0492) and IL-1β (P = 0.0222) compared with CON. P-selectin, a key thromboinflammatory marker, is stored in the α-granules of platelets and Weibel-Palade bodies of endothelial cells and is translocated to the surface of the cell upon stimulation [bib_ref] P-selectin primes leukocyte integrin activation during inflammation, Wang [/bib_ref]. This membrane glycoprotein is expressed on activated platelets and endothelial cells and contributes to the localized inflammation that ultimately eliminates pathogens and clears cell debris. P-selectin is involved in the initial attachment of platelets and endothelial cells to leukocytes and the rolling of immune cells to injured regions [bib_ref] Leukocyte rolling and extravasation are severely compromised in P selectin-deficient mice, Mayadas [/bib_ref]. In addition, P-selectin on activated platelets has been shown to stimulate exposure of tissue factor on monocytes, which may promote intravascular hemostasis and thrombosis [bib_ref] Platelet P-selectin triggers rapid surface exposure of tissue factor in monocytes, Ivanov [/bib_ref]. Although P-selectin is essential for optimal immune responses and the rapid elimination of infectious agents and foreign particles, under conditions of unresolved disease or chronic insult, P-selectin is related to the propagation and amplification of the inflammatory response. Importantly, P-selectin has also been linked to COVID-19 complications. Authors have reported 6%-12% declines in sP-selectin following pharmaceutical interventions in patient populations. Nomura et al. [bib_ref] Effects of efonidipine on platelet and monocyte activation markers in hypertensive patients..., Nomura [/bib_ref] noted a 6% reduction in sP-selectin following administration of efonidipine to patients with hypertension and diabetes and suggested that this may prevent the development of cardiovascular complications caused by cell adhesion molecules or activated platelets and monocytes. Riondino et al. [bib_ref] Platelet hyperactivity in hypertensive older patients is controlled by lowering blood pressure, Riondino [/bib_ref] noted a 12% reduction in sP-selectin following drug-induced normalization of blood pressure in older adults who were hypertensive. Thus, the 8% reduction in sP-selectin noted herein is within the ranges cited in other intervention trials, suggesting its clinical relevance. Procoagulant responses and thrombosis are augmented in patients with COVID-19 [bib_ref] High risk of thrombosis in patients with severe SARS-CoV-2 infection: a multicenter..., Helms [/bib_ref]. These derangements in hemostasis persist in moderate and severe COVID-19 [bib_ref] Altered platelet and coagulation function in moderate-to-severe COVID-19, Litvinov [/bib_ref] and are associated with the signature inflammatory imbalances of the cytokine storm and severity of disease outcomes [bib_ref] Clinical features of patients infected with 2019 novel coronavirus in Wuhan, China, Huang [/bib_ref] [bib_ref] Endothelial injury in COVID-19 and acute infections: putting the pieces of the..., Levy [/bib_ref]. P-selectin expression on the surface of platelets was recently shown to be significantly elevated in hospitalized patients with COVID-19, regardless of severity, compared with healthy donors [bib_ref] Platelet gene expression and function in patients with COVID-19, Manne [/bib_ref]. Furthermore, P-selectin is a key receptor for formation of platelet-leukocyte aggregation, which is increased in those infected with COVID-19 [bib_ref] Platelet gene expression and function in patients with COVID-19, Manne [/bib_ref] , and a marker of in vivo platelet activation during viral illness [bib_ref] Measuring and interpreting platelet-leukocyte aggregates, Finsterbusch [/bib_ref]. Moreover, serum sPselectin levels were found to be higher in moderate and severe cases of COVID-19 than in a healthy control group, suggesting the potential of sP-selectin as a prognostic marker for COVID-19 disease [bib_ref] Soluble P-selectin as a potential diagnostic and prognostic biomarker for COVID-19 disease:..., Karsli [/bib_ref]. It has been noted that sP-selectin, the form in circulation, is a useful biomarker and potential contributor to vascular complications [bib_ref] Role of platelet P-selectin and microparticle PSGL-1 in thrombus formation, Furie [/bib_ref]. Additionally, as patients with COVID-19 have highly stimulated circulating platelets, it has been suggested that P-selectin can be utilized as a marker of platelet activation in COVID-19 infection [bib_ref] Association of neutrophil activation, more than platelet activation, with thrombotic complications in..., Petito [/bib_ref]. Studies have further reported that crizanlizumab, a monoclonal antibody that targets sP-selectin, can decrease inflammation by binding to it and blocking leucocyte and platelet adherence to the vessel wall [bib_ref] Leukocyte adhesion to P-selectin and the inhibitory role of crizanlizumab in sickle..., Man [/bib_ref]. Consistent with these findings, blocking of sP-selectin has been demonstrated to attenuate the hyperinflammatory and hyperthrombotic state characteristic of COVID infections [bib_ref] The inhibition of P-Selectin reduced severe acute lung injury in immunocompromised mice, Liu [/bib_ref]. Given the prothrombotic consequences of increases in P-selectin on COVID-19 immunothrombosis, interventions targeting Pselectin may be favorable in dually targeting platelet and TABLE 3 Status of inflammatory cytokines and chemokines in serum and RNA levels of inflammatory markers in whole blood determined by qRT-PCR and linear regression analyses on change from baseline scores in adults recently diagnosed with COVID-19 in the LEV and CON groups (n = 30/group) 1 endothelial cell mechanisms of this disease. In the present study, ingestion of LEV for 4 wk induced a significant reduction in sPselectin from baseline in adults recently diagnosed with COVID-19, as opposed to an increase in this marker in the CON group. Additionally, it has been noted that whereas inflammatory mediators play a crucial role in viral defense, disproportionate immune responses are associated with the severity of COVID-19. Several proinflammatory cytokines, chemokines, and markers of infection have been indicated to be upregulated in COVID-19: C-reactive protein, IL-6, IL-1β, IL-2, TNF-α, monocyte chemoattractant protein 1, granulocyte-macrophage colony-stimulating factor, and IFN-γ [bib_ref] COVID-19: inflammatory profile, Wang [/bib_ref]. Of note, several of these factors are upregulated by increases in NF-κB expression. Therefore, it has been suggested that the NF-κB pathway could be a beneficial therapeutic target for mitigating COVID-19 severity [bib_ref] The role and therapeutic potential of NF-kappa-B pathway in severe COVID-19 patients, Hariharan [/bib_ref]. In the present study, we also noted significant mean decreases in IL-1β (P = 0.0222) and IL-2 (P = 0.0492) associated with LEV treatment. Evidence from a 2016 study suggested that PEA administration reduces P-selectin, neutrophil infiltration, cytokine production (TNF-α and IL-1β), and NF-κB expression in rats following myocardial ischemia reperfusion injury [bib_ref] Protective effects of ultramicronized palmitoylethanolamide (PEA-um) in myocardial ischaemia and reperfusion injury..., Paola [/bib_ref]. It also suggested, in support of the PEA mechanisms previously reported, that the effect of NF-κB may be associated with PPARα activation and expression. This finding agrees with a prior study demonstrating that the anti-inflammatory effects of PEA are in part mediated by PPAR-α activation [bib_ref] Palmitoylethanolamide reduces early renal dysfunction and injury caused by experimental ischemia and..., Paola [/bib_ref]. Importantly, NF-κB and mast cell activation are linked to production of inflammatory cytokines, which can induce the expression of adhesion molecules on the endothelium [bib_ref] Protective effects of ultramicronized palmitoylethanolamide (PEA-um) in myocardial ischaemia and reperfusion injury..., Paola [/bib_ref]. Regarding the results of the present study, although systemic reductions in sP-selectin, IL-2, and IL-1β were observed in the experimental group (LEV), we did not observe reductions in expression of NF-κB, IL-6, CD177, and other circulating inflammatory factors. However, it is possible that these markers would be modestly upregulated in those with mild-moderate COVID-19 infections, as participants in this study were not hospitalized for COVID-19 illness. It is also plausible that whereas sP-selectin may be a simple diagnostic tool related to COVID-19 diagnosis and severity, alterations in markers associated with sP-selectin reduction at the level of specific target tissues (e.g., the lungs) may not be revealed by our analysis. Importantly, a recent investigation suggested that murine alveolar macrophages challenged with the SARS-CoV-2 spike viral protein in vitro exhibited significant increases in proinflammatory markers, including NF-κB, IL-6, TNF-α, and IL-1β, and inflammasome expression, with this response being reduced significantly by treatment with PEA [bib_ref] Ultramicronized palmitoylethanolamide inhibits NLRP3 inflammasome expression and pro-inflammatory response activated by SARS-CoV-2..., Del Re [/bib_ref]. It was additionally demonstrated that this response was mediated by PPAR-α. Moreover, in a study of LPS-induced acute lung injury in rats, PEA administration reduced markers of neutrophil infiltration, immune cells quantity, and mast cell degranulation in the lungs and inhibited bronchoalveolar lavage fluid levels of the proinflammatory cytokines IL-6, TNF-α, IL-1β, and IL-18. Furthermore, PEA was suggested to block inflammatory cytokine production through inhibition of NF-κB activation in lung tissue [bib_ref] Management of acute lung injury: palmitoylethanolamide as a new approach, Peritore [/bib_ref]. It is also possible that the reductions in sP-selectin and cytokines found in this study could be related to implications of PEA in the entourage effect. It has been postulated that PEA can inhibit degradation of endocannabinoids, thereby potentiating the effects at their targets. In line with this potential contributory mechanism, Zhao et al. [bib_ref] Activation of cannabinoid CB2 receptor ameliorates atherosclerosis associated with suppression of adhesion..., Zhao [/bib_ref] demonstrated that activation of CB2 receptors can inhibit the expression of P-selectin in an animal model of atherosclerosis, which was associated with reductions in macrophage infiltration. It has also been demonstrated that PEA may enhance the release of 2-arachidonoyl glycerol, an endocannabinoid that has been shown to reduce IL-2 secretion [bib_ref] A novel crosstalk within the endocannabinoid system controls GABA transmission in the..., Musella [/bib_ref] [bib_ref] Interleukin-2 suppression by 2-arachidonyl glycerol is mediated through peroxisome proliferator-activated receptor γ..., Rockwell [/bib_ref] [bib_ref] The anti-inflammatory mediator palmitoylethanolamide enhances the levels of 2-arachidonoyl-glycerol and potentiates its..., Petrosino [/bib_ref]. Although the mechanisms attributable to PEA that are associated with reductions in P-selectin and inflammation have not been fully elucidated, the findings of this study support a growing body of evidence that PEA may be beneficial in modulating thromboinflammatory markers in various models and disease states. Our findings within the current context are especially promising, as P-selectin has recently been associated with COVID-19 coagulopathy in both the short term and the long term (i.e., "long COVID") [bib_ref] Soluble P-selectin as a potential diagnostic and prognostic biomarker for COVID-19 disease:..., Karsli [/bib_ref] [bib_ref] Persistent clotting protein pathology in long COVID/post-acute sequelae of COVID-19 (PASC) is..., Pretorius [/bib_ref]. A limitation to this study was that our participants were allowed to the test site following CDC return-to-work guidance; therefore, we were unable to capture responses immediately following viral infection. Yet, it is important to note that the LEV and CON groups were comparable in terms of the types of COVID-19 symptoms reported and the total number of symptoms reported. Furthermore, the CDC stated that inflammatory reactions and health problems related to COVID-19 can be experienced weeks after viral infection, even in those who experienced no or minimal symptoms during infection. With all of this noted, secondary outcome data were not adjusted for multiplicity, and caution should be exercised in interpreting findings of secondary outcome analysis. Finally, it is a limitation of this trial that plasma concentrations of PEA were not measured; however, a previous report documented the pharmacokinetics and bioavailability of the Levagen + brand of PEA as well as the standard PEA formulation in humans [bib_ref] Increased absorption of palmitoylethanolamide using a novel dispersion technology system (LipiSperse) study..., Briskey [/bib_ref]. Elevations in plasma PEA were evident at 30 min postdose and remained elevated for at least 4 h postdose for both preparations, and bioavailability was elevated nearly 75% for the Levagen + PEA formulation compared with the standard PEA formulation [bib_ref] Increased absorption of palmitoylethanolamide using a novel dispersion technology system (LipiSperse) study..., Briskey [/bib_ref]. Though inflammatory mechanisms are crucial to an optimal immune response, unchecked secretion of cytokines and thromboinflammatory markers can promote the development of the inflammatory response in unresolved disease states and is implicated in COVID-19 complications. Therefore, the reduction in inflammatory markers noted herein suggests that PEA may exert anti-inflammatory actions and possibly reduce the severity of COVID-19. P-selectin and inflammatory cytokines are also elevated in many chronic conditions linked to inflammation, including obesity, atherosclerosis, asthma, and cancer; hence, Levagen + administration may offer a degree of relief from inflammatory symptoms for many chronic conditions. [fig] FIGURE 1: Study participant flow. CON, placebo; LEV, 600 mg Levagen + twice daily. [/fig] [table] TABLE 1: Baseline characteristics of study participants (N = 60) in the LEV and CON groups 1 [/table] [table] TABLE 2: Status of primary outcome serum inflammatory biomarkers at baseline and after the 4-wk intervention and linear regression analyses on change from baseline scores in adults recently diagnosed with COVID-19 in the LEV and CON groups (n = 30/group) 1 Data are presented as mean ± SD. Data were analyzed by multiple linear regression models with the change from baseline as the response variable regressed on the treatment group, adjusting for age, sex, BMI, and existing conditions. CON, placebo; COVID-19, coronavirus disease 2019; CRP, C-reactive protein; ICAM-1, intercellular adhesion molecule 1; LEV, 600 mg Levagen + twice daily; N/L, neutrophil/lymphocyte ratio; sP-selectin, soluble P-selectin.2 Data are presented as the regression estimate (95% CI) of the estimate for the group assignment variable from each model. * Significance is retained after correction for multiple comparisons: Benjamini-Hochberg false discovery rate ≤ 0.10. [/table] [bib_ref] Could natural products modulate early inflammatory responses, preventing acute respiratory distress syndrome..., Amaral-Machado [/bib_ref]

Embodied Cognition in Meditation, Yoga, and Ethics—An Experimental Single-Case Study on the Differential Effects of Four Mind–Body Treatments # Introduction Yoga has been proposed as an embodied intentional practice and path toward eudaimonic well-being and a well-lived life of fulfilled meaning and purpose [bib_ref] Toward an Explanatory Framework for Yoga Therapy Informed by Philosophical and Ethical..., Sullivan [/bib_ref]. Theories on "embodiment" or "embodied cognition" assume a complex reciprocal relationship between bodily states, emotion, cognition, and behavior [bib_ref] Embodiment in attitudes, social perception, and emotion, Niedenthal [/bib_ref] [bib_ref] Embodiment of cognition and emotion, Winkielman [/bib_ref]. Nevertheless, the body is still overlooked in positive psychology, despite its undeniable interconnection with the overall experience of well-being and flourishing [bib_ref] The Role of Embodiment in Optimal Functioning, Hefferon [/bib_ref]. On the other hand, yoga is one of the most renowned mind-body exercises, which promotes engagement with body sensations and awareness [bib_ref] Minding the body: Yoga, embodiment, and well-being, Impett [/bib_ref]. Along with other mind-body practices, it has been proposed to connect body and mind and facilitate the integration of both bottom-up physiological and top-down cognitive processes [bib_ref] How does yoga reduce stress? Embodied cognition and emotion high-light the influence..., Francis [/bib_ref] [bib_ref] Potential self-regulatory mechanisms of yoga for psychological health, Gard [/bib_ref] [bib_ref] Neurophysiological and neurocognitive mechanisms underlying the effects of yoga-based practices: Towards a..., Schmalzl [/bib_ref]. These practices might enhance practitioners' experience of unity between the body and the self, or of an "embodied self". However, traditional yoga encompasses a multiplicity of components, and little is known about which differential health benefit is associated with which yoga component. In addition to the lack of differentiation and systematic investigation, the methodological quality of yoga studies has often been poor [bib_ref] Yoga for depression: A systematic review and me-ta-analysis, Cramer [/bib_ref] [bib_ref] Control group design: Enhancing rigor in research of mind-body thera-pies for depression, Kinser [/bib_ref]. Dismantling or comparative studies with several repeated measurements, such as experience sampling or single-case research designs [bib_ref] How mindfulness training promotes positive emotions: Dismantling acceptance skills training in two..., Lindsay [/bib_ref] [bib_ref] Tracking longitudinal changes in affect and mindfulness caused by concentration and loving-kindness..., May [/bib_ref] , provide a promising approach in this regard. Therefore, in the present study, we used a single-case multiple-baseline design to evaluate four different combinations of yoga components. The components were based on the new mind-body program Meditation-Based Lifestyle Modification (MBLM) [bib_ref] Meditation-Based Lifestyle Modification: Development of an Integrative Mind-Body Program for Mental Health..., Bringmann [/bib_ref]. We chose this design as it has strong internal validity and enables researchers to investigate in-depth how change processes develop over time and across different individuals. This makes it highly suitable for explorative research. With this approach, we hope to determine which combination of components might be particularly beneficial for improving body awareness, emotion regulation, affectivity, self-compassion, and distress tolerance in healthy participants. From theory and the literature reviewed above, we assume differential effects for different combinations and stronger (incremental) effects for combined interventions. It seems plausible that both physical yoga and meditation, and specifically their combination, enhance our variables of interest. The physical yoga component might even have a slightly greater impact on body awareness than meditation alone. Ethical education, and specifically its combination with other practices, might be particularly beneficial for developing emotion regulatory skills [bib_ref] Yoga and emotion regulation: A review of primary psychological outcomes and their..., Menezes [/bib_ref]. However, as we know little about how different mind-body interventions differentially and incrementally affect processes of embodied cognition, we chose an exploratory approach. Combining quantitative and qualitative methods, we examined interindividual differences, incremental differences between conditions, and interactions between our dependent variables. # Method ## Procedure The present study is a secondary analysis of data that was previously reported by Matko et al. [bib_ref] Differential effects of ethical education, physical Hatha yoga, and mantra meditation on..., Matko [/bib_ref] [bib_ref] Who Benefits Most? Interactions between Personality Traits and Outcomes of Four Incremental..., Matko [/bib_ref]. It expands the previously reported findings by providing a detailed analysis of affectivity, body awareness, emotion regulation, distress tolerance, and selfcompassion. The study employed a single-case multiple-baseline design to investigate the effects of different combinations of yoga practices over the course of four different treatments. Participants went through a baseline phase, where they received no treatment, and an 8-week treatment phase. The baseline phase could be 7, 14, or 21 days and treatments could be mantra meditation alone (MA), mantra meditation plus physical Hatha yoga (MY), mantra meditation plus ethical education (ME), and mantra meditation plus physical yoga and ethical education (MYE). The baseline length was determined by balancing the need for repeated measurements and the anticipated response burden of our participants. We randomized participants across these conditions and baseline lengths using a random number generator. Participants completed short daily online questionnaires throughout the entire study and longer online questionnaires at pre-test, post-test, and follow-up (8 weeks and 12 months). The study methods and treatments are described in more detail by Matko et al. [bib_ref] Differential effects of ethical education, physical Hatha yoga, and mantra meditation on..., Matko [/bib_ref]. ## Participants Fifty-seven participants were recruited from the Dresden general community through public and web-based advertising. Forty-two participants, 83% women, M (SD) age = 26.62 , completed the treatment and were included in the analysis. Individuals were excluded from participation if they had pre-existing psychiatric conditions or acute psychological issues, had a regular yoga or meditation practice during the last 6 months, were younger than 18 years, or did not have daily access to web-enabled devices. Participation was voluntarily and all participants provided written consent to participate in the study. The nature of the study was fully disclosed to participants before randomization and data collection, but they were not allowed to choose their treatment. Participants received no financial or other compensation for participating in the study. The institutional review 4 of 24 board of the Chemnitz University of Technology approved the experimental protocol. We analyzed the data of all 42 participants on the daily measurements. Of all participants, 41 completed the post-test, 40 the 2-month follow-up, and 33 the 12-month follow-up. For these data, we performed the analyses only on complete datasets. ## Treatment HCB, an accredited psychiatrist and psychotherapist, and KM, a psychologist and certified yoga instructor with 8 years of teaching experience, jointly led all treatments. The four treatments differed in the length of their weekly meetings due to the difference in the included components (total length MA: 60 min, MY: 105 min, ME: 135 min, MYE: 180 min). However, in each condition, the weekly sessions started and ended with the group sharing their experiences and included a collective 25-min silent mantra meditation practice. In the treatments that included Hatha yoga, participants practiced simple breathing techniques, the sun salutation, relaxation techniques, and a set of simple yoga postures. During ethical education, participants learned about and discussed the 10 yogic Yamas and Niyamas, following the protocol of the MBLM mind-body program [bib_ref] Meditation-Based Lifestyle Modification: Development of an Integrative Mind-Body Program for Mental Health..., Bringmann [/bib_ref]. In addition, and according to their treatment, participants should daily practice 20 min of mantra meditation, 20 min of yoga exercises, and/or engage in mindful living activities related to the ethical topic of the week. ## Measures All measures were obtained online using SoSci Survey. The measures that were collected daily or every few days had to be suitable for repeated measurement and, thus, be short and sensitive to changes. For this reason, we had to adapt the established questionnaires for body awareness and emotion regulation. We measured body awareness two times per week, and, for this purpose, developed a new questionnaire by extracting the 11 most suitable items from four questionnaires (FFMQ; [bib_ref] Using Self-Report Assessment Methods to Explore Facets of Mindfulness, Baer [/bib_ref] , MAIA; [bib_ref] The Multidimensional Assessment of Interoceptive Awareness (MAIA), Mehling [/bib_ref] , PBCS; [bib_ref] Consciousness of body: Private and public, Miller [/bib_ref] , BAQ; [bib_ref] The Body Awareness Questionnaire: Reliability and Validity, Shields [/bib_ref]. This questionnaire was thoroughly tested in two pilot studies. The final set of items can be found in Supplementary Material- . We used a shortened version of the Difficulties in Emotion Regulation Scale (DERS; [bib_ref] Multidimensional assessment of emotion regulation and dysregulation: Development, factor structure, and initial..., Gratz [/bib_ref] to weekly measure emotion regulation. This scale is a well-validated and widely used self-report measure with excellent psychometric properties (internal consistency α = 0.93). Higher scores indicate higher levels of emotion dysregulation. As the DERS is a very long scale for regular measurements, we decided to shorten it by extracting the two items with the highest factor loadings per subscale, resulting in 12 items altogether (see Supplementary Material- . The daily affective experience of participants was assessed with the economic single-item measure Affective Grid [bib_ref] Affect grid: A Singleitem Scale of Pleasure and Arousal, Russell [/bib_ref] , which was specifically developed for repeated observations. The Affective Grid assesses affect along the two dimensions, pleasure-displeasure and arousal-sleepiness, and has shown adequate reliability and validity [bib_ref] The Affect Grid: A Moderately Valid, Nonspecific Measure of Pleasure and Arousal, Killgore [/bib_ref]. We measured distress tolerance with the Distress Tolerance Scale (DTS; [bib_ref] The Distress Tolerance Scale: Development and Validation of a Self-Report Measure, Simons [/bib_ref]. It is a 15-item self-report questionnaire examining the degree to which individuals experience negative emotions as intolerable. The DTS has shown good internal consistency (α = 0.89). We assessed self-compassion with the German version of the Self-Compassion Scale (SCS-D; [bib_ref] Validierung einer deutschen Version der Self-Compassion Scale (SCS-D), Hupfeld [/bib_ref]. It evaluates a person's ability to be kind and forgiving to themselves in difficult circumstances. The SCS has been shown to be valid and reliable. To reduce response burden, we decided to shorten the original 26-item scale by again extracting the two items with the highest factor loadings per subscale, reducing it to 12 items (see Supplementary Material- . Distress tolerance was measured at pre-test, post-test, and 2-month follow-up. Self-compassion was assessed at pre-and post-test. Body awareness and emotion regulation were measured daily during the main study period and once at 2-and 12-month follow-up. For consistency reasons, all questionnaires (except for the Affective Grid) were rated on a 5-point Likert scale. At the beginning of the treatment phase, the daily questionnaire included questions that measure participants' home practice. They were asked to indicate how many minutes they spent meditating and practicing yoga, and whether they engaged in any of the mindful living activities, according to their condition. Furthermore, we assessed the subjective experiences participants had during each of these practices. We reported the detailed results regarding these variables in another publication [bib_ref] Differential effects of ethical education, physical Hatha yoga, and mantra meditation on..., Matko [/bib_ref]. Qualitative data were obtained from several free text items. In the daily questionnaire, participants could describe any special events that occurred on that day. At 2-and 12-month follow-up, we asked the participants to respond to several open-ended questions to reflect their experiences during and after the course and to describe whether and how they continued to practice the components of their course. # Data analysis Data analysis of single-case studies can be manifold [bib_ref] Special Issue on Advances in Single-Case Research Design and Analysis, Machalicek [/bib_ref] [bib_ref] Analysis and meta-analysis of single-case designs: An introduction, Shadish [/bib_ref] , as we reported in a previous publication [bib_ref] Differential effects of ethical education, physical Hatha yoga, and mantra meditation on..., Matko [/bib_ref]. Commonly, the progression of the dependent variables over time are plotted in graphs, which are then visually analyzed [bib_ref] Visual analysis in single case experimental design studies: Brief review and guidelines, Lane [/bib_ref]. In addition, multiple analysis strategies are chosen to increase the reliability and generalizability of findings. Accordingly, we analyzed data by visual inspection, calculating effect sizes, and multilevel modeling. Moreover, we supplemented our analyses with qualitative findings to explore possible reasons for interindividual differences. During visual inspection, we assessed trends in the baseline and the treatment phase, differences between the means and the variability of data in each phase, and the consistency of patterns across individuals and conditions. For an experimental single-case study, our sample was exceptionally large. Therefore, we kept our visual analyses relatively simple. Tau-U is a family of non-parametric effect size estimates that allow for controlling trends in both phases [bib_ref] Combining Nonoverlap and Trend for Single-Case Research: Tau-U, Parker [/bib_ref]. We calculated Tau-U coefficients for each participant and corrected trends in any phase when they were statistically significant or larger than 0.40. An effect size of less than 0.28 indicated a small effect; 0.29-0.47 a moderate effect; 0.48-0.57 a large effect; and 0.58 or above a very large effect [bib_ref] Critical Assumptions and Distribution Features Pertaining to Contemporary Single-Case Effect Sizes, Solomon [/bib_ref]. Multilevel modeling (or hierarchical linear modeling) is a powerful tool for examining both individual change and group differences [bib_ref] Multilevel Modeling: A Review of Methodological Issues and Applications, Dedrick [/bib_ref] [bib_ref] Quantitative methodology series, Hox [/bib_ref]. In this study, changes over time were modeled on one level and differences between individuals on a second level. We standardized all variables for the analyses. Time was coded with zero for the baseline phase and a logarithmic trend starting at the beginning of the treatment phase. We modeled time as a random slope and applied one-tailed tests as we expected all treatments to improve our dependent variables. To assess the proportion of explained variance in each model, we assessed marginal R 2 and conditional R 2 [bib_ref] A general and simple method for obtaining R2 from generalized linear mixed-effects..., Nakagawa [/bib_ref]. All models were estimated using the restricted maximum likelihood estimation procedure. We examined the incremental effects of the four conditions in an exploratory manner with (1) three dummy variables coding the four conditions (condition model), and (2) two other dummy variables coding the inclusion of the physical yoga or ethical education component (component model). These dummy variables were entered as predictors into two sets of regression models. First, we estimated simple regression models with the Tau-U effect sizes as criterion. Second, consistent with the first procedure, we modeled cross-level interactions between time and condition in our multilevel models with the continuous measures of our dependent variables as criterion. Both analyses produce comparable results. All models controlled for individual practice time, age, gender, occupation, and baseline length. Individual practice time was calculated by summing the reported length of each home practice. We hypothesized that combined interventions would have stronger effects than meditation alone and applied one-tailed tests of significance to the dummy variables. We considered p < 0.05 to be statistically significant. All statistical analyses were performed using R 4.2.0and the statistical packages lattice, ggpubr, scan, and nlme. Proportion of explained variance in multilevel models was calculated using the R-based online application mimosa [bib_ref] A modern graphical user interface for 2-level mixed models, Titz [/bib_ref]. All scripts and data that support the results can be found at https://osf.io/wnxk6/. This study's design was preregistered at clinicaltrials.gov under NCT04252976. # Results For the three continuously measured variables (body awareness, emotion regulation, affectivity), we report the results of our visual inspection, effect size estimation, and multilevel modeling on the general as well as incremental effects we observed. Then, we explore the relations between these three dependent variables before reporting the results on distress tolerance and self-compassion. Finally, we expand on the long-term effects and observations of specific single cases with possible explanations of these. [fig_ref] Figure 1: Body awareness scores in four conditions during baseline and treatment phases for... [/fig_ref] illustrates the development of body awareness scores of each participant over time. Most participants' body awareness did not significantly fluctuate, indicating that this construct was relatively stable and not subject to daily variation. Overall, there were observable interindividual differences in the general level and slope of the curves. ## Body awareness During baseline, most participants showed no change or a decrease in body awareness. A few participants' body awareness increased during baseline. Both increases and decreases in baseline could be due to the heightened sensitivity to the instrument. Nevertheless, the majority of participants showed a steady improvement in body awareness over the course of the treatment with effect sizes ranging from 0.26 to 0.87 (for full tables see Supplementary Material- . The increase was gradual for most participants starting with the beginning of the treatment or after a small delay. It seems as if being introduced to meditative practices influenced body awareness quite quickly, but continuous practice led to steady improvements. Only two participants showed no observable changes and had effect sizes close to zero. Nine participants experienced a moderate deterioration of body awareness with effect sizes ranging from −0.28 to −0.59. Moreover, we examined whether these changes persisted until follow-up two months after the study. Therefore, we compared the mean body awareness scores from phase B (M B = 3.45, SD B = 0.71) with the 2-month follow-up measurements (M FU2 = 3.62, SD FU2 = 0.84), and found significant improvements, t(39) = 2.58, p = 0.014. Interestingly, this did not apply to participants who stopped meditating after the study-their body awareness was generally lower (M B = 3.01, SD B = 0.75) and remained at follow-up (M FU2 = 3.06, SD FU2 = 0.99). In addition, we performed an analysis of variance for the 32 participants who completed all three measurements. While this model was not significant, F(2, 62) = 2.87, p = 0.064, there was a medium effect size, η 2 = 0.09. This suggests that body awareness improved from post-test (M B = 3.38, SD B = 0.77) to 2 months (M FU2 = 3.55, SD FU2 = 0.89) and 12 months after the study (M FU12 = 3.57, SD FU12 = 0.67). Then, we investigated possible differences between the four conditions. Participants whose body awareness did not change or declined were found in every condition, but particularly in the MA condition. The most improvements were found in the ME and MY conditions. We further explored the incremental effects of our treatments by grouping all Tau-U effect size estimates by condition and generating four box plots [fig_ref] Figure 2: Box plots for averaged Tau-U body awareness effect size estimates in each... [/fig_ref]. In summary, although there was substantial interindividual variation between participants, all four treatments helped the participants in enhancing their body awareness. This finding is supported by all of our three analyses and is independent of other potentially influential variables. In summary, although there was substantial interindividual variation between participants, all four treatments helped the participants in enhancing their body awareness. This finding is supported by all of our three analyses and is independent of other potentially influential variables. inferences on baseline trends, as there were only few baseline reference points. Three participants had to be excluded from the statistical analysis due to missing baseline data. fluctuate, but slopes and levels varied between participants. It was difficult to make reliable inferences on baseline trends, as there were only few baseline reference points. Three participants had to be excluded from the statistical analysis due to missing baseline data. When we tentatively compared both phases, we found that most participants exhibited a clear downward trend regarding their difficulties in emotion regulation. This indicates that they improved their ability to be aware of and regulate their feelings. A negative sign in effect sizes represented reduced difficulties in emotion regulation and, thus, a change in the expected direction. Three quarters of the effect size estimates ranged from −0.30 to −0.90 indicating moderate to very large effects (see Supplementary Material- . For most participants, this amelioration became gradually stronger over time. Very few participants showed no observable changes or increased difficulties in emotion regulation. All conditions comprised participants with large effect sizes, whereas the conditions ME and MY comprised the most. In every condition, we found one participant whose emotion regulation skills weakened over time. When we tentatively compared both phases, we found that most participan ited a clear downward trend regarding their difficulties in emotion regulation. T cates that they improved their ability to be aware of and regulate their feelings. A sign in effect sizes represented reduced difficulties in emotion regulation and change in the expected direction. Three quarters of the effect size estimates rang −0.30 to −0.90 indicating moderate to very large effects (see Supplementary M . For most participants, this amelioration became gradually stronger ov Very few participants showed no observable changes or increased difficulties in regulation. When we compared the mean scores from phase B (MB = 2.21, SDB = 0 the 2-month follow-up measurements (MFU2 = 2.14, SDFU2 = 0.60), we found no si differences, t (37) All conditions comprised participants with large effect sizes, whereas the co ME and MY comprised the most. In every condition, we found one participan emotion regulation skills weakened over time. [fig_ref] Figure 4: depicts box plots with effect size estimates grouped by condition [/fig_ref] depicts box plots with e estimates grouped by condition. In the overall multilevel model, there was an effect of time, F(1, 334) = 30.62, p < 0.001, but no significant cross-level interaction, F(3, 334) = 0.60, p = 0.617. Both models predicted a significant improvement in emotion regulation over time, β = −0.26, SE = 0.05, p < 0.001. Once again, there were no significant predictors nor significant cross-level interactions, suggesting that all conditions enhanced regulatory skills equally well. Effect sizes for both models were marginal R 2 = 0.19 and conditional R 2 = 0.81. For full regression tables, see Supplementary Material-Tables S10-S13. Summarizing the three analyses, all four treatments strengthened participants' abilities to regulate their emotional responses. ## Emotion regulation ## Emotion regulation ## Affectivity The variable-by-time plots for valence and arousal are depicted in [fig_ref] Figure 5: Valence scores in four conditions during baseline and treatment phases for each... [/fig_ref] , respectively. A thin horizontal line was drawn into each graph at the score 5 to delineate a neutral valence or arousal that is neither high/positive nor low/negative [bib_ref] Affect grid: A Singleitem Scale of Pleasure and Arousal, Russell [/bib_ref]. The visual inspection of these plots indicated strong inter-and intra-individual fluctuations in both affective dimensions over time. In most cases, the baseline phase displayed a stronger upward or downward trend than the treatment phase. These trends reversed or leveled out during the treatment phase. The variability of scores was comparable across both phases. Effects were inconsistent across individuals. Both arousal and valence increased, decreased, or stayed the same from baseline to treatment. A majority of effect size estimates ranged from −0.25 to 0.37 for valence, and −0.30 to 0.40 for arousal, indicating a great variability in the results (see Supplementary Material- . Valence declined for a majority of participants from the MA condition, but improved for many participants from the ME condition. . We had to exclude one participant from the regression analyses of arousal as she was an outlier who was considerably older and had substantially elevated levels of arousal. The aggregated effect of all four conditions for valence and arousal was F(3, 33) = 1.17, p = 0.334 and F(3, 32) = 0.73, p = 0.543, respectively, suggesting no relevant differences between conditions. In the condition model of valence, receiving the ME treatment significantly predicted an increase in valence over time, β = 0.41, p = 0.041, suggesting that this condition improved positive affect. Neither ethical education or physical yoga nor any other variable significantly predicted changes in valence and arousal. However, in the condition model of arousal, being employed (compared to being a student) was a significant predictor of increased arousal over time, β = 0.45, p = 0.038. Multiple The visual inspection of these plots indicated strong inter-and intra-individual fluctuations in both affective dimensions over time. In most cases, the baseline phase displayed a stronger upward or downward trend than the treatment phase. These trends valence increased, decreased, or stayed the same from baseline to treatment. A majority of effect size estimates ranged from −0.25 to 0.37 for valence, and −0.30 to 0.40 for arousal, indicating a great variability in the results (see Supplementary Material- . Valence declined for a majority of participants from the MA condition, but improved for many participants from the ME condition. [fig_ref] Figure 7: displays the grouped effect sizes of all four treatments on both variables [/fig_ref] displays the grouped effect sizes of all four treatments on both variables. In the overall multilevel model, we found no interaction effect for valence or arousal, F(3, 2492) = 1.89, p = 0.129, and F(3, 2454) = 0.00, p = 0.708, respectively. There were no significant predictors in the arousal models. In the condition model of valence, there were two significant cross-level interactions between time and the ME condition, β = 0.08, SE = 0.03, p = 0.013, and time and the MY condition, β = 0.06, SE = 0.03, p = 0.033, indicating an increase in valence over time. Neither the yoga nor the ethical component predicted valence in the component model. Therefore, the positive effect was limited to the two conditions that involved meditation, in addition to one additional component, but did not depend on a specific component. Interestingly, the only (other) significant predictor of valence in both models was the total practice time that the individuals completed, β = 0.17, SE = 0.08, p = 0.033/0.031. Therefore, the more participants engaged in a regular home practice, the more their valence increased. This was independent of their specific condition, although total practice time was longer in the more extensive conditions. The effect sizes for the final component models for valence/arousal were marginal R 2 = 0.04/R 2 = 0.04 and conditional R 2 = 0.22/R 2 = 0.15. Full regression tables can be found in the Supplementary Material-Tables S15-S22. Next, we considered both dimensions simultaneously, visually and by placing effect sizes of both variables side by side in a combined [fig_ref] Figure 8: Stacked bar chart of Tau-U valence and arousal estimates with qualitatively classified... [/fig_ref]. As a result, we developed five main categories of affective responses to our treatment, which are depicted with different symbols in [fig_ref] Figure 8: Stacked bar chart of Tau-U valence and arousal estimates with qualitatively classified... [/fig_ref]. The majority of participants experienced a positive change in affect, a small group of participants showed no observable changes, and some participants felt more stressed. tary Material-Tables S15-S22. Next, we considered both dimensions simultaneously, visually and by placing effect sizes of both variables side by side in a combined [fig_ref] Figure 8: Stacked bar chart of Tau-U valence and arousal estimates with qualitatively classified... [/fig_ref]. As a result, we developed five main categories of affective responses to our treatment, which are depicted with different symbols in [fig_ref] Figure 8: Stacked bar chart of Tau-U valence and arousal estimates with qualitatively classified... [/fig_ref]. The majority of participants experienced a positive change in affect, a small group of participants showed no observable changes, and some participants felt more stressed. Next, we compared the frequencies of each category of affective response in each condition (see Supplementary Material- . Around two-thirds of participants in the MA condition felt more stressed over the course of time. Half of the participants from the MY condition experienced no change in their affect. However, no one from this group was classified as being more stressed, suggesting a stress-relieving effect of physical yoga. Participants from both groups that received ethical education (ME and MYE) were more energetic, more relaxed, or more stressed during the intervention. Therefore, the ethical education component elicited multi-layered responses. This contrasts and complements our previous analyses where the MY and specifically the ME condition had a favorable effect on valence, whereas the MA condition was less advantageous. ## Relations between affectivity, body awareness, and emotion regulation Research literature postulates a close connection between body awareness and emotion regulation [bib_ref] Interoceptive Awareness Skills for Emotion Regulation: Theory and Approach of Mindful Awareness..., Price [/bib_ref]. Therefore, we explored this connection by correlating effect size estimates of both variables. We found a moderate correlation (r = −0.24), indicating that when participants' body awareness increased, their difficulties in emotion regulation decreased, as well. We found similar patterns and trajectories for both variables during visual inspection, indicating that both skills develop simultaneously during contemplative training. Next, we identified whether there were any differences between the conditions regarding this relation. Interestingly, correlations were significantly higher in the MA (r = −0.53) and MY (r = −0.43) conditions compared to both conditions that received ethical education: ME (r = 0.20) and MYE (r = −0.15). Therefore, in treatments where participants were mainly focused on observing rather than actively changing themselves, increases in body awareness were more strongly related to increases in emotion regulation. Furthermore, we were interested in how changes in both skills were related to changes in daily affect. We found that decreases in difficulties in emotion regulation were higher when the affective response was classified as positive (Mdn = −0.52, IQR = 0.31), compared to no change (Mdn = −0.38, IQR = 0.37) or negative (Mdn = −0.29, IQR = 0.66). In contrast, enhancements in body awareness were highest when participants experienced no change in affectivity (Mdn = 0.59, IQR = 0.79), compared to positive (Mdn = 0.43, IQR = 0.36) or negative (Mdn = 0.24, IQR = 0.75). This indicates that emotion regulation and affective experience co-varied, whereas body awareness and affective experience did not. ## Distress tolerance For distress tolerance, a mixed two-way ANOVA showed a significant effect of time, F(2, 72) = 23.01, p < 0.001, with a large effect size, η 2 = 0.13. There were no significant differences between the four conditions, F(3, 36) = 1.27, p = 0.299, nor a significant interaction between time and condition, F(6, 72) = 0.51, p = 0.746. The mean distress tolerance across all conditions was M pre = 3.42 (SD = 0.74) before the study, M post = 3.95 (SD = 0.72) at completion of the study, and M fu = 3.99 (SD = 0.70) 2 months after the study. Therefore, all treatments markedly enhanced distress tolerance from pre-to post-test, which was sustained until 2-month follow-up. ## Self-compassion Another mixed two-way ANOVA yielded a significant effect of time for self-compassion, F(1, 37) = 26.25, p < 0.001, with a medium effect size, η 2 = 0.08. Once again, there were no significant differences between the four conditions, F(3, 37) = 1.23, p = 0.311, nor a significant interaction between time and condition, F(3, 37) = 1.57, p = 0.213. The mean self-compassion across all conditions increased from M pre = 3.00 (SD = 0.60) before the study to M post = 3.37 (SD = 0.64) at completion of the study. Accordingly, all treatments helped the participants in cultivating their ability to be kind and forgiving to themselves. ## Changes in home practice The reported long-term effects were found although the practice times declined considerably from the last month of the treatment to 2-and 12-month follow-up . We calculated the accumulated minutes of each practice per month to obtain comparable values. The engagement in all practices decreased over time, but meditation and ethical practice decreased more steeply than the yoga practice. Looking at the reported practice times at 12-month follow-up, we found another interesting result. Ethical practice was reported mainly in the two conditions that included ethical education. Meditation practice times were comparable across conditions. However, surprisingly, the physical yoga practice was reported consistently in all four conditions, with the highest mean practice time in the MYE condition (3 h per month). Similarly, of all 33 participants who completed the 12-month follow-up, 63.4% still meditated, 57.6% practiced ethics, and 72.7% practiced yoga. ## Selected cases Subsequently, we looked at potential explanations for the negative effects we found for body awareness and emotion regulation. Participants 5, 29, 31, and 35 experienced a considerable decline in body awareness throughout the treatment. Participant 5 (MA) was classified as becoming more stressed throughout the treatment, but markedly improved his distress tolerance over time. At 12-month follow-up, he stated being very grateful for the course, as he had learned a valuable relaxation technique that he could rely on, especially when things were getting stressful. He had meditated regularly during the course and had continued to meditate until the present time. We had considered excluding Participant 29 (MY) as she was not very conscientious in responding to our daily questionnaires and stopped responding by Week 5 of the treatment. Nevertheless, she regularly attended her class although she did not seem very enthusiastic during class and did not practice at home very often neither during the study nor afterwards. Interestingly, her affective experience did not change, her self-compassion declined, but her distress tolerance improved. On the other hand, participant 31 (MY) was very enthusiastic about her course, even in retrospective, and reported that particularly yoga became part of her daily routine up to the present time. She had high levels of distress tolerance from the beginning, her affective experience did not change, and her self-compassion increased over the course of the study. Participant 35 (MYE) conscientiously practiced yoga and meditation during the study, but lost motivation to continue although he was still interested. He found working out and being in nature more helpful. Throughout the study, he became more energetic, but his self-compassion or distress tolerance did not change significantly. In summary, although all four participants experienced a decline in body awareness, they responded quite differently (and often positively) to the treatment in other variables. . Accumulated monthly practice of meditation, physical yoga, and ethical practice in minutes for the last month of the treatment, the 2-and 12-month follow-up, and the 12-month follow-up in each condition. ## Meditation Physical Only two participants reported increased difficulties in emotion regulation over time. Interestingly, Participant 4 (MA) was the only participant who exhibited a deterioration in all three continuous dependent variables (affect, body awareness, emotion regulation). Nevertheless, her distress tolerance improved. Participant 38 (MYE) was classified as becoming more stressed and her distress tolerance declined, but her body awareness and self-compassion increased. We identified and discussed these two participants in another publication as they also experienced a considerable decline in well-being [bib_ref] Differential effects of ethical education, physical Hatha yoga, and mantra meditation on..., Matko [/bib_ref]. Indeed, effect sizes of difficulties in emotion regulation and well-being were highly correlated (r = −0.41). Nevertheless, both participants retrospectively appreciated their course and continued to meditate regularly. This might be an indication that the treatment led to an initial worsening during the process of change. # Discussion The present study investigated the specific and incremental benefits of adding physical Hatha yoga and ethical education to a mantra meditation intervention in healthy participants. It evaluated processes related to embodied cognition in four different combinations of yoga components, which are part of the new mind-body therapy MBLM [bib_ref] Meditation-Based Lifestyle Modification: Development of an Integrative Mind-Body Program for Mental Health..., Bringmann [/bib_ref]. The singlecase multiple-baseline design made it possible to evaluate individual responses and change processes over time. All four treatments had positive effects on body awareness, emotion regulation skills, distress tolerance, and self-compassion. Effects were strongest for emotion regulation and distress tolerance. Surprisingly, there were no significant between-group differences regarding these variables although the treatments differed considerably in session duration and complexity. In contrast, findings on daily affect were more differentiated, but also less conclusive. The MA condition had the least favorable effect on valence and overall affectivity, whereas the ME condition improved valence the most and the MY condition elicited no negative change in affectivity. However, due to the exploratory nature of our study, and the relatively small sample size, our findings should be regarded as preliminary. Nonetheless, these results complement and extend our previously reported findings on well-being and stress [bib_ref] Differential effects of ethical education, physical Hatha yoga, and mantra meditation on..., Matko [/bib_ref]. In this earlier publication, we elucidated that the simple meditation treatment was the least effective and led to increased stress and decreased well-being in a sizeable number of participants. Furthermore, the two less extensive combinations of yoga components, namely, the ME and MY condition were the most helpful in enhancing well-being (ME) and reducing stress (MY). This is in line with the abovementioned results on affectivity. Nevertheless, the condition-independent positive effects on the other dependent variables seem somewhat puzzling. These indicate that factors common to all four interventions were responsible for the observed changes in these variables. These factors might be unspecific, such as social support, group dynamics, or attention from the intervention teachers [bib_ref] Control group design: Enhancing rigor in research of mind-body thera-pies for depression, Kinser [/bib_ref]. On the contrary, most effects persisted until follow-up, some even until 12 months after the intervention had ended. Therefore, it seems more plausible to assume that these effects were caused by the specific shared component of mantra meditation. ## Mantra meditation might be the driving force behind subjective improvements Previous studies have substantiated the positive effects of mantra meditation on mental health and affectivity in healthy populations, although methodological flaws limit the validity of these findings [bib_ref] Mantra meditation for mental health in the general population: A systematic review, Lynch [/bib_ref] [bib_ref] The psychological effects of meditation: A meta-analysis, Sedlmeier [/bib_ref]. Our study provides methodologically sound evidence for the positive effects of mantra meditation on body awareness, emotion regulation, selfcompassion, and distress tolerance, but not necessarily on positive affect. Interestingly, the former variables were neither influenced by differences in session duration nor by the fact that participants were fully disclosed about the study design. Therefore, mantra meditation seems to have had a quite powerful impact. Fittingly, meditation is regarded as the key element in classical yoga (ashtanga yoga) as outlined by Patanjali, with ethics, postures, and breathing practices being considered as pre-requisites or preparatory exercises. Schmalzl et al. [bib_ref] Neurophysiological and neurocognitive mechanisms underlying the effects of yoga-based practices: Towards a..., Schmalzl [/bib_ref] proposed that meditation leads to an increase in interoception and metacognitive awareness. Accordingly, interoception has been suggested as one of the central mechanisms of meditation [bib_ref] Interoception, contemplative practice, and health, Farb [/bib_ref] [bib_ref] Interoception, and the Body: A Contemporary Perspective, Gibson [/bib_ref]. Furthermore, neuroscientific studies have shown consistent activations in brain regions associated with interoceptive processes (e.g., insula) across many different kinds of meditation [bib_ref] Functional neuroanatomy of meditation: A review and meta-analysis of 78 functional neuroimaging..., Fox [/bib_ref]. In a recent comparative study, walking meditation, humming meditation, concentration on an object, and observing-thoughts meditation led to similar increases in body awareness and emotion regulation in novice meditators. According to a recent classification system, mantra meditation does not have a strong orientation toward the body [bib_ref] What Is Meditation? Proposing an Empirically Derived Classification System, Matko [/bib_ref]. Nevertheless, it encourages meditators to turn their attention inwards, which might explain why it enhanced the processing of bodily and emotional stimuli in the present study. Similarly, the sustained focus on a simple, repetitive stimulus (mantra) resembles states of sensory deprivation and perceptual isolation [bib_ref] A Phenomenology of Meditation-Induced Light Experiences: Traditional Buddhist and Neurobiological Perspectives, Lindahl [/bib_ref]. These states have been shown to intensify or alter sensory perceptions and make them more salient [bib_ref] A qualitative analysis of sensory phenomena induced by perceptual deprivation, Lloyd [/bib_ref]. Additionally, participants were instructed to be kind and friendly toward themselves and their experiences during meditation. This might have enhanced their self-compassion and tolerance of negative emotions [bib_ref] How Does Mindfulness Meditation Work? Proposing Mechanisms of Action from a Conceptual..., Hölzel [/bib_ref]. Consequently, the strong focus of meditation on internal processes and on increasing self-awareness might be a central mechanism underlying the observed changes. ## Adding yoga components to meditation enhances positive affect In contrast, mantra meditation alone exhibited a partly negative effect on participants' daily affect. In this respect, adding other yoga components to the treatment proved to be a valuable extension and possibly provided participants with additional inspiration and a frame of reference [bib_ref] Differential effects of ethical education, physical Hatha yoga, and mantra meditation on..., Matko [/bib_ref]. While there were no significant between-group differences regarding body awareness, effect sizes were slightly higher in the groups with additional yoga components. Remarkably, it mostly did not matter which component was added or practiced in this regard. Physical Hatha yoga emphasizes becoming aware of the (moving) body and has been found to increase vagal/parasympathetic activity [bib_ref] Potential self-regulatory mechanisms of yoga for psychological health, Gard [/bib_ref] [bib_ref] How does yoga reduce stress? A systematic review of mechanisms of change..., Riley [/bib_ref]. This might explain its balancing effect on affectivity. On the other hand, ethical education encourages the development of value-oriented behavior [bib_ref] Toward an Explanatory Framework for Yoga Therapy Informed by Philosophical and Ethical..., Sullivan [/bib_ref] , and is associated with improved wellbeing, self-compassion, and anxiety. This is in accordance with its positive effects on valence as reported above. Intriguingly, physical yoga was the only component that was consistently practiced until 12-month follow-up, even in conditions that did not originally include physical yoga practice. This indicates that physical yoga might be easier to integrate into participants' lives, corresponding with earlier findings [bib_ref] Yoga in Arterial Hypertension, Cramer [/bib_ref]. ## Body awareness and emotion regulation are central mechanisms Body awareness and emotion regulation developed steadily and continuously over time, whereas daily affect fluctuated significantly. This indicates that the former are important skills and possible mechanisms of action that develop simultaneously over time. Both skills have been described as central to meditation [bib_ref] A qualitative analysis of sensory phenomena induced by perceptual deprivation, Lloyd [/bib_ref] , but more research is warranted to substantiate our findings. In addition, we found stronger correlations between changes in both skills in the non-ethical than in the ethical conditions. This supports the notion that postures, breathing, and meditation encourage an integrated and embodied way of processing and modulating bodily and emotional responses [bib_ref] Potential self-regulatory mechanisms of yoga for psychological health, Gard [/bib_ref]. On the other hand, the ethical education component might have strengthened top-down processing and, thus, a decoupling of body awareness and emotion regulation. This corresponds to the finding that an (embodied) yoga intervention was more effective than a (non-embodied) cognitivebehavioral stress-management course in enhancing interoceptive awareness and emotion regulation [bib_ref] Yoga and Cognitive-behavioral Interventions to Reduce Stress in Incoming College Students: A..., Park [/bib_ref]. We observed substantial interindividual variation in response to the treatments. While the majority of participants benefitted from the treatments, a few participants became more stressed or experienced a decline in their body awareness and/or emotion regulatory skills. As we know from psychotherapeutic contexts, change is not always easy and often leads to an initial impairment of symptoms [bib_ref] Shapes of early change in psychotherapy under routine outpatient conditions, Stulz [/bib_ref]. This might have happened to some of our participants, as well, since most of them were quite grateful for their course 12 months after the study. Moreover, a decrease in body awareness might indicate a shift in participants' frame of reference. Specifically, participants might be more aware or more mindful toward their inner experience during the treatment and, thus, adjust their responses [bib_ref] Challenging the Construct Validity of Mindfulness Assessment-a Cognitive Interview Study of the..., Belzer [/bib_ref]. Similarly, individual characteristics might influence a participant's response to the different yoga components [bib_ref] Neurophysiological and neurocognitive mechanisms underlying the effects of yoga-based practices: Towards a..., Schmalzl [/bib_ref] [bib_ref] Who Benefits Most? Interactions between Personality Traits and Outcomes of Four Incremental..., Matko [/bib_ref] [bib_ref] How Does Mindfulness Meditation Work? Proposing Mechanisms of Action from a Conceptual..., Hölzel [/bib_ref]. ## Limitations and future directions Although our study employed rigorous methodology, it had a few limitations, as well. First, participants in our study were presumably intrinsically motivated to participate in this study. Second, the intense data gathering was very demanding for participants and led to a relatively high number of dropouts and missing data. Therefore, providing financial or other compensation might help in enhancing participant commitment. Third, our sample size was somewhat limited for certain aspects of the evaluation. A sample of 42 participants is considered exceptionally large for an experimental single-case study. The large number of measurements (up to 85 for daily measures) makes our findings and effect size estimates very reliable. This applies less to measures collected less often and to between-group comparisons. However, a post-hoc power analysis indicated that our study was wellpowered to perform pre-post-follow-up analyses with medium to large effects. For the continuous measures, we performed both regression analyses and multilevel modeling to increase the reliability our findings. Finally, our results are hypotheses generating rather than testing. Therefore, they could guide future research with preferably larger subgroups. Furthermore, dependent variables should be collected more frequently with short, specific, and reliable instruments. In the meantime, researchers have developed short forms for measuring emotion regulation [bib_ref] The Difficulties in Emotion Regulation Scale Short Form (DERS-SF): Validation and Replication..., Kaufman [/bib_ref] and self-compassion [bib_ref] Construction and factorial validation of a short form of the Self-Compassion Scale, Raes [/bib_ref]. However, our body awareness questionnaire needs to be thoroughly validated to substantiate its usefulness beyond the present study. Daily affect could be assessed more specifically with simple questions, such as "To what extent did you feel happy/sad/calm/stressed today?" Moreover, specific emotion regulatory strategies, including the proposed mindful emotion regulation [bib_ref] Mindful emotion regulation: An integrative review, Chambers [/bib_ref] , or specific embodiment processes could be assessed in more detail. In addition, experiencesampling methods [bib_ref] Ecological momentary assessment, Shiffman [/bib_ref] represent an alternative research design, which is highly suited for capturing immediate experiences over the course of the day. Interestingly, the meditation-only condition, which served as an appropriate control group, produced the same effects for most variables as the more extensive conditions. Future studies could evaluate whether these effects generalize to different types of meditation [bib_ref] What Is Meditation? Proposing an Empirically Derived Classification System, Matko [/bib_ref]. Similarly, future studies could match conditions by session length and dismantle the effects of yoga components even further by comparing only ethics, only physical postures, only breathing techniques, or only meditation to diverse combinations of these. Likewise, they could include an active non-yoga and non-meditation (placebo) control group (see, e.g.,to expand on the specificity of the found effects. Furthermore, future studies should investigate how personality and motivational factors impact the outcomes of different treatments to enable personalized recommendations [bib_ref] Who Benefits Most? Interactions between Personality Traits and Outcomes of Four Incremental..., Matko [/bib_ref] [bib_ref] Towards an individual differences perspective in mindfulness training re-search: Theoretical and empirical..., Tang [/bib_ref]. In summary, all of these research efforts could help in developing a deeper understanding of the working mechanisms of the multifaceted practice of yoga and how it contributes to general well-being and flourishing. # Conclusions This experimental single-case study provided valuable insights into the working mechanisms of different yoga components. Although the investigated treatments differed in session length and components taught, they had a similarly positive impact on body awareness, emotion regulation, distress tolerance, and self-compassion. This points to the central role of the shared component of mantra meditation in enhancing the processing of bodily and emotional stimuli. Nevertheless, adding other yoga components had a favorable effect on daily affectivity of the participants. Most changes persisted until follow-up, even though participants did not continue to practice. More research is needed to substantiate our findings. [fig] Figure 1: Body awareness scores in four conditions during baseline and treatment phases for each participant with regression lines for each phase. Note. MA = Mantra meditation only; ME = meditation and ethical education; MY = meditation and physical yoga; MYE = meditation, physical yoga, and ethical education. [/fig] [fig] Figure 2: Box plots for averaged Tau-U body awareness effect size estimates in each condition. Individual estimates are scattered across the box plots. Note. MA = Mantra meditation only (dots); ME = meditation and ethical education (triangles); MY = meditation and physical yoga (squares); MYE = meditation, physical yoga, and ethical education (pluses). Whiskers represent the largest and lowest values within a distance of 1.5 times the interquartile range. According to Figure 2, all conditions led to an improvement in body awareness with the greatest enhancement in the MY condition (Mdn = 0.49, interquartile range [IQR] = 0.32), followed by the ME condition (Mdn = 0.43, IQR = 0.47) and the MYE condition (Mdn = 0.38, IQR = 0.32). The MA condition had the smallest effect (Mdn = 0.21, IQR = 0.71). When we estimated the total effect of group differences in multiple regression analysis, we found no differences between the conditions, F(3, 33) = 0.52, p = 0.675. None of the conditions or components nor practice time, age, gender, occupation, or baseline significantly predicted changes in body awareness over time. Multiple R 2 was 0.13 and 0.11 for the condition and component model, respectively. In the multilevel model estimating the overall effects, there was a significant effect of time, F(1, 677) = 10.23, p = 0.001, but no significant cross-level interaction, F(3, 677) = 0.47, p = 0.702. Both models predicted a significant enhancement of body awareness over time, β = 0.15, SE = 0.05, p = 0.001 (condition and component model). Nevertheless, there were no other significant effects and no crosslevel interactions. The effect sizes for both models were marginal R 2 = 0.07 and conditional R 2 = 0.86. All regression tables can be found in the Supplementary Material- [/fig] [fig] Figure 3: displays how the difficulties in emotion regulation developed over time for each participant. Similar to body awareness, emotion regulation did not significantly [/fig] [fig] Figure 4: depicts box plots with effect size estimates grouped by condition. Int. J. Environ. Res. Public Health 2022, 19, [/fig] [fig] Figure 7: displays the grouped effect sizes of all four treatments on both variables. The average effect sizes for arousal were close to zero and did not differ significantly between conditions-MA (Mdn = 0.07, IQR = 0.18), ME (Mdn = 0.05, IQR = 0.39), MY (Mdn = −0.02, IQR = 0.11), and MYE (Mdn = −0.09, IQR = 0.42). However, there was a discernible difference in valence effect sizes between the conditions MA (Mdn = −0.14, IQR = 0.31) and ME (Mdn = 0.10, IQR = 0.37). The average effect size for valence in the other two conditions was close to zero-MY (Mdn = −0.00, IQR = 0.18) and MYE (Mdn = 0.03, IQR = 0.28) [/fig] [fig] Figure 5: Valence scores in four conditions during baseline and treatment phases for each participant with regression lines for each phase. Note. Thin horizontal line represents a neutral valence (value = 5); MA = mantra meditation only; ME = meditation and ethical education; MY = meditation and physical yoga; MYE = meditation, physical yoga, and ethical education. [/fig] [fig] Figure 6: Arousal scores in four conditions during baseline and treatment phases for each participant with regression lines for each phase. Note. Thin horizontal line represents a neutral arousal (value = 5); MA = mantra meditation only; ME = meditation and ethical education; MY = meditation and physical yoga; MYE = meditation, physical yoga, and ethical education. [/fig] [fig] Figure 8: Stacked bar chart of Tau-U valence and arousal estimates with qualitatively classified affective responses per participant. Note. Numbers denote participants; MA = mantra meditation only; ME = meditation and ethical education; MY = meditation and physical yoga; MYE = meditation, physical yoga, and ethical education. [/fig]

Biochemical characterization of the xylan hydrolysis profile of the extracellular endo-xylanase from Geobacillus thermodenitrificans T12 Background: Endo-xylanases are essential in degrading hemicellulose of various lignocellulosic substrates. Hemicellulose degradation by Geobacillus spp. is facilitated by the hemicellulose utilization (HUS) locus that is present in most strains belonging to this genus. As part of the HUS locus, the xynA gene encoding an extracellular endo-xylanase is one of the few secreted enzymes and considered to be the key enzyme to initiate hemicellulose degradation. Several Geobacillus endo-xylanases have been characterized for their optimum temperature, optimum pH and generation of degradation products. However, these analyses provide limited details on the mode of action of the enzymes towards various substrates resulting in a lack of understanding about their hydrolytic potential. Results: A HUS-locus associated gene (GtxynA1) from the thermophile Geobacillus thermodenitrificans T12 encodes an extracellular endo-xylanase that belongs to the family 10 glycoside hydrolases (GH10). The GtxynA1 gene was cloned and expressed in Escherichia coli. The resulting endo-xylanase (termed GtXynA1) was purified to homogeneity and showed activity between 40°C and 80°C, with an optimum activity at 60°C, while being active between pH 3.0 to 9.0 with an optimum at pH 6.0. Its thermal stability was high and GtXynA1 showed 85% residual activity after 1 h of incubation at 60°C. Highest activity was towards wheat arabinoxylan (WAX), beechwood xylan (BeWX) and birchwood xylan (BiWX). GtXynA1 is able to degrade WAX and BeWX producing mainly xylobiose and xylotriose. To determine its mode of action, we compared the hydrolysis products generated by GtXynA1 with those from the well-characterized GH10 endo-xylanase produced from Aspergillus awamori (AaXynA). The main difference in the mode of action between GtXynA1 and AaXynA on WAX is that GtXynA1 is less hindered by arabinosyl substituents and can therefore release shorter oligosaccharides. Conclusions: The G. thermodenitrificans T12 endo-xylanase, GtXynA1, shows temperature tolerance up to 80°C and high activity to a variety of xylans. The mode of action of GtXynA1 reveals that arabinose substituents do not hamper substrate degradation by GtXynA1. The extensive hydrolysis of branched xylans makes this enzyme particularly suited for the conversion of a broad range of lignocellulosic substrates. # Background Many Geobacillus species contain a hemicellulose utilization (HUS) locus in their genome that encodes for multiple hydrolytic enzymes and sugar transporters responsible for the degradation of hemicellulose [bib_ref] A two-component system regulates the expression of an ABC transporter for xylo-oligosaccharides..., Shulami [/bib_ref] [bib_ref] Comparative analysis of the Geobacillus hemicellulose utilization locus reveals a highly variable..., De Maayer [/bib_ref]. The most well characterized HUS locus belongs to G. stearothermophilus T-6 and its proposed degradation mechanism is based on a limited number of secreted hydrolytic enzymes and further processing of the generated degradation products intracellularly [bib_ref] Multiple regulatory mechanisms control the expression of the Geobacillus stearothermophilus gene for..., Shulami [/bib_ref]. One of these secreted hydrolytic enzymes is an endo-xylanase that cleaves the xylan backbone of hemicellulose into shorter oligosaccharides [bib_ref] Multiple regulatory mechanisms control the expression of the Geobacillus stearothermophilus gene for..., Shulami [/bib_ref] [bib_ref] Purification and characterization of a thermostable xylanase from Bacillus stearothermophilus T-6, Khasin [/bib_ref]. The xylan backbone is composed of xylosyl residues linked by ß-(1 → 4)-glycosidic bonds, and is typically decorated with a varying degree of substitutions containing (4-O-methyl)-α-D-glucuronic acids, α-L-arabinofuranosyl residues and acetyl residues. The degree of substitution and polymerization is dependent on the origin of the substrate. Substrates from grasses and annual plants are rich in arabinoxylans, whereas hardwoods mostly consist of (4-O-methyl)-α-D-glucuronoxylan. This variety of substitutions requires multiple enzymes acting in synergy to fully degrade the xylans. Endo-(1 → 4)-ß-D-xylanases (E.C. 3.2.1.8) hydrolyse the ß-(1 → 4)-xylosidic bonds to release xylose and (substituted) xylo-oligosaccharides of various lengths that can be degraded further by other hemicellulases. Endo-xylanases belonging to glycoside hydrolase family 10 and 11 are the best characterized and described [bib_ref] Microbial hemicellulases, Shallom [/bib_ref]. Endo-xylanases belonging to family 10 (GH10) can tolerate the presence of substituents, such as arabinosyl residues, better than those belonging to glycoside hydrolase family 11 (GH11) [bib_ref] Impact and efficiency of GH10 and GH11 thermostable endoxylanases on wheat bran..., Beaugrand [/bib_ref]. An example of the mode of action of two endo-xylanases, belonging to families GH10 and GH11, from Aspergillus awamori (Aa) towards an arabinoxylan model substrate has been described [bib_ref] Mode of action of the xylan-degrading enzymes from Aspergillus awamori on alkali-extractable..., Kormelink [/bib_ref]. The AaGH10 was found to be able to cleave the xylan backbone next to a xylosyl residue substituted with one or two arabinosyl residues. In contrast, xylanases from family GH11 are hindered by arabinosyl substituents and cleave the xylan backbone only next to an unsubstituted xylosyl residue, producing longer oligosaccharides in comparison to those obtained from GH10 [bib_ref] Mode of action of the xylan-degrading enzymes from Aspergillus awamori on alkali-extractable..., Kormelink [/bib_ref]. Endo-xylanases from Geobacillus species belong to the GH10 family and have been described in multiple studies [bib_ref] Novel thermostable endo-xylanase cloned and expressed from bacterium Geobacillus sp. WSUCF1, Bhalla [/bib_ref] [bib_ref] Cloning, purification and characterization of an alkali-stable endoxylanase from thermophilic Geobacillus sp...., Canakci [/bib_ref] [bib_ref] Cloning, purification and characterization of a cellulase-free xylanase from Geobacillus thermodenitrificans AK53, Irfan [/bib_ref] [bib_ref] Characterization of a recombinant thermostable xylanase from hot spring thermophilic Geobacillus sp...., Liu [/bib_ref] [bib_ref] Properties of an alkali-thermo stable xylanase from Geobacillus thermodenitrificans A333 and applicability..., Marcolongo [/bib_ref] [bib_ref] Cloning, overexpression, and characterization of a novel alkali-thermostable xylanase from Geobacillus sp, Mitra [/bib_ref] [bib_ref] Thermostable and alkalistable endoxylanase of the extremely thermophilic bacterium Geobacillus thermodenitrificans TSAA1:..., Verma [/bib_ref]. These studies mainly focused on the characterization of activity ranges at higher temperatures and thermo-stability but lack detailed analyses of the mode of action of the enzyme on its substrate. Although some studies have described xylo-oligosaccharides formation upon hydrolysis of xylan by the endoxylanase, the position of hydrolysis and the possible hindrance of substituents present on the xylan backbone have not been characterized [bib_ref] Novel thermostable endo-xylanase cloned and expressed from bacterium Geobacillus sp. WSUCF1, Bhalla [/bib_ref] [bib_ref] Cloning, purification and characterization of a cellulase-free xylanase from Geobacillus thermodenitrificans AK53, Irfan [/bib_ref] [bib_ref] Properties of an alkali-thermo stable xylanase from Geobacillus thermodenitrificans A333 and applicability..., Marcolongo [/bib_ref] [bib_ref] Thermostable and alkalistable endoxylanase of the extremely thermophilic bacterium Geobacillus thermodenitrificans TSAA1:..., Verma [/bib_ref]. In this paper, we describe the cloning and overproduction in E.coli of GtXynA1, a GH10 family endo-xylanase isolated from Geobacillus thermodenitrificans. We also determined the enzyme's mode of action by identifying the generated hydrolysis products and propose a model for the degradation of substituted xylans. # Methods ## Media, strains and cultivation methods Wheat arabinoxylan (WAX) was of medium viscosity and supplied by Megazyme (Wicklow, Ireland). The chemical composition of this type of WAX has previously been described [bib_ref] Screening for distinct xylan degrading enzymes in complex shake flask fermentation supernatants, Van Gool [/bib_ref]. Beechwood xylan (BeWX) and all chemicals were purchased from Sigma-Aldrich (St Louis, MO, USA) unless otherwise specified. AaXynA was kindly provided by the laboratory of Food Chemistry (Wageningen University) as described elsewhere [bib_ref] Mode of action of the xylan-degrading enzymes from Aspergillus awamori on alkali-extractable..., Kormelink [/bib_ref]. Luria-Bertani (LB) medium contained per L: 10 g NaCl, 10 g tryptone (Oxoid), 5 g yeast extract (Roth). LB2 contains per liter: 10 g tryptone (Oxoid), 5 g yeast extract (Roth), 10 g sodium chloride and salts mix consisting of 1 g NH 4 Cl; 3 g NaCl; 1.50 g Na 2 SO 4 ; 0.08 g NaHCO 3 ; 1 g KCl; 1.8 g MgCl 2 × 6H 2 O; 0.30 g CaCl 2 × 2H 2 O. pH was set to 6.6 at room temperature and the medium was autoclaved for 20 min at 121°C, after which 1 mL K 2 HPO 4 (250 g/L) was added. For plasmid propagation and protein expression E. coli strains DH5α and BL21(DE3) were used respectively. E. coli strains were grown in LB medium at 37°C, pH 7.0 with an agitation speed of 150 RPM. Geobacillus thermodenitrificans T12 was isolated from compost [bib_ref] Isolation of a genetically accessible thermophilic xylan degrading bacterium from compost, Daas [/bib_ref]. Strain T12 was grown in LB2 medium at 65°C, 150 RPM and pH 7.0. ## Cloning of gtxyna1 Genomic DNA of strain T12 was obtained using the MasterPure™ Gram Positive DNA Purification Kit (Epicentre) according to manufacturer's protocol. The XynA coding sequence was cloned without the signal peptide, which was determined by SignalP4.1 to consist of the first 28 amino acids of the protein. Primers used for amplification of the xynA gene (GenBank accession number KX962565) were XynA-FW (5'-GCGCTCATG AAAACTGAACAATCATACGCTAAAAAG-3') and Xy nA-RV (5'-GCGCCTCGAGCTTATGATCGATAATAGC CCAATACG-3') that contain the BspHI and XhoI restriction sites, respectively. PCR amplification was performed on 50 ng of genomic DNA using PhusionHF DNA polymerase (Thermo Scientific) in the following conditions; 1 cycle 98°C 5 min., 30 cycles of 98°C 30 s, 50°C 30 s and 72°C 45 s followed by a final elongation step at 72°C for 5 min. For enzyme characterization studies GtXynA1 was cloned into pET24d(+) using the restriction sites of BspHI and XhoI. The reverse primer was designed in such a way that the stop codon was removed to include the His-tag from the pET24d vector. Plasmids were purified from E. coli DH5α using the GeneJET™ Plasmid Miniprep Kit (Thermo Scientific) and subsequently transformed into E. coli BL21(DE3). ## Overproduction and purification of gtxyna1 Protein production was induced by adding 0.1 mM IPTG to a 250 mL culture at the start of inoculation. Culture was provided with 50 μg/mL kanamycin as antibiotics and incubated O/N at 37°C at an agitation speed of 150 RPM. Cells were then centrifuged (4000×g, 15 min, 4°C) and cell pellet was resuspended in 25 mL phosphate buffer at a pH of 7.4. Cell suspension was disrupted using French Press and cell free extract (CFE) was obtained by centrifugation (30,000×g, 20 min, 4°C). Filtered CFE was applied to a HisPrep FF 16/10 Ni-NTA Sepharose column (GE Healthcare) equilibrated with 50 mM sodium phosphate buffer (pH 7.4) containing 300 mM NaCl. Bound GtXynA1 was eluted using a linear gradient of 0-500 mM imidazole in the same buffer. Purified GtXynA1 was then desalted by applying the protein solution to a HiPrep 26/10 desalting column (GE Healthcare) equilibrated with 50 mM sodium phosphate buffer (pH 7.4). Recombinant xylanase was eluted at a rate of 5 mL/min and stored at−20°C for further research. Protein purity was assayed by SDS-PAGE analysis using a 10% (w/v) polyacrylamide gel according to the method of [bib_ref] Cleavage of structural proteins during the assembly of the head of bacteriophage..., Laemmli [/bib_ref]. A PageRuler™ protein ladder was used and bands were visualized using PageBlue™ Protein Staining Solution (Thermo Scientific). ## Enzyme activity assays Wheat arabinoxylan (WAX), Beechwood xylan (BeWX) and Birchwood xylan (BiWX) were incubated with the purified GtXynA1 in 10 mM NaOAc buffer, pH 6.0 (1 mL, 10 mg substrate dry matter) at 60°C for 24 h. The enzymes were dosed at 0.1% (w/w) protein based on substrate dry matter. Next, 2 μl of 4 M HCl was added to stop the enzyme activity and the sample was centrifuged (10,000×g, 10 min, 10°C) prior to analysis. The pH and temperature optima were tested by incubating WAX with GtXynA1 in a range of pH 3 to 9. For the pH range of 3 to 8 we used 50 mM sodium citrate buffer and for the pH range of 8 to 10 we used 200 mM Tris-HCl buffer. Buffers were adjusted for correct pH at the temperature of incubation. Optimum temperature was determined in a range from 30 to 100°C. Incubation time, enzyme dosage and stopping the reaction were as described above. All samples collected were submitted to the PAHBAHassay, HPSEC, HPAEC and RP-UHPLC-UV-MS. ## Analysis of hydrolysis products The supernatants obtained were analysed for the amount of reducing ends present by the PAHBAH reducing assay in duplicate [bib_ref] A new reaction for colorimetric determination of carbohydrates, Lever [/bib_ref]. ## Hpsec High performance size exclusion chromatography (HPSEC) of WAX and BiWX before and after incubation with GtXynA1 was performed on an Ultimate 3000 HPLC system (Thermo Scientific, Sunnyvale, CA, USA) equipped with a set of three TSK-gel columns (6.0 mm × 15.0 cm per column) in series (SuperAW4000, SuperAW3000, SuperAW25000, Tosoh Bioscience, Stuttgart, Germany) in combination with a PWX-guard column (Tosoh Bioscience). HPSEC was controlled by the Chromeleon software (Thermo Scientific). Elution took place at 40°C with 0.2 M sodium nitrate at a flow rate of 0.6 mL/min. The eluate was monitored using a refractive index (RI) detector (Shoko Scientific Co., Yokohama, Japan). Calibration was made by using pullulan series (Polymer Laboratories, Union, NY, USA) with a molecular weight in the range of 0.18-788 kDa. ## Hpaec Oligosaccharides and monosaccharides released after enzymatic incubation of WAX and BeWX with GtXynA1 were analysed by HPAEC as described elsewhere [bib_ref] Importance of acid or alkali concentration on the removal of xylan and..., Martínez [/bib_ref]. ## Rp-uhplc-uv-ms Xylo-oligosaccharides produced by the enzymes were 2-AA labelled and analysed by RP-UHPLC-UV-MS as previously described by Murciano Martinez et al. [bib_ref] Importance of acid or alkali concentration on the removal of xylan and..., Martínez [/bib_ref]. # Results In this study, we cloned and overexpressed the Geobacillus thermodenitrificans T12 xynA1 gene in E.coli and determined the mode of action of the GtXynA1 xylanase enzyme on wheat arabinoxylan (WAX), beechwood xylan (BeWX) and birchwood xylan (BiWX). The xynA gene is 1224 bp long and encodes a protein of 407 amino acid residues, which belongs to GH family 10. The amino acid sequence contains a signal peptide of 28 amino acids and two catalytic residues were predicted at positions H264 and D295. Phylogenetic analysis revealed high amino acid identity to several Geobacillus endo-xylanases [fig_ref] Table 1: Sequence identity of several characterized endoxylanases to the G [/fig_ref]. ## Cloning, expression and characterization of gtxyna1 The gene coding for GtXynA1 (1224 bp) was PCR-amplified, cloned into the pET24d vector and successfully expressed in E.coli BL21(DE3). The endo-xylanase was purified to apparent homogeneity, as it appeared as a 50 kDa single band on SDS-PAGE after visualization with PageBlue stain (Additional file 1: [fig_ref] Figure 1: pH [/fig_ref] ; Additional file 2: . The purified endoxylanse GtXynA1 showed highest activity at pH 6.0 in a broad activity range from pH 4.0 to 9.0 [fig_ref] Figure 1: pH [/fig_ref]. GtXynA1 showed activity between 40°C to 80°C with an optimum at 60°C [fig_ref] Figure 1: pH [/fig_ref]. The thermostability of GtXynA1 was tested in the range of 30°C to 80°C. After 1 h at 50 and 60°C the enzyme retained 90 and 45% residual activity, respectively, whereas the residual activity after 1 h at 70°C was only 18% . ## Mode of action Activity of GtXynA1 towards wheat arabinoxylan (WAX), birchwood xylan (BiWX), and beechwood xylan (BeWX) was determined as outlined in materials and methods. Degradation profiles of BiWX and BeWX were similar (data not shown), therefore only the BeWX degradation patterns will be further explained. WAX contains β-(1 → 4) linked xylosyl residues substituted with α-(1 → 2) linked arabinosyl residues or with both α-(1 → 2) and α-(1 → 3) linked arabinosyl residues. BiWX and BeWX contain β-(1 → 4) linked xylosyl residues substituted with α-(1 → 2) linked 4-O-methylglucuronic acid (UA me ). Both WAX and BeWX were incubated with GtXynA1 at its optimum pH and temperature for 24 h . Both WAX and BeWX were degraded to oligomers and monomers, as deduced from the disappearance of soluble polymers (WAX around 55 kDa, BeWX around 12 kDa; and the increase of lower molecular weight products of around 1 kDa after 24 h incubation. The difference in molecular weight of the polymeric structures analysed at 0 h incubation is because WAX is completely soluble and BeWX is only partly soluble [bib_ref] Two GH10 endo-xylanases from Myceliophthora thermophila C1 with and without cellulose binding..., Van Gool [/bib_ref]. GtXynA1 incubated with WAX and BeWX released a series of linear and substituted xylo-oligosaccharides that were identified based on the known HPAEC elution of WAX digested with the well-characterized GH10 endoxylanase from Aspergillus awamori (AaXynA) [bib_ref] Mode of action of the xylan-degrading enzymes from Aspergillus awamori on alkali-extractable..., Kormelink [/bib_ref]. The oligomeric profile released by GtXynA1 is comparable to the profile produced by AaXynA, supporting that GtXynA1 is a GH10 endo-xylanase [fig_ref] Figure 4: HPAEC elution pattern of with AaXynA and GtXynA1 incubated with beech wood... [/fig_ref]. Only one substituted oligosaccharide was produced from BeWX, eluting at 16 min [fig_ref] Figure 4: HPAEC elution pattern of with AaXynA and GtXynA1 incubated with beech wood... [/fig_ref] , which was identified as a xylotriose substituted by 4-O-methylglucuronic acid located at the non-reducing end [fig_ref] Figure 4: HPAEC elution pattern of with AaXynA and GtXynA1 incubated with beech wood... [/fig_ref] after 2-AA labelling and analysis by UPLC-UV-MS. The UV chromatogram at 340 nm of 2-AA labelled degradation products from BeWX and the full MS spectra overlapped. The two predominant masses seen in the full MS spectra were 724 (m/z) and 666 (m/z) (data not shown), and they overlapped with UV chromatogram peaks eluting at 13 and 16.8 min, respectively. [fig_ref] Figure 4: HPAEC elution pattern of with AaXynA and GtXynA1 incubated with beech wood... [/fig_ref] also shows that GtXynA1 is able to degrade WAX and BeWX to smaller-sized linear xylo-oligosaccharides, mainly xylose and xylobiose. This observation is substantiated on the integrated peak area of linear xylo-oligosaccharides [fig_ref] Table 2: HPAEC integrated peak area of linear xylo-oligosaccharides produced from WAX and BeWX... [/fig_ref]. On the contrary, AaXynA is able to degrade WAX and BeWX producing mainly xylobiose and xylotriose, but also xylotetraose, xylopentaose and xyloheaxaose [fig_ref] Table 2: HPAEC integrated peak area of linear xylo-oligosaccharides produced from WAX and BeWX... [/fig_ref]. When comparing the diversity of substituted xylooligosaccharides produced [fig_ref] Figure 4: HPAEC elution pattern of with AaXynA and GtXynA1 incubated with beech wood... [/fig_ref] , it is clear that GtXynA1 produces shorter oligomers from both WAX and BeWX compared to AaXynA. Substituted xylooligosaccharides are identified in a range of DP from 2 to 6 by both HPAEC [fig_ref] Figure 4: HPAEC elution pattern of with AaXynA and GtXynA1 incubated with beech wood... [/fig_ref] and MS (not shown), suggesting that AaXynA is less tolerant towards 4-Omethylglucuronic acid substituted xylan. Degradation products formed during the degradation of WAX by AaXynA that where eluting at 15, 17 and 25 min are absent in the profile of GtXynA1 incubated on WAX [fig_ref] Figure 4: HPAEC elution pattern of with AaXynA and GtXynA1 incubated with beech wood... [/fig_ref]. These xylo-oligosaccharides are characteristic because of the presence of two arabinosyl residues located at the O-2 and O-3 of the same xylosyl residue. Hence, GtXynA1 cleaves next to an adjacent non-substituted residue, from the non-reducing end when a single substituted residue is present. In case the xylosyl residue is doubly substituted by arabinosyl residues, GtXynA1 needs two adjacent unsubstituted residues at the non-reducing end. In case the cleavage takes place from the reducing end, the enzyme can cleave directly next to the substituted (single or doubly substituted) residue in most cases [fig_ref] Figure 1: pH [/fig_ref] , . However, when the substituted residue (single or doubly substituted), is followed by a doubly substituted xylosyl residue, the enzyme cleaves next to an unsubstituted residue [fig_ref] Figure 2, Figure 3: Temperature stability of GtXynA1 [/fig_ref]. On the contrary, the presence of more than one doubly or single substituted xylosyl residues obstructs the action of AaXynA. Based on the information provided by the HPAEC chromatograms and RP-UV-UHPLC-MS profile on the linear and substituted xylo-oligosaccharides produced by the enzymes, an schematic mode of action of GtXynA1 is proposed in [fig_ref] Figure 4: HPAEC elution pattern of with AaXynA and GtXynA1 incubated with beech wood... [/fig_ref] * and d*. # Discussion The ability of Geobacillus species to degrade xylan relies on the activity of the hemicellulose utilization (HUS) locus. The HUS-locus contains a variety of xylan degrading enzymes, sugar transporters and metabolic pathways for the complete utilization of (substituted) xylans. Xylan degradation initiates by the action of an extracellular endoxylanase (XynA1) that cleaves the xylan backbone, thereby producing (substituted) xylo-oligosaccharides. The linear oligosaccharides generated by the action of XynA1 can then enter the cell through an ABC-transporter that is encoded by the xynEFG operon whereas the substituted oligosaccharides enter the cell via other substrate specific ABC transporters [bib_ref] A two-component system regulates the expression of an ABC transporter for xylo-oligosaccharides..., Shulami [/bib_ref] [bib_ref] Multiple regulatory mechanisms control the expression of the Geobacillus stearothermophilus gene for..., Shulami [/bib_ref] [bib_ref] The L-arabinan utilization system of Geobacillus stearothermophilus, Shulami [/bib_ref]. In this research, we cloned and expressed GtXynA1, the endo-xylanase of G. thermodenitrificans T12 and analysed its mode of action. The purified endo-xylanase GtXynA1 showed highest activity at pH 6.0 in a broad activity range from pH 4.0 to 9.0 [fig_ref] Figure 1: pH [/fig_ref] , which is in line with previous reports on endo-xylanases from geobacilli [bib_ref] Characterization of a recombinant thermostable xylanase from hot spring thermophilic Geobacillus sp...., Liu [/bib_ref] [bib_ref] Thermostable and alkalistable endoxylanase of the extremely thermophilic bacterium Geobacillus thermodenitrificans TSAA1:..., Verma [/bib_ref]. GtXynA1 showed activity between 40°C to 80°C with an optimum at 60°C [fig_ref] Figure 1: pH [/fig_ref] where previous reports on Geobacillus endo-xylanases demonstrated optimum temperatures from 60°C up to 75°C [bib_ref] Characterization of a recombinant thermostable xylanase from hot spring thermophilic Geobacillus sp...., Liu [/bib_ref] [bib_ref] Properties of an alkali-thermo stable xylanase from Geobacillus thermodenitrificans A333 and applicability..., Marcolongo [/bib_ref] [bib_ref] Thermostable and alkalistable endoxylanase of the extremely thermophilic bacterium Geobacillus thermodenitrificans TSAA1:..., Verma [/bib_ref]. The thermostability of GtXynA1 was tested in the range of 30°C to 80°C. After 1 h at 50 and 60°C the enzyme retained 90 and 45% residual activity, respectively, whereas the residual activity after 1 h at 70°C was only 18% . The endo-xylanase characterized from Geobacillus stearothermophilus T-6 remains 90% active (1 h time point) after incubation at 70°C [bib_ref] Purification and characterization of a thermostable xylanase from Bacillus stearothermophilus T-6, Khasin [/bib_ref]. This higher temperature stability of the T-6 endo-xylanase may be caused by its production from the native strain instead of a recombinant expression system in E.coli. A lower thermostability of an endo-xylanase produced in E. coli compared to the same protein produced in Geobacillus has been reported for Geobacillus sp. TC-W7 [bib_ref] Characterization of a recombinant thermostable xylanase from hot spring thermophilic Geobacillus sp...., Liu [/bib_ref] , which produces an endo-xylanase highly similar to GtXynA1 [fig_ref] Table 1: Sequence identity of several characterized endoxylanases to the G [/fig_ref]. Alternatively, the difference in thermostability between GtXynA1 and the T-6 endoxylanase may be (partly) caused by differences in their amino acid composition. The GtXynA1 enzyme was more thermostable in comparison with other bacterial endo-xylanases such as XynC from Bacillus subtilis 168, that shows around 20% residual activity at 50°C and no residual activity at 60°C [bib_ref] Characterization of XynC from Bacillus subtilis subsp. subtilis strain 168 and analysis..., St [/bib_ref]. The thermostability of GtXynA1 has great advantage for the degradation of xylan at higher temperatures (60°C) over enzymes produced by most mesophilic organisms. In previous studies, G. stearothermophilus T-6 was grown on 4-O-methyl -D-glucurono-D-xylan and the main end products found were xylotriose substituted with 4-O-methyl-D-glucuronic acid together with xylobiose and xylose [bib_ref] The glucuronic acid utilization gene cluster from Bacillus stearothermophilus T-6, Shulami [/bib_ref]. Although we found the same end products in the present study, we also detected xylotriose and xylotetraose as minor end products [fig_ref] Figure 4: HPAEC elution pattern of with AaXynA and GtXynA1 incubated with beech wood... [/fig_ref]. Only one substituted oligosaccharide was produced, identified as a xylotriose substituted by 4-O-methylglucuronic acid located at the non-reducing end [fig_ref] Figure 4: HPAEC elution pattern of with AaXynA and GtXynA1 incubated with beech wood... [/fig_ref]. When comparing the mode of action on WAX, the main difference between GtXynA1 and AaXynA is that the former enzyme is less hindered by arabinosyl substituents. This is deduced from the difference in degradation products formed between the incubations of GtXynA1 and AaXynA on WAX. Similar end products were found in previous studies using TLC analysis but, no quantification of the linear end products and no identification of substituted end products was performed [bib_ref] Properties of an alkali-thermo stable xylanase from Geobacillus thermodenitrificans A333 and applicability..., Marcolongo [/bib_ref] [bib_ref] Characteristics of thermostable endoxylanase and β-xylosidase of the extremely thermophilic bacterium Geobacillus..., Anand [/bib_ref]. Formed xylo-oligosaccharides are taken up by the cell via specific sugar transporters [bib_ref] Multiple regulatory mechanisms control the expression of the Geobacillus stearothermophilus gene for..., Shulami [/bib_ref]. These transporters are encoded by the operon xynEFG, located in the HUS locus, and have a preference for short xylooligosaccharides within a DP of 2 to 6. More specifically, highest affinity is towards trisaccharides and for other xylo-oligosaccharides the affinity is as follows: X2 < X3 > X4 > X5 > X6 [bib_ref] A two-component system regulates the expression of an ABC transporter for xylo-oligosaccharides..., Shulami [/bib_ref]. The relative quantity of different xylooligosaccharides formed by the action of GtXynA1 on BeWX shows that mostly xylobiose and xylose is formed [fig_ref] Figure 5: Peak area ratio of linear end products after 24 h incubation on... [/fig_ref]. The discrepancy between end products formed by GtXynA1 and the affinity of the sugar transporter XynEFG is most likely caused by the 24 h incubation as GtXynA1 is also active towards xylotriose. Previous studies demonstrated the formation of xylotriose within 30 min of incubation and prolonged incubation resulted in a decrease of xylotriose and an increase in xylobiose and xylose [bib_ref] Properties of an alkali-thermo stable xylanase from Geobacillus thermodenitrificans A333 and applicability..., Marcolongo [/bib_ref] [bib_ref] Characteristics of thermostable endoxylanase and β-xylosidase of the extremely thermophilic bacterium Geobacillus..., Anand [/bib_ref]. The mode of action of GtXynA1 creates linear and branched xylo-oligosaccharides of which most are in the DP range of 2 to 6. These small branched oligosaccharides are believed to enter the cell via specific ABC-transporters [bib_ref] Multiple regulatory mechanisms control the expression of the Geobacillus stearothermophilus gene for..., Shulami [/bib_ref]. For most of these transporters their affinity towards linear oligosaccharides was determined [bib_ref] A two-component system regulates the expression of an ABC transporter for xylo-oligosaccharides..., Shulami [/bib_ref] [bib_ref] The L-arabinan utilization system of Geobacillus stearothermophilus, Shulami [/bib_ref] [bib_ref] The glucuronic acid utilization gene cluster from Bacillus stearothermophilus T-6, Shulami [/bib_ref]. However, for the arabino-oligosaccharide transporter AbnEFJ it was shown that the binding constants of branched oligosaccharides where 2 orders of magnitude higher than those obtained for linear oligosaccharides [bib_ref] The L-arabinan utilization system of Geobacillus stearothermophilus, Shulami [/bib_ref]. It is likely that also other ABC transporters that are involved in the uptake of oligosaccharides have higher affinity for branched oligosaccharides in comparison to linear oligosaccharides. The possibility of transporting branched oligosaccharides enables the Geobacillus species to use natural substrates in a rapid and efficient way. [fig] Figure 1: pH (a) and temperature (b) profiles of the incubated GtXynA1 with wheat arabinoxylan (WAX) [/fig] [fig] Figure 2, Figure 3: Temperature stability of GtXynA1. The temperature stability was determined by incubating GtXynA1 at temperatures between 30 and 80°C for 1 h, 6 h, and 24 h after which the residual activity was measured against wheat arabinoxylan at 60°C and a pH of 6Molecular weight distributions of wheat arabinoxylan (WAX) (a) and beechwood xylan (BeWX) (b) incubated with GtXynA1 for 0, 1, 6 and 24 h [/fig] [fig] Figure 4: HPAEC elution pattern of with AaXynA and GtXynA1 incubated with beech wood xylan (BeWX) (a, b respectively) and wheat arabinoxylan (WAX) (c, d respectively) for 24 h, and respective cleavage pattern of GtXynA1 on BeWX (b*) and WAX (d*). • = xylosyl residue, ◊ = arabinosyl residue, ▲ = 4-O-methylglucuronic acid. (b) shows the identification of the main substituted xylo-oligosaccharide from B. The structure is identify based on its UV signal and the MS 2 fragmentation pattern [/fig] [fig] Figure 5: Peak area ratio of linear end products after 24 h incubation on beech wood xylan in comparison to oligosaccharide transporter (xynEFG) affinity. Green: ratio of affinity of the XynEFG oligosaccharide transporter of G. stearothermophilus towards different oligosaccharides. Red: ratio of the end product area after a 24 h incubation of AaXynA on beechwood xylan. Blue: ratio of the end product area after a 24 h incubation of GtXynA1 on beechwood xylan [/fig] [table] Table 1: Sequence identity of several characterized endoxylanases to the G. thermodenitrificans T12 endo-xylanase [/table] [table] Table 2: HPAEC integrated peak area of linear xylo-oligosaccharides produced from WAX and BeWX after 24 h incubation with GtXynA1 and AaXynA [/table]

Prevalence of peritonitis and mortality in patients treated with continuous ambulatory peritoneal dialysis (CAPD) in Africa: a protocol for a systematic review and meta-analysis ## Prevalence of peritonitis and mortality in patients treated with continuous ambulatory peritoneal dialysis (capd) in africa: a protocol for a systematic review and meta-analysis Mothusi Walter Moloi, 1,2,3 Shepherd Kajawo, [bib_ref] Peritoneal dialysis in developing countries, Nayak [/bib_ref] [bib_ref] Peritoneal dialysis in Cape Town, South Africa, Okpechi [/bib_ref] [bib_ref] Peritoneal dialysis in Africa, Abu-Aisha [/bib_ref] Jean Jacques Noubiap, [bib_ref] Continuous ambulatory peritoneal dialysis: an option in the developing world?, Zent [/bib_ref] Ikechukwu O Mbah, [bib_ref] An African community-based chronic ambulatory peritoneal dialysis programme, Katz [/bib_ref] Udeme Ekrikpo, [bib_ref] Peritoneal dialysis in developing countries, Nayak [/bib_ref] [bib_ref] Peritoneal dialysis in Cape Town, South Africa, Okpechi [/bib_ref] [bib_ref] Peritoneal dialysis-related peritonitis: twenty-seven years of experience in a Colombian medical center, Nieto-Ríos [/bib_ref] Andre Pascal Kengne, [bib_ref] Continuous ambulatory peritoneal dialysis: an option in the developing world?, Zent [/bib_ref] [bib_ref] ISPD peritonitis recommendations: 2016 update on prevention and treatment, Li [/bib_ref] Aminu K Bello, [bib_ref] Peritoneal dialysis associated infections: An update on diagnosis and management, Akoh [/bib_ref] Ikechi G Okpechi [bib_ref] Peritoneal dialysis in developing countries, Nayak [/bib_ref] [bib_ref] Peritoneal dialysis in Cape Town, South Africa, Okpechi [/bib_ref] To cite: Introduction Continuous ambulatory peritoneal dialysis (CAPD) is the ideal modality for renal replacement therapy in most African settings given that it is relatively cheaper than haemodialysis (HD) and does not require in-centre care. CAPD is, however, not readily utilised as it is often complicated by peritonitis leading to high rates of technique failure. The objective of this study is to assess the prevalence of CAPD-related peritonitis and all-cause mortality in patients treated with CAPD in Africa. Methods and analysis We will search PubMed, EMBASE, SCOPUS, Africa Journal Online and Google Scholar for studies conducted in Africa from 1 January 1980 to 30 June 2017 with no language restrictions. Eligible studies will include cross-sectional, prospective observational and cohort studies of patients treated with CAPD. Two authors will independently screen, select studies, extract data and conduct risk of bias assessment. Data consistently reported across studies will be pooled using randomeffects meta-analysis. Heterogeneity will be evaluated using Cochrane's Q statistic and quantified using I 2 statistics. Graphical and formal statistical tests will be used to assess for publication bias. Ethics and dissemination Ethical approval will not be needed for this study as data used will be extracted from already published studies. Results of this review will be published in a peer-reviewed journal and presented at conferences. The Preferred Reporting Items for Systematic reviews and Meta-Analyses for Protocols 2015 (PRISMA-P 2015) framework guided the development of this protocol. PrOsPErO registration number CRD42017072966. # Introduction Continuous ambulatory peritoneal dialysis (CAPD) ought to be the ideal modality for renal replacement therapy in most African settings, as it provides superior rehabilitation, better quality of life, averts need for expensive machines and allows for home-based therapy especially for patients in rural areas where immense distance from haemodialysis units exists. [bib_ref] Peritoneal dialysis in developing countries, Nayak [/bib_ref] Africa is home to 70% of the least developed countries in the world and has a high burden of patients with chronic kidney disease (CKD) with reported prevalence as high as 13.9%. In 2007, Africa's overall dialysis population accounted for 4.5% of the global dialysis population, with a total dialysis prevalence of 74 per million population (pmp), compared with a global average of 250 pmp. [bib_ref] Peritoneal dialysis in Africa, Abu-Aisha [/bib_ref] During the same period, the prevalence of CAPD in Africa was 2.2 pmp relative to global prevalence of 27 pmp, while 85% of all CAPD patients in Africa resided in South Africa. [bib_ref] Peritoneal dialysis in Africa, Abu-Aisha [/bib_ref] Challenges associated with utilisation of CAPD in Africa includes high cost of CAPD fluids (given that they are mostly imported), lack of expertise on utilisation of CAPD and several sociodemographic factors (including poor housing, overcrowding in houses, poor literacy, inadequate water and electricity strengths and limitations of this study ► To the best of our knowledge, there is no systematic review on the prevalence of continuous ambulatory peritoneal dialysis (CAPD)-related peritonitis in Africa, and this review will assist to fill the gap that exists in the literature. ► The findings of this review will aid policy and guidelines for the management of patients with CAPDrelated peritonitis. ► We will conduct a meta-analysis from the studies included. ► Limitation to this study is the low utilisation of CAPD in Africa and therefore few published studies are included in the analysis. ► Heterogeneity of studies carried out on the topic in Africa will be another limitation of this study. Open Access supplies) that have been associated with peritonitis from various studies. Peritonitis remains the most common complication of CAPD in treated patients with end-stage renal disease (ESRD). [bib_ref] An African community-based chronic ambulatory peritoneal dialysis programme, Katz [/bib_ref] [bib_ref] Peritoneal dialysis-related peritonitis: twenty-seven years of experience in a Colombian medical center, Nieto-Ríos [/bib_ref] [bib_ref] ISPD peritonitis recommendations: 2016 update on prevention and treatment, Li [/bib_ref] Globally, CAPD-related peritonitis rates are estimated to be 0.24-1.66 episodes per patient-year, [bib_ref] Peritoneal dialysis associated infections: An update on diagnosis and management, Akoh [/bib_ref] exceeding, in many low-income and middle-income countries, the guidelines recommendation of not more than 0.5 episodes per patient-year. [bib_ref] ISPD peritonitis recommendations: 2016 update on prevention and treatment, Li [/bib_ref] Although developed countries have observed a gradual decline in the prevalence of CAPD-related peritonitis due to improvements in the CAPD techniques, the same trends have not been reported in low-income and middle-income countries. [bib_ref] 32 years' experience of peritoneal dialysis-related peritonitis in a university hospital, Van Esch [/bib_ref] Various sociodemographic factors, including availability of electricity and tap water, number of occupants at home, 5 level of education, distance from the dialysis units 11 as well as clinical and biochemical factors [bib_ref] Peritonitis in children on peritoneal dialysis in Cape Town, South Africa: epidemiology..., Raaijmakers [/bib_ref] have been associated with CAPD-related peritonitis in Africans. CAPD-related peritonitis is independently associated with higher risk of all-cause, cardiovascular and infection-related mortality and this risk increases with prolonged use of CAPD. 14 It is estimated that for every 0.5 per-year increase in peritonitis rate, the risk of death increases by 4%-18%. [bib_ref] Peritoneal dialysis-related peritonitis: twenty-seven years of experience in a Colombian medical center, Nieto-Ríos [/bib_ref] In one study, [bib_ref] Peritonitis-related mortality in patients undergoing chronic peritoneal dialysis, Pérez Fontan [/bib_ref] although the rate of peritonitis was found to be 0.60 episodes per patientyears, mortality among CAPD patients was reported to be 5.9% while peritonitis was directly implicated in 15.2% of all deaths, and non-peritonitis-related infections caused 68.5% of deaths. [bib_ref] Peritonitis-related mortality in patients undergoing chronic peritoneal dialysis, Pérez Fontan [/bib_ref] Another study from China showed a much higher mortality rate (19.8%) despite reporting a low rate of peritonitis (0.16 episodes per patient-year). [bib_ref] The impact of peritoneal dialysisrelated peritonitis on mortality in peritoneal dialysis patients, Ye [/bib_ref] In the same study, peritonitis was independently associated with a 95% increase in all-cause mortality and up to fourfold risk of infection-related mortality. 14 Abu-Aisha et al from Sudan reported a peritonitis rate of 0.55 episodes per patient-year with an overall mortality 10%, of which 16.7% were related to peritonitis. [bib_ref] Rates and causes of peritonitis in a National Multicenter Continuous Ambulatory Peritoneal..., Abu-Aisha [/bib_ref] Although these studies show that the prevalence of CAPD-related peritonitis is high and that overall mortality and peritonitis-related mortality is high in these patients, prevalence of peritonitis and mortality in African patients with ESRD treated with CAPD has not been adequately characterised. The purpose of this review is therefore to determine the prevalence of peritoneal dialysis-associated peritonitis as well as mortality reported in CAPD patients in Africa. ## Objective To conduct a systematic review and meta-analysis to estimate the prevalence of peritonitis and mortality in African patients with ESRD treated with CAPD. ## Review questions This review aims to answer the following questions: 1. What is the prevalence of peritonitis among patients with ESRD treated with CAPD in Africa? 2. What is the mortality rate among patients treated with CAPD and in those who develop CAPD-related peritonitis? 3. What are the causative organisms for PD peritonitis? 4. What are the rates of modality switch to haemodialysis in patients with ESRD treated with CAPD who develop CAPD-related peritonitis? # Methods and analysis The Preferred Reporting Items for Systematic reviews and Meta-Analyses for Protocols 2015 (PRISMA-P 2015) was OR "aspergillus niger"[tw]))))))))))))))) #4 #1+ #2 + #3 #5 Limits #4: Humans, 1980/01/01 to 2017/06/30 Open Access used to develop the current protocol. [bib_ref] Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015 statement, Moher [/bib_ref] PROSPERO registration number: CRD42017072966. ## Inclusion criteria All human studies meeting the criteria listed below will be included: 1. Observational studies (cross-sectional studies, prospective observational studies and cohort studies) and letters to editors reporting on the prevalence of peritonitis and all-cause mortality among children and adults with ESRD treated with CAPD and residing in Africa. 2. Age: there will be no age restriction. 3. Duration: we will consider data from all studies published from the 1 January 1980 to 30 June 2017. 4. Language restrictions: there will be no language restriction. 5. Sample size restrictions: there will be no sample size restriction. ## Exclusion criteria The following types of studies will not be considered: 1. Case reports, case-control studies, randomised control trials and review articles. 2. Studies conducted among populations of African origin residing outside of Africa. 3. Duplicates: for studies published in more than one paper, the most comprehensive one reporting the largest sample size will be considered. 4. Studies from which the prevalence of PD-related peritonitis cannot be determined and whose full data could not be accessed even after request from the authors. ## Source of information This systematic review will be reported according to the Meta-Analysis of Observational Studies in Epidemiology Guidelines. [bib_ref] Meta-analysis of observational studies in epidemiology: a proposal for reporting. Meta-analysis Of..., Stroup [/bib_ref] search strategy for identifying related studies The search strategy will be conducted in two stages. ## Bibliographic database searches Relevant studies will be identified by searching PubMed, EMBASE, SCOPUS, Africa Journal Online and Google Scholar. The searches will only be limited to studies performed in Africa on African subjects and published from 1 January 1980 to 30 June 2017. Key search terms will be: 'peritoneal dialysis', 'peritonitis', 'Africa' with a filter Flow diagram for study selection. ## Open access to include all African countries (PubMed search strategy depicted on table 1). We will further use controlled vocabularies synonyms to identify related terms. Individual country names for all African countries will also be included in the search. Two authors (MWM and SK) will independently conduct the database searches and articles in French will be assessed by (JJN). ## Searching for other sources We will further manually scan through the reference lists of pertinent studies in search of more studies. Authors for studies which cannot be accessed through online searches will be directly contacted via email to request for the articles. Where we are unable to get feedback from the author, the studies will be excluded from the review. selection of studies for inclusion in the review An abstract of articles reporting on the prevalence and outcomes of CAPD peritonitis will be screened for inclusion into the review. Titles and abstracts of all eligible papers will be independently reviewed by two authors (MWM and SK). A joint review process will be undertaken for any disagreement during the selection process and one of the authors (IGO) will be asked to adjudicate the process in order to reach consensus. shows the process of study selection for the review. Assessment of methodological quality and reporting of data Two authors (MWM and SK) will assess the quality of eligible studies guided by the Joanna Briggs Institute Critical Appraisal Checklist for Studies Reporting Prevalence Data (2014) and the guidelines from the Cochrane Handbook for Systematic Reviews of Interventions V. 5.1.0. Patients and public involvement Patients and public were not involved in the marking of this study protocol. data extraction and management Two investigators (MWM and SK) will independently extract data from included studies, using a standardised and pretested data extraction form. Extracted data will include study characteristics (authors, country, year of publication, the study design, sample size, mean or median age of population, gender distribution, duration of study and mean or median duration of CAPD use) and specific features related to complications of CAPD (number of episodes of peritonitis and rates, causative organisms of peritonitis and rates of technique failure or modality switch to HD). Data on overall mortality and mortality related to peritonitis will also be extracted from each study. data synthesis and analysis All data will be analysed using the R statistical software (V.3.3.3 (2017-03-06), the R Foundation for statistical computing, Vienna, Austria) and 'meta' package. Interrater agreement for study inclusion will be assessed using Cohen's κ coefficient. [bib_ref] Cohen's coefficient κ, Sedgwick [/bib_ref] For data that we are unable to conduct a meta-analysis, we will provide a narrative description. These data will include study characteristics such as year of publication, sample size and country where study. Peritonitis rates will be reported as number of episodes per patient-year. [bib_ref] ISPD peritonitis recommendations: 2016 update on prevention and treatment, Li [/bib_ref] For the outcomes of interest consistently reported across studies, random effects model meta-analyses will be used to pool estimates across those studies. Stratified analyses will be consistently used to derive those pooled estimate separately in adult and paediatric populations. Heterogeneity across studies will be evaluated using Cochrane's Q statistic and quantified using I 2 statistics. Where substantial heterogeneity is detected, a subgroup analysis will be performed to explore the possible sources using the following grouping variables: age group, gender, geographical area (central, eastern, northern, southern and western Africa) and study quality. Graphical and formal statistical tests will be used to assess for publication bias.Potential outliers will be investigated in sensitivity analysis by dropping each study at a time. The Duval and Tweedie trim-and-fill will be used to adjust estimates for the effects of publication bias. ## Presentation and reporting of results The PRISMA flow chart will be used to show how the studies were included into the review and also explain why some of the studies were excluded. Data will be summarised in tables and forest plots. Where appropriate, pie charts will also be used. ## Ethics and dissemination Ethical approval will not be needed for this study as data used will be extracted from already published studies. Results of this review will be published in a peer-reviewed journal and presented at conferences. [table] Table 1: Search strategy for pubMed Burkina Faso"[tw] OR Burundi[tw] OR Cameroon[tw] OR "Canary Islands"[tw] OR "Cape Verde"[tw] OR "Central African Republic"[tw] OR Chad[tw] OR Comoros[tw] OR Congo[tw] OR "Democratic Republic of Congo"[tw] OR Djibouti[tw] OR Egypt[tw] OR "Equatorial Guinea"[tw] OR Eritrea[tw] OR Ethiopia[tw] OR Gabon[tw] OR Gambia[tw] OR [/table]

Revisiting the Fertility Transition in England and Wales: The Role of Social Class and Migration # Online appendix ## Fig. a1 Model estimates from fixed-effects models using consistent or changing administrative geography for the relationship between marital net fertility (number of children aged 0-4) and husband's social class, Note: Models control for age of woman, age difference between spouses, household status, and husband's social class based on estimation on registration districts (RDs) or registration sub-districts . and Day (2018b). (2014) and Day (2018b). [formula] - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - [/formula]

Image Alignment for Tomography Reconstruction from Synchrotron X-Ray Microscopic Images A synchrotron X-ray microscope is a powerful imaging apparatus for taking high-resolution and high-contrast X-ray images of nanoscale objects. A sufficient number of X-ray projection images from different angles is required for constructing 3D volume images of an object. Because a synchrotron light source is immobile, a rotational object holder is required for tomography. At a resolution of 10 nm per pixel, the vibration of the holder caused by rotating the object cannot be disregarded if tomographic images are to be reconstructed accurately. This paper presents a computer method to compensate for the vibration of the rotational holder by aligning neighboring X-ray images. This alignment process involves two steps. The first step is to match the ''projected feature points'' in the sequence of images. The matched projected feature points in the x-h plane should form a set of sine-shaped loci. The second step is to fit the loci to a set of sine waves to compute the parameters required for alignment. The experimental results show that the proposed method outperforms two previously proposed methods, Xradia and SPIDER. The developed software system can be downloaded from the URL, # Introduction Synchrotron X-rays possess several unique characteristics including high intensity, straight-line beam traveling, and low scattering [bib_ref] Basic characteristics of synchrotron radiation, Shenoy [/bib_ref] , which can be used to develop an X-ray microscope [bib_ref] Synchrotron radiation in radiology: radiology techniques based on synchrotron sources, Meuli [/bib_ref]. Researchers have developed many synchrotron X-ray microscopy techniques for various purposes including angiography [bib_ref] Synchrotron microangiography with no contrast agent, Hwu [/bib_ref] , X-ray lithography for nanofabrication [bib_ref] E-beam lithography and electrodeposition fabrication of thick nanostructured devices, Lo [/bib_ref] , and cell tomography [bib_ref] Imaging cells and tissues with refractive index radiology, Hwu [/bib_ref] [bib_ref] X-ray tomography generates 3-d reconstructions of the yeast, saccharomyces cerevisiae, at 60-nm..., Larabell [/bib_ref] [bib_ref] Soft x-ray tomography and cryogenic light microscopy: The cool combination in cellular..., Mcdermott [/bib_ref]. This study investigated a problem arising from 3D tomography reconstruction when the pixel size is in the nanoscale. Previous research has demonstrated that a series of X-ray projections around an object can be used to reconstruct 3D volume data by using the appropriate reconstruction algorithms [bib_ref] Practical cone beam algorithm, Feldkamp [/bib_ref] [bib_ref] Simultaneous algebraic reconstruction technique (sart): a superior implementation of the art algorithm, Andersen [/bib_ref] [bib_ref] Development and optimization of regularized tomographic reconstruction algorithms utilizing equally-sloped tomography, Mao [/bib_ref]. When the light source is a synchrotron X-ray, it cannot rotate around an object. Therefore, a rotatable object-holder was designed to hold an object to enable acquiring a series of projections from different angles while rotating the object [fig_ref] Figure 1: View of the rotational holder for acquiring the projection images for reconstructing... [/fig_ref]. The problem of mechanical imprecision arises when the resolution increases to a certain level, such as that required for cell tomography. When the pixel size is approximately 10 nm, a slight mechanical vibration can hinder accurate reconstruction. In the cell tomography experience of the authors, the pixel size of image is 11.78 nm. Rotating the object holder can cause a 5 to 30 pixel difference in position because of the mechanical instability. Although the position of the object holder can be calibrated during image acquisition, the calibration process can take an unacceptably long time, causing the object to receive excessive X-rays. The TXM controller provided by Xradia (hereinafter called ''Xradia'') was designed to solve the misalignment problem. Unfortunately, manual adjustments are generally required to obtain satisfactory tomography reconstruction. A similar problem also exists in electron microscopic tomography. Using the crosscorrelation function to align the electron microscopic images is a common solution to this problem. A software system, ''SPIDER'', was implemented based on the cross-correlation function [bib_ref] Spider image processing for single-particle reconstruction of biological macromolecules from electron micrographs, Shaikh [/bib_ref]. However, the cross-correlation function alignment does not considering the projection model, thus limiting the quality of tomography reconstruction. This study presents a feature-based alignment approach for calibrating the displacement caused by mechanical vibration. Because a synchrotron X-ray is a parallel beam projection, the resulting displacement can be decomposed into vertical and horizontal displacements. The proposed method aligns the images in the vertical direction by direct image alignment. Calibrating the images in the horizontal direction is more complex than that in the vertical direction. In addition to matching feature points, the matched feature points must form sine-wave shaped loci. This study proposes approximating the loci of the matched feature points in the x-h coordinate system according to sine waves by using the least square curve fitting. The deviation between the loci and the sine waves provides information for horizontal calibration. The remainder of this paper is organized as follows. The Results section presents the results. The Methods section presents the # Results The proposed algorithm was used to align projection images and then reconstruct the 3D volume data of HeLa cells from X-ray projections. To verify the correctness of the alignment algorithm, a phantom data set was used to simulate the HeLa cell stained using the gold nanoparticles. The following subsections present the construction details and results obtained. ## The phantom A volume datum containing 20 imitative gold nanoparticles was constructed, and X-ray projections of the shifted volume were generated to simulate machine vibrations. [fig_ref] Figure 2: Figure 2 [/fig_ref] shows the simulated X-ray image of the phantom cell, in which 180 projections of 5126512 pixels images were generated. The volume rotated is 1u between successive projections. Before each projection, the volume was shifted in the vertical and horizontal directions to simulate the mechanical imprecision of the object holder. The amount of displacement was determined according to a random number uniformly distributed over the range 620 pixels. The proposed alignment algorithm was used to calibrate the images, which were first aligned in the vertical direction. The detection and matching methods of projected feature points were then applied to determine 16 feature points. The locus of the horizontal position of a feature point over various angles should form a sine-shaped curve. [fig_ref] Figure 3: Feature loci of the phantom data [/fig_ref] shows the loci of the 16 points. This figure is in the x-h coordinate system: the vertical axis represents the rotation angle, and the horizontal axis represents the horizontal position of the projected feature point. The loci are not smooth because of horizontal displacement. The most suitable sine waves to fit the loci [fig_ref] Figure 3: Feature loci of the phantom data [/fig_ref] were then calculated, and the displacement of the feature points in the horizontal position were estimated [fig_ref] Figure 3: Feature loci of the phantom data [/fig_ref]. The filtered back-projection (FBP) algorithm was then used to reconstruct the 3D volume data based on the aligned X-ray images [bib_ref] Accelerating popular tomographic reconstruction algorithms on commodity pc graphics hardware, Xu [/bib_ref]. To verify the accuracy of the proposed alignment method, the positions and diameters of the spherical particles in the reconstructed images were compared with the original volume data, which included 20 spherical particles, and the diameter of each particle was 12 voxels. There were 20 particles found in the reconstructed volume. The average errors of the particle center and diameter were found to be 0.72 and 0.03 voxels, respectively. The reconstructed volume is close to the original volume. [fig_ref] Figure 4: One slice in the phantom data [/fig_ref] shows a slice in the original volume data. [fig_ref] Figure 4: One slice in the phantom data [/fig_ref] shows the tomography results of the same slice obtained using the proposed alignment method. For comparing the results of the proposed method with the SPIDER, the same phantom image was reconstructed using SPIDER. [fig_ref] Figure 4: One slice in the phantom data [/fig_ref] shows the tomography reconstructed results using SPIDER. The simulated X-ray image of the phantom data. 180 emulated X-ray images were generated. The size of each image is 5126512 pixels, the rotation angle between two consecutive images is 1 0 , and the ranges of the vertical and horizontal errors are +20 pixels. doi:10.1371/journal.pone.0084675.g002 The mean squared error (MSE) of the foreground voxels between the original volume data and the reconstructed volume was also calculated. A voxel was classified as a foreground voxel when its intensity exceeded a given intensity threshold. The MSE between the original volume and the reconstructed volume from the unaligned volume was found to be 0.029. The reconstructed volume obtained from the images aligned using SPIDER has an MSE of 0.002. For the volume reconstructed from the images aligned using the proposed method, the MSE was found to reduce to 0.0005. Because Xradia does not recognize the file format of the phantom images, the proposed method was not compared with Xradia. ## Hela cells Figs. 5(a) and (b) separately show the projection images of two HeLa cells. The first HeLa cell, named HeLa1, contained 84 identified projected feature points. Six reliable feature points were selected from the identified projected feature points. [fig_ref] Figure 6: The loci of the reliable projected feature points of HeLa1 and HeLa2 [/fig_ref] shows the loci of the selected points in the x-h coordinate system. (c, f) were obtained using Xradia. As shown in [fig_ref] Figure 7: One slice in the reconstructed tomographic images of HeLa1 and HeLa2 [/fig_ref] , the results of the proposed method were more favorable than those of SPIDER and Xradia. Comparing the results of the SPIDER and Xradia, the proposed method exhibits most well-defined membrane structures with least amount of artifacts, which is evident from 7(a) and (d). The texture-based volume rendering algorithmwas used to produce a 3D image of the volume data. The proposed algorithm was applied to process 10 other X-ray image sets of HeLa cells. Among these 10 image sets, eight were successfully reconstructed (A1-A8, Figs. 9 and 10) except the other two (B1 and B2, [fig_ref] Figure 1: View of the rotational holder for acquiring the projection images for reconstructing... [/fig_ref]. SPIDER and Xradia were also applied to the same sets of image data. Neither SPIDER nor Xradia could align the images in B1 and B2 for reconstruction. Specifically, SPIDER was effective for A3, A5, and A6; Xradia was effective for A1, A2, A5, A7, and A8. Although the reconstructions could be completed, comparing to our results, the artifacts produced due to the misalignment are more apparent. [fig_ref] Table 1: The test results. [/fig_ref] lists the experiments conducted in this study including the phantom, HeLa1, HeLa2, and the other 10 HeLa cells, as well as the computing time required. This table shows that the main factors affecting computational time are the image size, number of images, and number of identified projected features. The experiments in this study show that when the input data contains 180 images of 102461024 pixels, the alignment can be performed in 10 min. # Methods ## Sample preparation and image acquisition HeLa cells were used in this study. The cells were grown on Kapton film, and endocytosis [bib_ref] Quantitative analysis of nanoparticle internalization in mammalian cells by high resolution xray..., Chen [/bib_ref] was used to stain HeLa cells by absorbing gold nanoparticles of 250 mM (micromolar). The cells were then fixed in a container using a mixture of paraformaldehyde and glutaraldehyde [bib_ref] Quantitative analysis of nanoparticle internalization in mammalian cells by high resolution xray..., Chen [/bib_ref]. The synchrotron microscope used in this study was built at the National Synchrotron Radiation Research Center, Hsinchu, Taiwan. The CCD size is 204862048 pixels, and the field of view is 24 mm. Each projection image was taken after a 1u rotation. To prevent the object holder from becoming perceptible (occurring when the object holder is nearly parallel to the X-ray), the range of the rotation angle of the object holder was 670u. Only 140 projection images were acquired. The size of each image was 102461024 pixels, and the pixel size was 11.78 nm. In this study, the exposure time of each image was 1 second. # The alignment method A projected feature point is a pronounced mark in an X-ray projection image. Alignment is accomplished by first aligning the projected feature points in the vertical direction and then in the horizontal direction. The projected feature points should be maintained in the vertical direction. Thus, vertically aligning the projected feature points in the second image to the previous image is sufficient. However, the location of the feature points in the horizontal direction varies among projection images. Calculating the horizontal location of the feature points is a more difficult task than alignment in the vertical direction. The following subsections describe these steps. Vertical Direction Alignment. For each pair of projection images, the sum of the intensity values on each row is calculated. The sums of the rows form histograms that should be similar in a pair of consecutive images. The vertical displacement can be calculated by minimizing the difference between the histograms. Given an N6M image I(x,y), 0ƒxvN, 0ƒyvM and I(x,y)[[0, 1]. The vertical histogram h is calculated by Assume that I a is the unaligned image and I b is the reference image. The vertical correction of I a isŷ y, which can be estimated byŷ [formula] y~arg min y' X M{1 y~0 (h(I a ,yzy'){h(I b ,y)) 2 :ð2Þ [/formula] To achieve the most favorable results, the image is preprocessed to enhance the features. In this experiment, the estimated correction is more accurate when the images are enhanced by applying the edge detection method. Horizontal Direction Calibration. The horizontal calibration is based on the projected feature points forming a sine-shaped locus in the x-h coordinate system. This calibration involves three steps: detecting projected feature points, matching projected feature points to construct a set of loci from the matched projected feature points, and fitting the loci to sine curves to adjust the horizontal displacement of images. Detecting Projected Feature Points. Feature point extraction is a fundamental step in image stitching, object recognition, and feature-based image alignment. Researchers have proposed many feature detection methods. The corner detection method proposed by Harris and Stephen [bib_ref] A combined corner and edge detector, Harris [/bib_ref] is commonly used to extract corner-shape regions in an image. To achieve scale invariance, Kadir and Brady [bib_ref] Saliency, scale and image description, Kadir [/bib_ref] selected the salient region from the image scale-space as the feature that possesses the maximum entropy. Lowe [bib_ref] Distinctive image features from scale-invariant keypoints, Lowe [/bib_ref] proposed the scale-invariant feature transform (SIFT) algorithm to select local extrema from the differences of a Gaussian (DoG) pyramid in an image. The SIFT algorithm uses the gradient location-orientation histogram as a feature descriptor to achieve rotation invariance and illumination invariance. Researchers have proposed several improved versions of the SIFT algorithm. Bay et al. used the Haar wavelet to expedite feature detection [bib_ref] Surf: Speeded up robust features, Bay [/bib_ref]. Rady et al. proposed entropy-based feature detection [bib_ref] Entropy-based features for robust place recognition, Rady [/bib_ref] , and Suri et al. combined mutual information alignment with the SIFT algorithm [bib_ref] Combining mutual information and scale invariant feature transform for fast and robust..., Suri [/bib_ref]. In this study, a modified SIFT algorithm was employed to extract automatically the projected feature points contained in Xray images. The typical SIFT implementation involves describing a feature according to its location, size, the orientation of the sampling region, and the image gradient histogram in the sampling region. Because the proposed method matches the projected feature points in two X-ray images based on mutual information [bib_ref] Alignment by maximization of mutual information, Viola [/bib_ref] [bib_ref] Mutual information based registration of medical images: A survey, Pluim [/bib_ref] [bib_ref] Mutual information for lucas-kanade tracking: An inverse compositional formulation, Dowson [/bib_ref] , each projected feature point in this study contained the entropy of the sampling region rather than the image gradient histogram. To reduce the noise and the number of low-contrast projected feature points, the entropy of each selected projected feature point must exceed a given threshold. The experiments in this study entailed setting a threshold between 0.5 and 1.0. Because the features in the objects are gold nanoparticles, the size and orientation of the sampling region were fixed in this implementation. Matching Projected Feature Points. Let F i , i~1, . . . ,m be the sets of projected feature points in m projection images. The projected feature points are classified into k groups. In the ideal case, each group is the set of projected feature points, which are the projections of a feature (i.e., gold nanoparticle) in the object from various angles. Because the rotation angle of the object is small, the projected feature points are in proximity and have similar mutual information in two consecutive images. However, the distance between the two matched projected feature points depends on the distance between the feature and the rotation axis of the object. This means that an affine transform cannot match the projected feature points in two images. Therefore, this study presents a greedy method for classifying the projected feature points. For each pair of images, the random sample consensus (RANSAC) method [bib_ref] Random sample consensus: a paradigm for model fitting with applications to image..., Fischler [/bib_ref] was first applied to compute an initial alignment of the two images, and a tracking method was then employed to match the projected feature points in the next image. Several feature tracking methods are available. The proposed method is designed based on the Shafique and Shah's method [bib_ref] A noniterative greedy algorithm for multiframe point correspondence, Shafique [/bib_ref] , which is a greedy algorithm for tracking moving objects in videos, and the Tang and Tao's method [bib_ref] Probabilistic object tracking with dynamic attributed relational feature graph, Tang [/bib_ref] , which integrates the hidden Markov model to eliminate unreliable matches. Given the projected feature points sets F i{1 and F i , the RANSAC method was applied to compute a translation matrix T i,i{1 , so that a sufficient number of projected feature points p and q in F i{1 and F i respectively, jT i,i{1 (q){pj is less than a given threshold . Applying translation matrices T i,i{1 , i~2, . . . ,m to the consecutive images achieves the initial alignment of the m projection images. All of the images are aligned based on the first image. Given the initially aligned projected feature points F i ,i~1, . . . ,m, the following procedures yield a set of possible loci of the projected feature points produced by feature points in the object. compute T i{1,i (p) where p is the final point of l and T i{1,i is the inverse of T i,i{1 . Let L be a region in I i centered at T i{1,i (p). Search in L for the projected feature points [fig_ref] Figure 1: View of the rotational holder for acquiring the projection images for reconstructing... [/fig_ref]. If this region contains only one projected feature point q, then that point q is selected as the final point of l. If the region contains more than one projected feature point, 3. Reverse the image orders and repeat Step 2, but do not include 2b to backtrack all loci. The X-ray images used in this study measured 1024|1024 pixels, and 128|32 pixels was the size of search region L, and five previous points for t. Because two loci could intersect (i.e., two projected feature points on two loci could overlap or be extremely close), the average entropy must be computed to select the best-matching projected feature point in Step 2a. In this step, some projected feature points with a high entropy in F i are not included in any locus. These significant projected feature points should not be disregarded, and Step 2b entails creating a new locus for each of them. ## After Step 2, the forward feature tracking required to construct the set of loci is complete. To verify the correctness of the loci and complete the loci added in the Step 2b, backtrack all loci in the final step. ## Horizontal displacement estimation Let the set of k loci collected in the previous step be fl 1 ,l 2 , . . . ,l k g. Each locus, l j , 1ƒjƒk, consists of m projected feature points, vf 1 j ,f 2 j , . . . ,f m j w. These projected feature points are expected to be the projections of a feature, p j , in the object from various angles. Assume that these projected feature points are the projections from {p=2 and p=2. Recall that f i j is not a point, but a rectangular box. Take the X -coordinates of the center of the rectangular boxes, x, and transform the j to h, the direction of the projections. Then, draw (x,h) for all f i j in the x-h coordinate system. The locus of (x,h) corresponds to a sine curve, (i.e., the sinogram). Given a set of loci of features, the horizontal alignment is conducted by fitting the curves to a set of sine curves and then by computing the deviations. Consider a locus, l j , and assume that the projected feature points on the locus are the projections of the feature point p j . This feature point p j can be expressed as (c j ,v j ), where c is the radial coordinate and v is the angular coordinate. According to the Radon transform, the corresponding horizontal position x j in the i th image (rotation angle h i ), 1ƒiƒm, can be written as [formula] x i j~cj sin (h i zv j )ð3Þc j ( cos h i sin v j z sin h i cos v j ),ð4Þ [/formula] or in matrix form as x~Hu, where u~½u 1 ,u 2 T~½ c j sin v j ,c j cos v j T , x~½x 1 j ,x 2 j ,:::,x m j T , and : u 1 and u 2 can be solved by the least-squares method, [formula] u~(H T H) {1 H T x:ð5Þ [/formula] Finally (c j ,v j ), [formula] c j~ffi ffiffiffiffiffiffiffiffiffiffiffiffiffi u 2 1 zu 2 2 qð6Þ [/formula] and [formula] v j~s in {1 ( u 1 c j ):ð7Þ [/formula] The horizontal displacementx x i j is estimated bŷ [formula] x x i j~cj sin (h i zv j ){x i j :ð8Þ [/formula] Because both v j and p{v j are solutions to (7), choose the one that minimizes the sum of errors, P m i~1x x i j . To determine the horizontal displacement c i for the i th image from k feature loci, use the average of the k horizontal corrections: [formula] c i~1 k X k j~1x x i j :ð9Þ [/formula] Some loci are unreliable because of noises, out-of-view projected feature points, or bad projected feature points matching. These loci should be removed. For each locus l j , adjust the points on l j based on the estimated c i , 1ƒiƒm. l j is unreliable if the absolute peak distance between a point on the adjusted l j and its best-fitting sine wave is greater than a given threshold s. If there are no unreliable loci, stop the algorithm and output the aligned results. Otherwise, remove the unreliable loci and repeat the horizontal displacement estimation algorithm. # Discussion This paper presents an image alignment method for X-ray images produced by synchrotron-radiation microscopy. The proposed method enables reconstructing the 3D volume of small objects. The proposed method identifies projected feature points and classifies the projected feature points into a set of loci. The key idea of this method is that the projection of a point in the object should be a sine wave in the x-h coordinate system. Thus, fitting the set of loci to a set of sine waves can compute the parameters required for alignment. The proposed method was applied to 12 cases of HeLa cells, and only two of the 12 could not be reconstructed. Compared with the available software systems, SPIDER and Xradia, they could respectively construct three and five cells. Neither SPIDER nor Xradia could construct the two unsuccessful cases of the proposed method. The main reason for the unsuccessful cases is the insufficient number of projected feature points in the X-ray images. The most crucial factor affecting the performance of the proposed method is the number of reliable projected feature points. If there are enough reliable projected feature points, even if the projected feature points are not in the field of view in some projections, the method still works well because it also considers the set of partial loci. The proposed method performs most favorably if the features that produce the projected feature points are close to the rotational axis. Carefully adjusting the rotation axis before images acquisition can improve the quality of the reconstruction. Considering the shape of the projected feature points, aside from particle objects, the proposed method can manage any shape of object if it contains a sufficient number of distinct projected features. For examples, the corners of a square, the two tips of a rod, and the branch points of a tree-structured object can be used as feature points as long as the features do not deform during image acquisition. If the images satisfy these requirements, then the proposed method can successfully align the images. [fig_ref] Table 1: The test results. [/fig_ref] lists the computing time required for the 10 successful reconstructions of HeLa cells. A graphical user interface software system for the Windows system and Mac OS X 10.8 has been developed. The software system can be downloaded from the following URL: http://www.cs.nctu.edu.tw/ cheng~chc/SCTA , or http://goo.gl/s4AMx. [fig] Figure 1: View of the rotational holder for acquiring the projection images for reconstructing the 3D volume images using Synchrotron light source and for correcting the horizontal alignment due to vibration below 10 nm. doi:10.1371/journal.pone.0084675.g001 [/fig] [fig] Figure 2: Figure 2. The simulated X-ray image of the phantom data. 180 emulated X-ray images were generated. The size of each image is 5126512 pixels, the rotation angle between two consecutive images is 1 0 , and the ranges of the vertical and horizontal errors are +20 pixels. doi:10.1371/journal.pone.0084675.g002 [/fig] [fig] Figure 3: Feature loci of the phantom data: (a) loci of 16 projected feature points before horizontal alignment; (b) bestfit sine waves; and (c) the aligned loci. doi:10.1371/journal.pone.0084675.g003 [/fig] [fig] Figure 4: One slice in the phantom data: (a) original image; (b) result of proposed method; and (c) the result of SPIDER. doi:10.1371/journal.pone.0084675.g004 [/fig] [fig] Figure 5: X-ray projection images of HeLa cells. 140 synchrotron X-ray projection images were acquired for HeLa1 and HeLa2. The size of each image is 102461024 pixels, and the rotation angle between two projection images is 1 0 . (a) HeLa1, and (b) HeLa2. doi:10.1371/journal.pone.0084675.g005 [/fig] [fig] Figure 6: The loci of the reliable projected feature points of HeLa1 and HeLa2. (a) There were six reliable feature loci found in the acquired images of HeLa1. (b) Most suitable sine waves of HeLa1, and (c) aligned loci of HeLa1. (d) There were four reliable feature loci identified in the acquired images of HeLa2. (e) Best-fit sine waves of HeLa2, and (f) aligned loci of HeLa2. doi:10.1371/journal.pone. [/fig] [fig] Figure 7: One slice in the reconstructed tomographic images of HeLa1 and HeLa2. (a), (b), and (c) are the same slice in the tomographic images of HeLa1: (a) result of proposed method; (b) result of SPIDER, and (c) result of Xradia. (d), (e), and (f) are the same slice in the tomographic images of HeLa2: (a) result of proposed method; (b) result of SPIDER, and (c) result of Xradia. doi:10.1371/journal.pone.0084675.g007 [/fig] [fig] Figure 8: The 3D volume rendering of the reconstructed volume (a), (b), and (c) are HeLa1; (d), (e), and (f) are HeLa2: (a) and (d) Results of proposed method; (b) and (e) results of SPIDER, and (c) and (f) results of Xradia. doi:10.1371/journal.pone.0084675.g008 [/fig] [fig] Figure 9: Eight examples of successful alignment (1). doi:10.1371/journal.pone.0084675.g009 [/fig] [fig] 1: Every projected feature point in F 1 is the starting point of a locus. 2. Iteratively process F i , i~2, . . . ,m; (a) Let L be the set of the loci computed. For each locus l[L, [/fig]

Quantification of Effective Connectivity in the Brain Using a Measure of Directed Information Effective connectivity refers to the influence one neural system exerts on another and corresponds to the parameter of a model that tries to explain the observed dependencies. In this sense, effective connectivity corresponds to the intuitive notion of coupling or directed causal influence. Traditional measures to quantify the effective connectivity include model-based methods, such as dynamic causal modeling (DCM), Granger causality (GC), and information-theoretic methods. Directed information (DI) has been a recently proposed information-theoretic measure that captures the causality between two time series. Compared to traditional causality detection methods based on linear models, directed information is a model-free measure and can detect both linear and nonlinear causality relationships. However, the effectiveness of using DI for capturing the causality in different models and neurophysiological data has not been thoroughly illustrated to date. In addition, the advantage of DI compared to modelbased measures, especially those used to implement Granger causality, has not been fully investigated. In this paper, we address these issues by evaluating the performance of directed information on both simulated data sets and electroencephalogram (EEG) data to illustrate its effectiveness for quantifying the effective connectivity in the brain. # Introduction Neuroimaging technologies such as the electroencephalogram (EEG) make it possible to record brain activity with high temporal resolution and accuracy. However, current neuroimaging modalities display only local neural activity rather than large-scale interactions between different parts of the brain. Assessment of the large-scale interdependence between these recordings can provide a better understanding of the functioning of neural systems [bib_ref] Large-scale cortical networks and cognition, Bressler [/bib_ref] [bib_ref] Organization, development and function of complex brain networks, Sporns [/bib_ref]. Three kinds of brain connectivity are defined to describe such interactions between recordings: anatomical connectivity, functional connectivity, and effective connectivity [bib_ref] Organization, development and function of complex brain networks, Sporns [/bib_ref]. Anatomical connectivity is the set of physical or structural connections linking neuronal units at a given time and can be obtained from measurements of the diffusion tensor [bib_ref] An investigation of functional and anatomical connectivity using magnetic resonance imaging, Koch [/bib_ref] [bib_ref] Functional and effective connectivity: a review, Friston [/bib_ref]. Functional connectivity captures the statistical dependence between scattered and often spatially remote neuronal units by measuring their correlations in either time or frequency domain. Effective connectivity describes how one neural system affects another [bib_ref] Organization, development and function of complex brain networks, Sporns [/bib_ref] [bib_ref] Functional and effective connectivity: a review, Friston [/bib_ref] [bib_ref] Functional and effective connectivity in neuroimaging: a synthesis, Friston [/bib_ref] , which can provide information about both the magnitude and the direction of the interaction. The main approaches used to quantify the effective connectivity between two time series are model-based measures and information-theoretic measures [bib_ref] Nonlinear multivariate analysis of neurophysiological signals, Pereda [/bib_ref]. Granger-causalitybased methods and dynamic causal modeling [bib_ref] Dynamic causal modelling, Friston [/bib_ref] are two widely used model-based measures. Granger causality is a widely used measure to describe the causality between two time series. It defines a stochastic process X causing another process Y if the prediction of Y at the current time point, Y n , is improved when taking into account the past samples of X. This approach is appealing but gives rise to many questions on how to apply this definition to real data [bib_ref] Testing for causality. A personal viewpoint, Granger [/bib_ref]. Granger causality has been mostly applied within a linear prediction framework using a multivariate autoregressive (MVAR) model yielding methods such as directed transfer function (DTF), partial directed coherence (PDC), and directed partial correlation [bib_ref] Partial directed coherence: a new concept in neural structure determination, Baccala [/bib_ref] [bib_ref] The use of time-variant EEG Granger causality for inspecting directed interdependencies of..., Hesse [/bib_ref] [bib_ref] Evaluating causal relations in neural systems: Granger causality, directed transfer function and..., Kamiński [/bib_ref] [bib_ref] Investigating multivariate systems using directed partial correlation, Mader [/bib_ref]. For example, Hesse et al. applied time-varying Granger causality to EEG data and 2 Computational and Mathematical Methods in Medicine found that conflict situation generates directional interactions from posterior to anterior cortical sites [bib_ref] The use of time-variant EEG Granger causality for inspecting directed interdependencies of..., Hesse [/bib_ref]. DTF to EEG recordings of human brain during stage 2 sleep and located the main source of causal influence [bib_ref] Evaluating causal relations in neural systems: Granger causality, directed transfer function and..., Kamiński [/bib_ref]. Schelter et al. employed PDC to EEG recordings from a patient suffering from essential tremor [bib_ref] Testing for directed influences among neural signals using partial directed coherence, Schelter [/bib_ref]. The extensions of Granger-causality-based methods, such as kernel Granger causality, generalized PDC (gPDC), and extended PDC (ePDC), have also found numerous applications in neuroscience [bib_ref] Nonlinear connectivity by Granger causality, Marinazzo [/bib_ref] [bib_ref] Connectivity analysis of somatosensory evoked potentials to noxious intracutaneous stimuli in patients..., Leistritz [/bib_ref] [bib_ref] Extended causal modeling to assess partial directed coherence in multiple time series..., Faes [/bib_ref]. However, Granger causality-based methods, especially those developed from MVAR models, are limited to capturing linear relations or require a priori knowledge about the underlying signal models [bib_ref] Directed information flow-a model free measure to analyze causal interactions in event..., Hinrichs [/bib_ref]. These approaches may be misleading when applied to signals that are known to have nonlinear dependencies, such as EEG data [bib_ref] Interdependence of EEG signals: linear versus nonlinear associations and the significance of..., Lopes Da Silva [/bib_ref]. DCM, on the other hand, can quantify nonlinear interactions by assuming a bilinear state space model. However, DCM requires a priori knowledge about the input to the system [bib_ref] Dynamic causal modelling, Friston [/bib_ref] [bib_ref] Transfer entropy-a model-free measure of effective connectivity for the neurosciences, Vicente [/bib_ref] and is limited to a network with small size [bib_ref] Functional and effective connectivity: a review, Friston [/bib_ref]. Thus, a model-free measure detecting both linear and nonlinear relationships is desired. Information theoretic tools [bib_ref] Measuring information transfer, Schreiber [/bib_ref] [bib_ref] Causality, feedback and directed information, Massey [/bib_ref] , such as transfer entropy [bib_ref] Measuring information transfer, Schreiber [/bib_ref] , address the issue of model dependency and have found numerous applications in neuroscience [bib_ref] Directed information flow-a model free measure to analyze causal interactions in event..., Hinrichs [/bib_ref] [bib_ref] Complex network measures of brain connectivity: uses and interpretations, Rubinov [/bib_ref] [bib_ref] Mapping information flow in sensorimotor networks, Lungarella [/bib_ref]. "Transfer entropy" (TE) proposed by Schreiber computes causality as the deviation of the observed data from the generalized Markov condition and is defined as [bib_ref] Measuring information transfer, Schreiber [/bib_ref] TE(X −→ Y) = yn+1,yn−l+1:n,xn−m+1:n p y n+1 y n−l+1:n x n−m+1:n × log p y n+1 | y n−l+1:n x n−m+1:n p y n+1 | y n−l+1:n , where m and l are the orders (memory) of the Markov processes X and Y, respectively. p(y n+1 y n−l+1:n x n−m+1:n ) is the joint probability of random variables (Y n+1 , Y n−l+1:n , X n−m+1:n ), where Y n−l+1:n = (Y n−l+1 , . . . , Y n ) and X n−m+1:n = (X n−m+1 , . . . , X n ). Sabesan et al. employed TE to identify the direction of information flow for the intracranial EEG data and suggested that transfer entropy plays an important role in epilepsy research [bib_ref] Information flow and application to epileptogenic focus localization from intracranial EEG, Sabesan [/bib_ref]. Wibral et al. applied TE to magnetoencephalographic data to quantify the information flow in cortical and cerebellar networks [bib_ref] Transfer entropy in magnetoencephalographic data: quantifying information flow in cortical and cerebellar..., Wibral [/bib_ref]. Vicente et al. extended the definition of TE and measured the information flow from X to Y with a general time delay of u, that is, replaced y n+1 in the above equation with y n+u , and showed that TE has a better performance in detecting the effective connectivity for nonlinear interactions and signals affected by volume conduction such as real EEG/MEG recordings compared to linear methods [bib_ref] Transfer entropy-a model-free measure of effective connectivity for the neurosciences, Vicente [/bib_ref]. The performance of transfer entropy depends on the estimation of transition probabilities, which requires the selection of order or memory of the Markov processes X and Y [bib_ref] Information flow and application to epileptogenic focus localization from intracranial EEG, Sabesan [/bib_ref]. "Directed transinformation" (T) introduced by Saito and Harashimameasures the information flow from the current sample of one signal to the future samples of another signal given the past samples of both signals. Hinrichs et al. used this measure to analyze causal interactions in event-related EEG-MEG experiments [bib_ref] Directed information flow-a model free measure to analyze causal interactions in event..., Hinrichs [/bib_ref]. However, this measure does not discriminate between totally dependent and independent processes [bib_ref] The relationship between two directed information measures, Al-Khassaweneh [/bib_ref]. Recently, directed information proposed by Marko [bib_ref] The bidirectional communication theory-a generalization of information theory, Marko [/bib_ref] and later reformalized by Massey, Kramer, Tatikonda, and others have attracted attention for quantifying directional dependencies [bib_ref] Causality, feedback and directed information, Massey [/bib_ref] [bib_ref] The bidirectional communication theory-a generalization of information theory, Marko [/bib_ref] [bib_ref] Capacity results for the discrete memoryless network, Kramer [/bib_ref] [bib_ref] The capacity of channels with feedback, Tatikonda [/bib_ref] [bib_ref] On directed information theory and Granger causality graphs, Amblard [/bib_ref]. Directed information theory has been mostly aimed towards the study of communication channels with feedback. In recent years, new theoretical developments motivated the use of this measure in quantifying causality between two time series. In particular, Amblard and Michel [bib_ref] On directed information theory and Granger causality graphs, Amblard [/bib_ref] recently showed how directed information and Granger causality are equivalent for linear Gaussian processes and proved key relationships between existing causality measures and the directed information. Therefore, there has been a growing interest in applying this measure to applications in signal processing, neuroscience, and bioinformatics. For example, it has been successfully used to infer genomic networks [bib_ref] Using directed information to build biologically relevant influence, Rao [/bib_ref] and to quantify effective connectivity between neural spike data in neuroscience [bib_ref] On directed information theory and Granger causality graphs, Amblard [/bib_ref] [bib_ref] Estimating the directed information to infer causal relationships in ensemble neural spike..., Quinn [/bib_ref] [bib_ref] Information theoretic approach to quantify causal neural interactions from EEG, Liu [/bib_ref]. In order to detect both linear and nonlinear relationships, in this paper, we propose directed information as a powerful measure to quantify the effective connectivity in the brain. The theoretical advantages of DI over existing measures have been noted in literature [bib_ref] On directed information theory and Granger causality graphs, Amblard [/bib_ref] [bib_ref] Estimating the directed information to infer causal relationships in ensemble neural spike..., Quinn [/bib_ref] [bib_ref] Information theoretic approach to quantify causal neural interactions from EEG, Liu [/bib_ref]. However, until now the benefits of using DI for capturing the effective connectivity in the brain through neurophysiological data have not been illustrated thoroughly and formally. In addition, because of the relationship between Granger causality and directed information, in this paper, we mainly focus on the comparison between these two measures and investigate the advantage of DI over Granger-causality-based model measures. Theoretical developments only proved the equivalence between these two measures for the case that the time series are distributed as Gaussian in a linear model. However, to date, there has not been much work that compares the actual performance of DI and Granger-causality-based measures for realistic signal models, including both linear and nonlinear interactions. Moreover, most applications of DI to real data have been limited to using either parametric density models for the data or making assumptions about the time dependencies such as assuming a first-order Markov chain and have not considered the difficulties associated with estimating DI from a finite sample size [bib_ref] On the detection of gene network interconnections using directed mutual information, Mathai [/bib_ref]. For complex systems, the computational complexity and the bias of the DI estimator increase with the length of the signal. The main contribution of this paper is to address these issues by evaluating the performance of DI and Granger-causality-based methods under a common framework without making any assumptions about the data distribution. In this paper, we first give a brief introduction to directed information and its computation based on nonparametric estimation methods. We propose a modified time-lagged directed information measure that simplifies the DI computation by reducing the order of the joint entropy terms while still quantifying the causal dependencies. We then evaluate the performance of DI for quantifying the effective connectivity for linear and nonlinear autoregressive models, linear mixing models, single source models, and dynamic chaotic oscillators in comparison to existing causality measures, in particular with Granger causality. Finally, we apply our method to EEG data to detect the effective connectivity in the brain. # Materials and methods ## Definitions and notations. In this section, we will first review some common notations and information-theoretic definitions that will be used throughout this paper. Let X = X n = X 1:n = (X 1 , . . . , X n ) be a random process with length n and p(x 1 , . . . , x n ) = p(x n ) = p(x 1:n ) be the joint probability of random variables (X 1 , . . . , X n ). DX n = X n−1 = (0, X 1 , . . . , X n−1 ) will be used to define the time-delayed version of sequence X n , which is also equivalent to X 1:n−1 . Given two continuous random variables X and Y , the mutual information (MI) is defined as follows (All integrals in the paper are from −∞ to +∞ unless otherwise specified.): [formula] I(X; Y ) = p x, y log p x, y p x (x)p y y dx dy,(2) [/formula] where p(x, y) is the joint probability density function (pdf) of X and Y , and p x (x), p y (y) are the marginal pdfs of X and Y , respectively. I(X; Y ) ≥ 0 with equality if and only if X and Y are independent. In information theory, mutual information can be interpreted as the amount of uncertainty about X that can be reduced by observation of Y , or the amount of information Y can provide about X, that is, [formula] I(X; Y ) = H(X) − H(X | Y ). Since I(X; Y ) ≥ 0, H(X | Y ) ≤ H(X) [/formula] with equality if and only if X and Y are independent; that is, conditioning reduces entropy. For any three random variables X, Y , and Z, if the conditional distribution of Z depends only on Y and is conditionally independent of X, that is, p(z | y) = p(z | yx), then X, Y , and Z are said to form a Markov chain, denoted by X → Y → Z. In this case, the conditional mutual information between X and Y given Z defined as I(X; [formula] Z | Y ) = H(Z | Y ) − H(Z | X, Y ) is equal to 0 [36]. [/formula] ## Directed information. Mutual information can be extended to random vectors or sequences X N and Y N as I(X N ; [formula] Y N ), where I(X N ; Y N ) = H(X N ) − H(X N | Y N ) = H(Y N )− H(Y N | X N ) [/formula] . However, mutual information is a symmetric measure and does not reveal any directionality or causality between two random sequences. Massey addressed this issue by defining the directed information from a length N sequence X N = (X 1 , . . . , X N ) to Y N = (Y 1 , . . . , Y N ) [bib_ref] Causality, feedback and directed information, Massey [/bib_ref] as follows: [formula] DI X N −→ Y N = H Y N − H Y N X N = N n=1 I X n ; Y n | Y n−1 ,(3) [/formula] where H(Y N ||X N ) is the entropy of the sequence Y N causally conditioned on the sequence X N , and H(Y N ||X N ) is defined as [formula] H Y N X N = N n=1 H Y n | Y n−1 X n ,( 4 )which differs from H(Y N | X N ) = N n=1 H(Y n | Y n−1 X N ) [/formula] in that X n replaces X N in each term on the right-hand side of (4), that is, only the causal influence of the time series X up to the current time sample n on the process Y is considered. An alternative definition of the directed information is proposed by Tatikonda in terms of Kullback-Leibler (KL) divergence [bib_ref] The capacity of channels with feedback, Tatikonda [/bib_ref]. It shows that the difference between mutual information and directed information is the introduction of feedback in the definition of directed information [bib_ref] Causality, feedback and directed information, Massey [/bib_ref] [bib_ref] The capacity of channels with feedback, Tatikonda [/bib_ref] [bib_ref] On directed information theory and Granger causality graphs, Amblard [/bib_ref]. Mutual information and directed information expressed by KL divergence are written as [formula] I X N ; Y N = D KL p x N , y N p x N p y N , DI X N −→ Y N = D KL p x N , y N ← − p x N | y N p y N ,(5)where ← − p (x N | y N ) = N n=1 p(x n | x n−1 y n−1 ) [/formula] is the feedback factor influenced by the feedback in the system, that is, the probability that the input X at current time is influenced by the past values of both itself and Y. If there is no feedback, then p(x n | x n−1 y n−1 ) = p(x n | x n−1 ) and [formula] ← − p (x N | y N ) = p(x N ). In fact, p(x N , y N ) = ← − p (x N | y N ) p(y N | x N ), where p(y N | x N ) = N [/formula] n=1 p(y n | x n y n−1 ) and is defined as the feedforward factor affected by the memory of the system. If the system is memoryless, then p(y n | x n y n−1 ) = p(y n | x n ). ## Directed information versus granger causality. Granger quantifies causality so that the time series X N causes Y N if the variance of the prediction error for Y at the present time is reduced by including past measurements from X. Based on Granger's definition of causality, Geweke introduced the Geweke's indices to quantify the causal linear dependencies under Gaussian assumptions [bib_ref] Measurement of linear dependence and feedback between multiple time series, Geweke [/bib_ref]. Amblard and Michel proved that the directed information rate and Geweke's indices are equal for Gaussian processes [bib_ref] On directed information theory and Granger causality graphs, Amblard [/bib_ref] as indicated by [formula] DI ∞ DX N −→ Y N = 1 2 log ε 2 ∞ Y N | Y N−1 ε 2 ∞ (Y N | Y N−1 X N−1 ) =F X N → Y N ,(6) [/formula] where DX N stands for the time-delayed sequence (0, X 1 , . . . , X N−1 ) with N being the length of the signal, [formula] DI ∞ (X N → Y N ) is the directed information rate; that is, DI ∞ (X N → Y N ) = lim N → ∞ I(X N ; Y N | Y N−1 ), ε 2 ∞ (Y N | Y N−1 ) = lim N → ∞ ε 2 (Y N | Y N−1 ) [/formula] is the asymptotic variance of the prediction residue when predicting Y N from the observation of Y N−1 , and F X N → Y N refers to the linear feedback measure from random processes X N to Y N defined by Geweke [bib_ref] Measurement of linear dependence and feedback between multiple time series, Geweke [/bib_ref]. This equality shows that asymptotically the DI rate is equivalent to the gain in information by predicting Y using the past values of both Y and X compared to only using the past samples of Y, which is similar to the definition of Granger causality. Moreover, Amblard and Michel proved the equality of directed information and Granger's approach for multivariate time series in the case of Gaussian distributions [bib_ref] On directed information theory and Granger causality graphs, Amblard [/bib_ref]. ## Computation of directed Information. The definition of DI for two length N sequences X N = (X 1 , . . . , X N ) and Y N = (Y 1 , . . . , Y N ) can also be rewritten in terms of the total change of joint entropy or mutual information along time as follows: [formula] DI X N −→ Y N = N n=1 I X n ; Y n | Y n−1 = N n=1 H X n Y n−1 − H(X n Y n ) + H Y N = N n=1 I(X n ; Y n ) − I X n ; Y n−1 .(7) [/formula] From the above equations, we can observe that the computation of DI requires the estimation of joint probabilities of high-dimensional random variables over time. If X n and Y n are normally distributed, the joint entropy can be estimated based on the covariance matrices. However, for EEG data, the distribution is usually not Gaussian. The nonparametric entropy and mutual information estimators, such as plug-in estimator, m-spacing estimator, and Kozachenko and Leonenko (KL) estimator, have been extensively addressed in literature [bib_ref] A new class of entropy estimators for multidimensional densities, Miller [/bib_ref] [bib_ref] Estimation of the information by an adaptive partitioning of the observation space, Darbellay [/bib_ref]. In this paper, directed information estimation based on mutual information is used to estimate DI directly from EEG data by using adaptive partitioning method discussed in [bib_ref] Estimation of the information by an adaptive partitioning of the observation space, Darbellay [/bib_ref]. However, when the length of the signal increases, the computational complexity, the bias, and the variance of these estimators increase immensely with limited sample sizes. Methods that can reduce the dimension and simplify the computation of DI are needed. In order to simplify the estimation of DI, we first clarify the connection between the definition of DI used in information theory and the definition as it applies to physical time series. In a physical recording system, if X starts to influence Y after p 1 time points or with a delay of p 1 samples, we need to record at least N + p 1 time points to obtain N points of the time sequence Y that have been affected by X. The directed information rate from time series X N+p1 to Y N+p1 can be defined as [bib_ref] Capacity results for the discrete memoryless network, Kramer [/bib_ref]. We have [formula] DI ∞ X N+p1 −→ Y N+p1 = lim N+p1 → ∞ 1 N + p 1 N+p1 n=1 I X n ; Y n | Y n−1 (8) = lim N+p1 → ∞ I X N+p1 ; Y N+p1 | Y N+p1−1 (9) = lim N+p1 → ∞ H Y N+p1 | Y N+p1−1 −H Y N+p1 | X N+p1 Y N+p1−1 (10) = lim N+p1 → ∞ H Y N+p1 | Y p1+1:N+p1−1 −H Y N+p1 | X N+p1 Y p1+1:N+p1−1 (11) = lim N+p1 → ∞ H Y N+p1 | Y p1+1:N+p1−1 −H Y N+p1 | X 1:N Y p1+1:N+p1−1 (12) = lim N+p1 → ∞ I X 1:N ; Y N+p1 | Y p1+1:N+p1−1 (13) = lim N → ∞ 1 N N n=1 I X n ; Y n+p1 | Y p1+1:n+p1−1 (14) = DI ∞ X 1:N −→ Y p1+1:p1+N ,(15) [/formula] where (11) comes from the fact that Y 1:p1 is independent of Y N+p1 , and (12) is derived using the fact that X N+1:N+p1 has no effect on Y N+p1 because of the time delay p 1 between these two time series. For two physical recordings X and Y with length N + p 1 and a lag of p 1 , the last equation shows that DI rate for these two time series is equivalent to DI rate for two random processes with length N that are not synchronized in time. In fact, Y p1+1:p1+N may be indexed as Y 1:N when using the information theoretic indexing, which indexes the signal not according to the physical time point but based on when the receiver receives its first piece of information. Therefore, directed information rate computed by using physical time indices is equivalent to the directed information rate using information theoretic indices for two systems that interact through a time delay. Moreover, when the length of the signal is long enough, the directed information value using both indices will be equivalent. Once the definition of directed information is extended from random vectors to two physical time series, we propose a modified time-lagged DI to simplify the computation of DI, which is an extension of time-lagged DI proposed for every two samples of X N and Y N in [bib_ref] Time-lagged directed information, Liu [/bib_ref] to general signal models. As we mentioned before, as the length N of the signal increases, the computational complexity, the bias, and the variance of estimating DI increase immensely with limited sample sizes. In addition, the directed information defined for the physical system is actually a DI with a lag of p 1 samples over a time window with length N. Therefore, an intuitive way to simplify the computation is to apply DI with lag p 1 over a small window. We first give the definition of time-lagged DI for two time series X N and Y N with length N at the nth time sample for a block of two time samples with a time delay of p 1 (n > p 1 ): [formula] DI n X n−p1 X n−p1+1 −→ Y n Y n+1 = I X n−p1 ; Y n + I X n−p1 X n−p1+1 ; Y n+1 | Y n = H X n−p1 + H X n−p1 X n−p1+1 Y n + H(Y n Y n+1 ) − H X n−p1 Y n − H X n−p1 X n−p1+1 Y n Y n+1 ,(16) [/formula] where n = p 1 + 1, . . . , N − 1, p 1 is the time lag between the two time series, and N is the length of the whole time series. Computational and Mathematical Methods in Medicine 5 Therefore, the total directed information over the whole time series in terms of the time-lagged DI can be simplified as [bib_ref] Time-lagged directed information, Liu [/bib_ref] (the details of the derivation are given in [bib_ref] Time-lagged directed information, Liu [/bib_ref] , [formula] DI X N −→ Y N = p1 n=1 I X n ; Y n | Y n−1 + N − p 1 2 N − p 1 − 1 × N−1 n=p1+1 DI n X n−p1 X n−p1+1 −→ Y n Y n+1 .(17) [/formula] The time-lagged DI is equivalent to the original definition of DI when p 1 is equal to the actual time delay of the system, the signals X and Y follow a single-order model, and Y n only depends on one past sample of itself, Y n−1 . However, these assumptions are not always true. Therefore, we propose the modified time-lagged DI to address these issues. Consider a general Markov model, where X N and Y N are time series with a lag of p [bib_ref] Large-scale cortical networks and cognition, Bressler [/bib_ref] [bib_ref] Organization, development and function of complex brain networks, Sporns [/bib_ref] is the order of the process X, and p 3 is the order of the process Y. In this model, it is assumed that X starts to influence Y with a delay of p 1 samples, and the order of the model is p 2 − p 1 + 1. When the length of the signal N is large enough, then [bib_ref] Connectivity analysis of somatosensory evoked potentials to noxious intracutaneous stimuli in patients..., Leistritz [/bib_ref] can be further simplified as [formula] and p(Y n | X 1:n−p1 , Y p1+1:n−1 ) = p(Y n | X n−p2:n−p1 , Y n−p3:n−1 ), where p 2 ≥ p 1 , p 3 ≥ 1, pDI X N −→ Y N = DI X 1:N−p1 −→ Y p1+1:N = N n=p1+1 I X 1:n−p1 , Y n | Y p1+1:n−1 = N n=p1+1 H Y n | Y p1+1:n−1 −H Y n | X 1:n−p1 Y p1+1:n−1 .(18) [/formula] Since [formula] p(Y n | X 1:n−p1 , Y p1+1:n−1 ) = p(Y n | X n−p2:n−p1 , Y n−p3:n−1 ), X 1:n−p2−1 Y p1+1:n−p3−1 → X n−p2:n−p1 , Y n−p3:n−1 → Y n [/formula] follows a Markov chain. According to Markov chain property, [formula] I X 1:n−p2−1 Y 1:n−p3−1 ; Y n | X n−p2:n−p1 Y n−p3:n−1 = H Y n | X n−p2:n−p1 Y n−p3:n−1 − H Y n | X 1:n−p1 Y p1+1:n−1 = 0,(19) [/formula] which means H(Yn|Xn−p 2 :n−p 1 Yn−p 3 :n−1 )=H(Yn|X1:n−p 1 Yp 1 +1:n−1 ) . Therefore, [formula] DI X N −→ Y N = N n=p1+1 H Y n | Y p1+1:n−1 −H Y n | X 1:n−p1 Y p1+1:n−1 = N n=p1+1 H Y n | Y p1+1:n−1 −H Y n | X n−p2:n−p1 Y n−p3:n−1 ≤ N n=p1+1 H Y n | Y n−p3:n−1 −H Y n | X n−p2:n−p1 Y n−p3:n−1 = N n=p1+1 I X n−p2:n−p1 ; Y n | Y n−p3:n−1 ,(20) [/formula] where the second equality is using the Markov property, and the inequality comes from the fact that conditioning reduces entropy. For a general Markov model, where X N and Y N are stationary statistical processes without instantaneous interaction, such as p(Y n | X 1:n−p1 , Y p1+1:n−1 ) = p(Y n | X n−p2:n−p1 , Y n−p3:n−1 ), the modified time-lagged directed information (MDI) is defined as the upper bound of DI: [formula] MDI X N −→ Y N = N n=p+1 I X n−p · · · X n−1 ; Y n | Y n−p · · · Y n−1 ,(21) [/formula] where we let p 1 = 1, p = max(p 2 , p 3 ) to reduce the number of parameters. Note that letting p 1 = 1 does not lose any of the information flow compared to using the actual time delay, p 1 > 1. The only drawback of letting p 1 = 1 is that the computational complexity of estimating the joint entropies increases since the length of the window to compute MDI increases and the dimensionality increases. The main reason why we let p 1 = 1 is because estimating the actual value for the delay accurately is not practical when the amount of data is limited. In a lot of similar work such as in [bib_ref] Transfer entropy-a model-free measure of effective connectivity for the neurosciences, Vicente [/bib_ref] , different values of p 1 are tested to choose the best one which is not computationally efficient either. According to [bib_ref] Measuring information transfer, Schreiber [/bib_ref] , modified time-lagged directed information is the upper bound of directed information, that is, MDI ≥ DI. Moreover, MDI is a more general extension of time-lagged DI introduced in our previous work and has two major advantages. First, MDI considers the influence of multiple past samples of Y on the DI value. Second, it takes into account models with multiple orders; that is, Y is influenced by different time lags of X. The modified time-lagged directed information extends the length of the window from 2 to p, which is closer to the actual information flow. When X and Y are normally distributed, the computational complexity of the MDI is O(p 3 N) and the computational complexity of the original definition of DI is O(N 4 ) (using LU decomposition. Therefore, the computation of MDI is more efficient than that of the original definition of DI. ## Order selection. For the implementation of MDI, we need to determine the maximum order of the model p. Criterions such as Akaike's final prediction error (FPE) can be used to determine the order of the signal model p. However, this criterion is based on the assumption that the original signal follows a linear AR model and may lead to false estimation of the order when the underlying signal model is nonlinear. Therefore, model-free order selection methods, such as the embedding theorem, are needed. For the simplification of computation or parameter estimation, we are only interested in a limited number of variables that can be used to describe the whole system. Suppose we have a time series (X 1 , . . . , X n ), and the time-delay vectors can be reconstructed as (X n , X n−τ , X n−2τ , . . . , X n−(d−1)τ ). Projecting the original system to this lower-dimensional state spacedepends on the choice of d and τ, and the optimal embedding dimension d is related to the order of the model p = d [bib_ref] Transfer entropy-a model-free measure of effective connectivity for the neurosciences, Vicente [/bib_ref]. A variety of measures such as mutual information can be used to determine τ. For discrete time signals, usually the best choice of τ is 1 [bib_ref] Practical method for determining the minimum embedding dimension of a scalar time..., Cao [/bib_ref]. To determine d, Cao criterion based on the false nearest neighbor procedure [bib_ref] Transfer entropy-a model-free measure of effective connectivity for the neurosciences, Vicente [/bib_ref] is used to determine the local dimension. The underlying concept of nearest neighbor is that if d is the embedding dimension of a system, then any two points that stay close in the ddimensional reconstructed space are still close in the (d + 1)dimensional reconstructed space; otherwise, these two points are false nearest neighbors [bib_ref] Transfer entropy-a model-free measure of effective connectivity for the neurosciences, Vicente [/bib_ref] [bib_ref] Practical method for determining the minimum embedding dimension of a scalar time..., Cao [/bib_ref]. The choice of d, that is, the model order p, is important for DI estimation. If d is too small, we will lose some of the information flow from X to Y. If it is too large, the computational complexity of MDI will be very high, causing the bias and the variance of the estimators to increase. ## Normalization and significance [formula] Test. Since DI(X N → Y N ) + DI(Y N → X N ) = I(X N ; Y N ) + DI(X N → Y N ||DX N ) and DI(X N → Y N ) = DI(DX N → Y N ) + DI(X N → Y N ||DX N ) [29], then DI X N −→ Y N + DI Y N −→ X N = DI DX N −→ Y N + DI X N −→ Y N DX N + DI DY N −→ X N + DI Y N −→ X N DY N .(22) [/formula] Therefore, [formula] DI DX N −→ Y N + DI DY N −→ X N + DI Y N −→ X N DY N = I X N ; Y N ,(23)where DI(Y N → X N ||DY N ) = DI(X N → Y N ||DX N ) [/formula] indicating the instantaneous information exchange between processes X and Y. For a physical system without instantaneous causality, that is, [formula] I(X N → Y N ||DX N ) = 0, then DI(X N → Y N ) + DI(Y N → X N ) = I(X N ; Y N ) and 0 ≤ DI(X N → Y N ) ≤ I(X N ; Y N ) < ∞. Aρ DI X N −→ Y N = DI X N −→ Y N I(X N ; Y N ) = DI X N −→ Y N DI(X N −→ Y N ) + DI(Y N −→ X N ) ,(24) [/formula] where for a unidirectional system X → Y with no instantaneous interaction between X and Y, ρ DI (X N → Y N ) = 1 and ρ DI (Y N → X N ) = 0; otherwise, if there is no causal relationship between the two signals, the values of ρ DI (X N → Y N ) and ρ DI (Y N → X N ) are very close to each other. In order to test the null hypothesis of noncausality, the causal structure between X and Y is destroyed. For each process with multiple trials, we shuffle the order of the trials of the time series X 100 times to generate new observations X * m , m = 1, . . . , 100. In this way, the causality between X and Y for each trial is destroyed, and the estimated joint probability changes [bib_ref] Non-parametric significance estimation of joint-spike events by shuffling and resampling, Pipa [/bib_ref]. We compute the DI for each pair of data (X * m and Y). A threshold is obtained at a α = 0.05 significance level such that 95% of the directed information for randomized pairs of data (DI(X * m → Y)) is less than this threshold. If the DI value of the original pairs of data is larger than this threshold, then it indicates there is significant information flow from X to Y. ## Simulated data. To test the validity and to evaluate the performance of DI for quantifying the effective connectivity, we generate five different simulations. We use these simulation models to compare DI with classical Granger causality (GC) for quantifying causality of both linear and nonlinear autoregressive models, linear mixing models, single source models, and Lorenz systems. The Matlab toolbox developed by Seth is used to compute the GC value in the time domain. GC is also normalized to the [0, 1] range for comparison purposes [bib_ref] A MATLAB toolbox for Granger causal connectivity analysis, Seth [/bib_ref]. The performance of GC depends on the length of the signal, whereas the performance of DI relies on the number of realizations of time series. Therefore, for each simulation, the length of the generated signal for implementing GC is equal to the number of realizations for DI. The significance of DI values are evaluated by shuffling along the trials, while the significance of GC values are evaluated by shuffling along the time series. Example 1 (Multiple Order Bivariate Linear Autoregressive Model). In this example, we evaluate the performance of DI on a general bivariate linear model, [formula] X(n) = p4 i=1 α i X(n − i) + σ x η x (n − 1),(25)Y (n) = p3 i=1 β i Y (n − i) + γ p2 i=p1 X(n − i) + σ y η y (n − 1).(26) [/formula] Computational and Mathematical Methods in Medicine 7 In this bivariate AR model with a delay p 1 and order p 2 − p 1 + 1, γ controls the coupling strength between the signals X and Y. The initial values of X and Y and the noise η x and η y are all generated from a Gaussian distribution with mean 0 and standard deviation 1. All coefficients (α i , β i , σ x , and σ y ) are generated from Gaussian distributions with zero mean and unit variance with unstable systems being discarded. To evaluate the performance of directed information, we generate the bivariate model 4096 times with the same parameters but different initial values. γ is varied from 0.1 to 1 with a step size of 0.1, p 1 = 1 and p 2 = p 3 = p 4 = 5; that is, Y is influenced by X through multiple time lags. Without loss of generality, we repeat the simulation 10 times, and average DI(X N → Y N ) and DI(Y N → X N ) over 10 simulations for different γ values. For each simulation, the threshold is evaluated by trial shuffling, and the average threshold is obtained. For GC, the length of the generated signal is chosen as 4096, which is the same as the number of realizations for DI. The GC values in two directions and the corresponding thresholds at the 5% significance level are obtained. Example 2 (Multiple-Order Bivariate Nonlinear Autoregressive Model). In this example, we evaluate the performance of DI on a general bivariate nonlinear model [formula] X(n) = p4 i=1 α i X(n − i) + σ x η x (n − 1),(27)Y (n) = p3 i=1 β i Y (n − i) + γ p2 i=p1 1 1 + exp(b 1 + b 2 X(n − i)) + σ y η y (n − 1).(28) [/formula] For this bivariate nonlinear AR model, the setting for the coupling strength γ and the generation of X, Y, η x , η y , α i , β i , σ x , σ y , p 1 , p 2 , p 3 , and p 4 are the same as in Example 1. Y and X interact nonlinearly through the sigmoid function. Parameters of this function b 1 and b 2 control the threshold level and slope of the sigmoidal curve, respectively. We set b 1 = 0 and b 2 = 50. DI value and its threshold are averaged over 10 simulations for different γ. The GC values in two directions and the corresponding thresholds at 5% significance level are obtained. Example 3 (Linear Mixing Model). In this example, we test the effectiveness of DI in inferring effective connectivity when there is linear mixing between these two signals. Linear instantaneous mixing is known to exist in human noninvasive electrophysiological measurements such as EEG or MEG. Instantaneous mixing from coupled signals onto sensor signals by the measurement process degrades signal asymmetry [bib_ref] Transfer entropy-a model-free measure of effective connectivity for the neurosciences, Vicente [/bib_ref]. Therefore, it is hard to detect the causality between the two signals. For unidirectional coupled signal pairs X → Y described in [bib_ref] Information flow and application to epileptogenic focus localization from intracranial EEG, Sabesan [/bib_ref] to [bib_ref] The bidirectional communication theory-a generalization of information theory, Marko [/bib_ref] , we create two linear mixtures X and Y as follows: [formula] X (n) = (1 − )X(n) + Y (n), Y (n) = X(n) + (1 − )Y (n),(29) [/formula] where controls the amount of linear mixing and is varied from 0.05 to 0.45 with a step size of 0.05, and γ is fixed to 0.8 for both models. When = 0.5, the two signals are identical. Both DI and GC are used to quantify the information flow between X and Y in the two directions. Example 4 (Single-Source Model). A single source is usually observed on different signals (channels) with individual channel noises [bib_ref] Transfer entropy-a model-free measure of effective connectivity for the neurosciences, Vicente [/bib_ref] , which is common in EEG signals due to the effects of volume conduction. In this case, false positive detection of effective connectivity occurs for methods such as Granger causality [bib_ref] Robustly estimating the flow direction of information in complex physical systems, Nolte [/bib_ref] , which means that GC has low specificity. We generate two signals X and Y as follows to test the specificity of DI when there is no significant information flow from one signal to the other signal. We have [formula] S(n) = p4 i=1 α i S(n − i) + η S (n), X (n) = S(n), Y (n) = (1 − )S(n) + η Y (n),(30) [/formula] where S(n) is the common source generated by an autoregressive model, order p 4 = 5, α i and η S (n) are generated from a Gaussian distribution with mean 0 and standard deviation 1. S(n) is measured on both sensors X and Y . Y is further corrupted by independent Gaussian noise η Y (n) with 0 mean and unit variance. controls the signal to noise ratio (SNR) in Y and is varied from 0.1 to 0.9 with a step size of 0.1, corresponding to SNR in the range of −19 ∼ 19 dB. Example 5 (Nonlinear Dynamic System). In this example, we illustrate the applicability of DI to coupled Lorenz oscillators with a certain delay. The Lorenz oscillator is a threedimensional dynamic system that exhibits chaotic behavior. Synchronization of two Lorenz systems has been widely investigated for the analysis of EEG data because the dynamic interactions related to the behavior of the cortex can be exemplified by these coupled systems [bib_ref] Nonlinear phase desynchronization in human electroencephalographic data, Breakspear [/bib_ref]. In the following, we examined two asymmetric coupled Lorenz oscillators (X 1 , Y 1 , Z 1 ) and (X 2 , Y 2 , Z 2 ) as follows [bib_ref] Synchronization of delaycoupled nonlinear oscillators: an approach based on the stability analysis..., Michiels [/bib_ref] : [formula] X 1 (t) = −A(X 1 (t) − Y 1 (t)), Y 1 (t) = RX 1 (t) − Y 1 (t) − X 1 (t)Z 1 (t), Z 1 (t) = X 1 (t)Y 1 (t) − BZ 1 (t), X 2 (t) = −A(X 2 (t) − Y 2 (t)) + βX 1 t − t p , Y 2 (t) = RX 2 (t) − Y 2 (t) − X 2 (t)Z 2 (t), Z 2 (t) = X 2 (t)Y 2 (t) − BZ 2 (t),(31) [/formula] ## Computational and mathematical methods in medicine where each equation is a first-order differential equation. A = 10, R = 28, B = 8/3, and t p = 0.02 represents the time delay between two coupled components of these two oscillators, that is, X 1 and X 2 . β corresponds to the coupling strength and is varied from 0.1 to 1 with a step size of 0.2. The differential equations are numerically integrated with a time step of 0.01 using Euler's method, corresponding to a delay of 2 time samples between X 1 and X 2 . The initial conditions of these six components are randomly generated from a Gaussian distribution with zero mean and unit variance. We generate 100 samples, and the first 90 samples are discarded to eliminate the initial transients. We compute the information flow in two directions over 10 time points, and the significance of the obtained DI value is verified by trial shuffling. ## Biological data. In this paper, we examine EEG data from ten undergraduates at Michigan State University drawn from an ongoing study of relationships between the errorrelated negativity (ERN) and individual differences (Participants for the present analysis were drawn from samples reported on in [bib_ref] Mind your errors: evidence for a neural mechanism linking growth mindset to..., Moser [/bib_ref] [bib_ref] Parsing relationships between dimensions of anxiety and action monitoring brain potentials in..., Moser [/bib_ref] such as worry and anxiety. ERN is a brain potential response that occurs following performance errors in a speeded reaction time task [bib_ref] Effects of noise letters upon the identification of a target letter in..., Eriksen [/bib_ref]. All participants retained for analysis make at least six errors for computation of stable ERNs, as in [bib_ref] The stability of error-related brain activity with increasing trials, Olvet [/bib_ref]. Participants complete a letter version of the Eriksen Flanker task [bib_ref] Effects of noise letters upon the identification of a target letter in..., Eriksen [/bib_ref]. Stimuli are presented on a Pentium R Dual Core computer, using Presentation software (Neurobehavioral systems, Inc.) to control the presentation and timing of stimuli, the determination of response accuracy, and the measurement of reaction times. Continuous electroencephalographic activity is recorded by 64 Ag-AgCl electrodes placed in accordance with the 10/20 system. Electrodes are fitted in a BioSemi (BioSemi, Amsterdam, The Netherlands) stretch-lycra cap. All bioelectric signals are digitized at 512 Hz using ActiView software (BioSemi). For each subject, EEG data are preprocessed by the spherical spline current source density (CSD) waveforms to sharpen event-related potential (ERP) scalp topographies and eliminate volume conduction [bib_ref] Principal components analysis of Laplacian waveforms as a generic method for identifying..., Kayser [/bib_ref]. In addition, a bandpass filter is used to obtain signals in the theta band. In this study we focus on 33 electrodes corresponding to the frontal, central, and parietal regions of the brain. For each pair of 33 electrodes X and Y for each subject, the effective connectivity is quantified by computing the modified time-lagged DI over 70 trials and a model order of p in the theta band. The model order or the length of the time window p is determined by the Cao Criterion. We also apply Granger causality to the same data and compare its performance with directed information. Previous work indicates that there is increased synchronization associated with ERN for the theta frequency band [bib_ref] Functional and effective connectivity in neuroimaging: a synthesis, Friston [/bib_ref] [bib_ref] Nonlinear multivariate analysis of neurophysiological signals, Pereda [/bib_ref] [bib_ref] Dynamic causal modelling, Friston [/bib_ref] [bib_ref] Testing for causality. A personal viewpoint, Granger [/bib_ref] and ERN time window 25-75 ms after the response for error responses (ERN) in the anterior cingulate cortex (ACC), in particular between the lateral prefrontal cortex (lPFC) and medial prefrontal cortex (mPFC) [bib_ref] A phase synchrony measure for quantifying dynamic functional integration in the brain, Aviyente [/bib_ref]. In this paper, we wish to verify these existing findings using the proposed DI measure and to further infer the directional causality underlying these dependencies. # Results and discussion In this section, we first evaluate the effectiveness of directed information on quantifying both linear and nonlinear causal relationships through simulated data and compare the performance of directed information with GC. We then apply the directed information to real EEG data to reveal the pairwise information flow in the brain. ## Simulated data Example 1 (Multiple-Order Bivariate Linear Autoregressive Model). In this example, the DI value in two directions averaged across 10 simulations with different γ is shown in [fig_ref] Figure 1: Application of directed information and Granger causality to bivariate linear autoregressive model [/fig_ref]. The performance of GC is shown in [fig_ref] Figure 1: Application of directed information and Granger causality to bivariate linear autoregressive model [/fig_ref]. The estimated order of the model is p = 5, which is in accordance with the simulation model. γ controls the coupling strength between X and Y. We observe that DI(X N → Y N ) is significant for all values of γ. On the contrary, DI(Y N → X N ) is less than the threshold, which indicates the acceptance of the null hypothesis that there is no significant causal information flow from Y to X. Since GC uses a linear autoregressive framework for quantifying causality; in this example, GC detects the causality relationship between X and Y successfully; that is, the information flow from X to Y is significant for all γ while it is insignificant for the opposite direction. It is also interesting to note that GC and DI exhibit similar behavior across different values of γ, indicating the equivalency of the two measures for linear Gaussian signal models. Example 2 (Multiple-Order Bivariate Nonlinear Autoregressive Model). In this example, the performance of DI and GC for the nonlinear autoregressive model in [bib_ref] The relationship between two directed information measures, Al-Khassaweneh [/bib_ref] and [bib_ref] The bidirectional communication theory-a generalization of information theory, Marko [/bib_ref] averaged across 10 simulations with different γ are evaluated as shown in [fig_ref] Figure 2: Application of directed information and Granger causality to bivariate nonlinear autoregressive model [/fig_ref]. The estimated order of the model is 5. We observe that when γ is less than 0.3, the coupling strength between X and Y is weak and the DI value in both directions is not significant. As γ increases, DI(X N → Y N ) increases and becomes significant. DI(Y N → X N ) decreases with increasing γ and is still less than the threshold as expected. The results indicate increased unidirectional information flow from X to Y with increasing γ and show that detecting the information flow in nonlinear processes is more difficult especially when the coupling strength is low. GC fails to detect the information flow from X to Y for all γ. Since GC is implemented in a linear framework, the estimated order and the model itself do not match with the nonlinearity of the signal. Therefore, it cannot detect nonlinear causality. Example 3 (Linear Mixing Model). For this example, the DI value and GC value averaged across 10 simulations with changing linear mixing coefficient for both linear and nonlinear AR models are shown in [fig_ref] Figure 3: Application of directed information and Granger causality to linear mixing for both... [/fig_ref]. The estimated order of the model is 5 as before. When = 0.5, the two observed mixing signals are identical, and we expect to see no significant information flow in the two directions. We observe that, for the linear AR model, directed information detects the causality between X and Y when is smaller than 0.4. When is larger than 0.4, the causality between X and Y is hard to detect because of the strong mixing; that is, X and Y are almost identical, and the information flow in both directions becomes insignificant. Compared to DI, GC only detects the causality from X to Y when the mixing is weak ( < 0.2), indicating that GC is more vulnerable to linear mixing. It is probably due to the fact that GC is sensitive to the mixture of signals, and the assumed signal model does not match with the original signal [bib_ref] Robustly estimating the flow direction of information in complex physical systems, Nolte [/bib_ref]. For the nonlinear AR model, DI fails to detect causality when is larger than 0.1, which indicates that linear mixing of nonlinear source models makes it harder to detect effective Example 4 (Single-Source Model). We use the single source model to test the specificity of DI. The DI value and GC value averaged across 100 simulations for changing for a single source model are shown in [fig_ref] Figure 4: Application of directed information and Granger causality to single source model [/fig_ref]. The estimated order of the model is 5 as before. In addition, the false positive rate using both DI and Granger causality with increasing is also calculated. We observe that the information flow in two directions using DI is less than the threshold for all values of , which indicates the acceptance of the null hypothesis that there is no significant causal information flow from X to Y or Y to X. Note that DI is normalized by the mutual information. For a common source model, the instantaneous information exchange between X and Y contributes mostly to the mutual information between X and Y. Thus, according to (23) mutual information are close to 0 and less than the threshold from the randomized data pairs. The false positive rate of DI is 0 for all . Therefore, DI is able to discriminate between instantaneous mixing from actual causality and is very robust to noise. For GC, when is small (<0.2) or large (>0.9), the value of GC is less than or very close to the threshold in both directions thus indicating that there is no causal information flow between the two processes. However, GC fails to accept the null hypothesis when is between 0.3 to 0.9 and detects a nonexisting effective connectivity. GC reaches its maximum value when = 0.5. This is due to the fact that GC is close to 0 when two processes X and Y are independent or identical, that is, when = 1 and = 0. Based on the definition of GC, the prediction of Y at the current time point will not be improved by taking into account the past samples of X for these processes [bib_ref] Transfer entropy in magnetoencephalographic data: quantifying information flow in cortical and cerebellar..., Wibral [/bib_ref]. Therefore, as increases from 0 to 0.5, X becomes the most different from Y; therefore, it can provide more new information about Y and the GC increases. As increases from 0.5 to 1, X becomes independent of Y, and the GC decreases. The false positive rate of GC is not equal to 0 for all values of , which indicates that it has lower specificity compared to DI. Therefore, GC is not robust to the effect of a common source and may infer false positive effective connectivity. This simulation indicates that DI is more sensitive and discriminative about the information flow patterns in the presence of volume conduction, which means it is a more promising method to capture the effective connectivity for real EEG data. [formula] γ X → Y X → Y Y → X Y → X (a)γ X → Y X → Y Y → X Y → Xγ X → Y X → Y Y → X Y → Xγ X → Y X → Y Y → X Y → XX → Y X → Y Y → X Y → X ɛ (a)X → Y X → Y Y → X Y → X ɛ (b)X → Y X → Y Y → X Y → X ɛ (c)X → Y X → Y Y → X Y → X ɛ (d)X → Y X → Y Y → X Y → X ɛ (a)X → Y X → Y Y → X Y → X ɛ (b)X → Y X → Y Y → X Y → X(X → Y X → Y Y → X Y → X (b) [/formula] Example 5 (Nonlinear Dynamic System). In this example, the DI values and GC values between X 1 and X 2 of two asymmetric coupled Lorenz systems are computed with coupling strength β being set from 0.1 to 1. The estimated order of the model is 3. Though this is larger than the actual model order, our method will not lose any information except for the increased computational complexity. The results are shown in [fig_ref] Figure 5: Application of directed information and Granger causality to two asymmetric coupled Lorenz... [/fig_ref]. The results show that DI values from X 1 to X 2 increase with the coupling strength β and are significant for all values of β. In addition, there is no significant causal information flow from X 2 to X 1 . Therefore, DI can effectively detect the causality in a nonlinear dynamic system. On the contrary, GC cannot detect any significant information flow for all β values. It is due to the fact that the model selected for implementing GC is not consistent with the dynamic characteristics of the system. ## Eeg data. Previous work indicates that there is increased information flow associated with ERN for the theta frequency band [bib_ref] Functional and effective connectivity in neuroimaging: a synthesis, Friston [/bib_ref] [bib_ref] Nonlinear multivariate analysis of neurophysiological signals, Pereda [/bib_ref] [bib_ref] Dynamic causal modelling, Friston [/bib_ref] [bib_ref] Testing for causality. A personal viewpoint, Granger [/bib_ref] and ERN time window 25-75 ms for error responses compared to correct responses in particular between mPFC and lPFC regions [bib_ref] A phase synchrony measure for quantifying dynamic functional integration in the brain, Aviyente [/bib_ref]. In addition, Cavanagh et al. have shown that there is increased synchronization for error trials between electrode pairs, such as FCz-F5 and FCz-F6, compared to the synchrony between FCz-CP3 and FCz-CP4 [bib_ref] Prelude to and resolution of an error: EEG phase synchrony reveals cognitive..., Cavanagh [/bib_ref]. The DI and GC values for each pair of electrodes averaged over 10 subjects are computed over a time window of 53 time points (100 ms). The estimated order of the model for each electrode pairs is 3. In order to control the error rates for multiple hypothesis testing for all pairs of electrodes, the method proposed by Genovese et al. is used in this paper [bib_ref] Thresholding of statistical maps in functional neuroimaging using the false discovery rate, Genovese [/bib_ref]. To implement this procedure, for two electrodes with time series X and Y, we first shuffle the order of the trials of X 100 times to generate new observations X * m , m = 1, . . . , 100. The P value of DI(X → Y) is obtained by comparing it with DI values from randomized pairs of data DI(X * m → Y), m = 1, . . . , 100. We then obtain the threshold P r for all P values (33 × 33 × 10) by controlling the FDR bound q as 0.05. For DI(X → Y), if the P value is less than P r , then the directed information flow from X to Y is significant; otherwise, it is not significant. Electrode pairs between which the information flow is significant in at least one of the ten subjects are shown in [fig_ref] Figure 6: Application of directed information and Granger causality to EEG data [/fig_ref]. We also test the significance of Granger causality in the same way. When the FDR is controlled at 0.05, the information flow between electrode pairs is significant if the P-value of DI or GC is less than 0.01. Electrode pairs that have significant causality relationship using both measures are shown in [fig_ref] Figure 6: Application of directed information and Granger causality to EEG data [/fig_ref]. In [fig_ref] Figure 6: Application of directed information and Granger causality to EEG data [/fig_ref] and 6(c), each small circle shows the directed information and Granger causality from a particular electrode to other electrodes. In [fig_ref] Figure 6: Application of directed information and Granger causality to EEG data [/fig_ref] and 6(d), each small circle shows electrode pairs that have significant causality relationship. The results indicate that DI detects strong information flow from the frontal region (e.g., F5 and F6) to the frontal-central region (e.g., FC2 and FCz) corresponding to the lateral prefrontal cortex (lPFC) and medial prefrontal cortex (mPFC). In addition, the central-parietal region (e.g., CPz, CP1, and CP2) around the midline, corresponding to the motor cortex, has strong influence on the central and frontal regions (e.g., FCz and F6) since this is a speeded response task involving the motor cortex. The details of the significant electrode interactions are shown in [fig_ref] Table 1: Electrode pairs in the region of interest with significant DI values [/fig_ref]. These results are aligned with the previous work in [bib_ref] Prelude to and resolution of an error: EEG phase synchrony reveals cognitive..., Cavanagh [/bib_ref] , which shows that error processing is controlled by the communication between the lateral prefrontal cortex and medial prefrontal cortex. When GC is applied to the same data, the information flow pattern around the midline is similar to the DI. However, the information flow from the lateral prefrontal cortex to the rest of the brain is significant. On one hand, the similar patterns of connectivity using both measures verify the validity of proposed DI computation algorithm. On the other hand, GC shows significance over a wide region of the brain especially in the lateral areas compared to DI, which may be due to GC's low specificity to volume conduction in the form of a common source. Previous work and our simulation in Example 4 have indicated that Granger-causality-based measures may infer erroneous effective connectivity in the case of the common source as seen in EEG data [bib_ref] Transfer entropy-a model-free measure of effective connectivity for the neurosciences, Vicente [/bib_ref] [bib_ref] Robustly estimating the flow direction of information in complex physical systems, Nolte [/bib_ref]. However, without ground truth, we cannot confirm that some links reported as significant by GC are spurious and due to volume conduction in a conclusive manner, but the results from DI agree more with the suggestions in [bib_ref] Prelude to and resolution of an error: EEG phase synchrony reveals cognitive..., Cavanagh [/bib_ref] , that most of the increase in connectivity during cognitive control, that is, ERN, should be between medial prefrontal cortex and lateral prefrontal cortex, compared to the results of GC. Therefore, DI is more sensitive and discriminative about the information flow patterns compared to GC for real neurophysiological data. # Conclusions In this paper, we illustrated the advantages of a new directed information measure over Granger-causality-based measures for quantifying the effective connectivity in the brain. In order to illustrate the advantages of this measure, first, we applied directed information measure to identify the causality relationships for both linear and nonlinear AR models, linear mixing models, single source models, and Lorenz systems and compare its performance with Granger causality. Directed information is shown to be more effective in detecting the causality of different systems compared to Granger causality. We then applied the directed information measure on EEG data from a study containing the error-related negativity to infer the information flow patterns between different regions. The results showed that the directed information measure can capture the effective connectivity in the brain between the mPFC and lPFC areas as predicted by previous work. Directed information, as a model-free measure, is able to detect both linear and nonlinear causality relationships between two signals. However, other model-free entropybased measures would also detect effective connectivity such as transfer entropy and directed transinformation. Directed transinformation introduced by Saito measures the information flow from the current sample of one signal to the future samples of another signal given the past samples of both signals but does not discriminate between totally dependent and independent processes. Transfer entropy and directed information are very closely related to each other. Transfer entropy quantifies the information gained at each time step by measuring the deviation of the observed data from the generalized Markov condition. Therefore, the definition of transfer entropy implicitly assumes a stationary Markov process [bib_ref] On directed information theory and Granger causality graphs, Amblard [/bib_ref]. Compared to transfer entropy, directed information quantifies the sum of information obtained over the whole time seriesand does not make any assumptions about the underlying signal model. Thus, theoretically, the original definition of directed information can apply to any signal models. In real applications, in order to simplify the computation of directed information, we usually make certain assumptions about the underlying signal model such as the modified time-lagged DI proposed in this paper, which basically assumes a stationary Markov process similar to transfer entropy. In addition, Amblard and Michel proved that, for a stationary process, directed information rate can be decomposed into two parts, one of which is equivalent to the transfer entropy when l = m = n in (1) and the other to the instantaneous information exchange rate [bib_ref] On directed information theory and Granger causality graphs, Amblard [/bib_ref]. In another words, for a physical system without instantaneous interactions between its subsystems, the rate of these two measures, directed information and transfer entropy, is equivalent asymptotically as the length of the signal goes to infinity. There are still remaining issues with the implementation of directed information. First, the performance of directed information relies on accurate estimation from limited sample sizes that introduces bias to the estimated values. This problem can be addressed by either using parametric density models or improving existing mutual information and entropy estimators. Recently, Zhao et al. proposed an universal algorithm to estimate directed information for stationary ergodic processes by using sequential probability assignment, which may be used to improve the effective connectivity results discussed in this paper [bib_ref] Universal estimation of directed information, Zhao [/bib_ref]. Second, the performance of directed information relies on the selection of the model order. If the order of the model is too small, it will lose the information from X to Y. If it is too large, the computational complexity is very high. In addition to classical embedding dimension determination methods such as the Cao criterion used in this paper, Faes et al. proposed a sequential procedure to determine the embedding dimension of multivariate series [bib_ref] Information-based detection of nonlinear Granger causality in multivariate processes via a nonuniform..., Faes [/bib_ref]. This method is based on an information-theoretic technique and shows promising performances for various signal models, which may be extended to DI computation in the future. Third, directed information does not discriminate between direct and indirect interactions among multivariate time series. However, this is not a shortcoming of DI since DI does not assume any particular signal interaction model: bivariate or multivariate. Similar to other information theoretic measures, such as mutual information, whether the particular measure can identify interactions between multiple processes depends on how the measure is applied. For example, in the case of mutual information, though the original definition is for two random processes X and Y, it is possible to extend it to multiple processes [bib_ref] Inferring connectivity of genetic regulatory networks using informationtheoretic criteria, Zhao [/bib_ref]. Similarly, we can apply DI over multiple processes using conditional directed information such as the definition given by Kramer. We address this issue in a previous paper [bib_ref] Information theoretic approach to quantify causal neural interactions from EEG, Liu [/bib_ref] by using conditional directed information and develop an algorithm to infer the actual network. Similarly, GC originally is defined for two time series that a stochastic process X causing another process Y if the prediction of Y at the current time point, Y n , is improved when taking into account the past samples of X. However, in application it has been extended to multiple processes through the use of multivariate AR models. Future work will focus on the comparison of these two measures in a multivariate setting. [fig] Figure 1: Application of directed information and Granger causality to bivariate linear autoregressive model. (a) Directed information with different γ. (b) Granger causality with different γ. [/fig] [fig] Figure 2: Application of directed information and Granger causality to bivariate nonlinear autoregressive model. (a) Directed information with different γ. (b) Granger causality with different γ. [/fig] [fig] Figure 3: Application of directed information and Granger causality to linear mixing for both linear and nonlinear autoregressive models. (a) Directed information with different for the linear mixing of linear AR model. (b) Granger causality with different for the linear mixing of linear AR model. (c) Directed information with different for the linear mixing of nonlinear AR model. (d) Granger causality with different for the linear mixing of nonlinear AR model. connectivity compared to mixing of linear source models. On the other hand, GC fails to detect any causality even when = 0 since it cannot detect nonlinear interactions. [/fig] [fig] Figure 4: Application of directed information and Granger causality to single source model. (a) Directed information with different for the single source model. (b) Granger causality with different for the single source model. (c) False positive rate for directed information with different for the single source model. (d) False positive rate for Granger causality with different for the single source model. [/fig] [fig] Figure 5: Application of directed information and Granger causality to two asymmetric coupled Lorenz oscillators. (a) Directed information with different β. (b) Granger causality with different β. [/fig] [fig] Figure 6: Application of directed information and Granger causality to EEG data. (a) Pairwise directed information. (b) Electrode pairs with significant DI values. (c) Pairwise Granger causality. (d) Electrode pairs with significant GC values. For (b) and (d), green dots indicate the location of the particular node, and white regions correspond to significant information flow from that particular electrode to other electrodes. [/fig] [table] Table 1: Electrode pairs in the region of interest with significant DI values. [/table]

The microscopic structure of charge density waves in underdoped YBa2Cu3O6.54 revealed by X-ray diffraction Charge density wave (CDW) order appears throughout the underdoped high-temperature cuprate superconductors, but the underlying symmetry breaking and the origin of the CDW remain unclear. We use X-ray diffraction to determine the microscopic structure of the CDWs in an archetypical cuprate YBa 2 Cu 3 O 6.54 at its superconducting transition temperature B60 K. We find that the CDWs in this material break the mirror symmetry of the CuO 2 bilayers. The ionic displacements in the CDWs have two components, which are perpendicular and parallel to the CuO 2 planes, and are out of phase with each other. The planar oxygen atoms have the largest displacements, perpendicular to the CuO 2 planes. Our results allow many electronic properties of the underdoped cuprates to be understood. For instance, the CDWs will lead to local variations in the electronic structure, giving an explicit explanation of density-wave states with broken symmetry observed in scanning tunnelling microscopy and soft X-ray measurements. A charge density wave (CDW) is a periodic modulation of the electron density, associated with a periodic lattice distortion that may or may not be commensurate with the crystal lattice. The charge density modulation may be brought about by electron-phonon or electron-electron interactions [bib_ref] Classification of charge density waves based on their nature, Zhu [/bib_ref]. It is now clear that the CDW state is a ubiquitous phenomenon in cuprate high-critical-temperature (high T c ) superconductors, appearing in the underdoped region in both hole- [bib_ref] Long-range incommensurate charge fluctuations in (Y,Nd)Ba 2 Cu 3 O 6 þ..., Ghiringhelli [/bib_ref] [bib_ref] Distinct charge orders in the planes and chains of ortho-IIIordered YBa 2..., Achkar [/bib_ref] [bib_ref] Direct observation of competition between superconductivity and charge density wave order in..., Chang [/bib_ref] [bib_ref] Momentum-dependent charge correlations in YBa 2 Cu 3 O 6 þ x..., Blanco-Canosa [/bib_ref] [bib_ref] Resonant X-ray scattering study of charge-density wave correlations in YBa 2 Cu..., Blanco-Canosa [/bib_ref] [bib_ref] X-ray diffraction observations of a charge-density-wave order in superconducting ortho-II YBa 2..., Blackburn [/bib_ref] [bib_ref] Competing charge, spin, and superconducting orders in underdoped YBaCuO, Huecker [/bib_ref] [bib_ref] Charge Order Driven by Fermi-Arc Instability in Bi 2 Sr 2-x La..., Comin [/bib_ref] [bib_ref] Ubiquitous interplay between charge ordering and high-temperature superconductivity in cuprates, Da Silva Neto [/bib_ref] [bib_ref] Direct observation of bulk charge modulations in optimally doped Bi 1.5 Pb..., Hashimoto [/bib_ref] [bib_ref] Rotated stripe order and its competition with superconductivity in La 1.88 Sr..., Thampy [/bib_ref] [bib_ref] Charge order and its connection with Fermi-liquid charge transport in a pristine..., Tabis [/bib_ref] [bib_ref] Charge density wave fluctuations in La 2-x Sr x CuO 4 and..., Croft [/bib_ref] and electron-doped [bib_ref] Charge ordering in the electron-doped superconductor Nd 2-x Ce x CuO 4, Da Silva Neto [/bib_ref] materials at a temperature higher than T c , suggesting that the CDW is a characteristic instability of the CuO 2 plane. The CDW competes with superconductivity [bib_ref] Long-range incommensurate charge fluctuations in (Y,Nd)Ba 2 Cu 3 O 6 þ..., Ghiringhelli [/bib_ref] [bib_ref] Distinct charge orders in the planes and chains of ortho-IIIordered YBa 2..., Achkar [/bib_ref] [bib_ref] Direct observation of competition between superconductivity and charge density wave order in..., Chang [/bib_ref] [bib_ref] Magnetic-field-induced charge-stripe order in the high temperature superconductor YBa 2 Cu 3..., Wu [/bib_ref] , and pressure-dependent datasuggest that if the CDW can be suppressed in YBa 2 Cu 3 O y (YBCO), then an enhanced T c occurs in the nominally underdoped region rather than at optimum doping. Experiments on YBCO using resonant soft X-ray scattering suggest that the CDW is associated with significant d-wave components for charges on the oxygen bonds around the Cu site [bib_ref] Symmetry of charge order in cuprates, Comin [/bib_ref] , as proposed by Sachdev [bib_ref] Bond order in two-dimensional metals with antiferromagnetic exchange interactions, Sachdev [/bib_ref]. This conclusion is supported by scanning tunnelling microscopy (STM) observations of the surface of Bi 2 Sr 2 CaCu 2 O 8 þ x and Ca 2 À x Na x CuO 2 Cl 2 (ref. [bib_ref] Direct phase-sensitive identification of a d-form factor density wave in underdoped cuprates, Fujita [/bib_ref]. However, to understand the generic high-T c CDW phenomenon, discovering the actual structure of the CDW is vital. CDWs break the translation symmetry of the parent lattice, and have been observed by X-ray diffraction, and many other probes such as STM [bib_ref] Direct phase-sensitive identification of a d-form factor density wave in underdoped cuprates, Fujita [/bib_ref] and nuclear magnetic resonance [bib_ref] Magnetic-field-induced charge-stripe order in the high temperature superconductor YBa 2 Cu 3..., Wu [/bib_ref] [bib_ref] Incipient charge order observed by NMR in the normal state of YBa..., Wu [/bib_ref]. Signatures of the Fermi surface reconstruction believed to be associated with this include quantum oscillation measurements [bib_ref] Quantum oscillations and the Fermi surface in an underdoped high-T c superconductor, Doiron-Leyraud [/bib_ref] , which show unexpectedly small Fermi surface pockets in an underdoped sample, and transport measurements, which indicate a change from hole carriers in the overdoped region to electron-like transport in the underdoped region [bib_ref] Nernst and Seebeck coefficients of the cuprate superconductor YBa 2 Cu 3..., Chang [/bib_ref]. To relate these observations to the CDW, its actual structure needs to be known. The studies of CDWs by X-ray diffraction in numerous cuprates have generally concentrated on determining the wave vector of the CDW and the temperature and magnetic field dependence of the order parameter and correlation lengths, and therefore have considered only a handful of diffraction satellites arising from the CDW. The only way to determine the structure unambiguously is by measuring the intensities of as many CDW satellites as possible. Here we determine the structure of the CDWs in a bilayer cuprate. The material we have investigated is the well-studied material YBCO at a doping level where there is strong competition between superconductivity and the CDW, and the oxygen ordering in the crystal is most perfect. We find that the ionic displacements associated with the CDWs are maximum near the CuO 2 bilayers and break their mirror symmetry. They involve displacements of planar oxygens perpendicular to the layers; these displacements have a strong component with dsymmetry. These results allow a physical understanding of the changes in electronic structure, transport properties and quantum oscillation results in the normal state of this cuprate material that are associated with the CDWs. # Results X-ray diffraction measurements. We have used non-resonant X-ray diffraction to measure the intensities of all experimentally accessible CDW satellites near the (h, 0, c), (0, k, c) and (h, h, c) planes for both of the CDW modulation vectors, q a ¼ (d a , 0, 0.5) and q b ¼ (0, d b , 0.5). Throughout this paper, we express wave vectors in reciprocal space coordinates (h, k, c), where Q ¼ (ha* þ kb* þ cc*). Here a, b, c are the Cartesian vectors defining the YBCO crystal cell dimensions, and a*, b*, c* are the corresponding vectors in reciprocal space; q is used to denote the full wave vector of a CDW mode and d its basal plane part. By collecting a comprehensive data set, we deduce with great certainty the displacement patterns of the ions in the unit cell and hence the structure of the CDW in YBCO. Our experiment was carried out on an underdoped crystal with the ortho-II structure (meaning that the oxygen sites on alternate CuO chains are unoccupied). The crystal was the same as that used in ref. 7. Ortho-II was selected as it has been well-studied by multiple techniques and the satellites associated with the oxygen ordering have minimal overlap with the CDW satellites. Here d a B0.323, d b B0.328 for this underdoped YBCO crystal. Measurements were made at the superconducting T c of our sample (60 K), where the CDW intensity is a maximum in zero field [bib_ref] Direct observation of competition between superconductivity and charge density wave order in..., Chang [/bib_ref]. CDW signals in high-T c cuprates are observed with basal plane wave vectors along both a and b crystal directions. These modulations may be present in the crystal in separate domains having q a or q b modulation (a 1-q model); alternatively, both modulations could be present and superposed in the same region (a 2-q model). Intensity measurements at separate q a and q b do not interfere, so all qualitative features of the two CDW components that we may deduce from our results are independent of the 1-or 2-q state of the sample, which only affects numerical estimates of the absolute magnitudes of the displacements (by a factor of O2). [fig_ref] Figure 1 |: Typical observations of CDW satellites at 60 K and their temperature dependence [/fig_ref] -c shows some typical scans through CDW diffraction satellites. They peak at half-integral values of c [fig_ref] Figure 1 |: Typical observations of CDW satellites at 60 K and their temperature dependence [/fig_ref] , and are extremely weak (B10 À 7 of a typical crystal Bragg reflection). Therefore, the satellites are measured above a relatively large background, but due to their known position and shape, their intensities can be found and spurious signals ignored (see Methods). A compilation of some of the measured CDW intensities is displayed in [fig_ref] Figure 2 |: Sample data compared with the fits to the two possible models [/fig_ref] ; the area of the red semicircles is proportional to the measured peak intensity. X-rays are sensitive to ionic displacements. Non-resonant X-rays are primarily sensitive to the ionic displacements associated with a CDW, rather than changes in charge densities, although if one of these is present, so must the other [bib_ref] Charge modulations versus strain waves in resonant X-ray scattering, Abbamonte [/bib_ref]. (See Supplementary Note 1 for a simple model.) The CDW order gives rise to very weak diffraction satellites at positions in reciprocal space Q ¼ s ± q around lattice Bragg peaks s which are at integer h, k and c. The diffracted amplitude at wave vector Q due to an ion carrying a total of N electrons displaced by small distance u is BN Q.u. Hence, the variation of the intensities with Q reflects the directions and magnitudes of the different ion displacements throughout the unit cell, and by observing intensities of CDW diffraction signals over a wide range of directions and values of Q we can determine the CDW structure. (See also an illustration of this point in [fig_ref] Figure 1 |: Typical observations of CDW satellites at 60 K and their temperature dependence [/fig_ref] The full theory relating the CDW satellite intensities to the CDW structure is given in Supplementary Note 2. We may write the displacements u j , of the individual ions from their regular positions r 0 j as a sum of two terms, one of which is polarized along c ðu c j Þ and the other (u a j or u b j ) parallel to d, with mirror symmetry about the relevant layer of the crystal. [formula] r j ¼ r 0 j þ u c j cosðd:r 0 j þ jÞ þ u a;b j sinðd:r 0 j þ jÞ;ð1Þ [/formula] Symmetry 27 requires that the u c j and u a;b j displacements are p/2 out of phase, as expressed in equation (1). Ionic displacements obtained from the intensities. Group theory indicates which of the incommensurately modulated displacement patterns or irreducible representations (IRs) are consistent with the observed ordering wave vectors. There are four IRs for each ordering wave vector, labelled A 1 , A 2 , A 3 and A 4 for q a and similarly B 1 -B 4 for q b (ref. [bib_ref] ISODISPLACE: an internet tool for exploring structural distortions, Campbell [/bib_ref]. The even-numbered patterns have purely basal plane transverse displacements, and are therefore incompatible with our observations of satellites close to both the c* axis and the basal plane [bib_ref] Direct observation of competition between superconductivity and charge density wave order in..., Chang [/bib_ref]. The other IRs have longitudinal displacements in the basal plane parallel to d, combined with shear displacements parallel to the c axis. Only the A 1 pattern for q a (and B 1 for q b ) are consistent with our data. These IRs have equal c-axis shear displacements in the two halves of the CuO 2 bilayer region, combined with basal plane compressive displacements (and hence charge density modulations), which are equal and opposite in the two halves of a bilayer. Thus, these CDWs break the mirror symmetry of the bilayers. For these patterns, the CuO chain layer is a mirror plane of the CDW. For the patterns A 3 and B 3 , the yttrium layer is instead a mirror plane of the CDW, so that the basal plane compressive displacements would be equal on the two sides of the bilayer. (See also Supplementary Note 3 and [fig_ref] Figure 2 |: Sample data compared with the fits to the two possible models [/fig_ref] for a visualization of these symmetries.) Some ionic displacements in these IRs are zero by symmetry. This results in a detailed description of an IR that consists of 13 non-zero parameters representing displacement components of the 11 ions in the YBCO unit cell. In our model, we average over the half-occupied chain oxygen site in ortho-II YBCO, because we find no evidence for different responses in those cells having a full or empty CuO chain. (No CDW satellites were observed about s þ ½a* positions). Our most complete data set is for the q b satellites. The q a data are somewhat sparser and the results have larger errors, due to the tails of the peaks arising from the ortho-II oxygen ordering which give large and rapidly varying backgrounds. Models A 1 and B 1 always converged in a few iterations to a good fit and gave the same fitted values of displacements independent of the starting value of the parameters. In contrast, models A 3 and B 3 always gave poor fits (for example, w 2 (B 3 ) 410 Â w 2 (B 1 )), whatever the starting values of the fitting parameters. Sample data and fits are shown in [fig_ref] Figure 2 |: Sample data compared with the fits to the two possible models [/fig_ref] , with complete maps of the intensities measured and the fits in . We therefore conclude that the IRs A 1 and B 1 are close to an accurate representation of the CDW. In [fig_ref] Figure 3 |: Representation of the CDW ionic displacement motifs for an unmodulated unit cell [/fig_ref] , we represent the patterns of ionic displacements in a single unit cell as given by the data for both q a and q b modulations. The overall similarity of the two patterns is apparent. The spatial variation of the ionic displacements, shown in , is derived from the motifs in [fig_ref] Figure 3 |: Representation of the CDW ionic displacement motifs for an unmodulated unit cell [/fig_ref] by modulating the c-axis and basal plane displacements by cos(2pd a x/a) and sin(2pd a x/a) respectively for the q a mode, and similarly for q b . We have obtained an estimate of the absolute magnitude of ion displacements by comparing the satellite intensities with those of the Bragg peaks from the lattice. (See Supplementary Note 4 for details.) The fitted values of the ionic displacements are given in [fig_ref] Table 1 |: The values of fitted ionic displacements [/fig_ref]. The table gives signs, values and errors for the fitted displacements from models A 1 and B 1 for the q a and q b modes. Certain features of our fit, such as the c-axis motion of the yttrium layer moving with both the CuO 2 planes are expected on physical grounds, but were not imposed as constraints, giving extra confidence in the fit results. Most ionic displacement values are well-defined, but some oxygen horizontal displacements have large errors, particularly, that of the chain oxygen, which gives a small scattering amplitude, since it has a small number of electrons, is half-occupied and its amplitude falls off at large Q. The first pair of columns show the results for the q b mode if the chain oxygen displacement is left free. It refines to a very large value with an even larger uncertainty. In the next pair of columns this displacement is set to a physically reasonable value equal to the small displacement of the adjacent Cu. It can be seen that this makes very little difference to the fitted values of the other ion displacements and to the w 2 . (A small value of the chain oxygen displacement is justified on the grounds that soft X-ray measurements 3 indicate very small CDW charge build-up in the chain layer.) # Discussion Earlier measurements found differences in the magnitudes of the X-ray signals from the q a and q b modes 6,7 that suggested that the CDW in ortho-II YBCO might be essentially single-q, and dominated by the q b mode. The results presented here indicate that this is not the case; the two modes have similar displacement amplitudes, but the value of their ratio depends on which ion is chosen to make the comparison. For instance, if we consider the motion of the CuO 2 plane oxygens as key, we find that the relative c-motion of these oxygen ions is essentially identical for the two modes: in both cases, the amplitude is B4-5 Â 10 À 3 Å. Our fits do show differences in the heavy ion displacements, and even if these are small, they can make noticeable contributions to the X-ray signals because these ions carry many core electrons. Only this complete survey, rather than measurements of the intensities of a few q a and q b satellites, can reveal the similarities and differences between the two CDW modes. | Representation of modulated ionic displacements for the CDW modes. The displacements are B10 À 3 of interatomic distances, so have been exaggerated to make them visible. In a is shown the spatial variation of the fitted ionic displacements for the A 1 CDW mode with wave vector q a , and in (b) the fitted displacements for the B 1 CDW mode with wave vector q b . The shaded planes passing through the CuO chain layers are the mirror planes of the CDWs. If the structure of the CDW is 1-q, these displacement patterns would be located in different regions of the crystal. If 2-q, the total displacement of the ions in the crystal would be the sum of those associated with the q a and q b modulation vectors. (c) The displacement of any particular ion in a CDW lies on an ellipse: we give an example for an oxygen in the lowest CuO 2 plane of a. The deduced ionic displacements are maximal near the CuO 2 planes and weak near the CuO chains; this is in agreement with the observed competition between the CDW and superconductivity [bib_ref] Long-range incommensurate charge fluctuations in (Y,Nd)Ba 2 Cu 3 O 6 þ..., Ghiringhelli [/bib_ref] [bib_ref] Distinct charge orders in the planes and chains of ortho-IIIordered YBa 2..., Achkar [/bib_ref] [bib_ref] Direct observation of competition between superconductivity and charge density wave order in..., Chang [/bib_ref] [bib_ref] Magnetic-field-induced charge-stripe order in the high temperature superconductor YBa 2 Cu 3..., Wu [/bib_ref]. Surprisingly, the largest amplitudes are out-of-plane shear rather than compression of the CuO 2 planes, so that the CDW is not purely a separation of charge as commonly assumed. It may be that the lattice is deforming in this way because shear deformations cost less elastic energy than compressive ones. There are also CDW-modulated charges associated with the small longitudinal displacements in the two halves of a bilayer, but they are equal and opposite. This would be favoured by Coulomb effects within a bilayer. We note the similarity of some of the displacements to a soft phonon observed in optimally doped YBCO (ref. [bib_ref] Highly anisotropic anomaly in the dispersion of the copperoxygen bond-bending phonon in..., Raichle [/bib_ref]. However, in that mode, the c-motion is in antiphase for the two halves of the bilayer. Buckling of the CuO 2 planes is also seen in 214 compounds [bib_ref] Stripes and superconductivity in cuprates, Tranquada [/bib_ref] , where it mainly consists of tilts of rigid Cu-O octahedra. Here, however, the displacements in the CuO 2 layers are clearly inconsistent with tilts of a rigid arrangement of ions. We draw attention to the up/down butterfly nature of the displacements of the four oxygens around a Cu in the bilayers, which is seen for both q a and q b modes. The two oxygens in the d-direction around a copper are displaced in the same direction as the Cu along c, but the other perpendicular pair is displaced oppositely [fig_ref] Figure 3 |: Representation of the CDW ionic displacement motifs for an unmodulated unit cell [/fig_ref]. To an STM (ref. [bib_ref] Direct phase-sensitive identification of a d-form factor density wave in underdoped cuprates, Fujita [/bib_ref] this could appear as a d-charge density on the oxygens, since c-axis motion of an oxygen-relative to the yttrium and/or to the crystal surface would alter its local doping and electronic state. We note that the STM measurements are analysed in such a way as to emphasize the electronic states, rather than the positions of atoms. In [fig_ref] Figure 5 |: in absolute units, 10 À 3 Å, calculated, as described in Supplementary... [/fig_ref] , we show qualitatively what the effect on the local doping of the oxygen ions might be by assuming that the change is proportional the displacement along c. The pattern produced has the same symmetry as that observed by STM (ref. [bib_ref] Direct phase-sensitive identification of a d-form factor density wave in underdoped cuprates, Fujita [/bib_ref] in Ca 2 À x Na x CuO 2 Cl 2 [fig_ref] Figure 5 |: in absolute units, 10 À 3 Å, calculated, as described in Supplementary... [/fig_ref]. STM and azimuthal angle-dependent resonant X-ray studies [bib_ref] Symmetry of charge order in cuprates, Comin [/bib_ref] of the charge order have been analysed in terms of modulated states with local symmetry of three types with respect to a planar copper site: equal density on the copper atoms (s-symmetry); equal density on the neighbouring oxygen atoms (s 0 -symmetry); opposite-sign density on the neighbouring O x and O y sites (d-symmetry). Our measured copper and oxygen displacements, recorded in [fig_ref] Table 1 |: The values of fitted ionic displacements [/fig_ref] , can provide an explanation for the relative proportions of these components. In agreement with the STM and resonant X-ray studies, we find that the d-symmetry component is dominant. These results carry several important messages. First, they show that a strictly planar account of high-T c phenomena may miss important aspects of the physics, and that the third dimension and crystal lattice effects cannot be ignored. In our experiments, we have observed a charge density wave with a strong shear (c-axis) component. The butterfly pattern of oxygen shear displacements around the planar copper ions can simulate a d-charge density on the oxygens. It will be very interesting to repeat these X-ray measurements on other underdoped high-T c compounds to establish the generality (or otherwise) of these results, and to relate these results to the changes in the CDW that occur at high fields where quantum oscillation measurements are performed. Ultrasonic measurements [bib_ref] Thermodynamic phase diagram of static charge order in underdoped YBa 2 Cu..., Leboeuf [/bib_ref] show that changes occur at B18 T. Very recent measurements in pulsed field [bib_ref] Three-Dimensional Charge Density Wave Order in YBa 2 Cu 3 O 6.67..., Gerber [/bib_ref] in an YBCO sample with ortho-VIII oxygen ordering show that longerrange order with the same value of d b emerges at high field. This is clearly related to our zero-field structure, and leads to interesting questions 32 about the Fermi surface reconstruction (a) A representation of the spatial variation of the z co-ordinate of the bilayer oxygen atoms from our X-ray results is shown for one (q a ) of two modulation directions present in the crystal. In (b) is an STM R(r) image, (where R(r) ¼ I (r,E)/ À I (r, À E) is a measure of the asymmetry of the current for positive and negative bias) acquired from the lightly doped cuprate Ca 2-x Na x CuO 2 Cl 2 (NaCCOC). (Reproduced with permission from ref. [bib_ref] Direct phase-sensitive identification of a d-form factor density wave in underdoped cuprates, Fujita [/bib_ref] The R(r)-image is used to highlight spatial variation of doping or electronic structure. Both a and b have the same symmetry as a bond ddensity wave along the d a direction. Note that NaCCOC has a repeat period of approximately four unit cells, which is longer than that in YBCO. at low and high fields. It is clear that antiferromagnetic order, the CDW, pseudogap and superconductivity are all intertwined, since they all remove electron states near the antinodal regions of the Fermi surface. It appears that there is a quantum critical point underlying the superconducting dome; we trust that our results will help to achieve an overarching theory relating the relationship of all these phenomena to high-T c superconductivity. [formula] u c for q b u b for q b u c for q b u b for q b f(Q ¼ 0) Â u c (q b ) f(Q ¼ 0) Â u b (q b ) u c for q a u a for q a f(Q ¼ 0) Â u c (q a ) f(Q ¼ 0) Â u a (q [/formula] # Methods X-ray techniques. To obtain sufficient data required the flexibility of a four-circle diffractometer, which is provided at the XMaS beamline (XMaS-The UK material science beamline at the ESRF (2015): http://www.xmas.ac.uk) at the ESRF, Grenoble [bib_ref] The XMaS beamline at ESRF: instrumental developments and high resolution diffraction studies, Brown [/bib_ref]. The sample was mounted in a closed-cycle cryostat and all measurements were carried out in zero magnetic field in reflection from the flat c face of the crystal (of area B2 Â 2 mm 2 ) at an X-ray energy of 14 keV. This gives a penetration depth of 25 mm into the sample, so the results are not dominated by surface effects. For CDW intensity measurements, the sample temperature was controlled at T c to maximize the signal, and it was taken to 150 K to check for spurious signals, which did not go to zero. The diffractometer angles were set so that the incoming and detected beams were close to the same angle to the c face of the crystal, which allowed correction for sample absorption, as described in Supplementary Note 5. CDW intensity measurements were carried out near the (h, 0, c), (0, k, c) and (h, h, c) planes of reciprocal space over as wide a range of h, k and c allowed by the maximum scattering angle, and the avoidance of grazing incidence at low c. CDW peaks were scanned parallel to d, through positions of the form Q ¼ s±q. Measurement and fitting of CDW satellite intensities. The intensities of the CDW peaks were established by fitting each scan with a Gaussian of fixed width, with a smoothly varying cubic polynomial background. By examination of 150 K measurements, or by the w 2 of the fit, spurious peaks were removed from the list of measured satellites. As shown in [fig_ref] Figure 2 |: Sample data compared with the fits to the two possible models [/fig_ref] , the finite range of the CDW order results in satellites that are broad, particularly in the c* direction. However, all the intensity of any satellite is confined to a single Brillouin zone, allowing it to be integrated over reciprocal space. The resulting list of intensities, weighted by their errors, was fitted to our CDW models by varying the ionic displacements {u j } to minimize w 2 . Further details are in Supplementary Note 2. [fig] Figure 1 |: Typical observations of CDW satellites at 60 K and their temperature dependence. (a-c) are obtained for the CDW with modulation vector q b . They are scans parallel to the basal plane component of the modulation, through the points (h, k, c) ¼ (0, 1 À d, 16.5), (0, 1 À d, 15.5) and (0, 3 þ d, 4.5). The counts are normalized so that they are approximately per second, measured over 10 s per point, plotted versus wave vector along the b* direction, labelled k. (a) shows a strong satellite, along with the fit line which gives the intensity as the area under the peak. The CDW is clearly centred at an incommensurate position (d b B0.328), although the value 1 3 lies within the peak. (b) shows a weaker peak and (c) is taken at a position where the CDW signal is unobservably small, and the fitted area of the peak is controlled by Poisson errors. (d) The intensities of CDW satellites for both q a (d a B0.323) and q b (d b B0.328) modes, normalized to their intensities at T c , are plotted versus temperature; these track each other within errors. (e) The integrated intensity of the satellite (a) is plotted versus c. The width in c, which reflects the finite c-axis coherence of the CDW, is much larger than the instrument resolution. Since it is a property of the CDW, it is the same for all satellites. All error bars in the above plots represent Poisson counting s.d.'s. [/fig] [fig] Figure 2 |: Sample data compared with the fits to the two possible models. (a,b) Maps of the satellite intensities associated with CDW modulations q a and q b measured in the (h, 0, c) and (0, k, c) planes of reciprocal space, respectively. The measured intensities are proportional to the areas of the red semicircles on the right of each Q point. Two different models are allowed by group theory: one involves the modes A 1 (for q a ) and B 1 (for q b ), and the other the A 3 and B 3 modes. The blue semicircles show the results of a good fit to the A 1 /B 1 modes (to all measured data, not just that shown in theFigure). (c,d) The same data (red semicircles) and the analogous fits (blue semicircles) to the A 3 /B 3 modes; these give a very poor fit to the data. Blank spaces indicate inaccessible regions or where a spurious signal prevented measurements of the CDW order. [/fig] [fig] Figure 3 |: Representation of the CDW ionic displacement motifs for an unmodulated unit cell. The displacements are B10 À 3 of interatomic distances, so have been exaggerated to make them visible. In (a) and (b) are shown the a-and c-axis components of the q a (A 1 ) mode of the CDW. In (c) and (d) are shown the b-and c-axis components of the q b (B 1 ) mode. The basal plane and c-axis displacements have a p/2 phase difference and hence are shown in separate unit cells. The next crystal unit cells in the c-direction would be in antiphase with those shown here. These motifs are modulated as a function of position with the relevant wave vector. The oxygen sites in the CuO chains represented here are half-occupied in our sample. [/fig] [fig] Figure 5 |: in absolute units, 10 À 3 Å, calculated, as described in Supplementary Note 4 and 5, from the fit of the data for the qb and qa modulated CDWs in ortho-II YBCO and subject to an overall possible systematic error B50%, not included above. In the first two columns all variables are free, and in the second pair some values have been fixed and are marked by an error 0 in parentheses. Also given are the values multiplied by the scattering amplitude for each ion at Q ¼ 0 (the number of electrons on the ion) to emphasize the relative contributions of each ion to the amplitude. The c-components of the displacements are even about the yttrium layer of the crystal unit cell, and the horizontal displacements are odd. Displacements that are zero by symmetry are represented by 0. Below the displacements are given the fitted anisotropic Debye-Waller factors a and b, which appear in the expression: expð À aðQ 2x þ Q 2 y Þ À bQ 2 z Þ. This multiplies the calculated intensities, and slightly improves the fit to the data, although the fitted displacements are little altered by including it. The units of a and b are 10 À 3 Å 2 . The bottom row of the Table gives w 2 per degree of freedom for the fits. Comparison of the bilayer oxygen height with an STM R(r)image. [/fig] [table] Table 1 |: The values of fitted ionic displacements. [/table]

Multiscale simulation of thrombus growth and vessel occlusion triggered by collagen/tissue factor using a data-driven model of combinatorial platelet signalling During clotting under flow, platelets bind and activate on collagen and release autocrinic factors such as ADP and thromboxane, while tissue factor (TF) on the damaged wall leads to localized thrombin generation. Towards patient-specific simulation of thrombosis, a multiscale approach was developed to account for: platelet signalling [neural network (NN) trained by pairwise agonist scanning (PAS), PAS-NN], platelet positions (lattice kinetic Monte Carlo, LKMC), wall-generated thrombin and platelet-released ADP/thromboxane convection-diffusion (partial differential equation, PDE) and flow over a growing clot (lattice Boltzmann). LKMC included shear-driven platelet aggregate restructuring. The PDEs for thrombin, ADP and thromboxane were solved by finite element method using cell activation-driven adaptive triangular meshing. At all times, intracellular calcium was known for each platelet by PAS-NN in response to its unique exposure to local collagen, ADP, thromboxane and thrombin. When compared with microfluidic experiments of human blood clotting on collagen/TF driven by constant pressure drop, the model accurately predicted clot morphology and growth with time. In experiments and simulations at TF at 0.1 and 10 molecule-TF/μm 2 and initial wall shear rate of 200 s −1 , the occlusive blockade of flow for a 60-μm channel occurred relatively abruptly at 600 and 400 s, respectively (with no occlusion at zero TF). Prior to occlusion, intrathrombus concentrations reached 50 nM thrombin,~1 μM thromboxane and~10 μM ADP, while the wall shear rate on the rough clot peaked at~1000-2000 s −1 . Additionally, clotting on TF/collagen was accurately simulated for modulators of platelet cyclooxygenase-1, P2Y 1 and IP-receptor. This multiscale approach facilitates patient-specific simulation of thrombosis under hemodynamic and pharmacological conditions. # Introduction Platelets play a critical role in primary haemostasis to prevent blood loss due to vessel injury. Secondary haemostasis requires tissue factor (TF) to activate the coagulation protease cascade leading to thrombin generation and subsequent fibrin polymerization to stabilize the clot. Unfortunately, during atherosclerotic plaque rupture that initiates a heart attack, flowing blood is exposed to a highly procoagulant surface containing TF. Platelets are captured from flow and activate on collagen. Intracellular calcium mobilization results in integrin activation and firm arrest, dense granule release of Adenosine diphosphate (ADP), activation of cycloxygenase-1 (COX1) with consequent thromboxane (TXA 2 ) synthesis, and phosphatidylserine exposure. Both ADP and TXA 2 are autocrinic species that enhance platelet activation and are highly targetable by drugs (e.g. P2Y 12 inhibitors and aspirin) to reduce risk. Thrombosis is the result of complex biological processes that occur in space and time and are coupled to prevailing hemodynamic that dictate cell collision rates, dilution of reactive species and forces on clot ensembles. Computational modelling of thrombosis has advanced significantly [bib_ref] A multiscale model of thrombus development, Xu [/bib_ref] [bib_ref] Grow with the flow: a spatial-temporal model of platelet deposition and blood..., Leiderman [/bib_ref] [bib_ref] Modelling platelet-blood flow interaction using the subcellular element Langevin method, Sweet [/bib_ref] [bib_ref] Multiscale prediction of patientspecific platelet function under flow, Flamm [/bib_ref] [bib_ref] Multiscale models of thrombogenesis, Xu [/bib_ref] [bib_ref] Three-dimensional multi-scale model of deformable platelets adhesion to vessel wall in blood..., Wu [/bib_ref] [bib_ref] Continuous modeling of arterial platelet thrombus formation using a spatial adsorption equation...., Babushkina [/bib_ref] [bib_ref] Modelling of platelet-fibrin clot formation in flow with a DPD-PDE method, Tosenberger [/bib_ref] as reviewed by [bib_ref] Systems biology of coagulation, Diamond [/bib_ref] and [bib_ref] Fluid mechanics of blood clot formation, Fogelson [/bib_ref]. The opportunity now exists to use patient-specific hemodynamic [bib_ref] Computational fluid dynamics applied to cardiac computed tomography for noninvasive quantification of..., Taylor [/bib_ref] , platelet signalling models [bib_ref] Multiscale prediction of patientspecific platelet function under flow, Flamm [/bib_ref] [bib_ref] A human platelet calcium calculator trained by pairwise agonist scanning, Lee [/bib_ref] , and bottom-up models of coagulation [bib_ref] A model for the stoichiometric regulation of blood coagulation, Hockin [/bib_ref] for multiscale simulations of thrombosis [bib_ref] Multiscale prediction of patientspecific platelet function under flow, Flamm [/bib_ref]. Such multiscale simulations can be compared and validated against direct measurements of clotting reactions over defined reactive surfaces and shear rates in microfluidic flows [bib_ref] or P2Y1 inhibitors reduce platelet deposition in a microfluidic model of thrombosis..., Maloney [/bib_ref] [bib_ref] Thrombus growth and embolism on tissue factor-bearing collagen surfaces under flow: role..., Colace [/bib_ref] [bib_ref] FXIa and platelet polyphosphate as therapeutic targets during human blood clotting on..., Zhu [/bib_ref]. In prior work [bib_ref] Multiscale prediction of patientspecific platelet function under flow, Flamm [/bib_ref] , a multiscale platelet deposition model utilized a four-agonist (ADP, TXA 2 , collagen, prostacyclin) neural network (NN) to predict platelet dynamics under venous and arterial shear condition in the absence of thrombin. This NN was trained using calcium traces obtained for all single and pairwise combinations of agonists at low, medium and high concentration. In this pairwise agonist scanning (PAS) experiment, ADP and mimetics for collagen, thromboxane, and prostacyclin were used to quantify P2Y 1 /P2Y 12 , GPVI, TP and IP signalling, respectively, for NN training. Along with NN for platelet signalling, the multiscale platelet deposition model utilized a velocity field solved by lattice Boltzmann (LB) and prevailing concentration profiles obtained by finite element (FEM) solution of convection-diffusion PDEs. Platelet motion and binding/unbinding were solved stochastically by the LKMC method. In this study , the multiscale model has been expanded and refined in several important aspects: (i) the NN prediction of platelet calcium was expanded to include thrombin in the PAS training set averaged over 10 healthy donors (50% male) [bib_ref] A human platelet calcium calculator trained by pairwise agonist scanning, Lee [/bib_ref] ; (ii) the FEM mesh for solution of ADP, TXA 2 and thrombin spatiotemporal concentrations was adaptively refined based on platelet activation (Figs 3 and 4); (iii) individual platelets were allowed to find the most stable nearby position of greatest bonding after initial capture to the clot (Figs 5 and 6); (iv) thrombin was released from the wall into the clot assuming a parameterized curve with initiation and decaying stages [fig_ref] Figure 7: Comparison of different thrombin flux signals for constant flow rate simulations at... [/fig_ref] ; and (v) flow fields were calculated for either constant flow rate (non-physiologic but experimentally accessible) or constant pressure-drop (full channel occlusion possible) . The adaptive meshing algorithm allowed for efficient calculation of soluble species transport with ongoing platelet capture and activation. The remodelling algorithm was embedded in the LKMC model to achieve more realistic clot morphologies. The simulated platelet deposition results were compared with two different modes of microfluidic experiments (the constant flow and pressure relief mode) . In the constant flow mode, blood was perfused at constant flow rate through a microfluidic device where growing platelet deposits experience increasingly high shear as the clot grows across the channel. In the pressure relief mode , flow was diverted from the occluding channel to an open non-clotting channel so that clots were able to grow completely across the channel until flow stopped. The multiscale model accurately simulated the platelet morphology and clot growth rates observed in the two different modes and provided quantitative predictions on platelet deposition dynamics consistent with experiments with healthy human blood treated with the Factor XIIa inhibitor, corn trypsin inhibitor (CTI). To our knowledge, this is the first model to quantitatively predict dynamic platelet deposition in the presence of wall-generated thrombin and inhibitors of ADP or thromboxane, as validated by comparison with actual data obtained with flowing human blood. Such predictions are novel and relevant given the clinical use of direct thrombin inhibitors and antiplatelet agents targeting ADP and thromboxane 524 . Multiscale model of platelet activation and thrombus formation under flow. An 8-channel microfluidic device allows whole blood perfusion through each high aspect ratio channel (250 μm wide × 60 μm high) and over a discrete 250 μm-long collagen patch (a). A 2D multiscale model domain (500 μm long × 60 μm high) corresponds to the centre of the microfluidic channel which includes collagen as a uniform boundary (b). The biology and physics of thrombosis includes flowing platelets being captured to collagen or other bound platelets, platelet activation triggered by collagen (rectangular bar at bottom) and thrombin (Factor IIa, FIIa) modelled as a boundary condition [time-dependent flux J FIIa (t)] and ADP/TXA 2 , which are autocrinic molecules released by platelets reaching a critical level of activation (c). The multiscale simulation of platelet deposition under flow requires simultaneous solution of the instantaneous velocity field v over a complex and evolving platelet boundary by LB, concentration profiles of ADP, TXA 2 thrombin by FEM, individual intracellular platelet state ([Ca 2+ ] i ) dictated by NN model and activation-dependent release reactions R for ADP and TXA 2 , and all platelet positions and adhesion/detachment kinetics by LKMC (d, e). A 2-layer, 12-node NN architecture was employed for prediction of intracellular calcium in response to agonist exposure. Agonist concentrations at a given time point were fed into the processing layers. The layers then integrated the input signal with feedback at t = 1, 2, 4, 8, 16, 32, 64 and 128 s to calculate [Ca 2+ ] i at the next time point (f) (for related colour see the online version). 525 Y. LU ET AL. . Connection between experimental measurement and numerical prediction. PAS experiments were performed on 10 healthy donors (50% male) to train a NN ensemble with 6160 calcium traces to predict calcium response for an average healthy person and was incorporated into the multiscale model to predict platelet activation. An effective boundary flux term was imposed. Thrombus growth dynamics can be predicted by the multiscale model and were compared against microfluidic experiments on healthy human blood conducted under identical conditions. autocrinic signalling. While other models have predicted dynamic platelet deposition in the presence of thrombin generation [bib_ref] A multiscale model of venous thrombus formation with surface-mediated control of blood..., Xu [/bib_ref] [bib_ref] Grow with the flow: a spatial-temporal model of platelet deposition and blood..., Leiderman [/bib_ref] , to our knowledge, this is the first model to include inhibitors of ADP and thromboxane, agonists of the IP receptor, and be directly compared with experimental data. To our knowledge, this is also the first study to quantitatively predict platelet-mediated channel occlusion and to compare predicted occlusion times to actual measurements conducted under constant pressure drop conditions. Additionally, the model predicts a peak thrombin flux on the order of 10 −12 nmole/μm 2 -s testable by experimental measurement for a TF rich surface exposed to flowing blood. # Methods A 2D rectangular simulation domain (500 μm long × 60 μm high) was used for all simulations . At a location between 100 and 350 μm downstream of the simulation domain entrance, a 250-μm collagen patch was defined as a boundary condition . The 2D computational domain represented a centreline cross-section of an actual 3D microfluidic channel (250 μm wide × 60 μm high) used to perform the ex vivo whole blood perfusion experiments . 526 . Simulation domains and meshes for the different modules. In LKMC, each platelet (in circle) is resolved with multiple square elements (lattice spacing of 0.5 μm). In LB, the uniform grid had the same lattice spacing as LKMC, but was offset by 0.25 μm in each direction such that the LB nodes surround the LKMC pixels. In FEM, a refined triangular mesh was generated around the clotting region such that each platelet occupies ∼30 triangular elements. LKMC feeds platelet information such as position and activation states into LB and FEM. FEM computes concentration field for soluble species interpolated back onto the LKMC grids. LB computes the velocity field used by both FEM and LKMC (for related colour see the online version). The overall computational framework utilized four coupled modules. At the centre of the framework, a LKMC simulator was used to evolve platelet positions in time on a regular square grid. Platelets were assumed to be rigid circular particles with diameter of 3 μm. The LKMC simulation was informed by three other computational modules to fully specify rates for the various platelet events (e.g. translation, binding, unbinding). First, the velocity field that convects the platelets across the microchannel from entrance to exit was computed with the LB method. The LB simulation assumes a quasi-steady flow profile that depends on the instantaneous thrombus boundary location and is resolved at a resolution approaching the single platelet scale. Next, the concentration profiles of the soluble platelet agonists (ADP, TXA 2 and thrombin) were described by a system of convection-diffusion-reaction equations that were solved using finite element method (FEM). Finally, the activation state of each platelet, as defined using a metric based on intracellular calcium [Ca 2+ (t)] i was calculated in time using the NN model for the unique multicomponent agonist exposure experienced by each platelet. The connectivity between these various computational modules is shown in . Each of the four computational modules are described in detail in the following sections. ## Nn module PAS was used to obtain calcium traces for apixaban-treated, diluted and calcium dye-loaded platelet rich plasma (PRP) stimulated with all single and pairwise combinations of six different agonists used at low, 527 . Platelet activation-driven adaptive triangular meshing. The 2D Gaussian influence function and the resultant meshes for an example case of two adjacent and activated platelets (a). The triangular mesh is shown for a typical thrombus configuration and the resultant concentration fields for ADP and TXA 2 as well as each platelet's activation state of integrated intracellular calcium (b). medium and high concentrations (0.1, 1 and 10 × EC 50 ) . In this approach, the six agonists used were: ADP, U46619 (thromboxane mimetic), convulxin (collagen mimetic), thrombin, GSNO (NO donor) and iloprost (prostacyclin analog). NN training was then carried out using the NN Toolbox in the MATLAB software package to predict platelet calcium level for any agonist dose and combination. Output concentrations were mapped between 0 (resting calcium levels) and 1 (maximal response). The structure of the NARX (Nonlinear AutoRegressive network with eXogenous inputs) model utilized two-layer processing (eight nodes in the first layer and four nodes in the second layer) and a tapped delay line with 128 s of feedback to each layer . The hyperbolic tangent transfer function was used in all nodes and a linear transfer function was used in the output layer. Initial states of the feedback were set to 0 corresponding to the resting platelet. The NN predicted the calcium level for the next time step based on the feedback vector and 128 s prior to the current instant) and the current concentration input from the six agonists. A 100 NN-ensemble for PAS of 10 healthy donors and 10 trained NNs per donor predicted averaged calcium concentration output for human platelets exposed to any combinations of the six agonists. This approach captured the measured information from 6160 calcium traces containing 511,280 calcium data points. See [bib_ref] A human platelet calcium calculator trained by pairwise agonist scanning, Lee [/bib_ref] for detailed information validating PAS-NN prediction of [Ca 2+ (t)] i beyond the pairwise training set for trinary stimulations, 4-6 agonist stimulations and sequential agonist stimulations. ## Lb for blood flow The LB method was used to solve the flow field around the platelets satisfying the incompressible Navier-Stokes equation with continuity condition. A D2Q9 scheme with Zou-He Boundary conditions was applied [bib_ref] On pressure and velocity boundary conditions for the lattice Boltzmann BGK model, Zou [/bib_ref]. The top and bottom walls were treated as no-slip surfaces. Platelets were simulated as pixelated objects on the LKMC lattice . Free flowing platelets and their interactions with each other were ignored in the LB model of Newtonian blood flow. However, the hydrodynamic effect of a clot-bound platelet on the fluid was explicitly handled in the LB model as a no-slip surface. Two different experimental settings, constant flow rate mode and pressure relief mode, were simulated. In the constant flow rate mode, blood was perfused over the microfluidic channel at a constant flow rate regardless of the size of the developing platelet deposit on the wall. In the pressure relief mode , two identical channels in parallel with a single outlet were perfused. Only one channel experienced obstruction caused by deposited platelets, while the other channel had no wall deposit (EDTA-treated blood in the real experiment). With increasing flow resistance caused by platelet accumulation, flow was diverted to the free channel which eventually led to zero flow in the fully obstructed channel. In the constant flow rate mode, a parabolic flow profile with initial wall shear of 200 s −1 was maintained at the inlet. In the pressure relief mode, a similar parabolic flow allowing the same wall shear for both of the two channels was maintained at the outlet. ## Fem for soluble agonist transport and reaction The concentration fields of the agonists, C j (x, y, t) for j = ADP, TXA 2 and thrombin were obtained by solving a system of convection-diffusion-reaction equations using the FEM: [formula] ∂C j ∂t + v · ∇C j = D j ∇ 2 C j + R j ,(1) [/formula] where D j was the Brownian diffusion coefficient of ADP, TXA 2 and thrombin, and v was the fluid velocity provided by the LB module. The release of ADP and TXA 2 by the ith platelet was triggered by its internal activation state, [formula] ξ i (t) = t 0 ([Ca i ] − [Ca] baseline ) [/formula] dτ exceeding a critical level of ξ crit = 9 μms. Once sufficiently activated, the platelet released ADP and TXA 2 with rates governed by an exponential decay function, [formula] R j (t) = M j τ j exp − t − t release τ j ,( 2 ) [/formula] where M j was the total moles of species j for a platelet and τ j was a time constant. Published experiments have determined that each platelet can release 1 × 10 −8 nmol of ADP with a time constant of 5 s [bib_ref] Determination of ATP and ADP in blood platelets: a modification of the..., Holmsen [/bib_ref] [bib_ref] Platelet active concentration profiles near growing thrombi. A mathematical consideration, Hubbell [/bib_ref] and 4 × 10 −10 nmol of TXA 2 with a time constant of 100 s [bib_ref] Thromboxane-B2 generation during ex-vivo platelet aggregation, De Caterina [/bib_ref]. The element at the centre of the activated platelet was treated as the source element for the PDE calculation (FEM). Agonists were assumed to undergo transport by diffusion within the platelet mass. ## Thrombin release modelled as boundary flux Motivated by the work of [bib_ref] Surface-mediated control of blood coagulation: the role of binding site densities and..., Kuharsky [/bib_ref] and numerous observations of a thin thrombin/fibrin core that is immediately adjacent to wall-derived TF [bib_ref] Platelet-targeting sensor reveals thrombin gradients within blood clots forming in microfluidic assays..., Welsh [/bib_ref] [bib_ref] A systems approach to hemostasis: 1. The interdependence of thrombus architecture and..., Welsh [/bib_ref] [bib_ref] Side view thrombosis microfluidic device with controllable wall shear rate and transthrombus..., Muthard [/bib_ref] [bib_ref] Hierarchical organization in the hemostatic response and its relationship to the platelet-signaling..., Stalker [/bib_ref] [bib_ref] A systems approach to hemostasis: 3. Thrombus consolidation regulates intrathrombus solute transport..., Stalker [/bib_ref] [bib_ref] A systems approach to hemostasis: 2. Computational analysis of molecular transport in..., Tomaiuolo [/bib_ref] , thrombin was introduced into the computational domain as a time-dependent wall boundary flux condition (Figs 1 and 2). This boundary flux assumption is consistent with a fixed amount of Factor X (FX) being activated to FXa near the wall due to extreme transport limits on FX diffusion through a dense and activated platelet aggregate since FX binds to activated platelets [bib_ref] Lipid-bound factor Xa regulates tissue factor activity, Hathcock [/bib_ref]. Instead of resolving the full detail of the coagulation cascade, which involves~50 PDEs and~100 reactions where many of reaction kinetics has yet to be verified experimentally within dense platelet deposits formed under flow, thrombin was simulated as an imposed boundary flux at the wall with an initiation phase followed by a decay phase whose shape and peak value was motivated from both calibrated automated thrombography (CAT) and a previously validated ODE model . Since measurements are not available for intrathrombus thrombin generation per molecule of TF on a surface, a 400-s thrombin flux profile (J, nmole/μm 2 -s) was studied parametrically in three cases. Case I: strong thrombin flux for high [TF] wall = 10 molecule-TF/μm 2 ; Case II: weak thrombin flux for low [TF] wall = 0.1 molecule-TF/μm 2 ; and Case III: zero thrombin flux for TF-free wall condition. A twopiece cubic spline (8 degrees-of-freedom) for the thrombin flux was defined for the initiation and decay phases with the following imposed constraints: zero flux at the initial and final time of 400 s, continuity, slope and curvature matching at the joint, and zero slope at the final time. These constraints reduce the degrees-of-freedom to only two parameters: time to peak flux, T max , and peak flux, J max . The three cases of thrombin flux are shown in [fig_ref] Figure 7: Comparison of different thrombin flux signals for constant flow rate simulations at... [/fig_ref]. T max was set to 160 s, a time typical for onset of fibrin generation at high 531 Y. LU ET AL. concentration of [TF] wall [bib_ref] Platelet-targeting sensor reveals thrombin gradients within blood clots forming in microfluidic assays..., Welsh [/bib_ref]. The peak flux value imposed underneath the growing platelet deposit (where convection is minimal) was set to J max = 8 × 10 −12 nmole/μm 2 -s, which is 10 to 100-fold less than the imposed flux needed to polymerize fibrin from flowing fibrinogen over a membrane [bib_ref] Thrombin flux and wall shear rate regulate fibrin fiber deposition state during..., Neeves [/bib_ref]. Case II deployed a five-fold reduction in J max to simulate the situation of a typical low TF (~0.1/μm 2 ) stimulation. By testing against measured clot growth rates, Cases I and II represent predictions of the multiscale model that potentially can be compared to future measurements of thrombin generation by a growing thrombus. While the thrombin flux condition was inspired by the CAT assay (closed system) corresponding to a partially restricted inner clot 'compartment', clot formation over collagen/TF under flow is an open system. It remains to be determined if influx of prothrombin can lead to a plateau in thrombin production. Future measurements of dynamic thrombin flux from clots formed under flow will help refine this wall flux term. ## Cell activation-driven adaptive meshing An important feature of platelet deposition is that the clotting region, especially at early times, occupies only a small portion of the computational domain. Consequently, using a uniform mesh throughout the entire domain incurs a large computational cost where many mesh nodes must be applied to achieve a reasonable resolution around each platelet. An adaptive meshing scheme was designed to allow the mesh density to be refined in the clotting region of an evolving rough clot surface. The mesh generator DistMesh [bib_ref] A simple mesh generator in MATLAB, Persson [/bib_ref] with a Gaussian size function was used to produce an unstructured triangular mesh . For the ith platelet that reached the critical activation state, we computed a 2D Gaussian, g i , around the platelet centre: [formula] g i = exp − (x − x i ) 2 + (y − y i ) 2 2σ 2 ,(3) [/formula] where x i , y i are coordinates of the centre of the ith platelet, σ was a parameter that describes the spread off the Gaussian in both directions. The overall influence function G (x, y) was defined as the sum of all Gaussians of single platelets, i.e. [formula] G (x, y) = all activated platelets exp − (x − x i ) 2 + (y − y i ) 2 2σ 2 .( 4 ) [/formula] The size function h (x, y) defined the equilibrium length of the triangular element by the Gaussian influence level as follows: [formula] h (x, y) = h min if G (x, y) < G cutoff = min (h min + α (G (x, y) − G cutoff ) , h max ) otherwise,(5) [/formula] where h (x, y) was the minimum equilibrium length of triangles, h max was the maximum equilibrium length of triangles, α was the slope of change in the transition region and G cutoff was the cut-off level for the influence. ## Lkmc module The LKMC module was used to simulate the dynamics of platelet transport and deposition under blood flow. The LKMC simulation was executed on a square grid with lattice spacing, h LKMC = 0.5 μm, while platelets were assumed to be circular with radius (1.5 μm). Platelet transport events were defined as (i) [formula] diffusion with rate D = D platelet h 2 LKMC , [/formula] where D platelet was the effective diffusivity accounting for both Brownian motion and the red blood cell (RBC) dispersion effect [bib_ref] Platelet adhesion to a spinning surface, Turitto [/bib_ref] and (ii) advection with [formula] rate C = v i h LKMC , [/formula] where v i was the fluid velocity component along the lattice direction e i . The total rate of motion for each platelet was the sum of the convective and diffusive rates. This approach was validated in our earlier work [bib_ref] Lattice kinetic Monte Carlo simulations of convective-diffusive systems, Flamm [/bib_ref]. Flowing RBCs also generate an excess of platelet concentration in the plasma layer near the boundary, which was modelled as an additional radial drift velocity superimposed on the actual velocity that resulted in a three-fold wall excess of platelets [bib_ref] Model of platelet transport in flowing blood with drift and diffusion terms, Eckstein [/bib_ref]. In addition to the convection-dispersion-drift events for each platelet, we also considered the activation-dependent rates of attachment and detachment between two platelets or between a platelet and the collagen boundary. Both cumulative and recent-history activation levels were considered for each platelet. The cumulative internal activation state, ξ , of the ith platelet at time t was defined as the accumulated integral calcium concentration above the basal level: [formula] ξ i (t) = t 0 ([Ca i ] − [Ca] baseline ) dτ ,(6) [/formula] where[Ca] baseline = 100 nM [bib_ref] Molecular mechanisms of platelet activation, Siess [/bib_ref]. The recent-history activation level was defined as the accumulated calcium level between the current time t and previous time t − t ( t = 30 s): [formula] ξ t,i = t t− t ([Ca i ] − [Ca] baseline ) dτ .(7) [/formula] The recent-history activation metric allows platelet integrins to return to a resting state (non-adhesive) if calcium has returned to baseline for 30 s. One of the most widely used functional forms, the Hill functions were introduced to normalize, between α min = 0.001 and 1, the cumulative and 30-s recent history activation states: [formula] F (θ i ) = α min + (1 − α min ) θ n i θ n i + θ n 50 , θ i = ξ i , ξ t,i ,(8) [/formula] where α min = 0.001 represents the base level of activation, n represents the sharpness of the Hill function, and θ 50 was the critical level for 50% activation. An overall rate of attachment of the ith platelet to the collagen surface was defined by: [formula] collagen att = k collagen att F (ξ ) F ξ t,i .(9) [/formula] The attachment rate between the ith and jth platelet j, via fibrinogen depending on both the cumulative and transient calcium level of two particles was modelled as: The detachment rate between platelet and the reactive surface was modelled using the Bell exponential to describe the shear-dependent breakage of receptor-ligand bonds, [formula] fibrinogen att = k fibrinogen att F (ξ i ) F ξ j F ξ t,i F ξ t,j .(10)collagen det = k collagen det F (ξ i ) −1 F ξ t,i −1 exp γ i γ c ,(11) [/formula] where γ i was the local shear rate around the ith platelet and γ c was the characteristic shear rate required to initiate bond breakage. The detachment rate between the ith and jth platelet was given similarly by [formula] fibrinogen det = k fibrinogen det F (ξ i ) F ξ j −1 F ξ t,i F ξ t,j −1 exp γ i + γ j 2γ c .(12) [/formula] In the current model, platelets could only dissociate as singlets from the thrombus and fracture of large chunks of platelets was not considered due to the complexity of stress propagation within the random aggregated clot. Platelets displayed no bulk aggregation as they entered the concentration boundary layer due to short exposure times and rarity of collisions in the bulk flow. Therefore, aggregation for free-flowing platelets was not considered. At each time step, a specific event k with rate k was chosen from the rate catalogue with probability k / total . A specific time step τ = − ln (u) / total was also chosen by drawing a random number u from the unit interval (0, 1]. ## Clot remodelling scheme Platelets can transiently roll (translocate) on Von Willebrand factor (VWF) prior to firm arrest mediated by integrin activation. However, resolving molecular-level events of detailed adhesion and restructuring mechanics incurs too much computational cost given the spatial and temporal domain of thrombosis (1000s of cells over many minutes and millimetres). A remodelling algorithm was thus designed to mimic the remodelling behaviour without providing full descriptions of the inherent biology. Upon binding, each platelet was able to roll a certain distance along the trajectory of the thrombus contour in the flow direction. A typical rolling distance of 10 μm is observed by tracking single platelet movement under microscopes. Under the square lattice settings, each platelet occupied a collection of pixelated squares in the domain. As shown in [fig_ref] Figure 5: Remodelling algorithm for platelet deposition [/fig_ref] , whenever a quiescent platelet (in black) attempted to bind to the target platelet (in gray), a search was performed along all the potential binding sites (trajectory indicated by curved arrow) in the direction of the local flow. Upon binding, the algorithm allowed the platelet to search along the contour of the thrombus for a distance of 10 μm (observed under microscopes) based on local flow and morphology. A list of all the potential binding sites along the search trajectory and their corresponding potential binding rates were evaluated. Sites with higher binding rates are more likely to be chosen. The algorithm picks the final deposition site based the potential binding rates which are the sums of all the platelets acting on that site. For instance, a site with three neighbouring platelets (assuming all with the same activation levels, therefore same binding rates) will be three times more likely to be chosen than one with only one neighbour. Denser packing is an emergent outcome since sites that can form close packing are more likely to be chosen as these sites tend to have more platelets acting on them, thus higher potential binding rates. This remodelling algorithm circumvented the biological complexity of detailed receptor adhesive dynamics calculations, but allowed the platelet to search for a favoured site for binding, mimicking the rolling and translocation behaviour observed both in vitro and in vivo. ## Multiscale simulation of thrombosis ## Module integration Exchange of information between the four computational modules was performed at specified time intervals : t LKMC = 5 × 10 −3 s, t LB = 5 × 10 −3 s, t FEM = 10 −2 s and t NN = 10 −2 s. At the start of the simulation, the LKMC, LB, FEM and NN modules were all initialized. Each module was then stepped forward in time until the first coupling time was reached (t = 5 × 10 −3 s), and then information was exchanged between relevant modules. When the total time evolved by the LKMC module first exceeded t LKMC , each of the three modules, LB, NN and FEM was updated to the current system time. During the first update time (∼t = 5 × 10 −3 s), the LB solver received the current configuration of the bound platelets from the LKMC simulation and was used to update the velocity field to the current system time. Then, all methods resumed stepping forward in time until the next coupling time (t = 10 −2 s) occurred. At the time of update, all models shared information: the LKMC was used to update position of platelets in FEM, LB; LB updated the velocity field in LKMC and FEM; FEM updated the concentration field; and NN predicted the activation state of each platelet based on updated agonist concentrations to be used in LKMC and FEM. This process was repeated until the end of the specified simulation time. The overall multi-element model used 25 parameters , the first 15 of which are either well-characterized physical or biological properties (ρ blood , μ blood , etc.) and the remaining 10 parameters (n, ξ 50 , , , , , and γ c ) were determined from our previous model [bib_ref] Multiscale prediction of patientspecific platelet function under flow, Flamm [/bib_ref] and experimental measurements. # Experimental methods ## Blood collection and preparation Blood was collected via venipuncture into a syringe containing 40 μg/mL of a FXIIa inhibitor CTI (Haematologic Technologies, Essex Junction, VT) from healthy donors who self-reported as free of alcohol use and had refrained from oral medication for at least 72 h prior to donation. All donors provided informed consent under approval of the University of Pennsylvania Institutional Review Board. Platelets were labelled with anti-human CD61 antibody. Fluorescent fibrinogen was added (1 mg/mL stock solution, 1:80 v/v % in whole blood) for measuring fibrin generation. All experiments were initiated within 5 min after phlebotomy. ## Preparation and characterization of collagen/tf surface Glass slides were treated with Sigmacote (Sigma, St. Louis, MO) to impede protein adsorption and clotting on glass surface. A volume of 5 μL of collagen was perfused through the patterning channel (250 μm wide and 60 μm high) of a microfluidic device to create a single stripe of fibrillar collagen, as previously described by [bib_ref] Microfluidic focal thrombosis model for measuring murine platelet deposition and stability: PAR4..., Neeves [/bib_ref] Lipidated TF was then sorbed to the collagen surface by introduction of 5 μL of Dade Innovin PT reagent (20 nM stock concentration) (Duckers et al., 2010) diluted 300-, and 5-fold with N-2-hydroxyethylpiperazine-N9-2-ethanesulfonic acid buffered saline to obtain low and high TF surface concentrations of 0.1 and 10 molecules per μm 2 , respectively, as estimated by imaging of sorbed annexin V-fluorescein isothiocyanate-stained vesicles [bib_ref] FXIa and platelet polyphosphate as therapeutic targets during human blood clotting on..., Zhu [/bib_ref]. In all experiments, the Dade Innovin PT reagent was incubated with the collagen for 30 min without flow and the surface was then rinsed and blocked with 0.1% bovine serum albumin buffer. ## 535 ## Microfluidic clotting assay on collagen surfaces with or without tf An 8-channel polydimethylsiloxane flow device , described by [bib_ref] or P2Y1 inhibitors reduce platelet deposition in a microfluidic model of thrombosis..., Maloney [/bib_ref] , was vacuum-mounted perpendicularly to collagen/TF surfaces forming eight parallel-spaced prothrombotic patches 250 μm wide by 250 μm long. Blood was then perfused across the eight channels to be withdrawn through a single outlet. Initial wall shear rate was controlled by a syringe pump (Harvard PHD 2000; Harvard Apparatus, Holliston, MA) connected to the outlet on the flow device to maintain a venous shear rate of 200 s −1 (corresponding to a volumetric flow rate of 2 μL/min/channel). Thrombi were formed either under constant flow rate (constant Q; CTI-treated blood in all eight channels) or under constant pressure drop (constant DP) conditions as described by [bib_ref] Thrombus growth and embolism on tissue factor-bearing collagen surfaces under flow: role..., Colace [/bib_ref]. To achieve constant pressure drop (pressure relief mode), EDTA-treated blood was delivered into alternating channels to abolish thrombus formation, thus allowing CTI-treated blood to clot in assay channels and divert flow into the matched EDTA channels. Platelet and fibrin activity were monitored simultaneously by epifluorescence microscopy (IX81; Olympus America Inc., Center Valley, PA). Images were captured with a charged coupled device camera (Hamamatsu, Bridgewater, NJ) and were analysed with ImageJ software (National Institutes of Health). To avoid side-wall effects, fluorescence values were taken only from the central 75% of the channel. ## Confocal imaging of clot morphology and flow rate Real-time clot height was determined by platelet fluorescent intensity, which was linearly proportional to the fluorescent intensity of a fully occluded thrombus. Experiments were terminated when flow was in complete cessation (full channel occlusion) in assay channels for the pressure relief condition and final clot height was assumed to be the same as the height of the channel (60 μm). Profiles of clot morphology were generated using a custom MATLAB script, in which the average of 40 vertical line-scans (in the direction of flow) was calculated. The height was also assumed to be proportional to the fluorescence level and was calibrated by the maximum intensity for a fully occluded thrombus. Supplemented fluorescent fibrinogen provides a fluorescent background in assay channels while EDTA channels remain dark all the time. Flow rate in assay channel was determined by measuring the width of the fluorescent portion in the middle flow plane at the merged outlet of every two paired channels at each time point and calculating the corresponded volumetric flow rate within the fluorescent portion of the channel using a custom MATLAB script with assumptions that a parabolic flow profile was fully developed. # Results ## Cell activation-driven adaptive meshing The FEM mesh for calculation of agonist transport was generated by the remeshing algorithm described in Section 5. The adaptive remeshing algorithm allowed for automatic generation of mesh configurations for any specific platelet contour shape that resulted stochastically from LKMC. Starting from an initially coarse uniform mesh over the entire 500μ × 60 μm simulation domain, the remeshing algorithm generated a highly resolved mesh in proximity to the growing thrombus contour, as required to resolve the boundary layers associated with the agonist concentrations. The equilibrium side length of the smallest triangles in the clotting region was set to 0.7 μm. For the earlier stages, the clotting region only occupied a small portion of the domain, and a mesh with no more than 8000 nodes was usually sufficient. For a fully developed thrombus, the program generated an unstructured triangular mesh with about 30,000 nodes and 50,000 triangular elements . For the same domain of 500 μm by 60 μm, using a uniform square mesh with 1 μm spacing in both directions would require as many as 30,000 (500 × 60) nodes and 60,000 elements. Each activated 3-μm diameter platelet occupied about 30 triangular elements (side length = 0.7 μm) compared with nine elements for uniform meshing. Therefore, a three-fold increase in mesh resolution was achieved at the site of each platelet's release of ADP and TXA 2 while applying coarsened meshes elsewhere therefore maintaining a similar problem size. Computational power was especially focused during the early stages of clotting when only a few platelets were immobilized to the surface, with these initial events driving subsequent pattern formation. The refined regions occupied the concentration boundary layer extending up to ∼10 μm from the platelets on the outer most layer, thus coinciding with the regions where the sharp gradient change in agonist concentrations emerged from the simulation. A new mesh was generated only if the current mesh resolution for one or several activated platelets fails to reach the tolerance criteria. Over the course of one simulation, a total of 25 different meshes were used and mesh update was more frequent during the earlier development stage. The triangular elements were of high quality with appropriate aspect ratio [fig_ref] Figure 5: Remodelling algorithm for platelet deposition [/fig_ref] for stable and accurate finite element calculations. Overall, the adaptive remeshing scheme led to about 40% reduction in computational effort relative to the simulations with uniform meshes, with most of the savings occurring during the earlier stages of clotting. While adaptive meshing is a well-established technique to increase computational efficiency, particleactivation driven remeshing will be critical to multiscale simulation of human thrombosis where the flow field spans the millimetre to centimetre length-scale range, whereas the thrombotic processes near the wall involve micron-scale boundary layer phenomena. Although the adaptive meshing technique itself does not present new physics, it allows us to probe larger domains that would otherwise be computationally prohibitive. In fact, it would be a necessity if the model were to be applied for larger vessel simulations where creation of suitable mesh would be vital to numerical computations. The approach developed also is a versatile and general automatic mesh refinement based on evolving platelet deposition patterns, which are not known a priori and are affected by flow and, in turn, affect the flow. The adaptive meshing approach of Equation (3)-(5) represents a tractable and robust numerical implementation between the Monte Carlo, NNs and Finite Element calculations of the multiscale model. In future work, the mesh refinement within clots can be relaxed in regions where clot regions become inactive at longer times (full depletion of ADP, TXA 2 and thrombin). Even for clots that grow to mm-scale thicknesses to occlude a coronary artery, e.g. the current mesh refinement algorithm offers significant computational savings. ## Thrombin flux case studies To simulate thrombin generation, we imposed a time-dependent wall influx with the initiation and decaying stage parameterized by two parameters, T max and J max . A comparison of platelet deposition dynamics is shown in [fig_ref] Figure 7: Comparison of different thrombin flux signals for constant flow rate simulations at... [/fig_ref]. For the first case representing normal physiology, after an initial lag of 60 s, platelet mass started to grow linearly to 400 s which was consistent with fluorescence measurements in experiments. A typical averaged thrombin concentration within the thrombus was found to be ∼30 nM which agrees with both experimental observations and earlier modelling work [bib_ref] Surface-mediated control of blood coagulation: the role of binding site densities and..., Kuharsky [/bib_ref]. Platelets were able to grow a height of ∼50 μm in a typical constant blood flow simulation. Case II with five-fold reduction in maximum thrombin signal was also studied. Here, a delayed platelet accumulation was observed during the earlier lag phase. The thrombus height at 450 s of the simulation reached approximately ∼45 μm. Case III with no thrombin presence predicted a typical thrombus height of 20-30 μm which was consistent with our previous work [bib_ref] Multiscale prediction of patientspecific platelet function under flow, Flamm [/bib_ref]. the flow velocity becomes quite high in the narrowed lumen with peak wall shear rates exceeding ∼1000 s −1 . Platelet released ADP and TXA 2 and wall-derived thrombin form thin boundary layers as these species elute from the clot and are convected downstream. The simulation predicts the growth of the clot upstream of the collagen, a core of highly activated platelets in proximity with collagen, and an outer shell of relatively unactivated platelets. Inclusion of thrombin has interesting ramifications compared to other soluble agonists. Thrombin has a diffusivity that is much lower than ADP or thromboxane. Also platelet signalling in response to thrombin displays unique properties of fast-on but slow-off due to PAR-1 and PAR-4 signalling dynamics captured in the PAS training data. Differences in calcium signalling due to thrombin [fig_ref] Figure 6: Comparison of thrombus morphology with and without the remodelling algorithm [/fig_ref] highlight these effects. Without thrombin present, a clotting event does not cause 60-micron channel occlusion or reduction of flow [fig_ref] Figure 8: Fully resolved thrombosis for whole blood flow over collagen/TF after 400 s [/fig_ref]. The model predicted a two-fold increase in growth rate of thrombus height (∼0.13 μm/s) under the presence of thrombin as compared to the case without thrombin. ## Platelet remodelling upon deposition and thrombus morphology The remodelling algorithm was designed to capture platelet adhesion behaviour (VWF translocation followed by integrin mediated firm arrest) without a detailed description of single bond kinetics. In the absence of remodelling, the clot structure was found to be dendritic with dangling platelet clusters or mechanically-unlikely platelet assemblies. One of the reasons for addition of the remodelling was because the clots formed in the presence of thrombin were much larger than those studied in our prior work where thrombin was absent and dendritic growth was far less developed [bib_ref] Multiscale prediction of patientspecific platelet function under flow, Flamm [/bib_ref]. The remodelling algorithm produced compact clot structures with platelet deposits that mimicked the rough modulating thrombus morphology found in experiment [fig_ref] Figure 6: Comparison of thrombus morphology with and without the remodelling algorithm [/fig_ref]. For an initial wall shear rate of 200 s −1 , the predicted rough thrombus surface grew linearly with time reaching a thickness of ∼50-60 microns in 400 s. This overall clot growth rate of ∼8 micron/min emerges from the simulation as a clinically relevant and experimentally testable metric. The multiscale model simulated the contour evolution of platelet mass adherent to the collagen surface, the changing wall shear conditions over the rough clot, the activation and deposition dynamics of each 539 single platelet and the spatiotemporal concentration distribution of important agonists, ADP, TXA 2 and thrombin [fig_ref] Figure 8: Fully resolved thrombosis for whole blood flow over collagen/TF after 400 s [/fig_ref] , Supplementary Movie S1 and S2). The collagen surface created the initial boundary that captured the freely flowing resting platelets (note: F(0) = α min = 0.001 for resting platelets which can increase up to 1000-fold with activation). The intracellular calcium concentration rises slowly for the platelets bound to collagen but remains elevated, as driven by the NN calculator. Additionally, exposure to a strong thrombin flux enhanced activation (via PAR1/4 signalling within the trained NN calculator) for newly arriving platelets recruited to the growing deposit over the collagen. When the cumulative calcium level reached a threshold of ξ crit = 10μM-s at approximately 60 s, ADP and TXA 2 were released forming a concentration boundary layer that worked synergistically with the thrombin in the boundary layer to strengthen the platelet calcium signalling. The concentration boundary layer thickness for ADP, TXA 2 and thrombin extended to 10 μm beyond the clot boundary, which was consistent with our previous work [bib_ref] Multiscale prediction of patientspecific platelet function under flow, Flamm [/bib_ref]. The ADP and TXA 2 concentration levels were within the effective dynamic range (<10 × EC 50 ) and were significantly higher than those in the absence of the thrombin interaction in our prior work because significantly more platelets became activated. The ADP and TXA 2 concentrations ranged from 5 to 10 μM and 0.2 μM to 0.3 μM, respectively, in the clotting region for both constant flow and pressure relief settings. During clotting, the agonist exposure was multicomponent and time-varying. This spatial variation in agonist exposure created three different calcium signals [fig_ref] Figure 6: Comparison of thrombus morphology with and without the remodelling algorithm [/fig_ref]. For the first layer of deposition, the platelets exposed to both collagen and thrombin at the bottom surface became the most activated due to the sustained calcium mobilization induced by these two potent agonists. For the next few layers that were about 5-20 μm above the surface, the platelets were exposed to a sustained high level of thrombin stimulation especially before ADP and TXA 2 were produced. The calcium signal rose rapidly but decayed afterwards. For the outer most layers (>20 μm above the surface), platelets were mostly exposed to high ADP and TXA 2 concentrations but relatively low thrombin stimulation, which was reflected in a less pronounced increase in calcium signal followed by a similar decaying process as those undergoing major thrombin influence (typically in the 5−20 μm layer). Another emergent feature of the simulation was that arriving platelets could facilitate growth at the front of the clot in the upstream direction [fig_ref] Figure 8: Fully resolved thrombosis for whole blood flow over collagen/TF after 400 s [/fig_ref] Movie S1). The accumulation of upstream platelets was consistent with experimental observation of clot growth within the microfluidic clotting assay. As platelets in the near wall layer were convected by the flow in the simulation, the approaching platelets added to the front of the immobilized platelet deposit, even in the absence of underlying collagen. ## Constant flow and pressure relief In both experiments and corresponding simulations, two different flow modes were considered: the constant flow rate mode and the pressure relief mode . In the constant flow rate mode, blood was perfused at a constant flow rate to ensure the initial inlet wall shear rate of 200 s −1 was maintained during the entire course of the experiment and simulation. This is an easily and commonly achieved experimental condition by setting a constant pump speed at the start of the experiment, but this flow mode is not physiological. There was a considerable increase in wall shear rate along the plateletfluid boundary in the simulation. As the thrombus started to grow to the point where no more platelets could form a stable bond with the existing thrombus, the shear rate became highly variable along the thrombus contour and reached a maximum shear of about 3000 s −1 (a pathologically high shear rate). At ∼350 s, the platelet accumulation rate was significantly reduced. At 400 s, the thrombus grew to approximately 85% occlusion with a maximum height of about 50 μm. However in the body, as a blood vessel becomes fully occluded, flow is diverted to other vessels. This is achieved experimentally using microfluidics in the pressure relief mode. For paired channels (one with blood treated with EDTA to quench all clotting), blood flow was diverted to the open EDTAchannel as platelets accumulated in the thrombotic active channel. Initially, the wall shear rates along the platelet-fluid boundary grew similarly as in the constant flow mode [fig_ref] Figure 7: Comparison of different thrombin flux signals for constant flow rate simulations at... [/fig_ref] because a partial blockage creates little pressure drop. Only at times longer than 300 s did the growing platelet mass cause a significant pressure drop and thus reduce flow in both the experiment and in the pressure relief mode simulation [fig_ref] Figure 9: Comparison of experiment [/fig_ref]. As the channel approached full occlusion, the maximum shear rate experienced by the platelet mass was ∼1000 s −1 , considerably less than for the constant flow rate case. Because the unbinding kinetic rates were shear-dependent [Equation (13)], the pressure relief mode allowed the platelet mass to grow to full occlusion. Another noticeable feature for the pressure relief mode was that the agonist concentration was much higher than those under constant flow settings at the later stages of aggregation. This was also expected as the flow was essentially stopped at the later stages approaching full occlusion as resistance to flow caused by the thrombus increased. With convective dilution significantly reduced, the agonist transport was mainly diffusive, leaving a much thicker boundary layer than the one with constant flow rate. We have previously observed in experiments that autocrinic effects through ADP and TXA 2 can be particularly potent the moment when flow stops , effectively leading to a flow-dependent quorum sensing mechanism. Also, the multiscale model was able to capture platelet deposition dynamics and morphological information under different levels of TF stimulation consistent with microfluidic experiments conducted under pressure relief conditions [fig_ref] Figure 9: Comparison of experiment [/fig_ref]. At nearly complete occlusion, a total of about 1100 and 1300 platelets were deposited on the collagen patch for 0.1 molecules/μm 2 (low) and 10 molecules/μm 2 (high), respectively, which corresponded to about 300,000 platelets in a 0.0045 μL volume [(250+50 μm)long × 250-μm wide × 60-μm high], or ∼200-fold the concentration of PRP which has a concentration of ∼300,000 platelets/μL. In both simulation and experiment, after a lag of about 50 s, the thrombus mass grew linearly from 50 to ∼400 s, for both high and low TF level [fig_ref] Figure 9: Comparison of experiment [/fig_ref]. At low TF levels, a five-fold decrease in maximum thrombin flux was chosen. The typical maximum thrombin concentration [fig_ref] Figure 10: Evolution over 400 s of flow and thrombosis [/fig_ref] , Supplementary Movies S1 and S2) reached ∼90 and ∼20 nM for high and low TF cases, respectively, which was within an expected range of thrombin concentration reported for experiments [bib_ref] The effect of factor VIII deficiencies and replacement and bypass therapies on..., Onasoga-Jarvis [/bib_ref] [bib_ref] FXIa and platelet polyphosphate as therapeutic targets during human blood clotting on..., Zhu [/bib_ref]. Higher levels of TF saturated the clot thrombin level. This threshold behaviour in thrombin generation was also reported in other multiscale models with coagulation cascade embedded [bib_ref] Surface-mediated control of blood coagulation: the role of binding site densities and..., Kuharsky [/bib_ref] [bib_ref] Grow with the flow: a spatial-temporal model of platelet deposition and blood..., Leiderman [/bib_ref]. For the low and high TF cases in pressure relief mode, the simulated instantaneous volumetric flow rates matched accurately with the experimental measurements [fig_ref] Figure 9: Comparison of experiment [/fig_ref]. The multiscale model was able to predict important outcomes such as occlusion time for both low and high TF levels. Under high TF condition, the channel was occluded at around 450 s, about 100 s faster than under low TF stimulation. Individual clot morphology was also correctly predicted by the model. As seen in both simulation and experiment, the clot width extended upstream and downstream of the initial 250 μm collagen patch [fig_ref] Figure 9: Comparison of experiment [/fig_ref]. TF level did not seem to have a substantial effect on clot morphology but did affect the thrombus growth rate. [fig_ref] Figure 9: Comparison of experiment [/fig_ref] shows a typical 2D thrombus evolution profile under high TF condition, which compares well against the experimental height distribution averaged over a total of 40 confocal scans of the cross-section of the 3D channel [fig_ref] Figure 9: Comparison of experiment [/fig_ref]. At an earlier stage of clotting, platelets were shown to form individual clumps with sizes ranging from 20 to 60 μm that were quite consistent with individual line scans. ## Agonist study The model was also applied to different pharmacological conditions to study the role of ADP, TXA 2 or prostacyclin-like signalling alters platelet deposition dynamics. In a sequence of simulations, the concentration of each agonist, ADP or TXA 2 , was set to zero to isolate the effect of one agonist as compared to the control case. In addition, the prostacyclin analogue iloprost was set to maximum level to simulate the case of strong IP receptor stimulation. These computational conditions correspond precisely to microfluidic experiments using MRS2179 for P2Y 1 inhibition (removal of ADP), aspirin for COX-1 inhibition (removal of TXA 2 ) and ilopost for IP receptor stimulation. [fig_ref] Figure 11: Comparison of experiment [/fig_ref] and b shows a comparison between the simulated platelet number and experimental fluorescence for the three agonist conditions in contrast to the control case under the low TF stimulation. Iloprost had the strongest effect by strongly inhibiting the platelet function in the experiment and in the simulation. Removing ADP also significantly reduced platelet accumulation by almost 50% in both experiment and simulation. Inhibition of TXA 2 had the least pronounced effect, as expected for aspirin, which is a less potent antiplatelet agent. There was considerable agreement between experiments and simulations for different agonist conditions in the presence of wall generated thrombin, which represents an experimental validation of the functional forms of the adhesion dynamics which centred on the calcium signal as a metric and the thrombin wall flux used in the simulation. # Discussion A major challenge in modelling blood clotting is the strong coupling of intracellular signalling, extracellular protease cascades, hemodynamics and multicomponent mass transfer. Currently, massive computational expense would be associated with: (i) single molecule resolution of adhesion dynamics, (ii) detailed bottom-up models of single platelet signalling, (iii) wall-initiated thrombin generation and (iv) complex 3D pulsatile flows. To facilitate multiscale simulation of clotting in two dimensions at reasonable computational cost, we implemented cell activation-driven adaptive meshing , a rapid calculation of aggregate restructuring (Figs 5 and 6), a model of thrombin generation implemented as a boundary condition rather than as additional governing equations in the domain . Additionally, a data-driven NN model of platelet calcium mobilization used PAS , allowing a very rapid calculation of 'average healthy human' platelet response to multiple, time-dependent stimuli. Together, the resulting computational efficiency and the multiscale model structure using PAS-NN allows platelet phenotyping for patient-specific simulation of thrombosis. Also, the stochastic nature of the LKMC allowed the evolution of rough thrombus contours as observed in real clotting processes, a result impossible to obtain with continuum models. The model predicted platelet calcium responses based on spatiotemporal exposures and location within the clot [fig_ref] Figure 8: Fully resolved thrombosis for whole blood flow over collagen/TF after 400 s [/fig_ref]. Platelets in contact with collagen have sustained calcium mobilization. Platelets deposited subsequently become activated to a lesser extent than the first layer of platelets with the last arriving platelets displaying the least activation driven only by eluting ADP and thromboxane [fig_ref] Figure 6: Comparison of thrombus morphology with and without the remodelling algorithm [/fig_ref]. These differences in calcium signalling rate emerge from the multiscale simulation and represent tissue-scale gradients that may underlie the observed core-shell architecture observed in experiments in vivo in mice and with human blood ex vivo by [bib_ref] A systems approach to hemostasis: 1. The interdependence of thrombus architecture and..., Welsh [/bib_ref]. To simulate experimental studies in arterials (∼30 μm diameter) or microfluidics (60 μm channel height) with wall-derived TF, imposing a boundary flux parameterized by key parameters, maximum signal and time to reach maximum represents a suitable coarse-graining approach. Along with the aggregate restructuring model, the multiscale model predicted clot morphology [fig_ref] Figure 9: Comparison of experiment [/fig_ref] , upstream clot growth [fig_ref] Figure 9: Comparison of experiment [/fig_ref] , clot growth rate and occlusion time for the cases of high and low TF stimulation [fig_ref] Figure 9: Comparison of experiment [/fig_ref]. The use of apparent kinetics for cell-activation dependent binding/detachment [Equations (7)-(13)] and the platelet restructuring algorithm reduced by at least ∼10 3 -fold the numerical burden of single cell/single molecule resolution via adhesive dynamics simulations. Improved resolution of concentration boundary layers over a rough boundary was achieved through the use of adaptive remeshing without a significant increase in computational burden. The remeshing algorithm used a global influence function to distinguish regions with sharp gradient change from large regions of unreactive blood flow outside the boundary layer. This feature of the remeshing algorithm will be essential for larger, 3D simulations of coronary or carotid artery thrombosis where the flow field is substantially larger and more complex. This remeshing algorithm can be readily applied to a variety of hierarchical systems biology problems with a developing boundary layer/reaction front that is driven by discrete cell processes in the presence of prevailing transport physics. This is the first report of a data-driven platelet model of thrombosis under flow in the presence of surface collagen/TF and convective-diffusive transport of ADP, TXA 2 and thrombin. Importantly, the central predictions of the simulation were tested against microfluidic experiments conducted with human blood under the exact conditions of the simulations. Two types of operational modes, constant flow and pressure relief were studied. There was considerable agreement between experimental measurements and numerical prediction on platelet morphology and thrombus formation dynamics. Clearly, clot growth rate and final size were regulated by prevailing wall shear rates. Under constant flow for constant inlet wall shear rate of 200 s −1 , the thrombus typically reached ∼85% occlusion before the single platelet deposition on-rate was balanced by the single platelet erosion off-rate from the surface. Although the multiscale model did not treat whole-clot fracture from the surface (embolism), the steady state clot thickness of 50 μm (∼85% occlusion of the 60-μm channel) occurred when the wall shear rate experienced on the clot surface exceeded ∼6000 s −1 and the net addition of platelets was zero (not shown). More interestingly, the model was able to predict platelet deposition dynamics for a set of clinically relevant pharmacological conditions. One of the key features of this multiscale model was the use of intracellular calcium signalling to simulate cell-cell interactions. This signalling approach was shown to have excellent agreement with microfluidic experiments conducted on healthy donors in response to various blood modulators . This model has shown the ability to be used as a reliable prediction of blood function under hemodynamic conditions. In future work of thrombosis in arterial stenoses, the effect of extreme shear rates on von Willebrand structure and function [bib_ref] Enhanced platelet adhesion and aggregation by endothelial cell-derived unusually large multimers of..., Kumar [/bib_ref] [bib_ref] Factor VIII is synthesized in human endothelial cells, packaged in weibel-palade bodies..., Turner [/bib_ref] will require consideration and parameterization. Additionally, numerical integration of the local instantaneous thrombin concentration provides a metric for estimating fibrin monomer generation/transport as well as fibrin formation. Fibrin, in turn, would quench platelet detachment rates. Fibrin deposition has been observed in the region closest to the TF-laden surface and to display spatial variations, typically with more fibrin appearing downstream . In summary, flow affects clotting and clotting affects flow. The multiscale model honours this coupling by deploying data-driven kinetic information, full bottom-up reaction networks, and fully resolved hemodynamic and transport phenomenon. The computational efficiency of the algorithm is directed at supporting full simulation of thrombosis using patient-specific blood phenotypes in relevant arterial vasculatures. # Funding This work was supported by, NIH U01-HL-131053 (T.S., S.L.D.) and CBET-1404826 (T.S.). # Supplementary material Supplementary material is available at http://imammb.oxfordjournals.org. [fig] Figure 5: Remodelling algorithm for platelet deposition. Upon binding, the number of contacts the adhering platelet experiences (colour coded at each cell centre position along the trajectory) is calculated until forbidden overlap is achieved, whereby the most favourable binding site was chosen based on the cumulative potential binding rates. [/fig] [fig] Figure 6: Comparison of thrombus morphology with and without the remodelling algorithm. The remodelling algorithm resulted in physically realistic clot morphology (bottom) as compared to the highly dendritic and mechanically unlikely structure obtained without remodelling upon adhesion (top). [/fig] [fig] Figure 7: Comparison of different thrombin flux signals for constant flow rate simulations at initial inlet wall shear rate = 200 s −1 . Different thrombin flux conditions were imposed on the wall. Case I represents normal physiological condition (T max = 160 s, J max = 8E-12 nmol/μm 2 ) under high TF stimulation. Case II has a five-fold reduction in J max . Case III shows a base case with zero thrombin. (a) Thrombin boundary layer for three cases at 200 s. (b) Averaged thrombin for the thrombus as a function of time for three cases. (c) Snapshots of fully-developed thrombus configuration at 400 s for three different cases. (d) Comparison of platelet accumulation dynamics for three different cases. (e) Averaged shear rate at the outmost layer of platelets for three different cases. (f) [/fig] [fig] Figure 8: Fully resolved thrombosis for whole blood flow over collagen/TF after 400 s. At constant inlet wall shear rate of 200 s −1 , [/fig] [fig] Figure 9: Comparison of experiment (a, c) and multiscale simulation (b, d) for occlusive thrombosis. Platelet deposition (a, b) and flow rate (c, d) are shown for pressure relief mode (initial wall shear rate, 200 s −1 ) for whole blood flow over high and low TF (10 or 0.1-molecules/μm 2 ). Clot morphology and clot height over collagen (located between 100 and 350 μm) is shown at various times of clotting (e: simulation, top panel; experiment, bottom panel). [/fig] [fig] Figure 10: Evolution over 400 s of flow and thrombosis (ADP, TXA 2 , thrombin) under pressure relief mode (top) or at constant flow rate (bottom). Initial inlet wall shear rate = 200 s −1 . [/fig] [fig] Figure 11: Comparison of experiment (a, c, e) and multiscale simulation (b, d, f) for different agonist conditions. Measured platelet deposition in the presence of aspirin, MRS 2179 and iloprost treatment (a) and corresponding multiscale simulation (n = 8) with no TXA 2 , no ADP and iloprost stimulation (b). [/fig]

Epigenetic Regulation of microRNAs in Cancer: Shortening the Distance from Bench to Bedside Citation: Pajares, M.J.; Alemany-Cosme, E.; Goñi, S.; Bandres, E.; Palanca-Ballester, C.; Sandoval, J. # Introduction MicroRNAs (miRNAs) are 18-25-nucleotide-long, non-coding RNAs with critical roles in a variety of biological processes, such as proliferation, differentiation, or immune response. They exert their function through the regulation of gene expression, mostly at the post-transcriptional level. Most miRNA sequences are located within introns of coding genes or in intron and exons of non-coding RNAs. miRNAs regulation occurs mainly due to genetic or epigenetic mechanisms. Epigenetic mechanisms include DNA methylation, post-translational modification of histones, and RNA modification. In addition, a group of miRNAs, called epi-miRNAs, can modulate the expression of DNA methyltransferases (DNMTs), histone deacetylases (HDACs), or histone methyltransferases (HMTs), affecting the expression of both coding and non-coding genes and consequently having a clear impact on the global epigenome. Furthermore, miRNAs can interact with complementary sequences in gene promoters, representing a platform for the assembly of specific protein complexes that regulate gene expression through changes in the chromatin structure. Interestingly, aberrant expression of miRNAs is associated with different diseases, especially cancer. Besides, there is a growing list of reported miRNAs with an oncogenic function (referred to as "oncomiRs") as well as miRNAs with a tumoral suppressing function (namely "oncosuppressor miRs") in several neoplastic malignancies. Importantly, their . The biogenesis of microRNAs. MicroRNA genes are generally transcribed by RNA Polymerase II in the nucleus to form large pri-miRNA transcripts, which are capped and polyadenylated. These pri-miRNA transcripts are processed by the RNase III enzyme Drosha to release the ~70-nucleotide pre-miRNA precursor product. Exportin 5 (XPO5) transports the pre-miRNA into the cyto- . The biogenesis of microRNAs. MicroRNA genes are generally transcribed by RNA Polymerase II in the nucleus to form large pri-miRNA transcripts, which are capped and polyadenylated. These pri-miRNA transcripts are processed by the RNase III enzyme Drosha to release the~70nucleotide pre-miRNA precursor product. Exportin 5 (XPO5) transports the pre-miRNA into the cytoplasm. Subsequently, another RNase III enzyme, Dicer, processes the pre-miRNA to generate a transient~22-nucleotide miRNA: miRNA* duplex. This duplex is then loaded into the miRNAassociated multiprotein RNA-induced silencing complex (mi-RISC). The mature miRNA then binds to complementary sites in the target mRNA to induce an RNA-binding protein decoy, activation of translation, or inhibition of translation by mRNA degradation or translation repression. Created with Biorender.com. ## Epigenetic regulation of mirnas expression Epigenetics refer to inheritable features that are related to alterations or changes outside the DNA nucleotide sequence, giving rise to changes in gene expression patterns. As other coding and non-coding genes, miRNAs have been described to be regulated by epigenetic mechanisms such as DNA methylation, histone, and RNA modifications. Besides, TFs have been shown to facilitate the recruitment of epigenetic regulators to gene promoters, contributing to epigenetic control of gene expression in different scenarios. The work of Ozsolak and collaborators in 2008, identifying the miRNA promoter structures, was important to establish the miRNA expression regulatory mechanisms. In fact, high-throughput analysis of miRNA promoter structures by nucleosome mapping, and H3K4me3 and H3K9/14ac ChIP-Chip screening, confirmed the similarity of the RNAPIItranscribed miRNA promoters and mRNA-encoding promoters regarding the CpG island, TATA elements, TFIIB recognition elements (BRE), initiator (Inr), and other elements. Notably, the nucleosome occupancy information surrounding the miRNA transcription start site (TSS) was also used for the discovery of TF-mediated regulation of miRNAs. ## Dna methylation DNA methylation consists of the covalent addition of a methyl group in cytosine nucleotides (5-methylcytosine, 5-mC), usually within CpG dinucleotides that are concentrated in CpG islands. Around 60% of these CpG islands are located in gene promoter regions, where DNA methylation causes transcriptional repression, enabling the binding of repres-sor proteins and preventing the interaction between TFs and DNA. 5-mC can also be found in gene bodies and intergenic regions, encountering some differences regarding transcription regulation. Similar to promoters, 5-mC accumulation in intergenic regions and repetitive elements is associated with genomic integrity. Remarkably, many CpG islands have been found outside TSS, indicative of unannotated transcripts or enhancer elements. Regarding gene bodies, high levels of 5-mC have been found in highly expressed genes. This apparent paradox has been related to the protection of the gene body from spurious RNA polymerase II entry and cryptic transcription initiation, leading to the fidelity of gene transcription initiation. DNA methyltransferase family enzymes are responsible for the covalent addition of methyl groups, being specific for each substrate. For instance, DNMT1 acts on hemimethylated DNA and maintains the methylation of the newly synthesized strand in DNA replication. DNMT3A and DNMT3B are responsible for de novo methylation of unmethylated DNA. Significantly, DNA hypermethylation in promoter regions is associated with transcriptional repression by different mechanisms, such as the prevention of TF binding, the recruitment of histone deacetylases, and the recruitment of methyl-CpG-binding proteins a with repressive function. In contrast, DNA demethylation is mediated by the Tet methylcytosine dioxygenase (TET) family of enzymes, regardless of DNA replication. Importantly, the first evidence suggesting that miRNAs might be regulated by DNA methylation was published in 2006. Saito and collaborators treated bladder cancer cells with the demethylating agent 5-Aza-2 -deoxycytidine (5-aza-dC), leading to the upregulation of 17 out of the 313 human miRNAs characterized by a miRNA microarray. Notably, one of these miRNAs, miR-127, was embedded in a CpG island and was significantly upregulated upon treatment, proving for the first time that DNA methylation regulated miR-127 expression. Since then, around 50% of the miRNAs have been described to be located in CpG island-rich positions and therefore to be subjected to epigenetic regulation. The expression of neighboring mRNAs was not analyzed in this study, but Ozsolak and colleagues would confirm that there is no expression correlation between intronic miRNAs having distinct promoters, and their host's mRNAs. DNA methylation changes have been largely observed in different pathologies, particularly in cancer. The latter are characterized by global genomic demethylation, especially mobile genetic elements, and selective hypermethylation of regions exhibiting tumorsuppressor functions. ## Histone modifications Eukaryotic cells have their DNA highly packed into the nucleus as an assembly of DNA and DNA-interacting proteins, mainly histones. Histones are basic proteins that form the octamer structures (H3, H4, H2A, and H2B), which interact with DNA, leading to the formation of nucleosomes. These are no longer believed to be static entities but very dynamic in the regulation of gene expression. Every histone protein owns a characteristic side chain or tail, which is mainly composed of basic lysine and arginine residues. The histone tails experience extensive covalent post-translational modifications that will ultimately lead to changes in chromatin organization and packaging. The main modifications that have been studied so far are acetylation, methylation, and phosphorylation, but there are many others, such as citrullination, ubiquitination, ADP-ribosylation, deamination, formylation, O-GlcNAcylation, propionylation, butyrylation, crotonylation, and proline isomerization. ## Histone acetylation Acetyl groups can be added at lysine residues on histone tails, leading to the neutralization of the basic charge of histones, mostly localized at the enhancers, promoters, and gene bodies. This modification weakens the interaction between the negatively charged DNA and histones, promoting the active transcription of genes. The most frequent acetylated residues are lysine 27 and lysine 9 from the H3 histone. They are usually placed in enhancers and promoters by typical "writers" (histone acetyltransferases), such as the P300/CREB-binding protein (CBP), and removed by "erasers", such as HDACs/sirtuins. HDAC proteins have been reported to be altered in a wide variety of disorders, including cancer. Overexpression of HDAC leads to the compaction of chromatin and repression of transcription, and thus, the use of different HDAC inhibitors (HDACi) (4-sodium phenylbutyrate (PBA), trichostatin A (TSA), or suberoylanilide hydroxamic acid (SAHA) have led to the discovery of HDAC-repressed miRNAs in cancer. HDACi treatment in cancer cell lines led to the derepression of several miRNAs, which confirmed the role of HDACs in the regulation of miRNAs. For instance, the induction of miR-200aand miR-200cwas observed in breast cancer cells upon HDACi treatment, leading to the inhibition of cell proliferation, invasion, and migration. In pancreatic cancer cell lines, HDACi treatment led to the induction of miR-34a expression, causing the inhibition of tumor-progression-related features, such as cell proliferation, cell cycle progression, epithelial to mesenchymal transition (EMT), and invasion. In line with this, the combination of different HDACi was shown to induce the expression of miR-31 in breast cancer cells, resulting in cellular senescence. These and other examples discussed in Section 5 point out the critical role of HDACs in the regulation of miRNA expression and their dysregulation in cancer. ## Histone methylation The addition of a specific number of methyl groups (-CH 3 ) in specific lysine or arginine residues is tightly regulated, having a pivotal role in gene regulation. Lysine residues in histones can be mono-, di-, or tri-methylated and will have a different effect on gene regulation, depending on the position of the post-translational modification. Regarding lysine residues, H3K4 trimethylation (localized at gene promoter regions) and H3K4 monomethylation (enriched at enhancer and promoter regions), H3K36 trimethylation (mainly distributed within the gene body), and H3K79 dimethylation (localized both in the promoter and along the gene body) modifications are usually gene-activating marks. On the other hand, H3K9 trimethylation and H3K27 monomethylation marks repress gene expression. Importantly, each modification is associated with specific writers (histone methyltransferases) and erasers (histone demethylases), such as the couple SET Domain Containing 1A/D (SETD1A/D) (writers) and Lysine Demethylase 5A/B/C (KDM5A/B/C) (erasers) in the case of H3K4me3; for H3K36me3, SET Domain Containing 2D (SETD2) and Lysine Demethylase 4 (KDM4); and for H3K4me1, Mixed Lineage Leukemia protein-1-5 (MLL1-5) and Lysine Demethylase 1A/B (KDM1A/B). Finally, for H3K9me3, the writer Suppressor Of Variegation 3-9 Homolog 1/2 (SUV39H1/2) and eraser Lysine Demethylase 3/4(KDM3/4), and Enhancer Of Zeste 1/2 Polycomb Repressive Complex 2 Subunit (EZH2/EZH1) and Lysine Demethylase 6 A/B (KDM6A/B) in the case of H3K27me1. Noteworthy, several histone methylation modifiers have been related to miRNA regulation, such as EZH2 in ovarian cancer, which has been shown to induce the repressive histone mark H3K27me3 in target miRNAs: miR-101-3p, let-7e-5p, miR-26a-5p, miR-98-5p, and miR-141-3p. In lung cancer cells, KDM5B was shown to inhibit the miR-200 family via demethylation of H3K4me3, inducing EMT of cancer cells. KDM5A upregulation was proven to promote cervical cancer progression by repressing miR-424-5p through directly interacting with its promoter region and removing the H3K4 methyl groups. In Section 5, more details regarding histone modification and miRNA regulation in cancer will be discussed. ## Rna modifications To date, more than 140 RNA modifications have been discovered as adenosine methylation (m6A), cytosine methylation (m5C), ribose methylation (2 -O-Me), and pseudourylation (Ψ). Among them, m6A is one the most prevalent internal modification of mRNA. m6A has been found in around 7000 genes and is catalyzed by a complex of proteins with methyl-transferase activity, including methyltransferase-like 3 (METTL3) and methyltransferaselike 14 (METTL14). It has been shown that the presence of m6A might alter gene expression levels, mRNA stability, translation efficiency, and other relevant functions. Importantly, Alarcon and colleagues found that m6A was also localized in non-coding genes, such as the pri-miRNAs, promoting their recognition by DGCR8 and allowing the genesis of the pre-miRNAs. Besides, they demonstrated that downregulation of m6A by depletion of METTL3 led to a reduction in most mature miRNAs. ## Mirnas as epigenetic regulators In the previous section, the epigenetic regulation of miRNAs has been addressed. Conversely, many studies have proven that miRNAs themselves might act as epigenetic regulators (epi-miRNAs), which post-transcriptionally target the factors belonging to the epigenetic machinery, such as DNMTs or DNA-demethylases, histone acetylases or HDACs, and histone methyltransferases (EZH2) or demethylases (KDMs). Epi-miRNAs, likewise other miRNAs, target the 3 -UTRs of the mRNAs, inducing their degradation. The latter enables the regulation of DNA methylation, histone acetylation, and histone methylation, with the consequent changes in global gene transcription. Importantly, epi-miRNAs themselves can be epigenetically regulated, conforming to the regulatory circuits that often appear deregulated in several disorders. Many demethylases have been involved in epi-miRNA reciprocal regulatory loops, such as DNMT3A with miR-29a/b in lung cancerand miR-200c in gastric cancer. EZH2 has also been frequently involved in epi-miRNA regulatory loops. For instance, the previously mentioned study from Liu and collaboratorsconfirmed the reciprocal regulation of EZH2 and a set of five epi-miRNAs (miR-101-3p, let-7e-5p, miR-26a-5p, miR-98-5p, and miR-141-3p) in ovarian cancer, promoting malignant proliferation by maintaining the high expression of EZH2. Other examples will be further characterized in Section 5. In addition, deregulated m6A modification is an important hallmark of various diseases, including cancer. Notably, regarding the regulation of RNA modifications by miR-NAs, it was proved that miRNAs regulate m6A RNA methylation by modulating the binding of METTL3 to mRNAs. Outside of the post-transcriptional function of many miRNAs, it has been shown that miRNAs might remain retained in the nucleus, regulating gene activation and silencing, via recruitment of other epigenetic factors. In fact, several miRNAs have been reported to activate gene transcription via enrichment of markers for transcriptionally active promoters (H3K4me3) and recruitment of polymerase II (Pol II). Regarding gene silencing, many other miRNAs have been proven to repress gene transcription by the recruitment of Polycomb proteins and inducing the H3K27me3, which maintains the chromatin in a condensed form. All these mechanisms combine tightly in order to regulate gene expression and, thus, the relevant biological processes. The deregulation of any of the aforementioned mechanisms in cancer might have tremendous effects, disturbing the circuits that control important processes, including cell proliferation or apoptosis. ## Epigenetic alteration and microrna dysregulation in cancer ## The cancer epigenome landscape Cancer diseases are characterized by extensive epigenetic changes. The first epigenetic abnormality described in human tumors was the loss of DNA methylation back in 1983. It is well-established that genome-wide DNA hypomethylation is a frequent feature of human cancers, which can be found in the early stages of carcinogenesis and associated with tumor progression. The overall low levels of genomic DNA methylation is due to the hypomethylation of highly repeated DNA sequences (such as longinterspersed retrotransposable elements (LINEs), short-interspersed retrotransposable elements (SINEs), and long terminal repeats (LTRs)). In contrast, the genomic regions associated with hypermethylation are gene regions, mainly localized in promoterassociated CpG islands. In fact, the inactivation or downregulation of tumor-suppressor genes via promoter hypermethylation is commonly observed in most types of human cancers. Some examples of silenced genes by CpG-island hypermethylation include the inhibitor of the JAK-STAT pathway SOCS1 in the liver and myeloma tumors or the cell-cycle inhibitor Rb in retinoblastoma tumors. Although the impact of global DNA hypomethylation on cancer is less straightforward than that of the localized hypermethylation, it is also thought to contribute to cancer development by generating chromosomal instability, reactivating transposable elements, or causing the loss of genomic imprinting. On the other hand, an aberrant pattern in the histones' post-translational modifications in cancer has also been described, leading to the reconfiguration of the entire genome ## Epigenetic alteration and microrna dysregulation in cancer ## The cancer epigenome landscape Cancer diseases are characterized by extensive epigenetic changes. The first epigenetic abnormality described in human tumors was the loss of DNA methylation back in 1983. It is well-established that genome-wide DNA hypomethylation is a frequent feature of human cancers, which can be found in the early stages of carcinogenesis and associated with tumor progression. The overall low levels of genomic DNA methylation is due to the hypomethylation of highly repeated DNA sequences (such as long-interspersed retrotransposable elements (LINEs), short-interspersed retrotransposable elements (SINEs), and long terminal repeats (LTRs)). In contrast, the genomic regions associated with hypermethylation are gene regions, mainly localized in promoter-associated CpG islands. In fact, the inactivation or downregulation of tumor-suppressor genes via promoter hypermethylation is commonly observed in most types of human cancers. Some examples of silenced genes by CpG-island hypermethylation include the inhibitor of the JAK-STAT pathway SOCS1 in the liver and myeloma tumors or the cell-cycle inhibitor Rb in retinoblastoma tumors. Although the impact of global DNA hypomethylation on cancer is less straightforward than that of the localized hypermethylation, it is also thought to contribute to cancer development by generating chromosomal instability, reactivating transposable elements, or causing the loss of genomic imprinting. On the other hand, an aberrant pattern in the histones' post-translational modifications in cancer has also been described, leading to the reconfiguration of the entire genome during the tumor process. As previously outlined, these modifications in histones are due to alterations in the levels of the regulatory enzymes, such as histone deacety-lases (HDAC1, HDAC2) or histone demethylases (lysine-specific demethylase LSD1)and have a preponderant role during EMTand in the regulation of tumoral metastasis. In cancer, deregulation of the histone writers and erasers can lead to the histone hypoacetylation of oncosuppressor miRs or hyperacetylation of oncomiRs. Concerning miRNAs, their dysregulation is also a common feature of human cancers. In the past decades, their relevant role in tumor onset, growth, and metastasis has been demonstrated. Generally, the expression of miRNAs is downregulated in tumors compared to their corresponding healthy tissues. This leads to the idea that many miRNAs could be acting as oncosuppressor miRs. Nevertheless, overexpression of miRNAs functioning as oncogenes has also been described in human tumors. All in all, it has to be kept in mind that miRNAs can have multiple targets and can function as either tumor suppressors or oncogenes under different circumstances, depending on the tissue or cell type where they exert their function. Consequently, to understand the repercussion of miRNA dysregulation, it is crucial to pay attention to cancer-specific miRNA expression patterns. The major causes of miRNA dysregulation in malignant cells are the amplification, deletion, or translocation of the miRNA-encoding genes, abnormal epigenetic modifications, defects in the miRNA biogenesis machinery, or widespread transcriptional repression. It is also noteworthy that a significant number of miRNA genes are located within cancer-associated genomic regions or fragile sites. ## Mirnas in the control of critical cancer-related pathways Cancer diseases are characterized by the disruption of cellular homeostasis pathways, which ultimately result in uncontrolled cell growth, proliferation, and resistance to apoptosis. miRNAs function as fundamental and versatile gene regulators in cancer since they can target a large number of the pathways that sustain these essential cellular functions. Firstly, miRNAs acting as oncomiRs are typically overexpressed and enable cancerous cells to enter and progress through the cell cycle, whereas miRNAs functioning as oncosuppressor miRs, typically lost or downregulated during cancer, normally assist in the cell cycle arrest. For instance, the miR-17-92 cluster regulates the translation process of E2F transcription factor 1 (E2F1), E2F2, and E2F3, which are key cell proliferation protein regulators; in turn, a negative feedback mechanism regulates the expression of the miR-17-92 cluster. In cancer cells, miR-17-92 overexpression disrupts this negative feedback loop, leading to cell proliferation. Conversely, the miR-17-20 cluster, which represses cyclin D1 expression and suppresses breast cancer cell proliferation, has been found to be downregulated in breast tumors. miRNAs are also linked to the core apoptosis pathways in cancer. In fact, there is a growing list of identified miRNAs with both anti-apoptotic and pro-apoptotic properties, which target the central apoptotic genes such as Phosphatase and Tensin Homolog (PTEN), Caspase-9, or B-cell lymphoma 2 (BCL-2). Interestingly, involvement in apoptosis gives miRNAs a major role in cancer drug resistance. For example, miRNA-21 targets PTEN in stomach cancer and breast cancer, promoting cell resistance to a variety of drugs. On the other hand, a wide range of miRNAs has been revealed as modulators of the cellular pathways involved in senescence. Senescence is the irreversible state of cellular growth arrest and constitutes a barrier to tumorigenesis since it prevents the malignant proliferation of cells harboring oncogenic DNA mutations. Remarkably, miRNAs commonly associated with senescence have also been involved in human malignancies, such as let-7 miRNAs. In the same way, crucial genes involved in the DNA damage response, which is critical in cancer, are regulated by their specific miRNA. One great example is the linear signaling pathway of N-MYC→miR-421→Ataxia Telangiectasia Mutated (ATM), where the oncogenic transcription factor N-MYC upregulates miR-421, which targets the apical damage sensor kinase ATM. In this fashion, miR-421-mediated ATM downregulation is thought to contribute to N-MYC-induced tumorigenesis in neuroblastoma. Another important cellular process in which miRNAs play an important role is autophagy. Increasing studies have linked miRNAs to autophagic regulation during cancer initiation (such as miR-224 targeting SMAD Family Member 4 (SMAD4) in hepatocellular carcinoma (HCC)) and cancer development (e.g., miR-224-3p targeting RB1-inducible coiled-coil protein (RB1CC1) in cervical tumors). Autophagy is a multi-step lysosomal degradation process whereby a cell degrades long-lived proteins and damaged organelles. Especially in cancer cells, autophagy serves as a means of temporary survival, a relevant physiological mechanism. However, if cellular stress induces continuous or excessive autophagy, cell death ensues. All in all, miRNAs are involved in several autophagic stages in which they exert a function as oncomiRs or oncosuppressor miRs. ## Bidirectional relationship between epigenetic alterations and mirna dysregulation: cases with biological relevance in cancer diseases As indicated in Section 3, miRNA gene expression is subjected to epigenetic mechanisms, and at the same time, miRNAs have been proved to regulate the expression of epigenetic regulators. As a matter of fact, there is current evidence indicating that dysregulation of miRNAs can lead to aberrant DNA methylation in cancer diseases. Thus, a bidirectional relationship is established between epigenetic alterations and miRNA dysregulation in cancer, often being involved in regulatory loops. ## Mirnas and lung cancer miRNAs inactivation via promoter DNA methylation has shown biological significance, especially in lung cancer. For instance, aberrant CpG methylation downregulates the expression level of miR-145 in lung adenocarcinoma. miR-145 has been recognized to act as an oncosuppressor miR, having shown to be involved in tumor invasion and progression by targeting C-MYC, Astrocyte Elevated Gene-1 (AEG-1), Epidermal Growth Factor Receptor (EGFR), Nudix Hydrolase 1 (NUDT1), and Octamer-binding transcription factor 4 (OCT4) in LAC. To cite more examples, miR-127 and miR-9 promoter hypermethylation have also been proposed to play a role in non-small cell lung cancer (NSCLC) development and progression. Additionally, miR-34b/c promoter hypermethylation is a frequent event in lung adenocarcinoma, and low levels of miR-34b and miR-34c are associated with distant metastases. Paradoxically, it is important to note that although miR-34b/c downregulation in metastasizing lung adenocarcinomas can be a direct result of increased miR-34b/c promoter hypermethylation, the hypermethylation itself is not associated with metastasizing lung adenocarcinomas. This highlights the complex regulatory networks in which miRNAs play a role in cancer and the difficulty of addressing them. A previous study already stated that miR-34b/c promoter methylation and consequent downregulation is a frequent event in lung adenocarcinomas and that restoration of miR-34b/c expression suppresses cell proliferation, migration, and invasiveness. On the other hand, Brueckner et al. demonstrated that the let-7a-3 promoter could be hypomethylated in human lung cancer, leading to its epigenetic activation and therefore identifying let-7a-3 as a miRNA gene with oncogenic function in lung cancer. However, it is important to note that the vast majority of the literature currently focuses on localized hypermethylation instead of hypomethylation concerning miRNA dysregulation, just as more studies have been conducted on the potential tumor-suppressor genes hypermethylated in cancer rather than on the potential oncogenes hypomethylated in cancer. This is probably because, as we previously described, hypermethylation tends to be in localized gene-associated regions, whereas hypomethylation tends to be generalized to the whole genome, affecting mostly repeated DNA sequences. With regard to miRNAs contributing to DNA methylation dysregulation, it has been long described in lung cancer that the miR-29 family members target both DNMT3A and DNMT3B. In fact, the relevance of miR-29s was discovered after its reinforced expression in NSCLC cell lines restored the normal patterns of DNA methylation, inducing the re-expression of methylation-silenced tumor-suppressor genes, such as Fragile Histidine Triad Diadenosine Triphosphatase (FHIT) and WW Domain Containing Oxidoreductase (WWOX), and inhibiting tumorigenicity in vitro and in vivo. Interestingly, the expression of miR-29a and miR-29b could be partially regulated in a positive feedback loop by DNMT3A and DNMT3B. On the other hand, downregulation of miR-212 correlated to the severity of the disease in lung cancer, and its transcriptional silencing was found to be associated with H3K9me2 and H3K27me3 but not DNA hypermethylation. Furthermore, HDAC5 was found to be aberrantly overexpressed in lung cancer, negatively correlating with miRNA-589 expression. Remarkably, miR-589 was found to target HDAC5 mRNA, regulating important cell cycle and EMT-related genes. Interestingly, it is the hypermethylation of the miR-589 promoter that ultimately leads to the upregulation of HDAC5. ## Mirnas and gastric cancer Even though there is less research conducted on gastric cancer (GC) on this topic, it also serves to exemplify the epigenetics-microRNA regulatory networks. With regard to miRNAs activation induced by loss of DNA methylation in GC, Tsai and co-workers demonstrated that abnormal DNA hypomethylation induced overexpression of miR-196b. Later on, it was glimpsed that miR-196b upregulation promoted the proliferation and invasion ability of GC cells by regulating the Phosphatidylinositol 3 kinase (PI3K)/Protein kinase B (AKT)/Mechanistic Target of Rapamycin Kinase (mTOR) pathway. Very recently, miR-196b has also been shown to promote GC progression by targeting Augurin Precursor (ECRG4). On the other hand, Hashimoto et al. argued that miR-181c could be silenced through methylation in GC, activating its target genes Neurogenic Locus Notch Homolog Protein 2/4 (NOTCH2/4) and KRAS and therefore contributing to the pathogenesis of GC. Zabaglia et al. also supported that downregulation of miR-181c may play an important role in GC progression by controlling the important genes associated with apoptosis. Remarkably, recent research stated that miR-129-2 was hypermethylated in tumoral tissues of GC patients, suggesting that its methylation was involved in the development of the disease. Hypermethylation of miR-129-2 in primary GC tissues was already reported two decades ago altogether with that of the aforementioned miR-34b. Hypermethylation of miR-129-2 promoter has also been reported in other cancers, such as HCC, endometrial cancer, and ovarian cancer. Interestingly, a novel epigenetic feedback loop between miR-200c and DNMT3A has been described in the carcinogenesis and progression of GC. DNMT3A upregulation is responsible for the hypermethylation of the miR-200c gene promoter in GC, ultimately causing the downregulation of miR-200c. At the same time, miR-200c directly targets DNMT3A and induces endogenous pre-miR-200c and pri-miR-200c re-expression. ## Mirnas and ovarian cancer In ovarian cancer samples, Chen et al. observed that the increase in the promoter hypermethylation of miR-193a-3p was significantly correlated with the loss of miR-193a-3p expression and tumor stage. Remarkably, by conducting in vivo studies, they concluded that loss of miR-193a-3p could enhance oncogenic capacities via activation of MAPK/ERK signaling, facilitating tumor colonization of metastatic ovarian cancer in peritoneal metastases. Regarding cisplatin resistance in ovarian cancer, Deng and colleagues revealed miR-199a-3p as an upstream regulator of Discoidin Domain Receptor Tyrosine 1 (DDR1) (which confers the malignance and cisplatin resistance of ovarian cancer) that happens to be hypermethylated in ovarian cancer. Thus, the hypermethylated miR-199a-3p gene contributes to tumor aggressiveness and cisplatin resistance through promoting DDR1 expression. Besides, in connection with ovarian cancer metastasis, Vitaly et al. recently showed the involvement of some novel hypermethylated miRNA genes in ovarian metastasis and the inactivation of miR-191 via hypomethylation with a potentially associated oncogenic role. Furthermore, another bidirectional regulation has been described between DNA methyltransferases and miRNAs, with importance in ovarian cancer: a feedback loop between miR-30a/c-5p and DNMT1 that mediates cisplatin resistance. As the authors of this study claim, miR-30a/c-5p is aberrantly methylated and thus silenced by overexpressed DNMT1, which relieves the inhibitory effect of miR-30a/c-5p on DNMT1 and Snail (a key inducer of EMT), leading to cisplatin resistance and partial EMT in ovarian cancer in vitro. On the other hand, re-expression of miR-145 in ovarian cancer cells, which is usually downregulated in this cancer, was shown to inhibit the Warburg effect by targeting DNMT3A and hexokinase-2 (HK2). Moreover, DNMT3A regulated miR-145 expression through methylation, giving rise to a negative feedback loop. Interestingly, miR-137 mediated the functional link between c-MYC and EZH2, regulating cisplatin resistance in ovarian cancer. The downregulation of miR-137 (which targets EZH2 mRNA) leads to an increased expression of EZH2, which activates the cellular survival pathways, resulting in resistance to cisplatin. ## Mirnas and breast cancer Epigenetic-miRNAs regulatory networks have also been described in breast cancer. For instance, the miR-129-2 gene has been observed to be hypermethylated in breast cancer. Furthermore, downregulation of miR-129-2 by promoter hypermethylation has been shown to regulate cell proliferation and apoptosis. Another example can be illustrated by the work of Gacem and co-workers, who determined that miR-124a-1, miR-124a-2, and miR-124a-3 genes were frequently methylated in breast cancer and played a role in tumor growth and aggressiveness. On the other hand, Hu et al. found a hypomethylated miRNA, miR-663, whose overexpression could induce chemoresistance in breast cancer cells. More importantly, almost a decade ago, Xu et al., for the first time, described a negative regulatory circuit between DNMT1 and two miRNAs, miR-148a and miR-152, in breast cancer cells. The downregulation of miR-148a and miR-152 as a consequence of their promoter methylation was inversely correlated with tumor grades and lymph node status in breast cancer tissues. These miRNAs appeared to act as tumor suppressors by targeting Insulin-like Growth Factor 1 Receptor (IGF-1R) and Insulin Receptor Substrate 1 (IRS1), often overexpressed in breast cancer. In breast cancer, the epigenetic regulation of HOX genes is also remarkable. For instance, the overexpression of the HOXA1 gene is counteracted by the expression of miR-1469, miR-99a, and miR-100 in particular BC contexts. Promoter hypermethylation of HOX genes, such as HOXA5, can also lead to altered expression levels of this gene, causing its silencing. Regarding histone modifications that provoke miRNAs dysregulation, Ryu and coworkers identified miR-708 to be transcriptionally repressed by Polycomb Repressor Complex 2-induced H3K27me3 in metastatic breast cancer. Interestingly, in patients with breast cancer, miR-708 expression was decreased in lymph nodes and distal metastases, suggesting a metastasis-suppressive role. In summary, we have provided several examples of the most frequent epigenetic alterations and miRNA aberrant expression in common cancers and how they are interrelated. Their regulation is dependent on each other. As we have evidenced, the bidirectional regulation between epigenetic mechanisms (especially DNA methylation) and miRNAs has been described in several types of cancer. These epigenetic alterations have been reported at every stage of cancer, from initiation to progression, in metastasis, and in resistance to oncologic therapies. In addition, it has been proved to affect the course of these events profoundly. This highlights their importance in cancer, and thus, the need to take them into account when trying to improve our knowledge of tumoral malignancies. ## Targets/pathways affected by mirnas dysregulation with potential clinical implications Lung cancer miR-145; miR-127, miR-9; miR-34b/cmiR let-7a-3miR-29 ⇒DNMT3A&DNMT3B; [formula] miR-29a/b ⇔ DNMT3A & DNMT3B [93] [/formula] H3K9me2 & H3K27me3 ⇒ miR-212; miR 589 ⇒ HDAC5miR-145 ⇒ c-Myc, AEG-1, EGFR, NUDT1Gastric cancer miR-181cmiR-129-2miR-196bmiR-200c ⇔ DNMT3AHDAC ⇒ miR-127miR-196b ⇒PI3K/ AKT)/ mTOR pathway; miR-196b ⇒ ECRG4Ovarian cancer mir-193a-3p, miR-199a-3pmiR-191miR-30a/c-5p ⇔ DNMT1miR-145 ⇒ DNMT3A; miR-137⇒ EZH2; miR-101-3p, let-7e-5p, miR-26a-5p, miR-98-5p, miR-141-3p ⇔ EZH2miR-193a-3p ⇒ MAPK/ERK; miR-199a-3p ⇒ DDR1; miR-30a/c-5p ⇒ SNAIL; miR-145 ⇒ DNMT3A, HK2Breast cancer miR-129-2; miR-124a-1, miR-124a-2 & miR-124a-3miR-663miR-148a & miR-152 ⇔ DNMT1H3K27me3 ⇒ miR-708miR-129-2 ⇒ BCL2L2; miR-148a & miR-152 ⇒ IGF-1R, IRS1 ## Clinical applications: mirnas epigenetics in cancer Biomarkers are indicators of either physiological or pathological biological processes. An acceptable biomarker should be accurate and highly reproducible in standardized cost-effective assays. Besides, it should be preferably measured from minimally invasive samples and provide valuable information for the patient's clinical management. Many miRNAs have been found to be aberrantly expressed in different malignancies. As we have shown, epigenetic mechanisms, such as hypo/hypermethylation of promoter CpG islands or histone post-transcriptional modifications, regulate miRNA expression. The detection of these deregulated mechanisms may serve as promising diagnostic and prognostic biomarkers in cancer as well as novel therapeutic strategies. ## Mirnas methylation as diagnostic biomarkers Many studies have highlighted the potential benefits of implementing methods to evaluate aberrant miRNA promoter methylation patterns in biological samples as a strategy for early detection of cancer. However, few studies describing novel diagnostic biomarkers based on miRNAs methylation are reported in the literature to date. This section summarizes those that we considered most promising in the context of cancer diagnosis. Toiyama and co-workers evaluated the potential of miR-1, miR-9, miR-124, miR-137, and miR-34b/c methylation levels as diagnostic biomarkers in ulcerative colitis (UC)associated colorectal cancer. Methylation of the aforementioned miRNAs was increased in cancer tissues and dysplasia compared to UC non-neoplastic tissues. The combination of all miRNAs allowed for more robust discrimination of colorectal carcinoma patients. More importantly, they found that this signature could accurately identify patients with ulcerative colitis at risk of developing colorectal carcinoma (CRC), with high sensitivity and specificity. DNA methylation-based silencing of miR-124 was shown to be a marker for improved detection of cervical cancer and its high-grade precursor lesions. Subsequently, several studies have validated the use of a methylation-based signature composed of a combination of miR-124 and other genes (MAL/miR-124-2, FAM19A4/miR-124-2...) as a triage test for the identification of premalignant lesions (cervical intraepithelial neoplasia) in high-risk human papillomavirus-positive women. Furthermore, FAM19A4/miR-124-2 methylation analysis in large cohorts of patients confirmed its value as a high-sensitivity screening method for the diagnosis of cervical cancer. To date, many efforts have been made to identify alterations in the methylation patterns in tissue samples from cancer patients. However, given the costs and risks associated with surgical biopsy, identifying these biomarkers in liquid biopsy provides great benefit for the patients. In addition, biological fluids, such as plasma, serum, urine, saliva, or stool, have been shown to provide valuable information for the diagnosis of a wide range of tumors. Using urine sediments, methylation of miRNAs has also been demonstrated to be useful to diagnose different genitourinary carcinomas. For instance, miR-193b promoter methylation levels allow the detection of prostate cancer with 91.6% sensitivity and 95.7% specificity, providing an overall accuracy of 92.9%. Comparing the methylation levels of miR-30a-5p in urine from patients with renal clear carcinoma (RCC) and asymptomatic controls, Outeiro-Pinho and colleagues have also shown the potential utility of this biomarker in the diagnosis of RCC. The overall accuracy of this assay was not very high (67%), but the results were validated in the second cohort of 171 RCCs. Moreover, a panel of two microRNA methylated promoters composed of miR-663a and miR-129-2 was shown to accurately detect urothelial carcinomas in urine (85.85% accuracy). Lu and co-workers have also demonstrated that methylation levels of miR-129-2 were increased in HCC compared to adjacent normal tissue. Moreover, miR-129-2 methylation was detected in plasma from HCC patients but not in plasma from liver cirrhosis patients or healthy individuals, which implies a potential utility of this biomarker as an early diagnostic marker for HCC. Furthermore, miR-17-5p methylation level allows the discrimination of patients with pancreatic tumors from healthy controls with extremely high specificity and sensitivity. Additionally, serum-circulating DNA was used to demonstrate the value of miR-34b/c methylation for the diagnosis of malignant pleural mesothelioma with high specificity and moderate sensitivity. Abnormal methylation of CpG islands of miR-34b/c promoter has been proposed as a potential biomarker for detecting CRC using fecal samples. Kalimutho and co-workers found that 75% of fecal specimens from CRC patients were positive for promoter methylation of miR-34b/c, whereas only 16% of patients with high-grade dysplasia and 13% of healthy individuals showed this alteration. Similarly, Wu et al. compared the positive rate of miR-34b/c methylation in fecal samples from CRC patients and healthy individuals and showed that the sensitivity and specificity for screening CRC were very high (95% and 100%, respectively). Detection of miRNA gene promoter hypermethylation in oral rinses has also been investigated. Promoter methylation of miR-137 is a relatively common event in head and neck squamous cell carcinoma (HNSCC), and its presence in oral rinses from these patients has been demonstrated. Moreover, HNSCC patients had nearly five times the odds of having miR-137 promoter methylation compared to the normal oral mucosa of control subjects. Thus, the miR-137 promoter methylation level in oral rinses distinguished HNSCC patients from healthy individuals with high specificity but low sensitivity. ## Epigenetic regulation of mirnas as prognostic biomarkers Discrimination of cancer patients with an aggressive biology could assist the clinicians in the management of these patients. As a result, a high number of prognostic markers have been identified, but unfortunately, trials that validate and confirm the utility of these markers are still lacking in most cases. Thus, the identification of novel tools that allow a more accurate prognostication of the patient is needed. In 2008, Lujambio et al. showed that methylation of miRNAs was correlated with the metastatic behavior of tumors in different organs. Lujambio et al. demonstrated that hypermethylation of the miR-34b/c, miR-148a, and miR-9-3 CpG islands was significantly associated with the presence of lymph node metastasis in melanoma, lung, and breast cancer. Subsequent studies confirmedand extended these results to other miRNAs and malignancies. Specifically, the correlation between lymph node metastasis and aberrant methylation was also observed in CRC for miR-9-1and miR-34aand in invasive breast ductal carcinomas for miR-124a. Given that lymph node metastasis is often associated with tumor recurrence and poor survival, the prognostic value of a plethora of miRNAs methylation has been evaluated. In this context, hypermethylation of miR-124 and miR-9 was shown to be associated with an increased risk of recurrence in clear cell RCC. In contrast, in breast cancer, miR-124 methylation levels were associated with different survival rates according to the age of the patients. The study concluded that miR-124 hypomethylation was a poor prognostic marker in young breast cancer patients (≤35 years old) as opposed to the longer survival rates found in older patients (>50 years old). In tumors of the respiratory tract, the prognostic value of miR-34b/c methylation is one of the most frequently investigated. In 2011, Wang et al. showed that aberrant miR-34b/c DNA methylation was an independent prognostic marker of stage I NSCLC. In this study, the association between altered DNA methylation of miR-34b/c and shorter recurrence-free and overall survival was demonstrated in a large series of 161 patients. This proved to be very useful for selecting a subset of stage I tumors with poor outcomes, which could benefit from additional therapy after resection. Subsequently, these results were confirmed in 140 lung adenocarcinoma patients. They evaluated the prognostic value of miR-34b/c methylation in an exploratory set of 58 LAC lung adenocarcinomas and validated their results in a confirmatory cohort of 82 patients. Moreover, miR-34b/c methylation was also a prognostic marker for stage I lung adenocarcinoma patients. Besides, Kim et al. confirmed the prognostic value of miR-34b/c in NSCLC and demonstrated a combined effect of miR-34b/c and miR-124-3 methylation patterns for the prognosis of NSCLC. Overall survival decreased as the number of methylated miRNAs increased; i.e., patients with two methylations exhibited significantly poorer overall survival than patients with none or one methylation. In line with these results, a profile composed of the methylation of five genes (miR-152, miR-9-3, miR-124-1, miR-124-2, and miR-124-3) was analyzed in NSCLC. Longer progression-free survival was proved in patients with none or one methylation compared to two or more. The association between survival and hypermethylation of other miRNAs, such as miR-127 or miR-145, has also been shown in lung cancer patients. In addition, the hypermethylation of miR-137 was associated with a shorter overall survival in HNSCC. In CRC, hypermethylation of miR-34a promoter CpG islands was also strongly associated with metastasis to the liver. Moreover, the expression of miR-148a was inversely correlated with its promoter methylation status. Both markers were pointed out as independent predictors of survival in adjuvant-treated stage IV CRC patients. The prognostic value of the methylation status of different miRNAs has also been investigated in hematological malignancies. Specifically, epigenetic silencing of miR-124a has been found to correlate with a higher recurrence rate and mortality rate in Acute Lymphoblastic Leukaemia (ALL), being an independent predictor for disease-free survival and overall survival. Moreover, the same group analyzed the hypermethylation profile of 11 CpG islands associated with several miRNAs (miR-124a1, miR-124a2, miR-124a3, miR-34b/c, miR-9-1, miR-9-3, miR-10b, miR-203, miR-196b, miR-9-2, and miR-132/212) in the same cohort of ALL patients and they found statistically significant differences in outcome between non-methylated and methylated ALL patients (methylation in at least one miRNA). Other groups have also examined the prognostic role of various miRNAs methylation in different lymphoid malignancies and showed that miR-129-2 and miR-340 methylation adversely impacted the survival factors in chronic lymphocytic leukemia multiple myeloma, respectively. Moreover, aberrant miR-137 methylation was shown to be associated with shorter progression-free survival in myeloma. ## Epigenetic regulation of mirnas as a therapeutic strategy in cancer Modulation of miRNAs expression has emerged as a promising strategy for cancer management. With the aim of restoring the expression levels of oncosuppressor miRs in cancer cells, different epigenetic drugs have been tested on different tumor models, and their tumor-suppressive effects have been evaluated. Some of these drugs, such as 5-Aza derivatives (Azacitidine and 5-Aza-dC) or HDAC inhibitors (Vorinostat, Panobinostat, Belinostat, and Romidepsin), have already been approved by the U.S. Food and Drug Administration for the treatment of different hematologic malignancies. Nowadays, an extension of their usage to solid tumors is being pursued by different pharmaceutical companies. The discovery of novel epigenetic drugs is also receiving great attention. A wide range of epigenetic-based drugs are being tested in preclinical and clinical trials (for a review, see. These drugs trigger a global effect in coding and non-coding genes, and thus their impact is difficult to be attributed to specific genes. Few reports have linked specific responses to epigenetic drugs and miRNAs expression in treated patients. Recently, Berg et al. found that the hypomethylating agents azacitidine and decitabine significantly upregulated the expression of miR-125a associated with anti-leukemic effects. These data were validated in chronic myelomonocytic leukemia patients, where higher levels of miR-125a were observed after treatment with the hypomethylating agents. Importantly, the increase was particularly pronounced in the responders to these drugs. More insight into the link between the mechanism (mode of action) of the epigenetic drugs and the miRNAs in cancer will provide new opportunities in the development of new strategies for cancer therapy. # Conclusions This review highlighted the functional implications of the epigenetic alterations and miRNA dysregulation in cancer diseases. Given the implications of miRNA in cancerrelated pathways and their described oncogenic or tumor-suppressive roles, their dysregulation seems crucial to fully understand neoplasia. The existing bidirectional regulation between DNA methylation, histone modifications, and miRNAs places epigenetics as one of the central pillars of carcinogenesis. Deciphering the epigenetic regulation of mRNAs in cancer diseases provides more insight into tumor initiation and progression and gives rise to a wide range of potential clinical applications. As we have overviewed, this is especially reflected by the large number of miRNA genes with aberrant methylation that have been proposed as putative biomarkers for diagnosis or prognosis in cancer diseases. Funding: This study has been funded by Instituto de Salud Carlos III through the project "PI19/00572" (Co-funded by European Regional Development Fund, "A way to make Europe"). and INNVA1/2020/71 de la Línea 1 de Valorización y Transferencia a las empresas, de la Agencia Valenciana de la Innovación. ## Conflicts of interest: The authors declare no conflict of interest.

Threat of Staphylococcus aureus Pneumonia in Severe COVID-19 Patients Coronavirus disease 2019 has been spreading worldwide with unprecedented rapidity. Staphylococcus aureus is reported to frequently cause bacterial complications in patients with COVID-19. We herein present two additional cases of S. aureus pneumonia involving such patients. The first case was an obese 48-year-old man without any particular underlying diseases. The second case was another patient, a 72-year-old man, with hypertension, dyslipidemia, and steatohepatitis. Both patients developed methicillinsusceptible S. aureus pneumonia in the clinical course of COVID-19, to which antibiotic therapy with cefazolin was effectively administered. Through these cases, we emphasize that S. aureus secondary infections should be well cared with a high degree of caution in a case of critically ill COVID-19 patients. # Introduction Coronavirus disease 2019 (COVID-19), caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), has been affecting the whole world. Recent literature investigated the incidence of secondary bacterial infections in COVID-19 patients with mechanical ventilation due to acute respiratory distress syndrome [bib_ref] Characterization of secondary bacterial infections and antibiotic use in mechanically ventilated patients..., Risa [/bib_ref]. Among 126 cases, 61% developed bacterial infections such as ventilator-associated pneumonia (VAP), blood stream infections, and urinary tract infections in its order, in which Staphylococcus aureus was the most common pathogen. Especially, methicillin-sensitive S. aureus (MSSA) accounted for 30% of the causative organisms of the respiratory infections. Another literature also suggested S. aureus to be the most common etiology of co-infection or superinfection secondary to COVID-19 [bib_ref] Incidence of co-infections and superinfections in hospitalized patients with COVID-19: a retrospective..., Garcia-Vidal [/bib_ref]. We herein present two additional cases of MSSA pneumonia involving patients with severe COVID-19. ## Case presentation The first case was an obese 48-year-old man (body mass index, 30.6 kg/m 2 ) without any other underlying diseases. The patient underwent the SARS-CoV-2 polymerase chain reaction (PCR) test and was diagnosed with COVID-19 in other institution before hospitalization in our hospital. On admission, his body temperature was 37.2°C and oxygen saturation level was 92% on 4 L/min oxygen supplementation. Laboratory tests revealed elevations of serum levels of C-reactive protein (CRP, 18.7 mg/dL), lactate dehydrogenase (608 U/L), and procalcitonin (0.44 ng/mL). Chest computed tomography showed bilateral multiple ground-glass opacities, which were compatible with COVID-19 pneumonia. The patient received 250 mg per day of methylprednisolone for three days in combination with 100 mg per day of remdesivir for five days. However, his respiratory condition deteriorated and the patient was intubated on the third day of hospitalization. Thereafter, his general status including the inflammatory markers and the findings on chest X-ray ameliorated . On the seventh day of hospitalization, however, his sputum appeared purulent and the infiltrate shadows got worse , accompanying elevated serum levels of CRP (16.3 mg/dL) and procalcitonin (0.21 ng/mL). Blood culture was negative. Sputum culture detected a large number of MSSA along with leukocyte migration, leading to the diagnosis of MSSA-induced pneumonia. We treated the patient with 6 g per day of cefazolin for 10 days. His respiratory condition subsequently ameliorated , and the patient was discharged. The second case was a 72-year-old man with hypertension, dyslipidemia, and steatohepatitis. The patient also had a confirmed positive result of SARS-CoV-2 PCR test in other hospital and was transferred to our hospital. Due to deteriorated respiratory condition , the patient was intubated and administered with 125 mg per day of methylprednisolone for three days and a single dose of 8 mg per kg of tocilizumab. Subsequently, his respiratory status and serum inflammatory markers ameliorated. However, on the sixth day, the infiltrate shadows worsened and a purulent sputum emerged, which was positive for MSSA and active leukocytes. Blood culture resulted negative. We made a diagnosis of MSSA-induced pneumonia, to which antibiotic therapy with 6 g per day of cefazolin for 10 days was effectively administered. The patient recovered soon , and was discharged. ## Figure 1: imaging of methicillin-susceptible staphylococcus aureus pneumonia in two covid-19 patients. Chest X-rays of the first case on the fourth day of hospitalization (A), on the seventh day (B, before antibiotic therapy), and on the 16th day (C, after antibiotic therapy). Chest X-rays of the second case on the third day of hospitalization (D), on the sixth day (E, before antibiotic therapy), and on the 16th day (F, after antibiotic therapy). The infiltration shadows dramatically improved after antimicrobial treatment in both cases. # Discussion COVID-19 has been spreading worldwide with unprecedented rapidity. The disease potentially induces various complications, of which secondary bacterial infections should be regarded as the most harmful events [bib_ref] Characterization of secondary bacterial infections and antibiotic use in mechanically ventilated patients..., Risa [/bib_ref]. Among various bacterial etiologies, recent literature indicated S. aureus-associated infections to be of great importance. For instance, an observational cohort study investigated the incidence of coinfections and superinfections in hospitalized patients with COVID-19 [bib_ref] Incidence of co-infections and superinfections in hospitalized patients with COVID-19: a retrospective..., Garcia-Vidal [/bib_ref]. Overall, 74 (7.5%) of 989 patients had bacterial infection secondary to COVID-19, of which 44 cases (59.5%) were hospital-acquired superinfections complicating clinical course of patients with COVID-19. VAP accounted for 25% of the nosocomial infections (11/44 cases), in which S. aureus was reportedly the most common causative pathogen (36.4%). Another retrospective, single-facility study reported that 66 (5.52%) of 1,251 samples had bacterial infection secondary to COVID-19, of which S. aureus was reportedly the common causative pathogen (24.3%) [bib_ref] Bacterial coinfections in COVID-19 pneumonia in a tertiary care hospital: surfing the..., Ruiz-Bastián [/bib_ref]. Accordingly, bacterial complications, especially S. aureus-associated infections, do frequently occur in COVID-19 patients, although Gram-negative organisms, including Pseudomonas aeruginosa and Acinetobacter baumanii, are the representatives of VAP pathogens. Notably, S. aureus potentially causes various respiratory complications such as necrotizing pneumonia or pneumothorax [bib_ref] Post COVID-19 MSSA pneumonia, Chaudhry [/bib_ref] , and, thus, clinicians should be aware of the clinical importance of prevention and management of such cases. When treating patients with S. aureus infections, an empirical coverage against anti-methicillin-resistant S. aureus (MRSA) is always an issue to be addressed. Clinicians should be aware of antibiogram, or MRSA isolation rate, at each medical situation, and determine the need of starting broad-spectrum antibiotics. In our hospital, the MRSA isolation rate for nosocomial cases changes approximately at 40-50% (data not shown). In the present cases, we carefully observed the patient's clinical course without initiating anti-MRSA agents. Antimicrobial stewardship should have a priority even for the patients with COVID-19. To differentiate deterioration of COVID-19 pneumonia and involvement of the secondary bacterial infections, procalcitonin would be of use [bib_ref] Procalcitonin for patient stratification and identification of bacterial coinfection in COVID-19, Han [/bib_ref]. Results of a retrospective, single-facility cohort study indicated that, based on a procalcitonin-guided algorithm, antimicrobial usages could be safely reduced in non-critically ill COVID-19 patients [bib_ref] Evaluation of procalcitonin-guided antimicrobial stewardship in patients admitted to hospital with COVID-19..., Calderon [/bib_ref]. Emergence of antimicrobial resistance is another challenge that the global society is facing, which should be cared even amid the COVID-19 pandemic. Tanaka et al. Cureus 14(2): e22486. DOI 10.7759/cureus.22486 2 of 3 Tanaka et al. Cureus 14(2): e22486. DOI 10.7759/cureus.22486 3 of 3 ConclusionsWe described two severe cases of S. aureus pneumonia involving ventilated patients suffering from COVID-19. S. aureus is usually not such a common pathogen in VAP. Compiling related literature with our cases, we would like to highlight that S. aureus-targeted treatment should be empirically initiated when COVID-19 patients manifest findings suggesting bacterial pneumonia.Additional Information DisclosuresHuman subjects: Consent was obtained or waived by all participants in this study. Conflicts of interest: In compliance with the ICMJE uniform disclosure form, all authors declare the following: Payment/services info: All authors have declared that no financial support was received from any organization for the submitted work. Financial relationships: All authors have declared that they have no financial relationships at present or within the previous three years with any organizations that might have an interest in the submitted work. Other relationships: All authors have declared that there are no other relationships or activities that could appear to have influenced the submitted work.