text
stringlengths 66
377k
|
|---|
Membrane-cytoplasm translocation of annexin A4 is involved in the metastasis of colorectal carcinoma
Annexin A4 (ANXA4) is a Ca 2+ -and phospholipid-binding protein that belongs to the annexin family, which is involved in the development of multiple tumour types via NF-κB signalling. In this study, we verified the high expression and membrane-cytoplasm translocation of ANXA4 in colorectal carcinoma (CRC). Calcium/calmodulindependent protein kinase II gamma (CAMK2γ) was found to be important for high ANXA4 expression in CRC, whereas carbonic anhydrase (CA1) promoted ANXA4 aggregation in the cell membrane. An increased Ca 2+ concentration attenuated the small ubiquitin-like modifier (SUMO) modification of cytoplasmic ANXA4 and ANXA4 stabilization, and relatively high expression of ANXA4 promoted CRC tumorigenesis and epithelial-mesenchymal transition (EMT).
# Introduction
Colorectal cancer (CRC) is the second most common cancer in women and the third most common cancer in men, causing more than 900,000 deaths worldwide annually. In China, colon cancer represents the third most common cause of death (7.9%) among the deaths related to malignant neoplasms. Despite the availability of comprehensive diagnostic and treatment methods, the average 5-year survival rate of colon cancer patients does not exceed 60%, and metastasis and recrudescence are still the main causes of death in CRC patients.
Annexin A4 (ANXA4) is a member of the annexin family and a Ca 2 +and phospholipid-binding protein. The ANXA4 gene is located on human chromosome 2q13 and encodes a 36-kDa protein that aggregates on the cell membrane and contains four annexin repeats with one Ca 2 + -binding site and five α-helices in each repeat, which can bind phospholipids in a Ca 2 + -dependent manner. Studies have indicated that human ANXA4 is predominantly expressed in the secretory epithelia in the lungs, stomach, intestine, and kidneys. ANXA4 accelerates membrane repair, upregulates exocytosis, decreases mobilityand is involved in skin wound haemostasis, human heart failureand tumours.
It is conceivable that ANXA4 is highly expressed in multiple tumour types and is an indicator for tumour development, invasion, chemo-resistance, and poor AGING outcomes in cancer patients. ANXA4 biology may be a potential target for therapeutic intervention, such as the Annexin A4-NF-κB feedback circuit that activates malignant cell behaviour and tumour growth in gallbladder cancerand Annexin A4 fucosylation that enhances the interaction of ANXA4 with NF-kB p50 and promotes tumour progression in ovarian clear cell carcinoma; toosendanin mediates cisplatin sensitization through targeting Annexin A4/ATP7A in non-small cell lung cancer cells.
Calcium/calmodulin-depende nt protein kinase II (CaMKII, also known as CAMK2γ) is a multifunctional serine/threonine protein kinase that transmits calcium signals in various cellular processes. CAMK2γ is important in both haematopoietic malignancies and solid tumours.
SUMOylation is a posttranslational modification that leads to diverse biological consequences. The addition of small ubiquitin-like modifier (SUMO) affects target proteins by altering their subcellular localization, enzymatic activity and/or protein-protein/protein-DNA interactions. Three SUMO proteins are found in mammalian cells, with SUMO-2 and SUMO-3 sharing 97% sequence identity with one another but only approximately 50% identity with SUMO-1. Annexin A1 is regulated by domain cross-talk through posttranslational phosphorylation and SUMOylation.
Carbonic anhydrases (CAs) are a family of metalloenzymes that are found in almost every type of tissue. Carbonic anhydrases have 14 different isoforms, and the distribution patterns of the isoenzymes differ. Silencing carbonic anhydrase I enhances the malignant potential of exosomes secreted by prostate tumour cells.
In this study, we verified the high expression and membrane-cytoplasm translocation of ANXA4 in colorectal carcinoma. Next, we further showed that CAMK2ɤ was indispensable for high ANXA4 expression in colorectal carcinoma and that the location of ANXA4 location in the cell membrane was Ca 2 + dependent; CA1 could also promote ANXA4 aggregation in the cell membrane. Moreover, we also found that increasing the Ca 2+ level decreased SUMOylation of cytoplasmic ANXA4 and that ANXA4 promoted CRC tumorigenesis and metastasis.
# Results
## Cell membrane-cytoplasm translocation of anxa4 in crc
Our previous studies have revealed that ANXA4 is highly expressed in CRC tissues and is a candidate biomarker in serum extracellular vesicles for detecting early-stage CRC. Using Gene Expression Omnibus datasets (GEO) [GSE20916 (Poland), Skrzypczak Colorectal; GSE 8671 (Switzerland), and Sabates-Bellver Colon], we found that ANXA4 expression was significantly higher in colorectal cancer tissue than in the normal colonic mucosa . As shown in , ANXA4 expression was strongly positive in CRC, and ANXA4 was located in both the membrane and cytoplasm in CRC tissues compared with the corresponding normal tissues, which were moderately stained and exhibited ANXA4 localized mainly in the plasma membrane (n=138). Furthermore, CRC patient tumour tissue samples, tumour-adjacent tissue samples and normal colon tissue samples (n=20) were used to examine the protein level and subcellular distribution of ANXA4 by western blotting of membrane-cytoplasm fractionated lysates. ANXA4 expression was higher in the CRC tissue samples than in the tumour-adjacent tissue samples and was detected in both the plasma membrane and cytosolic fractions in the CRC tissue samples, whereas ANXA4 expression was minimal in the normal tissue samples . Comparison of endogenous ANXA4 levels between normal epithelial cells (NCM460) derived from the human colonic mucosa and three CRC cell lines (SW480, SW620, and HT-29) revealed that ANXA4 expression was significantly higher in HT-29 cells, which are derived from a highly metastatic CRC cell line, than in SW480 and SW620 cells and that ANXA4 was mainly distributed in the cytoplasm of HT-29 cells . Chi-square test analysis showed a positive correlation between high ANXA4 expression and lymph node metastasis in CRC patients (n=138) .
## Camk signalling affects the expression and location of anxa4 in crc
ANXA4 contains conserved ANX repeat domains and binds phospholipids in a Ca2 + -dependent manner. High intracellular Ca2 + levels promote interactions between ANXA4 and the NF-κB p50 subunit, thus inhibiting the transcriptional activity of NF-κB. We next wondered whether the calcium/CAMK signalling pathway has an effect on the regulation of ANXA4 in HT-29 cells. Western blotting was performed to detect ANXA4 expression after increasing the intracellular Ca2 + level by exogenously adding different concentrations of ionomycin (0 µM, 1 μM and 2 μM) for different times (2 h, 6 h, 14 h and 24 h) , 2E). We found that 1~2 μM ionomycin decreased the expression of ANXA4 after treatment for 14 h (p<0.001) and 24 h (p<0.05).
Subsequently, we tested whether Ca 2 + affects ANXA4 expression via CAMK. CaMKII is encoded by four AGING AGING different genes, α, β, γ and δ, and CAMK2γ is the major isoform in CRC cells. We therefore evaluated whether Ca2+ decreases ANXA4 expression via CAMK2γ. Our results indicated that an increase in the intracellular Ca2 + level promoted CAMK2γ expression in HT-29 cells. As shown in , 2F, CAMK2γ expression increased nearly 2-fold after 1~2 μM ionomycin treatment for 24 h. We also transfected an shCAMK2γ plasmid into HT-29 cells and then exogenously added ionomycin (0 µM, 1 μM, or 2 μM) and found that interfering with CAMK2γ reversed the negative regulatory effect of increased Ca2 + levels induced by ionomycin on ANXA4 in HT-29 cells . Results acquired after membranecytoplasm fractionation also clearly showed the translocation of ANXA4 into the membrane fraction after ionomycin treatment , 2H). The above data suggest that CAMK2γ is important for high ANXA4 expression in HT-29 cells and that the localization of ANXA4 in the cell membrane is Ca2 + dependent.
## Ca1 promoted anxa4 aggregation in the cell membrane
Our previous studies have revealed that CA1 is expressed at low levels and ANXA1 is highly expressed in clinical TNM stages I-IV by MALDI-TOF/TOF MS and immunohistochemical analyses of 103 CRC samples from Chinese patients. In this study, we identified 187 commonly changed genes in CRC samples from the GSE 25070 (USA), GSE 41258 (Israel), GSE 44076 (Spain) and GSE 44861 (USA) datasets, including CA1, which was also found to have low expression in a TCGA CRC dataset. There was a negative correlation between CA1 and ANXA4 expression in GSE 20916 and GSE 8671, and this negative correlation between CA1 (low) and ANXA4 (high) expression was further verified in 103 Chinese CRC samples by IHC. By mining TCGA data, we selected 50 upregulated and 50 downregulated genes correlated with ANXA4 or CA1 expression (p<0.05) to perform GO analysisand found that these genes are involved in the plasma membrane or integrated in the plasma membrane. This suggested that CA1 might be associated with the membrane-cytoplasm translocation of ANXA4. We used recombinant human CA1 to treat HT-29 cells and found that 1 µl CA1 promoted ANXA4 aggregation in the cell membrane. These data suggest that the relatively low CA1 expression may be one of the reasons why ANXA4 is highly expressed and located in the cytoplasm in CRC tissues, but the mechanism is not particularly clear and needs further exploration.
## Sumoylation stabilized anxa4 expression in the cytoplasm of crc cells
Some studies have confirmed that the molecular expression and subcellular distribution of the Annexin A family can be regulated by SUMOylation. However, the SUMOylation of ANXA4 and its effects on deallocation from the cell membrane to the plasma are still unknown, so we tested whether ANXA4 is a posttranslational modification target of SUMO.
We predicted potential SUMOylation sites in ANXA4 with bioinformatic software (SUMO plot TM Analysis and SUMOsp 2.0), and the three predicted conserved binding sites are found (data not shown). Then, we constructed the plasmids HA-SUMO1, HA-SUMO2, HA-SUMO3 and Flag-ANXA4; cotransfected HA-SUMO1, HA-SUMO2 or HA-SUMO3 and Flag-ANXA4 into HEK293T cells; and performed immunoprecipitation with anti-HA gel. The results indicated that exogenous ANXA4 was mainly modified by SUMO1, SUMO2 and SUMO3; however, the transfection of an HA-SUMO plasmid did not upregulate ANXA4 expressionor influence the membrane-cytoplasm translocation of ANXA4 in CRC cells (data not shown). These data indicated that SUMOylation was not associated with the membrane-cytoplasm translocation of ANXA4. However, SUMOylation of ANXA4 promoted ANXA4 stability, although the Ca2 + concentration increased, whereas Ca2 + decreased the stabilization of ANXA4. This result is consistent with Ca2 + downregulating ANXA4 expression.
## High anxa4 expression contributes to crc tumorigenesis and metastasis
We further explored the impacts of ANXA4 on the proliferation and metastatic potential of HT-29 cells. Overexpression of ANXA4 resulted in an increase in viability, as measured by using a CCK8 assay. A wound-healing assay and Matrigel invasion assay showed that ANXA4 promoted the cell migration and invasive potential of HT-29 cells (data not shown). In addition, HT-29 cells transfected with ANXA4overexpression or empty vectors were injected subcutaneously into nude mice, and tumour volume was evaluated at 7, 12, 18, and 25 days. The in vivo xenograft model indicated that the ANXA4 overexpression groups had significantly higher tumourigenicity with obvious haemorrhagic necrosis on the tumour surface than the control group (n=6, n1=3/6 VS n2=0/6). IHC identified the protein expression levels of ANXA4 and the EMT signalling pathway factors E-cadherin, Vimentin and Snail in tumours. Higher ANXA4 expression was accompanied AGING AGING by lower E-cadherin expression and higher Vimentin and Snail expression.
Although we did not find tumour metastasis in our subcutaneous xenograft tumour model because of time limitations (25 days), we assessed the relationships between clinicopathological characteristics and ANXA4 expression levels by immunohistochemical staining of 138 CRC samples and corresponding tumour-adjacent tissue samples. Higher expression of ANXA4 was correlated with tumour differentiation (p<0.001) and metastasis (p<0.05) , but it was not correlated with sex, age, tumour location or tumour size.
## Anxa4 caused the downregulation of the stromal response to promote emt in crc
As shown in, GO analysis was focused on the molecular function and biological pathway categories for ANXA4 expression-related genes, and we found that AGING ANXA4 was mainly involved in receptor signalling complex scaffold activity, extracellular matrix structural constituent, cell adherin molecule activity, integrin family cell surface interaction and epithelial-mesenchymal transition. Based on the ESTIMATE algorithm, stromal scores ranged from -1307.73 to 1112.11, and immune scores were distributed between -1097.42 and 1759.18. We found AGING that ANXA4 did not change immune scores, but ANXA4 decreased stromal scores, which was consistent with the GO analysis results. The above analysis implies that ANXA4 may be involved in CRC metastasis.
We further investigated the effects of ANXA4 on EMTrelated genes (E-cadherin, Vimentin, β-catenin, and Snail) in CRC cells by transfecting ANXA4-carrying plasmids into SW620 or HT-29 cells and evaluating the cells by western blotting. We found that ANXA4 increased the expression levels of vimentin and Snail 1 but decreased the expression of E-cadherin and βcatenin. The same results were obtained in the context of ANXA4 overexpression in the subcutaneously transplanted tumour nude mouse model.
# Discussion
Annexin A4 (ANXA4) can reversibly bind to membrane phospholipids in a calcium-dependent manner and is involved in regulating tumour cell proliferation, apoptosis, adhesion, invasion, metastasis, and chemotherapy resistance, as well as other biological processes. In the current study, western blot and immunohistochemistry analyses of ANXA4 expression and localization indicated that ANXA4 was highly expressed and located in the cell membrane and cytoplasm in CRC tissue compared to normal colon tissue and that the quantity of ANXA4 in the cytoplasm was obviously greater than that in the nucleus.
We next explored the mechanism underlying the cell membrane to cytoplasm shift in ANXA4 localization in CRC. Gaudio E et al.reported that the Fhit protein can interact with ANXA4 and that Fhit overexpression prevents ANXA4 translocation from the cytosol to the plasma membrane in A549 lung cancer cells treated with paclitaxel. Our results indicated that under stimulation with the Ca2 + activator ionomycin, increasing Ca2 + concentrations decreased the expression of ANXA4 in HT-29 cells but promoted ANXA4 assembly in the cell membrane. CAMK2γ in intestinal epithelial cells modulates colitis-associated colorectal carcinogenesis by enhancing STAT3 activation. CAMK2γ antagonizes growth factoror insulin-induced mTORC1 activation by inhibiting IRS1/AKT signalling. Our results indicated that Ca2 + stimulation alone decreased the stabilization of ANXA4to limit ANXA4 expression but CAMK2γ was indispensable for high ANXA4 expression in HT-29 cells and ANXA4 localization in the cell membrane in a Ca2 + -dependent manner. Moreover, we found that relatively low CA1 expression was accompanied by relatively high ANXA4 expression and promoted ANXA4 localization in the cytoplasm in CRC.
Small ubiquitin-like modifier, including SUMO 1-4 in humans, is a reversible protein modifier. SUMO modification (SUMOylation) is a newly discovered post-translational regulatory mechanism that plays critical roles in a variety of cellular processes by regulating the stability or translocation of target proteins. In particular, SUMOylation can impact the subcellular distributions of proteins; for example, SUMOylated KHSRP is more likely to translocate from the nucleus to the cytoplasm than is unmodified KHSRP in glioblastoma multiform cells. In colorectal carcinoma, Ca2 + did not affect the interaction between SUMO1 and ANXA4, and SUMOylation did not affect the membrane-cytoplasm translocation of ANXA4. After ANXA4 was translocated into the cytoplasm, it was further modified by SUMOylation to make it more stable, supporting its involvement in CRC tumorigenesis and progression.
# Materials and methods
## Patients and tissue samples
Samples from one hundred forty-three patients diagnosed with CRC were obtained from the Hunan Provincial Tumor Hospital and the affiliated tumour hospital of Xiangya Medical School. For all cases, pathological and clinical follow-up data were collected after obtaining ethical approval from the China Ethical Review Committee. CRC was diagnosed based on clinical symptoms, proctoscopy and biopsy.
## Cell culture and transfection
SW480, SW620, HT-29 and HEK293 cells were obtained from the Cell Center of Peking Union Medical College (Beijing, China) and were maintained in RPMI-1640 medium (Gibco) containing 10% foetal bovine serum (Sigma) and antibiotics (100 μg/ml penicillin and streptomycin) in 5% CO2 in a 37° C incubator.
The SW620 cell line is derived from a lymph node metastasis of the primary tumour from which the SW480 cell line is derived. HT 29 cells are derived from another primary colorectal adenocarcinoma. AGING HA-SUMO1, HA-SUMO2 and HA-SUMO3 were purchased from Addgene Company. pcDNA3.1-shRNA-CAMK2G, pcDNA3.1-Flag-ANXA4 and pcDNA3.1-empty control were purchased from Integrated Biotech Solutions Company. All the plasmids were confirmed by sequencing. Cell transfection was performed using Lipofectamine 3000 (Invitrogen) according to the manufacturer's instructions. Recombinant human carbonic anhydrase-1 (CA1) was purchased from ProSpec Company (China, ENZ-462).
## Intracellular ca 2+ level increase
Ionomycin was added to HT-29 cells at different concentrations (0 µM, 1 μM and 2 μM) for different times (2 h, 6 h, 14 h and 24 h) to increase the intracellular Ca 2 + level. HT-29 cell proteins were extracted for immunoblotting.
## Western blotting
Proteins were extracted from cell lines or tissue and lysed in GLB + supplemented with a protease inhibitor cocktail (Bimake) for 20 min on ice, followed by centrifugation (12,000 rpm, 4° C, 15 min). Membrane and cytoplasmic proteins were prepared with the Mem-PER™ Plus Membrane Protein Extraction Kit (Thermo, 89842). Then, 25-60 μg of protein extracts were loaded onto 8-12% electrophoresis gels and transferred to membranes. The membranes were blocked with 5% non-fat milk and subsequently incubated with primary antibodies. The antibodies used were anti-ANXA4 (Proteintech), anti-CAMK2G (Proteintech), anti-HA (Proteintech), anti-FLAG (Sigma), anti-E-cad (Cell Signaling Technology), anti-N-cad (Cell Signaling Technology), anti-Vimentin (Cell Signaling Technology), anti-Snail 1 (Proteintech), anti-β-catenin (Proteintech), anti-Tubulin (Proteintech), anti-GAPDH (Proteintech) and anti-Na+-K+-ATPase (Santa Cruz).
## Analysis of sumo-modified anxa4
In total, 1 × 10 7 HEK-293 cells and HT-29 cells were plated in 10-cm dishes and transfected with 10 μg of f1ag-ANXA4 and HA-SUMO1, HA-SUMO2, or HA-SUMO3. After 48 h of co-transfection, the cells were lysed in RIPA buffer. Next, we conducted IP experiments. The first antibody we used was anti-HA.
## Immunoprecipitation
HA-SUMO1, HA-SUMO2 or HA-SUMO3 was cotransfected with Flag-ANXA4 into HEK-293 cells and immunoprecipitated with anti-HA gel. ANXA4 immunoprecipitation from replicating chromatin under denaturing conditions was performed by resuspending the chromatin pellet in chromatin preparation buffer (CPB) (20 mM HEPES, pH 7.6; 100 mM KCl; 2% sucrose; and 5 mM MgCl2) supplemented with 1% SDS, followed by 10-fold dilution in CPB supplemented with 1% Triton X-100. The samples were then incubated with 500 ng of anti-ANXA4 antibodies with agitation at 4° C for 45 min, followed by the addition of protein A Dynabeads and incubation at 4° C for 1 h.
## Immunohistochemistry
CRC and normal colorectal tissue samples were sectioned into 7-mm-thick slides and incubated with appropriate antibodies at 4° C in a humidified container. A non-biotin horseradish peroxidase detection system and DAB substrate (Ultra Sensitive TM SP IHC Kit, Biotechnologies) were used for staining. A semi-quantitative scoring system was used to evaluate the ANXA4 protein levels, which considered both the intensity and extent. The ANXA4 protein levels were scored as follows: 0 (negative), 1 (weak), 2 (moderate), and 3 (strong). The scores were evaluated independently by two experienced pathologists blinded to the patients' clinical characteristics and outcomes, and the average ANXA4 IHC score was recorded as low (score<1) or high (score>2).
## Wound healing and cell invasion assays
Sixteen hours after transfection of an ANXA4-carrying plasmid, cells reached 70% confluence in 6-well plates. The cell layers were scratched using a 10-μl tip to form wound gaps, washed with PBS twice and cultured in serum-free medium. The wound gaps were imaged at different time points and analysed by measuring the distance of migrating cells in five different areas for each wound.
Transwell inserts (8-μm pore, Corning) were precoated with a 20-μL mixture of Matrigel (BD Biosciences) and 1640 medium (Sangon Biotech) at a ratio of 1:1 for 1 hour at 37° C. HT-29 cells were transfected with ANXA4/empty control vector for 36 h, and then 200 μl of cell suspension (105 cells) was placed into the upper chambers with 2% FBS 1640 medium. After 28 h of incubation, the cells inside the upper chamber were removed with cotton swabs. The invaded cells on the lower membrane surface were fixed and then stained with 10% crystal violet. Four randomly selected fields in each well were counted.
## Aging
## Xenograft tumour model
All animal experiments were approved by the Animal Care and Use Committee of Central South University. For subcutaneous tumour formation, cells (2×10 6 ) in 100 μl of serum-free medium were injected subcutaneously into the left flank of nude mice. A total of 12 mice were used for the intracranial xenograft tumour model, which included the ANXA4 overexpression groups and control groups. Six mice in each group were sacrificed after 25 days for IHC. Tumour volumes were determined according to the following formula: A×B2/2, where A is the largest diameter and B is the diameter perpendicular to A.
# Statistical analysis
All experiments were analysed with GraphPad Prism 5 (La Jolla, CA, USA). Inter-group differences were tested using Student's t-test or one-way ANOVA. Correlation analysis was performed with Spearman's rank test and the chi-square test. Data are expressed as the mean ± SEM. of at least three independent experiments. A probability value (p) <0.05 was considered statistically significant.
# Author contributions
YP and YL designed the study. ZXP, ALZ, CHL, ZYZ, YL conducted the experiments. ZYZ and ZXP wrote the article. YP revised of the manuscript. All authors read and approved the final manuscript.
## Conflicts of interest
|
Long-term effects of youth unemployment on mental health: does an economic crisis make a difference?
Background Ill health is a risk factor and a consequence of unemployment, which might vary depending on the national rate of unemployment. We investigated the long-term effect of youth unemployment on mental health and explored the possible interaction during periods of high (economic crisis) and low (non-crisis) unemployment rates. Methods A register-linked population-based cohort study was conducted including individuals aged 17-24 years. The crisis cohort (n=6410)
# Introduction
In the last decades, youth unemployment has become an increasing problem for many countries, including Sweden.Prior to 1991, the Swedish youth unemployment was at a comparatively low stable level, around 5%.During the years of 1991-1994, Sweden suffered a deep economic crisis. As a result, youth unemployment increased from 3.4% in 1990 to 19% in 1993. Today, the Swedish youth unemployment rate is at a high stable rate, above 20%.Leaving school to find employment is a central and difficult transition for young people. An extensive body of literature suggests that youth unemployment is related to a decrease in physical and mental health and an increase in smoking and alcohol consumption. [bib_ref] Unemployment and mental health among young people: a longitudinal study, Hammer [/bib_ref] [bib_ref] Nervous and depressive symptoms in a longitudinal study of youth unemployment-selection or..., Hammarström [/bib_ref] [bib_ref] The effects of unemployment on psychiatric illness during young adulthood, Fergusson [/bib_ref] [bib_ref] Unemployment pre-dates symptoms of depression and anxiety resulting in medical consultation in..., Montgomery [/bib_ref] [bib_ref] Unemployment and psychosocial adjustment in young adults: causation or selection?, Fergusson [/bib_ref] [bib_ref] Alcohol consumption among unemployed youths: results from a prospective study, Janlert [/bib_ref] [bib_ref] Unemployment and psychological distress among young adults in the Nordic countries: a..., Reneflot [/bib_ref] [bib_ref] Unemployment and psychosocial outcomes to age 30: a fixed-effects regression analysis, Fergusson [/bib_ref] [bib_ref] Unemployment and mental health scarring during the life course, Strandh [/bib_ref] [bib_ref] Early unemployment can contribute to adult health problems: results from a longitudinal..., Hammarström [/bib_ref] Further, youth appears to be a sensitive time period in life, as recent studies have found that the effect of youth unemployment on mental health remains in adulthood, independent of later unemployment experiences. The relationship between unemployment and ill health is, however, more complex as the effect could differ depending on the national rate of unemployment.
There are several theoretical frameworks that could potentially explain the contextual influence on the association between unemployment and health. If the health effect of unemployment is stronger during periods of low unemployment then the association might be more confounded by health selection or unemployment might be more stigmatised. Alternatively, being unemployed during a period of high unemployment might buffer the negative consequences of unemployment as it is easier to attribute ones situation to external causes. [bib_ref] Psychological and physical well-being during unemployment: a meta-analytic study, Mckee-Ryan [/bib_ref] Then again, high unemployment creates uncertainty among the unemployed about their possibilities on the labour market which could strengthen the effects of unemployment. There is, however, no consistent evidence in the current literature concerning the direction and magnitude of the influence of national rate of unemployment on the association between unemployment and mental health. [bib_ref] Excess mortality of unemployed men and women during a period of rapidly..., Martikainen [/bib_ref] In addition, the national rate of unemployment might affect health differently in certain subgroups. The area of youth unemployment has received little attention with this regard. One previous study that attempted to examine the effects of health selection by comparing two time periods, did not find a difference in somatic and psychological symptoms among the long-term unemployed. [bib_ref] Health hazards of unemployment-only a boom phenomenon? A study of young men..., Novo [/bib_ref] Previous research on youth unemployment has to a great extent relied on self-reported measures to assess the health or health behaviours. [bib_ref] Unemployment and psychological distress among young adults in the Nordic countries: a..., Reneflot [/bib_ref] Thus, there is a need to include more severe health outcomes, such as hospital diagnoses to more comprehensively address the effects of youth unemployment on health later in life. [bib_ref] Unemployment and psychological distress among young adults in the Nordic countries: a..., Reneflot [/bib_ref] The aims of the current study are (1) to investigate the effect of youth unemployment on mental health during periods of high and low unemployment rates and (2) to explore whether there is any interaction in mental health between labour force status and level of unemployment.
# Methods
Our population-based study is based on a data set created through record linkage of nationwide registers kept by Statistics Sweden and the National Board of Health and Welfare by using the Swedish unique personal identification number. Registers included are the Labour Force Survey (LFS), the National Population and Housing Censuses, The Cause of Death Register and National Hospital Discharge Register. Sampling and study population
The study was based on two separate cohorts who participated in the LFS between the ages of 17-24. The LFS is a telephonesupported interview. Included individuals are interviewed every 3 months for a total period of 2 years, resulting in eight different interviews regarding their current labour force status. The age span was selected since the majority have completed compulsory schooling by the age of 17 and the official upper limit of the Swedish definition of youth unemployment is 24 years.Individuals that are registered in Sweden (aged are randomly selected to participate in the LFS. [bib_ref] Arbetskraftsundersökningarna 1985-1994: Befolkningen i åldern 16-64 år (Labour Force Survey 1985-1994), Sweden [/bib_ref] The first cohort comprised all individuals born between 1969 and 1974 who completed the LFS during the period 1991-1994 (n=7208), the 'crisis cohort'. The second cohort comprised all individuals born between 1961 and 1966 who completed the LFS during 1983-1986 (n=9076), the 'non-crisis cohort'. To determine youth unemployment, we used several measurement points of the individuals' current labour force status, consequently individuals that had missing data on more than three out of the eight interviews from the LFS were excluded ( please see supplementary figure S1).
## Exposure: employment status
Information on employment status was collected from the LFS. The LFS divides the population into two main categories: individuals who are active in the labour force (ie, employed or unemployed) and individuals who are economically inactive (ie, students, conscripts, pensioners or individuals with long-term illness). The employed group consists of people who perform at least 1 hour of work each week. The unemployed group consists of people who work for <1 hour/week, actively looking for work and able to start a new job within 2 weeks of the interview. The study population was categorised into six mutually exclusive groups based on the LFS definitions: unemployed <3 months, unemployed 3-6 months, unemployed >6 months, economically inactive, unstable labour force status and full-time employed (the reference group). Individuals reporting being unemployed during one interview were categorised as <3 months unemployment, reporting being unemployed in two consecutive interviews were categorised as 3-6 months of unemployment and for three or more consecutive interviews were categorised as >6 months of unemployment. Individuals defined as being economically inactive for at least five of the eight interviews were categorised into this group. Individuals with a combination of labour force status or working part-time (<35 hours/week) were categorised as having an unstable labour force status. Individuals defined as employed and working 35 hours/week or more for at least five out of the eight interviews were defined as being full-time employed.
Outcome measure: mental diagnosis Diagnosis of mental health, according to the Swedish version of the International Statistical Classification of Disease (ICD) versions 9 and 10 was collected from the National Hospital Discharge Register. The outcome of mental diagnosis was defined through four categories of discharge diagnosis: alcohol or drug use disorders (ICD-9: 291, 292, 303-305; ICD-10: F10-F19), affective disorders (mood disorders) (ICD-9: 296, 311; ICD-10: F30-F39), nervous or stress-related disorders (ICD-9: 300, 306, 308, 309; ICD-10: F40-F48) and self-harm (ICD-9: E950-E959, E980-E989; ICD-10: X60-X84, Y10-Y34). Only first-time admissions were of interest during follow-up. The individual was considered as having the outcome of interest, if they had any of the diagnoses as either a principal or secondary discharge diagnosis.
## Covariates
The analyses were controlled for the following individual factors: sex, age (continuous), country of birth and prior mental diagnosis (before 1991 in crisis cohort and 1983 in the noncrisis cohort). In addition to the mental diagnosis chosen as the outcome, we included diagnosis of organic psychosis, schizophrenia, other non-affective disorders, personality disorders, childhood mental disorders and mental retardation as covariates (ICD-8: 292, 293, 294, 294.4, 294.8-9, 295, 297, 298, 301, 302, 306, 308-315 and the corresponding ICD-9 codes).
The Multi-Generation Register (MGR), which contains information on all individuals registered in Sweden, was used to identify the individual's biological parents.The analyses were controlled for the following parental risk factors that were obtained at baseline: highest level of parent's education, socioeconomic index (SEI) and any of the parents registered inpatient care or cause of death due a mental diagnosis. The covariates were categorised as indicated in table 1A, B.
# Statistical analysis
The analyses were stratified by cohort. HRs with 95% CIs were obtained by Cox proportional hazard regression analysis after verification of the proportional hazards assumption using loglog plots and plots of Schoenfeld residuals. Person-time was calculated from the end of the exposure window (crisis cohort: 1 January 1995; non-crisis cohort: 1 January 1986) until first date of a mental diagnosis, emigration, death or until end of follow-up (crisis cohort: 31 December 2012; non-crisis cohort; 31 December 2004).
Evidence of an interaction between labour force status and sex, as well as labour force status and level of unemployment, was assessed by fitting the regression models with and without the interaction term and performing a likelihood ratio test. In a sensitivity analysis, we excluded individuals with prior mental health problems to assess potential bias due to reverse causality. Additional analysis was conducted in order to explore relationship between youth unemployment and specific mental diagnoses. Missing values were coded as separate categories. All analyses were computed using Stata Statistical Software: Release 13.
# Results
The final sample consisted of 6410 participants in the crisis cohort and 8162 in the non-crisis cohort. In both cohorts, individuals excluded were more likely to be male and born outside of Sweden compared with the included individuals. Similar prevalence of prior mental health problems was found among the included and excluded individuals. , B presents the study population of the crisis and noncrisis cohort. A larger proportion of individuals were defined as unemployed in the crisis cohort compared with the non-crisis cohort, consequently there was a smaller proportion of full-time employed in the crisis cohort.
## Descriptive characteristics
The parents of the unemployed individuals had worse mental health compared with the other groups. Generally, a higher prevalence of prior mental health problems was found among the unemployed individuals in both cohorts compared with the other groups. Furthermore, in the non-crisis cohort a larger proportion of unemployed individuals were not born in Sweden compared with the other groups.
Comparing the two cohorts, the parent's level of education and SEI was consistently higher in the crisis cohort compared with the non-crisis cohort. More male individuals were defined as being unemployed in the crisis cohort compared with the non-crisis cohort.
## Labour force status and mental diagnosis
Analyses were combined for male and female individuals, since no significant interaction between labour force status and sex was found. Eleven individuals died and were lost to follow-up, the majority was due to fatal accidents and one committed Baseline characteristics of the study population that participated in the Labour Force Survey in 1991-1994 for the crisis cohort and 1983-1986 for the non-crisis cohort among male and female individuals 17-24 years of age suicide. During the follow-up, a total of 557 individuals received inpatient care due to a mental diagnosis, 252 (3.9%) in the crisis cohort and 305 (3.7%) in the non-crisis cohort. In the crisis cohort, the average follow-up time was 17.1 years and 16.8 years in the non-crisis cohort.
In the fully adjusted analysis for the crisis cohort, <3 months unemployment (HR: 1.69, 95% CI 1.14 to 2.49), 3-6 months unemployment (HR: 2.19, 95% CI 1.43 to 3.37) and >6 months unemployment (HR: 2.70, 95% CI 1.71 to 4.28) was associated with an increased risk of getting a mental diagnosis [fig_ref] Table 2: Crude and adjusted HRs with 95% CIs for the associations between labour... [/fig_ref]. No association was found between being economically inactive or having an unstable labour force status and mental diagnosis. Similar results were seen in the fully adjusted analysis for the non-crisis cohort; <3 months unemployment (HR: 1.92, 95% CI 1.40 to 2.63), 3-6 months unemployment (HR: 2.60, 95% CI 1.72 to 3.94) and >6 months unemployment (HR: 3.33, 95% CI 2.00 to 5.57). Further, no association was found between being economically inactive or having an unstable labour force status and mental diagnosis.
# Interaction analysis
After adjusting for prior mental diagnosis, the HRs attenuated marginally more in the non-crisis cohort compared with the crisis cohort (model 2). The interaction between labour force status and level of unemployment in the fully adjusted model was not statistically significant ( p=0.95). However, part of the health selection effects could have been removed after adjusting for prior mental health problems, consequently we conducted an additional interaction analysis after adjusting for sex, age and country of birth, which reported similar findings ( p=0.79).
# Sensitivity analysis
Poor mental health is a risk factor of unemployment. Thus, we performed a sensitivity analysis excluding the 238 individuals with registered mental health problems prior to participating in the LFS. The sensitivity analysis demonstrated similar effects of unemployment on mental health as in the main analysis (see online supplementary table S1).
# Additional analysis
Additional analyses were performed to explore the effect of unemployment on alcohol and drug use disorders, affective disorders, stress-related disorders and self-harm . Although the three unemployment groups were combined into one group for this analysis, the results should be interpreted with caution due to few observations in each group. The results suggest that youth unemployment is strongly associated with alcohol and drug use disorders in the crisis cohort (HR: 2.58, 95% CI 1.43 to 4.65) and the non-crisis cohort (HR: 2.34, 95% CI 1.54 to 3.57). In addition, in the non-crisis cohort unemployment was positively associated with stress-related disorders (HR: 2.08, 95% CI 1.64 to 4.76).
# Discussion
The results of this nationwide study found that youth unemployment was associated with an increased risk of getting a mental diagnosis during a long-term follow-up, irrespective of the overall national rates of unemployment. Further, youth unemployment appears to be strongly associated with alcohol and drug use disorders.
The current results support and further extend the long-term negative consequences of youth unemployment in a number of ways. [bib_ref] Unemployment and psychological distress among young adults in the Nordic countries: a..., Reneflot [/bib_ref] First, in line with previous longitudinal studies and meta-analysis on the working-age population, a positive association between youth unemployment and mental diagnosis requiring inpatient care was found. [bib_ref] Unemployment impairs mental health: meta-analyses, Paul [/bib_ref] Previous research has found that youth unemployment is associated with decreased self-reported mental health and well-being. Extending previous research on the positive association between youth unemployment and self-reported alcohol consumption and alcohol dependence, the current results suggest an increased risk of being hospitalised for an alcohol or drug use disorder. Similar to previous research, the length of unemployment appeared to increase the risk of getting a mental diagnosis, which could be an indication of a causal relationship. The results referring to the group of young people unemployed <3 months should be interpreted with caution. This group partly contains youths who are in between education and working life, and some long-term unemployed youths that did not participate in all eight interviews, since consecutive unemployment was of interest. In accordance to previous research, the current results did not find strong evidence that the experience of unemployment changes when the national rate of unemployment fluctuates. Scholars have proposed stronger adverse health effects of becoming unemployed during a period of low unemployment compared with a period of high unemployment. [bib_ref] Excess mortality of unemployed men and women during a period of rapidly..., Martikainen [/bib_ref] The potential social and psychological processes that vary with context might, however, be different among youths compared with the middle-aged population. Youths are more likely to seek employment as opposed to becoming unemployed, consequently periods of high unemployment might be more stressful for this group due to the limited employment opportunities. Owing to sampling variability, the proportion of unemployed was high in both cohorts, well above the national unemployment rates at that time. These statistics are based on country averages, thus there could be areas with higher and lower unemployment rates. [bib_ref] Psychological and physical well-being during unemployment: a meta-analytic study, Mckee-Ryan [/bib_ref] Youth tend to pursue a variety of educational and employment pathways, consequently measuring their employment status every 3 months would capture this large variation in their labour force status. This should not have influenced our results as the individuals included in our cohorts experienced unemployment during a period of high and low unemployment, thus any potential influence of context on the health effects of unemployment should remain.
Taking a life course perspective, research suggests that unemployment during youth, a very sensitive time period, can independently effect health later in life. As a result, changes in prevalence of alcohol consumption and illegal drug use, both risk factors of mental health, during the long-term follow-up should not have much influence on our results. However, the economic crisis in the 1990s directly affected the welfare policy system in Sweden and youths were a very vulnerable group with little social welfare protection and welfare resources available. [bib_ref] Young people and the great recession, Bell [/bib_ref] Previous research suggests that better financial resources are related to better mental health, thus this change in the welfare resources might potentially have biased our results. [bib_ref] Psychological and physical well-being during unemployment: a meta-analytic study, Mckee-Ryan [/bib_ref]
# Strengths and limitations
A long follow-up time, a large sample size and register-based linkage are major strengths. The exposure was well defined and measured prospectively at repeated assessment points with a 3-month interval. Previous research has relied on retrospectively collected data on employment status either yearly, 5 years or up to 10 years in between each interview. The employment statuses in the LFS are based on numerous questions, in order to minimise misclassification. As the LFS is used to obtain the official labour force statistics in Sweden, the individuals categorised as unemployed are defined in accordance with European definition of unemployment set by Eurostat. The outcome of mental diagnosis was collected from highquality and reliable registers decreasing the risk of biased selfreported results and attrition. Previous research has to a great extent relied on the General Health Questionnaire. This outcome measure is highly susceptible to bias, especially when repeated measurements are collected. For example, research using this measure has found that individuals rated their mental health better when re-employed, compared with before becoming unemployed. A limitation is, however, that the Hospital Discharge Register began in 1964 and reached full coverage with respect to mental health in 1973.Thus, there might be a lack of coverage in relation to the individual's parent's registered mental diagnosis and their own prior mental diagnosis. Furthermore, in Sweden only a fraction of all individuals with mental health problems require inpatient care,thus although the outcome of interest was well defined, it only captured a proportion of the potential cases with the most severe mental health problems. Owing to coverage constraints, we were unable to include the outpatient care register. In addition, due to changes within the medical care system in Sweden since the 1990s, there could be several sources of bias in relation to the amount of hospital beds available, the proneness of doctors to refer mental health patients to inpatient care and the health-seeking behaviour of these individuals that could have influenced the results in different directions. 34 Adjusted HRs with 95% CIs for the associations between labour force status and specific mental diagnosis, stratified by cohort CONCLUSION Youth unemployment in Sweden was associated with increased risk of mental health problems that needed inpatient care, in times of an economic crisis and when employment rates were substantially lower. This is important at a time when youth unemployment rates are stable on a high level in Sweden.
What is already known on this subject?
It is known that youth unemployment is associated with self-reported poor mental health, and poor mental health could be a risk factor of unemployment.
## What this study adds?
This study adds that youth unemployment is a long-term risk factor for hospitalisation for a mental diagnosis independently of the overall national rate of unemployment.
[table] Table 2: Crude and adjusted HRs with 95% CIs for the associations between labour force status and mental diagnosis, stratified by cohort [/table]
|
Transcranial direct current stimulation combined with robotic therapy for upper and lower limb function after stroke: a systematic review and meta-analysis of randomized control trials
Background: Transcranial direct current stimulation (tDCS) is a non-invasive brain stimulation method able to modulate neuronal activity after stroke. The aim of this systematic review was to determine if tDCS combined with robotic therapy (RT) improves limb function after stroke when compared to RT alone.Methods:A search for randomized controlled trials (RCTs) published prior to July 15, 2021 was performed. The main outcome was function assessed with the Fugl-Meyer motor assessment for upper extremities (FM/ue) and 10-m walking test (10MWT) for the lower limbs. As secondary outcomes, strength was assessed with the Motricity Index (MI) or Medical Research Council scale (MRC), spasticity with the modified Ashworth scale (MAS), functional independence with the Barthel Index (BI), and kinematic parameters.Results: Ten studies were included for analysis (n = 368 enrolled participants). The results showed a non-significant effect for tDCS combined with RT to improve upper limb function [standardized mean difference (SMD) = − 0.12; 95% confidence interval (CI): − 0.35-0.11)]. However, a positive effect of the combined therapy was observed in the lower limb function (SMD = 0.48; 95% CI: − 0.15-1.12). Significant results favouring tDCS combined with RT were not found in strength (SMD = − 0.15; 95% CI: − 0.4-0.1), spasticity [mean difference (MD) = − 0.15; 95% CI: − 0.8-0.5)], functional independence (MD = 2.5; 95% CI: − 1.9-6.9) or velocity of movement (SMD = 0.06; 95% CI: − 0.3-0.5) with a "moderate" or "low" recommendation level according to the GRADE guidelines.Conclusions: Current findings suggest that tDCS combined with RT does not improve upper limb function, strength, spasticity, functional independence or velocity of movement after stroke. However, tDCS may enhance the effects of RT alone for lower limb function. tDCS parameters and the stage or type of stroke injury could be crucial factors that determine the effectiveness of this therapy.
# Background
Globally, cerebrovascular accident or stroke is a leading cause of death and disability among the adult population according to the latest estimates by the Global Burden of Disease (GBD). Most stroke patients live with disabilities affecting their quality of life, such as limb weakness or paralysis; deficits in balance, vision, or speech; and cognitive and psychological impairments. The rehabilitation process of these patients shows a nonlinear evolution, the clinical recovery period is shorter and the prognosis is better during the first weeks after the event (subacute phase), and the recovery is minimal or non-significant after the sixth month (chronic phase). For this reason, an early and intensive neurorehabilitation approach, consisting of functional and repetitive movements, should be carried out to restore normal function.
The robotic devices can provide repetitive, highintensity and task-specific treatment of the affected limbs and measure and quantify patient progress. Along with previous identified advantages, robotic therapy (RT) allows stroke survivors to perform independent training with less supervision, receive timely feedback and greater adherence to treatment. However, it has been demonstrated that RT alone is not superior to other conventional rehabilitation methods, and it may be necessary to optimize its effectiveness by including complementary therapies.
Transcranial direct current stimulation (tDCS) is a non-invasive brain stimulation method that has been shown to be a promising neurorehabilitation intervention. Its principal action mechanism is to modulate neuronal excitability networks of the affected and nonaffected hemisphere after stroke through the application of low intensity direct current through superficial electrodes applied on the scalp. Previous systematic reviews and meta-analyses have investigated the effects of tDCS as therapy alone or in combination with other treatments. However, no meta-analyses have been conducted to specifically analyse the effects of tDCS as an adjunct of robotic therapy on upper and lower limb function after stroke.
The aim of this systematic review and meta-analysis was to determine whether the combined use of tDCS and robotic therapy enhances the function of the upper and lower limbs in people with stroke compared to robotic-assisted rehabilitation alone. The secondary objective was to assess the safety of tDCS and the effectiveness in combination with RT in improving strength, spasticity, functional independence and movement velocity.
# Methods
This systematic review and meta-analysis followed the protocol developed in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelinesand it was registered in PROSPERO (reference number: CRD42020186963).
## Search strategy
Two independent researchers (AMG and NCS) performed an independently searched in the following databases: PubMed, Physiotherapy Evidence Database (PEDro) and the Cochrane Library. Moreover, the reference lists of all relevant articles were manually searched to identify studies that may have not been identified by the database search (Additional file 1). The databases were searched for articles published from the start of the databases until July 15th, 2021.Combinations of the following keywords were used to search the abovementioned databases: "Transcranial direct current stimulation", "tDCS", "non-invasive brain stimulation", "robotic", "robot", "exoskeleton", "Lokomat", "neurological disease", and "stroke".
## Eligibility criteria and study selection
The study selection process is shown in the flowchart in . The studies were selected based on the PRISMA checklist's PICOS method (P-participants; I-interventions; C-comparators; O-outcome and S-study design). We included studies in accordance with the following criteria: (1) the patients were diagnosed with a cerebrovascular accident or stroke; (2) the study was a randomized control trial (RCT); (3) transcranial direct current stimulation combined with robotic therapy was performed; (4) the intervention was compared with a control or conventional therapy; (5) the function of the upper/lower limbs was measured; and (6) the article was written in English or Spanish. Studies were excluded if they met the following criteria: (1) abstracts or congress conference papers; (2) non-human studies or preclinical trials; and (3) studies applying additional electrical stimulation as therapy. Two independent researchers (AMG and NCS) selected the studies based on the inclusion/ exclusion criteria. Disagreements were resolved by consensus with a third researcher (JGS).
Keywords: Transcranial direct current stimulation, tDCS, Robotic, Neuromodulation, Stroke
## Data extraction
The data were extracted by two researchers (AMC and NCS) using a chart designed for this purpose. A third researcher (JAC) compared both charts and presented the final data collected. This information was divided into two tables: , which includes the sociodemographic and clinical characteristics of the subjects in each study.
Regarding the primary outcome, we analysed the effect of the combined therapy on limb function using scales and functional tests. According to our previous protocol, when a study included more than one functional scale for upper limbs, the Fugl-Meyer motor assessment of the upper extremities (FM/ue) was considered first, which is a scale designed to assess reflex activity, movement control and muscle strength in the upper limbs. For the lower limbs, the 10-m walking test (10MWT) was chosen preferably. During this test, the subject had to walk a distance of 10 m as quickly as possible. The secondary variables adverse effects and patients lost to follow-up were measured as the number of participants who suffered adverse effects within each group and the number of participants lost to follow-up. In addition, we analysed the strength using the Motricity Index (MI) or the Medical Research Council scale (MRC); spasticity with the modified Ashworth scale (MAS); functional independence with the Barthel Index (BI); and kinematic parameters were assessed. Regarding the kinematics, only the velocity of movement data (degrees per second or cm per second) could be extracted. To measure the intervention effect, post-intervention scores instead of change scores were chosen. In the studies where it was necessary to obtain or clarify missing data, the corresponding authors were contacted for additional information. Data that were only represented by graphs were extracted using Graph Grabber v 2.0.1 software for graph digitalization (https:// www. quint essa. org/ softw are/).
## Risk of bias assessment
The potential risk of bias was assessed on the basis of Cochrane Collaboration's guidelines. This questionnaire was performed by two independent reviewers (MAM and DSM). Disagreements were resolved by a third senior researcher (JGS). Review Manager (RevMan) software (computer program, version 5.3, Copenhagen:
The Nordic Cochrane Centre, The Cochrane Collaboration, 2014) was used to perform the analysis. Six items were addressed, and the relevant risk was expressed as three levels (unclear, low, and high). The researchers agreed prior to the assessment that for the item "blinding of the participants and personnel", the level of risk would be rated as unclear when the participants or personnel were not blinded, and for the item "selective reporting", studies without a registered protocol would be qualified as having unclear or high risk, depending on the final report. Additionally, funnel plots for the main variable (function) were assessed to evaluate publication bias.
## Data synthesis and analysis
The inverse variance method and a fixed-effects model were used for the 7 assessed variables. The standardized mean difference (SMD) was used to express the results for upper and lower limb function, strength and movement velocity since these variables are sometimes reported with different scales or units. Lower limb function was assessed by the 10MWT, which measures the time an individual takes to walk 10 m. A higher score indicates more severe disability, so this value was multiplied by − 1 to align the effect direction. The mean difference (MD) was used for the spasticity and functional independence results, which were expressed on the same scale in the included studies. The risk difference (RD) was calculated for the adverse events and loss to followup variables. The 95% confidence interval (95% CI) were calculated for all outcomes. Statistical heterogeneity was evaluated using the chi-squared test (with statistical significance set at p < 0.10) and was measured by calculating the I 2 , with values of 25%, 50%, and 75% representing low, moderate, and high heterogeneity, respectively.
The results corresponding to the longest follow-up period were analysed for each of the included studies. In the studies where the results were reported by the intention-to-treat and by protocol, the data from the intention-to-treat analysis were used. If participant data were available and the authors did not present the intention-to-treat results, this analysis was also performed. In the three-arm studies, the shared group was split according to the Cochrane Group guidelinesto avoid data being counted twice. In addition to the global analysis, an analysis by subgroups was conducted for the main variable, limb function (lower limb vs upper limb), the time from stroke onset (< 16 weeks vs ≥ 16 weeks), and the tDCS current density (≥ 0.05 mA/ cm 2 vs < 0.05 mA/cm 2 ). The analysis by subgroups was not performed based on the follow-up period as in the previous protocol because all but one study had a short follow-up period equal to or less than 3 months. RevMan software was used for quantitative analysis. The quality of evidence was classified for each outcome as high, moderate, low, or very low following the Grades of Recommendation Assessment, Development and Evaluation (GRADE) method.
# Results
After the duplicates were removed, 445 articles were identified as eligible, and 389 were excluded after the titles and abstracts were read. Finally, after the full texts were read, 10 RCTsthat met the inclusion criteria were included in this systematic review and metaanalysis . Additional information was requested from the authors of five studies, but a response was received from only one author.
## Qualitative summary of the included studies
All included studies were sham controlled. The effect of active anodal tDCS was compared with those of sham tDCS and both therapies combined with robot-assisted rehabilitation. Of the included studies, seven were aimed at treating the upper limbs, and threewere aimed at treating the lower limbs . The sample size comprised 368 enrolled participants (n = 207 in active tDCS groups and n = 173 in sham tDCS groups); n = 159 were women (43.2%), n = 299 had ischaemic stroke (81.25%), n = 122 had cortical lesions (33.15%) and n = 60 had subcortical lesions (16.3%). The average age ranged between 52.7 and 75.25 years. The time since stroke onset was < 6 months (subacute stroke) in three studies, ≥ 6 months (chronic stroke) in five studies, and both subacute and chronic stroke were assessed in two studies. For these studies on two types of stroke, we contacted the authors to request the results of separate subacute and chronic analysis. The sociodemographic and clinical characteristics of the patients in the included studies are shown in .
In all included studies, tDCS was performed with anodal stimulation over the primary motor cortex (M1) from the stroke-affected hemisphere. Hesse et al.included a third group in which a cathode over the M1 of the unaffected hemisphere was used as the active electrode in one of the three study arms. The cathode electrode was placed on the contralateral supraorbital area in all studies except in two studieswhere was applied over M1 of the unaffected hemisphere. In eight studies, the tDCS session lasted for 20 min, and in the two remaining studies, the session lasted for 7and 30 min. tDCS was delivered simultaneously during RT (online stimulation) in six studies. The most common frequency of sessions was five sessions per weeks. The total number of sessions ranged from two to thirty-six. The electrode area was 35 cm 2 in most of the studies, and only one studyused electrodes of 25 cm 2 . The current intensity ranged between 1 and 2 mA, and the current density ranged between 0.03 and 0.08 mA/cm 2 .
In the robot-assisted protocol, the duration of the session ranged from 20 to 60 min. The robots used for gait training were the Gait Trainer GT1 (Reha-Stim, Berlin, Germany), Lokomat (Hocoma Inc, Switzerland), and Walkbot (P&S Mechanics, Seoul, Republic of Korea). For upper limb therapy, the robot-assisted Bi-Manu Track (Reha-Stim Bi-Manu Track, Berlin, Germany), Armeo ® Spring (Hocoma AG, Switzerland), REO Therapy System (Motorika, LTD, Israel), InMotion wrist robot (Interactive Motion Technologies, Inc., Cambridge, MA, USA), REAplan robot (Axinesis, Wavre, Belgique), MIT Manus (planar robot)and InMotion WRIST robot (Bionik Laboratories Corp., Watertown, MA, USA)were used (see .
The change in upper limb function was measured with the FM/ue scale in six studiesand the Box & Block test in five studies. The effect on lower limb function was analysed by the 10 MWT in three studies. With regard to the secondary variables, six studies assessed strength by the MRC scale in upperand lower limbsor MI scale in two studies for upper limbsand one study for lower limbs, four studies assessed spasticity by the MAS scale in the upperand lower limbs, two studiesassessed functional independence for upper limbs by the BI scale, and three studiesassessed upper limb velocity. Additionally, some studies assessed other variables and/or scales that were not within the objectives of the registered protocol of this review (see . The longest follow-up period was 6 months. The follow-up period was less than or equal to three months (2 to 12 weeks) in five studies, and in the remaining four studies, there was no follow-up period . Five studies reported lost to follow-up : a total of n = 30 (8.6%), with n = 17 from the experimental group (tDCS) and n = 13 from the control group Adverse effects and/or complications were specifically stated in nine of the ten included studies. Of these nine studies, five did not report any adverse events, and the other fourreported mild adverse effects related to tDCS .shows the risk of bias for the ten included studies. Four trials presented an unclear selection biassince the way in which the participants were randomized to groups was not described in detail. In terms of performance bias, six studies had an unclear risk since the blinding of the participants but not the blinding of the personnel was possible. However, the study by Geroin et al.was assessed as having a high risk of bias regarding the blinding of the participants and personnel, as the researchers kept the device turned off throughout the session in the sham group instead of using ramps at the beginning and end of the session. Eight of the ten assessed trials were rated as having a low risk of detection bias.
## Risk of bias in the included studies
Only the two studies carried out by Mazzoleni et al.were rated as having an unclear and high risk of bias because the authors did not specify whether assessor blinding was conducted and because the study was single blinded, respectively. Regarding the outcome data, only the study published by Danz et al.was considered to have a high risk of attrition bias, as the authors reported only the change scores for the main variable to measure the intervention effect. Four studieswere considered to have a low risk of reporting bias due to the protocols being previously registered, and three studieswere considered to have an unclear risk since the protocols had not been previously registered. Three studies were rated as having a high risk; Geroin et al.did not report the results of the spasticity outcome, Triccas el al.reported some secondary variables that were different from those registered in the previous protocol, and Straudi et al.did not report spasticity or motorevoked potential outcomes, although these outcomes were included in the previous protocol.
The risk of publication bias was considered low since the distribution of the main variable (function) in the funnel plots was not asymmetrical.
## Quantitative summary: effects of active tdcs versus sham tdcs both methods combined with robotic-assisted rehabilitation
According with the objective of this meta-analysis and the protocol published in PROSPERO, a quantitative analysis of the main variable function was performed. This effect was investigated in ten studies, seven for upper limbsand three for lower limbs. The secondary outcome strength was investigated in six studies, the spasticity in three studies, the functional independence in two studies, and the velocity of upper limb movements in three studies.summarizes the trials that assessed the effect of the combination of active tDCS and robotic-assisted rehabilitation compared with that of the combination of sham tDCS and robotic-assisted rehabilitation on function.
## Effect on function
The study by Danz et al.was excluded from this analysis, as the authors reported the results as change scores. No differences were observed in the magnitude of improvement in function between the experimental group (active tDCS) and the control group (sham tDCS) (SMD = − 0.05; 95% CI: − 0.27-0.16), and there was a low level of heterogeneity (I2 = 0%, p = 0.61). In addition, no differences were observed between the experimental and control groups in the individual analysis of the included studies. In the subgroup analysis of function, a potential effect of the tDCS combined with robotic-assisted rehabilitation was observed in the lower limb function (SMD = 0.48; 95% CI: − 0.15-1.12), which was limited by only two studies. A non significant effect of the combined therapy was found in the upper limb function (SMD = − 0.12; 95% CI: − 0.35-0.11) . When this effect was compared in people with chronic stroke (≥ 6 months) and with subacute stroke (< 6 months), no differences were found (Chi2 = 0.8, p = 0.36) . For this analysis, the study by Straudi et al.was excluded since the results for subacute and chronic patients were reported together. In addition, no differences were observed in the results between different dosages or current densities applied regarding the subgroups of ≥ 0.05 mA/cm 2 and < 0.05 mA/cm 2 (Chi 2 = 0.0, p = 0.99) . The quality of the evidence for this outcome according to the GRADE guidelines was moderate, considering a serious risk of bias as a factor that rating down. The individual results of the included studies showed that only Geroin et al.reported differences between the active and sham groups favouring the sham group. The quality of this evidence, according to the GRADE guidelines, was considered low, considering a serious risk of bias and heterogeneity of results as factors that rating down.summarizes the trials that assessed the effect of the interventions on spasticity by the modified Ashworth scale. The overall effect on spasticity showed no differences between the active and sham tDCS groups (MD = − 0.15; 95% CI: − 0.8-0.5) and showed a low level of heterogeneity (I2 = 0%, p = 0.82). In addition, no differences were observed between the experimental and control groups in the individual results of the included studies. The quality of this evidence, according to the GRADE guidelines, was moderate, considering a serious risk of bias as a factor that rating down.summarizes the trials that assessed the effect of the interventions on functional independence by the Barthel Index. The overall effect on this outcome did not differ between the active tDCS and sham tDCS groups (MD = 2.5; 95% CI: − 1.9-6.9) and showed a low level of heterogeneity (I2 = 0%, p = 0.72). No differences were observed between the experimental and control groups in the individual results of the included studies.
# Effect on strength
## Effect on spasticity
## Effect on functional independence
The quality of this evidence, according to the GRADE guidelines, was moderate, considering a serious risk of bias as a factor that rating down.summarizes the trials that assessed the effect of the interventions on upper limb movement velocity. The overall effect on this outcome did not differ between the active tDCS and sham tDCS groups (SMD = 0.06; 95% CI: − 0.3-0.5) and showed a low level of heterogeneity (I2 = 0%, p = 0.80). In addition, no differences were observed between the experimental and control groups in the individual results of the included studies. The quality of this evidence, according to the GRADE guidelines, was moderate, considering a serious risk of bias as a factor that rating down. Regarding the individual results of the included studies, only the study carried out by Edwards et al.showed a high risk for adverse events in the active tDCS group. The quality of this evidence, according to the GRADE guidelines, was moderate, considering a serious risk of bias as a factor that rating down.
## Effect on velocity of upper limb movements
## Adverse events and lost to follow
# Discussion
This systematic review and meta-analysis included 10 RCTs and 368 participants and was conducted to investigate the effects of tDCS as an adjunct to robotic therapy on limb function after stroke. In addition, the safety of tDCS and its effectiveness to improve strength, spasticity, functional independence and movement velocity were analysed. Currently, the evidence about the effectiveness of tDCS in previous systematic reviews and clinical trials is contradictory. The results obtained in the present meta-analysis showed non-significant improvement for the main variable (function) and secondary variables (strength, spasticity, functional independence and movement velocity), with a "moderate" to "low" recommendation level according to the GRADE guidelines. These results reveal that tDCS does not have an additional effect to RT alone on these outcomes. A recently published guidelines and a meta-analysis on the use of tDCS for neurological and psychiatric disordersfound that when tDCS was combined with other therapies in the treatment of subacute and chronic stroke, patients showed improvements, and tDCS enhanced the effects of the adjuvant therapy. However, tDCS combined with intensive RT did not improve motor recovery to a greater extent than did RT. Our results are consistent with this guidelines, the subgroup analysis results showed no statistically significant differences between the experimental and the control tDCS groups, and both groups experienced improvements in clinical and kinematic variables.
Regarding other factors that may influence the effectiveness of tDCS, we can find the type and stage of the lesion, which can affect stroke evolution. A casecontrolled study showed that when two patients had the same basal neurological severity, same functional disability, age and sex, haemorrhagic stroke patients had better prognosis than ischaemic ones. In our review, of the 368 enrolled patients, 81.25% had ischaemic stroke, and most of the participants suffered from chronic-stage stroke and cortical lesions. Several studies have shown larger improvements with rehabilitation in the subacute stage (< 6 months) than in the chronic stage (> 6 months). These benefits may be related to spontaneous recovery, which is usually observed over third month after stroke onset. This period could be extended, depending on the severity, type and intensity of the intervention. Factors including tDCS parameters, electrode size, electrode location, stimulation duration and the number of sessions could also affect the effectiveness of the intervention. The tDCS protocols of the included studies in this review are too heterogeneous. The current density has been described as one of the main parameters that determine the effectiveness of tDCS. Two systematic reviews and meta-analyses showed a positive relationship between current density and the recovery of motor function. The current density determines the electrical field strength, which depends on the current intensity and electrode size. Higher current densities or smaller electrodes are associated with a higher efficacy of tDCS, which means a deeper penetration of the current into the scalp changing the excitability of the neurons under the electrode. Traditionally, intensities ranging between 1 and 2 mA are used in human studies. To date, the effects of current densities greater than 0.08 mA/cm 2 are unknown. Regarding adverse events, tDCS can be considered a safe therapy with some mild adverse effects observed in the included studies. It is necessary to perform studies where adverse events are actively assessed, as half of the included studies in this review did not report any adverse events, and one of the studies did not mention adverse events.
This systematic review and meta-analysis has limitations that could affect the obtained results: (1) due to differences in the patient characteristics, the study population is heterogeneous. The results obtained in this review cannot be generalized to haemorrhagic or subacute stroke patients because most of the enrolled participants had ischaemic or chronic stroke. (2) The sample sizes of the included studies were larger than 30 in only 3 studies. This factor may influence the results, as studies with larger sample sizes, in which tDCS combined with other therapies, found significant differences in the variables analysed in stroke patients.
(3) The heterogeneity in the tDCS parameters assessed made it difficult to compare the results. (4) There was variability in the number of sessions, the intervention protocol and the devices used for upper and lower limbs RT. (5)Most of the included studies had a short follow-up period, with only one study including a longer follow-up period of 6 months.The study outcomes monitored differed across studies, which limited the ability to compare outcomes across studies.
# Conclusion
The reported findings suggest that the application of tDCS as an adjunct to RT does not enhance the effect of RT alone on upper limb function after stroke. This meta-analysis revealed that tDCS combined with RT may favour the lower limb function, however these results should be interpreted with caution because there were analysed by only two studies. Furthermore, positive results favouring tDCS combined with RT were not found in strength, spasticity, functional independence or movement velocity with an evidence confidence |
Risk factors analysis and prevention of metabolic bone disease of prematurity
The present study aims to analyze the risk factors for metabolic bone disease (MBD) of prematurity.A total of 238 preterm infants who were born at <34 weeks of gestation and were hospitalized for at least 6 weeks in the Department of Neonatology, Fujian Maternity and Children Hospital between January 1, 2011 and November 30, 2015 were enrolled in the study. Sixteen preterm infants diagnosed with MBD were selected as the case group, and 32 non-MBD preterm infants were matched 2:1 at admission into the study. The 2 groups were compared to examine the differences in maternal obstetric conditions, conditions during parturition, neonatal conditions, and neonatal diseases and treatments. The risk factors for MBD of prematurity were analyzed using t tests, x 2 tests, and a logistic regression model.The mean gestational age and birth weight of the case group were significantly lower (P < .05) than those of the control group. Compared with the control group, the case group had a significantly higher ratios of small-for-gestational-age infants, antenatal maternal corticosteroids use, sedative use, ventilator use, aminophylline use, diuretic use, liver function impairment, vitamin D (VitD) supplementation at more than 14 days of age, achievement of total enteral nutrition (TEN) beyond 28 days of age, and feeding intolerance.Logistic regression analysis showed that birth at <30 weeks of gestation, VitD supplementation at >14 days of age, and achievement of TEN beyond 28 days of age were independent risk factors for MBD (P < .05).Level of Evidence: IVAbbreviations: ALP = alkaline phosphatase, MBD = metabolic bone disease, NICU = neonatal intensive care unit, PTH = parathyroid hormone, SGA = small for gestational age, TEN = total enteral nutrition, VitD = vitamin D.Medicine ® OPEN P < .05 was considered to indicate statistical significance. CI = confidence interval, OR = odd ratio, TEN = total enteral nutrition, VitD = vitamin D.
# Introduction
Metabolic bone disease (MBD) of prematurity is a metabolic disease in preterm infants due to disorders in calcium and phosphorus metabolism. The clinical manifestations include abnormal bone mineral content, decreased trabecular bone, cortical thinning, and other skeletal changes. Severe cases may present with rickets-like symptoms and even fractures. MBD occurs mainly in extremely preterm infant, especially in very-low-or ultra-low-birth-weight infants. The condition often occurs in preterm infants 6 to 8 weeks after birth. [bib_ref] Metabolic bone disease of prematurity, Vachharajani [/bib_ref] MBD is a serious complication in the late neonatal period, and it has a negative impact on the growth and development of preterm infants. Shortterm, this disease can cause ventilator dependence and increase the risk of fractures, and it can result in increased risk of myopia, kidney failure, and abnormal bone development, which impairs respiratory function, affects adult height or lead to senile osteoporosis in the long term. [bib_ref] Metabolic bone disease of prematurity, Vachharajani [/bib_ref] [bib_ref] Daily physical activity in low-risk preterm infants: positive impact on bone strength..., Tosun [/bib_ref] [bib_ref] Outcomes of standardised approach to metabolic bone disease of prematurity, Chin [/bib_ref] Early clinical diagnosis of MBD is difficult, and clinical manifestations often lag far behind blood biochemical changes or bone changes seen on x-ray. Therefore, a comprehensive diagnosis of MBD is largely made based on declines in serum calcium and phosphorus, elevation in alkaline phosphatase (ALP) and parathyroid hormone (PTH), and osteopenia or fractures revealed by x-ray, among other characteristics.
The understanding of MBD has grown in European and American countries in recent years. These developed countries have improved the nutrient management strategy for preterm infants, modified the clinical intervention methods, and strengthened the monitoring of high-risk infants, all of which help to gradually decrease the incidence of MBD and improve the quality of life for preterm infants. However, in-depth research and a systematic understanding of MBD of prematurity remain lacking in developing countries, from which too few relevant reports contribute to insufficient awareness and prevention of MBD. Owing to the rise and development of perinatal medicine, as well as the markedly improved technical levels in neonatal intensive care units (NICUs) in developing countries, the medical treatment and diagnosis capacity for preterm infants has significantly improved, and the survival rate of preterm infants has substantially increased. Cases of prematurity with MBD are not uncommon, and these require urgent attention. Preterm infants have insufficient innate reserves of minerals, and it is difficult to supplement mineral elements through the gastrointestinal tract in the early postnatal stage. Moreover, the primary disease treatment often requires extensive medical intervention. Together, these factors frequently lead to disorders in calcium and phosphorus metabolism and to bone loss, which together initiate the development of MBD. The present study analyzed the risk factors for MBD of prematurity to provide some suggestions for the prevention and control of this condition.
# Materials and methods
## Participants
A total of 249 preterm infants were retrospectively analyzed. All infants were born at <34 weeks of gestation and remained hospitalized in the Department of Neonatology of the authors' hospital, for at least 6 weeks between January 1, 2011 and November 30, 2015. We excluded patients who were admitted at >1 day of age and preterm births complicated by a congenital chromosomal abnormalities or an inherited metabolic disease. A total of 238 cases were included. Sixteen preterm infants (9 boys and 7 girls) with the following features during hospitalization were chosen for the case group: rickets-like clinical manifestations; serum phosphorus <1.3 mmol/L and/or ALP >800 mmol/L, and/ or PTH>130 pg/mL; x-ray findings including loss of metaphyseal sclerotic line; diaphyseal cortical thinning; thinning of the vertebral endplates with an "empty vertebral body"; evidence of osteopenia plus metaphyseal cupping, fraying or irregularity; physeal widening especially of the wrist or knee; cupping or enlargement of the anterior rib ends at the costochondral junction; and/or acute or healing fracture. [bib_ref] Comparison of intact parathyroid hormone, alkaline phosphatase, phosphate levels for diagnosing severe..., Tkach [/bib_ref] Thirty-two non-MBD preterm infants were included in the control group and were matched 2:1 at admission. The development of MBD of prematurity or hospital discharge was the outcome. The study was approved by the Ethical Committee of the authors' hospital. Informed consent was obtained from all individual participants (from their parents or guardians) included in the study.
# Methods
Medical records were examined retrospectively. The following data were recorded: maternal conditions, including the number of births, pregnancy complications (e.g., gestational hypertension and gestational diabetes), endocrine disorders (e.g., parathyroid dysfunction), prenatal abnormalities of calcium and phosphorus metabolism, and prenatal steroid hormone use; conditions during parturition, including the mode of delivery and the presence of birth asphyxia; neonatal conditions, including gestational age, sex, birth weight, and the presence of small for gestational age (SGA); and neonatal disease and treatment, including the use of a ventilator, parenteral nutrition, sedatives, corticosteroids, aminophylline, and diuretics, as well as early supplementation of calcium, phosphorus, and vitamin D (VitD). The case and control groups were compared to analyze the differences in the above characteristics.
# Statistical analysis
Data acquisition and data analysis were accomplished by different individuals. Data analysis was performed using SPSS [bib_ref] Evidence of aluminum loading in infants receiving intravenous therapy, Sedman [/bib_ref].0, by one of the authors, who was blinded to data acquisition and grouping. Data with a normal distribution are expressed as̄x ± sd; intergroup comparison was performed using the 2 independent samples t test. Count data are described using frequencies and percentages; the intergroup comparison of count data was performed using the x 2 test. Multifactorial analysis was conducted using forward stepwise logistic regression analysis. P < .05 was considered to indicate statistical significance.
# Results
## Incidence of mbd
The incidence of MBD in preterm infants was 6.7% (16/238). In the case group, 3 of 16 cases suffered fracture and x-ray examination indicated there was rickets-like changes in 13 cases. Their mean gestational age was 28.8 ± 1.5 weeks, and their mean birth weight was 1023.8 ± 192.5 g.
## Unifactorial analysis of risk factors for mbd
The case group had a significantly lower gestational age and birth weight than the control group, and they also had significantly higher use of antenatal maternal corticosteroids, sedatives, ventilators, aminophylline, and diuretics. In addition, the case group had a significantly higher ratio of SGA infants as well as significantly greater liver function impairment, VitD supplementation at >14 days of age, achievement of total enteral nutrition (TEN) at >28 days of age, and feeding intolerance [fig_ref] Table 1: Unifactorial analysis of risk factors for metabolic bone disease of prematurity between... [/fig_ref].
## Multiple logistic regression analysis of risk factors for mbd
Relevant factors selected by the unifactorial analysis were subjected to nonconditional multiple logistic regression analysis, and an a value of 0.05 and a b value of 0.10 were obtained. The results showed that gestational age of <30 weeks, achievement of TEN beyond 28 days of age, and VitD supplementation after 14 days of age were retained in the stepwise regression equation. These 3 factors were therefore identified as the main factors influencing the development of MBD in preterm infants [fig_ref] Table 2: Multiple logistic regression analysis of risk factors for metabolic bone disease of... [/fig_ref].
# Discussion
The development of MBD of prematurity is associated with gestational age at birth. The incidence of MBD of prematurity increases at younger gestational ages and lower birth weights. [bib_ref] Metabolic bone disease of prematurity, Vachharajani [/bib_ref] The incidence of MBD is up to 30% in preterm infants born at <28 weeks of gestation, and approximately 10% of preterm infants suffer fractures by a corrected gestational age of 36 to 40 weeks. [bib_ref] Metabolic bone disease of prematurity, Vachharajani [/bib_ref] [bib_ref] Metabolic bone disease of prematurity, Backström [/bib_ref] [bib_ref] Osteopenia of prematurity: a national survey and review of practice, Harrison [/bib_ref] Our results also show that young gestational age was a risk factor for the development of MBD of prematurity. Premature birth leads to MBD mainly because premature birth results in deficient mineral reserves in the fetus. The last 3 months of pregnancy are important because the fetus acquires 80% of calcium and phosphorus reserves between 25 and 40 weeks of gestation. [bib_ref] Clinical guidelines: nutrition support of neonatal patients at risk for metabolic bone..., Nehra [/bib_ref] [bib_ref] Metabolic bone disease of prematurity and secondary hyperparathyroidism, Lothe [/bib_ref] In this period, the mean deposition rates of calcium and phosphorus are 100 to 120 and 50 to 65 mg/kg/day, which provide the newborn with 20 and 10 g of calcium and phosphorus reserves. If preterm birth occurs in this period, the newborn will partially or completely miss the optimal stage of acquiring calcium and phosphorus reserves. [bib_ref] Calcium, phosphorus, magnesium and vitamin D requirements of the preterm infant, Mimouni [/bib_ref] A study by Henriksen et al [bib_ref] Fat-soluble vitamins in breast-fed preterm and term infants, Henriksen [/bib_ref] showed that serum 25-(OH)D 3 levels are significantly affected by gestational age in preterm infants born at 25 to 35 weeks of gestation; suggesting that preterm birth can also affect VitD reserves in newborns. In addition, the younger the gestational age, the less the organ systems have matured, and the likelihood of requiring drug intervention increases. The use of aminophylline, diuretics, and steroid hormones can also lead to the loss of bone minerals as well as to disorders in calcium and phosphorus metabolism. [bib_ref] Metabolic bone disease of prematurity and secondary hyperparathyroidism, Lothe [/bib_ref] [bib_ref] Metabolic bone disease: a continued challenge in extremely low birth weight infants, Viswanathan [/bib_ref] The present study found that late achievement of TEN was a risk factor for MBD of prematurity. Preterm infants, particularly those with very-low-birth weight or ultra-low-birth weight, are often unable to feed in the early postnatal period because of illness or have difficulties achieving TEN in the short term, and thus requiring long-term parenteral nutrition. However, the formulation of parenteral nutrition often fails to provide a sufficient or available mineral supply because of various factors, including a deficiency of corresponding mineral formulations, poor solubility of minerals, mutual antagonism of nutrients, impact of the pH value, and limit of the volume of intravenous fluid for preterm infants because of illness. [bib_ref] Committee on NutritionCalcium and vitamin d requirements of enterally fed preterm infants, Abrams [/bib_ref] Consequently, calcium and phosphorus depositions in preterm infants during the early postnatal period cannot meet the requirements of the intrauterine bone growth rate. [bib_ref] Calcium, phosphorus, magnesium and vitamin D requirements of the preterm infant, Mimouni [/bib_ref] [bib_ref] Metabolic bone disease in preterm newborn: an update on nutritional issues, Bozzetti [/bib_ref] Furthermore, it has been reported that aluminum contamination of parenteral nutrition can cause MBD. [bib_ref] Aluminum contamination in parenteral nutrition admixtures for lowbirth-weight preterm infants in Mexico, Lima-Rogel [/bib_ref] A study reported that the bone aluminum content of preterm infants receiving parenteral nutrition for >3 weeks was 10-fold that of the control group. [bib_ref] Evidence of aluminum loading in infants receiving intravenous therapy, Sedman [/bib_ref] [bib_ref] Prenatal deficiency of phosphate, phosphate supplementation, and rickets in very-low-birthweight infants, Holland [/bib_ref] Aluminum contamination of parenteral nutrition can result in excessive aluminum deposition at the mineralized surface of bone, which affects osteoblast activity and hinders bone formation, ultimately leading to osteomalacia. This seriously affects the bone health of infants in both the short and long terms.
Fortified formula has a suitable proportion of minerals and balanced nutrients, and the proportion of calcium and phosphorus is similar to that of breast milk. Thus, fortified formula can provide an optimal calcium and phosphorus deposition rate during intrauterine growth. For example, the gastrointestinal absorption rate of phosphorus can exceed 90% during enteral feeding. [bib_ref] Prophylactic effect of low dose vitamin d in osteopenia of prematurity: a..., Taheri [/bib_ref] Therefore, enteral nutrition has unparalleled advantages for ensuring the absorption efficiency of minerals, including calcium and phosphorus, in preterm infants during the early postnatal period.
A lack of VitD supplementation in the early postnatal period is associated with the development of MBD. [bib_ref] Prophylactic effect of low dose vitamin d in osteopenia of prematurity: a..., Taheri [/bib_ref] [bib_ref] Prematurity and bone health, Pieltain [/bib_ref] The European Society for Pediatric Gastroenterology, Hepatology, and Nutrition recommends 800 to 1000 IU/day of VitD for preterm infants in the first week after birth. This dose can ensure more than 80 nmol/L of 25-(OH)D 3 and satisfy the metabolic needs of preterm infants. [bib_ref] Osteopenia of prematurity: a national survey and review of practice, Harrison [/bib_ref] In the present study, 13 preterm infants in the case group received no VitD supplementation within 14 days of birth, and 4 cases only received attention and intervention after the development of MBD. The mechanism of VitD acts through 1, 25-(OH) 2 D 3 , a major hormone involved in maintaining the balance of calcium and phosphorus metabolism. The 1,25-(OH) 2 D 3 hormone acts on the intestine, kidney, and bone, among other target organs to regulate the transport, deposition, and release of calcium and phosphorus; 1,25-(OH) 2 D 3 is essential to ensure the bone conversion of calcium and phosphorus. Insufficient long-term intake of VitD can result in a series of pathophysiological changes, including decreased intestinal absorption of calcium and phosphorus, increased urinary excretion of calcium and phosphorus, and disorders in osteogenesis. The catch-up growth of preterm infants after birth requires a high nutrient demand. However, the VitD reserves of preterm infants are insufficient at birth, while a lack of sunlight due to long-term hospitalization also affects VitD production in the body. In addition, the primary disease condition often leads to delayed feeding or repeated fasting that also negatively affects the early administration of VitD. All of these factors can lead to a serious deficiency of VitD in preterm infants and thus harm the balance of calcium and phosphorus in vivo. This imbalance will cause a lack of calcium and phosphorus as well as disorders in bone metabolism, leading to the development of MBD. [bib_ref] Prematurity and bone health, Pieltain [/bib_ref]
# Limitations
This study has several limitations. First, our sample size seemed relatively small and that may influence the results. As we know, in general, the incidence of MBD is very low in newborns, but it increases at younger gestational ages and lower birth weights, particularly for the newborns with a gestational age <28 weeks. The level of technology in the NICU is markedly improved in China, but the expensive treatment cost and the possible sequel make many young couples suspend treatment of their extremely preterm infants. Moreover, due to the unbalanced development of the economy and society in China, it is still a common problem and cannot be solved in the short term. Therefore, long-term survival of extremely preterm infant is not always assured. However, the sample source in our study was collected from the authors' hospital, which is a top neonate emergency center in Fujian Province and the vast majority of the extremely preterm infants of this region were treated here. Although the sample size seems small in our study the data from the infants was precious and representative because it likely reflects the current state of MBD in Fujian Province, China. Second, our study should also investigate factors which may mediate the incidence of MBD, yielding a certainty about the relative contribution of nonspecific factors in the outcome and identifying specific and nonspecific factors which may help us to formulate intervention. Third, followup data should be included. Longitudinal samples of high risk of MBD are ideal for this type of study, as it is possible to evaluate the subjects on a baseline before they are exposed to risk factors.
# Conclusions
MBD is a serious disease affecting the short-term and long-term physical development of preterm infants. Premature birth, lack of TEN, and early VitD deficiency are the major risk factors for MBD. For the prevention and treatment of MBD, it is important to guarantee perinatal care, actively prevent premature birth, and appropriately increase the balanced intake of calcium, phosphorus, and VitD in the third trimester of pregnancy. In the early postnatal period, sufficient minerals should be supplied for preterm infants through parenteral nutrition, and make the transition to TEN as soon as possible. Oral supplementation of calcium, phosphorus, and VitD is needed when enteral nutrition is first established. For primary disease treatment, caution must be taken to avoid the use of drugs that may lead to decreased absorption or increased excretion of mineral elements and VitD, such as steroid hormones, aminophylline, diuretics, or sedatives. For preterm infants at risk, it is necessary to perform dynamic monitoring of blood biochemistry as well as to make an early diagnosis and undertake timely interventions to reduce complications and improve the prognosis, all of which improve the quality of life of preterm infants.
Data curation: Wenhao Chen, Changyi Yang. Formal analysis: Baoquan Zhang. Funding acquisition: Wenhao Chen.
[table] Table 1: Unifactorial analysis of risk factors for metabolic bone disease of prematurity between the case and control groups. P < .05 was considered to indicate statistical significance. BW = birth weight, CI = confidence interval, GA = gestational age, NICU = neonatal intensive care unit, OR = odd ratio, SGA = small for gestational age. [/table]
[table] Table 2: Multiple logistic regression analysis of risk factors for metabolic bone disease of prematurity. [/table]
|
Tumor microenvironment-oriented adaptive nanodrugs based on peptide self-assembly
The aberrant metabolism of tumor cells creates an inimitable microenvironment featuring acidic pH, high glutathione (GSH) levels, and overexpression of certain enzymes, which benefits the overwhelming progress of a tumor. Peptide self-assembly, emerging as a biofriendly and versatile fabrication strategy, harnesses multiple noncovalent interactions to obtain a variety of nanostructures tailored on demand.Orchestrating the reversible nature of noncovalent interactions and abnormal physiological parameters in the tumor microenvironment enables peptide-based nanodrugs to be targetable or switchable, thereby improving the drugs' bioavailability and optimizing the treatment outcome. This review will focus on peptide-modulated self-assembly of photosensitizers, chemotherapeutic drugs, immunoactive agents for tumor microenvironment-oriented adaptive phototherapy, chemotherapy, immunotherapy and combinatorial therapy. We will emphasize the building block design, the intermolecular interaction principle, adaptive structural transformation in the tumor microenvironment and corresponding therapeutic efficacy, and aim to elucidate the critical role of peptide-modulated, tumor microenvironment-oriented adaptive assemblies in improving the therapeutic index. Challenges and opportunities will be covered as well to advance the development and clinical application of tumor therapies based on peptide self-assembly materials and techniques.
# Introduction
Tumors, as a major burden of disease worldwide, feature a pathophysiological dysfunction of cells and their surrounding environment.The aberrant metabolism of tumor cells creates physiological variations compared to normal tissues and therefore forms a tumor microenvironment, which is inimitably characterized as acidic pH, 4-6 high glutathione (GSH) concentration, 7-9 overexpression of certain enzymes or ligands,awed vascular system and hypoxia,as well as other cells including immune cells, broblast cells, and macrophages nourished in an extracellular matrix.The interactions between the tumor and tumor microenvironment are complex and dynamic, which orchestrate cellular or molecular events and ultimately support tumor survival, growth, and metastasis.
## Shukun li is currently a phd student at the institute of process
Engineering (IPE), Chinese Academy of Sciences (CAS), under the supervision of Prof. Yan. She is conducting research on the peptidemodulated self-assembly of photosensitive molecules for tumor theranostics. Wenjia Zhang is currently a postgraduate student at the Radiology Department, Peking Union Medical College Hospital, under the supervisor of Prof. Xue. She is conducting research on both basic development and clinical research into applications of MRI to assess treatment response in cancer as well as multi-modality imaging of animal tumor models. Therefore, the task of fabrication of smart antitumor drugs with adaptive therapeutic efficacy in the tumor microenvironment is daunting.
Advanced nanotechnology has promoted the development of supramolecular self-assembly for construction of nanodrugs.Peptide self-assembly, in harnessing various noncovalent interactions including hydrophobic interactions, electrostatic forces, hydrogen bonds, p-stacking, van der Waals forces and metal coordination bonds, has emerged as an efficient and powerful strategy for fabrication of smart nanodrugs.Manipulating synergistic or reciprocal noncovalent interactions can obtain tailored nanodrugs. Simultaneously, the reversible nature of noncovalent interactions renders the resulting nanodrugs to be dynamic in terms of structures and functions once they respond to various stimuli (pH,GSH, 26-28 enzyme,etc.) in the tumor microenvironment. Such peptide-modulated, tumor microenvironment-oriented adaptive nanodrugs show unique superiorities: (i) with respect to regulation, selecting desired peptide building blocks (hydrophobic, amphiphilic, hydrogen bond-contained, etc.) and changing kinetic parameters (temperature, concentration, solvent, etc.) yield controlled structures in nanoscale that not only enhance the permeability and retention (EPR) effect,but also optimize the therapeutic index of drugs. Specically, the introduction of peptides can modulate the supramolecular photothermal effect of photosensitizers,optimize pharmacokinetic proles of chemotherapeutic drugs,and form a supramolecular self-adjuvant effect in immunotherapy.(ii) Stimuli-responsive triggers broadly exist in the tumor microenvironment and their signicance lie in two aspects. On the one hand, the sensitivity of nanodrugs in the tumor microenvironment is conducive for targeted drug release, permitting optimal therapeutic concentration at tumor sites while eliminating off-target side effects in normal tissues.On the other hand, stimuli in the tumor microenvironment play a switchable role in the morphological reorganization of nanodrugs, which promotes the uptake, the retention, and the efficacy of drugs to a large extent.In particular, integrating targetability and switchability may be a promising alternative for further improving treatment outcome. (iii) Peptides are naturallyoccurring building blocks with controllability, programmability, and biosafety.Owing to the biological capability of peptides fullled in organisms, introducing the biofunction of peptides into nanodrugs is easily achieved.For example, the immune bioactive peptides encompassing specic amino acid sequences that act as antigens, adjuvants and checkpoint blockades, have been recently studied as building blocks to construct nanodrugs for combinational antitumor therapy.This paradigm avoids the addition of ineffective components, simplifying the design and amplifying the therapeutic effect. However, elaboration of the relationship between adaptive assembly and function implementation, intrinsically the dynamic control of noncovalent interactions in the tumor microenvironment, still remains a challenge. Therefore, a perspective on peptide-modulated self-assembled nanodrugs with adaptive therapeutic functions in the tumor Huadan Xue is currently a professor and vice director of the Radiology Department in Peking Union Medical College Hospital. She is also a Doctor Tutor. She has been a resident, chief resident and attending doctor in the Radiology Department since 2003. Dr Xue has broad interests, but her main research eld is focused on diagnostic radiological methods of the digestive system and gynecology system malignancies, including CT, MRI and molecular imaging. microenvironment is imperative to advance the further development of nanodrugs.
In this review, we will focus on the peptide-modulated selfassembly of photosensitizers, chemotherapeutic drugs and immunoactive agents for tumor microenvironment-oriented adaptive (1) phototherapy, (2) chemotherapy, (3) immunotherapy and even (4) combinatorial therapy. The emphasis will put on building block design and intermolecular interaction principle in the tumor microenvironment and the corresponding function implementation, aiming to elucidate the critical roles of peptide-regulated, tumor microenvironment-oriented adaptive assemblies in improving the therapeutic index. Representative examples with tumor microenvironment (pH, GSH, enzyme)-responsiveness based on the above therapies will be presented, providing unique perspectives to advance the development of nanodrugs. Finally, we will lay out challenges and opportunities for the utilization of peptide-modulated, tumor microenvironment-oriented nanodrugs to optimize preclinical tests and accelerate their nal clinical applications.
## Phototherapy
Phototherapy, mainly including photodynamic therapy (PDT) and photothermal therapy (PTT), relies on the excitation of photosensitive molecules through light to generate reactive oxygen species (ROS) or local hyperthermia to kill tumor cells. Intriguingly, photosensitive molecules not only act as therapeutic drugs but also integrate diagnosis function, permitting them to be theranostic agents. Peptide self-assembly has become a facile method for fabrication of photosensitive theranostic nanodrugs. In essence, modulating the photophysical properties of photosensitive molecules by peptide selfassembly determines the therapeutic and imaging modalities. In the Jablonski diagram, the light excited photosensitive molecules decay back to the ground state through three pathways: uorescence emission (uorescence imaging), intersystem crossing to the triplet excited state followed by energy transfer with adjacent oxygen to generate reactive oxygen (ROS) species (PDT), and thermal deactivation via nonradiative relaxation (PTT and photoacoustic imaging). Normally, the excited photosensitive molecules in the molecular state are prone to emit uorescence with accompanying intersystem crossing that generates ROS. Once the photosensitive molecules were modulated into an aggregation state, the absorbed energy competitively dissipated through thermal relaxation. Therefore, the peptide-modulated photosensitive self-assemblies possess a supramolecular photothermal effect that can be applied for photoacoustic imaging and PTT. Combined with tumor microenvironment-oriented drug release or morphological reorganization, the resulting nanodrugs may restore their capabilities of uorescence emission and photodynamic effect, thereby maximizing therapeutic effect.
## Ph
The disordered metabolism of tumor cells contributes to the acidic microenvironment.Compared to the physiological pH 7.4 in blood and normal tissues, more acidic pH values are found in interstitial tumors (pH 6.5-7.2) and endosomes (pH 5.0-6.5), which can be exploited as triggers for photosensitive drug release to ensure the optimal concentration at tumor sites, especially favoring for uorescence imaging and PDT.
Protonation of building blocks occurring in the acidic tumor microenvironment is benecial for targeted drug release. For example, our group 58 co-assembled a short peptide (H-Phe-Phe-NH 2 $HCl, CDP) or an amino acid derivative (9-uorenylmethoxycarbonyl-L-lysine, Fmoc-L-Lys) with chlorin e6 (Ce6) to form photosensitive nanodrugs. The aromatic groups in CDP and the Fmoc groups in Fmoc-L-Lys can interact with the pyrrole rings in Ce6 through hydrophobic interactions or p-p stacking, which promoted the formation of Fmoc-L-Lys/Ce6 nanodrugs and CDP/Ce6 nanodrugs. Compared to free Ce6, higher tumor-selectivity with Fmoc-L-Lys/Ce6 nanodrugs and CDP/Ce6 nanodrugs was obtained due to the EPR effect. Once they reached into the tumor microenvironment, protonation of the carboxyl groups in Ce6 molecules occurred, thus weakening the electrostatic interactions of the building blocks, leading to disassembly of Ce6 molecules, achieving imaging-guided PDT. In another example, Yan and coworkers 59 conjugated a chemotherapeutic drug (mannose) to a pH-sensitive polypeptide copolymer (PE) through formation of Schiff base bonds, obtaining a pH-sensitive mannose-containing copolymer (PM). The photosensitive molecule iodinated boron dipyrromethene (BDPI) was encapsulated to obtain the nanodrug. In the acidic tumor microenvironment, the detachment of the Schiff base bonds and the protonation of the tertiary amine group in PM destroyed the nanostructure, resulting in release of BDPI and mannose for PDT and chemotherapy. Alternatively, acidic pH can cleave some specic bonds to achieve drug release. For instance, hydrazone and benzoic-imine are typical pH-responsive bonds, which are widely utilized in the design of targeted drug release in the tumor microenvironment.Besides targeted drug release, the structural transformation of nanodrugs can also be oriented by the tumor acidic microenvironment. Yu and coworkers 60 reported that pHresponsive isomerization spatially optimized cellular uptake of nanodrugs. In the design, Ce6 was conjugated to a pentapeptide to obtain AmpF-C. The 4-amino-proline (Amp) amide bonds contained in AmpF-C showed pH-responsive isomerization. In solution at pH 7.4, AmpF-C self-assembled into superhelices with a height of approximately 7 nm. Decreasing the solution pH to 6.5 and 5.5 resulted in the cis to trans isomerization of the Amp amide bonds, promoting the transformation of AmpF-C from superhelices to nanoparticles with diameters of 104.1 AE 14.6 nm, and 120.2 AE 17.2 nm, respectively. Such a pH gradient is parallel to that in the tumor microenvironment, implying their potential for endosome/lysosome-involved cellular uptake and PDT. Moreover, other pH-responsive structural transformations are able to optimize delivery processes. For instance, the size decrease that nanodrugs undergo in response to acidic pH in the tumor microenvironment can overcome the diffusion hindrance.The pH-responsive shape switch from spherical nanodrugs to rod-like nanodrugs results in enhancement of the cellular internalization.Surface charge is of signicance in determining the fate of nanodrugs in vivo. An ideal nanodrug should be negatively or neutrally charged during blood circulation as to avoid clearance by the reticuloendothelial system, while it should be positively charged in tumor tissues to enhance the cellular uptake. The "charge dilemma" was eliminated by the strategy of tumor acidity-specic charge reversal nanodrugs developed by Wang and coworkers.Generally, tumor acidity-cleavable maleic acid amide (TACMAA) was incorporated in the design, which was produced by a reaction of an amino group with 2,3-dimethylmaleic anhydride (DMMA). In the acidic tumor microenvironment, nanodrugs containing TACMAA realized their charge switching for cellular uptake enhancement. An expanded molecule of 2-propionic-3-methylmaleic anhydride (CDM) was utilised in the fabrication of the pH-facilitated charge reversal nanodrug, D m -NP.The building block of PEG-Dlink m -R9-PCL encompassed a poly(ethylene glycol) (PEG) corona for prolonging circulation time, a pH-responsive amide bond as a Dlink m linker, and a cell-penetration peptide poly(3-caprolactone)-R9 (PCL-R9), and self-assembled into micelles. Importantly, the R9 in the hydrophilic shell of D m -NP allowed for the encapsulation of siRNA to form micelleplexes, D m -NP siN.C. , with a surface charge of 23.9 mV at pH 7.4. Decreasing the pH to 6.8, the Dlink m linker was cleaved to remove the PEG corona and the PCL-R9 was exposed to transform the surface charge of D m -NP siN.C. into 34.8 mV, which facilitated the targeted cellular uptake in tumor tissues, and achieved superior tumor inhibition activity. Adopting the same tumor acidity induced charge reversal strategy, Yang and coworkers 64 used DMMA to modify amines in TAT lysine residues to avoid clearance during blood circulation. The photosensitizer Ce6 and contrast agent Gd 3+ were encapsulated into the modied TAT and formed DA TAT-NP nanostructures. Once accumulated in the acidic tumor microenvironment, the DMMA groups were detached from DA TAT-NP. The exposed TAT enhanced cell penetration and cellular uptake, resulting in enlarged uorescence and MR signals for imagingguided PDT.
## Gsh
Redox potential is a characteristic stimulus in the tumor microenvironment. The concentration of GSH at tumor sites is fourfold higher than that in normal physiological conditions. Furthermore, the concentration of GSH inside tumor cells is almost three orders of magnitude higher than that in the extracellular plasma. Therefore, the differences between GSH concentrations pave avenues for the design of peptidemodulated, GSH-oriented adaptive nanodrugs.
Taking advantage of the evaluated GSH concentration, disulde bonds become the most common chemical bonds that are incorporated into nanodrugs. Our grouphave developed disulde bonds cross-linked cationic dipeptide (CDP)-based nanodrugs. Electrostatic interactions and hydrophobic interactions promoted the co-assembly of CDP and Ce6. Glutaraldehyde assisted the co-assemblies to form disulde bonds to improve the structural stability. Importantly, GSH in the tumor microenvironment broke the disulde bonds in nanodrugs, achieving Ce6 release for PDT. Similarly, as another example of disulde bonds that respond to GSH in the tumor microenvironment, Tan and coworkers 66 designed a pH (low) insertion peptide (pHLIP) for a pH e -driven targeting conjugate, Ce6-pHLIP ss -AuNRs. pHLIP was a class of peptide that can fold and insert cross membrane at an acidic pH value, 67 leading to a unique way for targeted drug delivery. In the design, a side peptide chain containing cysteine residues was conjugated to the pHLIP by a disulde bond to obtain pHLIP ss . Then, the Ce6-pHLIP ss was synthesized. Finally, the Ce6-pHLIP ss and thiolterminated monomethoxyl poly(ethylene glycol) (mPEG-SH) were assembled on the surface of AuNRs by thiol chemistry, obtaining Ce6-pHLIP ss -AuNRs. AuNRs acted as a photothermal agent and Ce6 acted as a photodynamic agent. The reduction of the disulde bond in the conjugate by extracellular GSH in the tumor microenvironment enabled separation of Ce6 from Ce6-pHLIP ss -AuNRs. Owing to the pH e -driven targeting ability and GSH-driven activated PDT efficacy, the accumulation of Ce6-pHLIP ss -AuNRs was enhanced and a better synergistic efficacy of PDT and PTT was achieved.
Metal ions coordination in living organisms is critical for aspects of structure and function, which guides the fabrication of delicate nanostructures.More recently, our group 7 developed metallo-nanodrugs based on multicomponent metal coordination self-assembly, which is inspired by hemoglobin that integrates porphyrin derivatives and histidine residues via cooperative coordination. A histidine-containing dipeptide (N-benzyloxycarbonyl-L-histidine-L-phenylalanine, Z-HF) and amphiphilic histidine derivative (9-uorenylmethoxycarbonyl-L-histidine, Fmoc-H) were selected as building blocks. Carboxyl group-containing Ce6 was selected as a model photosensitive drug. Zinc ions coordination self-assembly with imidazole rings in the peptide and carboxyl groups in Ce6, together with hydrophobic interactions and p-p stacking, promoted the formation of metallo-nanodrugs, Fmoc-H/Zn 2+ / Ce6 (79 AE 21 nm) and Z-HF/Zn 2+ /Ce6 (76 AE 21 nm). Acidic pH rendered the protonation of imidazole rings and high GSH competitively coordinated with zinc ions within the metallo-This journal is © The Royal Society of Chemistry 2020
Chem. Sci., 2020, 11, 8644-8656 | 8647
## Perspective
Chemical Science nanodrugs. Dual-stimuli jointly weakened the metal coordination interactions of the nanodrugs, which promoted their disassembly in the tumor microenvironment. The disassembled Ce6 molecules were conducive for uorescence imaging and PDT effects. Since structural stability in blood circulation and ultra-sensitive responsiveness in the tumor microenvironment were integrated into the metallo-nanodrugs, tumors in mice injected with the metallo-nanodrugs were eradicated. As another optimization strategy, our group 26,69 introduced the imaging metal ion Mn 2+ that not only regulated structural stability but also bestowed the nanodrugs with magnetic resonance imaging (MRI) properties. Notably, exploitation of GSH cooperative coordination in the tumor microenvironment realized the dynamic disassembly of the nanodrugs. Meanwhile, it eliminated the intracellular GSH that is detrimental to ROS generation.
## Enzyme
Tumor cells overexpress certain enzymes, such as matrix metalloproteinases (MMP), alkaline phosphatases (ALP) and carboxylesterases (CES), which can cleave their corresponding substrates in peptide-based building blocks, leading to enzymeinduced drug release or morphological transformation of selfassemblies. Enzyme-instructed peptide self-assembly was pioneered by Xu and coworkers.A variety of hydrogelator-precursors with distinct enzyme responsiveness were designed and synthesized. Rao and coworkers 72 extended the concept as they employed a bioorthogonal cyclization reaction to control the self-assembly of small molecules and suggested its application in tumor enzyme activity. Wang and coworkers 73 applied enzyme-assisted peptide self-assembly in phototherapy. For example, they designed and synthesized a peptide-photosensitizer conjugate (compound 1) that was composed of three parts: photosensitizer (Purpurin 18, P18), gelatinase-responsive linker PLGVRG, and a targeting ligand RGD that bind to a v b 3 integrin on tumor cell membranes. Compound 1 diffused and extravasated readily in blood circulation. RGD promoted the circulation of compound 1 to the tumor microenvironment, and then the gelatinase overexpressed by tumor cells cleaved the PLGVRG linker. The enhanced hydrophobic interactions and the reduced steric hindrance resulted in the self-assembly of building blocks for the formation of nanobers. Such nanobers exhibited prolonged retention time and led to an enhanced photoacoustic signal and therapeutic efficacy. Zhang and coworkers 74 synthesized another peptide-photosensitizer conjugate for tumortargeting PDT. A cell-penetrating peptide (R 9 GPLGLAGE 8 , ACPP) is sensitive to MMP-2 that is overexpressed in the tumor microenvironment. Protoporphyrin (PpIX), as the photosensitizer, was conjugated to ACPP. In normal tissue, the polyanionic peptide (E 8 ) will block the cell-penetrating function of polycationic CPP (R 9 ) through intramolecular electrostatic interactions. Once they circulated to the tumor site, proteolysis of the oligopeptide linker (GPLGLAG) between R 9 and E 8 occurred, disassociating R 9 -PpIX from E 8 . ACPP-PpIX remained stable in blood circulation, while MMP-2 overexpressed on tumor cells can orient the targeted release of R 9 -PpIX for effective PDT.
Effective PDT requires the photosensitizer to be kept in a molecular state. By contrast, a considerable aggregation state of photothermal nanodrugs is preferred as to achieve thermal relaxation for PTT. A supramolecular effect can be readily realized by aggregating photosensitizer at the tumor site.Joining the self-assembly and tumor microenvironmentoriented disassembly of photosensitizers readily achieves the combination of PTT and PDT. For instance, our group 42 designed and synthesized a phthalocyanine-peptide conjugate (PF). Driven by the hydrophobic interactions and p-p stacking, PF self-assembled into nanoparticles (PF NPs) with a diameter of 54.8 AE 17.6 nm. Simultaneously, supramolecular photothermal effect was generated for PTT. While PF NPs interacted with the hydrophobic domains widely existing on the cell membrane, PF NPs disassembled into molecules for PDT, as evidenced by the restored uorescence intensity of molecular PF conjugates. The adaptive transformation of the phthalocyanine-peptide conjugate in the tumor microenvironment enabled the integration of multiple therapeutic modalities in one component, achieving a superior localized antitumor effect. The strategy is facile and effective, however, improvement in the targeting ability of the PF conjugate should be the next research interest.
## Chemotherapy
Chemotherapy is a traditional therapeutic modality in clinical use. Unfortunately, the treatment efficacy of chemotherapy is severely impeded by the occurrence of acquired multidrug resistance (MDR) in tumor cells and side effects to normal tissues, which are caused by insufficient drug concentration in tumor sites and premature drug release during whole body blood circulation. Peptide-modulated, tumor microenvironment-oriented chemotherapeutic self-assemblies possessing structural controllability, high loading efficiency, extended blood circulation and targeting ability tackled the above problems and thus optimized therapeutic outcome.
## Ph
Chilkoti and coworkers 76 reported a pH-dependent nanodrug that assembled by a chimeric polypeptide-doxorubicin (DOX) conjugate to abolish tumor cells via a single injection. A pH-responsive hydrazone bond linked the polypeptide and DOX. The hydrophobic DOX and hydrophilic polypeptide imparted sufficient amphiphilicity to the conjugates that promoted their self-assembly into a nanodrug. Intriguingly, the hydrazone bond was considerably stable during blood circulation at pH 7.4. While with decreased pH at 5.0 which is relevant to endosome/lysosomal trafficking in the tumor microenvironment, the hydrazone bond was cleaved and DOX was liberated for effective chemotherapy. Additionally, a pH-sensitive switchable nanodrug was reported by Hu and coworkers. 77 DOX molecules were encapsulated into liposomes (LS), and the LS were functionalized by a peptide STP (SKDEEWHKNNFPLSP). In the presence of protons within the acidic tumor microenvironment, the segment of KDEE in STP can form an a helix, which could improve the internalization. STP also acted as an affinity ligand for vessel marker endothelial growth factor receptor-2 (VEGFR-2) on the tumor, hence the formed STP-LS-DOX nanodrugs possessed the dual function of pH-triggered responsiveness and VEGFR-2-oriented targetability. Therefore, the STP-LS-DOX nanodrugs could enhance the antitumor effect.
Normally, the therapeutic efficacy of chemotherapeutic drug is determined by ve steps: including circulation, accumulation, retention, internalization, and release. In this regard, Chen and coworkers 37 developed a hierarchical responsive strategy to enhance the above ve steps of nanodrugs. A peptide conjugate, RGD-PC7A-POEG-PssCPT, was designed. Cyclic RGD peptide was applied as the targeting segment to bind to the tumor membrane. The triblock copolymer included a hydrophobic PC7A chain, hydrophilic poly(oligo-(ethylene glycol) monomethyl ether methacrylate) (POEG) segment and hydrophobic poly reduction-responsive camptothecin (CPT) prodrug (PssCPT) chain. During blood circulation at the neutral pH of 7.4, the RGD-PC7A and PssCPT segments were trapped inside the hydrophobic core and POEG was coated at the surface of the nanodrug, thus the targeting ability of RGD was shielded and the circulation time of the nanodrug was prolonged. While in the tumor microenvironment with decreased pH below 6.8, protonation of PC7A occurred and POEG became positively charged, resulting in exposure of RGD in the nanodrug. RGD oriented the nanodrug to bind on the cell membrane and therefore enhanced its tumor retention and cellular internalization. Furthermore, the high concentration of GSH within cells cleaved the disulde bond in PssCPT, further promoting release of CPT molecules. Such features-packed nanodrugs have demonstrated their targetability, potency and safety in vivo.
Exploiting the toxicity of peptide drugs may be another method to improve the therapeutic index. Cytotoxic peptide (KLAKLAK) 2 (KLAK) is a proapoptotic peptide that can disrupt mitochondria structures and then induce apoptosis. Wang and coworkers 78 synthesized amphiphilic hyperbranched poly(bthioester)s (PPHD-PK) conjugated with KLAK and PEG. In aqueous solution, PPHD-PK self-assembled into nanoparticles, whereas PPHD acted as an inner core for encapsulation of DOX, PEG and KLAK were displayed as the hydrophilic outer shell. Acidity in the tumor microenvironment broke the b-thiopropionate group in PPHD to induce release of DOX and KLAK, leading to potent efficacy for killing tumor cells. As a more precise strategy in vivo, they 79 recently designed a peptide conjugate with the property of acidity-increased hydrophobicity in the tumor microenvironment. KLAK was decorated with an acid-responsive moiety cis-aconitic anhydride (CAA). Cellpenetrating peptide TAT was conjugated to the poly(b-thioester)s backbones. Thus, PT-K-CAA was obtained that was negatively charged and kept as a single chain at pH 7.4, which can penetrate deeply into a tumor. Intriguingly, the hydrolysis of CAA in the tumor microenvironment at pH 6.5 led to increased hydrophobicity of the peptide conjugate, which enforced their self-assembly into nanoparticles. The strategy enhanced internalization efficiency and promoted therapeutic efficacy of KLAK for killing deep solid tumors.
## Gsh
Metal coordination-driven multicomponent self-assembly, as a novel fabrication of nanodrugs, is also applicable in chemotherapy to improve circulation stability and achieve targeted release of chemotherapeutic drugs. Our group 36 demonstrated that curcumin-based nanodrugs can be assembled through the multicomponent coordination interactions of curcumin, zinc ions and metal-binding peptides or amino acids. Intriguingly, the size of nanodrugs can be controlled through regulation of This journal is © The Royal Society of Chemistry 2020
Chem. Sci., 2020, 11, 8644-8656 | 8649
## Perspective
Chemical Science kinetic parameters (tension coefficients of the solvent) or dynamic parameters (concentration of curcumin) in the process of preparation. Concurrently, a high loading efficiency is readily obtained ($52%). Moreover, competitive ligands of GSH can recoordinate with zinc ions, resulting in curcumin release for cellular uptake enhancement and chemotherapy improvement. As another exible method to modulate chemotherapeutic drugs self-assembly, amphiphilic peptide-drug conjugates, named as prodrugs, were developed by Cui and coworkers.The design endowed chemotherapeutic drugs with the dual capabilities of regulation and chemoactivity. More recently, they suggested the signicant role of critical micellization concentration (CMC) in designing and developing chemotherapeutic drug delivery systems.Four prodrugs (SAPDs) were synthesized, which consisted of two hydrophobic CPT molecules and oligoethylene-glycol (OEG)-decorated peptides with various OEG numbers of 2, 4, 6, 8, named as SAPD 1, SAPD 2, SAPD 3, and SAPD 4, respectively. In addition, the CPT and the hydrophilic moiety were connected via a GSH-responsive disulfanylethyl carbonate linker (etcSS). Driven by the directional p-p stacking, self-assembled nanodrugs were obtained. SAPD 1 formed supramolecular laments with a diameter of 9 nm which were several micrometers in length. SAPD 2 formed shorter laments with a length less than 400 nm. SAPDs 3 and 4 both aggregated into micrometer-long nanoribbons of various widths because of the increase in hydrogen bonds supplied by OEG. The CMC values were estimated to be 2.7 and 10.1 mM for SAPDs 1 and 2, respectively, while the CMCs of SAPDs 3 and 4 exceeded 200 mM and cannot be accurately extracted. The results implied the structural stability of the SAPD SPs would be SAPD 1 > SAPD 2 > SAPDs 3 and 4. In the presence of GSH (10 mM), the drug release rate was SAPD 1 < SAPD 2 < SAPDs 3 and 4, which may be attributed to the fact that the more stable the structure, the less the drug release. An in vivo circulation study revealed that SAPD 1 maintained a lower degradation in plasma (86% remained in the bound form) compared to SAPD 2 (28% remained in the bound form) within 5 min, implying the importance of the CMC in determining the morphological and structural integrity of supramolecular self-assemblies in blood circulation. With the enhanced stability and prolonged circulation time, the antitumor effect of SAPD 1 was superior to other groups. The work established a correlation between the CMC of drug molecules and their in vivo performance. Similarly, Cao and coworkers 82 also investigated the important role of the CMC of drug molecules in supramolecular systems for improving targeting and therapeutic outcome.
## Enzyme
Enzyme-instructed drug-peptide amphiphilic hydrogelators can be developed as hydrogel networks for enhancing targetability and uptake of nanodrugs in the tumor microenvironment. Motivated by the principle, Xu and coworkers 83 have developed a tumor intracellular self-assembly strategy harnessing CES to overcome drug resistance. An enantiomeric pair of precursors containing taurine, L-DPT and D-DPT, were synthesized, where the naphthyl capped diphenylalanine at the N-terminal provided self-assembly forces including p-p stacking and hydrogen bonds. Aer the hydrolysis by CES overexpressed in tumor cells, the taurine moiety was detached and the hydrogelators, L-DP and D-DP, were obtained. Then, the hydrogelators self-assembled to form nanobrils, which disrupted the dynamics of actin laments in tumor cells and thus caused their necroptosis or apoptosis. Particularly, compared to normal cells, the inhibition capacity for tumor cells was specic. Additionally, the inhibition of tumor cells was selective, which was positively correlated with the CES activity in various tumor cells. The expression of enzymes, such as CES and ALP, differs in different cell lines. For example, comparing the two cell lines OVSAHO and HepG2, the activities of ALP overexpressed by them are comparable, while the activity of CES in the HepG2 cell line is higher than that in the OVSAHO cell line. The difference prompted Xu and coworkers 84 to couple the assembly-disassembly dynamic process for selectively killing tumor cells that downregulated CES. They designed peptidebased precursors as the substrates of both CES and ALP. The peptide-based precursors can turn into building blocks aer ALP hydrolysis and self-assemble into nanobrils. In tumor cells that upregulated CES, hydrolysis of ester bonds in the building blocks enabled disassembly of the nanobrils, thus being innocuous to the tumor. By contrast, in tumor cells with downregulation or dysfunction of CES, there was no occurrence of hydrolysis of the nanobrils, hence the nanobers caused tumor cell death.
Gianneschi and coworkers 85 exploited the MMP overexpressed in an array of tumor types as a targeting tool for the delivery of chemotherapeutic drugs. They designed a peptidepaclitaxel conjugate, where the peptide is hydrophilic and can respond to MMP. The amphiphilic conjugate can self-assemble into micellar nanodrugs with a peptide shell and drug core due to the hydrophobic interaction. Notably, the peptide shell was cleaved upon exposure to MMP in the tumor microenvironment, then the nanodrugs underwent a drastic change in morphology from discrete, spherical micelles with a diameter of 20 nm to form micrometer scale assemblies, leading to release of PTX and achieving a measurable therapeutic dose for effective chemotherapy. Despite the tremendous successes of MMPresponsive nanodrugs for chemotherapy, however, the responsiveness is still hindered by the accessibility of the MMP to the substrates that are oen embedded inside the assembled structures. Although the resistance of the assembled structures to enzymatic degradation was improved, inevitably, the responsiveness of the assembled nanostructures to MMP is less sensitive than their corresponding monomeric molecules. To address this problem, Cui and coworkers 86 developed a strategy of post-assembly crosslinking to promote the formation of enzyme-substrate complexes and further to facilitate the enzymatic reaction. The synthesized amphiphilic peptide MASP1 self-assembled into laments at pH 4.5, then, MMP-2 specic peptide substrates were utilized to crosslink above the laments. Of note, in the presence of MMP and at pH 7.5, peptide substrates on the lament surface were specically cleaved by MMP, disassociating the cross-linked laments. Integrating the improved stability and enzyme-induced instability at targeted sites, the strategy implied its great potential as drug carriers. From the perspective of molecular design, they 87 demonstrated the signicance of the unsymmetric reverse bolaamphiphiles (RBA) design in exposing the MMP-2 cleavable segments on the lament surface. The reverse peptide bolaamphiphile consists of four parts. A hydrophilic peptide sequence RGDR and an MMP cleavable peptide sequence PLGVR were in the middle, while two structurally different hydrophobic segments, a triple valine sequence and a twelve-carbon alkyl chain, were capped at the C-terminus and N-terminus, respectively. Hydrogen bonds and hydrophobic interactions promoted the RBA to selfassemble into supramolecular laments with MMP-2 cleavable sequence displayed on the surface, and then to entangle into a hydrogel. For the utilization of the design in chemotherapy, paclitaxel was conjugated to the N-terminus instead of the alkyl tail. Furthermore, a reducible linker buSS was used to link the paclitaxel and peptide. In a tumor microenvironment that overexpressed MMP-2 and enriched GSH, the supramolecular hydrogel was degraded into fragments and then into monomers, leading to the drug release for effective chemotherapy.
Enzyme-triggered morphology transitions are implicated in biological functions. For instance, MMP-2 triggered ber-tomicelle morphological transition of self-assemblies enhanced their cell-penetrating ability as reported by Azevedo and coworkers.Cell-penetrating peptide amphiphile (CPPA) underwent self-assembly into nanobers with long circulation times in blood. Upon accumulation in the tumor microenvironment, MMP-2 detached the sensitive linker, exposing the previously hidden CPP sequences on the surface of the assemblies and enhancing their cell-penetrating process. On the contrary, micelle-to-ber morphological transitions studied by Ulijn and coworkers 89 demonstrated their potential for sustainable chemotherapeutic drug release. A peptide amphiphile with MMP-9 responsiveness was synthesized and then self-assembled into spherical aggregates. DOX can be encapsulated into the hydrophobic inner core. Aer MMP-9 hydrolysis, the ber forming moiety remained and re-assembled into nanobers, providing a depot for sustainable drug release.
## Immunotherapy
The clinical successes of checkpoint blockade antibodies have revolutionized immunotherapy for treatment of tumors. In contrast to traditional therapy, immunotherapy aims to boost the host immune system and then kill tumor cells in a safe and effective manner with minimal side effects. Peptide-based, tumor microenvironment-oriented self-assemblies have attracted tremendous attention in the respect of vaccine design and immunosuppressive microenvironment remedy.
This journal is © The Royal Society of Chemistry 2020
Chem. Sci., 2020, 11, 8644-8656 | 8651 Perspective Chemical Science
## Vaccine design
Modulating noncovalent interactions of peptide-based selfassemblies allows for formation of immune nanodrugs showing supramolecular self-adjuvant effects, which promoted antitumor immunotherapy. The pioneer work was reported by Collier and coworkers.Model antigen OVA 323-339 was attached to a domain of Q11 (QQKFQFQFEQQ) to obtain peptide-based building blocks, OVA-Q11. Then, the OVA-Q11 self-assembled into nanobers. In vivo results demonstrated that nanobers increased the population of IgG titers in serum, thereby enhancing the antitumor immunotherapy. Further studies suggested that the immune response relied on conjugation of OVA 323-339 and Q11. Li and coworkers 91 also used the Q11 domain to conjugate epitopes. MUC1-derived epitopes with varied glycosylated threonine residues were conjugated to Q11 for the design of building blocks. Epitope-Q11 conjugates selfassembled into brous self-adjuvant vaccines. These strategies provided a novel vaccine candidate with a simple and chemically dened formulation. Our group 92 also reported an antitumor immune hydrogel through electrostatic coupling between poly-L-lysine (PLL) and a dipeptide derivative (Fmoc-FF). The formed helical bril hydrogels without addition of antigens, immune regulatory factors, and adjuvants, showed a promising perspective for the development of immune responsive nanodrugs for antitumor therapy. Distinct from the methods mentioned above, Yang and coworkers 93 created supramolecular hydrogels by virtue of enzymeassisted self-assembly, and further suggested their self-adjuvant function. In detail, the enantiomeric effect was investigated by the synthesis of peptide gelators Nap-GFFpY-OMe and Nap-G D F D F D pY-OMe. Both of them can co-assemble with OVA for the formation of hydrogels upon ALP catalysis. Effective cellular cytotoxic immune responses of the two enantiomeric hydrogel adjuvants were veried. Particularly, D-peptide based hydrogels possessed superior accumulation in lymph nodes, thereby preventing tumor growth effectively. The work provided a facile strategy to administer vaccine adjuvants with biosafety.
## Immunosuppressive microenvironment remedy
Coupling supramolecular self-assembly and the acidic tumor microenvironment may promote the release of immunoactive molecules, remedying the immunosuppressive microenvironment. For example, Xie and coworkers 94 synthesized responsive exosome (Exo) bioconjugates for immunotherapy. Firstly, the macrophages were polarized to anti-tumoral 1 macrophages (M1) and the cell membranes were modied with azide groups. Then, the dibenzocyclooctynes (DBCO)-modied anti-CD47 antibody (aCD47) and anti-signal regulatory protein alpha antibody (aSIRPa) were conjugated to azide-modied M1 Exo linked by pH-sensitive benzoic-imine bonds. aCD47 oriented the M1 Exo to target tumor by bonding to CD47 on tumor cell membranes while the acidic microenvironment cleaved the benzoic-imine bonds to further release aSIRPa and aCD47. The inhibitory receptors of SIRPa and CD47 were individually blocked, thus improved the phagocytosis of macrophages.
Simultaneously, M1 Exo re-educated the immune-suppressive M2 to immune-promotive M1 for tumor immunotherapy.
Other stimuli were also exploited to design smart nanodrugs for tumor immunotherapy. Liu and coworkers 95 designed a GSH-responsive nanodrug. A positively charged cellpenetrating peptide with the sequence CWWR 8 CR 8 CR 8 C was synthesized and co-assembled with negatively charged OVA antigen upon electrostatic interactions. The oxidization of the thiols in the cysteine residues promoted the formation of disulde bonds that cross-linked the structure to obtain stable peptide/OVA nanodrugs. Owing to the cell-penetrating ability of the peptide, the antigen uptake of the peptide/OVA nanodrugs was increased. Aer internalization, the antigen was released in the cytoplasm since the GSH cleaves the disulde bonds, resulting in potent CD8 + T cell immunity. Alternatively, as an example based on enzyme-responsiveness, Nie and coworkers 52 fabricated an immune nanodrug based on co-assembly of an amphiphilic peptide and NLG919, an inhibitor for immunosuppressive indoleamine 2,3-dioxygenase (IDO). The amphiphilic peptide contained a 3-diethylaminopropyl isothiocyanate (DEAP) segment, a PLGLAG domain with responsiveness to MMP-2, and a short D-peptide antagonist ( D PPA-1). The resulting nanodrug (NLG919@DEAP-D PPA-1) sequentially responded to the tumor microenvironment. The acidic pH protonated the DEAP molecules and MMP-2 cleaved the PLGLAG segments, releasing D PPA-1 and NLG919 for immunotherapy. For the testing of the antitumor efficacy, a similar amphiphilic peptide with a scrambled amino acid sequence (DEAP-D PPA-1-Scr) was synthesized as a negative control group. Compared to other groups, the NLG919@DEAP-D PPA-1 nanodrug showed superior antitumor efficacy. As expected, the relative proportions of CD8 + T cellsand IFN-g-producing cytotoxic T cellswere elevated. Consistently, higher productions of IFN-g and IL-2 were obtained. In particular, the survival rates of mice were prolonged as well. These results demonstrated that the co-delivery of DEAP-D PPA-1 and NLG919 can trigger robust antitumor immune response.
## Combinational therapy
Monotherapy is normally insufficient to eradicate tumors because of metastasis and recurrence of tumor cells. Combinational therapy, that integrates multiple therapeutic agents into one platform for enhanced antitumor efficacy, is of particular interest.With peptide-modulated, tumor microenvironment-oriented self-assemblies for combinational tumor therapy, the synergistic therapeutic effect can be realized.
The promoted effect may be optimized by integrating therapeutic modalities.Hyperthermia caused by PTT can accelerate the blood circulation and increase oxygen perfusion, beneting oxygen-dependent therapeutic modalities, such as PDT and radiotherapy. 14 Intriguingly, phototherapies not only kill primary tumor cells, but also generate tumor debris for promoting an antitumor immune response, named "immunogenic cell death (ICD)". ICD is associated with the release of damage-associated molecular patterns that are benecial for antigen uptake and subsequent immune system activation.For instance, Qian and coworkers 102 constructed multifunctional nanodrugs with the abilities of tumor-targeting, penetration enhancement, targeted drug release behavior, and immunomodulation mediated by checkpointed blockade (PD-1/ PD-L1). Nanodrugs were co-assembled from a chemotherapeutic drug (docetaxel, DTX), photothermal agent (IR820) and a predesigned peptide with 27 amino acids (CF27). CF27 functioned as a cross-linker in the nanodrug fabrication to yield high drug loading efficiency. CF27 is responsive to MMP-2 and GSH in the tumor microenvironment, Signicantly, CF27 contained a peptide sequence for blocking the PD-L1 immunosuppressive signal. The resulting nanodrug integrated photothermal therapy, chemotherapy and immunotherapy, strikingly inhibiting the tumor growth and metastasis.
Another typical example of combinational therapy was recently reported by Cui and coworkers.They utilized a drug-based supramolecular prodrug to achieve locally delivered immune checkpoint blockers (ICBs) for combination chemoimmunotherapy. By conjugating hydrophilic iRGD to two hydrophobic CPT molecules through an MMP-2 responsive peptide linker PLGLAG, the amphiphilic prodrug, diCPT-PLGLAGiRGD, was synthesized. In order to attach the CPT moiety and PLGLAG sequence, a reducible etcSS linker was applied to endow the amphiphilic prodrug with GSH responsiveness. The prodrug selfassembled into supramolecular nanotubes (P-NTs) in aqueous solution. Introducing PBS into the solution promoted the quick solgel transition for the formation of hydrogel due to the charge screening effect from the counterions. Simultaneously, by mixing the antibody of aPD1 with P-NTs, the P-NT-aPD1 hydrogel was obtained. In vivo results also demonstrated that localized delivery of aPD1 and a P-NT solution resulted in the formation of P-NT-aPD1 hydrogel within the injection site in mice. MMP-2 cleavage and GSH reduction in the tumor microenvironment promoted the degradation of P-NT-aPD1 hydrogel. Hence, the formed hydrogel played the role of a therapeutic reservoir for extended tumoral release of CPT and aPD1 which increased the frequency of T effs and reduced the population of immune suppressor cells, thereby provoking a long-term immune response against tumor. Combined with the chemotherapeutic drug CPT, the tumor regression was regressed, and the tumor recurrence and metastasis were inhibited.
# Conclusion
In summary, we have highlighted the recent advances of peptide-modulated, tumor microenvironment-oriented adaptive nanodrugs for improved tumor therapy. Programmable or functional peptides serving as building blocks regulate the selfassembly of drug molecules through multiple noncovalent interactions for fabrication of nanodrugs. Ubiquitous physiological variations in the tumor microenvironment, such as decreased pH, increased GSH and overexpressed enzymes, play the trigger roles to promote drug release or induce morphological transformation of nanodrugs and thereby improve the drug bioavailability and optimize the treatment outcome.
Notwithstanding the giant successes of peptide-modulated, tumor microenvironment-oriented adaptive nanodrugs that have been gained, the clinical translation remains a great challenge. Firstly, with peptides acting as regulating blocks, some chemical modications, such as introducing amphiphilic moieties, are inevitable. Although structural stability has been realized, concerns about immunogenicity and biosafety have been raised. Secondly, the tumor entity is heterogeneous, which determines the complexity of interactions between nanodrugs and tumor cells. Understanding the tumor microenvironment by virtue of several physiological parameters is oversimplied. Thirdly, an ideal adaptive nanodrug requires considerable stability in blood circulation and ultra-sensitive responsiveness in the tumor microenvironment, which is still a formidable challenge for peptide-modulated self-assembly and other supramolecular self-assembly techniques. Regarding the challenges mentioned above, further improvements of nanodrugs are suggested as follows. First and foremost, to guarantee the functionality and biosafety, it is of signicance to screen peptides with intrinsic bioactivity, explicit metabolic mechanism and multiple noncovalent interaction-donating groups.Besides, real-time monitoring of dynamic interactions between nanodrugs and tumor cells is important and urgent, which calls for in-depth understanding of tumor microenvironments and self-assembly mechanisms. Novel agents, methods and techniques should be developed to indicate the noncovalent interactions in vitro and in vivo. Last but not least, the stability and the responsiveness of nanodrugs should be optimized. Noncovalent interactions can be enhanced and manipulated by modulating thermodynamic or kinetic parameters. For example, applying metal coordination to stabilize the structural integrity in blood circulation is practical.Integrating complementary triggers into one platform improves the responsive sensitivity of nanodrugs in the tumor microenvironment. Combining and exploiting intracellular dysfunctional signals, such as ROS, 106 may improve the responsive sensitivity on the site of interest and enhance the specicity of tumor inhibition. Overall, the collaboration of researchers with multidisciplinary backgrounds is needed to construct smart nanodrugs and accelerate their clinical applications.
## Conflicts of interest
There are no conicts to declare. |
Universal Hepatitis B Vaccination in Adults Aged 19–59 Years: Updated Recommendations of the Advisory Committee on Immunization Practices — United States, 2022
[bib_ref] Prevention of hepatitis B virus infection in the United States: recommendations of..., Schillie [/bib_ref]
# Background
Hepatitis B is a vaccine-preventable, communicable disease of the liver caused by HBV. HBV is transmitted through percutaneous (i.e., puncture through the skin) or mucosal (i.e., direct contact with mucous membranes) exposure to infectious blood or body fluids. Since HepB vaccine was introduced in 1982, the number of reported hepatitis B cases has declined substantially. However, despite reductions in hepatitis B incidence during the past 4 decades, which were achieved through incremental expansion of groups for whom HepB vaccination is recommended, progress in recent years on further reducing acute hepatitis B cases has stalled (3). Incident hepatitis B declined from 26,654 reported cases (172,700 estimated actual cases) in 1985 to a low of 2,791 reported cases (18,100 estimated actual cases) in 2014 [bib_ref] Estimating acute viral hepatitis infections from nationally reported cases, Klevens [/bib_ref]. In 2019, a total of 3,192 cases of acute hepatitis B were reported to CDC, corresponding to 20,700 estimated acute infections (95% CI . The most commonly reported risk behaviors and exposures were injection drug use (35%), multiple sex partners (23%), and surgery (10%), followed by other sexual and bloodborne risk behaviors; risk behavior and exposure information were missing for 37.1% of cases. There are an estimated 880,000 (95% CI = 580,000-1,170,000) prevalent chronic HBV infections in the United States based on 2013-2018 National Health and Nutrition Examination Survey data, with a modeled estimate of 1.89 million (range = 1.49-2.40 million) that accounts for potential underrepresentation of the non-U.S.-born population [bib_ref] An updated assessment of chronic hepatitis B prevalence among foreign-born persons living..., Wong [/bib_ref]. In 2018, the reported HepB vaccination coverage (≥3 doses) was 30.0% among adults aged ≥19 years, only a small increase over the past 4 decades (7).
# Methods
During September 2019-October 2021, the ACIP* Hepatitis Work Group † (Work Group) held monthly conference calls to review and discuss scientific evidence relevant to the use of HepB vaccines in a universal adult vaccination recommendation. The Work Group identified the following outcomes of interest for evaluation: incidence of hepatitis B, morbidity related to hepatitis B, mortality related to hepatitis B, and vaccine-related serious adverse events. Data on universal HepB vaccination outcomes and safety were summarized based on findings from a systematic review of the literature completed on
## Summary of key findings
The scientific literature was searched through a systematic review using PubMed, Medline, Embase, CINAHL, and Cochrane Library databases from January 1, 2006, through September 10, 2020. Search terms included "hepatitis b vaccines," "adult," "routine," and "universal." To qualify as a candidate for inclusion in the review, a study had to discuss adult HepB vaccination. Studies were excluded if they did not address the adult population, were non-English language, discussed HepB vaccines not licensed in the United States, or if data could not be abstracted. The search identified 3,226 studies, 263 of which were deemed eligible and informed this review. Rates of reported acute hepatitis B have not notably decreased for over 1 decade, with 20,700 estimated infections in . None of the identified studies reported hepatitis B incidence, morbidity, and mortality when comparing universal and risk-based adult HepB vaccination. The safety of single-antigen 3-dose HepB vaccines has been established [bib_ref] Prevention of hepatitis B virus infection in the United States: recommendations of..., Schillie [/bib_ref]. PreHevbrio was approved by FDA in 2021 and recommended by ACIP in 2022. Little or no difference in seroprotection or occurrence of serious adverse events or mild adverse events (GRADE evidence type 3; low certainty evidence) was found for PreHevbrio in comparison with a 3-dose, singleantigen vaccine (Engerix-B), and serious adverse events were rare for both vaccines. The 2-dose HepB vaccine (Heplisav-B) was approved by FDA in 2017 and recommended by ACIP in 2018. No difference in occurrence of serious adverse events (GRADE evidence type 1; high certainty evidence) was found for Heplisav-B compared with a 3-dose vaccine (Engerix-B), and serious adverse events were rare for both vaccines [bib_ref] Recommendations of the Advisory Committee on Immunization Practices for use of a..., Schillie [/bib_ref].
## Rationale for recommendations
Approximately one half of acute hepatitis B cases reported in 2019 occurred among persons aged 30-49 years . The number of cases of acute hepatitis B has increased among adults aged ≥40 years, particularly among those aged 40-49 years, for whom the rate of reported cases increased from 1.9 per 100,000 population in 2011 to 2.7 per 100,000 population in 2019 . The rate among adults aged 50-59 years increased 45.5% during the same period (from 1.1 to 1.6 per 100,000 population) and accounted for 22.2% of reported cases in 2019. Acute HBV infections among adults leads to chronic hepatitis B disease in an estimated 2%-6% of cases.
HepB vaccination coverage among adults aged ≥19 years is low. In 2018, self-reported HepB vaccination coverage (≥3 doses) among adults aged ≥19 years was 30.0% (7). HepB vaccination coverage (≥3 doses) was 40.3% for adults aged 19-49 years and 19.1% for adults aged ≥50 years. During 2013-2018, 21.4% (95% CI = 20.2%-22.6%) of adults aged ≥25 years had vaccine-induced immunity to hepatitis B (5). HepB vaccination coverage among adults with risk factors has been suboptimal. In 2018, self-reported coverage (≥3 doses) was 33.0% among adults with chronic liver disease, 38.9% among travelers to countries where HBV infections have been endemic since 1995, 33.0% among adults with diabetes aged 19-59 years, and 67.2% among health care personnel (7). In a national survey of 433 family medicine physicians and 420 internal medicine physicians to assess their barriers to adult HepB vaccination, 68% of physicians cited patients' nondisclosure of risk factors as a barrier, and 44% felt there was inadequate time to routinely assess patients for risk factors [bib_ref] Physician practices regarding adult hepatitis B vaccination: a national survey, Daley [/bib_ref].
A universal recommendation for HepB vaccination could increase the number of persons who receive vaccination before the onset of chronic liver disease and other comorbidities (e.g., obesity or diabetes) that might make vaccination less effective. For example, patients with chronic liver disease are known to have decreased immune response to HepB vaccination [bib_ref] Safety and efficacy of hepatitis B vaccination in cirrhosis of liver, Roni [/bib_ref].
Among the 3,192 case reports of acute hepatitis B received by CDC for 2019, risk behavior and exposure data were missing for 1,183 (37.1%). Risk factors assessed under prior recommendations for HepB vaccination include potential criminal or stigmatizing behavior (e.g., injection-drug use, incarceration, or multiple sex partners), limiting the effectiveness of provider risk assessment [bib_ref] Hepatitis B vaccination uptake in hard-to-reach populations in London: a cross-sectional study, Taylor [/bib_ref] [bib_ref] A blind spot? Confronting the stigma of hepatitis B virus (HBV) infection-a..., Mokaya [/bib_ref]. A universal vaccination recommendation eliminates the need for risk assessment before vaccination.
Racial and ethnic disparities exist among those who become infected with HBV. In 2005, acute hepatitis B incidence among non-Hispanic Black Americans was approximately twice that among several other racial and ethnic populations (3). In 2019, the rate of HBV infection among non-Hispanic Black adults was triple that of Asian or Pacific Islander adults and approximately twice that of Hispanic adults (3). Rates of hepatitis B among children and adolescents of all races and ethnicities converged to a lower rate after a universal vaccination strategy was implemented for this age group (3).
## Resource use
An economic model was used to estimate the health improvements that are expected to result from universal adult HepB vaccination [bib_ref] Assessing the cost-utility of universal hepatitis B vaccination among adults, Hall [/bib_ref]. One measure of cost-effectiveness, the incremental cost-effectiveness ratio (ICER), was calculated at $153,000 per quality-adjusted life-year (QALY) gained for all adults aged ≥19 years. A sub-analysis performed for adults aged 19-59 years yielded an ICER of $117,000 per QALY gained. § Increased vaccination coverage resulting from the modeled vaccination intervention strategies resulted in better § https://www.cdc.gov/vaccines/acip/meetings/downloads/slides-2021-11-2-3/02-HepWG-weng-508.pdf health outcomes; the average QALYs gained, life-years gained, number of acute HBV infections averted, and number of hepatitis B-related deaths averted all increased as vaccination coverage in the intervention strategy increased [bib_ref] Assessing the cost-utility of universal hepatitis B vaccination among adults, Hall [/bib_ref]. Among the cohort aged ≥60 years, hepatitis B incidence is markedly lower (0.6 cases per 100,000 population in 2019); thus, the number of preventable HBV infections in that age group is lower than for those aged 19-59 years.
## Recommendations
HepB vaccination is recommended for adults aged 19-59 years and adults aged ≥60 years with risk factors for hepatitis B. Adults aged ≥60 years without known risk factors for hepatitis B may also receive HepB vaccines (Box). Infants and all other persons aged <19 years are already recommended to receive HepB vaccines (2).
## Clinical guidance
ACIP recommends that adults aged 19-59 years and adults aged ≥60 years with risk factors for hepatitis B should receive HepB vaccines, and that adults aged ≥60 years without known risk factors for hepatitis B may receive HepB vaccines. In previous HepB vaccine recommendations, providers were advised to administer HepB vaccine to all patients who requested it. The new language for adults aged ≥60 years without known risk factors is intended to prompt all providers to offer HepB vaccination to patients in that cohort, rather than wait for a patient to request vaccination, thus shifting the responsibility of initiating the consideration of HepB vaccination from the patient to the provider.
Persons who have completed a HepB vaccination series at any point or who have a history of HBV infection should not receive additional HepB vaccination, although there is no evidence that receiving additional vaccine doses is harmful. ¶ However, there are cases where revaccination might be indicated as specified in the 2018 ACIP recommendation (e.g., nonresponder infants born to persons testing positive for hepatitis B surface antigen [HBsAg], health care providers, and persons on hemodialysis) [bib_ref] Prevention of hepatitis B virus infection in the United States: recommendations of..., Schillie [/bib_ref]. Providers should only accept dated records as evidence of HepB vaccination. Vaccination of persons immune to HBV infection because of current or previous infection or HepB vaccination does not increase the risk for adverse events. However, in settings in which the patient population has a high rate of previous HBV infection,** prevaccination testing, which may be performed concomitantly with administration of the first dose of vaccine, might reduce costs by avoiding complete vaccination of persons who are ¶ https://www.cdc.gov/vaccines/hcp/acip-recs/general-recs/index.html ** https://cdafound.org/polaris/ (Accessed November 19, 2021).
## Box. persons recommended to receive hepatitis b vaccination
## All infants
## Persons aged <19 years
## Adults aged 19-59 years
Adults aged ≥60 years with risk factors for hepatitis B:
- Persons at risk for infection by sexual exposure ï Sex partners of persons testing positive for HBsAg ï Sexually active persons who are not in a long-term, mutually monogamous relationship (e.g., persons with more than one sex partner during the previous 6 months) ï Persons seeking evaluation or treatment for a sexually transmitted infection ï Men who have sex with men - Persons at risk for infection by percutaneous or mucosal exposure to blood ï Persons with current or recent injection drug use ï Household contacts of persons testing positive for HBsAg ï Residents and staff members of facilities for persons with developmental disabilities ï Health care and public safety personnel with reasonably anticipated risk for exposure to blood or blood-contaminated body fluids ï Persons on maintenance dialysis, including incenter or home hemodialysis and peritoneal dialysis, and persons who are predialysis ï Persons with diabetes at the discretion of the treating clinician - Others ï International travelers to countries with high or intermediate levels of endemic hepatitis B virus infection (HBsAg prevalence of ≥2%) ï Persons with hepatitis C virus infection ï Persons with chronic liver disease (including, but not limited to, persons with cirrhosis, fatty liver disease, alcoholic liver disease, autoimmune hepatitis, and an alanine aminotransferase or aspartate aminotransferase level greater than twice the upper limit of normal) ï Persons with HIV infection ï Persons who are incarcerated
## Adults aged ≥60 years without known risk factors for hepatitis b may receive hepatitis b vaccines
Abbreviation: HBsAg = hepatitis B surface antigen.
## Summary
What is already known about this topic?
Vaccination with hepatitis B (HepB) vaccines shows wellestablished safety and efficacy. However, because of risk factor-based approaches of previous vaccination recommendations, coverage among adults has been suboptimal.
What is added by this report?
In addition to groups for whom HepB vaccination is already recommended, the Advisory Committee on Immunization Practices recommends that all adults aged 19-59 years should receive HepB vaccines.
What are the implications for public health practice? Universal adult HepB vaccination through age 59 years removes the need for risk factor screening and disclosure and could increase vaccination coverage and decrease hepatitis B cases. already immune. Prevaccination testing consists of testing for HBsAg, antibody to HBsAg (anti-HBs), and antibody to hepatitis B core antigen (anti-HBc). The presence of HBsAg indicates current HBV infection. The presence of anti-HBs is generally interpreted as indicating immunity, either from HepB vaccination after a complete series or after recovery from HBV infection. The presence of total anti-HBc indicates previous or ongoing infection with HBV. Detailed interpretations of serologic markers for HBV infection are available [bib_ref] Prevention of hepatitis B virus infection in the United States: recommendations of..., Schillie [/bib_ref]. Lack of access to serologic testing should not be a barrier to vaccination of susceptible persons, especially in populations that are difficult to reach. Testing is not a requirement for vaccination, and in settings where testing is not feasible, vaccination of persons recommended to receive the vaccine should continue [bib_ref] Prevention of hepatitis B virus infection in the United States: recommendations of..., Schillie [/bib_ref].
The safety and effectiveness of Heplisav-B and PreHevbrio have not been established in adults on hemodialysis . Data are not available to assess the effects of Heplisav-B and PreHevbrio on the breastfed infant or on milk production and excretion. Data on Heplisav-B and PreHevbrio are currently insufficient to inform vaccine-associated risks in pregnancy. Thus, providers should vaccinate pregnant women needing HepB vaccination with Engerix-B, Recombivax HB, or Twinrix. Abbreviations: ACIP = Advisory Committee on Immunization Practices; HepA = hepatitis A; HepB = hepatitis B. * If the HepB vaccination schedule is interrupted, the series does not need to be restarted. If a 3-dose series is interrupted after the first dose, the second dose should be administered as soon as possible; the second and third doses should be separated by an interval of ≥8 weeks. If only the third dose has been delayed, it should be administered as soon as possible. The final dose of a 3-dose series must be administered ≥8 weeks after the second dose and ≥16 weeks after the first dose; the minimum interval between the first and second doses is 4 weeks. Inadequate doses of hepatitis B vaccine or doses received after a shorter-than-recommended dosing interval should be readministered, using the correct dosage or schedule. Vaccine doses administered ≤4 days before the minimum interval or age are considered valid. Because of the unique accelerated schedule for Twinrix (https://www.fda.gov/media/119351/download), the 4-day guideline does not apply to the first 3 doses of this vaccine when administered on a 0-day, 7-day, 21-30-day, and 12-month schedule. PreHevbrio (https://www.fda.gov/media/154561/download) is a three-antigen HepB vaccine approved by the Food and Drug Administration in 2021 and recommended by ACIP in 2022. † A 2-dose schedule of Recombivax HB adult formulation (10 μg) (https://www.fda.gov/media/74274/download) is licensed for children and adolescents aged 11-15 years. When scheduled to receive the second dose, persons aged ≥16 years should be switched to a 3-dose series, with doses 2 and 3 consisting of the pediatric formulation administered on an appropriate schedule. § Engerix-B (https://www.fda.gov/media/119403/download) for adults on hemodialysis and is administered as a series of 4 doses (2 mL each) as a single 2-mL dose or as two 1-mL doses on a 0-, 1-, 2-, and 6-month schedule. Recombivax HB for adults on dialysis is a 3-dose series. ¶ The safety and effectiveness of Heplisav-B and PreHevbrio have not been established in adults on hemodialysis. Data are not available to assess the effects of Heplisav-B and PreHevbrio on breastfed infants or on maternal milk production and excretion. Data on Heplisav-B (https://www.fda.gov/media/108745/download) and PreHevbrio are currently insufficient to inform vaccine-associated risks in pregnancy. Thus, providers should vaccinate pregnant persons needing HepB vaccination with Engerix-B, Recombivax HB, or Twinrix.
[fig] FIGURE.: Rates of reported acute hepatitis B virus infection, by age group -United States, 2004Source: https://www.cdc.gov/hepatitis/statistics/2019surveillance/Figure2.4.htm [/fig]
[table] TABLE: Recommended doses and schedules of hepatitis B vaccine for adults aged ≥18 years and persons aged 11-19 years, by vaccine type and age group* HepB vaccine*/Age group, yrsDose (μg) Volume (mL) Schedule [/table]
|
Childhood undernutrition in three disadvantaged East African Districts: a multinomial analysis
Background: Undernutrition is an important public health indicator for monitoring nutritional status and survival. In spite of its importance, undernutrition is a significant problem health problem in many East African communities. The aim of this study was to identify factors associated with childhood undernutrition in three disadvantaged East African Districts. Methods: We examined data for 9270 children aged 0-59 months using cross-sectional survey from Gicumbi District in Rwanda, Kitgum District in Uganda and Kilindi District in Tanzania. We considered the level of undernutrition (stunting, wasting and underweight) as the outcome variables with four ordinal categories (severely undernourished, moderately undernourished, mildly undernourished, and nourished). Generalized linear latent and mixed models (GLLAMM) with the mlogit link and binomial family that adjusted for clustering and sampling weights were used to identify factors associated with undernutrition among children aged 0-59 months in three disadvantaged East African Districts. Results: After adjusting for potential confounding factors, the odds of a child being stunted were higher in Gicumbi District in Rwanda while the odds of a child being wasted and underweight were higher in Kitgum District in Uganda. Having diarrhoea two weeks prior to the survey was significantly associated with severe undernutrition. Wealth index (least poor household), increasing child's age, sex of the child (male) and unavailability of water all year were reported to be associated with moderate or severe stunting/wasting. Children of women who did not attend monthly child growth monitoring sessions and children who had Acute Respiratory Infection (ARI) symptoms were significantly associated with moderate or severe underweight. Conclusions: Findings from our study indicated that having diarrhoea, having ARI, not having water availability all year and not attending monthly child growth monitoring sessions were associated with undernutrition among children aged 0-59 months. Interventions aimed at improving undernutrition in these disadvantaged communities should target all children especially those children from households with poor sanitation practices.
# Background
Undernutrition is a major public health problem and an important health indicator for monitoring nutritional status and survival of children under-5 years in many developing countries around the world. Measures of childhood undernutrition are used to track development progress and socioeconomic inequalities in many low and middle-income countries. Suboptimal nutrition in the first 1000 days of life could lead to impaired physical development, which has a long-term impact on cognitive ability thus resulting in reduced educational performance and economic productivity in adulthood [bib_ref] Contextualising complementary feeding in a broader framework for stunting prevention, Stewart [/bib_ref]. Undernutrition lowers immunity thereby predisposing a child to the higher risk of infections, as well as increases the frequency and severity of such infections, and also delay recovery. The relationship between undernutrition and infection creates a vicious cycle of worsening illness and deteriorating nutritional status as undernutrition can make a child more susceptible to infection, and infection also contributes to undernutrition [bib_ref] The interaction between nutrition and infection, Katona [/bib_ref].
Sub-Saharan Africa bears one of the highest burdens of undernutrition. In 2016, more than one-third of stunted children (38%) and more than one-quarter of wasted (27%) children lived in sub-Saharan Africa. However, a more detailed look into the distribution of undernutrition within sub-Saharan Africa shows that Eastern Africa (36.7%) has a higher prevalence of stunting compared to Western Africa (21.4%), Central Africa (32.5%), and Southern Africa (28.1%). While Western Africa (8.5%) has a higher rate of wasting than Central Africa (7.3%), Southern Africa (5.5%), and Eastern Africa (6.5%), these estimates reveal regional disparities in the distribution of undernutrition, thus the need to identify region-specific factors contributing to the distribution of undernutrition across the regions.
The predictors of childhood undernutrition are well-researched and could be classified as either proximate or distal, and affect the nutritional status of children at different levels [bib_ref] Malnutrition among children under the age of five in the Democratic Republic..., Kandala [/bib_ref] [bib_ref] Stunting and severe stunting among children under-5 years in Nigeria: a multilevel..., Akombi [/bib_ref]. Proximate factors operate at the individual and child level which includes age, sex, birth size, birth order, birth interval and infections [bib_ref] Multilevel analysis of factors associated with wasting and underweight among children underfive..., Akombi [/bib_ref] [bib_ref] Prevalence and predictors of undernutrition among infants aged six and twelve months..., Medhin [/bib_ref]. The distal factors include a wider range of conceptual factors within the socio-cultural, economic, environmental, climatic and political context which influence food (in) security, sanitation, access to health care services and education at the household and community level. Studies have reported that these factors tend to vary spatially depending on geographic location and climatic conditions. Previous studies have reported a relationship between undernutrition and geographical region, clearly acknowledging that a child's geographic location is an important modifier of known determinants of undernutrition [bib_ref] Malnutrition among children under the age of five in the Democratic Republic..., Kandala [/bib_ref] [bib_ref] Multilevel analysis of factors associated with wasting and underweight among children underfive..., Akombi [/bib_ref]. However, despite the huge body of evidence on the relationship between undernutrition and geographical region, there is limited evidence of analysis of the factors associated with childhood undernutrition across communities within highly burdened countries in the same sub-region in order to identify the most consistent factors. Given the high rate of undernutrition in Eastern Africa, this study aims to identify the specific factors associated with childhood undernutrition in three East African Districts in order to drive targeted interventions within these communities. With a decrease in undernutrition within communities, there will be a corresponding decrease at the national and sub-regional level thus setting Eastern Africa on the path to achieving the World Health Organization (WHO) global nutrition target by 2025.
# Methods
## Study area
Gicumbi district is situated in the Northern Province of Rwanda and has a population of 395,606 residents. Gicumbi District covers 21 sectors, 109 cells, 630 villages (Imidugudu). It has 23 health centres, 1 district hospital and 1 prison clinic. Rwanda has a health development strategy based on decentralized management and district-level care. Kilindi district is located in the northern zone of Tanzania with a population of 236,833 residents. Kilindi District comprises of 16 rural wards and 102 villages. It has 30 dispensaries, 3 health centres and one hospital. Tanzania has a hierarchical health system made up of the dispensaries found in every village, health centres at the ward level, district hospital at the district level, the regional referral hospital at the regional level, the zone hospitals at the tertiary level and the national hospital at the national level. There are also some specialized hospitals which do not fit directly into this hierarchy and therefore are directly linked to the ministry of health. Kitgum district is located in the northern region of Uganda and has a population of 247,800 residents. It comprises 51 parishes and 437 village councils . Kitgum district has 2 hospitals and 23 health centres; 21 are government owned while 4 are owned by non -government Organizations. Health services delivery in Uganda is decentralized within national, districts and health sub-districts levels with referral hospitals at the national level and health centres at the district and sub-district levels.
The sample was in two stages. In the first stage, a total of 20 villages (clusters) were selected from cells for Gicumbi, wards for Kilindi and Parishes for Kitgum. In the second stage, 32 households were randomly selected in each selected villages (clusters). The detailed sampling procedure for Gicumbi in Rwanda has been reported elsewhere [bib_ref] Moderate and severe household food insecurity predicts stunting and severe stunting among..., Agho [/bib_ref]. For district-level results, sample weights will be used, and sampling weight was calculated by the product of the reciprocal of the sampling fractions employed in the selection of (cells for Gicumbi, wards for Kilindi and Parishes for Kitgum). For the combined analysis of the three datasets, we re-normalised our sampling weights by computing the total sum of weights for each district and divide each district survey sampling weights with the total sum of weights.
## Data source
Our dataset was obtained from a survey conducted during the harvest period, from 21st-31st of January, 2016 in Gicumbi district in Rwanda, Kitgum district in Uganda and Kilindi district in Tanzania. The survey was commissioned as part of World Vision Rwanda, Uganda and Tanzania funding service agreement to generate evidence to influence maternal and child health programmes which aimed to reach 36,250 disadvantaged beneficiaries in these East African districts. The Maternal Newborn Child Health (MNCH) Project aimed to collect health and related indicators to identify the health needs of women and children and to establish priorities for evidence-based planning, decision-making in these regions. The program was an opportunity for World Vision to embed knowledge and action of the organisation's '7-11' interventions for maternal and child survival in the Region [bib_ref] Bridging the gap for maternal newborn and child health human resources in..., Geoffrey [/bib_ref]. World Vision uses the 7-11 approach to prevent maternal and child mortality and morbidity through 7 key interventions for a mother and 11 interventions for the child. The intervention for the mother are: diet, deworming and iron supplements, prevention of infectious diseases, malaria prevention and treatment, appropriate pregnancy spacing, birth preparedness, and access to antenatal and postnatal maternity services. The 11 interventions for the child are appropriate breastfeeding, newborn care, timely complementary feeding, age-appropriate immunisation, sufficient iron intake, consistent hand washing prevention and treatment for acute malnutrition, prevention and treatment of malaria, and acute respiratory infection. Others are timely administration of oral rehydration therapy to treat diarrhoea, prevention and care for pediatric Human Immunodeficiency Virus (HIV), and timely deworming.
## Study outcomes
The nutritional status of children under five years of age was measured anthropometrically. We considered height-for-age (stunting), weight-for-height (wasting) and weight-for-age (underweight). The height-for-age index is an indicator of linear growth retardation and cumulative growth deficits in children, Weight-for-height index measures body mass in relation to height and reflects the current nutritional status of the child. Weight-for-age takes into account both acute malnutrition (wasting) and chronic malnutrition (stunting), but it does not distinguish between stunting and wasting. The index is calculated using growth standards published by WHO in 2006. These growth standards were generated through data collected in the WHO Multicentre Growth Reference Study and expressed in standard deviation units from the Multicentre Growth Reference Study median. Child undernutrition status was categorized into four categories -severe undernutrition (< − 3.0 Z-score), moderate undernutrition (− 3.0 to − 2.0 Z-score) and mild undernutrition (− 2.0 to − 1.0 Z-score) and proper nutrition (≥ − 1.0 Z-score). The level of childhood undernutrition (stunting, wasting and underweight) was considered as the outcome variables with four ordinal categories (severely undernourished, moderately undernourished, mildly undernourished, and nourished).
## Potential confounders
The potential confounding factors were organised into four distinct groups: Socio-economic and demographic (Districts, primary caregiver, education level, marital status, household wealth index and food security); child (fever in two weeks preceding each survey, acute respiratory infection (ARI) in two weeks prior to each survey, diarrhoea in the two weeks preceding each survey, sex of baby and child's age in months); maternal and child health (antenatal care and attended child monthly growth monitoring sessions); health services and environmental factors (quality of care from health services, place of delivery, water available all year, sources of drinking water and type of toilet facility).
The household wealth index variable measures basic household needs for all children 5-18 years. The household wealth index was constructed by assigning weights to three basic household needs for children 5-18 years (i.e. difficulty providing at least two sets of clothes for all children aged 5-18 years living in the household, difficulty providing a pair of shoes for all children aged 5-18 years living in the household and difficulty paying school fees or school contribution for all children aged 5-18 years living in the household) using principal components analysis. The household wealth index was divided into three categories: poorest, middle and least poor [bib_ref] Estimating wealth effects without expenditure data-or tears: an application to educational enrolments..., Filmer [/bib_ref]. Improved and unimproved sources of drinking water and type of toilet facility were categorised based on the WHO and UNICEF Joint Monitoring Programme guidelines. Birth order and number of children per household were not collected. Higher birth order and number of children per household may result in higher susceptibility to undernutrition due to impact on income.
# Statistical analysis
Re-normalised weight was used for Survey (SVY) tabulation that adjusts for clustering, and sampling weights were used to determine the percentage and frequency count of all selected characteristics and, district-specific weights were used for the Taylor series linearization method in the surveys when estimating 95% confidence intervals around prevalence estimates of undernutrition by severity in each district.
Multivariate multinomial logistic regression model guided by the conceptual framework was used to determine factors associated with childhood undernutrition (stunting, wasting and underweight) with nourished children used as reference category.
In the multivariate analyses, a four-stage model was carried out, in the first stage model, socio-economic and demographic factors were entered into the model, and a stepwise backward elimination method was used to remove the non-significant factors (p > 0.05). In the second stage model, the significant factors in the first stage model were added to the child level factors, and this was followed by another stepwise backward elimination procedure which retained all the significant factors. A similar procedure was employed for the third stage model which included the individual (maternal and child) level factors as well as health services factors and the final stage model which introduced environmental factors. After completion of all four modelling stages, the factors that were significantly associated with the outcomes were retained. All statistical analyses were conducted using STATA/MP Version.14.1 (StataCorp, College Station, Texas, USA) and adjusted odds ratios (AORs) and their 95% confidence intervals (CIs) obtained from the adjusted multivariate multinomial logistic regression model were used to measure the factors associated with childhood undernutrition.
# Results
Distribution of children aged 0-59 months in three East African Districts is presented in [fig_ref] Table 1: Distribution of children aged 0-59 months in three disadvantaged East African Districts [/fig_ref]. The majority of women were from Kitgum district in Uganda, and the proportion of respondents who were educated up to secondary level was about 9%. Women who were never married made up 57% while 89% of the women reported being the primary caregiver.
The proportion of male and female children was relatively constant, and 51.2% of respondents were from poorest households. Over two-thirds of women reported having water availability all year and over 80% had no diarrhoea or ARI. The proportion of mildly, moderately and severely stunted were 25.1, 16.8 and 11.8%, respectively while a high proportion of children reported either not wasted or not underweight. Percentage distribution of childhood undernutrition aggregated by three East African Districts was reported Additional file 1: [fig_ref] Table 1: Distribution of children aged 0-59 months in three disadvantaged East African Districts [/fig_ref]. [fig_ref] Figure 2: Prevalence and 95% confidence intervals [/fig_ref] show the prevalence and 95% confidence Intervals of undernutrition (stunting, wasting and underweight) by severity. In the figures, if the 95% confidence intervals overlap implies non-significant different at the 95% confidence level. In [fig_ref] Figures 1 ,: Figures 1, 2 and 3 show the prevalence and 95% confidence Intervals... [/fig_ref] , the prevalence of severe and moderate stunting was significantly higher in Gicumbi district and Kilindi district but lower in Kitgum district while the prevalence of mild stunting was higher in Kitgum district compared with the other two districts (Gicumbi and Kilindi).
## Prevalence of undernutrition
In [fig_ref] Figure 2: Prevalence and 95% confidence intervals [/fig_ref] , the prevalence of mild, moderate and severe wasting was significantly higher in Kitgum district compared with Gicumbi and Kilindi districts while the prevalence of underweight was also higher in Kitgum district than the other two districts (Gicumbi and Kilindi) as represented in [fig_ref] Figure 3: Prevalence and 95% confidence intervals [/fig_ref].
Factors associated with mild, moderate and severe stunting [fig_ref] Table 2: Factors associated with mildly, moderately and severely stunted in children aged 0-59... [/fig_ref] shows the factors associated with mild, moderate and severe stunting in children aged 0-59 months in three disadvantaged East African districts. The odds of a child being moderately or severely stunted significantly decreased in Kitgum and Kilindi districts compared to Gicumbi district. Significantly increased odds of mild, moderate and severe stunting were reported among household wealth index (least poor), primary caregiver (mothers), male child, child aged 24-59 months and children who reported having diarrhoea.
Factors associated with mild, moderate and severe wasting [fig_ref] Table 3: Factors associated with mildly, moderately and severely wasted in children aged 0-59... [/fig_ref] shows the factors associated with mild, moderate and severe wasting in children aged 0-59 months in three disadvantaged East African districts. Mild, moderate and severe wasting significantly increased in Kitgum district compared with Kilindi and Gicumbi districts. Significantly increased odds of mild and severe wasting were observed among the middle household wealth index while significantly increased odds of moderate and severe wasting were reported among the least poor. Maternal education (no schooling), primary giver (mothers), male child, no water availability all year and children who reported having diarrhoea were significantly associated with mild, moderate and severe wasting. The odds of moderate and severe wasting were significantly higher in children aged 24-59 months while the source of drinking water (unimproved) was significantly associated with mild wasting.
Factors associated with mild, moderate and severe underweight [fig_ref] Table 4: Factors associated with mildly, moderately and severely underweights in children aged 0-59... [/fig_ref] shows the factors associated with mild, moderate and severe underweight in children aged 0-59 months in three disadvantaged East African districts. The odds of a child being mildly, moderately or severely underweight were significantly higher in Kitgum district compared with Kilindi and Gicumbi districts. Increased odds of mild, moderate and severe wasting were observed Acute respiratory infection (ARI) Weighted total = 9270 unless otherwise given in parenthesis among mother of children who did not attend child monthly growth monitoring sessions, a child aged 0-23 months and children who reported having diarrhoea. Children who had ARI symptoms were 1.18 times and 1.82 times more likely to report higher odds of moderately or severely underweight, respectively.
# Discussion
This study identified the factors associated with childhood undernutrition in three East African Districts. The main factors associated with moderate or severe stunting/wasting were: wealth index (poorest households), increasing child's age, sex of the child (male) and unavailability of water all year. Children of women who did not attend monthly child growth monitoring sessions and children who had ARI symptoms were significantly associated with moderate or severe underweight while having diarrhoea two weeks prior to the survey was significantly associated with severe undernutrition (stunting, wasting and underweight). In this study, children under-five years residing in Gicumbi District in Rwanda were more predisposed to stunting while the odds of a child being wasted and underweight were higher in Kitgum District in Uganda. Though all three districts analysed were in East Africa, disparities still existed in the distribution of child undernutrition, and this might be due to specific features associated with the district's geographic location. This report shows that geographic location may influence child nutrition and is a determinant of childhood undernutrition. Geographic location affects the cultivation of food crops, which in turn affects access and availability of food for household consumption and sale. The impact of geographical location could be seen in the environmental variability and predominant occupation practised in the region which could influence food (in) security and consequently affect child nutrition, growth and development [bib_ref] Correlates of stunting among children in Ghana, Darteh [/bib_ref]. Geographic locations with more favourable climatic conditions and whose residents are predominantly farmers have greater access to variable food types and are more food secured [bib_ref] Correlates of stunting among children in Ghana, Darteh [/bib_ref]. However, studies have shown that cultural beliefs and practices unique to a region influence the food given to the growing child despite its overall nutritional value and availability [bib_ref] Stunting and severe stunting among children under-5 years in Nigeria: a multilevel..., Akombi [/bib_ref] [bib_ref] Multilevel analysis of factors associated with wasting and underweight among children underfive..., Akombi [/bib_ref] [bib_ref] Determinants of appropriate child health and nutrition practices among women in rural..., Mwangome [/bib_ref].
In this study, the wealth index was used as a proxy to access socioeconomic status. Children from poor households were more susceptible to childhood stunting and wasting when compared to their counterparts from richer households in all three East African Districts.
Evidence from previously conducted studies shows the inverse relationship between socioeconomic status and childhood stunting/wasting [bib_ref] Stunting and severe stunting among children under-5 years in Nigeria: a multilevel..., Akombi [/bib_ref] [bib_ref] Prevalence and determinants of under-nutrition among children under six: a cross-sectional survey..., Kavosi [/bib_ref] [bib_ref] Determinants of child nutritional status in Zambia: an analysis of a national..., Masiye [/bib_ref] [bib_ref] Determinants of stunting and severe stunting among under-fives: evidence from the 2011..., Tiwari [/bib_ref]. These studies confirm that children from households with low socioeconomic status lack access to sufficient food of adequate quality, basic health care services, and have a higher risk of infection as a result of compromised immunity due to poor nutrition and suboptimal living conditions [bib_ref] Stunting, wasting and underweight in sub-Saharan Africa: a systematic review, Akombi [/bib_ref].
Child's age was also reported as a major factor associated with stunting and wasting in the study area. Suboptimal growth increased with age as older children were reported to be more predisposed to childhood stunting/wasting. This increase in child undernutrition with age could be as a result of an increase in energy expenditure by the older child without sufficient and adequate food intake. It could also result from an increased interaction of the older child with its immediate environment which may lead to increased risk of infections and exposure to childhood diseases either through drinking water from unimproved sources, consumption of contaminated foods, poor hygiene or poor environmental sanitation [bib_ref] Stunting, wasting and underweight in sub-Saharan Africa: a systematic review, Akombi [/bib_ref]. Furthermore, as the child's age increases, s/he requires more energy (calories) and nutrients for proper growth and development. The inability of a growing child to meet required daily energy and nutrient needs could result in undernourishment [bib_ref] Multilevel analysis of factors associated with wasting and underweight among children underfive..., Akombi [/bib_ref].
Male children were more prone to childhood stunting/ wasting than their female counterparts. Male children tend to be more physically active thereby expending large amounts of energy which should be channelled towards proper growth and development. This finding is consistent with results from other studies carried out in Nigeria [bib_ref] Stunting and severe stunting among children under-5 years in Nigeria: a multilevel..., Akombi [/bib_ref] [bib_ref] Multilevel analysis of factors associated with wasting and underweight among children underfive..., Akombi [/bib_ref] , Iran [bib_ref] Prevalence and determinants of under-nutrition among children under six: a cross-sectional survey..., Kavosi [/bib_ref] , Kenya [bib_ref] Trends and determinants of undernutrition among young Kenyan children: Kenya demographic and..., Masibo [/bib_ref] , Indonesia [bib_ref] Prevalence and risk factors for stunting and severe stunting among under-fives in..., Ramli [/bib_ref] , Tanzania [bib_ref] Determinants of Child Malnutrition in Tanzania: a Quantile Regression Approach, Shiratori [/bib_ref] Ghana [bib_ref] Correlates of stunting among children in Ghana, Darteh [/bib_ref] , Ethiopia [bib_ref] Determinants of wasting among under-five children in Ethiopia: (a multilevel logistic regression..., Dabale [/bib_ref] and South Africa [bib_ref] Risk factors of poor anthropometric status in children under five years of..., Lesiapeto [/bib_ref]. The studies reported a higher prevalence of childhood undernutrition among male children than females. However, a biological reason for this is still unknown.
Unavailability of water was reported as a significant factor associated with moderate and severe stunting. Lack of access to clean and safe water from improved water sources could have a significant detrimental effect on child growth and development resulting from sustained exposure to enteric pathogens and diarrhoeal disease [bib_ref] Can water, sanitation and hygiene help eliminate stunting? Current evidence and policy..., Cumming [/bib_ref]. Thus, improvements in the quantity and microbial quality of water available to the growing child is essential if the child is to meet its nutritional needs and reduce the risk of infection or dehydration. Additionally, the unavailability of water for irrigation purposes affects farming leading to low food production and food insecurity especially in locations with low rainfall. Household unavailability of water may result from the inability of households with low socio-economic status to afford the cost of clean drinking water. Therefore, to improve access to water and the nutritional status of children, consistent access to affordable, clean water, especially in disadvantaged communities, is vital [bib_ref] Can water, sanitation and hygiene help eliminate stunting? Current evidence and policy..., Cumming [/bib_ref]. Having diarrhoeal episodes and manifesting ARI symptoms two weeks prior to the survey were significantly associated with severe undernutrition. Studies have confirmed the relationship between ARI, diarrhoea and undernutrition which results in a vicious cycle [bib_ref] Prevalence of undernutrition and associated factors among children aged between six to..., Asfaw [/bib_ref] [bib_ref] Maternal and child undernutrition and overweight in low-income and middle-income countries, Black [/bib_ref] [bib_ref] Malnutrition and Gastrointestinal and Respiratory Infections in Children: A Public Health Problem, Rodriguez [/bib_ref] [bib_ref] Nutritional predictors of acute respiratory infections among children born to HIV-infected women..., Mwiru [/bib_ref] [bib_ref] Factors associated with acute respiratory infection in children under the age of..., Geberetsadik [/bib_ref]. Gastrointestinal and respiratory infections reduce appetite, increases catabolism and inhibit intestinal absorption of nutrients from food thus leading to increased susceptibility to severe undernutrition especially underweight [bib_ref] Malnutrition and Gastrointestinal and Respiratory Infections in Children: A Public Health Problem, Rodriguez [/bib_ref] [bib_ref] Nutritional predictors of acute respiratory infections among children born to HIV-infected women..., Mwiru [/bib_ref]. Consequently, child undernutrition leads to immune dysfunction, such as the impairment of cell-mediated immunity, cytokine and immunoglobulin production which lowers immunity and predisposes a child to infectious diseases [bib_ref] Malnutrition and Gastrointestinal and Respiratory Infections in Children: A Public Health Problem, Rodriguez [/bib_ref] [bib_ref] Factors associated with acute respiratory infection in children under the age of..., Geberetsadik [/bib_ref] [bib_ref] Child nutrition and lower respiratory tract disease burden in New Zealand: a..., Grant [/bib_ref]. Diarrhoea and ARI often occur due to suboptimal child feeding practices and an unhealthy environment. Studies have also shown that exclusive breastfeeding in the first 6 months of life and the timely introduction of appropriate complementary foods (6-23 months) reduces the risk of diarrhoea and ARI [bib_ref] Malnutrition and Gastrointestinal and Respiratory Infections in Children: A Public Health Problem, Rodriguez [/bib_ref] [bib_ref] Nutritional predictors of acute respiratory infections among children born to HIV-infected women..., Mwiru [/bib_ref]. Proper child feeding practices build the child's immunity through the transfer of innate immune components such as secretory IgA, lactoferrin and lysozyme from mother to child during breastfeeding and enhance antibody response to pathogens thus preventing infections [bib_ref] Malnutrition and Gastrointestinal and Respiratory Infections in Children: A Public Health Problem, Rodriguez [/bib_ref] [bib_ref] Nutritional predictors of acute respiratory infections among children born to HIV-infected women..., Mwiru [/bib_ref]. Therefore, to prevent ARI and the occurrence of diarrhoea, interventions targeted at promoting exclusive breastfeeding, vaccination, Los-osmolarity ORS, zinc and Vitamin A supplementation, availability of safe water, improved sanitation practices and preventing household pollution should be effected.
Children of women who did not attend monthly child growth monitoring sessions were significantly associated with moderate or severe underweight. Attending monthly child growth monitoring sessions allows for the evaluation of child growth in order to assess the appropriateness of the child's nutrient intake and sanitation.
Additionally, the high prevalence of undernutrition in children under-5 years reported in this study may be attributed to the main lean period (October-December) before the survey when food prices are unreasonably high due to seasonal variability, and this continues up to harvest period before gradually decreasing.
# Strengths and limitations
This study had several strengths. First, the study was population-based with a large sample size which is representative of the study area. Second, it applied appropriate statistical adjustments to data obtained from the survey and identified the most vulnerable subpopulation affected by childhood undernutrition in a large sample. However, this study also had some limitations. First, due to the cross-sectional nature of the study design, a causal relationship between the observed risk factors and childhood undernutrition cannot be established. Second, this study did not include validity assessments of undernutrition; no actual dietary assessments were made. Third, despite the use of a comprehensive set of variables in the analysis, the effect of genetics, birth defects, and early food allergies as well as the effect of residual confounding as a result of unmeasured covariates were not addressed. Finally, the measure of household wealth index using a small number of variables may misrepresent household expenditure and the actually socioeconomic status of the household.
## Policy implications
Findings from this study are useful for public health planning to improve the nutritional status of children under-five years in East Africa. Intervention strategies geared towards reducing undernutrition in East African communities should focus on children from poor households with suboptimal sanitation practices. Communitybased educational sessions which educate women on the benefits of attending postnatal monthly child growth monitoring sessions should also be initiated.
# Conclusions
Findings from our study indicate that having diarrhoea, ARI, unavailability of water supply all year and failure to attend monthly child growth monitoring sessions increases the odds of childhood undernutrition, especially among socioeconomically disadvantaged households. Thus interventions to reduce childhood undernutrition should focus on improving household sanitation and access to adequate water supply, as well as initiating community-based educational campaigns on the merits of attending child growth monitoring sessions. These interventions will improve the nutritional status of children under-five years in these East African communities which will invariably result in an overall decline in childhood undernutrition in Sub-Saharan Africa hereby setting the region on the path to achieving the WHO global nutrition target by 2025.
## Additional file
Additional file 1:
## Availability of data and materials
The dataset analysed in this study is available as an additional supporting file.
Authors' contributions KEA was involved in study conceptualization, data analysis and drafting of the manuscript. BJA was involved in drafting, editing and revising of the manuscript. JKK revised and enhanced the intellectual content of the manuscript. FJA revised the manuscript and made important contributions. IM was involved in editing and enhancing the intellectual content of the manuscript. All authors read and approved the final manuscript.
Ethics approval and consent to participate Ethical clearance was obtained from Ministry of Health in Kigali, Kampala and Dar es Salaam, and necessary permission was also obtained from the Gicumbi, Kilindi and Kitgum regional health office and the Gicumbi, Kilindi and Kitgum health office and local administrators. Participants were given informed consent to sign before taking part in the survey, including assurance of anonymity and a description of how the data would be used. For illiterate participants, informed consent information was read aloud and signed. Mothers and children with serious illness were referred to the nearby health facilities. The data in this article presented as aggregate to ensure all respondents' identification information is obscured.
## Consent for publication
Not applicable.
[fig] Figures 1 ,: Figures 1, 2 and 3 show the prevalence and 95% confidence Intervals of undernutrition (stunting, wasting and [/fig]
[fig] Figure 2: Prevalence and 95% confidence intervals (CIs) of wasting by severity [/fig]
[fig] Figure 3: Prevalence and 95% confidence intervals (CIs) of underweight by severity [/fig]
[fig] Figure 1: Prevalence and 95% confidence intervals (CIs) of stunting by severity [/fig]
[table] Table 1: Distribution of children aged 0-59 months in three disadvantaged East African Districts (n = 9270) [/table]
[table] Table 2: Factors associated with mildly, moderately and severely stunted in children aged 0-59 months in three disadvantaged East African Districts [/table]
[table] Table 3: Factors associated with mildly, moderately and severely wasted in children aged 0-59 months in three disadvantaged East African Districts [/table]
[table] Table 4: Factors associated with mildly, moderately and severely underweights in children aged 0-59 months in three disadvantaged East African Districts Confounding factors adjusted for are: Socio-economic and demographic factors; child factors; child undernutrition status, maternal, health services and environmental factors and child's health factors [/table]
[table] Table S1: Percentage distribution of undernutrition in children aged 0-59 months by three East African Districts (N= 8,706). (DOCX 15 kb) Abbreviations ANC: Antenatal Care; AOR: Adjusted Odd Ratios; ARI: Acute Respiratory Infection; CIs: Confidence Intervals; HIV: Human Immunodeficiency Virus; MNCH: Maternal Newborn Child Health; UNICEF: United Nations Children's Fund; WHO: World Health Organization Funding This study was part of end line evaluation for the East Africa maternal, newborn child health project. This project was funded by the Department of Foreign Affairs and Trade of the Australian Government through (AusAID Funding Order 62/37923 under the Australia Africa Community Engagement Scheme (AACES). [/table]
|
Electro-clinical and neurodevelopmental outcome in six children with early diagnosis of tuberous sclerosis complex and role of the genetic background
# Introduction
Tuberous Sclerosis Complex (TSC) is a multisystem neurocutaneous disorder caused by heterozygous pathogenic variants in TSC1 (Chr. 9q34.or TSC2 (Chr. 16p13.3). It is characterized by hamartomas affecting the brain, skin, eye, heart, lung, and kidney. The typical Central Nervous System (CNS) lesions are present in more than 90% of the patients and consist of cortical tubers, white matter radial migration lines, subependymal nodules (SENs), and subependymal giant cell astrocytomas (SEGAs). Up to 85% of the affected individuals have a diagnosis of epilepsy, and seizure onset in the first year of life is very common (67%). TSC may therefore be considered a genetic developmental and epileptic encephalopathy, as both genetic factors and epileptic activity contribute to the phenotype. Early onset epilepsy usually presents with focal seizures or infantile spasms, but all seizure types have been clinically described.
There is a strong association between intellectual disability (ID) and epilepsy, and it is recognized that early seizure onset may negatively contribute to worsening the final developmental outcome. On the other hand, the influence of subclinical seizures has yet to be clarified.
Recent studies have shown that many infants develop a progressive deterioration of the EEG during the first months of life, before the onset of seizures. Furthermore, clinically silent seizures can be present in patients with TSC, and are difficult to detect unless an ictal EEG is available.
Close EEG monitoring (i.e. every 4-6 weeks) during the first months of life allows for early detection of electroencephalographic seizures and for prompt treatment to minimize the deleterious impact of early-onset seizures.
Nowadays, a diagnosis of TSC may be suspected prenatally if cardiac rhabdomyomas are detected on fetal ultrasound, and is subsequently confirmed soon after birth by the presence of skin findings, brain findings, or molecular testing. Thus, early diagnosis has allowed clinicians to follow children with TSC before the development of neurological signs. To the best of our knowledge, there are only few reports of the neurological manifestations of TSC in infants before seizure onset. In this study, we describe early EEG activity, clinical and genetic data, and developmental outcome in order to identify prognostic factors that could lead to a better management of these children.
# Material and methods
We performed a retrospective study of children referred to our TSC clinic, which now comprises over 200 individuals. We selected those who had been suspected to have TSC prenatally, perinatally, or in the first 6 months of life and had been followed at our clinic on a regular basis since 2013. A definite diagnosis of TSC was established after birth according to the 2012 revised criteria.
Our standard clinical practice consists of serial Video-EEG recordings every 4-8 weeks during the first two years of life. In addition to video-EEG monitoring, parents are educated on how to recognize and video record subtle seizures and spasms, in case seizures develop in between two recordings.
If seizures occur, antiepileptic treatment with Vigabatrin is introduced, starting from 50 mg/Kg/day up to 100-150 mg/Kg/day, according to the current clinical recommendations. Patients with normal EEG recordings are followed up without any antiepileptic medications.
Scalp-EEGs with synchronized video are recorded according to the International 10-20 system, modified for newborns using our Institution's clinical EEG software (Micromed, System plus). All EEG recordings include at least the following channels, along with electrocardiography (ECG), deltoid muscles, and breathing recordings. In infants, EEGs are recorded applying the International 10-20 system of electrode placement directly or by cable telemetry with more than 16 channels and at least 3 polygraphic channels (ECG and deltoid muscles).
Prolonged recordings (at least 90 min) during wakefulness and sleep were obtained for all patients.
We reviewed and analyzed each EEG for: We analyzed EEG abnormalities applying the classification proposed by Domanska-Pakiela et al., which is based on interictal epileptiform discharges' (IEA) distribution and their intensity. EEGs were analyzed by two epileptologists (AM and AV), who were blinded to the patients' management. Details of the methods used for analyzing EEGs are presented in. All the patients underwent brain MRI studies performed with a 1.5 Tesla magnet, including fluid-attenuated inversion recovery (FLAIR) sequences to detect the presence of typical cerebral lesions (cortical tubers, subependymal nodules, white matter migration lines, and SEGA).
[formula] Background [/formula]
Psychomotor development (total Developmental Quotient: DQ) or cognitive level (total Intelligence Quotient: IQ) was assessed through standardized scales (Griffiths' Scales of Infant Development, GMDS-ER, Wechsler Preschool and Primary Scales of Intelligence, WPPSI). Behavioral problems were investigated through prolonged clinical observations, caregivers' reports, and standardized tests in children.
Molecular genetic testing for TSC1/TSC2 was available for all the patients, and included analysis for point mutations and deletions/duplications in both genes. All variants identified were submitted to a publicly available database, whenever not already present (http://chromium.lovd.nl/LOVD2/TSC/home.php).
# Results
The study cohort consisted of 6 children (4 boys and 2 girls). The diagnostic criteria that were present in each patient at time of enrolment are reported in.
The diagnosis was suspected following the detection of cardiac rhabdomyomas (3 prenatally, 2 perinatally, and 1 at age 6 months) in all the patients. Subsequent skin examination, neuroimaging, eye examination and abdominal ultrasonography showed the presence of hypomelanotic macules, retinal hamartomas, cortical dysplasia (including tubers and cerebral white matter radial migration lines) and subependymal nodules as the second major diagnostic criterion. No renal angiomyolipomas have been detected in these young patients yet, whereas in three individuals renal cysts were present at the time of evaluation (Patients n. 3, 5 and 6).
Genetic testing documented pathogenic variants in TSC2 in all but one patient. In one individual no mutations in TSC1/TSC2 were identified.
Age at first evaluation ranged from 4 weeks to 6 months, depending on different referrals, with a followup of 2-4 years.
## Patient 1
Patient 1 was born at 37 gestational weeks by caesarean section for breech presentation, after an uneventful pregnancy. Multiple cardiac rhabdomyomas were detected in the neonatal period. Cardiac arrhythmias were recorded on 24-h EKG Holter monitoring, and treatment with Beta-blockers and Flecainide acetate was started.
The first EEG was performed at age 4 months and was normal. EEGs performed every 8 weeks showed the presence of normal background activity with multifocal spikes and sharp waves in sleep, involving the occipital left region (since age 8 months), the right frontotemporal region (since age 12 months), or the left fronto-temporal region (18 months). We classified Patient 1 as C1 []. Interictal discharges disappeared by age 24 months. She never experienced epileptic seizures.
Neurological examination and psychomotor development were normal at two years of age (DQ 98, Griffiths Scales). Genetic testing showed a pathogenic variant in exon 24 of TSC2: c.2771_2772del (p.Phe924*).
## Patient 2
Patient 2 was born at 32 weeks by caesarean section, after a pregnancy complicated by maternal hypertension and diabetes mellitus type 1. Apgar score was 5-9 and weight was 2360 g. Delivery was complicated by perinatal respiratory distress and pneumothorax requiring continuous positive airways pressure (CPAP) ventilation.
Multiple cardiac tumors were detected at 6 months of age.
The first EEG was performed at age 6 months, and was normal. Follow up EEGs were done every 8 weeks, and remained normal until age 2 years.
At age 10 months he presented with an isolated seizure associated with fever. The seizure was characterized by Molecular analysis for TSC1 and TSC2 point mutations (through next generation sequencing) and deletions/duplications was negative.
## Patient 3
Patient 3 was born at 37 weeks of gestation by vaginal delivery. Multiple cardiac tumors had been detected on prenatal ultrasonography at 32 weeks of gestation, and were confirmed by fetal MRI. The first EEGs were performed during the neonatal period and at 3 months of age at a different institution, and were normal.
Seizures occurred at age 4 months and the patient was therefore referred to our TSC clinic. They were characterized by series of infantile spasms, with focal motor signs and asymmetric involvement of the upper limbs (more on the left side) documented on parents' video recordings. EEG was performed, and showed interictal discharges characterized by frequent multifocal spikes and spike-and-waves complexes during wakefulness and sleep involving mainly the right centro-temporal region. We did not record hypsarrhythmia. We classified this patient's EEG features as C3. The patient was started on Vigabatrin at 75 mg/Kg/day, and seizures have been controlled since the first day of treatment. He has been seizure free to the time of last assessment.
Follow-up EEG performed at age 6 months (2 months after starting Vigabatrin) was normal, and remained normal through age 17-24 months.
At two years of age, neurological examination and neuropsychological assessment were normal (DQ 89, Griffiths Scales), with mild language delay.
Genetic tests identified a large deletion involving exons 17-22 of TSC2.
## Patient 4
Patient 4 was born at 40 weeks by vaginal delivery after an uneventful pregnancy. Multiple cardiac rhabdomyomas had been detected prenatally. The first EEG was performed at age 8 weeks and was normal. Follow up EEG performed 8 weeks later showed focal spikes and spike waves complexes during sleep over the right central region. The option of preventive antiepileptic treatment with Vigabatrin was proposed, but the family decided to wait and delayed the appointment for EEG recordings.
Three months later, at age 7 months, the parents reported a series of infantile spasms on awakening.
EEG was performed, and showed frequent interictal discharges, characterized by multifocal spike and wave discharges with diffusion during sleep. We never recorded the presence of hypsarrhythmia. This patient was classified as D3 [.
She was treated with Vigabatrin (100 mg/Kg/day), with good seizure control since the first week of treatment.
Follow-up EEGs showed progressive decrease of epileptic discharges at age 8, 10 and 12 months, and were completely normal at serial recordings after the first year of age.
Neuropsychological development at age 28 months was normal (DQ 104, Griffiths Scales). Neurological evaluation was normal.
A pathogenic variant was identified in exon 30 of TSC2: c.3626 T > C (p.Leu1209Pro), inherited from her affected mother.
## Patient 5
The child was born by vaginal delivery after an uneventful pregnancy. He presented with cardiac arrhythmias during the neonatal period, and subsequent echocardiography showed multiple cardiac rhabdomyomas. The first EEG was performed at age 4 weeks and was abnormal, showing focal interictal discharges over the right frontotemporal region. The first EEG was performed at our clinic at 4 months of age, and showed an increase in epileptic discharges with frequent multifocal spike-andwaves complexes during sleep. While the parents were deciding whether or not to start treatment with Vigabatrin, the family noticed a series of infantile spasms four days after the first EEG recording. The patient was admitted to the hospital, and seizures were recorded at video EEG monitoring. They were characterized by series of spasms with ictal EEG correlates of pseudo periodic high voltage generalized slow waves. Interictal EEG showed multifocal spike and spike-and-waves discharges, with bilateral activation and diffusion during sleep. This patient was classified as D3. Treatment with Vigabatrin (titrated to 150 mg/Kg/day) was started, with control of seizures from the 10th day of treatment.
At 8 months of age he experienced seizures associated with fever; valproic acid was added, with good seizure control.
At age 20 months the patient started to present polymorphic seizures, characterized by prolonged focal clonic seizures involving the left arm during febrile illnesses, with or without bilateral tonic-clonic evolution. Relapsing epileptic spasms coexisting with focal subtle seizures were documented from age 20 months. He was treated with several AEDs with unsatisfactory results. Epilepsy is still drug-resistant at the time of last assessment.
Follow up EEG performed at age 6, 8 and 12 months showed persistent multifocal discharges.
At age 4 years a SEGA of 13 × 10 mm located near the left foramen of Monro was detected on brain MRI, which was stable at neuroradiological follow-up.
At the time of last assessment, the patient's psychomotor development was delayed and moderate cognitive impairment was diagnosed (IQ 48, WPPSI at age 4 years). Autism spectrum disorder (ASD) was also diagnosed (ADOS evaluation).
Genetic testing showed a pathogenic variant in exon 34 of TSC2: c.4544_4547delACAA (p.Asn1515Serfs*60).
## Patient 6
The patient was born at term after an uneventful pregnancy. Multiple cardiac tumors had been detected prenatally. Brain MRI showed bilateral cortical and The first EEG performed at age 8 weeks was normal. At age 4 months, he started to present focal seizures characterized by staring, head and eye deviation, with sparing of the left arm.
Focal seizures arising from the right frontal region were recorded at Video-EEG. Interictal EEG was characterized by focal spike and spike-and-waves discharges on the right fronto-temporal region.
At 7 months of age the patient started to present infantile spasms associated with subtle focal seizures. Sleep EEG was characterized by diffuse slow activity with sporadic spike and wave discharges predominantly on the right fronto-temporal region, without hypsarrhythmia.
The patient was classified as B2 []. Treatment with Vigabatrin initially introduced for focal seizures, was titrated up to 120 mg/Kg/day. Due to drug-resistant seizures, Levetiracetam, Topiramate and Carbamazepine in different combinations were introduced with unsatisfactory results.
Follow-up EEGs confirmed the presence of persisting interictal discharges that were predominant over the right fronto-temporal regions. Patient 6's psychomotor development was delayed based on neuropsychological testing performed at age 24 months (DQ 69, Griffiths Scales), with predominant speech delay.
Genetic testing showed a de novo pathogenic variant in exon 35 of TSC2: c.4570dup (p.Ser1524Serfs*5).
The course of EEG activity and epilepsy for each patient is reported in.
# Discussion
Most Authors recognize a link between early seizure onset and poor long-term neurological and behavioural outcome. This has stimulated several investigators to develop strategies to prevent seizure onset. Assuming that interictal epileptiform abnormalities (IEA) in children with TSC can be considered a marker of the dynamics of the epileptogenic process, multicenter prospective randomized double-blind clinical trials are currently ongoing, aiming at verifying the impact of the treatment of preclinical vs. clinical seizures (EPISTOP project in Europe and PREVENT trial in the United States [NCT02849457]). However, data about early EEG monitoring in TSC patients are limited in the literature.
While waiting for the results of these seminal studies, in clinical practice we are faced with newborns referred for TSC due to the improvement of ultrasound techniques during pregnancy. Nowadays TSC is increasingly more frequently diagnosed in the prenatal period or in early infancy, thus offering the possibility of monitoring children before the onset of seizures and/or neurodevelopmental delay.
According to the clinical recommendations for management of epilepsy in TSC, treatment should be initiated in infants and children within 24 months of age if ictal discharges occur, with or without clinical manifestations. Since 2013, we have had the opportunity to follow 6 infants with TSC starting from the first months of life. All of them received video-EEG recordings every 4-8 weeks. The aim of the present study was to describe the electro-clinical and neurodevelopmental outcome in 6 children with early diagnosis of TSC.
Although limited by the small number of individuals included in this study and the relatively short follow-up, neither age at EEG abnormalities appearance nor EEG characteristics at onset were sufficient for determining long-term prognosis. Indeed, one of the patients described herein showed EEG abnormalities for months and never developed seizures (Patient 1). Her neurodevelopment is normal. On the other hand, disappearance of EEG abnormalities was related to good prognosis for both epilepsy and cognitive outcome.
Mean age at EEG abnormalities appearance was 4 months, in line with that reported in larger cohorts by Jozwiack et al..
According to the EEG classification proposed by Domanska-Pakiela et al., irregular and continuous spike-and-waves complexes, either localized on one hemisphere or multifocal, were more frequently seen also in our group of patients. Clear hypsarrhythmia was not identified on the EEGs even though four infants showed IS.
The lack of EEG deterioration could be explained by the close EEG monitoring leading to early treatment of IS, which did not allow the progression of epileptiform activity. This finding should be considered a typical feature in TSC, and clinicians should not wait to start antiepileptic treatment.
In addition, seizure type -either focal seizures or ISdid not correlate with epilepsy or intellectual prognosis in our sample.
The major contribution to neurodevelopment seems not only linked to seizure onset but rather to the opportunity of their prompt control, as seen in patients 3 and 4 from our cohort. This finding is in line with the recent work by Capal et al., demonstrating that timing of complete seizure control, in particular at age 12 months, is predictive of global development and ASD behaviors in the long-term.
Moreover, our data further confirm that the refractoriness of seizures is associated with cognitive disability, as seen in Patients 5 and 6 and reported by Humphrey et al., who demonstrated that IQ decline is associated with duration of infantile spasms in infants with TSC.
Nevertheless, it is known that seizures are not the only determinant of cognitive status in TSC. Genetic factors, i.e. pathogenic variants in TSC2 rather than TSC1, have been recognized as a significant risk factor for earlier and more severe epilepsy, as well as for a higher rate of cognitive impairment, with due exceptions.
Specifically, TSC2 pathogenic variants have been associated with a significantly higher occurrence of IS and other epilepsy types. However, certain missense mutations located in the central region of TSC2 (exons 23-33) have been reported to be associated with a reduced incidence of IS. In fact, we are aware of patients with central missense mutations presenting with IS. Of note, Patient 4 in our cohort exhibited a missense variant in exon 30 of TSC2 and presented with infantile spasms, but had normal development. This suggests that, although infantile spasms can be seen in patients with central missense mutations (yet, significantly less frequently than in patients with different TSC2 pathogenic variants), developmental outcome could be favorable.
Our small group of children with TSC was genetically homogenous. A pathogenic variant in TSC2 was detected in all but one: the latter did not develop seizures and has normal neurodevelopment. On the other hand, the two patients with the worst outcome, in terms of seizure control and intellectual/behavioral disability, both carried pathogenic variants in exons 34 and 35 of TSC2.
As hypothesized by Kothare and colleagues, changes in the TSC2 GAP domain may lead to a more pronounced mTOR hyperactivation and a more severe phenotype. We therefore hypothesize that pathogenic variants in exons 34-35 of TSC2 (GAP domain) are responsible for the worst outcome in the two infants in our sample, but this finding should be confirmed in bigger cohorts.
Taken together, these observations led us to infer that genotype itself could influence the outcome in patients with TSC, at times independent from epilepsy onset, seizure type, and seizure control. However, further studies on larger numbers of patients are warranted to verify this hypothesis.
In addition to the influence of seizures and genetics on the outcome, neuropathological findings, such as a higher tuber/brain proportionor the number of radial migration lineshave been demonstrated to contribute to cognitive and behavioural phenotype. All our patients showed multiple cortical tubers and subependymal nodules, as well as white matter radial migration lines on brain MRI, so that this finding was difficult to correlate to the patients' outcome. The fact that intellectual dysfunction has been reported in patients without epilepsyor with no apparent brain lesions, as well as in animal models of TSC, suggests that the neurological and behavioral problems in TSC may somehow emerge independently from seizure onset, and be presumably related to underlying altered CNS development and neuronal connectivity.
Nevertheless, epileptic activity and brain structural alterations may also interact with each other, and seizures may result as a trigger for predisposed structures or provide additional insult worsening intellectual disability, as recently suggested.
Both early onset seizures and CNS abnormalities in TSC are genetically determined through a cascade of events involving early mTOR overactivation.
This consideration may open the way to the possibility of early treatment with mTOR inhibitors for TSCrelated neuropsychiatric disorders. However, in experimental models, animals that were treated prenatally with mTOR inhibitors exhibited long-term cognitive sequelae. On the other hand, administration of mTOR inhibitors in the early postnatal period resulted in preventing seizure onset or in making seizures cease if given after epilepsy onset.
Only limited human data on early postnatal use of mTOR inhibitors are available, suggesting this treatment is scarcely effective in preventing seizure onset.
Vigabatrin is recognized to be the most effective treatment for IS and focal seizures starting in infants with TSC. In addition to inhibiting GABA-transaminase and increasing the synaptic concentration of GABA in the brain, it also plays a role in inhibiting mTOR overactivation, and may have an impact on glutamatergic transmission, which is essential for neurodevelopment.
Inflammatory mechanisms involving specific cytokines and chemokines are abnormally activated in mouse models of TSC, and inhibition of these mechanisms was associated with a decrease in seizures and improved survival in these mice. Starting from this proof-ofconcept, other modulating factors should be further considered in children with TSC, and anti-inflammatory treatments may represent potential therapy for this genetic epilepsy.
# Conclusion
Our report supports the importance of EEG monitoring before seizure onset in TSC patients, and the correlation between seizure control and positive neurodevelopmental outcome. Although we could not find specific predictors, mainly because of the small number of patients analysed, we hypothesize a role of the genetic background in influencing the outcome. Hopefully, ongoing studies on large populations of children with TSC will delineate the key players of neurodevelopmental outcome in the near future. |
The association between vaccination confidence, vaccination behavior, and willingness to recommend vaccines among Finnish healthcare workers
Information and assurance from healthcare workers (HCWs) is reported by laypeople as a key factor in their decision to get vaccinated. However, previous research has shown that, as in the general population, hesitancy towards vaccines exists among HCWs as well. Previous studies further suggest that HCWs with a higher confidence in vaccinations and vaccine providers are more willing to take the vaccines themselves and to recommend vaccines to patients. In the present study with 2962 Finnish HCWs (doctors, head nurses, nurses, and practical nurses), we explored the associations between HCWs' vaccination confidence (perceived benefit and safety of vaccines and trust in health professionals), their decisions to accept vaccines for themselves and their children, and their willingness to recommend vaccines to patients. The results showed that although the majority of HCWs had high confidence in vaccinations, a notable share reported low vaccination confidence. Moreover, in line with previous research, HCWs with higher confidence in the benefits and safety of vaccines were more likely to accept vaccines for their children and themselves, and to recommend vaccines to their patients. Trust in other health professionals was not directly related to vaccination or recommendation behavior. Confidence in the benefits and safety of vaccines was highest among doctors, and increased along with the educational level of the HCWs, suggesting a link between confidence and the degree of medical training. Ensuring high confidence in vaccines among HCWs may be important in maintaining high vaccine uptake in the general population.
[formula] a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 [/formula]
# Introduction
The introduction of vaccinations has led to improvements in global health by dramatically decreasing the spread of infectious diseases. Global health organizations, such as the WHO, give high priority to the development and implementation of effective immunization programs. In spite of the indisputable benefits of vaccines, vaccination uptake has decreased in some parts of the world, and some vaccine-preventable diseases are on the increase. The question of why some individuals choose to reject vaccines, even though they are safe, easily accessible, and often free of charge or at least affordable, is becoming an increasingly important topic of research.
Several studies have found that a variety of contextual, social, individual, and vaccine-specific factors are associated with accepting, postponing, or refusing vaccination. From a psychological perspective, an important correlate of vaccination behavior is vaccination confidence, which refers to attitudes and beliefs related to the benefits and safety of vaccines, as well as trust in vaccine providers, such as healthcare workers (HCWs), health authorities, and policy makers. Systematic reviews show that individuals who perceive vaccines as less beneficial and safe, more often reject scheduled vaccinations for their children and influenza vaccinations for themselves. Systematic reviews also suggest that greater trust in HCWs is associated with a higher likelihood of accepting vaccines. HCWs play a key role in maintaining and increasing vaccine uptake, as they are involved in the immunization process and communicate with patients about vaccines. Studies based on self-reports from laypeople show that, in general, HCWs are considered to be the most reliable source of vaccine information. Furthermore, advice given by HCWs is the most frequently reported reason for vaccine acceptance among the general public, while lack of recommendation is mentioned as a reason for non-vaccination. HCWs' recommendation behavior therefore seems to play a pivotal role in vaccination decisions.
Although the vast majority of HCWs endorse vaccination, negative attitudes towards vaccination can be found among HCWs as well. According to a recent systematic review, HCWs with lower confidence in the benefits and safety of vaccines are less willing to recommend vaccines to their patients; see also,, and less likely to accept vaccinations for themselves. An important aspect to consider when investigating vaccine attitudes in HCWs, is that compared with the general public, HCWs are expected to have acquired evidence-based information on vaccines. The HCWs' perception about the benefits and safety of vaccines should, thus, not only be considered as a measure of their attitudes to vaccines, but also as an indicator of their vaccine-related knowledge.
Trust in their fellow health professionals is another aspect of vaccination confidence worthy of exploration among HCWs, as they too can experience a lack of trust in the health system, even though they are a part of it themselves. However, the association between trust in the health system and vaccination and recommendation behavior has rarely been studied in HCW samples. Učakar et alsurveyed 897 Slovenian physicians and found that those who regularly or occasionally take the influenza vaccine, more often reported that they trust professional vaccine recommendations, compared with those who no longer or never vaccinate themselves against influenza. A French study; see also,on 2586 general practitioners showed that lower trust in official and scientific sources of vaccine information (such as health authorities, scientific sources, or specialist colleagues) was associated with a reduced likelihood of recommending vaccines. The general practitioners were also less likely to recommend vaccines when they believed that severe adverse effects from vaccines were likely and had doubts about the utility of vaccines. Thus, the results from this study suggested that all three components of vaccination confidence were related to general practitioners' willingness to recommend vaccines to their patients.
In the present study, we investigated whether the perceived benefit of vaccines, perceived safety of vaccines, and/or the degree of trust in health professionals, were related to whether the HCWs' accept vaccinations for themselves and for their children, and whether they recommend childhood and influenza vaccines to patients who are hesitant towards vaccines. More specifically, we hypothesized that HCWs who perceive vaccines as less beneficial, less safe, and/or have less trust in health professionals, are 1) more likely to have hesitated, postponed, or rejected a vaccination for their children, 2) more likely to have rejected the influenza vaccine for themselves, and 3) are less likely to recommend childhood and influenza vaccines to vaccine-hesitant patients.
# Methods
## Study context
In Finland, where the present study was carried out, childhood vaccinations are voluntary and free of charge and almost all vaccines are administered by a public health nurse at a child health clinic in accordance with the national vaccination program . The influenza vaccines are administered to HCWs at their place of work and are paid for by their employer. Since March 2017, the new Infectious Disease Act 48 §5 requires HCWs working with patients belonging to risk populations to be immunized for measles, chickenpox, and influenza. HCWs who do not accept annual influenza vaccinations may be assigned to other tasks.
# Ethics statement
The project received ethical approval from the ethics committee of the Hospital District of Helsinki and Uusimaa. Before completing the questionnaire, respondents were presented with information about the purpose of the study and management of the data. Participants were asked to indicate their understanding of the information provided and expressed consent by checking a box. The respondents were informed that participation was voluntary and that they could withdraw from the study at any time.
## Study population
We sourced participants from the Finnish Public Sector study, which is a large on-going cohort study among municipal and hospital employees. We sent an invitation to participate in an electronic survey to the 8770 hospital staff members who had participated in the study wave the preceding year (2017). For practical reasons, the survey was sent out in two phases. In the first round (February 28 th , 2018), the survey was sent to hospital personnel in the regions of Forssa (n = 916), Kanta-Häme (n = 1201), Pietarsaari (n = 761), and Vaasa (n = 1321), and in the second round (March, 13 th , 2018) to hospital personnel in Pirkanmaa (n = 4571).
## Measures
The survey questions were developed by the authors of the present study after a literature review and discussions with experts working within a nursing education programme in Finland. The dimensions of the survey relevant for the aim of the present study concerned: 1) beliefs about the benefits of vaccines, 2) beliefs about the safety of vaccines, 3) general trust towards healthcare professionals, 4) own vaccination behavior (i.e., whether respondents had accepted childhood vaccines for their own children and influenza vaccines for themselves), and 5) recommendation behavior in cases where patients are hesitant towards vaccines (S1 Appendix).
The survey questions concerning perceived benefit of vaccines, perceived safety of vaccines, and trust in health professionals, can be seen in. The questions on the benefits and safety of vaccines were created according to current official vaccination guidelines and evidence-based information.
Perceived benefits of vaccines. Nine statements were created to measure the HCWs' perceptions and knowledge of the benefits of vaccinations. Items were of varying polarity; e.g., "Childhood vaccines are effective in protecting against disease" vs. "It is not worth getting the influenza vaccine, as the influenza symptoms are not serious". Some of the questions queried knowledge about vaccines in general, whereas others concerned knowledge about specific vaccines or vaccine-preventable diseases (measles or influenza). The respondents were asked to rate how much they agreed with each statement on a scale from 1 (strongly disagree) to 5 (strongly agree).
Perceived safety of vaccines. Six statements were formulated to measure the perceived safety of vaccines (e.g., "Childhood vaccines are safe", "The risk of side effects outweighs the protective benefits of the influenza vaccines"). As with the questions regarding the perceived benefits of vaccines, some statements were related to vaccines in general, and others to specific vaccines. Two of the specific questions concerned common misinformation about vaccines; that vaccines could cause autism and that vaccines contain dangerous quantities of mercury. Again, the respondents rated how much they agreed with each statement on a scale from 1 (strongly disagree) to 5 (strongly agree).
Trust in health professionals. Four statements were created to measure the HCWs' trust in the intentions and professional competence of doctors and health professionals in general (e.g., "When healthcare professionals make medical decisions, they have the patients' best interest in mind"). The respondents were asked to rate how much they agreed with each statement on a scale from 1 (strongly disagree) to 5 (strongly agree).
## Own vaccination behavior.
To understand the HCWs' own vaccination behavior, the survey included 1) three questions on whether or not they had their children vaccinated with the childhood vaccines ("Have you ever hesitated in letting your child(ren) receive any of the childhood vaccines?", "Have you ever postponed a vaccination for your child(ren) with any of the childhood vaccines?", and "Have you ever decided not to let your child(ren) receive any of the childhood vaccines?"), and 2) one question probing whether the respondents had taken the influenza vaccine during the preceding influenza season. The questions concerning childhood vaccinations were administered only to respondents who reported having children.
Recommendation behavior. Respondents who reported discussing or administering vaccines to patients on a weekly basis were asked two questions about how they recommended vaccines to vaccine-hesitant patients. These questions were "How do you proceed if a parent is unsure about a vaccination decision concerning the childhood vaccines (and the child does not have any medical contraindications)?" and "How do you proceed if a patient is unsure about a vaccination decision concerning an influenza vaccine (and the patient does not have any medical contraindications)?" For each question, the respondents could choose between three response options; "I try to guide the parent/patient towards vaccinating", "I do not try to guide the parent/patient in any direction", or "I try to guide the parent/patient towards not vaccinating".
# Statistical analysis
We analyzed the data using structural equation modeling (SEM), or more specifically, structural regressions (SR). The four outcome measures constituted: 1) childhood vaccination decisions concerning own children, 2) own influenza vaccinations, 3) recommendation behavior concerning childhood vaccines, and 4) recommendation behavior concerning influenza vaccines. The first outcome variable was created based on the three questions regarding the HCWs' decisions on vaccines for their child(ren) (i.e., whether they had hesitated in a vaccination decision, postponed a vaccination, or rejected a vaccination altogether). The responses on the new variable were coded as an ordered factor where: 0 = no hesitancy, no postponing, and no rejection of a vaccination; 1 = hesitancy, but no postponing or rejection of a vaccination; 2 = postponing, but no rejection of a vaccination; or 3 = rejection of a vaccination. However, if a respondent had indicated that the reason for postponing a vaccination was that the child had been ill at the initial vaccination occasion, their response was coded as 0 (= no hesitancy). The response was coded as no hesitancy also if the reason for hesitating or rejecting a childhood vaccination was allergies or contra-indicative medical conditions. The outcome variable on the HCWs' own influenza vaccinations was coded as: 0 = had received the vaccine against influenza; or 1 = had not received the vaccine against influenza. If a respondent had indicated that the reason for not receiving the vaccination was allergies or other contra-indicative medical conditions, their response was again coded as 0. The two variables concerning recommendation behavior were coded as ordinal factors where: 0 = try to guide the parent/patient towards vaccinating; 1 = do not try to guide the parent/patient in any direction; and 2 = try to guide the parent/patient towards not vaccinating.
The outcome measures were included in separate models, and, hence, four SR models were fitted to the data. In the initial model specification, each model included an outcome measure as an observed variable that was regressed on three latent factors: 1) perceived benefits of vaccines (Benefit; nine indicators), 2) perceived safety of vaccines (Safety; six indicators), and 3) trust in health professionals (Trust; four indicators; see,for indicators specified to load on the factors). Before fitting the full SR models, we evaluated the fit of the measurement part of the models (the latent factors) by confirmatory factor analysis (CFA). The three latent factors were first analyzed in separate CFA models to investigate the unidimensionality of indicators, and then the three-factor model was evaluated.
We conducted the analyses using the lavaan package (version 0.6-3)in R (version 3.5.0). Due to the ordinal and categorical nature of the response variables, as well as the non-normal distribution of responses, the analyses were conducted using robust WLS (WLSMV) estimation with delta parameterization. The WLSMV estimator analyzes the asymptotic covariance matrix generated from the polychoric correlations between the indicators. Regression coefficients were estimated using the probit link function. The probit regression coefficient represents the change in the standard normal distribution (z-score) of the outcome variable, given a one-unit increase in the predictor. Missing data was handled with pair-wise deletion.
# Results
Altogether, 4286 individuals responded to the survey (response rate 49%). The present study only included those hospital personnel who may work with vaccinations, such as doctors, head nurses, nurses, and practical nurses. Other health professionals (e.g., psychologists and physiotherapists) and other personnel working at hospitals (e.g., within administration and human resource management) were excluded from the sample of respondents. Therefore, the final sample included 2962 HCWswith the mean age of 44.80 years (SD = 11.14, range = 20-67).
Only 0.91% of the total observations on the 19 questions measuring vaccination confidence (benefit, safety, and trust) were missing, while between 0.04% and 2.22% of the responses on the vaccination and recommendation behavior questions were missing (see, S1 .presents the HCWs' responses to the questions concerning perceived benefit and safety of vaccines, and trust in health professionals. In the table, reverse-scored items have been recoded so that all items have the same polarity (e.g., % Pos = positive attitudes towards vaccines or healthcare professionals). In general, the HCWs perceived vaccines to be beneficial and safe, and reported trust in health professionals. However, depending on the item, 1.7%-38.1% of the HCWs did not consider vaccines beneficial, 4.6%-25.7% reported that they did not consider the vaccines safe, and 3.9%-42.5% reported negative attitudes on the measures of trust. Compared with the whole sample, the HCWs who have the right to administer vaccines (doctors, head nurses, nurses), and who reported that they either discuss or administer vaccines on a weekly basis (n = 751), considered vaccines more beneficial and safe (see, S2 . Between 0.9% and 31.2% of those HCWs responded that they did not perceive vaccines as beneficial, and between 4.0% and 17.9% did not perceive vaccines as safe. Finally, between 1.8% and 38.6% reported low trust in health professionals.
The HCWs' responses to the questions on whether they accepted childhood vaccines for their own children and influenza vaccines for themselves, as well as whether they recommend childhood and influenza vaccines to hesitant patients, are shown in. Of the HCWs who reported having children (n = 2234), the great majority had never hesitated in a vaccination decision, or postponed or rejected a vaccination for their child (81.6%). The proportion of HCWs who had received the influenza vaccine the previous season was also high (86.2%).
Most of the HCWs whose duties involved discussing and/or administering vaccines on a weekly basis (n = 792 for childhood vaccines; n = 802 for influenza vaccines) indicated that they guide vaccine-hesitant patients towards accepting the childhood and influenza vaccines (85.7% and 73.6%, respectively).
## Main analyses
Evaluation of the measurement models. The fit of the three factors Benefit, Safety, and Trust, when analyzed in separate models, can be seen in S3 In the three-factor model, three error correlations were re-specified as freely estimated based on investigation of modification indices and theoretical considerations. The fit indices of the final measurement model indicated adequate fit, χ 2 (146) = 2328.94, CFI = .945, TLI = .936, RMSEA = .071, SRMR = .057, and all indicators had significant standardized factor loadings above .40. However, the correlation between the factors Benefit and Safety was very strong (r = .93). We therefore decided to collapse the two factors into one factor labeled BenefitSafety. The two-factor model also showed adequate fit, χ 2 (148) = 2424.56, CFI = .943, TLI = .934, RMSEA = .072, SRMR = .058. Again, all indicators had standardized loadings above .40. Statistical comparison of the three-factor and the two-factor models showed that the three-factor solution was a significantly better fit to the data, χ 2 diff (2) = 93.83, p < .001. However, due to the strong correlation between Benefit and Safety that indicated low discriminant validity between the factors, as well as the small difference in fit, we decided to retain the two-factor model. The correlation between the two remaining factors, BenefitSafety and Trust, was .74.
## Structural regression models
After having confirmed an acceptable measurement model, we fitted the four SR models to the data. All four SR models included the two latent variables BenefitSafety and Trust as predictors, as well as one of the four outcome variables that was regressed on the two latent factors. The model concerning influenza vaccination behavior was fit to the whole sample of HCWs (N = 2962). In this SR model, χ 2 (165) = 2630.38, CFI = .941, TLI = .932, RMSEA = .071, SRMR = .062, BenefitSafety was significantly related to whether the HCWs had received the influenza vaccine the previous season or not, β = -.79, 95% CI [-.89, -.69], SE = 0.05, Z = 16.16, p < .001. This indicated that individuals who perceived vaccines as beneficial and safe were more likely to have taken the influenza vaccine. Furthermore, Trust showed a statistically significant association with influenza vaccination behavior, β = .12, 95% CI [.01, .24], SE = 0.06, Z = 2.04, p = .041, indicating that, contrary to our hypothesis, the HCWs with lower trust were more likely to have taken the influenza vaccine. However, the association was very weak, and therefore, we will not consider this result further.
The models including the outcome variables related to recommendation behavior were fitted using the sample of HCWs who reported that their work duties involved discussing and/or administering vaccines on a weekly basis (n = 810; 27. We also fitted each SR model to the data excluding the practical nurses to investigate whether the pattern of associations would differ when including only individuals who have the right to administer vaccines. Practical nurses do not have the right to administer vaccines in Finland. The results from these control analyses were in line with the abovementioned findings for all other associations except for Trust, which was not significantly associated with the HCWs' own influenza vaccination status (see, S4 .
Follow-up analyses for professional groups. Because the results from the abovementioned SR models showed that BenefitSafety predicted both the HCWs' own vaccination behavior and recommend vaccines to hesitant patients, we investigated whether the professional groups differed in their perceptions of vaccines as beneficial and safe. In this follow-up SR analysis, BenefitSafety was set as the outcome variable, while the professions were compared using repeated contrasts according to the level of education (practical nurses vs. nurses; nurses vs. head nurses; head nurses vs. doctors). The analysis was run using the full sample of HCWs (n = 2962). The model, χ 2 (129) = 1832.25, CFI = .926, TLI = .940, RMSEA = .067, SRMR = .070, revealed significant differences for all comparisons. Practical nurses perceived vaccines less beneficial and safe than nurses did, β = -.16, 95% CI [-.19, -.12], SE = 0.02, Z = 8.82, p < .001, nurses considered vaccines less beneficial and safe than head nurses did, β = -.13, 95% CI [-.18, -.08], SE = 0.03, Z = 4.63, p < .001, and head nurses perceived vaccines less beneficial and safe than doctors did, β = -.23, 95% CI [-.28, -.18], SE = 0.03, Z = 8.79, p < .001.the HCWs' attitudes related to the statements included in the factor Benefit-Safety, divided by whether the statements queried general or specific knowledge. The bars represent the percentage of HCWs that report positive attitudes, averaged over the items. An inspection of the mean percentages for each profession separately, indicated that all professions agreed more with general statements compared with specific ones. However, the difference between the proportions of individuals agreeing with general and specific statements was smallest for doctors and increased as the educational level of the HCWs decreased.
# Discussion
In the present study we investigated if Finnish HCWs' perceptions of the benefits and safety of vaccines, and trust in health professionals, is related to their decisions to accept vaccinations for themselves and their children, and to their willingness to recommend vaccines to vaccinehesitant patients. The majority of the HCWs perceived vaccines to be beneficial and safe, and trusted health professionals. Nonetheless, a notable share of the HCWs questioned the benefits and safety of vaccines, and expressed distrust in the professional competence and intentions of health professionals. We review and explore the implications of the core findings.
## Own vaccination behavior
Most HCWs reported accepting all childhood vaccinations for their children (95.9%). However, 13.0% reported that they had hesitated in a vaccination decision, 6.3% that they had postponed a vaccination, and 4.1% that they had rejected a vaccination for their children. These results are in line with previous research suggesting that hesitancy towards vaccines exists among HCWs as well. The uptake of the influenza vaccine was high in the present sample of HCWs (86.2% vaccinated) compared to uptake rates reported in other countries. This might be due to the new Infectious Diseases Actand the work-related consequences the Finnish HCWs might face if they do not take the vaccine against influenza. However, data on the uptake of influenza vaccines among Finnish HCWs at Turku University Hospital (the personnel from this hospital was not included in the present study) collected before the act came into effect (influenza season 2015-2016) indicate a vaccination rate of 63%. In European countries where influenza vaccinations for HCWs are not mandatory, uptake rates are typically lower than 30%.
The results from the present study further showed that HCWs' perception and knowledge about the benefits and safety of vaccines were related to their own vaccination behavior, both in terms of allowing their children to get vaccinated in accordance with the national vaccination program, and in terms of accepting the vaccine against seasonal influenza for themselves. In agreement with previous research in the general population, as well as in HCW samples Vaccination confidence in Finnish healthcare workers, respondents with more positive attitudes towards vaccines were less likely to have hesitated, postponed or rejected vaccinations for their children, and were more likely to have accepted the influenza vaccine during the previous influenza season. The HCWs' trust in health professionals was not directly associated with the decision to vaccinate their children, when their perceptions of the benefits and safety of the vaccines were taken into account (see below).
## Recommendation behavior
Very few HCWs reported that they guide hesitant patients towards not vaccinating (0.5% and 0.4% for childhood and influenza vaccination respectively). However, 13.8% reported that they do not guide the patient in any direction when the hesitancy concerns the childhood vaccines. This figure was approximately twice as high (26.1%) when the patients' hesitancy concerned the influenza vaccine. The fact that a non-negligible proportion of the HCWs do not guide vaccine-hesitant patients towards taking the vaccine is worrying, as a lack of recommendation from health professionals is reported by laypeople as a reason for not taking the vaccine. In an attempt to clarify why some HCWs do not guide their patients in immunization decisions, we investigated whether recommendation behavior was associated with their own attitudes towards vaccines and other health professionals. This was found to be the case for vaccine attitudes, as HCWs with more positive views on vaccine benefits and safety were also more likely to guide hesitant patients towards accepting childhood and influenza vaccines. This relationship between vaccine attitudes and recommendation behavior has been observed in previous research. Trust in health professionals was not, however, directly related to recommendation behavior.
## Trust in health professionals
As previously mentioned, and contrary to our hypothesis, the HCWs' trust in health professionals was not directly associated with the HCWs' own vaccination behavior or their willingness to recommend vaccines to vaccine-hesitant patients. This finding stands in apparent conflict with the results from previous studies suggesting that HCWs with a higher trust in the health system are also more likely to take the influenza vaccine and to recommend vaccines to patients. The main difference between the present study and the two previous ones investigating this association is the operationalization of trust. In the present study, the statements designed to measure trust concerned the HCWs' trust in the professional competence and intentions of doctors and health professionals in general. In previous studies, on the other hand, trust was operationalized as trust in health authorities, scientific sources, and experts, or professional recommendations, concerning vaccine-related circumstances in particular. The difference between results might, hence, be due to the different aspects of trust that were measured. Another possible reason for the discrepancy in results relates to how the data were analyzed. In the present study, the correlation between the HCWs' perception of the benefits and safety of vaccines and their trust in health professionals was strong (r = .74), indicating that trust is strongly related to vaccine attitudes. In the multivariate regression model, however, where the two predictors (BenefitSafety and Trust) compete with one another, perceptions of the benefit and safety of vaccines showed a unique association with the outcome measures, while there was no unique relation of trust in health professionals on the outcome measures. In the study by Učakar et al, other variables were not controlled for. The analyses in the study by Collange et alincluded several predictors, of which some were not investigated in the present study (such as personal and professional characteristics). Furthermore, cultural differences between the investigated populations might explain the conflicting results, as trust issues are contextual and vary between countries. Also, the samples utilized by the abovementioned studies were more homogenous, including doctors only, while here, four different professional groups were included in the analyses.
## Profession
As the HCWs' perception and knowledge about the benefits and safety of vaccines predicted both their own vaccination behavior and their willingness to recommend vaccines to hesitant patients, we conducted follow-up analyses to investigate if the levels of confidence differed between the various professional groups. Those analyses showed statistically significant differences between all professional groups, with doctors being most confident, followed by head nurses, nurses, and practical nurses. This finding suggested that the degree to which the HCWs' perceptions were in line with scientific evidence, meaning that they perceived vaccines to be beneficial and safe, was positively related to their level of education. Also previous research has found that, in general, doctors have more positive attitudes towards vaccines than nurses. Furthermore, the difference between professions seemed to be greater for statements that required specific knowledge (i.e., related to specific vaccines or diseases) than for more general statements (i.e., related to vaccines in general;. In other words, the results indicated some uncertainty among HCWs when it comes to more specific knowledge about vaccines or the diseases they prevent. This was the case particularly for HCWs with lower levels of education. Taken together, these findings might indicate that confidence in the benefits and safety of vaccines among HCWs is related to the amount of medical training they have received. If this is the case, higher confidence may be achieved by providing HCWs with more vaccine-related education. The relationship between confidence and professional group can, however, also be explained by other factors, such as differences in professional identity or amount or type of vaccine-related work. This possible link between profession and vaccine confidence is an important question to be addressed in future research.
# Limitations
There are some important issues to take into account when interpreting the results of the present study. First, the results are based on self-reported attitudes and behaviors. Therefore, the responses might be affected by, for example, desirability bias or memory issues. Using official vaccination records might lead to more accurate measures of vaccination behavior in terms of uptake, and observing HCWs during their encounters with patients would yield more accurate and detailed information about their communication behavior. However, the methods utilized in the present study enabled the collection of information from a large sample of HCWs, which increases the generalizability of the results.
Second, the questionnaire employed was developed for the purpose of the present study and has not been validated in other samples. However, face validity of the questionnaire was assessed by experts in the field and factor analysis was utilized to evaluate the degree to which the questions loaded on the constructs and to handle measurement error.
Third, the data collection was cross-sectional and, hence causality cannot be inferred with certainty. There are, however, good reason to suppose that the confidence HCWs have in vaccines and healthcare, determines their vaccination and recommendation behavior, rather than the other way around.
Fourth, our study population did not include HCWs working at child health clinics. This limits the generalizability of our findings especially when it comes to the recommendation behavior related to childhood vaccines, as most childhood vaccinations in Finland are carried out at child health clinics by public health nurses. To what degree the present results apply also to HCWs working at child health clinics, is therefore an important topic for future studies.
# Conclusions
The results from the present study suggest that the majority of Finnish HCWs have high confidence in the benefits and safety of vaccines and show trust in other health professionals. However, low vaccination confidence was found among a non-negligible proportion of the HCWs. The results further showed that HCWs who perceived vaccines as less beneficial and safe were also less likely to have accepted vaccines for themselves and for their children, and were less willing to recommend vaccines to vaccine-hesitant patients. Trust in health professionals was not directly associated with the HCWs' own vaccination decisions or willingness to recommend vaccines.
Confidence in evidence-based information on vaccinations seemed to be related to the level of education among the HCWs, as the degree of confidence increased along with the educational level. This was the case in particular for statements requiring knowledge about specific vaccines or diseases. Further research should examine whether vaccination confidence can be increased by more vaccine-related education or training. Ensuring that HCWs have high confidence in vaccines may be important for maintaining high vaccine uptake in the general population, as assurance by HCWs is reported by laypeople as a key factor in their decision to get vaccinated.
## Supporting information
|
Respiratory Syncytial Virus Fusion Protein Promotes TLR-4–Dependent Neutrophil Extracellular Trap Formation by Human Neutrophils
Acute viral bronchiolitis by Respiratory Syncytial Virus (RSV) is the most common respiratory illness in children in the first year of life. RSV bronchiolitis generates large numbers of hospitalizations and an important burden to health systems. Neutrophils and their products are present in the airways of RSV-infected patients who developed increased lung disease. Neutrophil Extracellular Traps (NETs) are formed by the release of granular and nuclear contents of neutrophils in the extracellular space in response to different stimuli and recent studies have proposed a role for NETs in viral infections. In this study, we show that RSV particles and RSV Fusion protein were both capable of inducing NET formation by human neutrophils. Moreover, we analyzed the mechanisms involved in RSV Fusion protein-induced NET formation. RSV F protein was able to induce NET release in a concentration-dependent fashion with both neutrophil elastase and myeloperoxidase expressed on DNA fibers and F proteininduced NETs was dismantled by DNase treatment, confirming that their backbone is chromatin. This viral protein caused the release of extracellular DNA dependent on TLR-4 activation, NADPH Oxidase-derived ROS production and ERK and p38 MAPK phosphorylation. Together, these results demonstrate a coordinated signaling pathway activated by F protein that led to NET production. The massive production of NETs in RSV infection could aggravate the inflammatory symptoms of the infection in young children and babies. We propose that targeting the binding of TLR-4 by F protein could potentially lead to novel therapeutic approaches to help control RSV-induced inflammatory consequences and pathology of viral bronchiolitis.
# Introduction
Respiratory Syncytial Virus (RSV)-induced acute bronchiolitis is the most prevalent respiratory disease in children under age 2 years, and its seasonal epidemics are associated with a significant number of hospital admissions, with a huge burden to communities worldwide [bib_ref] Respiratory syncytial virus. Brief review, Stott [/bib_ref]. Almost 70% of all children are infected with RSV during the first year of life, and by age 3, practically all children will have experienced at least one infection with this virus [bib_ref] Review of epidemiology and clinical risk factors for severe respiratory syncytial virus..., Welliver [/bib_ref] [bib_ref] Respiratory syncytial virus infections: old challenges and new opportunities, Mejías [/bib_ref]. RSV is a single stranded RNA virus, whose genome encodes up to 11 proteins [bib_ref] Host immunity during RSV pathogenesis, Bueno [/bib_ref]. The Fusion (F) protein, present at the virion surface, mediates fusion of the viral envelope with the target cell membrane during virus entry. Only membrane-bound F protein is indispensable for virus replication in vitro and in vivo [bib_ref] Respiratory syncytial virus (RSV) SH and G proteins are not essential for..., Karron [/bib_ref] , and this protein is the primary target for both antiviral drug and vaccine developments [bib_ref] Antigenic relatedness between glycoproteins of human respiratory syncytial virus subgroups A and..., Johnson [/bib_ref] [bib_ref] Structural characterization of the human respiratory syncytial virus fusion protein core, Zhao [/bib_ref]. It has been demonstrated that RSV F protein activates pattern recognition receptors TLR-4 and CD14, inducing pro-inflammatory cytokine secretion [bib_ref] Pattern recognition receptors TLR4 and CD14 mediate response to respiratory syncytial virus, Popova [/bib_ref]. In addition, it has been recently shown that RSV F protein directly interacts with the MD-2-TLR-4 complex, thus activating the transcription factor NF-κB [bib_ref] Respiratory syncytial virus fusion protein-induced toll-like receptor 4 (TLR4) signaling is inhibited..., Rallabhandi [/bib_ref]. These studies highlight the importance of specific signaling pathways activated by F protein to stimulate inflammation.
One of the characteristic features of RSV infection is the large amounts of neutrophils in the lower airways once infection is established [bib_ref] Bronchoalveolar lavage cellularity in infants with severe respiratory syncytial virus bronchiolitis, Mcnamara [/bib_ref]. It is also well recognized that neutrophils and their products are present in the airways of patients and animal models with RSV bronchiolitis [bib_ref] Bronchoalveolar lavage cellularity in infants with severe respiratory syncytial virus bronchiolitis, Mcnamara [/bib_ref] [bib_ref] Respiratory syncytial virus affects pulmonary function in BALB/c mice, Van Schaik [/bib_ref] [bib_ref] Human neutrophil elastase in RSV bronchiolitis, Emboriadou [/bib_ref] , and also in virus-induced asthma [bib_ref] Role of nasal interleukin-8 in neutrophil recruitment and activation in children with..., Teran [/bib_ref] [bib_ref] Respiratory syncytial virus infection results in airway hyperresponsiveness and enhanced airway sensitization..., Schwarze [/bib_ref]. This body of evidence suggests that neutrophils play an important role in the pathogenesis observed in the airways of affected children [bib_ref] The interaction of neutrophils with respiratory epithelial cells in viral infection, Wang [/bib_ref] [bib_ref] Neutrophils induce damage to respiratory epithelial cells infected with respiratory syncytial virus, Wang [/bib_ref].
Aside from the traditional mechanisms of phagocytosis, generation of reactive oxygen species (ROS), and degranulation, neutrophils can also produce neutrophil extracellular traps (NETs), an important strategy to immobilize and kill pathogens [bib_ref] Neutrophil extracellular traps kill bacteria, Brinkmann [/bib_ref]. NETs are formed by decondensed chromatin fibers decorated with antimicrobial proteins, such as neutrophil elastase and myeloperoxidase [bib_ref] Neutrophil extracellular traps kill bacteria, Brinkmann [/bib_ref]. NET-inducing stimuli include cell surface components of bacteria, such as LPS, whole bacteria, fungi, protozoan parasites, cytokines, and activated platelets, among others [bib_ref] Neutrophil extracellular traps kill bacteria, Brinkmann [/bib_ref] [bib_ref] Leishmania amazonensis promastigotes induce and are killed by neutrophil extracellular traps, Guimarães-Costa [/bib_ref] [bib_ref] Platelet TLR4 activates neutrophil extracellular traps to ensnare bacteria in septic blood, Clark [/bib_ref] [bib_ref] NETs formed by human neutrophils inhibit growth of the pathogenic mold Aspergillus..., Mccormick [/bib_ref] [bib_ref] Neutrophil extracellular traps capture and kill Candida albicans yeast and hyphal forms, Urban [/bib_ref]. More recently, studies have demonstrated that viruses are also capable of inducing NET formation. In vitro, the production of NETs is modulated in neutrophils isolated from cats infected with feline immunodeficiency virus [bib_ref] Characterization of neutrophil extracellular traps in cats naturally infected with feline leukemia..., Wardini [/bib_ref]. NETs activated after infection by Human Immunodeficiency Virus (HIV-1) are crucial for the elimination of virus [bib_ref] Neutrophil extracellular traps mediate a host defense response to human immunodeficiency virus-1, Saitoh [/bib_ref]. NET release in the liver vasculature also protects host cells from poxvirus infection [bib_ref] Neutrophils recruited to sites of infection protect from virus challenge by releasing..., Jenne [/bib_ref]. However, an excessive production of NETs contributes to the pathology of respiratory viral infections. NET formation is potently induced in lungs of mice infected with Influenza A virus, in areas of alveolar destruction [bib_ref] Excessive neutrophils and neutrophil extracellular traps contribute to acute lung injury of..., Narasaraju [/bib_ref] , suggesting a putative role for NETs in lung damage.
We show that RSV virion was able to induce NET formation by human neutrophils and RSV F protein stimulated NET formation dependent on TLR-4 receptor activation. Moreover, F protein-induced NETs were decorated with neutrophil elastase and myeloperoxidase, granule proteins that can damage tissues. F protein potently induced NADPH Oxidase-derived ROS production and this was crucial for NET generation. Also, F protein induced NET production in an ERK and p38 MAPK phosphorylation-dependent manner. Together, these results provide compelling evidence to support a signaling mechanism activated by RSV F protein to induce NET formation. The massive production of NETs in the airways of children infected with RSV may worsen lung pathology and impair lung function.
## Materials and methods reagents
RSV A2 strain was provided by Dr. Fernando Polack (Vanderbilt University School of Medicine, USA). Human recombinant RSV Fusion protein was purchased from Sino Biological Inc. According to the manufacturer, the glycosylated protein purity is >95% and endotoxin level is <1.0 EU per 1 μg of the protein, as determined by the LAL method. PMA and Protease-free DNase 1 were from Promega. Dextran, LPS O111:B4 from Escherichia coli, Diphenyleneiodonium (DPI), N-acetyl-L-cysteine (NAC), and Histopaque-1077 were obtained from Sigma-Aldrich. ECORI and HINDIII were from Invitrogen. PD98059 and SB203580 were from Cayman Chemical. Polymyxin B and anti-RSV F protein (131-2A) were from Millipore. Blocking anti-TLR-4 (HTA125) and mouse IgG2a isotype control were from eBioscience. The 5-(and-6)-chloromethyl-2'-7'-dichlorodihydrofluorescein diaceate, acetyl ester (CM-H 2 DCFDA) was from Molecular Probes. RPMI 1640 was from Cultilab, and FCS was from Gibco.
## Human neutrophil isolation
Whole blood (20 mL) was collected from healthy volunteer donors (with a mean age of 29 years, from both sexes) with documented verbal consent into heparin-treated tubes. Erythrocytes were removed using Dextran sedimentation followed by two rounds of hypotonic lysis. Neutrophils were isolated from the resulting cell pellet using Histopaque-1077 density centrifugation and then resuspended in RPMI 1640 medium. Neutrophil purity was evaluated by flow cytometry using FACSCanto II (Becton Dickinson), based on morphology and a granulocyte marker expression, resulting in around 97%. Only singlet cells were verified by gating on granulocytes size on the basis of forward scatter (FSC) and side scatter (SSC), followed by CD66b and CD3 expression discrimination. Cell viability was always higher than 99%, as examined by Trypan Blue exclusion assay.
## Rsv preparation and neutrophil stimulation
The RSV A2 strain was grown in Hep-2 cells. Virus was purified from cell culture supernatant and the viral titer was determined by infection of Hep-2 cell monolayers followed by a carboxymethylcellulose plaque assay. The virus aliquots were stored in -80°C. Human neutrophils (2 x 10 6 /mL) were stimulated with RSV (10 2 -10 4 PFU/mL) for 3 h at 37°C under 5% CO 2 atmosphere. These RSV concentrations were used because higher concentrations were cytotoxic to neutrophils (data not shown). After this period, culture supernatants were collected and NETs were quantified using Quant-iT dsDNA HS kit (Invitrogen), according to manufacturer's instructions.
## Quantification of net release
Neutrophils (2 x 10 6 /mL) were stimulated with F protein (0.1-5 μg/mL), LPS (100 ng/mL), PMA (50 nM) or medium alone. After 1 h, 20 U/ml of each restriction enzyme (ECORI and HINDIII) was added to the cultures, and then kept for 2 h at 37°C, under 5% CO 2 atmosphere [bib_ref] Leishmania amazonensis promastigotes induce and are killed by neutrophil extracellular traps, Guimarães-Costa [/bib_ref]. NETs were quantified in culture supernatants using Quant-iT dsDNA HS kit (Invitrogen), according to manufacturer's instructions. To evaluate the involvement of TLR-4, NADPHoxderived ROS, and MAPK (ERK and p38) on F protein-induced NET formation, neutrophils were pretreated with selective inhibitors at 37°C under 5% CO 2 , as indicated in figure legends. The Trypan Blue exclusion assay was used to evaluate the viability of cells treated with these inhibitors, and at the end of incubation, the cellular viability was always higher than 97%.
## Immunofluorescence
Neutrophils (2 x 10 5 /300 μL) were incubated with F protein (1 μg/mL), LPS (100 ng/mL), PMA (50 nM) or medium alone for 3 h at 37°C under 5% CO 2 in 8-chamber culture slides (BD Falcon). After this period, cells were fixed with 4% paraformaldehyde (PFA) and stained with anti-elastase (1:1000; Abcam), followed by anti-rabbit Cy3 antibodies (1:500; Invitrogen) or anti-myeloperoxidase PE antibody (1:1000; BD Biosciences) and Hoechst 33342 (1:2000; Invitrogen). Confocal images were taken in a Zeiss LSM 5 Exciter microscope.
## Assay of intracellular ros generation
The determination of intracellular ROS generation was based on the oxidation of 0.5 μM 5-(and-6)-chloromethyl-2',7'-dichlorodihydrofluorescein diacetate, acetyl ester (CM-H 2 DCFDA) to yield an intracellular fluorescent compound. Neutrophils (2 x 10 6 cells/microtube) were pretreated with NAC (1 mM) or DPI (10 μM) and stimulated with F protein for 60 minutes at 37°C under 5% CO 2 . Afterwards, cells were incubated with CM-H 2 DCFDA for 30 minutes at 37°C under 5% CO 2 . Cytosolic ROS production was measured by flow cytometry using FACSCanto II flow cytometer (Becton Dickinson) with the BD FACSDiva software and analyzed with FlowJo v 7.5.
## Expression of phospho-erk and phospho-p38
The expression of phospho-ERK 1/2 and phospho-p38 in human neutrophils was measured by flow cytometry using BD Phosflow (BD Biosciences) protocol for human whole blood samples. Neutrophils were stimulated with F protein (1 μg/mL) for 5 minutes. Briefly, cells were fixed in Phosflow Buffer I for 10 minutes at 37°C. After washing, permeabilization was performed with Phosflow Perm Buffer II for 30 minutes on ice. Then, neutrophils were washed twice and stained with APC anti-phospho-ERK 1/2 and Alexa 488 anti-phospho-p38 antibodies for 30 minutes on ice. Also here, data were accessed by flow cytometry using FACSCanto II cytometer (Becton Dickinson) with BD FACSDiva software and analyzed with FlowJo v 7.5.
## Statistical analyses
Data were presented as mean ± SEM. Results were analyzed using GraphPad Prim 5.0 statistical software package. Statistical differences among the experimental groups were evaluated by analysis of variance with Newman-Keuls correction or with Student's t Test. The level of significance was set at p 0.05.
# Ethics statement
This study was reviewed and approved by the Research Ethics Committee of Pontifícia Universidade Católica do Rio Grande do Sul (CEP/PUCRS) under protocol nr. CEP 310.623. CEP/ PUCRS approved the use of verbal consent for this study and blood donors provided their verbal informed consent before blood collection. The authors have documented the verbal consent provided by the donors.
# Results
## Rsv particles and rsv fusion protein induce net formation
It has been previously shown that neutrophils and their products are present in the airways of patients and animals infected with RSV [bib_ref] Respiratory syncytial virus affects pulmonary function in BALB/c mice, Van Schaik [/bib_ref] [bib_ref] Human neutrophil elastase in RSV bronchiolitis, Emboriadou [/bib_ref] [bib_ref] Role of nasal interleukin-8 in neutrophil recruitment and activation in children with..., Teran [/bib_ref]. Furthermore, recent studies demonstrated that viruses are able to induce NET formation [bib_ref] Neutrophil extracellular traps mediate a host defense response to human immunodeficiency virus-1, Saitoh [/bib_ref] [bib_ref] Neutrophils recruited to sites of infection protect from virus challenge by releasing..., Jenne [/bib_ref]. Therefore, we sought to investigate whether RSV would be able to induce NET formation in human neutrophils by stimulating neutrophils with increasing concentrations of RSV and quantifying extracellular DNA after 3 h. Indeed, RSV was able to induce NET production in a concentration-dependent manner [fig_ref] Fig 1: RSV particles and RSV Fusion protein induce NET formation [/fig_ref]. RSV Fusion protein is essential for viral replication [bib_ref] Respiratory syncytial virus (RSV) SH and G proteins are not essential for..., Karron [/bib_ref] and it is known to activate human monocytes, inducing a pro-inflammatory response [bib_ref] Pattern recognition receptors TLR4 and CD14 mediate response to respiratory syncytial virus, Popova [/bib_ref]. We hypothesized that RSV F protein could play a role on NET production. To test that, human neutrophils were stimulated with different concentrations of F protein in vitro and after 3 h of incubation extracellular DNA was quantified in culture supernatants. RSV F protein induced NET formation in a dose-dependent manner, with the concentration of 1 μg/mL inducing the strongest response [fig_ref] Fig 1: RSV particles and RSV Fusion protein induce NET formation [/fig_ref]. In an alternative approach to demonstrate the production of extracellular DNA by F protein, we stimulated neutrophils with medium alone, LPS, PMA or F protein and performed confocal laser scanning microscopy analysis. All stimulants (F protein, PMA and LPS) were able to induce NET formation compared to medium alone [fig_ref] Fig 1: RSV particles and RSV Fusion protein induce NET formation [/fig_ref]. The expression of antimicrobial proteins on NETs is induced by different stimuli, including bacteria, fungi, and virus [bib_ref] Neutrophil extracellular traps kill bacteria, Brinkmann [/bib_ref] [bib_ref] Neutrophil extracellular traps mediate a host defense response to human immunodeficiency virus-1, Saitoh [/bib_ref] [bib_ref] Requirements for NADPH oxidase and myeloperoxidase in neutrophil extracellular trap formation differ..., Parker [/bib_ref] [bib_ref] Neutrophil extracellular traps contain calprotectin, a cytosolic protein complex involved in host..., Urban [/bib_ref]. We sought to characterize the composition of NETs induced by RSV F protein, analyzing it by immunostaining. F protein induced the formation of NETs containing the proteins from azurophilic granules, neutrophil elastase (NE) [fig_ref] Fig 1: RSV particles and RSV Fusion protein induce NET formation [/fig_ref] and myeloperoxidase (MPO) [fig_ref] Fig 1: RSV particles and RSV Fusion protein induce NET formation [/fig_ref] , which colocalized with extracellular DNA.
## Effect of different treatments on f protein-induced nets generation
To ensure that the structures visualized and quantified were in fact NETs, we stimulated neutrophils with F protein or LPS, as a control, and treated the cells with protease-free DNase. DNase treatment was able to dismantle NETs induced by both LPS and F protein [fig_ref] Fig 2: Fig 2 [/fig_ref] , indicating that those structures were made of DNA, and consequently NETs. A major concern when characterizing any putative activator of TLR is the possible presence of microbial-derived contaminants. LPS is the prototype TLR-4 agonist and it is among the most potent proinflammatory stimuli both in vivo and in vitro. To test whether the effect of F protein could be due to LPS contamination, we stimulated neutrophils with F protein or LPS in the presence or absence of polymyxin B and quantified extracellular DNA in culture supernatants. As expected, LPS-induced NET release was inhibited by polymyxin B, which has been previously shown to bind and neutralize LPS [bib_ref] Lipopolysaccharide: neutralization by polymyxin B shuts down the signaling pathway of nuclear..., Tsuzuki [/bib_ref]. In contrast, F protein was able to induce NET formation in the presence of polymyxin B [fig_ref] Fig 2: Fig 2 [/fig_ref] , indicating that the effect of F protein is not attributable to LPS contamination. Next, we treated F protein with proteinase K for 90 minutes, to digest the protein structure, or boiled F protein for 10 minutes at 100°C, to further exclude the possibility that the effect could be due to other heat-resistant contaminant. Both treatments profoundly inhibited F protein-induced NET formation [fig_ref] Fig 2: Fig 2 [/fig_ref] , confirming that only integral F protein is capable of inducing NET production. Finally, we treated the F protein solution with either a neutralizing antibody directed to RSV F protein or with an isotype-matched antibody and stimulated neutrophils with these preparations. The neutralized F protein was not able to induce NET release compared to the protein treated with the control antibody [fig_ref] Fig 2: Fig 2 [/fig_ref] , confirming its role on NET production.
## F protein-induced net formation is dependent on tlr-4 activation
RSV F protein activates the pattern recognition receptors TLR-4-CD14-MD-2 to induce the activation of the transcription factor NF-κB and proinflammatory cytokine secretion [bib_ref] Pattern recognition receptors TLR4 and CD14 mediate response to respiratory syncytial virus, Popova [/bib_ref] [bib_ref] Respiratory syncytial virus fusion protein-induced toll-like receptor 4 (TLR4) signaling is inhibited..., Rallabhandi [/bib_ref]. We hypothesized that F protein could activate TLR-4 to induce NET production. A blocking antibody against TLR-4 was used to define the involvement of this receptor on F protein-induced NET formation. Pretreatment of neutrophils with anti-TLR4 significantly inhibited the effect of F protein on NET release [fig_ref] Fig 3: Fig 3 [/fig_ref]. As an alternative approach to show the role of TLR-4 on NET formation by F protein, we visualized DNA fibers after pretreatment of cells with anti-TLR4. The release of DNA induced by F protein after pretreatment with the antibody is completely blocked [fig_ref] Fig 3: Fig 3 [/fig_ref]. These results indicate that RSV F protein induces NET formation via a TLR-4 activation pathway.
Essential role for NADPH Oxidase-derived ROS on F protein-induced NET generation NET release induced by various agents has been previously shown to depend on ROS generation [bib_ref] Requirements for NADPH oxidase and myeloperoxidase in neutrophil extracellular trap formation differ..., Parker [/bib_ref] [bib_ref] Novel cell death program leads to neutrophil extracellular traps, Fuchs [/bib_ref] [bib_ref] Neutrophil extracellular trap cell death requires both autophagy and superoxide generation, Remijsen [/bib_ref]. To characterize the involvement of ROS on F protein-induced NET production, neutrophils were treated with inhibitors of ROS generation. Treatment with NAC blocked Effect of different treatments on F protein-induced NETs generation. Human neutrophils (2 x 10 6 / mL) were stimulated with: (A) F protein (1 μg/mL) or LPS (100 ng/mL) in the presence or absence of DNase-1 (100U/mL); (B) F protein (1 μg/mL) or LPS (100 ng/mL) in the presence or absence of polymyxin B (Pmx B, 1 μg/mL); (C) F protein (1 μg/mL), boiled F protein (1 μg/mL, 10 min at 100°C) or F protein (1 μg/mL) treated with proteinase K (1 mg/mL for 90 min) for 3 h at 37°C with 5% CO 2 . (D) F protein solution was treated with monoclonal anti-F protein (10 μg/mL) or isotype-matched (10 μg/mL) antibody and neutrophils (2 x 10 6 /mL) were stimulated with these preparations for 3 h at 37°C with 5% CO 2 . NETs were quantified in culture supernatants using Quant-iT dsDNA HS kit. Data are representative of at least 2 independent experiments performed in triplicates and represent mean ± SEM. *p<0.05; ***p<0.001 when compared to negative control (NCtrl); #p<0.05 when compared to LPS-or F protein-treated cells. protein-induced NET formation is dependent on TLR-4 activation. (A) Human neutrophils (2 x 10 6 /mL) were pretreated with monoclonal anti-TLR4 (10 μg/mL) or isotype-matched (10 μg/mL) antibody for 1 h and stimulated with F protein (1 μg/mL) or medium for 3 h at 37°C with 5% CO 2 . NETs were quantified in culture supernatants using Quant-iT dsDNA HS kit. Data are representative of at least 3 separate experiments performed in triplicates and represent mean ± SEM. *p<0.001 when compared to negative control (NCtrl); #p<0.05 when compared to F protein-treated cells. (B) Neutrophils (2 x 10 5 /300 μL) were pretreated with anti-TLR4 (10 μg/mL) for 1 h at 37°C with 5% CO 2 and stimulated with F protein (1 μg/mL) or medium for 3 h at 37°C with 5% CO 2 . Cells were fixed with 4% PFA and stained with Hoechst 33342 . Confocal images were taken in a Zeiss LSM 5 Exciter microscope. Image is representative of 2 independent experiments. Scale bars = 50 μm. NET formation induced by F protein [fig_ref] Fig 4: Essential role for NADPH Oxidase-derived ROS on F protein-induced NET generation [/fig_ref] and abrogated F protein-induced ROS generation [fig_ref] Fig 4: Essential role for NADPH Oxidase-derived ROS on F protein-induced NET generation [/fig_ref]. Similarly, treatment with DPI, a NADPH Oxidase inhibitor, significantly inhibited F protein-stimulated NET production [fig_ref] Fig 4: Essential role for NADPH Oxidase-derived ROS on F protein-induced NET generation [/fig_ref] and abolished ROS generation induced by F protein [fig_ref] Fig 4: Essential role for NADPH Oxidase-derived ROS on F protein-induced NET generation [/fig_ref]. Together, these results indicate that F protein stimulates NET production dependent on NADPH Oxidase-derived ROS generation.
## F protein activates erk and p38 mapk to induce net formation
Recent studies have shown that ERK and p38 MAPK are indispensable for NET production [bib_ref] Activation of the Raf-MEK-ERK pathway is required for neutrophil extracellular trap formation, Hakkim [/bib_ref] [bib_ref] Reactive oxygen species-induced activation of ERK and p38 MAPK mediates PMA-induced NETs..., Keshari [/bib_ref]. To investigate the role of these MAPK on F protein-induced NET formation, we treated neutrophils with selective inhibitors of ERK and p38 MAPK. Pretreating neutrophils with PD98059 and SB203580, ERK and p38 inhibitors respectively, profoundly decreased DNA release induced by F protein [fig_ref] Fig 5: F protein activates ERK and p38 MAPK to induce NET formation [/fig_ref] , pointing to a critical role for these MAPK on F protein-induced NET formation. We also evaluated whether treatment of neutrophils with F protein would activate these signaling pathways, analyzing phosphorylation of ERK 1/2 and p38. F protein rapidly activated phosphorylation of these signaling pathways [fig_ref] Fig 5: F protein activates ERK and p38 MAPK to induce NET formation [/fig_ref] leading to NET release, thus supporting the results obtained with the inhibitors.
# Discussion
Neutrophils are key players in microbial containment due to their phagocytic properties, being able to deliver antimicrobial molecules in the phagolysosome and release neutrophil extracellular traps that entrap and kill a multitude of microorganisms [bib_ref] Neutrophil extracellular traps kill bacteria, Brinkmann [/bib_ref] [bib_ref] How neutrophils kill microbes, Segal [/bib_ref]. NETs are formed by a variety of stimuli, including bacteria, fungi, parasites, cytokines and endogenous proteins [bib_ref] Neutrophil extracellular traps kill bacteria, Brinkmann [/bib_ref] [bib_ref] HMGB1 promotes neutrophil extracellular trap formation through interactions with Toll-like receptor 4, Tadie [/bib_ref] [bib_ref] Neutrophil extracellular traps: double-edged swords of innate immunity, Kaplan [/bib_ref] [bib_ref] NETs: a new strategy for using old weapons, Papayannopoulos [/bib_ref]. Recent studies proposed a role for NETs in the control of viral infections [bib_ref] Neutrophil extracellular traps mediate a host defense response to human immunodeficiency virus-1, Saitoh [/bib_ref] [bib_ref] Neutrophils recruited to sites of infection protect from virus challenge by releasing..., Jenne [/bib_ref]. Neutrophilderived NETs were able to capture HIV-1 particles and this effect was dependent on TLR-7 and TLR-8 activation [bib_ref] Neutrophil extracellular traps mediate a host defense response to human immunodeficiency virus-1, Saitoh [/bib_ref]. Systemic injection of viral TLR ligands or poxvirus infection led to accumulation of neutrophils in liver sinusoids that formed aggregates with platelets and released NETs into the vessels [bib_ref] Neutrophils recruited to sites of infection protect from virus challenge by releasing..., Jenne [/bib_ref]. These studies point out to a beneficial role for NETs in controlling and neutralizing viral infection. However, the excessive formation of NETs could be pathogenic to the host, mainly in respiratory viral infections, because NETs could expand more easily in the pulmonary alveoli, causing lung injury. It has been recently shown that Influenza A virus induced the formation of NETs, entangled with alveoli in areas of tissue injury, suggesting a potential link with lung damage [bib_ref] Excessive neutrophils and neutrophil extracellular traps contribute to acute lung injury of..., Narasaraju [/bib_ref].
In this study we were able to demonstrate that RSV particle and one of its membraneexpressed glycoproteins potently induced NET formation. RSV F protein caused the release of NETs coated with granular proteins NE and MPO. These proteins have been shown to be important for NET formation [bib_ref] Neutrophil elastase and myeloperoxidase regulate the formation of neutrophil extracellular traps, Papayannopoulos [/bib_ref] [bib_ref] Myeloperoxidase is required for neutrophil extracellular trap formation: implications for innate immunity, Metzler [/bib_ref] and to possess microbicidal activities [bib_ref] NETs formed by human neutrophils inhibit growth of the pathogenic mold Aspergillus..., Mccormick [/bib_ref] [bib_ref] Neutrophil extracellular traps mediate a host defense response to human immunodeficiency virus-1, Saitoh [/bib_ref] [bib_ref] Requirements for NADPH oxidase and myeloperoxidase in neutrophil extracellular trap formation differ..., Parker [/bib_ref]. MPO present in NETs provides the bactericidal activity against S. aureus [bib_ref] Requirements for NADPH oxidase and myeloperoxidase in neutrophil extracellular trap formation differ..., Parker [/bib_ref] and promotes the elimination of HIV-1 [bib_ref] Neutrophil extracellular traps mediate a host defense response to human immunodeficiency virus-1, Saitoh [/bib_ref]. NE expressed in NETs induced by the pathogenic mold A. fumigatus helps to inhibit its growth [bib_ref] NETs formed by human neutrophils inhibit growth of the pathogenic mold Aspergillus..., Mccormick [/bib_ref]. However, the antimicrobial proteins released with NETs are directly toxic to tissues and the massive production of NETs may damage host tissues [bib_ref] NET balancing: a problem in inflammatory lung diseases, Cheng [/bib_ref] , as is the case for elastase, which cleaves host proteins at the site of inflammation or infection [bib_ref] Release of neutrophil elastase and its role in tissue injury in acute..., Fujie [/bib_ref]. Neutrophils actively producing NETs in the lung tissue disturb microcirculation and elicit pulmonary dysfunction [bib_ref] The clinical value of neutrophil extracellular traps, Lögters [/bib_ref]. Moreover, NETs directly induce epithelial and endothelial cell death [bib_ref] Neutrophil extracellular traps directly induce epithelial and endothelial cell death: a predominant..., Saffarzadeh [/bib_ref]. NE and MPO expressed on DNA fibers stimulated by F protein could exacerbate lung pathology induced by RSV infection, through the destruction of connective tissue, degradation of endothelial cell matrix heparan sulfate proteoglycan, resulting in post infection tissue injury [bib_ref] The clinical value of neutrophil extracellular traps, Lögters [/bib_ref].
The fibrous structure of NETs is essential for providing high local concentrations of antimicrobial proteins [bib_ref] Beneficial suicide: why neutrophils die to make NETs, Brinkmann [/bib_ref] , but it can also be detrimental for host tissues, since it can impair lung function [bib_ref] CXCR2 mediates NADPH oxidase-independent neutrophil extracellular trap formation in cystic fibrosis airway..., Marcos [/bib_ref]. Furthermore, the characterization of NETs structure is a great concern when studying these DNA lattices and their function. With two different approaches, the quantification of extracellular DNA and fluorescence microscopy, we demonstrated that RSV F protein-induced NETs were dismantled by DNase treatment, confirming that their structural backbone is chromatin.
Together with G protein, F protein comprises the major glycoprotein on RSV surface and these proteins are the main targets of neutralizing antibodies against RSV. F protein mediates the fusion of virus with the target cell and it is essential for viral replication both in vivo and in vitro [bib_ref] Respiratory syncytial virus (RSV) SH and G proteins are not essential for..., Karron [/bib_ref] , being considered the primary target for vaccine and antiviral drug development. Monoclonal antibodies to F protein passively protect against RSV challenge in an animal model and reduce the severity of infection in premature and newborn babies [bib_ref] humanized respiratory syncytial virus monoclonal antibody, reduces hospitalization from respiratory syncytial virus..., Palivizumab [/bib_ref] [bib_ref] Development of a humanized monoclonal antibody (MEDI-493) with potent in vitro and..., Johnson [/bib_ref]. A major feature of RSV infection is the large numbers of neutrophils recruited to the airways of patients and animals [bib_ref] Bronchoalveolar lavage cellularity in infants with severe respiratory syncytial virus bronchiolitis, Mcnamara [/bib_ref] [bib_ref] Respiratory syncytial virus affects pulmonary function in BALB/c mice, Van Schaik [/bib_ref] [bib_ref] Human neutrophil elastase in RSV bronchiolitis, Emboriadou [/bib_ref] [bib_ref] Local IL-17A potentiates early neutrophil recruitment to the respiratory tract during severe..., Stoppelenburg [/bib_ref]. This phenomenon is more profound than in any other respiratory viral infection in childhood, in which mostly alveolar macrophages and T cells prevail.
Although neutrophils are essential effector cells of the innate immune system and have a crucial role in the clearance of microorganisms [bib_ref] Neutrophils and immunity: challenges and opportunities, Nathan [/bib_ref] , it has been suggested that neutrophils may contribute to the pathology observed in the airways of patients and animals infected with RSV [bib_ref] The respiratory syncytial virus fusion protein and neutrophils mediate the airway mucin..., Stokes [/bib_ref]. Moreover, it has been shown that RSV is able to activate neutrophils, inducing degranulation and IL-8 secretion [bib_ref] Respiratory syncytial virus stimulates neutrophil degranulation and chemokine release, Jaovisidha [/bib_ref] and also inhibit neutrophil spontaneous apoptosis [bib_ref] Respiratory syncytial virus inhibits granulocyte apoptosis through a phosphatidylinositol 3-kinase and NF-kappaBdependent..., Lindemans [/bib_ref]. It is plausible to reason that these effects could be mediated by F protein binding to TLR-4, once it has been demonstrated that F protein binds to TLR-4/CD14 and physically interacts with MD-2, an essential accessory molecule for TLR-4 activation [bib_ref] Pattern recognition receptors TLR4 and CD14 mediate response to respiratory syncytial virus, Popova [/bib_ref] [bib_ref] Respiratory syncytial virus fusion protein-induced toll-like receptor 4 (TLR4) signaling is inhibited..., Rallabhandi [/bib_ref]. F protein induced NET formation in a TLR-4-dependent manner, since the treatment of neutrophils with a blocking antibody against TLR-4 profoundly inhibited extracellular DNA production. Our findings are in agreement with studies showing the activation of TLR-4 by different stimuli to induce NET generation [bib_ref] Neutrophil extracellular traps kill bacteria, Brinkmann [/bib_ref] [bib_ref] Platelet TLR4 activates neutrophil extracellular traps to ensnare bacteria in septic blood, Clark [/bib_ref] [bib_ref] HMGB1 promotes neutrophil extracellular trap formation through interactions with Toll-like receptor 4, Tadie [/bib_ref]. Importantly, the activation of TLR-4 by F protein was not attributable to LPS contamination, since the treatment with polymyxin B did not inhibit NET formation induced by the protein, but did inhibit the effect of LPS on NET generation. Furthermore, the native conformation of F protein was required in order to stimulate NET formation, once the boiled or proteinase K-digested protein lost this effect, as well as the neutralized protein by a monoclonal antibody directed against RSV F protein.
Stimulation of TLR-4 initiates a signal transduction cascade that induces the assembly of NADPH Oxidase complex. Several studies indicate that ROS are required for NET formation [bib_ref] Requirements for NADPH oxidase and myeloperoxidase in neutrophil extracellular trap formation differ..., Parker [/bib_ref] [bib_ref] Novel cell death program leads to neutrophil extracellular traps, Fuchs [/bib_ref] [bib_ref] Neutrophil extracellular trap cell death requires both autophagy and superoxide generation, Remijsen [/bib_ref]. Then, we sought to investigate whether F protein would be able to stimulate ROS production in neutrophils and whether this induction would be necessary for NET generation. Treatment with the ROS scavenger NAC abolished F protein-induced ROS and extracellular DNA production. Also, the oxidase inhibitor DPI, at the typical concentration needed to block the respiratory burst, completely blocked ROS production and NET formation induced by F protein. Thus, F protein-induced NET release is mediated by ROS generation. How ROS production contributes to DNA release is a question still open for debate. One possibility is that they promote the morphological changes seen in neutrophils secreting NETs [bib_ref] Neutrophil elastase and myeloperoxidase regulate the formation of neutrophil extracellular traps, Papayannopoulos [/bib_ref]. In addition, it has been suggested that ROS can act as second messengers [bib_ref] Redox redux: revisiting PTPs and the control of cell signaling, Tonks [/bib_ref]. The requirement of ROS for NET generation induced by RSV F protein indicate that ROS act as second messengers for this stimulus, likely promoting downstream events that culminate in DNA release.
Recent evidence shows that NET formation needs additional signaling, with the involvement of ERK and p38 MAPK. Furthermore, activation of these MAP kinases is downstream of NADPH Oxidase-derived ROS production [bib_ref] Activation of the Raf-MEK-ERK pathway is required for neutrophil extracellular trap formation, Hakkim [/bib_ref] [bib_ref] Reactive oxygen species-induced activation of ERK and p38 MAPK mediates PMA-induced NETs..., Keshari [/bib_ref]. We hypothesized that F protein would activate ERK and p38 MAPK to stimulate extracellular DNA release. Treatment of neutrophils with selective inhibitors of ERK and p38 MAPK almost abolished NET induction by F protein. Importantly, F protein was able to activate the phosphorylation of these MAP kinases. Taken together, these results indicate that RSV F protein-induced NET formation is mediated by the phosphorylation of p38 MAPK and ERK.
In conclusion, our study demonstrates that RSV particle stimulates NET formation by human neutrophils and RSV F protein is able to induce NET release through specific signaling pathways. This induction occurs through activation of TLR-4 and it is dependent on NADPH Oxidase-derived ROS generation, and on ERK and p38 MAPK phosphorylation . Neutrophils play an important role in the immunopathology during RSV infection and are continuously recruited from the bone marrow and blood stream to the lungs. The binding of RSV F protein to TLR-4 on neutrophils could induce the massive production of NETs, which can fill the lungs and impair lung function and consequently aggravate the inflammatory symptoms of the infection in young children and babies. We propose that targeting the binding of TLR-4 by F protein or even the associated use of DNase could potentially lead to novel therapeutic approaches to help control RSV-induced inflammatory consequences and pathology of viral bronchiolitis, which has a major disease burden among infants, worldwide.
[fig] Fig 1: RSV particles and RSV Fusion protein induce NET formation. (A) Human neutrophils (2 x 10 6 /mL) were stimulated with RSV (10 2 -10 4 PFU/mL) or left unstimulated for 3 h at 37°C with 5% CO 2 . (B) Neutrophils (2 x 10 6 /mL) were stimulated with RSV F protein (0.1-5 μg/mL), PMA (100 nM) or medium alone for 3 h at 37°C with 5% CO 2 . NETs were quantified in culture supernatants using Quant-iT dsDNA HS kit. Data are representative of at least 3 independent experiments performed in triplicates and represent mean ± SEM. *p<0.05; **p<0.01; ***p<0.001 when compared to negative control (NCtrl). (C-F) Neutrophils (2 x 10 5 / 300 μL) were stimulated with (C) medium, (D) LPS (100 ng/mL), (E) PMA (100 nM) or (F) F protein (1 μg/mL) for 3 h at 37°C with 5% CO 2 . Cells were then fixed with 4% PFA and stained with Hoechst 33342 (1:2000). Images are representative of at least 4 independent experiments. (G-L) Neutrophils (2 x 10 5 /300 μL) were stimulated with F protein (1 μg/mL) for 3 h at 37°C with 5% CO 2 . Cells were fixed with 4% PFA and stained with: (G-I) Hoechst 33342 (1:2000), anti-elastase (1:1000), followed by anti-rabbit Cy3 (1:500) antibodies; (J-L) Hoechst 33342 (1:2000), anti-myeloperoxidase PE (1:1000) antibody. Overlay of the fluorescence images are shown in the last panels (I,L). Images are representative of 2 independent experiments. Images were taken in a Zeiss LSM 5 Exciter microscope. Scale bars = 50 μm. doi:10.1371/journal.pone.0124082.g001 [/fig]
[fig] Fig 2: Fig 2. Effect of different treatments on F protein-induced NETs generation. Human neutrophils (2 x 10 6 / mL) were stimulated with: (A) F protein (1 μg/mL) or LPS (100 ng/mL) in the presence or absence of DNase-1 (100U/mL); (B) F protein (1 μg/mL) or LPS (100 ng/mL) in the presence or absence of polymyxin B (Pmx B, 1 μg/mL); (C) F protein (1 μg/mL), boiled F protein (1 μg/mL, 10 min at 100°C) or F protein (1 μg/mL) treated with proteinase K (1 mg/mL for 90 min) for 3 h at 37°C with 5% CO 2 . (D) F protein solution was treated with monoclonal anti-F protein (10 μg/mL) or isotype-matched (10 μg/mL) antibody and neutrophils (2 x 10 6 /mL) were stimulated with these preparations for 3 h at 37°C with 5% CO 2 . NETs were quantified in culture supernatants using Quant-iT dsDNA HS kit. Data are representative of at least 2 independent experiments performed in triplicates and represent mean ± SEM. *p<0.05; ***p<0.001 when compared to negative control (NCtrl); #p<0.05 when compared to LPS-or F protein-treated cells. [/fig]
[fig] Fig 3: Fig 3. F protein-induced NET formation is dependent on TLR-4 activation. (A) Human neutrophils (2 x 10 6 /mL) were pretreated with monoclonal anti-TLR4 (10 μg/mL) or isotype-matched (10 μg/mL) antibody for 1 h and stimulated with F protein (1 μg/mL) or medium for 3 h at 37°C with 5% CO 2 . NETs were quantified in culture supernatants using Quant-iT dsDNA HS kit. Data are representative of at least 3 separate experiments performed in triplicates and represent mean ± SEM. *p<0.001 when compared to negative control (NCtrl); #p<0.05 when compared to F protein-treated cells. (B) Neutrophils (2 x 10 5 /300 μL) were pretreated with anti-TLR4 (10 μg/mL) for 1 h at 37°C with 5% CO 2 and stimulated with F protein (1 μg/mL) or medium for 3 h at 37°C with 5% CO 2 . Cells were fixed with 4% PFA and stained with Hoechst 33342 (1:2000). Confocal images were taken in a Zeiss LSM 5 Exciter microscope. Image is representative of 2 independent experiments. Scale bars = 50 μm. [/fig]
[fig] F: ig 3. protein-induced NET formation is dependent on TLR-4 activation. (A) Human neutrophils (2 x 10 6 /mL) were pretreated with monoclonal anti-TLR4 (10 μg/mL) or isotype-matched (10 μg/mL) antibody for 1 h and stimulated with protein (1 μg/mL) or medium for 3 h at 37°C with 5% CO 2 . NETs were quantified in culture supernatants using Quant-iT dsDNA HS kit. Data are representative of at least 3 separate experiments performed in triplicates and represent mean ± SEM. *p<0.001 when compared to negative control (NCtrl); #p<0.05 when compared to protein-treated cells. (B) Neutrophils (2 x 10 5 /300 μL) were pretreated with anti-TLR4 (10 μg/mL) for 1 h at 37°C with 5% CO 2 and stimulated with protein (1 μg/mL) or medium for 3 h at 37°C with 5% CO 2 . Cells were fixed with 4% PA and stained with Hoechst 33342 (1:2000). Confocal images were taken in a Zeiss LSM 5 Exciter microscope. Image is representative of 2 independent experiments. Scale bars = 50 μm. [/fig]
|
Guidelines for Micro–Computed Tomography Analysis of Rodent Dentoalveolar Tissues
Micro-computed tomography (μCT) has become essential for analysis of mineralized as well as nonmineralized tissues and is therefore widely applicable in the life sciences. However, lack of standardized approaches and protocols for scanning, analyzing, and reporting data often makes it difficult to understand exactly how analyses were performed, how to interpret results, and if findings can be broadly compared with other models and studies. This problem is compounded in analysis of the dentoalveolar complex by the presence of four distinct mineralized tissues: enamel, dentin, cementum, and alveolar bone. Furthermore, these hard tissues interface with adjacent soft tissues, the dental pulp and periodontal ligament (PDL), making for a complex organ. Drawing on others' and our own experience analyzing rodent dentoalveolar tissues by μCT, we introduce techniques to successfully analyze dentoalveolar tissues with similar or disparate compositions, densities, and morphological characteristics. Our goal is to provide practical guidelines for μCT analysis of rodent dentoalveolar tissues, including approaches to optimize scan parameters (filters, voltage, voxel size, and integration time), reproducibly orient samples, define regions and volumes of interest, segment and subdivide tissues, interpret findings, and report methods and results. We include illustrative examples of analyses performed on genetically engineered mouse models with phenotypes in enamel, dentin, cementum, and alveolar bone. The recommendations are designed to increase transparency and reproducibility, promote best practices, and provide a basic framework to apply μCT analysis to the dentoalveolar complex that can also be extrapolated to a variety of other tissues of the body.
# Introduction
M icro-computed tomography (μCT) analysis has evolved into the gold standard for evaluating mineralized tissue microarchitecture in rodent research models, complementing and in some respects surpassing the capabilities of traditional histomorphometry by yielding data on two-dimensional (2D) and three-dimensional (3D) morphology, volume, and mineral density of skeletal elements. [bib_ref] Micro-CT of rodents: state-of-the-art and future perspectives, Clark [/bib_ref] [bib_ref] Micro-CT analysis of the rodent jaw bone micro-architecture: a systematic review, Faot [/bib_ref] [bib_ref] Analysis of bone architecture in rodents using micro-computed tomography, Van't Hof [/bib_ref] [bib_ref] Quantitative analysis of bone and soft tissue by micro-computed tomography: applications to..., Campbell [/bib_ref] [bib_ref] Small animal imaging and examination by micro-CT, Vasquez [/bib_ref] [bib_ref] Morphometric analysis of human bone biopsies: a quantitative structural comparison of histological..., Muller [/bib_ref] [bib_ref] Validation of cortical bone mineral density distribution using micro-computed tomography, Mashiatulla [/bib_ref] The nondestructive approach, relatively short turnaround time, high throughput, and volumetric analyses are particularly appealing for those researchers analyzing bone cortical and trabecular architecture. In the early years of μCT analysis in skeletal research, lack of consistent methods and reporting made it difficult to critique analytical approaches, interpret some findings, and compare results across multiple studies. A timely publication in 2010 articulated specific and practical guidelines to standardize scanning parameters, analysis approaches, and reporting nomenclature for long bone (ie, typically femur or tibia) μCT analysis. [bib_ref] Guidelines for assessment of bone microstructure in rodents using micro-computed tomography, Bouxsein [/bib_ref] That work has become a resource for those learning skeletal μCT analysis, planning μCT-based studies, and citing standard methods.
As μCT use has expanded, analyses have grown to encompass complex structures featuring mineralized and nonmineralized tissues. An example is the dentoalveolar complex, which is unique and challenging to analyze because it features four distinct, sometimes contiguous, mineralized tissues: enamel, dentin, cementum, and alveolar bone. These hard tissues interface with adjacent soft tissues, the dental pulp and periodontal ligament (PDL), adding further complications. Some publications have focused on applications of μCT in the craniofacial region or oral cavity [bib_ref] Estimating mineral changes in enamel formation by ashing/BSE and microCT, Schmitz [/bib_ref] [bib_ref] Micro-computed tomography for evaluating alveolar bone resorption induced by hyperocclusion, Tsutsumi [/bib_ref] [bib_ref] Three-dimensional microcomputed tomographic imaging of alveolar bone in experimental bone loss or..., Park [/bib_ref] [bib_ref] Using micro-computed tomography to evaluate the dynamics of orthodontically induced root resorption..., Xu [/bib_ref] [bib_ref] A robust methodology for the quantitative assessment of the rat jawbone microstructure, Chatterjee [/bib_ref] [bib_ref] Microcomputed tomography of craniofacial mineralized tissue: a practical user's guide to study..., Cox [/bib_ref] [bib_ref] State of the art of micro-CT applications in dental research, Swain [/bib_ref] ; though, to our knowledge only one publication has addressed more specific protocols for application in odontogenesis studies. [bib_ref] Microcomputed tomography imaging in odontogenesis studies, Verdelis [/bib_ref] However, despite current advances, the lack of standardized approaches for scanning, analyzing, and reporting data makes it difficult to understand how analyses were performed in many projects, how results should be interpreted, and if findings can be broadly compared to other models and studies, as well as making entry into μCT analyses challenging for researchers who are novices in the approach. We review considerations associated with analyzing the dentoalveolar complex, which is composed of tissues with similar and disparate compositions, densities, and morphological characteristics. To address these challenging complexities, we introduce innovative techniques that can also be extrapolated to other areas of the body.
Our goal in this report is to provide detailed and practical guidelines for μCT analysis of the dentoalveolar complex in order to increase transparency and reproducibility, promote best practices, and provide a basic framework for researchers who may benefit from such considerations. Using illustrative examples, we include a discussion of how to optimize scan parameters, reproducibly reorient samples, define regions and volumes of interest, segment and subdivide dentoalveolar tissues, interpret findings, and report methods and results. We focus on Mus musculus because of the many preclinical advantages, including ease in testing of therapeutic agents, reduced experimental time, flexibility in genetic manipulation, availability of established challenge and wound healing models, and accessibility of molecular reagents targeting murine models. Furthermore, the μCT concepts presented here can be broadly applied to large animal models and human cone beam computed tomography (CBCT) imaging. [bib_ref] Effects of daily treatment with parathyroid hormone on bone microarchitecture and turnover..., Dempster [/bib_ref] [bib_ref] Laboratory x-ray micro-computed tomography: a user guideline for biological samples, Du Plessis [/bib_ref] [bib_ref] 3D micro-computed tomography of trabecular and cortical bone architecture with application to..., Laib [/bib_ref] [bib_ref] Translation from murine to human lung imaging using x-ray dark field radiography:..., Vignero [/bib_ref] [bib_ref] Genetic engineering a large animal model of human hypophosphatasia in sheep, Williams [/bib_ref] [bib_ref] Micro-computed tomography-current status and developments, Ritman [/bib_ref] [bib_ref] Hypercementosis associated with ENPP1 mutations and GACI, Thumbigere-Math [/bib_ref] We rely on previous publications [bib_ref] Microcomputed tomography of craniofacial mineralized tissue: a practical user's guide to study..., Cox [/bib_ref] [bib_ref] Microcomputed tomography imaging in odontogenesis studies, Verdelis [/bib_ref] to supply detailed technical explanations of concepts in order to focus on recommendations tailored to special complexities and challenges featured in dentoalveolar tissues.
## Overview of the dentoalveolar complex
In this report, we focus on analysis of the mouse dentoalveolar complex, which includes the teeth and their supporting periodontal structures. We focus on the rodent mandibular first molar, the most commonly analyzed tooth and a reasonable model for human tooth development. [bib_ref] Molecular regulation of tooth development, Thesleff [/bib_ref] [bib_ref] Genetic basis for tooth malformations: from mice to men and back again, Mitsiadis [/bib_ref] [bib_ref] Three-dimensional analysis of molar development in the mouse from the cap to..., Lesot [/bib_ref] Mice have a dental pattern of 1 j 3 in each quadrant, representing one (continuously erupting) incisor, a large diastema (toothless region), and three molars [fig_ref] Fig 1: Murine dentoalveolar anatomy [/fig_ref]. As in human teeth, the mouse molar is divided into the crown, that portion visible above the gingiva, and roots that extend into the sockets formed by the maxilla or mandible. Four mineralized tissues are found in the dentoalveolar complex: enamel, dentin, cementum, and alveolar bone [fig_ref] Fig 1: Murine dentoalveolar anatomy [/fig_ref]. Other anatomical locations in the mouse craniofacial complex are amenable to μCT analysis, but will not be featured in this report, including the continually erupting incisors, condyle and temporomandibular joint, basal bone not directly associated with the dentition, and the remainder of the cranium.
## Enamel
Enamel is the hardest and the most highly mineralized substance in the body, composed of >95% mineral by weight. [bib_ref] Enamel formation and amelogenesis imperfecta, Hu [/bib_ref] [bib_ref] Dental enamel formation and implications for oral health and disease, Lacruz [/bib_ref] [bib_ref] The structural biology of the developing dental enamel matrix, Fincham [/bib_ref] Enamel comprises the outermost layer of the tooth crown. During amelogenesis, the initial organic matrix deposited by ameloblasts achieves its full thickness and hydroxyapatite ribbons thicken and replace protein content. [bib_ref] Enamel formation and amelogenesis imperfecta, Hu [/bib_ref] In its final mineralized state, ameloblasts are no longer present. Unlike humans, in which the enamel covers the entire crown, in rodents, enamel only covers part of the molar crown. Clearly visible on dental radiographs due to its radiopacity associated with its extremely high mineral content [fig_ref] Fig 1: Murine dentoalveolar anatomy [/fig_ref] , the borders of fully mature enamel are also easily identifiable on μCT scans.
## Dentin
Dentin forms the bulk of the tooth crown and root, lies under the enamel, and surrounds the pulp chamber. [bib_ref] Dentin: structure, composition and mineralization, Goldberg [/bib_ref] Dentin is composed of a network of radiating tubules created by odontoblasts that lie at the dentin-pulp border. Dentin is approximately 70% mineral by weight, however mineral content changes with location within dentin, age, maturation, and disease processes. [bib_ref] Dentin: structure, composition and mineralization, Goldberg [/bib_ref] [bib_ref] Maturational changes in dentin mineral properties, Verdelis [/bib_ref] Dentin is easily differentiated from enamel by radiograph due to relatively lower mineral content [fig_ref] Fig 1: Murine dentoalveolar anatomy [/fig_ref]. Dentin encloses the pulp chamber that includes unmineralized collagen matrix, cells, blood vessels, and nerves.
## Cementum
Cementum envelopes the root dentin and is essential for anchoring PDL collagen fibers to the tooth, providing attachment and mechanical stability. [bib_ref] Advances in defining regulators of cementum development and periodontal regeneration, Foster [/bib_ref] [bib_ref] Are cementoblasts a subpopulation of osteoblasts or a unique phenotype?, Bosshardt [/bib_ref] Cementum is found primarily in two forms: acellular cementum covers the cervical portions of roots, and cellular cementum is localized to apical portions of roots and includes osteocyte-like cementocytes. Cementum is approximately 45% to 50% mineral by weight, slightly less mineralized than dentin. Their adjacent position and similar mineral content make dentin and cementum difficult to differentiate in radiographs [fig_ref] Fig 1: Murine dentoalveolar anatomy [/fig_ref]. On the exterior, cementum is surrounded by unmineralized PDL, and therefore the cementum-PDL border is easily defined.
## Alveolar bone
The tooth-associated bone that forms the sockets is referred to as alveolar bone, distinct from underlying basal bone of the maxilla or mandible [fig_ref] Fig 1: Murine dentoalveolar anatomy [/fig_ref]. [bib_ref] Molecular and cellular biology of alveolar bone, Sodek [/bib_ref] Like cementum on the root surface, PDL fibers insert into the bundle bone layer as Sharpey's fibers. Bone mineral density can vary widely by site, age, maturation, and other factors; however, on average, bone is estimated to be 50% to 60% mineral by weight. [bib_ref] Molecular and cellular biology of alveolar bone, Sodek [/bib_ref] [bib_ref] Bone composition: relationship to bone fragility and antiosteoporotic drug effects, Boskey [/bib_ref] As bone is separated from dental tissues by the unmineralized PDL, it is easily distinguished from the other three hard tissues. The cementum-PDLalveolar bone (along with gingiva) forms the periodontal complex that attaches and supports the tooth. [bib_ref] The periodontal ligament: a unique, multifunctional connective tissue, Beertsen [/bib_ref]
## Image acquisition of dentoalveolar tissues
Because the dentoalveolar complex features four distinct mineralized tissues with different compositions, it is imperative to optimize μCT scanning parameters for assessment of tissue-specific effects resulting from modulation of genes, proteins, environment, trauma, aging, etc. Optimal parameters to visualize one tissue may not work well for others because of differences in mineral densities. Numerous reviews have described in detail the concepts underlying μCT scanning [bib_ref] Quantitative analysis of bone and soft tissue by micro-computed tomography: applications to..., Campbell [/bib_ref] [bib_ref] Small animal imaging and examination by micro-CT, Vasquez [/bib_ref] [bib_ref] Microcomputed tomography of craniofacial mineralized tissue: a practical user's guide to study..., Cox [/bib_ref] [bib_ref] Laboratory x-ray micro-computed tomography: a user guideline for biological samples, Du Plessis [/bib_ref] [bib_ref] Micro-computed tomography-current status and developments, Ritman [/bib_ref] [bib_ref] Micro-computed tomography: a method for the non-destructive evaluation of the three-dimensional structure..., Stauber [/bib_ref] [bib_ref] Microcomputed tomography: approaches and applications in bioengineering, Boerckel [/bib_ref] [bib_ref] Spectral optimization for micro-CT, Hupfer [/bib_ref] [bib_ref] Application of micro-CT in small animal imaging, Schambach [/bib_ref] ; therefore, we will introduce these only briefly in order to focus on specific challenges related to scanning the dentoalveolar complex. There are several approaches for sample preparation; however, for results shown here we used a standard method employing formalin-fixed tissues scanned under aqueous conditions in 70% ethanol, in order to maintain samples at biologically relevant levels of hydration and avoid cracking of mineralized tissues. There are a variety of μCT scanners available for purchase or use in core facilities. All results included here are obtained from a Scanco μCT 50 ex vivo cabinet system (Scanco Medical, Brüttisellen, Switzerland). Each scanner model will feature different capabilities and limitations (eg, resolution, filters, space dictating size of analyzed samples, density capacities, and fields of view), and therefore will require its own optimization for dentoalveolar tissues. However, the optimization steps outlined below can be adapted to all scanners based on our detailed evaluation criteria of scan results shown in [fig_ref] Fig 2: Optimization of μCT scan parameters for dentoalveolar structures [/fig_ref]. There are numerous options for analysis software, including proprietary applications sold with scanners. [bib_ref] Microcomputed tomography imaging in odontogenesis studies, Verdelis [/bib_ref] [bib_ref] Laboratory x-ray micro-computed tomography: a user guideline for biological samples, Du Plessis [/bib_ref] [bib_ref] Microcomputed tomography: approaches and applications in bioengineering, Boerckel [/bib_ref] [bib_ref] Application of micro-CT in small animal imaging, Schambach [/bib_ref] [bib_ref] Automated quantitative bone analysis in in vivo X-ray micro-computed tomography, Behrooz [/bib_ref] These also have their own capabilities and limitations that dictate to some degree the ease and efficiency of analysis. All analyses and images included in this report were accomplished using Scanco proprietary software for scan reconstructions (default settings) and AnalyzePro 1.0 medical imaging software (Analyze Direct, Overland Park, KS), a software system used in our studies. [bib_ref] Hypercementosis associated with ENPP1 mutations and GACI, Thumbigere-Math [/bib_ref] [bib_ref] Loss of discoidin domain receptor 1 predisposes mice to periodontal breakdown, Chavez [/bib_ref] [bib_ref] Genetic and pharmacologic modulation of cementogenesis via pyrophosphate regulators, Chu [/bib_ref] [bib_ref] MMP20 overexpression disrupts molar ameloblast polarity and migration, Shin [/bib_ref] [bib_ref] Reduced orthodontic tooth movement in Enpp1 mutant mice with hypercementosis, Wolf [/bib_ref] [bib_ref] Dental and craniofacial defects in the Crtap( −/− ) mouse model of..., Xu [/bib_ref] [bib_ref] Dentoalveolar defects in the Hyp mouse model of X-linked hypophosphatemia, Zhang [/bib_ref] CT images are created using X-rays directed and rotated around a sample (or a sample rotating in relation to the X-ray source), creating a series of projections. X-rays are a form of electromagnetic radiation with high photon energy, and image generation relies on the partial transmission of X-rays; ie, differential absorption of X-rays by tissues. Based on previous reports, common scanning parameters optimized for image acquisition are: inclusion of a filter, voltage, voxel size, and integration time, all of which alter the absorption of X-rays by a sample. [bib_ref] Analysis of bone architecture in rodents using micro-computed tomography, Van't Hof [/bib_ref] [bib_ref] Quantitative analysis of bone and soft tissue by micro-computed tomography: applications to..., Campbell [/bib_ref] [bib_ref] Small animal imaging and examination by micro-CT, Vasquez [/bib_ref] [bib_ref] Microcomputed tomography imaging in odontogenesis studies, Verdelis [/bib_ref] [bib_ref] Laboratory x-ray micro-computed tomography: a user guideline for biological samples, Du Plessis [/bib_ref] [bib_ref] Micro-computed tomography-current status and developments, Ritman [/bib_ref] [bib_ref] Microcomputed tomography: approaches and applications in bioengineering, Boerckel [/bib_ref] [bib_ref] Spectral optimization for micro-CT, Hupfer [/bib_ref] [bib_ref] Application of micro-CT in small animal imaging, Schambach [/bib_ref] [bib_ref] Effect of micro-computed tomography voxel size and segmentation method on trabecular bone..., Christiansen [/bib_ref] [bib_ref] Effect of integration time on the morphometric, densitometric and mechanical properties of..., Oliviero [/bib_ref] In order to compare images acquired from different scanning parameters, we scanned the same 6-week-old mouse mandible, systematically varying filter, voltage, voxel size, and integration time [fig_ref] Fig 2: Optimization of μCT scan parameters for dentoalveolar structures [/fig_ref] , [fig_ref] Table 1: Selection of μCT Scan ParametersVariables including filter, voltage, voxel size, and integration... [/fig_ref]. Scans were calibrated against a set of five hydroxyapatite (HA) phantoms of known density (0, 100, 200, 400, and 800 mg HA/cm 3 ) scanned with the same parameters. This density calibration using calibration phantoms is routine and essential for quantitative μCT, mimicking the attenuation of different tissue types and densities, and providing a standard curve to calculate absolute or relative density units for samples, allowing cross-study comparisons and helping detect and control for potential changes in the scanner X-ray source over time. Calibration phantoms are often composed of resin-embedded HA. Although the μCT scanner measures only X-ray attenuation and does not identify HA per se, HA is the major mineral component in bones and teeth. Thus, the assumption is made that attenuation from vertebrate skeletal and dental materials will match that of HA phantoms in linear fashion (in some situations, the nature of the mineral may need to be confirmed by additional techniques such as spectroscopy or electron microscopy). Importantly, calibration curves have been validated even for highly mineralized enamel. [bib_ref] Microcomputed tomography calibration using polymers and minerals for enamel mineral content quantitation, Alyahya [/bib_ref] [bib_ref] Characterization of a novel calibration method for mineral density determination of dentine..., Zou [/bib_ref] The limitations of density calibration phantoms, their correlation to tissue mineral content, and other factors have been tested and discussed in detail elsewhere, and these sources can be consulted for more detailed information on the subject. [bib_ref] Micro-CT of rodents: state-of-the-art and future perspectives, Clark [/bib_ref] [bib_ref] Validation of cortical bone mineral density distribution using micro-computed tomography, Mashiatulla [/bib_ref] [bib_ref] Estimating mineral changes in enamel formation by ashing/BSE and microCT, Schmitz [/bib_ref] [bib_ref] Quantitative microcomputed tomography: a non-invasive method to assess equivalent bone mineral density, Nazarian [/bib_ref] [bib_ref] Quantitative assessment of bone tissue mineralization with polychromatic microcomputed tomography, Burghardt [/bib_ref] [bib_ref] Preparation and characterization of calibration standards for bone density determination by micro-computed..., Schweizer [/bib_ref] [bib_ref] Further improvements on the factors affecting bone mineral density measured by quantitative..., Entezari [/bib_ref] [bib_ref] Accuracy of volumetric bone mineral density measurement in high-resolution peripheral quantitative computed..., Sekhon [/bib_ref] [bib_ref] Specimen size and porosity can introduce error into microCT-based tissue mineral density..., Fajardo [/bib_ref] [bib_ref] Quantitative CT for determination of bone mineral density: a review, Cann [/bib_ref] Geometric calibration of the μCT scanner is also critical for proper reconstruction of 2D sample data, reduction of scan artifacts, and optimization of resolution. Scanners typically have a geometric calibration protocol that should be operated on a regular schedule, and additional references can be consulted for more information on this topic. [bib_ref] Geometric calibration for a dual tube/detector micro-CT system, Johnston [/bib_ref] [bib_ref] Accurate technique for complete geometric calibration of cone-beam computed tomography systems, Cho [/bib_ref] Here we illustrate the effects of changing scanning parameters in optimization scans of mouse mandibles by showing Xray absorption as a grayscale image, a density heat map, and a voxel frequency distribution in mg HA/cm 3 . Each reconstructed grayscale image is composed of millions of voxels, which vary depending on scan parameters. Heat maps were generated to demonstrate density distributions, with 500 mg HA/cm 3 corresponding to red and 1600 mg HA/cm 3 to lilac [fig_ref] Fig 2: Optimization of μCT scan parameters for dentoalveolar structures [/fig_ref]. When counts assigned to each value (in mg HA/cm 3 ) are plotted on a frequency graph, an overwhelming number of voxels are considered background noise. Each frequency graph depicts two main peaks: the taller peak at a lower density range corresponds to dentin/bone/cementum, and the shorter peak at a higher density range corresponds to enamel (top right of each graph in [fig_ref] Fig 2: Optimization of μCT scan parameters for dentoalveolar structures [/fig_ref]. To optimize scanning parameters, images were evaluated based on the amount of noise, contrast between mineralized tissues, and clarity of structures.
## Filter
Physical filters can be applied during μCT scans, and by concentrating the X-ray beam, they act to reduce noise and increase contrast. Additionally, applying a filter can reduce artifacts caused by low energy X-rays. [bib_ref] Laboratory x-ray micro-computed tomography: a user guideline for biological samples, Du Plessis [/bib_ref] [bib_ref] The impact of spectral filtration on image quality in micro-CT system, Ren [/bib_ref] [bib_ref] High-throughput microCT scanning of small specimens: preparation, packing, parameters and post-processing, Hipsley [/bib_ref] The strength of the filter (ie, composition and thickness) should be considered during selection; stronger filters are indicated for denser samples; eg, mineralized tissues. Holding the voltage (70 kVp), voxel size (10 μm), and integration time (300 ms) constant, we compared no filter to a 0.1-mm copper (Cu) filter or 0.5-mm aluminum (Al) filter [fig_ref] Fig 2: Optimization of μCT scan parameters for dentoalveolar structures [/fig_ref] , [fig_ref] Table 1: Selection of μCT Scan ParametersVariables including filter, voltage, voxel size, and integration... [/fig_ref]. The scan with no filter appeared relatively crisp in grayscale; however, the heat map revealed numerous red and orange (low density) voxels and very poor separation of crown dentin and enamel. This is shown in the histogram as an expanded single peak where the enamel peak has been absorbed into the dentin/bone/cementum peak that skewed right [fig_ref] Fig 2: Optimization of μCT scan parameters for dentoalveolar structures [/fig_ref] , red *). There are also numerous artifacts of very high-density voxels well above biologic density values.
The scan with the Cu filter did not show any improvement in dentin and enamel separation and distribution of voxel densities, which was exacerbated by an unacceptable amount of noise in the grayscale and heat map images. The heat map and flattening of peaks in the histogram indicated that the noise holds density values similar to those of dentoalveolar tissues, preventing digital filtering of noise from enamel, dentin, and cementum and rendering the scan unusable for analysis.
The scan using the 0.5-mm Al filter yielded clearer images without excess noise, better distribution of voxel densities in the heat map, and better demarcation between enamel and dentin. The more compressed dentin/bone/cementum peak was better separated from the enamel peak [fig_ref] Fig 2: Optimization of μCT scan parameters for dentoalveolar structures [/fig_ref] , blue arrows). Because of these advantages, subsequent scans were performed with the 0.5-mm Al filter.
## Voltage
Voltage corresponds to the energy of the photons passing through a sample and is expressed in kiloelectron volts (kEV). In μCT scanning, the X-ray tube potential is specified by the user. Expressed in kVp, the X-ray tube potential corresponds to the applied peak electron potential of the X-ray tube that accelerates electrons for generating X-ray photons. [bib_ref] Guidelines for assessment of bone microstructure in rodents using micro-computed tomography, Bouxsein [/bib_ref] Generally, lower voltages provide better contrast between tissues, particularly for tissues with lower densities. [bib_ref] Small animal imaging and examination by micro-CT, Vasquez [/bib_ref] [bib_ref] Laboratory x-ray micro-computed tomography: a user guideline for biological samples, Du Plessis [/bib_ref] [bib_ref] Micro-computed tomography-current status and developments, Ritman [/bib_ref] [bib_ref] Microcomputed tomography: approaches and applications in bioengineering, Boerckel [/bib_ref] [bib_ref] Spectral optimization for micro-CT, Hupfer [/bib_ref] [bib_ref] Application of micro-CT in small animal imaging, Schambach [/bib_ref] However, voltages that are too low result in decreased transmission of X-rays through the sample that reach the detector. In contrast, voltages that are too high lead to increased transmission of X-rays through the sample that reach the detector, which can result in increased scatter and reduced resolution of less dense materials (i.e. decreased signal to noise ratio). For dentoalveolar tissues, the voltage would optimally be high enough to pass through enamel and minimize noise but low enough to generate contrast between tissues with similar densities; eg, cementum and dentin. Holding the filter (0.5-mm Al), voxel size (10 μm), and integration time (300 ms) constant, we compared scan voltages of 55, 70, and 90 kVp [fig_ref] Fig 2: Optimization of μCT scan parameters for dentoalveolar structures [/fig_ref] , [fig_ref] Table 1: Selection of μCT Scan ParametersVariables including filter, voltage, voxel size, and integration... [/fig_ref]. Changing kVp resulted in altered brightness and contrast in grayscale images and shifts in density values in the heat maps. The scan acquired with 55 kVp was not ideal for dentoalveolar tissues. The photon energy was too low and unable to penetrate high density tissues, leading to lower numbers of photons reaching the detector and a truncated histogram around 2000 mg HA/cm 3 [fig_ref] Fig 2: Optimization of μCT scan parameters for dentoalveolar structures [/fig_ref] , pink *). With increasing voltage at 70 or 90 kVp, the dentin/cementum/bone peak narrowed, and the enamel peak emerged. Thus, 70 and 90 kVp scans both showed acceptable image contrast and enamel separation. A voltage of 70 kVp was used in subsequent scans due to acceptable enamel separation and better separation potential and detection of the lower end of density for dentin/cementum/bone.
## Voxel size
Voxel size is the size of a 3D pixel in the reconstructed image following μCT scanning. Smaller voxel sizes correspond to higher resolution scans. Smaller voxel sizes allow for the visualization of a greater number of structures in more detail and accuracy, particularly small features. However, in practical terms, smaller voxel sizes require increased time per scan and produce increased file sizes. Bouxsein and colleagues [bib_ref] Guidelines for assessment of bone microstructure in rodents using micro-computed tomography, Bouxsein [/bib_ref] recommended that the minimum ratio of voxels to objects of interest should be 2 (eg, a minimum of two voxels per bone trabecula, thickness of cementum, or other feature), with higher ratios resulting in more accurate measurements. [bib_ref] Microcomputed tomography imaging in odontogenesis studies, Verdelis [/bib_ref] [bib_ref] Laboratory x-ray micro-computed tomography: a user guideline for biological samples, Du Plessis [/bib_ref] [bib_ref] Effect of micro-computed tomography voxel size and segmentation method on trabecular bone..., Christiansen [/bib_ref] [bib_ref] Spatial resolution characterization of a X-ray microCT system, Rueckel [/bib_ref] [bib_ref] Multiscale and multimodality computed tomography for cortical bone analysis, Ostertag [/bib_ref] [bib_ref] Micro-CT examinations of trabecular bone samples at different resolutions: 14, 7 and..., Peyrin [/bib_ref] Holding the filter (0.5-mm Al), scan voltage (70 kVp), and integration time (300 ms) constant, we compared voxel sizes of 20, 10, and 6 μm [fig_ref] Fig 2: Optimization of μCT scan parameters for dentoalveolar structures [/fig_ref] , [fig_ref] Table 1: Selection of μCT Scan ParametersVariables including filter, voltage, voxel size, and integration... [/fig_ref].
The grayscale image revealed that a 20-μm voxel size was wholly inadequate for the mouse dentoalveolar complex, with substantial blurring and noise in the soft tissues of the pulp and PDL. The corresponding heat map depicted dentin as nearly homogenous in density, and the y axis of the frequency distributions showed that voxel size dramatically influenced the counts per density. Borders of enamel and dentin, tissues with significantly different densities, were poorly resolved in grayscale, heat maps, and on the histogram. In contrast, higher resolution (lower voxel size) resulted in a clearer dentinoenamel junction, enhanced contrast of structures (eg, we could begin to resolve cellular cementum from dentin), and visualization of much greater detail in bone structures. Smaller voxel sizes also resulted in higher counts and sharper dentin/cementum/bone and enamel peaks, more amenable for segmentation of the tissues from one another for analysis.
In the scanner we used (Scanco μCT 50), the voxel size is restricted by the sizes of the sample holders, enclosed tubes in which samples are loaded and can remain hydrated during scanning. Other scanners use alternative methods for sample stabilization during scanning. [bib_ref] High-throughput microCT scanning of small specimens: preparation, packing, parameters and post-processing, Hipsley [/bib_ref] In Scanco scanners, smaller sample holders that reduce the distance between the X-ray source, sample, and X-ray detector, are capable of higher resolutions. To reduce scan time and file sizes, we optimized scans based on a 19.0-mm (diameter) sample holder. This was the smallest sample holder that allowed for mandibles to be loaded where molar occlusal plane was perpendicular to sample holder length (ie, mandibles were laid on their sides and stacked in a column with each separated from the next by a foam spacer, efficiently allowing multiple mandibles to be scanned per sample holder). A smaller sample holder would require mouse mandibles to be loaded in a different orientation (increasing scanning time and file size further) or trimmed (requiring additional preparation with the possibility of damaging target tissues). According to the manufacturer's recommended settings, the smallest voxel size for the 19.0-mm sample holder was 6 μm, which we have found to be adequate for dentoalveolar tissues (see below under Examples of Analyses). Because no substantial advantage for enamel, dentin, and bone was noted with smaller for the 2-μm voxel setting, a voxel size of 6 μm was used in subsequent scans. There is one exception for segmentation of acellular cementum (described below under Examples of Analyses: Cementum and PDL) where a 6.0-mm sample holder was necessary for 2-μm voxel scans to accurately separate murine acellular cementum from dentin.
## Integration time
Integration time refers to the time spent on each projection (in milliseconds; ms), and along with frame averaging (the number of times each projection is repeated), can change the number of photons directed at the sample. [bib_ref] Laboratory x-ray micro-computed tomography: a user guideline for biological samples, Du Plessis [/bib_ref] [bib_ref] Effect of integration time on the morphometric, densitometric and mechanical properties of..., Oliviero [/bib_ref] Therefore, increasing integration time increases overall scan time, but does not increase file size. Holding the filter (0.5-mm Al), scan voltage (70 kV), and voxel size (6 μm) constant, we compared integration times of 300, 600, and 900 ms [fig_ref] Fig 2: Optimization of μCT scan parameters for dentoalveolar structures [/fig_ref] , [fig_ref] Table 1: Selection of μCT Scan ParametersVariables including filter, voltage, voxel size, and integration... [/fig_ref].
In grayscale images, increasing integration time reduced noise in pulp and PDL soft tissues, and removed speckling in mineralized tissues. Although there are only subtle differences between the three integration times in the grayscale images, in heat maps and frequency distributions, dentin and cellular cementum were better differentiated by density when 900 ms integration time was used. Increased integration time did not shift locations of peaks in the frequency graphs, but instead sharpened the dentin/cementum/bone and enamel peaks. We did not attempt integration times greater than 900 ms for practical reasons. However, there is room for improvement, although the detector could potentially become saturated at some point, with no further benefit to increasing time.
## Summary of selection of scan parameters
In the previous sections under Image Acquisition of Dentoalveolar Tissues, we demonstrated the impact of filter, voltage, voxel size, and integration time on optimization of scan parameters, landing on 0.5-mm Al filter, 70 kVp, 6 μm voxel size, and 900 ms integration time [fig_ref] Fig 2: Optimization of μCT scan parameters for dentoalveolar structures [/fig_ref] ; [fig_ref] Table 1: Selection of μCT Scan ParametersVariables including filter, voltage, voxel size, and integration... [/fig_ref]. Filter had the greatest impact on scans, with Cu filter and no filter resulting in unusable scans. Next, voltage dramatically altered density distributions and visualization of mineralized tissue; 70 kVp was determined to be the best for resolving structures within 500 to 1600 mg HA/cm 3 . For qualitative and quantitative assessments, 6 μm is suitable for separating enamel, bone, dentin, and cellular cementum (see subsequent sections). Last, we chose an integration time of 900 ms based on our goal to segment dentin and cellular cementum. Each one of the aforementioned parameters can alter quality of scans. At a minimum, the make and model of scanner, the filter, voltage, voxel size (resolution), and integration time should be reported in methods sections of manuscripts employing μCT.
## Image processing
Tissues of the oral cavity present several unique challenges for μCT analysis in comparison to sites more frequently analyzed, eg, lumbar vertebrae or long bones, though their successful analysis shares a few key steps: orientation, region of interest (ROI), and segmentation.
## Orientation
Digital orientation of samples after scanning is the first step toward reproducible results. This is true for femur and tibia analyses, where the long axis of the bone is used to realign the sample prior to selection of trabecular and cortical ROI. [bib_ref] Guidelines for assessment of bone microstructure in rodents using micro-computed tomography, Bouxsein [/bib_ref] The mandible and maxilla are more complex shapes, and landmarks must be carefully chosen. [bib_ref] Microcomputed tomography imaging in odontogenesis studies, Verdelis [/bib_ref] Though orientation is often not described in detail, many papers reporting dentoalveolar analyses use one or more of the molars as the most consistent anatomical landmarks to orient mandibles/maxillae, as can be deduced from their figures. [bib_ref] Three-dimensional microcomputed tomographic imaging of alveolar bone in experimental bone loss or..., Park [/bib_ref] [bib_ref] A robust methodology for the quantitative assessment of the rat jawbone microstructure, Chatterjee [/bib_ref] [bib_ref] Hypercementosis associated with ENPP1 mutations and GACI, Thumbigere-Math [/bib_ref] [bib_ref] Loss of discoidin domain receptor 1 predisposes mice to periodontal breakdown, Chavez [/bib_ref] [bib_ref] Genetic and pharmacologic modulation of cementogenesis via pyrophosphate regulators, Chu [/bib_ref] [bib_ref] Dental and craniofacial defects in the Crtap( −/− ) mouse model of..., Xu [/bib_ref] [bib_ref] Use of micro-computed tomography for the assessment of periapical lesions in small..., Kalatzis-Sousa [/bib_ref] [bib_ref] Osteopontin regulates dentin and alveolar bone development and mineralization, Foster [/bib_ref] [bib_ref] Sclerostin antibody stimulates bone regeneration after experimental periodontitis, Taut [/bib_ref] [bib_ref] Effects of spaceflight on the murine mandible: possible factors mediating skeletal changes..., Ghosh [/bib_ref] [bib_ref] Micro-anatomical responses in periodontal complexes of mice to calibrated orthodontic forces on..., Pal [/bib_ref] [bib_ref] The use of live-animal microcomputed tomography to determine the effect of a..., Cantley [/bib_ref] [bib_ref] Endodontic infection-induced inflammation resembling osteomyelitis of the jaws in toll-like receptor 2/interleukin..., Sasaki [/bib_ref] [bib_ref] IL-17 receptor A signaling is protective in infection-stimulated periapical bone destruction, Alshwaimi [/bib_ref] Incisors or whole mandible/maxilla approaches are also sometimes used for orientation. [bib_ref] Micro-anatomical responses in periodontal complexes of mice to calibrated orthodontic forces on..., Pal [/bib_ref] [bib_ref] Accelerated enamel mineralization in Dspp mutant mice, Verdelis [/bib_ref] [bib_ref] A novel method to detect 3D mandibular changes related to soft-diet feeding, Kono [/bib_ref] Here, we describe the use of the mandibular first molar (M1) as a guide for orientation. To achieve consistency, orientation in sagittal, frontal/coronal, and transverse/axial planes should be standardized across samples using anatomical landmarks. In the sagittal view, the plane generated by the mesial and distal aspects of the cementoenamel junction (CEJ) was oriented parallel to the transverse plane [fig_ref] Fig 3: Orientation and ROIs in the dentoalveolar complex [/fig_ref]. In the transverse view, the plane generated by the center of the mesial and distal root canals of the first molar were oriented parallel to the sagittal plane. In the coronal view, the plane bisecting the pulp of the mesial root was oriented parallel to the sagittal plane. Although orientation is important for consistent imaging of M1, it becomes even more critical when measuring surrounding tissues like alveolar bone (as described in more detail below under Examples of Analyses: Alveolar Bone). This digital orientation approach yielded highly reproducible results during training of multiple users in our laboratories (<0.1% difference in volume measurements using the same mandible scan), validating this approach (data not shown). Experimental factors may dictate which teeth serve as landmarks, eg, if a ligature is placed on the second molar to induce periodontal breakdown, then the second molar may be a better choice as the central landmark for mandible orientation. When reporting μCT analysis of the mouse mandible, the approach to sample orientation should be described in enough detail and/or images should be included to illustrate the approach. [bib_ref] Osteopontin regulates dentin and alveolar bone development and mineralization, Foster [/bib_ref] ROI After orientation, the ROI is defined based on experimental questions to be answered. In long bones, there are prescribed ROIs to analyze trabecular and cortical bone based on anatomical landmarks (eg, proximal or distal growth plate) and recommended minimum numbers of slices or μm. [bib_ref] Effect of micro-computed tomography voxel size and segmentation method on trabecular bone..., Christiansen [/bib_ref] [bib_ref] A new high-resolution computed tomography (CT) segmentation method for trabecular bone architectural..., Scherf [/bib_ref] For dentoalveolar analysis, the tooth is a well-defined and self-contained organ that can be analyzed in whole, an approach not usually taken for long bones. Due to limitations in software, computing power, and/or time, some investigators may opt to analyze more limited ROIs rather than the entire tooth. For example, rather than segmenting the entire dentin of the molar, a cubic, ring, or other shaped ROI of a certain number of voxels or μm 3 may be chosen within dentin and used as a representative sample. Although this approach may be necessary in some circumstances, it should be avoided if possible and used with caution as it can bias results and provide density values that inaccurately represent the bulk tissue properties; eg, densities of the dentoalveolar tissues vary by location. [bib_ref] Mineral density volume gradients in normal and diseased human tissues, Djomehri [/bib_ref] If this restricted ROI approach is used, locations must be chosen stringently and consistently across samples to minimize selection bias. In terms of creating the ROI, more details on segmenting dental tissues from one another are discussed below under the Segmentation section.
Orientation and landmarks used to define the ROI must be anatomically similar across samples to capture potential differences between experimental and control groups. In our studies, the ROI was determined using a reoriented M1 and included all alveolar bone buccal and lingual to the molar and extending the region 480 μm forward from the mesial root and 480 μm backward to the distal root, using the most mesial and distal root locations, respectively [fig_ref] Fig 3: Orientation and ROIs in the dentoalveolar complex [/fig_ref]. This ROI was selected to include the alveolar rise mesial to M1 and the interdental bone between M1 and M2, and this approach has been applied successfully to a wide range of mouse mandibles aged 2 weeks to over 1 year old. [bib_ref] Hypercementosis associated with ENPP1 mutations and GACI, Thumbigere-Math [/bib_ref] [bib_ref] Loss of discoidin domain receptor 1 predisposes mice to periodontal breakdown, Chavez [/bib_ref] [bib_ref] Genetic and pharmacologic modulation of cementogenesis via pyrophosphate regulators, Chu [/bib_ref] [bib_ref] Reduced orthodontic tooth movement in Enpp1 mutant mice with hypercementosis, Wolf [/bib_ref] [bib_ref] Dental and craniofacial defects in the Crtap( −/− ) mouse model of..., Xu [/bib_ref] [bib_ref] Dentoalveolar defects in the Hyp mouse model of X-linked hypophosphatemia, Zhang [/bib_ref] [bib_ref] Osteopontin regulates dentin and alveolar bone development and mineralization, Foster [/bib_ref] [bib_ref] Dental defects in the primary dentition associated with hypophosphatasia from biallelic ALPL..., Kramer [/bib_ref] [bib_ref] Overlapping functions of bone sialoprotein and pyrophosphate regulators in directing cementogenesis, Ao [/bib_ref] The ROI we describe here would suffice for many studies of alveolar bone but should be experimentally determined for the hypothesis being examined and the specific needs of the study. In publications, the ROI should be described in detail, including anatomical landmarks used and precise numbers of slices or μm for tissues like alveolar bone.
## Segmentation
Following orientation and ROI definition, segmentation is the next step toward establishing consistency in μCT analyses. Segmentation, or isolation of structures, typically begins with setting thresholds to separate tissues by mineral density. [bib_ref] Effect of micro-computed tomography voxel size and segmentation method on trabecular bone..., Christiansen [/bib_ref] [bib_ref] A new high-resolution computed tomography (CT) segmentation method for trabecular bone architectural..., Scherf [/bib_ref] [bib_ref] A comparative study of automatic thresholding approaches for 3D x-ray microtomography of..., Gomez [/bib_ref] Setting appropriate thresholds is imperative to promote accuracy and reproducibility, because measurements can be greatly influenced by the choice of threshold values. Thresholds for all target tissues should be defined in HA/mg, Hounsfield units, grayscale values, or similar absolute or relative units. Although relative units (eg, grayscale values) are adequate for comparisons within a given study (same scanner, settings, and short time span between scans), calibrated absolute units (eg, mg HA per volume and Hounsfield units) allow more robust comparisons between studies and compilation of normal and abnormal data sets across the published literature. [bib_ref] Analysis of bone architecture in rodents using micro-computed tomography, Van't Hof [/bib_ref] [bib_ref] Quantitative analysis of bone and soft tissue by micro-computed tomography: applications to..., Campbell [/bib_ref] [bib_ref] Three-dimensional microcomputed tomographic imaging of alveolar bone in experimental bone loss or..., Park [/bib_ref] [bib_ref] A robust methodology for the quantitative assessment of the rat jawbone microstructure, Chatterjee [/bib_ref] [bib_ref] Effect of micro-computed tomography voxel size and segmentation method on trabecular bone..., Christiansen [/bib_ref] [bib_ref] Preparation and characterization of calibration standards for bone density determination by micro-computed..., Schweizer [/bib_ref] As shown in [fig_ref] Fig 2: Optimization of μCT scan parameters for dentoalveolar structures [/fig_ref] , use of calibrated absolute units allows comparison of scans acquired from different sets of parameters.
Recommendations for thresholding bone tissue in skeletal analysis have been discussed. [bib_ref] Guidelines for assessment of bone microstructure in rodents using micro-computed tomography, Bouxsein [/bib_ref] However, the dentoalveolar complex offers the challenge of four unique mineralized tissues, requiring further considerations. Thresholding of dentoalveolar tissues is based on anatomical structures and inherent density differences between tissues. Because average enamel mineral density is considerably greater than dentin, bone, and cementum, simple thresholding is largely sufficient to segment enamel [fig_ref] Fig 1: Murine dentoalveolar anatomy [/fig_ref]. The significance of threshold values can be appreciated using tissue volume as a readout. There is no absolute correct threshold for any of the tissues, and each researcher should optimize their thresholds according to their scanner and scan conditions. However, we provide examples of how to evaluate thresholds and how their choice directly affects enamel, dentin, and bone volumes calculated. Based on anatomical landmarks examined over several studies, 1600 mg HA/cm 3 was determined to most accurately label enamel in our scans. [bib_ref] Loss of discoidin domain receptor 1 predisposes mice to periodontal breakdown, Chavez [/bib_ref] [bib_ref] Genetic and pharmacologic modulation of cementogenesis via pyrophosphate regulators, Chu [/bib_ref] [bib_ref] MMP20 overexpression disrupts molar ameloblast polarity and migration, Shin [/bib_ref] [bib_ref] Reduced orthodontic tooth movement in Enpp1 mutant mice with hypercementosis, Wolf [/bib_ref] [bib_ref] Dental and craniofacial defects in the Crtap( −/− ) mouse model of..., Xu [/bib_ref] [bib_ref] Dentoalveolar defects in the Hyp mouse model of X-linked hypophosphatemia, Zhang [/bib_ref] [bib_ref] Osteopontin regulates dentin and alveolar bone development and mineralization, Foster [/bib_ref] [bib_ref] Overlapping functions of bone sialoprotein and pyrophosphate regulators in directing cementogenesis, Ao [/bib_ref] Qualitatively, a lower threshold (1400 mg HA/cm 3 ) resulted in areas of dentin being identified as enamel (visualized as speckling in dentin areas), whereas a higher threshold (1800 mg HA/cm 3 ) resulted in enamel volume (EV) underestimation (seen as visible gaps in 3D renderings, [fig_ref] Fig 3: Orientation and ROIs in the dentoalveolar complex [/fig_ref]. Quantitatively, these lower and higher thresholds resulted in substantially different EV values at +32.4% and −39.5%, respectively.
Thresholding for dentin and bone was similarly evaluated. We selected a lower threshold of 650 mg HA/cm 3 as most accurately segmenting bone and dentin volumes (BV and DV, respectively) in our scans. A lower threshold (450 mg HA/cm 3 ) included areas that did not anatomically correspond to dentin; for example, an accessory canal was mislabeled as dentin [fig_ref] Fig 4: Strategies for segmentation of dentoalveolar tissues [/fig_ref]. Conversely, a higher threshold (850 mg HA/cm 3 ) excluded apical regions of dentin [fig_ref] Fig 4: Strategies for segmentation of dentoalveolar tissues [/fig_ref]. While these differences (A7-A12) the second row demonstrates use of filters to accomplish segmentation of cementum from dentin, particularly how scan resolution and median filters can substantially improve fidelity of segmentation. Details are described in the text. (B1-B6) Threshold techniques are available to segment tissues quickly and reproducibly, including software operations: object separator, region grow, lock layers, fill void spaces, and dilate object. (C1-C6) Semi-automation of corrections after segmentation is recommended to save time and increase reproducibility. These include: edge detection (software uses filters to identify borders; red arrow), walls (user sets boundaries to help software define borders; red arrow and red dotted line), morph functions (close function is highlighted, where software dilates then erodes the object to original position unless the dilated object was fully encompassed by the function, thus correcting small internal errors; red arrows), mask (software uses a previously generated mask, registers it to same position and allows for segmentation only within/outside the mask), and manual correction (here, we show removal of areas that do not anatomically correspond to cementum). (D1-D3) Subregions of tissues can be defined for targeted analysis. For example, mandibular bone is separated into basal bone and buccal and lingual aspects of alveolar bone. Proximity can be used to define areas within a certain anatomical structure, such as bone within a certain distance from the tooth (magenta). Thicknesses can be measured, such as in enamel, different regions of dentin, AC, and CC. AC = acellular cementum; BV = bone volume; CC = cellular cementum; DV = dentin volume; EV = enamel volume. appear subtle in 2D images, the resultant volumetric changes were not insignificant. Quantitatively, the lower threshold resulted in values +12.3% in BV and + 6.2% in DV, and the higher threshold resulted in values −11.2% in BV and −9.4% in DV [fig_ref] Fig 4: Strategies for segmentation of dentoalveolar tissues [/fig_ref].
Because of their similar densities, cementum is particularly challenging to segment from dentin, but the two can be distinguished with appropriate resolution and image processing filters. Although acellular cementum in mice is extremely thin (less than 10 μm thickness at ages typically used, eg, up to 6 months postnatal), cellular cementum on the apical portions of roots grows rapidly after tooth eruption. [bib_ref] Are cementoblasts a subpopulation of osteoblasts or a unique phenotype?, Bosshardt [/bib_ref] [bib_ref] Initial formation of cellular intrinsic fiber cementum in developing human teeth. A..., Bosshardt [/bib_ref] In our experience, a high resolution (2 μm voxel size) and a high integration time (1200 ms) must be used in conjunction with proper filter and voltage to reliably segment mouse dentin from acellular cementum [fig_ref] Fig 1: Murine dentoalveolar anatomy [/fig_ref]. [bib_ref] Hypercementosis associated with ENPP1 mutations and GACI, Thumbigere-Math [/bib_ref] [bib_ref] Genetic and pharmacologic modulation of cementogenesis via pyrophosphate regulators, Chu [/bib_ref] As seen in the 6-μm and 2-μm heat maps with no filter, the cementum layer has a lower average density than dentin [fig_ref] Fig 4: Strategies for segmentation of dentoalveolar tissues [/fig_ref] ,A9, as indicated by red/orange/yellow voxels), which becomes clearer with higher resolutions (compare [fig_ref] Fig 4: Strategies for segmentation of dentoalveolar tissues [/fig_ref] to [fig_ref] Fig 4: Strategies for segmentation of dentoalveolar tissues [/fig_ref]. However, a general threshold alone cannot properly differentiate the tissues as dentin also includes areas with lower density similar to cementum [fig_ref] Fig 1: Murine dentoalveolar anatomy [/fig_ref]. To further enhance native density differences, a median filter was applied, which resamples voxels to the median density of their neighbors [fig_ref] Fig 1: Murine dentoalveolar anatomy [/fig_ref]. This results in a clearer demarcation between dentin and cementum, allowing for improved segmentation. With the median filter, cellular cementum is readily viewed in 6-μm and 2-μm scans; however, 6 μm is inadequate for isolating acellular cementum. Even with the median filter in a 2-μm scan, newly mineralized dentin immediately adjacent to pulp had similar density values to cementum [fig_ref] Fig 1: Murine dentoalveolar anatomy [/fig_ref]. Sequential steps are required to manually correct this situation, as outlined in the following paragraphs.
The thresholding values described here were optimized in normal adult mice, however, optimal thresholding values should be determined for the project and scientific question(s). For example, in some contexts (eg, very young mice, models of bone healing, or genetically modified mice with defective mineralization), a threshold value for alveolar bone at 650 mg HA/cm 3 may exclude hypomineralized bone and report a low tissue volume even if there is expansion of poorly mineralized (osteoid) bone. [bib_ref] Dentoalveolar defects in the Hyp mouse model of X-linked hypophosphatemia, Zhang [/bib_ref] Threshold values should always be included in publications and any limitations or alterations from the norm should be discussed.
After optimal threshold values are determined, a myriad of computer software-based histometric thresholding techniques is available. The most familiar and least selective is a global threshold where one value is set for all tissues in a sample. Global thresholds are useful for separating out structures that are homogeneous in density or for separating out objects from background noise, but they are inadequate when numerous tissues present have overlapping densities; eg, dentin and bone [fig_ref] Fig 1: Murine dentoalveolar anatomy [/fig_ref]. With computer and software advances, new histometric thresholding techniques are available. For all segmentation procedures described in the following paragraphs and shown in [fig_ref] Fig 4: Strategies for segmentation of dentoalveolar tissues [/fig_ref] and 4C, we used semi-automatic and manual tracing features in AnalyzePro to segment tissues quickly and reproducibly. Highlighted in [fig_ref] Fig 4: Strategies for segmentation of dentoalveolar tissues [/fig_ref] are useful software operations: object separator [fig_ref] Fig 2: Optimization of μCT scan parameters for dentoalveolar structures [/fig_ref] ; objects that are spatially distanced are separated), region grow [fig_ref] Fig 3: Orientation and ROIs in the dentoalveolar complex [/fig_ref] ; object definition is based on defined threshold value and connected components), lock layers [fig_ref] Fig 4: Strategies for segmentation of dentoalveolar tissues [/fig_ref] ; a segmented layer is locked from changes in subsequent steps), fill void spaces [fig_ref] Fig 4: Strategies for segmentation of dentoalveolar tissues [/fig_ref] ; spaces enclosed within a segmented object are identified as a new object), and dilate object [fig_ref] Fig 4: Strategies for segmentation of dentoalveolar tissues [/fig_ref] ; a labeled region is expanded by a user defined distance).
Although the majority of segmentation can be accurately achieved with threshold techniques illustrated in [fig_ref] Fig 4: Strategies for segmentation of dentoalveolar tissues [/fig_ref] , correction is required. Many semiautomatic correction tools are available, which save time and increase reproducibility, including: edge detection [fig_ref] Fig 1: Murine dentoalveolar anatomy [/fig_ref] ; software uses filters to identify borders, red arrow), walls [fig_ref] Fig 2: Optimization of μCT scan parameters for dentoalveolar structures [/fig_ref] ; user sets boundaries to help software define borders; red arrow and red dotted line), morph functions [fig_ref] Fig 3: Orientation and ROIs in the dentoalveolar complex [/fig_ref] ; close function is highlighted here where software dilates then erodes the object to original position unless the dilated object was fully encompassed by the function, thus correcting small internal errors, red arrows), mask [fig_ref] Fig 4: Strategies for segmentation of dentoalveolar tissues [/fig_ref] ; software uses a previously generated mask, registers it to same position and allows for segmentation only within/outside the mask), and finally, manual correction, ie, non-automatic, operatorspecific corrections based on previous anatomical knowledge [fig_ref] Fig 4: Strategies for segmentation of dentoalveolar tissues [/fig_ref] ; here, we show removal of areas that do not anatomically correspond to cementum).
Our approach includes multiple sequential steps. For tissues with disparate densities, eg, dentin and enamel, region grow for a single threshold can be used [fig_ref] Fig 3: Orientation and ROIs in the dentoalveolar complex [/fig_ref]. Because region grow searches for connected components, high density voxels in the dentin are excluded when a seed point is set on enamel. Areas in dentin that are incorrectly assigned to enamel can be eliminated with software operations such as morphological transformations [fig_ref] Fig 3: Orientation and ROIs in the dentoalveolar complex [/fig_ref]. Once the enamel layer is defined, locking the layer allows a second threshold to be set for dentin/bone without altering the enamel layer [fig_ref] Fig 4: Strategies for segmentation of dentoalveolar tissues [/fig_ref]. For tissues with a similar density, semiautomatic segmentation requires anatomic spatial separation, eg, dentin and bone, which are separated by an unmineralized PDL that typically spans 50 to 100 μm in width [bib_ref] Reduced orthodontic tooth movement in Enpp1 mutant mice with hypercementosis, Wolf [/bib_ref] [bib_ref] Dental and craniofacial defects in the Crtap( −/− ) mouse model of..., Xu [/bib_ref] [fig_ref] Fig 2: Optimization of μCT scan parameters for dentoalveolar structures [/fig_ref]. For contiguous tissues with similar densities (eg, dentin and cementum), additional processing methods can be employed. We applied a median filter with a kernel size of seven to reconstructed images better identified density differences between dentin and cementum (compare [fig_ref] Fig 1: Murine dentoalveolar anatomy [/fig_ref]. A threshold between 250 to 1100 mg/cm 3 was applied to capture all possible cementum tissue, with minor manual corrections to remove less dense dentin adjacent to the pulp chamber [fig_ref] Fig 4: Strategies for segmentation of dentoalveolar tissues [/fig_ref]. This object map was overlaid to the original scan (non-median filter) and cementum was then segmented at 650 mg/cm 3 only within the region traced in the median filter mask [fig_ref] Fig 4: Strategies for segmentation of dentoalveolar tissues [/fig_ref] , mask, manual correction). Unmineralized tissues, ie, pulp and PDL, can be segmented semiautomatically using density boundaries [fig_ref] Fig 4: Strategies for segmentation of dentoalveolar tissues [/fig_ref] , B6,C2, fill void objects, dilate object, and walls/computer assist). It is essential to accurately describe segmentation strategies in publications as these can dramatically change results.
## Subregions
Further digital subdivision of tissues can be invaluable for dentoalveolar tissues, which exhibit a great degree of heterogeneity. Analyses can be targeted toward specific tissue anatomical subregions to detect changes that would otherwise be missed with whole tissue analyses. For example, subdivision of dentin volume allows comparison of crown versus root dentin or mesial versus distal molar roots. Alveolar bone is especially amenable to subdivision into anatomical regions and presents opportunities to address tissue-specific scientific questions. For example, subdivisions may include separation of the mandible into basal versus alveolar bone, and alveolar bone can be subdivided into mesial versus distal, buccal versus lingual, or even tripartite division into cervical, middle, and apical alveolar bone [fig_ref] Fig 1: Murine dentoalveolar anatomy [/fig_ref]. Further, proximity can be used to define areas within a certain anatomical structure. In the example shown in [fig_ref] Fig 2: Optimization of μCT scan parameters for dentoalveolar structures [/fig_ref] , we labeled alveolar bone within 240 μm from the tooth root surface [fig_ref] Fig 2: Optimization of μCT scan parameters for dentoalveolar structures [/fig_ref] , magenta) to indicate bone most involved with periodontal attachment (alveolar bone proper or ABP, as discussed in more detail below under Examples of Analyses: Alveolar Bone). Accordingly, the periodontal apparatus can be examined regionally or functionally, eg, the PDL, cementum, and alveolar bone can be subdivided into apical, middle, and coronal regions in parallel. Because subdivision can be specifically tailored to the experimental question, the approach must be described in detail when published, and a figure displaying subdivisions can be illustrative. [bib_ref] Osteopontin regulates dentin and alveolar bone development and mineralization, Foster [/bib_ref] In addition to determination of volume and density, thickness can be a useful measurement. Methods include taking multiple linear measurements and using software applications to measure average thickness over a given area. [bib_ref] Cartilage morphology assessed by high resolution micro-computed tomography in non OA knees, Delecourt [/bib_ref] [bib_ref] Assessment of the accuracy of dental enamel thickness measurements using microfocal X-ray..., Olejniczak [/bib_ref] Single linear measurements are problematic because they are susceptible to high user bias and lack of reproducibility within and between users. Strategies to decrease bias include averaging multiple measurements, assigning a single user, calibrating multiple users, and ensuring reproducible orientations. In some software packages, the same formulas used to measure cortical thickness in long bones can be repurposed to measure thickness in dental tissues given that a hollow cylinder can be approximated by the tissue (eg, crown dentin). Some examples of thickness measurements are illustrated [fig_ref] Fig 3: Orientation and ROIs in the dentoalveolar complex [/fig_ref].
## Examples of analyses
To address specific research questions, application of appropriate scan settings [fig_ref] Fig 2: Optimization of μCT scan parameters for dentoalveolar structures [/fig_ref] , regions of interest [fig_ref] Fig 3: Orientation and ROIs in the dentoalveolar complex [/fig_ref] , and segmentation techniques [fig_ref] Fig 4: Strategies for segmentation of dentoalveolar tissues [/fig_ref] , are required to provide appropriate data to test the hypothesis. In the next sections, we provide examples of how these recommended μCT practices were applied to real analyses of enamel, dentin, cementum, and alveolar bone in genetically edited mouse models.
## Enamel
Optimal evaluation of enamel depends on proper thresholding to avoid overestimating or underestimating enamel volume [fig_ref] Fig 1: Murine dentoalveolar anatomy [/fig_ref]. In mouse studies where molars are erupted and amelogenesis has been completed, we typically threshold enamel at >1600 mg HA/cm 3 . [bib_ref] Loss of discoidin domain receptor 1 predisposes mice to periodontal breakdown, Chavez [/bib_ref] [bib_ref] Genetic and pharmacologic modulation of cementogenesis via pyrophosphate regulators, Chu [/bib_ref] [bib_ref] Dental and craniofacial defects in the Crtap( −/− ) mouse model of..., Xu [/bib_ref] [bib_ref] Dentoalveolar defects in the Hyp mouse model of X-linked hypophosphatemia, Zhang [/bib_ref] After thresholding and confirming segmentation of enamel from dentin, enamel volume and average mineral density are easily calculated. Although enamel is typically the most straightforward tissue to threshold and segment, severe defects create challenges for analysis; eg, when enamel and dentin densities become close or overlapping, when the enamel structure is severely disrupted (as in the example in [fig_ref] Fig 5: Legend on next page. [/fig_ref] , or when enamel is too unstable to remain attached (also as in [fig_ref] Fig 5: Legend on next page. [/fig_ref].
Several mouse models of amelogenesis imperfecta (AI), an inherited disease affecting enamel formation, have been created and evaluated by μCT. [bib_ref] Simmer JP. MMP20, KLK4, and MMP20/KLK4 double null mice define roles for..., Hu [/bib_ref] [bib_ref] Maturation stage enamel malformations in Amtn and Klk4 null mice, Nunez [/bib_ref] Mice genetically ablated for matrix metalloproteinase 20 (MMP20) phenocopy a form of AI with hypomaturation defects that reflect secretory stage defects in amelogenesis. [bib_ref] Dental enamel development: proteinases and their enamel matrix substrates, Bartlett [/bib_ref] In studies analyzing the functions of MMP20, mice overexpressing this metalloproteinase (Mmp20 +/+ Tg + ) exhibited severe enamel defects marked by a patchy, thin, poorly defined enamel layer that delaminated from dentin [fig_ref] Fig 5: Legend on next page. [/fig_ref]. [bib_ref] MMP20 overexpression disrupts molar ameloblast polarity and migration, Shin [/bib_ref] Molar enamel volume was decreased 60% and mineral density reduced 40%. Because enamel on the cusp regions of Mmp20 +/+ Tg + mice appeared to include ectopic blebs and flaking dentin, it was difficult to confidently define enamel borders and estimate thickness. In this scenario, we opted to measure enamel thickness at lateral locations on the first molar, where it would be less likely to break due to occlusal forces. Molar enamel thickness was calculated using cortical bone algorithms for the most median 25 axial slices (150 μm), as measured from the cementum-enamel junction to the highest cusp tip. This approach revealed 70% decrease in enamel thickness in Mmp20 +/+ Tg + versus control mice molars.
## Dentin and pulp
As the mineralized tissue composing the bulk of the tooth, dentin has been frequently analyzed by μCT. Although nonmineralized dental pulp is not directly analyzed by μCT, morphometric inferences can be made because it is enclosed by dentin. In [fig_ref] Fig 5: Legend on next page. [/fig_ref] analyses of dentoalveolar tissues in genetically engineered mouse models. Approaches described in this work are illustrated here for analysis of specific dentoalveolar tissues. (A) Enamel analysis was performed on Mmp20 +/+ Tg + overexpressing mice. This severe AI-like phenotype is characterized by reduced enamel volume and density compared to WT. Enamel thickness was measured at lateral locations (yellow arrows) to avoid area of destruction on occlusal surfaces. (B) Dentin and pulp analyses were performed on the Hyp mutant mouse model of XLH. Compared to WT, Hyp mice exhibit decreased dentin volume and density (yellow arrows) corresponding with an increase in pulp volume (yellow *). (C) Cementum and PDL analyses were performed on the Enpp1 mutant mouse model of GACI. Enpp1 mutant mice feature dramatically increased AC volume and thickness compared to WT (green arrow). Increased PDL volume and thickness is also detected (red arrow and green *) in Enpp1 mutant versus WT mice. (D) Alveolar bone analysis was performed on Ddr1 mutant mice that feature periodontitis-like bone loss at 9 months of age. Three μCT approaches were employed to quantify bone loss (yellow arrows and *). First, an adaptation of the classic linear measurement approach shows bone loss as increased distances up to 0.4 mm from CEJ-ABC at multiple locations on buccal and lingual aspects around Ddr1 −/− versus WT molars. Locations are indicated by tooth [bib_ref] Micro-CT of rodents: state-of-the-art and future perspectives, Clark [/bib_ref] [bib_ref] Micro-CT analysis of the rodent jaw bone micro-architecture: a systematic review, Faot [/bib_ref] and direction or feature (M, mesial; D; distal; B; buccal; L, lingual; C, cusp; G, groove), as described in Chavez and colleagues. [bib_ref] Loss of discoidin domain receptor 1 predisposes mice to periodontal breakdown, Chavez [/bib_ref] Second, measurement of total alveolar bone loss around molars reveals 14% reduction in Ddr1 −/− versus WT molars. Third, definition of ABP using a proximity technique (as described in the text and in [fig_ref] Fig 4: Strategies for segmentation of dentoalveolar tissues [/fig_ref] indicated a 30% reduction in ABP in Ddr1 −/− versus WT molars. Statistical analysis was performed by independent samples t test; *p < .05; **p < .01; ***p < .001; ****p < .0001. (A) Reproduced and adapted with permission from Shin and colleagues. [bib_ref] MMP20 overexpression disrupts molar ameloblast polarity and migration, Shin [/bib_ref] (B) Reproduced and adapted with permission from Zhang and colleagues. [bib_ref] Dentoalveolar defects in the Hyp mouse model of X-linked hypophosphatemia, Zhang [/bib_ref] (C) Reproduced and adapted from Thumbigere-Math and colleagues. [bib_ref] Hypercementosis associated with ENPP1 mutations and GACI, Thumbigere-Math [/bib_ref] (D) Reproduced and adapted from Chavez and colleagues. [bib_ref] Loss of discoidin domain receptor 1 predisposes mice to periodontal breakdown, Chavez [/bib_ref] ABP = alveolar bone proper; AC = acellular cementum; AI = amelogenesis imperfecta; CEJ-ABC = cementum-enamel junction to alveolar bone crest; GACI = generalized arterial calcification in infancy; PDL = periodontal ligament; XLH = X-linked hypophosphatasia. mouse studies where molars are erupted and primary dentin formation is established, we typically threshold dentin at >650 mg HA/cm 3 . [bib_ref] A robust methodology for the quantitative assessment of the rat jawbone microstructure, Chatterjee [/bib_ref] [bib_ref] State of the art of micro-CT applications in dental research, Swain [/bib_ref] [bib_ref] A validated method for titanium implant anchorage analysis using microCT and biomechanical..., Gabet [/bib_ref] [bib_ref] A metal artifact reduction method for a dental CT based on adaptive..., Hegazy [/bib_ref] After thresholding and confirming segmentation of dentin from enamel and bone, dentin volume and average mineral density can be calculated. Dentin and cementum volumes or densities are often reported as one collective measurement because of similar densities and proximity (though this volume is sometimes incorrectly identified simply as "dentin" in publications). This may result in misrepresentation of root dentin changes because alterations in cementum would also contribute to combined root tissue measurements. Additional μCT measurements or combination with histological or other approaches can clarify the situation. Subdivision can be useful in μCT analyses of dentin and pulp, eg, separating crown and root dentin for analysis, and this can be done by identifying the CEJ to demarcate crown versus root tissues.
Several inherited disorders affect dentinogenesis, including dentinogenesis imperfecta/dentin dysplasia, caused by mutations in dentin sialophosphoprotein (DSPP). Other mineralization disorders may affect dentin, including multiple forms of osteogenesis imperfecta (OI), hypophosphatasia (HPP), and genetic/ congenital forms of hypophosphatemic rickets. [bib_ref] Dental and craniofacial defects in the Crtap( −/− ) mouse model of..., Xu [/bib_ref] [bib_ref] Conditional Alpl ablation phenocopies dental defects of hypophosphatasia, Foster [/bib_ref] [bib_ref] Craniofacial and dental defects in the Col1a1Jrt/+ mouse model of osteogenesis imperfecta, Eimar [/bib_ref] [bib_ref] Ultrastructural organization of dentin in mice lacking dentin sialo-phosphoprotein, Fang [/bib_ref] [bib_ref] Mineral and matrix changes in Brtl/+ teeth provide insights into mineralization mechanisms, Boskey [/bib_ref] Hyp mutant mice featuring mutations in phosphate-regulating endopeptidase homolog X-linked gene (Phex) represent a mouse model of X-linked hypophosphatemia (XLH). In a study to define the precise dentoalveolar pathology associated with XLH, Hyp mice were found to feature severe mineralization disorders in multiple dentoalveolar tissues [fig_ref] Fig 5: Legend on next page. [/fig_ref] , notably expanded pulp chambers and dentin defects. [bib_ref] Dentoalveolar defects in the Hyp mouse model of X-linked hypophosphatemia, Zhang [/bib_ref] Molar dentin volume was decreased 30% to 40%, whereas dentin mineral density was decreased 4% in Hyp versus control mice. Conversely, pulp volume was increased 300% in Hyp mice compared to WT. These dentin measurements included the entire molar after segmenting out enamel and pulp space; therefore, they included a small contribution from cementum. In the case of Hyp mice, the dentin defects were so severe, and cementum was so reduced and/or hypomineralized, that there was no practical alternative but to calculate measurements for combined dentin/cementum, and this was noted in the Materials and Methods and Results sections in Zhang and colleagues, [bib_ref] Dentoalveolar defects in the Hyp mouse model of X-linked hypophosphatemia, Zhang [/bib_ref] and explored further by other approaches, including histology, histomorphometry, scanning electron microscopy (SEM), and nanoindentation.
## Cementum and pdl
The cementum layer is diminutive in mice, particularly the acellular cementum on the cervical region of the root. To date, very few studies have used μCT to analyze either acellular or cellular cementum, therefore cementum volume, thickness, and mineral density have not been routinely included in publications. This is in spite of the fact that acellular cementum is critical for tooth attachment and cellular cementum comprises a significant portion of root volume. The primary reason for this is that segmentation of cementum from dentin presents several technical challenges not easily overcome. In histological slides, both acellular and cellular cementum are clearly identifiable from dentin based on selective staining (eg, H&E, toluidine blue, or picrosirius red stain viewed under polarized light, where cementum and dentin exhibit different organization and orientation of collagen fibers). [bib_ref] Methods for studying tooth root cementum by light microscopy, Foster [/bib_ref] In order to accurately define cementum by μCT, scan parameters must be sensitive enough to take advantage of the morphological, organizational, and relatively small density differences between cementum and dentin (as outlined above and in [fig_ref] Fig 4: Strategies for segmentation of dentoalveolar tissues [/fig_ref]. As with pulp, nonmineralized PDL is not usually analyzed by μCT, but as an essential component of the periodontal complex that can adapt to and reflect changes in cementum. Analysis of PDL volume or thickness can be achieved after segmenting the adjacent hard tissues.
We have optimized scan parameters and analytical approaches for murine cementum. Parameters judged the most optimal for cellular cementum were 0.5-mm Al filter, 70 kVp, 6 μm voxel size, and 900 ms integration time. For acellular cementum, mouse molars were dissected from the mandible and scanned in a smaller diameter sample holder (Scanco 6.0 mm) with settings of 0.5-mm Al filter, 70 kVp, 2 μm voxel size, and 1200 ms integration time. In [fig_ref] Fig 4: Strategies for segmentation of dentoalveolar tissues [/fig_ref] , we show application of median filters and masks to segment cementum, and these techniques have been integrated in our studies. [bib_ref] Hypercementosis associated with ENPP1 mutations and GACI, Thumbigere-Math [/bib_ref] [bib_ref] Genetic and pharmacologic modulation of cementogenesis via pyrophosphate regulators, Chu [/bib_ref] [bib_ref] Reduced orthodontic tooth movement in Enpp1 mutant mice with hypercementosis, Wolf [/bib_ref] In a mouse model of generalized arterial calcification in infancy (GACI), mice genetically ablated for ectonucleotide pyrophosphatase phosphodiesterase 1 (Enpp1 −/− ) featured ectopic calcification due to reduced levels of mineralization inhibitor, inorganic pyrophosphate (PP i ). Compared to controls, Enpp1 −/− mice featured dramatically increased cementum formation [fig_ref] Fig 5: Legend on next page. [/fig_ref]. [bib_ref] Hypercementosis associated with ENPP1 mutations and GACI, Thumbigere-Math [/bib_ref] [bib_ref] Genetic and pharmacologic modulation of cementogenesis via pyrophosphate regulators, Chu [/bib_ref] Acellular and cellular regions were defined by examination of μCT and histology, identifying cellular cementum in the apical one third of the root and acellular cementum in the cervical two thirds of the root. With this separation, Enpp1 −/− mice exhibited a 500% increase in acellular cementum volume, whereas cellular cementum increased 50%, compared to controls. We further evaluated the PDL space for alterations as a result of cementum expansion. [bib_ref] Reduced orthodontic tooth movement in Enpp1 mutant mice with hypercementosis, Wolf [/bib_ref] PDL was segmented using a combination of semi-automated and manual functions (see [fig_ref] Fig 4: Strategies for segmentation of dentoalveolar tissues [/fig_ref] , dilate object, walls). When PDL volumes from WT and Enpp1 −/− mice were overlaid, we noted that PDL width was maintained and in fact volume was increased in Enpp1 −/− compared to WT mice [fig_ref] Fig 5: Legend on next page. [/fig_ref].
## Alveolar bone
Alveolar bone in rodents is studied for many reasons, including to advance our understanding of inherited and acquired bone disorders, therapies to increase bone quality or quantity, orthodontic tooth movement, and the aging process. Prescribed approaches for μCT analysis of long bones are well established. [bib_ref] Guidelines for assessment of bone microstructure in rodents using micro-computed tomography, Bouxsein [/bib_ref] Using current image analysis software packages, detailed quantitative data can be generated, including bone volumes, two dimensional measurements (eg, cross-sectional areas), and characteristics specific to different regions of bone (eg, cortical porosity, trabecular thickness, and periosteal perimeter). Unlike long bones, alveolar bone does not feature easily identifiable regions that are primarily trabecular (like the distal portion of a femur) or cortical (like the midshaft of a long bone). Instead, the alveolar bone is composed of a dense layer of cortical bone with an inner trabecular network of bone in some locations. Consequently, μCT analysis of rodent alveolar bone has been applied with a wide variety of approaches. [bib_ref] Three-dimensional microcomputed tomographic imaging of alveolar bone in experimental bone loss or..., Park [/bib_ref] [bib_ref] Loss of discoidin domain receptor 1 predisposes mice to periodontal breakdown, Chavez [/bib_ref] [bib_ref] Characterization of mandibular bone in a mouse model of chronic kidney disease, Lee [/bib_ref] These inconsistent approaches can be attributed to attempts to account for the heterogeneity of bone structure, organization, and function in the mandible. Earlier, we outlined several ways to analyze alveolar bone (eg, Figs. 4A-D and 5D), including linear measurements, volumetric measurements, and region-specific measurements that can be tailored to specific scientific questions. Additionally, measurements such as trabecular connectivity degree, structure model index, degree of anisotropy, and cortical minimum moment of inertia can be obtained, which can be used in finite elements analysis. In all cases, ROIs must be carefully defined, and limitations should be considered and discussed.
Periodontal diseases are among the most prevalent on earth, causing cementum, PDL, and alveolar bone destruction and tooth loss, significantly affecting oral and overall health and quality of life. [bib_ref] Impact of periodontal disease on quality of life: a systematic review, Ferreira [/bib_ref] [bib_ref] Update on prevalence of periodontitis in adults in the United States: NHANES..., Eke [/bib_ref] [bib_ref] The global burden of periodontal disease: towards integration with chronic disease prevention..., Petersen [/bib_ref] There is great utility in μCT to analyze periodontal pathology, repair, and regeneration. In a study to define the dentoalveolar functions of discoidin domain receptor 1 (DDR1), a collagen receptor tyrosine kinase that regulates cell functions and collagen fibrillogenesis and mineralization, mice genetically ablated for Ddr1 (Ddr1 −/− ) exhibited severe alveolar bone loss with age [fig_ref] Fig 5: Legend on next page. [/fig_ref]. [bib_ref] Loss of discoidin domain receptor 1 predisposes mice to periodontal breakdown, Chavez [/bib_ref] In that study, we employed μCT in three ways to quantify alveolar bone loss in this model: linear measurements, total alveolar bone volume, and a novel approach to measure "alveolar bone proper" (ABP) immediately adjacent to molar roots. In a modified method of the traditional caliper based of CEJ to alveolar bone crest (CEJ-ABC) linear measurement, [bib_ref] Optimization of the ligature-induced periodontitis model in mice, Abe [/bib_ref] there was significant vertical bone loss up to 0.4 mm at five locations around Ddr1 −/− molars. In a 3D volumetric analysis of alveolar bone surrounding all three mandibular molars, a 14% reduction in bone volume was detected in Ddr1 −/− versus WT mice. To better approximate the actual loss of attachment that accompanies periodontal disease, we defined ABP as bone within 240 μm of the tooth root, as measured radially from tooth root surfaces to include buccal, lingual, radicular, and interproximal alveolar bone [fig_ref] Fig 2: Optimization of μCT scan parameters for dentoalveolar structures [/fig_ref]. ABP, also called bundle bone in humans, corresponds to the lamina dura of dental radiographs and includes a high density of Sharpey's fibers that provide continuity between tooth-PDL-alveolar bone tissues.Ddr1 −/− mice exhibited 30% reduction in ABP versus WT. Although μCT can approximate linear measurements of alveolar bone loss, 3D measurement of total alveolar bone better approximates volumetric changes. A strategy like that described for ABP increases sensitivity to detect bone loss even further, putting the focus on the functionally important bundle bone that anchors PDL fibers.
## Artifacts in scans of dentoalveolar tissues
There are several circumstances where μCT scanning artifacts may be created by appliances or materials, including dental implants or springs, wires, and composite materials used for orthodontic tooth movement (OTM) experiments. [bib_ref] Reduced orthodontic tooth movement in Enpp1 mutant mice with hypercementosis, Wolf [/bib_ref] The scan optimization recommendations in this paper were optimized for biologic ranges in density; however, these materials typically have significantly higher densities than oral mineralized tissues. This situation is particularly difficult to correct because one approach to reduce these artifacts would be to increase voltage; however, this will simultaneously reduce quality of biologically relevant densities. Dual-energy and multi-energy X-ray tomography are technologies that would potentially reduce this problem, as well as provide many other advantages in the segmentation of biological tissues. These techniques are currently being developed, and clinical applications are emerging.Dual-energy quantitative X-ray tomography, in which a dual-energy X-ray source is combined with the resolution of individual photon energy in the detector, is currently being developed and has an even greater potential in resolving materials that feature disparate densities within one sample. [bib_ref] Quantitative dual-energy micro-CT with a photon-counting detector for material science and nondestructive..., Sellerer [/bib_ref] However, practical uses of these technologies are still under development. Alternatively, imaging processing algorithms have been proposed to reduce metal and other artifacts in CT scans, [bib_ref] A metal artifact reduction method for a dental CT based on adaptive..., Hegazy [/bib_ref] [bib_ref] Evaluation of two iterative techniques for reducing metal artifacts in computed tomography, Boas [/bib_ref] [bib_ref] Metal artifact reduction on cervical CT images by deep residual learning, Huang [/bib_ref] although, to our knowledge, these methods have been primarily implemented for CT, and few studies have translated them to μCT analysis. [bib_ref] A validated method for titanium implant anchorage analysis using microCT and biomechanical..., Gabet [/bib_ref] Because there may not be a way to optimally prevent or ameliorate artifacts, care should be taken in analysis and interpretation, and limitations should be discussed in publications.
## Summary of guidelines
μCT has become an essential tool for analysis of mineralized tissues in mouse models; however, analyses of murine dentoalveolar tissues present unique challenges. Because of these inherent challenges, important considerations must be taken. Based on our optimization and application of strategies in dentoalveolar studies, we provide a list of recommendations for inclusion in studies utilizing μCT.
- Scanning parameters should be optimized for analysis of dentoalveolar tissues, or at least the target tissue(s) in the study, such that results are not tainted by artifacts or poor image acquisition, leading to spurious results. Parameters should be listed in the methods section, including at a minimum the make and model of scanner, the filter, voltage, and voxel size (resolution). Integration time has not been widely reported in studies, although we found this setting useful for scan optimization. - Calibration to standards should be addressed; eg, if and how samples underwent calibration, and whether absolute or relative units were used in quantitative data. - Orientation of samples should be described using appropriate anatomical landmarks. - ROI should be carefully defined and/or shown by a figure.
- Segmentation techniques should be detailed in description of methods. Details should include thresholds applied, use of automated or semi-automated approaches, and application of manual corrections. If tissues are subdivided for analysis, this must be carefully described, possibly with inclusion of images depicting subregions. - Analysis software should be cited, and digital tools used for segmentation should be listed.
Our goals for this review parallel those outlined by Bouxsein and colleagues [bib_ref] Guidelines for assessment of bone microstructure in rodents using micro-computed tomography, Bouxsein [/bib_ref] and Vardelis and Salmon [bib_ref] Microcomputed tomography imaging in odontogenesis studies, Verdelis [/bib_ref] in their indispensable reports. Specifically, this review is not meant to dictate specific approaches for assessing dentoalveolar tissues in mouse models, but rather to promote increased transparency and reproducibility, encourage best practices, and provide a basic framework that can be adapted to apply μCT analysis to dentoalveolar tissues. The methods and strategies described here for mouse dentoalveolar tissues may be extrapolated to dentoalveolar tissues of larger animals or to other skeletal regions. This requires understanding the use of X-ray scanning instruments across lengthscales that are sensitive to space, density, and fields of view, but can be accomplished using methods described here.
# Disclosures
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 manuscript.
[fig] Fig 1: Murine dentoalveolar anatomy. (A) 3D μCT reconstruction of murine skull highlighting key landmarks including mandibular molars (M1-M3). (B) Schematic of mouse molar, incisor, and surrounding structures (coronal plane). (C,D) Radiograph of 90 days postnatal murine mandible highlighting dentoalveolar structures distinguishable by density and X-ray absorbance: enamel, dentin/cementum, AB, and radiolucent unmineralized tissues including the PDL and pulp. AB = alveolar bone; PDL = periodontal ligament. [/fig]
[fig] Fig 2: Optimization of μCT scan parameters for dentoalveolar structures. All panels depict the same 6-week-old mouse mandible that has been scanned and calibrated. Each optimization scan includes a grayscale image, a heat map of densities, and a density histogram reporting the number of voxels detected in the scan. Inserts in histogram panels show a higher magnification of densities corresponding to enamel-like densities. Color scale for heat maps and histograms is at bottom offigure.(A) To test the effects of prismatic filters, scans were performed with no filter, 0.5-mm Al filter, or 0.1-mm Cu filter.(B) To test effects of voltage, scans were performed with 55, 70, or 90 kVp. (C) To test the effects of resolution, scans were performed at voxel size of 20, 10, or 6 μm. (D) To test the effects of integration time, scans were performed at 300, 600, or 900 ms. Discussion of scanning results and optimization of parameters are described in the text. JBMR Plus (WOA) n 4 of 17 CHAVEZ ET AL. [/fig]
[fig] Fig 3: Orientation and ROIs in the dentoalveolar complex. (A) 3D and 2D representative images demonstrating schematic of sagittal, frontal/coronal, and transverse/axial planes anatomical planes for consistent orientation of mouse mandible. Red lines indicate intersecting planes for optimal orientation and consistency. (B) ROI was determined after mandible reorientation by finding the M1 mesial and distal edges (red dotted lines) and expanding the region 480 μm mesial and distal (black dotted lines) to include the alveolar bone in these regions. The shaded area indicates regions excluded from analysis. ROI = region of interest. [/fig]
[fig] Fig 4: Strategies for segmentation of dentoalveolar tissues. (A1-A12) Selection of threshold values (mg HA/cm 3 ) will affect calculation of EV, DV, and BV measurements. (A1-A6) Thresholds and resulting 2D images and calculated volumes are reported in the top row; [/fig]
[fig] Fig 5: Legend on next page. [/fig]
[table] Table 1: Selection of μCT Scan ParametersVariables including filter, voltage, voxel size, and integration time were optimized for murine dentoalveolar tissues. [/table]
|
Challenges in Noninvasive Skin Biomarker Measurements in Daily Practice: A Longitudinal Study on Skin Surface Protein Detection by the Transdermal Analysis Patch in Pediatric Psoriasis
Introduction: Skin surface proteins are potential biomarkers in psoriasis and can be measured noninvasively with the transdermal analysis patch (TAP). This study aimed to assess markers measured by TAP over time in daily clinical practice, explore their correlation with disease severity in pediatric psoriasis, and compare the TAP and tape stripping detection capability. Methods: In this prospective observational daily clinical practice study, pediatric psoriasis patients (aged >5 to <18 years) were followed during 1 year. At each visit, TAPs were applied to lesional (n = 2), peri-lesional (n = 2), and nonlesional (n = 1) sites. Post-lesional skin was sampled if all lesions on the arms, legs, or trunk cleared. Treatment and psoriasis severity data were collected.
bility of the TAP was compared to tape stripping in a separate cohort of adult psoriasis patients. Results: 32 patients (median age 15.0 years, median Psoriasis Area and Severity Index [PASI] 5.2) were followed for a mean of 11.3 (±3.4) months with a total of 104 visits. In lesional skin (n = 197), significantly higher IL-1RA, hBD-2, IL-8, VEGF, CXCL-1/2, IL-23, hBD-1, IL-22, CCL-27, and IL-17A levels were found compared to non-lesional skin (n = 104), while IL-1α was higher in non-lesional skin. Marker levels were highly variable over time and did not correlate with disease severity measured by PASI or SUM scores. Comparison of the TAP and tape strip detection capability in adult psoriasis patients (n = 10) showed that lesional hBD-2, IL1-α, IL-8, and VEGF and nonlesional IL-1RA, hBD-2, IL-8, and VEGF were more frequently detected in tape extracts than TAPs. Conclusion: Due to the lack of correlation with clinical disease severity and the current detection capability of the markers measured by TAP in psoriasis, its use in regular practice is still a bridge too far.
# Introduction
Psoriasis is a common chronic inflammatory skin disease with an onset during childhood in almost one-third of all cases. Its disease course is characterized by exacerbations and remissions and is unpredictable: some patients have a mild and stable disease for many years, while in other patients, psoriasis might progress quickly [bib_ref] The importance of early treatment in psoriasis and management of disease progression, Kerdel [/bib_ref]. Our understanding of the pathophysiology of (plaque) psoriasis and involved cytokines (e.g., IL-17, IL-23, TNF-α) has rapidly increased over the past decades [bib_ref] Pathophysiology, clinical presentation, and treatment of psoriasis: a review, Armstrong [/bib_ref]. If biomarkers were able to predict disease course or response to treatment, their added value to daily practice would be considerable, especially in facilitating early intervention and improving treatment decisions. The skin surface provides a unique source of potential biomarkers as cytokines, chemokines, growth factors, and antimicrobial peptides can be measured directly at the site of the disease and are thought to reflect the underlying pathophysiology [bib_ref] β-Defensin-2 protein is a serum biomarker for disease activity in psoriasis and..., Jansen [/bib_ref] [bib_ref] Cytokines in psoriasis, Baliwag [/bib_ref] [bib_ref] Noninvasive proteome analysis of psoriatic stratum corneum reflects pathophysiological pathways and is..., Mehul [/bib_ref] [bib_ref] Thongboonkerd V. Proteomics in psoriasis, Chularojanamontri [/bib_ref].
Proteins in psoriatic skin are conventionally sampled through skin biopsies [bib_ref] Expression of Th17 cytokines in skin lesions of patients with psoriasis, Li [/bib_ref] [bib_ref] Proteogenomic analysis of psoriasis reveals discordant and concordant changes in mRNA and..., Swindell [/bib_ref] [bib_ref] Skininfiltrating, interleukin-22-producing T cells differentiate pediatric psoriasis from adult psoriasis, Cordoro [/bib_ref] [bib_ref] β-Defensin 2 is a responsive biomarker of IL-17A-driven skin pathology in patients..., Kolbinger [/bib_ref] , which are associated with a risk of pain, scarring, and infection [bib_ref] Why minimally invasive skin sampling techniques? A bright scientific future, Wang [/bib_ref]. Noninvasive biomarker measurements provide a patient-friendly alternative by avoiding discomfort and fear for interventions, especially in pediatric patients, and can be performed repeatedly, making them more suitable for use in daily clinical practice. Tape stripping, during which corneocytes of the superficial stratum corneum are sampled with adhesive tape, is regarded as the golden standard for noninvasive skin protein sampling and is widely performed [bib_ref] Noninvasive proteome analysis of psoriatic stratum corneum reflects pathophysiological pathways and is..., Mehul [/bib_ref] [bib_ref] An analysis of select pathogenic messages in lesional and non-lesional psoriatic skin..., Benson [/bib_ref] [bib_ref] Use of tape strips to detect immune and barrier abnormalities in the..., Guttman-Yassky [/bib_ref]. Previous tape stripping studies have underlined the value and power of local biomarkers [bib_ref] Noninvasive proteome analysis of psoriatic stratum corneum reflects pathophysiological pathways and is..., Mehul [/bib_ref] [bib_ref] Use of tape strips to detect immune and barrier abnormalities in the..., Guttman-Yassky [/bib_ref]. In psoriasis and/or atopic dermatitis, numerous proteins were successfully detected in the stratum corneum using tape stripping, such as interleukins (e.g., IL-1α, IL-1β) or chemokines (e.g., CXCL-1, CXCL-8) [bib_ref] Noninvasive proteome analysis of psoriatic stratum corneum reflects pathophysiological pathways and is..., Mehul [/bib_ref] [bib_ref] Use of tape strips to detect immune and barrier abnormalities in the..., Guttman-Yassky [/bib_ref]. Recently, the transdermal analysis patch (TAP) and the FibroTx patch were described [bib_ref] Development of TAP, a non-invasive test for qualitative and quantitative measurements of..., Orro [/bib_ref] [bib_ref] Non-invasive assessment of soluble skin surface biomarkers in atopic dermatitis patients-effect of..., Røpke [/bib_ref] [bib_ref] Skin surface protein detection by transdermal analysis patches in pediatric psoriasis, Schaap [/bib_ref]. These methods sample soluble proteins from an intact stratum corneum. Previous research by our group revealed that the TAP can detect skin proteins in lesional, peri-lesional, and nonlesional skin in pediatric psoriasis patients treated with systemic and/or topical agents and is regarded as patient friendly [bib_ref] Development of TAP, a non-invasive test for qualitative and quantitative measurements of..., Orro [/bib_ref] [bib_ref] Skin surface protein detection by transdermal analysis patches in pediatric psoriasis, Schaap [/bib_ref].
The consistency of proteins measured by TAP over time and the correlation with disease severity are not studied to date but are relevant to identify potential biomarkers for use in clinical practice. Additionally, the skin surface protein detection capability of the TAP might dif-fer from that of tape stripping since these methods sample at different depths of the stratum corneum and include different detection methods. The primary objective of this study, performed in a daily clinical practice setting, was to explore the value of TAP measurements in daily practice in (pediatric) psoriasis patients by (i) comparing the marker levels measured by TAP in lesional, peri-lesional, non-lesional, and post-lesional skin, (ii) assessing the course of skin surface markers measured by TAP over time in pediatric psoriasis patients, and (iii) exploring the correlation between lesional marker levels and disease severity. In addition, the marker detection capability of TAP was compared to tape stripping extraction in a pilot study in adults with psoriasis.
# Materials and methods
## Study design and population
In this prospective observational daily clinical practice study, pediatric and adolescent psoriasis patients aged 5-18 years were recruited between June 2018 and July 2019 at the outpatient clinic of the Department of Dermatology of the Radboud University Medical Center, Nijmegen, The Netherlands. Inclusion criteria were a dermatologist-confirmed plaque psoriasis diagnosis and sufficient psoriasis plaques to apply the TAPs at baseline. Patients with another concurrent inflammatory skin disease or other types of psoriasis were excluded. Treatment occurred as part of regular care without a washout phase and could consist of topical and/or systemic treatment. Patients were followed during 1 year, and visits were planned every 3 months for systemically treated patients and every 6 months for solely topically treated patients. However, no visits occurred between March and August 2020 due to COVID-19 restrictions. Written informed consent was given by all participants aged ≥16 years. Participants aged 12-16 years gave written informed assent. Legal guardians of participants aged under 16 years gave written informed consent before enrollment. The study was approved by the Ethics Committee of the Radboud University Medical Center, region Arnhem-Nijmegen (NL60952.091.17).
# Transdermal analysis patch
The TAP is developed and commercialized by FibroTx and consists of a multiplex capture-antibody microarray supported by an adhesive bandage. The TAP method is described in the online supplementary Materials (for all online suppl. material, see www. karger.com/doi/10.1159/000527258) and previous publications [bib_ref] Development of TAP, a non-invasive test for qualitative and quantitative measurements of..., Orro [/bib_ref] [bib_ref] Skin surface protein detection by transdermal analysis patches in pediatric psoriasis, Schaap [/bib_ref]. Two TAPs were used with different preset protein panels. One TAP measured CXC chemokine ligand (CXCL)-1/2, interleukin (IL)-1RA, IL-23, IL-1α, IL-8, vascular endothelial growth factor (VEGF), and human beta-defensin (hBD)-2, and a second TAP measured CC chemokine ligand (CCL)-27, IL-4, IL-22, IL-17A, hBD-1, and kallikrein-related peptidase (KLK)-5. After application, four drops of phosphate-buffered saline (pH 7.4) were added to microarray reservoir. TAPs were applied to the skin for 20 min to capture skin-derived proteins through immune recognition and were stored at −20°C until quantification with spot-enzyme-linked immunosorbent assay (spot-ELISA). Levels within half of the low-Skin Pharmacol Physiol 2022;35:319-327 DOI: 10.1159/000527258 er detection limit were taken unchanged, lower values were substituted by zero. The TAP detection limits are provided in online supplementary [fig_ref] Table 1: Baseline characteristics Switched from the therapy group during follow-up, n [/fig_ref].
## Study procedures
TAPs were applied to two lesional, two peri-lesional, and one non-lesional skin sites (no psoriasis within a distance of 10 cm). The two lesions of interest were preferably located on different body parts, including the arms, legs, or trunk, with preferably a similar SUM score. TAPs were applied to the same skin sites at each visit during follow-up if possible. However, if all lesions on the arms, legs, or trunk cleared, TAPs were applied to post-lesional skin instead of lesional skin. Patients did not use topical agents on the investigated sites on the day of the visit.
After TAP removal, pediatric patients rated experienced discomfort on a simplified 10-point visual analogue scale (VAS). A VAS score of 0 and 10 corresponded to no and maximal discomfort, respectively. Psoriasis severity was measured with the Psoriasis Area and Severity Index (PASI; range 0-72) score, Physician Global Assessment (PGA; range 0-5), and affected Body Surface Area (BSA) [bib_ref] Severe psoriasis: oral therapy with a new retinoid, Fredriksson [/bib_ref] [bib_ref] Psoriasis disease severity measures: comparing efficacy of treatments for severe psoriasis, Weisman [/bib_ref] [bib_ref] The 5-point Investigator's Global Assessment (IGA) Scale: a modified tool for evaluating..., Langley [/bib_ref]. SUM scores (0-12), defined as the sum of the severity scores for erythema (0-4), induration (0-4), and desquamation (0-4), were determined for the lesions of interest [bib_ref] Topical treatment of mild to moderate plaque psoriasis with 0.3% tacrolimus gel..., Vissers [/bib_ref]. Additionally, demographics and information on current treatment were collected. All procedures were performed by two physicians (MJS or FMB).
## Tap versus tape stripping
To compare the marker detection capability of TAP with tape stripping, we performed a pilot study in adult psoriasis patients from September to October 2020. Inclusion criteria were diagnosis of plaque psoriasis confirmed by a dermatologist, a plaque large enough to perform both TAP and tape stripping, and no other concurrent inflammatory skin disease. TAP and tape stripping were consecutively performed at one time point in a randomized order on one lesional and non-lesional skin site. Directly after each method, experienced discomfort was rated by the patient on a 100mm VAS. TAP sampling and analysis were performed as previously described. For tape stripping, eight successive round adhesive tapes with a diameter of 2.2 cm (DSquame, CuDerm, USA) were pressed to the skin for 10 s using a pressure device (CuDerm, USA). Tapes were stored in cryovials at −80°C until analysis. Proteins were extracted from the tapes using ultrasonication with PBS and were either quantified by conventional ELISA (hBD-2) or Luminex multiplex-based platform (CXCL-1/2, IL-1RA IL-1α, IL-8, and VEGF) and normalized to the total amount of protein. Values exceeding the detection limits were not included in the analysis. Detailed methods are provided in the online supplementary Material. All patients gave written informed consent. The study was approved by the Ethics Committee of the Radboud University Medical Center, region Arnhem-Nijmegen (NL73363.091.20).
# Statistical analysis
Patient characteristics and marker levels were first analyzed with descriptive statistics and presented as frequencies and percentages, means and standard deviations (±SD), or medians and interquartile ranges (IQRs). Marker levels on both the two lesional and peri-lesional sites within 1 patient were compared on group level with a Wilcoxon signed rank test. Mann-Whitney U tests were performed to compare marker levels in lesional, peri-lesional, non-lesional, and post-lesional skin. Mean lesional marker lev-els and the PASI score were plotted over time in each patient. Spearman rank correlation tests were computed for the PASI score and mean lesional marker levels and for lesion severity (total SUM score, desquamation, erythema, induration) and lesional marker levels. The TAP and tape stripping detection capability was assessed by comparing the number of samples that were detected within the limits of the assay for each marker. Statistical package SPSS, version 25 (IBM, Armonk, NY) and SAS 9.4 (SAS Institute, Inc., Cary, NC, USA) were used to perform analyses. A two-sided p < 0.05 was regarded statistically significant.
# Results
## Patient characteristics
Thirty-two patients were included with a median age of 15.0 years (IQR 10.9-16.9) and a median PASI score of 5.2 (IQR 3.7-8.7) at baseline [fig_ref] Table 1: Baseline characteristics Switched from the therapy group during follow-up, n [/fig_ref]. A mean of 3.25 BSA, body surface area; PGA, Physician Global Assessment. a Unless stated otherwise. b The 25th and 75th percentile are shown. c Psoriasis Area and Severity Index score (range 0-72). d Physician Global Assessment (range 0-5). e Body surface area (%). f All lesions of interest combined, resulting in a total of 64 lesions (2 lesions per patient) at baseline. g Patients may also use topical treatments (corticosteroids and/or vitamin D derivatives). DOI: 10.1159/000527258 (±1.0) visits were performed per patient, resulting in a total of 104 visits and a mean follow-up duration of 11.3 (±3.4) months. Due to COVID-19 restrictions, 10 patients dropped out before 1 year of follow-up, longer visit intervals occurred toward the end of follow-up, and follow-up exceeded 1 year in 14 patients. Of the 10 patients that dropped out early, still two to four visits were performed and data of these visits were included in the analysis. At baseline, 19 patients received solely topical treatment, while 13 patients received systemic treatment. During follow-up, 4 patients switched from topical to systemic treatment, while 2 patients switched from systemic to solely topical treatment [fig_ref] Table 1: Baseline characteristics Switched from the therapy group during follow-up, n [/fig_ref]. Mild adverse events of TAP were reported during six visits, including transient local erythema (n = 2), itch (n = 3) and a burning sensation (n = 1). The median VAS score after TAP removal was 1.0 (IQR 0.0-1.0).
## Marker level differences between lesional, peri-lesional, non-lesional, and post-lesional skin
When including all samples on both the two lesional and two peri-lesional sites, marker levels were considered similar (online suppl. . Therefore, all data were taken together for further analysis, resulting in 197 lesional, 197 peri-lesional, 104 non-lesional, and 6 post-lesional TAP measurements. The TAP was able to distinguish lesional from non-lesional skin: significantly higher levels of IL-1RA, hBD-2, IL-8, VEGF, CXCL-1/2, CCL-27, IL-23, hBD-1, and IL-17A were found in lesional skin, whereas higher levels of IL-1α were found in non-lesional skin . In peri-lesional skin, IL-1RA, hBD-2, IL-8, and VEGF levels remained significantly higher compared to non-lesional skin. IL-4 was only (and barely) detected in two lesional samples and did not show any significant relations. In addition, if all lesions on the arms, legs, or trunk cleared during follow-up, TAPs were applied to post-lesional skin (online suppl. [fig_ref] Figure 1: PASI scores and lesional marker levels measured by TAP over time [/fig_ref]. Intriguingly, post-lesional marker levels were equal to lesional levels or even significantly higher (for hBD-2 and KLK-5, . Moreover, compared to non-lesional skin, levels of hBD-2, IL-8, VEGF, IL-17A, and KLK-5 were significantly higher in post-lesional skin. A post hoc analysis solely including the patients in which post-lesional skin was sampled during follow-up revealed similar trends.
## Lesional marker levels over time and their correlation with disease severity
Detected lesional marker levels highly varied between analyzed skin surface proteins . To explore the course of marker levels during follow-up in each patient, mean lesional marker levels were plotted over time. PASI scores were added to these figures to explore if the trend of marker levels followed the overall disease severity. Regarding the association between marker concentrations and disease severity (PASI score) over time, inter-and intra-patient differences were seen. Four representative patients are depicted in [fig_ref] Figure 1: PASI scores and lesional marker levels measured by TAP over time [/fig_ref]. To further explore this association, correlations between marker levels and the PASI score were calculated [fig_ref] Table 3: Correlation between lesional marker concentrations and disease severity [/fig_ref]. These correlations were weak (<0.30) and not statistically significant (except for KLK-5). Mostly inverse correlations were obtained, indicating lower marker levels for higher PASI scores [fig_ref] Table 3: Correlation between lesional marker concentrations and disease severity [/fig_ref]. Furthermore, the association between marker concentrations and lesion severity scores (SUM scores) was assessed. Analysis showed mostly inverse correlations between desquamation severity and marker levels, reaching statistical significance for IL-1α, hBD-1, and KLK-5 [fig_ref] Table 3: Correlation between lesional marker concentrations and disease severity [/fig_ref]. The possible association between lower marker levels and more desquamation was further sub-stantiated by calculating mean marker levels for each local desquamation score. As the amount of lesional desquamation increased, a clear trend of decreasing IL-1α, VEGF, CCL-27, IL-23, hBD-1, IL-22, IL-17A, and KLK-5 levels was observed (online suppl. [fig_ref] Table 3: Correlation between lesional marker concentrations and disease severity [/fig_ref]. This trend was less distinct for the local erythema and induration scores (data not shown). Of note, only for IL-8, a weak positive correlation coefficient was found with lesion severity scores (<0.3; p < 0.01, [fig_ref] Table 3: Correlation between lesional marker concentrations and disease severity [/fig_ref].
## Tap versus tape stripping detection capability
To compare the detection capability of TAP with tape stripping, an additional study was performed including noninvasive protein profiling of stratum corneum by both TAP and tapes. Both methods were performed on the same lesion and non-lesional skin site in 10 adult psoriasis patients (median age 58.0, median PASI 2.0) at one time point. Most patients were treated with systemic agents (n = 7; 70.0%). Full population characteristics are [fig_ref] Table 4: Marker levels in lesional and non-lesional skin of adult psoriasis patients [/fig_ref]. In general, markers were more often detected, or detected within the detection limits, in stratum corneum extracts from tape strip samples. Specifically, lesional hBD-2, IL1-α, IL-8, and VEGF and non-lesional IL-1RA, hBD-2, IL-8, and VEGF were more frequently detected in tape extracts (Table 4). Regarding the implementation of the sample techniques in daily practice, patients reported a low median VAS score for discomfort for both procedures: 0.0 mm (IQR 0.0-0.0 mm) for the TAP versus 0.00 (IQR 0.0-6.0 mm) for tape stripping. No adverse events were reported.
# Discussion
This study aimed to explore the value of TAP measurements in daily practice in (pediatric) psoriasis patients. We assessed skin surface proteins measured by TAP over time, determined their correlation with psoriasis severity, and compared the TAP detection capability with tape stripping. We showed that marker levels measured by TAP can differentiate between lesional and non-lesional skin. Surprisingly, marker levels in post-lesional skin were equal to or higher than lesional levels and signifi-cantly higher than non-lesional levels. Lesional marker levels varied highly over time, and no convincing correlations were found between these marker levels and PASI or SUM scores. In addition, the protein detection capability by tape stripping appeared superior to the TAP for five out of six quantified markers in a separate cohort of 10 adults with psoriasis. In the next paragraphs, we will discuss the implication of these results.
In line with our previous publication on the baseline analysis of this cohort [bib_ref] Skin surface protein detection by transdermal analysis patches in pediatric psoriasis, Schaap [/bib_ref] , IL-1RA, hBD-2, IL-8, VEGF, and CXCL-1/2 levels were significantly increased in lesional skin, while the level of IL-1α was increased in nonlesional skin compared to lesional skin. In this follow-up analysis, IL-23, IL-22, hBD-1, CCL-27, and IL-17A were also significantly increased in lesional skin. These results are in line with a previous study by Mehul et al. [bib_ref] Noninvasive proteome analysis of psoriatic stratum corneum reflects pathophysiological pathways and is..., Mehul [/bib_ref] that used tape stripping for stratum corneum proteome profiling in adult psoriasis patients, which also found higher levels of VEGF, CXCL-1/2, IL-8, CCL-27, and IL-17A and lower levels of IL-1α in lesional skin. In contrast, we found higher levels of IL-22 in the pediatric psoriatic stratum corneum, while this was not increased in adults in the study by Mehul et al. [bib_ref] Noninvasive proteome analysis of psoriatic stratum corneum reflects pathophysiological pathways and is..., Mehul [/bib_ref]. It has previously been suggested that IL-22 could differentiate pediatric from adult psoria- Spearman correlations are shown for lesion or disease severity and lesional marker concentrations. All lesional data were included in this analysis. No correction for multiple measures within 1 patient was performed. a For the correlation with lesion severity, the two lesional measurements at each patient visit were included separately and post-lesional samples were excluded, resulting in a total of 197 included measurements. Total SUM scores ranged from 2 to 9, desquamation scores ranged from 0 to 3, erythema scores ranged from 1 to 3, and induration scores ranged from 0 to 4. b For the correlation with the Psoriasis Area and Severity Index score, the mean lesional protein concentration was used for each patient visit. PASI scores ranged from 1. [bib_ref] Skininfiltrating, interleukin-22-producing T cells differentiate pediatric psoriasis from adult psoriasis, Cordoro [/bib_ref]. As expected from the pathophysiology of psoriasis and previous studies [bib_ref] Noninvasive proteome analysis of psoriatic stratum corneum reflects pathophysiological pathways and is..., Mehul [/bib_ref] [bib_ref] Skin surface protein detection by transdermal analysis patches in pediatric psoriasis, Schaap [/bib_ref] , IL-4 was barely detected: only in 2 out of 197 lesional samples.
In the pediatric cohort, a low correlation between TAP marker levels and disease severity was found. Since TAPs measured lower marker concentrations as the amount of lesional desquamation increased, excessive desquamation in psoriasis seems to hamper sampling of soluble skin surface proteins. Our results indicate that tape stripping has a greater protein detection capability and a greater fold change between non-lesional and lesional samples, which further substantiates potential detection issues using the TAP in psoriasis. The greater protein detection capability of tape stripping could be explained by sampling deeper layers of the stratum corneum, and a greater sensitivity of quantification methods used for the tape extracts (herein used Luminex platform) than the spot-ELISA used for the TAPs. As the pilot study in adults with psoriasis solely aimed to compare the detection capability of TAP and tape stripping, statements about the potential clinical value of markers derived from tape strips in psoriasis cannot be made. Further research is needed regarding the consistency of these marker levels over time and their correlation with disease severity. A second explanation for the low correlation between TAP marker levels and disease severity is the fact that the marker set captured with the TAP was a preset general inflammatory panel, thus not specific for psoriasis. Hence, other psoriasis-specific markers might yield better correlations with disease severity. Another hypothesis to explain the low correlation between TAP marker levels and disease severity might be that the marker levels detected on the stratum corneum surface reflect a delayed representation of the activity in the skin. In line with this hypothesis, measured post-lesional marker levels were equal to or higher than lesional levels. However, we postulate that actual lesional levels were even higher than post-lesional levels, but excessive desquamation resulted in an under-detection of lesional marker levels. Nonetheless, hBD-2, IL-8, VEGF, IL-17A, and KLK-5 levels were significantly higher in post-lesional compared to non-lesional skin. Multiple studies have revealed that molecular and cellular imprints of psoriasis do not fully disappear in clinically cleared skin [bib_ref] Resolved psoriasis lesions retain expression of a subset of disease-related genes, Suárez-Fariñas [/bib_ref] [bib_ref] Clinically resolved psoriatic lesions contain psoriasis-specific IL-17-producing αβ T cell clones, Matos [/bib_ref] [bib_ref] Resolution of plaquetype psoriasis: what is left behind (and reinitiates the disease), Benezeder [/bib_ref]. However, given the small number of post-lesional samples (n = 6), we are unable to make valid statements regarding the presence of a residual inflammatory scar, and results have to be confirmed in more post-lesional samples.
We note several limitations. Given the explorative study design, we did not perform correction for multiple measurements within 1 patient since we did not want to miss any potential differences or correlations. Moreover, the marker set captured with the TAP was a preset general inflammatory panel, thus not specific for psoriasis. Additionally, a relatively low number of participants were included: 32 pediatric and 10 adult psoriasis patients. Given the daily practice setting, it was not feasible to study parameters for which a controlled setting is required (e.g., sensitivity or specificity). Lastly, the daily practice setting resulted in uncontrolled variables (such as treatment), possibly disturbing correlations between marker levels and psoriasis severity. However, it is important that correlations with disease severity are still present in a daily practice setting in order to be of value for clinical practice. Therefore, we underline the relevance of biomarker studies in a daily practice setting. This study highlights important challenges for noninvasive biomarker measurements by TAP in pediatric psoriasis in a daily practice setting. Highly variable marker levels during the course of the disease were seen, and robust correlations between lesional marker levels measured by TAP and psoriasis severity could not be established. Moreover, detection issues related to desquamation were seen. Tape stripping may be considered instead of TAP for protein sampling of lesions with excessive desquamation, like in psoriasis. In the future, noninvasive biomarker measurements in daily practice could be of added value for managing psoriasis in the individual patient. However, in its current form, the use of biomarker measurements by TAP in pediatric psoriasis patients in daily clinical practice is still a bridge too far.
[fig] 2: Marker concentrations measured in lesional, peri-lesional, non-lesional, and post- [/fig]
[fig] Figure 1: PASI scores and lesional marker levels measured by TAP over time. PASI scores (depicted on the left y axis) and mean lesional marker levels (depicted on the right y axis) are visualized over time for four representative patients. The follow-up duration is plotted on the x axis with intervals of 3 months. Marker levels are shown in ng/mL. DOI: 10.1159/000527258 shown in online supplementary [/fig]
[table] Table 1: Baseline characteristics Switched from the therapy group during follow-up, n (%) [/table]
[table] Table 3: Correlation between lesional marker concentrations and disease severity [/table]
[table] Table 4: Marker levels in lesional and non-lesional skin of adult psoriasis patients (N = 10) measured by TAP and tape stripping [/table]
|
Mimicking associative learning using an ion-trapping non-volatile synaptic organic electrochemical transistor
Associative learning, a critical learning principle to improve an individual's adaptability, has been emulated by few organic electrochemical devices. However, complicated bias schemes, high write voltages, as well as process irreversibility hinder the further development of associative learning circuits. Here, by adopting a poly(3,4-ethylenedioxythiophene):tosylate/ Polytetrahydrofuran composite as the active channel, we present a non-volatile organic electrochemical transistor that shows a write bias less than 0.8 V and retention time longer than 200 min without decoupling the write and read operations. By incorporating a pressure sensor and a photoresistor, a neuromorphic circuit is demonstrated with the ability to associate two physical inputs (light and pressure) instead of normally demonstrated electrical inputs in other associative learning circuits. To unravel the non-volatility of this material, ultraviolet-visible-near-infrared spectroscopy, X-ray photoelectron spectroscopy and grazingincidence wide-angle X-ray scattering are used to characterize the oxidation level variation, compositional change, and the structural modulation of the poly(3,4-ethylenedioxythiophene):tosylate/Polytetrahydrofuran films in various conductance states. The implementation of the associative learning circuit as well as the understanding of the non-volatile material represent critical advances for organic electrochemical devices in neuromorphic applications.
T he human brain is a complex computing machine that contains~10 [bib_ref] Short-term plasticity and long-term potentiation mimicked in single inorganic synapses, Ohno [/bib_ref] neurons interconnected by~10 15 synapses [bib_ref] Do we have brain to spare?, Drachman [/bib_ref] [bib_ref] Artificial synapses emulated by an electrolyte-gated tungstenoxide transistor, Yang [/bib_ref]. It works in a highly parallel, fault-tolerant, and energy-efficient manner that can easily outperform modern computers in some complex and unstructured tasks, such as pattern recognition, motor control, and multisensory integration [bib_ref] The blue brain project, Markram [/bib_ref] [bib_ref] Building the human brain, Machens [/bib_ref] [bib_ref] An energy budget for signaling in the grey matter of the brain, Attwell [/bib_ref] [bib_ref] The attention system of the human brain, Posner [/bib_ref]. Inspired by the operation of human brain, synaptic devices have attracted widespread interest in recent years [bib_ref] Synaptic electronics: materials, devices and applications, Kuzum [/bib_ref] [bib_ref] Electric-double-layer transistors for synaptic devices and neuromorphic systems, He [/bib_ref] [bib_ref] Nanoionics-enabled memristive devices: strategies and materials for neuromorphic applications, Wang [/bib_ref] [bib_ref] Organic electronics for neuromorphic computing, Van De Burgt [/bib_ref] , and have great potential to serve as building blocks for non-von Neumann neuromorphic computing in the nextgeneration artificial intelligence. As one of the most wellstudied components of synaptic devices, inorganic memristors based on various switching mechanisms have been developed and extensively adopted as artificial synapses [bib_ref] Short-term plasticity and long-term potentiation mimicked in single inorganic synapses, Ohno [/bib_ref] [bib_ref] Learning abilities achieved by a single solid-state atomic switch, Hasegawa [/bib_ref] [bib_ref] Demonstration of synaptic behaviors and resistive switching characterizations by proton exchange reactions..., Chang [/bib_ref] [bib_ref] Self-assembled networked PbS distribution quantum dots for resistive switching and artificial synapse..., Yan [/bib_ref] [bib_ref] Stochastic phase-change neurons, Tuma [/bib_ref] [bib_ref] Brain-like associative learning using a nanoscale nonvolatile phase change synaptic device array, Eryilmaz [/bib_ref]. However, application challenges including non-linearity in the write process, energy-costly switching, and lack of biocompatibility for tissue interfacing still exist [bib_ref] Flexible ionic-electronic hybrid oxide synaptic TFTs with programmable dynamic plasticity for brain-inspired..., John [/bib_ref] [bib_ref] Low-power, electrochemically tunable graphene synapses for neuromorphic computing, Sharbati [/bib_ref] , limiting their implementation. Organic synaptic transistors present a natural alternative: they show low voltage writing [bib_ref] Dynamically reconfigurable short-term synapse with millivolt stimulus resolution based on organic electrochemical..., Ling [/bib_ref] , continuous tuning of conductance states 20 , materials tunability, and biocompatibility 21 as well as potential for low-cost and large-area manufacturing [bib_ref] Neuromorphic device architectures with global connectivity through electrolyte gating, Gkoupidenis [/bib_ref]. Based on these unique properties, organic synaptic transistors can not only simulate basic synaptic functions like short-term plasticity (STP) [bib_ref] Artificial synapse network on inorganic proton conductor for neuromorphic systems, Zhu [/bib_ref] , long-term plasticity (LTP) [bib_ref] A carbon nanotube synapse with dynamic logic and learning, Kim [/bib_ref] , and spiking timedependent plasticity (STDP) [bib_ref] Organic core-sheath nanowire artificial synapses with femtojoule energy consumption, Xu [/bib_ref] but are also able to function as biomimetic devices to directly interface with living tissue [bib_ref] A bioinspired flexible organic artificial afferent nerve, Kim [/bib_ref] [bib_ref] Organic bioelectronics for electronic-to-chemical translation in modulation of neuronal signaling and machine-to-brain..., Larsson [/bib_ref] which is critical for next-generation bioelectronics capable on high order signal processing and analysis.
Simulating the learning process of the human brain such as associative learning is important in bioelectronics and brain-computer interface (BCI) research. As a fundamental learning principle, associative learning is particularly important for an individual's adaptability [bib_ref] Associative learning of social value, Behrens [/bib_ref]. The best-known associative learning experiment is the Pavlov's dog experiment, which claims that the conditioned stimulus (CS, ring of bell) can only trigger the unconditioned response (UR, salivation of dog) after the dog has undergone training sessions incorporating both CS and unconditioned stimulus (US, sign of food). To simulate the associative learning, an organic synaptic transistor with nonvolatile memory property whose conductance state can be tuned during training is essential. van de Burgt et al. developed an electrochemical neuromorphic organic device (ENODe) whose conductance can be stabilized by decoupling the write and read process, which can be achieved by either physically disconnecting the gate electrode [bib_ref] A non-volatile organic electrochemical device as a lowvoltage artificial synapse for neuromorphic..., Van De Burgt [/bib_ref] , or by introducing an access device like conductive bridge memory (CBM) at the gate terminal [bib_ref] Parallel programming of an ionic floating-gate memory array for scalable neuromorphic computing, Fuller [/bib_ref]. Their device shows extreme low switching energy <10 pJ. However, in this device, the presynaptic terminal (gate) must be decoupled from the channel in order to maintain non-volatility when simulating associative learning. As a result, additional access devices or high input impedance on gate terminal is necessary, which complicates the circuit design. Gerasimov et al. achieved associative learning by adopting the in situ polymerization of monomer to alter the conductance in the channel of the transistor [bib_ref] An evolvable organic electrochemical transistor for neuromorphic applications, Gerasimov [/bib_ref]. However, the monomer contained electrolyte and the irreversible polymerization process limits the potential of this approach. Another approach has been employed whereby a large bias (~5 V) is usually necessary to induce the heavily electrochemical doping in the active channel of the transistor, the slow ion migration kinetics, or even the irreversible electrochemical doping change the conductance of the transistor and enable the associative learning [bib_ref] Restickable oxide neuromorphic transistors with spike-timing-dependent plasticity and pavlovian associative learning activities, Yu [/bib_ref] [bib_ref] Flexible neuromorphic architectures based on self-supported multiterminal organic transistors, Fu [/bib_ref] [bib_ref] Gelatin-hydrogel based organic synaptic transistor, Lai [/bib_ref]. However, large write bias, incomplete reversibility, and limited retention time are drawbacks. Based on the above-mentioned concerns about using organic synaptic transistors to simulate associative learning, a non-volatile transistor with write bias <1 V, coupled write and read operation, reversible conductance tuning and long retention time is highly desired in an associative learning circuit.
Herein, an organic electrochemical transistor (OECT), a mixed ionic/electronic device often employed for applications in bioelectronics [bib_ref] Highly sensitive metabolite biosensor based on organic electrochemical transistor integrated with microfluidic..., Ji [/bib_ref] [bib_ref] In vivo recordings of brain activity using organic transistors, Khodagholy [/bib_ref] [bib_ref] Self-powered ultra-flexible electronics via nano-gratingpatterned organic photovoltaics, Park [/bib_ref] , has been used as the pivotal component in the associative learning circuit with a previously reported vapor phase polymerized poly : tosylate (PEDOT:Tos)/ Polytetrahydrofuran (PTHF) composite as the active layer [bib_ref] New one-pot poly (3, 4-ethylenedioxythiophene): poly (tetrahydrofuran) memory material for facile fabrication..., Winther-Jensen [/bib_ref] [bib_ref] Synaptic plasticity functions in an organic electrochemical transistor, Gkoupidenis [/bib_ref]. The PEDOT:Tos/PTHF-based OECT with photolithographically patterned channel show much faster response time (~1 ms) compared with previously demonstrated device due to over 2000-fold channel size reduction [bib_ref] Synaptic plasticity functions in an organic electrochemical transistor, Gkoupidenis [/bib_ref]. At the same time, the OECT shows non-volatile characteristics with write bias <0.8 V, reversible and continuous conductance tuning ability as well as long retention time (>200 min) in a coupled write and read testing scheme. Forms of synaptic plasticity like paired-pulse facilitation (PPF), post-tetanic potentiation (PTP), and short-term memory (STM) to long-term memory (LTM) transition have been simulated and found critically related to the mass ratios between PEDOT:Tos and PTHF. Different from other neuromorphic device-based associative learning circuits where two electrical inputs [bib_ref] A non-volatile organic electrochemical device as a lowvoltage artificial synapse for neuromorphic..., Van De Burgt [/bib_ref] [bib_ref] Restickable oxide neuromorphic transistors with spike-timing-dependent plasticity and pavlovian associative learning activities, Yu [/bib_ref] [bib_ref] Mimicking classical conditioning based on a single flexible memristor, Wu [/bib_ref] or one electrical input together with one physical input [bib_ref] Synergistic gating of electro-iono-photoactive 2D chalcogenide neuristors: coexistence of hebbian and homeostatic..., John [/bib_ref] [bib_ref] A multi-input light-stimulated synaptic transistor for complex neuromorphic computing, He [/bib_ref] are usually involved, the synaptic circuit described herein can associate two physical inputs (light and pressure) by integrating a pressure sensor and a photoresistor with a volatile and a non-volatile OECT. Finally, the oxidation level, composition, and microstructure of PEDOT:Tos/PTHF composites under electrical bias were intensively evaluated by spectroscopy and scattering to unravel the origin of non-volatility. The non-volatile OECT with superior memory retention and the synaptic circuit with associative learning functions will have a significant impact in the next-generation neuromorphic devices and bioelectronics such as neuroprosthetics, which could help patients with selfperception and learning challenges.
# Results
Electrical characterization of non-volatile OECT. A synapse is a structure through which a neuron can transmit signals to another neuron. The action potential in the presynaptic neuron triggers the release of neurotransmitters from synaptic vesicles into the synaptic cleft. Neurotransmitters will then bind to receptors at the plasma membrane of the postsynaptic neuron and generate the postsynaptic current. Analogously, OECT gate voltage and channel current can be regarded as the action potential and postsynaptic current, respectively. Similarly, the electrochemical doping and dedoping process in an OECT channel by the voltagecontrolled injection and extraction of ions is quite similar to the neurotransmitter release and uptake at the synaptic cleft [fig_ref] Figure 1: Electrical non-volatility of OECT [/fig_ref]. Before simulating synaptic functions using OECT, the nonvolatility of OECTs is evaluated first. OECTs were fabricated photolithographically (Supplementary [fig_ref] Figure 1: Electrical non-volatility of OECT [/fig_ref] with PEDOT:Tos/ PTHF composites (insert in [fig_ref] Figure 1: Electrical non-volatility of OECT [/fig_ref] , which is deposited by vapor phase polymerization (VPP, [fig_ref] Figure 2: Mimicking STP using an OECT artificial synapse [/fig_ref] , as active channel. Different mass ratios of PTHF were adopted and the resulting films are annotated as P-x% PTHF (x = 0, 20, 50, 80, 90) (in the film, x% is PTHF by mass). The morphology and thickness of P-x% PTHF films are shown in [fig_ref] Figure 3: Mimicking LTP using an OECT artificial synapse [/fig_ref]. The transfer characteristics of the OECTs were measured by cycling V G from −0.8 to 0.8 V while maintaining V DS at −0.2 V. By comparing the transfer curves of the PEDOT:Tos-based OECT with the P-80% PTHF-based OECT in [fig_ref] Figure 1: Electrical non-volatility of OECT [/fig_ref] , a significant hysteresis enhancement with a memory window around 0.39 V could be observed in the P-80% PTHF-based OECT. [fig_ref] Figure 4: OECT-based neuromorphic circuit for demonstrating associative learning [/fig_ref] confirms that OECTs of higher PTHF blend ratios show larger hysteresis, i.e., memory effect. To validate the cycling stability of the hysteresis, a P-80% PTHF-based OECT was consecutively scanned 50 times between −0.8 V and 0.8 V and showed repeatable characteristics [fig_ref] Figure 5: Spectroelectrochemistry of the P-x% PTHF films [/fig_ref]. By carefully controlling the VPP process, the reproducibility of the fabrication is good and the P-80% PTHF-based OECTs show coherent performance with maximum current values of 1:12 ± 0:27 mA and a memory window of 0:34 ± 0:07 V (N = 6, over 3 separate VPP syntheses). To investigate the non-volatile feature of the PEDOT:Tos/PTHF-based OECTs, the charge retention property was further studied by applying 1 s pulsed gate voltage with various amplitudes and polarities. The memory level indicated in [fig_ref] Figure 1: Electrical non-volatility of OECT [/fig_ref] was defined as the magnitude of the difference in the steady-state channel current after being programmed at a specific gate voltage (I DS SteadyðVgÞ ) relative to the steady-state channel current after being programmed at the initial −0.2 V gate voltage (I DS SteadyðÀ0:2Þ ) [fig_ref] Figure 6: XPS and GIWAXS of the P-x% PTHF films [/fig_ref]. As shown in [fig_ref] Figure 1: Electrical non-volatility of OECT [/fig_ref] , the memory level increased with increasing positive programmed gate voltage and increasing PTHF blend ratio, indicating the continuous conductance tunability of the non-volatile OECT with write bias <0.8 V. Charge retention time longer than 200 min was also demonstrated in OECTs with varied biasing condition (0.6, 0.7, and 0.8 V) and channel composition (P-50% PTHF and P-80% PTHF) . The reversibility of the device is critical and was verified in P-80% PTHF-based OECT by programming the device with discrete pulsed voltage or a pulse train as shown in . The conductance of the device can be reliably tuned to discrete levels between high and low [fig_ref] Figure 1: Electrical non-volatility of OECT [/fig_ref] , which implies good reversibility. More importantly, our non-volatile OECT retains its memory effect even when the gate terminal remains connected and the bias is returned to 0 V. As such, this device avoids the use of a transistor or memristor-based access device, which can simplify circuit design in more complicated 1-transistor-based active arrays [bib_ref] Emerging trends in flexible active multielectrode arrays, Lee [/bib_ref].
Mimicking STP using an OECT artificial synapse. Taking advantage of the true non-volatile behavior of PEDOT:Tos/PTHF composite, artificial synapses based on OECTs were constructed to mimic the synaptic function in the brain. STP, which is critical in information processing and various computational tasks in human brain, was emulated by OECT first. Two notable forms of STP are PPF and PTP, which play key roles in decoding temporal information in auditory or visual signals. To simulate PPF and PTP in non-volatile OECTs, a pulsed voltage on the Ag/AgCl gate electrode was applied with different patterns defined as shown in [fig_ref] Figure 2: Mimicking STP using an OECT artificial synapse [/fig_ref]. Ten voltage pulses were applied to the gate terminal with an amplitude, V p , of 0.6 V, a duration (t p ) of 1 ms, and a frequency of 800 Hz. The positive gate voltage de-doped the active channel and decreased the channel current. For simplicity in comparing among different devices, the 4I DS was defined here as the absolute value of the channel current variation between the steady-state channel current before applying the gate voltage (I DS;0 ) and the channel current at a specific time (I DS;t ),
[formula] 4I DS ¼ I DS;0 À I DS;tð1Þ [/formula]
Fig. 2b shows the triggered 4I DS in the P-80% PTHF-based OECT by a pulsed gate voltage. The two performance indicators, PPF and PTP indexes, were defined from [fig_ref] Figure 2: Mimicking STP using an OECT artificial synapse [/fig_ref] as: The amplitudes of the second current pulse in PPF and the tenth current pulse in PTP were compared to that of the first one. This facilitation effect was stronger when increasing the frequency of the presynaptic voltage pulses or when decreasing their interval time, as illustrated in . The PPF and PTP behaviors were highly correlated with the relaxation time of the OECT devices. When the relaxation time was longer than the pulse interval time, the cations in the OECT channel triggered by the first pulse were unable to completely diffuse back to the electrolyte, and these accumulated cations continued to maintain the de-doped state of the channel and tune its current with the addition of cations injected by successive pulses [bib_ref] Neuromorphic functions in PEDOT: PSS organic electrochemical transistors, Gkoupidenis [/bib_ref]. As a result, the next current pulse was facilitated when the presynaptic voltage pulses had a short interval time. Because the charge relaxation time varied markedly with the different PTHF blend ratios in the OECTs' active channels [fig_ref] Figure 2: Mimicking STP using an OECT artificial synapse [/fig_ref] , the PPF and PTP behaviors were strongly dependent on the PTHF ratio. The relationship between the PPF and PTP indexes with the presynaptic pulse voltage interval time, and the OECT channel composition are shown in [fig_ref] Figure 2: Mimicking STP using an OECT artificial synapse [/fig_ref] and e. Both the PPF and PTP showed a typical two-phase exponential behavior and can be fitted by a double-exponential function:
[formula] PPF index ¼ ðA 2 =A 1 100Þ%ð2ÞPTP index ¼ ðA 10 =A 1 100Þ%ð3ÞPPF index; PTP index ¼ 1 þ C 1 exp Àt τ 1 þ C 2 exp Àt τ 2ð4Þ [/formula]
where t is the pulse interval time, C 1 and C 2 were the initial facilitation magnitudes of each phase, and τ 1 and τ 2 were the respective relaxation time constant. Interestingly this two-phase exponential decay is coherent with the decay of synaptic facilitation in a biological synapse; τ 1 and τ 2 can be used to mimic the relaxation time of rapid and slower decay phase in a synapse [bib_ref] Short-term synaptic plasticity, Zucker [/bib_ref]. However, the fundamental physical meaning of these two time constants in this particular system requires further investigation. From fitting parameters listed in , one can notice that the facilitation magnitude (C 1 and C 2 ) and relaxation time (τ 1 and τ 2 ) generally follow the trend that the elevated value can be obtained with the increment of the PTHF blend ratio. However, one exception is the relaxation time τ 2 in P-90 PTHF sample. This phenomenon was attributed to the slow response time of P-90% PTHF which will be discussed further below.
Mimicking LTP using an OECT artificial synapse. In contrast to STP, in which temporary modifications of synaptic weight occur, LTP plays its role in memory or learning by changing the synaptic weight more persistently. LTP can last longer from minutes to days or even years. The process of memory formation in the brain (from sensory memory (SM) to STM and finally to LTM) can be aided by repeated rehearsal or training and is illustrated in [fig_ref] Figure 3: Mimicking LTP using an OECT artificial synapse [/fig_ref].
The transition process from STM to LTM was simulated in our non-volatile OECTs with the repeated application of presynaptic gate voltage pulses, which was analogous to the rehearsal process, and the LTM level in OECTs was defined as shown in [fig_ref] Figure 3: Mimicking LTP using an OECT artificial synapse [/fig_ref]. The dedoping level of the channel material during 20 pulsing cycles is believed to be increased after plotting I DS on a log-scale and considering the magnitude of the leakage current (I G ), which begins to dominate at the applied potentials of interest. This effect will influence the precise evaluation of electronic current inside the channel (indicative of the dedoping level) [fig_ref] Figure 1: Electrical non-volatility of OECT [/fig_ref]. A series of pulsed gate voltages with fixed pulse amplitude (0.7 V), duration (10 ms), and interval time (2.5 ms) were applied to the P-80% PTHF-based OECT. The changes in channel current (4I DS ) of the OECT device after four representative training cycles are shown in [fig_ref] Figure 3: Mimicking LTP using an OECT artificial synapse [/fig_ref]. As observed in [fig_ref] Figure 3: Mimicking LTP using an OECT artificial synapse [/fig_ref] , the memory level, defined as the steady-state channel current change before and after application of gate voltage, increased with each additional training cycle of pulsed gate voltage, which indicated a transition from STM to LTM. This phenomenon is very similar to the process that occurs in a neural synapse, in which a large number of neurotransmitters can be released when the presynaptic neuron is triggered by several action potentials, enhancing the transmission efficiency between neurons. Besides the effect of training cycles on the LTM, we also study the influence of channel composition on the LTM. A comparison among the PEDOT:Tos, the P-50% PTHF and the P-80% PTHF-based OECT after training with 50 consecutive pulses [fig_ref] Figure 3: Mimicking LTP using an OECT artificial synapse [/fig_ref] shows a clear STM to LTM transition for the P-50% and P-80% PTHF devices, while in the device without PTHF, almost all STM was lost and was not transferred to LTM after removing the pulsed gate voltage. The slightly observable memory effect in PEDOT: Tos-based OECTs may be due to a small population of residual ions or to a bias-stress effect during testing. The relation between memory level and channel composition at 50 training cycles is shown in [fig_ref] Figure 3: Mimicking LTP using an OECT artificial synapse [/fig_ref]. As opposed to the monotonic increase of memory level vs composition we observed in [fig_ref] Figure 1: Electrical non-volatility of OECT [/fig_ref] , memory level here increased non-monotonically with larger PTHF fractions. When the blend ratio of PTHF increased from 0 to 80%, the memory level increased with the PTHF blend ratio. Beyond P-80%, the memory level decreased. Based on this observation and PPF and PTP results obtained for the P-90% devices (the relaxation time τ 2 in the P-90% PTHF-based OECT is smaller than the one in the P-80% PTHF-based OECT), we deduce that an excessively high PTHF blend ratio (90%) results in a thicker film (~1.6 μm, [fig_ref] Figure 3: Mimicking LTP using an OECT artificial synapse [/fig_ref] and thus slower response time (~7 ms, [fig_ref] Figure 2: Mimicking STP using an OECT artificial synapse [/fig_ref]. Given the short pulse duration (10 ms) in STM to LTM measurement, this gives the illusion of lower memory levels, which can be recovered with longer input pulse duration. Thus, balancing the observed memory effect with the response time at this transistor size scale, P-80% PTHF is selected as the optimal material for non-volatile OECTs in the subsequent synaptic circuits. OECT-based neuromorphic circuit for demonstrating associative learning. Based on the demonstrated synaptic behavior of PEDOT:Tos/PTHF-based OECT, we advanced the single synaptic transistor into a synaptic circuit to simulate associative learning. A pressure sensor, a photoresistor, a PEDOT:PSS-based volatile OECT, and a P-80% PTHF-based non-volatile OECT were integrated into a synaptic circuit [fig_ref] Figure 4: OECT-based neuromorphic circuit for demonstrating associative learning [/fig_ref]. Pressure and light were sensed by the corresponding sensors and were transferred to the channel conductance change of the non-volatile OECT as shown in [fig_ref] Figure 1: Electrical non-volatility of OECT [/fig_ref] , which allowed the simulation of associative learning between two physical inputs. Light from a LED bulb and pressure from a finger press were considered as CS and US, respectively [fig_ref] Figure 4: OECT-based neuromorphic circuit for demonstrating associative learning [/fig_ref] , while the current change in the drain-source circuit (4I DS ) was considered as UR. The LED bulb, the photoresistor, and the pressure sensor were assembled into an associative learning circuit dock as shown in [fig_ref] Figure 1: Electrical non-volatility of OECT [/fig_ref]. The complete associative learning process is shown in [fig_ref] Figure 4: OECT-based neuromorphic circuit for demonstrating associative learning [/fig_ref]. First, pulsed light (CS, [fig_ref] Figure 1: Electrical non-volatility of OECT [/fig_ref] was applied on the circuit; in this case, only a small part of the gate voltage could be applied on both the volatile and non-volatile OECT as the 13 MΩ resistor was connected in series to their gate inputs [fig_ref] Figure 4: OECT-based neuromorphic circuit for demonstrating associative learning [/fig_ref]. As a result, only a slight increase of 4I DS was achieved without reaching memory threshold. Then, the pressure sensor was pressed by the finger as the US signal [fig_ref] Figure 1: Electrical non-volatility of OECT [/fig_ref]. Due to the low resistance of the pressure sensor while pressing with a finger and the high resistance of the photoresistor in a dark environment, most of the gate voltage was applied on the volatile device and very little was applied on the non-volatile one. Therefore, although a high 4I DS was attained to trigger the UR (salivating dog), it later decayed leaving no memory effect because of the untriggered non-volatile OECT. After solely applying the US on our circuit, the CS was implemented again and the amplitude of 4I DS was comparable with the first time it was triggered by the CS, which was too small to induce UR. This result confirmed that training only with the US did not help our circuit to associate the CS with the US. Next, the CS and US were applied together as the training process [fig_ref] Figure 1: Electrical non-volatility of OECT [/fig_ref] ; in this case, the gate voltage could be ideally applied on both the volatile and non-volatile OECT. Consequently, a high enough 4I DS to trigger UR was achieved and it was retained due to the memory property of the non-volatile OECT [fig_ref] Figure 1: Electrical non-volatility of OECT [/fig_ref]. This memory effect (4I DS1 ) helped the 4I DS approach the memory threshold when triggered by the CS only after the first training. By performing 5 training cycles under the simultaneous application of CS and US, the obtained memory level increased significantly (4I DS5 , [fig_ref] Figure 1: Electrical non-volatility of OECT [/fig_ref]. Finally, the CS alone was able to trigger UR [fig_ref] Figure 1: Electrical non-volatility of OECT [/fig_ref] without the existence of the US, which confirmed the neuromorphic circuit successfully associated the CS with the US and enabled the learning process in the neuromorphic circuit. The demonstration of associative learning is the first important step, our proposed circuit can be further extended to more sensory inputs (available from a suit of possible sensors) on the nonvolatile OECTs and can be potentially integrated with electronics skin, smart robotics, or even in implantable devices to directly interface with neurons and synapses, which has great impact in neuroprosthetics and brain-machine interface.
Spectroelectrochemistry of the P-x% PTHF films. Beyond demonstrating the synaptic functions of the non-volatile OECT, it is key to understand the origin of the non-volatility. Although previous research proposed the structural collapse of PTHF result in the non-volatility of vapor phase polymerized PEDOT with PTHF 37,38 , further characterizations are needed to investigate how the film properties change optically, compositionally, and structurally under various gate bias. Spectroelectrochemistry of the P-x% PTHF films under different biases was investigated by operando ultraviolet-visible-near-infrared (UV-Vis-NIR) spectroscopy as shown in [fig_ref] Figure 5: Spectroelectrochemistry of the P-x% PTHF films [/fig_ref]. P-x% PTHF films were deposited on ITO glass and functioned as the working electrode in an electrochemical cell. A stepwise voltage with 10 s duration was applied between the P-x% PTHF films and an Ag/AgCl pellet (reference/counter electrode) through 0.1 M NaCl electrolyte [fig_ref] Figure 5: Spectroelectrochemistry of the P-x% PTHF films [/fig_ref] while the UV-Vis-NIR absorbance was continuously monitored. The UV-Vis-NIR absorbance during electrical bias of PEDOT:Tos and P-80% PTHF are shown in [fig_ref] Figure 1: Electrical non-volatility of OECT [/fig_ref]. As expected, two samples with a different blend ratio of PTHF showed decreased π-π* transition absorbance (~620 nm) and increased polaronic state absorbance, including polaron (~900 nm) and bipolaron (~1350 nm), which is consistent with electrical properties of two samples being converted from an insulating state (low oxidation level) to a highly conductive state (high oxidation level) when the bias increase from −0.8 V (dedoping) to 0.8 V (doping). However, when comparing the UV-Vis-NIR absorbance after the bias has been returned to 0 V, the samples [fig_ref] Figure 5: Spectroelectrochemistry of the P-x% PTHF films [/fig_ref] displayed different behavior. Pure PEDOT:Tos film did not retain the spectral signature of the various oxidation levels, while for P-80% PTHF, the distinct residual spectral signatures of the oxidation levels were maintained without bias. When comparing the state with the lowest oxidation level (−0.8 V, 10 s) to the state with the highest oxidation level (0.8 V, 10 s), the modulation of π-π* transition, polaron, and bipolaron absorbance of PEDOT:Tos were all <5%. Conversely, the π-π* transition absorbance of P-80% PTHF increased by 75.9%, the polaron absorbance of P-80% PTHF decrease by 14.1% and bipolaron absorbance of P-80% PTHF decrease by 25.7%, respectively, as shown in [fig_ref] Figure 1: Electrical non-volatility of OECT [/fig_ref]. This result indicates that the film with 80% PTHF shows strong non-volatility, in agreement with the electrical behavior of the devices above.
To confirm the reversibility of the oxidation level change caused by the electrical bias in P-80% PTHF, a cyclic doping (0.8 V) and dedoping (−0.8 V) bias was applied and the absorbance of P-80% PTHF was recorded after bias. As shown in Supplementary [fig_ref] Figure 1: Electrical non-volatility of OECT [/fig_ref] , the highly repeatable absorbance of P-80% PTHF in both low oxidation level (−0.8 V) and high oxidation level (0.8) confirmed the high reversibility during electrical bias of different polarity. Besides, long-term charge retention of P-80% PTHF was also examined optically as a supplement of the above-mentioned charge retention results that has been electrically demonstrated. Through tracking the intensity of π-π* transition with respect to time, a quick erase of memory state right after the removal of bias can be observed [fig_ref] Figure 1: Electrical non-volatility of OECT [/fig_ref]. However, after charge stabilization, distinct difference in low oxidation level (−0.8 V) and high oxidation level (0.8 V) can still be observed after the bias was removed for 1 h, which confirms the charge retention time can be longer than 1 h when the film is still immersed in the electrolyte with gate terminal connected [fig_ref] Figure 1: Electrical non-volatility of OECT [/fig_ref]. The state retention time of the P-80% PTHF film was as long as 8 days after being removed from electrolyte and stored in ambient [fig_ref] Figure 1: Electrical non-volatility of OECT [/fig_ref].
## Article
The persistence of a stable low oxidation level in P-80% PTHF films implies the inability of positive electronic charge carriers to return to the PEDOT chains. This could arise electronically, that despite the thermodynamic preference of the PEDOT to return to the doped state, electronic charge transport may be so frustrated that positive charge carriers could not reach the PEDOT chains on experimentally relevant timescales. Alternatively, the nonvolatility could arise ionically due to the cations being trapped near the PEDOT chains in the film and anions being impeded from reaching those PEDOT chains, thus preventing the dopant stabilization of electronic charge on the PEDOT chains. Either of these processes, or a combination of both, would maintain the dedoped state, resulting in an enhanced π-π* transition absorbance, diminished polaron and bipolaron absorbance, and diminished film conductivity. Frustrated electronic charge transport alone could not account for the non-volatile nature of the low oxidation state, as the conductivity, though low, is non-zero as shown in [fig_ref] Figure 6: XPS and GIWAXS of the P-x% PTHF films [/fig_ref]. Considering the capacity of the PEDOT:
Tos/PTHF film and film conductivity in low oxidation state, if electronic charge transport was the limiting step, the film would re-dope on much shorter timescales. A far more likely source of non-volatility is frustrated ionic transport, which results in a large number of trapped cations inside the film driven by electrical bias.
XPS of the P-x% PTHF films. To track the ions inside the P-x% PTHF under different conductance states, X-ray photoelectron spectroscopy (XPS) was used to identify the relative quantity of Tos − , Cl − , and Na + in the films. The low resistance state (LRS) and high resistance state (HRS) of the films were achieved with 1 min bias of −0.8 and 0.8 V, respectively. As shown in [fig_ref] Figure 6: XPS and GIWAXS of the P-x% PTHF films [/fig_ref] , pristine P-80% PTHF shows distinct S 2p peak from Tos − (166~170 eV) while no Cl 2p and Na 1s peaks, which is consistent with the composition of pristine P-80% PTHF that the Tos − is the only counterion to dope the PEDOT chain. However, for P-80% PTHF in both LRS and HRS, the S 2p peaks from Tos − are diminished while the Cl 2p peaks are observed, which is consistent with ion exchange of the more bulky Tos − with the smaller Cl − during electrochemical operation [bib_ref] Acido-basic control of the thermoelectric properties of poly (3, 4-ethylenedioxythiophene) tosylate (PEDOT-Tos)..., Khan [/bib_ref]. The same phenomenon is also shown in PEDOT:Tos [fig_ref] Figure 6: XPS and GIWAXS of the P-x% PTHF films [/fig_ref]. More importantly, a very strong Na 1s peak is observed in P-80% PTHF in HRS compared with the minor Na 1s peak in P-80% PTHF in LRS. From the relatively peak area of Na 1s and Cl 2p in P-80% PTHF in both LRS and HRS, we can conclude that a large amount of Na + is trapped in P-80% PTHF and a part of Cl − is removed during biasing process. At the same time, no Na 1s peak is observed in PEDOT:Tos in HRS [fig_ref] Figure 6: XPS and GIWAXS of the P-x% PTHF films [/fig_ref] , which confirms that PTHF is indeed the component responsible for Na + trapping. The relative ratio of Na + over Cl − in the bulk area was also calculated from XPS depth profiling [fig_ref] Figure 2: Mimicking STP using an OECT artificial synapse [/fig_ref] ; higher Na + /Cl − ratio (~0.3) in HRS compared with (<0.1) in LRS confirms that the trapping of Na + occurs in the bulk of the P-80% PTHF film, and not only near the polymer/electrolyte interface.
GIWAXS of the P-x% PTHF films. GIWAXS was further used to understand the mechanism of non-volatility of P-x% PTHF induced by the trapped Na + from structural point of view. Before conducting GIWAXS on films in different conductance states, understanding the microstructure of the at-cast P-x% PTHF is important [fig_ref] Figure 1: Electrical non-volatility of OECT [/fig_ref]. In general, pristine PEDOT: Tos shows a preferred edge-on structure with a lamellar stacking distance 13.7 Å and π-π stacking distance 3.5 Å. The relatively large breadth of the PEDOT:Tos scattering peaks indicate relatively short coherence lengths (<10 nm) of polymer crystallites, which are likely embedded in an amorphous matrix (Supplementary [fig_ref] Figure 2: Mimicking STP using an OECT artificial synapse [/fig_ref]. Introducing 50% PTHF revealed no new scattering peaks due to PTHF crystallites. Further, PEDOT:Tos scattering peaks showed no appreciable broadening, indicating that the length scales of coherent PEDOT:Tos ordering were not diminished due to the presence of PTHF. The presence of PTHF did bring about shifts in the position of the PEDOT:Tos scattering indicating changes in the PEDOT:Tos d-spacing. The outof-plane (100) peak shifted to lower q z , as shown in Supplementary [fig_ref] Figure 2: Mimicking STP using an OECT artificial synapse [/fig_ref] , indicating an increased lamellar stacking distance around 14.4 Å for P-50% PTHF, while the in-plane π-stack peak position remained unchanged. The lamellar expansion may result from PTHF chains being closely integrated with PEDOT chains and partially occupying the inter-lamella space (Supplementary [fig_ref] Figure 2: Mimicking STP using an OECT artificial synapse [/fig_ref]. However, this lamellar expansion does not account for the total PTHF content added to the composite, suggesting most of the PTHF chains must reside outside the PEDOT crystallites in the amorphous regions along with the amorphous fraction of PEDOT:Tos. When PTHF content was further increased to 80%, PEDOT crystallite scattering peaks could no longer be observed, indicating extensive disruption of PEDOT ordering, or diminished relative scattered intensity that is not resolvable due to isotropic PTHF scattering ((110)*, (020) *) [bib_ref] CH 2 ) m -O-] n . III 1 . Molecular and..., Imada [/bib_ref] [bib_ref] Physical properties of poly (tetrahydrofuran)-block-poly (2-ethyl-2-oxazoline) triblock copolymer, Motokucho [/bib_ref]. With an 80% PTHF loading, PTHF diffraction shows no shifts in peak position compared to pure PTHF. This result indicates that PEDOT does not disrupt semicrystalline PTHF packing (in P-80% PTHF) and is confined to the amorphous regions of the film [fig_ref] Figure 2: Mimicking STP using an OECT artificial synapse [/fig_ref].
Based on the results from as-cast P-x% PTHF, we then conduct GIWAXS on P-x% PTHF films in LRS and HRS which are achieved by 1 s bias of −0.8 and 0.8 V, respectively. Because PEDOT:Tos and P-50% PTHF show predominant PEDOT scattering while P-80% PTHF shows intense PTHF scattering, these three compositions are informative for studying the iontrapping effect both in ordered PEDOT region and PTHF region. The 2D-GIWAXS patterns of PEDOT:Tos, P-50% PTHF, and P-80% PTHF in LRS and HRS are shown in [fig_ref] Figure 2: Mimicking STP using an OECT artificial synapse [/fig_ref]. Consistent with the lack of non-volatile state stability, PEDOT:Tos displayed no persisting differences in lattice dspacing [fig_ref] Figure 6: XPS and GIWAXS of the P-x% PTHF films [/fig_ref] , indicating that the rapid electronic relaxation includes a concomitant structural relaxation as illustrated in [fig_ref] Figure 2: Mimicking STP using an OECT artificial synapse [/fig_ref]. However, the P-50% PTHF showed a marked shift in (100) and (200) peak positions [fig_ref] Figure 6: XPS and GIWAXS of the P-x% PTHF films [/fig_ref] , consistent with an increase of 0.2 Å in the lamellar spacing in the HRS over the LRS as schematically shown in [fig_ref] Figure 2: Mimicking STP using an OECT artificial synapse [/fig_ref] , which is proposed to be induced by trapped Na + . The πstack associated d-spacing was unchanged for both PEDOT:Tos and P-50% PTHF [fig_ref] Figure 2: Mimicking STP using an OECT artificial synapse [/fig_ref]. Consistent with HRS P-50% PTHF, operando PEDOT:PSS measurements have also shown a lamellar expansion upon dedoping 47 , which indicates electrical stability of the HRS in non-volatile P-50% PTHF should be coupled with a long-lived structural modification. In addition, (020)* and (110)* peaks of PTHF in P-80% PTHF are unchanged in both LRS and HRS [fig_ref] Figure 6: XPS and GIWAXS of the P-x% PTHF films [/fig_ref] which indicates that the PTHF lattice parameters are insensitive to different conductance states. Since switching states necessitates significant ion transport, the lack of change to the PTHF crystallite structure strongly suggests that cations transport and reside within the disordered or aggregated PEDOT:Tos/PTHF fraction. At the same time, the P-80% PTHF has already been demonstrated with even stronger memory effect than P-50% PTHF, implying the crystalline PTHF imposes a further barrier to ionic transport compounding the trapping of Na + in the non-crystalline fraction of PTHF [fig_ref] Figure 6: XPS and GIWAXS of the P-x% PTHF films [/fig_ref]. The relevance of the P-50% PTHF samples is thus to highlight the importance of the PTHF chains that are closely interfaced with PEDOT, which presumably exist within P-80% PTHF in a manner not readily resolvable with GIWAXS.
# Discussion
The hindered ionic transport that effectively (reversibly) traps Na + in PTHF may be due to the following reasons: In general, the relatively low oxygen density in the backbone of PTHF (O:C=1:4) reduces the number and proximity of coordination sites for Na + compared with the analog polymer electrolyte polyethylene glycol (PEG) (O:C=1:2). The diminished cation-polymer coordination results in poorer ion solubility and slower ionic transport in the solvent-free condition [bib_ref] Enhanced Li + conduction within single-ion conducting polymer gel electrolytes via reduced..., Ford [/bib_ref]. This is usually overcome in PTHFbased polymer electrolytes with additional solvent like ethylene carbonate (EC)/diethyl carbonate (DEC) that can swell PTHF [bib_ref] Enhanced Li + conduction within single-ion conducting polymer gel electrolytes via reduced..., Ford [/bib_ref]. The loosely coordinated cation and polymer chain in PTHF will facilitate cation-solvent coordination which enables a faster ionic transport due to the vehicular motion of solvated cation by solvent. However, this phenomenon is not applicable in our device. PTHF is relatively hydrophobic 51 , thus the aqueous electrolyte employed to gate PEDOT:Tos/PTHF OECTs likely imparts minimal swelling or plasticization of the PTHF phase which could facilitate facile ionic transport. Due to the diminished ionsolvent coordination, ion hopping in the backbone of PTHF will dominate. Furthermore, the ion hopping efficiency is reduced due to the low density of ion coordination sites and large distance between them in the PTHF backbone. In addition, the linear PTHF we used also favors crystallization which is determinantal to ionic transport because of the hindered segmental motion of polymer chains [bib_ref] Poly (ethylene oxide)-based electrolytes for lithiumion batteries, Xue [/bib_ref]. The interpretation here and the abovementioned characterization provide evidence that the frustrated ionic transport in the non-crystalline PTHF fraction, compounded with increased PTHF crystallinity, cause the trapping of Na + ions, which leads to the observed non-volatility. This mechanism thus provides a materials route to explore new composites for longer state retention, faster response, or selective memory owing to certain ionic species.
In summary, a non-volatile OECT device was successfully fabricated with a VPP deposited PEDOT:Tos/PTHF composite active channel. This device shows fast response time (~1 ms), stable memory window, write bias <0.8 V, reversible and continuous conductance tuning ability as well as long retention time (>200 min) in coupled write and read testing scheme. By varying the pattern of the pulsed gate voltage used as presynaptic input, different synaptic functions from STP, like PPF and PTP, to LTP, like STM to LTM transition, have been emulated. The PTHF ratio in the composite is shown to play a critical role in defining the STP and LTP properties of the device. For STP, a higher blend ratio of PTHF is preferred as reflected by the increased PPF and PTP indexes with elevated PTHF ratios. While for LTP, the memory behavior is not monotonically enlarged with the increased PTHF ratio; the device with 80% PTHF showed the best-attained memory compared with other devices after equivalent training. Based on the non-volatile nature of the OECT, a neuromorphic circuit has been successfully demonstrated for simulating associative learning from two physical inputs by combining a pressure sensor, a photoresistor, a volatile OECT, and a non-volatile OECT. Besides the application of this device, in-depth study of device physic was conducted. A multimodal ex-situ and operando characterization suite is used to investigate the oxidation level variation, compositional change, and the structural modulation of P-x% PTHF films in various biasing conditions. Non-crystalline PTHF is believed to be capable of trapping cations which is the origin of the memory behavior of OECT devices, with PTHF crystallites enhancing the effect. The understanding of the non-volatile material and the application of the associative learning circuit are significant for both fundamental material selection as well as hardware implementation in the area of neuromorphic computing and bioelectronics.
# Methods
Fabrication of non-volatile OECT device. The device was fabricated by a standard microfabrication technique. 50 nm Au and 5 nm Cr were deposited on a photoresist patterned glass slide by thermal evaporation and the following immersion in DMSO was used to remove the photoresist for defining drain-source pattern of OECT. Two consecutive parylene layers with the same thickness of 1.5 μm were deposited to encapsulate the metal contact as well as pattern the active layer. The substrate was first treated with 0.5% saline 174 solution to improve the adhesion with the first parylene layer. Before the deposition of the second parylene layer, 2% micro 90 solution was spin-coated as an anti-adhesive layer which enables the delamination of the second parylene layer. Positive photoresist AZ 9260 was patterned on parylene and used as a template to pattern the underneath double-layer parylene by oxygen plasma etching. Before active layer deposition, the photoresist AZ 9260 was stripped by the AZ 400T solution. The device with metal electrodes and patterned parylene layers was transferred to N 2 filled glove box to avoid moisture-inducing crystallization of oxidant for later polymerization. A mixture solution of oxidant Fe(III):Tos and PTHF in butanol with predefined ratio was spin-coated on device surface followed by 1-min 70°C thermal annealing. The device was immediately transferred to an EDOT filled tube furnace and the vapor phase polymerization (VPP) of EDOT took place at 70°C for 1 h with continuous purged N 2 (100 SCCM). Finally, after the VPP process, the sacrificial parylene layer was peeled off to define the active channel of PEDOT:Tos/PTHF composite. The device was rinsed in DI water for 1 h to remove excess oxidant before doing the electrical measurement.
Preparation of PEDOT:Tos/PTHF composite precursor. P-80% PTHF is used as an example here. The amount of PTHF was determined by the fact that it needs 2.25 moles of Fe (III) to produce 1 mole PEDOT 53 . 0.37 g PTHF and 1 g Fe(III)Tos were added into 3.63 g butanol to maintain a 20 wt% of Fe(III)Tos in the mixed solution. The mixture was magnetically stirred and heated at 60°C for 30 min to thoroughly dissolve the PTHF and Fe(III)Tos. Afterward, 60 μL pyridine was added into the mixture as a base inhibitor in the VPP process [bib_ref] Base inhibited oxidative polymerization of 3, 4-ethylenedioxythiophene with iron (III) tosylate, Winther-Jensen [/bib_ref] Fabrication of pressure sensor. The detailed fabrication process can be found in our previous publication [bib_ref] High sensitivity, wearable, piezoresistive pressure sensors based on irregular microhump structures and..., Wang [/bib_ref]. In short, the sandpaper was treated with oxygen plasma first and then coated with a self-assembled monolayer (SAM) of Trichloro (1H, 1H, 2H, 2H-perfluorooctyl) saline by vapor phase deposition. This additional hydrophobic SAM can help the demolding of Polydimethylsiloxane (PDMS) later. The mixture of SYLGARD 184 and curing agent (10:1) was spin-coated on sandpaper at 700 rpm and followed by 30-min thermal annealing. After that, the PDMS thin film with micro-hump structure was peeled off and treated with oxygen plasma for 10 min. PEDOT:PSS (CLEVIOS PH 1000) was spin-coated on a PDMS surface at 700 rpm and then dried at 30°C for 30 min. The PDMS coated with PEDOT:PSS was placed on screen-printed interdigitated silver electrodes to form a pressure sensor. To make sure the pressure is evenly distributed on the pressure sensor, a square of Si wafer was mounted on the back of the PDMS.
Operando UV-Vis-NIR spectroscopy. Spectroelectrochemistry of P-x% PTHF coated on ITO glass electrode was carried out in 0.1 M aqueous NaCl in a PMMA cuvette with an Ag/AgCl pellet (Warner Instruments) reference/counter electrode. Potential control was carried out with a potentiostat (Ivium). Simultaneous absorption spectroscopy was carried out with a halogen white light source (Ocean Optics, DH-2000-BAL) and an optical fiber light path split to separate UV-visible (Ocean Optics, FLAME-S) and near-infrared (Ocean Optics, NQ512) spectrometers. Spectroscopic data were recorded with OceanView software. All analysis was executed with MATLAB software.
X-ray photoelectron spectroscopy (XPS). The XPS spectrums of P-x% PTHF were taken using Thermo Scientific ESCALAB 250Xi equipped with a monochromatic KR Al X-ray source (spot size around 500 μm) at the Northwestern University Atomic and Nanoscale Characterization Experimental center (NUANCE). A flood gun was used for charge compensation. The analysis of the spectrum was performed using the Avantage (Thermo Scientific) software.
Grazing-incidence wide-angle X-ray scattering (GIWAXS). 2D GIWAXS was carried out at the Advanced Photon Source at Argonne National Laboratory on beamline 8-ID-E at room temperature under vacuum with 10.92 keV (λ = 1.135 Å) synchrotron radiation with a 0.14°incident angle and measured with a Pilatus 1 M hybrid pixel array detector during 10 s exposures. Data analysis was carried out with GIXSGUI Matlab toolbox 56 .
Electrical characterization. All the devices are gated by Ag/AgCl electrode through 0.1 M NaCl electrolyte. Basic characterizations like transfer curves and output curves are measured by Keithley 2636B source meter with custom-made LabVIEW programs. For synaptic function characterizations, pulsed gate voltage was generated by Agilent 33210 arbitrary function generator. A Keithley 7001 digital switch was used to change the gate bias between source meter and a function generator. For transient measurement, a DMM 6500 digital multimeter was connected to the source terminal to measure the drain-source current with a sampling rate of 50 kHz.
## Data availability
The data that support the findings of this study are available from the corresponding author upon reasonable request.
Received: 28 August 2020; Accepted: 16 March 2021;
[fig] Figure 1: Electrical non-volatility of OECT. a Analogy of synapse and OECT and the chemical structure of PEDOT + , Tos − , and PTHF. b Transfer characteristics of the PEDOT:Tos-based OECT and the P-80% PTHF-based OECT (W ¼ 500 μm; L ¼ 10 μm). Optical micrograph of channel is shown inSupplementary Fig. 4a. c Memory level with respect to channel composition of OECT and programmed gate voltage. d Reversible conductance change of P-80% PTHF-based OECT in response to a pulse train with different polarity (conductance value calculated fromSupplementary Fig. 8b). [/fig]
[fig] Figure 2: Mimicking STP using an OECT artificial synapse. a Patterns of pulsed gate voltage with different parameters; V p : voltage amplitude, 4t: pulsed voltage interval time, t p : pulsed voltage duration, T p : pulsed voltage period. b Channel current modulation induced by 10 consecutive gate voltage pulses in the P-80% PTHF-based OECT. c Transient response of the PEDOT:Tos/PTHF-based OECT with different PTHF blend ratio. All the dashed lines are exponential decay fitting curves. The rise-time of the current pulse when the gate voltage was applied is defined as response time, while the decay-time of the current pulse after the gate voltage was removed is defined as ion relaxation time.The response time of devices with different PTHF blend ratio are τ res 0 ¼ 0:143 ms; τ res 20 ¼ 0:269 ms; τ res 50 ¼ 0:387ms; τ res 80 ¼ 2:104 ms; τ res 90 ¼ 7:113 ms. The ion relaxation time of devices with different PTHF blend ratio are τ rel 0 ¼ 0:288 ms; τ rel 20 ¼ 0:241 ms; τ rel 50 ¼ 0:692ms; τ rel 80 ¼ 2:824 ms; τ rel 90 ¼ 5:659 ms. d, e Paired-pulse facilitation (PPF) and post-tetanic potentiation (PTP) indexes as a function of OECT channel composition and interval time of gate pulses, under t p of 1 ms. Gray dashed lines are the fitting curves. [/fig]
[fig] Figure 3: Mimicking LTP using an OECT artificial synapse. a Scheme of the memory formation process in the human brain. b Short-term memory (STM) to long-term memory (LTM) transition in the P-80% PTHF-based OECT triggered by20 consecutive gate voltage pulses. c Channel current change comparison of the P-80% PTHF-based OECT after biasing by four different cycles of gate voltage pulses. d Memory level as a function of pulsed gate voltage training cycles. e Channel current change comparison of OECT among three different PTHF ratio after biasing by the 50 pulsed gate voltage. f Memory level with respect to channel composition of OECT at 50 pulsed gate voltage. | (2021) 12:2480 | https://doi.org/10.1038/s41467-021-22680-5 | www.nature.com/naturecommunications [/fig]
[fig] Figure 4: OECT-based neuromorphic circuit for demonstrating associative learning. a Circuit diagram of the neuromorphic circuit for simulating associative learning. V G = 0.7 V and V DS = −0.2 V. b Analogy of the CS (conditioned stimulus) and US (unconditioned stimulus) in Pavlov's dog experiment and our neuromorphic circuit. c Complete associative learning process. After five times training, the CS can trigger UR (unconditioned response, salivation of dog). [/fig]
[fig] Figure 5: Spectroelectrochemistry of the P-x% PTHF films. a Set up of operando spectroelectrochemistry study of PEDOT:Tos/PTHF film with different compositions. b The pattern of the stepwise bias that was applied on the working electrode (P-x% PTHF). c, d UV-Vis-NIR absorbance of PEDOT:Tos and P-80% PTHF after the electrical bias has been removed. [/fig]
[fig] Figure 6: XPS and GIWAXS of the P-x% PTHF films. a XPS of the S 2p, Cl 2p, and Na 1s peaks of P-80% PTHF under different conductance states. b XPS of the S 2p, Cl 2p, and Na 1s peaks of PEDOT:Tos under different conductance states. In XPS data, gray lines are the original data before fitting. Blue (For P-80% PTHF) and red (for PEDOT:Tos) lines are overall fitted data. Dashed green and yellow lines are fitted doublet peaks of S 2p from PEDOT + and Tos − , respectively. Dashed purple lines are fitted doublet peaks of Cl 2p. c Out-of-plane linecuts for PEDOT:Tos and P-50% PTHF in LRS and HRS; and in-plane linecuts for P-80% PTHF in LRS and HRS. d Cartoon of the microstructure and composition of PEDOT:Tos and P-80% PTHF in LRS and HRS illustrating the mechanism of non-volatility. PEDOT:Tos has similar microstructure and composition in LRS and HRS, which results in comparable electrical current. For P-80% PTHF, trapped Na + in non-crystalline PTHF fraction leads to a diminished hole current in HRS compared with LRS. [/fig]
|
Augmenting Microarray Data with Literature-Based Knowledge to Enhance Gene Regulatory Network Inference
Gene regulatory networks are a crucial aspect of systems biology in describing molecular mechanisms of the cell. Various computational models rely on random gene selection to infer such networks from microarray data. While incorporation of prior knowledge into data analysis has been deemed important, in practice, it has generally been limited to referencing genes in probe sets and using curated knowledge bases. We investigate the impact of augmenting microarray data with semantic relations automatically extracted from the literature, with the view that relations encoding gene/protein interactions eliminate the need for random selection of components in non-exhaustive approaches, producing a more accurate model of cellular behavior. A genetic algorithm is then used to optimize the strength of interactions using microarray data and an artificial neural network fitness function. The result is a directed and weighted network providing the individual contribution of each gene to its target. For testing, we used invasive ductile carcinoma of the breast to query the literature and a microarray set containing gene expression changes in these cells over several time points. Our model demonstrates significantly better fitness than the state-of-the-art model, which relies on an initial random selection of genes. Comparison to the component pathways of the KEGG Pathways in Cancer map reveals that the resulting networks contain both known and novel relationships. The p53 pathway results were manually validated in the literature. 60% of non-KEGG relationships were supported (74% for highly weighted interactions). The method was then applied to yeast data and our model again outperformed the comparison model. Our results demonstrate the advantage of combining gene interactions extracted from the literature in the form of semantic relations with microarray analysis in generating contribution-weighted gene regulatory networks. This methodology can make a significant contribution to understanding the complex interactions involved in cellular behavior and molecular physiology.
# Introduction
Gene regulatory networks (GRNs) are DNA-encoded regulatory subsystems in the genome that coordinate input from activator and repressor transcription factors to control various biological functions, including development, cell differentiation, and response to environmental cues. They provide a systems level illustration of physiological function and are composed of modules at varying hierarchical levels [bib_ref] Emerging properties of animal gene regulatory networks, Davidson [/bib_ref]. A GRN provides the pathways of gene interactions within the context of location and time [bib_ref] Gene regulatory networks for development, Levine [/bib_ref]. For instance, when a receptor on a particular cell receives a signal and initiates the activation of a transcription factor, the transcription factor increases the expression of the target gene, which in turn alters the production and activation of other pathway components in a cascading manner, changing the behavior of the cell.
The development of microarray technology allowed the discovery of gene regulation to move from individual interactions to thousands of interactions in parallel [bib_ref] Microarrays: biotechnology's discovery platform for functional genomics, Schena [/bib_ref]. However, system-level techniques like microarrays produce large datasets, requiring efficient computational methods to identify significant changes in gene expression and their correlation. One area of systems biology where these computational methods are increasingly applied is for the inference of GRNs [bib_ref] Inferring gene networks from time series microarray data using dynamic Bayesian networks, Kim [/bib_ref] [bib_ref] Lessons from a gene regulatory network: echinoderm skeletogenesis provides insights into evolution,..., Ettensohn [/bib_ref] [bib_ref] Module-based analysis of robustness tradeoffs in the heat shock response system, Kurata [/bib_ref] [bib_ref] Dynamic modeling of cis-regulatory circuits and gene expression prediction via cross-gene identification, Lin [/bib_ref] [bib_ref] Deciphering complex mechanisms in neurodegenerative diseases: the advent of systems biology, Noorbakhsh [/bib_ref] [bib_ref] Computational toxicology-a state of the science mini review, Kavlock [/bib_ref]. Although microarray databanks contain a wealth of data in support of the elucidation of GRNs, mammalian datasets are often limited by low or nonexistent replication and too few time points to allow for reliable results. There have been suggestions (e.g. Sîrbu et al. [bib_ref] Comparison of evolutionary algorithms in gene regulatory network model inference, Sîrbu [/bib_ref] that the wealth of biological knowledge on possible interactions available in the literature coupled with the limits on available microarray data warrant an attempt to implement an integration of the two to improve reliability of GRN inferencing results.
We propose utilizing knowledge extracted from publications as a network of interactions and then applying microarray data to provide a measure of the quantitative effect of each of the individual interaction components. The qualitative knowledge at the core of our approach is provided by SemRep, a natural language processing system that extracts textual meaning from the biomedical literature in the form of semantic relations called predications (subject-relation-object triples). These predications mostly represent mammalian and specifically human interactions, therefore we use quantitative data provided by analysis of publicly available human breast cancer microarray datasets using a genetic algorithm. We compare the fitness of our model to that of a highperforming model from the literature to determine the performance of our technique as well as another model based on timedelay correlation. We also compare our results to KEGG pathways and find not only included interactions but also interactions not included but validated in the literature. We then modified the components of our method to be compatible with yeast data and again compared with the state-of-the-art model. Our method provides a novel approach to enhancing microarray data analysis with knowledge from the literature, and is the first such approach to incorporate semantic relations. Combining microarray data with semantic relations provides a more accurate model of gene interactions directing the behavior of cells, and consequently an enriched understanding of molecular physiology. This supports the discovery of disease mechanisms, which can then be exploited for diagnostic and therapeutic development in medicine.
# Related work
Previous computational models used to reconstruct gene regulatory networks from microarray data have employed techniques such as Hidden Markov, Bayesian network, and stochastic differential equation models [bib_ref] The application of hidden markov model in building genetic regulatory network, Ji [/bib_ref] [bib_ref] Hidden Markov models in computational biology: Applications to protein modeling, Krogh [/bib_ref] [bib_ref] Stochastic delay differential equations for genetic regulatory networks, Tian [/bib_ref] [bib_ref] A stochastic differential equation model for quantifying transcriptional regulatory network in Saccharomyces..., Chen [/bib_ref]. In a Hidden Markov Model, the real states of genes are treated as hidden variables, and the gene expression values are observed. This permits the states of genes at a given time t to be considered as depending only on the previous time point t-1. A Bayesian network is a representation of a joint probability distribution. When applying Bayesian networks to genetic regulatory systems, nodes are identified with genes and their expression levels, edges indicate interactions between genes, and conditional distributions describe these interactions. In the stochastic differential equation, the change of expression of a given target gene over two adjacent time points is equal to the accumulation of the weighted results of the sigmoid function of its regulator genes plus a random error. These computational models all reflect one or more aspects of the nature of genes and gene regulatory networks, but computational complexity limits the dimensionality of the modeled networks, and sensitivity to noise in gene expression measurement largely reduces their accuracy.
Genetic algorithms (GA)have also been widely used for the inference of gene regulatory networks [bib_ref] Inferring a system of differential equations for a gene regulatory network by..., Sakamoto [/bib_ref] [bib_ref] Identifying gene regulatory network as differential equation by genetic programming, Sakamoto [/bib_ref] [bib_ref] Inference of gene regulatory model by genetic algorithms, Ando [/bib_ref]. Genetic algorithms are inspired by Darwin's theory of evolution. Within this methodology, a population of candidate solutions to a problem is created and then evolves over a specified number of generations using phenomena such as cross-over (swapping components from other candidates) and mutation (internal, random changes). The best candidates are propagated through with incremental changes in the overall structure and the final generation becomes the solution. At each generation, fitness functions are used to determine the fittest candidates. A genetic algorithm is a stochastic algorithm and therefore it is highly likely to find global optima and can easily escape local maxima. Since the genetic algorithm execution technique is not dependent on the error surface, it is capable of solving multi-dimensional, non-differential, non-continuous, and even non-parametrical problems, which is the nature of gene expression data. Keedwell et al. used small random gene subsets evaluated by an artificial neural network (ANN) that is optimized by gradient descent to form the population of the genetic algorithm [bib_ref] Discovering gene networks with a neuralgenetic hybrid, Keedwell [/bib_ref] ; Liu and Wu used a differential equation to model the GRNs and genetic programming for optimization [bib_ref] Modeling Gene Regulatory Network Based on Genetic Programming, Liu [/bib_ref] ; Sîrbu et al. compared various evolutionary algorithms for GRN inferencing and found that Keedwell et al.'s method performs the best overall [bib_ref] Comparison of evolutionary algorithms in gene regulatory network model inference, Sîrbu [/bib_ref] , prompting us to use their method for the basis of our model. They also note that using literature-derived knowledge offers the potential for significant improvement over techniques that probe component genes randomly, and our use of such knowledge forms the basis of our deviation from the Keedwell et al.'s model.
With the rapid rate of growth in the biology literature, text mining is increasingly seen as indispensable in managing and discovering new biological knowledge [bib_ref] Natural language processing and systems biology, Cohen [/bib_ref]. An active area of research in biological text mining has been extraction of interactions between biomolecular entities (genes, proteins, etc.) from the research literature. Many systems, adopting various representational means (binary interactions, events, etc.) and using a variety of rule-based and machine learning-based techniques, have been proposed for this task. Early systems that focused only on co-occurrence of entities were soon replaced by systems that relied on shallow parsing and hand-crafted syntactic rules to extract binary interactions . These methods generally provided high precision at the expense of lower recall, in contrast to co-occurrence based methods. More recently, dependency parsing has become the predominant syntactic tool in extracting biological relations, as evidenced by the BioNLP Shared Task competitions [bib_ref] Overview of BioNLP Shared Task, Nédellec [/bib_ref]. These competitions have also signaled the increasing focus on events as the representational means for biological relations. Most commonly, dependency relations have formed the basis for syntactic features (shortest paths, dependency n-grams, etc.) for machine learning methods, along with lexical (tokens, n-grams, part-of-speech tags, etc.) and semantic (entity types, hypernyms, etc.) features. Best machine learning approaches have included pipeline models based on support vector machines [bib_ref] University of Turku in the BioNLP'11 Shared Task, Björne [/bib_ref] [bib_ref] Event extraction with complex event classification using rich features, Miwa [/bib_ref] as well as model combination techniques [bib_ref] Combining joint models for biomedical event extraction, Mcclosky [/bib_ref] and joint inference [bib_ref] Fast and robust joint models for biomedical event extraction, Riedel [/bib_ref]. Some rule-based systems have reported competitive results in this task, as well [bib_ref] Biological event composition, Kilicoglu [/bib_ref]. Recently, coreference resolution has also been beneficially integrated into several event extraction systems [bib_ref] Coreference based event-argument relation extraction on biomedical text, Yoshikawa [/bib_ref] [bib_ref] Boosting automatic event extraction from the literature using domain adaptation and coreference..., Miwa [/bib_ref].
As these text mining methods mature, they are increasingly applied to practical needs of biologists, in tasks such as database
## Author summary
We have developed a methodology that combines standard computational analysis of gene expression data with knowledge in the literature to identify pathways of gene and protein interactions. We extract the knowledge from PubMed citations using a tool (SemRep) that identifies specific relationships between genes or proteins. We string together networks of individual interactions that are found within citations that refer to the target pathways. Upon this skeleton of interactions, we calculate the weight of the interaction with the gene expression data captured over multiple time points using state-of-theart analysis algorithms. Not surprisingly, this approach of combining prior knowledge into the analysis process significantly improves the performance of the analysis. This work is most significant as an example of how the wealth of textual data related to gene interactions can be incorporated into computational analysis, not solely to identify this type of pathway (a gene regulatory network) but for any type of similar biological problem.
curation and pathway generation. For example, two tasks in the recent BioNLP 2013 Shared Task competition [bib_ref] Overview of BioNLP Shared Task, Nédellec [/bib_ref] (Pathway Curation [bib_ref] Overview of the Pathway Curation (PC) task of BioNLP Shared Task, Ohta [/bib_ref] and Gene Regulation Network in Bacteria [bib_ref] BioNLP Shared Task 2013 -An overview of the Genic Regulation Network Task, Bossy [/bib_ref] investigated the feasibility of automatically constructing such networks from the literature alone. Given a set of relevant biomolecular entities, the former focused on extracting pathwayrelevant events (e.g. gene expression, regulation, binding, regulation), while the latter focused on constructing a gene regulation network for the model bacterium Bacillus subtilis involving these entities. For the latter task, participating groups could either directly construct a regulation network for the provided entities or extract the interactions from which such a network could be derived using a predefined algorithm that the organizers provided. While the results were encouraging, these tasks remain challenging as evidenced by the limited participation and the fact that both tasks presupposed that the entities involved were already known, thus, addressing only a fraction of the problem of network construction from the literature. From an opposing viewpoint, Miwa et al. [bib_ref] A method for integrating and ranking the evidence for biochemical pathways by..., Miwa [/bib_ref] focused on linking interactions in biological pathways to supporting evidence from the literature; however, their work does not address the task of pathway construction.
There have been several attempts at improving gene regulatory network modeling by incorporating existing knowledge. For example, Steele et al. used the correlation between different gene concept profiles to calculate the probability of edges in GRNs modeled by Bayesian networks [bib_ref] Literature-based priors for gene regulatory networks, Steele [/bib_ref]. These profiles are determined by the occurrence of terms in the literature, using the Unified Medical Language System (UMLS) for normalization. Gutierrez-Rios et al. used regulatory interactions described in RegulonDB, a database of the regulatory network of Escherichia coli K-12, to establish the network of causal relationships to evaluate the congruence between the literature and whole-genome expression profiles [bib_ref] Regulatory network of Escherichia coli: consistency between literature knowledge and microarray profiles, Gutierrez-Rios [/bib_ref]. Additionally, the literature contains examples of efforts to combine literature-derived networks and microarray analysis in contexts other than GRN inferencing. Duarte et al. reconstructed the human metabolomic network depending largely on a manual literature review combined with knowledge extracted from genomic databases and use gene expression analysis to fill in the gap for a subset of network components [bib_ref] Global reconstruction of the human metabolic network based on genomic and bibliomic..., Duarte [/bib_ref]. text mining to derive biological pathways from literature relevant to in-stent restenosis and analyzed which of these were most relevant to the expression profiles identified by microarray analysis of tissue samples from patients with this condition [bib_ref] Network analysis of human in-stent restenosis, Ashley [/bib_ref]. All of these techniques either use human review to identify asserted interactions or automated approaches to infer them based on cooccurrence of terms. Our approach combines automation, allowing for an exponentially greater survey of the literature, with the identification of assertions in the text by SemRep, moving beyond mere term co-occurrence and increasing the validity of the extracted relations.
Several books, book chapters, and journal reviews are available detailing the pros and cons of various modeling approaches for GRN inferencing [bib_ref] Approaches to modeling gene regulatory networks: a gentle introduction, Schlitt [/bib_ref] [bib_ref] Gene regulatory network inference: data integration in dynamic models-a review, Hecker [/bib_ref] [bib_ref] Modelling and analysis of gene regulatory networks, Karlebach [/bib_ref] [bib_ref] Modeling and simulation of genetic regulatory systems: a literature review, De Jong [/bib_ref] [bib_ref] Computational studies of gene regulatory networks: in numero molecular biology, Hasty [/bib_ref]. In addition to a description of various approaches, Karlebach and Shamir [bib_ref] Modelling and analysis of gene regulatory networks, Karlebach [/bib_ref] also provide a comparative summary of the relative advantages of different types of models for various features. They align GRN models along a spectrum from logical models (e.g. Boolean networks) to continuous models (e.g. linear differential equations) to single-molecule level models (e.g. stochastic simulation models). They identify the logical end of the spectrum as having a decreased detail, less faithfulness to biological reality, lower data quantity needs, and reduced ability to model dynamics, while having greater model size, computational speed, and inferencing ability. Models at the single-molecule level are positioned at the other end of the spectrum, with a higher level of detail, increased faithfulness to biological reality, greater data quantity needs, and increased ability to model dynamics and decreased model size, computational speed, and inferencing ability. Continuous models are positioned in the middle of the spectrum, with moderate levels of the assessed model's characteristics. Although not providing a complete picture of the comparative characteristics of Step 1) a network is formed from semantic predications extracted from MEDLINE by SemRep for each set of citations related to a given pathway;
Step 2) each network is filtered by predication argument distance and frequency;
Step 3) a genetic algorithm uses gene expression data to quantify the weight of the interactions of the gene regulatory network. doi:10.1371/journal.pcbi.1003666.g001 GRN modeling techniques, they give an easily accessible summarization.
# Background
SemRep. The idea that the literature holds a wealth of data that can be systematically used in the development of hypotheses has been a point of discussion in biology for over a decade [bib_ref] Conceptual biology: unearthing the gems, Blagosklonny [/bib_ref]. The challenge is facilitating the extraction of relevant information from the literature. SemRep [bib_ref] The interaction of domain knowledge and linguistic structure in natural language processing:..., Rindflesch [/bib_ref] addresses this challenge by extracting salient content from titles and abstracts of MEDLINE citations in the form of semantic predications, thereby facilitating large-scale analysis of biomedical knowledge. A semantic predication is a subject-predicate-object triple, whose elements are largely derived from ontological knowledge in the UMLS [bib_ref] The unified medical language system (UMLS): integrating biomedical terminology, Bodenreider [/bib_ref]. The subject and object arguments correspond to concepts in the UMLS Metathesaurus and the predicates to relations in an extended version of UMLS Semantic Network. For example, SemRep extracts the predication in (2) from the sentence in (1):
(1) In addition to the upregulation of death receptors, p53
induced the pro-apoptotic Bcl-2 family members Bik and Bak and downregulated the anti-apoptotic Bcl-xL protein.
[formula] (PMID: 11313989) (2) p53-STIMULATES-Bik [/formula]
SemRep relies on the UMLS SPECIALIST Lexicon [bib_ref] Lexical methods for managing variation in biomedical terminologies, Mccray [/bib_ref] , MedPost part-of-speech tagger [bib_ref] MedPost: (2004) a part-of-speech tagger for biomedical text, Smith [/bib_ref] , and a partial syntactic parser (''chunker'') for syntactic analysis. Noun phrases identified by the parser are mapped to UMLS Metathesaurus concepts by the MetaMap program [bib_ref] An overview of MetaMap: historical perspective and recent advances, Aronson [/bib_ref]. With respect to gene/protein terms, ABGene [bib_ref] Tagging gene and protein names in biomedical text, Tanabe [/bib_ref] and EntrezGene [bib_ref] Entrez Gene: gene-centered information at NCBI, Maglott [/bib_ref] serve as supplementary resources to MetaMap and the UMLS Metathesaurus, respectively. To identify predicate mentions in text, SemRep uses indicator rules that map lexical/syntactic constructions, such as verbs, nominalizations, and modifier-head structures, to UMLS Semantic Network relations. For example, one of those rules stipulates that the verb induce may indicate the relation STIMU-LATES, and this rule licenses mapping induced in the example above to this relation. The subject and object arguments (p53 and Bik, respectively) are recognized using MetaMap and ABGene and are then normalized to official EntrezGene symbols. Several studies have evaluated predications extracted by SemRep. For example, Ahlers et al. reported 73% precision and 55% recall in pharmacogenomics relations [bib_ref] Extracting semantic predications from Medline citations for pharmacogenomics, Ahlers [/bib_ref]. Another evaluation focused on a specific linguistic structure (nominalizations) and reported 75% precision and 57% recall [bib_ref] Arguments of nominals in semantic interpretation of biomedical text, Kilicoglu [/bib_ref]. The entire MEDLINE database has been preprocessed with SemRep for efficient access, resulting in the SemMedDB database, which contains more than 57 million predications extracted from 21 million citations, as of June 30, 2012 [bib_ref] SemMedDB: a PubMed-scale repository of biomedical semantic predications, Kilicoglu [/bib_ref]. By normalizing free text to semantic predications, SemRep provides the ability to represent biomedical content as a network of relations. In our model, the nodes of the network represent genes and the edges represent interactions (INTER-ACTS_WITH, INHIBITS, and STIMULATES) between the genes.
## Microarray data
The breast cancer microarray data used in our experiments comes from NCBI's Gene expression omnibus (GEO: http:// www.ncbi.nlm.nih.gov/geo/) [bib_ref] NCBI GEO: archive for high-throughput functional genomic data, Barrett [/bib_ref]. GEO provides free access to raw and processed microarray and sequencing data submitted by researchers based on their published work. For the yeast study, we used the cdc15 time course data from the Yeast Cell Cycle Analysis Project website (http://genome-www.stanford.edu/ cellcycle/) originally used in [bib_ref] Comprehensive identification of cell cycle-regulated genes of the yeast Saccharomyces cerevisiae by..., Spellman [/bib_ref].
## Kegg
For assessment of our methodology, we use the Kyoto Encyclopedia of Genes and Genomes (KEGG: http://www. genome.jp/kegg/) [bib_ref] KEGG for integration and interpretation of large-scale molecular datasets, Kanehisa [/bib_ref] , which provides gold standard sets of molecular pathways. KEGG pathways are manually curated networks based on a review of protein-gene and protein-protein interactions described in the literature. It is worth noting that KEGG pathway maps are an abbreviated representation of known interactions, focusing on those considered to be best supported by evidence and most relevant. As a consequence, the interactions in a KEGG pathway form a subset of those in the literature, and it is necessary to assess separately the validity of interactions not in the KEGG pathway. A KEGG pathway can be accessed as a downloadable kgml file or an online map, which have slight differences regarding which genes are included and which interactions are specified. To mediate these differences we use both formats for our assessment and consider whether genes and interactions are included in either version.
# Methods
We infer gene regulatory networks in three steps, illustrated in [fig_ref] Figure 1: Literature-based gene regulatory network discovery schema [/fig_ref]. N Effect quantification uses a genetic algorithm with an ANN fitness function on microarray data to quantify gene-gene interaction strength in the network created in previous steps.
To facilitate the establishment and assessment of the system model, we select the well-studied disease breast cancer as the starting point, due to the availability of relevant citations, microarray data, and established KEGG pathways. The 13 pathways contained in the KEGG Pathways of Cancer (human) map (http://www.genome.jp/kegg-bin/show_pathway?hsa05200) were used to guide the predication network generation and also in evaluating the resulting GRNs. These pathways are the p53 (see [fig_ref] Figure 2: The KEGG p53 signaling pathway entry showing the pathway map and some... [/fig_ref] , Apoptosis, Cell Cycle, PPAR, VEGF, MAPK, Wnt, TGF-beta, mTOR, jak-STAT, ErbB, Focal Adhesion, and Adherens Junction pathways.
## Predication network generation
In the first step, predication network generation, predications containing gene-gene interactions (INTERACTS_WITH, STIM-ULATES, and INHIBITS predications) are extracted from the MEDLINE citations that supported manual curation of each KEGG pathway; these citations are listed in the pathway entry on the KEGG website. These predications are augmented with those from additional relevant citations, which we identify by first extracting the Medical Subject Heading (MeSH) terms for each citation identified in KEGG and then manually refining that list to eliminate non-specific terms such as Humans, Male, or Biological Models. The resulting MeSH terms formed the basis of our PubMed searches. The terms that occurred in a significant distribution across the citations were grouped with an ''OR'' in the query when they were roughly equivalent or part of a set of different subtopics. As an example, the query for p53 citations was '' .'' This procedure was repeated to provide a citation list for each pathway. The number of citations for each pathway is given in [fig_ref] Table 1: Number of predications for each pathway [/fig_ref] in 'Citation' column. The predications extracted from these citations were then retrieved from SemMedDB. The number of resulting predications for each pathway is shown under the column heading 'Raw' in [fig_ref] Table 1: Number of predications for each pathway [/fig_ref].
## Network filtering and normalization
To improve the reliability of our approach, we used three predication filtering mechanisms, explained below. These mechanisms result in a smaller set of predications extracted for each pathway, which forms subnetworks. These subnetworks, together, serve as the predication network. provides an example of a subnetwork created from predications related to p53. This example has been pruned to provide a small network for simplicity.
## Argument-predicate distance filtering
Argument-predicate distance is the number of intervening noun phrases between an argument and its predicate in the sentence from which they were extracted. It has been shown that smaller argument distance leads to higher precision in extracting interactions [bib_ref] Argument-predicate distance as a filter for enhancing precision in extracting predications on..., Masseroli [/bib_ref]. In the example below, both Bax and procaspase-3 are potential objects for p53 activity.
(3) Furthermore, the up-regulation of p53 promoted Bax expression, which led to the activation of pro-caspase-3 and Bax has an object distance of 1 since it is the first noun phrase subsequent to the predicate 'promoted' and pro-caspase-3 has a distance of 2. Preferring an argument-predicate distance of 1 selects the predication p53 STIMULATES Bax, while eliminating the predication p53 STIMULATES pro-caspase-3. In our experiments, we limit both the predicate-subject distance and the predicate-object distance to 1, which decreased the number of predications to approximately 27% of the initial number of predications on average. The number of predications for each pathway after filtering using argument-predicate distance of 1 is shown in [fig_ref] Table 1: Number of predications for each pathway [/fig_ref] (under 'Dist.').
## Normalization
After argument-distance filtering is applied, we normalize the set of predications by removing duplicate predications and mapping the subjects and objects to formal gene symbols based on the standard gene name dataset (HUGO, http://www. genenames.org/). If a predication argument cannot be mapped to a formal gene symbol, the predication is pruned. The numbers of predications after duplicate removal and normalization are given in [fig_ref] Table 1: Number of predications for each pathway [/fig_ref] (under 'Uniq.' and 'Norm.', respectively).
## Document frequency filtering
Document frequency for a predication is the number of citations in which the predication occurs. Document frequency filtering is based on the hypothesis that the confidence credential of a predication is in direct ratio to its occurrence in documents. In our experiments, we discard all predications that occur in fewer than two articles. The result is shown in [fig_ref] Table 1: Number of predications for each pathway [/fig_ref] ('Freq.'). Note that network filtering discards all relevant predications for the Focal Adhesion pathway, which was not considered in subsequent steps.
## Effect quantification
In the final step we quantify gene-gene interaction strength in the network generated in previous steps, using a genetic algorithm, depicted in [fig_ref] Figure 4: Genetic algorithm for effect quantification [/fig_ref]. In the initial population of chromosomes, each chromosome contains a candidate set of interactions between all genes in the pathway. The predication network determines the initial gene-gene interactions, while the strength of interaction is initially randomly generated for each of the 2000 chromosomes in the population. Then for each generation, the fittest candidates are replicated into the next generation and the balance of the population is filled through reproduction of the current generation, using random crossover and mutation to introduce novel diversity into the population. We compute the fitness of a chromosome, as compared to the time series microarray data, using an artificial neural network. These procedures are described in more detail in the following subsections.
## Genetic algorithm
With the genetic algorithm used to infer the gene regulatory networks, we train the weights of the inbound interactions for every gene separately. A population of chromosomes is created where each chromosome is represented as a matrix of interaction weights between each possible pairing of genes identified from the predications. Each weight, valued between 21 (greatest inhibition) and +1 (greatest stimulation), indicates the strength of interaction. A weight of zero indicates no interaction. The predications in the network define which genes have interactions and whether the interaction is inhibitory or stimulatory, but do not contribute to the determination of the weight. The absence of an interaction is represented in the matrix with a weight of zero but is not altered through subsequent generations. The direction of the interaction is maintained from the subject-object relation in the predication. We randomly generate 2000 chromosomes that contain different weights for the interactions in the predication network. The gene structure in each chromosome is the same, as well as the noninteracting/zero-weighted gene pairs, but the weights representing the strength of inhibition or stimulation of pairs found in the predications are varied. The population of chromosomes is then evaluated with the fitness function in , and the fittest 20 percent of the chromosomes are copied into the next generation directly. The rest (80 percent) of the chromosomes for the next generation are generated by crossover at a specific (randomly selected) gene pairing between pairs of randomly selected chromosomes in the current generation with a mutation rate of 0.25 percent. This mutation rate approximates those commonly used to facilitate convergence of the chromosomes toward a fittest result and to avoid excessive intergenerational fluctuation. After evolving for 200 generations (an empirically determined limit), the chromosome with the highest fitness value is selected as the final result.
## Fitness test function
In our experiments, we use an artificial neural network (ANN) to model the interaction of the genes in each pathway. In this model, the gene expression level of a given gene at a given time point is a function of all other genes at the previous time point (see [fig_ref] Figure 5: Population and chromosome weight matrix [/fig_ref]. In a comparative study by Sîrbu et al., this model outperformed 4 other recent GRN models (GA + evolutionary strategy, differential evolution + Akaike's Theoretic criterion, genetic local search, and an iterative algorithm based on GA) in a combined score of 6 performance measures (data fit, parameter quality, noise, sensitivity, specificity, and scalability) [bib_ref] Comparison of evolutionary algorithms in gene regulatory network model inference, Sîrbu [/bib_ref]. We follow the most common implementation of such a model, using a nonlinear weighted sum. Equation (1) shows the fitness test function based on the ANN model.
[formula] f i i~2 1ze ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi X R r~1 X T{1 t~1 (g g tz1 i {g tz1 i ) 2 (T{1)|R v u u t ð1Þ whereg g tz1 i~K ( X n j~1w w ji g t j )ð2Þ [/formula]
and T indicates the number of time points in the microarray dataset, R is the number of microarray replicates at each time point, r is the current replicate, n is the number of genes in the pathway, K is the activation function, w ! ji is the weight of the interaction between gene j and gene i, and g t i is the gene expression value in the microarray set for gene i at time point t. Since the difference of the expression value of a gene over each time course is considered to be a direct function of only the joint effect of other genes in the pathway and the gene expression values used have been normalized between 21 and 1, we use a linear function as the activation function, effectively dropping the term. A visual representation of the ANN model is presented in [fig_ref] Figure 6: Artificial neural network fitness function [/fig_ref]. Inference of yeast cell cycle GRN MEDLINE contains many more citations related to human genes than yeast genes. A simple PubMed search for ''yeast and gene'' returns a little more than 86,000 citations while the query ''human and gene'' returns over a million. When combining either species term with ''cell cycle'' (a concept with a heavy amount of research done in yeast), there are over 96,000 citations related to human and just over 15,000 for yeast. In addition to this limitation, SemRep was designed to identify human genes and uses Entrez Gene database entries specific to humans for genes and proteins. Although this led to an initial study with human genes, we made modifications to SemRep to support a study with yeast. This required changing the data source to the Entrez Gene fungi dataset that contains Saccharomyces species and processing relevant citations (resulting from the PubMed query ''Saccharomyces[mh] AND cell cycle[major]'') to extract predications. All subsequent procedures were consistent with the breast cancer study except that the predications were not filtered by argument distance or document frequency due to the lower initial number of predications. The number of citations and predications at each step are given in [fig_ref] Table 2: Number of predications for yeast cell cycle pathway [/fig_ref].
# Results
Evaluating a gene regulatory network inference model for human data is challenging, since there is no gold standard providing a complete reference of gene connectivity. Therefore, we evaluated the implementation of our approach using breast cancer in two ways: (1) comparing the accuracy of the model against the highest-performing model reported in the literature, [bib_ref] Gene regulatory networks for development, Levine [/bib_ref] comparing the resulting networks to KEGG pathways and literature that formed the basis for their extraction, paying particular attention to the p53 pathway. Comparison to a KEGG pathway provides an assessment of the contribution of our natural language processing techniques (i.e., SemRep).
## Comparison of model accuracy
As a measure of accuracy of the model, we determine how well the model fits the data, in line with previous research [bib_ref] Discovering gene networks with a neuralgenetic hybrid, Keedwell [/bib_ref]. For this purpose, we use microarray data from a human breast cancer experimental set (GEO: GSE29917), which contains expression values for 7 time points and 6 replicates (two microarrays are missing from the set for a total of 40 microarrays). We compare the accuracy of our model with that of the Repeated Genetic Algorithm with Neural Network model described by , the best performing model in a recent comparison of evolutionary algorithms [bib_ref] Comparison of evolutionary algorithms in gene regulatory network model inference, Sîrbu [/bib_ref] and the basis for the interaction quantification component of our algorithm. Since we use the same algorithm for interaction quantification, the comparison helps isolate the effect of literature-derived knowledge on GRN inference. We downloaded their source code to be able to run their algorithm on our data, with only minimal modification for data format differences. Additionally we included a second comparison model, with a different type of algorithm (time-delayed Spearman Rank-correlation or TDSRC) to help identify any biases in our methodology [bib_ref] A computational framework for gene regulatory network inference that combines multiple methods..., Gupta [/bib_ref]. This model was included by Gupta et al. as a part of a composite model and available for download and use within MatLab. For this comparison, we limit the microarray data to 78 genes included in the KEGG P53 pathway (listed in . We use a leave-oneout approach based on time points, sequentially using gene expression values for each time point as test data against models trained with values from all of the other time points. Fitness is defined as the standard deviation of predicted values from the average microarray expression of all 6 replicates in the test set. Because the microarrays for 2 different replicates at two different timepoints were missing, the total number of microarrays was 40.
The root mean square error (RMSE) combines fitness for all genes in the given test set. [fig_ref] Figure 7: Comparison of fitness to microarray data [/fig_ref] shows the root mean square errors over the 78 genes for each time point for each model. We used a paired t-test to assess the statistical significance of the differences between models. The p-values for each comparison are included in
## Assessment of interactions
We assessed the value of gene regulatory networks generated by our approach by comparing them to KEGG pathways, both the kgml format downloaded from the KEGG website and manually using the search function in the online pathway map. Nodes, representing genes, and edges, representing interactions, were independently compared and the results are shown in . The 'Predication' column represents the number of nodes and edges in the network generated by our approach after removing any genes that were not included in the microarray set, the 'KEGG' column provides the equivalent information for the kgml network, 'Common' indicates the intersection of our results and the corresponding kgml network, and 'New' provides interactions exclusive to our results. As the numbers indicate, new nodes and edges significantly outnumber those in common between the two networks. The p53 pathway had the highest ratio of common edges compared to the kgml network (19:65, 29.2%), so was chosen for further validation at the predication level.
A comparison was made for each of the new interactions with the online KEGG p53 signaling pathway map. 49 of the 92 interactions contained a gene not included in the map, 7 had both genes present but no interaction, and 9 existed in the map but were not included in the kgml version.
We also validated each new interaction in the resulting p53 network against the literature in three ways: a) whether the literature asserts an interaction between the two at either a gene or protein level (i.e., the precision of our natural language processing techniques), b) whether we capture correctly the direction of the effect, i.e. stimulatory or inhibitory, as compared to the weight generated by GRN analysis of the microarray data, and c) combining the two above, whether we capture correctly both the presence of the interaction and the direction. In addition to assessing the accuracy of SemRep in capturing relevant interactions, this validation allows us to establish the specific contribution of using semantic predications over term co-occurrence. The comparison was limited to citations from which the predications were extracted, ranging from 1 to 87 citations for each interaction. Although there may exist another citation that would validate the resulting interaction, this approach was taken to limit the manhours required to a reasonable amount while still allowing a reasonable possibility for validation. As seen in [fig_ref] Table 5: Literature validation of newly included interactions [/fig_ref] , 78.3% of the new interactions in the resulting p53 network [fig_ref] Figure 8: The resulting gene regulatory network for the p53 pathway [/fig_ref] were asserted in at least one of the source citations. When comparing the sign of the interaction weight to the direction of the effect provided in the source literature, 76.4% were consistent. Those interactions that were both consistent with the literature and were weighted in the appropriate direction numbered 55 out of 92 (59.8%).
We additionally focused on those interactions having a weight with absolute value .0.1 [fig_ref] Table 5: Literature validation of newly included interactions [/fig_ref] , bottom), exploring the hypothesis that stronger interactions should be less affected by noise. Within these interactions a total of 32 interactions (84.2%) were stated in the literature and 34 (87.5%) were correct in the direction of effect, yielding 28 (73.7%) correct on both counts.
Finally, we investigated whether argument-predicate distance filtering had a detrimental effect on the results, since long distance syntactic dependencies are common in biomedical literature [bib_ref] RelEx-relation extraction using dependency parse trees, Fundel [/bib_ref]. For this purpose, we focused on the Jak-STAT pathway and checked whether any of the 27 KEGG interactions, none of which appeared in our predication network, could be derived from the initial set of unfiltered, non-normalized predications. We found three such predications for the Jak-STAT pathway; one (Jak1 INTERACTS_WITH PTPN11) was eliminated due to argumentdistance filtering, while the other two (Jak1 INTERACTS_WITH STAT1 and GRB2 INTERACTS_WITH PTPN11) passed the argument-distance filter but not the document frequency filter. Note that the former (Jak1 INTERACTS_WITH PTPN11) occurred in a single document and, therefore, would also be eliminated in the subsequent document frequency filtering step, had it passed the predicate-argument distance filter.
## Performance on yeast data
We compared the performance of our model against that of Keedwell et al. in the same manner as mentioned in our results section Comparison of model accuracy, but now on the yeast cellcycle dataset, which contains 24 time points. shows the root mean-square errors for each time interval for the two models. We assessed statistical significance of the differences between models by calculating the p-values (included in using a paired t-test. Using the yeast data, our model had increased fitness at every interval over the Keedwell et al. model, but this time every difference was significantly different (p,0.05).
# Discussion
In this work, we augmented microarray data with literature knowledge to infer gene regulatory networks, replacing the random selection of component genes generally used in similar modeling efforts. We use SemRep to extract information from the literature, which forms the backbone of the network. Stateof-the-art genetic algorithm-based analysis of the microarray data was used to determine the strengths of the effects between gene-gene interactions in the network. Our model provides better fitness than the state-of-the-art model used as the basis for the genetic algorithm and fitness function components of our model and the difference between error rate of the models is statistically significant (p,0.05), demonstrating that litera- ture-derived knowledge provides a significant advantage over random selection of genes in this task, as suggested in [bib_ref] Comparison of evolutionary algorithms in gene regulatory network model inference, Sîrbu [/bib_ref]. Another advantage of our approach is that it maximizes the best possible networks without being limited to relationships that are already well established and considered important by a curator, as would be the case if a standard interaction database were used. These pathways provide targets for novel therapeutic interventions and a mechanistic understanding of current therapeutic approaches with poorly understood mechanisms.
## Interactions identified by our model not included in kegg
Although some of the genes in the KEGG pathways did not appear in the results due to the strict filtering process, the presence of essential member genes in the resulting predication networks (for example, TP53, MDMD2, BAX, CDKN1A, GADD45A, and CDK2 in the p53 network) and known interactions (p53 STIMULATES Gadd45a, E2F1 STIMULATES cdkn2a, and RCHY1 INHIBITS tp53) demonstrate the potential of this technique to replicate ''known'' pathways. Perhaps more importantly, we were able to identify new interactions that were not included in the KEGG maps, which is not surprising since these maps are curated and therefore provide only the most thoroughly established and important interactions as determined by the curators.
Two of the highest-weighted new interactions included in the resulting p53 pathway but not in the kgml file are p53 STIMULATES BIK and MDM2 INHIBITS CDKN1A. The weight of the interaction between p53 and BIK was very strongly stimulatory at 0.977. BIK (BCL-2 interacting killer) is not included in the KEGG P53 pathway map. It is a proapoptotic protein discovered in 1995 and interacts strongly with BCL-2 and BCL-xL [bib_ref] BIK, the founding member of the BH3-only family proteins: mechanisms of cell..., Chinnadurai [/bib_ref]. P53 has been shown to induce expression of BIK under certain conditions, especially in breast cancer cells, as in our microarray dataset [bib_ref] Characterization of two types of endothelial progenitor cells and their different contributions..., Hur [/bib_ref]. Our p53 pathway result also included an interaction between MDM2 an CDKN1A with a weight of 20.987, i.e., very strongly inhibitory. MDM2 is shown to inhibit p53 in the KEGG p53 map but there is only an indirect inhibitory action of MDM2 on CDKN1A through P53 (by diminishing p53's activation of CDKN1A). However, a direct interaction as a negative regulator of CDKN1A is present in the literature [bib_ref] MDM2 is a negative regulator of p21WAF1/CIP1, independent of p53, Zhang [/bib_ref] , supporting our result.
## Limitations and future directions
A major limitation of this method is the accumulation of system errors. As shown in [fig_ref] Table 5: Literature validation of newly included interactions [/fig_ref] , the accuracy of interactions identified using SemRep with the filtering used in our method is 78.3% overall and 84.2% for significantly weighted interactions (.0.1). The reproducibility of the microarray data between platforms and even with different algorithms on the same platform has been determined to be as low as 50-60% [bib_ref] Evaluation of gene expression measurements from commercial microarray platforms, Tan [/bib_ref]. At a result, the accuracy of both the structure and the weights of the resulting GRNs is limited if viewed at the most precise level of granularity. The training of the weights is also limited by an insufficient time course in the microarray data, especially for large numbers of arguments (i.e. genes in a pathway). Although it is beyond our means to improve microarray reproducibility or experimental design of publicly available datasets, improvements to SemRep with regard to gene and protein interactions can potentially improve results and the incorporation of multiple microarray datasets as training data may be able to overcome some of the inherent limitations. Additionally, although these limitations affect the use of these techniques for determining the precise quantitative nature of interactions, this is true generally of such studies, and does not prevent their use for hypothesis generation.
Currently our network backbone is limited to genes included in the predications, but this set could be expanded by incorporating genes from the microarrays that are similar to the genes provided by the predications, using an ensemble of machine learning techniques such as support vector machines, random forests, and prediction analysis of microarrays. This approach of classifying genes from the microarray into each pathway would additionally reduce bias in the technique by not limiting interactions to what is already known. It would be valuable to maintain a distinction between the literature-based genes and the expansion set because a significant expansion in the final resulting networks would suggest that what is known in these pathways is only a fraction of what is waiting to be discovered.
Another technique to extend the set of genes would be to adopt more sophisticated predication filtering mechanisms. The current mechanisms were effective in significantly reducing the computational complexity; however, our limited analysis of the effect of argument-distance filtering showed that some relevant predications were also eliminated due to filtering. A potentially useful approach would be to train a classifier that can identify 'good' predications, based on predication features, such as the predicate type, indicator types, etc. Predicate-argument distance could also serve as a feature for such a classifier.
Additionally, although we used interactions from KEGG to evaluate our GRN, such interactions from this and similar databases could be incorporated into our interaction network to augment the proposed interactions from literature with established interactions, expanding the utility of prior knowledge.
Although our current method of selecting source citations for predication extraction yielded usable results, there are many possible permutations in the method and a systematic comparison of source citations and their resulting networks should be explored. Our current search method facilitates the generation of hypothetical member genes for established pathways, but this approach can also be used to generate novel pathways by specifying a gene set and/or biological functions in the predication citation search, thereby providing an appropriate predication network.
SemRep representation of biomolecular interactions (subject-predicate-object triples) is simple, intuitively accessible and lends itself easily to downstream applications. On the other hand, more complex representations, particularly event representation (as discussed in Related work section), have been gaining in popularity in the BioNLP community, mostly due to available corpora [bib_ref] Corpus annotation for mining biomedical events from literature, Kim [/bib_ref] and shared task competitions [bib_ref] Overview of BioNLP Shared Task, Nédellec [/bib_ref]. An obvious question is whether such complex representations could provide an advantage over or could complement SemRep representation in the task of gene regulatory network inference. It seems very likely that generating a seed network from more complex representations would require some nontrivial post-processing along the lines of the algorithm described by Bossy et al. [bib_ref] BioNLP Shared Task 2013 -An overview of the Genic Regulation Network Task, Bossy [/bib_ref]. Since that algorithm essentially breaks down complex relations to simpler SemRep-style triples, it seems safe to assume that SemRep representation can adequately capture the complexity of biomolecular interactions. However, this needs further testing and validation.
To assess the precision of interactions for the yeast study, we performed an automated comparison using Cytoscape (http:// www.cytoscape.org) against a Biogrid (http://thebiogrid.org) yeast interaction dataset. We used the Biogrid Saccharomy-ces_cerevisiae-3.2.109 tab file, which contains 339,921 interactions and experimental information from multiple database sources including Saccharomyces Genome Database, MINT, IntAct, and Pathway Commons. As seen in [fig_ref] Table 6: Precision of yeast interactions [/fig_ref] , although 346 out of 349 genes are found in the reference set, only 147 out of 520 interactions are included. That is a relatively low precision of 28%, which only increases to 31% for interactions weighted 0.1 or higher. This is not particularly meaningful without appropriate context. Since this is a relatively straightforward and easily undertaken evaluation approach, it would be worthwhile to conduct a study applying the same evaluation across various published models to see how precision compares among them.
# Conclusions
We present a methodology of gene regulatory network inference that combines literature knowledge and microarray data. Using SemRep, we extract gene and protein interactions from citations related to the pathways included in the KEGG Pathways of Cancer. These predications are linked together to form a network and a genetic algorithm is used on a breast cancer time sequence microarray dataset to determine the weight of contribution of each stimulating or inhibiting gene on its target, thereby providing a weighted gene regulatory network. Our model performs significantly better than comparable models in terms of fitness of predictive output to microarray results. The resulting networks contain both interactions included in the appropriate KEGG pathways and interactions not included but validated through a literature search. The accuracy of these interactions increases from 60% overall to 74% when minimally-weighted interactions are excluded. Our model also performed better in terms of fitness against a comparison model when modified for yeast predications and microarray data. This approach offers significant potential in elucidating new interactions in existing pathways as well as the possibility of identifying novel pathways. In a broader sense, it also provides a blueprint for incorporating automatically extracted knowledge from literature with largescale biological analysis. Incorporating such knowledge underpins more accurate understanding of both normal and disturbed molecular physiology, leading to advances in diagnosis and treatment of disease.
## Supporting information
[fig] Figure 1: Literature-based gene regulatory network discovery schema. [/fig]
[fig] Figure 2: The KEGG p53 signaling pathway entry showing the pathway map and some reference citations. The PMIDs for each reference citation were used to identify MeSH terms with PubMed (see section Predication network generation). http://www.genome.jp/kegg-bin/ show_pathway?hsa04115. doi:10.1371/journal.pcbi.1003666.g002 [/fig]
[fig] N: etwork generation relies on SemRep and results in an initial predication-based network. etwork filtering and normalization involves normalizing gene names to formal gene symbols and selecting the most reliable interactions from the predication network. [/fig]
[fig] Figure 4: Genetic algorithm for effect quantification. An initial set of 2000 chromosomes, each containing the full complement of genes paired with every other gene, is created with an initial random assignment of interaction weights. The fitness of a chromosome is compared against the expression profile in the microarray data. The fittest 20% of chromosomes are replicated into the next generation, and the complement of the population is formed by random crossover and point mutation of interaction weights of members of the current generation. After 200 generations, the fittest candidate chromosomes are identified. ANN = artificial neural network. doi:10.1371/journal.pcbi.1003666.g004 eventually to apoptosis in MCF10A-ras cells. (PMID: 5566875) [/fig]
[fig] Figure 5: Population and chromosome weight matrix. A population of 2000 chromosomes is initially generated and maintained at each new generation. Each chromosome contains a matrix of interaction weights between every pairing of genes. P = population, C = chromosome, G = gene, w 2,1 = the weight of the interaction from gene 1 to gene 2. doi:10.1371/journal.pcbi.1003666.g005 [/fig]
[fig] Figure 6: Artificial neural network fitness function. Each contributory gene's expression in the current time point is magnified by its weight of interaction with the target gene. The sum of these contributions determines the expression level of the target gene at the subsequent time point. t indicates the time point in the microarray dataset, n is the number of genes in the pathway, w ! ji is the weight of the interaction between gene i and gene j, g t i is the gene expression value in the microarray set for gene i at time point t. doi:10.1371/journal.pcbi.1003666.g006 [/fig]
[fig] Figure 7: Comparison of fitness to microarray data. Fitness of our model is compared to that of the Keedwell et al. model and the TDSRC model on our microarray dataset. Fitness is significantly better (p,0.05) in 4 out of 6 time points and for all time points combined (overall). doi:10.1371/journal.pcbi.1003666.g007 [/fig]
[fig] Table 4: Comparison of resulting network and kgml network. number of nodes and edges in the predication network; KEGG: number of nodes and edges in the KEGG network; Common: number of nodes and edges in both predication and KEGG pathway; New: number of nodes and edges found in predication but not KEGG network. doi:10.1371/journal.pcbi.1003666.t004 [/fig]
[fig] 7: Our model shows improvement at every test data set/ time point over the Keedwell et al. model and is significantly different (p,0.05) in 4 out of 6 time points and for all time points combined (overall p = 8.26610 28 ). The RSP model performed significantly worse than both of the other models. This demonstrates that our model is significantly better in terms of fitting the microarray data. [/fig]
[fig] Figure 8: The resulting gene regulatory network for the p53 pathway. Interactions are defined by predications extracted from MEDLINE citations and weighted based on microarray data. Red node color signifies induction or activation of p53 while blue signifies suppression or deactivation and yellow indicates genes that do not act directly onto p53. Red arrows show induction/activation. Blue arrows show inhibition. Size of a node corresponds to the number of connections to other nodes. doi:10.1371/journal.pcbi.1003666.g008 [/fig]
[table] Table 1: Number of predications for each pathway. Citations: the number of citations returned from the PubMed search using related MeSH terms. Raw: the number of gene interaction predications SemRep extracted from the returned citations. Dist.: the predications after filtering out predications with distance .1. Uniq.: the number of unique predications after eliminating duplicates. Norm.: the number of predications after normalizing against the HUGO database. Freq.: the number of predications after filtering for a frequency of at least 2 predications. doi:10.1371/journal.pcbi.1003666.t001Figure 3. Example network created from semantic predications, edited for simplicity. Arrows indicate stimulation; bars indicate inhibition; loops indicate self-stimulation or self-inhibition.Interactions are directed, so an interaction from atm to tp53 is distinct from one from tp53 to atm. A single interaction can represent a single or multiple predications. doi:10.1371/journal.pcbi.1003666.g003 [/table]
[table] Table 2: Number of predications for yeast cell cycle pathway. Citations: the number of citations returned from the PubMed search using related MeSH terms. Raw: the number of gene interaction predications SemRep extracted from the returned citations. Unique: the number of unique predications eliminating duplicates. Normalized: the number of predications after normalizing against the [/table]
[table] Table 5: Literature validation of newly included interactions. [/table]
[table] Table 6: Precision of yeast interactions. Each gene and interaction predicted by our model was compared against interactions contained in the Biogrid dataset. Predicted: number of genes and interactions predicted by our model. In Biogrid: number of genes and interactions predicted by our model and found in the Biogrid dataset. Precision: number found/number predicted. doi:10.1371/journal.pcbi.1003666.t006Figure 9. Comparison of fitness to yeast cell cycle microarray data. Fitness of our model is compared to that of the Keedwell et al. model on a yeast cell cycle microarray dataset. Fitness is significantly better (p,0.05) over all points. doi:10.1371/journal.pcbi.1003666.g009 [/table]
[table] Table S1: PubMed queries for each pathway. The query submitted to PubMed for each pathway in the breast cancer study is provided. Each search provided a citation list relevant to the 13 pathways in the KEGG Pathways of Cancer map. mh: MeSH heading. sh: subject heading. (DOCX) [/table]
|
Familial hypercholesterolaemia is underdiagnosed and undertreated in the general population: guidance for clinicians to prevent coronary heart disease
# Introduction
Familial hypercholesterolaemia (FH) is a common genetic cause of premature coronary heart disease (CHD) (i.e. ischaemic heart disease), namely myocardial infarction and angina pectoris, due to lifelong elevated plasma low-density lipoprotein (LDL) cholesterol levels.If left untreated, men and women with heterozygous FH (later referred to simply as FH, unless specified as heterozygous or homozygous FH) with total cholesterol levels of 8-15 mmol/L (310-580 mg/dL) typically develop CHD before age 55 and 60, respectively, while homozygotes with total cholesterol levels of 12-30 mmol/L (460 -1160 mg/dL) typically develop CHD very early in life and if untreated die before age 20. However, once diagnosed, heterozygotes can readily be treated with cholesterollowering medication to attenuate development of atherosclerosis and to prevent CHD.The extent of underdiagnosis and undertreatment of individuals in the general population with FH is largely unknown. It is generally believed that among whites, 1/500 are heterozygous for FH and 1/1 000 000 are homozygous 1,2 ; however, even these individuals are not diagnosed in most countries.Furthermore, these prevalences likely represent underestimates and as cardiovascular disease is the leading cause of death worldwide.Indeed, many individuals and families with FH may simply be overlooked among the huge number of individuals with any CHD caused by more common risk factors, and as a consequence be underdiagnosed and undertreated for genetically elevated cholesterol levels.
The aim of the present consensus paper is to critically evaluate the extent to which FH is underdiagnosed and undertreated worldwide. Based on a consensus of the opinions of the experts in this panel and/ or on small studies, retrospective studies, and registries (level of evidence C 6 ), the EAS Consensus Panel proposes recommendations on (i) how better to diagnose individuals and families with FH and (ii) therapeutic strategies for best practice aimed to prevent CHD in these extremely high-risk individuals and families. Importantly, the effect of LDL cholesterol lowering on reduction in CHD and allcause mortality in individuals without FH is based on multiple randomized clinical trials and meta-analyses 7 (level of evidence A 6 ). Details of the levels of evidence specifically derived from FH studies can be found elsewhere.This Consensus Statement is aimed at cardiologists, endocrinologists, internists, paediatricians, general practitioners, clinical biochemists, public health practitioners, health service planners, other health professionals, and healthcare providers worldwide.
## Underdiagnosis
FH was not attributed an independent code in the World Health Organisation International Classification of Diseases, making reliable estimates of the number of individuals diagnosed with this condition difficult. Of the roughly 200 countries/territories in the world, we have therefore only been able to obtain estimates of the number of individuals diagnosed with FH for the 22 countries/territories shown in. Upon inclusion of the 180 countries/territories not listed in, ,1% are diagnosed in most countries. The few exceptions are 71% diagnosed in the Netherlands, 43% in Norway, 19% in Iceland, 13% in Switzerland, 12% in the UK, and 6% in Spain. Even these numbers are somewhat unreliable; for example, it has been estimated in Norway that roughly 1/300 have FH,and applying this prevalence, instead of 1/500, would mean that 26% rather than 43% of individuals with FH are diagnosed in Norway.
To date, the prevalence of FH has not been assessed directly in an unselected sample from the general population. Using the Copenhagen General Population Study,an unselected European general population sample comprising 69 016 participants with heterozygous FH was diagnosed using the Dutch Lipid Clinic Network (DLCN) criteria. The prevalence of individuals classified with definite or probable FH combined (DLCN criteria, .5 points) was 1/200 .Interestingly, prevalences for definite or probable FH were similar for women and men below age 60; in contrast, above age 60, more women than men were in this category. These findings suggest that many men with FH had died at an earlier age, as was also observed in a UK prevalence study.Based on extrapolations from these estimated 1/500 -1/200 prevalences, there are between 14 and 34 million individuals with FH worldwide . Furthermore, even higher prevalences are observed in subpopulations with founder effects. 2 Taken together, these data strongly suggest that FH is vastly underdiagnosed in most countries.
## Undertreatment
Hitherto, data have not been reported either on the risk of CHD, or on frequency of statin treatment, in FH individuals in a large sample from the general population not subject to ascertainment bias. Using the Copenhagen General Population Study, 10 the prevalence of CHD among definite/probable FH participants was 33%; only 48% of subjects with FH received statins. The risk of CHD was increased 13-fold (95% CI: 10-to 17-fold) among individuals with Guidance for clinicians to prevent coronary heart disease definite/probable FH not receiving statins; similar findings have been reported in FH cohort studies.Furthermore, the corresponding increase in risk for CHD among individuals with FH on statin was ten-fold (8-to 14-fold), suggesting that the dose of statin therapy provided resulted in insufficient cholesterol-lowering medication, and was introduced too late in life, when severe atherosclerosis had already developed. Other studies also support massive undertreatment of individuals with FH.Pathophysiology and genetics FH is caused by mutations in genes encoding key proteins involved in the LDL receptor endocytic and recycling pathways, leading to decreased cellular uptake of LDL and increased plasma LDL cholesterol concentrations 1. Within hepatocytes, cholesterol is recycled or synthesized de novo, with 3-hydroxy-3-methylglutaryl coenzyme A reductase being rate-limiting; statins block the activity of this enzyme. Cholesterol is packaged into apolipoprotein B-containing very low-density lipoproteins (VLDL), the intravascular precursors of LDL, which in turn transports most cholesterol from the liver to peripheral tissues. Regulated endocytosis of LDL via apolipoprotein B by peripheral cells and hepatocytes occurs through the LDL receptor and an adaptor protein (LDLRAP, alias ARH). 14 Most LDL receptors recycle, although when proprotein convertase subtilisin/kexin type 9 (PCSK9) is complexed to the LDL receptor, it short-circuits its intracellular recycling from the endosome, thereby reducing receptor numbers.
The life-threatening effects of both heterozygous and homozygous FH are related to the resulting elevation in plasma LDL cholesterol, with consequent cholesterol retention in the arterial wall and foam cell formation within the intima of arteries; such early lesions typically progress to occlusive atherosclerosis with angina pectoris and/or plaque rupture with CHD (i.e. myocardial infarction).
Heterozygous FH is caused either by heterozygous loss-offunction mutations in LDLR, heterozygous mutations in APOB that affect the LDL receptor-binding domain of apolipoprotein B, or heterozygous gain-of-function mutations in PCSK9. 14 Currently, .1200 mutations have been documented worldwide in LDLR 15 ; these affect all functional domains of the LDL receptor protein and include single-nucleotide mutations, copy number variations, and splicing mutations throughout the LDLR gene. A single mutation, Arg3500Gln, is the only common FH-related mutation in APOB, while .20 different mutations have been detected in PCSK9. Heterozygous LDLR, APOB, and PCSK9 mutations are found in .90, 5, and 1%, respectively, of heterozygous FH subjects with a causative mutation.The prevalence varies geographically.
Homozygous FH results from homozygous, or more often, from compound heterozygous mutations in either the LDLR or ARH genes.Some rare subjects are 'double heterozygotes', which means they carry mutations in two of the above-mentioned four genes, usually leading to a phenotype that is intermediate between heterozygous and homozygous FH.
## Clinical vs. mutation diagnosis
Historically, heterozygous FH was diagnosed clinically and the phenotypically most severe cases were detected, that is, those with severe LDL cholesterol elevations, premature and familial CHD, and tendon xanthomas. 1 However, with the increased understanding of the genetic causes of this disease, direct detection of mutations in the LDLR, APOB, PCSK9, and LDLRAP genes is now available in many countries. Such progress has led to the understanding that some 10-40%, depending on referral criteria, of those with a clinical diagnosis will not have a detectable causal mutation; rather, they have a clinical diagnosis of FH, but not a mutation diagnosis.There may therefore be yet other key genes implicated in this disease; alternatively, these individuals may present a polygenic basis for their LDL cholesterol elevation without contributions from any of the classical FH genes.
Conversely, genetic cascade testing from FH subjects with a detected causative mutation has shown that, while on average, relatives who carry the causative mutation have two-fold higher mean LDL cholesterol levels compared with non-carrier relatives, a significant proportion are below the clinical diagnostic cut-off and thus they have a mutation diagnosis but not a clinical diagnosis of FH.Such individuals may possess other favourable genes and/or a lifestyle that reduces the impact of the mutation, but because of their lifetime LDL cholesterol exposure, they should still be offered appropriate lipid-lowering therapy according to the LDL cholesterol targets given later.
## Whom to screen: how do we recognize index cases?
Probands (index cases) should be identified according to the following criteria:
(i) plasma total cholesterol ≥8 mmol/L (≥310 mg/dL) in an adult or adult family member(s) (or .95th percentile by age and gender for country), (ii) premature CHD in the subject or family member(s), (iii) tendon xanthomas in the subject or family member(s), (iv) sudden premature cardiac death in a family member.
For child probands, criteria #(ii) -(iv) are identical to those of adults, but criterion #(i) should be a plasma total cholesterol ≥6 mmol/L (≥230 mg/dL) in a child or child family member(s) (or .95th percentile by age and gender for country). The highest likelihood of detecting FH is in those with very high LDL cholesterol levels, tendon xanthomas, and/or premature CHD in a family member.Drawing a family pedigreeis essential to evaluate the likelihood of FH. In cases of probable or definite FH, cascade screening using LDL cholesterol measurement in the family should be conducted and the subject referred for genetic testing if available, with subsequent cascade testing in the family if a causative mutation is found. Initial family members to be tested are biological first-degree relatives, namely parents, siblings, and children. Biological seconddegree relatives including grandparents, grandchildren, uncles, aunts, nephews, nieces, and half-siblings should also be considered. 'Premature CHD' signifies CHD before age 55 in males and before age 60 in female first-degree relatives, while in second-degree relatives, the corresponding ages are 50 and 55.
## Diagnosis
Diagnosis of FH relies on five criteria: family history, clinical history of premature CHD, physical examination for xanthomas and corneal arcus, very high LDL cholesterol on repeated measurements, and/ or a causative mutation detected by molecular genetics. Secondary causes of hyperlipidaemia must be excluded by determining that liver enzymes, renal function, and thyroid hormones are normal and that there is no hyperglycaemia or albuminuria.
In addition to drawing a family pedigree, a systematic physical examination for the presence of tendon and tuberous xanthomas and corneal arcus must be performed. Sonographic evaluation of the Achilles tendons increases the rate of xanthoma detection.Although total cholesterol levels are ≥8 mmol/L (≥6 mmol/L in children), triglyceride and HDL cholesterol levels are generally unremarkable. The presence of hypertriglyceridaemia does not Guidance for clinicians to prevent coronary heart disease exclude the FH diagnosis; other reasons for hypertriglyceridaemia should however be evaluated and treated if necessary.The DLCN criteria are recommended in order to establish the clinical diagnosis of FH. Among individuals with a definite or probable diagnosis of FH (DLCN . 5), and particularly those with an obvious clinical diagnosis with xanthoma and/or high cholesterol plus a family history of premature CHD, molecular genetic testing is strongly recommended. When a causative mutation is found in the index case, a genetic test should be offered to all firstdegree relatives.
## Lifetime risk assessment and risk factors
It should be stressed that risk calculators such as the European SCORE or the US Framingham Risk Score are not appropriate for FH subjects, as such individuals are at considerably higher risk due to lifelong elevated LDL cholesterol levels. Nevertheless, whether diagnosed clinically or through a causative mutation, not all individuals with FH develop atherosclerosis and CHD to the same extent. Thus, as observed for the development of any CHD,other risk factors besides elevated LDL cholesterol act to determine the threshold for CHD, and risk factor counting is critical to assess CHD risk.Importantly, as elevated LDL cholesterol is the major problem in FH, this condition is dominated by CHD, whereas cerebrovascular disease is more common in individuals with hypertension and atherosclerosis in the lower limbs is more common among smokers.
The concept of a cumulative LDL cholesterol burdenillustrates the importance of early treatment. The cumulative LDL cholesterol burden of a 55-year-old person without FH is typically 160 mmol, a burden sufficient for CHD to develop; data derived fromalso an important cardiovascular risk factor in FH.Because elevated Lp(a) significantly enhances risk of premature cardiovascular disease in those already at extremely high risk due to FH, additional, aggressive LDL lowering with statins and other drugs should be initiated in FH individuals with high levels of Lp(a). In such individuals, and in extreme cases with severe atherosclerosis and/or CHD, lipoprotein apheresis should be considered.Asymptomatic atherosclerosis
To improve risk assessment, imaging techniques are recommended to detect asymptomatic atherosclerosis in individuals at intermediate risk in the 2012 European Society of Cardiology (ESC) Guidelines for Cardiovascular Prevention.Although subjects with FH on average are at much higher risk, risk within FH is sufficiently variable that assessment of atherosclerosis should also be considered in asymptomatic FH subjects or in those whose family history is unclear. Imaging techniques are also useful in detecting aortic valve calcification and stenosis, which is particularly relevant in homozygous FH and in individuals with elevated Lp(a) levels.Techniques available to identify asymptomatic coronary atherosclerosis include exercise electro-and echocardiography, coronary calcium score, and angiography by computed tomography. Some,but not all,guidelines underscore the value of noninvasive imaging of atherosclerosis in assessing and managing asymptomatic FH subjects.
Exercise electro-or echocardiography should be considered for risk assessment in adults with FH at very high risk; symptomatic patients should be referred urgently for cardiac specialist review. Coronary artery calcification is a surrogate marker for atherosclerosis, with the calcium score being proportional to atherosclerotic plaque burden and cardiovascular disease risk.With the latest techniques, radiation exposure is as low as 1 mSV. Coronary artery calcification and the presence and severity of atherosclerosis detected by computed tomography can identify FH subjects with increased cardiovascular risk who may need more intensive cholesterol-lowering therapy; however, absence should not preclude cholesterol-lowering treatment, because there would likely be diffuse, non-calcified plaques in such individuals. Importantly, presence of coronary calcium is not identical with presence of relevant coronary lesions, because its specificity regarding the potential presence of ≥50% stenosis is only 50%.Angiography by computed tomography is at present not recommended for risk assessment.Cascade, opportunistic, and universal screening
The most cost-effective approach for identification of new FH subjects is cascade screening of family members of known index cases. Index cases can be detected by opportunistic or targeted systematic screening in primary care guided by a family history of premature CHD and hypercholesterolaemia, and among patients aged ,55/60 in men/women with CHD in hospital settings; the DLCN criteria should be used to establish the clinical diagnosis. Universal screening of children has often been suggested but has so far only been implemented in Slovenia and at the age of 5.Cascade screening using the protocol outlined in has been found to be feasible and acceptable to subjects with FH and to physicians.To be maximally cost-effective, cascade screening should be systematic, centrally co-ordinated in a specialized centre and carried out using a combination of plasma lipid profiles and genetic testing. However, if the causative mutation is not known or genetic testing is not available, screening can be performed using the plasma lipid profile alone. Cascade testing in families with a known causative mutation has been carried out very successfully in the Netherlands over the last 15 years using trained genetic field workers. Guidance for clinicians to prevent coronary heart disease Another promising but untested targeted approach would be to screen all children for hypercholesterolaemia,for example at a time of infant immunization, and then, for children with total cholesterol .6 mmol/L (or .95th percentile), to perform 'reverse' cascade screening by testing their parents. This approach is based on the fact that LDL cholesterol values differentiate much better between mutation-negative and mutation-positive FH in children as compared with adults.However, as parents may not always be the blood relatives, DNA testing of children presents an ethical dilemma.Also, given the relatively small fraction of the population with FH, it is unclear whether such an approach would be feasible; moreover, it would likely be associated with prohibitively high cost and possibly with a high false positive rate.
## Ldl cholesterol targets
We recommend the following LDL cholesterol targets in FH, in accordance with recent ESC/EAS guidelines:
(i) children ,3.5 mmol/L (,135 mg/dL), (ii) adults ,2.5 mmol/L (,100 mg/dL), 6 (iii) adults with CHD or diabetes ,1.8 mmol/L (,70 mg/dL).These targets are for both heterozygous and homozygous FH regardless of age. However, in children and adults with homozygous FH, these values are extremely difficult to achieve with current treatments.
Due to ethical reasons, no randomized trial has been conducted documenting the benefit of lipid-lowering drug therapy specifically in FH subjects; however, treatment targets are based on large outcome lipid-lowering trials in persons without FH. 7 LDL cholesterol is the primary target of therapy and the reduction in both cardiovascular and total mortality is proportional to the degree of LDL cholesterol reduction, with every 1 mmol/L reduction being associated with a corresponding 22% reduction in cardiovascular mortality and a 12% reduction in total mortality over 5 years. 7 All untreated individuals with FH above age 40 should be considered to be at very high cardiovascular risk, as they have been exposed to elevated LDL cholesterol levels since birth.
## Treatment
All subjects with FH and their families should undergo intensive education targeting lifestyle management,including intervention on smoking, diet, and physical activity. It is imperative that smokers quit smoking, and such individuals should be referred to a specialized tobacco unit/programme when necessary. Advice to children and young adults not to start smoking is especially important.
A certified dietitian/nutritionist should support implementation of a healthy diet with the involvement of the whole family. A complete record of dietary habits must be obtained, and recommendations for a healthy diet should be individualized. Functional foods known to lower LDL cholesterol, such as plant sterols and stanols, may be considered. The main objective of the nutritional advice is to avoid overweight and to reduce the amount of food and beverages with high cholesterol, saturated fat, and transfat content. Regular physical exercise must be implemented. In adults with FH, assessment of cardiovascular function is advisable before starting any significant exercise programme.
Cholesterol-lowering drugs should be initiated immediately at diagnosis in adults and strongly considered starting at age 8-10 in childhood, along with lifestyle management. The priority for pharmacotherapy should be as follows:
Children:
(i) Statin, (ii) Ezetimibe, (iii) Bile acid-binding resin, (iv) Lipoprotein apheresis in homozygotes.
Statins for children should only be those that have been shown to be safe in this group. Adults:
(i) Maximal potent statin dose, (ii) Ezetimibe, (iii) Bile acid-binding resins, (iv) Lipoprotein apheresis in homozygotes and in treatmentresistant heterozygotes with CHD.
## Table 2 cascade testing issues in familial hypercholesterolaemia
Notification of relatives at risk of familial hypercholesterolaemia should generally not be instituted without the consent of the index case.
National and local healthcare service protocols concerning disclosure of medical information without consent should be consulted.
A proactive approach that respects privacy, justice, and autonomy is required. All material communicated to relatives and the telephone approach should be comprehensible and not cause alarm.
Pre-testing counselling should be offered to at risk family members of an index case prior to phenotypic or genetic testing.
If genetic testing detects a causative mutation, a definitive diagnosis of familial hypercholesterolaemia can be made in the tested individual particularly when the phenotype also suggests familial hypercholesterolaemia;: clinical diagnosis and mutation diagnosis).
If genetic testing does not detect a causative mutation, the diagnosis of familial hypercholesterolaemia can be excluded, except when the clinical phenotype is highly suggestive of familial hypercholesterolaemia: clinical diagnosis without mutation).
If genetic testing detects a causative mutation but the phenotype does not suggest familial hypercholesterolaemia, then a definitive diagnosis of familial hypercholesterolaemia should not be made; however, the person and family should be monitored every 2-5 years for LDL cholesterol levels: mutation without clinical diagnosis).
Genetic testing may have implications for insurance cover in certain countries.
## B.g. nordestgaard et al.
Maximal potent statin dose should be started at first consultation in adults and could be either atorvastatin 80 mg, rosuvastatin 40 mg, or pitavastatin 4 mg; simvastatin 80 mg should not be used, as this dose is associated with elevated risk of myositis and rhabdomyolysis. We recommend initiation with maximal potent statin dose in adults with FH because among subjects with FH:
(i) ,1/20 achieve the recommended LDL cholesterol targets;
(ii) most need to decrease LDL cholesterol by at least 50%; (iii) many receive statin doses insufficient to attain LDL cholesterol targets; (iv) many physicians do not uptitrate statin doses despite suboptimal treatment.
Clinical assessment of efficacy and safety is advisable 4 -6 weeks after initiating treatment.
Statins are the drug of first choice because of the huge and robust body of evidence for statin-mediated reduction in major cardiovascular events.The introduction of statins has also reduced CHD events in individuals with FH in observational studies,such that with treatment before onset of CHD, survival without CHD can be similar to that in the general population.
Despite use of the highest doses of potent statins, many subjects with FH will not achieve the LDL cholesterol target with monotherapy alone. Under these conditions, and despite lack of demonstrated clinical benefit in FH of coadministration of the cholesterol absorption inhibitor, ezetimibe, we recommend this agent as an add-on to statin therapy in view of few side effects and high compliance. The statin-ezetimibe combination will decrease LDL cholesterol by 60 -70%. For subjects at very high risk with established CHD or type 2 diabetes and with LDL cholesterol .1.8 mmol/L (.70 mg/ dL), a bile acid-binding resin (cholestyramine, colestipol, or colesevelam) as a third drug is advised. In some FH patients and in some countries, use of pure niacin (up to 3 g/day) in association with a statin, ezetimibe, or a bile acid-binding resin may be an option for additional reduction of LDL and/or Lp(a). However, niacin in the form of Tredaptive is no longer available.
In FH subjects with elevated triglycerides and low HDL cholesterol or with triglycerides .5.7 mmol/L (.500 mg/dL), maximal potent statin dose combined with fibrates can be considered, in particular fenofibrate given its satisfactory drug-drug interaction profileand effect on LDL cholesterol reduction in FH.Fenofibrate can also lower LDL cholesterol when triglycerides are normal and may be used if other drugs cannot be tolerated or are unavailable. Details of the efficacy, safety, and management of lipid-lowering drugs are described elsewhere.FH individuals with statin intolerance require specialized management to ensure that several different statins have been tested when possible and to combine (depending on the individual case) low dose of statin, ezetimibe, and resins.
In individuals with FH in extreme cases at very high cardiovascular risk with CHD, and with very high LDL cholesterol levels despite drug therapy or because of statin intolerance, adjunctive treatment with lipoprotein apheresis should be considered; this is particularly relevant for children with homozygous FH. Weekly or bi-weekly lipoprotein apheresis can decrease LDL cholesterol and Lp(a) by 50-75% and has clinical benefits in individuals with severe FH. Lipoprotein apheresis can be appropriately conducted at specialized lipid clinics, and at centres for haemodialysis and for blood transfusion, and elsewhere. Clinical thresholds for initiation of lipoprotein apheresis may vary between countries.
## Children with familial hypercholesterolaemia
In FH, elevated cholesterol is already present at birth and results in early atherosclerotic lesions. Our recommendations in children with this condition rely on intervention trials in children showing good tolerance and efficacy of statins in terms of reduction in LDL cholesterol,together with reduced progression of subclinical atherosclerosis.The optimal age range for screening is between 2 and 10 years, as determined by optimal discrimination using cholesterol measurement between children with and without FH. Currently, it is considered unreasonable to start a low-fat diet before age 2, and there are no safety data on the use of statins before age 8 -10. On the other hand, the earlier screening and treatment are initiated, the greater the benefit and compliance in the future.If total or LDL cholesterol is high, a second lipid profile after 2 or 3 months of dietary guidance, along with other biochemical analyses to exclude secondary hyperlipidaemia (see above) and other risk factors such as Lp(a), should be performed. Once hypercholesterolaemia has been detected in the child, it is important to establish its vertical transmission through a family pedigree, as awareness of the genetic nature may improve the compliance to treatment of both the parents and the child.
For diagnosis in children with one parent with FH, an LDL cholesterol level .3.5 mmol/L (.135 mg/dL) is strongly suggestive. Genetic tests should be performed in all children of parents with FH and a causative mutation, irrespective of whether or not they Guidance for clinicians to prevent coronary heart disease have high LDL cholesterol; the ethical issues involved in genetic testing of children must also be borne in mind.The absence of a positive genetic test in the parents does not exclude FH in a child with high cholesterol, and a clinical diagnosis is likely if a child with LDL cholesterol .3.5 mmol/L has one parent with a DLCN score .5. Importantly, many children in FH families are on a healthy diet and thus have lower LDL cholesterol than expected. In children, xanthomas and corneal arcus are not reliable clinical criteria as they only appear later; however, if present, they are suggestive of homozygous FH. When lifelong drug treatment is under consideration, then genetic demonstration of a causative mutation in LDLR, PCSK9 or APOB is optimal for the diagnosis of FH.
Dietary advice from a certified dietitian/nutritionist given to the parents should start after age 2 of the child. Dietary recommendations are similar to those given to adults with FH; particular caution is however needed to avoid caloric restriction (if weight is normal) and to monitor the growth curve.
Priorities for cholesterol-lowering drugs in children are given earlier. However, it is unknown at what age atherosclerotic lesions become irreversible, only short-term follow-up data are available that low dosages of statins are safe in children, and there is no longterm study evaluating the cardiovascular benefit of cholesterollowering drugs in children. Therefore, as is common in paediatrics, the therapeutic decision is based on extrapolation from adult studies and on short-term paediatric studies evaluating the safety and the efficacy of pharmacotherapy on LDL cholesterol lowering or intermediate endpoints.Thus, a registry of statin-treated children to collect meaningful follow-up data is urgently needed. Studies in children have shown this medication to be safe when started from age 8 to 10 and, based on the graphical assumption in, we therefore recommend initiation of statin therapy at age 8 -10, that is, when the diagnosis of FH is supported by a genetic test or by strong clinical arguments including LDL cholesterol .3.5 mmol/L (.135 mg/dL). As mentioned earlier, the LDL cholesterol target in children is ,3.5 mmol/L (,135 mg/dL); however, the presence of very high LDL cholesterol or additional cardiovascular risk factorsmay lower this target or the age at initiation of statin therapy. Importantly, lipoprotein apheresis should be offered in children with homozygous FH. Despite initial enthusiasm, the therapeutic potential of double heart -liver transplantation in children with homozygous FH should be considered with caution.
## Cost-effectiveness
Individuals with FH will incur costs to the healthcare system over their lifetime; if unidentified, these may include the cost of the premature CHD they are likely to suffer. If treated however, such costs will include the budget for cholesterol-lowering therapies and for the healthcare professionals that diagnose and treat them. Healthcare economic modelling has demonstrated that there are considerable overall savings in identifying and appropriately treating subjects with FH.For individuals in whom the causative mutation has been found, cascade testing of their relatives using genetic testing is highly cost-effective, as roughly 50% will have inherited the mutation. Because of their lifelong burden of LDL cholesterol accumulation, subjects with FH warrant intensive cholesterol-lowering therapy and, even if more expensive agents are used, it remains costeffective.The cost per Life Year Gained for genetic cascade testing and intensive statin therapy in FH is E3 -4000, which compares very favourably with mammography for breast cancer screening.High-intensity lipid-lowering statin therapy would lead to 101 fewer cardiovascular deaths per 1000 FH individuals treated, and when extrapolating to the 500 million population of the EU (with an estimated 1 000 000 FH subjects), roughly E4700 million could be saved from avoidance of cardiovascular events if all relatives of index cases were identified and treated optimally over a 55-year period, equating to an economy of E86 million per year.
## Novel therapies
Attainment of LDL cholesterol targets over time is imperative in FH subjects, to reduce cumulative lifetime risk. Statin therapy is often inadequate for this goal.Novel, well-tolerated therapeutic strategies as add-ons to statin therapy or as sole drugs in case of statin intolerance are therefore essential in FH. New classes of efficacious, LDL-and Lp(a)-lowering agents are currently at advanced stages of development, including therapies targeting PCSK9, anti-sense oligonucleotides targeting APOB, microsomal triglyceride transfer protein inhibitors, and cholesteryl ester transfer protein inhibitors. Additional studies on their long-term safety and efficacy, together with tolerability over time are, however, needed.
## Diagnostic and treatment summary
We recommend that most individuals with FH should be treated in primary care, preferably in a family context, while complex cases including children should be referred to specialized lipid or FH clinics. However, as FH management in primary care poses major challenges such as appropriate use of genetic testing, frequent requirement for polypharmacy, specialist knowledge of non-invasive testing, and complex organizational requirements for family cascade screening, 'shared care' between primary care and specialized lipid or FH clinics is another attractive option. In most countries, there is an urgent need for education of physicians to deal with FH, to establish networks of Lipid/FH clinics, and to establish laboratories for genetic screening and testing, preferably in a concerted manner.
# Authors' contributions
EAS Consensus Panel members were nominated by the Co-chairs M.J.C. and H.N.G. to represent expertise across FH clinical management and research from across the World. The Panel met twice, organized and chaired by M.J.C. and H.N.G. The first meeting critically reviewed the literature while the second meeting reviewed additional literature and scrutinized the first draft of the consensus paper. All panel members agreed to conception and design; contributed to interpretation of available data; suggested revisions for this document; and approved the final document before submission. |
Regulation of Sox2 and stemness by nicotine and electronic-cigarettes in non-small cell lung cancer
Background: Lung cancer is the leading cause of cancer related deaths and its incidence is highly correlated with cigarette smoking. Nicotine, the addictive component of tobacco smoke, cannot initiate tumors, but can promote proliferation, migration, and invasion of cells in vitro and promote tumor growth and metastasis in vivo. This nicotine-mediated tumor promotion is facilitated through the activation of nicotinic acetylcholine receptors (nAChRs), specifically the α7 subunit. More recently, nicotine has been implicated in promoting self-renewal of stem-like side-population cells from lung cancers. This subpopulation of cancer stem-like cells has been implicated in tumor initiation, generation of the heterogeneous tumor population, metastasis, dormancy, and drug resistance. Here we describe the molecular events driving nicotine and e-cigarette extract mediated stimulation of self-renewal of stem-like cells from non-small cell lung cancer. Methods: Experiments were conducted using A549 and H1650 non-small cell lung cancer cell lines and human mesenchymal stem cells according to protocols described in this paper. 2 μM nicotine or e-cigarette extracts was used in all relevant experiments. Biochemical analysis using western blotting, transient transfections, RT-PCR and cell biological analysis using double immunofluorescence and confocal microscopy, as well as proximity ligation assays were conducted. Results: Here we demonstrate that nicotine can induce the expression of embryonic stem cell factor Sox2, which is indispensable for self-renewal and maintenance of stem cell properties in non-small cell lung adenocarcinoma (NSCLC) cells. We further demonstrate that this occurs through a nAChR-Yap1-E2F1 signaling axis downstream of Src and Yes kinases. Our data suggests Oct4 may also play a role in this process. Over the past few years, electronic cigarettes (e-cigarettes) have been promoted as healthier alternatives to traditional cigarette smoking as they do not contain tobacco; however, they do still contain nicotine. Hence we have investigated whether e-cigarette extracts can enhance tumor promoting properties similar to nicotine; we find that they can induce expression of Sox2 as well as mesenchymal markers and enhance migration and stemness of NSCLC cells.Conclusions: Our findings shed light on novel molecular mechanisms underlying the pathophysiology of smokingrelated lung cancer in the context of cancer stem cell populations, and reveal new pathways involved that could potentially be exploited therapeutically.
# Background
Despite growing insights into the mutational events that drive the genesis of NSCLC and the development of novel therapeutic strategies, lung cancer remains the leading cause of cancer related deaths. Lung cancer accounts for more deaths than breast, prostate, and colon cancers combined [bib_ref] Cancer statistics, Siegel [/bib_ref] [bib_ref] Lung Cancer statistics, Torre [/bib_ref]. Nicotine is the major addictive component of tobacco smoke; while it is not a carcinogen and cannot initiate tumors itself, nicotine has been shown to possess a number of tumor promoting properties in multiple tumor types, both in vitro and in vivo [bib_ref] Connections of nicotine to cancer, Grando [/bib_ref] [bib_ref] Nicotine induces cell proliferation, invasion and epithelial-mesenchymal transition in a variety of..., Dasgupta [/bib_ref] [bib_ref] Nicotine stimulates angiogenesis and promotes tumor growth and atherosclerosis, Heeschen [/bib_ref] [bib_ref] Nicotine/cigarette smoke promotes metastasis of pancreatic cancer through alpha7nAChR-mediated MUC4 upregulation, Momi [/bib_ref] [bib_ref] Nicotine induces selfrenewal of pancreatic cancer stem cells via neurotransmitter-driven activation of..., Al-Wadei [/bib_ref] [bib_ref] Nicotine-mediated cell proliferation and tumor progression in smoking-related cancers, Schaal [/bib_ref] [bib_ref] Nicotine promotes tumor growth and metastasis in mouse models of lung cancer, Davis [/bib_ref]. Nicotine exerts its tumor promoting functions through the activation of nicotinic acetylcholine receptors (nAChRs), which are typically expressed on neuronal cells; they are also expressed on cells of endothelial and epithelial origin, including tumor cells [bib_ref] Recent progress in revealing the biological and medical significance of the nonneuronal..., Grando [/bib_ref] [bib_ref] Nicotine activates cell-signaling pathways through muscle-type and neuronal nicotinic acetylcholine receptors in..., Carlisle [/bib_ref] [bib_ref] Nicotinic acetylcholine receptor signaling in tumor growth and metastasis, Singh [/bib_ref]. Our lab and others have shown that nicotine can promote proliferation, angiogenesis, epithelial-to-mesenchymal transition (EMT), migration, invasion, and survival of cultured non-small cell lung cancer cells. In addition, nicotine could also promote the growth and metastasis of lung and pancreatic cancers in mouse xenograft models, primarily through the α7 subunit of nAChRs [bib_ref] Nicotine induces cell proliferation, invasion and epithelial-mesenchymal transition in a variety of..., Dasgupta [/bib_ref] [bib_ref] Nicotine stimulates angiogenesis and promotes tumor growth and atherosclerosis, Heeschen [/bib_ref] [bib_ref] Nicotine promotes tumor growth and metastasis in mouse models of lung cancer, Davis [/bib_ref] [bib_ref] Beta-Arrestin-1 mediates nicotine-induced metastasis through E2F1 target genes that modulate epithelial-Mesenchymal transition, Pillai [/bib_ref] [bib_ref] High content screening for inhibitors of protein interactions and post-translational modifications in..., Leuchowius [/bib_ref] [bib_ref] Nicotine induces inhibitor of differentiation-1 in a Src-dependent pathway promoting metastasis and..., Trevino [/bib_ref]. More recently, we have reported that nicotine can enhance the self-renewal of a subset of lung adenocarcinoma cells enriched in stem-like cell populations, through the induction of c-Kit ligand/Stem Cell Factor (SCF). SCF is known to promote self-renewal and differentiation of multiple stem cell types through the binding of its receptor, c-Kit [bib_ref] Nicotinic acetylcholine receptors induce c-kit ligand/stem cell factor and promote stemness in..., Perumal [/bib_ref] [bib_ref] The c-kit receptor, stem cell factor, and mast cells, Galli [/bib_ref] [bib_ref] The expression of c-kit protein during oogenesis and early embryonic development, Horie [/bib_ref] [bib_ref] Stem cell factor is a chemoattractant and a survival factor for CNS..., Erlandsson [/bib_ref] , and this finding reveals a novel mechanism by which nicotine might be promoting tumor progression.
Tumors were traditionally thought to be a disease of clonal origin where a single transformed cell has the ability to give rise to heterogeneous tumor cell populations, with each daughter cell having the same capacity to give rise to more tumor cells. More recently, growing evidence supports the cancer stem cell model, which indicates that cancer stem-like cells (CSCs) arise through reprogramming of adult stem cells or progenitor cells, and these cells are responsible for tumor initiation, maintenance, progression and metastasis [bib_ref] Tumoral stem cell reprogramming as a driver of cancer: theory, biological models,..., Vicente-Duenas [/bib_ref] ; in addition, the tumor stem-like cells have also been shown to contribute to drug resistance, dormancy, recurrence, and metastasis [bib_ref] EMT, cancer stem cells and drug resistance: an emerging axis of evil..., Singh [/bib_ref] [bib_ref] Anti-Cancer stem-like cell compounds in clinical development -an overview and critical appraisal, Marcucci [/bib_ref]. The model proposes that only CSCs are able to initiate tumors; these stem-like cells resemble traditional stem cells in that they are able to self-renew, divide asymmetrically, and are slow cycling [bib_ref] EMT, cancer stem cells and drug resistance: an emerging axis of evil..., Singh [/bib_ref]. Given these properties, understanding and targeting CSCs has become an important area of cancer research.
CSCs have been characterized and isolated using cell surface markers, which are differentially expressed on CSCs compared to non-stem-like, differentiated cancer cells. Various markers such as aldehyde dehydrogenase 1 (ALDH1) and CD133 positivity are effectively used for cancers such as breast, colon, brain, pancreas, head and neck [bib_ref] CD133, selectively targeting the root of Cancer, Schmohl [/bib_ref] [bib_ref] Aldehyde dehydrogenase activity is a cancer stem cell marker of tongue squamous..., Zou [/bib_ref] [bib_ref] Cancer stem cells: a step toward the cure, Boman [/bib_ref] [bib_ref] Cells with characteristics of cancer stem/progenitor cells express the CD133 antigen in..., Rutella [/bib_ref] ; however, there is no single marker ubiquitously expressed and used to identify lung cancer CSCs. A subset of tumor cells enriched in CSCs can be isolated based on their ability to efflux Hoechst 33342 dye out of their nuclei through the ABCG2 drug transporter expressed on the cell membrane of the stem-like cells, and have been termed as side-population cells, based on their distribution in flow cytometric sorting [bib_ref] EGFR/Src/Akt signaling modulates Sox2 expression and self-renewal of stemlike side-population cells in..., Singh [/bib_ref] [bib_ref] The ABC transporter Bcrp1/ABCG2 is expressed in a wide variety of stem..., Zhou [/bib_ref] [bib_ref] Side population in human lung cancer cell lines and tumors is enriched..., Ho [/bib_ref]. Our lab and others have shown that non-small cell lung cancer CSCs can be isolated using SP phenotype from cell lines as well as human tumor xenografts. Such cells were highly tumorigenic and produced highly invasive tumors in mice compared to the non-SP cells, and displayed stem-like properties such as the ability to self-renew, expression of epithelial-to-mesenchymal transition (EMT) markers as well as the classic embryonic stem cell transcription factors Sox2, Oct4, and Nanog [bib_ref] EGFR/Src/Akt signaling modulates Sox2 expression and self-renewal of stemlike side-population cells in..., Singh [/bib_ref] [bib_ref] YAP1 regulates OCT4 activity and SOX2 expression to facilitate self-renewal and vascular..., Bora-Singhal [/bib_ref]. Sox2 itself was critical to maintain self-renewal of SP cells from NSCLC cell lines, compared to Oct4 and Nanog [bib_ref] EGFR/Src/Akt signaling modulates Sox2 expression and self-renewal of stemlike side-population cells in..., Singh [/bib_ref]. Since we find that Sox2 transcription factor is required to maintain NSCLC CSC stemness and that nicotine acts to enhance stemness, we sought to determine whether this nicotine-mediated promotion of stemness occurs through the induction of Sox2. Here we report that nicotine can induce Sox2 through a Yap1/E2F1/Oct4 signaling axis.
More recently, electronic cigarettes or e-cigarettes have been marketed as a healthy alternative to traditional cigarette smoking, as they do not contain tobacco, which contains multiple carcinogens such as polycyclic hydrocarbons, tobacco specific nitrosamines, and aldehydes [bib_ref] Tobacco smoking and lung Cancer: perception-changing facts, Furrukh [/bib_ref] [bib_ref] NIH electronic cigarette workshop: developing a research agenda, Walton [/bib_ref]. While e-cigarettes do not contain tobacco carcinogens, they contain nicotine in addition to other components such as propylene glycol, glycerol, and flavorings [bib_ref] NIH electronic cigarette workshop: developing a research agenda, Walton [/bib_ref]. These devices are typically used by pressing a button which activates an internal heating coil which brings the e-liquid containing nicotine to a boil, which is then delivered as a vapor to the user [bib_ref] NIH electronic cigarette workshop: developing a research agenda, Walton [/bib_ref]. The concentration of nicotine present in e-cigarettes varies by brand and container, but is typically represented as percent nicotine by volume (NBV). How nicotine present in e-cigarettes impacts the pathophysiology and health of users remains unclear, and whether the additional components of the e-liquid might abrogate or amplify the effects of nicotine in the context of cancer has not been determined. Here we report that e-cigarette extracts can promote self-renewal in a manner similar to nicotine; further, e-cigarette extracts could induce Sox2 expression, suggesting that exposure to nicotine, either through tobacco smoke or through the use of e-cigarettes, might have deleterious effects.
# Methods
## Cell lines
Human non-small cell lung adenocarcinoma cell lines A549 and H1650 were obtained from the American Type Culture Collection (ATCC). A549 cells were maintained in Ham's F12K medium (Cellgro, Mediatech, Inc.) supplemented with 10% fetal bovine serum (Atlas Biologicals), and H1650 cells were maintained in RPMI 1640 (Gibco, Life Technologies, Thermo Fisher Scientific Inc.) containing 10% fetal bovine serum. Normal human bone marrow derived mesenchymal stem cells (hMSCs) were purchased from Lonza and maintained in their mesenchymal stem cell basal growth medium (MSCGM) designed to maintain these cells in a proliferative but not differentiated state. A549 and H1650 cell lines have been validated by ATCC and were validated again on May 25, 2016. hMSCs were pre-validated by Lonza, and only used upto passage 10.
## Generation of stable cell lines
A549 cell line was used for generating stable overexpression cells. The Sox2-core-luc (Bora-Singhal et al., 2015) and YAP1 (Addgene #15682), Oct4 (Addgene #17964) [bib_ref] Generation of human induced pluripotent stem cells from dermal fibroblasts, Lowry [/bib_ref] and E2F1 expression vectors [bib_ref] Beta-Arrestin-1 mediates nicotine-induced metastasis through E2F1 target genes that modulate epithelial-Mesenchymal transition, Pillai [/bib_ref] were transfected using FugeneHD reagent (Promega) per manufacturer's protocol. The transfected cells were selected using G418 and puromycin and maintained in Ham's F12 K medium; single colonies were selected and expanded for use in experiments.
Nicotine, E-cigarettes, and inhibitor studies (−)-nicotine (N3876; Sigma-Aldrich) or e-cigarettes (local stores) were used in these studies. A549 or H1650 cells were rendered quiescent by serum starvation in media containing 0.1% fetal bovine serum for 24 h, following which cells were stimulated with 2 μM nicotine or e-cigarette extracts for the indicated time points. For studies using nicotine or e-cigarette extracts in hMSCs, cells were maintained in stem cell media and stimulated with 2 μM nicotine or e-cigarette extracts 24 h after plating, for indicated time points. For studies using signal transduction inhibitors/anti-cancer drugs, cells were rendered quiescent by serum starvation for 24 h, were treated with inhibitors for 30 min prior to stimulation with 2 μM nicotine; the cells were maintained in the serum free medium during nicotine stimulation. The inhibitors used were AZD0530/Saracatinib (Sellekchem) at 10 μM, NVP-BKM120/Buparsilib (Chemietek) at 20 μM, GSK1120212/Trametinib (Chemietek) at 10 μM, LEE001/Ribociclib (Chemietek) at 20 μM, RRD251 at 10 μM, α-bungarotoxin (Sigma) at 10 μM, or visudyne (Sigma) at 2 μM.
Three different brands of e-cigarettes were used to demonstrate the effects; these included Fin, Njoy, and Mistic (which are referred to as E-cig 1, E-cig2, or E-cig3, respectively). E-cigarette liquid was obtained through extraction of an internal liquid-soaked sponge within the devices for E-cig 1 and 2, or by syringe extraction for E-cig 3. E-cig 1, 2, and extracts were 1.6% nicotine by volume (NBV) or 16 mg/ml, 1.5% NBV or 15 mg/ml, and 1.8% NBV or 18 mg/ml respectively as indicated on the manufacturer's packaging. Molarity of extracts from each brand was calculated based on the molecular weight of nicotine of 162.23, and the working concentration of 2 μM was achieved by serial dilutions of 1:10, 1:9, or 1:11 for E-cig 1, 2, or 3 respectively, to achieve 10 mM, then diluted 1:50 for a final concentration of 2 μM. siRNAs and antibodies siRNAs used were purchased from Santa Cruz Biotechnology including Oct3/4 (sc36123), TEF4/Tead2 (sc45232), α7 nAChR (sc42532), E2F1 (sc29297), c-Src (sc44250), Sox2 (sc38408), Yap1 (sc38637), c-Yes (sc29860), and β-arr-1 (sc29741). Antibodies used for western blot against Sox2 (3579s), p-Src (2101s), p-AKT (9018p), pan-AKT (C67E7), Oct4 (2750s), p-ERK1/2 (9101s), and total ERK1/2 (9102s) were purchased from Cell Signaling Technologies; against c-Src (05-184) from EMD Millipore; against E2F1 (sc251) from Santa Cruz Biotechnology; against Yap1 (53-161) from Abnova, α7 nAChR (ab23832 and ab10096) from Abcam; and Actin from Sigma Aldrich (A1978).
Antibodies used for chromatin immunoprecipitation (ChIP) assays included E2F1 (sc193), E2F2 (sc633), E2F3 (sc879), E2F4 (sc1082), E2F5 (sc999), Rb (sc50), from Santa Cruz Biotechnology; Yap1 (ab56701), H327Kme3 (ab9045), and H327Kme1 (ab6002) from Abcam; a rabbit anti-mouse secondary antibody from Pierce was used as a negative control.
Antibodies used for immunofluorescence included Sox2 (3579s) from Cell Signaling Technologies; ZO-1 (339100) from Invitrogen; E2F1 (sc193), Crm1 (sc5595), and E-cadherin (sc8426) from Santa Cruz Biotechnology; α7 nAChR (ab10096) Yap1 (ab56701) from Abcam.
## Chip-pcr experiments
Chromatin immunoprecipitation assays were conducted using previously described protocols [bib_ref] Chromatin immunoprecipitation assays: analyzing transcription factor binding and histone modifications in vivo, Pillai [/bib_ref] [bib_ref] ARRB1-mediated regulation of E2F target genes in nicotine-induced growth of lung tumors, Dasgupta [/bib_ref]. Interactions of the proteins with specific regions of the Sox2 promoter were detected by PCR amplification using the following primer sequences:
## Transient transfections and luciferase assays
Cells transfected in Opti-MEM medium (Gibco, Life Technologies) using Fugene HD (Promega) transfection reagent following the manufacturer's protocol. The mutSox2-core-luc construct containing a mutated Oct4 binding site was previously generated using Quikchange Lightening multi-site-directed mutagenesis kit (Agilent Technologies), as previously reported from our lab [bib_ref] YAP1 regulates OCT4 activity and SOX2 expression to facilitate self-renewal and vascular..., Bora-Singhal [/bib_ref].
To confirm the role of E2Fs in regulation of the Sox2 promoter, the 7 E2F consensus binding sites present in the 500 bp region upstream of TSS (where TSS = 0) were mutated. This was done by mutating each of the 4 bp CGCG consensus sites to AATT within the Sox2-core-promoter at the following positions: -37 bp through -40 bp, − 50 bp through -53 bp, − 107 bp though -110 bp, − 119 bp through -122 bp, − 336 bp through -339 bp, − 361 bp through -364 bp, and -476 bp through -479 bp. Mutation of the E2F sites was outsourced to Genscript USA, Inc., and the E2F-mutant Sox2-core promoter was then cloned into pGL3 expression vector by our lab, for use in transient transfection experiments. The expression vectors used were pcDNA3-HA-E2F1, pcDNA3-E2F2, pcDNA3-E2F3, pcDNA3-E2F4, pcDNA3-E2F5, and Yap1 (Addgene #18978). Empty vector pcDNA3 was used as a control. Luciferase assays were conducted 24 to 48 h after transfection per manufacturer's protocol using the Dual Luciferase Assay system (Promega). Results are reported as relative luciferase activity (RLA) based on the ratio of RLUs1 (firefly luciferase) to RLUs2 (Renilla luciferase: normalization control) values as measured on a Turner Biosystems luminometer.
siRNA transfections and quantitative real-time PCR Cells were transfected in Opti-MEM (Gibco Life Technologies) with 100 pmol of siRNAs using Oligofectamine reagent (Invitrogen) as per the manufacturer's protocol. Media was replaced by complete medium containing 10% FBS 4-6 h after transfection. RNA was isolated using Qiagen RNEasy miniprep kit (Hilden, Germany) according to manufacturer's protocol. First strand cDNA was synthesized using Bio-Rad iScript cDNA synthesis kit (Hercules, CA). mRNA expression was assessed using qRT-PCR (Bio-Rad CFX96 Real Time System) and data were analyzed using the CFX96 software. RT-primers used were as follows:
GAPDH
[formula] Yap1(F): 5′-CCCAAGACGGCCAACGTGCC-3′; Yap1(R): 5′-ACTGGCCTGTCGGGAGTGGG-3′; Sox2(F): 5′ -GGGAAATGGGAGGGGTGCAAAAG A-3′; Sox2(R): 5′- TTGCGTGAGTGTGGATGGG ATTGG-3′; ZEB1(F): 5'-AGCAGTGAAAGAGAAGGGAATGC-3′; ZEB1(R): 5'-GGTCCTCTTCAGGTGCCTCAG-3′; ZEB2(F): 5'-ATCTGCTCAGAGTCCAATGCAGCA C-3′; ZEB2(R): 5'-AACAGTATTGTCCACAATC TGTAG-3′. [/formula]
Data was normalized using GAPDH as an internal control, and fold change was determined using the 2 -ΔΔCT method.
## Lysate preparation and ip/western blotting
Cell lysates were prepared and processed for western blotting as described in our previous work [bib_ref] YAP1 regulates OCT4 activity and SOX2 expression to facilitate self-renewal and vascular..., Bora-Singhal [/bib_ref] [bib_ref] ARRB1-mediated regulation of E2F target genes in nicotine-induced growth of lung tumors, Dasgupta [/bib_ref]. Protein was detected using ECL reagent from GE Healthcare or Pierce Biotechnology according to standard protocols; actin was used as a control.
For co-immunoprecipitation assays, 200 μg of total protein lysate from A549 and H1650 cells were incubated with 4 μg of indicated antibodies. An equal amount of non-specific IgG from rabbit or mouse serum (Sigma-Aldrich) was used as a negative control. The interacting proteins were detected by western blotting.
## Immunofluorescence analysis and confocal microscopy
Immunofluorescence assays were conducted as previously described [bib_ref] YAP1 regulates OCT4 activity and SOX2 expression to facilitate self-renewal and vascular..., Bora-Singhal [/bib_ref] [bib_ref] Tank binding kinase 1 is a centrosome-associated kinase necessary for microtubule dynamics..., Pillai [/bib_ref]. Cells were visualized with a DM16000 inverted Leica TCS SP5 tandem scanning confocal microscope at 630× or 1890× magnification.
## Proximity ligation assays (pla)
PLA studies were conducted as previously described using Duolink assay system (Sigma-Aldrich) [bib_ref] High content screening for inhibitors of protein interactions and post-translational modifications in..., Leuchowius [/bib_ref] [bib_ref] YAP1 regulates OCT4 activity and SOX2 expression to facilitate self-renewal and vascular..., Bora-Singhal [/bib_ref] [bib_ref] Tank binding kinase 1 is a centrosome-associated kinase necessary for microtubule dynamics..., Pillai [/bib_ref]. The images were taken using Leica TCS SP5 confocal microscope (Leica Microsystems) at 630x and 1890x magnification.
## Isolation of side-population (sp) cells and self-renewal
SP cells were isolated from heterogenous cell populations using flow cytometry based on Hoechst 33342 dye efflux, and were then plated for self-renewal assays on low-adherence plates in stem-cell selective media, using protocols described in detail earlier [bib_ref] EGFR/Src/Akt signaling modulates Sox2 expression and self-renewal of stemlike side-population cells in..., Singh [/bib_ref] [bib_ref] Gli1-mediated regulation of Sox2 facilitates self-renewal of stem-like cells and confers resistance..., Bora-Singhal [/bib_ref]. For experiments involving nicotine or e-cigarette extracts, these were added directly into stem-cell media at the time of plating. For depletion experiments, cells were transfected using siRNA and SP cells were isolated 48 h later.
## Wound healing assays
Wound healing or scratch assays were conducted as previously described [bib_ref] Nicotine induces cell proliferation, invasion and epithelial-mesenchymal transition in a variety of..., Dasgupta [/bib_ref] [bib_ref] ID1 facilitates the growth and metastasis of nonsmall cell lung cancer in..., Pillai [/bib_ref]. Images were taken every 24 h for 48 h, using EVOS FL microscope system (Life Technologies) at 10× magnification.
# Statistical analysis
All data have been statistically analyzed using Microsoft Office Excel 2010 (Microsoft Corporation, Redmond, WA). The data presented here is with ± standard deviation (SD) values derived from three independent experiments unless otherwise stated. The statistical comparisons between the groups were carried out by unpaired two tailed Student's t-test or one-way ANOVA to calculate the p value for statistical significance. *p < 0.05, **p < 0.01 and ***p < 0.001.
# Results
## Nicotine and e-cigarette extracts enhance self-renewal of sp cells and promote emt
We had previously found that nicotine could enhance the self-renewal ability of SP cells, so we next determined whether this was observed with e-cigarette extracts. SP cells were isolated from A549 and H1650 cell lines by FACS, and SP and MP cells were plated separately on low adherence plates in stem cell selective media. Nicotine or extracts from each of the three brands of e-cigarettes was added to the media in the corresponding wells at the time of plating. E-Cigarette extracts were added equivalent to 2 μM nicotine, as described in the Materials and Methods. After 10 days cells were imaged to assess sphere formation as an indication of self-renewal using microscopy. It was found that e-cigarette extracts could enhance sphere formation and self-renewal in a manner similar to nicotine, in both A549 and H1650 cells . Our earlier studies had demonstrated that Sox2 was necessary for the self-renewal of SP cells from lung adenocarcinoma cell lines. To examine if Sox2 was also needed for nicotine and E-cigarette mediated enhancement of self-renewal, Sox2 was depleted using siRNA. While transfection of a control siRNA did not affect the self-renewal of SP cells in response to nicotine or E-cigarette extracts, depletion of Sox2 abolished sphere formation in both the cell lines .
Epithelial-to-mesenchymal transition (EMT) is a normal process during development that is frequently activated during tumor progression allowing for a more migratory and invasive phenotype [bib_ref] New insights into the crossroads between EMT and Stemness in the context..., Fabregat [/bib_ref] [bib_ref] Regulatory roles of Dclk1 in epithelial Mesenchymal transition and Cancer stem cells, Chandrakesan [/bib_ref]. Growing evidence suggests that activation of EMT induces the acquisition of stem cell properties in epithelial cells and that the induction of EMT and emergence of CSCs is strongly linked [bib_ref] New insights into the crossroads between EMT and Stemness in the context..., Fabregat [/bib_ref] [bib_ref] Regulatory roles of Dclk1 in epithelial Mesenchymal transition and Cancer stem cells, Chandrakesan [/bib_ref]. We have previously reported that, in addition to enhancing CSC properties, nicotine can induce a number of factors involved in EMT including mesenchymal markers such as vimentin, fibronectin, Zeb1, and Zeb2; further, it could suppress epithelial markers such as E-cadherin and disrupt the tight junction protein ZO-1, thereby enhancing cell motility, cell migration and invasion [bib_ref] Nicotine induces cell proliferation, invasion and epithelial-mesenchymal transition in a variety of..., Dasgupta [/bib_ref] [bib_ref] Beta-Arrestin-1 mediates nicotine-induced metastasis through E2F1 target genes that modulate epithelial-Mesenchymal transition, Pillai [/bib_ref]. Since we also find e-cigarette extracts are capable of enhancing CSC properties, we next interrogated whether e-cigarette extracts could enhance EMT phenotypes, perhaps as a precursor to the acquisition of CSC phenotype. Immunofluorescence assays demonstrated that e-cigarette extracts could disrupt ZO-1 tight junction protein with a concordant decrease in ZO-1 staining overall, and could additionally reduce levels of E-cadherin in A549 cells and f; red fluorescence). At the mRNA level, we found that each of the three e-cigarette brands could induce the mesenchymal markers vimentin, fibronectin, Zeb1, and Zeb2 . Wound healing assays additionally demonstrated that e-cig 1 could enhance cell migration to a certain extent ; the results are quantified in . Overall, this data suggests that e-cigarette extracts induce EMT phenotypes in a manner similar to what was previously observed with nicotine.
## Nicotine and e-cigarette extracts enhance sox2 expression
Our lab had reported that the embryonic stem cell transcription factor Sox2 is indispensable for the self-renewal of SP cells from lung adenocarcinomas [bib_ref] EGFR/Src/Akt signaling modulates Sox2 expression and self-renewal of stemlike side-population cells in..., Singh [/bib_ref] , while Oct4 and Nanog had a relatively lesser role. Further, stimulation with nicotine could enhance the self-renewal of SP cells in a nicotinic acetylcholine receptor dependent manner [bib_ref] Nicotinic acetylcholine receptors induce c-kit ligand/stem cell factor and promote stemness in..., Perumal [/bib_ref]. We first sought to elucidate whether nicotine enhances self-renewal through induction of Sox2 in A549 and H1650 lung adenocarcinoma cell lines, and whether e-cigarette extracts had similar effects compared to nicotine. Cells were serum starved for 24 h and stimulated with 2 μM nicotine; induction of Sox2 was examined by immunofluorescence microscopy. Nicotine or extracts from each of three brands of e-cigarettes, E-cig 1, E-cig 2, and E-cig 3, could induce Sox2 after 21 h of nicotine stimulation in A549 lung cancer cells. Similar induction was also observed in human mesenchymal stem cells (hMSCs), suggesting that the induction of Sox2 by nicotine is not restricted to cancer cells [fig_ref] Figure 2 a: An immunofluorescence experiment showing the induction of Sox2 by nicotine and E-cigarette... [/fig_ref].
Western blot and qRT-PCR experiments confirmed that E-cig 1, E-cig 2, or E-cig 3 could induce Sox2 mRNA and protein expression in A549 cells after 21 h of stimulation, at levels comparable to nicotine [fig_ref] Figure 2 a: An immunofluorescence experiment showing the induction of Sox2 by nicotine and E-cigarette... [/fig_ref] and c). To further confirm this induction, a Sox2 core promoter luciferase construct (Sox2-luc) was transiently transfected into cells, followed by nicotine and e-cigarette extract stimulation. This construct contained the region -530 bp upstream through + 238 of the transcription start site (TSS) (where TSS = 0) on the Sox2 gene promoter driving the luciferase reporter. Nicotine and each of the three brands of e-cigarette extracts induced Sox2-core-luc activity after 21 h in A549 and H1650 cells [fig_ref] Figure 2 a: An immunofluorescence experiment showing the induction of Sox2 by nicotine and E-cigarette... [/fig_ref].
To determine the time course of the nicotine-mediation induction of Sox2, A549 cells were stimulated with nicotine for 18, 24, 48, 72, 96, and 120 h and Sox2 expression was assessed by western blotting a Nicotine and E-cigarette extracts promote the self-renewal of stem-like SP cells from A549 cells; similar results were obtained on H1650 cells (b). SP represents side population and MP represents main population cells. Depletion of Sox2 using a siRNA abrogates the self-renewal of SP cells from A549 cells (c) or H1650 cells (d). Nicotine and E-cigarette extracts induce EMT-like changes in A549 cells. Immunofluorescence experiments using an antibody to ZO1 shows that nicotine and E-cigarette extracts reduce the tight junctions effectively (red fluorescence); CRM1 and DAPI are used to visualize the cells (e). Similar results were observed with E-cadherin (f). g A RT-PCR experiment shows the induction of mesenchymal markers vimentin and fibronectin as well as ZEB1 and ZEB2. The data presented here is with ± standard deviation (SD) values derived from three independent experiments. The statistical comparisons between the groups were carried out by unpaired two tailed Student's t-test. h-i Nicotine and E-cigarette extracts promote the migration of A549 cells on plastic, as seen by a wound-healing assay (h) and the healing of the wound was represented graphically as % wound area over time (i). The statistical comparison between the groups was carried out by unpaired two tailed Student's t-test and qRT-PCR. 2 μM nicotine could induce Sox2 protein as well as mRNA at 18 and 24 h, an effect that was diminished by 48 h and completely abolished by 72 h [fig_ref] Figure 2 a: An immunofluorescence experiment showing the induction of Sox2 by nicotine and E-cigarette... [/fig_ref]. The disappearance of Sox2 by 72 h of nicotine treatment could be due to the cells acquiring sufficient downstream targets to maintain stemness and may no longer require Sox2. Since we saw the peak induction of Sox2 occurring after 18-24 h of nicotine stimulation, we used 21 h as the time point for the majority of the remaining experiments.
Nicotine-mediated induction of Sox2 occurs through a nAChR-Yap1-E2F1 axis Given our finding that nicotine induces Sox2, and since Sox2 is necessary for the self-renewal of SP cells, we next sought to elucidate the mechanism by which nicotine induces expression of Sox2. Previous studies have shown that nicotine exerts a number of tumor promoting properties such as proliferation, migration, and invasion through the binding to and activation of α7 nAChR, and that the nicotine-mediated activation of these receptors results in the transcriptional activity of E2F1 transcription factor [bib_ref] Nicotine induces cell proliferation, invasion and epithelial-mesenchymal transition in a variety of..., Dasgupta [/bib_ref] [bib_ref] Nicotine promotes tumor growth and metastasis in mouse models of lung cancer, Davis [/bib_ref] [bib_ref] Beta-Arrestin-1 mediates nicotine-induced metastasis through E2F1 target genes that modulate epithelial-Mesenchymal transition, Pillai [/bib_ref] [bib_ref] The Rb-E2F transcriptional regulatory pathway in tumor angiogenesis and metastasis, Schaal [/bib_ref]. We have also previously reported that Yes Associated Protein 1 (Yap1), which is a transcriptional co-activator and the major effector of the Hippo signaling pathway, binds to Oct4 embryonic stem cell transcription factor on the Sox2 promoter to regulate both Sox2 expression as well as the stem-like functions of cancer stem-like cells. Induction of Sox2 by Yap1 occurred independent of TEAD2 transcription factor, which is a well-documented binding partner and mediator of Yap1 functions [bib_ref] YAP1 regulates OCT4 activity and SOX2 expression to facilitate self-renewal and vascular..., Bora-Singhal [/bib_ref]. To evaluate the relative contributions of these proteins to the induction of Sox2, we depleted α7 nAChR, E2F1, Yap1, Oct4, or Tead2 using siRNA in A549 cells, stimulated with nicotine, and conducted western blot experiments. It was found that depletion of α7 nAChR, E2F1, or Yap1 could abrogate nicotine-mediated induction of Sox2 at the protein level . To verify that the siRNAs used could reduce the expression of the indicated factors, transient transfection experiments were conducted in untreated A549s cells demonstrating that siRNAs targeting α7 nAChR, E2F1, Yap1, Oct4, Sox2, or TEAD2 effectively reduced their corresponding protein levels, as seen by western blotting (Additional file 1: A-E). This led us to hypothesize that perhaps nicotine-mediated activation of α7 nAChR elevates Yap1's binding to E2F1 on the Sox2 promoter to induce its expression. Based on our results, we conclude that nicotine mediated activation of α7 nAChR leads to the downstream signaling events resulting in the induction of Sox2 and self-renewal.
Sox2 gene expression is regulated by E2F1, and nicotine enhances E2F1 binding to Sox2 promoter
To determine whether E2F transcription factors can bind to the Sox2 gene promoter and contribute to its expression, Genomatix MatInspector Analysis Software was used to analyze a 1500 bp region upstream of the Sox2 transcriptional start site for potential E2F binding sites; 19 predicted E2F binding sites were detected on this promoter region . Primers were designed spanning four different regions of the Sox2 promoter a Depletion of E2F1, YAP1 or α7 nAChR by siRNA prevents the nicotine-mediated induction of Sox2. Depletion of Oct4 or TEAD2 did not have any impact on the induction. b A schematic showing the location of the E2F binding sites that were tested by ChIP assays; the location of forward and reverse primers for each site are indicated by arrows. c A ChIP assay shows the association of E2F1 with all the tested binding sites; E2Fs 2 and 3 were mainly associated with the binding site spanned by primers F2 and R2. There was no E2F associated with c-Fos promoter. d Nicotine stimulation induces the association of E2F1 with sites spanned by primers F1R1 and F2R2; surprisingly, Rb could also be detected on site F1R1. e A transient transfection experiment conducted on A549 cells showing the induction of Sox2-Luc reported by E2F family members. The data represented is with ± standard deviation (SD) values derived from three independent experiments. The statistical comparison was carried out by unpaired two tailed Student's t-test. f YAP1 and E2F1 can induce the Sox2 promoter, and can show an additive effect in transient transfection experiments. The data represented is with ± standard deviation (SD) values derived from three independent experiments. The statistical comparison was carried out by one-way ANOVA. g Nicotine and E-cigarette extracts induce YAP1 levels in A549 cells and human mesenchymal stem cells, as seen by an immunofluorescence experiment and chromatin immunoprecipitation assays were carried out to verify if E2Fs bound to the predicted regions. E2F1, and to a lesser extent E2F2 and E2F3, could bind to a region − 104 through -259 bp upstream of TSS, denoted as F1R1, in A549 cells; E2F1, E2F2, and E2F3 could also bind to a region could bind to a region − 266 through − 497 bp upstream of TSS denoted as F2R2; E2F1 and to a lesser extent E2F2 and E2F3 could bind could to a region − 1078 through − 1233 bp upstream of TSS denoted as F3R3; and E2F1 and to some extent E2F2 could bind to a region − 1414 through -1433 bp upstream of TSS denoted as F4R4. An irrelevant IgG antibody was used as a negative control for IP in these experiments, which showed no amplification for any region, and c-Fos was used as a negative promoter control . A549 cells were serum starved and treated with nicotine, and ChIP assays were conducted to assess whether the association of E2F1 with the Sox2 promoter was responsive to nicotine stimulation. It was found that nicotine could enhance recruitment of E2F1 to the Sox2 promoter after 18 h of stimulation, and both the F1R1 region − 104 through -259 bp and the F2R2 region − 266 through − 497 bp upstream of TSS .
Given that E2Fs could bind to the Sox2 promoter, transient transfection experiments were conducted to determine whether E2Fs could regulate Sox2 expression; a luciferase reporter driven by a Sox2 core promoter was used for this purpose. It was found that the transcriptionally active E2F family members (E2F1-3) could induce Sox2-core promoter-luciferase activity by 2.3, 1.8, and 1.8-fold respectively, while little or no effect was seen with the inactive E2F4 and E2F5 (0.7 and 1.0 fold respectively) in A549 cells . These results show that E2F1 can induce Sox2, and nicotine is possibly inducing Sox2 through this transcription factor.
Since we had previously reported that Yap1 could induce Sox2 expression and here we find that E2F1 also induces Sox2 expression, we next sought to determine whether Yap1 and E2F1 cooperatively induce the Sox2 promoter. Transient transfection experiments showed that Yap1 or E2F1 alone could induce Sox2-core-luc activity by 1.8 and 2.7-fold respectively; when both were transfected together, they showed an 6.2 fold induction of Sox2-core-luc activity, indicating a co-operative effect .
## Nicotine and e-cigarette extracts enhance yap1 expression
Interestingly, we observed that the expression of Yap1 was also enhanced upon nicotine stimulation. Indeed, it has been reported that nicotine could induce Yap1 in certain cell types. Additional immunofluorescence microscopy experiments in A549 cells showed that nicotine and all three brands of e-cigarette extracts induced Yap1 levels after 21 h. Nicotine could induce Yap1 expression in hMSCs to a lesser extent as well, demonstrating that the induction is not specific to NSCLC cells .
Binding of Yap1 to E2F1 is enhanced by nicotine or ecigarette extracts
It has been suggested that Yap1 could co-operate with E2F1-mediated transcription programs in certain cell types [bib_ref] Yap1 activation enables bypass of oncogenic Kras addiction in pancreatic cancer, Kapoor [/bib_ref] [bib_ref] YAP inhibition restores hepatocyte differentiation in advanced HCC, leading to tumor regression, Fitamant [/bib_ref]. Given that Yap1 can regulate Sox2 expression, and since E2F1 was found to associate with the Sox2 promoter, we next examined whether Yap1, could co-localize or interact with E2F1 and if this was sensitive to nicotine or e-cigarette extract stimulation. Double immunofluorescence experiments showed that Yap1 and E2F1 co-localized in A549 cells; treatment with nicotine or extracts from each of the three e-cigarette brands enhanced this interaction . To further confirm the physical interaction of Yap1 and E2F1 proteins, immunoprecipitation-western blot experiments were conducted; these results showed that Yap1 could physically interact with E2F1 in untreated A549 and H1650 cell lines . Yap1 could be detected in E2F1 immunoprecipitates, by western blotting, and vice versa. An irrelevant IgG was used as a negative control in both cases to establish the specificity of the assay .
Similarly, proximity ligation assays (PLA) [bib_ref] High content screening for inhibitors of protein interactions and post-translational modifications in..., Leuchowius [/bib_ref] [bib_ref] Analysis of protein interactions in situ by proximity ligation assays, Koos [/bib_ref] [bib_ref] Direct observation of individual endogenous protein complexes in situ by proximity ligation, Soderberg [/bib_ref] [bib_ref] Characterizing proteins and their interactions in cells and tissues using the in..., Soderberg [/bib_ref] [bib_ref] Proximity ligation assays: a recent addition to the proteomics toolbox, Weibrecht [/bib_ref] were conducted to determine whether Yap1 and E2F1 proteins exist in close proximity within the cell, and at what point in time this occurs. PLA is able to detect proteins within a 40 nm range of one another, which generally indicates a direct physical interaction, and this can be visualized as individual foci by confocal microscopy. These assays showed that Yap1 and E2F1 interacted with one another, and the interaction was enhanced by nicotine as well as extracts from each of the three brands of e-cigarettes. We further find that this effect increased from 6 to 12 to 18 to 24 h of nicotine or e-cigarette extract stimulation, coinciding with the time points at which Sox2 is induced .
## Nicotine-mediated induction of sox2 occurs through src kinases
The Yap1 protein is known to be oncogenic as it promotes growth while inhibiting apoptosis, and is amplified or overexpressed in a number of cancer types [bib_ref] Yes-associated protein (YAP65) is a proline-rich phosphoprotein that binds to the SH3..., Sudol [/bib_ref] [bib_ref] The hippo signaling pathway coordinately regulates cell proliferation and apoptosis by inactivating..., Huang [/bib_ref] [bib_ref] Transforming properties of YAP, a candidate oncogene on the chromosome 11q22 amplicon, Overholtzer [/bib_ref]. It is named Yes-associated-Protein 1 due to its ability to associate with SH3 domains of Src family tyrosine kinases, which include Src, Yes, and Fyn [bib_ref] Yes-associated protein (YAP65) is a proline-rich phosphoprotein that binds to the SH3..., Sudol [/bib_ref]. Typically in quiescent cells, Yap1 is phosphorylated downstream of the tumor suppressive Hippo pathway, by Lats1 and Lats2 kinases; phosphorylation leads to its sequestration in the cytoplasm by 14-3-3 proteins resulting in proteasomal degradation, thereby preventing its nuclear import [bib_ref] Regulation of the hippo-YAP pathway by G-proteincoupled receptor signaling, Yu [/bib_ref]. Conversely, Yap1 is phosphorylated by Yes1 in embryonic stem cells, leading to its activation. Multiple studies have now reported that Yap1 can function independent of the Hippo pathway, and it has further been shown that Yap1 is a direct phosphorylation target of Src in a number of cancer cell lines, independent of the canonical Hippo pathway [bib_ref] Nicotine induces selfrenewal of pancreatic cancer stem cells via neurotransmitter-driven activation of..., Al-Wadei [/bib_ref] [bib_ref] TGF-beta targets the hippo pathway scaffold RASSF1A to facilitate YAP/ SMAD2 nuclear..., Pefani [/bib_ref] [bib_ref] A gp130-Src-YAP module links inflammation to epithelial regeneration, Taniguchi [/bib_ref]. Our lab has previously reported that when nicotine binds to α7 nAChR, the scaffolding protein β-arrestin-1 (β-arr-1) is recruited and activates Src kinase (p-Src), which mediates a number of downstream pathways, including E2F1 transcriptional activity [bib_ref] Beta-Arrestin-1 mediates nicotine-induced metastasis through E2F1 target genes that modulate epithelial-Mesenchymal transition, Pillai [/bib_ref] [bib_ref] ARRB1-mediated regulation of E2F target genes in nicotine-induced growth of lung tumors, Dasgupta [/bib_ref]. We have additionally reported that inhibition of the EGFR pathway including Src and PI3K could strongly inhibit Sox2 expression, thereby suppressing the self-renewal of SP cells; and this occurred in a β-arrestin-1 dependent manner [bib_ref] EGFR/Src/Akt signaling modulates Sox2 expression and self-renewal of stemlike side-population cells in..., Singh [/bib_ref] [bib_ref] β Arrestin-1 and Mcl-1 modulate self-renewal growth of cancer stem-like side-population cells..., Singh [/bib_ref]. In this context, we next sought to determine if nicotine-mediated regulation of Sox2 by Yap1 and E2F1 was a result of the upstream activation of Src or Yes kinases. Initial studies were carried out in A549 cells using inhibitors to Src/Yes/Fyn (Saracatinib), PI3K (Buparsilib), MEK1/2 (Trametinib), CDK4/6 (Ribociclib), Rb-Raf interaction (RRD251), α7 nAChR (α-bungarotoxin), and Yap1 (visudyne), to determine which molecules downstream of nAChR facilitate Sox2 induction by nicotine. Cells were treated with inhibitors as described in materials and methods, stimulated with nicotine for 21 h, and Sox2 expression was assessed by western blotting. It was a Nicotine and E-cigarette extracts promote the co-localization of E2F1 and YAP1 in A549 cells, as seen by a double immunofluorescence experiment followed by confocal microscopy. b An immunoprecipitation-western blot experiment showing the association of YAP1 with E2F1 in A549 and H1650 cells; the immunoprecipitation was conducted by an E2F1 antibody followed by western blotting with a YAP1 antibody. c An IP-western blot experiment in the reverse direction, where a YAP1 antibody was used for IP followed by western blotting with an E2F1 antibody, further confirms the association of YAP1 with E2F1. IP with an antibody to Rb was used as a positive control in this experiment. d A proximity ligation assay showing the enhanced association of YAP1 with E2F1 in A549 cells upon treatment with nicotine or E-cigarette extracts. The interaction was maximal at 24 h found that treatment with inhibitors to Src/Yes/Fin, PI3K, CDK4/6, or Yap1 could significantly abrogate nicotine-mediated induction of Sox2 .
To further confirm whether Src or Yes played a role, we conducted depletion experiments in A549 cells using siRNA targeting β-arr-1, Src, or Yes, following which cells were stimulated with nicotine for 21 h; subsequently, protein levels of Sox2 were assessed by western blotting. Depletion of β-arr-1, Src, or Yes could reduce the induction of Sox2 following nicotine treatment . Western blotting was conducted to confirm that siRNA targeting β-arrestin-1, c-Src, or c-Yes1 could deplete the corresponding proteins; it was found that siRNA transfections resulted in reduced levels of protein expression, as expected (Additional file 1: .
Double immunofluorescence experiments showed that depletion of Src using siRNA or inhibition using Saracatinib could reduce the co-localization of Yap1 and E2F1 in response to nicotine stimulation after 21 h ; there was a reduction in the levels of Yap1 as well, in both the cases. These results were recapitulated in PLA experiments conducted in the same manner . These results were supported by an IP-western blot experiment . Together, this data suggests that Src family kinases act upstream of Yap1 and E2F1 to regulate induction of Sox2 expression in response to nicotine stimulation. a Treatment with selected inhibitors prevents nicotine-mediated induction of Sox2; the Src family kinases appear to be especially vital for the induction. b Depletion of Src, Yes or β-arrestin-1 reduces the nicotine-mediated induction of Sox2 in A549 cells (c) Treatment with a siRNA to Src reduces YAP1 levels and its interaction with E2F1; similar results were obtained upon treatment with a Src inhibitor (d). e Proximity ligation assay showing that depletion of Src abrogates the interaction of YAP1 with E2F1; similar results were obtained upon treatment with a Src inhibitor (f). g An IP-western blot experiment showed that depeltion of Yes and perhaps Src reduces the icotine-mediated interaction of YAP1 with E2F1
## Oct4 contributes to nicotine-mediated induction of sox2
We had previously shown that Yap1 was elevated in cancer stem-like cells from NSCLC and was necessary for their self-renewal and ability to form angiogenic tubules; and these effects of Yap1 were mediated through the induction of Sox2 [bib_ref] YAP1 regulates OCT4 activity and SOX2 expression to facilitate self-renewal and vascular..., Bora-Singhal [/bib_ref]. Further, depletion of Yap1 resulted in the inability of NSCLC cell lines to form tumors and metastasize in murine orthotopic lung implantation models, and the overexpression of Sox2 could rescue this effect [bib_ref] YAP1 regulates OCT4 activity and SOX2 expression to facilitate self-renewal and vascular..., Bora-Singhal [/bib_ref]. While this supports an important role for Yap1 in modulation of stem-like functions through the regulation of Sox2, it led us to question whether Yap1 interaction with Oct4 might also play a role in nicotine-mediated induction of Sox2, in addition to Yap1-E2F1 interaction. PLA experiments were conducted to determine whether nicotine or e-cigarette extracts had an effect on the interaction of Yap1 with Oct4 in A549 cells . These assays showed that Yap1 and Oct4 existed in close proximity of one another, and the interaction was enhanced by nicotine as well as extracts from each of the three brands of e-cigarettes similar to what we found for Yap1-E2F1 interaction.
We next sought to determine whether Oct4 in addition to E2F1, had a role in nicotine induction of Sox2 itself. This was first assessed by determining whether depletion of Oct4 protein in addition to depletion of E2F1 protein could abrogate nicotine-mediated induction of Sox2, further implicating their roles in this process downstream of Yap1. Transient transfections were conducted in A549 cells where Yap1, E2F1, or Oct4 were depleted using siRNA and transiently transfected using the Sox2-luc reporter, serum starved for 24 h, and stimulated with 2 μM nicotine for 21 h. These experiments demonstrated that depletion of each of the three proteins could abrogate nicotine-mediated induction of Sox2, further supporting their roles in this signaling cascade . Since these results demonstrated the inability of nicotine to induce Sox2 expression when Oct4 or E2F1 were depleted, we next sought to determine whether disruption of the binding sites of these transcription factors on the Sox2 promoter had similar effects. Experiments were conducted using A549 cells stably expressing either a wild type Sox2-luc reporter (Sox2_WT), a Sox2-luc reporter with the Oct4 site deleted at a region − 95 through -105 bp upstream of the TSS of the Sox2 promoter to prevent Oct4 binding (Sox2_mutOct4), or each of two clones of a Sox2-luc reporter with seven E2F sites mutated to prevent E2F binding (Sox2_mutE2F1_1 or Sox2_mutE2F1_2). Cells were serum starved and subsequently stimulated with 2 μM nicotine for 21 h. These assays showed that while nicotine could enhance Sox2_WT expression, it could not enhance the expression of the Sox2 reporter in cells lacking an Oct4 binding site or in cells with mutated E2F binding sites, suggesting that the Oct4 site might contribute to nicotine-mediated induction of Sox2-luc activity in addition to the E2F sites . To further validate the importance of the Oct4 and E2F binding sites on the Sox2 promoter and assess whether there may be interplay between the two factors, transient transfections were the conducted using each of the stably expressing cell lines to see whether overexpression of E2F1 could still induce Sox2 in cells with altered Oct4 or E2F binding sites, or conversely if Oct4 overexpression could still induce cells with altered Oct4 or E2F binding sites. These experiments demonstrated that when E2F binding sites were mutated on the Sox2 promoter, E2F1 could no longer induce Sox2 but Oct4 could; and conversely when Oct4 sites where deleted on the Sox2 promoter Oct4 could no longer induce Sox2 while E2F1 could . Overall, Yap1 seems to have an integral role in the regulation of Sox2 and its induction by nicotine, an effect which is moderated through is transcriptional collaboration with E2F1 and Oct4; however, whether these two transcription factors function independently or together in this context remains elusive.
# Discussion
CSCs represent a subpopulation of tumor cells with increased tumor-initiating capability. They can divide asymmetrically to replenish the heterogenous tumor bulk, and are highly efficient in initiating tumors upon implantation in animal models [bib_ref] Anti-Cancer stem-like cell compounds in clinical development -an overview and critical appraisal, Marcucci [/bib_ref]. CSCs are resistant to various treatment modalities in part due to their enhanced ability to efflux drugs; additional reasons include the fact that they are slower cycling, and they express higher levels of anti-apoptotic proteins. At the same time, complete mechanisms underlying the drug resistance of these cells are not fully understood [bib_ref] Anti-Cancer stem-like cell compounds in clinical development -an overview and critical appraisal, Marcucci [/bib_ref]. Additionally, these cells are thought to remain dormant and facilitate tumor recurrence and metastasis [bib_ref] Anti-Cancer stem-like cell compounds in clinical development -an overview and critical appraisal, Marcucci [/bib_ref] [bib_ref] Cancer stem cells: a step toward the cure, Boman [/bib_ref]. Not surprisingly, based on these properties of CSCs, efforts are being made to elucidate mechanisms underlying the biology of CSCs in order to target this subpopulation. CSCs have different gene regulatory programs, including epigenetic changes, than the bulk tumor cells; understanding what these differences are, how the programs are regulated, will open up new opportunities for therapeutic targeting.
We have previously reported that nicotine could enhance self-renewal of NSCLC SP cells [bib_ref] Nicotinic acetylcholine receptors induce c-kit ligand/stem cell factor and promote stemness in..., Perumal [/bib_ref]. The schematic In represents the proposed mechanism of nicotine mediated induction of Sox2 and possibly stemness [bib_ref] Nicotinic acetylcholine receptors induce c-kit ligand/stem cell factor and promote stemness in..., Perumal [/bib_ref]. Our earlier studies [bib_ref] Nicotinic acetylcholine receptor signaling in tumor growth and metastasis, Singh [/bib_ref] [bib_ref] Beta-Arrestin-1 mediates nicotine-induced metastasis through E2F1 target genes that modulate epithelial-Mesenchymal transition, Pillai [/bib_ref] [bib_ref] ARRB1-mediated regulation of E2F target genes in nicotine-induced growth of lung tumors, Dasgupta [/bib_ref] [bib_ref] Nicotine-mediated regulation of nicotinic acetylcholine receptors in non-small cell lung adenocarcinoma by..., Schaal [/bib_ref] as well as the current study showed that nicotine binding to the α7 nicotinic acetyl choline receptor recruits β-arr-1 and Src . This results in the activation of Yap1, which has been shown to be a target of Src and Src family members [bib_ref] Nicotine induces selfrenewal of pancreatic cancer stem cells via neurotransmitter-driven activation of..., Al-Wadei [/bib_ref] [bib_ref] TGF-beta targets the hippo pathway scaffold RASSF1A to facilitate YAP/ SMAD2 nuclear..., Pefani [/bib_ref] [bib_ref] A gp130-Src-YAP module links inflammation to epithelial regeneration, Taniguchi [/bib_ref]. α7 nAChR-mediated activation of Src leads to the phosphorylation of Rb and its dissociation from E2F1, enhancing the transcriptional activity of E2F1 [bib_ref] Nicotinic acetylcholine receptor signaling in tumor growth and metastasis, Singh [/bib_ref] [bib_ref] Beta-Arrestin-1 mediates nicotine-induced metastasis through E2F1 target genes that modulate epithelial-Mesenchymal transition, Pillai [/bib_ref] [bib_ref] Nicotine induces inhibitor of differentiation-1 in a Src-dependent pathway promoting metastasis and..., Trevino [/bib_ref] [bib_ref] ARRB1-mediated regulation of E2F target genes in nicotine-induced growth of lung tumors, Dasgupta [/bib_ref] [bib_ref] The Rb-E2F transcriptional regulatory pathway in tumor angiogenesis and metastasis, Schaal [/bib_ref] [bib_ref] Nicotine-mediated regulation of nicotinic acetylcholine receptors in non-small cell lung adenocarcinoma by..., Schaal [/bib_ref]. As mentioned earlier, Yap1 has been found to interact with E2F1 and promote the expression of its downstream targets. Our study suggests a potentially new mechanism by which nicotine induces Sox2 expression in NSCLC cells through Yap1 and its interaction with transcription factors like E2F1 or Oct4 . We also find that nicotine induces expression of Yap1 itself, and that the nicotine-mediated induction of Sox2 and Yap1 is not just specific to lung cancer cells but is also observed in human mesenchymal stem cells. One previous report has demonstrated the ability of nicotine to induce Yap1 in esophageal squamous cell carcinoma (ESCC), and this occurred through nAChRs. Interestingly, they find that Yap1 physically interacts with nAChRs and stimulation with nicotine could induce nuclear translocation and activation of Yap1 by disrupting its association with a negative regulatory complex in the cytoplasm composed of α-catenin, β-catenin, and 14-3-3 proteins. The molecular mechanisms regulating this process are not completely understood.
Our prior studies have shown that Yap1 regulates Sox2 through the binding to Oct4 transcription factor, facilitating self-renewal and vascular mimicry [bib_ref] YAP1 regulates OCT4 activity and SOX2 expression to facilitate self-renewal and vascular..., Bora-Singhal [/bib_ref]. Here we report that E2F1 transcription factor can regulate the Sox2 promoter, and that Yap1 binds to E2F1 likely modulating this effect. Further, we also find that nicotine a Nicotine can induce the association of Oct4 with YAP1 in A549 cells, as seen by a proximity ligation assay. b Depletion of Oct4, E2F1 or YAP1 prevents the nicotine-mediated induction of Sox2-Luc in A549 and H1650 cells. c Similarly, mutating Oct4 or E2F binding sites on the Sox2 promoter prevents its induction by nicotine in A549 cells, as seen in a transient transfection experiment. d Mutating Oct4 binding sites prevents the induction of Sox2-Luc by Oct4, but the promoter is responsive to E2F1; conversely, mutating E2F binding site 1 or site 2 prevents the induction of the Sox2-Luc reporter by E2F1, but conserves the response to Oct4 in a transient transfection experiment. The graphical data represented in this figure has ± standard deviation (SD) values derived from three independent experiments. The statistical comparisons between the representative groups were carried out by one-way ANOVA to determine the statistical significance. e Schematic representing the nicotine mediated upregulation of Sox2. Nicotine binds to the α7 nAChR receptor and activates Src in a βArr1 dependent manner which promotes the binding of to Oct4 and E2F1, resulting in the induction of Sox2 or e-cigarette extracts can increase the binding of Yap1 to both E2F1 and Oct4. Nicotine has been shown to induce E2F1 transcriptional activity through a sequence of signaling events mediated downstream of nAChRs [bib_ref] ARRB1-mediated regulation of E2F target genes in nicotine-induced growth of lung tumors, Dasgupta [/bib_ref]. Upon nicotine binding, β-arrestin-1 scaffolding protein is recruited to the receptor and activates Src kinase, which subsequently activates Raf-1. Raf-1 then acts to phosphorylate the Rb tumor suppressor protein, which is typically bound to E2F1 during cellular quiescence; but dissociation of hyperphosphorylated Rb from E2F1 allows it to turn on a number of promoters involved in proliferation and survival [bib_ref] Rb function in the apoptosis and senescence of non-neuronal and neuronal cells:..., Dasgupta [/bib_ref]. We now find that this pathway might contribute to the induction of stemness, by facilitating the expression of Sox2 . The downregulation of Sox2 expression after 72 h of nicotine treatment is intriguing; the possibility exists that the cells undergo a transition to a more differentiated state, which might not require the presence of Sox2 by that time point. Alternately, the cells might have acquired sufficient levels of downstream targets of Sox2 to maintain stemness and self-renewal and my not require Sox2 per se by that later time point. It is also likely that the cells might have undergone metabolic changes that allows them to survive in the absence of Sox2.
Our studies also suggest that Yap1 is induced by a non-canonical signaling mechanism in response to nicotine. The Hippo signaling pathway has been demonstrated to have tumor suppressive roles, but is aberrantly altered in multiple cancers including those of the lung [bib_ref] The hippo signaling pathway and stem cell biology, Ramos [/bib_ref]. Typically the activation of this pathway by upstream mediators Mst1/2 and Lats1/2 results in the inactivation of Yap1 through its phosphorylation, leading to cytoplasmic sequestration and degradation by 14-3-3 protein [bib_ref] The hippo signaling pathway and stem cell biology, Ramos [/bib_ref]. Our results in NSCLC cells suggest that Yap1 is activated through Src and Yes kinases in response to nicotine; the role of the canonical Hippo signaling pathway in the induction remains unclear.
Overall these studies suggest that upon nicotine binding to α7 nAChR, Src is activated and subsequently leads to Yap1 binding to E2F1 and/or Oct4, upregulating Sox2 expression, thereby enhancing self-renewal of CSCs. However, the role of Oct4 in this process is not fully clear. When endogenous expression of Oct4 is knocked down, nicotine could still induce Sox2; in contrast, in cell lines stably expressing a Sox2 promoter containing a mutation of the Oct4 site prevented nicotine-mediated induction of Sox2-luciferase. The molecular basis for the difference in the induction of endogenous Sox2 versus artificially induced Sox2-luciferase remains elusive at this time. It could be that other proteins are forming complexes with Oct4 or E2F1 to regulate Sox2, and these are disrupted by mutation of the Oct4 binding site. It is also possible that post-translational modifications of the proteins involved or histone modifications on Sox2 promoter around the Oct4 binding site play a role.
Our lab had shown that nicotine induces the translocation of β-arrestin-1 scaffolding protein to the nucleus where it binds to E2F1 transcription factors to enhance transcription of E2F target genes [bib_ref] ARRB1-mediated regulation of E2F target genes in nicotine-induced growth of lung tumors, Dasgupta [/bib_ref]. This was found to occur through the formation of an oligomeric complex consisting of β-arrestin-1, E2F1, and p300 histone acetyltransferase proteins which facilitated the acetylation of histones and E2F1, acting to induce transcription of genes involved in proliferation and survival [bib_ref] ARRB1-mediated regulation of E2F target genes in nicotine-induced growth of lung tumors, Dasgupta [/bib_ref]. Our initial experiments show that depletion of β-arrestin-1 reduced endogenous levels of Sox2; this raises the possibility that Yap1 is recruited to a complex with β-arrestin-1 and E2F1 on the Sox2 promoter. Alternately, β-arrestin-1 might be recruiting p300 to E2F1 independent of Yap1. It is additionally worth noting that other E2F family transcription factors may be involved, and their role is worth investigating. These are novel findings that might have a significant impact on our understanding of how nicotine promotes selfrenewal of stem-like cells from non-small cell lung cancer. Full elucidation of these mechanisms will shed light on the pathophysiology of smoking-related cancers, and reveal new pathways involved in promotion of CSC populations that can potentially be therapeutically exploited.
# Conclusions
The studies presented here provide compelling evidence that nicotine and components of E-cigarettes can promote the self-renewal of lung adenocarcinoma stem-like cells. This occurs through the mediation of Oct4, Yap1 and E2F1, in response to signaling events from the α7 nAChR. Targeting these pathways and molecules might offer a viable strategy to prevent the self-renewal of stem-like cells and perhaps tumor growth that is promoted by nicotine and E-cigarettes.
## Additional file
Additional file 1: .
[fig] Figure 2 a: An immunofluorescence experiment showing the induction of Sox2 by nicotine and E-cigarette extracts in A549 cells (top panels) and human mesenchymal stem cells (bottom panel). b. The induction of Sox2 by the same agents as seen by western blotting. c An RT-PCR experiment shows the induction of Sox2 by nicotine and E-cigarette extracts occurring at the transcriptional level in A549 cells. d Treatment of A549 and H1650 cells that are transiently transfected with a Sox2-Luciferase reporter with nicotine and E-cigarette extracts induces promoter activity, confirming the induction of Sox2 at the transcriptional level. e A western blot showing the induction of Sox2 protein by nicotine from 18 h to 48 h post treatment (f) A RT-PCR experiment showing the induction of Sox2 message at the same time points. The graphical data represented in this figure has ± standard deviation (SD) values derived from three independent experiments. The statistical comparisons between the groups were carried out by unpaired two tailed Student's t-test [/fig]
|
Rationale, design and methods of the Study of Work and Pain (SWAP): a cluster randomised controlled trial testing the addition of a vocational advice service to best current primary care for patients with musculoskeletal pain (ISRCTN 52269669)
# Background
Musculoskeletal pain and in particular acute back pain are major contributors to short term (less than 20 working days) and long term (greater than 20 working days) work absence, accounting for 38% and 37% of short-term absence respectively in manual jobs and 37% and 28% respectively in non-manual jobs. However, around one third of all work absence is attributable to long-term musculoskeletal conditions accounting for long-term absence in 37% of manual and 34% of non-manual jobs.
## Current policy regarding health and work
The health service costs and lost capacity in the workplace have made health and work a key target for public policy. In the UK the Government is actively aiming to reduce the number of employees signed off sick each year. Provision of occupational health in the workplace in the UK is currently limited. Even when occupational health services are broadly defined, only 15% of UK employers provide such a service and these are generally the larger organisations. Occupational health services are even less likely to be provided in Small and Medium Enterprises (SMEs), which employ an estimated 13.5 million people. For the vast majority of SME employees, in the UK and elsewhere, the first line of occupational health care is their primary care practitioner and there is a strong case for primary care services being involved in work-related health interventions by providing more options to refer patients.
## Limitations of current occupational health care for musculoskeletal pain
The benefits of remaining active despite pain have been well documented in workers with musculoskeletal pain and back pain in particular, leading to less sick leave, less time on modified duties and a reduction in pain recurrence [bib_ref] Population based intervention to change back pain beliefs and disability: three part..., Buchbinder [/bib_ref] [bib_ref] Evidence-based care for low back pain in workers eligible for compensation, Mcguirk [/bib_ref] [bib_ref] Systematic reviews of bed rest and advice to stay active for acute..., Waddell [/bib_ref]. A review of vocational rehabilitation highlighted primary care as a key arena in which to address the issue of work with patients. Although there are guidelines in place to support primary care practitioners in providing appropriate advice about work, many GPs have limited training in work issuesand they often report that they feel ill-equipped to deal with patients' concerns about work. In the UK this is particularly important given the introduction of the 'Statement of Fitness for Work' which replaces the sickness certificate, requiring GPs to assess fitness for work and provide their patients with more specific advice regarding activities (e.g. altered hours or modified activities) that may facilitate successful return to work.
## Interventions to facilitate return to work
Initiatives addressing health and work have been predominantly policy driven, such as Job Centre Plus, the Job Retention and Rehabilitation Pilot and the Pathways to Work initiatives in the UKand are often directed towards people who have extended work absence (greater than 6 months). Yet evidence from back pain research suggests that the longer an individual is out of work, the harder it is for them to get back into work [bib_ref] European guidelines for the management of acute nonspecific low back pain in..., Van Tulder [/bib_ref] , therefore it is logical to tackle absence before it becomes long-term. Evidence suggests that intervening in the early stages of sickness absence may be effective for many people with musculoskeletal conditions and yet most initiatives currently are directed towards longer-term absence from work.
In the research arena there are a range of interventions addressing shorter term absence that have been tested to examine their effects on work absence, these include but are not limited to back schools, exercise programmes, work hardening programmes and educational programmes [bib_ref] Workplace-based return-to-work interventions: a systematic review of the quantitative literature, Franche [/bib_ref] [bib_ref] The Costs and Benefits of Active Case Management and Rehabilitation for Musculoskeletal..., Hanson [/bib_ref]. However, these interventions have mostly been undertaken in the workplace, and they have been tailored to the specific needs of the organisations in which they have taken place.
There are methods by which the impact of health on work may be addressed on an individual level, rather than a policy level or organisational level, to ensure that patients receive support in managing their health in the context of their work. In Denmark a multidisciplinary intervention including case management has been evaluated in the rehabilitation of employees sick-listed for 4-12 weeks due to low back pain [bib_ref] The Costs and Benefits of Active Case Management and Rehabilitation for Musculoskeletal..., Hanson [/bib_ref] and "Fit for Work" services, based on case managed, multidisciplinary approaches providing treatment, advice and guidance for people in the early stages of sickness absence have been recommended in the UK. Case management can be defined as a "goal oriented approach to keeping employees at work and facilitating an early return to work" [bib_ref] Subgroup analyses on return to work in sick-listed employees with low back..., Stapelfeldt [/bib_ref]. Given that early intervention is advocated, that musculoskeletal conditions are a common cause of work absence and that many individuals seek their healthcare initially from their primary care practitioner, testing a service located in primary care that can address work issues early on in patients with musculoskeletal conditions is appropriate. However, such a service needs to have a broad enough scope to ensure appropriate advice for the majority of patients whilst still providing a tailored service, therefore the case management approach is the most appropriate model.
## Aim
This paper describes the rationale, design and methods for a cluster randomised controlled trial and linked qualitative interviews, to investigate the clinical and cost-effectiveness of introducing a vocational advice service into general practice, with the aim of providing a structured approach to managing work related issues for primary care patients with musculoskeletal pain who are absent from work or struggling to remain in work.
The principal research question of the SWAP trial is: What is the effect of the addition of a vocational advice service to best current care compared to best current care alone, for adults with musculoskeletal conditions in primary care absent from or struggling to remain in work? The secondary questions are:
1. Is the addition of a vocational advice service to best current care compared to best current care alone for adults with musculoskeletal conditions in primary care absent from or struggling to remain in work cost effective? 2. What are the experiences of patients, GPs/NPs and vocational advisors of a primary care based vocational advice service?
# Ethical approval
Ethical approval was obtained from NRES Committee West Midlands -Staffordshire in April 2012 (REC reference: 12/WM/0020).
# Methods
## Trial design
SWAP is a pragmatic cluster randomised controlled trial with two parallel arms and incorporates economic evaluation and linked qualitative interviews. The unit of randomisation is the general practice with data collected from individual participants.
## Settings and clusters
This cluster trial will take place in six general practices in the South Staffordshire area of the Staffordshire and Stoke-on-Trent Partnership NHS Trust in the UK. Informed consent for practices to participate will be provided by the senior GP partner. Patients will follow the care to which their practice is randomised with identical participant information for both arms explaining that their local musculoskeletal services are being evaluated using patient self-complete questionnaires and medical record review. A second information sheet was used to inform participants about the interview study. Individual patients will be able to opt-out of the questionnaire data collection and the interview study.
## Randomisation and allocation concealment
GP practices are the unit of randomisation. Practices recruited to the cluster trial will be matched based on list size, with matched practices subsequently randomly allocated to the intervention or control arms. Allocation concealment for participating GPs and vocational advisors is not possible but individual participants will not know the allocation of their practice. In this cluster RCT individual participants will not know they are in a trial as the patient information will not mention randomisation of practices and will simply inform participants that local musculoskeletal services are being evaluated. In addition, data entry staff who input data from study questionnaires will be blind to allocation. Analysis of the primary outcome will be carried out by two statisticians (one of which will be blinded to treatment arm). The results will be reviewed and agreed by both statisticians, with one statistician remaining blind until agreement on final estimates is reached.
## Participant eligibility criteria
Adults aged 18 to 70 years consulting in primary care with musculoskeletal pain will be eligible to take part if they are:
Currently employed (paid) Current sickness absence of less than 6 months duration (either GP or self-certified absence) due to musculoskeletal pain OR Patients considered by the GP (or a nurse practitioner (NP)), during the consultation, to be struggling with work due to musculoskeletal pain Exclusion criteria are:
Patients with symptoms indicative of possible serious pathology, requiring urgent medical attention Patients unable to read and speak English Patients with serious mental health problems who are vulnerable and for whom participation in the study would be detrimental (at the GP's discretion) Those who have long term work absence (greater than 6 months) Pregnancy or those patients on maternity leave
## Participant recruitment
Potential participants will be identified when they consult their GP practice with musculoskeletal pain. When a Read code for a musculoskeletal pain problem is entered in the electronic medical record, a computer template will be activated. The template will prompt the GP or NP to record whether the patient is struggling to remain in work or absent from work. Patients who are present when the GP or NP completes the computerised template and express an interest in the research will be given a SWAP information pack at the GP practice. The records of patients who are not present with the GP or NP when the computer template is completed, will be 'tagged' and downloaded on a weekly basis. The local NIHR Clinical Research Network Primary Care administrator will post an information pack to these patients. Electronic templates have been successfully implemented in previous studies carried out by the Arthritis Research UK Primary Care Centre at Keele University and are now routinely used to identify participants for research studies based in general practice [bib_ref] A primary care back pain screening tool: Identifying patient subgroups for initial..., Hill [/bib_ref]. The information pack will include a letter of invitation, a participant information sheet, consent form, self-completion questionnaire (baseline data collection) and a pre-paid reply envelope. The letter of invitation will invite potential participants to take part by completing a consent form and returning the baseline questionnaire. The information sheet will provide further details about the trial. As participants will not be individually consented to randomisation, participants in both arms of the trial will be asked to give written consent to take part in a study investigating work related musculoskeletal problems and local health services by completing three questionnaires (at baseline, 4 months and 12 months) and to allow the research team access to their medical records to identify GP certified Fit Notes in the 6 months prior to consent and during the follow-up period, and to review further health care utilisation for cost analysis. The same procedure will be followed for both the intervention and control practices. A flowchart illustrating the SWAP trial is shown in [fig_ref] Figure 1: SWAP trial flowchart [/fig_ref].
## Description of intervention and control arms
All GPs and NPs working in both the intervention and control practices will be invited to participate in an evidence update session discussing best current care for the management of musculoskeletal pain and work. This aims to ensure that all patients receive the same level of best current care, allowing the added benefit of the vocational advice service to be assessed. The evidence update session will centre on providing GPs and NPs with information to ensure that the correct advice is provided to patients about working with musculoskeletal pain. It will focus on the key messages that a) work is usually good for people with musculoskeletal pain, b) long periods of absence from work are harmful, c) musculoskeletal pain can often be accommodated at work with appropriate adjustments and support, if necessary d) planning and supporting return to work are important parts of clinical management. In addition to these key messages GPs and NPs will be provided with advice about how to approach discussing difficult issues with patients such as negotiating absence or modified duties in the workplace.
## Control practices
Control practices will provide best current care by GPs and NPs in addition to all other usual care that patients may require for their musculoskeletal pain.
## Intervention practices
Intervention practices will also provide best current care and all other care as usual. In addition to best current care a vocational advice service will be available to intervention practices. Patients who require help and support in remaining at or returning to work may be referred to the vocational advice service by their GP or NP, irrespective of whether they also consent to participate in the research evaluation. Patients who are referred to the vocational advice service will be contacted by a vocational advisor, seven days after receipt of the referral who will help the patient to identify and overcome obstacles to remaining at or returning to work. The vocational advisor will wait for seven days before contacting the patient to minimise the number of participants who provide baseline data after having contact with the vocational advisor. It is expected that obstacles to return to work or remaining at work will fall into several categories, and the Flags model of managementof the health and work interface will be used to structure the vocational advice service. The Flags model focuses on the identification of obstacles to working with health conditions, development of a plan to manage health and work, taking action to address the issues each individual patient is facing with respect to managing their musculoskeletal condition in the workplace and re-evaluating the patient's situation regularly until a sustained return to work is achieved. The model is a "light touch" approach based around the principles of case management and stepped care, with vocational advisors providing a goal oriented approach to return to work or remaining in work and with patients being able to "step up" the support they receive when necessary [fig_ref] Figure 2: Model of stepped care provided by the vocational advisor [/fig_ref]. Stepped care has been used successfully in the management of mental health conditions and has begun to be used successfully in pain management [bib_ref] Optimized antidepressant therapy and pain self-management in primary care patients with depression..., Kroenke [/bib_ref]. Patients will be eligible for continued vocational advice until they have a sustained return to work, feel able to manage their health condition in the context of their work, or until they have been absent from the workplace for a total of six months, at which point they will be directed towards other appropriate services.
## Audit of intervention
The vocational advisors will complete case report forms for each participant in the intervention arm, recording basic demographic details, assessment findings and their management plan, and the type and number of contacts each participant has with the vocational advisor. An audit on the completion of the case report forms against the vocational advisor clinical case notes will assess whether patient demographics, contacts with the vocational advisor, details of the assessment and management plan, and any contact with other stakeholders (e.g. healthcare providers, employers) are consistently and accurately recorded on the case report forms.
## Training and mentoring of vocational advisors
Four health care practitioners have been recruited to vocational advisor posts for the trial. They attended a four day training programme on managing work issues within primary care for patients with musculoskeletal conditions. The training programme was based on stepped care and case management principles. The vocational advisors also attended a half day update just prior to the start of the vocational advice service. Monthly mentoring meetings will be scheduled throughout the study where the vocational advisors have the opportunity to request further clarification on any aspect of the teaching, and discuss individual cases both with colleagues and the trainers (a consultant physiotherapist and clinical psychologist who are experienced in managing work related issues).
## Sample size
In summary, 330 recruited participants (165 per arm) in the SWAP trial will give 80% power to detect at least a mean difference of 10 days (days off work between baseline and 4 months) given an expected standard deviation of 25 [bib_ref] Acupuncture for chronic low back pain in routine care, Weidenhammer [/bib_ref] , and 5% two-tailed significance level. The primary analysis method is described below and does not directly involve computing mean difference (but incidence rate ratio via a Poisson/negative binomial process). However, the above calculation holds when applying Normal approximation to the binomial distribution as is generally accepted when the combination of rate of occurrence and sample size is sufficiently large (i.e. both np and n(1-p) exceed 5, where p in this case denotes the probability of taking time off work in any given day, and generally for any Poisson process where the mean/rate is 10 or greater). The model proposed for the analysis of the primary endpoint (number of days off work) is more suited for analysing discrete data than general linear models which are suited to continuous data.
The total sample size requirement for analysis of an unclustered unadjusted analysis based on detecting a mean difference of 10 with 80% power and two-tailed 5% significance level (assuming an SD of 25) is 200 individuals (100 per study arm). The above calculated total sample size requirement of 330 participants for the SWAP trial takes into account three levels of inflation (of the unadjusted : (i) 20% (×1.2 magnification) through clustering of data (at practitioner-level) based on an ICC for betweenpractitioner effects of 0.05 [bib_ref] Measuring practitioner/therapist effects in randomised trials of low back pain and neck..., Lewis [/bib_ref] with anticipated average cluster size of 5 per practitioner; (ii) 15% owing to variation in expected recruitment rates between GPs (based on an expected coefficient of variation of 0.65) [bib_ref] Sample size for cluster randomized trials: effect of coefficient of variation of..., Eldridge [/bib_ref] , and (iii) 20% allowance for loss to follow-up at 4 months.
## Participant (baseline) characteristics
Baseline data by trial arm will be summarised and presented. Baseline characteristics are to be compared between arms, and presented at the level of: (i) GP practice clusters, and (ii) Patient characteristics.
Baseline data for GP practice characteristics include data on the stratified variable for randomisationi.e. practice list size. Also, number of GP practitioners, median index level of deprivation for the practice, mean age, and gender (male/female) distribution of practice populations will be described.
Also, comparison will be made between participants' demographic, pain/disability and quality of life characteristics. Mean (SD) and median (IQR) will be applied to normal and skewed numerical data respectively. Frequency counts and percentages will be presented for nominal and ordered data.
Balance of baseline characteristics is particularly important to establish for cluster trials given the (higher level) unit of randomisation. A lack of balance is indicative of differential selection of patients to the trial across the treatment arms (though appreciating that random difference will occur due to randomisation and between-practice variations). A further limitation of the design is that 'baseline' assessment occurs after initial GP or NP consultation and therefore a difference in pain management responses may occur within that periodi.e. prior to baseline assessment (so baseline, defined as the date the participant completes the first questionnaire, in this context is not a true 'baseline' of where baseline is usually considered to be, prior to the start of treatment). We may expect differences in treatment (such as issuing of sickness certificates) to occur by this baseline assessment and in particular there may already be differences in approach that may influence the primary outcome by the time of this first self-report baseline assessment. Therefore, the primary analysis will not adjust for baseline pain intensity (though this adjustment will be carried out as a sensitivity analysis -see Analysis section for further details). No formal statistical testing will be carried out for differences in baseline characteristics as this is not an 'outcome' for the trial.
## Assessment of potential bias
Selection bias: Over the period of recruitment the number of patients who consult with musculoskeletal pain and are potentially eligible for the trial as coded by the GP or NP on the computer prompt will be recorded in the intervention and control practices. Any evidence of selection bias in rate of uptake to the research and in baseline descriptive statistics between the control and intervention practices will be explored. Demographic comparisons will be drawn between trial participants, non-participants and screened patients who do not take part.
Attrition bias: Differences between individuals that are followed up and those who dropout is a concern, and may result in between-arm bias in estimates particularly if dropout is unequal between the two trial arms and analysis fails to take into account appropriate adjustment for missing data. Thus, we will compare: (i) baseline characteristics of those who are successfully followed up at 4 months against those who dropout to assess whether those missing are related to observed baseline factors, (ii) attrition rate between trial arms to assess whether there is a differential dropout rate. Statistical adjustment will be carried out to help address issues of imbalance in characteristics.
## Outcome assessment primary outcome measure
The primary outcome measure is number of days off work over 4 months from entry into the trial. This is based on response to the following questions in the 4month self-report questionnaires: "Have you taken time off work during the last 4 months (since your last questionnaire) because of your pain? If yes, please write in the number of days, weeks or months you were off work due to your pain in the last 4 months. (i.e. between baseline and 4 month follow up assessments)". Days off work in this context jointly captures sick leave issued by the GP and shorter length self-certified absences that don't require GP sign-off.
## Secondary outcome measures
Self-reported time off work (in binary form (yes/no)) will be a secondary outcome. We will also undertake a separate analysis to compare the proportion of participants in the two trial arms that are issued a GP sickness certificate in the first 4 months (through review of medical records for those who provide consent to medical record review). Secondary evaluation will also look at self-reported time off work and medical record review based sick certification periods over 12 months follow-up.
Other secondary outcome measures include the Selfefficacy to Return to Work Questionnaire [bib_ref] Development of the return-to-work self-efficacy (RTWSE-19) questionnaire -psychometric properties and predictive validity, Shaw [/bib_ref] , pain intensity (0-10 rating scales), bothersomeness (1-5 rating scale), global assessment of change and work performance (SPS6). [fig_ref] Table 1: Outcome measures and timing of data collection [/fig_ref] summarises the outcome measures and their respective time-points of data collection.
# Analysis
Data will be analysed after the 4 month follow-up and the 12 month analysis, which will include Health Economic data, will then follow. For the primary analysis (time (days) off work in the first 4 months), the proposed analysis is by hierarchical negative binomial regression adjusting for age, gender, and GP practice size (at the GP-cluster level). Multi-level robust Poisson and zero-inflated models taking into account clustering of data and over-dispersion in the spread of data will also be scrutinised (alongside the negative binomial model). The goodness of fit of each model for observed versus predicted values will be scrutinised (comparisons will be drawn through a likelihood-ratio test and Akaike's Information Criterion (AIC)). Mixed-models (linear-or generalised-as appropriate to numerical and categorical outcome data, respectively) will be fitted to estimate and test for between arm effects across primary and secondary outcome measuresadjusting for baseline covariates (as indicated above). An intention-to-treat approach analysing participants as per randomised allocation will be followed.
A limitation to the design/methods of this trial is the small number of GP practice clusters (i.e. units of randomisation). It has been reported that a minimum number of clusters for a valid methodological evaluation is four per arm (our trial has three GP practices per arm) [bib_ref] Patterns of intra-cluster correlation from primary care research to inform study design..., Adams [/bib_ref]. Much of this concern centres on the assumptions for the hierarchical model, and the fact that any cluster level analysis (only) will fail to detect a statistically significant pvalue (at the level of the customary 5% two tail testing). Hence, we propose to carry out the hierarchical model with individual practitioners (GP/NPs as opposed to GP practice) as the upper-level random factor. GP/NPs are likely to be the main contributors to the variation in sickness certification between GP practices and may therefore be considered to be reasonable substitutes [bib_ref] General practitioner sickness absence certification for low back pain is not directly..., Watson [/bib_ref]. Descriptive statistics on numbers of participants and proportion of participants who take time off work in each arm will be reported for the primary outcome. The adjusted effect estimate (incidence rate ratio), 95% confidence interval and p-value for the test of association for the primary measure will be presented. Similarly, mean scores (SDs) for numerical outcomes and frequency counts and percentages for categorical data will be presented for secondary outcome measuresas appropriate to the scale of the data. Mean differences and 95% CIs and odds ratios with 95% CIs will be presented for all secondary outcomesas appropriate to the scale of the data.
# Sensitivity analysis
The following sensitivity analyses are planned:
1. Evaluation of the primary outcome measure (number of days off work) by robust Poisson and zero-inflated models. 2. The main evaluations will utilise minimal covariate adjustment (owing to the fact that differences may already be inherent in baseline assessment due to the time-scale of return of questionnaires following the initial GP/NP consultation). However, we would not anticipate any real difference in outcomes in such a short time period particularly as the patients in the intervention practices will not have been contacted by a vocational advisor until at least 7 days after receiving the baseline questionnaire. Greater covariate adjustment is also relevant in that it helps safeguard the analysis against major selection bias and/ confounding bias. Thus, as a sensitivity analysis, we will carry out statistical modelling that includes additional baseline adjustment by further including pain intensity, time off work at baseline (at the individual level) as well as corresponding baseline score (if applicable). 3. A second sensitivity analysis will be carried out at the upper cluster level (individual practitioners) using non-parametric sum rank test and permutations test. Individual-level regression methods may not be reliable and the distributional assumptions difficult to verify when the number of units of analysis are smallin such circumstances, as is the case in this trial, it is recommended to carry out a simple crude analysis that is not dependent on distributional assumptions (in this case a simple non parametric comparison since the primary outcome of interest is likely to be highly skewed). 4. Per protocol evaluation (further sensitivity analysis of the primary outcome): A per protocol evaluation will be undertaken comparing the primary outcome for those participants in the intervention practices who engaged with any aspect of the vocational advice service (at least one contact by telephone) versus 'comparable' participants in the control practices. A complier average causal effect analysis will be performed to provide an unbiased estimate of 'per protocol' effect by adjusting the per protocol estimate (on the assumption that a similar level of non-compliance would be expected for the control arm).
Subgroup analyses: Evaluation of the primary outcome measure will be carried out to examine whether time off work/number of days absenteeism is different across different baseline subgroups by: return to work self-efficacy, location of pain (spinal pain versus pain in other areas), and duration of work absence. Statistical estimates will be obtained through including interaction terms in the statistical model of treatment effect.
## Economic evaluation
The economic evaluation conducted alongside the SWAP trial will determine the cost-effectiveness and return on investment of the vocational advice service (cost-benefit analysis) in comparison to best current care.
A cost-consequence analysis will initially be reported, describing all the important results relating to costs and consequences (across the full range of clinical outcomes). Subsequently, two methods of economic evaluation will be used. A cost-effectiveness analysis will be undertaken from a healthcare perspective to determine the cost per additional day of work absence avoided. A cost-benefit analysis will also be undertaken from a broader societal perspective to calculate the net societal benefit of the vocational advice service, by subtracting the difference in direct health care costs (costs) between the groups from the difference in indirect productivity costs (benefits) between the trial arms.
## Costs
Information on time off work will be collected from the postal questionnaires completed by patients at 4 months and 12 months. Health care resource use will be collected in the 12 month questionnaire. Health sector costs will include primary and secondary care contacts, investigations, medication and contacts with other health care professionals such as physiotherapists and occupational therapists (both through the NHS and private). Data on musculoskeletal pain related time off work and health care resource use will also be available from the medical record review. Questions on patients' personal expenditure will concentrate on private health care use and over-thecounter treatments. Questions on time off work and occupation will provide information required to calculate the indirect (productivity) costs (benefits). In order to obtain the cost of the vocational advice service, information on the type and number of contacts with the vocational advisor (telephone calls or visits) will be obtained and unit costs applied to calculate overall cost of the intervention.
Resource use will be multiplied by unit costs obtained from standard (national) sources and health care providers [bib_ref] Unit Costs of Health and Social Care, Curtis [/bib_ref]. Due to the lack of nationally representative unit cost estimates for private health care, this care will be costed as the NHS equivalent. Patient reported costs for over-the-counter treatments will be used.
## Health economic outcomes
The outcome measure for the cost-effectiveness analysis is self-reported number of days absent from work. In the cost-benefit analysis, benefits will be estimated from the productivity losses. These will be calculated using data collected on employment status at every time point and number of days off work due to their musculoskeletal pain problem. Information on occupation, further details of typical work activities and the nature of their employment (full time or part time) will be sought in follow-up questionnaires. The average wage for each respondent will be identified using UK Standard Occupational Classification coding and annual earnings data for each job type. The analysis will use the human capital approach, and the self-reported days of absence will be multiplied by the respondent-specific wage rate. The human capital approach assumes that the value of lost work is equal to the amount of resources an individual would have been paid to do that work, and values productivity losses as a result of morbidity (or mortality) by measuring time lost from work and multiplying this with the gross wage of the person.
## Cost-consequence, cost-effectiveness and cost-benefit analysis
The health economic analysis will estimate the incremental cost-effectiveness and the cost-benefit of the intervention in comparison with best current care. Costs for the trial arms will be presented for each broad cost category (health care costs, patient-incurred costs, productivity costs) and disaggregated within each of these cost categories. An incremental cost-effectiveness analysis will be conducted from a healthcare perspective using information on time off work to calculate the cost per additional day of work absence avoided. A cost-benefit analysis from a broader societal perspective will calculate the net societal benefit of the intervention in monetary terms, by subtracting the difference in costs from the difference in benefits (productivity losses). Subsequently, a return on investment will be calculated by dividing the net benefits of the vocational advice service (gain minus cost) by the net costs of the intervention. The base-case analyses will use self-reported patient information on health care utilisation over a 12 month period.
The data for costs is likely to have a skewed distribution therefore a non-parametric comparison of means (e.g. bootstrapping) will be undertaken to estimate confidence intervals around costs. Mean substitution techniques (for individual-item missing resource use data) and multiple imputation techniques (resource use data) will be carried out to ensure that all trial participants are included in the final analysis. Clustering of data by GP/NP will be taken into account through a multi-level approach, in line with the main statistical analysis. Adjustment for baseline covariates will focus on the same variables as outlined for the primary clinical evaluation.
The robustness of the base-case results will be explored using sensitivity analysis. This will explore uncertainties in the trial based data itself and the methods employed to collect and analyse the data. An available case analysis will be conducted as an alternative to using a multiple-imputed data set. A further sensitivity analysis will be undertaken using health care resource use data solely obtained from the medical record review. Uncertainty will be explored through the use of cost-effectiveness acceptability curves (CEACs); these plot the probability that the addition of a vocational advice service is cost-effective against threshold values for cost-effectiveness.
## Qualitative research qualitative methods
In linked qualitative interviews we will explore experiences of the vocational advice service from the perspectives of GPs and NPs in the intervention practices who can refer patients to the service, patients who access the service with work related problems and vocational advisors who are delivering the service. GPs and NPs (up to n = 15) will be interviewed both prior to the start of the new service and 12 months later. Patients (n = 20) who have consented to the research evaluation will be opportunistically invited for interview following discharge from the care of the vocational advisors. Vocational advisors (n = 4) will be interviewed four times, prior to the start of the new service and at 1, 6 and 12 months after the vocational advice service commences. These longitudinal interviews will explore how their knowledge, confidence and experience of providing the vocational advice service evolves over time.
# Qualitative analysis
Interviews with GPs/NPs, patients and vocational advisors will initially be coded in N-vivo 9 and subsequently analysed in search of common themes and differences, using the constant comparative framework, based on the broad principles of grounded theory. Although each set of interviews will be coded separately (using separate coding frameworks), each dataset will subsequently be analysed as a whole in search of similarities and differences across GPs/NPs, patients and vocational advisors. Samples of early interviews will be independently coded by members of the multidisciplinary research team, coding frameworks agreed and coded data analysed in search of themes at multidisciplinary research analysis meetings. The themes will be analysed through in-depth discussion to examine plausibility and validity to develop a robust thematic framework (or conceptual model); specifically patients' , vocational advisors' and GPs/NPs' perceptions of the acceptability, benefits and limitations of the vocational advice service. The constant comparative method will provide a means of identifying similarities and differences in the qualitative data, whilst the longitudinal dimension of the qualitative interviews with VAs (baseline, 1 month, 6 months and 12 months) and GPs/NPs (baseline and 12 months) will identify changes over time in attitudes and experiences towards the acceptability and added value (or otherwise) of the vocational advice service.
Trial timeline Trial recruitment commenced in July 2012. We aim to recruit 330 participants into the trial over an 18 month period from 6 general practices. Follow-up is targeted for completion by January 2015 and analysis will follow.
# Discussion
The SWAP trial is investigating the clinical and costeffectiveness of the addition of a vocational advice service to best current primary care to provide a structured approach to managing work related issues in primary care patients with musculoskeletal pain who are absent from work or struggling to remain in work. Given that early intervention is advocated, that musculoskeletal conditions are a common cause of work absence and that the majority of individuals in the UK seek their healthcare initially from their GP, we have developed and are testing a service located in primary care to address the issues of health and work early on in patients with musculoskeletal conditions. This is the first such trial in the UK. The results will provide evidence to inform primary care practice and may guide the development of services to provide support for musculoskeletal pain patients with work-related issues.
The main strength is the cluster randomised controlled trial design. The primary outcome is self-reported number of days off work over 4 months. A range of secondary outcomes will also be assessed and qualitative interviews will explore the value of a vocational advice service to GPs and NPs, patients and vocational advisors.
[fig] Figure 1: SWAP trial flowchart. [/fig]
[fig] Figure 2: Model of stepped care provided by the vocational advisor (VA) in addressing return-to-work (RTW). [/fig]
[table] Table 1: Outcome measures and timing of data collection [/table]
|
Effect of a Run‐In Period on Estimated Treatment Effects in Cardiovascular Randomized Clinical Trials: A Meta‐Analytic Review
BACKGROUND: A run-in period may increase adherence to intervention and reduce loss to follow-up. Whether use of a run-in period affects the magnitude of treatment effects is unknown.METHODS AND RESULTS:We conducted a meta-analysis comparing treatment effects from 11 systematic reviews of cardiovascular prevention trials using a run-in period with matched trials not using a run-in period. We matched run-in with non-run-in trials by population, intervention, control, and outcome. We calculated a ratio of relative risks (RRRs) using a random-effects meta-analysis. Our primary outcome was a composite of cardiovascular events, and the primary analysis was a matched comparison of clinical trials using a run-in period versus without a run-in period. We identified 66 run-in trials and 111 nonrun-in trials (n=668 901). On meta-analysis there was no statistically significant difference in the magnitude of treatment effect between run-in trials(relative risk [RR], 0.83 [95% CI, 0.80-0.87]) compared with non-run-in trials (RR, 0.88 [95% CI, 0.84-0.91]; RRR, 0.95 [95% CI, 0.90-1.01]). There was no significant difference in the RRR for secondary outcomes of all-cause mortality (RRR, 0.97 [95% CI, 0.91-1.03]) or medication discontinuation because of adverse events (RRR, 1.05 [95% CI, 0.85-1.21]). Post hoc exploratory univariate meta-regression showed that on average a run-in period is associated with a statistically significant difference in treatment effect (RRR, 0.94 [95% CI, 0.90-0.99]) for cardiovascular composite outcome, but this was not statistically significant on multivariable meta-regression analysis (RRR, 0.95 [95% CI, 0.90-1.0]).CONCLUSIONS:The use of a run-in period was not associated with a difference in the magnitude of treatment effect among cardiovascular prevention trials.A prerandomization run-in period is intended to improve the precision of treatment-effect estimates by excluding participants who are nonadherent with study interventions or trial protocols. 1-5 During a run-in period, participants receive an intervention (either active or placebo), and only those deemed adherent and tolerant of the intervention during the run-in period are randomly assigned in the trial. The main proposed advantage of a run-in period is improved average drug adherence rates of randomly assigned participants and improved rates of follow-up.6However, if populations who are nonadherent with medications are also those at higher cardiovascular risk (eg, higher prevalence of comorbidities), use of a
run-in period may result in a trial population with lower event rates. [bib_ref] Medication nonadherence: an unrecognized cardiovascular risk factor, Munger [/bib_ref] An unproven concern of run-in trials, given the systematic exclusion of nonadherent individuals, is the introduction of a selection bias, thereby affecting the external validity of the trial. [bib_ref] External validity of randomised controlled trials: "to whom do the results of..., Rothwell [/bib_ref] This issue is most relevant to phase III randomized clinical trials where analyses are based on the intention-to-treat principle, in which the purpose is to represent summary treatment effects for a general population, including those who are nonadherent with the trial intervention. As such, one may speculate that the treatment effects reported in run-in trials may be closer to an "on-treatment" analysis than an intention-to-treat analysis given that it is expected to increase the proportion adherent with treatment. Moreover, if participants who fail the run-in period are systematically different from those who succeed the run-in period, especially if differences relate to treatment efficacy, the estimates of treatment efficacy and safety may also differ between run-in and non-run-in trials. To date, evidence to support or refute differences in treatment effect estimates is lacking despite the widespread use of run-in periods. [bib_ref] Application and impact of run-in studies for the evaluation of statin efficacy..., Fralick [/bib_ref] ,10 It has also not been established whether the proposed benefits of run-in, that is, improved adherence with the intervention and trial protocols, are substantiated in clinical practice. Employing a run-in period increases the complexity of randomized clinical trials, adds more burden to study participants and researchers, and may increase trial costs. [bib_ref] Use of run-in periods in randomized trials, Huo [/bib_ref] Accordingly, it is necessary to confirm the proposed benefits of run-in periods and reliably exclude any association with biased treatment estimates.
In this study, we evaluated whether the use of a run-in period in cardiovascular prevention trials is associated with differences in relative treatment effects and the rates of adverse events resulting in drug discontinuation, cardiovascular events, mortality, and loss to follow-up.
# Methods
## Data sources
The data that support the findings of this study are available from the corresponding author upon reasonable request. A protocol detailing the design and methods of the current meta-analysis has been published previously. [bib_ref] The impact of a run-in period on treatment effects in cardiovascular prevention..., Murphy [/bib_ref] In summary, we identified 11 systematic reviews published between 2010 and 2019 that reported both primary and secondary prevention trials of effective therapies for cardiovascular prevention, namely, antihypertensive, lipid-lowering, and glucose-lowering medications [bib_ref] Efficacy and safety of more intensive lowering of LDL cholesterol: a meta-analysis..., Baigent [/bib_ref] [bib_ref] Association of blood pressure lowering with mortality and cardiovascular disease across blood..., Brunstrom [/bib_ref] [bib_ref] Statins for prevention of cardiovascular disease in adults: evidence report and systematic..., Chou [/bib_ref] [bib_ref] The effects of blood pressure reduction and of different blood pressure-lowering regimens..., Czernichow [/bib_ref] [bib_ref] Blood pressure lowering for prevention of cardiovascular disease and death: a systematic..., Ettehad [/bib_ref] [bib_ref] Proprotein convertase subtilisin/kexin 9 inhibitors in reducing cardiovascular outcomes: a systematic review..., Du [/bib_ref] [bib_ref] Effect of statins and non-statin LDL-lowering medications on cardiovascular outcomes in secondary..., Koskinas [/bib_ref] [bib_ref] Cardiovascular, mortality, and kidney outcomes with GLP-1 receptor agonists in patients with..., Kristensen [/bib_ref] [bib_ref] Antihypertensive treatment and secondary prevention of cardiovascular disease events among persons without..., Thompson [/bib_ref] [bib_ref] Blood pressure-lowering drugs and secondary prevention of cardiovascular disease: systematic review and..., Xie [/bib_ref] [bib_ref] SGLT2 inhibitors for primary and secondary prevention of cardiovascular and renal outcomes..., Zelniker [/bib_ref] ; [fig_ref] Table 1: Characteristics of Run-In and Non-Run-In Trials Matched by Cardiovascular Composite Outcome [/fig_ref]. We selected systematic reviews of proven cardiovascular prevention benefit where the overall summary estimate of the meta-analyses was significant to test the hypothesis that the run-in trial might bias estimates of efficacy. The strategy of sourcing trials for inclusion from a range of high-quality published systematic reviews allowed us to reduce research waste. [bib_ref] Accumulating research: a systematic account of how cumulative meta-analyses would have provided..., Clarke [/bib_ref] We included all phase III randomized clinical trials from these systematic reviews where an effective medicinal product for cardiovascular prevention (ie recommended by current cardiovascular prevention guidelines) was compared with a placebo or standard-of-care control. Given that none of these published meta-analyses selected trials based on run-in status, we considered this approach to be associated with a lower risk of selection bias. We conducted primary data extraction independently from the primary trial report with data double checked by a second independent researcher. Any discrepancies were reviewed by both reviewers and resolved by consensus of the data extraction team, or if required, review with a senior author.
## Clinical perspective
What Is New?
- A run-in period is used in randomized clinical trials to increase adherence to the intervention and reduce loss to follow-up in clinical trials. - It is not known whether the use of a run-in period affects the magnitude of treatment effects in cardiovascular prevention trials. - In this meta-analysis of cardiovascular prevention clinical trials, differences in treatment effects for cardiovascular composite outcomes between trials with run-in periods What Are the Clinical Implications?
- Our analysis reports that the use of a run-in period, compared with matched trials that did not use a run-in period, was not significantly associated with a difference in treatment estimates in cardiovascular prevention randomized clinical trials.
## Nonstandard abbreviations and acronyms
RRR ratio of relative risk UL upper bound of CI
## Matching process
We integrated a methodological approach based on previous meta-analyses evaluating the effect of loss to follow-up and early trial stopping rules on effect estimates of interventions in randomized clinical trials. [bib_ref] Potential impact on estimated treatment effects of information lost to follow-up in..., Akl [/bib_ref] [bib_ref] Stopping randomized trials early for benefit: a protocol of the Study Of..., Briel [/bib_ref] We completed the matching of run-in trials with non-run-in trials using a 3-step approach. In step 1, we matched run-in and non-run-in trials by intervention, as we considered matching on drug class to be an essential criterion (eg, match angiotensin-converting enzyme inhibitor with angiotensin-converting enzyme inhibitor). Therefore, a mandatory matching criterion was an exact match for intervention (ie, needed to be the within the same drug class). Following step 1, we generated a score for each potential run-in/non-run-in trial pairing based on similarity of population, control, and outcome. A score of 0 was defined as not a match, 1 was defined as an acceptable match, 2 was defined as a close match, and 3 was defined as an exact match based on prespecified criteria [fig_ref] Table 2: Summary Estimates of Relative Treatment Estimates [/fig_ref]. In step 2, we included all potential matches from step 1 and applied a population score criterion of ≥1 (ie, minimum acceptable match). Therefore, following step 2, all potential matches had a minimum score of 4. Following this step, we had 1359 unique potential matches between run-in and non-run-in trials. In step 3, we assigned a matching score for all potential matched trial pairs, ranging from a minimum of 4 (a score of "acceptable match" in each population, intervention, control, and outcome domain) to a maximum of 12 (a score of "exact match" in each population, intervention, control, and outcome domain; [fig_ref] Figure 2: Pooled ratio of relative risks and 95% CIs for run-in versus non-run-in... [/fig_ref].
# Statistical analysis
For each run-in and non-run-in trial, we calculated the individual trial relative risk (RR) for each outcome measure. If multiple non-run-in trials matched to a single run-in trial, we performed a random-effects meta-analysis of the non-run-in trials to calculate the summary estimate to give a single non-run-in RR to compare with the run-in trial. This meant that for each matched comparison, a run-in trial RR could either be matched against a single non-run-in trial RR or a meta-analysis of multiple non-run-in RRs. Then for each matched comparison, we calculated a ratio of RRs (RRRs) and 95% CI by subtracting the log(nonrun-in trial RR) from the log(run-in trial RR) with SEs calculated as the square root of the following: SE(Run-in RR) 2 +SE(Non-run-in RR) 2 . We conducted a randomeffects meta-analysis of the RRRs between run-in and matched non-run-in trials to obtain a summary estimate of the RRRs. A ratio <1 indicates a greater treatment effect in the run-in trials (compared with matched non-run-in trials). We used a random-effects model because run-in trials were matched to variable numbers of non-run-in trials, and we considered this approach to best represent the "average" log-RR in this setting. We used random-effects meta-analysis because we considered that the true effect sizes being combined were exchangeable. [bib_ref] A re-evaluation of random-effects meta-analysis, Higgins [/bib_ref] In addition, we performed sensitivity analysis using fixed-effects metaanalysis to ensure consistent findings regardless of the exact form of meta-analysis used. We used Cochran's Q test to test for heterogeneity and the I 2 statistic to estimate the percentage of variability.
Our primary analysis was a comparison of the RR reported in run-in trials compared with matched nonrun-in trials for the primary outcome (cardiovascular composite) and was confined to "best-matched" trials (ie, maximum population, intervention, control, and outcome matched score) for this outcome. Non-run-in trials could only be matched once to prevent trials with large treatment effects biasing the estimate. For a non-run-in trial prematched to several run-in trials, we selected the run-in and non-run-in pair with the largest population, intervention, control, and outcome score and run-in trial sample size [fig_ref] Figure 3: Pooled ratio of relative risks and 95% CIs for run-in versus non-run-in... [/fig_ref]. We completed analyses of secondary clinical outcomes, all-cause mortality, adverse events leading to medication discontinuation, nonfatal myocardial infarction, nonfatal stroke, and loss to follow-up.
Sensitivity analyses were conducted where nonrun-in trials were permitted to match multiple times to different run-in trials. We also completed a robustness analysis using a parametric bootstrapping method with 1000 iterations. The bootstrap algorithm takes into account variability (at the study level), variability attributed to the selection of studies, and correlation structure attributed to nonindependence of matched pairs. During each iteration, we resampled each RR from a normal distribution with mean of the log(RR) and SD equal to the RR variance to account for the sampling error of each trial RR estimate. For each possible run-in and non-run-in match, we calculated the log-RRR by subtracting the corresponding log-RR and resampled all pairs with replacement. In this analysis, we calculated the mean, 2.5% quantile, and 97.5% quantile from the 1000 summary estimates of RRRs to obtain a parametric bootstrapping summary estimate and CI.
To address the degree to which our matching ameliorates confounding, we completed an analysis without individual trial matching (ie, comparison of meta-analytic estimate of the 2 groups of trials). First, we calculated the summary run-in and non-run-in RR estimates using random-effects meta-analysis, and then we calculated the log(run-in RR)−log(non-run-in RR) to obtain an RRR. We then calculated the corresponding E-value [bib_ref] Sensitivity analysis in observational research: introducing the E-value, Vanderweele [/bib_ref] for the upper bound of CI: UL*=1/ UL and E-value=UL*+sqrt(UL*×[UL*−1]).
We additionally performed a random-effects metaregression analysis with log-transformed treatment effect as our outcome and run-in status as our predictor variable, with univariate and multivariable adjusting for mean age, sex, trial sample size, duration of follow-up, primary versus secondary prevention, and year of publication.
To explore differences in clinical event rates between run-in and non-run-in trial populations, we compared incidence rates of clinical events and loss to follow-up in the control groups of trials using an inverse variance weighting meta-analysis to give higher importance to studies of larger sample sizes.
## Assessment of the quality of the studies: risk of bias
We used the Cochrane Risk of Bias Tool to assess methodological quality of eligible trials including random sequence generation, allocation concealment, blinding of participants and health care personnel, blinded outcome assessment, completeness of outcome data, evidence of selective reporting, and other biases. [bib_ref] The Cochrane Collaboration's tool for assessing risk of bias in randomised trials, Higgins [/bib_ref] Risk-of-bias assessments were performed independently by 2 reviewers, and disagreements resolved by a third reviewer.
# Results
## Comparison of characteristics of run-in and non-run-in trials
From the 11 identified systematic reviews, a total of 177 randomized clinical trials were identified, with 66 eligible run-in trials and 111 eligible non-run-in trials with a total sample size of 668 901 patients [fig_ref] Table 1: Characteristics of Run-In and Non-Run-In Trials Matched by Cardiovascular Composite Outcome [/fig_ref]. [fig_ref] Table 1: Characteristics of Run-In and Non-Run-In Trials Matched by Cardiovascular Composite Outcome [/fig_ref] describes the characteristics of the run-in and non-run-in studies included in our primary cardiovascular composite outcome analysis with 32 run-in trials and 76 non-run-in trials (ie, best matches). The majority (88%) were parallel group trials. Run-in trials Data are provided as mean±SD or number (percentage). *Between-group significant differences were calculated using t tests and Pearson chi-squared tests for categorical variables. had older populations (mean±SD age, 64.5±6.28 versus 61.0±6.20; P=0.01) and larger samples sizes, with the mean number of participants in the run-in trials (mean±SD sample size, 6604±7243) significantly larger than in the non-run-in trials (mean±SD sample size, 3471±4036; P=0.03). Overall, 78% of run-in trials were reported in high-impact journals compared with 59% of non-run-in trials (P=0.09). Risk-of-bias comparisons were similar between run-in and non-run-in studies [fig_ref] Table 1: Characteristics of Run-In and Non-Run-In Trials Matched by Cardiovascular Composite Outcome [/fig_ref] and . The mean±SD duration of follow-up were similar: 42.8±17.5 months versus 42.7±17.7 months (P=0.97). A meta-analysis of the matched runin and non-run-in trials found no significant difference for loss to follow-up (RRR, 1.04 [95% CI, 0.84-1.30]). Differences between placebo run-in and active run-in studies are included in [fig_ref] Table S3: Comparison of Placebo Run-in and Active Run-in Studies [/fig_ref]. Active run-in was more commonly performed in antihypertensive studies than in studies of lipid-lowering or glucose-lowering agents (P=0.05).
## Cardiovascular composite outcome
A total of 32 run-in trials matched with ≥1 non-run-in trials, of which 15 were matched with multiple nonrun-in trials. There was no significant difference in treatment effect for cardiovascular events between run-in trials (RR, 0.83 [95% CI, 0.80-0.87]) compared with non-run-in trials (RR, 0.88 [95% CI, 0.84-0.91]; RRR, 0.95 [95% CI, 0.90-1.01]; . Sensitivity analyses including multiple-matches analysis and parametric bootstrapping analysis increased the precision of our estimate but were not statistically significant [fig_ref] Table 2: Summary Estimates of Relative Treatment Estimates [/fig_ref]. Our latent confounding sensitivity analysis suggested that an unmeasured confounder, . Pooled ratio of relative risks (RRR) and 95% CIs for run-in versus non-run-in randomized clinical trials (RCTs): cardiovascular composite. Single best population, intervention, control, and outcome score analysis. Pooled ratio of relative risks (RRR) for matched run-in and non-run-in RCTs that reported a cardiovascular composite outcome. "No. of Studies" refers to the number of non-run-in trials that have been matched to each individual run-in trial. The size of the data markers indicates the weight of the comparison in the metaanalysis. A summary estimate <1 represents an exaggerated treatment effect of run-in. df indicates degrees of freedom; I 2 , degree of heterogeneity; Q, Cochran's Q; RE, random-effects meta-analysis; and RR, relative risk. associated with run-in and RR, with a risk ratio of 1.16 could explain the upper confidence limit, but weaker confounding could not. Post hoc exploratory univariable meta-regression analysis showed that on average a run-in period was associated with a statistically significant difference in treatment effects (RRR, 0.94 [95% CI, 0.90-0.99]) for cardiovascular composite outcome. This was not statistically significant on multivariable meta-regression analysis adjusting for mean age, sex, trial sample size, duration of follow-up, primary versus secondary prevention, and year of publication (RRR, 0.95 [95% CI, 0.90-1]). Sensitivity analysis using fixed-effects meta-analysis of the non-run-in trials did not alter our findings (RRR, 0.95 [95% CI, 0.90-1.01]).
## All-cause mortality outcome
A total of 34 run-in trials matched with ≥1 run-in trials, of which 24 were matched with multiple non-run-in trials. There was no difference in mortality incidence rate in control groups of run-in studies (24.7 per 1000 person-years) compared with non-run-in studies [fig_ref] Figure 2: Pooled ratio of relative risks and 95% CIs for run-in versus non-run-in... [/fig_ref]. Sensitivity analyses including multiple-matches analysis, bootstrapping analysis, or meta-regression did not alter the findings [fig_ref] Table 2: Summary Estimates of Relative Treatment Estimates [/fig_ref].
## Adverse events leading to permanent medication discontinuation outcome
A total of 26 run-in trials were matched with non-run-in trials, of which 15 were matched with multiple nonrun-in trials. There was no statistically significant difference in the incidence rate of adverse events leading to permanent medication discontinuation among control groups between run-in studies (12.8 per 1000 personyears) and non-run-in studies (34.9 per 1000 personyears; incidence difference, −9.49 [95% CI, −28.1 to 9.1]; . There was no significant difference in adverse events leading to permanent discontinuation between run-in trials [fig_ref] Figure 3: Pooled ratio of relative risks and 95% CIs for run-in versus non-run-in... [/fig_ref]. Sensitivity analyses including multiple-matches analysis, bootstrapping analysis, or meta-regression did not alter the findings [fig_ref] Table 2: Summary Estimates of Relative Treatment Estimates [/fig_ref].
## Nonfatal myocardial infarction and nonfatal stroke outcome
A random-effects meta-analysis of the RRRs between run-in and matched non-run-in studies showed no [fig_ref] Table 2: Summary Estimates of Relative Treatment Estimates [/fig_ref]. Sensitivity analyses including multiple-matches analysis, bootstrapping analysis, or meta-regression did not alter the findings [fig_ref] Table 2: Summary Estimates of Relative Treatment Estimates [/fig_ref].
## Quality of reporting of run-in
Authors reported the reason for a run-in design in 48 of 66 trials (72% . Of the run-in trials, 38 (58%) reported the absolute number of exclusions during the run-in period. Individual reasons for participant exclusions during run-in was reported in 30 run-in trials (45%), with 20 trials clearly reporting the total number of exclusions during run-in attributed to adverse events during the run-in period. Placebo run-in alone was conducted in 42 trials, active run-in alone was conducted in 14 trials, and 9 trials had both an active and placebo run-in period. In 1 trial it was unclear whether placebo or active run-in was used. Loss to follow-up during the run-in period was reported in 38 trials (58%). The median reported duration of placebo run-in was 4 weeks (range, 1 week-4 months). The median reported duration of active run-in was 4 weeks (range, 1 day-3 months). The median reported percentage of patients excluded during active run-in was 13.5%. The median reported percentage of patients excluded during placebo run-in was 12.3%. Single best population, intervention, control, and outcome score analysis. Pooled ratio of relative risks for matched run-in and nonrun-in RCTs that reported mortality outcomes. "No of Studies" refers to the number of non-run-in trials that have been matched to each individual run-in trial. The size of the data markers indicates the weight of the comparison in the meta-analysis. A summary estimate <1 represents an exaggerated treatment effect of run-in. df indicates degrees of freedom; I 2 , degree of heterogeneity; Q, Cochran's Q; RE, random-effects meta-analysis; and RR, relative risk.
# Discussion
We did not find a significant difference in the relative treatment effect of run-in trials compared with matched non-run-in trials for several clinical outcomes-a composite of cardiovascular outcomes, mortality, adverse events, nonfatal myocardial infarction, and nonfatal stroke. We also failed to identify differences in adverse events leading to medication withdrawal or the proportion of participants lost to follow-up, which are expected advantages of using run-in periods in randomized clinical trials.
Our central hypothesis was that treatment effects reported in clinical trials that used run-in may be larger than those reported in trials that did not use a run-in period, which may have implications for guideline recommendations. The rationale to suspect differences is based on the contention that populations included in trials that used run-in would differ from those in nonrun-in trials, as suspected and reported by previous researchers. [bib_ref] Estimating the effect of the run-in on the power of the Physicians'..., Lang [/bib_ref] [bib_ref] External validity of randomised controlled trials: "to whom do the results of..., Rothwell [/bib_ref] In previous meta-analyses, comparisons of run-in and non-run-in trials have been confined to smaller groups of trials. [bib_ref] Physicians' health study: aspirin and primary prevention of coronary heart disease, Khaw [/bib_ref] In our larger analyses, the lower bound of the CI for RRR ranged from 0.90 to 0.94 based on approaches of bootstrapping analysis and meta-regression, meaning that a treatment difference in this range is still possible but unlikely. The only analysis reporting a significant association was our univariable meta-regression, which reported a 6% difference in RR associated with run-in but was not significant on adjusting for factors that differed between run-in and non-run-in trials (eg, sample size). By comparison, a previous analysis evaluating the effect of early stopping of clinical trials reported an RRR of 30%, 31 resulting in revision to the Grading of Recommendations, Assessment, Development, and Evaluations system for guideline recommendations, with rating down quality because of risk of bias for trials that stopped early because of efficacy. [bib_ref] GRADE guidelines: 4. Rating the quality of evidence-study limitations (risk of bias), Guyatt [/bib_ref] In our analyses, we did not find a significant difference in the RRR between trials with and without run-in periods, and our findings suggest that if a difference does exist, it is of a magnitude that should not affect the quality of evidence assessments in guideline recommendations or influence physician prescribing patterns.
We expected to find differences in clinical event rates, adverse events leading to drug discontinuation, and loss to follow-up given that a run-in period Single best population, intervention, control, and outcome score analysis. Pooled ratio of relative risks for matched run-in and nonrun-in RCTs that reported adverse events leading to discontinuation of study medication. "No. of Studies" refers to the number of nonrun-in trials that have been matched to each individual run-in trial. The size of the data markers indicates the weight of the comparison in the meta-analysis. A summary estimate <1 represents an exaggerated treatment effect of run-in. df indicates degrees of freedom; I 2 , degree of heterogeneity; Q, Cochran's Q; RE, random-effects meta-analysis; and RR, relative risk.
is intended to exclude participants at higher risk of adverse events and nonadherence with trial protocol. Although adverse events were numerically more common in non-run-in trials, it was not statistically significant, and the RR of adverse events was similar in both trial groups. These observations provide empirical evidence that relative adverse treatment effects derived from run-in trials are unlikely to be biased compared with nonuse of a run-in period. However, our findings would also appear to challenge whether a run-in period is truly associated with the intended effect on trial populations given that we did not observe significant differences in adverse event rates or loss to follow-up. Moreover, the similar rates of mortality and cardiovascular events does not support a major selection bias in cardiovascular risk in populations included in trials employing a run-in period. Against this observation, individual trial analyses of populations excluded during run-in report differences in risk and event rates, which may reflect different effects depending on the clinical population. Of note, our analysis may have been underpowered to show differences in control incidence rates, which may become more apparent and significant in a larger study. [bib_ref] Run-in periods in randomized trials: implications for the application of results in..., Pablos-Méndez [/bib_ref] Reviews of the reporting quality of published randomized clinical trials have shown inconsistent quality of reporting, and some trials failed to record participant flow clearly, which is particularly common in the stages before randomization. [bib_ref] Reporting of participant flow diagrams in published reports of randomized trials, Hopewell [/bib_ref] We also found inconsistencies in reporting the proportion of patients excluded during the run-in period and the reasons for their exclusion. Our findings are in keeping with similar findings for randomization, blinding, and attrition. [bib_ref] The quality of reports of randomised trials in 2000 and 2006: comparative..., Hopewell [/bib_ref] [bib_ref] Adequacy and reporting of allocation concealment: review of recent trials published in..., Hewitt [/bib_ref] [bib_ref] Improving the reporting of randomised trials: the CONSORT Statement and beyond, Altman [/bib_ref] Previous research has found that when improved reporting standards have been adopted, such as the Consolidated Standards of Reporting Trials checklist for randomized clinical trials, this has been associated with improved reporting of randomized clinical trials. [bib_ref] Does the CONSORT checklist improve the quality of reports of randomised controlled..., Plint [/bib_ref] [bib_ref] explanation and elaboration: updated guidelines for reporting parallel group randomised trials, Moher [/bib_ref] It would appear reasonable that the Consolidated Standards of Reporting Trials statement include requirements for standardized descriptions of the run-in periods.
The strengths of our study include a robust methodology with use of a matching strategy, resulting in a comprehensive meta-analytic examination of run-in periods among cardiovascular prevention trials. We included a large number of randomized clinical trials (n=177) that were systematically and independently reviewed across a range of journals and covering a broad period from 1990 to 2019. We chose causespecific clinical outcomes that can be classified with reasonable reproducibility, which reduces the risk of ascertainment bias. [bib_ref] Problems with use of composite end points in cardiovascular trials: systematic review..., Ferreira-González [/bib_ref] [bib_ref] Choice of clinical outcomes in randomized trials of heart failure therapies: disease-specific..., Yusuf [/bib_ref] We also completed a range of secondary analyses completed for each clinical outcome. [bib_ref] Does a placebo run-in or a placebo treatment cell affect the efficacy..., Trivedi [/bib_ref] [bib_ref] Association of run-in periods with weight loss in obesity randomized controlled trials, Affuso [/bib_ref] Limitations of our study include the number of clinical trials included in our analyses and the fact that our analyses were restricted to cardiovascular prevention trials of medicines, meaning that these findings may not be extended to other populations or interventions. However, confining our analyses to 1 clinical area may also be considered a strength as it reduced heterogeneity in clinical trial design, population, and outcome measures. In some of our analyses, a low proportion of trials were included, such as loss to follow-up. We did not include an outcome for all adverse events because of inconsistent reporting, and we were therefore unable to report on the association of run-in period with rates of all adverse events. We do report the more valid and reliable outcome of adverse event leading to drug discontinuation. Finally, although our findings have implications to all clinical trials employing a run-in period, they are only directly relevant to cardiovascular prevention trials of established preventive medications.
# Conclusions
In conclusion, the use of a run-in period was not associated with a significant difference in the magnitude of treatment effect among randomized controlled trials evaluating medications to prevent cardiovascular events. If placebo is compared with standard treatment and an additional medicationscored 2 -
If placebo is compared with standard treatment -scored 1
## Outcome
- Similar CV Composite/Primary Outcome -scored 3 -
Report multiple outcomes for mortality/stroke/MI -scored 2 -
If report one of mortality/stroke/myocardial infarction -scored 1 -
If report no similar outcome -score 0 Step 3: Intervention, Control, Outcome Matching Essential Criterion -matches must score > 1 in each
Step 2: Population Matching Essential Criterion -matches must have a population match score > 1
Step 1: Match on Drug Intervention Essential Criterion -matches must be from the same drug class Autogenerate All Potential Matches [fig_ref] Figure 3: Pooled ratio of relative risks and 95% CIs for run-in versus non-run-in... [/fig_ref]. Visual Representation of Different PICO Matching Scenarios.
Scenario 1: Highest PICO scenario Non-run-in trials were matched to their corresponding highest PICO score run-in trial(s) e.g. Trial A and Trial C with total PICO 12/12.
## Scenario 2: equal pico scenario
If several trials have equal PICO matches, we selected the trials with the largest non-run-in sample size e.g. Trial B and Trial C with run-in sample size of 8000.
## Figure s4. risk of bias overall summary for trials included in cardiovascular
Composite Outcome Analysis.
[fig] Figure 2: Pooled ratio of relative risks and 95% CIs for run-in versus non-run-in randomized clinical trials (RCTs): all-cause mortality. [/fig]
[fig] Figure 3: Pooled ratio of relative risks and 95% CIs for run-in versus non-run-in randomized clinical trials (RCTs): adverse events leading to medication discontinuation. [/fig]
[fig] Figure S2: PICO [/fig]
[table] Table 1: Characteristics of Run-In and Non-Run-In Trials Matched by Cardiovascular Composite Outcome [/table]
[table] Table 2: Summary Estimates of Relative Treatment Estimates [/table]
[table] Table S2: Run-in and Non-run-in Matching Score Considerations. Population • Main considerations: prevention type (primary vs secondary), population summary, mean age, gender match • Studies that compare primary prevention and secondary prevention populations are not a match • Chronic stable secondary prevention populations do not match with studies that include populations who recruit patients with acute events • Inclusion criteria should be similar in age, cardiovascular comorbidity population and inclusion criteria Control • Main considerations: placebo control vs standard treatment. • If both studies are placebo controlled -scored 3 • [/table]
[table] Table S3: Comparison of Placebo Run-in and Active Run-in Studies. [/table]
|
Successful Cardiopulmonary Resuscitation in a Sevoflurane Anaesthetized Horse That Suffered Cardiac Arrest at Recovery
A 17-year-old mare undergoing dental surgery suffered a cardiac arrest while being transferred from the surgical theatre to the recovery box. This complication was diagnosed early, thus allowing a prompt start to the cardiopulmonary resuscitation maneuvers. External thoracic compressions, intermittent positive pressure ventilation, and adrenaline administration were at the core of this successful resuscitation. Although it was not possible to confirm the cause of cardiac arrest in this horse, a Bezold-Jarisch reflex due to potential decrease on venous return because of postural change and drug interactions was hypothesized. Based on this report, it appears advisable to smoothly change the position of anaesthetized patient; furthermore, the administration of drugs affecting cardiovascular hemodynamics or sympatho-vagal balance to animals while changing their recumbency should be avoided.
# Introduction
A perioperative mortality rate from 0.12 to 0.9% has been reported in non-colic horses [bib_ref] Equine perioperative fatalities associated with general anaesthesia at a private practice -a..., Bidwell [/bib_ref]. Despite the advances and improvements in equine anaesthesia, mortality rate remains still high [bib_ref] Equine anaesthesia-associated mortality: where are we now?, Dugdale [/bib_ref]. A third of these fatalities is due to cardiac arrest, as reported in sick and healthy horses [bib_ref] The confidential enquiry into perioperative equine fatalities (CEPEF): mortality results of Phases..., Johnston [/bib_ref]. Severe debilitating diseases and anaesthetic drugs-related factors have been described as leading causes of cardiac arrest in horses [bib_ref] The confidential enquiry into perioperative equine fatalities (CEPEF): mortality results of Phases..., Johnston [/bib_ref]. The occurrence of cardiac reflexes following change of hemodynamic conditions may also lead to this complication.
Peri-anaesthetic cardiac arrest, despite being uncommon, carries a poor outcome in horses [bib_ref] Equine perioperative fatalities associated with general anaesthesia at a private practice -a..., Bidwell [/bib_ref] [bib_ref] Twenty years later: a single-centre, repeat retrospective analysis of equine perioperative mortality..., Dugdale [/bib_ref] , mainly because of the difficulties performing cardiopulmonary resuscitation (CPR) maneuvers in this specie. An early detection of the complication and a trained staff are key-points for the potential success of CPR. The present case aimed to describe the successful detection and treatment of a cardiac arrest occurring after a postural change during the recovery phase of anaesthesia in a horse.
# Background
Written informed consent was obtained from the owner for the publication of this case report. A 17-year-old Hanoverian mare, 550 kg, was admitted to the Equine Hospital for surgical removal of a fractured tooth (tooth #308, modified Triadan Tooth Numbering System of equine dental nomenclature). Preoperative physical examination, electrocardiography (ECG) and blood tests were unremarkable. Therefore, the American Society of Anaesthesiologists (ASA) physical status was classified as two. Due to the complexity of the surgical procedure, the mare was scheduled for tooth removal under general anaesthesia the following day.
Food was withheld for 12 h with free access to water. In the day of surgery, a 14-gauge catheter was inserted percutaneously into the left jugular vein.
Acepromazine (Calmivet R , Vetoquinol, France) 0.04 mg kg −1 was administered intramuscularly (IM), phenylbutazone (Phenylarthrite R , Vétoquinol, France) 2.2 mg kg −1 and trimethoprim-sulfadoxine (Borgal R , Virbac, France) 15 mg kg −1 were administered intravenously (IV) 1 h before anaesthesia. In the induction box, the patient was sedated with romifidine (Sedivet R , Boehringer Ingelheim, France) 0.04 mg kg −1 IV followed 5 min later by morphine (Morphine Clorhydrate Aguettant R , Aguettant, France) 0.1 mg kg −1 IV. Anaesthesia was induced with diazepam (Valium R , Roche, France) 0.05 mg kg −1 and ketamine (Imalgène1000 R , Merial, France), 2.2 mg kg −1 IV, given in separate syringes. Endotracheal intubation was performed with a 26 mm internal diameter silicone tube and the patient hoisted to the theatre and positioned in right lateral recumbency.
Once in theatre, the horse was connected to a circle breathing system and anaesthesia was maintained with sevoflurane (Sevoflo R , Axience, France) in 100% oxygen. The inspired fraction of sevoflurane was titrated to effect to maintain an adequate depth of anaesthesia based on clinical signs (palpebral reflexes, position of the eye, absence of nystagmus). A 20gauge cannula was placed in the left metatarsal artery for invasive blood pressure (IBP) monitoring and regular arterial blood sampling for blood gas analyses. Vital signs monitoring was performed with a multivariable monitor (Datex S/5, GE Healthcare, UK) and consisted in continuous ECG, heart rate (HR), oxygen saturation (SpO 2 ), inspired and expired fraction of carbon dioxide (P E ′ CO 2 ), inhaled and end-tidal concentration of oxygen and sevoflurane and IBP. Ringer Lactate (Ringer Lactate Aguettant R , Aguettant Laboratories, France) was infused at 10 mL kg −1 hour −1 . Dobutamine (Dobutamine Panpharma R , Panpharma, France) IV was administered at 2-10 µg kg −1 min −1 to effect, to maintain mean arterial pressure (MAP) above 70 mmHg. Intermittent positive pressure ventilation (IPPV) was provided using a volume cycled -pressure controlled ventilator (Stephan Respirator-GT; F. Stephan GmbH, Germany), and settings were adapted to maintain a P E ′ CO 2 between 4.6 and 6.0 kPa (35 and 45 mmHg). Additional analgesia was provided with lidocaine (Lurocaine R , Vétoquinol SA) 1.5 mg kg −1 IV administered over a 20-min period, followed by a constant rate infusion (CRI) (50 µg kg −1 min −1 ).
Trepanation for removal of the tooth #308 was performed, followed by a cleaning and a curettage of the dentary alveoli into which a silicone temporary prosthesis was placed.
Anaesthesia lasted 130 min, with a surgery duration of 75 min. No surgical complications were reported. Except for a period of 10 min after induction in which MAP values were around 60 mmHg, no other events or abnormalities on the ECG, HR or MAP were observed during anaesthesia. Ninety minutes after induction, morphine 0.1 mg kg −1 IM was administered, and the lidocaine CRI was stopped 20 min before the end of anaesthesia. Ten minutes later, the horse was weaned from the ventilator. At the end of the surgical procedure, 5 min before the end of anaesthesia, the intra-arterial catheter was removed, fluids were stopped but the IV catheter was kept for recovery. The vaporizer and oxygen were switched off; the horse was disconnected from the monitoring devices and from the breathing system but remained orotracheally intubated. While the horse was attached to a hoist and positioned on dorsal recumbency, a romifidine 0.02 mg kg −1 IV bolus was given, and the horse moved thereafter to the recovery box. Once there, the horse was positioned on right lateral recumbency on a padded floor, with the anaesthetist at its head to check vital signs and avoid premature attempts of rising. The time from the end of anaesthesia to this point was approximately 2 min. Despite the horse was breathing spontaneously before its transfer, apnoea was noticed when positioned in the recovery box. At physical examination, pulse was absent, mucous membranes were greyish, and pupils mydriatic. Cardiac auscultation confirmed the absence of cardiac beats. The time was noted, and thoracic compressions were immediately started by an operator jumping with his knees on the mare's thorax. Three persons (weighting 60, 80, and > 90 kg, respectively) rotated every 2 min to perform the external massage. The third heavier operator performed massage by rhythmically and energetically sitting on the horse's thorax. Meanwhile, 6 mg of adrenaline (Adrenaline Aguettant R , Aguettant, France) was administered IV, followed by 5 mL of heparinised saline. Mechanical ventilation was provided with a demand valve at a rate of 10 breaths min −1 , with 100 % oxygen and Ringer Lactate was administered at 10 mL kg −1 hour −1 . Vital parameters were continuously monitored by mandibular pulse palpation, eye reflexes evaluation. While the third operator was performing the external massage, mandibular pulse was detectable and synchronous to the thoracic compressions. A multiparameter monitor was then connected for ECG, HR, and non-invasive blood pressure measurements, with a cuff on the left metacarpal bone. However, due to the movements on the patient, monitoring assessment was difficult. Five minutes after the start of CPR, the anaesthetist detected a stronger mandibular pulse, asynchronous to the external massage. Thoracic compressions were stopped, and the anaesthetist confirmed the presence of normal QRS complexes on the ECG, whereas the horse was still apnoeic. Mechanical ventilation was continued with the demand valve with a respiratory rate of 6 breaths min −1 . The mare was initially tachycardic (HR of 60 beats min −1 ), with a sinus rhythm, and MAP of 80 mmHg. Within the next 10 min, HR decreased to 37 beats min −1 and MAP dropped to 50 mmHg with a poor pulse quality. Dobutamine CRI was thus administered to effect at 0.5 to 2 µg kg −1 min −1 IV for 5 min, until MAP reached 70 mmHg, and stopped thereafter. Ten minutes after the return of spontaneous circulation, spontaneous breathing reappeared. Afterwards, IPPV was stopped but oxygen supplementation was continued using a flow-by method, with an oxygen supply tubing positioned in the endotracheal tube and an oxygen flow set at 12 L min −1 . At this time, pupillary reflex was present, but not palpebral reflex. Capillary refill time was less than 2 sec and SpO 2 100%, but mucous membranes remained pale pink and sweating was present. An arterial blood gas analysis was carried out and revealed a non-compensated respiratory acidosis (pH 7.3; arterial pressure of carbon dioxide (PaCO 2 ) 67 mmHg; bicarbonate, 29 mmol L −1 ; anion gap 14 mmol L −1 , base excess 0 mmol L −1 ). Palpebral reflex and nystagmus were noticed 15 min after the return to spontaneous circulation. Ten minutes later, reflexes became stronger and the horse started presenting some movements, which were controlled by two operators at the head to avoid premature standing. The endotracheal tube was secured to the horse's mouth and all equipment was removed from the box to prepare for the recovery, which was assisted with ropes. The mare stood up at the first attempt 1 h after the start of CPR, and remained quiet thereafter, although trembling. The patient was then extubated and kept in the recovery box for close observation. Two hours later, the mare was transferred to the hospitalization box and received phenylbutazone 4.4 mg kg −1 IV and omeprazole (Gastrogard, Merial, France) 2.2 mg kg −1 , orally. Venous blood sample analysis revealed a lactate into the normal range (1.3 mmol L −1 ) and mild increased creatinine kinase (751 IU L −1 ). Neurological examination was normal thereafter: the horse was alert with normal pupillary reflexes, no apparent blindness, deafness or ataxia.
The postoperative period was uneventful. Two days after anaesthesia, a cardiac ultrasound was performed, which did not reveal any abnormality. Due to the favorable outcome, the patient was discharged from the hospital 1 week later.
# Discussion
This case reports the successful resuscitation of a 17-year-old, 550 kg mare undergoing tooth removal under general anaesthesia that suffered cardiac arrest while transferred to the recovery box.
Cardiac arrest is a complication of equine anaesthesia that has been poorly studied. Risk factors regarding perioperative mortality include an increased ASA physical status, age, surgery type, prolonged duration of anaesthesia and emergency procedure (1). In the present case, it was debatable if a 17 years old horse could be considered as geriatric. If so, the patient could have presented a decreased ability to respond to circulatory changes or stress and, therefore, at a higher risk to anaesthetic complications (5). However, despite its age, the mare was considered overall healthy with no detected systemic abnormalities and no exercise intolerance; vital signs remained remarkably stable during anaesthesia. Therefore, in this case, the occurrence of the cardiac arrest was difficult to predict but, fortunately, its early detection allowed the prompt start of CPR maneuvers.
Cardiopulmonary resuscitation aims to restore spontaneous circulation and breathing and avoid irreversible hypoxic damages to organs. The probability of success for CPR in adult horses is considered as poor [bib_ref] Cardiopulmonary resuscitation, Muir [/bib_ref]. The size of the animal and the physical effort required to provide cardiac massage render this procedure complicated to perform. In addition, the lack of advanced monitoring during recovery may delay the detection of cardiac arrest and worsen the outcome.
Although cardiac arrest involves one third of equine perioperative mortality (1), there is a lack in literature regarding its occurrence and treatment in adult horses. Successful CPR after direct cardiac massage in a pony and a horse was reported by De Moor et al. [bib_ref] Intrathoracic cardiac resuscitation in the horse, De Moor [/bib_ref] , however, both animals died in post resuscitation period. Hubbell et al. [bib_ref] Cardiovascular effects of thoracic compression in horses subjected to euthanasia, Hubbell [/bib_ref] evaluated the effects of thoracic compression rate on cardiac output in horses with induced cardiac arrest. They reported that thoracic compressions at a rate of 80 compressions min −1 allowed a better cardiac output, in comparison with lower rates. Moreover, cardiac output was higher when the operator was heavy. In this last study, horses were on right lateral recumbency and the operator delivered a blow to the chest wall immediately posterior to the left elbow with his knees. In the present case, thoracic compressions were performed in a similar way. Despite the aim was to perform 80 compressions min −1 , this rate seemed very difficult to achieve in practice and the rate observed in our case was probably closer to 40 to 60 compressions min −1 . The third operator was the heaviest and most experienced surgeon; instead of compressing the horse's thorax with his knees, he used his whole core body by sitting on it. A better pulse quality was subjectively achieved with this way of performing the external massage, but it could also have been attributed to the heavier weight of the operator.
In addition to these physical maneuvers, adrenaline was administered to the horse. This drug is recommended for asystole in horses [bib_ref] Cardiopulmonary resuscitation, Muir [/bib_ref] and small animals [bib_ref] RECOVER Advanced Life Support Domain Worksheet Authors. RECOVER evidence and knowledge gap..., Rozanski [/bib_ref]. Adrenaline is a synthetic catecholamine with strong α 1 -and β 1 -, and moderate β 2adrenergic receptor activity which produces vasoconstriction and an increase in HR and contractility [bib_ref] Drugs for cardiovascular support in anaesthetized horses, Schauvliege [/bib_ref]. In the present case, a single low dose (0.01 mg kg −1 ) IV was used. Despite it was difficult to evaluate which part of the resuscitation maneuvers contributed to the return of spontaneous circulation, it was probably the combination of both, thoracic compressions and adrenaline that contributed to the successful CPR.
In addition to the external massage, the basic life support consists in ventilation. As previously described [bib_ref] Sudden cardiac arrest in an anaesthetised horse associated with low venous oxygen..., Mcgoldrick [/bib_ref] , IPPV using a demand valve was performed early during CPR, at a rate of 6-10 breaths min −1 . Even though we cannot be certain of the minute ventilation provided, it probably allowed sufficient oxygenation of the animal, as no clinical signs of hypoxic brain damage were noticed thereafter.
In this case, the lack of close monitoring during the horse transfer made difficult to determine the precise moment and the real cause of the cardiac arrest. It was unlikely to be related to the animal health status, although individual idiosyncrasy could not be excluded, but may have been due to the surgical procedure, the occurrence of cardiovascular reflex or drug effect. Firstly, there was a potential risk of embolism associated with bleeding. However, no bleeding was noticed, nor was any sudden variation of P E ′ CO 2 during the procedure. Second, a cardiac reflex following a postural change could be considered in this case. Severe cardiovascular depression in similar circumstances has been reported in a dog [bib_ref] Potential bezold jarisch reflex secondary to a 180 • postural change in..., Mcmillan [/bib_ref] and in humans [bib_ref] Postural changes in the peripheral blood flow of normal subjects with observations..., Brigden [/bib_ref] [bib_ref] Bezold-Jarisch reflex caused by postural change, Kim [/bib_ref]. In this case, the horse experienced a rapid postural change for its transfer to the recovery box. This could have produced a compression of the caudal vena cava by abdominal viscera, leading to a decreased venous return for a few seconds, the so-called Bezold-Jarisch reflex (BJR). The BJR is a complex neuro-cardiogenic reflex mediated by ventricular receptors sensitive to chemical and / or mechanical stimuli in response to a decreased ventricular filling [bib_ref] The bezold jarisch reflex revisited: clinical implications of inhibitory reflexes originating in..., Allyn [/bib_ref] , that is not necessarily related to a hypovolaemic state. The afferent pathway is vagally mediated and terminate at the nucleus tractus solitary in the central nervous system [bib_ref] Bezold-Jarisch reflex attenuates dynamic gain of baroreflex neural arc, Kashishana [/bib_ref]. The efferent response produces an increased parasympathetic tone, leading to bradycardia, vasodilation and apnoea [bib_ref] Role of cardiac vagal C-fibers in cardiovascular control, Thoren [/bib_ref]. In addition to a postural change, an interaction of a variety of anaesthetic drugs may trigger this reflex [bib_ref] Perioperative bradycardia and asystole: relationship to vasovagal syncope and the Bezold-Jarisch reflex, Kinsella [/bib_ref].
Therefore, the drugs used in the present case may also have participated to the observed complication. After its use for premedication without noticeable adverse effect, romifidine 0.02 mg kg −1 IV was administered at the time the horse was attached to the hoist. In our practice, this drug is routinely administered at the end of anaesthesia to improve the quality of recovery of horses [bib_ref] Effects of postanesthetic sedation with romifidine or xylazine on quality of recovery..., Woodhouse [/bib_ref]. As an α 2 -agonist, romifidine produces an initial increase in blood pressure with a subsequent bradycardia; second degree atrioventricular blocks and a decrease in stroke volume are commonly reported [bib_ref] The effects of detomidine, romifidine or acepromazine on echocardiographic measurements and cardiac..., Buhl [/bib_ref]. Despite cardiovascular effects of romifidine may be dose dependent, to the authors' knowledge there are no studies that evaluated the cardiovascular effect of low dose of romifidine in anaesthetized horses. Even though it seems unlikely that romifidine alone would have been responsible for this complication, we cannot exclude a role in the development of the cardiac arrest.
Among the other drugs administered, acepromazine, a phenothiazine derivate and α 1 -adrenoreceptor antagonist, was used in the premedication. In addition to hypotension (21, 22), phenothiazine derivates have been associated in humans to a decrease in myocardial contractility [bib_ref] Effects of myocardial alpha 1-adrenergic receptor stimulation and blockade on contractility in..., Landzberg [/bib_ref] however, this has not been reported in horses. Conversely, it has been associated to a protective cardiac effect and a decreased anaesthetic mortality [bib_ref] The confidential enquiry into perioperative equine fatalities (CEPEF): mortality results of Phases..., Johnston [/bib_ref]. Even though it may have a prolonged effect on blood pressure, it was unlikely to have promoted the cardiac arrest as blood pressure was overall well maintained during anaesthesia in this case. Sevoflurane was used for maintenance of anaesthesia. In healthy horses, it produces a dose dependent decrease in MAP, cardiac index and systemic vascular resistance (SVR) [bib_ref] Effects of sevoflurane dose and mode of ventilation on cardiopulmonary function and..., Steffey [/bib_ref]. This decrease in SVR could have contributed to worsen the potential decrease in venous return hypothesized in this case. Similarly, lidocaine, a local anaesthetic commonly used in equine anaesthesia, may produce bradycardia and hypotension at high doses [bib_ref] Local anaesthetics, Rioja [/bib_ref] and may have participated to the altered cardiovascular response to the postural change, even though it was stopped 20 min before the end of anaesthesia. Dobutamine, a β 1 -adrenoreceptor agonist with inotropic properties, commonly used for treating hypotension in anaesthetized horses, was also associated with the occurrence of BJR in a dog [bib_ref] Dobutamine-induced bradycardia in a dog, Hoffmeister [/bib_ref]. As it was sparely used in this case and particularly not at the time of recovery, it was unlikely that it had any influence on the onset of the reported complication. Finally, fatal reactions to trimethoprim-sulfadoxine have been reported in anaesthetized horses [bib_ref] Possible potenciated sulphonamide-associated fatality in an anaesthetised horse, Dick [/bib_ref] [bib_ref] Trimethoprim/sulfonamide combination in the horse a review, Van Duijkeren [/bib_ref]. However, in this case, trimethoprim-sulfadoxine was administered 1 h before anaesthesia, which made it unlikely as a cause of the adverse event observed.
Based on the chronology of events, we hypothesized that the postural change combined with the cardiovascular effects of the different anaesthetic drugs used triggered the occurrence of a cardiac arrest, due to a possible BJR. However, the absence of monitoring during the transfer of the animal rendered difficult to confirm this assumption. The early detection of the complication and the presence of trained staff allowed a successful outcome. The present case also underlined that the equine patient is at high risk during the recovery period not only regarding the risk of trauma but also regarding the risk of cardiovascular and respiratory instability.
## Concluding remarks
The early detection of cardiac arrest and start of CPR maneuvers permitted the successful resuscitation of a 17-year-old mare that suffered cardiac arrest during her transfer to the recovery box.
In the light of this report, a continuous palpation of the peripheral pulse should be performed during the horse transfer to the recovery box. It seems advisable to administer drugs with a potential depressive cardiovascular effect in a time frame that does not overlap with postural changes.
# Author contributions
CC and SJ participated in the development of the case and wrote or contributed to the writing of the manuscript.
Frontiers in Veterinary Science | www.frontiersin.org
June 2018 | Volume 5 | Article 138
ACKNOWLEDGMENTSAuthors would like to thank Professor Michael Schramme for his assistance in the development of this case report.Conflict of Interest Statement:The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.Copyright © 2018 Conde Ruiz and Junot. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. |
Examining the Factor Structure of a Risk Assessment Inventory in Young Offenders: FER-R, Risk and Resource Assessment Form
Citation: Alarcón, P.; Pérez-Luco, R.; Chesta, S.; Wenger, L.; Concha-Salgado, A.; García-Cueto, E. Examining the Factor Structure of a Risk Assessment Inventory in Young Offenders: FER-R, Risk and Resource Assessment Form. Int. J. Environ. Res.
# Introduction
The study of social behavior during adolescence, and especially adolescent-offender evaluation, is a special challenge; it offers a window of opportunity to enhance their development and psychosocial well-beingand, likewise, allows us to recognize multiple variables involved in the onset and persistence of antisocial behaviors in this phase of the life cycle.
The increase in evidence-based practices and the modernization of justice systems in different countries, and recently in Ibero-America, has led to the incorporation of riskassessment tools to support judicial decision-making and the management of intervention programs.
Risk-assessment tools have evolved from the first and second generation, focused almost exclusively on the prediction of recidivism, towards tools that, on the one hand integrate structured professional judgment, and on the other, concentrate on intervention management (third and fourth generation). For this, the central objectives in the measurements are the dynamic factors, susceptible to change, conceptualized as the youth's criminogenic needs, with these being the key focal points of the interventions.
Risk-assessment tools have diversified and offer important advantages over traditional clinical evaluations; they serve to establish evidence-based practices both in specialized In the Hispanic population, a high degree of concordance has been reported in YLS/CMI and SAVRY measurements for predicting recidivism; however, the SAVRY also measures protective factors, and these show a negative relationship with all YLS/CMI factors.
Viljoen et al.indicate that the SAVRY measures protective factors, but its measurement is more sensitive to recognizing their absence than their variability in predicting non-recidivism.
In summary, in the Latin American context, there is little development and application of risk-assessment instruments in adolescents, and if they are applied, they do not have an exhaustive psychometric study, with follow-up and verification of evidence of validity to give robustness to the measurements and guide the management and decision-making of the intervention.
Based on the previous background, it is imperative to study the risk-assessment tools used with Latin American adolescents, particularly in Chile, and to consider the conditions described in the evidence as a priority, integrating criminogenic risk factors and protective factors to manage intervention programs with young offenders, as well as to support with robust psychometric evidence the scales that are most sensitive to the needs of local adolescents. Responding to these needs, the present study shows empirical evidence of the validity of a structured professional judgment inventory, developed and studied in Chile.
The Risk and Resource Assessment Form, FER-R (its Spanish acronym, from "Ficha de Evaluación de Riesgos y Recursos"), is a rationally constructed inventory based on evidence from the RNR model and developmental criminology, together with direct supervisory experience of professional teams intervening with Chilean young offenders since 2002. It consists of 8 factors and 57 items distributed in 2 domains: Risks and Resources. The first factor (interventions) assesses static risk (non-modifiable), the next five (education, peers, family, drugs, and attitudes) assess dynamic risks, susceptible to change; and the last two (protective resources and youth's interests) assess personal and contextual adaptative resources that act as protective factors. The FER-R has a 24-month follow-up study of 100 adolescents, which shows overall predictive validity in 68.3% of cases and the effect size for predicting recidivism is robust (AUC = 0.73), with a 95% confidence interval (0.63-0.83). It also has an inter-rater reliability studyby Intra-class Correlation Coefficient analysis, obtaining a CCI value = 0.858 and reliability by internal consistency with satisfactory indices for risks (alpha 0.70-0.98) and for resources (alpha 0.80-0.94).
These results show an adequate and robust predictive utility of the FER-R for criminal recidivism, and if the magnitude of the partial prediction is considered, three dynamic criminogenic risk factors of greater prediction are highlighted: (1) education (AUC = 0.74), and specifically progressive school disengagement (AUC = 0.78); (2) peers with high criminal commitment (AUC = 0.71); and (3) family (AUC = 0.66), in particular the weakness of parental supervision (AUC = 0.66); drugs and negative attitudes have a lower predictive weight (AUC = 0.63). In resources, similar indices are obtained. The youth's interests possess good predictive capacity (AUC = 0.67) of non-recidivism (desistance). Finally, the FER-R has a characterization studywith 486 adolescents, which demonstrates the capacity of the inventory to discriminate according to sex (male and female), type of delinquency exhibited (persistent or transitory), and criminal recidivism detected two years after the assessment.
The FER-R has also been used to evaluate the impact of family interventions in treatment programs for adolescents with delinquent behavior. It adds a greater number of items to the family risk factor, which is very relevant as it delves into an area with distinctive cultural particularities of the Latin American context that may affect adolescents differently. All of the above add evidence on the impact that this inventory can have for the design of pertinent intervention programs.
The general objective of this study was to examine the psychometric properties of the FER-R in young offenders in Chile, and the specific objectives were: (a) to examine its factorial validity, (b) to examine convergent, divergent, and discriminant validity, (c) to determine reliability, and (d) to comparatively characterize the presence of risks and resources in the participants of the sample studied.
# Materials and methods
## Participants
The study population was young people convicted under the Adolescent Criminal Responsibility Law in Chile, enrolled in intervention programs in four regions of the central-south of the country between 2008 and 2012 (The data correspond to two different studies financed by the Chilean National Council for Science and Technology -CONICYT, currently ANID-, FONDECYT project Nº 1070397, and FONDEF project D08i-1205. The adolescent population in Chile for the 15-19 age cohort in 2010 was estimated to be 1,426,634, and 11,256 were convicted between 2008 and 2012. A total of 649 adolescent response protocols were selected by non-probabilistic convenience sampling in two research stages; in the first stage, 263 valid protocols were obtained from male adolescents in 2008, and in the second stage, 386 protocols were obtained from male (326) and female (60) adolescents between 2011 and 2012 distributed according to type of sanction in the four study regions (see. The time period in which the data were collected includes two generations of adolescents. However, both groups of study participants were equivalent in age (t = −0.11; p = 0.913) and schooling (t = −1.25; p = 0.212) and their proportional distributions did not differ in rural or urban origin (χ 2 = 3.015; p = 0.221) and Mapuche or non-Mapuche origin (χ 2 = 0.090; p = 0.765). All of them met the inclusion criteria: (a) criminal record in judicial files; (b) being serving a sanction, and; (c) voluntariness in their participation, manifested by signing an informed consent form.
## Design
Two sub-samples were taken from the total sample with complete protocols for the different analyses: 633 for factorial validity and internal consistency, and 442 for convergent, divergent, and discriminant validity, using an instrumental designfor the final resolution of the study.
## Instruments
Risk and Resource Assessment Form, FER-R, the acronym in Spanish for "Ficha de Evaluación de Riesgos y Recursos", is an inventory of structured professional judgment that measures criminogenic risks, static and dynamic, along with adaptation resources in young offenders. It was developed in Chile by Paula Alarcónand has previously published evidence of predictive reliability and validity. It consists of 57 items, 39 that measure criminogenic risks (0-39 points), and 18 that measure adaptation resources (0-18 points). The scores on the FER-R are expressed in eight areas: (F1) impact of previous interventions (convictions); (F2) education, which includes the domains F2a, school disengagement, and F2b, behavior problems; (F3) relationship with peers; (F4) family, which includes the domains F4a, weak supervision, F4b, approval of criminal behavior, and F4c, abuse; (F5) drugs; (F6) clear attitudes or tendencies; (F7) personal and family protective resources; and (F8) the youth's interests. The FER-R is scored by presence (1) or absence (0) and has a structured coding guide.
In the risks domain, the scores are standardized as low, moderate, and high in each facet, creating a risk profile that can identify the criminogenic needs of greatest urgency for the intervention (see; and the total risk index that varies between 0 and 39 is estimated with cut-off points categorized as low-risk, <11, moderate-risk, 12-17, and high-risk >18, values derived from a two-year follow-up study. different analyses: 633 for factorial validity and internal consistency, and 442 for convergent, divergent, and discriminant validity, using an instrumental designfor the final resolution of the study.
## Instruments
Risk and Resource Assessment Form, FER-R, the acronym in Spanish for "Ficha de Evaluación de Riesgos y Recursos", is an inventory of structured professional judgment that measures criminogenic risks, static and dynamic, along with adaptation resources in young offenders. It was developed in Chile by Paula Alarcónand has previously published evidence of predictive reliability and validity. It consists of 57 items, 39 that measure criminogenic risks (0-39 points), and 18 that measure adaptation resources (0-18 points). The scores on the FER-R are expressed in eight areas: (F1) impact of previous interventions (convictions); (F2) education, which includes the domains F2a, school disengagement, and F2b, behavior problems; (F3) relationship with peers; (F4) family, which includes the domains F4a, weak supervision, F4b, approval of criminal behavior, and F4c, abuse; (F5) drugs; (F6) clear attitudes or tendencies; (F7) personal and family protective resources; and (F8) the youth's interests. The FER-R is scored by presence (1) or absence (0) and has a structured coding guide.
In the risks domain, the scores are standardized as low, moderate, and high in each facet, creating a risk profile that can identify the criminogenic needs of greatest urgency for the intervention (see; and the total risk index that varies between 0 and 39 is estimated with cut-off points categorized as low-risk, <11, moderate-risk, 12-17, and highrisk >18, values derived from a two-year follow-up study. In the resources domain, the score is also standardized, but independently for each facet. The graph generated for protective resources goes from 0 to 13 points, with a cutoff score of 8, and the graph for the youth's interests from 0 to 5 points, with a cut-off score In the resources domain, the score is also standardized, but independently for each facet. The graph generated for protective resources goes from 0 to 13 points, with a cut-off score of 8, and the graph for the youth's interests from 0 to 5 points, with a cut-off score of 3 (see. This allows a differentiated consideration in the design of interventions, using interests for the activation of protective resources and risk reduction. of 3 (see. This allows a differentiated consideration in the design of interventions, using interests for the activation of protective resources and risk reduction. The YLS/CMI is also an inventory of structured professional judgment, derived from a tool used in the adult judicial system, the LSI-R, which has vast evidence in different countries to evaluate criminogenic risks. It consists of 42 items and is divided into eight subscales: previous and current convictions, family situation and parental role, education and employment, relationships with peers, substance abuse, free-time use, personality and behavior, and attitudes/tendencies. Each item is scored as present or absent, giving a total score between 0 and 42 points. According to this total score, the adolescents are categorized into four recidivism risk levels: low, moderate, high, or very high. The YLS/CMIhas an adaptation in Chile derived from its French-Canadian version, the IRNC, and a version revised and adapted through a preliminary validity study in Chile, YLS/CMI.
The MACI, the Millon Adolescent Clinical Inventory, is a rationally constructed, clinical self-report instrument to assess personality patterns and difficulties inherent in adolescence. It is comprised of 160 true-false items that make up a total of 31 scales, grouped into 12 personality patterns, 8 expressed concerns, and 7 clinical syndromes. In addition, it has three control scales and one validity scale. Its psychometric properties are described by its author for adolescents between 13 and 19 years. In Chile, it is a widely studied instrument and has a version adapted to Chilean standards.
## Procedure
The assessment of the FER-R and YLS/CMI inventories was carried out by the professionals responsible for each case within the intervention programs, responding simultaneously to both inventories, for which each professional received a minimum of 16 h of direct training in risk-assessment methodology and in the assessment guides of each instrument, and were subsequently supervised by the researchers to review their scores and sources of information (at least three). The MACI was applied by the same professionals, The YLS/CMI is also an inventory of structured professional judgment, derived from a tool used in the adult judicial system, the LSI-R, which has vast evidence in different countries to evaluate criminogenic risks. It consists of 42 items and is divided into eight subscales: previous and current convictions, family situation and parental role, education and employment, relationships with peers, substance abuse, free-time use, personality and behavior, and attitudes/tendencies. Each item is scored as present or absent, giving a total score between 0 and 42 points. According to this total score, the adolescents are categorized into four recidivism risk levels: low, moderate, high, or very high. The YLS/CMIhas an adaptation in Chile derived from its French-Canadian version, the IRNC, and a version revised and adapted through a preliminary validity study in Chile, YLS/CMI.
## Qualitative
## Indicators of presence of
The MACI, the Millon Adolescent Clinical Inventory, is a rationally constructed, clinical self-report instrument to assess personality patterns and difficulties inherent in adolescence. It is comprised of 160 true-false items that make up a total of 31 scales, grouped into 12 personality patterns, 8 expressed concerns, and 7 clinical syndromes. In addition, it has three control scales and one validity scale. Its psychometric properties are described by its author for adolescents between 13 and 19 years. In Chile, it is a widely studied instrument and has a version adapted to Chilean standards.
## Procedure
The assessment of the FER-R and YLS/CMI inventories was carried out by the professionals responsible for each case within the intervention programs, responding simultaneously to both inventories, for which each professional received a minimum of 16 h of direct training in risk-assessment methodology and in the assessment guides of each instrument, and were subsequently supervised by the researchers to review their scores and sources of information (at least three). The MACI was applied by the same professionals, but was answered directly by the adolescents through self-report in the intervention context.
# Ethical safeguards
The data acquisition adhered to the fundamental principles of bioethics (autonomy, beneficence, non-maleficence, and justice), and was reviewed ethically by the Chilean National Council for Science and Technology (CONICYT), which awarded and funded the projects in a national public competition. The implemented ethical protocol involved, prior to the data collection, managing the authorizations of the entity responsible for the execution of penal sanctions of adolescents in Chile, and also of the leadership of the organizations participating in the study and in charge of the direct intervention with young people. The University signed formal research collaboration agreements with each of them (The three national public institutions responsible for juvenile justice in Chile: National Service for Children, SENAME, Ministry of Justice, MINJU and Undersecretary for Crime Prevention, SPD; and two collaborating State agencies responsible for the implementation of sanctions in the study regions: Land of Hope Foundation, FTDE and Children's Defense Council, CODENI). Given that the assessments were always performed by the professionals responsible for the intervention in each case, each was asked to sign a confidentiality agreement. The evaluated adolescents were invited to participate in the study within the framework of the sentence they were serving, and signed a consent form in which they agreed to participate in the study. Finally, an anonymized database was consolidated that was safeguarded, ensuring confidentiality for the analysis under the responsibility of the principal investigators.
# Data analysis
For the examination of the internal structure of the FER-R, two confirmatory factor analyses were performed with a bifactor model (CFA-BF) using the MPLUS program version 7.3and the method of mean-and variance-adjusted weighted least squares (WLSMV) with the polychoric correlation matrix. The following were used to evaluate the fit of the model: comparative fit index (CFI), Tucker-Lewis index (TLI), and root mean square error of approximation (RMSEA). An adequate fit of the model considered CFI and TLI values greater than 0.90 and RMSEA values equal to or less than 0.08. Once the factorial structure had been determined, the reliability was estimated for each domain.
To obtain evidence of convergent validity of the FER-R, the nonparametric Spearman's Rho correlation with the YLS/CMI was estimated; for divergent validity, the total risk score of the FER-R was correlated with the same test with two personality-pattern scales of the MACI Inventory in its version adapted in Chile, (e4) dramatization and (e5) egocentrism, which measure general characteristics of most adolescents without constituting a construct related to criminogenic risks. For discriminant validity, the sample was divided into recidivists and non-recidivists with a retrospective criterion, using as a measure the presence of previous offenses (with and without antecedents) and estimating mean differences (Student's t) and effect size (Cohen's d) for the contrast between both groups. These analyses were performed using SPSS 23 software (IBM, Chile, License UFRO) and the online calculator of Lenhard and Lenhard.
# Results
## Evidence of validity
## Based on the internal structure
To evaluate the fit of the factorial structures of the risks and resources domains on the FER-R inventory, the scores of 633 participants were taken, with which the CFA-BF was performed, with the 39 risk items and then separately with the 18 resource items. In both cases, unidimensional, oblique, and bifactor models were estimated using the WLSMV estimator to account for the ordinal data.
In the risks domain, the unidimensional structure was confirmed with a bifactor model of 39 items, distributed in six facets: (i) interventions, 4 items, (ii) education, 7 items, (iii) peers, 5 items, (iv) family, 13 items, (v) drugs, 6 items, and (vi) attitudes, 4 items. The CFI was = 0.965, the TLI was = 0.961, and the RMSEA was = 0.036 (0.033-0.040).
For the resources domain, the unidimensional structure was confirmed with a bifactor model of 18 items and 2 facets: (vii) protective resources, 13 items, and (viii) the youth's interests, 5 items, where the CFI was = 0.978, the TLI was = 0.971, and the RMSEA was = 0.042 (0.035-0.049).
As seen in, the indicators show a good fit between the conceptual framework and the observed data for both models, which is illustrated in. This evidence confirms that the FER-R has a conceptually consistent internal structure with the empirically obtained measurement model.
## Based on the relation to other variable: convergent validity
The convergent validity of the FER-R was estimated with the Youth Level Service Case Management Inventory in its Chilean version, YLS/CMI. Using the SPSS 23 program, the two domains and eight facets of the FER-R were correlated with the eight indicators on the YLS/CMI, obtaining values of significant association between the two inventories in the eighty intersections, all with p < 0.001. The intersection of total risks on both inventories stands out, with a Rho = 0.919. All the values correlate in a statistically significant way and in the expected direction, i.e., positively between risk factors and negatively between resources on the FER-R and the different risks measured by the YLS/CMI (see. The findings reveal a strong association between the variables measured by the two instruments, with moderate-to-high effect sizes. These results are important evidence of the convergent validity of the FER-R when using the most disclosed test in the scientific literature as a criterion.
## Based on the relation to other variables: divergent validity
The divergent validity was estimated by correlating the total risk score from the FER-R with two personality-pattern scales from the MACI, not theoretically related to criminogenic risks: (e4) dramatization and (e5) egocentrism. The values of association in both cases tend to be negative and not statistically significant (p > 0.05), which shows the absence of an empirical relation between the evaluated constructs (see. Discriminant validity was estimated by dividing the sample into two groups according to the criterion of retrospective criminal recidivism (presence or absence of judicial antecedents prior to the current sanction) and then comparing the means of both groups in protective resources, youth's interests, and criminogenic risks using Student's t statistic, and finally estimating the effect size of the observed evidence using Cohen's d (see.
## Evidence of reliability
The reliability was obtained with Cronbach's alpha and Omega Hierarchical (ωH), with both being estimated for the two domains. In the risks domain, a ωH = 0.870 and an alpha = 0.847 were obtained, and in the resources domain a ωH = 0.716 and an alpha = 0.839 were obtained. The results in reliability are satisfactory and can be interpreted unidimensionally, but a deeper analysis of the internal consistency of each facet is required.
## Comparative characterization of the sample
An exploratory analysis of the scores obtained by males and females in the sample was performed on the risks and resources domains, comparing the medians of each group using the nonparametric Mann-Whitney U test. As noted in, the results show statistically significant differences in only two of the eight facets that the inventory measures: interventions (p < 0.05) where the males scored higher than the females, but with a Cohen's d value that means absence of effect; and the youth's interests (p < 0.01), where the females scored higher, but with a small effect size.
# Discussion
This study examined the behavior of the FER-R as a risk-and-resources-assessment inventory for a group of 633 Chilean adolescents aged 14 to 19 years, all convicted of crimes. The study of the psychometric properties confirmed that the FER-R is a valid and reliable instrument for the measurement of criminogenic risk factors and protective resources in young offenders .
In relation to the need for evidence of factorial validity for the risk-assessment tools, it was of special interest to recognize the unidimensionality through the bifactor model in the factorial structure of the FER-R for the risks and resources domains. A general factor was determined that combines a static risk facet with five dynamic risk facets was determined in the risks domain. As Mei et al.indicate, examining the construct validity in risk tools is imperative, if in practice they are treated as if they existed without demonstrating it; therefore, this study advances with evidence of internal validity for this construct. Furthermore, the FER-R, being an instrument built in a Latin American context, allows progress in the cultural relevance of these evaluations, especially due to the greater relevance of the family-risk dimension, being able to monitor its factorial structure in different groups and countries of the region if this study is replicated with diverse samples.
In the resources domain, unidimensionality is also confirmed; however, the reliability is slightly lower than in risks. Abbiati et al.indicate that there is little consensus as to how protective factors are operationalized and this may produce less internal consistency, but the FER-R advances the assessment of these factors through its two facets -protective resources and the youth's interests-with differentiated items, which can act in association with an overall response of adaptation resources.
The reports on instruments that integrate the perspective of risk and resources in their predictive capacity, as described by Viljoen. Et al.,, are promising in the field of youth justice. Consequently, the evidence of external validity with the YLS/CMI shows a strong association with moderate-to-large effect sizes, both in a positive association for the domain of risks and in a negative one with resources, coinciding with the negative association of risk factors in the YLS/CMI and protective factors of the SAVRY reported by Ortega-Campos et al. (2020). This consolidates a tool that is consistent with the construct of the RNR model. This evidence goes together with the predictive validity reported previously on a sample of 101 adolescent offenders that showed a robust effect size (AUC = 0.73), with a 95% confidence interval (0.63-0.83), comparable to the data reported by Olver et al.and the JAIS system reported by Baird et al..
Finally, it is necessary to highlight that the characterization of the risks and resources of the studied sample did not allow the detection of diversity or differences according to gender or age groups. This can be explained by the limitations of this study, with a higher representation of male adolescents and a very small sample of female adolescents, in addition to a heterogeneous distribution of age groups, given the selectivity of sanction programs, with a higher concentration in adolescents over 16 years of age. To obtain more evidence of validity, it is suggested that the size and statistical power of the sample is increased, follow-up in repeated measures is provided, data in other Latin American countries are collected, and analyses are replicated with more-recent samples, as the data were collected a decade ago. This is a limitation of the study given the far-reaching changes that Chilean society has undergone in recent years that could affect younger generations.
# Conclusions
It can be concluded that the FER-R is a state-of-the-art inventory, based on structured professional judgment, with solid evidence of internal and external validity, which simultaneously includes the evaluation of criminogenic risks and protective factors, and critical aspects considered for assessment instruments of risks (Fazel & Wolf, 2018). From a practical perspective, the determination of the risk profile and the outlining of resources for the intervention are an aid in the targeting of actions according to the criminogenic needs of each case, which facilitates the management of differentiated interventions.
This study is a pioneer given the size and Latin American origin of the sample, and provides evidence that supports the FER-R as a valid instrument for the measurement of criminogenic risk factors and protective resources in Chilean adolescents, the results being encouraging for its international transfer to other Latin American countries. However, it is necessary to increase the size of the sample of females to more accurately observe its functioning with this group. In addition, there is concern about the temporal stability of their structure given the high sensitivity of adolescent behavior to social changes, which requires replicating the study with more-contemporary samples. Finally, it is very important to study the relationship between the variables measured by the FER-R with other variables of the adolescents' vital context such as emotional regulation, mental health, adverse experiences in childhood and, especially, the presence of criminal organizations in the living environment during an adolescent's development. 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 due to privacy restrictions. |
Clinical Researches on Auscultation of the Respiratory Organs, and on the First Stage of Pulmonary Phthisis
## 311
of the inspiratory; the latter often falls to 4 or 5, while the former retains its normal value, 2. But if the cause of the depression of the inspiratory murmur continue in action beyond this point?in cases of pleuritic effusion, for instance,?the expiratory sound gradually falls to 0. In respect of the relation subsisting between the two murmurs in the morbid state, our author asserts the accuracy of the following general propositions. All affections which produce a simultaneous and proportional decrease in the intensity and duration of inspiration, are attended with diminished expiration; for example, pneumonia, pleurisy, pulmonary oedema, and apoplexy. On the other hand, when the intensity and duration of inspiration undergo opposite modifications, the one of increase, the other of diminution, the expiratory murmur is augmented; of this fact, pulmonary emphysema and tuberculization supply illustrations. A parallel instituted between the morbid variations of the two murmurs, discloses a curious opposition in the laws regulating them; a few of these may be stated, as indicating the general character of the whole. The intensity and duration of the expiratory sound almost always rise or fall coevally; these characters frequently undergo opposite changes in inspiration. Modifications of quality belong more especially to expiration ; those affecting the properties of dryness and softness, to inspiration. Diminution, affecting both murmurs, appears always first, and advances to its extreme degree more rapidly in inspiration than in expiration; in the return to the healthy standard the expiratory, on the contrary, precedes the inspiratory sound. In the particular facts related regarding each murmur, we have furnished the reader with materials for extending the parallel. The exaggerated importance attached by Laennec to the distinction of morbid vocal resonance, to pectoriloquy and aegophony, for instance, as signs of excavation in the substance of the lung, and of the presence of a thin stratum of liquid in the pleura, is matter of daily demonstration. M. Fournet, who with even greater boldness than his predecessors affirms that it is in some cases impossible to determine whether a given ringing of the voice be segophony, bronchophony, or pectoriloquy, has judiciously arranged the arguments proving the comparative futility of their distinction. We recommend this chapter to that rather numerous class of practitioners, who somewhat perversely cling to the correctness of Laennec's doctrine,?to persons who fancy themselves justified in invariably distinguishing hepatization from effusion, by the absence or presence of an segophonic twang in the vocal resonance, or predicate the existence or non-existence of caverns from that of pectoriloquy. But while we express ourselves thus strongly against the habit of trusting mainly to these signs, more especially to pectoriloquy, we by no means deny their occasional value, and dispute not their general importance as allied signs. The accurate information derived from certain forms of pectoriloquy, has been satisfactorily pointed out by Dr. Williams; and, as M. Fournet observes, the strange, mysterious, ventriloquous sound produced in caverns, when the laryngeal voice is reduced by the progress of the tuberculous disease to a mere whisper, can hardly be confounded with any other known phenomenon.
The augmentation of natural adult respiration, termed puerile by Laennec, because resembling that of children; supplementary by Andral, [April, from its making up for diminished energy in some other part of the lungs; undergoes, upon inadequate motives, a second change of title in M. Fournet's volume: the idea of the new epithet, exaggerated, is contained in Andral's term, which, besides, marks the most important signification, practically speaking, of such respiration. The chief character of this form of respiration is an increase of intensity and duration of both murmurs, proportionally greater in expiration, and accompanied with a trace of clear or blowing quality. The comparatively greater increase of the second murmur, and the quality of both sounds in supplementary respiration, make it difficult to distinguish this form from cases in which the expiratory murmur has singly undergone a morbid increase. This difficulty, as we shall hereafter show, arises in the first stage of phthisis. In emphysema it may also be encountered, and the distinction is hardly to be made without the aid of other signs, e. g. hardness and dryness, in the murmurs. In truth it seems highly probable that, as M. Fournet hints, supplementary respiration, and the state characterized by prolonged expiration, are one and the same, fundamentally, and acquire their distinctive peculiarities from the combination of other changes with the latter.
Supplementary respiration is sometimes of no mean importance as a diagnostic sign; of this fact, strikingly remarkable in certain cases of pneumonia, M. Fournet gives a useful illustration. In three individuals expectorating the rusty viscid sputa, and exhibiting the general symptoms of pneumonia, the ordinary physical signs could not be detected. The only unusual phenomenon was marked supplementary respiration on one side of the chest; in two of these cases the ordinary signs subsequently appeared in the situation where the augmented respiration had existed; in the third, the latter phenomenon disappeared with the rusty sputa and general symptoms. Here were indubitable examples of central inflammation ; in the latter instance, persisting to the end in its original situation; in the two former, spreading to the periphery of the lung. In cases where the characteristic sputa are wanting, as well as the usual physical signs,?an exceedingly rare coincidence however,? M. Fournet's observation may, we doubt not, be found of utility. Allied with this subject is the important question of the earliest signs of pneumonic inflammation, upon which a few words here will not be misplaced. Dr. Stokes, as our readers are aware, teaches the existence of a distinct stage of pneumonia prior to that of engorgement, and is led to believe that " an intense puerility of respiration in the affected part" will be found to accompany the earliest seizure of the pulmonary tissue, and hence point out the precise spot in cases of spreading inflammation about to become the seat of the crepitant ronchus. Dr. Williams, in imitation of Dr. Stokes, is of opinion that " the natural respiratory murmur will be rendered rough, and perhaps sharper, before the crepitation begins." But on the other hand, M. Grisolle, an observer of respectable Parisian reputation, declares " he has been enabled to follow the gradual extension of the inflammation, by the loss of force and purity on the part of the respiratory murmur, in a circumscribed point adjoining the manifestly inflamed part."* Now the particulars above related, from M. Fournet's work, testify to the accuracy of Dr. Stokes's obser~ * Memoire sur la Pneumonie, p. 11. Paris, 1836.
## 1840.]
and Curability of Phthisis Pulmonalis. 313 vation respecting the sign of imminent attack of a particular part by the inflammatory action: but there is this fundamental difference between the two observers, that the one considers the puerile character produced in a still healthy part, the other in tissue in an incipient state of inflammation ; in the one case it announces the actual existence of adjacent engorgement, in the other it implies that engorgement is ready to occur. That Dr. Stokes's rationale is probably (judging from close analogy, as well as from the third case alluded to by M. Fournet,) erroneous may, we think, be assumed; and, as we shall presently see, M. Fournet's experience of the signs produced in actually congested tissue, tallies more closely with that of M. Grisolle than of our countryman.
Under the name of pulmonary crumpling sound (bruit de froissement pulmonaire), the author describes a phenomenon apparently occurring, as he alleges, in the substance of the lung, and differing from all auscultatory signs hitherto noticed. In its most intense form, this sound very closely resembles the new-leather creak of pericarditis, differing therefrom solely in its greater acuteness of quality; in a less marked degree it constitutes a sort of plaintive noise, varying in tone with the rapidity of respiration; and, in the weakest and most common form, bears a close similitude to the gentle, quick, dry sound, elicited by blowing on fine paper. This phenomenon is generally produced in a very confined space, ordinarily coexists with inspiration only, and has been noticed by M. Fournet in the first stage of phthisis, in one case of encephaloid cancer of the mediastinum, and in another of non-tuberculous cavity of the summit of the lung. Some phthisical subjects experience a feeling of discomfort in the situation where the sound is audible; but this is rare. The anatomical conditions under which it arises are stated to be, a mechanical obstacle to the expansion of the lung, lobular induration of the pulmonary texture, and alternate flapping backwards and forwards of a fibrous lamina forming the wall of a cavern. As a sign of intra-thoracic tumour it possesses no value; it is of greater utility in cases of phthisis, as it existed in one eighth of tuberculous subjects attentively examined.
We shall recur to the subject of this sound, when engaged with the early diagnosis of phthisis. Meanwhile we may venture to express our doubts, founded very much on his own description, of M. Fournet's correctness regarding the location, mode of production, and close relationship of these modifications of sound. We can discover no proof of their being connected in the manner alleged; there is no attempt even made to point out the stage of transition from one to another, and that a sound like the pericarditic leather creak should be produced by the mere compression of indurated pulmonary tissue by the inspiratory rush of air into the neighbouring vesicles is, prima facie, most questionable. It is very probable that the most delicate form of crumpling sound may be thus developed; but the intense variety referred to appears to us an uncommon and peculiar species of friction sound, produced in pleural false membrane, or similar structure. In the case of cavern adverted to, such was confessedly its mode of production. We have before us notes of a case of chronic pleurisy in a tuberculous subject, observed in August, 1835, in which we find these words : " Under the left clavicle the inspiration is masked by a well-marked friction sound, which does not exist in expiration, though when the patient breathes expansively, there is a Fournet on Auscultation, and on the Diagriosis [April, well-marked expiratory murmur. In the supra-spinata fossa a sound resembling the creaking of new leather coexists with inspiration and expiration ; opposite the spine of the scapula this changes into a slight crackling sound, still masking the respiratory murmurs; these are heard about and within the angle of the scapula, the inspiratory sound is feeble, unaccompanied with ronchus, the expiratory prolonged and blowing." Now let us contrast the characters assigned to the creaking and friction sounds:
Creaking variety of crumpling sound, Coexists with inspiration only.
Is continuous, and unaccompanied with any sensation of displacement.
Is most frequently heard at the summit of the lungs.
Does not produce any sensation appreciable by the patient himself.
Is accompanied with no thoracic fremitus.
Friction sound, Almost always coexists with both movements.
Is jerking, and attended with a sensation of elevation and depression.
Is most commonly met with at the central part of the posterior surface.
Is not unusually felt by the patients themselves.
Gives rise to distinct fremitus.
The reader will observe, that the phenomenon under the clavicle possessed the majority of the characters of the crumpling sound, while the sound in the back was that of friction. Now the similarity in their characters appears presumptive evidence of the similarity of the tissue in which they are produced, and all the differences noted between them may be accounted for on the supposition, that the one arises from the unfolding; of indurated pleural tissue, while the other originates in the collision of granulations, or laminar false membrane. It must be remembered, too, that although the distinctive peculiarities of the friction and crumpling sounds generally exist as they are stated above, yet each character may be absent, or those represented as peculiar to the one coexist with the other. The subject requires further examination.
Under the title of humid ronchus with continuous bubbles, M. Fournet describes a morbid sound which, he states, existed in twenty-three subjects, the only ones carefully examined, affected with active sanguineous congestion of the lungs. This is a vesicular ronchus, composed of bubbles with a peculiar character of viscidity and humidity, and which, in the language of our author, instead of attaining a completely spherical shape before bursting, form a third or a half only of a sphere,?an imperfection of development very distinctly traceable to the viscous quality of the liquid forming them. They hold a mean rank, in point of size, between those of the crepitant ronchus of pneumonia and of mucous ronchus. They are few in number, three or four only occurring during each inspiration (with which movement they coexist exclusively), and are slowly formed. Each new bubble is produced before the full development and explosion of its predecessor; hence the sensation of continuousness with which this ronchus is accompanied.
We have closely reproduced the more striking characteristics of this ronchus, as detailed by M. Fournet; the minuteness of analysis gives it an apparently novel character, but we have no doubt it has been observed by many, and confounded with the subcrepitant ronchus of acute capillary bronchitis. The latter is however, according to the author, distinguishable from it by the larger size, greater number, more perfect sphericity, and comparatively rapid production of its bubbles, from its existing over a wider space, and at the posterior bases of both lungs, while the ronchus of active sanguineous congestion is usually confined to one side.
In the normal state, the movements of expansion and retraction of the lungs do not give rise to any appreciable sound, having its seat in the pleurae. Of this fact, ascertainable by auscultating healthy individuals, and demonstrated on horses by M. Andral, the author adduces new proof from the examination of the thoracic and pulmonary movements in two rabbits. These experiments would not have been worth notice, were they not cited as proving?for M. Fournet affirms that M. Piorry and himself saw this through the flayed walls of the animals' chests? that friction of the pleural surfaces against each other regularly occurs in the healthy state, during inspiration and expiration. The observation clashes with some physical principles, and the nice deductions of M. Woillez on the reciprocal influence of the thorax and its contents; but as we have not space to enter on a consideration of the subject, we must content ourselves with referring the reader to our review of that gentleman's recent work.
Were M. Fournet's experiments more numerous, and of unimpeachable accuracy, no a priori consideration should interfere with the inference they furnish; but they are few in number, and (as all who have examined the curious question of the effects of penetrating wounds on the motions of the lungs will at once recognize on perusal of their description) performed by a person most imperfectly acquainted with the complex nature of the phenomena he was about to analyze.
The morbid phenomena originating in collision of the pleural surfaces are examined with characteristic minuteness. Three varieties of friction sound are here admitted : 1, The grazing sound (bruit de frolement) ; 2, Friction sound properly so called (bruit de frottement) ; 3, Grating sound (bruit de raclement). They are always audible in inspiration, not appearing in expiration unless strongly marked; thus the grazing variety is not perceived in the latter movement, while the others manifest themselves in both; under all circumstances they appear first in, and disappear last from, inspiration. Their occurrence over an extensive surface is of favorable augury, as such state usually announces rapid and uniform absorption of pleuritic effusion. M. Fournet's experience confirms the conclusion already arrived at by Andral, Louis, Dr. Stokes, and others, respecting the non-existence of friction signs in interlobular emphysema. Indeed, the error of Laennec, in assigning this causation to the sounds in question, is now generally recognized.
The rarity of friction signs, compared with the frequency of pleuritis, has not failed to strike our author. Upon the more familiar causes of this disparity it is needless to dwell; we may however notice that M. Fournet considers diminished facility of locomotion in the lung, either from general pleuritic adhesion or extensive engorgement of its tissue, as capable of preventing their development. That they do cease, under the circumstances, had already been noticed by Dr. Stokes, who tendered, as an hypothesis, the explanation now announced as matter of certainty. Indeed, the fact that the supervention of friction sounds, in some cases of pleuro-pneumonia is coeval with the period of resolution, not of the Fournet on Auscultation, and on the Diagnosis [April, pleuritic effusion, but of the coexisting pneumonia, seems otherwise inexplicable. According to M. Fournet, some trace of respiratory murmur always coexists with friction signs, because the expansion of the lung, and the rubbing which follows, cannot be effected without the penetration of some air into the vesicles. We are not unwilling to admit the accuracy of the latter clause, but question the truth of the inference from it, and deny the propriety of the mode in which that inference is drawn. Dr. Stokes was, Ave believe, the first to point out the interesting fact that friction phenomena may coexist with " great and universal dulness," and consequently with extensive liquid effusion; now, though Dr. Stokes does not mention precisely the condition of the respiratory murmur under these circumstances, yet if the dulness were at once great and universal, that these were absent seems the only reasonable conclusion. Nevertheless, it may be urged in favour of M. Fournet's opinion that, as in a case of abundant effusion attended with friction sound, he traced their production to some membranous adhesions retaining one part of the lung in rather close connexion with the costal pleura, the respiratory murmurs must have been audible opposite those adhesions, and are likely to have been similarly perceptible in Dr. Stokes's cases.
We can hardly assent to the correctness of another of our author's " laws,"?that the duration and intensity of the respiratory murmur generally decrease in proportion as the friction sounds reach a higher type. The occasional coexistence of the most intense form of friction sound (grating variety) with minute granular elevations, either on the surface of a false membrane or of the pleura itself, a circumstance adverted to by himself, is clearly at variance with the supposition of proportional diminution of the respiratory murmurs.
The grazing variety differs, in some important particulars, from allied phenomena. Though remarkable for mobility, it occurs more frequently at the upper part of the pleura than elsewhere; the changeableness of its situation depends on its being produced by local and transitory attacks of pleurisy, themselves determined by eccentric tuberculization of the subjacent lung. It is ordinarily of short duration ; a single day suffices for its production, development, and termination, and this series of phases may, as M. Fournet has observed, be accomplished several times successively, in the course of a few days. Should these statements be verified, they must give a new character of precision to the diagnosis of local pleurisy, and perhaps help to determine the still unsettled question of the cause of the wandering thoracic pains of phthisis.
On the causes of the variable duration of friction signs, nothing novel is related. M. Fournet however states, and the statement is to us new, that a blister, applied in the situation where the rubbing sound is audible, hastens its disappearance. We are unfortunately left in the dark as to the number of times the fact was observed ; but, admitting the occurrence to be a common one, how is the effect produced ? " In one case, two blisters were applied successively; on both occasions the friction sound disappeared during the period of suppuration, and reappeared when they were suffered to dry up." We confess we cannot comprehend these changes, unless by supposing that the blisters produced slight effusion in the subjacent portion of pleura, and so deprived the false membrane of the property of dryness essential to the development of the sounds in question. Dr. Macartney was in the habit of remarking, in his Lectures, that the cutaneous irritation produced by flogging, often brings on pleurisy; now we should not wish to draw any decisive inference from the analogy here presented, even were the alleged fact clearly proved, but, coupled with the phenomenon detected by auscultation, it rather warrants doubts, which are justified upon other scores too, of the real therapeutical power of blisters applied over the affected part, in pleurisy. We are the more anxious to draw attention to this point, as young practitioners, deceived by the marked relief of pain afforded by the application of blisters, imagine that in prescribing these epispastics they are really curing the disease, and are hence not unlikely to neglect means of less contestable efficacy. By therapeutic power, we mean that of diminishing the mortality, or shortening the mean duration of a disease. The sudden disappearance of friction signs, originally developed on the absorption of an effusion, announces the return of the latter, or as a corollary from the proposition already stated, the supervention of pneumonia ; M. Fournet affirms that in one instance this change in the pleural signs first awakened his suspicions, confirmed by the event, of the latter disease having appeared. There are several other particulars in this section well worthy of notice; but we must allow the reader to discover them for himself.
In his examination of vesicular ronchi, M. Fournet confirms the general observation, particularly dwelt on by Dr. Stokes, respecting the occasional absence of crepitant ronchus, both previous to pneumonic solidification and after its resolution.
While upon this point, we may take the opportunity of stating our belief that Dr. Stokes has rather undervalued this ronchus as a sign of pneumonia. The coarse auscultator, to whose ill-educated sense all crepitating murmurs appear the crepitant ronchus, will no doubt sin frequently in his judgment respecting the presence of pneumonia, when no such affection really exists, if he presume to found his diagnosis on his acoustic perceptions alone. But of the pathognomonic character of the crepitant ronchus, formed of minutely delicate, dry, uniform, closely-pressed, regularly-rounded bubbles, bursting as with a sudden puff under the ear, we, with M. Fournet, entertain not a shadow of doubt.
The distinction of the second and third stages of pneumonia has long, confessedly, been one of the most puzzling points of semeiology. M. Bouillaud has, it is true, lately written, apparently with the view of enhancing the excellence of his marvellous saignees coup sur coup, as if the distinction were to him matter of facility; but he has convinced none, even of his countrymen, of his prowess, except those whose natural vocation it is to bow the head when " an authority" opes his lips. Dr. Stokes, however, some time since, announced his persuasion that the supervention of the third stage (that of interstitial suppuration) might be detected, with tolerable certainty, by the coexistence of " a sharp and peculiar muco-crepitant ronchus with bronchial respiration;" and M. Fournet, believing himself, as usual, original, confirms the accuracy of this statement.
He adds the important observation, that the ronchus coexisted with inspiration only: we call this important, because it shows that the phenomenon was not produced in the bronchi, and is therefore not the Fournet on Auscultation, and on the Diagnosis [April, same as that which Laennec, & priori, as it is suspected, assigned as characteristic of pulmonary suppuration. In his fourth chapter the author conveys, in a tabular form, the substance of his remarks on the healthy and morbid characters of respiration, and classifies the various ronchi. The presumed seat of the phenomenon forms the main groundwork of his classification; the character of dryness or humidity gives rise to the formation of sub-classes. The idea of this arrangement is far from new, but is here carried out more completely than elsewhere.
In respect of their seat, they are divided into six classes,?intra-vesicular, extra-vesicular, bronchial, tracheal, laryngeal, and bucco-pharyngeal. The intra-vesicular class contains five species,? the humid crepitant, with continuous babbles, already described at length; the sub-crepitant of pulmonary oedema, the sub-crepitant of acute capillary bronchitis, the ronchus crepitans redux of pneumonia, and the primary crepitation of the same disease. Under the head extravesicular, are included those morbid sounds, presumed to originate in the actual texture or parenchyma of the organ. The term is evidently illjudged, for it is plain that tracheal ronchi have quite as fair a claim to such title as those on which our author confers it. Be this as it may, the class contains two subclasses, founded on the fact of the organ having undergone excavation, or being free from solution of continuity. To the latter belong the pulmonary crumpling sound, and the dry crackling ronchus (of which we shall presently give a full description), produced, as M. Fournet imagines, by the crumpling of the pulmonary tissue against tuberculous deposits. With the former, rank the humid crackling ronchus, the clear mucous, or cavernulous of Hirtz, the dry cavernous, and the humid cavernous, or gurgling.
With the remaining classes we find no reason to detain the reader; and pass on to the chapter developing the laws of coexistence of the morbid phenomena of respiration with the inspiratory and expiratory sounds respectively. The separate appreciation of the two sounds, under all conditions of disease, has led M. Fournet to the conclusion, that such laws do prevail, and that intimate acquaintance with them frequently furnishes a valuable guide in practice. Of this we have, indeed, already given one rather striking illustration. We have shown too, that, according to our author, the progress of certain affections may be traced by the manifestation of a morbid quality in both sounds, which had previously accompanied one only. Of the constancy of these " laws," we are of course, for the present, obliged to accept M. Fournet's assurance. The greater number of morbid characters belong to inspiration; but those of highest diagnostic importance appear in expiration ; none are peculiar to the latter murmur only. The laws of coexistence may be traced to a certain number of general principles, of which the following are the most striking and important. 1. The closer the proximity of the seat of any ronchus to the pulmonary vesicles, the more marked is the tendency to exclusive coexistence with inspiration. 2. On the contrary, the further the seat of any ronchus is from the vesicles, the more uniformly does its production coincide with both movements. As a corollary from these two propositions it follows, that the proneness to exclusive coexistence with inspiration increases in the inverse ratio of the caliber of the tubes in which the ronchus originates; and this holds good in respect of the ronchi of cavities of new formation and of increasing size,?witness the successive series of the humid crackling, the cavernulous and cavernous ronchi. 3. The larger the bubbles of humid ronchi, the more frequently do these coexist with both movements, and vice versS,. 4. The greater the intensity of dry ronchi, the more commonly do they affect both murmurs; the weaker they are, the more exclusively do these appear in inspiration only : as an instance of this, dry bronchial ronchi may be compared with the crumpling and dry crackling sounds of phthisis. 5. Ronchi of grave tone coexist more especially with expiration; those of acute tone, with inspiration. 6. Ronchi which are exclusively inspiratory appertain, in general, to more serious diseases than those audible with both movements; the humid crackling and humid cavernous ronchi form the only exceptions to this rule.
In the following table of the author's we present our readers with the means of increasing the list of general principles commenced above, and of more accurately estimating the originality and character of M. Fournet's researches on this section of his subject, than we could even by the fullest commentary. 14. Metallic tinkling and echo/ 15. Blowing character.
* In some cases of hydropneumotborax with perforation, observes M. Fournet, " the metallic tinkling of Laennec is not to be discovered ; but, instead, the amphoric character of the respiration seems to reverberate in a sort of vague diffused echo, which rings like the voice under an archway; this phenomenon may be called rdsannance metallique ; it often accompanies the voice and cough." The English reader will here recognize the precise description, almost the very words of Dr. Williams, in reference to the phenomenon of tinkling echo. In confirmation of the utility of the distinction of ronchi, &c., into inspiratory and expiratory, and those common to both movements, our author collects together some of the most striking examples of differential diagnosis obtainable thereby. We have already, however, touched on the majority of these, and shall only notice two more points. There is no small difficulty, we have said, in distinguishing supplementary respiration from respiration with the expiratory murmur morbidly prolonged; now, if change of quality of the bronchial type exist in expiration only, all hesitation is at an end?the case is certainly not one of supplementary respiration. But important as this distinction is, scientifically, in all cases, and practically in many, there are circumstances in which it may have little value in the latter point of view. Let us suppose a respiration of the doubtful character alluded to, audible under the clavicle: to whichever species it really belongs, it may announce the presence of tubercles; if it be supplementary, we conclude that the foreign bodies are rather deeply seated in the lung; if of the other species, that they are close to the periphery : the practical man will hardly thank us for the delicate distinction.
'321 If M. Fournet be correct, the primary crepitant ronchus coexists with inspiration only ; the ronchus redux with inspiration chiefly, but slightly with expiration : the inference is, that the period of the disease may thus be ascertained in a patient observed for the first time, and under unfavorable circumstances for learning the previous course of the complaint. But the observations of two distinguished home auscultators do not coincide with those of M. Fournet. Thus Dr. Williams, speaking of the primary ronchus, states that " it is first heard at the commencement of inspiration and the end of expiration ; but it soon accompanies the whole respiratory act, and, in advanced degrees of the first stage, is heard only at the end of inspiration and the beginning of expiration."* And Dr. Stokes, describing the ronchus redux, affirms that " it is to be heard during the whole inspiration, and, in a diminished degree, during expirationand, so far agreeing with the French observer, adds, that " in other cases the first part of the inspiration is pure, and the rale only appears at the termination," and has once observed " rale first, and then pure vesicular murmur." The impetus which the promulgation of M. Fournet's doctrines will, we are convinced, give to the study of the two murmurs, must soon settle questions of this character; and it is on this account we abstain from noticing numerous points in the table which our own experience would lead us to regard as of doubtful accuracy. We would, however, distinctly state that in all probability the characters or phenomena placed lowest in each division very frequently infringe the law to which they are referred.
A brief examination of the physical signs of the different affections of the lungs follows. Here there is little original matter, except the careful statement of the respective conditions of the two murmurs ; to these we have already, as far as is practicable, referred. The exposition of the signs of emphysema is rendered curious by the author's fancying himself the first to study the movements of the chest in that affection with precision.f He has, however, fairer claims to the character of an original observer, in respect of a disease termed active congestion of the lungs; and as both symptomatically and anatomically this affection probably possesses some importance, we shall condense the more essential particulars related about it. If, as M. Fournet conceives, it be really the preparatory stage of pneumonia, and by early recognition of it the development of the latter severe disease may occasionally at least be prevented, it needs no stronger claim to the notice of the English practitioner.
This affection, which is described from the observation of twenty-three cases, chiefly occurred in subjects between the ages of seventeen and twenty-five; was apparently little influenced by sex; most commonly attacked muscular subjects of sanguineous temperament; acknowledged in many instances the influence of hereditary predisposition ; [how this was ascertained, we are at some loss to understand ;] or of previous [April* congestive affections; appeared most frequently in the hottest months of the year, and was in four instances immediately caused by " carbonic acid asphyxia," prolonged muscular effort, or insolation. The local signs were humid viscous respiration (see experiments with sponge) ; fall of the inspiratory murmur to 4 or 5, of the expiratory to 1*; the ronchus described as humid with continuous bubbles ; slight dry cough ; slight bronchophonic echo, and dulness on percussion, in extreme cases ; sputa rare, white, aerated, slightly viscous; (18 times in 23 cases) all coexisting with a peculiar sensation of discomfort and oppression. Though most commonly developed at the central and posterior part of the lungs, it may occur in any situation, and is neither in its primitive seat nor mode of extension subservient to the laws of gravitation.
Several series of characters, putatively distinguishing this affection from the first stage of pneumonia, capillary bronchitis, passive oedema, and passive congestion, are very industriously laid down ; but we cannot help suspecting that M. Fournet's imagination has not been passive in the construction of these well-ordered categories. They are, at all events, such as d, priori reasoning would easily supply ; and mans imaginings and nature's doings so rarely agree, that we cannot avoid regarding this nice harmony with singular mistrust.
The terminations of active congestion are in resolution, hemorrhage, or pneumonia. The second, it is said, may be either pulmonary or bronchial. The disease merged into pneumonia in eleven of the twentythree cases, and it is presumable that pneumonic inflammation is always preceded by active congestion : this at least is M. Fournet's opinion; but even analogy does not support it. He conceives the proposition proved, however, by the following result: Having, while intrusted with the care of M. Andral's patients, received two subjects of similar age, constitution, and strength, and both presenting the signs of active pulmonary congestion, he bled one freely, and prescribed some simple ptisan for the other: on the morrow, the signs of congestion had almost completely disappeared in the former, and were converted into those of pneumonia in the latter. This proceeding has been since repeated in a similar manner with a like result. M. Fournet calls it an " innocent experiment;" but, unless this ingenious gentleman has the power of saying to a pneumonia, " Thus far shalt thou go, and no farther," the innocence of the pastime seems somewhat of an assumption.* M. Fournet's observations on the anatomical characters of active congestion are founded on a single case; they are manifestly in great part those of engorgement, and are completely different from those attributed by Dr. Stokes to the preliminary changes in pneumonia. Dr. Stokes's description has not yet received the sanction of general experience; but that gentleman cannot at least be charged with the absurdity of describing as novel, appearances already familiarly known.
We cannot find room for an account of three chapters occupied with the relation of experiments and hypotheses respecting the connexion of * M. Bouillaud some time since, as we took care to inform our readers at the period, made the important disclosure that Broussais was the " Messiah of Medicine.'' Broussais has since departed this life, and, for aught we know, the mantle has fallen on M.
Fournet; if so, we must confess it would be heresy to persist in our doubts of the safety of his therapeutical gambols.
## 1840.]
and Curability of Phthisis Pulmonalis. the pulmonary murmurs and thoracic movements, the seat and mechanism of those murmurs, and the reason of the coexistence of morbid variations with inspiration or expiration. We confess we regret this but little, as, on the whole, these chapters possess an anticipatory character : they are busy in imagining causes for phenomena of which the very existence is still in many instances problematical.
We shall here take leave of our author's researches on auscultation generally, and introduce the reader to the second part of the work, containing his contributions to the diagnosis of the earliest stage of phthisis. And here, before we enter into a critical examination of these, let us award M. Fournet his due meed of praise for the apparent zeal with which he has entered on his task.
If, as Mr. Farr's tables show, 19*6 per cent, of the specified mortality of England is from phthisis, neither the philanthropist nor the medical philosopher can devote their energies better than in searching, by well-ordered observation and close induction, for some method of arresting its ravages. He who shall one day cry " tvptjKa "?and is it Utopian to believe that day will come ?? will appear in the records of fame not second even to Jenner. The incurability of phthisis by art in the second and third periods of the disease is unfortunately a demonstrated fact; but, as M. Fournet enquires, where are the proofs, and who has supplied them, that in its earliest anatomical condition the seal of incurability is stamped upon it?
The truth is, they cannot exist; for we have been to the present hour without the means of detecting the presence of tubercles with surety and precision, while yet in their actually incipient state; the opportune time for action slips through our fingers before we are aware of it, and we commence the attack when the enemy is assured of victory. The first point, then, in an enquiry into the possibility of curing tubercle must be an attempt to establish its early diagnosis ; and with this view the author has investigated minutely 192 cases of phthisis in the different aspects which the progress of clinical observation and his own special acquaintance with auscultation permitted him to do. [April, limited resources?wherein he exhibits unwonted correctness of plan and execution. He sets out with the principle?of which we grant the truth ?that phthisis is either inherited or acquired ; and where hereditary transmission cannot be made out, diligently seeks out some anti-hygiological condition to which the patient may have been previously exposed : if successful in his search, he straight concludes that what comes last is the consequence of what went before. Glorious system, this post hoc propter hoc mode of argumentation, which tells us, for instance, that night is the consequence of day : so long as we are wise enough to appreciate its excellence, we shall be in no want of etiological instruction. A consumptive patient consults us; we obtain an admission that he has been immoderately given to sexual indulgence, and can discover nothing else to our fancy to account for the disease, therefore that indulgence is its cause ; and as it is here a cause, in the next individual we meet, whose appetites have similarly sinned, the frailty constitutes a sign of phthisis.
But this fundamental vice is not all we have to expose in this chapter; M. Fournet displays marked ignorance of the principles of statistical medicine. He informs us, for example, that a particular form of lymphatic temperament was the most common among his phthisical patients, and hence infers its efficacy in inducing the disease. But this statistician forgets that, in order to justify any such conclusion, lymphatic subjects should not only superabound absolutely, but proportionally to the quota of individuals belonging to each temperament, forming the floating population of the Parisian hospitals in which he observed. If (we of course put the case hypothetically, though the probabilities are in favour of its reality,) lymphatic individuals are more numerous in those hospitals than those of other temperaments, will not the necessary effect be an excess of such patients among the consumptive, unless that form of constitution be actually prophylactic against the disease ? M. Fournet teaches his countrymen that, according to Sir James Clark, the period of life during which the greatest number of phthisical patients are observed, is comprised between the twentieth and thirtieth years : now Sir James Clark has not made this assertion, for the table referred to in that pathologist's volume relates altogether to the period of death. Though a misquotation, this and numerous other passages show that M. Fournet has not only perused but studied the work in question; yet, with singular hardihood, he produces as original the important opinion that diseases of the digestive and secretory systems of parents are those most prone to influence the health of children. But he inferred the fact from his own researches?possibly he did ; but who taught him to search for it ? who proved the search would not be in vain ?
He has no more claim here to the character of a discoverer than the traveller who, provided with the chart of the first explorer of a previously unknown region, follows confortably in the track of his enterprising predecessor.
We have been induced to expose the imperfections of this part of M. Fournet's volume, not on account of any assumed importance of his etiological principles in the diagnosis of phthisis?for he admits generally that little is to be learned in this way, but in order to warn the reader, who in the interest of perusal might forget the fallacious manner in which many of them were obtained, from being imposed upon by the air of scientific exactitude wherewith this author has managed to invest his notions on the causation of the disease. This is the more necessary, as M. Fournet has thefolle vanite to propose making a book, based on his 192 cases, on the etiology and nature of phthisis. Besides, if, as we conscientiously believe, in minute observation and the numerical method lie the main hopes of improving medicine, it is our first duty to expose such application of numbers to pathology as must tend not only to perpetuate but to give authority to error.
In the author's article on the influence of diseases of the lungs and pleura on pulmonary tuberculization, we find an affirmation that he has several times found tuberculous miliary granulations in the substance of pleuritic false membrane, while the subjacent lung contained not a trace of such product, or simply presented a few granulations or slight tuberculous infiltration of its peripheric substance. He has also seen a large cavern in false membrane similarly situated, while there were only a few traces of commencing excavation in the pulmonary tissue; and known the tuberculous deposition to advance from the false membrane outwards and affect the subcostal cellular membrane only. Under these circumstances the opposite lung, uninvested by false membrane, had occasionally remained perfectly free from tubercle. From these alleged facts the conclusion is drawn, that tuberculization of pleuritic tissue may occur independently of preexisting pulmonary disease; may be primitive and induce consecutively a similar morbid process in the lung. The author's observations tend also, he assures us, to prove that a similar mode of transmission from the serous membranes of the subjacent structure occasionally prevails in the intestines and brain. As we are promised a distinct publication upon this point also, we for the present register it without comment; we shall, on the appearance of the new tome, have occasion to examine the proofs given of the tuberculous nature of the granulations referred to.
M. Fournet's observations lead to the result long since announced by M. Louis, that the existence of double pleurisy with effusion involves, in the immense majority of cases, that of tubercles in the lungs. His conclusion on the influence of thoracic inflammations generally on the production and evolution of tubercle is the same as that adopted by Louis, Andral, Dr. Forbes, Sir James Clark, and all those who have observed nature, unbiassed by the dogma of irritation : as to the diagnostic value of intercurrent thoracic inflammations, derived from their peculiar course and character, we perceive nothing novel.
Here is examined at much length the symptomatic bearing of hemoptysis ; but the degree of confidence, which the enquiries of preceding observers have taught us to repose in that symptom, is neither increased nor diminished thereby. Its occurrence as the first indication of the disease and in a state of sound health is justly regarded by the author as a circumstance of exceeding rare occurrence. But the fact that in these cases " the aspect of the patient is generally by no means indicative of perfect health," no matter how earnest his own protestations of robustness may be, had already been pointed out to the profession by Sir James Clark.
M. Fournet is of opinion that " active congestion and hemoptysis tend, under the influence of a special predisposition to tuberculous formation, to attract and to concentrate in the lungs the process of tuberculization ; and when this is once established to perpetuate and accelerate its progress." These opinions are not new; but as one or two arguments of a novel kind are adduced in their support, we shall put the reader in possession of these. The period of lite at which active hyperemia of the lungs most frequently occurs is comprised, as we have seen, between the seventeenth and twenty-fifth years: comparing this result with those obtained in different quarters respecting the age of maximum frequency of hemoptysis and pulmonary tubercle, M. Fournet observes that the period of greatest frequency of hyperemia comes first, then that of hemoptysis, and lastly of tubercle. The catenation in respect of cause and effect is immediately inferred. Now we do not object to the principle by which this conclusion is obtained, but we pray the reader to forget M. Fournet's general terms and attend to the state of the individual facts.
He observed altogether twenty-three cases of hyperemia, and yet talks of his statistical inferences therefrom : in two subjects only did this hyperemia terminate in hemoptysis, in one of these there were already manifest phthisical symptoms, in the other the disease was suspected; and finally, M. Louis has shown that hemoptysis was most common from forty to sixty-five, and in males of nearly the same frequency at all ages.
Besides, has it not been proved by M. Louis, that the existence of tubercle in any organ in subjects past the age of fifteen, involves almost with mathematical certainty their presence in the lungs; in other words, that those organs are essentially the natural primary seat of the morbid formation, and this without the influence of any previous hemoptysis to "attract the process of tuberculization to them?"* But while we deny that these clumsy attempts at numerical induction prove in the least the point on which they are intended to bear, we do not intend to question the probable secondary influence of congestive movement on tuberculization.
The possible production of heterologous substances generally in the interior of fibrinous coagula is a question of high import. M. Andral, on the evidence of a single case, originally maintained the possibility of the direct transformation of an apoplectic clot in the lung into tuberculous matter.
Criticised by M. Meriadec Laennec, he acknowledged his inference had been premature; and has never since met with a similar appearance in the lungs, but still affirms that such is the apparent mode of origin of splenic tubercles. Our author considers M. Andral's concession uncalled for, because he has himself observed a fibrinous mass in the centre of a clot in the aorta, studded with and almost wholly formed of small whitish, nonvascularized granulations. It is not a little curious how important a part the fibrinous portion of clots has been made to play by different writers: M. Andral, finding what he took for pus thus situated,in an instant converted the containing coagulum into an absorbing and secreting agent, endowed it with life, and with this fact for his text, ? M. Fournet marvels elsewhere why, in a subject addicted to sexual debauchery, the lungs, which had not been locally excited, were the principal seat of the disease ; the testes and penis no doubt ought to have been the chief sufferers. But he most philosophically masters the difficulty by hinting, that the acceleration of respiration accompanying the act of coition may have been the cause of the pulmonary deposition !
## 1840.]
and Curability of Phthisis Pulmonalis. 327 sermonized on the wonders of organization : the microscope has happily dispelled those crude imaginings by showing the presumed pus to be nothing more than some of the substance of the clot itself in a softened state.
What ingenuity, again, has been wasted in explaining the origin of the central " pus " of the coagula of phlebitis. And now we are required to credit a doctrine of tuberculization founded, the probabilities are a thousand to one, on some similar illusion. We know that an anatomist of somewhat greater experience than our author, namely M. Cruveilhier, was deceived into the belief that he had produced tubercles by injecting mercury into the bronchi; he mistook pus for that morbid matter. When our advocate of fibrinous tuberculization provides himself with some half dozen cases in which the microscopical character of the alleged tubercles are accurately enumerated, we shall give him a patient hearing.
Before entering into M. Fournet's exposition of the physical signs of the first stage of phthisis?a subject on which he is entitled to a greater share of attention than on those just adverted to?it is natural to enquire how these were ascertained; as the affection does not usually cut off its victims at so early a period of its existence as to allow a comparison of the signs of that stage with its anatomical characters. First, this observer carefully noted the signs produced on the confines of manifestly diseased parts, and compared these signs with the incipient disorganization by which they had been induced. Secondly, he traced the gradual advancement of the disease to the rest of the lung in patients presenting the signs of the second stage in the upper portion. Thirdly, he followed the gradual changes undergone by the respiratory murmurs in patients admitted into the hospital for affections unconnected with the chest, but who therein became phthisical, and presented a successive series of signs and symptoms, from the most trifling and obscure to those of excavation.
Such being the judicious mode (for the possible fallacies likely to spring from one method are avoided and rectified by the others) in which the signs are stated to have been ascertained, we proceed to lay an abridged account of these before our readers.
Respiration. The inspiratory murmur increases in intensity to 12, 15, or 18 ; though less uniformly, diminishes in duration to 8, 7, 6, 5, and becomes dry, hard, and rough. The latter character is invariable in its existence, until marked by alterations of quality ; the former two are not unfrequently absent. The expiratory sound undergoes an increase in intensity and duration by successive steps from 2 to 20; both characters in the majority of cases alter pari passu, but in rare instances the intensity remains unaltered, though the prolongation is marked: a sensation of dryness and hardness is at the same time conveyed to the ear. The extent of alteration of both murmurs furnishes a tolerably exact estimate of the degree of advancement of the first stage. These changes are accompanied with modifications of quality, which, exhibiting themselves first in the expiratory murmur, successively pass through the ascending series already described, the first four grades corresponding to the incipient stage of the disease. While the change of quality belongs to one or other of these four grades, it invariably presents itself in a more intense form in expiration than in inspiration ;.the cavernous and amphoric qualities, on the contrary, affect both murmurs to nearly a similar amount. With respect to the order of development and respective importance of these changes, it may be stated that modifications of duration and intensity present themselves first, and are therefore the very earliest physical results of tuberculization. Modified duration and intensity of the expiratory murmur, and changes of qvality in both, have respectively much greater diagnostic value than similar alterations of the inspiratory sound, or than a dry and hard character in either murmur. The morbid variations now passed in review generally coexist with each other; so true is this that were some of them detected without the others, their dependence on tubercles should be called in question. It is important to add that to the description now given M. Fournet has encountered but very rare exceptions.
Ronchi. The ronchi of the first stage of phthisis are stated to be the pulmonary crumpling sound, and the dry and humid crackling. The former we have already minutely described ; we shall now similarly dispose of the latter. The two crackling ronchi are successive degrees of the same phenomenon, the first accompanied with a character of dryness, the latter of humidity. The dry species is composed of a series of two or three crackling sounds; the humid form is similarly constituted, but its component parts gradually acquire the bubbling character more and more distinctly, until they actually produce a bubbling ronchus. The drier it is, the more completely is it confined to the inspiratory movement, when humid it accompanies both murmurs. The seat of this ronchus is primitively the summit of the lung, but it always originates, wherever there is tubercle in the first stage. The dry form is generally heard over a less extensive surface than the humid. Once developed the crackling ronchus rarely disappears temporarily (it did so in 9 out of 55 cases), but is distinguished by a persistence, which is itself more marked the longer the ronchus has existed. While possessing the dry character it is peculiar to tuberculization, and can be confounded with no other sound even in respect of its audible properties: the humid form may be mistaken for mucous ronchus, if its acoustic characters only be taken into account, but the course and seat of the phenomenon will clear up any doubts as to its nature. The crackling ronchus does not appear until the respiratory murmurs have undergone some slight alteration in intensity and duration, and, while other changes of quality, &c., seize those murmurs, passes successively to the humid form, the bubbling, the cavernulous, and the cavernous or gurgling. The time occupied in this series of mutations varies: M. Fournet found, in the majority of cases of acute phthisis, that the passage from the dry to the humid form was effected in from eight to twenty days; in chronic tuberculization in from twenty days to three months. The dry form occurs when a good number of crude tubercles either isolated or collected in groups form the anatomical evidence of the disease ; under peculiar circumstances, however, it would appear that a very few minute tubercles scattered through the pulmonary tissue are capable of giving rise to it.
The pulmonary crumpling sound is far from possessing the important diagnostic bearing of the ronchi we have been examining; and really sig-riifies nothing, unless coexistent with modifications of the respiratory murmurs ; but if thus detected, it augments the surety of the diagnosis precisely at the period, the first or second phases of the first stage, when the difficulty of detecting the disease by its sensible signs is greatest. The change to the humid character announces the establishment of softening, and the advancement of this process is measured by the degree of humidity, and rapid or slow assumption of the bubbling character. Sonorous and sibilant rtmchi depending upon local bronchitis sometimes coexist in the same region with the phenomena just described.
The regularity of occurrence of the changes traced by M. Fournet is of course a most important element for consideration in their diagnostic utility. He appears to regard those appearing to affect the inspiratory murmur as inseparable from the presence of tubercles, and gives it as his belief that if the dry crackling ronchus be not always observed, this arises from an examination not having been instituted at a sufficiently early period. We have already mentioned his ideas regarding the immediate seat and mechanism of the crackling ronchi; by others the humid form, at least, has been supposed to originate in the minute bronchial tubes, and such seems from the following passage to be the decided opinion of Dr. Hughes : " There is not unfrequently added to the signs already mentioned, either a small thin mucous rattle, resulting from the admixture of air with the glairy fluid existing in the minute tubes, or a! single clicking sound, arising apparently from portions of thick viscid mucus, presenting an obstacle to the passage of air through those of larger caliber."* We may here notice that Dr. Hughes speaks of " the bronchial irritation preceding the formation of tubercles," as if such were an usual or even a constant phenomenon. He appears to be ignorant that M. Louis long since proved that the bronchi are as frequently healthy in phthisis, except when communicating with softened tubercle, as in individuals dying of all diseases indiscriminately ; and that Major Tulloch has shown by the irrefragable testimony of figures, that while 13 per thousand of the whole military force are annually under treatment for consumption in Jamaica, and only 5 or 6 per thousand at home, catarrhal complaints only affect 65 per thousand of the force in the former place, while they are 122 per thousand in the latter. Pneumonia is still more rare in Jamaica.f?But to return to the signs of phthisis.
Voice. With the second phasis of the first period, in about one half the cases, a peculiar change occurs, according to M. Fournet, in the characters of the voice, rendering its tone grave, its quality less clear, its general character veiled, pectoral, and slightly hoarse : the whole constituting the incipient state of the well-known peculiar condition of the voice of subjects in an advanced stage of the disease. This change appeared to advance with a degree of rapidity proportionate to the frequency and duration of the colds under which the patient had suffered.
We observe nothing of the least novelty in this observer's analysis of the auscultatory phenomena. M. Fournet gives the recommendation to make the patient, instead of answering desultory questions, count a cer-* Guy's Hospital Reports, No. ix. f Vide British and Foreign Medical Review. Vol. VIII. p. 220. [April, tain number of figures, while the comparison is instituted between the resonance on both sides. In this way the changes of intonation produced by different words, and which probably affect the phenomena in question in a slight degree, are avoided. We have long been in the habit of employing this plan, (proposed, if not by Laennec, in his time,) but not, we confess, with M. Fournet's motive, which seems to us important.
On cough and the transmission of the heart's sounds, as indicative of tuberculization, there is nothing to detain us: the transmission of subclavian murmur by solidified lung, first noticed by Dr. Stokes, has escaped M. Fournet. There seems, however, to be some doubt as to the exact period of the disease at which this phenomenon becomes perceptible. By percussion no appreciable change of sonorousness is detected until the second phasis; on the application of this method of diagnosis, the volume before us falls short of the productions of Piorry, Dr. Stokes, and others. Under the title acouophonia or cophonia, M. Donne has designated a mode of investigation in which the observer places his ear to the chest and analyzes the sounds produced by percussing different parts of that cavity. This plan would perhaps appear better fitted a priori than any other to furnish a correct notion of the different degrees of density of the contained organs. But as M. Fournet has found by experience that it is really deceptive, that the perceived sound bears no precise relation to the condensation or rarefaction of the subjacent parts, we shall not seek to rescue M. Donne's novelty from the oblivion into which it is fast subsiding.
The modifications of locomotive and vibratile movement of the chest ascertainable by palpation or manual examination are delicately analyzed by the author; though the result, as respects the former, is rather to be noted from its declared connexion with the first period of phthisis and diminished volume of the lung, than from its newness. " In cases of tuberculization inducing a notable increase of density in the summits of the lungs, the partial motion of the corresponding ribs undergoes a diminution proportional to the induration and decreased size of the organ. The motion of totality (of general ascent and descent of the thorax) is, if at all modified, rather augmented than lessened in extent. As this sign cannot, generally speaking, be detected, until the disease is sufficiently advanced to be readily discoverable by auscultation and percussion, it might be considered a diagnostic superfluity, did it not seem to possess more negative than positive value. As in cases of disseminated miliary tuberculization the costal movements are not sensibly affected, the absence of morbid change in those movements may become a positive sign of that particular anatomical form of phthisis, when the other physical signs and the symptoms announce an advanced period of the first stage.
To us it has always appeared exceedingly difficult to draw any satisfactory conclusion from what is termed the vibratile motion or fremitus of the thoracic parietes. We shall probably be told that our tactile perception is of imperfect quality; and we should submit to the aspersion without demur, did we not detect a very notable contradiction on the subject in the statements of observers who, we presume, deem their fingers most happily moulded. If we believe Dr. Williams, liquid in the pleura will abolish the natural vocal fremitus completely, while in cases of hepatization, and under all circumstances of consolidation, the vibration is '?'?much stronger than on the healthy side." According to Dr. Stokes, the absence of vibration over the dull portion of surface is an exceedingly useful sign in the diagnosis of pleural effusion, hepatization, and enlarged liver, though traces of vibration may be detected in some cases of hepatization. But a very different version is that of M. Fournet. This observer affirms in the most satisfied tone that the maximum of vocal and tussive fremitus obtains in the normal state of the organ; that the phenomenon is always in the inverse ratio of the density or rarefaction of the pulmonary tissue and is totally abolished alike in hepatization and advanced emphysema. As these latest observations appear to have been made with patience, the following excerpta from the general conclusions respecting the healthy fremitus may possess interest for some of our readers. The vocal vibration is stronger in decumbiture than in the erect position; when the individual observed speaks in monosyllables than in polysyllables ; in the front than in the back of the chest, on the right than on the left side anteriorly (noticed by Dr. Stokes also), most strongly developed in the lower part of the sternal region ; extremely little marked in the clavicular region, hardly better in the sub-clavicular zones; and in many subjects altogether wanting. The difficulty of appreciating slight modifications in a sign thus characterized is sufficiently obvious; but M. Fournet's enthusiasm is nothing daunted, for he affirms that if the hand be laid over a mass of lung partly healthy, partly diseased, " the difference of vibration corresponding to the different parts of the hand may be easily distinguished." With such perfect specimens of tactile organization as this gentleman, it is of course vain to argue. It follows, however, from the general position above stated, that, a& indeed is elsewhere distinctly affirmed, the vibration is diminished opposite tuberculated lung in the exact ratio of the extent of disease : an idea opposite as night to day to that promulgated by Dr. Williams. We are further told that a very slight difference in density of the summits will induce a manifest difference of vibration ; yet this situation is pointed out as one distinguished by the very trifling degree of vibration producible in it when at the maximum, namely, in the normal state of the lungs. Why, if M. Fournet's statements be correct, the laws regulating the decrease of tactile should differ so widely from those of acoustic vibration we are unable to comprehend.
The author's enquiry into the signs of the first stage of phthisis, furnished by inspection, opens with an enumeration of some of the principal facts thus ascertainable in healthy subjects. The only point here stated bearing on this particular subject, which had not been elicited by the researches of M. Woillez,* seems to be that a piece of tape, stretched ? M. Fournet, with his usual happy ignorance of the proceedings of his contemporaries, knew nothing of these researches, of which an account was first published in May, he set about publishing his own work" in 1839! Truly the flood of light, which burst on M. Fournet from all quarters at that particular moment, must have been dazzling. Here, too, he manages to show us how easy it is to descry the VOL. IX. no. xvm.
'3 from the nipple to the most prominent part of the corresponding clavicle, will lie in immediate contact with the skin?a fact of which we shall presently learn the application.
The uniform transverse retraction of the upper portion of the chest, which induces the well-known cylindrical form of thorax, may, according to M. Fournet, be observed before or after the deposition of tubercle. This we admit; general experience acknowledges the fact, and M. Woillez has proved it. But we question if such preexisting conformation indicates, as the author affirms that it does, a predisposition to the growth of tubercle; the accurate observer just named has, it is true, shown that such malformation is much more frequent in phthisical subjects than in others, but where is the proof that in the particular individuals furnishing this result, the thoracic malformation preceded the tubercles in order of development? M. Woillez, we are aware, classes the defect of form in question with his physiological heteromorphisms ; but as he appears to have been ignorant of the atrophy, occurring in tuberculous lungs, which distinctly leads to sinking in of the parietes, we cannot accept his decision in the matter. This species of contraction exists, according to our author, in one fourth, or at the utmost one third, of phthisical subjects; he estimates as follows the diagnostic bearing of different degrees of this deformity. If we discover such contraction in a subject presenting the other signs of the first phasis of the first stage, it must evidently belong to the preexisting species. In the second phasis, if the deformity be much developed, the inference will be the same; but in the third, we may fairly conclude that such abnormal appearance is an effect of tuberculization. The preexisting may, it is said, usually be distinguished with ease from the consecutive contraction, in the following way : the former species exists to the same amount on both sides of the chest, and in its entire height, though gradually diminishing from above downwards ; the latter is particularly observed on the most diseased side, and at the summit. All this looks amazingly clear on paper, but we strongly doubt that these distinctions are founded on observation, or that the author has ever come to a conclusion in practice from their application. In justice we must add, that he does not regard thisheteromorphism as possessing absolute diagnostic force, but merely as warranting suspicions, which it is the business of the other signs to confirm or invalidate : to maintain a different opinion would be absurd, for this precise conformation is observed in persons who descend to the grave without a tubercle in their frame.
M. Fournet again proclaims his literary deficiency, by affirming that mote in our brother's eye, while we dream not of the beam that obstructs our own. He objects to M. Woillez's conclusion, that individuals following trades, which require particular exercise of the upper extremities, have on an average less development of thorax than others, because it is founded on too limited data (133 cases). Yet he himself preaches the law on every possible,point of the subject of phthisis on the foundation of scarcely a greater number of observations, resolves with the most perfect composure problems which the wisest statisticians regard as insoluble in the existing state of national censuses, and deduces statistical inferences from twenty-three facts. But the true cause of his objection is, that the conclusion in question opposes some of his favorite theories.
## 1840.]
and Curability of Phthisis Pulmonalis. 333 depression of the sub-clavicular region has not been noticed in cases of phthisis, except as a consequence of pulmonary excavation, and in supposing himself original in the announcement that it may be induced by crude tuberculization. " Decrease of the antero-posterior diameter of the thorax, with flattening or slight hollowing under the clavicles," says Dr. Stokes, " may be appreciated at a much earlier period than has been supposed and is found a most important aid in diagnosis in the earlier stages of phthisis." But this modification has been more minutely studied by M. Fournet than his predecessor. In every subject examined after death in whom he had observed it, the summit of the lung was found to be invested by a fibro-cellular false membrane uniting it closely to the adjacent wall of the chest. The sign is not, as he believes, available until the close of the first period, but he has occasionally observed it at the outset of the disease, when preexisting pleurisy had led to the deposition of firmer and more resisting pseudo-membranous tissue at the summit of the lung than elsewhere. This sunken appearance may usually be detected by simple inspection, but we are recommended to measure its degrees, by stretching a piece of tape from the nipple to the clavicle : the distance intervening between it and the thoracic surface gives the measure of the deformity. Dr. Stokes uses a pair of callipers moving on a graduated arc, one branch of which is fixed on the scapula, the other below the clavicle; this method will also give the measurement of any posterior depression which may be present.
M. Fournet adds nothing to Sir James Clark's accurate description of the visible changes in the expansion of the ribs opposite tuberculous accumulations; he states, however, that he has not observed this species of change except when there was sub-clavicular depression, and that both these signs may be detected in about one fourth of phthisical subjects.
Having accomplished his task in respect of the physical signs, M. Fournet turns to the local symptoms of the first stage. Admitting the observation made by Kuhn, that the microscopical tuberous tissue, of which tubercle is, according to that writer, composed, may be detected in the sputa of the consumptive, he affirms that the fact cannot be available for the purposes of diagnosis, until the period of commencing softening; as before that change, the passage of tuberculous matter into the expectoration cannot be supposed to take place. But this is an assumption which, though probably well founded, requires the attestation of direct experience; it is irreconcileable with Dr. Carswell's opinions respecting the seat of tuberculous secretion.
In speaking of dyspnoea as a symptom of the incipient stage, M. Fournet avails himself of M. Louis's cases, as countenancing the opinion that that symptom, when originating in early life, proceeds from the fancied imperfection of pulmonary development already spoken of. But he omits to mention, that M. Louis himself has ascribed it to the coexistence of emphysema with the tubercular disease.* The description of this symptom generally adds nothing to the graphic sketch of Sir James * British and Foreign Medical Review. Vol. VI. p. 38. Fournet on Auscultation, and on the Diagnosis [April,
Clark; the notice of the dyspnoea of acute phthisis may, however, be usefully consulted. Regarding the seat of tuberculous accumulations in the lungs, M. Fournet touches upon a point respecting which a curious difference obtains in the statements of observers?we mean the relative frequency with which the two lungs suffer and present the greater share of disorganization. As this is not a point of mere curiosity, but one which has close connexion with the application of the principle of comparison in diagnosis, it deserves investigation. The majority of writers assert that the left organ is most frequently and extensively implicated; Laennec seems to have stood almost alone in maintaining a contrary opinion. M. Louis, among others, has given the authority of his recognition to the former notion, and is cited to this effect by M. Fournet and others ; but, curiously enough, the cases of that observer do not, on analysis, warrant his inference, for here is the result in figures of his fifty tuberculous subjects:
Most extensive in right lung in -13 cases Cavities in both lungs in 33 cases Of 170 of M. Fournet's patients, 109 presented the greatest share of disease on the right side, forty-six on the left; in fifteen cases both organs were equally affected. M. Hirtz also speaks, in his recent thesis, as if he had obtained a similar result. As a small minority, twenty-nine only, of our author's patients were examined after death (although the ratio was among these as 21 : 8 in favour of predominance of the right lung), the question must be allowed to stand in need of further examination. Whatever be the decision, we, however, see no objection to conceding that it is unusual to find the disease equally developed on both sides?indeed, the tabular view we have given of M. Louis's cases proves this. The author elicits no novelty in his considerations founded on this fact; Dr. Stokes appears to have exhausted the topic.
Before extracting some passages from the author's estimate of the value of the physical signs, we shall exhibit, tabularly, the physical signs of the three phases of the first stage; we shall thus, we conceive, facilitate the reader's comprehension of the subject generally. In drawing up this synopsis, we have noted each phenomenon with the characters and in the position most frequently assigned to it by M. Fournet.
## 1840.]
and Curability of Phthisis Pulmonalis. ' ' quality, bronchial. Strong bronchophony, or imperfect pectoriloquy. Sound more obscure, or even dull.
Vocal and tussive fremitus much diminished.
Diminution of partial movements of ribs corresponding to indurated mass.
Transverse retraction of corresponding part of the chest. Subclavicular flattening.
1 But the facility with which these signs may be ascertained, the constancy of their presence, the regularity of their progress and catenation, form almost as important elements in judging of the confidence to which they are in practice entitled, as the actual reality of their existence. Upon some of these points, in so far as they are capable of being generalized on, M. Fournet speaks with just appreciation of the difficulties of clinical medicine.
"The observer must, while investigating the local signs, carefully guard against being influenced by any preconceived opinions, originating in the external appearance, or in the commemorative or general symptoms of his patient. If dominated by an opinion, in great measure already formed, the very best auscultator is not unlikely to become so careless and inattentive in his examination, that the impressions received by the senses are, without his being aware of it, converted into so many confirmations of his preformed judgment. This is the more likely to happen, because the phenomena appreciated by the senses and the mind are very numerous, and only distinguishable by very delicate differences from others of wholly distinct diagnostic force. When a first Fouk.net on Auscultation, and on the Diagnosis [April, examination has been made, under circumstances like these, we are sometimes astonished at the variation in the result of a second. . . . " The physical signs are not, as might be believed, pathognomonic of phthisis; for such signs alone merit that title as belong exclusively to a certain anatomical condition. . . . Some observers consider a prolonged expiratory murmur peculiar to phthisis, and almost all that has been written on the sound of expiration bears the impress of this idea; the belief is a groundless one, for the expiratory murmur manifests itself in a multitude of different conditions. . . . But from mutual combination, from coexistence with some signs, and from their appearance, independently of others, the phenomena described may lead to a precise diagnosis of tuberculization ; their primitive or absolute value goes no further than pointing out the presence of induration or foreign bodies in the pulmonary tissue. . . . " At a very early age no inference could, 1 think, be drawn from the existence of the morbid characters of the respiratory murmur, which I have assigned to the first phasis,?in many instances even to those of the second." Fully as we assent to the justness of these remarks generally, we would, nevertheless, point out a slight contradiction between one of them and the author's previous assertion, that the dry crackling ronchus is peculiar to tuberculization,?if peculiar to, it must, when present, be pathognomonic of the disease.
Nothing can be clearer than that the likelihood of forming a correct diagnosis through the physical signs increases, the more advanced the phasis of the first stage during which the examination is made. We shall not be so traitorous to the principles we have elsewhere professed as to affirm that the diminution of confidence in them, during the first phasis, must often amount to negation, unless they receive strong confirmation from the general and commemorative symptoms; but we should ourselves be disinclined to trust them, unless so supported. We, in truth, in this opinion echo, as we shall presently see, M. Fournet himself. Who shall affirm his capability of surely estimating changes of intensity and duration, amounting only to one or two tenths of the normal sum? who, if persuaded that such delicate variations are estimable by his senses, can declare, that such departure from the normal mean may not be the effect of individual idiosyncracy ? And again, is it easily credible that the complicated system of respiratory changes and ronchi, exhibited in the above column, works with the precision and regularity of a mechanical contrivance?that a morbid change once set in action never vacillates, but continues its onward course without interruption, till it reaches its maximum? Assuredly it is not; and however M. Fournet's general tone may be calculated to inspire conviction of the regularity of these changes, he is honest enough to confess, occasionally, that exceptions occur. In fact, the division into precise phases, and the ascending and descending notation of respiratory duration and intensity, are to be taken as approximations to the truth; unless this be admitted, numerous trifling contradictions, which escape the author in different parts of his volume, cannot be accounted for. But whatever be the irregularity to which it is probable the course and progress of these signs are liable, the reader must not forget that there scarcely exists a known combination of symptoms in any disease secure from similar unsteadiness.
Our author has, he alleges, observed, that the pulse of a great number of phthisical patients presents, even during the first stage, a peculiar softness. To this observation he, however, attaches little importance; and we notice the point merely to introduce to the reader a curious disclosure, recently made by Dr. Guy, respecting the diagnostic information to be derived from the pulse at an early period of the disease.* We have space for the conclusions only, but the striking character of these will, we doubt not, attract our readers to the original paper. "1. In cases of phthisis pulmonalis the frequency of the pulse varies within wide limits; the difference between the extremes amounting to 90 beats. "2. In the same individual, the frequency of the pulse undergoes remarkable fluctuations, passing, in a few days, through a range of upwards of 60 beats. " 3. In five out of six cases, the frequency of the pulse in phthisis exceeds the highest frequency observed in health.
"4. The difference between standing and sitting in phthisis is nearly the same for all frequencies of the pulse.
"5. The maximum difference between standing and sitting in all cases of phthisis pulmonalis, falls short of the mean difference in health. " 6. From the average results of a considerable number of cases, it appears that the mean difference in health is six times as great as the mean, and three times as great as the maximum difference in phthisis. " 7-On the supposition that the slight effect produced by change of posture is peculiar to phthisis pulmonalis, it forms one of the most constant and certain of its symptoms. "8 .On the supposition that the slight effect produced by change of posture is common to more than one disease characterized by increased frequency of pulse,it will distinguish these diseases from others with which they may be confounded ; and these diseases themselves are easily distinguished from phthisis pulmonalis, either by the peculiar character of the symptoms or by other physical signs.'1'
We fear Dr. Guy forms too favorable an estimate of the diagnostic utility of the " slight effect produced on the pulse by change of posture in phthisisshould experience prove us wrong, he will have fairly distanced M. Fournet, for, certainly, there are more persons capable of counting to 100 than of appreciating the delicate phenomena described by the latter observer. It is obvious that Dr. Guy is only entitled to affirm, from observations hitherto made, that the influence of posture is different in health and in phthisis; he cannot at present declare, that in many other diseases the pulse is not similarly affected, as in consumption. In his own practice he evidently sets at nought this difficulty; for he gives an instance of his diagnosticating phthisis on the evidence of this condition of pulse, and such appears, from the narration, to be not very uncommonly his habil. In the seventh and eighth propositions, a sort of provision is made for the objections of those unwilling to go the length of Dr. Guy, upon the imperfect data recorded.
But to return to M. Fournet.
This observer next presents us with a close investigation of the course and progress of the disease. The question, whether the local or general symptoms appear first, is one of no mean importance; it is one, also, on which we are willing to hear M. Fournet, because he appears to have comprehended fully, and, as far as possible, eschewed the circumstances likely to involve erroneous inferences on the subject?we mean imperfect analysis of the local signs of the earliest phases of the disease, and careless interrogation of patients ? Observations on the Pulse in Phthisis Pulmonalis. By Dr. Guy, Guy's Hospital Rejwrts, Oct. 1839. regarding the symptoms they may have experienced at or previous to the supposed outset of the complaint. His conclusions are, that in essentially chronic phthisis, the sensible phenomena of tuberculization may be sometimes detected in individuals presenting, in a marked manner, the evidences of predisposition to phthisis, at a period when neither they themselves, nor the medical observer, can discover any general symptoms; that in other cases, the constitutional symptoms lead the van, at a period when, though demonstrative proof cannot be given that microscopical tuberculous matter does not inhabit the lungs, it is yet, if it really does exist, in too minute quantity to affect, in anywise, the respiratory murmurs; that, in a third series of subjects, the local and general manifestations appear to originate simultaneously. In acute phthisis the general phenomena appear to set in at least as early as the local changes. These general statements may be considered to leave the matter just as their author found it, and, for the reason we have already given, some confidence may be placed in them. In commenting on the second variety, M. Fournet observes, that the general phenomena cannot be considered a mere effect of the reaction of deposited tubercle on the system generally, but that, in a number of cases, " they appear to be the direct effect of the tuberculous cachexia." This is evidently appropriated, without acknowledgment, from Sir James Clark, and, at the same time, unskilfully borrowed, for under these circumstances they are not the effect, but the actual essence of the cachexia itself, as it is termed by our distinguished countryman.
We need not dwell upon the author's description of the course of the symptoms and signs ; the order of development of the latter is distinctly shown in the table we have given of them; the account of the former is a periphrasis of Sir James Clark's chapter on the subject, and, indeed, corresponds therewith to a marvellous degree in some of its phraseology.
Two chapters and forty-nine pages are given by the author to the circumstantial and differential diagnosis of the affection. Under the former of these heads, he presents us with various combinations of local and general symptoms, which may occur in individual cases, and endeavours to estimate the probabilities for and against the existence of phthisis furnished by each. Here is exhibited much clinical sagacity, as, we think, the following specimens, selected chiefly from their reference to the value of the physical signs, will fully show. It is necessary to premise, that the accuracy of the diagnosis, under the circumstances supposed, has, in almost every instance, been ascertained, either by post-mortem examination, or by the gradual passage of the disease to its most advanced stages. " Case i. Being given an adult patient of strong constitution, sanguineous temperament, originally of robust health, of healthy parentage, never having been subject to colds or cough, nor having had hemoptysis, but who, after exposure for a certain time to anti-hygienic influences, perceived a change in his state ot health, we may, with perfect confidence, diagnosticate phthisis, provided we can detect distinctly and persistently the sum of local and general signs attributed to the second and third phases of the first period, and if there exist no other morbid condition capable of accounting for these signs.
"Case ii. Under the same circumstances, the same diagnosis may be made, but with less certainty of correctness, if the local signs be only those of the first phasis; provided the whole series of signs and symptoms acknowledge precisely, or nearly so, the laws I have traced as ordinarily regulating the development, coarse, and symptoms of accidental or acquired phthisis. "Case hi. If, under the circumstances supposed, the general symptoms were wanting, the local signs of the first phasis would only warrant a simple suspicion of the existence of phthisis. Again, if the general symptoms were absent or of doubtful character, and the signs of the second phasis existed in a distinct and regular manner, the fact of tuberculization would become probable. Finally, if the local signs of the third phasis are observed in a marked degree, and although there be no distinct general signs actually present, but something suspicious in the character of those, either existing or not long past, we may be pretty certain of the presence of pulmonary tubercles "Case x. The existence of the general symptoms of hectic fever, which I have said belong to the first, second, or third phases of the first period, if unaccompanied with any of the local signs of that period, could not, even if there were hereditary predisposition to the disease, justify us in concluding, that phthisis was actually present; but would warrant the opinion, that tuberculization will sooner or later take place. If, in addition to labouring under hereditary predisposition, the patient have had hemoptysis, colds, and cougli, with the course observed in phthisis, the absence of local signs should be set down as an exceptional occurrence, depending on a particular organic condition of the lung, and the case considered one of phthisis "Case xiv. If a subject born of apparently healthy parents, of only moderately strong constitution, and possessing delicate health, who has had frequent colds and chronic cough, and presents the morbid state of the voice elsewhere described, were attacked with pleurisy with effusion, which effusion becomes chronic, without any apparent cause for such termination, and entails deterioration of the general health, that subject should be considered affected with tubercles."
A similar position had already been established by M. Louis; but its vast practical importance justifies its repetition, whenever an opportunity occurs.
Fournet on Auscultation, and on the Diagnosis [April, phthisis; the following account, bearing on this question, is sufficiently curious to merit extraction.
Its accuracy rests on the authority of M. J. Guerin, the orthopaedist. " A female died in the third stage of phthisis, having shared her husband's bed till her last moments. The latter, who was originally of robust constitution, and sprung from a family, none of the members of which had ever been phthisical, married a second time; the subject of this union was of good constitution and healthy parentage. Eighteen months after marriage he died of confirmed phthisis; his wife, who cohabited with him to the last, remarried shortly after, and in two years died of consumption. Her second husband, of strong constitution, and belonging to a family in which phthisis had never been heard of, perished of the same affection soon after the death of his wife." Upon this tale we shall only remark, se non e vero, ? ben trovato, for the contagionists.
A disquisition on the curability of phthisis, during its third and first periods, presents some novel and speciously supported opinions. M. Fournet admits the possibility of a natural cure during the third stage, but endeavours to show that Laennecwasin error, both in respect of the frequency of its occurrence and in ascribing the fact to the cicatrization of caverns.
Curiously enough, this doctrine is put forward at the precise period that another Parisian observer, M. Rogee, publishes an elaborate and needlessly prolix paper, to prove that pathologists have no conception of the vast number of cases thus cured. The substance of M. Rogee's paper is elsewhere given ; and we must here limit ourselves to an abstract of M. Fournet's arguments: to institute a critical comparison between the rival doctrines would occupy more space than we can devote to the subject. Our author commences with a verbatim exposition of Laennec's theory of cicatrization, from the works of that observer, and then describes the very different manner in which he views the same phenomena. He remarks, founding his statements upon the dissection of lungs in which no distinct trace of tuberculous disease was discoverable, that when pleurisy with effusion takes place, leading to compression of the lung and exudation of membranous matter, if the false membrane is thicker and firmer in some parts than others, the surface of the lung is always depressed and furrowed in the corresponding spots. The furrows thus produced are sometimes so much as an inch in depth, and always proportional in dimensions to the general or local decrease in size of the lung. The false membrane on the pulmonary surface sends prolongations of different extent into the interior of the sulci. At a late period, the superficial patches of pseudo-membrane, corresponding to the depressed portions of lung, are converted into fibrous or fibro-cartilaginous tissue; which is sometimes incrusted with calcareous matter, and undergoes constant decrease of size from interstitial contraction.
At a still more advanced period, this fibrous disc is absorbed, and nothing remains but a few small fibrous bands, connecting together the well-known irregular mammillary prominences, which the affected parts of the lung present under these circumstances. Meanwhile, the prolongations placed in the sulci, and sunk, as it were, in the interior of the organ, similarly undergo the fibrous, fibrocartilaginous, or even petrous transformations. Sometimes they retain the character of white fibrinous lamellae, traversing the adjoining indurated pulmonary tissue in various directions; and occasionally, all trace of connexion with the external membrane is lost by the perfect closure and adhesion of the outer edges of the sulcus in which they are contained.
The adjoining portion of lung becomes indurated and thickly set with black colouring matter. Now these prolongations of pleural false membrane are, according to M. Fournet, what Laennec took for fibrous cicatrices of tuberculous cavities. Again, these fibrous prolongations sometimes become the seat of a secretion, an excavation forms in their interior, and puriform fluid appears therein ; such is the anatomical state which Laennec has, in some instances, taken for the evidence of a cavity in progress of cicatrization. Without following M. Fournet seriatim through the very close and plausible train of reasoning, by which these positions are sought to be established, we have given a sufficiently complete outline of his mode of argument to show that his objections to Laennec's theory are not wholly visionary; the reader will, on perusal of the original, find stronger evidence of this fact. The critical scrutiny of the cases published by Laennec and Andral, as affording proofs of the fact of cicatrization, which is appended by our author to the account of his own observations, unquestionably proves that both these writers have ascribed to this process effects otherwise easily explicable, and predicated the tuberculous nature of cavities upon insufficient grounds. "Whether M. Fournet's arguments are conclusive of the opinion he desires to substitute for those of his predecessors, further experience must decide. Meanwhile, his general conclusions may be abridged as follows: Pulmonary phthisis is, in extremely rare cases, susceptible of cure during the third stage, but has not been proved to depend upon complete cicatrization of cavities; nor is the mode, in which the cure is effected, yet understood, though it seems more likely to be by conversion of the caverns into fistulse than by their closure. The tendency of caverns to contraction and obliteration, although a probable circumstance, has not yet received anatomical demonstration ; and, finally, the only cases apparently capable of cure are those in which the tuberculous matter occupies a very limited extent of the lung, and gives rise to the formation of a single, or at most of two closely contiguous cavities.
The author shows, in a closing chapter, to what extent legitimate inference from anatomical phenomena supports the notion of the curability of the disease in its crude stage. He commences by examining the truth of Laennec's dogma, that the inevitable tendency of a tubercle is to increase in size and soften,?that the efforts of nature are always pointed in a direction opposed to its cure. This he proves, from Laennec's own lips, to be a perversion of the truth ; that author himself actually describes tubercles, bearing all the appearances of having been partially absorbed and replaced by inorganic salts. But to dwell upon what is now a recognized fact?the absorption of tuberculous matter in the lungs ?is unnecessary. There is this difference, however, in the opinions of M. Fournet and Dr. Carswell, who has most elaborately traced the anatomical evidences of curability, that the former regards the accidental fibrous tissue, formed occasionally in the substance of the lung, and investing particles or masses of cretaceous matter, as evidences, not of Fouiinet on Auscultation, and on the Diagnosis [April, the preexistence of cavities, but simply of that of crude tubercles transformed into that chalky substance. This is, it must be admitted, a very serious modification of previous doctrines, as it transfers the anatomical proofs of cure from the third to the first stage. Our space is exhausted, and we unfortunately cannot condense sufficiently, without weakening their force, the arguments adduced in support of this new view. In respect of the mode of formation of the accidental tissue round tuberculous matter, its successive transformations, its contractile power, and its occasional total disappearance, whereby the pulmonary tissue is left in what, to a superficial observer, appears a state of perfect health, this author follows exactly in the track of Dr. Carswell. Having thus, as he conceives, shown that all the anatomical probabilities are in favour of cure during the first stage, instead of, as has been believed, during the third, M. Fournet emphatically dwells on the encouragement given thereby to our therapeutic efforts; and, in imitation of Sir James Clark, enforces the necessity of general treatment, in the hope of securing such condition of the system at large, as may promote the absorption of the foreign bodies themselves, and obviate a renewed deposition of them. Before we lay by our pen, we deem it advisable to take a rapid general survey of the work we have just analyzed, of its character, its pretensions, and its originality. If we ask, in the first place, to what extent the tone of the author, and the evidences he adduces in its favour testify to his accuracy of observation, without which quality, in scientific research, all others are as dross, the answer must be of a mixed kind. The cases scattered through and appended to the volume have been minutely observed, and ordinarily bear upon them the stamp of faithful narration; but, on the other hand, how shall we repose implicit trust in the statements of a person, whose passion for distinction is such as to lead him into the error, alike silly and despicable, of appropriating to himself, as an original discoverer, truths unfolded by the researches of others ? Again, although he lauds the numerical method with emphasis, he is vastly chary of employing it; and going further even in dogmatism, than authors who use the vague terms " often," " seldom," " little," " much," &c., as expressive of the amount of their experience of particular facts, simply affirms, without hint of their comparative rarity or frequency, that such and such things are. When he does state his experience numerically, the smallness of the number of cases referred to is most striking, and justifies the belief, that when he has not dared to use it, their paucity must be singularly subversive of the importance he would fain have attached to his general announcements. The number of subjects examined after death is not even stated in a straightforward manner; and is, rather accidentally than otherwise, discovered to have been only twentynine.
The volume is spun out to a needless length by hyperdivision of topics, occasionally by useless wordiness and by frequent repetitions. The preface and some scattered sentences, in which the author's estimate of his own doings appears, prove him affected with the quintessence of the national failing?vanity. These singular outbreaks are, however, rather amusing than otherwise.
But, on the other hand, M. Fournet has given a new character of pre-cision to many phenomena already imperfectly known, added some new ones, and taken a more complete and philosophical view of the system of respiratory signs than any previous author. He has pointed out the route, in his diligent investigation of the two murmurs, to, we doubt not, numerous important additions to semeiology, and taught us how much remains unexplored in this, as was hitherto supposed, almost exhausted field : multum adhuc restat operis, multumque restabit, neque ulli nato prsecludetur occasio aliquid adjiciendi. The exact character of his improvements, in the early diagnosis of phthisis, requires explanation. He has assuredly himself learned the means of detecting the disease earlier than his predecessors; some cases of the affection diagnosticated in the very incipient stage (and correctly, as the event proved,) drew this admission from some of the members of the Academy of Medicine. But how has he done this, and to what extent has he taught others the secret? When we knew the contents of his volume only by the preface, we imagined by a number of new signs. So sanguine were we in this respect, that we fancied our future difficulty in the diagnosis of phthisis would be of a very novel kind, that the very multiplicity of decisive signs would almost generate confusion, and, " Like a rich armour worn in heat of day, That scalds with safety," betray us into one evil by the superabundance of protection against another. But we were quickly undeceived. His superior power arises not from the discovery of any new sign, but from his accurate appreciation of signs and symptoms already more or less familiarly known; his careful estimate of their mode of progress ; his judicious combination of local signs, symptoms, and general phenomena; and the perfection to which he has brought his acoustic faculty by laborious training. And from this it follows, that they who turn to M. Fournet's volume, in expectation of finding therein a sign or two, in itself easily ascertained and decisive of the existence of the disease, of here discovering, without expenditure of time or exercise of thought, an easy road to a diagnostic millennium, will be sorely disappointed.* On the contrary, before his pages will serve them at the bedside, they must devote their energies to the attainment of a just notion of the value of each category of symptoms, of the progress of these, and tutor their ear into the nicest power of perception. Finally, M. Fournet has the merit, and it is his highest perhaps, of attempting to turn the attention of his countrymen seriously to therapeutics. Should he succeed in diverting thereto, from its usual sole * Some persons seem to fancy that the mere application of the stethoscope to the chest is quite enough for purposes of diagnosis, that this, by a sort of mesmeric power, discloses all that goes on beneath. And we must in candour admit that these individuals work wonders with the instrument. At least, we lately met a gentleman in consultation, who had hardly placed the stethoscope close to the upper and anterior edge of the precordial region when he detected pleurisy, though the patient, whose pleura was as sound as his own, really laboured under nothing more than capillary bronchitis of the posterior bases of both organs. Unfortunately for persons of this stamp, the stethoscope is not a thinking, speaking, diagnosing being, but simply "1st wie das todte Sprachrohr, das den Schall Empfangt und wiedergiebt, und selbst nicht horet."
The Registrar-GeneraVs Report [April, object, pure pathological anatomy, even a tithe of the energy the French enquirers are wont to exhibit, something may be hoped from his innovation.
And while it will afford no small satisfaction to the profession in this country to find that the branch of medicine, to which they have always directed their chiefest attention, begins to assert its due rank among their continental brethren, the distinguished physician, from whose work on Phthisis M. Fournet's scheme of treatment is derived, must behold with honest pride the sterling improvement to which the dissemination of his labours shall eventually give rise.
[table] Table ,: showing the mode of coexistence of the Morbid Phenomena of Respiration with Inspiration and Expiration.[The order in wbich the different phenomena are set down in each division, exhibits the degree to which they relatively acknowledge the law regulating them all.]A. Morbid Characters coexisting exclusively, or almost exclusively, with inspiration. [/table]
|
BARM and BalticMicrobeDB, a reference metagenome and interface to meta-omic data for the Baltic Sea
The Baltic Sea is one of the world's largest brackish water bodies and is characterised by pronounced physicochemical gradients where microbes are the main biogeochemical catalysts. Meta-omic methods provide rich information on the composition of, and activities within, microbial ecosystems, but are computationally heavy to perform. We here present the Baltic Sea Reference Metagenome (BARM), complete with annotated genes to facilitate further studies with much less computational effort. The assembly is constructed using 2.6 billion metagenomic reads from 81 water samples, spanning both spatial and temporal dimensions, and contains 6.8 million genes that have been annotated for function and taxonomy. The assembly is useful as a reference, facilitating taxonomic and functional annotation of additional samples by simply mapping their reads against the assembly. This capability is demonstrated by the successful mapping and annotation of 24 external samples. In addition, we present a public web interface, BalticMicrobeDB, for interactive exploratory analysis of the dataset.DesignType database creation objective - time series design - observation design - biodiversity assessment objective - gene prediction objective Measurement Type(s) genome assembly - sequence annotation Technology Type(s) DNA sequencing - digital curation Factor Type(s) temporal_instant - depth - independent data collection method Sample Characteristic(s) marine metagenome - Baltic Sea - epeiric sea biomeThe Baltic Sea is a semi-enclosed inland sea characterized by strong physicochemical gradients, in particular horizontal and vertical salinity and oxygen gradients, and pronounced seasonal dynamics 1 . This ecosystem is also heavily impacted by anthropogenic eutrophication, manifested in e.g. harmful phytoplankton blooms and large areas with anoxic bottom waters 2 . Due to their key roles in biogeochemical cycles, microbial communities are particularly interesting to study in this ecosystem 3-11 . One of the most comprehensive methods to characterize the taxonomic and functional composition of microbial communities is through metagenomics, and specifically by metagenomic assembly, which enables high precision and sensitivity for both taxonomic and functional annotation 12 . These annotations can be quantified in individual samples by mapping short reads from samples that either were included in the assembly or constitute external samples. For some microbiomes, particularly those associated with the human body, extensive sequencing efforts have been undertaken to construct reference gene catalogues that are publicly available and can be utilized by others[13][14][15]. Large-scale metagenomic datasets also exist for the global ocean, such as the Tara Oceans dataset 15 . However, although the brackish Baltic Sea is composed of a mixture of marine-and freshwater like lineages 3,5,7,10 , these are genetically distinct from their relatives 8 , which hinders efficient read mapping to fresh-and marine water metagenomes. We here present a Baltic Sea metagenome co-assembly (BARM; BAltic sea Reference Metagenome) with annotated genes constructed from three sets of samples, selected to cover variation over geography, depth and season(Table 1(available online only),Fig. 1; Data Citation 1). After preprocessing of the reads, the 81 samples combined contained 586 billion bases in 2.6 billion read pairs. To allow the assembly of genes also from genomes having low abundance in individual samples, data from all samples were co-assembled. The resulting co-assembly consisted of 14 billion bases distributed over 22 million contigs. Out of these contigs, 2.4 million contigs were longer or equal to 1 kilobase. Functional and taxonomic annotation of genes is computationally demanding. For this reason, and since longer contigs were deemed to be more trustworthy, only genes found on the contigs >1 kilobase were subjected to functional and taxonomic annotations; 6.8 million genes were found on these.For functional analysis, several database sources were chosen; Pfam 16 , TIGRFAM (http://www.jcvi. org/cgi-bin/tigrfams/index.cgi), EggNOG 17 and dbCAN 18 . Additionally, enzyme commission (EC) numbers 19 were extracted based on the EggNOG assignments. Through mapping, the short reads were then used to quantify the individual genes over all the different samples, which were summarized per annotation identifier (ID) for each respective annotation source. The mapping rates for the different sample groups and annotation sources are summarized inFig. 2 and Fig. 3, where inFig. 2also 24 samples from a published metagenomic study 20 (Data Citation 2) of the Baltic Sea are included to illustrate the capabilities for BARM to work as a reference gene catalogue for the Baltic Sea.Along with the dataset, a public web interface (BalticMicrobeDB) was constructed to facilitate exploratory analysis of the data (https://barm.scilifelab.se). Through this it is possible to view counts of functional and taxonomic annotations over the different sample groups. Moreover, it is possible to search for functional annotations based on their descriptive texts and choose to view or download the counts for only those matching the search query.The annotated assembly presented here is a rich resource for further exploitation of the published datasets, facilitated through the web interface, but could also function as a reference metagenome assembly for the Baltic Sea, decreasing the computational demands for the analysis of new metagenome and metatranscriptome samples, and serve as reference for metaproteome analyses.
The samples were sequenced at the National Genomics Infrastructure at Science for Life Laboratory, Stockholm, Sweden, using a full HiSeq 2500 high-output flowcell producing an average of 69.5 million pair-end reads per sample.
The redoxcline samples consist of samples from station Boknis Eck (Data Citation 4), located at the entrance of the Eckernforde Bay in the southwestern Baltic Sea, and from station TF0271 at the Gotland Deep in the eastern Gotland Basin. The Boknis Eck station was sampled on September 23, 2014 on the R/ V Littorina during routine monitoring activities performed monthly by the GEOMAR Helmholtz Centre for Ocean Research Kiel. Due to windy conditions before the sampling day, the water at the Boknis Eck station was mixed over most of the water column and only the bottom water was sulfidic. Water was sampled from the mixed oxygenated layer and from the sulfidic bottom water, which was captured on a 3 μm pore size membrane filters (Whatman, Maidstone, UK) followed by 0.2 μm pore size Sterivex-GV filters (Millipore Billerica, Massachusetts, USA). The Gotland Basin was sampled during the cruise EMB087 on the R/V Elisabeth Mann Borgese on October 18 and October 26, 2014. The samples from October 18 were taken in the context of an experiment close to the oxic-anoxic interface from suboxic and anoxic water layers and were captured directly on 0.2 μm pore size Durapore membrane filters (Whatman, Maidstone, UK). The samples from October 26 were taken to cover different zones in the redox gradient (suboxic, oxic-anoxic interface, upper sulfidic, lower sulfidic) and were captured first on a 3 μm pore size membrane filters (Whatman, Maidstone, UK) followed by 0.2 μm pore size Sterivex-GV filters. DNA from water captured on 3 μm pore size membrane filters and 0.2 μm Sterivex-GV filters was extracted using the QIAmp DNA Mini Kit (Qiagen, Hilden, Germany): ATL buffer was added to filter pieces together with 200 μm low-binding Zirconium beads (OPS Diagnostics, Lebanon, NY, USA) and the suspension was vortexed for 5 minutes at maximum speed. Subsequently proteinase K was added and the suspension was incubated for approximately 1h at 56°C before continuing DNA extraction by following the manufacturer's instructions. Nucleic acids from Gotland Basin water sampled on October 18 on 0.2 μm pore size membrane filters were extracted using the AllPrep DNA/RNA Mini Kit (Qiagen, Hilden, Germany). Similar as before, filters were vortexed together with Zirconium beads in RTL buffer before continuing nucleic acid extraction by following the manufacturer's protocol. The concentration and quality of the eluted DNA was assured by gel electrophoresis. The samples were sequenced on a single HiSeq 2500 lane producing an average of 20.7 million pair-end reads per sample.
All sequencing libraries (including LMO) were prepared with the Rubicon ThruPlex kit (Rubicon Genomics, Ann Arbor, Michigan, USA) according to the instructions of the manufacturer.
## Preprocessing and assembly
The quality of the reads were checked and visualized with FastQC (http://www.bioinformatics.babraham. ac.uk/projects/fastqc/) through MultiQC 22 and trimmed from low quality bases with cutadapt 23 using Phred score 15 as a cutoff. Adapter sequences were also removed using cutadapt, keeping only read pairs where both reads in the pair were longer than 31 bases. Preprocessed reads were then assembled using Megahit 24 version 1.0.2 with default parameters including kmers 21,41,61,81 and 99.
Exclusively to the 30 samples from the transect cruise, genomic material (20 ng per L of seawater) from a known genome of Thermus thermophilus (strain HB8), which is not expected to be present in the Baltic Sea naturally, was added after filtration but prior to the DNA extraction, serving as internal standard to enable absolute quantifications. Aligning all contigs from the metagenome assembly against this reference genome showed that 84.1% of the genome was recovered within contigs aligning with average 99.82% identity. These additional genome contigs were kept in the reference assembly but reads aligning to the reference genome were filtered out before the quantification steps, and before uploading the processed reads to the European Nucleotide Archive (ENA) (Data Citation 5).
## Functional annotation
Genes were predicted on all contigs using Prodigal 25 version 2.6.3 with the '--meta' tag which potentially uses different coding tables for different contigs. Genes located on contigs longer or equal to 1 kilobase, identified with the script toolbox/scripts/fasta_lengths.py, were used for functional and taxonomic annotation. For functional annotation, the databases EggNOG 17 , Pfam 16 , TIGRFAM (http://www.jcvi. org/cgi-bin/tigrfams/index.cgi) and dbCAN [bib_ref] dbCAN: a web resource for automated carbohydrate-active enzyme annotation, Yin [/bib_ref] were chosen. Furthermore, EC-numberswere extracted from the EggNOG annotations.
To annotate genes with EggNOG 17 IDs, the EggNOG hmm file for all organisms, NOG.hmm.tar.gz, version 4.5 was downloaded from http://eggnogdb.embl.de/download/eggnog_4.5/data/NOG/. For performance reasons, hmmsearch was used instead of hmmscan [bib_ref] Accelerated Profile HMM Searches, Eddy [/bib_ref] , initially removing all hits with an E-value >0.0001. To select a maximum of one annotation per gene, the hit with highest score was chosen using the script toolbox/scripts/hmmer_filtering/keep_top_score.py. Information about each annotation was downloaded from http://eggnogdb.embl.de/download/eggnog_4.5/data/NOG/NOG.annotations.tsv.gz.
An Enzyme Commision (EC) numberwas assigned to each EggNOG through the Uniprot 27 proteins included in the EggNOG model, if a majority of its EC-assigned members were assigned to that EC. Note that proteins could have multiple EC numbers assigned and therefore some EggNOGs were assigned multiple EC numbers. The files needed for the conversion were eggnog4.protein_id_conversion. tsv.gz (downloaded from http://eggnogdb.embl.de/download/eggnog_4.5/ on January 9th 2017) and NOG.members.tsv.gz (downloaded from http://eggnogdb.embl.de/download/eggnog_4.5/data/ NOG/ on January 9th 2017). The protein ID conversion file gives EC numbers per reference protein and the members file gives the reference proteins that build each model. The protein with taxaid 400682 and protein ID "PAC" was removed from the protein ID conversion file since it had 695 EC entries. Likewise for taxaid 7070 and protein ID "TCOGS2", with 686 EC entries. The protein ID with the third most entries had 6 entries and therefore the two others were deemed as outliers. The suspected reason is that these entries belong to different genes for these genomes but there were no way to resolve this and the EC-number assignment for each EggNOG was deemed to not be affected by this. Given the assignment of EC-numbers per EggNOG, the assignment per gene was done with toolbox/scripts/assign_ec_from_nog.py.
Annotation against the dbCAN 18 (DataBase for automated Carbohydrate-active enzyme ANnotation) database was performed using version 5 (downloaded from http://csbl.bmb.uga.edu/dbCAN/download. php). Following the instructions from dbCAN (downloaded from http://csbl.bmb.uga.edu/dbCAN/ download/readme.txt), hmmscan [bib_ref] Accelerated Profile HMM Searches, Eddy [/bib_ref] was used with the option --domtblout and the result was further treated with the recommended script hmmscan-parser.sh (reference of used script available within toolbox/third_party_scripts/dbcan/hmmscan-parser.sh) from dbCAN requiring a covered fraction of the HMM larger than 0.3 and keeping long alignments (>80 amino acids) if the E-value was less than 1e-5 and short alignments if the E-value was less than 1e-3. An additional script toolbox/hmmer_filtering/ dbcan_strict_filtering.py was applied, choosing to follow recommendations for bacteria from dbCAN, keeping annotations with e-value less than 1e-18 and alignment coverage greater than 0.35. To allow for more than a single domain within a gene, any annotation which fulfilled these criteria was kept. Information about each annotation was collected (downloaded from http://csbl.bmb.uga.edu/dbCAN/ download/FamInfo.txt).
Annotation against Pfam 16 version 30.0 was conducted with the script pfam_scan.pl supplied from the ftp://ftp.ebi.ac.uk/pub/databases/Pfam/Tools for version 28.0, using hmmer version 3.1b1 (ref. [bib_ref] Accelerated Profile HMM Searches, Eddy [/bib_ref]. To allow for more than a single domain within a gene, any annotation which fulfilled these criteria was kept. Information about each annotation was collected as columns 1,2 and 4 from the file pfamA.txt.gz downloaded from ftp://ftp.ebi.ac.uk/pub/databases/Pfam/releases/Pfam30.0/database_files/ on January 11th 2017.
Annotation against TIGRFAM version 15, was performed using hmmsearch (v. 3.1b2) [bib_ref] Accelerated Profile HMM Searches, Eddy [/bib_ref] with --cut_tc argument to filter models by trusted cutoff. For each protein sequence, the best scoring HMM was selected using hmmparse.py available at https://github.com/johnne/biotools/blob/master/scripts/ hmmparse.py
## Taxonomic annotation
The method used to assign taxonomy was chosen in order to assign as many contigs as possible to a taxonomy while still keeping false positives to a low level. As the number of sequences in reference databases closely related to the genomes in our samples was expected to be low 8 , amino acid sequences from the assembly were used to compare against other amino acid sequences in the reference database, enabling higher sensitivity (due to the more conserved nature of amino acid sequences). This comparison was done using Diamond version 0.8. with the parameters "--seg yes","--sensitive" and "--top 10" against the NCBI nr database downloaded December 2nd 2016.
The code used to assign taxonomy from the Diamond search was based on an original available in the DESMAN package 29 and the modified version of the code is available as the script toolbox/scripts/ taxonomy_from_genes_to_contigs/lca_per_contig.py. The assignment was done as follows: all reported hits from the Diamond search were given a weight based on the aligned fraction of the query and the percentage identity of the alignment. At each taxonomic level, if the sum of the weights for one taxon was greater than half the sum of all weights, the gene was assigned to that taxon as long as the percentage identity was high enough. The levels for the percentage identity were set to 40% at superkingdom level, 50% at phylum level, 60% at class level, 70% at order level, 80% at family level, 90% at genus level and 95% at species level.
Taxonomic assignments were set per contig to the most detailed level where consensus for at least 50% of the weights of the preliminary gene assignments could be achieved. Genes without taxonomic annotation were ignored. The shared assignment was propagated to all genes present on that contig. In this way, all genes present on one contig will always share the taxonomic assignment. If no single superkingdom accounted for a majority of the gene assignment weights for a contig, the contig was left unassigned.
## Quantification and normalization
To use the metagenome assembly as a reference assembly, individual samples are functionally and taxonomically annotated by quantifying the different annotations present in the assembly. This is done by mapping all short reads against the assembly and quantifying genes, and thereby any associated annotation, with the number of reads mapping to them. More specifically bowtie2 (ref. [bib_ref] Fast gapped-read alignment with Bowtie 2, Langmead [/bib_ref] version 0.6.1 was used to get raw counts per gene. Counts per annotation was achieved by summing all counts for genes annotated with each respective annotation.
When quantifying annotation types where multiple annotations were allowed for a single gene (dbCAN and Pfam), some genes contributed several times to the quantities. This was kept in order to facilitate analysis of differential abundance for the individual annotations.
Along with raw counts of reads for each annotation type and taxonomy, a count normalized by gene length and number of mapped reads was also calculated. Analogously to the formula for Transcripts Per Million used in transcriptomics (ref 33), we calculate TPM for gene counts:
[formula] TPM ¼ r g UrlU10 6 f l g UT T ¼ X g A G r g Url f l g [/formula]
Where r g is the number of reads mapped to gene g from the sample, rl is the average read length for the sample, fl g is the length of the gene and G is the set of all genes. T is a convenience variable for the indicated sum over all genes.
## Code availability
Code used to preprocess reads, assemble contigs and annotate genes is publicly available at https://github. com/EnvGen/BLUEPRINT_pipeline, containing the pipeline definition of the workflows used, https:// github.com/EnvGen/snakemake-workflows, where the snakemake rules are specified in order to build the command used for each step, and the branch BARM_publication of https://github.com/EnvGen/toolbox, for custom scripts. Scripts within the latter repository that have been used have been indicated throughout the text.
## Data records
The preprocessed sequencing reads from the Transect and Redoxcline samples were submitted to ENA hosted by EMBL-EBI under the study accession number PRJEB22997 (Data Citation 5). The raw reads
## Technical validation
The mapping rates for all samples included in the reference assembly are shown in [fig_ref] Figure 2: Mapping rates divided on different sample groups [/fig_ref] , where the majority of samples included in the assembly reaches a level above 80%. This serves as a validation of the completeness of the metagenome assembly. The fraction of reads that did not map to the coassembly, and were hence not assembled past the 200 bases length cutoff most likely originate from low abundance species, or species with high intraspecies diversity generating fragmented assemblies. The mapping rate of the external samples shows the capability for this assembly to serve as a reference metagenome assembly for the Baltic Sea. These external samples [bib_ref] Picocyanobacteria containing a novel pigment gene cluster dominate the brackish water Baltic..., Larsson [/bib_ref] Assignment rates for different annotation types, as shown in [fig_ref] Figure 3: Fraction of reads mapping to genes annotated with respective database [/fig_ref] , are in the majority of cases below 10% of the total number of reads, which is expected since only genes on long contigs (representing 40% of the bases of the total assembly) were predicted and subjected to annotation. The fraction of reads annotated among reads mapping to genes included in the annotation procedure reaches well over 30% for Pfam and shows the generality of that database as compared to i.e. dbCAN, a much more niched resource, which reaches only around 2% of reads mapping to genes included in the annotation. The functional annotation was further validated through an NMDS plot [fig_ref] Figure 4: Non-metric dimensional scaling [/fig_ref] based on the EggNOG annotations of the transect data. Depth was found to be negatively correlated with the first dimension (Spearman's rank correlation ρ = − 0.73, P = 5.4 − 06 ) and salinity was negatively correlated with the second dimension (Spearman's rank correlation ρ = − 0.77, P = 2.4 − 06 ). These two environmental parameters have previously been found to correlate strongly with the microbial community in the Baltic Sea 5 which strengthens our trust in the EggNOG annotations. Furthermore, analyzing a single annotation with a known function, namely the photosynthetic reaction centre protein (PF00124), we could see a strong negative correlation with sampling depth over the thirty transect samples (Spearman correlation coefficient ρ = − 0.87, P = 3.1 -10 ).
The taxonomic annotation was validated by inspecting the taxonomic profile of the transect samples. The same dominant prokaryotic taxonomic groups were observed as in previous pan-Baltic amplicon sequencing and metagenomic studies [bib_ref] Functional tradeoffs underpin salinity-driven divergence in microbial community composition, Dupont [/bib_ref] [bib_ref] Diversity of Pico-to Mesoplankton along the 2000 km Salinity Gradient of the..., Hu [/bib_ref] [bib_ref] Phylogenetic Signals of Salinity and Season in Bacterial Community Composition Across the..., Herlemann [/bib_ref] , and the overall trends were conserved with an increase in Alpha-and Gammaproteobacteria and a decrease in Actinobacteria and Betaproteobacteria with increasing salinity levels [fig_ref] Figure 5: Taxonomic profiles of the 10 transect samples obtained from surface waters [/fig_ref].
Among the predicted proteins in BARM, 98% lacked hits with amino acid identities above 95%, hence potentially representing species for which sequenced genomes are lacking [bib_ref] Towards a genome-based taxonomy for prokaryotes, Konstantinidis [/bib_ref]. 31% of the sequences lacked significant hits (E-value >1) and potentially correspond to novel protein families.
## Usage notes
A publicly available repository at https://github.com/EnvGen/BARM_tools hosts instructions and a pipeline on how to quantify genes and their annotations within BARM for any kind of Baltic Sea metagenomic and metatranscriptomic samples.
The web interface BalticMicrobeDB, available to the public at http://barm.scilifelab.se, can be used to explore and access data for the three sample sets that the assembly is based upon. At the index page, the user can choose whether to access functional annotations or taxonomic annotations. For the functional annotations, the user can select specific annotation sources and identifiers and select the sample groups for which the counts will be displayed. Furthermore, a text search over the identifiers and the descriptions of the annotations can be used to create a custom table of counts over the selected samples. For taxonomic annotations, counts for the top level superkingdom are first presented but the user can unfold a taxonomic tree to select any taxon to view counts for.
[fig] Figure 1: A map showing the locations for all stations where samples were taken. The three sample groups included in the assembly (Transect, LMO and Redoxcline) are displayed together with the external sample set 20 (External), all groups indicated with different markers. The colour of the marker indicates the salinity of the water sample while the size indicates the depth at which it was taken. The background color indicates depth (from white to dark blue), with contour lines drawn with 50 m intervals. The map was generated using the Marmap package36 in R 37 with bathymetric data from the ETOPO1 dataset hosted on the NOAA server38 . [/fig]
[fig] Figure 2: Mapping rates divided on different sample groups. Mapping rates are calculated by Bowtie2 (ref. 30) as the "overall alignment rate". The three first sample groups; LMO 2012 (N = 37, 0.2-3.0 μm), Baltic Transect 2014 (N = 30, >0.2 μm) and Baltic Redoxcline 2014 (N = 6, 0.2-3.0 μm; N = 6, >3.0 μm; N = 2, >0.2 μm) were included in the assembly, while the four last sample groups; External Samples o 0.1 μm (N = 6), External Samples 0.1-0.8 μm (N = 6), External Samples 0.8-3.0 μm (N = 6) and External Samples 3.0-200 μm (N = 6) were not. The size intervals of the external samples indicate filter pore sizes used to tentatively separate viruses, free-living prokaryotes, and small and larger particles as well as Eukaryotic cells, respectively 34 . Created using Matplotlib 39 and Seaborn 40 . [/fig]
[fig] Figure 3: Fraction of reads mapping to genes annotated with respective database. Only genes identified on contigs longer than 1 kilobase were subjected to annotation, defining the 'included genes' category. N = 81 for all categories. Created using Matplotlib39 and Seaborn 40 . www.nature.com/sdata/ SCIENTIFIC DATA | 5:180146 | DOI: 10.1038/sdata.2018.146from LMO were published elsewhere 8 and are accessible at NCBI (Data Citation 3). Contig, gene and protein sequences from the co-assembly of the Transect, Redoxcline and LMO samples, as well as quantification tables, contextual data for the samples, and the annotations for each gene are accessible on Figshare (Data Citation 1). The raw sequencing reads from the external samples used for evaluation were also published elsewhere34 and are accessible at NCBI (Data Citation 2). [/fig]
[fig] Figure 4: Non-metric dimensional scaling (NMDS) of the 30 samples included in the Transect sample group based on EggNOG annotation. Samples are colored and sized according to salinity and depth, respectively. Created using Matplotlib 39 and Seaborn 40 . www.nature.com/sdata/ SCIENTIFIC DATA | 5:180146 | DOI: 10.1038/sdata.2018.146 [/fig]
[fig] Figure 5: Taxonomic profiles of the 10 transect samples obtained from surface waters. Numbers on x-axis indicate salinity, given in practical salinity units (PSU), and are sorted with increasing salinity to the right. Created using Matplotlib 39 and Seaborn 40 . www.nature.com/sdata/ SCIENTIFIC DATA | 5:180146 | DOI: 10.1038/sdata.2018.146 [/fig]
|
Antimicrobial peptides and nitric oxide production by neutrophils from periodontitis subjects
Neutrophils play an important role in periodontitis by producing nitric oxide (NO) and antimicrobial peptides, molecules with microbicidal activity via oxygen-dependent and -independent mechanisms, respectively. It is unknown whether variation in the production of antimicrobial peptides such as LL-37, human neutrophil peptides (HNP) 1-3, and NO by neutrophils influences the pathogenesis of periodontal diseases. We compared the production of these peptides and NO by lipopolysaccharide (LPS)stimulated neutrophils isolated from healthy subjects and from patients with periodontitis. Peripheral blood neutrophils were cultured with or without Aggregatibacter actinomycetemcomitans-LPS (Aa-LPS), Porphyromonas gingivalis-LPS (Pg-LPS) and Escherichia coli-LPS (Ec-LPS). qRT-PCR was used to determine quantities of HNP 1-3 and LL-37 mRNA in neutrophils. Amounts of HNP 1-3 and LL-37 proteins in the cell culture supernatants were also determined by ELISA. In addition, NO levels in neutrophil culture supernatants were quantitated by the Griess reaction. Neutrophils from periodontitis patients cultured with Aa-LPS, Pg-LPS and Ec-LPS expressed higher HNP 1-3 mRNA than neutrophils from healthy subjects. LL-37 mRNA expression was higher in neutrophils from patients stimulated with Aa-LPS. Neutrophils from periodontitis patients produced significantly higher LL-37 protein levels than neutrophils from healthy subjects when stimulated with Pg-LPS and Ec-LPS, but no difference was observed in HNP 1-3 production. Neutrophils from periodontitis patients cultured or not with Pg-LPS and Ec-LPS produced significantly lower NO levels than neutrophils from healthy subjects. The significant differences in the production of LL-37 and NO between neutrophils from healthy and periodontitis subjects indicate that production of these molecules might influence individual susceptibility to important periodontal pathogens.
# Introduction
Periodontitis is a chronic inflammatory disease characterized by connective tissue and alveolar bone destruction, eventually leading to tooth loss. The recruitment of neutrophils and other leukocytes in the periodontal pocket is an important feature of the inflammatory process in periodontal disease [bib_ref] Detection of individual human neutrophil alpha-defensins (human neutrophil peptides 1, 2 and..., Lundy [/bib_ref]. Important periodontopathogens, e.g., Aggregatibacter actinomycetemcomitans and Porphyromo-nas gingivalis, seem to induce destruction of periodontal tissues through lipopolysaccharide (LPS)-dependent mechanisms or by eliciting the production of a variety of biologically active substances by host immune cells [bib_ref] Bacterial DNA activates human neutrophils by a CpG-independent pathway, Trevani [/bib_ref]. In general, LPS activates different cells, including epithelial cells, fibroblasts, neutrophils and macrophages, by activating Toll-like receptor-4 (TLR4) signaling pathways [bib_ref] Toll-like receptor-4 is involved in eliciting an LPS-induced oxidative burst in neutrophils, Remer [/bib_ref] , promoting phagocytosis, production of reactive oxygen species, cytokines, and release of antimicrobial peptides from azurophil granules [bib_ref] Human variability in innate immunity, Kinane [/bib_ref].
Neutrophils respond to bacterial products or inflammatory mediators by chemotaxis, phagocytosis, and microbial killing through oxygen radical-and non-oxygen-dependent mechanisms [bib_ref] Regulation of the neutrophil-mediated inflammatory response to infection, Kobayashi [/bib_ref]. The non-oxidative antibacterial mechanisms involve a diverse group of antimicrobial peptides including alpha-defensins [bib_ref] Dying and necrotic neutrophils are anti-inflammatory secondary to the release of alpha-defensins, Miles [/bib_ref] and human cathelicidin, LL-37 [bib_ref] Analysis of neutrophil-derived antimicrobial peptides in gingival crevicular fluid suggests importance of..., Puklo [/bib_ref] , which are stored in azurophil granules. The defensins are divided into alpha, beta and theta defensins and are found in vertebrates and invertebrates. Alpha-defensins include a group of six types of small peptides [bib_ref] Dying and necrotic neutrophils are anti-inflammatory secondary to the release of alpha-defensins, Miles [/bib_ref] , four of which are produced by human neutrophils, thus being designated human neutrophil peptides (HNP) -1, -2, -3, and -4 [bib_ref] Dying and necrotic neutrophils are anti-inflammatory secondary to the release of alpha-defensins, Miles [/bib_ref]. HNP 1-3 are the most abundant, and individual analysis of these peptides is difficult because of the high similarity of their amino acid sequences [bib_ref] Primate defensins, Lehrer [/bib_ref]. HNP 1-3 are detected in whole saliva [bib_ref] Determination of defensin HNP-1, HNP-2, and HNP-3 in human saliva by using..., Goebel [/bib_ref] and gingival crevicular fluid of both healthy and periodontal disease subjects [bib_ref] Detection of individual human neutrophil alpha-defensins (human neutrophil peptides 1, 2 and..., Lundy [/bib_ref]. LL-37 is known to play an important role in the lysis of periodontal pathogens, which is consistent with the frequent development of severe periodontal diseases in subjects with Morbus Kostmann syndrome. In this congenital disease, neutrophils are deficient in LL-37, and this antimicrobial peptide is also present in abnormally low levels in saliva [bib_ref] Deficiency of antibacterial peptides in patients with Morbus Kostmann: an observation study, Putsep [/bib_ref]. Additionally, neutrophils from these subjects produce reduced amounts of HNP 1-3 [bib_ref] Analysis of neutrophil-derived antimicrobial peptides in gingival crevicular fluid suggests importance of..., Puklo [/bib_ref]. LL-37 impairs the in vitro growth of several bacterial species of the oral cavity, including the periodontal pathogens A. actinomycetemcomitans [bib_ref] Susceptibility of various oral bacteria to antimicrobial peptides and to phagocytosis by..., Ji [/bib_ref] , Fusobacterium nucleatum and Prevotella intermedia [bib_ref] Susceptibilities of periodontopathogenic and cariogenic bacteria to antibacterial peptides, {beta}-defensins and LL37,..., Ouhara [/bib_ref]. Additionally, LL-37 shows high affinity for LPS from different bacteria and thus can neutralize these endotoxins [bib_ref] Antimicrobial and chemoattractant activity, lipopolysaccharide neutralization, cytotoxicity, and inhibition by serum of..., Ciornei [/bib_ref].
Defensins and LL-37 have the ability to kill and/or inactivate several bacterial species (including A. actinomycetemcomitans and P. gingivalis) [bib_ref] Susceptibility of various oral bacteria to antimicrobial peptides and to phagocytosis by..., Ji [/bib_ref] , fungi, and some enveloped viruses [bib_ref] Peptides in oral diseases, Lucchese [/bib_ref]. These microbicidal properties lie in the ability of these peptides to form pores in the membranes, promoting bacterial lysis or affecting viral envelopes [bib_ref] Antimicrobial peptides: pore formers or metabolic inhibitors in bacteria?, Brogden [/bib_ref]. In addition, HNP 1-3 and LL-37 induce and/or amplify subsequent innate and adaptive immune responses against pathogens, such as mast cell degranulation, production of interleukins (IL) such as IL-10 and tumor necrosis factor, and dendritic cell activation [bib_ref] Salivary defense proteins: their network and role in innate and acquired oral..., Fábián [/bib_ref] [bib_ref] Cationic host defence peptides: multifaceted role in immune modulation and inflammation, Choi [/bib_ref]. The production of nitric oxide (NO) represents another mechanism of pathogen destruction in activated neutrophils [bib_ref] Nitric oxide and the immune response, Bogdan [/bib_ref]. Production of NO or expression of inducible NO synthase (iNOS) by peripheral neutrophils or in gingival tissues was associated with periodontal disease in human and animal models, respectively [bib_ref] Neutrophils in chronic and aggressive periodontitis in interaction with Porphyromonas gingivalis and..., Guentsch [/bib_ref] [bib_ref] The essential role of IFN-gamma in the control of lethal Aggregatibacter actinomycetemcomitans..., Garlet [/bib_ref]. Other studies have shown that production of NO and reactive oxygen species is enhanced by peripheral neutrophils stimulated in vitro with A. actinomycetemcomitans and P. gingivalis [bib_ref] Oxidative and nonoxidative killing of Actinobacillus actinomycetemcomitans by human neutrophils, Miyasaki [/bib_ref] [bib_ref] Deficiency of iNOS contributes to Porphyromonas gingivalisinduced tissue damage, Alayan [/bib_ref]. Despite the massive presence of neutrophils and their enhanced activity at sites of periodontal disease [bib_ref] Neutrophil-mediated host response to Porphyromonas gingivalis, Kantarci [/bib_ref] , the roles of these leukocytes and their antimicrobial prod-ucts in the susceptibility and/or pathogenesis of periodontal disease remain to be examined. Although neutrophils have been considered to be responsible for the destruction of periodontal tissues [bib_ref] Neutrophil-mediated host response to Porphyromonas gingivalis, Kantarci [/bib_ref] , some studies have suggested that neutrophils play protective roles in controlling pathogenic bacteria involved in periodontal disease [bib_ref] Role of the polymorphonuclear leukocyte in periodontal health and disease, Miller [/bib_ref] [bib_ref] Neutrophils in periodontal inflammation, Scott [/bib_ref].
In the present study, we investigated the expression of genes encoding HNP 1-3, LL-37, NO, and HNP 1-3/LL-37 production in neutrophils isolated from generally healthy subjects with and without periodontitis, in response to bacterial LPS from species considered or not to be periodontal pathogens.
# Subjects and methods
## Subjects
All subjects enrolled in this study were clinically and radiographically examined at the Graduate Clinic of the Faculdade de Odontologia de Piracicaba, Universidade Estadual de Campinas (UNICAMP), Brazil. Periodontal examination included full-mouth probing depth, plaque index, gingival index, and gingival recession. Inclusion criteria were: subjects diagnosed with generalized moderate chronic periodontitis (at least 4 teeth with probing depth ≥5 mm) who had not received periodontal treatment and/ or antibiotics in the last 6 months preceding the study, with at least 14 natural teeth. Exclusion criteria were: systemic modifying factors affecting the immune response, such as diabetes mellitus, immune and hormone disorders, smokers and former smokers, alcoholics and former alcoholics, pregnant and lactating women, and those taking oral contraceptive drugs.
The subjects were separated into two experimental groups. The chronic periodontitis group included 6 patients, 3 males and 3 females (mean age: 47.5 ± 11.8 years) with moderate chronic periodontitis according to the criteria proposed by the 1999 International Workshop for a Classification of Periodontal Diseases and conditions [bib_ref] Development of a classification system for periodontal diseases and conditions, Armitage [/bib_ref]. The healthy control group included 6 subjects, three males and three females (mean age: 31.4 ± 3.4 years). All subjects were clinically healthy (probing depth <3 mm, without bleeding on probing, with no detectable radiographic alveolar bone loss on radiography). All procedures performed were approved by the Ethics Committee of Faculdade de Odontologia de Piracicaba, UNICAMP, and all volunteers involved in this study signed a consent form.
## Bacterial strains, growth conditions and antigens
Periodontopathogens A. actinomycetemcomitans strain Y4 and P. gingivalis strain ATCC 33277 were obtained from the Bacterial Collection of the Research Group in Oral Biology (GREB), School of Dentistry, Laval University, Quebec, Canada. These microorganisms were cultured at 37°C in brain heart agar (Difco Co., USA) supplemented with 7% defibrinated sheep blood, 5 mg/mL hemin and 1 mg/ [bib_ref] The effects of nicotine and cotinine on Porphyromonas gingivalis colonisation of epithelial..., Cogo [/bib_ref]. Purification of LPS from A. actinomycetemcomitans and P. gingivalis was performed as described [bib_ref] Procedure for isolation of bacterial lipopolysaccharides from both smooth and rough Pseudomonas..., Darveau [/bib_ref]. LPS samples were freeze-dried and stored at -20°C. Protein contamination of these samples was lower than 0.001% in all preparations, as evaluated using a Quick Start™ Bradford protein assay with bovine serum albumin standard (Bio-Rad Laboratories, Canada). Escherichia coli-LPS (Ec-LPS) obtained from Sigma was used as positive control.
## Isolation and in vitro stimulation of human neutrophils
Whole blood (20 mL) from control and periodontitis subjects was collected into lithium heparin tubes (BD Vacu-tainer™, USA). Neutrophils purified from 5 mL Histopaque 1119 (Sigma-Aldrich Brazil Ltda., Brazil) were poured into a 15-mL round bottom tube and overlaid with 3 mL Histopaque 1083 (Sigma-Aldrich Brazil Ltda.), and 6 mL whole blood was layered over the gradients. Tubes containing gradients and blood were centrifuged at 460 g for 28 min at 25°C. The layer containing the neutrophils was aspirated and washed twice with cold RPMI 1640 (Invitrogen™, Brazil), supplemented with 10% fetal bovine serum and antibiotics. The erythrocytes were eliminated by hypotonic lysis for 30 s. The viability of blood neutrophils was determined by Trypan blue exclusion using a Neubauer chamber. One x 10 7 cells/ well were incubated at 37°C in the presence of 5% CO 2 for 1 h. Neutrophils were then stimulated with 100 ng/mL of A. actinomycetemcomitans-LPS (Aa-LPS), P. gingivalis-LPS (Pg-LPS) or Ec-LPS. In order to examine the effects of LPS on gene expression and production of HNP 1-3 or LL-37 by neutrophils, the cells isolated from blood were stimulated with LPS for 6 and 12 h. Neutrophils from similar cultures not exposed to LPS were used as negative controls. After incubation, cells were centrifuged (290 g, at 4°C for 10 min), and the culture supernatants were stored at -70°C until use for the determination of HNP 1-3 and LL-37 production. The neutrophils were used for the total mRNA extraction for the analysis of HNP 1-3 and LL-37 gene expression.
## Rna extraction and reverse transcription
Neutrophils from cultures exposed or not to LPS were washed twice with PBS and the total RNA was immediately extracted using Trizol reagent (Invitrogen, USA), according to manufacturer instructions. After treatment of RNA samples with DNase (Turbo DNA-free, Ambion Inc., USA), a total of 0.5 µg RNA was used for cDNA synthesis. Reverse transcription was carried out using the Transcriptor First-Strand cDNA Synthesis kit (Roche Diagnostic Co., USA) according to manufacturer instructions.
## Quantitative real-time polymerase chain reaction (qrt-pcr)
The levels of HPN 1 and HPN 3 transcripts were assessed using a primer set specific for a conserved sequence of these peptides [bib_ref] High alpha-defensin levels in patients with systemic lupus erythematosus, Sthoeger [/bib_ref]. [fig_ref] Table 1: Primer sequences for the genes, amplification profile, and amplicon length studied here [/fig_ref] shows the primer sequences applied in qRT-PCR for HPN 1-3 and LL-37.
The sequences of the primers for qRT-PCR analysis were obtained using the Light-Cycler software (Roche Diagnostics GmbH, Germany). The qRT-PCR assays were performed using the Light Cycler system (Roche Diagnostics GmbH). The reaction mixture (15 µL) included template cDNA (0.5 µg), 0.75 µM of each primer, and 1X SYBR-Green mix (Fast Start DNA Master plus, Roche Diagnostic Co.). The thermal cycling conditions were: 95°C for 10 min for the initial denaturation, followed by 40 cycles of three steps consisting of denaturation at 95°C for 10 s, primer annealing at 59°C for 4 s, and primer extension at 72°C for 5 s. Controls included reaction mixtures without template cDNA. The housekeeping gene, glyceraldehyde-3-phosphate dehydrogenase (GAPDH), was amplified in parallel to the gene of interest. Relative copy numbers compared with GAPDH were calculated using 2 ∆Ct . Assays with RNA samples were performed in triplicate.
## Enzyme-linked immunosorbent assay (elisa)
The amounts of HNP 1-3 and LL-37 produced by neutrophils exposed or not to LPS samples were quantified by ELISA using kits from Hycult Biotechnology (The Netherlands), according to manufacturer instructions. Individual quantification of each HNP subtype (1, 2, and 3) is not possible because of the high structural similarity of these defensins. The concentrations of HNP 1-3 and LL-37 in culture fluids are reported as pg/mL and ng/mL, respectively. Assays were performed in triplicate with the supernatants of neutrophil culture.
# Flow cytometry analysis
Cell acquisition was performed with a FACSort flow cytometer using the CellQuest software (BD Biosciences, USA). The absolute leukocyte values/20 mL peripheral blood were calculated according to characteristics of granularity and size (side scatter vs forward scatter) based on the percentage obtained by FACS (50,000 events), and the amount of cells was determined in a Neubauer chamber.
## Nitric oxide production
Nitrate was reduced to nitrite with nitrate reductase and the concentration of nitrite was determined by mixing culture supernatants of neutrophils exposed or not to LPS with Griess reagent, as described below. Briefly, neutrophils (1 x 10 7 cells/mL) from subjects with periodontal disease or healthy ones were cultured in the presence or absence of Aa-LPS, Pg-LPS or Ec-LPS (100 ng/mL) for 6 and 12 h, at 37°C and in the presence of 5% CO 2 . A total of 50 µL culture supernatant was then incubated at room temperature with an equal volume of the Griess reagent on 96-well plates (Corning, USA) for 30 min. The absorbance (A 540 nm ) of the samples was measured using a plate scanner (VersaMax Tunable Microplate Reader; Molecular Devices, USA). The amounts of NO 2 in the samples were then calculated using a standard curve of NaNO 2 (1-200 µM) within a linear range. Assays were performed in triplicate with supernatants of neutrophil cultures from each subject.
## Statistical analysis and software
The Student t-test was used to determine the statistical significance of the differences observed between the volunteers with and without periodontitis and periodontally healthy volunteers. The statistical analyses were performed with the help of the GRAPHPAD PRISM software version 5.00 for Windows (http://www.graphpad.com). Differences were considered to be statistically significant at P values of <0.05.
# Results
## Expression of hnp 1-3 and ll-37 in peripheral blood neutrophils
The purity of cells isolated from peripheral blood of patients with periodontitis and healthy subjects was assessed by flow cytometry [fig_ref] Figure 1: HNP 1-3 and LL-37 mRNA expression in peripheral neutrophils from periodontitis and... [/fig_ref]. Approximately 94% total gated cells were polymorphonuclear neutrophils, thus ensuring a minimum influence of cell contaminants in the experiments performed. Subsequently, it was verified that Aa-LPS, Pg-LPS and Ec-LPS could induce different levels of HNP 1-3 and LL-37 mRNA in neutrophils from patients and healthy subjects. According to our data, neutrophils from patients with periodontitis express higher amounts of mRNA corresponding to HNP 1-3 than control cells after Pg-LPS stimuli for 6 and 12 h [fig_ref] Figure 1: HNP 1-3 and LL-37 mRNA expression in peripheral neutrophils from periodontitis and... [/fig_ref]. Similarly, Aa-LPS and Ec-LPS induced high levels of HNP 1-3 transcripts in neutrophils from periodontitis patients compared to neutrophils from control subjects after 12 h of incubation [fig_ref] Figure 1: HNP 1-3 and LL-37 mRNA expression in peripheral neutrophils from periodontitis and... [/fig_ref]. However, no significant difference was observed in the levels of HNP 1-3 transcripts between neutrophils from the two groups that were cultured with Aa-LPS for a shorter period of time . No significant differences in HNP 1-3 transcripts were detected between neutrophils from the periodontitis and healthy groups when cells were cultured without the LPS stimulus for 12 h [fig_ref] Figure 1: HNP 1-3 and LL-37 mRNA expression in peripheral neutrophils from periodontitis and... [/fig_ref].
The LL-37 transcript levels were closely similar for neutrophils from periodontitis patients and healthy subjects after stimulation or not with LPS for 6 and 12 h [fig_ref] Figure 1: HNP 1-3 and LL-37 mRNA expression in peripheral neutrophils from periodontitis and... [/fig_ref]. An exception were the neutrophils of periodontitis patients exposed to Aa-LPS for 12 h, which showed significantly higher levels of LL-37 transcripts compared to neutrophils from healthy controls exposed to Aa-LPS under the same conditions [fig_ref] Figure 1: HNP 1-3 and LL-37 mRNA expression in peripheral neutrophils from periodontitis and... [/fig_ref].
## Production of hnp 1-3 and ll-37 proteins by peripheral blood neutrophils
The neutrophils from healthy subjects tended to produce higher concentrations of alpha-defensin proteins than cells from periodontitis subjects, but the differences were not statistically significant [fig_ref] Figure 2: HNP 1-3 and LL-37 production by peripheral neutrophils [/fig_ref]. Additionally, similar levels of LL-37 protein were produced by neutrophils from periodontitis and healthy subjects regardless of LPS stimulation for 6 h [fig_ref] Figure 2: HNP 1-3 and LL-37 production by peripheral neutrophils [/fig_ref]. Neutrophils from periodontitis subjects stimulated with Pg-LPS and Ec-LPS for 12 h produced significantly higher amounts of LL-37 protein compared to neutrophils from healthy subjects similarly stimulated [fig_ref] Figure 2: HNP 1-3 and LL-37 production by peripheral neutrophils [/fig_ref]. However, neutrophils from both groups produced similar amounts of total protein after a 12-h stimulation with Aa-LPS or without LPS exposure [fig_ref] Figure 2: HNP 1-3 and LL-37 production by peripheral neutrophils [/fig_ref].
## Differentiated production of no by neutrophils
Neutrophils from periodontitis patients cultured with Pg-LPS or Ec-LPS for 6 h or cultured without LPS stimulation produced significantly lower levels of NO levels when compared to neutrophils from healthy subjects [fig_ref] Figure 3: NO production by neutrophils stimulated with LPS [/fig_ref]. After 12 h of incubation, NO production was still significantly lower in cultures of neutrophils from periodontitis patients stimulated with Pg-LPS and Ec-LPS compared to cells of control subjects. However, equivalent NO levels were produced by neutrophils isolated from periodontitis and healthy subjects stimulated with Aa-LPS for 6 and 12 h. No significant differences in the amounts of NO were observed between neutrophils from the two groups not stimulated with LPS [fig_ref] Figure 3: NO production by neutrophils stimulated with LPS [/fig_ref].
# Discussion
In periodontal disease, neutrophil deficiency may be associated with the worsening of clinical symptoms of periodontitis [bib_ref] Neutrophils in periodontal inflammation, Scott [/bib_ref]. Neutrophils are also recognized as the major source of alpha-defensins (or HNP) and LL-37. In humans, these antimicrobial peptides seem to play a key role in protecting the periodontal tissues against oral bacteria [bib_ref] Susceptibility of various oral bacteria to antimicrobial peptides and to phagocytosis by..., Ji [/bib_ref] [bib_ref] Susceptibilities of periodontopathogenic and cariogenic bacteria to antibacterial peptides, {beta}-defensins and LL37,..., Ouhara [/bib_ref] neutrophils to normal blood counts and reduces recurrent infections, but does not prevent severe periodontitis [bib_ref] Periodontal disease in patients from the original Kostmann family with severe congenital..., Carlsson [/bib_ref]. Peripheral neutrophils from these patients are deficient in LL-37 production and produce reduced amounts of HNP 1-3 defensins. However, whether neutrophils from periodontitis patients have atypical responses to components of periodontal pathogens (e.g., LPS) compared to periodontally healthy subjects remains to be investigated. In the present study, we show that neutrophils from subjects with periodontitis respond differently to stimuli with LPS from different bacterial sources (periodontal pathogens and E. coli) compared to neutrophils from healthy subjects, expressing significantly higher levels of HNP 1-3 and LL-37 genes and showing reduced amounts of NO.
Although the number of subjects with periodontitis studied was relatively small, given the complex nature of chronic periodontitis, the data presented here open a new line of investigation about the influence of neutrophil phenotypes on the course of periodontal diseases. Previous studies using a similar number of patients with aggressive and chronic periodontitis (6 and 12 subjects, respectively) were able to detect significant differences in neutrophil phagocytosis and killing of A. actinomycetemcomitans and P. gingivalis when compared to neutrophils from healthy subjects [bib_ref] Neutrophils in chronic and aggressive periodontitis in interaction with Porphyromonas gingivalis and..., Guentsch [/bib_ref]. In addition, the profiles of transcription and protein production of LL-37 were previously determined in gingival tissues of the same number of volunteers with chronic periodontitis [bib_ref] Antimicrobial peptide hCAP-18/LL-37 protein and mRNA expressions in different periodontal diseases, Turkoglu [/bib_ref]. Further studies will be necessary to compare the phenotypes of peripheral and periodontal neutrophils between subjects with and without periodontal diseases.
It is known that the amounts of mRNA from peripheral blood neutrophils can be consistently quantified by molecular biology methods [bib_ref] Increased levels of leukocyte-derived MMP-9 in patients with stable angina pectoris, Jonsson [/bib_ref]. We measured the amounts of HNP 1-3 and LL-37 transcripts in neutrophils from blood samples exposed to different types of LPS. We observed that neutrophils harvested from peripheral blood of periodontitis patients expressed higher HNP 1-3 mRNA levels when stimulated with Aa-LPS, Pg-LPS and Ec-LPS, specifically during 12 h of exposure. However, the protein levels of HNP 1-3 produced did not vary significantly between neutrophils from healthy versus periodontitis subjects. In a previous study, the amounts of HNP 1-3 were determined in the gingival crevice fluid of sites affected or not by periodontitis from periodontally healthy subjects or subjects with periodontitis using mass-spectrometry analysis [bib_ref] Detection of individual human neutrophil alpha-defensins (human neutrophil peptides 1, 2 and..., Lundy [/bib_ref]. The authors demonstrated no significant differences in the amounts of these peptides detected between healthy and diseased sites, although these defensins seem to be more abundant in healthy gingival crevicular fluid (1). In another study, no association was observed between HNP 1-3 levels in gingival fluid and periodontal status [bib_ref] Evaluation of gingival crevicular fluid adrenomedullin and human neutrophil peptide 1-3 levels..., Turkoglu [/bib_ref]. On the other hand, low levels of HNP 1-3 in neutrophils from Morbus Kostmann syndrome patients were associated with the development of chronic periodontal disease in these patients [bib_ref] Deficiency of antibacterial peptides in patients with Morbus Kostmann: an observation study, Putsep [/bib_ref]. The discrepancies between these studies might be due to different models used to obtain clinical samples and protein quantification. It has been recognized that periodontal infections elicit systemic inflammatory responses likely due to transient bacteremias from periodontitis-affected sites [bib_ref] Systemic markers of inflammation in periodontitis, Loos [/bib_ref]. These events may ultimately influence peripheral neutrophil concentration and activation [bib_ref] Systemic markers of inflammation in periodontitis, Loos [/bib_ref]. The present study shows that circulating neutrophils from periodontitis-affected subjects are more responsive to LPS, expressing higher levels of antimicrobial peptides compared to neutrophils from periodontally healthy subjects. Although no significant differences were observed in the amounts of HNP 1-3 in the supernatants of neutrophil cultures between these two groups, it is possible that alpha-defensin produced after LPS stimulation remained stored in neutrophil azurophil granules, which contain up to 50% HNP 1-3 [bib_ref] Defensins: antimicrobial peptides of innate immunity, Ganz [/bib_ref].
To our knowledge, there are no published studies that correlate the cellular expression and release of antimicrobial peptides in neutrophil culture supernatants. Furthermore, no correlation between levels of transcripts and antimicrobial peptides released by neutrophils may be expected, since these cells even at rest have significant amounts of preformed antimicrobial peptides in cytosolic granules. Thus, newly activated neutrophils can secrete LL-37 and HNP 1-3 without de novo synthesis of these peptides. Further studies will be necessary to characterize the contents of azurophilic and specific granules of peripheral neutrophils from subjects with periodontitis. High LL-37 expression and production was previously reported for local neutrophils from gingival tissues of periodontitis patients, as evaluated by qRT-PCR and immunohistochemistry [bib_ref] Evaluation of gingival crevicular fluid adrenomedullin and human neutrophil peptide 1-3 levels..., Turkoglu [/bib_ref] [bib_ref] Innate immune peptide LL-37 displays distinct expression pattern from beta-defensins in inflamed..., Hosokawa [/bib_ref] , as also observed in our study. Concentrations of LL-37 in gingival tissue homogenates were also positively correlated with the depth of the gingival pockets [bib_ref] Innate immune peptide LL-37 displays distinct expression pattern from beta-defensins in inflamed..., Hosokawa [/bib_ref]. LL-37 has variable bactericidal activity against diverse bacterial species such as A. actinomycetemcomitans, P. gingivalis, F. nucleatum, Streptococcus sobrinus, and Streptococcus mutans [bib_ref] Susceptibilities of periodontopathogenic and cariogenic bacteria to antibacterial peptides, {beta}-defensins and LL37,..., Ouhara [/bib_ref]. There is evidence that LPS from different bacterial species differs regarding the ability to activate TLR. P. gingivalis expresses a heterogeneous mixture of lipid A species that can induce cell activation through TLR2 or TLR4, while Aa-LPS and Ec-LPS interact with TLR4 only [bib_ref] Aggregatibacter actinomycetemcomitans lipopolysaccharide regulates bone sialoprotein gene transcription, Li [/bib_ref]. The heterogeneity that characterizes P. gingivalis lipid A moieties is environmentally regulated and the resulting changes in Pg-LPS structure may determine the type of host response [bib_ref] Analysis of neutrophil-derived antimicrobial peptides in gingival crevicular fluid suggests importance of..., Puklo [/bib_ref]. Differences in biological activity of lipid A from various organisms have been recognized [bib_ref] Expression of a Porphyromonas gingivalis lipid A palmitylacyltransferase in Escherichia coli yields..., Bainbridge [/bib_ref]. However, the key structural differences between E. coli lipid A and P. gingivalis lipid A that account for the differences in activity have not been defined [bib_ref] Expression of a Porphyromonas gingivalis lipid A palmitylacyltransferase in Escherichia coli yields..., Bainbridge [/bib_ref]. In the present study, we observed some differences in neutrophil stimulation between different bacterial sources of LPS, i.e., Aa-LPS vs Pg/Ec-LPS. However, the biological basis of these differences remains to be elucidated, since there is no information regarding specific structural traits of Aa-LPS compared to Pg-LPS and Ec-LPS. The similar effects of Pg-LPS and Ec-LPS further suggest that the phenotypic changes induced in neutrophils by these two types of LPS might involve a common signal pathway of neutrophil activation. Besides differences in LPS structures, neutrophil responses to these components might be a result of a complex variation in the array of TLR/ ligand complexes.
In addition to the antimicrobial peptides, neutrophils also apply oxygen-dependent microbicidal mechanisms, with the NO production during pathogen phagocytosis being one of the most studied [bib_ref] Nitric oxide and the immune response, Bogdan [/bib_ref]. Periodontopathogenic bacteria such as A. actinomycetemcomitans, P. gingivalis and F. nucleatum are able to induce iNOS enzyme activation, resulting in high NO production in gingival tissues and blood of healthy subjects [bib_ref] Oxidative and nonoxidative killing of Actinobacillus actinomycetemcomitans by human neutrophils, Miyasaki [/bib_ref] [bib_ref] Deficiency of iNOS contributes to Porphyromonas gingivalisinduced tissue damage, Alayan [/bib_ref]. Corroborating these data, we showed that neutrophils from healthy subjects were able to produce elevated quantities of this toxic radical after Pg-LPS and Ec-LPS stimulation. In contrast, low NO levels were produced by neutrophils from chronic periodontitis patients cultured in the presence of Pg-LPS and Ec-LPS or in the absence of LPS stimuli. These results contrast with previous studies in that NO production was enhanced in peripheral neutrophils from chronic periodontitis subjects after exposure of these cells with opsonized bacteria by the components from the patient's own serum [bib_ref] Neutrophils in chronic and aggressive periodontitis in interaction with Porphyromonas gingivalis and..., Guentsch [/bib_ref]. Also, animals with experimental periodontitis expressed more iNOS in periodontal tissue than controls [bib_ref] The essential role of IFN-gamma in the control of lethal Aggregatibacter actinomycetemcomitans..., Garlet [/bib_ref]. Although the criteria for selection of subjects with or without periodontitis were similar to those adopted in other studies [bib_ref] Neutrophils in chronic and aggressive periodontitis in interaction with Porphyromonas gingivalis and..., Guentsch [/bib_ref] , difference in methods for NO quantification and in conditions/time used for neutrophil exposure to antigens could influence the results.
The precise role of NO during the development of inflammatory processes such as periodontitis remains unknown. We worked with the hypothesis that variation in NO production by neutrophils might influence individual susceptibility to periodontal infections, i.e., subjects with neutrophils with low NO synthesis might have increased susceptibility to periodontal disease. This hypothesis is strengthened by studies showing that iNOS inhibition or deficiency allows uncontrolled growth of diverse periodontopathogenic bacteria and periodontal disease development [bib_ref] The essential role of IFN-gamma in the control of lethal Aggregatibacter actinomycetemcomitans..., Garlet [/bib_ref] [bib_ref] Oxidative and nonoxidative killing of Actinobacillus actinomycetemcomitans by human neutrophils, Miyasaki [/bib_ref] [bib_ref] Deficiency of iNOS contributes to Porphyromonas gingivalisinduced tissue damage, Alayan [/bib_ref]. However, to determine the clinical/biological relevance of the data presented here, further studies are being conducted by our research group.
This study indicates that peripheral neutrophils from subjects with chronic periodontitis differ from neutrophils from periodontally healthy subjects regarding their responses to different LPS types, which affects the expression and production of antimicrobial peptides and NO production. Low NO production by neutrophils of periodontitis patients might ultimately represent an increased susceptibility to periodontal pathogen infections.
[fig] Figure 1: HNP 1-3 and LL-37 mRNA expression in peripheral neutrophils from periodontitis and healthy subjects. Neutrophils were isolated from peripheral blood and analyzed by flow cytometry, showing more than 94% purity according to side scatter (SSC) vs forward scatter (FSC) parameters (A). Neutrophils were incubated for 6 and 12 h with culture medium only or with medium containing Aa-LPS (100 ng/ mL), Pg-LPS (100 ng/mL) or Ec-LPS (100 ng/mL). The relative mRNA expression of HNP 1-3 (B) and LL-37 (C) in neutrophils was assessed by qRT-PCR. Data are reported as means ± SD for N = 6 in each group. Aa-LPS = Aggregatibacter actinomycetemcomitans-lipopolysaccharide; Pg-LPS = Porphyromonas gingivalis-LPS; Ec-LPS = Escherichia coli-LPS. *P ≤ 0.05, periodontitis compared to healthy subjects (Student t-test). [/fig]
[fig] Figure 2: HNP 1-3 and LL-37 production by peripheral neutrophils. Neutrophils were incubated for 6 and 12 h with culture medium only or with medium containing Aa-LPS (100 ng/mL), Pg-LPS (100 ng/mL) or Ec-LPS (100 ng/mL), and production of HNP 1-3 (A) and LL-37 (B) was determined by ELISA. Data are reported as means ± SD for N = 6 in each group. *P ≤ 0.05, periodontitis compared to healthy subjects (Student t-test). For abbreviations, see legend to Figure 1. www.bjournal.com.br Braz J Med Biol Res 45(11) 2012 [/fig]
[fig] Figure 3: NO production by neutrophils stimulated with LPS. Neutrophils isolated from peripheral blood of periodontitis and healthy subjects were cultured for 6 and 12 h in the presence and absence of Aa-LPS (100 ng/mL), Pg-LPS (100 ng/mL) and Ec-LPS (100 ng/mL). Nitric oxide (NO) concentrations were measured by the Griess reaction in the cell culture supernatant. Data are reported as means ± SD for N = 6 in each group. *P ≤ 0.05, periodontitis compared to healthy subjects (Student t-test). For abbreviations, see legend toFigure 1. www.bjournal.com.br Braz J Med Biol Res 45(11) 2012 [/fig]
[table] Table 1: Primer sequences for the genes, amplification profile, and amplicon length studied here. [/table]
|
Cleavage of Oligonucleotides Containing a P3’→N5’ Phosphoramidate Linkage Mediated by Single-Stranded Oligonucleotide Templates
[fig] Figure S2, Figure S3: Cleavage profiles of ON-1-ON-8 in the presence of PSD (A, B), ASD (C, D) and ASR (E, F) at pH 4.0, 40°C. Hydrolysis experiments were performed in a solution buffered to pH 4.0 containing 140 mM KCl, 10 mM MgCl 2 , 1.0 mM sodium phosphate, 10 mM sodium citrate-HCl, 3.35 µM each strand, at 40°C.with PSD with PSD with ASD with ASD with ASR with ASR Cleavage profiles of ON-1-ON-8 in the presence of PSD (A, B) and ASR (E, F) at pH 4.0, 20°C. [/fig]
[fig] Figure S5: 0×50 mm) [buffer A, 0.1 M TEAA (pH 7.0); buffer B, 0.1 M TEAA (pH 7.0)/MeCN = 1:1; linear gradient, B 14 to 30%/7 min; flow rate, 1.0 mL/min; detection, 260 nm]. Molecular models of the duplex ON-1•ASR (A) and the triplex ON-1•PDD (B). [/fig]
|
Resource management: ways to sustain the environmental gains of COVID-19 lockdown
Natural resources are under constant exploitation due to industrialization and urbanization. Ecological disturbance caused by over exploitation of resources is one of the possible reasons for the outbreak of COVID-19 pandemic. Due to the highly infectious nature of this disease, countries across the world have taken self-imposed isolation measures such as lockdown, quarantine, curfew, etc., to limit human-to-human spread. Though this pandemic has shaken the world and left millions suffering, it has also caused surprising positive effects to environment. Due to reduced human pressure on ecosystems during the lockdown, betterment of air, water quality and biodiversity along with reduced consumption of natural resources have been reported. It is necessary to maintain this improvement in order to avoid the environmental benefits slipping away once the world limbs back to normalcy. The benefits acquired in terms of resource conservation prompt us to avoid unnecessary human interference and adopt sustainable life styles. Wide usage of information and communication technologies (viz. work from home, teleconferencing, e-learning and e-commerce) during the pandemic revealed their potential in meeting the needs of human livelihood and played a significant role in improvement in air quality and reduced resource consumption. Implementing them should be a policy measure during an environmental crisis. Active government involvement is necessary for coordinating institutional and policy aspects of resource conservation. Smooth transitioning to more sustainable post-COVID world thus requires coordinated action at individual, local, national and international levels. Restoring environmental resources is essential to prevent future pandemics.
Natural disasters and disease outbreaks frequently occur due to over exploitation of resources for industrialization, urbanization, etc. The recent pandemic caused by COVID-19 has shaken the earth leaving millions infected with deaths exceeding 760,000 within a few months of its outbreak. It also has the potential to infect million more. It has taken a toll on the global economy and livelihood. Nations all over the world are struggling to protect public health, economies, trade, etc., from this perilous pandemic.
Respiratory droplets of size greater than 5-10 μm in diameter from the infected person are the main cause of corona virus infection . It is reported to be transmitted by human-to-human contact and also by indirect contact with infected persons [bib_ref] Early transmission dynamics in Wuhan, China, of Novel Coronavirus-infected pneumonia, Li [/bib_ref] [bib_ref] Community transmission of severe acute respiratory syndrome coronavirus 2, Liu [/bib_ref]. Social distancing and sanitation are the essential elements that control the infection rate of corona virus. So as to control the transmissivity of this infection, governments across the world have implemented containment measures such as total shut down, quarantine and curfews to ensure social distancing. By the end of April 2020, almost 90% of world's population was under different forms of confinement.
Though COVID-19 pandemic has left millions suffering, it has caused surprising positive effects too. It seemed impossible to save our ailing oceans, suffocated atmosphere and scared animal resources from the shackles of human exploitation. Now there is a ray of hope to conserve our natural resources due to the pause caused by the pandemic. Environment is breathing a sigh of relief as industries and transportation sectors came to a complete halt due to the lock down and self-quarantine measures imposed by governments. Nature has gained in terms of reported improvement in quality of its various components. It is necessary to maintain this improvement in order to avoid the environmental benefits slipping away once the lock down is lifted.
In this article, we have given a comprehensive evaluation of the environmental gains acquired in conserving natural resources such as air, water and biodiversity during shutdown. Among the four major sub-systems of earth (air), hydrosphere (water), biosphere (living things) and lithosphere (land)), the positive effect on environment during COVID-19 imposed lockdown was more evident on air, water and biocomponents (as endorsed by a majority of the publications), suggesting that these key environmental indicators have been exploited to a larger extent. Hence, these are chosen as an evaluation index. Scientific evidences on improvement in quality of air, water and biodiversity observed by various research groups have been compiled in this paper. Adopting eco-friendly measures, technological strategies, change in mindset, policy intervention, etc. are suggested as ways to sustain these vital resources. This review would help regulators, policy makers and the common man to reorient and march toward conserving the natural resources for achieving the sustainable development goals (SDGs) of the United Nations.
## Environmental gains of covid-19 lockdown
## Air and atmosphere
Atmosphere, which consists of layers of gases, is vital for existence of life forms. Atmospheric gases surrounding the earth give protection from dangerous UV rays of the sun. Oxygen, nitrogen and carbon dioxide the major components of earth's atmosphere are necessary for survival. Over the years, the air quality has been deteriorating due to over exploitation by mankind. Air pollution is one of the major causes of illness and death in 1 3 many countries. According to WHO data, 90% of world's population breathes polluted air. Transport and industrial sectors and construction activities are the major contributors of air pollution.
Due to the shutdown, improvement in air quality, reduction in noise pollution, reduction in travel-related accidents, reduction in demand for coal and oil and reduction in global carbon emissions were observed. Blue skies with breathable air were reported in many nations which otherwise face severe air pollution problems. Short-term improvement in air quality with reduced emission of NO 2 was reported in China due to the decreased economic activities during the pandemic. Reduction in levels of nitrogen dioxide and carbon monoxide to an extent of 0.00002 mol m −2 and < 0.03 mol m −2 in the pandemic affected hotspots was reported during Feb-Mar' 2020 [bib_ref] The dark cloud with a silver lining: Assessing the impact of the..., Lal [/bib_ref]. Due to the imposed restrictions, 50% reduction in air pollution levels in New York and 40% reduction in consumption of coal by china's thermal power plants have been reported [bib_ref] Environmental perspective of COVID-19, Saadat [/bib_ref]. Reduced NO 2 emission levels were observed in Europe . Among the south Asian countries, Delhi, the national capital of India, which is the most polluted city in the world (WHO 2016) with hazardous air quality has been reported to have 40 to 50% reduction in criteria air pollutants such as NO 2 , CO, PM 10 , PM 2.5 , SO 2 , O 3 along with ammonia. Ambient air quality analysis over a 04-year duration (2017 to 2020) revealed lowest levels of NO 2 (20 µg/m 3 ) and PM 2.5 (30 µg/m 3 ) in air during 2020 as against the maximum levels of around 55 and 150 µg/m 3 observed, respectively, during 2017. During the lockdown, improvement in air quality has been reported across the world [fig_ref] Table 1: Improvement in air quality observed across the world during COVID-19 lock down... [/fig_ref]. NASA satellite images showed 30% reduction in NO 2 levels in the pandemic-infected .
Fossil fuels are major source of energy and an important component of global economy. Worldwide, the dependency on fossil fuel derived energy varied from 32.1 to 100% (BP Year-wise comparisons of NO 2 and PM 2.5 levels in Indian cities (Average)Statistical . Fossil fuels also have the dubious distinction of contributing to air pollution. Carbon dioxide, carbon monoxide, nitrogen oxides (NOx), sulfur oxides (SOx) and particulate matter are the primary pollutants emitted during fossil fuel burning. It also contributes to negative environmental impacts such as global warming, ozone layer depletion, acid rain, etc. Greenhouse gas emission from fossil fuel combustion has proven physical and psychological health impacts [bib_ref] Multiple threats to child health from Fossil fuel combustion: Impacts of air..., Perera [/bib_ref]. Though the trends in consumption of fossil fuels have been ever increasing pre-COVID-19, decline in fossil fuel and oil consumption has been observed across the globe during the lockdown. As the demand for fossil fuel reduced, huge price cut down has been reported globally. Lowering trends in carbon emission was observed due to decline in world consumption of oil and oil-derived products.
## Water
Water, one of the fundamental resources for sustaining life, has been subjected to continuous deterioration by human activities. As the borders of nations were closed and tourist's numbers have got reduced drastically, and industries have come to a complete halt, water bodies have been reported to have lesser pollution load. Water canals of Venice in Italy had noticeable decrease in pollution levels with increased visibility of fishes [bib_ref] Environmental perspective of COVID-19, Saadat [/bib_ref]. Satellite remote sensing studies reported improved water quality in Vembanad lake, one of the longest fresh water lakes in India. The images gave evidence of reduction in suspended particulate matter in water to an extent of 34.6%. Time series analysis of data (2013 to 2020) revealed lowest SPM levels in 50% of the zones during the month of April' 2020 [bib_ref] COVID-19 and surface water quality: Improved lake water quality during the lockdown, Yunus [/bib_ref]. Lockdown imposed during COVID-19 could provide a unique opportunity to environmental scientist for collecting data on point source of pollution in rivers. As industrial operations have come to a complete halt, systematic testing and data collection on water quality could serve as baseline through which point source of pollution could be detected [bib_ref] Research opportunities in pandemic lockdown, Saraswat [/bib_ref]. Enhanced phytoplankton growth was reported in Hooghly estuary of the Ganges River in India due to decline in pollution load during lockdown [bib_ref] Eco-restoration of River Ganga water quality during COVID19 lockdown period using Total..., Mukherjee [/bib_ref]. [fig_ref] Table 2: Improvement in water quality observed across the world during COVID-19 lock down... [/fig_ref] summarizes few reported studies on improvement in water quality observed across the world.
## Biodiversity
Biodiversity and animal habitat have been diminishing drastically over the years due to the disturbance caused by anthropogenic activities. Ever-increasing global population, habitat conversion, deforestation and climate change due to pollution and invasion by exotic organisms and diseases are the major reasons, which put the biodiversity under constant threat. Tropical rainforest and marine ecosystems hold more than 90% of biodiversity. Noise from air, water and land transport causes tress to flora and fauna. Oil pollution in oceans and human coastal activities endanger the survivability of marine habitats.
Though there seems to be no hope for reviving the biodiversity, terrestrial and aquatic ecosystems have found a silver lining during the lockdown. Marine, land and wildlife species pushed their boundaries and started returning to their natural habitats due to the nil and minimum disturbance. As the transport and industrial activities near the beaches and on the sea have been halted, reduced levels of noise and reduction in ocean pollution loads have been reported by news agencies. Sighting of coyotes and foxes in the US cities, South Asian countries (Pakistan and India) coastal water quality, phytoplankton densities and algal blooms investigated using changes in chlorophyll (Chl-a) and turbidity 50% reduction in Coastal Chl-a and 29% reduction in turbidity was observed [bib_ref] Understanding temporary reduction in atmospheric pollution and its impacts on coastal aquatic..., Shafeeque [/bib_ref] wild boars in Spain, deers in Japan's subway have been reported. Due to reduced human mobility, wild life and sensitive animal species were seen freely moving in rural and urban areas [bib_ref] Impacts of the coronavirus pandemic on biodiversity conservation, Corlett [/bib_ref]. Over 1.5 lakh migratory flamingo birds visited Mumbai wetlands during April 2020 breaking their earlier records. According to news reports, the endangered hawksbill sea turtle and Olive ridley sea turtles were seen hatching on deserted beaches of Brazilian and Indian coasts. Uninhibited animal behavior patterns have been reported from many places [fig_ref] Table 3: Improvements in biodiversity observed across the world during COVID-19 lock down Sl [/fig_ref].
## Lessons learnt and ways to sustain the environmental benefits of covid-19 lockdown
Findings on improved environmental quality during corona pandemic lockdown have made us think about how to sustain these gains from slipping away. Redefining the environment in post-COVID requires unified action at individual, local, national and global levels. Technological interventions, individual life style changes and institutional and policy approaches are key drivers in sustaining the environmental benefits acquired during the lockdown.
## Technological interventions
Harnessing technological innovations for sustainable development leads to effective resource management with environmental and economic benefits. Societal changes were felt necessary to curtail the spread of virus, and hence, the world witnessed large-scale transition to information and communication technological (ICT) options such as work from home (WFH), E-learning, video conferencing and E-commerce for telecommuting. These technologies have never been used on a global scale before COVID pandemic. Apart from preventing the spread of virus, application of these technological interventions can greatly contribute to resource conservation [fig_ref] Figure 2: Technological interventions for environmental resource conservation [/fig_ref]. Improved air quality and reduced fuel consumption and traffic congestion are the major benefits of telecommuting [bib_ref] The role of organizational support in teleworker wellbeing: A socio-technical systems approach, Bentley [/bib_ref]. [bib_ref] Broadband and telecommuting: Helping the US environment and the economy, Fuhr [/bib_ref] reported that telecommuting could reduce the greenhouse gas emissions in the USA to an extent of 588.2 tones by 2021 due to reduced patterns of energy consumption, construction and travel. Effect of WFH concept on urban traffic and air quality was studied in Switzerland. Analysis of data over duration of 10 years (2002 to 2013) revealed reduction in traffic congestion and air pollution to an extent of 2.7 and 4.1 percent, respectively [bib_ref] The relationship between teleworking, traffic and air pollution, Giovanis [/bib_ref]. The advantages of digitization were more evident in the urban setting. Marginalized and disadvantaged lot of the global population was unable to make use of these e-platforms. Accessibility to mobile phones, electricity and internet connectivity have hampered the efficient utilization of digital technology in rural areas. Socio-economic status, resource availability, territorial difference and literacy level are some key factors, which amplified the existing inequalities during the pandemic. Disadvantages of vulnerable population to digital inequality lead to increase in unemployment, devoid of technology-oriented health care and quality education and poor accessibility to products and services. An intersectional approach is necessary to bridge the digital divide. Evidence of increased usage of these scientific innovations for networking during the lockdown is presented in this section.
## Work from home
To stop the spread of virus and to ensure social distancing, companies/ organizations get their job accomplished by asking their employees to work from home. WFH has both environmental and economic benefits. It is efficient in making people cope-up with their workrelated responsibilities while staying at home . Multinational companies such as Google, Amazon, Microsoft, etc., have rolled out WFH policy to curtail the spread of virus. As per government's order, during the lockdown, Indian information technology (IT) industries started implementing WFH model, and 90% of employees successfully worked from home out of which 65% worked from homes in metros and the remaining from homes in small towns strictly adhering to quality output (The economic times, 2020). This transition on WFH should be encouraged post-COVID-19, which could contribute to reduction in travel-related emissions and also is economic as there will be a considerable reduction in travel and infrastructure-related expenditure.
## E-education and video conferencing
Across the globe educational institutions were forced to close down to prevent the spread of COVID-19. During this time of crisis, virtual learning platforms came to the rescue and ensured continuation of learning process. Academic institutions have started switching over to on-line platforms such as 'Zoom'/'Google Meet'/'Teams'/'Webex' for inter-active sessions, 'Voxvote' for quizzes, 'Testmoz' for tests, 'Canvas' and Google Classroom for academically engaging students during the lockdown (Zayapragassarazan, 2020). Apart from virtual meeting platforms, there was a marked surge in usage of other Internet platforms such as WhatsApp and Skype for enabling smooth learning during the lockdown. A customized e-learning framework implemented at Polito University in Italy during February to April 2020 revealed enhancement in learning patterns and online collaborations through internet. An increase in ingoing (10 times) and outgoing (2.5 times) Internet traffic was observed with a marked surge in remote classes being attended even during weekends. COVID outbreak lockdown has made institutions and companies to organize conferences and meetings on virtual mode. Many planned conferences have got cancelled or held through video conferencing, which is a catalyst for reducing the carbon footprint. It will become a new normal for business and scientific communities as it covers wider audience and discussion prospects are better through video conferencing. Resources consumption in terms of making logical arrangements for meetings and providing infrastructure can be minimized by virtual conferencing, as it may lead to less waste generation. McConnell, 2013 compared the efficiency of face-to-face meetings with collaborative professional learning through video conferencing. Findings reveled that though teachers prefer face-to-face interactions, video conferencing could be an effective tool when time and distance factors become barriers.
Video call services also helped to maintain emotional and mental health of societies, as it enabled families, friends and relatives to work closely in a virtual mode during the lockdown. Network and mobile data traffic have increased manifolds across nations. A 70% increase in usage of video calls, voice calls and messaging app was reported in Italy during the lockdown. Spanish telecommunications company Telefonica has reported increase in mobile data traffic up to 50%. Video conferencing also played a major role in telemedicine, exchange of medical diagnostics and treatment related discussions during the viral outbreak. In Singapore hospitals, coordination and interaction of radiological experts between inter-and intradepartments to manage the COVID infection were promoted through video conferencing [bib_ref] Planning and coordination of the radiological response to the coronavirus disease 2019..., Tsou [/bib_ref].
## E-commerce
Executing business through digital means, referred as E-commerce, is reported to reduce carbon emissions to an extent of 209 mt of CO 2, and online procurement of standard electronic goods was found to reduce carbon footprint to an extent of 84% in comparison with purchase through conventional e-retailing [bib_ref] Measuring transport related CO 2 emissions induced by online and brick-and-mortar retailing, Carling [/bib_ref]. A study conducted to comparatively evaluate the CO 2 emission from conventional retail mode of business and e-commerce in Shenzhen, China, revealed that environmental cost implications were less in E-commerce and the total CO 2 emission difference between the two modes was estimated to be 124 million tonnes in 2016. Infrastructure and travel were the source of CO 2 emission in retail mode, and packaging was responsible for CO 2 emission from E-commerce. E-commerce could play a significant role during COVID-crisis. As visiting malls and big shopping complexes are discouraged as a precautionary measure, e-commerce platforms take a lead in meeting the needs of resource supply and demand. These online shopping strategies have marked environmental contributions in terms of reduced emissions and resource consumption. E-commerce was described as a sustainable and safe business option for vulnerable people and health workers during COVID emergency situation, and food orders were executed mostly through online in Italy and Spain during the pandemic [bib_ref] Competing during a pandemic? Retailers' ups and downs during the COVID-19 outbreak, Pantano [/bib_ref].
## Life style changes and society-responsible consumption and production
The ever-increasing global population, which is expected to reach 9.6 billion by 2050, would require 3 more earth planets to meet the demands of human needs. The lockdown served as a catalyst for coming up with plans and recovery policies for transforming a human society that adopts a need-based consumption pattern. A human society which instills reduce, recycle, reuse and recovery policies as its voluntary life style option could leave the environment to future generations in a better form. Sustainable resource production and consumption are one of the central themes of united nation's sustainable development goals, which discourage resource-intensive lifestyle patterns [bib_ref] Making sustainable consumption and production the core of sustainable development goals, Akenji [/bib_ref]. Understanding how to co-exist with other components of nature will enable human civilization to move in a matured and positive manner. Environmental impacts of European lifestyle aspects were assessed using an environmentally extended input-output simulation model framework wherein the carbon foot print of sustainable life style models was compared with the already existing scenarios. Land and water usage patterns, greenhouse gas emissions and toxicity potential were the environmental factors assessed. Adopting vegan lifestyles, durable clothing, reduced transport and sharing mentality in humans could contribute to reduction in carbon footprints to an extent of 14, 3, 26 and 18%, respectively [bib_ref] The environmental impact of green consumption and sufficiency lifestyles scenarios in Europe:..., Vita [/bib_ref]. Thus, human lifestyle patterns have significant environmental effects.
Corona lockdown has paved way for net-zero emission concept prompting humans to change their life style in a positive way. Forced home containment has enabled many humans to opt for healthy lifestyle changes such as improved physical activity, eating healthy home cooked food, spending quality family time, etc. that indirectly contributes to environmental conservation. Healthy life is an important component for a country's economy as poor health leads to spending more for maintaining health care. Disrupted supply chains and the resultant economic crisis created by the pandemic have prompted societies to make efficient use of local resources for maintaining industrial production and adopt need-based consumption patterns.
The pandemic has made us change our life styles in a positive way. Humans have learnt to adapt to optimum resource consumption, resource substitution, reduced waste generation, reduced travel and recycle as much as possible. Governments and businesses should take lead in advocating economic, health and environmental benefits obtained from adopting a low-carbon lifestyle. These behavioral and institutional changes if carried over post-COVID could contribute in minimizing the carbon footprint. The lockdown can accelerate transition toward need-based consumption with a sense of unity and purpose.
## Institutional and policy aspects
Natural resource management is intertwined with environmental protection. Institutional and policy aspects play a key role in management of natural resources. They are instrumental in conceptualization and implementation of environmentally conscious decisions, balancing traditional versus alternative approaches of resource consumption and mitigating natural hazards. Decisions on resource management could be at local, national or international level. Temporary benefits obtained on conservation of natural resources due to COVID-19 lockdown can be leveraged to long-term gains by coordinated actions at individual, local, national and international levels.
## Promoting affordable and clean energy
Significant reduction in climate change impacts could be accomplished by switching over to renewable energy options. Renewable energy resources have environmental benefits in terms of reduced emissions of particulate matter and greenhouse gases. Policy proposals should encourage production of energy efficient fuels from waste biomass, provide incentives for low-carbon transport strategies and promote eco-friendly electric power, bioethanol, compressed natural gas and other renewable energy options, as global trends in consumption of renewable energy are low presently (IRENA, 2019). Hydropower occupies the major share (44.7%) of renewable energy consumption followed by equal shares contributed by solar photovoltaic (22.9%) and onshore wind energy (23.4).
Other renewable options such as biogas, geothermal energy, liquid biofuels, municipal waste, solar thermal energy, etc. contribute the remaining 10% share. Policy interventions are essential for enhancing the pace of renewable energy transition. Polices should adopt a system-based approach and evaluate the renewable energy options based on their cost effectiveness and environmental benefits. Integrated planning that defines roles and responsibilities for every stakeholder in the energy sector is necessary for effective transition to renewable energy options. Regulators should encourage innovative approaches, which phase out fossil fuel-based energy systems. Hydrogen Breakthrough Steel Making (HYBRIT) project undertaken by Sweden is an example of government supported project that aims to replace fossil fuels by hydrogen-based energy for steel production [bib_ref] Potential transitions in the iron and steel industry in Sweden: Towards a..., Karakaya [/bib_ref].
Reduction in greenhouse gas emissions (up to 80% by 2050) is forecasted as one of the necessary steps to be taken for achieving the sustainable development goal on climate action [bib_ref] Policy: Sustainable development goals for people and planet, Griggs [/bib_ref]. Transitioning to clean energy alternatives and phasing out fossil fuels are vital for obtaining climate and air quality gains [bib_ref] Climate and air-quality benefits of a realistic phase-out of fossil fuels, Shindell [/bib_ref]. Electric vehicles (EVs) are emerging as one of the most sustainable transport modes. Potential of E-bike in reducing the greenhouse gas emission was assessed under varying temperature, distance and rainfall conditions in Switzerland, and the energy demand was modeled accordingly. Findings showed that savings in energy and GHG emission to an extent of 17.5 and 10% could be achieved by e-bike deployment (Bucher et al., Global trends in renewable energy consumption by International renewable energy agency, 2019 2019), though disposal of the batteries is a concern. Effective deployment of EVs by smart grid control and opting for electricity generation by renewable sources are the major factors, which need to be considered for optimal utilization of electric powered vehicles.
Discouraging the use of motorcar and transition to walking or cycling mode for short distance travel causes improved health and environmental gains. Bicycles are referred as the most economic and sustainable zero carbon emission transport mode [bib_ref] Contemplating cycling to work: Attitudes and perceptions in different stages of change, Gatersleben [/bib_ref] which has many economic and health benefits. In many developing countries, the roads are hostile for cyclists. Dedicating separate lanes for cyclists will ensure safety and encourage more utilization of this most ecofriendly transport mode.
Waste to energy (WtE) is yet another sustainable option that conserves fossil fuels, generates electricity and also disposes off waste. Agricultural waste, municipal waste, liquid and solid waste from industries and energy crops are the major sources of biomass-derived energy. Waste to energy routes (WtER) have the potential for distributed energy supply [bib_ref] Waste-to-energy: A way from renewable energy sources to sustainable development, Kothari [/bib_ref]. At present, incineration is the most adopted technology for generating energy from waste. Switching over to technologically advanced anaerobic digestion/ composting process for energy generation from waste could prove to be an eco-friendly option. Efficient energy recovery is a challenge when opting for WtE, and the economic viability of energy recovery from MSW is highly dependent on the availability and characteristics of the waste feedstock [bib_ref] Importance of waste composition for Life Cycle Assessment of waste management solutions, Bisinella [/bib_ref]. Policies should strengthen research and development for improved energy recovery and provide subsidies for their industrial level production.
## Clean water and sanitation
Access to good quality water, sanitation and hygiene are fundamental necessities in the fight against any disease outbreak. Urban slums have become the potential hotspots of rapid spread of COVID-19 infection. Water hygiene in urban slums can be regulated by providing improved water supply infrastructure and funding support. Sustenance of improved quality of water resources can be accomplished by establishing more real-time water quality monitoring network [bib_ref] Toward urban sustainability and clean potable water: Prediction of water quality via..., Dawood [/bib_ref] , improved technology and infrastructure for water treatment and pollution abatement, adopting ecologically focused low energy-intensive treatment, engineered natural systems for water pollutant removal and ecosystem-based approaches (EbA) such as eco-bioremediation that utilizes the natural degradation abilities of both plants and microorganisms with civil engineering and biological interventions for optimal performance. Adopting internationally agreed water quality standards and guidelines will ensure safe water for all [bib_ref] Improved water resource management framework for water sustainability and security, Ahmed [/bib_ref].
## Biodiversity conservation (life below water and life on land)
Biodiversity conservation could be managed by imposing /strengthening policies that regulates infringement of protected animal habitats, exploitation of animal species for human consumption, invasion of exotic organisms and diseases, resource allocation, and restoration of species habitat. Environmental impact assessment a tool that integrates environmental and social aspects in decision-making should have biodiversity conservation as one of its essential components. Establishing ways for co-existence and cohabitation with natural species is the most sustainable approach for biodiversity conservation. Marine biodiversity conservation is essential for protecting the health of ecosystem. Policies should regulate over exploitation of marine resources and marine pollution. The COVID-19 lock down gave an opportunity for the revival of marine resources. Governments should come up with a road map for retaining these temporary benefits and focus toward rebuilding a sustainable ocean economy.
As per the Global Assessment Report on Biodiversity and Ecosystem Service (2019), around a million living species face the threat of extinction. Conservation of forest resources will reduce the climate change impacts and ensure a healthy ecosystem. United Nations Environment Programme (UNEP) in 2016 warned an increase in incidence of zoonotic diseases, which is possibly due to the over exploitation of natural habitats. Government policies should come up with an action plan to restore land resources in a scientific manner.
## Climate action-sustaining the improved air quality
Implementing short-term shut down measures after taking in to consideration of the aspects of economy could be an effective solution for controlling emergency environmental pollution scenarios. As the air quality levels became hazardous in Delhi, the Indian capital city, community health emergency was declared in the year 2017 [bib_ref] Indian annual ambient air quality standard is achievable by completely mitigating emissions..., Chowdhury [/bib_ref]. Temporary shutdown thus could be a viable step during emergency situations. Governments should implement realistic decarbonization strategies to reduce global GHG emissions. Observing no vehicle day, encouraging people to use cycle, phasing out polluting vehicles and old power plants, strict enforcement of polluter pays principle, capacity building, strengthening research and development are other policy interventions which if implemented can bring immediate benefits for the climate as well as the health and livelihoods of millions [fig_ref] Figure 4: Institutional and policy aspects for resource conservation in post-COVID world [/fig_ref].
Governments should mandate adoption of austerity measures in the tourism system by way of cutting down on unplanned travel, extravagant logistical expenses and encouraging virtual conferences. The learning from the pandemic should lead the global tourism industry to reorient toward achieving SDGs which can play a major role in reducing GHG emissions and climate change impacts [bib_ref] Pandemics, tourism and global change: A rapid assessment of COVID-19, Gössling [/bib_ref].
## Sustainable cities and communities and sustainable economic growth
Small actions can lead to big changes. Individuals have a major role to play in resource conservation. Encouraging optimum utilization of resources and avoiding wasteful consumption should be the way of life. During lockdown, humans have learnt to learn with the available resources. These habits need to become new normal in the post-COVID world.
Technological innovations widely used during lockdown can be translated to long-term approaches. E-education can be made as an integral component of academic curriculum. Telecommuting using ICT is an effective resource management strategy. This strategy should be incorporated as a viable solution in environmental policies, which can be implemented as a response to emergency situations wherein environmental pollution has caused public health crisis. These web-based technological strategies technologies have advantages, but also the dubious distinction of having negative impacts such as poor social interaction and reduced physical activity, etc. [fig_ref] Table 4: Sociocultural and environmental advantages and disadvantages of web-based technologies Sl [/fig_ref]. Therefore, utilizing them in an optimized manner would make a balance of its positive and negative impacts.
Policy differences before and after the pandemic have both positive and negative environmental impacts. During the pandemic policy measures, viz. lock downs, travel restrictions, curfews, quarantine measures, closures of academic institutions and social distancing have been enforced to minimize the viral spread. Though effective in confining the spread, these policy interventions have put financial burden on every sector of the economy. Governments across the world have come up with recovery policies and packages to minimize the socio-economic impact. Measures related to work-from home, teaching by on-line platforms and procurement of goods and services through e-platforms have been rampantly implemented to reduce human-to-human contact. It is necessary to analyze the economic and institutional challenges and the magnitude of changes that these transitions would bring when implemented on a large scale. For smooth functioning of these e-platforms, it is necessary to ensure secured and robust network communication system, which is a major challenge in many developing countries. There is also a need for establishing low carbon enabling green ICT network strategies such as cloud computing, green metrics, etc. for sustainable use of these e-platforms [bib_ref] Energy efficiency and low carbon enabler green IT framework for data centers..., Uddin [/bib_ref] and international guidelines to improve the quality of services with respect to cross-border movement of goods and services through e-commerce. Utilizing them in an optimized manner would make a balance of its positive and negative impacts. With governments all over the world mandating the use of personal protective accessories such as masks gloves, etc. for preventing the spread of the disease, roll back/lenience of policies related to single-use plastics has been evidenced during the pandemic, posing waste management challenges. Green transport, control of invasive species, forest conservation and promoting nature-based solutions for waste management should be the critical components of policy measures.
# Discussions and conclusion
Human well-being is always interlinked with a healthy and resilient ecosystem as over exploitation of natural resources leads to ecological destruction, habitat disturbance and spill-over of disease from wildlife to humans [bib_ref] Impacts of the coronavirus pandemic on biodiversity conservation, Corlett [/bib_ref]. Corona pandemic has left millions dying and suffering. Saving human lives and controlling the pandemic should be the foremost aim of our society. Containment measures introduced during COVID outbreak have made us realize that reduced human pressure on ecosystems leads to revival of natural resources. Sustainable resource consumption and self-resilient behavior had a positive effect on the environment. From previous research, it is evident that the COVID pandemic served as a catalyst for positive changes in food, travel and working habits and enabled shift to less energy-intensive platforms and resource consumption patterns. Resilience and coping strategies during the pandemic were found to have a positive effect on the environment, whereas human centric beliefs had a negative effect. Encouraging eco-sensitive actions, enhancing public awareness about the linkage between public behavioral patterns and climate change can sustain the environmental gains in the post-pandemic era. Despite the positive environmental effects, COVID pandemic has also exposed the existing physical, economic and social inequalities in the world that limit the vulnerable population of the society to take advantage of advanced technologies. These inequalities have exacerbated the issues of unemployment, gender and digital divide, etc. Brave steps within and between countries are necessary to address these inequalities.
The insights obtained from the reported scientific evidences on the short-term environmental gains during the lock down prompt us to establish newer ways of conserving natural resources in the post-COVID world and work toward achieving the sustainable development goals. The following recommendations are proposed for sustaining the environmental gains:
- Capacity building and effective communication of environmental perils of climate change and pollution - Optimum utilization of environmental resources - Investing on renewable energy future - Encouraging behavioral change in day-to-day life - Personal and political commitment and international cooperation for alleviating inequalities - Prioritizing digitalization and developing techno-centric smart cities - Effective implementation and enforcement of industry, travel and environment-related regulations - The policy interventions to have a strong research base, with clear targets and timelines for effective implementation Environmental issues are directly linked with sustainable global development. The pandemic provided a wake-up call for scientists, government and policymakers to ensure judicious utilization of natural resources. Coordinated efforts at various levels with a long-term strategy are the need of the hour.
Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
# Declarations
## Conflicts of interest
The authors declare that they have no conflict of interest.
[fig] Figure 2: Technological interventions for environmental resource conservation [/fig]
[fig] Figure 4: Institutional and policy aspects for resource conservation in post-COVID world [/fig]
[table] Table 1: Improvement in air quality observed across the world during COVID-19 lock down Sl. No Name of the country and study details Air pollution parameters studied Reductions/ improvements achieved [/table]
[table] Table 2: Improvement in water quality observed across the world during COVID-19 lock down Sl. No Name of the country and study details [/table]
[table] Table 3: Improvements in biodiversity observed across the world during COVID-19 lock down Sl. No Name of the country and study details Biodiversity/ India Fish densities in the reef areas Significantly increased from 406 no. 250/m 2 [/table]
[table] Table 4: Sociocultural and environmental advantages and disadvantages of web-based technologies Sl. No Technology Saving and documenting of lectures is possible which provides repeated access to students b. Provides support study materials viz., FAQs, tutorials, online feedback, etc. c. Enables online collaboration d. Infrastructure and energy saving e. Reduced traffic f. Savings in expenditure g. Lesser use of paper, as materials are delivered electronically a. Growing anxiety around the pandemic leading to poor mental health and creating attention deficit in children b. Lack of self-discipline due to over-availability of resource c. Unfair burden for students in developing countries due to slow bandwidth and low internet connectivity d. Expensive tools for conduction of lectures/ meetings e. Isolation and lack of social contact and lack of co-curricular activities Work from home a. Flexible working hours b. Increase in productivity and job satisfaction c. Less distraction d. Employees spend less on traveling which can save their money e. Recruit larger pool of potential employees, which decreases the need for managing visas and work permits, and paying relocation expenses for overseas hires a. Remote working causes security issues which left users vulnerable to having their webcams and microphones hijacked b. Social and professional isolation overtime c. Disturbance in work-life and home-life balance d. Difficulty in monitoring performance and maintaining professional development and upgrading skills E-commerce features low barriers of entry in terms of initial capital requirements b. Online vendors can get a wide audience through presence and ads on search engines and social networks c. Saves travel and time d. Effective comparison of models, prices and discounts 6. Easy handling of return and reimbursement a. Need for more energy intensive computers b. Faster delivery options can lead to increased fuel consumption c. Waste generation (packing material) [/table]
|
Mapping the coverage, availability and uptake of External Quality Assessment programmes across One Health sectors in Asia
Introduction:Establishing effective external quality assessment (EQA) programmes is an important element in ensuring the quality of, and building capacity for, antimicrobial resistance (AMR) laboratory surveillance.Objectives: To understand the current coverage of, and challenges to participation in, EQAs in National Reference Laboratories (NRLs) across One Health (OH) sectors in Asia.Methods: Current EQA coverage was evaluated through desktop review, online surveys and interviews of both EQA participants and providers. EQA coverage was mapped and summarized by laboratory type and 'readiness' level and identified challenges evaluated qualitatively.Results: Of the 31 identified NRLs [16 Human Health (HH) and 15 Animal/Food Safety laboratories (A/FS)], 14 HH and 7 A/FS laboratories currently participated in international EQA schemes and several participated in two or more different schemes. Seven laboratories were currently not participating in any EQA scheme and two of these (one HH and one A/FS) do not currently perform microbiology; six HH NRLs provided national EQAs. Of the eight surveyed international EQA providers, three were based in Asia and all offered varying programmes in terms of pathogens, frequency and support mechanisms for reporting and follow-up. Only one provider currently served laboratories across all OH sectors.Conclusions:The current coverage of EQA programmes for AMR in Asia was heterogeneous across countries but especially across OH sectors. This updated overview of the coverage and challenges associated with participation in, and provision of, EQAs for AMR suggest the benefit and relevance of introducing one comprehensive and high-quality EQA programme across OH sectors in Asia.
# Introduction
Antimicrobial resistance (AMR) remains a global challenge and burden despite increased focus and efforts to control and prevent it in the past decade.It is estimated that by 2050, mortality attributed to AMR could reach 10 million annually, with the majority of these occurring in low-and middle-income countries (LMICs). This AMR disease burden will be associated with an estimated economic cost of up to 100 trillion USD if no action is taken to effectively prevent the continued emergence of AMR.These estimates are, however, based on modelling using incomplete data of questionable scientific accuracy, particularly from LMICs.LMICs exhibit a relatively larger burden of infectious diseases combined with the lack of AMR data of consistently good quality for surveillance, in part due to inadequate resources in laboratory diagnostics.Therefore, it is recommended to focus on improving laboratory surveillance capacity to ensure that robust data inform decision-making.An important part of building capacity for AMR laboratory surveillance is the establishment of effective external quality assessment (EQA) schemes.Effective EQA programmes are important to ensure correct identification of pathogens, antimicrobial susceptibility testing (AST)and to verify that laboratory procedures conform to acceptable international standards. 8,Since AMR is a cross-sectoral problem, the benefit of applying a One Health (OH) approach in the context of quality assurance for AMR activities is evident and has been increasingly gaining momentum.The Fleming Fund (FF) was established by the UK Government to support activities that improve AMR surveillance in LMICs in Asia and sub-Saharan Africa.Part of their early scoping to inform funding priorities was a comprehensive review of current surveillance networks, including specific analysis of existing initiatives relevant to quality management.Despite identifying a large number of initiatives within quality assurance, there was no clear or common framework in terms of programme content or governance, and very few identified coordinated efforts outside the human (public) health sector.
Furthermore, although previous global and regional EQA programmes have been targeting most of the WHO's Global Antimicrobial Resistance and Use Surveillance System (GLASS) pathogens,with some specific programmes also targeting specific foodborne pathogens, 14 they have had disproportionate participation from higher-income countries and highlighted significant limitations in both content, geographical coverage and larger deviations in results among laboratories in LMICs. As such, despite significant progress achieved in AMR surveillance efforts in Asia in recent years,gaps in available EQA schemes, as well as variation in capacity between countries and across OH sectors, still remain.Therefore, improving the quality of bacteriology diagnostics through implementing EQA provision across all OH sectors was identified as one of the specific priorities to be addressed within the FF portfolio of regional grants.Starting in January 2020, the 'Strengthening External Quality Assessment' in Asia (EQASIA) project was awarded one of these grants to implement EQA schemes in national reference laboratories (NRLs) in the Asian region.
To inform the appropriate approach to implementing a comprehensive regional EQA programme for AMR, EQASIA conducted an initial mapping exercise to understand the current EQA coverage across OH sectors in Asia. Here, EQASIA describes the findings from this exercise, including current uptake and provision of EQA programmes for AMR, as well as the identified challenges associated with implementation of these programmes.
# Methods
## Existing information sources
A desktop review was conducted to identify existing reference laboratories and current EQA providers in the 12 FF-designated priority countries in Asia: Bangladesh, Bhutan, India, Indonesia, Laos, Myanmar, Nepal, Pakistan, Papua New Guinea, Sri Lanka, Timor-Leste and Vietnam.However, active implementation of the project was only in 11 of the 12 FF-designated countries, which are the focus of this analysis. As an initial step, laboratories designated as AMR reference laboratories by the national AMR coordinating committees, as part of the National Action Plans (NAPs) in the respective countries, as well as other relevant regional/global reference centres and/ or EQA providers across the OH sectors, were identified where possible. The desktop review was complemented with data from other FF activities, including initial scoping visits conducted by Mott MacDonald between August 2018 and February 2020, input from communications with other ongoing FF country grants in Indonesia, Pakistan and Vietnam,and FF regional Round 1 grants.The WHO Global AMR team and the Food and Agriculture Organization Regional Office for Asia and the Pacific (FAO-RAP) also provided relevant information from their recent, respective survey activities on AMR surveillance capacities in the region. Finally, additional, and some newly appointed, reference laboratories were identified through communication and engagement with country stakeholders during three project symposia held in Quarter 3 (Q3) of 2020.
## Jac
## Survey tools and interviews
Two online surveys addressed to 'EQA participants' and 'EQA providers' were developed using survey software and distributed by sharing the URLs. All survey questions and answers were written/captured in English. The 'EQA participant survey' (Survey S1, available as Supplementary data at JAC Online), was sent to 18 Human Health (HH) and Animal/Food safety (A/FS) NRLs identified in eight countries. Supporting information was also collected through communication with FF country grantees. Using the 'EQA participant survey', the identified NRLs were queried on their general microbiology capacities, participation in EQA schemes and challenges associated with participation, as well as whether they already provided any EQAs nationally. For laboratories that were not yet participating in any EQA schemes, their perceived challenges for future participation were assessed.
The 'EQA provider survey' (Survey S2) was administered to the identified global/regional EQA providers. Additionally, if an NRL responded in the 'EQA participant survey' that they offered EQA within the country, they were asked to respond to the 'EQA provider survey'. Each EQA provider was expected to respond to the survey once. EQA providers were queried on the details and the quality/comprehensiveness of their programmes based on: (1) coverage of GLASS pathogens; (2) frequency of EQAs; (3) ISO/IEC 17043 accreditation; (4) use of IT modules for reporting (and feedback); (5) content and quality of follow-up exercises; and (6) funding structure of the EQAs, including funding sources and capacity for expansion of current programmes. Data from the remaining 3 of the 11 countries were obtained via different sources (as described above) or through direct communication at a later stage.illustrates the combined processes of consultations and data collection comprising the final, complete data sources.
# Data analysis
The survey information received from both the participant and provider laboratories were compared for initial cross-validation of the replies. Where details of EQA participation and/or provision were inconsistent or unclear, additional follow-up queries and/or interviews were attempted to provide more granular assessments.
A map to provide overviews and characterize the current EQA provision geographically and across sectors was created for an initial display of how well the designated priority recipients (NRLs) were covered through currently running programmes. Maps were created using R software (https://rstu dio.com/).In addition to mapping out the coverage of programmes, challenges to participation and general questions on microbiological diagnostic capacity and laboratory quality management procedures were also summarized. To identify the most commonly cited challenges by participants, qualitative data analysis within four prespecified domains (communication with the EQA provider, capacity, resources and logistics) was conducted using Microsoft Word 2016 version 16.0 and Displayr: Analysis and Reporting Software for Survey data (https://www.displayr.com/). 20
## Ethics
The collection of data in this project did not require human subject involvement. As such, the project was granted exemption from ethical review by the Institutional Review Board of the International Vaccine Institute (IVI). The purpose of collecting and storing information was explained, and consent to do so obtained from the respective respondents at the beginning of the survey.
# Results
## Eqa participation and provision
Of the 31 laboratories identified, 14 HH and 7 A/FS laboratories stated that they currently participated in international EQA schemes. Three of the HH laboratories and one A/FS laboratory mentioned that they participated in two different schemes and two HH laboratories listed participation in three different EQA schemes. Seven NRLs (one HH and six A/FS) were currently not participating in any EQA scheme and two of these (one HH and one A/FS) do not currently perform microbiology. Three laboratories remained unknown because they did not provide details of their EQA participation. Six of the HH NRLs but none of the A/FS NRLs provided national EQAs.summarizes the current EQA participation and national provision across all surveyed OH reference laboratories. Nine international EQA providers were identified through desktop review and consultations (eight of these replied to the survey) and six HH reference laboratories indicated that they provided national EQAs on the 'EQA participant survey' (four of these replied to the 'EQA provider survey'). None of the surveyed A/FS NRLs provided EQAs. Of the international providers who responded to the survey, three were based in Asia and only one global provider currently served laboratories across all OH sectors, whereas the remaining seven were sector-focused.
## Challenges of nrls participating in eqa programmes
When trying to summarize the challenges identified by laboratories of participating in EQAs, laboratories were grouped into three 'participation categories' in order to best characterize the challenges identified ('no EQA participation', 'EQA participation only' and 'EQA participation and provision'). Information on challenges was collected under four thematic areas: communication with EQA providers; capacity; resources; and logistics . For the laboratories that did not participate in EQA, in terms of resources, challenges included 'cost of participating in EQA' and 'lack of qualified staff'. Logistically, the laboratories mentioned 'difficulty with customs clearance' and 'lack of internet access'. In terms of capacity, 'lack of training for staff' was mentioned, whereas for communication with the provider, 'lack of knowledge' and 'access to EQA schemes' were mentioned. In the category of 'EQA participant only', the A/FS laboratories did not highlight any challenges. The HH laboratories cited 'limited knowledge' and 'limited access to EQA schemes' as challenges within the area of communication with the EQA provider. 'Lack of training', 'limited human resource' and 'high workload' were mentioned within the resource and capacity domains. In the 'EQA participation and provision' category, 'financial constraints' was a common resource challenge, with one laboratory being specific on the 'annual fee' and 'lack of access to current CLSI guidelines'. Laboratories experienced logistical challenges relating to 'difficulty in importing samples' and also mentioned 'difficulty to access the internet'. When it came to communication with the EQA provider, 'limited knowledge of available EQA schemes in the region and internationally' was mentioned. In terms of capacity, 'lack of support to implement corrective actions' was commonly mentioned.
## Quality of programmes and identified challenges for eqa providers
The general impression of current EQA schemes was rather heterogeneous and while most of the providers conducted EQAs across the spectrum of WHO GLASS priority pathogens, none currently provided comprehensive schemes covering all pathogens or Mogeni et al. External Quality Assessment across One Health sectors in Asia JAC the full range of antimicrobials. While almost all provided EQAs for pathogen identification, some only offered AST for selected isolates. Only five of eight international EQA providers were ISO/IEC 17043 accredited. Four of eight international and one of four national EQA providers indicated using information technology (IT) systems for reporting and analysis. One of the national providers only provided an EQA scheme for pathogen identification and not for AST. The frequency of the EQA distributions varied across providers, with 4 of 12 sending them three times per year. Some (3 of 12) also stated that they provided EQA distributions monthly, but whether this was a misconception of the nature of the programmes rather than being based on re-testing of isolates was not fully ascertained. Three of eight international providers and three of four national providers had a dedicated EQA budget. On the other hand, five of eight international EQA providers and one of four national EQA providers asked the EQA participants to pay an annual subscription fee. When highlighting challenges, five of eight international and one of four national EQA providers commented on 'lack of resources to support comprehensive follow-up activities for participating laboratories'. Three of the national EQA providers described the nature of their EQA programmes as 'facilitation of another international EQA programme'.
# Discussion
The current coverage of EQA programmes for AMR in Asia appears rather heterogeneous across countries, but especially across OH sectors, likely due to a wide range of current capacity levels and hence a wide range in readiness to participate in EQA programmes. Among current programmes, the coverage is variable in both content and frequency and there are redundancies due to participation in multiple programmes seen in several laboratories. The main shortcomings and challenges identified by current EQA providers are lack of IT solutions to support programme reports and evaluations, limited financial resources for sustaining participation, and follow-up exercises for underperforming laboratories. In 2017, a comprehensive review of existing programmes related to quality management in AMR surveillance identified 27 different initiatives in LMICs functioning at a regional or global level. Eleven of these were coordinated by a supranational body and the remaining ones by a mix of academic groups, commercial entities and governmental or non-governmental institutions. The content of the programmes also varied, with about half offering EQA programmes only and the other half offering different combinations of EQA programmes with training, assessments, standard/policy setting and accreditation or mainly functioning as a reference material provider.Such heterogeneity in other aspects of country efforts within AMR surveillance has also been evident in previous reviews from across the Asian region. Since these publications, multiple efforts have been initiated in the region, including establishment of NAPs to enable commitment from countries to support and share data with the WHO GLASS.In combination with the substantial efforts invested through the FF country grant portfolio, as well as other important initiatives, 23 the laboratory capacity across bacteriology reference laboratories in the region has also been strengthened. However, many of the generated data are still Mogeni et al.
not sufficiently quality assured and therefore still of limited value in surveillance efforts in LMICs.To appropriately guide the implementation of an EQA programme across the Asian region, the EQASIA project has provided an updated overview of the current coverage and challenges faced in participating in these programmes. This scoping activity has served to illustrate not only the coverage of existing EQA activities in 11 countries in Asia but also highlighted some of the challenges that laboratories face, either as EQA participants or providers. While most of the HH laboratories participated in EQA programmes, only half of the A/FS laboratories that provided feedback reported participation in an EQA programme. This shows a critical gap that needs to be addressed, particularly in this age of emerging and re-emerging zoonotic communicable diseasesand in the face of limited resources.Only one of the surveyed EQA providers already served both HH and A/FS laboratories in the region. Combining this EQA provider's expertise with the relevant and already-present EQA provider capacities working with WHO within the HH sector and with FAO within the A/FS sector in Asia represents a unique opportunity to bring together local expertise and presence while leveraging an existing global EQA platform to introduce a One Health EQA in line with both WHO 22 and FAO 27 priorities to strengthen existing regional programmes.
At the national level, several of the HH NRLs reported that they also provided EQA, while none of the A/FS laboratories did.
This demonstrates recent achievements and development within the HH sector but also highlights the need for similarly building the capacities of A/FS NRLs to become national EQA providers while ensuring that all national EQA providers are trained and supported in assuring quality in their further national implementation of EQAs in line with ISO/IEC 17043.It is important to ensure EQA providers are ISO/IEC 17043 certified to ensure the schemes they run are competent; its application applies to planning and design, management, personnel, equipment, quality assurance and confidentiality of the EQA.No global EQA scheme is currently comprehensively supporting the WHO GLASS. There are multiple programmes that offer EQAs worldwide but the associated costs are a barrier to participation for laboratories in LMICs.This obstacle was also highlighted by laboratories in the surveys. To maintain commitment to the EQAs, and compliant and equitable participation from all countries in the region, EQASIA deems it important to offer EQAs free of charge. Limited financial resources also affect the ability to access the needed support materials, as described previously by other EQA programmes. 8 This supports the idea that not only should participation in EQAs be offered free of charge, but programmes should also provide support for procurement of guidelines (e.g. CLSI documents) and financial support for associated training activities wherever feasible.
In addition to financial constraints, laboratories also face diverse challenges that can hinder participation in EQA, such as limited knowledge of providers, limited human resources and inadequate training.For any EQA programme to be truly effective, associated training activities and follow-up of underperforming laboratories is crucial. There is still limited use of IT for electronic microbiology data capture and collation from laboratories in the region,which also represents an opportunity that needs to be considered in the design and implementation of a regional EQA programme.
Finally, all of the abovementioned challenges jointly underline the importance of sufficient funding and political support to sustain comprehensive EQA programmes in Asia in the future. As such, the success and sustainability of a regional EQA programme will depend on the adoption and/or support by supranational organizations, such as WHO, FAO and the World Organisation for Animal Health (OIE), as well as commitment from global funding bodies.
# Conclusions
The current coverage of EQA programmes for AMR in Asia is rather heterogeneous across countries but especially across OH sectors, with a wide range of current capacity levels and readiness to participate in EQA schemes. Variable content and identified redundancies between programmes, as well as both technical and financial limitations, further challenges the successful implementation and optimal impact of EQA programmes in the region.
These findings suggest the benefit and relevance of introducing one comprehensive and high-quality EQA programme for AMR, free of charge to all NRLs across the OH sectors in the Asian region. Such a programme should include an IT solution for efficient reporting, a follow-up regimen for underperforming laboratories and also a comprehensive capacity-building programme to support participants' continuous development, allowing them to benefit from participation in the EQA programme. To ensure inclusion of all laboratories, regardless of current capacity levels, the training programme should be broad but flexible. Further studies to determine the optimal number of reference laboratories required by sector and by country is needed to sufficiently address EQA needs. A sustainability plan informed by realistic costing and forecasting analyses for the continued provision of the programme in the Asian region is also needed. Finally, the adoption and support of the programme by relevant supranational tripartite organizations, as well as committed funding bodies, are imperative for its future success and sustainability. |
Bax/Mcl-1 balance affects neutrophil survival in intermittent hypoxia and obstructive sleep apnea: effects of p38MAPK and ERK1/2 signaling
Background: Prolonged neutrophil survival is evident in various cardiovascular and respiratory morbidities, in hypoxic conditions in-vitro and in patients with obstructive sleep apnea (OSA) characterized by nightly intermittent hypoxia (IH). This may lead to persistent inflammation, tissue injury and dysfunction. We therefore investigated by a translational approach the potential contribution of the intrinsic stress-induced mitochondrial pathway in extending neutrophil survival under IH conditions. Thus, neutrophils of healthy individuals treated with IH in-vitro and neutrophils of OSA patients undergoing nightly IH episodes in-vivo were investigated. Specifically, the balance between pro-apoptotic Bax and anti-apoptotic Mcl-1 protein expression, and the potential involvement of p38MAPK and ERK1/2 signaling pathways in the control of Mcl-1 expression were investigated. Methods: Purified neutrophils were exposed to IH and compared to normoxia and to sustained hypoxia (SH) using a BioSpherix-OxyCycler C42 system. Bax and Mcl-1 levels, and p38MAPK and ERK1/2 phosphorylation were determined by western blotting. Also, Bax/Mcl-1 expression and Bax translocation to the mitochondria were assessed by confocal microscopy in pre-apoptotic neutrophils, before the appearance of apoptotic morphology. Co-localization of Bax and mitochondria was quantified by LSM 510 CarlZeiss MicroImaging using Manders Overlap Coefficient. A paired two-tailed t test, with Bonferroni correction for multiple comparisons, was used for statistical analysis.Results: Compared to normoxia, IH and SH up-regulated the anti-apoptotic Mcl-1 by about 2-fold, down-regulated the pro-apoptotic Bax by 41% and 27%, respectively, and inhibited Bax co-localization with mitochondria before visible morphological signs of apoptosis were noted. IH induced ERK1/2 and p38MAPKs phosphorylation, whereas SH induced only p38MAPK phosphorylation. Accordingly, both ERK and p38MAPK inhibitors attenuated the IH-induced Mcl-1 increase. In SH, only p38MAPK inhibition decreased Mcl-1 expression. Similar to neutrophils of healthy subjects exposed to IH (0.97± 0.2), in OSA neutrophils, Bax/Mcl-1 ratio was significantly lower compared to normoxic controls (1.0±0.5 vs.1.99±0.3, p=0.015), and Bax did not co-localize with mitochondria.Conclusions: These findings suggest that decreased Bax/Mcl-1 balance promotes neutrophil survival in IH in-vitro as well as in OSA patients. Moreover, Bax/Mcl-1 protein function in IH and SH might be regulated by different signal transduction pathways, highlighting a novel regulatory function through ERK1/2 signaling in IH.
# Background
Neutrophils are bone marrow derived short-lived cells which provide a unique model to study survival signaling. Once released into the circulation, neutrophils undergo constitutive apoptosis. However, their lifespan is prolonged in coronary syndromes such as unstable angina and acute myocardial infarction and in respiratory diseases such as chronic obstructive pulmonary disease (COPD) and neonatal and adult respiratory distress syndrome (RDS) [bib_ref] Delay of neutrophil apoptosis in acute coronary syndromes, Garlichs [/bib_ref] [bib_ref] Granulocyte apoptosis in the pathogenesis and resolution of lung disease, Bianchi [/bib_ref] [bib_ref] Dysregulated apoptosis and NFkappaB expression in COPD subjects, Brown [/bib_ref] [bib_ref] Acute lung injury:apoptosis and signaling mechanisms, Chopra [/bib_ref]. Prolonged neutrophil survival is also evident in patients with obstructive sleep apnea (OSA), characterized by repeated nightly episodes of intermittent hypoxia (IH) [bib_ref] Delayed neutrophil apoptosis in patients with sleep apnea, Dyugovskaya [/bib_ref]. Of note, increased neutrophil survival within tissues or in the circulation can promote persistent inflammation resulting in tissue injury and dysfunction.
Unlike in other cells, sustained hypoxia (SH) as well as IH were shown to profoundly inhibit neutrophil apoptosis in-vitro [bib_ref] Delayed neutrophil apoptosis in patients with sleep apnea, Dyugovskaya [/bib_ref] [bib_ref] Hypoxia prolongs neutrophil survival in vitro, Hannah [/bib_ref] [bib_ref] Hypoxia-induced neutrophil survival is mediated by HIF-1alpha-dependent NF-kappaB activity, Walmsley [/bib_ref] [bib_ref] Involvement of a ferroprotein sensor in hypoxia-mediated inhibition of neutrophil apoptosis, Mecklenburgh [/bib_ref] [bib_ref] Molecular pathways of spontaneous and TNF-{alpha}-mediated neutrophil apoptosis under intermittent hypoxia, Dyugovskaya [/bib_ref] and in-vivo [bib_ref] Delayed neutrophil apoptosis in patients with sleep apnea, Dyugovskaya [/bib_ref] [bib_ref] Acute hypoxemia in humans enhances the neutrophil inflammatory response, Tamura [/bib_ref]. Specifically in SH several signaling pathways and a number of family molecules that regulate apoptosis are activated. B-cell lymphocytic-leukaemia proto-oncogene (Bcl)-2 family members are one such family, which can be either proapoptotic (Bax and Bak, as well as BH3-only proteins, Bid, Bim, Bad, Bik, Noxa, PUMA) or anti-apoptotic (Mcl-1, Bcl-2, Bcl-X and A1) [bib_ref] Molecular control of neutrophil apoptosis, Akgul [/bib_ref] [bib_ref] Regulation of neutrophil apoptosis by Mcl-1, Edwards [/bib_ref] [bib_ref] Phosphorylation of Bax Ser184 by Akt regulates its activity and apoptosis in..., Gardai [/bib_ref] [bib_ref] Regulating neutrophil apoptosis: new players enter the game, Witko-Sarsat [/bib_ref]. The Bcl-2 family members are integrated in cell functions at the level of the mitochondria and participate in the regulation of stress-induced apoptosis [bib_ref] BCL-2 family expression in human neutrophils during delayed and accelerated apoptosis, Moulding [/bib_ref] [bib_ref] Bid truncation, bid/bax targeting to the mitochondria, and caspase activation associated with..., Maianski [/bib_ref] [bib_ref] The interplay between BCL-2 family proteins and mitochondrial morphology in the regulation..., Soriano [/bib_ref].
Bcl-2 associated X protein (Bax) is necessary for inducing apoptosis [bib_ref] Bcl-Xl-and Bax-alpha-mediated regulation of apoptosis of human neutrophils via caspase-3, Weinmann [/bib_ref] and its translocation and redistribution to the mitochondria is essential for implementing the apoptotic program [bib_ref] BCL-2 family expression in human neutrophils during delayed and accelerated apoptosis, Moulding [/bib_ref] [bib_ref] Bcl-Xl-and Bax-alpha-mediated regulation of apoptosis of human neutrophils via caspase-3, Weinmann [/bib_ref] [bib_ref] Apoptosis of neutrophils, Maianski [/bib_ref]. Therefore, Bax is considered a quantitative marker of early apoptotic events [bib_ref] Phosphorylation of Bax Ser184 by Akt regulates its activity and apoptosis in..., Gardai [/bib_ref] [bib_ref] Bid truncation, bid/bax targeting to the mitochondria, and caspase activation associated with..., Maianski [/bib_ref] [bib_ref] Neutrophil apoptosis pathways and their modifications in inflammation, Simon [/bib_ref] [bib_ref] Mitochondria as the target of the pro-apoptotic protein Bax, Er [/bib_ref] [bib_ref] Granulocyte colony-stimulating factor delays neutrophil apoptosis by inhibition of calpains upstream of..., Van Raam [/bib_ref]. Anti-apoptotic stimuli inhibit Bax insertion into the mitochondrial membrane, thereby inhibiting its pro-apoptotic activity [bib_ref] Phosphorylation of Bax Ser184 by Akt regulates its activity and apoptosis in..., Gardai [/bib_ref]. On the other hand, myeloid cell leukemia 1 (Mcl-1) promotes neutrophil survival by binding and sequestering Bak and Bax, which are capable of forming pores in the mitochondrial membrane [bib_ref] Mcl-1; the molecular regulation of protein function, Thomas [/bib_ref]. Mcl-1 is up-regulated in response to various survival stimuli [bib_ref] Mcl-1; the molecular regulation of protein function, Thomas [/bib_ref] [bib_ref] Cooperative regulation of Mcl-1 by Janus kinase/stat and phosphatidylinositol 3-kinase contribute to..., Epling-Burnette [/bib_ref] [bib_ref] Granulocyte macrophage colony-stimulating factor signaling and proteasome inhibition delay neutrophil apoptosis by..., Derouet [/bib_ref] [bib_ref] Mcl-1 expression in human neutrophils: regulation by cytokines and correlation with cell..., Moulding [/bib_ref] [bib_ref] Mechanisms involved in interleukin-15-induced suppression of human neutrophil apoptosis: role of the..., Pelletier [/bib_ref] [bib_ref] The anti-apoptotic effect of leukotriene B4 in neutrophils: a role for phosphatidylinositol..., Petrin [/bib_ref] [bib_ref] The dual effects of TNFalpha on neutrophil apoptosis are mediated via differential..., Cross [/bib_ref] and is required for neutrophil viability under SH [bib_ref] Suppression of PMN apoptosis by hypoxia is dependent on Mcl-1 and MAPK..., Leuenroth [/bib_ref] [bib_ref] The loss of Mcl-1 expression in human polymorphonuclear leukocytes promotes apoptosis, Leuenroth [/bib_ref]. Importantly, decrease in Mcl-1 levels precedes the appearance of the apoptotic morphology [bib_ref] Sodium salicylate promotes neutrophil apoptosis by stimulating caspase-dependent turnover of Mcl-1, Derouet [/bib_ref].
MAPKs, in particular p38 [bib_ref] Henson PM: p38 mitogen-activated protein kinase-dependent and -independent intracellular signal transduction pathways..., Frasch [/bib_ref] [bib_ref] Role of p38-mitogenactivated protein kinase in spontaneous apoptosis of human neutrophils, Aoshiba [/bib_ref] [bib_ref] Andersson T: p38-MAPK signals survival by phosphorylation of caspase-8 and caspase-3 in..., Alvarado-Kristensson [/bib_ref] and ERK [bib_ref] Granulocyte-macrophage colony-stimulating factor delays neutrophil constitutive apoptosis through phosphoinositide 3-kinase and extracellular..., Klein [/bib_ref] [bib_ref] Extracellular acidosis induces neutrophil activation by a mechanism dependent on activation of..., Martinez [/bib_ref] regulate the apoptotic program in neutrophils. Specifically, Mcl-1 expression might be regulated by signal transduction through ERK [bib_ref] The anti-apoptotic effect of leukotriene B4 in neutrophils: a role for phosphatidylinositol..., Petrin [/bib_ref]. ERK is also responsible for increasing Mcl-1 through protein stabilization by granulocyte-macrophage colony-stimulating factor (GM-CSF) [bib_ref] Granulocyte macrophage colony-stimulating factor signaling and proteasome inhibition delay neutrophil apoptosis by..., Derouet [/bib_ref]. Sustained hypoxia can increase neutrophil survival by activating p38MAPK signaling, thereby inducing Mcl-1 proteins [bib_ref] Suppression of PMN apoptosis by hypoxia is dependent on Mcl-1 and MAPK..., Leuenroth [/bib_ref].
Previously we have shown that NF-κB, its downstream gene IL-8, CXCR2 receptor expression, and p38MAPK signaling pathways are essential for controlling neutrophil survival in healthy individuals treated with IH in-vitro via the extrinsic pathway which is Fas receptors and TNF-α dependent [bib_ref] Molecular pathways of spontaneous and TNF-{alpha}-mediated neutrophil apoptosis under intermittent hypoxia, Dyugovskaya [/bib_ref]. To further elucidate the mechanisms involved in prolonging neurtophil survival under IH in-vitro as well as in patients with OSA, herein we investigated the intrinsic stress-induced mitochondrial pathway. These effects of IH were investigated during the early pro-apoptotic events, which occurred in neutrophils before the appearance of morphological changes and caspase's cascade activation. Thus, we show that Bax expression was decreased and its translocation to the mitochondria was inhibited under IH in-vitro. Concurrently, Mcl-1 expression was up-regulated via activation of ERK1/2 and p38MAPK dependent signaling pathways. Finally, we ascertained the involvement of the mitochondrial network in prolonging the survival of neutrophils in patients with OSA. Similar to the IH in-vitro model, in OSA patients which represent an IH in-vivo model, Bax did not co-localize with the mitochondria and Bax/Mcl-1 ratio was significantly lower than in healthy controls.
# Methods
## Neutrophil isolation and treatment
Blood samples were obtained from 10 healthy volunteers (age=35.8±11.9 yr, BMI=25.3±2.6 Kg/m 2 ) and from 7 OSA patients (age=51.4±15.4 yr, BMI=30.2±5.5 Kg/m 2 , apnea-hypopnea index (AHI)=35.7±20 events/hrs). All control subjects and OSA patients were free from cardiovascular disease or diabetes and had normal blood pressure values (not higher than 140/90 mm Hg). All controls and most OSA patients did not take medications for at least 2 weeks before the study was conducted. Two OSA patients used irregularly low-dose acetyl salicylic acid (micropirin-75). In 7/10 healthy-controls, AHI (2.1±1.8 events/hrs) was determined by a validated home monitored device (watch-PAT-100 Itamar Medical, Caesarea, Israel) [bib_ref] Evaluation of a portable device based on peripheral arterial tone for unattended..., Bar [/bib_ref] and 3/10 controls underwent full-night polysomnography (AHI 8.0 ±1.7 events/hrs) as all OSA patients (Technion Sleep Medicine Center, Haifa). OSA diagnosis was based on the recommendations of the American Academy of Sleep Medicine Task Force with a cutoff point of AHI≥10. Lipid profile and high sensitivity C-reactive protein (CRP) were determined in patients and controls as previously described [bib_ref] Delayed neutrophil apoptosis in patients with sleep apnea, Dyugovskaya [/bib_ref]. The protocol was approved by the local Human Rights Committee, and all participants signed an informed consent form.
Blood samples were withdrawn under fasting conditions and polymononuclear cells (PMNs) were isolated using a two layer Ficoll-Histopaque density gradient centrifugation (Histopaque 1.077 and 1.119, Sigma-Aldrich, Inc., St. Louis, MO, USA). PMN purity was greater than 96%, and viability was greater than 99%, as determined by trypan blue exclusion. Purified PMNs were resuspended in RPMI-1640 medium, supplemented with 10% FCS and 1 mM L-glutamine, plated without/with inhibitors and exposed to normoxia, SH or IH using the BioSpherix-OxyCycler C42 system as we described previously [bib_ref] Delayed neutrophil apoptosis in patients with sleep apnea, Dyugovskaya [/bib_ref] [bib_ref] Molecular pathways of spontaneous and TNF-{alpha}-mediated neutrophil apoptosis under intermittent hypoxia, Dyugovskaya [/bib_ref].
## Light-microscopy assessment of neutrophil apoptosis
Purified neutrophils' cytospin preparations were fixed, and stained with May Grunwald-Giemsa. Slides were read blindly by Axiovert 25 (Zeizz) light microscope. At least 300 cells/slide were analyzed. Cells showing apoptotic morphology were identified according to the following criteria: nuclear condensation in the form of a single nucleus or nuclear fragments not connected by strands [bib_ref] Sequential morphologic events during apoptosis of human neutrophils. Modulation by lipoxygenase-derived eicosanoids, Hebert [/bib_ref] [bib_ref] Cross-linking of human FcgammaRIIIb induces the production of granulocyte colony-stimulating factor and..., Durand [/bib_ref].
## In-vitro ih and sh protocol
Purified PMNs (0.6 ml per well; 3-4 × 10 6 cells/ml) were plated into 24 well plates and then were exposed to normoxia, SH or IH in custom-designed incubation chambers which are attached to an external O 2 -CO 2 -N 2 computer-driven controller using BioSpherix-OxyCycler-C42 system (Redfield, NY, USA). This system which enables to create periodic changes in external O 2 concentrations that control air gas levels in each chamber individually was described in detail previously [bib_ref] Delayed neutrophil apoptosis in patients with sleep apnea, Dyugovskaya [/bib_ref] [bib_ref] Molecular pathways of spontaneous and TNF-{alpha}-mediated neutrophil apoptosis under intermittent hypoxia, Dyugovskaya [/bib_ref]. Briefly, for IH, the O 2 saturation in the medium was kept at 2% for 6.6 ± 3.6 min durations, out of each 1 hr cycle. In each experiment 6 IH cycles were run. SH was employed for a comparable time at 2% actual oxygen in the medium for the entire period. Control purified PMNs were maintained in normoxic conditions for the same durations. Oxygen levels in the medium were determined by a fiber-optic dissolved oxygen electrode (BioSpherix, Redfield, NY, USA).
# Western blot analysis
PMNs cultured in normoxia, IH or SH, were lysed in Tris buffered Saline Triton -X (TBST) at pH 7.4 (50 mM Tris-HCl, 150 mM NaCl, 0.5% Triton X-100, 0.2 mM sodium vanadate), supplemented with a mixture of protease inhibitors (Roche Diagnostics GmbH, Roche Applied Science, Mannheim, Germany), and stored at −80°C until use. Cell lysates were centrifuged at 16,000 × g for 15 min and protein concentration was determined by Bradford reagent (Bio-Rad Laboratories GmbH, Munich, Germany). Cell lysates (50 μg) were run on 12% SDS-PAGE and transferred onto Hybond nitrocellulose membranes (Amersham Biosciences Europe GmbH, Freiburg, Germany). Membranes were blocked and incubated with primary rabbit polyclonal antibodies against Thr180/Tyr182-phosphorylated p38MAPK (Cell Signaling Technology, Inc., Beverly, MA, USA), Thr202/Tyr204-phosphorylated ERK1/2 (Cell Signaling Technology, Inc., Beverly, MA, USA), Bax (N20, sc493) and Mcl-1 (S-19, sc819, Santa Cruz Biotechnology Inc., CA, USA), followed by goat anti-rabbit IgG incubation (Amersham Biosciences Europe GmbH, Freiburg, Germany). Then membranes were washed six times with TBST buffer and incubated with horseradish peroxidaseconjugated secondary antibody (goat anti-rabbit IgG; Amersham Biosciences Europe GmbH, Freiburg, Germany) for 1 hr at room temperature. Densitometric analysis was performed using TotalLab TL100 v.2006c software (Nonlinear Dynamics Ltd., Newcastle Upon Tyne, UK), and is expressed in arbitrary units.
## Confocal laser scanning microscopy
Viable neutrophils were stained with 100 nM Mito-Tracker Orange CMTMRos (Invitrogen, Molecular Probes, Eugene, Oregon, USA) for mitochondria [bib_ref] Granulocyte colony-stimulating factor inhibits the mitochondria-dependent activation of caspase-3 in neutrophils, Maianski [/bib_ref]. Then fixed, Triton-permeabilized and labeled with anti-Bax (N20, sc493) or anti-Mcl-1(S-19, sc819) polyclonal antibodies (Santa Cruz Biotechnology Inc., CA, USA) followed by Cy TM 2-conjugated Goat anti-Rabbit IgG incubation (Jackson ImmunoResearch Laboratories, Inc., Baltimore Pike, PA, USA). Nuclei were stained with TO-PRO-3 (Invitrogen Inc., Carlsbad, CA, USA). Slides were mounted with fluorescence mounting medium (Vectashield H-1000, Vector lab Inc., Burlingame, CA, USA) and were analyzed by confocal laser scanning fluorescence system (Radiance 2000) with Nikon E600 (Japan) camera. Controls for staining included a primary nonspecific rabbit IgG, secondary antibodies and fivefold excess Mcl-1 blocking peptide (sc819 P, Santa Cruz Biotechnology Inc., CA, USA).
## Quantitative fluorescence intensity and co-localization analysis
Relative quantitation of green (Bax) and red (Mitochondrial staining) fluorescence of each cell was accomplished by acquiring grayscale images and fluorescence intensities were integrated using ImageJ 1.42q (Wayne Rasband National Institute of Health, USA). Co-localization of Bax and mitochondria was quantified by LSM 510 CarlZeiss MicroImaging GmbH v.4.2 R&D in collaboration with EmBl, Heidelberg, Germany using Manders Overlap Coefficient (MOC) [bib_ref] Measurement of co-localization of objects in dual-colour confocal images, Manders [/bib_ref]. Only neutrophils with MOC>0.6 were considered as cells with significant co-localization. At least 50 pre-apoptotic neutrophils from different fields were counted in each sample.
## Inhibitor experiments
MAPK inhibitors included: U0126 (10μM) for MEK1/2 blocking (Signal Transduction, Beverly, MA, USA) and SB202190 (30μM) for p38MAPK blocking (Calbiocem, EMD Chemicals, Inc., NJ, USA) [bib_ref] Granulocyte macrophage colony-stimulating factor signaling and proteasome inhibition delay neutrophil apoptosis by..., Derouet [/bib_ref] [bib_ref] Phosphatidylinositol 3-kinase and p38 mitogen-activated protein kinase regulate induction of Mcl-1 and..., Saffar [/bib_ref].
# Statistical analysis
Data are expressed as mean ± SD. A paired two-tailed t test was used for single comparison of parametric data. Values of p<0.05 were considered significant. A paired two-tailed t test with Bonferroni correction was used to compare the effects of IH and SH vs. normoxia. Therefore, for multiple comparisons only values of p<0.017 were considered significant. The NCSS 2004 statistical package, Kaysville, Utah, USA was used.
# Results
## Ih attenuates bax translocation to the mitochondria and its levels
To determine the effects of IH on neutrophil survival, apoptosis was quantified morphologically by light microscopy. The percentage of apoptotic neutrophils, as determined by a single nucleus with dense chromatin condensation, or nuclear fragments not connected by strands, was 25.0±6.3% in normoxia. Exposing neutrophils to 6 IH cycles or to 6 hrs of SH significantly decreased the percentage of apoptotic neutrophils (14.5 ±6.5%, p=0.0002 IH vs. normoxia; 19.5±5.6%, p=0.016 SH vs. normoxia). These baseline values confirmed our earlier findings that IH in-vitro increased neutrophil survival [bib_ref] Delayed neutrophil apoptosis in patients with sleep apnea, Dyugovskaya [/bib_ref] [bib_ref] Molecular pathways of spontaneous and TNF-{alpha}-mediated neutrophil apoptosis under intermittent hypoxia, Dyugovskaya [/bib_ref].
Under confocal microscopy apoptotic neutrophils were identified by the typical morphology of dense nuclei. The apoptotic neutrophils were also characterized by a very high Bax expression (green fluorescence, from 120,000 to 140,000 units), and its fusion with mitochondria (MOC>0.6; yellow-orange dots), as depicted in [fig_ref] Figure 1: Bax expression and co-localization with mitochondria in neutrophils exposed to various oxygen... [/fig_ref]. Such apoptotic neutrophils, which are more prevalent in normoxia, were not investigated in further experiments, since we focused on earlier mechanisms that trigger the apoptotic program before visible signs of apoptosis can be detected. Therefore Bax expression and its translocation to the mitochondria under the three oxygen conditions were examined only in pre-apoptotic neutrophils, characterized by normal nuclear morphology [fig_ref] Figure 1: Bax expression and co-localization with mitochondria in neutrophils exposed to various oxygen... [/fig_ref] , C).
Neutrophils of healthy subjects were exposed to IH and compared by quantitative immunofluorescence to those exposed to SH and normoxia. In normoxia, preapoptotic neutrophils demonstrated intensive fusion of Bax with mitochondria, with a shift in fluorescence to yellow-orange (MOC>0.6), as depicted in [fig_ref] Figure 1: Bax expression and co-localization with mitochondria in neutrophils exposed to various oxygen... [/fig_ref]. In contrast, in IH and SH treated neutrophils Bax and mitochondria were located separately with diffuse Bax distribution (green fluorescence) and the mitochondria remained dotted (red fluorescence) in the cytoplasm [fig_ref] Figure 1: Bax expression and co-localization with mitochondria in neutrophils exposed to various oxygen... [/fig_ref]. MOC in IH and SH treated neutrophils was lower than <0.4, indicating that less than 40% of both components overlapped. [fig_ref] Figure 1: Bax expression and co-localization with mitochondria in neutrophils exposed to various oxygen... [/fig_ref] summarizes the translocation of Bax and its co-localizion with mitochondria for 10 separate experiments. In normoxia, Bax translocation/co-localization (MOC>0.6) was noted in 53.7±12.9% of the neutrophils. In contrast, after treatment with IH or SH, the percentage of neutrophils with Bax translocating to the mitochondria was significantly decreased as compared to normoxia. However, it did not differ significantly between IH and SH treatments. Bax expression under the three oxygen conditions is summarized in [fig_ref] Figure 1: Bax expression and co-localization with mitochondria in neutrophils exposed to various oxygen... [/fig_ref]. Since the average fluorescence intensity of Bax expression per cell varied depending on the blood donor investigated, Bax expression in normoxia in each subject was considered as 100%, and the changes induced by IH or SH were plotted as a relative percentage of this value. The cumulative data [fig_ref] Figure 1: Bax expression and co-localization with mitochondria in neutrophils exposed to various oxygen... [/fig_ref] show that Bax expression was significantly down-regulated in neutrophils treated by either IH (by 41%) or SH (by 27%), compared to normoxia. [fig_ref] Figure 1: Bax expression and co-localization with mitochondria in neutrophils exposed to various oxygen... [/fig_ref] -H depicts representative confocal microscope photomicrographs of Bax expression in normoxia, IH, and SH. After 6 hrs of normoxia the intensity of Bax expression in pre-apoptotic neutrophils was slightly and nonsignificantly increased by 12% as compared to Time 0.
The decrease in Bax expression in the hypoxic conditions was also confirmed by protein levels as determined by western blot analysis. The following relative values for Bax expression over β-actin were obtained: normoxia, 1.82±0.7 units; IH, 0.96±0.2 units (p=0.03 vs. normoxia) and SH, 0.97±0.5 units (p=0.04 vs. normoxia). A representative immunoblot of Bax protein levels over β-actin from 6 independent experiments is depicted in [fig_ref] Figure 2: Mcl-1 expression in neutrophils exposed to various oxygen conditions [/fig_ref].
## Ih up-regulates the levels of mcl-1 protein
Similar to Bax, total Mcl-1 expression was also assessed at the protein level by western blotting as illustrated in [fig_ref] Figure 2: Mcl-1 expression in neutrophils exposed to various oxygen conditions [/fig_ref]. The average densitometric analysis from 6 independent experiments is presented in [fig_ref] Figure 2: Mcl-1 expression in neutrophils exposed to various oxygen conditions [/fig_ref]. In neutrophils exposed to IH or to SH, Mcl-1 protein foldincrease over β-actin was significantly higher by about 2-fold compared to normoxia. Also Mcl-1 up-regulation was observed by confocal microscopy, as illustrated in [fig_ref] Figure 2: Mcl-1 expression in neutrophils exposed to various oxygen conditions [/fig_ref]. Similar to Bax analysis, the intensity of Mcl-1 expression in normoxia for each subject was considered as 100% and the changes induced by IH or SH were plotted as a relative percentage of this value. The specificity of Mcl-1 was confirmed using five-fold excess of the Mcl-1 blocking peptide, which abolished Mcl-1 fluorescent staining. Representative confocal microscope
## Erk and p38mapk activation in response to hypoxia
To directly asses ERK1/2 and p38MAPK activation by IH, their phosphorylation was determined by western blotting. As depicted in , only IH but not SH significantly triggered the phosphorylation of ERK1/2. This pattern of ERK1/2 activation was consistently seen in each separate experiment performed with neutrophils isolated from 6 different donors. For comparison with ERK1/2, we also confirmed our earlier findings showing that p38MAPK phosphorylation was induced in response to both IH and SH [bib_ref] Molecular pathways of spontaneous and TNF-{alpha}-mediated neutrophil apoptosis under intermittent hypoxia, Dyugovskaya [/bib_ref]. is a representative immunoblot depicting ERK1/2 and p38MAPK phosphorylation. Non-phosphorylated controls of ERK1/ 2 and p38MAPK did not differ between the treatments.
## Bax expression and co-localization in neutrophils of osa patients
Bax expression and translocation to the mitochondria was also assessed in neutrophils of OSA patients. Neutrophils cultured for 6 hrs in normoxia or 6 cycles of IH were compared to controls. Three out of seven studied patients were obese having a BMI>30. Three out of ten healthy controls were investigated concurrently with the OSA patients. All underwent full-night polysomnography after which blood samples were taken. Individual demographic, blood chemistry and sleep data for OSA patients and the controls are presented in [fig_ref] Table 1: Demographic, blood chemistry and sleep apnea measures for patients and controls participating... [/fig_ref].
The pre-apoptotic neutrophils of these control subjects expressed Bax translocation to the mitochondria under normoxia as described earlier for healthy controls [fig_ref] Figure 5: Bax expression and co-localization with mitochondria in neutrophils of OSA patients and... [/fig_ref] , and treatment with IH inhibited Bax/mitochondria co-localization [fig_ref] Figure 5: Bax expression and co-localization with mitochondria in neutrophils of OSA patients and... [/fig_ref]. In contrast, in patients with OSA there was little, if any, Bax translocation and co-localization to the mitochondria in normoxia [fig_ref] Figure 5: Bax expression and co-localization with mitochondria in neutrophils of OSA patients and... [/fig_ref] , as well as in IH [fig_ref] Figure 5: Bax expression and co-localization with mitochondria in neutrophils of OSA patients and... [/fig_ref]. These findings were noted in non-obese patients with low CRP levels [fig_ref] Figure 5: Bax expression and co-localization with mitochondria in neutrophils of OSA patients and... [/fig_ref] , D) as well as in obese patients with high CRP levels [fig_ref] Figure 5: Bax expression and co-localization with mitochondria in neutrophils of OSA patients and... [/fig_ref] , F). As stated above, the fluorescence intensity of Bax and Mcl-1 expression was an individual trait. We therefore used Bax/Mcl-1 ratio for comparing the redistribution of pro-/anti-apoptotic proteins between OSA and healthy controls. The average Bax/Mcl-1 ratio in normoxia was 2-fold higher in healthy controls as compared to OSA patients and was significantly decreased by about 60% and 50% after treatment with IH and SH, respectively [fig_ref] Table 2: Bax/Mcl-1 ratio in controls and OSA patients [/fig_ref]. In OSA patients, the Bax/Mcl-1 ratio was already low at normoxia (1.0±0.5) and was further decreased after exposure to IH as depicted in [fig_ref] Table 2: Bax/Mcl-1 ratio in controls and OSA patients [/fig_ref]. Similar values were obtained for Bax/Mcl-1 ratio in normoxia immediately after harvesting the cells (data not shown).
# Discussion
Neutrophils survival was shown to increase in response to IH in-vitro as well as in-vivo [bib_ref] Delayed neutrophil apoptosis in patients with sleep apnea, Dyugovskaya [/bib_ref] [bib_ref] Molecular pathways of spontaneous and TNF-{alpha}-mediated neutrophil apoptosis under intermittent hypoxia, Dyugovskaya [/bib_ref] , however, the underlying mechanisms are not entirely understood. In the present study we investigated the contribution of the mitochondrial stress-induced pathway in prolonging neutrophil survival under IH treatment in-vitro and in a human IH model in-vivo. In neutrophils treated by IH patients undergoing nightly IH, Bax did not co-localize with the mitochondria and Bax/Mcl-1 ratio was significantly lower than in healthy controls. The Bcl-2 family of proteins is one of the key regulators of cell death at the mitochondrial level [bib_ref] The interplay between BCL-2 family proteins and mitochondrial morphology in the regulation..., Soriano [/bib_ref] [bib_ref] Apoptosis of neutrophils, Maianski [/bib_ref] [bib_ref] Functional characterization of mitochondria in neutrophils: a role restricted to apoptosis, Maianski [/bib_ref] [bib_ref] The mitochondrial network of human neutrophils: role in chemotaxis, phagocytosis, respiratory burst..., Fossati [/bib_ref] [bib_ref] Early events in spontaneous neutrophil apoptosis, Scheel-Toellner [/bib_ref] [bib_ref] Neutrophil apoptosis and the resolution of infection, Kennedy [/bib_ref] , and Bax is the best known pro-apoptotic protein. In most cell types, the expression and activity of anti-apoptotic Bcl-2 members is higher than pro-apoptotic members. By contrast, mature neutrophils constitutively express pro-apoptotic proteins and the expression of the antiapoptotic Bcl-2 members is very low [bib_ref] Molecular control of neutrophil apoptosis, Akgul [/bib_ref] [bib_ref] NADPH oxidase-derived ROS: key modulators of hemeinduced mitochondrial stability in human neutrophils, Arruda [/bib_ref]. Thus, the balance between pro-and anti-apoptotic members determines the fate of cells [bib_ref] New insights into the mechanisms controlling neutrophil survival, Cabrini [/bib_ref].
High Bax expression and its fusion with mitochondria were noted in apoptotic neutrophils by confocal microscopy analysis. Bax was also abundantly expressed, to a [bib_ref] Phosphorylation of Bax Ser184 by Akt regulates its activity and apoptosis in..., Gardai [/bib_ref] [bib_ref] Mcl-1 expression in human neutrophils: regulation by cytokines and correlation with cell..., Moulding [/bib_ref] [bib_ref] Functional characterization of mitochondria in neutrophils: a role restricted to apoptosis, Maianski [/bib_ref] [bib_ref] Constitutive neutrophil apoptosis: mechanisms and regulation, Luo [/bib_ref]. Its translocation to the mitochondria leads to the release of pro-apoptotic factors such as cytochrome c, which complexes with apoptotic protease-activating factor 1 (Apaf-1) and pro-caspase-9 to form a protein complex the 'apoptosome' which is involved in caspase-3 activation. The latter is responsible for the visible signs of apoptosis [bib_ref] Bid truncation, bid/bax targeting to the mitochondria, and caspase activation associated with..., Maianski [/bib_ref] [bib_ref] Apoptosis of neutrophils, Maianski [/bib_ref] [bib_ref] Early events in spontaneous neutrophil apoptosis, Scheel-Toellner [/bib_ref] [bib_ref] New insights into the mechanisms controlling neutrophil survival, Cabrini [/bib_ref] [bib_ref] Constitutive neutrophil apoptosis: mechanisms and regulation, Luo [/bib_ref]. Accordingly, our findings demonstrate that changes in Bax/Mcl-1 expression and translocation to the mitochondria were noted before the appearance of apoptotic morphology, as expected, but also before caspase activation, as indicated by flow cytometry and confocal microscopy using FAN-FLICA Poly Caspase Kit (binds the activated caspases-1, -3, -4, -5, -6, -7, -8, -9) (Dyugovskaya L, Berger S, unpublished observation).
Mcl-1 has a short half-life (~3 hrs), and spontaneous apoptosis is accompanied by Mcl-1 degradation [bib_ref] The loss of Mcl-1 expression in human polymorphonuclear leukocytes promotes apoptosis, Leuenroth [/bib_ref]. However, its expression may be increased depending on the stimuli exerted [bib_ref] BCL-2 family expression in human neutrophils during delayed and accelerated apoptosis, Moulding [/bib_ref]. Mcl-1 expression was increased in IH and SH in-vitro treated neutrophils compared to normoxia. It is likely that both IH and SH may induce Mcl-1 stabilization by preventing its degradation, and/or possibly by up-regulating its production. Besides in IH and SH demonstrated in this study, increased Mcl-1 levels have been previously implicated in neutrophil survival induced by LPS, cytokines such as GM-CSF, IL-1, TNF-α, IL-15 [bib_ref] Cooperative regulation of Mcl-1 by Janus kinase/stat and phosphatidylinositol 3-kinase contribute to..., Epling-Burnette [/bib_ref] [bib_ref] Granulocyte macrophage colony-stimulating factor signaling and proteasome inhibition delay neutrophil apoptosis by..., Derouet [/bib_ref] [bib_ref] Mcl-1 expression in human neutrophils: regulation by cytokines and correlation with cell..., Moulding [/bib_ref] [bib_ref] Mechanisms involved in interleukin-15-induced suppression of human neutrophil apoptosis: role of the..., Pelletier [/bib_ref] [bib_ref] The dual effects of TNFalpha on neutrophil apoptosis are mediated via differential..., Cross [/bib_ref] , leukotriene B4 [bib_ref] The anti-apoptotic effect of leukotriene B4 in neutrophils: a role for phosphatidylinositol..., Petrin [/bib_ref] , Toll-likereceptor agonist [bib_ref] Inhibition of neutrophil apoptosis by TLR agonists in whole blood: involvement of..., Francois [/bib_ref] and SH for at least 8 hrs at less than 2% oxygen [bib_ref] The loss of Mcl-1 expression in human polymorphonuclear leukocytes promotes apoptosis, Leuenroth [/bib_ref] , as obtained in this study for 6 hrs of SH.
Mcl-1 expression in neutrophils is regulated by a diverse array of signal transduction pathways which depend on the stimuli exerted [bib_ref] Granulocyte macrophage colony-stimulating factor signaling and proteasome inhibition delay neutrophil apoptosis by..., Derouet [/bib_ref] [bib_ref] Phosphatidylinositol 3-kinase and p38 mitogen-activated protein kinase regulate induction of Mcl-1 and..., Saffar [/bib_ref]. Here, we demonstrate for the first time that only under IH the up-regulation of Mcl-1 coincided with p-ERK1/2 activation, and by inhibiting ERK1/2, the expression of Mcl-1 was inhibited. In contrast, p38MAPK was up-regulated by both IH as well as by SH as previously shown [bib_ref] Molecular pathways of spontaneous and TNF-{alpha}-mediated neutrophil apoptosis under intermittent hypoxia, Dyugovskaya [/bib_ref] , and its inhibition affected Mcl-1 expression under both hypoxic conditions. Also, like in our SH experimental conditions, similar findings were reported for neutrophils exposed to 12 hrs of SH. Inhibition of p38MAPK led to a significant decrease in Mcl-1 expression, whereas inhibiting ERK1/2 led only to a slight, but not significant decrease in Mcl-1 levels [bib_ref] Suppression of PMN apoptosis by hypoxia is dependent on Mcl-1 and MAPK..., Leuenroth [/bib_ref].
The selective ERK1/2 phosphorylation in human neutrophils by IH suggests that Mcl-1 activity might be regulated by different signal transduction pathways in various hypoxic conditions, such as in IH and SH demonstrated here. We should note however, that other pathways not investigated in this study, in addition to p38MAPK and ERK1/2 could be involved in the upregulation of Mcl-1 under IH. For instance, the NF-κBdependent up-regulation of IL-8 levels described earlier for IH [bib_ref] Molecular pathways of spontaneous and TNF-{alpha}-mediated neutrophil apoptosis under intermittent hypoxia, Dyugovskaya [/bib_ref] may control the expression of survival genes of Bcl-2 family members [bib_ref] The survival effect of TNF-alpha in human neutrophils is mediated via NF-kappa..., Cowburn [/bib_ref] by increasing antiapoptotic and decreasing pro-apoptotic proteins levels in neutrophils [bib_ref] Anti-IL-8 autoantibody:IL-8 immune complexes suppress spontaneous apoptosis of neutrophils, Fudala [/bib_ref].
Finally we showed for the first time that in OSA patients Bax translocation to the mitochondria was minimal in neutrophils maintained at normoxic conditions, and it was further reduced in response to IH in-vitro in all patients investigated regardless of weight differences. Moreover, the normoxic values obtained for OSA were similar to those of control neutrophils exposed to IH in-vitro, illustrating the similarities between in-vitro and in-vivo IH. Additionally, the ratio Bax/Mcl-1 was significantly lower in OSA patients at normoxia as compared to control subjects clearly demonstrating that proapoptotic Bax was low whereas the anti-apoptotic Mcl-1 protein was high. Collectively, these finding suggest that the IH-dependent prolonged neutrophil survival in OSA is largely affected by the mitochondrial stress-induced pathway.
Elucidating potential mechanisms which might suppress neutrophil apoptosis by IH in-vivo, is of a great importance to OSA and sleep disordered breathing (SDB). OSA is a prevalent syndrome associated with cardiovascular morbidity and mortality [bib_ref] Sleep apnea and cardiovascular disease: an American Heart Association/American College of Cardiology..., Somers [/bib_ref]. It affects at least 4% and 2% of men and women in the adult population. However, the prevalence of SDB (IH without the characteristic excessive daytime sleepiness) is estimated to be as high as 24% and 9% in men and women. This value may rise to 60-90% in obese individuals [bib_ref] Obstructive sleep apnoea, Malhotra [/bib_ref]. Moreover, SDB is also highly prevalent in more than 60% of patients with acute myocardial infarction and in more than 50% (44-93%) of patients with stroke [bib_ref] Sleep apnea and cardiovascular disease: an American Heart Association/American College of Cardiology..., Somers [/bib_ref]. Furthermore, OSA and SDB are also associated with increased vascular inflammation, endothelial dysfunction and atherosclerosis [bib_ref] Sleep apnea and cardiovascular disease: an American Heart Association/American College of Cardiology..., Somers [/bib_ref] [bib_ref] Molecular mechanisms of cardiovascular disease in OSAHS: the oxidative stress link, Lavie [/bib_ref].
# Study limitations
The limitations of our study regarding the patients with OSA should be acknowledged. First, the IH experienced by OSA patients was not mimicked in our IH in-vitro model due to technical constrains. The short intervals as experienced by patients with OSA are rather difficult to reproduce with neutrophils in culture conditions because they are non-adherent cells. Therefore, the medium cannot be replaced at short intervals with alternating preconditioned hypoxic or normoxic medium. Despite that, the findings in OSA patients were similar to those obtained by IH in-vitro. Yet, it would be interesting to repeat this type of a study in an animal model of IH, that allows mimicking the time patterns of IH more closely to OSA patients and in a dose dependent manner, and to avoid confounding factors and comorbidities [bib_ref] Non-invasive system for applying airway obstructions to model obstructive sleep apnea in..., Carreras [/bib_ref].
Second, the number of OSA patients investigated in the current study is relatively small and there are significant differences in age and BMI between patients and controls. However, in a previous study [bib_ref] Delayed neutrophil apoptosis in patients with sleep apnea, Dyugovskaya [/bib_ref] meticulously and rigorously investigating more than 100 patients we have shown by linear regression analysis that all apnea measures (AHI, oxygen desaturation index of 3%, minimal oxygen desaturation, and % time spent below 90% saturation) were significantly correlated with decreased neutrophil apoptosis, clearly attesting to the importance of OSA severity. Moreover, using multivariate analysis we could verify that this relationship was independent of BMI, age, CRP, triglycerides, and adiponectin. Thus, although some of the patients investigated in the present study were obese and had higher CRP levels than controls, their results with respect to Bax/Mcl-1 ratio and the lack of Bax translocation to the mitochondria were identical to the results obtained in non-obese OSA patients with low CRP levels and like in the IH in-vitro, and unlike the normoxic controls' findings. It is noteworthy that in the previous study [bib_ref] Delayed neutrophil apoptosis in patients with sleep apnea, Dyugovskaya [/bib_ref] we also demonstrated the contribution of the IH and the AHI severity measures, regardless of confounding factors, in attenuating neutrophil apoptosis by investigating OSA patients with and without treatment with nasal continues positive airway pressure (nCPAP), which ameliorates the apneas. Patients on nCPAP treatment were enrolled and were investigated on two consecutive nights, one with and one without nCPAP. Patients demographics, BMI, lipid profile or CRP were unaffected by omitting the treatment, yet all apnea severity measures were increased and neutrophil apoptosis was decreased as compared to the treatment night. These earlier data clearly emphasize the role played by the IH in increasing neutrophils lifespan, therefore allowing to investigate the mechanisms of neutrophil apoptosis under conditions of IH in-vivo in patients as well as in-vitro in healthy controls.
# Conclusions
This study demonstrates that Bax/Mcl-1 ratio was significantly lowered in neutrophils treated by IH in-vitro and in patients with OSA by up-regulating the anti-apoptotic Mcl-1 and down-regulating the pro-apoptotic Bax. As a result of the IH, Bax translocation to the mitochondria was prevented. Thus, IH converts the proapoptotic phenotype into an anti-apoptotic one by modulating the Bcl-2 family members Bax and Mcl-1. This effect of IH is specifically mediated through ERK1/ 2 and p38MAPK signaling pathways whereas in SH it is mediated only through p38MAPK. Hence, identifying neutrophil survival pathways affected by IH may lead to new approaches in treating some sleep apnea complications associated with endothelial dysfunction and atherosclerosis. Moreover, these findings may bear relevance to other conditions and co-morbidities associated with components of IH such as physical activity, brief ascents to altitude, myocardial infarction and cancer.
[fig] Figure 1: Bax expression and co-localization with mitochondria in neutrophils exposed to various oxygen conditions. Bax expression and co-localization with mitochondria were analyzed by confocal laser scanning microscopy as described in methods. The cytoplasmic distribution of Bax (green fluorescence) and mitochondria (red fluorescence) was studied by double immunofluorescence labeling. (A) A representative photomicrograph of an apoptotic neutrophil with a very high Bax expression (140,000 arbitrary units) and intensive fusion of Bax with mitochondria (yellow), Manders Overlap Coefficient (MOC)>0.6. (B) A pre-apoptotic neutrophil with intensive Bax expression (17,000 arbitrary units). Bax is translocated and co-localized with mitochondria (yellow and orange), MOC>0.6 representing neutrophils predominantly seen at normoxia (Nox). (C) A pre-apoptotic neutrophil with low Bax expression (9,500 arbitrary units). Bax (green) and mitochondria (red) are located separately (MOC<0.6), representing neutrophils predominantly seen in intermittent hypoxia (IH) and sustained hypoxia (SH).(D). The percentage of pre-apoptotic neutrophils with fusion of Bax and mitochondria (MOC>0.6) in Nox, IH and SH. Values represent the means ± standard deviations (n=10). P values represent significance of IH or SH vs. Nox. (E) Relative percentage of Bax expression, assessed by confocal laser scanning microscopy (n=10). The average fluorescence intensity unit per cell, detected by immunofluorescent quantitation at normoxia was considered as baseline (100%), and the effects of hypoxia were calculated as relative decrease of Bax expression. P values represent significance of IH or SH vs. Nox. (F-H) Representative images of Bax expression (green) out of ten experiments in Nox (F), IH (G), and SH (H).photomicrographs of Mcl-1 expression in normoxia, IH, and SH are presented in Figure 3A-C. After 6 hrs of normoxia the intensity of Mcl-1 expression in pre-apoptotic neutrophils was decreased by about 20% compared to Time 0.Effects of ERK and p38MAPK inhibition on Mcl-1 expressionMAPKs, including ERK1/2 and p38MAPK control neutrophil survival under certain conditions[27,28,36,44,45]. Specific p38MAPK and ERK1/2 inhibitors were used to determine whether Mcl-1 expression was dependent on the activation of MAPK signaling pathways under IH. Neutrophils were incubated in normoxia, SH or IH (as described in methods) with or without 10 μM U0126 (a specific inhibitor of MEK1/2, blocks ERK1/2 activation) or 30 μM SB202190 (a competitive p38MAPK inhibitor). Mcl-1 distribution was determined in pre-apoptotic neutrophils by confocal laser scanning microscopy. Representative images (from 1 out of 4 experiments performed) of Mcl-1 expression in the three oxygen conditions without or with the inhibitors are presented inFigure 3A-I. Figures 3-K depict the average values of Mcl-1 expression for ERK1/2 inhibitor and p38MAPK inhibitor by relative percent when normoxia without the inhibitor was considered as 100%. Blocking ERK/MEK activity slightly increased (p=0.16) Mcl-1 expression in normoxia but significantly decreased the IH-mediated Mcl-1 up-regulation by 40%. In contrast, in SH, Mcl-1 expression was not affected by the ERK/MEK inhibitor (Figure 3D-F, J). Inhibiting p38MAPK also slightly increased (p=0.1) Mcl-1 expression in normoxia, but the hypoxia-induced enhanced Mcl-1 expression, was significantly attenuated in both IH (by 30%) and SH (34%) conditions (Figure 3G-I, K). [/fig]
[fig] Figure 2: Mcl-1 expression in neutrophils exposed to various oxygen conditions. (A) A representative western blot depicting Bax and Mcl-1 expression in normoxia (Nox), intermittent hypoxia (IH) and sustained hypoxia (SH) in one out of six experiments performed. (B) The normalized values of Mcl-1, obtained by densitometric analysis of western blots are presented as arbitrary units, and depicted as fold induction over β-actin in Nox, IH and SH. (C) Mcl-1 expression was assessed by confocal laser scanning microscopy (n=8). The average fluorescence intensity unit per cell, detected by immunofluorescence quantitation at normoxia was considered as baseline (100%), and the effects of IH and SH were calculated as relative Mcl-1increase. P values represent significance of IH or SH vs. Nox. [/fig]
[fig] Figure 3, Figure 4: Effects of ERK and p38MAPK inhibition on Mcl-1 expression in various oxygen conditions. Neutrophils were incubated for 6 hrs at normoxia (Nox), sustained hypoxia (SH), or 6 cycles of intermittent hypoxia (IH) without or with 10 μM U0126 (a specific inhibitor of MEK1/2, which blocks ERK1/2 activation), and 30 μM SB202190 (a competitive inhibitor of the p38 kinase). Mcl-1 expression was assessed by confocal laser scanning microscopy.(A-I) Representative photomicrographs of Mcl-1 expression (green) from 1 out of 4 experiments performed. (A) Neutrophils incubated in Nox, (B) IH and (C) SH. (D) Neutrophils incubated in Nox with MEK1/2 inhibitor, (E) in IH with MEK1/2 inhibitor, and (F) in SH with MEK1/2 inhibitor. (G) Neutrophils incubated in Nox with p38MAPK inhibitor (H) in IH with p38MAPK inhibitor, and (I) in SH with p38MAPK inhibitor. (J) The average values of Mcl-1 expression for ERK1/2 inhibitor and (K) for p38MAPK inhibitor by relative percent. The intensity of Mcl-1 expression at normoxia without inhibitors was considered as 100%, and the changes induced by the IH and SH with or without the inhibitors were plotted as a relative percentage of this value. P values represent significance of untreated vs. inhibitor treated neutrophils in each oxygen condition.in-vitro the expression of the pro-apoptotic protein Bax was decreased, Bax translocation to the mitochondria was inhibited and the anti-apoptotic protein Mcl-1 was up-regulated via activation of ERK1/2 and p38MAPK dependent signaling pathways. In SH treated neutrophils, unlike in IH, Mcl-1 up-regulation was only dependent on p38MAPK but not on ERK1/2 activation. Moreover, using a quantitative confocal microscopy analysis we have shown that the hypoxia-induced changes in Bax/Mcl-1 expression and translocation were noted in neutrophils before the appearance of apoptotic morphology. Similarly to the in-vitro findings, in OSA ERK1/2 and p38MAPK phosphorylation in neutrophils exposed to various oxygen conditions. (A) The average (± SD) normalized values of p-ERK1/2 (n=6), obtained by densitometric analysis of western blots, are presented as fold increase over β-actin in arbitrary units in normoxia (Nox), intermittent hypoxia (IH) and sustained hypoxia (SH). (B) A representative western blot depicts the dual phosphorylated (Thr202/Tyr204) form of ERK1/2 (p-ERK1/2) and dual phosphorylated (Thr180/Tyr182) form of p38MAPK (p-p38) compared to non-phosphorylated ERK1/2 and p38MAPK in Nox, IH, and SH. [/fig]
[fig] Figure 5: Bax expression and co-localization with mitochondria in neutrophils of OSA patients and controls. Bax expression and co-localization with mitochondria was analyzed by confocal laser scanning microscopy in neutrophils of OSA patients and controls in normoxia (Nox) and intermittent hypoxia (IH). The cytoplasmic distribution of mitochondria (red fluorescence) and Bax (green fluorescence) was determined by double immunofluorescence labeling as described in methods. (A, B) Representative images of a control subject (age=38 yr, BMI=25.3 Kg/m 2 , AHI=6.0 events/hrs, CRP=1.48mg/L, 19% apoptotic neutrophils were detected by Giemsa staining). (A) After 6 hrs of normoxia (Nox) Bax is translocated to and colocalized with mitochondria (yellow and orange). Manders Overlap Coefficient (MOC) >0.6 in neutrophils, (B) IH in-vitro inhibited Bax translocation to the mitochondria. (C) Representative images of a non-obese OSA patient (age=37 yr, BMI=24.7 Kg/m 2 , AHI=35.0 events/hrs, CRP=1.99 mg/L, 8% apoptotic neutrophils were detected by Giemsa staining) in normoxia and (D) in IH in-vitro. (E) Representative images of an obese OSA patient (age=48 yr, BMI=34.2 Kg/m 2 , AHI=33.0 events/hrs, CRP=8.09 mg/L, 9% apoptotic neutrophils were detected by Giemsa staining) in normoxia and in (F) IH in-vitro.In obese and non-obese OSA patients Bax was not translocated to the mitochondria; Bax (green) and mitochondria (red) were located separately in normoxia (C, E) and in IH (D, F). [/fig]
[table] Table 1: Demographic, blood chemistry and sleep apnea measures for patients and controls participating in the experiments [/table]
[table] Table 2: Bax/Mcl-1 ratio in controls and OSA patients [/table]
|
Phosphodiesterase 4 Inhibitors in Allergic Rhinitis/Rhinosinusitis
# Introduction
Allergic rhinitis (AR) is an inflammatory disorder of the nasal epithelium. Anterior and posterior rhinorrhea, nasal congestion, sneezing, and itching are typical clinical symptoms, which are reversible spontaneously or in response to therapy.
AR is not considered to be the most serious medical condition compared with other respiratory diseases. However, patients with AR have a reduced quality of life, lack of or worsened quality of sleep, cognitive function impairment, irritability, and fatigue, especially during peak pollen season. The most common form of AR is caused by exposure of sensitized patients to allergens such as house dust mites, grass pollen, tree pollen, weed pollens, cat and dog fur, and mold.
The phenotypes of this disease appear to actively change, with about 50-70% of patients suffering from mixed rhinitis (non-AR and AR) and approximately 10-20% suffering from the pharmacologically resistant phenotype. AR with localized nasal allergic response in the absence of systemic atopy has also been identified.
Standard treatment of AR includes the combination of oral antihistamines and glucocorticoids (GCs), which may not provide sufficient relief from symptoms; thus, novel therapies are needed. In recent years, various AR phenotypes caused by non-T helper 2 (Th2) cellmediated inflammation have been identified. Therefore, it is important to develop more effective drugs with different mechanisms of action such as inhibition of Janus kinase, phosphodiesterase (PDE), and chemoattractant receptor homologous molecule of Th2 cells.
PDEs are responsible for the hydrolysis of intracellular cyclic adenosine monophosphate (cAMP) and cyclic guanosine monophosphate (cGMP), which have emerged as novel therapeutic targets. PDE inhibitors allow the elevation of cAMP and cGMP, which leads to a variety of cellular effects such as relaxation of airway smooth muscle and inhibition of cellular inflammation. Special attention has been paid to PDE4 inhibitors, which are potent anti-inflammatory agents that have shown encouraging results in the treatment of chronic obstructive pulmonary disease (COPD) and asthma. Their potential in the management of upper airway diseases has also been suggested. This review summarizes current evidence on the beneficial as well as harmful effects of the clinical use of PDE4 inhibitors in allergic rhinitis/rhinosinusitis.
## Definition and epidemiology of ar
AR is clinically defined as a symptomatic disorder of the nose induced after allergen exposure by immunoglobulin E (IgE)mediated inflammation of the nasal membranes. AR is an extremely common disease that persists throughout lifeand occurs very frequently. Real-time data are difficult to interpret. Statistics only show physician-diagnosed AR, but the number of patients suffering from AR is a bit higher. During the last 20-30 years, the prevalence of allergic diseases has increased. The prevalence of self-reported AR has also increased and is approximately 2%-25% in children and 1% to more than 40% in adults. The prevalence of confirmed AR in adults oscillates in Europe from 17% to 28.5%, meaning that 100-150 million individuals in Europe suffer from AR. In the United States, 10-30% of the population has been diagnosed with AR. In Asia, AR affects a large part of the population, from 27% in South Korea to 32% in the United Arab Emirates. The incidence of AR seems to be comparable worldwide. AR is frequently associated with asthma; 15-38% of patients with AR have asthma and nasal symptoms are present in 6-85% of asthmatic patients. Thus, AR is a risk factor for asthma and successful management of AR can improve asthma symptoms.
## Pathophysiology of ar
As aforementioned, the most causal antigens for AR are inhalant allergens such as house dust mites, animal dander, and pollen. Many allergens have protease activity and are able to impair the epithelial barrier and penetrate into nasal mucosa. Allergenspecific Th2 cells are produced in patients with AR, whereas allergen-specific Th1 cells are produced in healthy individuals. The results of long-term studies have shown that all allergic diseases are primarily characterized by enhanced Th2 pathway activation, increased serum levels of IgE, blood eosinophil count, and allergen reactivity. However, in the last several years, the concept of AR has shifted from increased Th2 signaling and downstream inflammation to disease phenotypes with non-Th2 mediated inflammation. AR is a largely heterogenous group of airway diseases, and as such, research should not only focus on immunosuppressive agents (e.g., corticosteroids) but should also include targeted immunomodulatory pathways.
Basophil cells produce cytokines such as interleukin 4 (IL-4) and thymic stromal lymphopoietin in response to allergens with protease activity, and may contribute to Th2 differentiation. Th2 cells produce IL-4/IL-13 and express CD40 ligand, which are responsible for converting B cells to IgEproducing B cells. Inhalation of allergens by sensitized patients is followed by antigen passing through the epithelial tight junctions in the nasal mucosa and binding to IgE receptors on the surface of mast cells in the nasal mucosa. Chemical mediators such as histamine and prostaglandins are released. Histamine increases paracellular permeability by regulating tight junctions through coupling with H1 receptors. Interactions among chemical mediators, sensory nerve terminals, and blood vessels are responsible for the early phase of AR (sneezing, rhinorrhea, nasal congestion). Late-phase inflammation develops within 6-10 h after allergen stimulation.
Dysregulation of dendritic cells (DCs) plays an important role in the pathogenesis of AR. Understanding DC biology and its functions in physiologic and pathologic conditions may provide an understanding of the role of DCs in the initiation and perpetuation of inflammatory sinonasal diseases, and also lead to the development of potentially novel therapeutic strategies targeting DCs. DCs influence inflammation and adaptive responses of the body by releasing proinflammatory cytokines and other mediators. They are antigen-presenting cells (APCs) able to polarize T effector cells.
Activation of the cytotoxic activity of natural killer and CD8 + T cells protects against intracellular pathogens, but can also lead to tissue injury. DCs produce IL-23, and together with IL-6 and IL-1b, are involved in the expansion of Th17 cells, which are responsible for strong pro-inflammatory responses against extracellular pathogens, possibly leading to chronic inflammation and autoimmune diseases. Thus, for novel therapeutic strategies, it is important to focus on pathological conditions characterized by the expansion of Th1 and/or Th17 responses.
## Therapy
The primary non-pharmacological strategy in AR management is avoiding allergens and irritants. In relation to the elimination of allergens, topical saline irrigations help to reduce symptoms of AR when used alone or together with other drugs.
As first-line therapy for AR, GCs and antihistamines administered orally or intranasally are used. GCs have become a clinical mainstay for the suppression of numerous inflammatory and autoimmune diseases. They reduce the infiltration of inflammatory cells into the nasal mucosa, maturation of immunocompetent cells, cytokine production, and mediator release by mast cells. They also inhibit histamine release, induce apoptosis of eosinophils, and reduce the recruitment of APCs. Inhibition of B cells through GCs is limited, but GCs are able to inhibit classswitching to IgE in the nasal mucosa. GCs are also known as anti-inflammatory drugs that affect nasal constitutive, epithelial, vascular endothelial cells, and glands. GCs signal through genomic and non-genomic pathways. The classic, genomic actions of GCs are mediated through a single GC receptor (GR) with many isoforms, which expands the heterogeneity of hormonal signaling. GRs are expressed throughout the body and are localized in the cytoplasm of target cells. The main anti-inflammatory effects of GC are based on their ability to inhibit the transcription of cytokine and chemokine genes participating in inflammation.
Three generations of antihistamines that specifically block the effects of histamine on histaminergic receptors are available. First-generation oral antihistamines such as cyproheptadine are no longer recommended, because of their sedative and anticholinergic effects. Second-generation antihistamines such as loratadine or rupatadine lead to good improvements in symptoms without side effects to the central nervous system. They are rapid and well tolerated for the symptomatic treatment of AR.
The combination of oral antihistamines and intranasal corticosteroids is a reasonable choice because of the significantly faster onset of treatment effects. However, the combination of intranasal antihistamines and intranasal corticosteroids does not alleviate symptoms as well as the combination of oral antihistamines with intranasal corticosteroids.
During the above-mentioned long-term therapy, some side effects can occur such as atrophy of the nasal mucosa or systemic side effects, especially in children. Antihistamines can cause drowsiness or prolongation of the QT interval on electrocardiogram. Common pharmacological treatment with intranasal steroids, oral or intranasal antihistamines does not completely relieve the symptoms of AR. This is why more specific substances for disease control are needed.
## Pdes as a therapeutic target
PDEs are a group of enzymes consisting at least of 11 members that cause hydrolysis and subsequent inactivation of cAMP and cGMP serving as intracellular second messengers. They modulate many cellular functions such as contractility of the myocardium, airway and vascular smooth muscle tone, platelet aggregation, and the release of inflammatory mediators. PDE3, PDE4, PDE5, and PDE7 isoforms are the most well studied selective PDEs in relation to immune responses in the respiratory system.
PDE4 is the major cAMP-metabolizing enzyme expressed inside the inflammatory cells. The PDE4 family in mammals consists of four subfamilies, encoded by four genes (PDE4A, PDE4B, PDE4C, and PDE4D). PDE4A, PDE4B, and PDE4D are expressed by neutrophils, eosinophils, B cells, T cells, DCs, monocytes, and macrophages. Expression of PDE4C is typically minimal or absent. Expression is triggered by epithelial damage, microbial invasion, and allergen sensitization. Its inhibition is a promising target for suppressing inflammatory responsesas was demonstrated for roflumilast in an animal model of acute lung injury. PDE4 inhibitors are wellcharacterized pharmaceutical agents with a broad range of anti-inflammatory activities also in various inflammatory conditions including allergic diseases.
As aforementioned, PDE4 inhibitors may serve as potential therapeutic agents for various respiratory diseases, as well as for non-Th2 mediated AR. These effects have been evaluated in vitro and in animal models of allergic asthma. Their antiinflammatory activity results from blocking the degradation of cAMP in lymphocytes, eosinophils, neutrophils, and monocytes, leading to the attenuated release of histamine and leukotrienes as well as the release of several cytokines including IL-4, IL-5, IL-10, and granulocyte-macrophage colony-stimulating factor. Inhibition of PDE4 increases accumulation of intracellular cAMP and helps to balance antiand pro-inflammatory effects. PDE4 inhibitors such as apremilast, roflumilast, and crisaborole have been tested in clinical trials for various inflammatory diseases. Apremilast is now approved for the treatment of adults with moderate to severe plaque psoriasis and/or psoriatic arthritis. Roflumilast has shown initial efficacy for treating asthma, COPD, and asthma-COPD overlap. Based on phase III trials, crisaborole is considered an efficacious topical agent with a safety profile and limited systemic exposure. It is promising candidate for the treatment of atopic dermatitis.
PDE4 inhibitors, such as roflumilast, are able to suppress various inflammatory responses. A clinical study with oral doses of roflumilast in patients with COPD took place in III trial phases demonstrated its anti-allergic and anti-inflammatory benefits. Roflumilast is now approved for treatment of COPDand is recommended at a dose of 500 mg once daily. Many more studies have tested the effects of roflumilast on COPD than on AR.
The efficacy and safety profile of roflumilast shows that it leads to more side effects than other PDE4 inhibitors administered intranasally in patients with COPD. The most frequent side effects are nausea, diarrhea, appetite loss and weight loss, abdominal pain, headache, gastrointestinal, and sleep disturbance. These side effects limit the use of roflumilast in clinical practice. Therefore, real-world studies on the clinical use of this PDE4 inhibitor are limited. No effects on the cardiovascular system have been observed.
The first study to evaluate the efficacy of roflumilast in the treatment of AR was conducted by. In that study, roflumilast was found to be safe and well tolerated at the same dose as that used in COPD, 500 µg once a day. Headache was the most common side effect, and was followed by nausea and dizziness in some patients. Three days after the onset of treatment, increased airflow at rhinomanometry was recorded. After 4 d of treatment, subjective improvement of the obstruction was reported by the patients. Taken together, this study provided evidence that orally administered roflumilast is an effective anti-allergy therapy in AR. No further studies have been conducted on the efficacy of roflumilast in AR.
## Novel pde4 inhibitors
The clinical development of PDE4 inhibitors has recently focused on the novel PDE4 inhibitor CHF6001. It has been designed for intranasal delivery which allows CHF6001 to reach therapeutic concentration in the lung and reduce systemic side effects. CHF6001 administered twice a day decreases the levels of various inflammatory biomarkers in the sputum and reduces systematic adverse reactions in COPD patients. It seems to be well tolerated and able to reduce the allergen challenge response in asthmatic patients. CHF6001 is characterized by anti-inflammatory effects through the ability to reduce the activation of innate immune cells. This substance can effectively inhibit the release of TNF-a and oxidative burst.
GSK256066, another PDE4 inhibitor for intranasal use, was administered for 7 d in a double blind, placebo controlled, crossover study. It significantly reduced the response to allergen challenge and attenuated the decrease in lung function in patients with asthma. Moreover, it is associated with low systemic effects. A clinical trial on cilomilast was disappointing because strong side effects such as nausea, vomiting, headache, and diarrhea were observed.
## Therapeutic use of pde4 inhibitors in disorders other than ar
PDE4 inhibitors can also be used for the treatment of Th1mediated autoimmune diseases. They inhibit T-cell proliferation. Compared to GCs, the activity and effects of PDE4 inhibitors appear to be better in patients with a predisposition to atopic reactions. For skin diseases, apremilast, a selective PDE4 inhibitor administered orally, is approved for the treatment of adults with plaque psoriasis and psoriatic arthritis in Japan and United States.
In a prospective study, genetic analysis revealed significant differences between the PDE4D (rs1588265) gene in patients with no history of nasal polyposis (a type of AR) and patients with nasal polyposis, indicating the potential of using a PD4 inhibitor for the management of patients with nasal polyposis.
The effects of PDE inhibitors on ciliary beating frequency in rabbit maxillary sinus and trachea have been also investigated. The authors compared rolipram, a cAMP-specific PDE4 inhibitor with bronchodilator and anti-inflammatory activities with milrinone, a cGMP-specific PDE3 inhibitor with good bronchodilator activity but weak anti-inflammatory activity and zaprinast, a cGMP-specific PDE5 inhibitor with weak bronchodilator effects. All of them stimulated ciliary beating frequency in explants from the maxillary sinus as well as from the tracheal mucosa.
# Conclusions
AR is a common health problem with a rising incidence. Inadequate treatment may cause chronic AR and contribute to lower respiratory tract diseases such as bronchial asthma. In general, the disease can be well controlled with adequate treatment with the exception of AR, which is not mediated by the Th2-inflammatory pathway and does not respond to current available treatment modalities such as corticosteroids and antihistamines. PDE4 inhibitors can suppress a variety of inflammatory cell functions that are known to contribute to AR. However, use of systemically delivered PDE4 inhibitors, such as roflumilast, are limited by systemic side effects. Topical treatment with inhaled PDE4 inhibitors may be a viable alternative to increase tolerability and determine the maximum therapeutic potential of PDE4 inhibition in AR.
# Author contributions
VJ and HP wrote the substantial part of the manuscript. EB contributed to the epidemiology and pathophysiology of allergic rhinitis. VC contributed to the final version of the manuscript. AC conceived of the presented idea. AC and AH reviewed the article and made contributions according to their expertise.
# Funding
VEGA 1/0055/19 and APVV-17-0250.
Frontiers in Pharmacology | www.frontiersin.org July 2020 | Volume 11 | Article 1135
Conflict of Interest:The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.Copyright © 2020 Janosova, Calkovsky, Pedan, Behanova, Hajtman and Calkovska. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. |
Programmed Cell Death-Ligand 1 in Head and Neck Squamous Cell Carcinoma: Molecular Insights, Preclinical and Clinical Data, and Therapies
Citation: Meliante, P.G.; Barbato, C.; Zoccali, F.; Ralli, M.; Greco, A.; de Vincentiis, M.; Colizza, A.; Petrella, C.; Ferraguti, G.; Minni, A.; et al.
# Introduction
The programmed cell death protein 1 (PD-1) was isolated in 1992 [bib_ref] Induced expression of PD-1, a novel member of the immunoglobulin gene superfamily,..., Ishida [/bib_ref]. It is a transmembrane protein mainly expressed by immune system cells such as T lymphocytes with a crucial role in immune self-tolerance [bib_ref] Regulation and Function of the PD-L1 Checkpoint, Sun [/bib_ref] [bib_ref] A critical role for the programmed death ligand 1 in fetomaternal tolerance, Guleria [/bib_ref] [bib_ref] Facilitation of β selection and modification of positive selection in the thymus..., Nishimura [/bib_ref] [bib_ref] Development of lupus-like autoimmune diseases by disruption of the PD-1 gene encoding..., Nishimura [/bib_ref] [bib_ref] The programmed death-1 (PD-1) pathway regulates autoimmune diabetes in nonobese diabetic (NOD)..., Ansari [/bib_ref]. The interaction with its ligands, the programmed cell death protein ligand 1 and 2 (PD-L1 and PD-L2), induces T-cell inhibition [bib_ref] Regulation and Function of the PD-L1 Checkpoint, Sun [/bib_ref] [bib_ref] Differential binding properties of B7-H1 and B7-DC to programmed death-1, Youngnak [/bib_ref]. Physiologically, T-cells destroy aberrant cells, such as pathogen-infected and cancer cells, by binding the T-cell receptor (TCR) to the major histocompatibility complex (MHC) [bib_ref] Regulation and Function of the PD-L1 Checkpoint, Sun [/bib_ref]. The aberrant expression of PD-L1 by cancer cells inhibits the T-cells' cytotoxic activity, a leading mechanism of cancer immune evasion [bib_ref] Current Understanding of the Mechanisms Underlying Immune Evasion From PD-1/PD-L1 Immune Checkpoint..., Kok [/bib_ref]. Following this discovery, new drugs active on PD-1/PD-L1 have been produced, tested, and approved [bib_ref] Pembrolizumab versus methotrexate, docetaxel, or cetuximab for recurrent or metastatic head-and-neck squamous..., Cohen [/bib_ref] [bib_ref] Pembrolizumab alone or with chemotherapy versus cetuximab with chemotherapy for recurrent or..., Burtness [/bib_ref] [bib_ref] Nivolumab for Recurrent Squamous-Cell Carcinoma of the Head and Neck, Ferris [/bib_ref].
Head and neck squamous cell carcinomas (HNSCC) are the sixth most common cancer worldwide [bib_ref] Cetuximab plus chemotherapy versus chemotherapy alone in recurrent or metastatic head and..., Lang [/bib_ref] [bib_ref] Evaluation of surgical and functional outcomes of supracricoid laryngectomy and rehabilitation protocols, Ralli [/bib_ref]. The first-line regimen for recurrent or metastatic HNSCC, before the introduction of immunotherapy, was the EXTREME protocol with Cetuximab + Cisplatin or Carboplatin + 5-Fluorouracil . The results of this protocol were not optimal.
In 1992, Ishida Y. et al. isolated the PD-1 gene. The nucleotide sequence encodes for a 288 amino acids protein with two hydrophobic regions, the N-terminal acting as a signal peptide and the other in the middle as the transmembrane segment. After cleavage of the signal peptide, the mature form of the PD-1 protein is 268 amino acids, with an extracellular part of 147, a transmembrane of 27, and a cytoplasmic part of 94. The PD-1 extracellular domain is similar to that of the immunoglobulin superfamily. The cytoplasmatic tail is like that of most of the polypeptides associated with antigen receptors and Fc receptors (proteins found on the surface of certain cells binding to antibodies that are attached to infected cells or invading pathogens). The protein is expressed on the surface of antigen-stimulated T-cells [bib_ref] Induced expression of PD-1, a novel member of the immunoglobulin gene superfamily,..., Ishida [/bib_ref]. PD-1 interacts with its ligands PD-L1 and PD-L2. Despite the PD-1/PD-L2 having a 2-6-fold higher affinity than the PD-1/PD-L1 interaction, PD-L1 is the primary inhibitor of T-cells via PD-1 binding [bib_ref] Regulation and Function of the PD-L1 Checkpoint, Sun [/bib_ref] [bib_ref] Differential binding properties of B7-H1 and B7-DC to programmed death-1, Youngnak [/bib_ref].
## Pd-1 and pd1-l1 structures
PD-1 has a weight of about 55 kDa. It has 3 domains, an extracellular Ig-V-like domain (20 aa), a transmembrane domain, and a cytoplasmic domain with a signal transduction system with two tyrosine kinases . The intracellular domain also contains two motifs, an immunoreceptor tyrosine-based switch motif (ITSM) and an immunoreceptor tyrosine-based inhibitory motif (ITIM). ITSMs usually deliver inhibitory signals [bib_ref] Induced expression of PD-1, a novel member of the immunoglobulin gene superfamily,..., Ishida [/bib_ref] [bib_ref] Structural and functional analysis of the costimulatory receptor programmed death-1, Zhang [/bib_ref]. The C-terminal tyrosine is related to the Src homology region 2 domain-containing phosphatase-1 (SHP-1) and SHP-2 [bib_ref] B7-H1, a third member of the B7 family, co-stimulates T-cell proliferation and..., Dong [/bib_ref] [bib_ref] PD-1 and its ligands in tolerance and immunity, Keir [/bib_ref] [bib_ref] The PD-1/PD-L1 complex resembles the antigen-binding Fv domains of antibodies and T..., Lin [/bib_ref] [bib_ref] A family of cytokine-inducible inhibitors of signalling, Starr [/bib_ref].
## Pd-1/pd-l1 interaction
The binding between PD-1 and PD-L1 induces a PD-1 conformation change that activates the kinase cascade of the SRC family and leads to the phosphorylation of the cytoplasmic ITIM and ITSM. Which in turn activates SHP-1 and SHP-2, reducing the activation signals of T-cells. Recent studies have suggested that the main target of dephosphorylation is the CD28 co-stimulator receptor [fig_ref] Figure 1: Figure 1 [/fig_ref] [bib_ref] Genetic evidence for the involvement of the lck tyrosine kinase in signal..., Straus [/bib_ref] [bib_ref] Interactions of p59fyn and ZAP-70 with T-cell receptor activation motifs: Defining the..., Gauen [/bib_ref] [bib_ref] Influence of risk factors on stomal recurrence after total laryngectomy for laryngeal..., Wang [/bib_ref] [bib_ref] Src homology 2 domains of protein tyrosine phosphatase are associated in vitro..., Ugi [/bib_ref] [bib_ref] PD-1 inhibits T-cell receptor induced phosphorylation of the ZAP70/CD3ζ signalosome and downstream..., Sheppard [/bib_ref] [bib_ref] SHP-1 and SHP-2 associate with immunoreceptor tyrosine-based switch motif of programmed death..., Chemnitz [/bib_ref]. PD-L1 also interacts with CD80 with an inhibitory effect on activated T cells [bib_ref] Programmed death-1 ligand 1 interacts specifically with the B7-1 costimulatory molecule to..., Butte [/bib_ref]. [bib_ref] Induced expression of PD-1, a novel member of the immunoglobulin gene superfamily,..., Ishida [/bib_ref] [bib_ref] Structural and functional analysis of the costimulatory receptor programmed death-1, Zhang [/bib_ref]. The C-terminal tyrosine is related to the Src homology region 2 domain-containing phosphatase-1 (SHP-1) and SHP-2 [bib_ref] B7-H1, a third member of the B7 family, co-stimulates T-cell proliferation and..., Dong [/bib_ref] [bib_ref] PD-1 and its ligands in tolerance and immunity, Keir [/bib_ref] [bib_ref] The PD-1/PD-L1 complex resembles the antigen-binding Fv domains of antibodies and T..., Lin [/bib_ref] [bib_ref] A family of cytokine-inducible inhibitors of signalling, Starr [/bib_ref].
## Pd-1/pd-l1 interaction
The binding between PD-1 and PD-L1 induces a PD-1 conformation change that activates the kinase cascade of the SRC family and leads to the phosphorylation of the cytoplasmic ITIM and ITSM. Which in turn activates SHP-1 and SHP-2, reducing the activation signals of T-cells. Recent studies have suggested that the main target of dephosphorylation is the CD28 co-stimulator receptor [fig_ref] Figure 1: Figure 1 [/fig_ref] [bib_ref] Genetic evidence for the involvement of the lck tyrosine kinase in signal..., Straus [/bib_ref] [bib_ref] Interactions of p59fyn and ZAP-70 with T-cell receptor activation motifs: Defining the..., Gauen [/bib_ref] [bib_ref] Influence of risk factors on stomal recurrence after total laryngectomy for laryngeal..., Wang [/bib_ref] [bib_ref] Src homology 2 domains of protein tyrosine phosphatase are associated in vitro..., Ugi [/bib_ref] [bib_ref] PD-1 inhibits T-cell receptor induced phosphorylation of the ZAP70/CD3ζ signalosome and downstream..., Sheppard [/bib_ref] [bib_ref] SHP-1 and SHP-2 associate with immunoreceptor tyrosine-based switch motif of programmed death..., Chemnitz [/bib_ref]. PD-L1 also interacts with CD80 with an inhibitory effect on activated T cells [bib_ref] Programmed death-1 ligand 1 interacts specifically with the B7-1 costimulatory molecule to..., Butte [/bib_ref]. [fig_ref] Figure 1: Figure 1 [/fig_ref]. PD-1/PD-L1 mechanism of action and PD-L1 expression in cancer cells. ITIM, immunoreceptor tyrosine-based inhibitory motif; ITSM, immunoreceptor tyrosine-based switch motif; SHP-1, SHP-2, tyrosine phosphatases; ZAP70, zeta-chain associated protein kinase 70; IL-2, interleukin-2; JAK-2, Janus kinase 2; IFN-γ, interferon γ; IRFs, interferon responsive factors; LPS, lipopolysaccharide; TLR-4, toll-like receptor 4; NF-κB, nuclear factor kappa-light-chain-enhancer of activated B cells; HIF, hypoxia-inducible factors; HRE, hypoxia responsible elements; CDK5, kinase cyclin-dependent kinase 5; EGFR, epidermal growth factor receptor; ALK, anaplastic lymphoma kinase; miRNA, micro-RNA; CSN5-COP9, signalosome complex subunit 5.
## Biological role
In PD-1 gene-deficient or PD-L1 genes inhibited by antibodies in animal models, it has been observed the development of autoimmune diseases and cardiomyopathies, the onset of diabetes, alteration of thymic T-cells, and impaired feta maternal tolerance [bib_ref] Regulation and Function of the PD-L1 Checkpoint, Sun [/bib_ref] [bib_ref] A critical role for the programmed death ligand 1 in fetomaternal tolerance, Guleria [/bib_ref] [bib_ref] Facilitation of β selection and modification of positive selection in the thymus..., Nishimura [/bib_ref] [bib_ref] Development of lupus-like autoimmune diseases by disruption of the PD-1 gene encoding..., Nishimura [/bib_ref] [bib_ref] The programmed death-1 (PD-1) pathway regulates autoimmune diabetes in nonobese diabetic (NOD)..., Ansari [/bib_ref]. Furthermore, it has been shown that PD-1 inhibition modulates the cellular T response with an increase in their responsiveness [bib_ref] Enhancing virus-specific immunity in vivo by combining therapeutic vaccination and PD-L1 blockade..., Liu [/bib_ref] [bib_ref] PD-1 on dendritic cells impedes innate immunity against bacterial infection, Yao [/bib_ref] [bib_ref] The PD-1/PD-L costimulatory pathway critically affects host resistance to the pathogenic fungus..., Lázár-Molnár [/bib_ref] [bib_ref] Viral acute lower respiratory infections impair CD8+ T cells through PD-1, Erickson [/bib_ref]. PD-1/PD-L1 mechanism of action and PD-L1 expression in cancer cells. ITIM, immunoreceptor tyrosine-based inhibitory motif; ITSM, immunoreceptor tyrosine-based switch motif; SHP-1, SHP-2, tyrosine phosphatases; ZAP70, zeta-chain associated protein kinase 70; IL-2, interleukin-2; JAK-2, Janus kinase 2; IFN-γ, interferon γ; IRFs, interferon responsive factors; LPS, lipopolysaccharide; TLR-4, toll-like receptor 4; NF-κB, nuclear factor kappa-light-chain-enhancer of activated B cells; HIF, hypoxia-inducible factors; HRE, hypoxia responsible elements; CDK5, kinase cyclin-dependent kinase 5; EGFR, epidermal growth factor receptor; ALK, anaplastic lymphoma kinase; miRNA, micro-RNA; CSN5-COP9, signalosome complex subunit 5.
## Biological role
In PD-1 gene-deficient or PD-L1 genes inhibited by antibodies in animal models, it has been observed the development of autoimmune diseases and cardiomyopathies, the onset of diabetes, alteration of thymic T-cells, and impaired feta maternal tolerance [bib_ref] Regulation and Function of the PD-L1 Checkpoint, Sun [/bib_ref] [bib_ref] A critical role for the programmed death ligand 1 in fetomaternal tolerance, Guleria [/bib_ref] [bib_ref] Facilitation of β selection and modification of positive selection in the thymus..., Nishimura [/bib_ref] [bib_ref] Development of lupus-like autoimmune diseases by disruption of the PD-1 gene encoding..., Nishimura [/bib_ref] [bib_ref] The programmed death-1 (PD-1) pathway regulates autoimmune diabetes in nonobese diabetic (NOD)..., Ansari [/bib_ref]. Furthermore, it has been shown that PD-1 inhibition modulates the cellular T response with an increase in their responsiveness [bib_ref] Enhancing virus-specific immunity in vivo by combining therapeutic vaccination and PD-L1 blockade..., Liu [/bib_ref] [bib_ref] PD-1 on dendritic cells impedes innate immunity against bacterial infection, Yao [/bib_ref] [bib_ref] The PD-1/PD-L costimulatory pathway critically affects host resistance to the pathogenic fungus..., Lázár-Molnár [/bib_ref] [bib_ref] Viral acute lower respiratory infections impair CD8+ T cells through PD-1, Erickson [/bib_ref].
In T-cells, the interaction between PD-1 and PD-L1 leads to suppressive signals and dephosphorylation of the TCR pathway [bib_ref] The PD-1/PD-Ls pathway and autoimmune diseases, Dai [/bib_ref]. When PD-1 binds, its ligand causes the tyrosine phosphorylation of the cytoplasmic region and the binding of SHP-2 to the C-terminal of tyrosine in the ITSM region. Hence, SHP-2 dephosphorylates both the zetachain associated protein kinase 70 (ZAP70) and the CD28 PI3K pathway proteins, initiating inhibitory signaling downstream of the cascade. Furthermore, the inhibition of PI3K activation leads to reduced IL-2 production and glucose metabolism with the induction of T-cells energy (see [fig_ref] Figure 1: Figure 1 [/fig_ref] [bib_ref] Depletion of the programmed death-1 receptor completely reverses established clonal anergy in..., Bishop [/bib_ref] [bib_ref] PD-1-mediated suppression of IL-2 production induces CD8+ T cell anergy in vivo, Chikuma [/bib_ref] [bib_ref] Soluble PD-1 and PD-L1: Predictive and prognostic significance in cancer, Zhu [/bib_ref].
It has been observed that PD-L1 does not appear to be just a ligand. Still, its binding with PD-1 seems to initiate a cascade of signals in cancer cells that induces a decreased resistance to some forms of apoptosis, reducing the activity of both mTOR and glycolytic metabolism that protects against the cytotoxic effect of interferons I and II and cytolysis mediated by cytotoxic T lymphocytes. Although it is not yet crystal, what is the intracellular signal transduction mechanism [bib_ref] B7-H1 is a ubiquitous antiapoptotic receptor on cancer cells, Azuma [/bib_ref] [bib_ref] Metabolic Competition in the Tumor Microenvironment Is a Driver of Cancer Progression, Chang [/bib_ref].
## Pd-l1 tumor expression
In most cases, the increase in membrane PD-L1 levels gives cancer cells an advantage in evading immune defenses. But when these subjects are treated with immunotherapy at higher levels of PD-L1, there is also a greater susceptibility to drugs [bib_ref] Regulation and Function of the PD-L1 Checkpoint, Sun [/bib_ref] as observed for other types of cancer [bib_ref] Immunotherapy in the Treatment of Metastatic Melanoma: Current Knowledge and Future Directions, Ralli [/bib_ref] [bib_ref] New insights into CAR T cell-mediated killing of tumor cells, Espie [/bib_ref].
## Genetic modifications
PD-L1 and PD-L2 are only 42 kilobases away on chromosome 9 (9p24.1). The mechanisms adopted by cancer cells to increase their transcription are both amplification and translocation (Table 1, [fig_ref] Figure 1: Figure 1 [/fig_ref] [bib_ref] Regulation and Function of the PD-L1 Checkpoint, Sun [/bib_ref]. Their proximity, therefore, allows a single mutation of this type to lead to an increase in the expression of both genes. Analyzing chromosome 9, Green et al. observed that Janus kinase 2 (JAK2) is located at the 9p locus. If the modification of the DNA also involves the gene that decodes for it, a further increase in the expression of PD-L1 would be obtained thanks to the action on the IFN-γ receptor pathway (Table 1, [fig_ref] Figure 1: Figure 1 [/fig_ref] [bib_ref] Integrative analysis reveals selective 9p24.1 amplification, increased PD-1 ligand expression, and further..., Green [/bib_ref]. PD-L1 levels are also regulated by microRNAs (miRNAs) that bind the 3 'UTR of the gene. miRNAs play an inhibitory role at the post-transcriptional level. Its loss or modification was related to an increase in PD-L1 expression, and its deletion via CRISPR Cas9 led to greater stability of PD-L1 mRNA (Table 1, [fig_ref] Figure 1: Figure 1 [/fig_ref] [bib_ref] Regulation and Function of the PD-L1 Checkpoint, Sun [/bib_ref] [bib_ref] Aberrant PD-L1 expression through 3'-UTR disruption in multiple cancers, Kataoka [/bib_ref] [bib_ref] MicroRNA Landscape in Alzheimer's Disease, Cogoni [/bib_ref].
## Inflammatory signaling
Physiologically PD-1/PD-L1 serves to regulate the action of T cells, and the discovery of PD-L1 regulation by inflammatory signaling is not surprising. Tumors, in addition to determining DNA modifications as systems to increase their expression, also exploit inflammatory molecules. IFN-γ is considered the main inducer of PD-L1 expression. The binding to the receptor activates the Janus Kinases-signal transducer and activator of transcription proteins (JAK-STAT) pathway, especially STAT1, causing the expression of interferon-responsive factors (IRFs) [fig_ref] Figure 1: Figure 1 [/fig_ref] , [fig_ref] Table 1: PD-L1 tumor hyperexpression mechanisms [/fig_ref] [bib_ref] Interferon regulatory factor-1 is prerequisite to the constitutive expression and IFN-γ-induced upregulation..., Lee [/bib_ref] [bib_ref] Mechanisms of type-I-and type-II-interferon-mediated signalling, Platanias [/bib_ref]. The fact that IFN-γ is not the only one responsible for PD-L1 expression is also demonstrated by the experiments of Noguchi et al., who observed how inhibition of IFN-γ by antibodies reduces but does not eliminate the expression of PD-L1 [bib_ref] Temporally Distinct PD-L1 Expression by Tumor and Host Cells Contributes to Immune..., Noguchi [/bib_ref].
Lipopolysaccharide has also been observed to increase PD-L1 levels. Its mechanism of action occurs throughout the toll-like receptor (TLR) 4, which activates the nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB). The latter increases the expression of type I interferons [bib_ref] TLR4 signaling induces B7-H1 expression through MAPK pathways in bladder cancer cells, Qian [/bib_ref] [bib_ref] LPS/TLR4 signal transduction pathway, Lu [/bib_ref]. TLR3 also appears to increase PD-L1 expression. The same has been observed for numerous other inflammatory factors such as IL-17, IL-10, TNF-α, IL-4, IL-1b, IL-6, and IL-27 (Table 1, [fig_ref] Figure 1: Figure 1 [/fig_ref] [bib_ref] TLR3 triggering regulates PD-L1 (CD274) expression in human neuroblastoma cells, Boes [/bib_ref] [bib_ref] Inflammatory cytokines IL-17 and TNF-α up-regulate PD-L1 expression in human prostate and..., Wang [/bib_ref] [bib_ref] Synergistic effects of IL-4 and TNFα on the induction of B7-H1 in..., Quandt [/bib_ref] [bib_ref] IL-27 renders DC immunosuppressive by induction of B7-H1, Karakhanova [/bib_ref] [bib_ref] ERK/p38 MAP-kinases and PI3K are involved in the differential regulation of B7-H1..., Karakhanova [/bib_ref].
## Oncogenic pathways
The regulation of PD-L1 production is also regulated by some oncogenic transcription factors. MYC contributes to the tumorigenesis of 70% of all neoplasms. Its pharmacological inhibition significantly reduces PD-L1 expression. The mechanism of action appears to be the link between the MYC proto-oncogene-bHLH transcription factor (MYC) and the PD-L1 promoter which causes an increase in gene transcription. Consistent with this observation, MYC levels correlate with PD-L1 expression (Table 1, [fig_ref] Figure 1: Figure 1 [/fig_ref] [bib_ref] MYC expression correlates with PD-L1 expression in non-small cell lung cancer, Kim [/bib_ref].
Neoplasms often overcome the obstacle of hypoxia by producing hypoxia-inducible factors (HIFs), which promote angiogenesis. HIFs also increase PD-L1 expression by the interaction between HIF-1 A and HIF-2 B with the hypoxia response element (HRE) on the promoter of the PD-L1 gene (Table 1, [fig_ref] Figure 1: Figure 1 [/fig_ref] [bib_ref] A mechanism of hypoxia-mediated escape from adaptive immunity in cancer cells, Barsoum [/bib_ref] [bib_ref] Hypoxia-inducible factor-1α is a key regulator of metastasis in a transgenic model..., Liao [/bib_ref] [bib_ref] Renal Cell Carcinoma Programmed Death-ligand 1, a New Direct Target of Hypoxia-inducible..., Messai [/bib_ref] [bib_ref] Exploiting tumour hypoxia in cancer treatment, Brown [/bib_ref]. STAT3 is a molecule found downstream of many signal cascades and is known to be active in many neoplasms. It acts directly on the PD-L1 promoter and increases its expression [bib_ref] PD-L1 is commonly expressed and transcriptionally regulated by STAT3 and MYC in..., Atsaves [/bib_ref] [bib_ref] Activated STAT signaling in human tumors provides novel molecular targets for therapeutic..., Buettner [/bib_ref]. In addition to activating type I interferons, as mentioned above, NF-κB has a direct action on the PD-L1 gene. This transcription factor can be activated by both oncogenic mutations and inflammatory factors. One of its subunits, p65 (or RELA), directly binds the PD-L1 promoter and increases its expression [bib_ref] PD-L1 is upregulated by EBV-driven LMP1 through NF-κB pathway and correlates with..., Bi [/bib_ref].
Another suspected molecule is serine-threonine kinase cyclin-dependent kinase 5 (CDK5), known to destabilize the competitive IRF1 repressor called IRF2 and cause the induction of PD-L1 [fig_ref] Figure 1: Figure 1 [/fig_ref] , [fig_ref] Table 1: PD-L1 tumor hyperexpression mechanisms [/fig_ref] [bib_ref] Cdk5 disruption attenuates tumor PD-L1 expression and promotes antitumor immunity, Dorand [/bib_ref]. As discussed above, the role of IFN-γ in inducing PD-L1 expression has been described by many authors. Consequently, the molecules acting on IFN-γ can also be indirectly active in this sense [fig_ref] Figure 1: Figure 1 [/fig_ref] , [fig_ref] Table 1: PD-L1 tumor hyperexpression mechanisms [/fig_ref]. The AKT-mTOR cascade, located downstream of the phosphatidylinositol 3-kinase (PI3K) signaling, can be directly activated by type I and II IFN. Some authors have observed how suppression of the AKT-mTOR pathway reduces IFN-γ-induced PD-L1 expression. Moreover, the suppression of the Phosphatase and tensin homolog (PTEN) gene, which negatively regulates the signaling of PI3K-AKT, increases the expression of PD-L1 [fig_ref] Figure 1: Figure 1 [/fig_ref] , [fig_ref] Table 1: PD-L1 tumor hyperexpression mechanisms [/fig_ref] [bib_ref] Interferon-γ engages the p70 S6 kinase to regulate phosphorylation of the 40S..., Lekmine [/bib_ref] [bib_ref] Role of the Akt pathway in mRNA translation of interferon-stimulated genes, Kaur [/bib_ref] [bib_ref] PD-L1 induced by IFN-γ from tumor-associated macrophages via the JAK/STAT3 and PI3K/AKT..., Zhang [/bib_ref] [bib_ref] PTEN loss increases PD-L1 protein expression and affects the correlation between PD-L1..., Song [/bib_ref].
The increase in RAS GTPase and/or BRAF tyrosine kinase activity typical of many tumors activates the MEK-ERK pathway. Numerous studies have indicated that this pathway regulates PD-L1 expression. Again, the action is further confirmed by the observation that their inhibition reduces the transcription of PD-L1 [fig_ref] Figure 1: Figure 1 [/fig_ref] , [fig_ref] Table 1: PD-L1 tumor hyperexpression mechanisms [/fig_ref] [bib_ref] Plasma cells from multiple myeloma patients express B7-H1 (PD-L1) and increase expression..., Liu [/bib_ref] [bib_ref] Targeting the Raf-MEK-ERK mitogen-activated protein kinase cascade for the treatment of cancer, Roberts [/bib_ref]. K-RAS, epidermal growth factor receptor (EGFR) and anaplastic lymphoma kinase (ALK) can induce PD-L1 expression in cancer cells [bib_ref] KRAS mutation-induced upregulation of PD-L1 mediates immune escape in human lung adenocarcinoma, Chen [/bib_ref] [bib_ref] RAS-Mitogen-Activated Protein Kinase Signal Is Required for Enhanced PD-L1 Expression in Human..., Sumimoto [/bib_ref]. EGFR acts through mTOR and ERK-dependent mechanisms; ALK uses STAT3 and MEK-ERK [fig_ref] Figure 1: Figure 1 [/fig_ref] , [fig_ref] Table 1: PD-L1 tumor hyperexpression mechanisms [/fig_ref] [bib_ref] Upregulation of PD-L1 by EGFR Activation Mediates the Immune Escape in EGFR-Driven..., Chen [/bib_ref] [bib_ref] Control of PD-L1 Expression by Oncogenic Activation of the AKT-mTOR Pathway in..., Lastwika [/bib_ref] [bib_ref] B7-H1 expression is regulated by MEK/ERK signaling pathway in anaplastic large cell..., Yamamoto [/bib_ref].
## Mirna-mediated regulation
It is well known from the literature that miRNA acts as post-transcriptional regulators of gene expression targeting mRNA [bib_ref] Regulation and Function of the PD-L1 Checkpoint, Sun [/bib_ref]. The expression of PD-L1 is also regulated by its action. miR-513 was the first to be identified. It binds the PD-L1 3 UTR and inhibits IFN-γ induced PD-L1 expression; IFN-γ, instead, suppresses miR-513 expression. Subsequently, many inhibitory miRNAs have been isolated, such as miR-155, -34a, 142-5p, -93, -106b, -138-5p, -217 (laryngeal cancer), -200, -152, -570, -17-5p, -15a, -193a, -16, and -197 [bib_ref] MicroRNA-513 regulates B7-H1 translation and is involved in IFN-γ-induced B7-H1 expression in..., Gong [/bib_ref] [bib_ref] Cryptosporidium parvum induces B7-H1 expression in cholangiocytes by down-regulating microRNA-513, Gong [/bib_ref] [bib_ref] MicroRNA-155 induction via TNF-α and IFN-γ suppresses expression of programmed death ligand-1..., Yee [/bib_ref] [bib_ref] Tumor suppressor miR-34a targets PD-L1 and functions as a potential immunotherapeutic target..., Wang [/bib_ref] [bib_ref] miR-142-5p regulates tumor cell PD-L1 expression and enhances anti-tumor immunity, Jia [/bib_ref] [bib_ref] The miR-25-93-106b cluster regulates tumor metastasis and immune evasion via modulation of..., Cioffi [/bib_ref] [bib_ref] The tumor suppressor miR-138-5p targets PD-L1 in colorectal cancer, Zhao [/bib_ref] [bib_ref] miR-217 inhibits laryngeal cancer metastasis by repressing AEG-1 and PD-L1 expression, Miao [/bib_ref] [bib_ref] Metastasis is regulated via microRNA-200/ZEB1 axis control of tumour cell PD-L1 expression..., Chen [/bib_ref] [bib_ref] Helicobacter Pylori Promote B7-H1 Expression by Suppressing miR-152 and miR-200b in Gastric..., Xie [/bib_ref] [bib_ref] A miR-570 binding site polymorphism in the B7-H1 gene is associated with..., Wang [/bib_ref] [bib_ref] PD-L1 up-regulation in melanoma increases disease aggressiveness and is mediated through miR-17-5p, Audrito [/bib_ref] [bib_ref] Tumor Suppressor microRNAs Contribute to the Regulation of PD-L1 Expression in Malignant..., Kao [/bib_ref] [bib_ref] The clinical relevance of the miR-197/CKS1B/STAT3-mediated PD-L1 network in chemoresistant non-small-cell lung..., Fujita [/bib_ref]. By contrast, miR-20, -21, and -16 increase their expression (Table 1, [fig_ref] Figure 1: Figure 1 [/fig_ref] [bib_ref] MiR-20b, -21, and -130b inhibit PTEN expression resulting in B7-H1 over-expression in..., Zhu [/bib_ref].
## Protein level regulation
CMTM6 and 4 are the positive regulators of PD-L1 expression. The entity of the CMTM6 effect seems to vary among the tissues studied. CMTM6 binds PD-L1 and prevents ubiquitination and lysosomal degradation. CMTM6 allows PD-L1 molecules to clump together (Table 1, [fig_ref] Figure 1: Figure 1 [/fig_ref] [bib_ref] Regulation and Function of the PD-L1 Checkpoint, Sun [/bib_ref] [bib_ref] Identification of CMTM6 and CMTM4 as PD-L1 protein regulators, Mezzadra [/bib_ref] [bib_ref] CMTM6 maintains the expression of PD-L1 and regulates anti-tumour immunity, Burr [/bib_ref]. Starting from the observation that the level of PD-L1 varies during the cell cycle, Zhang et al. noted that the cascade of cyclin D-CDK4, through the phosphorylation of Speckle-type POZ protein (SPOP), elicits the ubiquitination, therefore the degradation, of PD-L1 [fig_ref] Figure 1: Figure 1 [/fig_ref] , [fig_ref] Table 1: PD-L1 tumor hyperexpression mechanisms [/fig_ref] [bib_ref] Cyclin D-CDK4 kinase destabilizes PD-L1 via cullin 3-SPOP to control cancer immune..., Zhang [/bib_ref]. The glycosylation of PD-L1 by glycogen synthase kinase 3b (GSK3b) increases its degradation and influences the interaction with PD-1 [bib_ref] Glycosylation and stabilization of programmed death ligand-1 suppresses T-cell activity, Li [/bib_ref]. NF-κB, in addition to the transcriptional regulation of PD-L1, increases PD-L1 protein levels by removing ubiquitin chains via COP9 signalosome 5 (CSN5) [bib_ref] Deubiquitination and Stabilization of PD-L1 by CSN5, Lim [/bib_ref].
## Mechanisms of pd-l1 overexpression observed in hnscc
Some of the mechanisms listed above have also already been observed in HNSCC; however, as regards the others, this does not exclude that they can be identified in head and neck tumors with future studies. Furthermore, some mechanisms have been discovered in mouse cell models and only subsequently studied in human cells; we cannot exclude that this may also happen for HNSCC.
Gene amplification is a mechanism of PD-L1 overexpression that has been observed in HNSCC. Straub et al. observed PD-L1 expression in 45% of oral cavity carcinomas and gene amplification in 19% (with high levels in 15% and low in 4%) [bib_ref] CD274/PD-L1 gene amplification and PD-L1 protein expression are common events in squamous..., Straub [/bib_ref]. IFN-γ and EGFR both use JAK2 to transmit signals of extrinsic or intrinsic origin, respectively. In HNSCC, overexpression of EGFR correlated with that of JAK2 and PD-L1. Furthermore, PD-L1 expression is dependent on that of EGFR and JAK2/STAT1, and JAK2 inhibition prevents PD-L1 upregulation [bib_ref] Identification of the Cell-Intrinsic and -Extrinsic Pathways Downstream of EGFR and IFNγ..., Concha-Benavente [/bib_ref].
miRNA-217 has been observed to have a role in esophageal, ovarian, and glioma carcinoma. Subsequently, Miao et al. have also observed its action in laryngeal carcinoma, i.e., that its expression is significantly lower in neoplastic cells than in healthy ones. Insertion of miRNA-217 into cells of the Hep2 lineage reduces their ability to migrate and invade tissues, as well as their ability to proliferate while increasing apoptosis and cell necrosis. The authors, therefore, concluded its fundamental role in inhibiting metastatic cell traits and, at the same time, that its downregulation is one of the mechanisms by which laryngeal carcinoma cells become metastatic [bib_ref] miR-217 inhibits laryngeal cancer metastasis by repressing AEG-1 and PD-L1 expression, Miao [/bib_ref]. This demonstrates that many of the mechanisms observed in vivo or in other cell populations must be studied in HNSCC.
STAT3 appears to have a very important role in regulating PD-1-PD-L1 in HNSCC. Their levels are associated, and inhibition of STAT3 downregulates that of PD-L1 [bib_ref] STAT3 Induces Immunosuppression by Upregulating PD-1/PD-L1 in HNSCC, Bu [/bib_ref]. A significant share of discoveries concerning the PD-1-PD-L1 axis is inherent in immune cells such as lymphocytes, dendritic cells, monocytes, macrophages, neutrophils, or even endothelial cells. These populations are not specific to a single neoplasm, so it is necessary to evaluate whether these findings can be generalized to all tumors in which PD-L1 is overexpressed. Among the inflammatory signaling molecules, IFN-γ is involved in the PD-1-PD-L1 axis in monocytes, neutrophils, dendritic cells, macrophages, and endothelial cells [bib_ref] B7-H1 is expressed by human endothelial cells and suppresses T cell cytokine..., Mazanet [/bib_ref] [bib_ref] Blockade of programmed death-1 ligands on dendritic cells enhances T cell activation..., Brown [/bib_ref] [bib_ref] Cutting Edge: IFN-γ Enables APC to Promote Memory Th17 and Abate Th1..., Kryczek [/bib_ref] [bib_ref] IFN-γ-stimulated neutrophils suppress lymphocyte proliferation through expression of PD-L1, De Kleijn [/bib_ref].
The involvement of INF-α and -β was also observed in the latter three cell populations. [bib_ref] Interferon Receptor Signaling Pathways Regulating PD-L1 and PD-L2 Expression, Garcia-Diaz [/bib_ref] [bib_ref] Interferon-β enhances monocyte and dendritic cell expression of B7-H1 (PD-L1), a strong..., Schreiner [/bib_ref] TLR4 has a role in the PD-1-PD-L1 axis in macrophages, monocytes, and dendritic cells, and TLR3 in dendritic and endothelial cells [bib_ref] Identification of CMTM6 and CMTM4 as PD-L1 protein regulators, Mezzadra [/bib_ref] [bib_ref] PD-L1 and PD-L2 are differentially regulated by Th1 and Th2 cells, Loke [/bib_ref] [bib_ref] NF-κB plays a key role in inducing CD274 expression in human monocytes..., Huang [/bib_ref] [bib_ref] TLR3-stimulated dendritic cells up-regulate B7-H1 expression and influence the magnitude of CD8..., Pulko [/bib_ref] [bib_ref] Unexpected protective role for Toll-like receptor 3 in the arterial wall, Cole [/bib_ref]. Interleukins have also been widely identified as related to the action of PD-1-PD-L1, such as with IL-12 in endothelial cells or IL-27 in dendritic cells [bib_ref] IL-27 renders DC immunosuppressive by induction of B7-H1, Karakhanova [/bib_ref]. It has been observed that human endothelial cells produce molecules such as PD-L1 itself, capable of activating T cells. PD-L1 expression on endothelial cells is not constitutive but induced by IFN-γ and TNF-α. Furthermore, PD-L1 also has a negative feedback function on the production of these cytokines; in fact, its blockage increases their production. Furthermore, PD-L1 expression is an active regulator of T-cell-activated cytokine synthesis [bib_ref] B7-H1 is expressed by human endothelial cells and suppresses T cell cytokine..., Mazanet [/bib_ref]. Even though monocytes are subject to the action of PD-L1, it has been observed that this action is regulated by NF-κB [bib_ref] NF-κB plays a key role in inducing CD274 expression in human monocytes..., Huang [/bib_ref].
## Immunotherapy in hnscc
The treatment of recurrent or metastatic HNSCC before the introduction of immunotherapy had unsatisfactory results. The mOS among platinum-sensitive patients treated with the EXTREME protocol was only 10.1 months. The second line of treatment had an even significantly lower mOS of 6 months with a response rate between 3 and 13% [bib_ref] Immunotherapy breakthroughs in the treatment of recurrent or metastatic head and neck..., Borel [/bib_ref]. More than 50% of advanced HNSCC had disease recurrence within 3 years. Furthermore, chemotherapy may be associated with significant toxicity and adverse effect incidence [bib_ref] Patients treated with antitumor drugs displaying neurological deficits are characterized by a..., De Santis [/bib_ref]. The introduction of immunotherapy gave new impetus to HNSCC treatment [bib_ref] Pembrolizumab versus methotrexate, docetaxel, or cetuximab for recurrent or metastatic head-and-neck squamous..., Cohen [/bib_ref] [bib_ref] Pembrolizumab alone or with chemotherapy versus cetuximab with chemotherapy for recurrent or..., Burtness [/bib_ref] [bib_ref] Nivolumab for Recurrent Squamous-Cell Carcinoma of the Head and Neck, Ferris [/bib_ref] [bib_ref] An open-label single-arm, phase II trial of zalutumumab, a human monoclonal anti-EGFR..., Saloura [/bib_ref]. Following two Phase 3 clinical trials, two molecules have been approved for treating HNSCC, Nivolumab, and Pembrolizumab [bib_ref] Pembrolizumab versus methotrexate, docetaxel, or cetuximab for recurrent or metastatic head-and-neck squamous..., Cohen [/bib_ref] [bib_ref] Pembrolizumab alone or with chemotherapy versus cetuximab with chemotherapy for recurrent or..., Burtness [/bib_ref] [bib_ref] Nivolumab for Recurrent Squamous-Cell Carcinoma of the Head and Neck, Ferris [/bib_ref] [bib_ref] Anti-PD-1 and Anti-PD-L1 in Head and Neck Cancer: A, Botticelli [/bib_ref].
Nivolumab is a fully human monoclonal antibody anti-PD-1. It was approved according to a phase III trial called Checkmate-141 that compared it to single chemotherapy agents (methotrexate, docetaxel, or cetuximab) in locally advanced HNSCC that progressed within 6 months of platinum-based therapy. The nivolumab group showed better outcomes in terms of mOS, which was 7.5 months versus 5.1 months in the standard therapy group. The overall survival (hazard ratio for death, 0.70) and 1-year survival rate were longer in the Nivolumab group than with standard therapy, 36.0 vs. 16.6, respectively [bib_ref] Nivolumab for Recurrent Squamous-Cell Carcinoma of the Head and Neck, Ferris [/bib_ref]. The Nivolumab group demonstrated better outcomes in median progression-free survival (2.0 months vs. 2.3 months), in the rate of progression-free survival at 6 months (19.7% vs. 9.9%), and in response rate (13.3 % vs. 5.8%). The 2-year survival rate was almost tripled in the Nivolumab group compared to standard therapy (16.9% vs. 6.0%), with a bigger difference in CPS PD-L1 patients. The outcomes were not influenced by the HPV status. The authors also observed a lower incidence of treatment-related adverse events in grades 3 or 4 (13.1% N 35.1%) [bib_ref] Nivolumab for Recurrent Squamous-Cell Carcinoma of the Head and Neck, Ferris [/bib_ref].
After a short time, another molecule has shown robust results, Pembrolizumab. It is an anti-PD-1 humanized monoclonal immunoglobulin. Its interaction with PD-1 inhibits binding to its ligand PD-L1 expressed by cancer cells allowing the action of T lymphocytes against neoplastic cells. The Keynote-040 was a randomized multicentric phase III study that compared Pembrolizumab with standard therapy in patients with recurrent or metastatic HNSCC that progressed after platinum-containing treatment or patients with locally advanced disease with recurrent or progressed cancer within 3-6 months of previous multimodal therapy containing platinum. They observed an mOS in the Pembrolizumab group of 8.4 months vs. 6.9 months in the standard of the care group. The mortality at the end of the observation period was 10% lower in the Pembrolizumab group (73% vs. 83%). The incidence of treatment-related adverse events was lower in the experimental group than in the standard-of-care one (13% vs. 36%). The immunotherapy treatment also had a more durable response (18.4 vs. 5 months). The differences were more significant in the PD-L1 CPS ≥ 1 sub-population [bib_ref] Pembrolizumab versus methotrexate, docetaxel, or cetuximab for recurrent or metastatic head-and-neck squamous..., Cohen [/bib_ref].
According to the Checkmate-141 and Keynote-040 trials, FDA and EMA approved Nivolumab and Pembrolizumab in the HNSCC treatment [bib_ref] Anti-PD-1 and Anti-PD-L1 in Head and Neck Cancer: A, Botticelli [/bib_ref]. Pembrolizumab was then investigated as a first-line agent alone in the Keynote-048 study. This phase III multicentric clinical trial compared three different treatment protocols, the first one with Pembrolizumab alone, the second with Pembrolizumab and Platinum or 5-FU, and the third with Cetuximab and Platinum or 5-FU. The mOS in the two groups treated with Pembrolizumab was greater than the one treated with chemotherapy alone (11.5 months vs. 10.7 months) [bib_ref] Pembrolizumab alone or with chemotherapy versus cetuximab with chemotherapy for recurrent or..., Burtness [/bib_ref]. The grade 3 and 4 treatment-related adverse events occurred in 55% of patients treated with Pembrolizumab alone, 85% of the Pembrolizumab and chemotherapy group, and 83% of the Cetuximab + chemotherapy population. The percentage of deaths among the groups was 8% for Pembrolizumab alone, 12% for Pembrolizumab + chemotherapy, and 10% for Cetuximab + chemotherapy-treated patients. The response rate was higher in patients treated with the EXTREME protocol (36% vs. 19.6% in the Pembrolizumab group). However, the duration response was greater in Pembrolizumab-treated patients (22.6 months vs. 4.5 months). The authors concluded that Pembrolizumab alone is an appropriate first-line treatment for PD-L1 positive recurrent or metastatic HNSCC, and the association between Pembrolizumab and chemotherapy is an appropriate first-line treatment for recurrent or metastatic HNSCC [bib_ref] Pembrolizumab alone or with chemotherapy versus cetuximab with chemotherapy for recurrent or..., Burtness [/bib_ref].
Considering the findings in the Keynote-048 study, Pembrolizumab was approved as a first-line agent alone or in combination with Cisplatin or 5-Fu in HNSCC patients with unresectable PD-L1 combined positive score (CPS) ≥ 1 [bib_ref] Pembrolizumab alone or with chemotherapy versus cetuximab with chemotherapy for recurrent or..., Burtness [/bib_ref] [bib_ref] Anti-PD-1 and Anti-PD-L1 in Head and Neck Cancer: A, Botticelli [/bib_ref]. The CPS, or immunohistochemistry combined positive score, for PD-L1 is calculated as 100 times the number of PD-L1 positive cancer cells, lymphocytes, and macrophages, divided by the number of viable tumor cells [bib_ref] Current Understanding of the Mechanisms Underlying Immune Evasion From PD-1/PD-L1 Immune Checkpoint..., Kok [/bib_ref]. Two other molecules, Durvalumab and Atezolizumab, have also been studied in the context of anti-PD-1-PD-L1 drugs. The results of the trials have not led to their approval in the treatment of HNSCC.
Durvalumab is an anti-PD-L1 high-affinity IgG1. In the Hawk phase II trial, authors studied its efficacy in platinum-refractory recurrent o metastatic HNSCC with high PD-L1 expression not previously treated with immunotherapy. Durvalumab demonstrated an overall response ratio of 16.2% with better efficacy in HPV-positive cancers (overall response ratio of 29.4% in the HPV+ sub-population and 10.9% in the HPV-sub-population). The mOS was 7.1 months with particularly low toxicity demonstrated by an 8% incidence rate of treatment-related adverse events with grade ≥ 3 [bib_ref] Durvalumab for recurrent or metastatic head and neck squamous cell carcinoma: Results..., Zandberg [/bib_ref]. Durvalumab was also compared to Tremelimumab alone and associated with it. Tremelimumab is an inhibitor of cytotoxic T-lymphocyte-associated antigen 4 (CTLA-4). In the 3-arm phase II trial called Condor, the authors compared the association of Durvalumab + Tremelimumab, Durvalumab alone, and Tremelimumab in patients affected by recurrent or metastatic HNSCC with low or negative CPS for PD-L1. The association group had an mOS of 7.8%, the Durvalumab alone group of 9.2%, and the Tremelimumab alone group of 1.6%. Tremelimumab was associated with a greater incidence of adverse events in grades 3 or 4. The Tremelimumab alone group and the Durvalumab + Tremelimumab group had an incidence of treatmentrelated adverse events of grade ≥ 3 of 16.9% and 15.8%, respectively, compared to the Durvalumab alone group with an incidence of 12.3%. The authors concluded that there was no advantage in mOS with Durvalumab alone or in combination with Tremelimumab vs. chemotherapy. However, they observed a median 1-year and 2-year survival with Durvalumab comparable to that obtained in the Checkmate-141 trial with Nivolumab and Keynote-040 with Pembrolizumab [bib_ref] Pembrolizumab versus methotrexate, docetaxel, or cetuximab for recurrent or metastatic head-and-neck squamous..., Cohen [/bib_ref] [bib_ref] Nivolumab for Recurrent Squamous-Cell Carcinoma of the Head and Neck, Ferris [/bib_ref] [bib_ref] Durvalumab with or without tremelimumab in patients with recurrent or metastatic head..., Ferris [/bib_ref]. Surely the choice made in the Condor study to have a population of patients with CPS for PD-L1 negative significantly influenced the outcome of the active drug against this axis. This makes comparing Pembrolizumab, Nivolumab, and Durvalumab trials difficult [bib_ref] Pembrolizumab versus methotrexate, docetaxel, or cetuximab for recurrent or metastatic head-and-neck squamous..., Cohen [/bib_ref] [bib_ref] Nivolumab for Recurrent Squamous-Cell Carcinoma of the Head and Neck, Ferris [/bib_ref] [bib_ref] Durvalumab with or without tremelimumab in patients with recurrent or metastatic head..., Ferris [/bib_ref]. Another drug active against PD-L1 is Atezolizumab, which showed a very tolerable safety profile in a phase I trial and encouraging data about its activity with mOS of 6.0 months and median progression-free survival of 2.6 months. Its action was unrelated to PD-L1 aspiration or HPV [bib_ref] Safety and clinical activity of atezolizumab in head and neck cancer: Results..., Colevas [/bib_ref].
## Immunotherapy resistance mechanisms in hnscc
Pembrolizumab and Nivolumab outperformed traditional chemotherapy in HNSCC treatment [bib_ref] Regulation and Function of the PD-L1 Checkpoint, Sun [/bib_ref] [bib_ref] Pembrolizumab alone or with chemotherapy versus cetuximab with chemotherapy for recurrent or..., Burtness [/bib_ref] [bib_ref] Nivolumab for Recurrent Squamous-Cell Carcinoma of the Head and Neck, Ferris [/bib_ref]. However, a significant proportion of patients do not respond to immunotherapy or develop resistance quickly. The identification of the mechanisms responsible for this phenomenon could help us to better identify patients eligible for therapy and to identify new therapeutic targets.
The resistance mechanisms to anti-PD-1/PD-L1 immunotherapy in HNSCC were divided into four categories: tumor cell adaptation, T-cell function and proliferation, change in the tumor microenvironment, and use of alternative immune checkpoint [bib_ref] Nivolumab for Recurrent Squamous-Cell Carcinoma of the Head and Neck, Ferris [/bib_ref]. Cancer cells respond to the selective pressure induced by immunotherapy by selecting those populations in which DNA modifications make them less sensitive to drugs. The HPV+ HNSCC cells are more prone to TRAF3 and β-2-microglobulin (β2M) mutations. The latter is part of the MHC Class I complex heavy chain, and its mutation hesitates in reducing T-cell recognition of cancer cells [bib_ref] Characteristic Hallmarks of Aging and the Impact on, Terracina [/bib_ref] [bib_ref] Therapeutic Insights from Genomic Studies of Head and Neck Squamous Cell Carcinomas, Hammerman [/bib_ref]. Some authors hypothesized that this mutation could be one of the mechanisms underlying the immune escape of cancer cells to anti-PD-1-PD-L1 immunotherapy [bib_ref] Interferon regulatory factor-1 is prerequisite to the constitutive expression and IFN-γ-induced upregulation..., Lee [/bib_ref] [bib_ref] Mechanisms of type-I-and type-II-interferon-mediated signalling, Platanias [/bib_ref]. IKZF1 is a transcription factor that induces recruitment of the immune infiltrate to tumors and increased sensitivity to PD-1 and CTLA4 inhibitors, including HNSCC. Its loss of function, on the other hand, has the opposite effect and could constitute a mechanism of resistance to immunotherapy (Table 2, [fig_ref] Figure 2: Immunotherapy resistance mechanisms in HNSCC [/fig_ref] [bib_ref] TLR4 signaling induces B7-H1 expression through MAPK pathways in bladder cancer cells, Qian [/bib_ref] [bib_ref] LPS/TLR4 signal transduction pathway, Lu [/bib_ref] [bib_ref] IKZF1 Enhances Immune Infiltrate Recruitment in Solid Tumors and Susceptibility to Immunotherapy, Chen [/bib_ref]. We know that the action of drugs against PD-1/PD-L1 is to remove the inhibition that these molecules have against the action of T lymphocytes [bib_ref] Nivolumab for Recurrent Squamous-Cell Carcinoma of the Head and Neck, Ferris [/bib_ref] , and a possible mechanism of resistance originates from the inhibition of the T cells themselves. Cyclic GMP-AMP (cGAMP) synthase (cGAS) stimulator of interferon genes (STING) activates innate immunity against infected or neoplastic cells. This molecule is suppressed by histone H3K4 lysine demethylases KDM5B and KDM5C, and activated by H3K4 methyltransferase. In HNSCC HPV+, the values of KDM5B are inversely related to those of STING, to CXCL10 (one of the interferon-induced chemokines that promotes inflammatory infiltrate in the tumor microenvironment), and to the amount of CD8+ infiltrate. CD8+ T-cell values in the T infiltrate are directly correlated with cancer survival. Consequently, high levels of KDM5B are correlated with poor prognosis indicating this molecule is a potential target for immunotherapy (Table 2, [fig_ref] Figure 2: Immunotherapy resistance mechanisms in HNSCC [/fig_ref] [bib_ref] KDM5 histone demethylases repress immune response via suppression of STING, Wu [/bib_ref]. The tumor microenvironment plays a fundamental role in immunotolerance. CD44+ stem-like cells are part of it in HNSCC immune microenvironment; they are capable of inhibiting T-lymphocytes proliferation and Th1 response while inducing immunosuppressive T-regulatory cells and myeloid-derived suppressor cells (MDSC) [bib_ref] TLR3 triggering regulates PD-L1 (CD274) expression in human neuroblastoma cells, Boes [/bib_ref] [bib_ref] Inflammatory cytokines IL-17 and TNF-α up-regulate PD-L1 expression in human prostate and..., Wang [/bib_ref] [bib_ref] Synergistic effects of IL-4 and TNFα on the induction of B7-H1 in..., Quandt [/bib_ref] [bib_ref] IL-27 renders DC immunosuppressive by induction of B7-H1, Karakhanova [/bib_ref] [bib_ref] ERK/p38 MAP-kinases and PI3K are involved in the differential regulation of B7-H1..., Karakhanova [/bib_ref] [bib_ref] Immunoregulatory properties of CD44+ cancer stem-like cells in squamous cell carcinoma of..., Chikamatsu [/bib_ref]. Their action is not dependent on the PD-1/PD-L1 pathway. Their increase in the tumor microenvironment is suspected to be an immunotherapy escape mechanism (Table 2, [fig_ref] Figure 2: Immunotherapy resistance mechanisms in HNSCC [/fig_ref] [bib_ref] Current Understanding of the Mechanisms Underlying Immune Evasion From PD-1/PD-L1 Immune Checkpoint..., Kok [/bib_ref]. We know that the action of drugs against PD-1/PD-L1 is to remove the inhibition that these molecules have against the action of T lymphocytes [bib_ref] Nivolumab for Recurrent Squamous-Cell Carcinoma of the Head and Neck, Ferris [/bib_ref] , and a possible mechanism of resistance originates from the inhibition of the T cells themselves. Cyclic GMP-AMP (cGAMP) synthase (cGAS) stimulator of interferon genes (STING) activates innate immunity against infected or neoplastic cells. This molecule is suppressed by histone H3K4 lysine demethylases KDM5B and KDM5C, and activated by H3K4 methyltransferase. In HNSCC HPV+, the values of KDM5B are inversely related to those of STING, to CXCL10 (one of the interferon-induced chemokines that promotes inflammatory infiltrate in the tumor microenvironment), and to the amount of CD8+ infiltrate. CD8+ T-cell values in the T infiltrate are directly correlated with cancer survival. Consequently, high levels of KDM5B are correlated with poor prognosis indicating this molecule is a potential target for immunotherapy (Table 2, [fig_ref] Figure 2: Immunotherapy resistance mechanisms in HNSCC [/fig_ref] [bib_ref] KDM5 histone demethylases repress immune response via suppression of STING, Wu [/bib_ref]. The tumor microenvironment plays a fundamental role in immunotolerance. CD44+ stem-like cells are part of it in HNSCC immune microenvironment; they are capable of inhibiting T-lymphocytes proliferation and Th1 response while inducing immunosuppressive T-regulatory cells and myeloid-derived suppressor cells (MDSC) [bib_ref] TLR3 triggering regulates PD-L1 (CD274) expression in human neuroblastoma cells, Boes [/bib_ref] [bib_ref] Inflammatory cytokines IL-17 and TNF-α up-regulate PD-L1 expression in human prostate and..., Wang [/bib_ref] [bib_ref] Synergistic effects of IL-4 and TNFα on the induction of B7-H1 in..., Quandt [/bib_ref] [bib_ref] IL-27 renders DC immunosuppressive by induction of B7-H1, Karakhanova [/bib_ref] [bib_ref] ERK/p38 MAP-kinases and PI3K are involved in the differential regulation of B7-H1..., Karakhanova [/bib_ref] [bib_ref] Immunoregulatory properties of CD44+ cancer stem-like cells in squamous cell carcinoma of..., Chikamatsu [/bib_ref]. Their action is not dependent on the PD-1/PD-L1 pathway. Their increase in the tumor microenvironment is suspected to be an immunotherapy escape mechanism (Table 2, [fig_ref] Figure 2: Immunotherapy resistance mechanisms in HNSCC [/fig_ref] [bib_ref] Current Understanding of the Mechanisms Underlying Immune Evasion From PD-1/PD-L1 Immune Checkpoint..., Kok [/bib_ref].
## Molecule
## Mechanism of action references
## Increase of cd44+ stem-like cells cd44+ stem-like cells inhibit t-cells and enhance
immunosuppressive T-reg cells [bib_ref] Current Understanding of the Mechanisms Underlying Immune Evasion From PD-1/PD-L1 Immune Checkpoint..., Kok [/bib_ref] CD69 sufficient state T-cells exhaustion [bib_ref] The metabolite BH4 controls T cell proliferation in autoimmunity and cancer, Cronin [/bib_ref] [bib_ref] Dendritic Cells and Programmed Death-1 Blockade: A Joint Venture to Combat Cancer, Versteven [/bib_ref]
## Gcp1 inhibition causes t-cells maturation prevention bh4
Reduction of BH4 inhibited by kineurine T-cells inhibiting
## Ido1
Increase of IDO1 reduces T-cells and inflammatory cells proliferation [bib_ref] Expression and prognostic impact of indoleamine 2,3-dioxygenase in oral squamous cell carcinomas, Laimer [/bib_ref] Arg-1 Arg-1 increase expression leads to greater degradation of L-arginine, a key nutrient for lymphocytes [bib_ref] Tryptophan metabolism as a common therapeutic target in cancer, neurodegeneration and beyond, Platten [/bib_ref] [bib_ref] The immune microenvironment of HPV-negative oral squamous cell carcinoma from never-smokers and..., Foy [/bib_ref] TGF-β Decrease dendritic cells in drainage lymph nodes and CD8+ cells. [bib_ref] Transforming growth factor-β1 immobilises dendritic cells within skin tumours and facilitates tumour..., Weber [/bib_ref] [bib_ref] TGFβ drives immune evasion in genetically reconstituted colon cancer metastasis, Tauriello [/bib_ref] [bib_ref] TGFβ attenuates tumour response to PD-L1 blockade by contributing to exclusion of..., Mariathasan [/bib_ref] [bib_ref] Targeting the TGFβ pathway with galunisertib, a TGFβRI small molecule inhibitor, promotes..., Holmgaard [/bib_ref] CD38, CD39 CD8+ cells inhibition via adenosine receptor [bib_ref] CD38-Mediated Immunosuppression as a Mechanism of Tumor Cell Escape from PD-1/PD-L1 Blockade, Chen [/bib_ref] [bib_ref] Overcoming Acquired PD-1/PD-L1 Resistance with CD38 Blockade, Mittal [/bib_ref] PMN-MDSC PMN-MDSC activates the nitric oxide pathway which inhibits the proliferation and function of T-cells [bib_ref] Overcoming Acquired PD-1/PD-L1 Resistance with CD38 Blockade, Mittal [/bib_ref] [bib_ref] Human Papilloma Virus Specific Immunogenicity and Dysfunction of CD8+ T Cells in..., Krishna [/bib_ref] [bib_ref] Inhibition of the adenosine A2a receptor modulates expression of T cell coinhibitory..., Leone [/bib_ref] [bib_ref] Specific blockade CD73 alters the "exhausted" phenotype of T cells in head..., Deng [/bib_ref] NRLP3 NRLP3 activation increases MDSCs, T-regs, and TAMs, and reduces IL-1 β [bib_ref] Blockage of the NLRP3 inflammasome by MCC950 improves anti-tumor immune responses in..., Chen [/bib_ref] STAT-pathway and cytokines Alteration of STAT-pathway leads to a dendritic cell loss of function [bib_ref] Dendritic Cells and Programmed Death-1 Blockade: A Joint Venture to Combat Cancer, Versteven [/bib_ref] [bib_ref] A natural killer-dendritic cell axis defines checkpoint therapy-responsive tumor microenvironments, Barry [/bib_ref] [bib_ref] Molecular regulation of dendritic cell development and function in homeostasis, inflammation, and..., Chrisikos [/bib_ref] LAG3, pathway of T-cell immunoglobulin, ITIM domain, TIM-3, VISTA Alternative immune checkpoint [bib_ref] Current Understanding of the Mechanisms Underlying Immune Evasion From PD-1/PD-L1 Immune Checkpoint..., Kok [/bib_ref] [bib_ref] Expression of VISTA correlated with immunosuppression and synergized with CD8 to predict..., Wu [/bib_ref] [bib_ref] Differential contribution of three immune checkpoint (VISTA, CTLA-4, PD-1) pathways to antitumor..., Kondo [/bib_ref] [bib_ref] Adaptive resistance to therapeutic PD-1 blockade is associated with upregulation of alternative..., Koyama [/bib_ref] [bib_ref] Phase Ib Study of Immune Biomarker Modulation with Neoadjuvant Cetuximab and TLR8..., Shayan [/bib_ref] [bib_ref] The head and neck cancer immune landscape and its immunotherapeutic implications, Mandal [/bib_ref] [bib_ref] LAG3 (CD223) as a cancer immunotherapy target, Andrews [/bib_ref] ATRA and IFB-β ATRA and IFB-β increase CD38 production via CD30-CD203a-CD73 axis. CD38 transforms NAD + and NADP + into cyclic ribose ADP (zADPR), NAADP, and ADPR, which act on calcium signaling [bib_ref] Current Understanding of the Mechanisms Underlying Immune Evasion From PD-1/PD-L1 Immune Checkpoint..., Kok [/bib_ref] [bib_ref] Inhibition of the adenosine A2a receptor modulates expression of T cell coinhibitory..., Leone [/bib_ref] [bib_ref] Specific blockade CD73 alters the "exhausted" phenotype of T cells in head..., Deng [/bib_ref] [bib_ref] Phase Ib Study of Immune Biomarker Modulation with Neoadjuvant Cetuximab and TLR8..., Shayan [/bib_ref] [bib_ref] Evolution and Function of the ADP Ribosyl Cyclase/CD38 Gene Family in Physiology..., Malavasi [/bib_ref] [bib_ref] Ectonucleotidases CD39 and CD73 on OvCA cells are potent adenosine-generating enzymes responsible..., Häusler [/bib_ref] CD73 dephosphorylates extracellular AMP which leads to the production of adenosine. Adenosine binds to the A2a and A2b receptors of T and NK lymphocytes, neutrophils, dendritic cells, and macrophages with an immunosuppressive action [bib_ref] CD38-Mediated Immunosuppression as a Mechanism of Tumor Cell Escape from PD-1/PD-L1 Blockade, Chen [/bib_ref] [bib_ref] Overcoming Acquired PD-1/PD-L1 Resistance with CD38 Blockade, Mittal [/bib_ref] [bib_ref] Human Papilloma Virus Specific Immunogenicity and Dysfunction of CD8+ T Cells in..., Krishna [/bib_ref] [bib_ref] Ectonucleotidases CD39 and CD73 on OvCA cells are potent adenosine-generating enzymes responsible..., Häusler [/bib_ref] [bib_ref] Targeting Cancer-Derived Adenosine: New Therapeutic Approaches, Young [/bib_ref] HLA, β-2-macroglobulin and TRAF3 mutation common in HPV+ HNSCC, whereas they are found in less than 10% of HPV− cancers. HPV antigens also enhance cytotoxic T-lymphocytes dysregulation. [bib_ref] Therapeutic Insights from Genomic Studies of Head and Neck Squamous Cell Carcinomas, Hammerman [/bib_ref] The study of the tumor microenvironment has led to the discovery of numerous molecules involved in resistance to immunotherapy with action on T-cells [bib_ref] The metabolite BH4 controls T cell proliferation in autoimmunity and cancer, Cronin [/bib_ref] [bib_ref] Dendritic Cells and Programmed Death-1 Blockade: A Joint Venture to Combat Cancer, Versteven [/bib_ref]. It has been observed that the increase in the expression of indoleamine 2-3-dioxygenase-1 (IDO1) reduces the proliferation not only of T-cells but also of other elements of the inflammatory infiltrate in oral squamous cell carcinoma [bib_ref] Expression and prognostic impact of indoleamine 2,3-dioxygenase in oral squamous cell carcinomas, Laimer [/bib_ref]. These neoplastic cells, in addition to acting on the signals that regulate functioning and proliferation, can also act on the nutrients that the cells of the immune system need to survive. A potential mechanism observed is the increase in the expression of arginase-1 (Arg -1) by cancer cells which leads to greater degradation of L-arginine, which is a key nutrient for T and NK cells (Table 2, [fig_ref] Figure 2: Immunotherapy resistance mechanisms in HNSCC [/fig_ref] [bib_ref] Tryptophan metabolism as a common therapeutic target in cancer, neurodegeneration and beyond, Platten [/bib_ref] [bib_ref] The immune microenvironment of HPV-negative oral squamous cell carcinoma from never-smokers and..., Foy [/bib_ref].
The production of TGF-β has a role in both intracellular and extracellular environments. Its secretion by cancer-associated fibroblasts inhibits CD8+ T cells and decreases the dendritic cells in draining lymph nodes [bib_ref] Transforming growth factor-β1 immobilises dendritic cells within skin tumours and facilitates tumour..., Weber [/bib_ref] [bib_ref] TGFβ drives immune evasion in genetically reconstituted colon cancer metastasis, Tauriello [/bib_ref] [bib_ref] TGFβ attenuates tumour response to PD-L1 blockade by contributing to exclusion of..., Mariathasan [/bib_ref] [bib_ref] Targeting the TGFβ pathway with galunisertib, a TGFβRI small molecule inhibitor, promotes..., Holmgaard [/bib_ref]. CD8+ lymphocytes are also inhibited by CD38 via the adenosine receptor signaling and CD39 [bib_ref] CD38-Mediated Immunosuppression as a Mechanism of Tumor Cell Escape from PD-1/PD-L1 Blockade, Chen [/bib_ref] [bib_ref] Overcoming Acquired PD-1/PD-L1 Resistance with CD38 Blockade, Mittal [/bib_ref]. In addition to direct inhibition of their action, the reduction of T lymphocyte activity is also induced through the activation of polymorphonuclear myeloid-derived suppressor cells (PMN-MDSC) which, through the nitric oxide pathway, inhibits the proliferation and function of T-cells [bib_ref] Overcoming Acquired PD-1/PD-L1 Resistance with CD38 Blockade, Mittal [/bib_ref] [bib_ref] Human Papilloma Virus Specific Immunogenicity and Dysfunction of CD8+ T Cells in..., Krishna [/bib_ref] [bib_ref] Inhibition of the adenosine A2a receptor modulates expression of T cell coinhibitory..., Leone [/bib_ref] [bib_ref] Specific blockade CD73 alters the "exhausted" phenotype of T cells in head..., Deng [/bib_ref]. The activation of the nucleotide-binding domain leucine-rich repeat and of the pyrin domain containing receptor 3 (NRLP3) inflamed induced an increase in MDSCs, T-regs, and TAMs, and a reduction of IL-1β synthesis, leading to an immunosuppressive effect [bib_ref] Blockage of the NLRP3 inflammasome by MCC950 improves anti-tumor immune responses in..., Chen [/bib_ref]. The action of immunotherapy escape mechanisms is also active against dendritic cells, alteration in STAT-mediated pathways, and cytokines production leading to a loss of function of dendritic cells [fig_ref] Table 2: Immunotherapy resistance mechanisms in HNSCC [/fig_ref] [bib_ref] Dendritic Cells and Programmed Death-1 Blockade: A Joint Venture to Combat Cancer, Versteven [/bib_ref] [bib_ref] A natural killer-dendritic cell axis defines checkpoint therapy-responsive tumor microenvironments, Barry [/bib_ref] [bib_ref] Molecular regulation of dendritic cell development and function in homeostasis, inflammation, and..., Chrisikos [/bib_ref] [bib_ref] The role of cytokines in head and neck squamous cell carcinoma: A..., Ralli [/bib_ref].
PD-L1 is not the only immune checkpoint in HNSCC; indeed, activating alternative tolerance mechanisms leads to cancer cells' immune escape. The lymphocyte activation gene-3 (LAG3), the pathway of the T-cell immunoglobulin and the ITIM domain, the T-cell immunoglobulin mucin-3 (TIM-3), and the V tomorrow-containing IG suppressor of T-cell activation (VISTA) are some examples of the alternative immune checkpoint that cancer cells use [bib_ref] Current Understanding of the Mechanisms Underlying Immune Evasion From PD-1/PD-L1 Immune Checkpoint..., Kok [/bib_ref] [bib_ref] Expression of VISTA correlated with immunosuppression and synergized with CD8 to predict..., Wu [/bib_ref] [bib_ref] Differential contribution of three immune checkpoint (VISTA, CTLA-4, PD-1) pathways to antitumor..., Kondo [/bib_ref] [bib_ref] Adaptive resistance to therapeutic PD-1 blockade is associated with upregulation of alternative..., Koyama [/bib_ref] [bib_ref] Phase Ib Study of Immune Biomarker Modulation with Neoadjuvant Cetuximab and TLR8..., Shayan [/bib_ref] [bib_ref] The head and neck cancer immune landscape and its immunotherapeutic implications, Mandal [/bib_ref] [bib_ref] LAG3 (CD223) as a cancer immunotherapy target, Andrews [/bib_ref]. Under immunotherapy selective pressure, neoplastic cells produce all-trans retinoic acid (ATRA) and IFN-β, which increase CD38 production via the CD30-CD203a-CD73 axis. CD38 transforms NAD(+) and NADP(+) into cyclic ribose ADP (zADPR), NAADP, and ADPR, which act on calcium signaling [bib_ref] Current Understanding of the Mechanisms Underlying Immune Evasion From PD-1/PD-L1 Immune Checkpoint..., Kok [/bib_ref] [bib_ref] Inhibition of the adenosine A2a receptor modulates expression of T cell coinhibitory..., Leone [/bib_ref] [bib_ref] Specific blockade CD73 alters the "exhausted" phenotype of T cells in head..., Deng [/bib_ref] [bib_ref] Phase Ib Study of Immune Biomarker Modulation with Neoadjuvant Cetuximab and TLR8..., Shayan [/bib_ref] [bib_ref] Evolution and Function of the ADP Ribosyl Cyclase/CD38 Gene Family in Physiology..., Malavasi [/bib_ref] [bib_ref] Ectonucleotidases CD39 and CD73 on OvCA cells are potent adenosine-generating enzymes responsible..., Häusler [/bib_ref]. Furthermore, CD73 dephosphorylates extracellular AMP which leads to the production of adenosine. Adenosine binds to the A2a and A2b receptors of T and NK lymphocytes, neutrophils, dendritic cells, and macrophages with an immunosuppressive action [bib_ref] CD38-Mediated Immunosuppression as a Mechanism of Tumor Cell Escape from PD-1/PD-L1 Blockade, Chen [/bib_ref] [bib_ref] Overcoming Acquired PD-1/PD-L1 Resistance with CD38 Blockade, Mittal [/bib_ref] [bib_ref] Human Papilloma Virus Specific Immunogenicity and Dysfunction of CD8+ T Cells in..., Krishna [/bib_ref] [bib_ref] Ectonucleotidases CD39 and CD73 on OvCA cells are potent adenosine-generating enzymes responsible..., Häusler [/bib_ref] [bib_ref] Targeting Cancer-Derived Adenosine: New Therapeutic Approaches, Young [/bib_ref]. HLA, β-2-macroglobulin, and TRAF3 mutation are common in HPV+ HNSCC, whereas they are found in less than 10% of HPV cancers [bib_ref] Therapeutic Insights from Genomic Studies of Head and Neck Squamous Cell Carcinomas, Hammerman [/bib_ref]. HPV antigens also enhance cytotoxic T-lymphocytes dysregulation (Table 2, [fig_ref] Figure 2: Immunotherapy resistance mechanisms in HNSCC [/fig_ref] [bib_ref] Human Papilloma Virus Specific Immunogenicity and Dysfunction of CD8+ T Cells in..., Krishna [/bib_ref]. Individuals with locally advanced HNSCC have a high incidence of local infections. Therefore, they are often treated with antibiotic therapy. The consequent alteration of the microbiome seems to be related to PD-1/PD-L1 drug resistance, although the mechanism by which dysbiosis leads to it is still unclear [bib_ref] Gut microbiome influences efficacy of PD-1-based immunotherapy against epithelial tumors, Routy [/bib_ref] [bib_ref] Antibiotics are associated with attenuated efficacy of anti-PD-1/PD-L1 therapies in Chinese patients..., Zhao [/bib_ref].
# Discussion
## Biomarkers of immunotherapy response in hnscc
Although immunotherapy outcome results were encouraging, ∼60% of patients with recurrent or metastatic HNSCC do not respond to anti-PD-1/PD-L1 therapy. [bib_ref] Current Understanding of the Mechanisms Underlying Immune Evasion From PD-1/PD-L1 Immune Checkpoint..., Kok [/bib_ref] Identifying molecular markers that, together with the CPS for PD-L1, allow us to predict the possible response to therapy early could help us in the selection of patients. As well as the identification of molecules that allow us to predict the response to treatment during the course of the same could help us manage therapies. IFN-γ upregulates PD-L1 and PD-L2 to downregulate the cytotoxic response. IFN-γ active signaling is associated with anti-PD-L1 therapy response, and IFN-γ-related mRNA profile predicts clinical response to PD-1 blockade in HNSCC [bib_ref] IFN-γ-related mRNA profile predicts clinical response to PD-1 blockade, Ayers [/bib_ref].
Hypoxia within the tumor microenvironment leads to an "immune desert", a decrease in immune cells that is a further immune evasion mechanism. The resulting paucity of T-cells explains the poor response to PD-1/PD-L1 immunotherapy. The biomarkers that can highlight this phenomenon in HNSCC are hypoxia-inducible factor-1α (HIF-1α) and its signaling [bib_ref] A mechanism of hypoxia-mediated escape from adaptive immunity in cancer cells, Barsoum [/bib_ref] [bib_ref] Development and validation of a combined hypoxia and immune prognostic classifier for..., Brooks [/bib_ref] [bib_ref] Loss of function JAK1 mutations occur at high frequency in cancers with..., Albacker [/bib_ref]. The tumor microenvironment and the cells contained in it might have a role in immunotherapy resistance. Intriguingly, head and neck cancer-associated fibroblasts (HNCAF) modulate the immune response to HNSCC and could be used as potential immunotherapy response biomarkers [bib_ref] Immunostimulatory Cancer-Associated Fibroblast Subpopulations Can Predict Immunotherapy Response in Head and Neck..., Obradovic [/bib_ref].
Usually, the higher the number of DNA mutations, the more neoantigens are presented to the antigen-presenting cells (APCs). The more the cancer cells are susceptible to cytotoxic killing by T-lymphocytes. A defect in DNA mismatch repair genes (hMLH1 and hMSH2) causes the accumulation of DNA mutations and microsatellite instability. It leads to a high tumor mutation burden (TMB-high). It has been observed that higher TMB predicted response to anti-PD-1/PD-L1 in head and neck cancers [bib_ref] Frameshift events predict anti-PD-1/L1 response in head and neck cancer, Hanna [/bib_ref]. The tumor cell mutation burden is also a predictor of patient survival in HNSCC when measured as "peripheral blood tumor cell mutation burden" (bTMB). bTMB, TMB, and inflammatory biomarkers are considered independent predictors of Pembrolizumab efficacy [bib_ref] Plasma-based tumor mutational burden (bTMB) as predictor for survival in phase III..., Li [/bib_ref] [bib_ref] Pan-tumor genomic biomarkers for PD-1 checkpoint blockade-based immunotherapy, Cristescu [/bib_ref].
## Future perspectives in hnscc
Despite the improvement that immunotherapy has brought to treating HNSCC, there is a significant percentage of patients who have no long-term benefit. Currently, we perform the patient selection by evaluating the PD-L1 expression only. A wider marker-based patient selection, also based on other molecules, could help define the best therapeutic approach for each patient [bib_ref] Anti-PD-1 and Anti-PD-L1 in Head and Neck Cancer: A, Botticelli [/bib_ref]. In this way, a personalized treatment protocol could be identified for each individual's oncological profile. Indeed, PD-L1 levels have a predictive value of the response to immunotherapy. In some patients, although the CPS for PD-L1 is greater than 1, there is no response to therapy. Seiwert et al. proposed measuring, together with the expression of PD-L1, the levels of IFN-γ in HNSCC, as they directly influence the expression of PD-L1, and, likewise, to ensure that PD-L1 expression is related to T cell activity and not inflammation of the tumor microenvironment [bib_ref] Safety and clinical activity of pembrolizumab for treatment of recurrent or metastatic..., Seiwert [/bib_ref].
Numerous additional factors seem to influence the response to PD-1/PD-L1 immunotherapy. One of these is the amount of non-synonymous DNA mutations whose increase causes a greater presence of neoantigens with a consequent greater cellular T response. Tumors with a higher rate of these mutations appear to have a greater susceptibility to PD-1 and PD-L1 inhibitors [bib_ref] Profound Immunotherapy Response in Mismatch Repair-Deficient Breast Cancer, Kok [/bib_ref].
Another advanced hypothesis was to use CMTM6 as a target, together with PD-L1, to reduce the expression of PD-L1 induced by IFN-γ [bib_ref] Regulation and Function of the PD-L1 Checkpoint, Sun [/bib_ref]. Despite the enthusiasm regarding Pembrolizumab and Nivolumab, most HNSCC patients will not have longterm benefits from immunotherapy treatment. Several trials test drug associations with multiple immunotherapies or a combination of them with traditional chemotherapy [bib_ref] Immunotherapy for head and neck squamous cell carcinoma, Lecocq [/bib_ref]. The rationale behind the association of immunotherapy and traditional chemotherapy is the observation that the latter makes cancer cells more recognizable by the immune system. By combining them with the anti-PD-1/PD-L1 drugs, we obtain the combination of two treatments. One that makes cells more visible to the immune system and one that takes away the inhibition of T-cells. This effect was also observed with lower doses of cytotoxic drugs (such as Cisplatin). Reducing the amount of drug administered induces fewer adverse events, especially bone marrow hematopoiesis inhibition [bib_ref] Metronomic chemotherapy: Changing the paradigm that more is better, Scharovsky [/bib_ref] [bib_ref] Immune-based mechanisms of cytotoxic chemotherapy: Implications for the design of novel and..., Bracci [/bib_ref] [bib_ref] Evidence that nerve growth factor promotes the recovery of peripheral neuropathy induced..., Aloe [/bib_ref]. The TPextreme trial studied the association between taxane chemotherapy followed by a secondline treatment with immunotherapy in R/M HNSCC. The observed mOS was 21.9 months. These results have not been compared with those of immunotherapy alone [bib_ref] TPExtreme randomized trial: Quality of Life (QoL) and survival according to second-line..., Guigay [/bib_ref].
Another possible target that appears to be correlated with surviving immunotherapy is KDM5B. It regulates the levels of STING, CXCL10, and therefore inflammatory infiltrates of the tumor microenvironment, in particular CD8+ T lymphocytes, which in turn correlate with survival. For this reason, some authors have indicated KDM5B as a potential target for immunotherapy in HNSCC [bib_ref] KDM5 histone demethylases repress immune response via suppression of STING, Wu [/bib_ref].
Obviously, not all associations have shown results superior to the treatments already approved. According to the Eagle trial, it was observed that there was no synergy between the two molecules in R/M HNSCC; probably because Tremelimumab is an IgG2 that does not induce cell death via an antibody-dependent mechanism, while NK cells are the most numerous lymphocytes in the HNSCC tumor microenvironment [bib_ref] Immunotherapy breakthroughs in the treatment of recurrent or metastatic head and neck..., Borel [/bib_ref]. In any case, the combination of anti-PD-L1 and anti-CTLA4 drugs is still under study. We are waiting for the Checkmate-651, Checkmate-714, and Kestrel trials. The former compares the association between Nivolumab (anti-PD-1) and Ipilimumab (anti-CTLA4) vs. standard therapy (EXTREME protocol). Checkmate-714 compares the same drug association (Nivolumab and Ipilimumab) with Nivolumab alone in platinum-sensitive and platinum-resistant diseases. The Kestrel trial is a tree-arm study in HNSCC platinum-sensitive patients that evaluates the association between Tremelimumab (anti-CTLA4) and Durvalumab (anti-PD-L1) vs. Durvalumab alone vs. EXTREME protocol. Obviously, anti-CTLA4 molecules are not the only ones to be studied in association with anti-PD-L1 drugs. The vascular endothelial growth factor (VEGF) is also an immunosuppressive molecule. The association between its inhibitors and anti-PD-1 therapies is under examination in the LEAP-010 trial. The results seem encouraging, with a superior anti-tumor activity compared with single molecules [bib_ref] Phase III LEAP-010 study: First-line pembrolizumab with or without lenvatinib in recurrent/metastatic..., Siu [/bib_ref] [bib_ref] Lenvatinib mesilate (LEN) enhanced antitumor activity of a PD-1 blockade agent by..., Kato [/bib_ref] [bib_ref] Cancer stem cells-driven tumor growth and immune escape: The Janus face of..., Triaca [/bib_ref].
## New molecules in hnscc
Several new molecules are currently in various stages of study. Monalizumab is an anti-NKG2A receptor humanized antibody. This receptor is expressed on CD8+ and NK lymphocytes. The Upstream trial is currently evaluating its efficacy alone and in combination with Durvalumab vs. standard care protocols in R/M HNSCC [bib_ref] A phase II study of monalizumab in patients with recurrent/metastatic squamous cell..., Galot [/bib_ref]. The Interlink-1 study analyzes the outcome of the association between Monalizumab and Cetuximab [bib_ref] Assessment of Efficacy and Safety of Monalizumab Plus Cetuximab Compared to Placebo..., Identifier [/bib_ref]. GSK609 is an anti-T-cell inducible co-stimulatory receptor (ICOS) monoclonal antibody. ICOS is involved in T-cells proliferation, differentiation, and survival. We are evaluating the results of a phase III trial INDUCE-3 which compares the association of Pembrolizumab, Platinum/5-FU, and GSK609 vs. Pembrolizumab and Platinum/5-FU.
## Hnscc therapy
The mOS observed with Pembrolizumab-based protocols is greater than standard chemotherapy in patients with recurrent or metastatic HNSCC. At higher PD-L1, CPS values correspond to a major treatment response [bib_ref] Pembrolizumab alone or with chemotherapy versus cetuximab with chemotherapy for recurrent or..., Burtness [/bib_ref]. Immunotherapy also has a better safety profile, with 2.7 times less incidence of treatment-related adverse events. Furthermore, it has greater durability [bib_ref] Pembrolizumab versus methotrexate, docetaxel, or cetuximab for recurrent or metastatic head-and-neck squamous..., Cohen [/bib_ref]. The two approved drugs in HNSCC treatment, Pembrolizumab and Nivolumab, had almost equal 1-year survival rates (37% and 36%). However, this result is achieved by applying protocols with very different doses. The Pembrolizumab posology was 75 mg/m 2 every 3 weeks in the Keynote-040 trial, and the Nivolumab one was 30-40 mg/week in the Checkmate-141 trial. Furthermore, the populations of the two trials had slightly different eligibility criteria. The Keynote-040 considered patients with disease progression between 3 and 6 months, and the Checkmate-141 patients with disease progressed within 6 months of platinum-based therapy. Hence the population in this study was also composed of patients that never responded to therapy, usually considered with a worse prognosis [bib_ref] Pembrolizumab versus methotrexate, docetaxel, or cetuximab for recurrent or metastatic head-and-neck squamous..., Cohen [/bib_ref] [bib_ref] Nivolumab for Recurrent Squamous-Cell Carcinoma of the Head and Neck, Ferris [/bib_ref].
The populations of Checkmate-141 and Keynote-040 trials also differed in PD-L1 CPS values. In Keynote-040, over 75% of patients had CPS for PD-L1 ≥ 1; in Checkmate-141, it was 72%. This small difference could have influenced outcomes because it is well-known from the literature that a value greater than this threshold significantly influences the immunotherapy response [bib_ref] Pembrolizumab versus methotrexate, docetaxel, or cetuximab for recurrent or metastatic head-and-neck squamous..., Cohen [/bib_ref] [bib_ref] Pembrolizumab alone or with chemotherapy versus cetuximab with chemotherapy for recurrent or..., Burtness [/bib_ref] [bib_ref] Nivolumab for Recurrent Squamous-Cell Carcinoma of the Head and Neck, Ferris [/bib_ref]. In the Eagle trial, there was an over-performance of the standard of care group compared to the Pembrolizumab and Nivolumab trials. There was also an over-mortality in the initial period of the immunotherapy administration [bib_ref] Immunotherapy breakthroughs in the treatment of recurrent or metastatic head and neck..., Borel [/bib_ref]. The over-mortality and the high rate of progressive disease have also been observed in Keynote-048 trials. The Pembrolizumab alone survival curve was below chemotherapy one over the first eight months. Then they crossed with the stabilization of a better outcome for immunotherapy-treated patients. This high rate of progressive disease at the beginning of the treatment could be explained by the "hyper-progression" phenomenon, a faster growth of cancer cells after immunotherapy initiation [bib_ref] Immunotherapy breakthroughs in the treatment of recurrent or metastatic head and neck..., Borel [/bib_ref]. The excess of early deaths in Pembrolizumab alone patients was eliminated in the Keynote-048 study with the addition of chemotherapy to it. However, it comes at a price of higher toxicity. This association has a significant advantage only in CPS PD-L1 ≥ 1 compared to chemotherapy only [bib_ref] Pembrolizumab alone or with chemotherapy versus cetuximab with chemotherapy for recurrent or..., Burtness [/bib_ref] [bib_ref] Immunotherapy breakthroughs in the treatment of recurrent or metastatic head and neck..., Borel [/bib_ref].
The findings of the Keynote-048 trial led to the approval of Pembrolizumab alone for CPS ≥ 1 patients, and in association with chemotherapy for any CPS. The EMA approved it alone or in association only with CPS ≥ 1 patients. It has been observed that early Pembrolizumab-based therapy in PD-L1 CPS ≥ 1 may sensitize the tumor to subsequent therapy thanks to microenvironment modification. The outcome of therapy given after immune checkpoint inhibition is greater than predicted historically in patients who both respond and do not respond to checkpoint inhibition [bib_ref] Response rates to single-agent chemotherapy after exposure to immune checkpoint inhibitors in..., Schvartsman [/bib_ref] [bib_ref] In the immuno-oncology era, is anti-PD-1 or anti-PD-L1 immunotherapy modifying the sensitivity..., Aspeslagh [/bib_ref] [bib_ref] Increased Response Rates to Salvage Chemotherapy Administered after PD-1/PD-L1 Inhibitors in Patients..., Park [/bib_ref] [bib_ref] Response Rate to Chemotherapy After Immune Checkpoint Inhibition in Metastatic Urothelial Cancer, Szabados [/bib_ref] [bib_ref] Response to salvage chemotherapy after progression on immune checkpoint inhibitors in patients..., Saleh [/bib_ref]. Anti-PD-L1 therapy has higher efficacy in male and smoker patients. A possible explanation for this finding is that smoking induces greater immunogenicity by increasing genetic mutations. It makes neoplastic cells more recognizable by T cells whose action is not stopped for the immunotherapy inhibition of PD-L1 [bib_ref] Clinical and molecular characteristics associated with the efficacy of PD-1/PD-L1 inhibitors for..., Weng [/bib_ref] [bib_ref] The immune response-related mutational signatures and driver genes in non-small-cell lung cancer, Chen [/bib_ref] [bib_ref] Mutational signatures associated with tobacco smoking in human cancer, Alexandrov [/bib_ref] [bib_ref] A dormant TIL phenotype defines non-small cell lung carcinomas sensitive to immune..., Gettinger [/bib_ref] [bib_ref] Overall survival and long-term safety of nivolumab (anti-programmed death 1 antibody, BMS-936558,..., Gettinger [/bib_ref] [bib_ref] Smokers or non-smokers: Who benefits more from immune checkpoint inhibitors in treatment..., Mo [/bib_ref] [bib_ref] Tumor B7-H3 (CD276) expression and smoking history in relation to lung adenocarcinoma..., Inamura [/bib_ref]. Clinical staging of HNSCC plays an important role in immunotherapy efficacy. Indeed, Botticelli et al. observed that anti-PD-1 is more effective in metastatic diseases and anti-PD-L1 in recurrent ones. The possible explanation for this finding is the systemic effect given by PD-1 inhibition that strikes every circulating lymphocyte. The metastases can be attacked by active T-cells that are no longer stoppable by cancer PD-L1. The anti-PD-L1 drugs showed greater local efficacy in cancer with lower heterogeneity [bib_ref] Anti-PD-1 and Anti-PD-L1 in Head and Neck Cancer: A, Botticelli [/bib_ref].
## Adverse events in hnscc immunotherapy
The most common side effect of anti-PD-L1 agents is autoimmune endocrinopathies [bib_ref] Nivolumab for Recurrent Squamous-Cell Carcinoma of the Head and Neck, Ferris [/bib_ref]. Particularly, the most commons were fatigue and hypothyroidism when associated with Pembrolizumab alone therapy. The association with chemotherapy and the association of Cetuximab and chemotherapy had a higher incidence of anemia and nausea [bib_ref] Pembrolizumab alone or with chemotherapy versus cetuximab with chemotherapy for recurrent or..., Burtness [/bib_ref]. Bleeding is sometimes a side effect of immunotherapy. The bleeding incidence for Pembrolizumab and the association of Pembrolizumab and chemotherapy were 7% and 9%, respectively; the Cetuximab and chemotherapy group had a bleeding incidence of only 5% [bib_ref] Pembrolizumab alone or with chemotherapy versus cetuximab with chemotherapy for recurrent or..., Burtness [/bib_ref].
# Conclusions
The introduction of anti-PD-L1 immunotherapy has significantly improved cancer pharmacotherapy. But the way to achieve satisfactory results is still impervious. The prospects regarding the new molecules, the new association protocols, and the new biomarkers are encouraging. Thanks to them, it is possible to move in the direction of personalized medicine, in which the choice of therapy is not based on the patient's pathology but on the molecular characteristics of each patient. To date, immunotherapy constitutes a fundamental weapon in the oncology of the cervical and facial area and beyond.
[fig] Figure 1: Figure 1. PD-1/PD-L1 mechanism of action and PD-L1 expression in cancer cells. ITIM, immunoreceptor tyrosine-based inhibitory motif; ITSM, immunoreceptor tyrosine-based switch motif; SHP-1, SHP-2, tyrosine phosphatases; ZAP70, zeta-chain associated protein kinase 70; IL-2, interleukin-2; JAK-2, Janus kinase 2; IFN-γ, interferon γ; IRFs, interferon responsive factors; LPS, lipopolysaccharide; TLR-4, toll-like receptor 4; NF-κB, nuclear factor kappa-light-chain-enhancer of activated B cells; HIF, hypoxia-inducible factors; HRE, hypoxia responsible elements; CDK5, kinase cyclin-dependent kinase 5; EGFR, epidermal growth factor receptor; ALK, anaplastic lymphoma kinase; miRNA, micro-RNA; CSN5-COP9, signalosome complex subunit 5. [/fig]
[fig] Figure 2: Immunotherapy resistance mechanisms in HNSCC. [/fig]
[fig] Author: Contributions: Conceptualization, P.G.M., C.B., A.M., G.F., M.F.; investigation, P.G.M., C.B.; writing original draft preparation, P.G.M.; writing-review and editing, C.P., P.G.M., C.B., A.M., M.F.; visualization, P.G.M., F.Z., C.B., A.M., A.C.; M.R., A.G., C.P., M.d.V.; supervision, C.B., A.M., M.F.; project administration, P.G.M., F.Z., A.C., C.P. All authors have read and agreed to the published version of the manuscript.Funding: This research was funded by DSB.AD007.256/TRANSLATIONAL BIOMEDICINE: MULTI-ORGAN PATHOLOGY AND THERAPY to C.B. Institutional Review Board Statement: The study did not require ethical approval. Informed Consent Statement: Not applicable since this is a review paper. Data Availability Statement: Not applicable. [/fig]
[table] Table 1: PD-L1 tumor hyperexpression mechanisms. [/table]
[table] Table 2: Immunotherapy resistance mechanisms in HNSCC. [/table]
|
Nonalcoholic Fatty Liver Disease and Endocrine Axes—A Scoping Review
Citation: Von-Hafe, M.; Borges-Canha, M.; Vale, C.; Leite, A.R.; Sérgio Neves, J.; Carvalho, D.; Leite-Moreira, A. Nonalcoholic Fatty Liver Disease and Endocrine Axes-A Scoping Review. Metabolites 2022, 12, 298. https://doi.
# Introduction
Nonalcoholic fatty liver disease (NAFLD) is the leading cause of chronic liver disease. NAFLD is a metabolic liver disease that encompasses a wide spectrum from simple steatosis to steatohepatitis (NASH) and fibrosis to cirrhosis and hepatocarcinoma [bib_ref] The diagnosis and management of nonalcoholic fatty liver disease: Practice guidance from..., Chalasani [/bib_ref]. It has also been termed a "barometer of metabolic health" due to its metabolic roots [bib_ref] Personalizing care for nonalcoholic fatty liver disease patients: What are the research..., Sookoian [/bib_ref]. A group of experts on the theme have recently reached a consensus, saying that this entity might be known as MAFLD (metabolic-associated fatty liver disease) and be diagnosed by positive criteria rather than being an exclusion diagnosis (requiring the exclusion of other causes chronic liver diseases before diagnosing NAFLD) [bib_ref] A new definition for metabolic dysfunction-associated fatty liver disease: An international expert..., Eslam [/bib_ref]. Its high prevalence, complex pathogenesis, and lack of approved therapies make this disease a hot topic of scientific research [bib_ref] Systematic review: The epidemiology and natural history of non-alcoholic fatty liver disease..., Vernon [/bib_ref].
NAFLD often occurs associated with endocrinopathies, as it has been increasingly recognized [bib_ref] Association between Nonalcoholic Fatty Liver Disease and Endocrinopathies: Clinical Implications, Singeap [/bib_ref] [bib_ref] Endocrine causes of nonalcoholic fatty liver disease, Marino [/bib_ref]. Digging into these associations may improve the current knowledge on this disease [bib_ref] Non-alcoholic fatty liver disease: Causes, diagnosis, cardiometabolic consequences, and treatment strategies, Stefan [/bib_ref]. Moreover, emerging evidence suggests that endocrine dysfunction may play an important role in NAFLD development, progression, and severity [bib_ref] Non-alcoholic fatty liver disease in common endocrine disorders, Hazlehurst [/bib_ref]. Moreover, a variety of rare hereditary liver and intestinal diseases as well as several drugs may trigger or worsen NAFLD; this further highlights the complexity in understanding NAFLD [bib_ref] Diagnosis and management of secondary causes of steatohepatitis, Liebe [/bib_ref].
## Hyperprolactinemia
Prolactin is a polypeptide hormone produced by lactotroph cells in the anterior pituitary [bib_ref] Prolactin-A pleiotropic factor in health and disease, Bernard [/bib_ref]. Prolactin release is mainly controlled by hypothalamic inhibitory tone through dopamine and the stimulatory influences of thyroid stimulating hormone (TSH)-releasing hormone and circulating estrogens. Its major functions are related to pregnancy and lac- Given the crucial role of GH in hepatic lipid metabolism, there are some clinical trials examining the impact of low-dose GH supplementation in patients with hepatic steatosis and NASH without known hypothalamic/pituitary disease. A new clinical trial had its results recently published, showing that treatment with recombinant human GH may have the potential to reduce liver fat content in adolescents with NAFLD and obesity [bib_ref] Effect of recombinant human growth hormone on liver fat content in young..., Pan [/bib_ref]. Other clinical trials studying the impact of GH supplementation on NAFLD are underway, such as the clinical trial named Growth Hormone and Intrahepatic Lipid Content in Patients With Nonalcoholic Fatty Liver Disease (NCT02217345). Lastly, IGF-1 replacement is also being considered as an option to treat patients with liver diseases [bib_ref] IGF-I induces senescence of hepatic stellate cells and limits fibrosis in a..., Nishizawa [/bib_ref]. Experimental studies show that treatment with IGF-1 is particularly beneficial in the reduction of liver fibrosis, although positive effects in hepatic steatosis and inflammation can also be seen [bib_ref] GH-independent IGF-I action is essential to prevent the development of nonalcoholic steatohepatitis..., Nishizawa [/bib_ref] [bib_ref] Somatotropic Axis Dysfunction in Non-Alcoholic Fatty Liver Disease: Beneficial Hepatic and Systemic..., Cabrera [/bib_ref].
## Acromegaly
Acromegaly is characterized by excessive GH and, consequently, IGF-1 and is the most frequent cause of GH-secreting pituitary adenoma. Increased levels of GH are associated with increased lipolysis and favorable body composition, with increased lean body mass and decreased visceral and subcutaneous adipose tissue [bib_ref] Acromegaly at diagnosis in 3173 patients from the Liege Acromegaly Survey (LAS)..., Petrossians [/bib_ref]. However, acromegaly also promotes insulin resistance, with consequent hyperglycemia, hyperinsulinemia, hypertriglyceridemia, and an increased risk of overt diabetes [bib_ref] Effects of growth hormone on glucose, lipid, and protein metabolism in human..., Moller [/bib_ref]. These paradoxical effects may justify contradictory evidence on this topic. While some studies including patients with active acromegaly found that intrahepatic lipid, measured by magnetic resonance spectroscopy, is relatively low in comparison to healthy subjects [bib_ref] Fat content in liver and skeletal muscle changes in a reciprocal manner..., Madsen [/bib_ref] [bib_ref] No evidence of ectopic lipid accumulation in the pathophysiology of the acromegalic..., Winhofer [/bib_ref] , others showed that hepatic steatosis is a common comorbidity in acromegaly, hypothesizing that lipotoxicity and insulin resistance may outweigh the direct hepatic effects of GH [bib_ref] Hepatic steatosis in patients with acromegaly, Koutsou-Tassopoulou [/bib_ref]. Acromegaly treatment with surgery or medical therapy improves metabolic risk by increasing insulin sensitivity [bib_ref] No evidence of ectopic lipid accumulation in the pathophysiology of the acromegalic..., Winhofer [/bib_ref]. However, GH, IGF-1, insulin-like growth factor binding proteins (IGFBPs), and medical treatment have a complex relationship with insulin sensitivity and hepatic steatosis. GHR antagonists (GHRA) induce an improvement in acromegaly glycemic control through the decrease of glucose and the normalization of insulin secretion [bib_ref] Insulin sensitivity and glucose tolerance improve in patients with acromegaly converted from..., Drake [/bib_ref]. This effect enables one to understand the important effect of GH on hepatic and peripheral IGF-1 action. Hepatic GH-induced IGF-1 production is regulated by portal insulin levels, as insulin promotes the translocation of hepatic GHR to the surface. When portal insulin levels are high, the liver becomes GH sensitive, regardless of the cause of the increased production of insulin. In addition, portal insulin also inhibits hepatic IGFBP1 production, which may increase the bioavailability of circulating IGF-1. Insulin suppression by somatostatin analogs (SSA) also selectively results in hepatic GH resistance, which itself decreases hepatic IGF-1 production [bib_ref] Insulin regulation of human hepatic growth hormone receptors: Divergent effects on biosynthesis..., Leung [/bib_ref]. Therefore, the consequent reduction in circulating IGF-1 does not necessarily reflect GH activity in peripheral tissues. It, thus, makes sense that the normalization of serum IGF-1 levels during SSA does not necessarily imply the control of disease's activity in peripheral tissues, which is a condition that Neggers coined as being "extra-hepatic acromegaly" [bib_ref] Hypothesis: Extra-hepatic acromegaly: A new paradigm?, Neggers [/bib_ref]. This concept received support in a study that evaluated surgically and SSA-treated acromegalics. Despite the normalization of IGF-1, SSA-treated patients had less suppressed GH levels and less symptom relief [bib_ref] Conventional and novel biomarkers of treatment outcome in patients with acromegaly: Discordant..., Rubeck [/bib_ref]. On the other hand, GHRA do not block all tissues with equal effectiveness for the GH actions. Adipose tissue seems to require less GHRA to reduce GH actions when compared to the liver, where more GHRA are required to reduce IGF-1 production [bib_ref] Inhibitory effect of a growth hormone receptor antagonist (G120K-PEG) on renal enlargement,..., Flyvbjerg [/bib_ref]. This could be a reason for local GHRA-induced lipomatosis. In further support of this hypothesis, it was recently reported that short-term GHRA administration in healthy subjects can suppress lipolysis without affecting either circulating or local IGF-1 [bib_ref] Impact of growth hormone receptor blockade on substrate metabolism during fasting in..., Moller [/bib_ref]. Accordingly, it is possible that peripheral suppression of GH activity is obtained prior to the normalization of hepatic IGF-1 production. Therefore, GHRA-treated patients with acromegaly and normal peripheral IGF-1 can have peripheral GH deficiency [bib_ref] Hypothesis: Extra-hepatic acromegaly: A new paradigm?, Neggers [/bib_ref]. If all this is true, then patients with acromegaly and diabetes should need higher GHRA doses to normalize IGF-1 compared to patients without diabetes. Recently, this was demonstrated by Droste et al. [bib_ref] Therapy of acromegalic patients exacerbated by concomitant type 2 diabetes requires higher..., Droste [/bib_ref].
In a cross-sectional study including patients previously treated for acromegaly, hepatic steatosis, measured by magnetic resonance spectroscopy, was found to be increased compared to healthy controls, even several years after successful treatment [bib_ref] Reduced basal ATP synthetic flux of skeletal muscle in patients with previous..., Szendroedi [/bib_ref]. In a recent prospective study, Ciresi et al. found no differences in the prevalence of steatosis after 12 months of treatment with somatostatin analogs, measured by abdominal ultrasonography [bib_ref] Hepatic Steatosis Index in Acromegaly: Correlation with Insulin Resistance Regardless of the..., Ciresi [/bib_ref]. The heterogeneity in the chosen imaging method to assess hepatic steatosis and in the treatments for acromegaly may explain these differences.
## Hyperprolactinemia
Prolactin is a polypeptide hormone produced by lactotroph cells in the anterior pituitary [bib_ref] Prolactin-A pleiotropic factor in health and disease, Bernard [/bib_ref]. Prolactin release is mainly controlled by hypothalamic inhibitory tone through dopamine and the stimulatory influences of thyroid stimulating hormone (TSH)-releasing hormone and circulating estrogens. Its major functions are related to pregnancy and lactation [bib_ref] Time for a New Perspective on Prolactin in Metabolism, Macotela [/bib_ref]. In addition, prolactin is involved in the regulation of the immune system, food intake, and bone formation [bib_ref] Prolactin-A pleiotropic factor in health and disease, Bernard [/bib_ref].
A growing body of evidence supports prolactin as an active contributor to human metabolic health [bib_ref] Time for a New Perspective on Prolactin in Metabolism, Macotela [/bib_ref]. Namely, experimental animal studies found that it stimulates pancreatic β cell proliferation and insulin gene transcription, modulates lipid metabolism in adipose tissue, and induces adipogenesis [bib_ref] Targeted deletion of the PRL receptor: Effects on islet development, insulin production,..., Freemark [/bib_ref] [bib_ref] Prolactin Promotes Adipose Tissue Fitness and Insulin Sensitivity in Obese Males, Ruiz-Herrera [/bib_ref]. Despite the beneficial roles of prolactin on metabolic homeostasis, pathological increases in prolactin levels have been frequently associated with metabolic disturbances, namely, weight gain, obesity, hyperinsulinemia, and reduced insulin sensitivity, all considered important players in the pathogenesis of NAFLD [bib_ref] The influences of hyperprolactinemia and obesity on cardiovascular risk markers: Effects of..., Serri [/bib_ref]. Normalization of prolactin levels with dopamine agonists correlated with weight loss, although some studies have shown a more pronounced weight loss in men, suggesting a gender difference [bib_ref] Insulin sensitivity and lipid profile in prolactinoma patients before and after normalization..., Berinder [/bib_ref]. Furthermore, treatment with dopamine agonists improves insulin sensitivity, glycemic control, and lipid profile, reducing triglycerides and total and LDL cholesterol [bib_ref] BMI and metabolic profile in patients with prolactinoma before and after treatment..., Dos Santos Silva [/bib_ref].
The impact of prolactin on liver function and structure is poorly understood [fig_ref] Figure 2: Impact of prolactin levels on NAFLD development and progression [/fig_ref]. The presence of functional prolactin receptors in hepatocytes has also been previously demonstrated [bib_ref] Tissue distribution and regulation of rat prolactin receptor gene expression. Quantitative analysis..., Nagano [/bib_ref]. Recent in vitro and in vivo studies suggest that prolactin protects the liver against lipid accumulation by decreasing the expression of CD36 and stearoylcoenzyme A desaturase 1 (SCD1), an enzyme involved in fatty acid biosynthesis [bib_ref] Prolactin improves hepatic steatosis via CD36 pathway, Zhang [/bib_ref] [bib_ref] Ablation of prolactin receptor increases hepatic triglyceride accumulation, Shao [/bib_ref]. In rodents, prolactin is thought to be involved in the regulation of liver insulin sensitivity [bib_ref] PRLR regulates hepatic insulin sensitivity in mice via STAT5, Yu [/bib_ref]. In human studies, lower prolactin levels were found in patients with more severe hepatic steatosis, suggesting a possible involvement of prolactin in the progression of this disease [bib_ref] Prolactin improves hepatic steatosis via CD36 pathway, Zhang [/bib_ref]. Although prolactin is thought to reduce liver fat content, it is plausible that chronic hyperprolactinemia is involved in the development of NAFLD. Despite the absence of studies evaluating liver function and structure in patients with hyperprolactinemia, a few animal studies support this hypothesis. Luque et al. suggest that prolactin may be directly involved in changes in the signaling pathways of de novo lipogenesis, which lead to fatty liver [bib_ref] Chronic hyperprolactinemia evoked by disruption of lactotrope dopamine D2 receptors impacts on..., Luque [/bib_ref]. In diabetic murine models, the triglyceride content in the liver increased with the administration of high doses of prolactin [bib_ref] Serum prolactin concentrations determine whether they improve or impair beta-cell function and..., Park [/bib_ref].
disease [bib_ref] Prolactin improves hepatic steatosis via CD36 pathway, Zhang [/bib_ref]. Although prolactin is thought to reduce liver fat content, it is plausible that chronic hyperprolactinemia is involved in the development of NAFLD. Despite the absence of studies evaluating liver function and structure in patients with hyperprolactinemia, a few animal studies support this hypothesis. Luque et al. suggest that prolactin may be directly involved in changes in the signaling pathways of de novo lipogenesis, which lead to fatty liver [bib_ref] Chronic hyperprolactinemia evoked by disruption of lactotrope dopamine D2 receptors impacts on..., Luque [/bib_ref]. In diabetic murine models, the triglyceride content in the liver increased with the administration of high doses of prolactin [bib_ref] Serum prolactin concentrations determine whether they improve or impair beta-cell function and..., Park [/bib_ref].
## Vasopressin disturbances
Vasopressin (V), also known as antidiuretic hormone (ADH), has a crucial role in the regulation of water balance, vascular tone, and the endocrine stress response [bib_ref] Vasopressin and Copeptin in health and disease, Christ-Crain [/bib_ref]. Due to its short half-life, very low concentration, small size, and poor stability in plasma samples, clinical studies indirectly determine the levels of ADH in the circulation through the measurement of copeptin, which is produced in equimolar quantities with ADH [bib_ref] Vasopressin and Copeptin in health and disease, Christ-Crain [/bib_ref].
A role of ADH in the regulation of glucose and lipid metabolism has been acknowledged by several epidemiological and experimental studies [bib_ref] Vasopressin and metabolic disorders: Translation from experimental models to clinical use, Nakamura [/bib_ref]. Patients with diabetes mellitus have markedly increased levels of ADH in comparison with healthy subjects [bib_ref] The Vasopressin System in the Risk of Diabetes and Cardiorenal Disease, and..., Enhorning [/bib_ref]. Whether copeptin is predictive of the development of diabetes-induced NAFLD remains unknown. Recent studies demonstrated a strong association between high copeptin levels and the prevalence and severity of both NAFLD and NASH [bib_ref] The Vasopressin System in the Risk of Diabetes and Cardiorenal Disease, and..., Enhorning [/bib_ref] [bib_ref] Elevated plasma copeptin levels identify the presence and severity of non-alcoholic fatty..., Barchetta [/bib_ref].
Several ex vivo and in vitro studies have shown that ADH intensifies hyperglycemia. It enhances hepatic gluconeogenesis and glycogenolysis through the activation of V1a receptors [bib_ref] Stimulation by vasopressin of glycogen breakdown and gluconeogenesis in the perfused rat..., Hems [/bib_ref]. ADH also induces vasoconstriction, contributing to liver hypoxia, and further stimulates glycogenolysis [bib_ref] Stimulation of glycogenolysis and vasoconstriction in the perfused rat liver by the..., Fisher [/bib_ref]. and stimulates the release of pituitary adrenocorticotropin hormone via V1b receptors, increasing the release of cortisol, which is thought to be an important contributor to ADH-induced hyperglycemia and insulin resistance [bib_ref] Reduced water intake deteriorates glucose regulation in patients with type 2 diabetes, Johnson [/bib_ref].
## Vasopressin disturbances
Vasopressin (V), also known as antidiuretic hormone (ADH), has a crucial role in the regulation of water balance, vascular tone, and the endocrine stress response [bib_ref] Vasopressin and Copeptin in health and disease, Christ-Crain [/bib_ref]. Due to its short half-life, very low concentration, small size, and poor stability in plasma samples, clinical studies indirectly determine the levels of ADH in the circulation through the measurement of copeptin, which is produced in equimolar quantities with ADH [bib_ref] Vasopressin and Copeptin in health and disease, Christ-Crain [/bib_ref].
A role of ADH in the regulation of glucose and lipid metabolism has been acknowledged by several epidemiological and experimental studies [bib_ref] Vasopressin and metabolic disorders: Translation from experimental models to clinical use, Nakamura [/bib_ref]. Patients with diabetes mellitus have markedly increased levels of ADH in comparison with healthy subjects [bib_ref] The Vasopressin System in the Risk of Diabetes and Cardiorenal Disease, and..., Enhorning [/bib_ref]. Whether copeptin is predictive of the development of diabetes-induced NAFLD remains unknown. Recent studies demonstrated a strong association between high copeptin levels and the prevalence and severity of both NAFLD and NASH [bib_ref] The Vasopressin System in the Risk of Diabetes and Cardiorenal Disease, and..., Enhorning [/bib_ref] [bib_ref] Elevated plasma copeptin levels identify the presence and severity of non-alcoholic fatty..., Barchetta [/bib_ref].
Several ex vivo and in vitro studies have shown that ADH intensifies hyperglycemia. It enhances hepatic gluconeogenesis and glycogenolysis through the activation of V1a receptors [bib_ref] Stimulation by vasopressin of glycogen breakdown and gluconeogenesis in the perfused rat..., Hems [/bib_ref]. ADH also induces vasoconstriction, contributing to liver hypoxia, and further stimulates glycogenolysis [bib_ref] Stimulation of glycogenolysis and vasoconstriction in the perfused rat liver by the..., Fisher [/bib_ref]. and stimulates the release of pituitary adrenocorticotropin hormone via V1b receptors, increasing the release of cortisol, which is thought to be an important contributor to ADH-induced hyperglycemia and insulin resistance [bib_ref] Reduced water intake deteriorates glucose regulation in patients with type 2 diabetes, Johnson [/bib_ref].
Despite being controversial and less understood, most studies suggest that ADH decreases plasma non-esterified fatty acids [bib_ref] Hypermetabolism of fat in V1a vasopressin receptor knockout mice, Hiroyama [/bib_ref] [bib_ref] Metabolic effects of vasopressin infusion in the starved rat. Reversal of ketonaemia, Rofe [/bib_ref]. ADH promotes lipogenesis in the hepatic tissue through the V1a receptor and inhibits lipolysis in adipocytes [bib_ref] Hypermetabolism of fat in V1a vasopressin receptor knockout mice, Hiroyama [/bib_ref] [bib_ref] The agouti gene product inhibits lipolysis in human adipocytes via a Ca..., Xue [/bib_ref]. The Vinduced decrease in plasma non-esterified fatty acids may reduce its supply to the liver [bib_ref] Metabolic effects of vasopressin infusion in the starved rat. Reversal of ketonaemia, Rofe [/bib_ref]. On the other hand, one study found the role of vasopressin in fatty acid synthesis and lipogenesis varied with different incubating media of rat hepatocytes, suggesting that vasopressin function may vary according to the nutritional status [bib_ref] Inhibition of lipogenesis by vasopressin and angiotensin II in glycogendepleted hepatocytes, Palmer [/bib_ref]. Taveau et al. found reduced levels of hepatic cholesterol and triacylglycerol and a lower expression of genes involved in lipogenesis in well-hydrated obese rats with low levels of ADH [bib_ref] Vasopressin and hydration play a major role in the development of glucose..., Taveau [/bib_ref].
## Phosphocalcic metabolism disturbances
## Vitamin d deficiency
Vitamin D is classically recognized for its role in phosphocalcium metabolism and bone health. Notwithstanding, its receptors are present ubiquitously, and it is associated with several effects in various organs and systems [bib_ref] Vitamin D: Deficiency, sufficiency and toxicity, Alshahrani [/bib_ref] [bib_ref] Overview of general physiologic features and functions of vitamin D, Deluca [/bib_ref]. Vitamin D deficiency seems to be associated with several metabolic disturbances [bib_ref] Relationship between adipose tissue dysfunction, vitamin D deficiency and the pathogenesis of..., Cimini [/bib_ref] [bib_ref] Associations of vitamin D with insulin resistance, obesity, type 2 diabetes, and..., Wimalawansa [/bib_ref] [bib_ref] Interplay of vitamin D and metabolic syndrome: A review, Prasad [/bib_ref] , namely, NAFLD, although inconsistently.
Fundamental research studies have shown that vitamin D exerts anti-inflammatory, anti-proliferative, and anti-fibrotic effects in the liver [bib_ref] Vitamin D inhibits proliferation and profibrotic marker expression in hepatic stellate cells..., Abramovitch [/bib_ref]. In vitro and in vivo studies suggest that this protective effect may be partially mediated by stellate cells through the inhibition of fibrogenesis [bib_ref] Relationship between adipose tissue dysfunction, vitamin D deficiency and the pathogenesis of..., Cimini [/bib_ref] [bib_ref] Vitamin D inhibits proliferation and profibrotic marker expression in hepatic stellate cells..., Abramovitch [/bib_ref] [bib_ref] Vitamin D counteracts fibrogenic TGF-beta signalling in human hepatic stellate cells both..., Beilfuss [/bib_ref] Using mouse hepatocytes, Dong B. et al. showed that vitamin D receptor activation in macrophages by vitamin D ligands ameliorates liver inflammation and steatosis [bib_ref] Vitamin D Receptor Activation in Liver Macrophages Ameliorates Hepatic Inflammation, Steatosis, and..., Dong [/bib_ref]. Moreover, obese animals with vitamin D deficiency presented with greater NAFLD progression, and vitamin D supplementation improved hepatic morphology and function [bib_ref] Relationship between adipose tissue dysfunction, vitamin D deficiency and the pathogenesis of..., Cimini [/bib_ref]. Contrariwise, Bozic et al. showed that hepatic vitamin D receptor activation promotes high-fat diet-associated liver steatosis in a mouse model [bib_ref] Hepatocyte vitamin D receptor regulates lipid metabolism and mediates experimental diet-induced steatosis, Bozic [/bib_ref].
Human studies are also conflicting. Several epidemiological studies lead towards an association between low vitamin D levels and NAFLD, though no causal relationship has been found [bib_ref] Meta-analysis: Vitamin D and non-alcoholic fatty liver disease, Eliades [/bib_ref] [bib_ref] Serum vitamin D concentrations and unexplained elevation in ALT among US adults, Liangpunsakul [/bib_ref] [bib_ref] Vitamin D: A new player in non-alcoholic fatty liver disease?, Eliades [/bib_ref]. Namely, Targher et al. showed that patients with biopsy-proven NAFLD had lower vitamin D levels compared to controls [bib_ref] Associations between serum 25-hydroxyvitamin D3 concentrations and liver histology in patients with..., Targher [/bib_ref]. Furthermore, an analysis of data from the National Health and Nutrition Examination Survey (NHANES III) claims that vitamin D levels are inversely associated with the severity of NAFLD [bib_ref] Association between Vitamin D Status and Non-Alcoholic Fatty Liver Disease: A Population-Based..., Liu [/bib_ref]. A recent study from our group associated vitamin D deficiency with a higher risk of hepatic steatosis in individuals with morbid obesity [bib_ref] The Impact of Vitamin D in Non-Alcoholic Fatty Liver Disease: A Cross-Sectional..., Borges-Canha [/bib_ref]. An analysis from the Sixth Korea National Health and Nutrition Examination Survey (KNHANES VI) argued against such association [bib_ref] The Association between Vitamin D Insufficiency and Nonalcoholic Fatty Liver Disease: A..., Ha [/bib_ref]. In a Mendelian randomization analysis, Wang et al. found no causal association between vitamin D and NAFLD in a Chinese population with over 9000 participants. Additionally, Barchetta et al. performed a randomized, double-blind, placebo-controlled trial (RCT) and concluded that vitamin D supplementation, in a high oral dose for a period of 24 weeks did not improve hepatic steatosis in patients with NAFLD and type 2 diabetes [bib_ref] No effects of oral vitamin D supplementation on non-alcoholic fatty liver disease..., Barchetta [/bib_ref] [bib_ref] A role of 1,25(OH)2D3 supplementation in rats with nonalcoholic steatohepatitis induced by..., Han [/bib_ref]. In another RCT in NAFLD patients, no beneficial effects on liver function were seen when comparing supplementation with vitamin D, calcitriol, and placebo [bib_ref] Vitamin D supplementation for the treatment of non-alcoholic fatty liver disease: A..., Dabbaghmanesh [/bib_ref]. Therefore, no strong recommendations currently exist concerning vitamin D supplementation in patients with NAFLD [bib_ref] Vitamin D supplementation for chronic liver diseases in adults, Bjelakovic [/bib_ref].
Pacifico L. et al. dwelled on the possible confounders of a NAFLD/NASH and vitamin D deficiency association, namely, the influence of the host and environment in vitamin D levels as well as the great variability in laboratory methods and intra-individual variability in this vitamin's level [bib_ref] Association between Vitamin D Levels and Nonalcoholic Fatty Liver Disease: Potential Confounding..., Pacifico [/bib_ref]. More studies, particularly well-powered randomized controlled trials, are needed to evaluate the potential role of vitamin D supplementation in the management of NAFLD.
## Other disturbances of bone metabolism
The deterioration of bone homeostasis has been incongruously associated with NAFLD, and the pathophysiology of such association remains to be elucidated [bib_ref] Novel insights into the relationship between nonalcoholic fatty liver disease and osteoporosis, Filip [/bib_ref]. It is hypothesized that very complex mechanisms are involved, and a plausible perspective centralizes the causal pathway in the liver. The dysfunctional visceral adipose tissue that often exists concurrently with NAFLD is a great source of pro-inflammatory, pro-coagulant, and profibrogenic factors that may contribute to this association. The state of insulin resistance commonly present in NAFLD patients may also play an important role as well as the vitamin D deficit that may relate to such hepatic disorders, as previously mentioned [bib_ref] Novel insights into the relationship between nonalcoholic fatty liver disease and osteoporosis, Filip [/bib_ref] [bib_ref] Nonalcoholic fatty liver disease and decreased bone mineral density: Is there a..., Targher [/bib_ref] [bib_ref] Non-alcoholic fatty liver disease connections with fat-free tissues: A focus on bone..., Poggiogalle [/bib_ref].
An observational study in a Chinese population showed that the liver fat content was inversely correlated with bone mineral density (BMD) in middle-aged and elderly men but not in women [bib_ref] The association of liver fat content and serum alanine aminotransferase with bone..., Xia [/bib_ref]. Additionally, in a cohort of Chinese participants that were 40 years old or older, the prevalence of osteoporotic fractures was significantly greater in men with NAFLD but not in women [bib_ref] Association between nonalcoholic fatty liver disease (NAFLD) and osteoporotic fracture in middle-aged..., Li [/bib_ref]. Moreover, a Korean population-based study presented a detrimental effect of NAFLD on BMD in men but an unexpected positive effect in postmenopausal women [bib_ref] Association between bone mineral density and nonalcoholic fatty liver disease in Korean..., Lee [/bib_ref]. There are some studies that show similar harmful effects for both genders (namely postmenopausal women), but others evidence a detrimental effect in postmenopausal women [bib_ref] Low bone mineral density in chinese adults with nonalcoholic Fatty liver disease, Cui [/bib_ref] [bib_ref] Association of nonalcoholic fatty liver disease with low bone mass in postmenopausal..., Moon [/bib_ref]. Bhatt et al. stated that PTH levels are independently associated with NAFLD in Asian Indians [bib_ref] Independent associations of low 25 hydroxy vitamin D and high parathyroid hormonal..., Bhatt [/bib_ref]. Similarly, a metanalysis that included observational studies focusing on bone mineral density claimed that no correlation exists between bone mineral density and NAFLD [bib_ref] Nonalcoholic fatty liver disease and osteoporosis: A systematic review and meta-analysis, Upala [/bib_ref]. It is important to recognise that most of the studies mentioned are in Asian populations, suggesting an ethnic difference in the impact of the phosphocalcic axis on the liver.
Despite being a controversial theme of the debate, it is of great interest to clarify the existence of such links in future prospective studies, namely, addressing gender, ethnic, and age differences. The routine screening of BMD in NAFLD patients may become an important addition to the management of these complex patients.
## Thyroid dysfunction
Thyroid dysfunction, explicitly, hypothyroidism, has been proposed as a possible contributory mechanism for the pathophysiology of NAFLD [bib_ref] Endocrine and liver interaction: The role of endocrine pathways in NASH, Loria [/bib_ref] [fig_ref] Figure 3: Thyroid-liver axis [/fig_ref]. A large 2018 metanalysis that included a total of 15 studies and 44,140 individuals suggested that hypothyroidism is significantly associated with the presence and severity of NAFLD [bib_ref] Association Between Primary Hypothyroidism and Nonalcoholic Fatty Liver Disease: A Systematic Review..., Mantovani [/bib_ref]. It is biologically plausible that the thyroid axis plays an important role in NAFLD development, as thyroid hormones (TH) are crucial in the regulation of numerous metabolic processes, such as cholesterol and lipid metabolism and intra-hepatic concentration, circulating lipoprotein levels, body weight, and insulin resistance [bib_ref] Association of non-alcoholic fatty liver disease with thyroid function: A systematic review..., Guo [/bib_ref] [bib_ref] Direct effects of thyroid hormones on hepatic lipid metabolism, Sinha [/bib_ref] [bib_ref] NAFLD in Some Common Endocrine Diseases: Prevalence, Pathophysiology, and Principles of Diagnosis..., Lonardo [/bib_ref] [bib_ref] Pathogenesis of hypothyroidism-induced NAFLD: Evidence for a distinct disease entity? Dig, Lonardo [/bib_ref]. TH regulate the expression of several hepatic lipogenic genes, and recent studies have shown that several genes whose expression are altered in NAFLD are also regulated by TH [bib_ref] Transcriptional regulation of hepatic lipogenesis, Wang [/bib_ref] [bib_ref] Repair-related activation of hedgehog signaling in stromal cells promotes intrahepatic hypothyroidism, Bohinc [/bib_ref]. Nevertheless, there is still controversy on this subject.
At a local level, both human and animal studies show that TH levels are decreased in the livers of individuals with NAFLD, and defective intrahepatic deiodinase expression has recently been proposed as a hallmark of NASH. Namely, a study in mice showed that myricetin, a dietary antioxidant, improved hepatic steatosis, and this was associated with increased hepatic type 1 deiodinase activity [bib_ref] Myricetin alleviated hepatic steatosis by acting on microRNA-146b/thyroid hormone receptor b pathway..., Xia [/bib_ref]. Moreover, the literature suggests that hepatic fatty acids in NAFLD may impair TH receptor activity [bib_ref] Fatty acyl-CoAs are potent inhibitors of the nuclear thyroid hormone receptor in..., Li [/bib_ref]. The local hypothyroid status decreases hepatic lipase activity, which promotes triglyceride accumulation [bib_ref] Targeting thyroid hormone receptor-beta agonists to the liver reduces cholesterol and triglycerides..., Erion [/bib_ref] [bib_ref] Reduction of hepatic steatosis in rats and mice after treatment with a..., Cable [/bib_ref] [bib_ref] 3,5-diiodo-l-thyronine, by modulating mitochondrial functions, reverses hepatic fat accumulation in rats fed..., Mollica [/bib_ref]. Furthermore, animal-based studies in models of MS and NAFLD have shown that the administration of both TH and TH agonists ameliorates hepatic steatosis [bib_ref] Targeting thyroid hormone receptor-beta agonists to the liver reduces cholesterol and triglycerides..., Erion [/bib_ref] [bib_ref] Reduction of hepatic steatosis in rats and mice after treatment with a..., Cable [/bib_ref] [bib_ref] 3,5-diiodo-l-thyronine, by modulating mitochondrial functions, reverses hepatic fat accumulation in rats fed..., Mollica [/bib_ref]. Although further studies are needed, human data are beginning to appear. Low-dose TH therapy in humans leads to safe and effective results in NAFLD patients [bib_ref] Low-Dose Levothyroxine Reduces Intrahepatic Lipid Content in Patients With Type 2 Diabetes..., Bruinstroop [/bib_ref]. A recent randomized double-blind placebo-controlled trial with a selective thyroid hormone receptor-β agonist showed a significant decrease in hepatic fat content in patients with NASH [bib_ref] MGL-3196) for the treatment of non-alcoholic steatohepatitis: A multicentre, randomised, double-blind, placebo-controlled,..., Harrison [/bib_ref].
Concerning systemic TH, Liu et al. showed that both free triiodothyronine (FT3) and TSH levels were positively correlated with the risk of NAFLD in euthyroid individuals [bib_ref] Thyroid Function and Risk of Non-Alcoholic Fatty Liver Disease in Euthyroid Subjects, Liu [/bib_ref]. Van der Bergh et al. also studied this topic on a euthyroid population, and the results evidenced that NAFLD patients presented with higher FT3 and lower free thyroxin (FT4) levels, and no differences were recorded concerning TSH [bib_ref] Higher free triiodothyronine is associated with non-alcoholic fatty liver disease in euthyroid..., Van Den Berg [/bib_ref]. In a recent meta-analysis, Guo et al. concluded that TSH levels may be positively correlated with NAFLD, independent of TH levels, and that TSH levels increase with the progression of NAFLD [bib_ref] Association of non-alcoholic fatty liver disease with thyroid function: A systematic review..., Guo [/bib_ref]. This association was also recently observed by our group in a morbidly obese population [bib_ref] Thyroid Function and the Risk of Non-Alcoholic Fatty Liver Disease in Morbid..., Borges-Canha [/bib_ref]. Although there are numerous other studies in agreement with this line of thought, some studies argue that no correlation exists [bib_ref] Nonalcoholic Fatty Liver Disease Is Not Associated with Thyroid Hormone Levels and..., Jaruvongvanich [/bib_ref].
Since both overt and subclinical hypothyroidism are widely associated with cardiovascular risk factors that are, in turn, associated with NAFLD [bib_ref] Subclinical hypothyroidism and cardiovascular risk factors, Delitala [/bib_ref] , this association may be the result of confounding indirect effects. However, the Rotterdam study evidenced a close relationship between NAFLD and overt hypothyroidism, independent of other metabolic risk factors [bib_ref] Thyroid Function and the Risk of Nonalcoholic Fatty Liver Disease: The Rotterdam..., Bano [/bib_ref].
thyroidism and showed that the administration of T3 compromised HCC progression, even when given at late stages, lifting the veil for a possible approach to this poor-prognosis-associated carcinoma [bib_ref] Thyroid hormone inhibits hepatocellular carcinoma progression via induction of differentiation and metabolic..., Kowalik [/bib_ref].
In contrast to the considerable evidence already available in hypothyroidism, the effect of hyperthyroidism in NAFLD has been considerably less well-studied. In a case report of a patient with NASH in whom Graves' disease (GD) was installed, the liver enzyme levels improved after the onset of GD and consequent hyperthyroidism and worsened after starting the treatment and returning to an euthyroid state [bib_ref] Hyperthyroidism Improves the Pathological Condition of Nonalcoholic Steatohepatitis: A Case of Nonalcoholic..., Miyake [/bib_ref]. Therefore, it seems that TH may stand out in a way that, even in pathological states, they seem to exert positive effects regarding NAFLD. Even if this is in line with the previously hypothesized role of TH in NAFLD pathophysiology, more studies are needed to confirm. The data from Kowalik MA et al. is also very interesting; these authors explored the recent hypothesis of an association between hepatocellular carcinoma (HCC) and hypothyroidism and showed that the administration of T3 compromised HCC progression, even when given at late stages, lifting the veil for a possible approach to this poor-prognosisassociated carcinoma [bib_ref] Thyroid hormone inhibits hepatocellular carcinoma progression via induction of differentiation and metabolic..., Kowalik [/bib_ref].
In contrast to the considerable evidence already available in hypothyroidism, the effect of hyperthyroidism in NAFLD has been considerably less well-studied. In a case report of a patient with NASH in whom Graves' disease (GD) was installed, the liver enzyme levels improved after the onset of GD and consequent hyperthyroidism and worsened after starting the treatment and returning to an euthyroid state [bib_ref] Hyperthyroidism Improves the Pathological Condition of Nonalcoholic Steatohepatitis: A Case of Nonalcoholic..., Miyake [/bib_ref]. Therefore, it seems that TH may stand out in a way that, even in pathological states, they seem to exert positive effects regarding NAFLD. Even if this is in line with the previously hypothesized role of TH in NAFLD pathophysiology, more studies are needed to confirm.
## Reproductive system dysfunction and nafld
It is widely appreciated that disorders causing sex hormone dysfunction may predispose patients to the development of MS [bib_ref] Metabolic Syndrome in Male Hypogonadism, Rastrelli [/bib_ref] [bib_ref] The Metabolic Syndrome in Central Hypogonadotrophic Hypogonadism, Dwyer [/bib_ref] [bib_ref] Metabolic Syndrome in Polycystic Ovary Syndrome, Pasquali [/bib_ref]. Therefore, it is plausible that changes in the reproductive axis may lead to the development of NAFLD [fig_ref] Figure 4: Sexual hormones and NAFLD [/fig_ref].
## Hypogonadism
Hypogonadism is defined as the reduction of sex hormones. There are various pathophysiological mechanisms underlying hypogonadism, but the presence of NAFLD seems to be an ubiquitous comorbidity in these patients [bib_ref] Hypogonadism and non-alcoholic fatty liver disease, Mintziori [/bib_ref]. The association between hypogonadism and NAFLD seems to be bidirectional, and causality is difficult to establish. There is a positive association between lower levels of testosterone and the prevalence of NAFLD [bib_ref] Low testosterone and non-alcoholic fatty liver disease: Evidence for their independent association..., Barbonetti [/bib_ref] [bib_ref] Prediction of prevalent but not incident non-alcoholic fatty liver disease by levels..., Seo [/bib_ref]. Men with biopsy-proven NASH have diminished dehydroepiandrosterone (DHEA) levels, and this decrease parallels an increasing severity of hepatic fibrosis [bib_ref] Endocrine and liver interaction: The role of endocrine pathways in NASH, Loria [/bib_ref] [bib_ref] Prediction of prevalent but not incident non-alcoholic fatty liver disease by levels..., Seo [/bib_ref] [bib_ref] Low circulating levels of dehydroepiandrosterone in histologically advanced nonalcoholic fatty liver disease, Charlton [/bib_ref] DHEA modulates oxidative stress sensitivity, as it is a potential mediator of reactive oxygen species scavengers, enhances insulin sensitivity, and increases peroxisome proliferator-activated receptor expression [bib_ref] Peroxisome proliferator-activated receptor alpha required for gene induction by dehydroepiandrosterone-3 beta-sulfate, Peters [/bib_ref]. Serum DHEA levels might be used for the identification of patients at risk for the development of NASH with advanced fibrosis [bib_ref] Low circulating levels of dehydroepiandrosterone in histologically advanced nonalcoholic fatty liver disease, Charlton [/bib_ref].
Additionally, the literature advocates that hypogonadism is associated with important cardiovascular risk factors, such as general and visceral obesity, impaired insulin sensibility, hypertension, and dyslipidemia [bib_ref] Endocrine and liver interaction: The role of endocrine pathways in NASH, Loria [/bib_ref]. These are also crucial contributors to the development of NAFLD and can be potential confounders in some studies in this topic. A study by Sakr et al. evidenced that male rats with decreased gonadal hormones have higher serum and hepatic levels of triglycerides, cholesterol, and LDL and have increased levels of enzymes involved in hepatic steatogenesis, such as sterol regulatory element-binding transcription factor 1 (SREBP-1) and 2 (SREBP-2) [bib_ref] Possible mechanisms underlying fatty liver in a rat model of male hypogonadism:..., Sakr [/bib_ref]. Androgen receptor knockout male mice acquire late-onset obesity and have an increased risk of NAFLD development and progression [bib_ref] Androgen receptor null male mice develop late-onset obesity caused by decreased energy..., Fan [/bib_ref]. Hormonal replacement therapy (HRT) with testosterone seems to ameliorate these changes and may be considered a protective strategy to be taken into account [bib_ref] Long-term testosterone therapy in hypogonadal men ameliorates elements of the metabolic syndrome:..., Traish [/bib_ref]. Clinical studies show that long-term testosterone replacement improves MS components and ameliorates liver enzymes changes in men with hypogonadism [bib_ref] Long-term testosterone therapy in hypogonadal men ameliorates elements of the metabolic syndrome:..., Traish [/bib_ref]. Sex hormones play a central role in regulating glucose metabolism and insulin sensitivity [bib_ref] An emerging regulator of insulin action and mitochondrial function, Gupte [/bib_ref]. The decrease in insulin sensitivity, particularly in the liver, leads to an increase in hepatic gluconeogenesis, glucose production, and lipogenesis, which may exacerbate hepatic steatosis [bib_ref] The role of hepatic lipids in hepatic insulin resistance and type 2..., Perry [/bib_ref]. Several studies suggest that sex hormones are, in part, responsible for liver energy homeostasis and the regulation of hepatic fat accumulation [bib_ref] Sex Hormones and Their Receptors Regulate Liver Energy Homeostasis, Shen [/bib_ref]. In association with metabolic dysfunction, a sex hormone imbalance may be responsible for inducing hepatic injury. Estrogens and androgens are present in both men and women, although in different concentrations. A growing body of evidence suggests that estrogens have important metabolic functions in both women and men [bib_ref] Gonadal steroids and body composition, strength, and sexual function in men, Finkelstein [/bib_ref]. Estrogens seem to play a protective role against hepatic fat accumulation by promoting lipolysis and decreasing lipogenesis, mostly through the induction of acetyl-CoA carboxylase and, consequently, improving mitochondrial function [bib_ref] Differential effects of estrogen/androgen on the prevention of nonalcoholic fatty liver disease..., Zhang [/bib_ref]. Estrogens also maintain the hepatic cholesterol balance, suppressing its synthesis and promoting cholesterol clearance. Namely, estrogen receptor signaling is essential for appropriate hepatic lipid metabolism and the evidence supports that both estrogen receptor α and the G protein-coupled estrogen receptor are important regarding this matter [bib_ref] Sex Hormone-Dependent Physiology and Diseases of Liver, Kur [/bib_ref]. Studies in animal models have shown that androgens also have an important role in protecting the liver from fat accumulation and in decreasing the risk of NAFLD development [bib_ref] Differential effects of estrogen/androgen on the prevention of nonalcoholic fatty liver disease..., Zhang [/bib_ref]. On the contrary, in women, higher androgen levels can increase the risk of NAFLD risk, as described later in the polycystic ovary syndrome (POCS) subhead [bib_ref] Polycystic ovary syndrome with hyperandrogenism is characterized by an increased risk of..., Jones [/bib_ref]. These opposite described effects might be explained by sex differences in androgenic action. In fact, in men, normal serum androgen levels prevent hepatic lipid accumulation, whereas androgen deficiency is associated with fatty liver accumulation [bib_ref] MECHANISMS IN ENDOCRINOLOGY: The sexually dimorphic role of androgens in human metabolic..., Schiffer [/bib_ref]. A possible explanation is that serum testosterone concentration in these two spectrums (females with excess androgen and males with androgen deficiency) may overlap. This range of androgens might be the concentration window responsible for metabolic dysfunction, and some authors even designate this adverse concentration range of circulating androgen as the "metabolic valley of death" [bib_ref] The striking similarities in the metabolic associations of female androgen excess and..., Escobar-Morreale [/bib_ref]. However, the cellular and systemic mechanisms of this phenomenon are still poorly understood.
## Hypogonadism
Hypogonadism is defined as the reduction of sex hormones. There are various pathophysiological mechanisms underlying hypogonadism, but the presence of NAFLD seems to be an ubiquitous comorbidity in these patients [bib_ref] Hypogonadism and non-alcoholic fatty liver disease, Mintziori [/bib_ref]. The association between hypogonadism and NAFLD seems to be bidirectional, and causality is difficult to establish. There is a positive association between lower levels of testosterone and the prevalence of NAFLD [bib_ref] Low testosterone and non-alcoholic fatty liver disease: Evidence for their independent association..., Barbonetti [/bib_ref] [bib_ref] Prediction of prevalent but not incident non-alcoholic fatty liver disease by levels..., Seo [/bib_ref]. Men with biopsy-proven NASH have diminished dehydroepiandrosterone (DHEA) levels, and this decrease parallels an increasing severity of hepatic fibrosis [bib_ref] Endocrine and liver interaction: The role of endocrine pathways in NASH, Loria [/bib_ref] [bib_ref] Prediction of prevalent but not incident non-alcoholic fatty liver disease by levels..., Seo [/bib_ref] [bib_ref] Low circulating levels of dehydroepiandrosterone in histologically advanced nonalcoholic fatty liver disease, Charlton [/bib_ref] DHEA modulates oxidative stress sensitivity, as it is a potential mediator of reactive oxygen species scavengers, enhances insulin sensitivity, and increases peroxisome proliferator-activated receptor α expression [bib_ref] Peroxisome proliferator-activated receptor alpha required for gene induction by dehydroepiandrosterone-3 beta-sulfate, Peters [/bib_ref]. Serum DHEA levels might be used for the identification of patients at risk for the development of NASH with advanced fibrosis [bib_ref] Low circulating levels of dehydroepiandrosterone in histologically advanced nonalcoholic fatty liver disease, Charlton [/bib_ref].
Additionally, the literature advocates that hypogonadism is associated with important cardiovascular risk factors, such as general and visceral obesity, impaired insulin sensibility, hypertension, and dyslipidemia [bib_ref] Endocrine and liver interaction: The role of endocrine pathways in NASH, Loria [/bib_ref]. These are also crucial contributors to the development of NAFLD and can be potential confounders in some studies in this topic. A study by Sakr et al. evidenced that male rats with decreased gonadal hormones have higher serum and hepatic levels of triglycerides, cholesterol, and LDL and have increased levels of enzymes involved in hepatic steatogenesis, such as sterol regulatory element-binding transcription factor 1 (SREBP-1) and 2 (SREBP-2) [bib_ref] Possible mechanisms underlying fatty liver in a rat model of male hypogonadism:..., Sakr [/bib_ref]. Androgen receptor knockout male mice acquire late-onset obesity and have an increased risk of NAFLD development and progression [bib_ref] Androgen receptor null male mice develop late-onset obesity caused by decreased energy..., Fan [/bib_ref]. Hormonal replacement therapy (HRT) with testosterone seems to ameliorate these changes and may be considered a protective strategy to be taken into account [bib_ref] Long-term testosterone therapy in hypogonadal men ameliorates elements of the metabolic syndrome:..., Traish [/bib_ref]. Clinical studies show that long-term testosterone replacement improves MS components and ameliorates liver enzymes changes in men with hypogonadism [bib_ref] Long-term testosterone therapy in hypogonadal men ameliorates elements of the metabolic syndrome:..., Traish [/bib_ref].
Regarding women, the most common cause of primary hypogonadism is Turner syndrome. This syndrome is characterized by a premature ovarian failure that causes estrogen deficiency [bib_ref] Turner syndrome: A pediatric diagnosis frequently made by non-pediatricians, Carvalho [/bib_ref]. Hypoestrogenism has an important role in some of the most common comorbidities of these patients, such as bone mineralization disorders and hepatic dysfunction, with increased hepatic enzymes [bib_ref] Hormone replacement therapy may improve hepatic function in women with Turner's syndrome, Elsheikh [/bib_ref] [bib_ref] Liver dysfunction in Turner's syndrome: Prevalence, natural history and effect of exogenous..., Koulouri [/bib_ref]. Besides other benefits, HRT may improve hepatic function and other components of metabolic dysfunction in women with Turner's syndrome. Animal studies are also in line with this association. A study showed that estrogen deficiency worsens the hepatic histological injury in mice fed a high-fat and high-cholesterol diet [bib_ref] Estrogen deficiency worsens steatohepatitis in mice fed high-fat and high-cholesterol diet, Kamada [/bib_ref]. In mice fed a high-fat and high-cholesterol diet, lower estrogen levels were associated with higher levels of macrophage infiltration and the enhanced expression of hepatic inflammatory genes, and HRT was able to reverse these changes [bib_ref] Estrogen deficiency worsens steatohepatitis in mice fed high-fat and high-cholesterol diet, Kamada [/bib_ref]. Both human and animal studies evidence the benefits of HRT on hepatic function and other components of metabolic dysfunction [bib_ref] Hormone replacement therapy may improve hepatic function in women with Turner's syndrome, Elsheikh [/bib_ref] [bib_ref] Nonnuclear Estrogen Receptor Activation Improves Hepatic Steatosis in Female Mice, Chambliss [/bib_ref]
## Impact of menopause on liver disease
As discussed throughout this manuscript, an imbalance in estrogen levels might have a pronounced impact on hepatic homeostasis. Menopause is a physiological condition of estrogen deficiency with an important impact on women's health. The risk of development and progression of NAFLD increases with the duration of estrogen deficiency [bib_ref] A longer duration of estrogen deficiency increases fibrosis risk among postmenopausal women..., Klair [/bib_ref]. Accordingly, women with premature menopause are at increased risk of severe liver fibrosis [bib_ref] A longer duration of estrogen deficiency increases fibrosis risk among postmenopausal women..., Klair [/bib_ref]. Thus, both menopausal status and menopause onset age should be taken into account when determining fibrosis risk among women with NAFLD. An epidemiological study evidenced a higher prevalence of NAFLD in postmenopausal female patients when compared with men [bib_ref] Increase in the prevalence of fatty liver in Japan over the past..., Kojima [/bib_ref]. Another study found a prevalence of NAFLD of approximately 60% in postmenopausal women, compared to 32% in premenopausal patients [bib_ref] Prevalence of non alcoholic fatty liver disease in premenopausal, posmenopausal and polycystic..., Gutierrez-Grobe [/bib_ref]. The increased risk of NAFLD is mainly caused by estrogen deficiency, similar to women with oophorectomy or premature ovarian failure [bib_ref] Surgical menopause and increased risk of nonalcoholic fatty liver disease in endometrial..., Matsuo [/bib_ref]. Moreover, women with NAFLD have a higher risk of fibrosis progression as they age [bib_ref] Independent predictors of liver fibrosis in patients with nonalcoholic steatohepatitis, Angulo [/bib_ref]. Postmenopausal women have a higher severity of fibrosis at any given hepatocyte ballooning and portal inflammation [bib_ref] Gender and menopause impact severity of fibrosis among patients with nonalcoholic steatohepatitis, Yang [/bib_ref]. The fat redistribution associated with menopause increases the risk of insulin resistance, dyslipidemia, hypertension, and diabetes and, consequently, may increase the risk of NAFLD [bib_ref] Nonalcoholic fatty liver disease in women, Suzuki [/bib_ref]. Among post-menopausal women, HRT is associated with a reduced risk of NAFLD and fibrosis progression. The administration of HRT among post-menopausal women appears to be protective against NAFLD development, but whether it affects fibrosis progression is still unclear [bib_ref] A longer duration of estrogen deficiency increases fibrosis risk among postmenopausal women..., Klair [/bib_ref]. Further studies evaluating the potential impact of HRT on hepatic fibrosis in postmenopausal women are needed, including the adequate dose, route of administration, duration, and age of initiation.
## Polycystic ovary syndrome (pcos)
Concerning diseases with increased levels of sex hormones, it is important to address PCOS. Recent guidelines define PCOS as a clinical and/or biochemical hyperandrogenism, chronic oligo-anovulation, and polycystic ovarian morphology [bib_ref] Revised 2003 consensus on diagnostic criteria and long-term health risks related to..., Rotterdam [/bib_ref]. The diagnosis requires two out of three of these criteria after the exclusion of other endocrine disorders [bib_ref] Revised 2003 consensus on diagnostic criteria and long-term health risks related to..., Rotterdam [/bib_ref]. A high percentage of women with PCOS present with obesity and MS [bib_ref] Prevalence, phenotype and cardiometabolic risk of polycystic ovary syndrome under different diagnostic..., Yildiz [/bib_ref]. Evidence suggests that the prevalence of NAFLD is increased in women with PCOS, regardless of weight and MS. The prevalence of NAFLD in women with PCOS is 35 to 70%, compared with 20 to 30% in age-and body mass index (BMI)-matched control women. PCOS is a prevalent condition among patients with biopsy-confirmed NAFLD (approximately 50-70%), and the risk of NAFLD development in POCS patients is two-fold higher compared with control women [bib_ref] A potential link between polycystic ovary syndrome and non-alcoholic fatty liver disease:..., Wu [/bib_ref]. Women with POCS are also more likely to have more severe histological features, such as NASH, advanced fibrosis, and cirrhosis [bib_ref] Nonalcoholic steatohepatitis and nonalcoholic Fatty liver disease in young women with polycystic..., Setji [/bib_ref]. The pathophysiological mechanism underlying the increased risk of NAFLD in PCOS is thought to be multi-factorial. Both genetic and acquired factors are involved, with contributions from abdominal adiposity, systemic insulin resistance, chronic inflammation, and hyperandrogenism. Whether NAFLD is associated with POCS as a consequence of shared risk factors or PCOS independently supports NAFLD development remains to be elucidated. Hyperandrogenism may be considered a central contributor to NAFLD development [bib_ref] A potential link between polycystic ovary syndrome and non-alcoholic fatty liver disease:..., Wu [/bib_ref] [bib_ref] Polycystic ovary syndrome, androgen excess, and the risk of nonalcoholic fatty liver..., Kumarendran [/bib_ref] [bib_ref] Role of Androgen in Liver Fat Content in Women: Metabolically Advantageous or..., Wang [/bib_ref] Jones et al. evidenced that women with POCS and higher androgen levels have greater intra-hepatic fat content compared with women with POCS and lower androgen levels [bib_ref] Polycystic ovary syndrome with hyperandrogenism as a risk factor for non-obese non-alcoholic..., Kim [/bib_ref]. A recent meta-analysis showed that women with POCS but without hyperandrogenism were not associated with an increased prevalence of NAFLD [bib_ref] A potential link between polycystic ovary syndrome and non-alcoholic fatty liver disease:..., Wu [/bib_ref]. Increased levels of testosterone in women can enhance lipogenic gene expression and de novo lipogenesis in hepatocytes. Therefore, the liver injuries seen in NAFLD may also be caused by lipotoxicity, which is commonly present in women with POCS [bib_ref] MECHANISMS IN ENDOCRINOLOGY: The sexually dimorphic role of androgens in human metabolic..., Schiffer [/bib_ref]. Qu et al. stated that women with PCOS and NAFLD had higher BMI, waist/hip ratio, worse insulin resistance, increased triglycerides, and lower HDL cholesterol levels than women with POCS without NAFLD. Moreover, co-existing NAFLD seems to worsen the metabolic profile of women with POCS [bib_ref] The clinical characteristics and etiological study of nonalcoholic fatty liver disease in..., Qu [/bib_ref].
Current guidelines for the treatment of POCS consist of lifestyle intervention. When a pharmacologic approach is needed to improve the metabolic profile, metformin remains the drug of choice [bib_ref] Diagnosis and treatment of polycystic ovary syndrome: An Endocrine Society clinical practice..., Legro [/bib_ref]. However, metformin does not seem to affect liver histology in patients with NAFLD and PCOS [bib_ref] Effect of liraglutide on ectopic fat in polycystic ovary syndrome: A randomized..., Frossing [/bib_ref]. Oral contraceptive pills (OCP) increase serum lipid levels and, consequently, might worsen NAFLD [bib_ref] Hormonal contraception, Baird [/bib_ref]. Interestingly, according to Liu SH. et al., women using OCP have a lower incidence of NAFLD, but this might be mediated or confounded by adiposity [bib_ref] Oral contraceptive pill use is associated with reduced odds of nonalcoholic fatty..., Liu [/bib_ref]. There are no studies evaluating the impact of OCP on the liver profile in women with PCOS. Preliminary results show that liraglutide and other glucagon-like peptide-1 receptor agonists can decrease the intra-hepatic fat content and visceral adipose tissue among obese women with PCOS [bib_ref] Effect of liraglutide on ectopic fat in polycystic ovary syndrome: A randomized..., Frossing [/bib_ref]. Additionally, the prevalence of NAFLD was reduced by two thirds in obese women with PCOS treated with liraglutide [bib_ref] Effect of liraglutide on ectopic fat in polycystic ovary syndrome: A randomized..., Frossing [/bib_ref]. More accurate diagnostic methods, such as liver biopsy, in NASH patients have evidenced a reduction in hepatic inflammation and the prevalence of NASH during liraglutide treatment. However, there is still not enough evidence to support the use of these drugs in these patients. More studies are needed to evaluate whether treatment with anti-androgenic drugs may reduce the risk of NAFLD in women with PCOS.
## Important sex differences in nafld development and severity
As mentioned earlier, sex hormones have an important impact on the pathophysiology and progression of NAFLD and should be taken into account when studying this topic [bib_ref] Sexual Dimorphism of NAFLD in Adults. Focus on Clinical Aspects and Implications..., Lonardo [/bib_ref] [bib_ref] Sex Differences in Nonalcoholic Fatty Liver Disease: State of the Art and..., Lonardo [/bib_ref]. Sex differences and other factors, such as dietary patterns, exercise, and quality of life, are equally important to consider when studying NAFLD [bib_ref] Gender Differences in Exercise Habits and Quality of Life Reports: Assessing the..., Craft [/bib_ref]. The understanding of the impact of sex in NAFLD remains scarce. Only a few studies describe sex differences in NAFLD, even though some of the major risk factors for NAFLD are profoundly impacted by sex. This hepatic disorder is more prevalent and has worse histological features in men at reproductive ages. The difference fades after menopause, when the prevalence of NAFLD in women equals or exceeds that of men, as discussed earlier [bib_ref] Gender and menopause impact severity of fibrosis among patients with nonalcoholic steatohepatitis, Yang [/bib_ref]. Moreover, sex differences have a significant impact on hepatic outcomes in patients with biopsy-confirmed NAFLD and advanced hepatic fibrosis, with a higher incidence of HCC and a worse survival rate in male patients compared with premenopausal women [bib_ref] Fibrosis Severity as a Determinant of Cause-Specific Mortality in Patients With Advanced..., Vilar-Gomez [/bib_ref]. Men with NASH cirrhosis have an almost 2-fold to 7-fold higher risk of having HCC than women [bib_ref] The incidence and risk factors of hepatocellular carcinoma in patients with nonalcoholic..., Ascha [/bib_ref]. It seems that men develop HCC at earlier liver fibrosis stages than women, which may explain the higher prevalence of HCC in men [bib_ref] Characteristics of patients with nonalcoholic steatohepatitis who develop hepatocellular carcinoma, Yasui [/bib_ref]. Finally, animal studies evidenced that male mice have innate immune cells that promote liver inflammation and fibrosis, while macrophages from female mice have a more protective and antifibrotic phenotype [bib_ref] Sex-based differences in immune function and responses to vaccination, Klein [/bib_ref]. In fact, Kupffer cells from male animals produce more IL-6, which increases inflammation, tissue injury, and subsequent hepatocyte proliferation when compared with female animals. Furthermore, estrogen has an inhibitory effect on IL-6 production, stellate cell activation, and fibrogenesis [bib_ref] Gender disparity in liver cancer due to sex differences in MyD88-dependent IL-6..., Naugler [/bib_ref]. Stellate cells incubated with estrogen have a significant decrease in collagen synthesis, which partially explains the delay in the progression of liver fibrosis in pre-menopausal or post-menopausal women treated with HRT [bib_ref] Opposing effects of oestradiol and progesterone on intracellular pathways and activation processes..., Itagaki [/bib_ref]. Evidence proves that males and females respond differently to lifestyle modifications and drug therapies [bib_ref] Men and women respond differently to rapid weight loss: Metabolic outcomes of..., Christensen [/bib_ref]. Therefore, a patient's sex and hormonal status are crucial factors to take into account when studying NAFLD pathophysiology and management.
Regardless of menopausal status, serum adiponectin is higher in women than men [bib_ref] The role of adiponectin in the pathogenesis and treatment of non-alcoholic fatty..., Polyzos [/bib_ref] [bib_ref] Menopause impacts the relation of plasma adiponectin levels with the metabolic syndrome, Henneman [/bib_ref]. Serum adiponectin levels are inversely associated with body weight and central obesity and have been proposed as a biomarker of MS [bib_ref] Adiponectin in insulin resistance: Lessons from translational research, Ziemke [/bib_ref]. Data regarding the beneficial effects of adiponectin on NAFLD are well documented. Adiponectin can reduce insulin resistance and has anti-steatotic and anti-inflammatory effects [bib_ref] Adiponectin and leptin in the diagnosis and therapy of NAFLD, Boutari [/bib_ref]. There is an association between lower levels of adiponectin and an increased risk and severity of NAFLD [bib_ref] Adiponectin and leptin in the diagnosis and therapy of NAFLD, Boutari [/bib_ref]. Although the causality of this association is difficult to establish, adiponectin may be considered a biomarker of NAFLD progression to NASH and cirrhosis [bib_ref] Serum total adiponectin in nonalcoholic fatty liver disease: A systematic review and..., Polyzos [/bib_ref]. A recent systematic review showed that the administration of pioglitazone increases serum adiponectin levels and improves histological features in NASH patients [bib_ref] Adiponectin as a target for the treatment of nonalcoholic steatohepatitis with thiazolidinediones:..., Polyzos [/bib_ref]. Although pioglitazone has been associated with an increased risk of heart failure, the use of selective PPARγ modulators may be promising [bib_ref] Thiazolidinedione use, fluid retention, and congestive heart failure: A consensus statement from..., Elasy [/bib_ref]. A deeper knowledge of the dynamic crosstalk between adipokines and the liver may result in a better and personalized treatment option in the future.
## Adrenal gland disorders and nafld
## Renin-angiotensin-aldosterone system (raas)
While there are some studies in animal models and/or patients with primary aldosteronism, most research has been conducted in the context of hyperstimulation of the RAAS. Kumagai et al. found that hyperaldosteronism results in the development of insulin resistance in patients with previously normal insulin metabolism (10-year follow-up) [bib_ref] Plasma aldosterone levels and development of insulin resistance: Prospective study in a..., Kumagai [/bib_ref]. Activation of the RAAS leads to altered insulin/IGF-1 signaling pathways in several tissues, namely, the liver [bib_ref] Renin-angiotensin-aldosterone system and oxidative stress in cardiovascular insulin resistance, Cooper [/bib_ref]. Local hepatic increased insulin resistance may lead to inadequate lipid accumulation and, eventually, to NAFLD [bib_ref] Insulin resistance and risk of chronic kidney disease in nondiabetic US adults, Chen [/bib_ref]. The association between an inappro-priately active RAAS and insulin resistance in the setting of the MS is an area of growing interest [bib_ref] Role of aldosterone and angiotensin II in insulin resistance: An update, Lastra-Lastra [/bib_ref]. Although aldosterone has been more extensively studied in the context of NAFLD, growing evidence suggests other hormones closely related to aldosterone may have significant and possibly independent roles in the development of NAFLD. Both angiotensin II and the angiotensin II type 1 receptor and mineralocorticoids and mineralocorticoids receptor activation contribute to insulin signaling attenuation [bib_ref] Renin-angiotensin-aldosterone system and oxidative stress in cardiovascular insulin resistance, Cooper [/bib_ref] [bib_ref] Role of aldosterone and angiotensin II in insulin resistance: An update, Lastra-Lastra [/bib_ref]. Moreover, RAAS-mediated reactive oxygen species formation not only appears to play a vital role in insulin signaling impairment but also leads to endothelial dysfunction, which contributes to hypertension, atherosclerosis, chronic kidney disease, and cardiovascular disease [bib_ref] Renin-angiotensin-aldosterone system and oxidative stress in cardiovascular insulin resistance, Cooper [/bib_ref] [bib_ref] Role of aldosterone and angiotensin II in insulin resistance: An update, Lastra-Lastra [/bib_ref]. Additionally, even ionic disturbances, such as the hypokalemia that is found in overactive RAAS states, seem to contribute to NAFLD [bib_ref] Non-Alcoholic Fatty Liver Disease and Hypokalemia in Primary Aldosteronism Among Chinese Population, Chen [/bib_ref]. Whether they do so in a synergistic or independent manner remains unknown.
Noguchi et al. showed that selective aldosterone blocker (SAB) treatment inhibited liver fibrogenesis and carcinogenesis by suppressing neovascularization and activated hepatic stellate cells in a NASH rat model [bib_ref] Selective aldosterone blocker ameliorates the progression of non-alcoholic steatohepatitis in rats, Noguchi [/bib_ref]. Polyzos et al. showed in a preliminary report that spironolactone improved insulin resistance in patients with NAFLD [bib_ref] Effect of spironolactone and vitamin E on serum metabolic parameters and insulin..., Polyzos [/bib_ref]. Another similar study by Pizarro et al. found that eplerenone effectively ameliorated histological steatosis and hepatic fibrosis in a mouse model of NASH, corroborating the aforementioned evidence [bib_ref] Beneficial effects of mineralocorticoid receptor blockade in experimental non-alcoholic steatohepatitis, Pizarro [/bib_ref]. Hence, aldosterone plays an important role in the progression of NASH and SAB, which are widely used and safely represent a potential new therapeutic strategy for NAFLD/NASH. Spironolactone may have anti-inflammatory and anti-fibrogenic effects on the liver and, therefore, may be an inexpensive and effective therapeutic target for NAFLD. However, a large-scale human trial is needed to further explore this hypothesis.
Most of the evidence sustaining that aldosterone has an important role in the development of NAFLD is based on its effects on insulin resistance. One study examined aldosterone's direct and independent effects in NAFLD. Kumar et al. found a positive association between serum aldosterone levels and fatty liver in African American women [bib_ref] Fatty Liver Disease, Women, and Aldosterone: Finding a Link in the Jackson..., Kumar [/bib_ref]. The association was independent of alcohol intake and persisted after insulin resistance and high-sensitivity C-reactive protein were added to the multivariable analysis [bib_ref] Fatty Liver Disease, Women, and Aldosterone: Finding a Link in the Jackson..., Kumar [/bib_ref]. These results suggest that the effect of aldosterone in fatty liver is partially independent of insulin resistance and high-sensitivity C-reactive protein. Although the results are promising, a few limitations of the study should be considered, namely, the inability to prove causality and the need to study different ethnicities and to clarify sex differences with larger samples. Similarly, a study by Lu et al. was able to isolate the angiotensinogen effects from those of angiotensin II and renin and found it contributes to body weight gain and liver steatosis.
Angiotensin II receptor blockers are generally considered to have no significant metabolic effects. However, experimental and clinical evidence suggest telmisartan is an exception. Besides blocking the angiotensin II type 1 receptor, telmisartan is a selective PPARγ modulator [bib_ref] Treating the metabolic syndrome: Telmisartan as a peroxisome proliferator-activated receptor-gamma activator, Kurtz [/bib_ref]. Telmisartan's selectiveness results in the well-established therapeutic benefits of PPARγ modulation in lipid and glucose metabolism without the undesirable side effects of conventional PPARγ activators, such as fluid retention, weight gain, and edema [bib_ref] Treating the metabolic syndrome: Telmisartan as a peroxisome proliferator-activated receptor-gamma activator, Kurtz [/bib_ref]. Moreover, the routine hypertension treatment dosage is enough to obtain the aforementioned effects [bib_ref] Treating the metabolic syndrome: Telmisartan as a peroxisome proliferator-activated receptor-gamma activator, Kurtz [/bib_ref]. Hence, telmisartan could play an important role in the prevention and treatment of MS, diabetes, and atherosclerosis [bib_ref] Treating the metabolic syndrome: Telmisartan as a peroxisome proliferator-activated receptor-gamma activator, Kurtz [/bib_ref]. A randomized controlled study by Hirata et al. explored the effect of telmisartan and losartan in improving steatosis in hypertensive NAFLD patients with type 2 diabetes [bib_ref] Effect of Telmisartan or Losartan for Treatment of Nonalcoholic Fatty Liver Disease:..., Hirata [/bib_ref]. While significant liver function improvement was not found in either group, the serum free fatty acid levels were significantly decreased in the telmisartan group. In addition, telmisartan also improved the liver/spleen ratio. By improving hepatic fat deposition, telmisartan may be a potential new therapeutic strategy for NAFLD [bib_ref] Effect of Telmisartan or Losartan for Treatment of Nonalcoholic Fatty Liver Disease:..., Hirata [/bib_ref]. Yokohama et al. showed that treatment with losartan improved not only serum liver enzyme levels but also reduced hepatic necroinflammation and fibrosis in patients with NASH [bib_ref] Therapeutic efficacy of an angiotensin II receptor antagonist in patients with nonalcoholic..., Yokohama [/bib_ref]. More specifically, losartan decreased transforming growth factor beta 1(TGF-β1) plasma levels, a known marker of hepatic fibrosis. The authors of the study hypothesized that angiotensin II receptor antagonists may be a safe and efficacious therapeutic option for NASH [bib_ref] Therapeutic efficacy of an angiotensin II receptor antagonist in patients with nonalcoholic..., Yokohama [/bib_ref]. showed that patients with primary hyperaldosteronism and hypokalemia were at a greater risk of developing metabolic and liver disease [bib_ref] Nonalcoholic fatty liver disease in primary aldosteronism: A pilot study, Fallo [/bib_ref]. A higher prevalence of NAFLD and insulin resistance was found in the hypokalemic patient subgroup compared to those with normokalemia [bib_ref] Nonalcoholic fatty liver disease in primary aldosteronism: A pilot study, Fallo [/bib_ref]. Sun et al. found that hypokalemia was independently associated with the prevalence of MS in a large Chinese population and hypothesized that potassium-correcting therapies could alter the progression of this disease [bib_ref] Serum potassium level is associated with metabolic syndrome: A population-based study, Sun [/bib_ref].
## Cushing's syndrome and nafld
Glucocorticoids (GC) are produced by the adrenal gland under the control of ACTH secreted by pituitary. The effects of GC on lipid metabolism, fat accumulation, and NAFLD development are complex [fig_ref] Figure 5: Impact of cortisol imbalance and NAFLD pathogenesis [/fig_ref]. Furthermore, hepatic dysfunction may impair GC metabolism and alter the adrenal axis. A small study with 50 patients reported NAFLD in 20% of patients with Cushing's syndrome, results similar to a retrospective study with a prevalence between 26 and 33% [bib_ref] Hepatic steatosis in Cushing's syndrome: A radiological assessment using computed tomography, Rockall [/bib_ref] [bib_ref] Demographic Characteristics, Etiology, and Comorbidities of Patients with Cushing's Syndrome: A 10-Year..., Zhou [/bib_ref]. Additionally, Rockall et al. found that 20% of patients with Cushing's syndrome had hepatic steatosis in a computed tomography scan [bib_ref] Hepatic steatosis in Cushing's syndrome: A radiological assessment using computed tomography, Rockall [/bib_ref]. Contrariwise, patients with NAFLD did not present with Cushing's syndrome, though some abnormalities of GC metabolism, such as increased urinary free cortisol and decreased dexamethasone suppression of plasma cortisol, have been shown [bib_ref] Relationship of nonalcoholic hepatic steatosis to overnight low-dose dexamethasone suppression test in..., Zoppini [/bib_ref] [bib_ref] Associations between liver histology and cortisol secretion in subjects with nonalcoholic fatty..., Targher [/bib_ref]. GC may contribute to the development of NAFLD through their lipolytic effects on adipose tissue, which result in more readily available free fatty acid for liver uptake [bib_ref] Non-alcoholic fatty liver disease in common endocrine disorders, Hazlehurst [/bib_ref]. The association between GC and hepatic lipid accumulation has been observed in both clinical and basic research. Auer et al. found a daily hydrocortisone dose to be positively associated with hepatic lipid accumulation in humans [bib_ref] Investigating the role of cortisol and growth hormone in fatty liver development:..., Auer [/bib_ref] , and D'Souza A et al. showed that exogenous corticosterone treatment in rats induced hepatic steatosis [bib_ref] Consumption of a high-fat diet rapidly exacerbates the development of fatty liver..., D'souza [/bib_ref]. Targher et al. found chronic hypothalamic-pituitary-adrenal (HPA) axis hyperactivity and subclinical hypercortisolemia in NAFLD patients [bib_ref] Associations between liver histology and cortisol secretion in subjects with nonalcoholic fatty..., Targher [/bib_ref]. In a more recent study, Hayashi et al. hypothesized that activation of adipocyte GC receptors by high levels of GC in patients with Cushing's syndrome generates a cascade of biomolecular events, culminating in hepatic steatosis and insulin resistance [bib_ref] Adipocyte GR Inhibits Healthy Adipose Expansion Through Multiple Mechanisms in Cushing Syndrome, Hayashi [/bib_ref]. By accelerating lipolysis and restricting healthy adipose expansion, adipocyte GC receptor activation leads to ectopic lipid accumulation and insulin resistance. This study suggests that the adipocyte GC receptor is a promising target for the treatment of metabolic dysfunction/diseases present in Cushing's syndrome [bib_ref] Adipocyte GR Inhibits Healthy Adipose Expansion Through Multiple Mechanisms in Cushing Syndrome, Hayashi [/bib_ref]. It is important to acknowledge that cortisol's influence on adipose tissue is dependent on insulin status. During fasting, insulin levels are low, and cortisol stimulates lipolysis. Paradoxically, when insulin levels are high, cortisol stimulates lipogenesis [bib_ref] Hepatic lipid accumulation: Cause and consequence of dysregulated glucoregulatory hormones, Geisler [/bib_ref]. Furthermore, it has been hypothesized that sympathetic nervous system activation may underlie the hypercortisolemia commonly observed in NAFLD patients through the stimulation of cortisol release [bib_ref] Hepatic lipid accumulation: Cause and consequence of dysregulated glucoregulatory hormones, Geisler [/bib_ref].
One enzyme of the GC cascade has sparked great interest in this field. 11βhydroxysteroid dehydrogenase type 1 (11β-HSD1) catalyzes the conversion of cortisone into cortisol. Chronic hypercortisolemia, obesity, and NAFLD increase local adipose tissue levels of cortisol by upregulating the expression of 11β-HSD [bib_ref] Overexpression of 11beta-hydroxysteroid dehydrogenase type 1 in visceral adipose tissue and portal..., Candia [/bib_ref]. Studies have shown a decrease in hepatic steatosis when 11β-HSD1 is pharmacologically inhibited or genetically absent [bib_ref] Carbenoxolone treatment ameliorated metabolic syndrome in WNIN/Ob obese rats, but induced severe..., Sakamuri [/bib_ref] [bib_ref] Effects of antisense-mediated inhibition of 11beta-hydroxysteroid dehydrogenase type 1 on hepatic lipid..., Li [/bib_ref]. The role of A-ring reductases (5aR and 5bR), which inactivate cortisol, has not yet been thoroughly studied [bib_ref] Non-alcoholic fatty liver disease in common endocrine disorders, Hazlehurst [/bib_ref]. However, a study found an increased risk of hepatic steatosis with the deletion of 5aR type 1 [bib_ref] Andrew, R. 5alpha-Reductase type 1 deficiency or inhibition predisposes to insulin resistance,..., Livingstone [/bib_ref].
Interestingly, one study by Ahmed et al. has defined two seemingly protective phases of altered hepatic cortisol metabolism in progressive NAFLD [bib_ref] A switch in hepatic cortisol metabolism across the spectrum of non alcoholic..., Ahmed [/bib_ref]. In steatosis, increased cortisol clearance leads to lower local levels of this hormone, consequently preserving the hepatic metabolic phenotype and limiting lipid accumulation. On the other hand, increased cortisol regeneration and, therefore, higher local cortisol levels are present in NASH, possibly to limit hepatic inflammation. This distinction is particularly pertinent when looking at 11β-HSD1 as a potential therapeutic target. Inhibition of 11β-HSD1 might be favorable in steatosis since it would further reduce local levels of cortisol. However, 11β-HSD1 inhibition in NASH could be detrimental, as it would worsen the inflammatory response. Therefore, the histological stage of NAFLD may dictate whether 11β-HSD1 inhibition is beneficial.
## Pheochromocytoma
Human primary hepatic stellate cells are dependent on catecholamines for their survival and fibrogenic effects [bib_ref] Sympathetic nervous system catecholamines and neuropeptide Y neurotransmitters are upregulated in human..., Sigala [/bib_ref]. Sigala et al. studied these cells and reported an upregulation of fibrogenic α/β-adrenoreceptor and neuropeptide Y receptors in human cirrhotic NAFLD [bib_ref] Sympathetic nervous system catecholamines and neuropeptide Y neurotransmitters are upregulated in human..., Sigala [/bib_ref]. Furthermore, the authors suggested adrenoreceptor and neuropeptide Y antagonists as potential anti-fibrotic agents for NAFLD. Although the effects of excessive systemic catecholamines levels are still widely unknown, one can speculate that patients with pheochromocytoma may be at higher risk of NAFLD based on the results of this study. More studies are necessary to further clarify the role of catecholamines and their related disorders in NAFLD.
# Conclusions
The prevalence of NAFLD is rising and the understanding of its pathophysiology is crucial for better management of this condition. It is biologically plausible that there is a link between several endocrine disorders and NAFLD, other than the typically known type 2 diabetes mellitus, obesity, and MS. Namely, our review presents that there is evidence supporting that hyperprolactinemia, excessive or defective GH, hypothyroidism, hypercortisolism, activation of the RAAS, PCOS, and hypogonadism are endocrine disorders that exert a negative/facilitative effect on NAFLD initiation or progression. The contribution of the other endocrine entities that we explored remain fairly controversial or uncertain. [fig_ref] Table 1: Effect of endocrine hormones on the different factors involved in nonalcoholic fat... [/fig_ref] and [fig_ref] Figure 6: Summary figure of the crosstalk between the endocrine axes and NAFLD [/fig_ref] summarize the main conclusions of the studies included in this review.
It is important to keep in mind when seeing patients with endocrine diseases that NAFLD may develop or that a pre-existing condition may exacerbate it with immeasurable consequences. Furthermore, NAFLD is a heterogeneous metabolic disorder, and we agree that the recognition of such associations is crucial towards the implementation of a tailored approach to our patients [bib_ref] Insights into Nonalcoholic Fatty-Liver Disease Heterogeneity, Arrese [/bib_ref] [bib_ref] Perspectives on Precision Medicine Approaches to NAFLD Diagnosis and Management, Lonardo [/bib_ref]. As described throughout the text, a better
## Pheochromocytoma
Human primary hepatic stellate cells are dependent on catecholamines for their survival and fibrogenic effects [bib_ref] Sympathetic nervous system catecholamines and neuropeptide Y neurotransmitters are upregulated in human..., Sigala [/bib_ref]. Sigala et al. studied these cells and reported an upregulation of fibrogenic α/β-adrenoreceptor and neuropeptide Y receptors in human cirrhotic NAFLD [bib_ref] Sympathetic nervous system catecholamines and neuropeptide Y neurotransmitters are upregulated in human..., Sigala [/bib_ref]. Furthermore, the authors suggested adrenoreceptor and neuropeptide Y antagonists as potential anti-fibrotic agents for NAFLD. Although the effects of excessive systemic catecholamines levels are still widely unknown, one can speculate that patients with pheochromocytoma may be at higher risk of NAFLD based on the results of this study. More studies are necessary to further clarify the role of catecholamines and their related disorders in NAFLD.
# Conclusions
The prevalence of NAFLD is rising and the understanding of its pathophysiology is crucial for better management of this condition. It is biologically plausible that there is a link between several endocrine disorders and NAFLD, other than the typically known type 2 diabetes mellitus, obesity, and MS. Namely, our review presents that there is evidence supporting that hyperprolactinemia, excessive or defective GH, hypothyroidism, hypercortisolism, activation of the RAAS, PCOS, and hypogonadism are endocrine disorders that exert a negative/facilitative effect on NAFLD initiation or progression. The contribution of the other endocrine entities that we explored remain fairly controversial or uncertain. [fig_ref] Table 1: Effect of endocrine hormones on the different factors involved in nonalcoholic fat... [/fig_ref] and [fig_ref] Figure 6: Summary figure of the crosstalk between the endocrine axes and NAFLD [/fig_ref] summarize the main conclusions of the studies included in this review. Legend: NAFLD: Nonalcoholic fatty liver disease; NASH: Nonalcoholic steatohepatitis; IGF1: Insulin Like Growth Factor 1; RAAS: renin-angiotensin-aldosterone system; TGF-1: transforming growth factor beta 1; ROS: reactive oxygen species; IL-6: Interleukin 6. Text in italic represents controversial results. gender: ♀female, ♂male.
understanding of the relationship between these hormonal imbalances and the development and progression of NAFLD can lead to treatment advances. Nevertheless, controversial and insufficient evidence in this area of knowledge preclude us from drawing definite conclusions. Prospective and well-designed studies are lacking but are crucial and must be carried out in the near future for a better understanding and treatment of patients with NAFLD. It is important to keep in mind when seeing patients with endocrine diseases that NAFLD may develop or that a pre-existing condition may exacerbate it with immeasurable consequences. Furthermore, NAFLD is a heterogeneous metabolic disorder, and we agree that the recognition of such associations is crucial towards the implementation of a tailored approach to our patients [bib_ref] Insights into Nonalcoholic Fatty-Liver Disease Heterogeneity, Arrese [/bib_ref] [bib_ref] Perspectives on Precision Medicine Approaches to NAFLD Diagnosis and Management, Lonardo [/bib_ref]. As described throughout the text, a better understanding of the relationship between these hormonal imbalances and the development and progression of NAFLD can lead to treatment advances. Nevertheless, controversial and insufficient evidence in this area of knowledge preclude us from drawing definite conclusions. Prospective and well-designed studies are lacking but are crucial and must be carried out in the near future for a better understanding and treatment of patients with NAFLD.
[fig] Figure 1: Possible mechanisms of GH deficiency on NALFD development and progression. NAFLD: Nonalcoholic fatty liver disease; NASH: Nonalcoholic steatohepatitis; GH: Growth hormone; IGF1: Insulin-like growth factor 1. [Viktoriia Kasyanyuk] © 123RF.com (accessed on 22 February 2022) and Servier Medical Art by Servier. [/fig]
[fig] Figure 2: Impact of prolactin levels on NAFLD development and progression. NAFLD: Nonalcoholic fatty liver disease; NASH: Nonalcoholic steatohepatitis; GH: Growth hormone; IGF1: Insulinlike growth factor 1. [Viktoriia Kasyanyuk] © 123RF.com (accessed on 22 February 2022) and Servier Medical Art by Servier. [/fig]
[fig] Figure 3: Thyroid-liver axis: The impact of thyroid hormones on liver. NAFLD: Nonalcoholic fatty liver disease; NASH: Nonalcoholic steatohepatitis; TH: Thyroid hormone. [Viktoriia Kasyanyuk] © 123RF.com (accessed on 22 February 2022) and Servier Medical Art by Servier. [/fig]
[fig] Figure 4: Sexual hormones and NAFLD. NAFLD: Nonalcoholic fatty liver disease; NASH: Nonalcoholic steatohepatitis; PCOS: Polycystic ovary syndrome. [Viktoriia Kasyanyuk] © 123RF.com (accessed on 22 February 2022) and Servier Medical Art by Servier. [/fig]
[fig] Figure 5: Impact of cortisol imbalance and NAFLD pathogenesis. NAFLD: Nonalcoholic fatty liver disease; NASH: Nonalcoholic steatohepatitis. [Viktoriia Kasyanyuk] © 123RF.com (accessed on 22 February 2022) and Servier Medical Art by Servier. [/fig]
[fig] Figure 6: Summary figure of the crosstalk between the endocrine axes and NAFLD. NAFLD: Nonalcoholic fatty liver disease; NASH: Nonalcoholic steatohepatitis; GH: Growth hormone; IGF1: Insulin-like growth factor 1; ADH: antidiuretic hormone; PCOS: Polycystic ovary syndrome; TH: Thyroid hormones; RAAS: Renin-Angiotensin-Aldosterone System. [Viktoriia Kasyanyuk] © 123RF.com (accessed on 22 February 2022) and Servier Medical Art by Servier. [/fig]
[fig] Author: Contributions: Conceptualization, methodology, and visualization, M.V.-H. and M.B.-C.; writing-original draft, M.V.-H., M.B.-C., A.R.L. and C.V.; writing-review and editing, J.S.N., D.C. and A.L.-M.; supervision, J.S.N., D.C. and A.L.-M. All authors have read and agreed to the published version of the manuscript. [/fig]
[fig] Funding: This research received no external funding. [/fig]
[table] Table 1: Effect of endocrine hormones on the different factors involved in nonalcoholic fat liver disease development.Sos, Harris et al. 2011, Nishizawa, Takahashi et al. 2012, Cordoba-Chacon, Majumdar et al. 2015 ↓ Hepatic inflammation and fibrosis Sumida, Yonei et al. 2015 ↑ proinflammatory cytokines and adipokines Ukropec, Penesova et al. 2008 Restoration of GH levels ↑ lean body mass and ↓ body fat Newman, Carmichael et al. 2015 Improvement of lipid profile Newman, Carmichael et al. 2015 ↓ insulin sensitivity and hyperinsulinemia Serri, Li et al. 2006 ↑ hepatic triglyceride content and de novo lipogenesis Park, Kim et al. 2011, Luque, Lopez-Vicchi et al. 2016 Decreased ↓ hepatic cholesterol and triacylglycerol Zhang, Ge et al. 2018 [/table]
|
Perinatal Exposure to Low Doses of Dioxin Can Permanently Impair Human Semen Quality
## Research | children's health
For the past 50 years, scientists have debated the issue of a reported decline in human semen quality, especially in Western countries [bib_ref] Evidence for decreasing quality of semen during past 50 years, Carlsen [/bib_ref] [bib_ref] The question of declining sperm density revisited: an analysis of 101 studies..., Swan [/bib_ref]. There is now a general consensus that human sperm quality has declined over time in different areas of the world [bib_ref] Decline in semen quality among fertile men in Paris during the past..., Auger [/bib_ref] [bib_ref] Regional differences in semen quality in Europe, Jørgensen [/bib_ref] [bib_ref] Geographic differences in semen quality of fertile U.S. males, Swan [/bib_ref] , especially in younger men [bib_ref] Geographical differences in semen quality in a population of young healthy volunteers..., López-Teijòn [/bib_ref] [bib_ref] Is semen quality related to the year of birth among Danish infertility..., Zheng [/bib_ref] , with indications of concurrent male sub-or infertility [bib_ref] Relation between semen quality and fertility: a population-based study of 430 firstpregnancy..., Bonde [/bib_ref] [bib_ref] Semen quality in sub-fertile range for a significant proportion of young men..., Paasch [/bib_ref] [bib_ref] Is semen quality related to the year of birth among Danish infertility..., Zheng [/bib_ref].
No clear causes have been identified; however, selected widespread persistent environmental endocrine-disrupting chemicals, such as 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) and other dioxin-like chemicals, are suspected to be potential etiologic agents, as demonstrated in experimental animals, but no definitive data are available for men.
Recently, we showed that exposure in infancy to low doses of dioxin induces a sperm concentration reduction of about 25%; however, no reduction is observed if similar exposures occur during adulthood [bib_ref] Dioxin exposure, from infancy through puberty, produces endocrine disruption and affects human..., Mocarelli [/bib_ref]. These data are consistent with animal models, which have demonstrated that the most sensitive times for these effects were exposures while in utero and during breastfeeding [bib_ref] Reproductive toxicity and tissue concentration of low doses of 2,3,7,8-tetrachlorodibenzo-p-dioxin in male..., Faqui [/bib_ref].
In infants, after 4-5 months of breastfeeding, dioxin serum concentrations increase up to 2-fold [bib_ref] Time course of PCDD/PCDF/PCB concentrations in breast-feeding mothers and their infants, Abraham [/bib_ref]. In that period, there is a hormonal surge [bib_ref] Evidence of testicular activity in early infancy, Forest [/bib_ref] [bib_ref] Editorial: the postnatal gonadotropin and sex steroid surge-insights from the androgen insensitivity..., Quigley [/bib_ref] , defined as "neonatal minipuberty," in which an increase of folliclestimulating hormone (FSH) and inhibin B [bib_ref] Longitudinal reproductive hormone profiles in infants: peak of inhibin B levels in..., Andersson [/bib_ref] induces and reflects a strong proliferation of Sertoli cells [bib_ref] Proliferation of Sertoli cells during development of the human testis assessed by..., Cortes [/bib_ref] , the number of which determines how many germ cells can develop into spermatozoa [bib_ref] Proliferation and functional maturation of Sertoli cells, and their relevance to disorders..., Sharpe [/bib_ref]. In addition to exposure during in utero development, exposure to TCDD within this developmental window could induce a direct or indirect toxic effect (i.e., imbalance of the hormonal equilibrium), impairing such proliferation and reducing sperm production in adults.
In this study, we assessed the association between human dioxin exposure in utero and during breast-feeding and adverse human adult male reproductive outcomes, as measured in men whose mothers were exposed to dioxin as a result of the trichlorophenol plant explosion near Seveso, Italy, in July 1976 [bib_ref] Serum concentrations of 2,3,7,8-tetrachlorodibenzo-p-dioxin and test results from selected residents of Seveso, Mocarelli [/bib_ref].
# Materials and methods
Participants. to 73 Caucasian women who lived in the most dioxin-polluted areas near [bib_ref] Serum concentrations of 2,3,7,8-tetrachlorodibenzo-p-dioxin and test results from selected residents of Seveso, Mocarelli [/bib_ref] were invited to participate. Thirty-nine of these men, who are the sons of 36 of the 73 women (mean age, 24.8 ± 5.6 years at the date of accident), actually participated; 21 were breastfed, and 18 were formula-fed [fig_ref] Table 1: Characteristics of study participants according to dioxin exposure of the mother and... [/fig_ref]. This further subdivision allowed us to separate the exposed men into two groups: men who had been exposed in utero only (those formula-fed and not breast-fed), and men who had been exposed perinatally-that is, both in utero and during nursing (breast-fed) [fig_ref] Table 1: Characteristics of study participants according to dioxin exposure of the mother and... [/fig_ref].
Background: In recent decades, young men in some industrialized areas have reportedly experienced a decrease in semen quality. oBjective: We examined effects of perinatal dioxin exposure on sperm quality and reproductive hormones. Methods: We investigated sperm quality and hormone concentrations in 39 sons (mean age, 22.5 years) born between 1977 and 1984 to mothers exposed to dioxin after the accident in Seveso, Italy (1976), and 58 comparisons (mean age, 24.6 years) born to mothers exposed only to background dioxin. Maternal dioxin levels at conception were extrapolated from the concentrations measured in 1976 serum samples. results: The 21 breast-fed sons whose exposed mothers had a median serum dioxin concentration as low as 19 ppt at conception had lower sperm concentration (36.3 vs. 86.3 million/mL; p = 0.002), total count (116.9 vs. 231.1; p = 0.02), progressive motility (35.8 vs. 44.2%; p = 0.03), and total motile count (38.7 vs. 98 million; p = 0.01) than did the 36 breast-fed comparisons. The 18 formulafed exposed and the 22 formula-fed and 36 breast-fed comparisons (maternal dioxin background 10 ppt at conception) had no sperm-related differences. Follicle-stimulating hormone was higher in the breast-fed exposed group than in the breast-fed comparisons (4.1 vs. 2.63 IU/L; p = 0.03) or the formula-fed exposed (4.1 vs. 2.6 IU/L; p = 0.04), and inhibin B was lower (breast-fed exposed group, 70.2; breast-fed comparisons, 101.8 pg/mL, p = 0.01; formula-fed exposed, 99.9 pg/mL, p = 0.02). conclusions: In utero and lactational exposure of children to relatively low dioxin doses can permanently reduce sperm quality. key words: breast-feeding, dioxin, environmental endocrine disrupters, human sperm impairment, human sperm quality, perinatal exposure, reproductive hormones, TCDD. The exposure status of the mothers of the exposed men (either participating or refusing) was verified by measuring dioxin levels in their serum, which was collected soon after the explosion. There were no significant differences in median TCDD serum concentrations in mothers of refusing men compared with the mothers of participating ones (36 vs. 51.7 ppt). Twenty exposed mothers breast-fed 21 sons, and 17 exposed mothers formula-fed 18 sons. Of these mothers, three each had two sons: One mother had a son in 1977 who was formula-fed, and another in 1981 who was breast-fed; one mother had a son in 1978 and another in 1980, both of whom were breast-fed; and one mother had a son in 1977 and another in 1983, both of whom were formula-fed. Sons of dioxin-exposed women whose spouses were also exposed, as well as all men with reported diseases [fig_ref] Table 1: Characteristics of study participants according to dioxin exposure of the mother and... [/fig_ref] , were excluded.
As a control, a comparison group of 123 consecutive healthy volunteer permanent blood donors (in Italy, blood donors have medical examinations performed every 3 months) of the same age and similar socioeconomic status as the exposed men, but whose mothers had not lived in the TCDD-contaminated areas, were invited to participate; they lacked additional accidental dioxin exposure (but did have the background exposure), and they were comparable in terms of breast-feeding and formula feeding. Fifty-eight of these men actually participated; 36 were breast-fed, and 22 were formula-fed [fig_ref] Table 1: Characteristics of study participants according to dioxin exposure of the mother and... [/fig_ref].
All participants completed a questionnaire on health, socioeconomic status, smoking and drinking habits, and working conditions. Participants also provided information on their weight at birth and how long they nursed as infants. They were screened for hidden diseases by clinical laboratory tests (aspartate aminotransferase, alanine aminotransferase, γ-glutamyl transferase, total cholesterol, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, C-reactive protein, glucose, creatinine, complete blood cell count and differential, hemoglobin, hepatitis B surface antigen, hepatitis B core antibody, hepatitis C, urine analysis) and donated blood and semen samples.
The study protocol was approved by institutional human subjects committee (Istituto di Ricovero e Cura a Carattere Scientifico Don Gnocchi, Milano, Italy). All study participants gave written informed consent.
Laboratory data. Semen samples were collected from participants as a postmasturbatory at-home semen sample, which was transported to the Desio Hospital laboratory at about 37°C and kept at that temperature until examination, which occurred within 1 hr from ejaculation. Blind tests were performed by the same two technicians according to the World Health Organization (WHO 1999) recommendations. Sperm percentages of A + B grades of motility, that is, progressive motility, were assessed at 400× magnification on a microscope heating stage (37°C) in duplicate, and concentration was measured using a Bürker-Türk chamber at phase contrast (400×). Morphology was evaluated by the same observer on 300 Papanicolaou-stained sperm per slide [bib_ref] Anomalies morphologiques du spermatozoïde human. 1. Proposition pour un système de classification, David [/bib_ref] [bib_ref] Male factors and the likelihood of pregnancy in infertile couples. 1. Study..., Jouannet [/bib_ref].
For serum hormone analysis, fasting blood samples were obtained on the same morning as the semen collection. An aliquot of serum was stored frozen at -80°C, and hormone levels were measured in large batches in order to reduce interassay variability. FSH, inhibin B, serum 17-β-estradiol (E 2 ), and luteinizing hormone (LH) were measured according to established immunometric methods. Testosterone was measured by radioimmunoassay. Quality control protocols were strictly applied for all tests using Westgard's multirule. Intra-and inter assay coefficients of variation were, respectively, 2.7% and 4.5% for FSH, 4.2% and 6.7% for inhibin B, 2.5% and 4.6% for E 2 , 1.8% and 3.4% for LH, and 3.4% and 6.6% for testosterone. Sensitivity was 0.1 IU/L for FSH, 7.0 pg/mL for inhibin B, 15 pg/mL for E 2 , 0.05 mIU/mL for LH, and 0.02 ng/mL for testosterone.
All serum TCDD measurements taken from maternal serum samples stored frozen since 1976-1977 were conducted by isotopedilution high-resolution mass spectrometry at the Centers for Disease Control and Prevention (Atlanta, GA, USA) [bib_ref] High-resolution gaschromatography/high-resolution mass spectrometric analysis of human serum on a whole weight..., Patterson [/bib_ref]. The estimated individual maternal serum TCDD concentration at the time of conception was calculated for 20-42 years of age by using the TCDD serum concentration in 1976-1977 and the TCDD half-life for Mothers of comparisons were considered as exposed to average background TCDD level of 10 ppt in 1976 . a Values are parts per trillion TCDD concentration expressed on serum lipid basis. b After breast-feeding for 4-5 months, the TCDD value in the child roughly doubles compared with the conception value [bib_ref] Time course of PCDD/PCDF/PCB concentrations in breast-feeding mothers and their infants, Abraham [/bib_ref]. c The formula-fed men were exposed to dioxin only in utero.
women of that age, estimated to be an average of 4 years (range, 2.1-6.7 years) [bib_ref] 3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD) and congeners in infants: a toxicokinetic model of human lifetime..., Kreuzer [/bib_ref] A general database was established and maintained using SAS software (version 8.2; SAS Institute Inc., Cary, NC, USA). Sperm and hormone data were fitted with a general linear model that included group, lactation class, and group × lactation class interaction as terms. Covariates included age at time of tests, length of abstinence in days, smoking habits (the total number of cigarettes smoked per day during months of habitual smoking), chemical exposures (mostly organic solvents, adhesives, paints, colors, and powders), body mass index (BMI) at the moment of the study, alcohol use (grams per day), educational level, and employment status. Scale transformations were applied to approximate normal distribution and homoscedasticity. Sperm concentration, total sperm count, progressive motile sperm count, and E 2 , testosterone, and FSH concentrations were log transformed, whereas semen volume and LH and inhibin B concentrations were square root transformed. Results were expressed as back transformation of least squares means (i.e., the means adjusted for all the terms in the model). Two types of comparisons were made: among groups within lactation class, and among lactation classes within group.
# Results
Dioxin exposure. In 1976, the median serum TCDD concentrations in women whose eligible sons did and did not participate in the study were similar (51.7 vs. 36 ppt). We assumed the serum TCDD concentrations in the comparison group mothers at conception [fig_ref] Table 2: Serum TCDD concentrations [/fig_ref] were 10 ppt in . All mothers at the time were also exposed to a background of dioxinlike chemicals [polychlorinated dibenzofurans (PCDFs) and some polychlorinated biphenyls (PCBs)], constituting a total of about 90 ppt of dioxin toxic equivalent (TEQ) concentration, which is added to the concentration of TCDD .
Median maternal TCDD concentrations [fig_ref] Table 2: Serum TCDD concentrations [/fig_ref] measured in serum collected in 1976 and extrapolated to conception levels for mothers who breast-fed or formula-fed their children were 19.0 and 27.9 ppt, respectively, and were not statistically different. Serum TCDD concentrations in sons after 4-5 months of breast-feeding are considered 2-fold those of the mothers [bib_ref] Time course of PCDD/PCDF/PCB concentrations in breast-feeding mothers and their infants, Abraham [/bib_ref].
Serum TCDD concentrations measured at the time of this study in the exposed breast-and formula-fed groups (average, 2.4 ppt and 1.1 ppt, respectively) and their respective comparison groups (average, 1.8 ppt and 1.0 ppt, respectively) were not statistically different.
Semen quality. About 50% of both the exposed and comparison group men recruited for this study complied with semen examination [fig_ref] Table 1: Characteristics of study participants according to dioxin exposure of the mother and... [/fig_ref]. and list variables for exposed men and their comparisons. The 39 men exposed in utero and during nursing had significantly decreased sperm concentration (p = 0.01), total sperm count (p = 0.03), and number of total motile sperm (p = 0.05) relative to the 58 men in the comparison group (Table 3, .
A stronger significant effect appeared when comparing the 21 breast-fed exposed sons and the 36 breast-fed comparisons (sperm concentration, p = 0.002; total sperm count, p = 0.02; total motile sperm count, p = 0.01; progressive motility, p = 0.03; see . In contrast, we observed no statistical differences between the 18 formula-fed exposed men and the 22 formula-fed comparisons or between the 36 breast-fed and 22 formula-fed comparisons . On the . Differences in sperm variables, FSH, and inhibin B between in utero and breast-or formula-fed men exposed to dioxin and comparisons. (A) Lower sperm concentration and the lower total motility of all TCDD-exposed men (median maternal TCDD at conception, 26 ppt) with respect to comparisons was related mainly to breast-feeding (see . (B) The nine breast-fed exposed men belonging to the two highest quartiles (median maternal TCDD at conception, 58.9 ppt) had both lower median sperm concentration (23.8 × 10 6 /mL) and lower motility (p = 0.0003, p = 0.003, respectively) than did the 36 breast-fed comparisons (82.5 × 10 6 /mL). Also, the 12 exposed breast-fed men belonging to the two lowest quartiles (median maternal TCDD at conception, 13.1 ppt) had lower (p = 0.05) sperm concentration (47.3 × 10 6 /mL) than did the 36 breast-fed comparisons (82.5 × 10 6 /mL). (C) Sperm variables did not statistically differ among all the formulafed men exposed to more or less than 26 ppt of TCDD and the formula-fed comparisons. (D) The effect of dioxin through breast-feeding on FSH (increasing) and inhibin B (decreasing) values, compared with exposed formula-fed and comparison men, was significant and permanent. (E) The low sperm concentration of TCDD in exposed breast-fed men was strongly correlated to low inhibin B concentration (a marker of Sertoli cells). The formula-fed exposed men and comparisons had normal inhibin B concentration. Breast-fed exposed (n = 21) (median TCDD: 19 ppt)
Formula-fed exposed (n = 18) (median TCDD ≤ 27.9 ppt)
Breast-fed exposed (n = 9) (TCDD ≥ 26 ppt: median 58.9 ppt)
Breast-fed exposed (n = 12) (TCDD < 26 ppt; median 13.1 ppt)
Formula-fed exposed (n = 10) (TCDD ≥ 26 ppt; median 54.6 ppt)
Formula-fed exposed (n = 8) (TCDD < 26 ppt; median 20.6 ppt) volume 119 | number 5 | May 2011 - Environmental Health Perspectives other hand, we observed an almost significant (p = 0.07) sperm concentration decrease in the 21 breast-fed exposed men compared with the 18 formula-fed exposed men. summarizes these observations, with TCDD values. In examining any differences in outcome, we found that the nine breast-fed men belonging to the two highest quartiles of TCDD serum concentrations (median TCDD, 58.9 ppt) did show lower sperm concentration (p = 0.0003) and total motility (p = 0.003) than the 36 breast-fed comparisons ; in the equivalent formula-fed groups, we observed no statistical differences between the 10 exposed men (median TCDD, 54.6 ppt) and the 22 comparisons . We observed similar effects for total sperm motility . For the men in the two lowest quartiles, who had serum TCDD concentrations < 26 ppt (median TCDD, 13.1 ppt), the only difference was a lower sperm concentration (p = 0.05) in the 12 exposed breast-fed men with respect to the 36 comparison breast-fed men .
These data also showed that a much lower semen concentration was present in the breast-fed men with the highest exposure, but the dose-response model did not reach statistical significance; we observed only a tendency, possibly because of the small number of cases. We did not observe this result with the formula-fed exposed men.
We observed no statistically significant difference for sperm morphology parameters between exposed and comparison groups.
Hormones. The 21 breast-fed exposed men had higher FSH concentrations than did both the 36 breast-fed comparison men (p = 0.03) and the 18 formula-fed exposed men (p = 0.04; , whereas inhibin B was significantly decreased (p = 0.01 and p = 0.02, respectively). ,E, illustrates the relationships between inhibin B and FSH, and inhibin B and sperm concentration in the different conditions. We observed no statistically significant differences for E 2 , testosterone, and LH between exposed and comparison groups.
# Discussion
The results of this study clearly indicate that continuous exposure of males starting from low concentrations of dioxin before and after birth due to the mother being exposed during the Seveso accident, and then due to breast-feeding during "neonatal minipuberty," results in a permanent impairment of the reproductive system (reduction of about 50% of sperm concentration and total sperm count and about 20% of sperm progressive motility, and increase of FSH with decrease of inhibin B) in young adulthood.
This impairment is not seen in males who were born to similarly exposed Seveso mothers who did not breast-feed. In fact, this group of men had sperm counts similar to those of the breast-fed and formula-fed comparison group.
The breast-fed exposed men had lower sperm variables than did the breast-fed . Differences in sperm and hormone data between men exposed to dioxin in utero (i.e., formula-fed) or also through breast-feeding and same-age comparisons.
## Breast-fed
Formula-fed p-Value, a breast-fed versus formula-fed comparison and formula-fed exposed men.
In the latter case, statistical significance was not reached, probably because of the small number of cases. These observations on semen quality are considerably reinforced by corresponding changes in hormone levels: decreased inhibin B and increased FSH found in the breast-fed exposed group. We had the rather unique opportunity to discriminate TCDD exposure with respect to background levels. Men exposed to rather high maternal serum TCDD concentrations at birth, such as 54.6 ppt (the 75th percentile of formula-fed exposed children), plus the then background value of 90 ppt of dioxinlike chemicals, did not show differences compared with the breast-and formula-fed comparison group members.
At the age of 4-5 months, formula-fed infants almost double their birth weight and halve the dioxin concentration, to about 27 ppt for the formula-fed exposed group and 5 ppt for the formula-fed comparison group, whereas the breast-fed comparison group approximately doubles the concentration to 20 ppt because of background dioxin exposure. On the other hand, the median of 19 ppt of dioxin in the breast-fed exposed group increases with breast-feeding to about 40 ppt. Our data show that impaired semen quality in adulthood results from exposure in utero, as well as continuous exposure through breast-feeding during "neonatal minipuberty," with adverse effects starting with a dioxin exposure range of 19-40 ppt and higher (the concentrations of background dioxin-like chemicals must also be added). Similar doses of TCDD administered to rhesus monkeys during pregnancy and lactation have recently been shown to produce lower semen quality in their young adult offspring [bib_ref] In utero and lactational exposure to 2,3,7,8-tetra chlorodibenzo-p-dioxin (TCDD) induces a reduction..., Arima [/bib_ref]. It would be interesting to investigate the effect of TCDD administration only during lactation.
Our observations show for the first time that continuous exposure of the developing human male in utero, and even more so during lactation, the "neonatal minipuberty period," to modest elevations of dioxin above background exposure levels can permanently impair later adult sperm production.
In fact, the only similar data [bib_ref] Semen quality after prenatal exposure to polychlorinated biphenyls and dibenzo furans, Guo [/bib_ref] are related to in utero and nursing exposure of children to high doses of burned PCBs/ PCDFs, but not to TCDD. That study demonstrated a reduction in sperm motility, but not density, as well as reduced egg penetration ability. Unfortunately, data concerning the specific concentration of the toxicants in maternal blood at conception were not available.
Mechanism of action. Dioxin and the structurally related PCDF chemicals mediate their toxic effect (but with different potencies) mainly as a ligand to the same aryl hydrocarbon receptor (AHR)/AHR nuclear translocator (ARNT) complex that binds to the xenobiotic responsive element (XRE; also called dioxin-responsive element, DRE) on target DNA. The widely expressed AHR/ARNT, when ligand activated, interacts with several transcription factors, steroids, and growth factors and possibly plays a critical constitutive role, especially in early tissue development. Through targeted gene networks, AHR/ARNT regulates or directly intervenes in reproductive system development [bib_ref] Expression of aryl hydrocarbon receptor and aryl hydrocarbon receptor nuclear translocator messenger..., Schultz [/bib_ref].
As our study shows, the untimely interference of dioxin in utero and within the first 4-5 months of life (the "neonatal minipuberty") [bib_ref] Evidence of testicular activity in early infancy, Forest [/bib_ref] [bib_ref] Editorial: the postnatal gonadotropin and sex steroid surge-insights from the androgen insensitivity..., Quigley [/bib_ref] results in a permanent impairing effect. It is during this time that the human testes almost double their volume [bib_ref] Proliferation of Sertoli cells during development of the human testis assessed by..., Cortes [/bib_ref] , due in great part to Sertoli cell proliferation, which determines spermatogenic potential [bib_ref] Proliferation and functional maturation of Sertoli cells, and their relevance to disorders..., Sharpe [/bib_ref]. An indication of possible Sertoli cell number (or functionality) reduction is given by the decrease of inhibin B, a marker of their activity, associated with the decrease of sperm concentration, not only in the exposed breast-fed with respect to the comparison breast-fed group, but also in the exposed breast-fed compared with the exposed formula-fed group. The increase of FSH associated with the decrease of inhibin B (as observed here) has been reported to be a very sensitive marker for impaired spermatogenesis in men [bib_ref] Serum inhibin B in combination with serum follicle-stimulating hormone (FSH) is a..., Von Eckardstein [/bib_ref].
It is well established that sons born to mothers who smoked heavily in pregnancy have a 20-50% reduction in sperm concentration in adulthood [bib_ref] Lower sperm counts following prenatal tobacco exposure, Jensen [/bib_ref] [bib_ref] Association of in utero exposure to maternal smoking with reduced semen quality..., Jensen [/bib_ref] [bib_ref] Is prenatal exposure to tobacco smoking a cause of poor semen quality?..., Ramlau-Hansen [/bib_ref] , reduced blood levels of inhibin B, and increased FSH [bib_ref] Does smoking during pregnancy affect sons' sperm counts, Storgaard [/bib_ref] , compared with unexposed males. No information is provided about smoking during breast-feeding in these studies, but it is likely that exposure to the chemicals in tobacco smoke occurred both prenatally and postnatally, similar to our dioxin study. This is relevant to our data because a common mechanism of the phenomenon we observed and the one due to smoking could be the action on XRE [bib_ref] High levels of dioxin-like potential in cigarette smoke evidence by in vitro..., Kasai [/bib_ref] of some polycyclic aromatic hydrocarbon (PAH) compounds contained in cigarette smoke, using the same AHR as dioxin.
It has also been shown that PAHs have a negative impact on rat and human Sertoli cell function [bib_ref] Environmental toxicant-induced germ cell apoptosis in the human fetal testis, Coutts [/bib_ref] [bib_ref] Polycyclic aromatic hydrocarbon-induced cytotoxicity in cultured rat Sertoli cells involves differential apoptotic..., Raychoudhury [/bib_ref].
Relevance to general population and individuals. These observed effects of perinatal dioxin exposure on lowering adult sperm counts might also explain, in part, the wider semen quality reduction reported in young men in some industrialized regions, such as in Spain, Denmark, and Germany [bib_ref] High frequency of sub-optimal semen quality in an unselected population of young..., Andersen [/bib_ref] [bib_ref] Geographical differences in semen quality in a population of young healthy volunteers..., López-Teijòn [/bib_ref] [bib_ref] Semen quality in sub-fertile range for a significant proportion of young men..., Paasch [/bib_ref] [bib_ref] Is semen quality related to the year of birth among Danish infertility..., Zheng [/bib_ref]. The men observed in these studies were in fact born in the 1970s and 1980s to mothers born in the 1950s and 1960s, a period when environmental concentrations of dioxin and dioxin-like chemicals peaked [bib_ref] Time trends in levels, patterns and profiles for PCDD/PCDF in sediment cores..., Hagenmaier [/bib_ref]. The bioaccumulation of these compounds with age is well demonstrated, and a concurrent effect of maternal smoking could also contribute to this phenomenon.
Because of environmental policies instituted throughout most of Europe and the United States, TCDD and total TEQ concentrations in women 20-40 years of age have recently decreased sharply, to less than 2 ppt and 12 ppt, respectively [bib_ref] PCDD/PCDF: human background data for Germany, a 10-year experience, Päpke [/bib_ref] [bib_ref] Total TEQ reference range (PCDDs, PCDFs, cPCBs, mono-PCBs) for the US population, Patterson [/bib_ref]. Thus, current serum dioxin levels in infants in these areas, even after being increased 2-to 3-fold because of breast-feeding, are far from the concentrations that would yield adverse effects.
# Conclusions
Our data indicate for the first time that environmental chemical exposure to dioxin continuously throughout the perinatal period permanently reduces semen quality and sperm counts in young men and demonstrates that the male reproductive system is dramatically sensitive to the action of dioxin, starting from rather low doses (~ 19 ppt) in utero and doubled by breast-feeding during "neonatal minipuberty," resulting in permanent impairment in the adult (reduction of ~ 50% of sperm concentration and total sperm count and ~ 20% of sperm progressive motility, with increased FSH and decreased inhibin B).
Although our findings are derived from examining a unique cohort of men exposed perinatally to supranormal dioxin levels, they may have relevance to the wider population of young men in industrialized countries and may partly explain the reported widespread occurrence of low sperm counts in young men in Europe. Our findings also raise concern for those areas of the world in which rapid industrial development may cause widespread distribution of environmental endocrine-disrupting chemicals, such as dioxin and dioxin-like chemicals.
Furthermore, our data should encourage studies of semen quality in men of appropriate ages, comparing breast-fed and formulafed groups, especially if this could be linked to proximity to polluted areas at the time of perinatal life.
[table] Table 1: Characteristics of study participants according to dioxin exposure of the mother and to the type of nursing. [/table]
[table] Table 2: Serum TCDD concentrations (percentile distribution) of mothers at exposure in 1976 and extrapolated values at conception of their sons. [/table]
|
Ovarian Cancer Molecular Stratification and Tumor Heterogeneity: A Necessity and a Challenge
# Introduction
Only two new drugs have been licensed for the treatment of epithelial ovarian cancer in the last 5 years (bevacizumab and olaparib). These are also the only two molecularly targeted agents licensed in this disease. As we continue to move into the genomic era of cancer therapy, it is clear that optimal therapy is going to depend on molecular stratification and that the stratification itself is going to need to contend with tumor heterogeneity. In this article, we discuss molecular stratification and tumor heterogeneity in the context of high-grade serous ovarian cancer.
The development of bevacizumab and olaparib has provided contrasting examples of stratification in molecularly targeted agents. Bevacizumab is licensed as an unselected agent, currently without molecular (or indeed histological) stratification. However, emerging data may be able to help us refine which patients may benefit the most from this agent (and which may not require it). Any such refinement can be expected to increase the median benefit in the selected population and reinforce the cost:benefit advantage. Conversely, olaparib is licensed as a highly selected agent, currently by genomic or somatic BRCA1/BRCA2 mutation in high-grade serous cancer. However, emerging data may be able to help us expand its role into tumors with other homologous recombination deficits (while also determining if all BRCA1/BRCA2 mutations respond equally). For both agents, however, cancers progress even on continuous therapy and targeting the resistant clones that have emerged from tumor heterogeneity will be key to extending benefit for these patients.
## Bevacizumab
Bevacizumab is the first vascular endothelial growth factor (VEGF)-targeted therapy to have been licensed in ovarian cancer; by the European Medicines Agency (EMA) in 2011 and the United States Food and Drug Administration (FDA) in 2014. These licenses differ although in both cases the bevacizumab is given in combination with chemotherapy. The EMA license was the result of first line data from the GOG 218 (1) and ICON7 (2) studies, second line platinum sensitive data from OCEANS (3) and second line platinum resistant data from AURELIA (4) all of which demonstrated a statistically significant increase in progression free survival compared to chemotherapy alone. The FDA license relates to the platinum resistant setting only and was dependent on the data from the AURELIA study. The optimal setting for this treatment is unknown (5), as is the value of treating through progression or utilizing combinations of anti-angiogenic therapies. VEGF-targeted therapy is clearly an active approach in ovarian cancer (at least in combination with chemotherapy) and other anti-angiogenic agents have been investigated in this setting including pazopanib (6), cedirinib (7), Ovarian cancer molecular stratification Frontiers in Oncology | www.frontiersin.org nintedanib (8), aflibercept (9), trebananib (10), sunitinib (11), sorafenib [bib_ref] Paclitaxel/carboplatin with or without sorafenib in the first-line treatment of patients with..., Hainsworth [/bib_ref] , and (PDGFR) imatinib [bib_ref] Weekly paclitaxel with intermittent imatinib mesylate (Gleevec): tolerance and activity in recurrent..., Safra [/bib_ref].
Given the mode of action, it was not unreasonable to seek potential broad activity for bevacizumab and, as with all previous agents in ovarian cancer, these trials had no molecularly or histological stratification. However, potential biomarkers are now emerging for benefit from bevacizumab [bib_ref] Markers of response for the antiangiogenic agent bevacizumab, Lambrechts [/bib_ref] , building on the extensive translational work incorporated in ICON7. For blood biomarkers, a link is evident between circulating Ang1 plus Tie2 levels and progression free survival [bib_ref] The combination of circulating Ang1 and Tie2 levels predicts progression-free survival advantage..., Backen [/bib_ref] , with most of the benefit from bevacizumab in the high Ang1-low Tie2 group (HR 0.27), no significant effect in the low Ang1 group and a detriment in the high Ang1-high Tie2 group (HR 3.6). A possible link noted with plasma VEGF-A in other tumor types [bib_ref] Evaluation of plasma VEGFA as a potential predictive pan-tumour biomarker for bevacizumab, Jayson [/bib_ref] , was not seen in this ovarian dataset. For tissue biomarkers, a signature made up of tissue mesothelin, FLT4 and AGP and blood CA125 also has potential to strongly differentiate between benefit and harm from bevacizumab but was limited by patient numbers in the analysis [bib_ref] Predicting response to bevacizumab in ovarian cancer: a panel of potential biomarkers..., Collinson [/bib_ref]. Recently, a transcriptomic signature has been presented [bib_ref] Molecular subgroup of high-grade serous ovarian cancer (HGSOC) as a predictor of..., Gourley [/bib_ref] , which identifies distinct molecular subgroups of high-grade serous ovarian cancer that respond very differently to bevacizumab. In this analysis, the two proangiogenic subgroups had a poorer overall survival but appeared to contain all the benefit from bevacizumab. The other, immune subgroup had a superior prognosis but had a detriment (HR 2.0) from bevacizumab. This data will need confirmation in further datasets but the above examples suggest that we are getting closer to a molecular biomarker for bevacizumab benefit (and resistance). The next step will of course be identifying if these molecular signals also emerge in acquired resistance and if they indicate a druggable pathway to improve or extend the activity of VEGF-directed therapies to resistant tumors (or resistant clonal subpopulations). The story may be complicated by the fixed duration of bevacizumab in some studies (such as ICON7) but continuous maintenance therapy is the expected direction of travel for the future.
Hopefully, the above approaches can help address the mystery of why VEGF-directed therapy does not yet seem to be living up to its clear potential. In the phase 2 single agent studies [bib_ref] Phase II trial of bevacizumab in persistent or recurrent epithelial ovarian cancer..., Burger [/bib_ref] [bib_ref] Phase II study of bevacizumab in patients with platinum-resistant ovarian cancer or..., Cannistra [/bib_ref] , bevacizumab had a roughly 20% objective response rate (ORR), a figure matched by the additional ORR benefit seen compared to chemotherapy alone in phase 2 and phase 3 combination studies, regardless of setting [fig_ref] TABLE 1 |: pivotal phase 2 and phase 3 bevacizumab studies in ovarian cancer [/fig_ref]. It is unclear why this clear ORR benefit has not translated into more impressive PFS or OS benefits in the phase 3 first line studies (PFS benefit 3.8m in GOG 218 and 2 months in ICON7, with updated analysis of the latter non-significant and no OS benefit in the ITT population). It seems likely that there is a subset benefiting greatly from therapy (and this subset will be evident when response rate is the primary endpoint, as in some phase 2 studies) but that the effect is being diluted by the current lack of a selection biomarker and is therefore harder to detect when PFS is the primary endpoint (as in the phase 3 studies). Identifying this subset may be the key to widening its licensed indications. Identifying the tumor heterogeneity that leads to resistance to VEGF-directed therapy may be the key to improving and prolonging the benefit.
## Olaparib
Olaparib is the first poly ADP ribose polymerase (PARP)inhibitor to have been licensed in ovarian cancer (2014). The EMA license is as post-chemotherapy maintenance in patients with germline or somatic BRCA1 or BRCA2 mutations. This was as a result of the subgroup analysis (21) of BRCA1/2 mutant cancers (germline or somatic), in the molecularly unselected Study 19 of relapsed platinum sensitive high-grade serous ovarian cancer [bib_ref] Olaparib maintenance therapy in platinum-sensitive relapsed ovarian cancer, Ledermann [/bib_ref]. The FDA license is as monotherapy in patients with deleterious or suspected deleterious germline BRCA mutated advanced ovarian cancer who have been treated with three or more prior lines of chemotherapy. PARP inhibitors have demonstrated strong activity in molecularly selected populations and other agents include rucaparib [bib_ref] Therapeutic potential of the poly(ADP-ribose) polymerase inhibitor rucaparib for the treatment of..., Ihnen [/bib_ref] , niraparib [bib_ref] Niraparib: a Poly(ADP-ribose) polymerase (PARP) inhibitor for the treatment of tumors with..., Jones [/bib_ref] , veliparib [bib_ref] A phase II evaluation of the potent, highly selective PARP inhibitor veliparib..., Coleman [/bib_ref] , and talazoparib [bib_ref] 01A phase 3 study of the oral PARP inhibitor talazoparib (BMN 673)..., Roche [/bib_ref]. The target population was a clear priority from the original phase 1 study, where BRCA1/BRCA2 mutation carriers were the predominant responders, leading to a BRCA1/BRCA2 expansion [bib_ref] Inhibition of poly(ADP-ribose) polymerase in tumors from BRCA mutation carriers, Fong [/bib_ref] , with promising activity in ovarian cancer [bib_ref] ADP)-ribose polymerase inhibition: frequent durable responses in BRCA carrier ovarian cancer correlating..., Fong [/bib_ref] , which was subsequently confirmed in phase 2 [bib_ref] Oral poly(ADP-ribose) polymerase inhibitor olaparib in patients with BRCA1 or BRCA2 mutations..., Audeh [/bib_ref]. However, BRCA1/BRCA2 mutation is known not to be an absolute biomarker for sensitivity and other homologous recombination defects have been strongly implicated as additional predictive biomarkers. Study 19 represented a pragmatic approach to enrich for this whole group without genetic testing by looking at a platinum-sensitive, high-grade serous histopathology, ovarian population [bib_ref] Olaparib maintenance therapy in platinum-sensitive relapsed ovarian cancer, Ledermann [/bib_ref] , given the known link between platinum sensitivity, high-grade serous histopathology and BRCA1/2 or wider homologous recombination gene defects [bib_ref] BRCAness" syndrome in ovarian cancer: a case-control study describing the clinical features..., Tan [/bib_ref] [bib_ref] Association of BRCA1 and BRCA2 mutations with survival, chemotherapy sensitivity, and gene..., Yang [/bib_ref] , even in non-hereditary cases [bib_ref] BRCA1 is both a prognostic and predictive biomarker of response to chemotherapy..., Carser [/bib_ref] [bib_ref] Somatic mutations in BRCA1 and BRCA2 could expand the number of patients..., Hennessy [/bib_ref]. However, although this study did identify activity in the non-BRCA mutant subgroup (HR 0.54), the dramatic effect was really in the BRCA1/2 mutant subgroup (HR 0.18), and the diluted overall signal risked being a barrier to licensing of these important, active agents. Therefore, olaparib is now licensed for BRCA1/2 mutant cancers and steps are being taken to identify additional predictive biomarkers, including other known homologous recombination defects. However, predictive biomarkers are seldom binary and olaparib provides a good example where molecular heterogeneity between patients and within tumors leads to significant variation in activity and resistance.
When comparing different patients with different homologous recombination deficient tumors, we are beginning to realize that all BRCA molecular deficits are not equal -epigenetic changes are clearly a different biology to genetic [bib_ref] Association of BRCA1 and BRCA2 mutations with survival, chemotherapy sensitivity, and gene..., Yang [/bib_ref] [bib_ref] Ovarian carcinomas with genetic and epigenetic BRCA1 loss have distinct molecular abnormalities, Press [/bib_ref] but what about different specific mutations or the difference between germline and somatic? Today, it is generally accepted that the different histological subgroups of epithelial ovarian cancer represent different diseases [bib_ref] The genesis and evolution of high-grade serous ovarian cancer, Bowtell [/bib_ref]. However, as with other cancers, it is also clear that different molecular subgroups (30) of high-grade serous ovarian cancer can have different phenotypes [bib_ref] Increased incidence of visceral metastases in Scottish patients with BRCA1/2-defective ovarian cancer:..., Gourley [/bib_ref] and outcomes [bib_ref] Association between BRCA1 and BRCA2 mutations and survival in women with invasive..., Bolton [/bib_ref] [bib_ref] Germline mutation in BRCA1 or BRCA2 and ten-year survival for women diagnosed..., Candido-Dos-Reis [/bib_ref] , with homologous recombination defects a clear example. It is natural to extend this to specific homologous recombination proteins and, even further, to specific mutations/defects of individual proteins.
Molecular heterogeneity within an individual's cancer may be just as important. While adaptive epigenetic changes have been implicated in platinum resistance [bib_ref] Candidate DNA methylation drivers of acquired cisplatin resistance in ovarian cancer identified..., Zeller [/bib_ref] , a more striking mechanism of resistance may underlie some of the cross-resistance of platinum therapy and PARP inhibition that has lead to the focus on platinum sensitive disease in the clinic [bib_ref] ADP)-ribose polymerase inhibition: frequent durable responses in BRCA carrier ovarian cancer correlating..., Fong [/bib_ref] [bib_ref] Functional restoration of BRCA2 protein by secondary BRCA2 mutations in BRCA2-mutated ovarian..., Sakai [/bib_ref]. This is the phenomenon where inactivating mutations within the BRCA1/ BRCA2 genes revert to functional genes [bib_ref] Functional restoration of BRCA2 protein by secondary BRCA2 mutations in BRCA2-mutated ovarian..., Sakai [/bib_ref] [bib_ref] Resistance to therapy caused by intragenic deletion in BRCA2, Edwards [/bib_ref] [bib_ref] Secondary BRCA1 mutations in BRCA1-mutated ovarian carcinomas with platinum resistance, Swisher [/bib_ref] [bib_ref] Secondary somatic mutations restoring BRCA1/2 predict chemotherapy resistance in hereditary ovarian carcinomas, Norquist [/bib_ref] [bib_ref] Secondary mutations as a mechanism of cisplatin resistance in BRCA2-mutated cancers, Sakai [/bib_ref] , clearly demonstrating the strong selection pressures, which drive outgrowth of a resistant subclone that lacks the one main feature that defined, or even induced, the original cancer but which subsequently had become its Achilles heel. In effect, the tumor is doing whatever it can to evade the agents used against it, even if that means reexpressing the gene that it had to lose in order to become a cancer cell in the first place.
Of course, not every individual's tumor will contain the resistant subclones that drive this response and much can be learned from the super-responders [bib_ref] ADP)-ribose polymerase inhibition: frequent durable responses in BRCA carrier ovarian cancer correlating..., Fong [/bib_ref] [bib_ref] Characterization of ovarian cancer long-term responders on olaparib, Lheureux [/bib_ref] who give an example of what we might hope to achieve if we could overcome that heterogeneity.
## The future
Clearly, future optimal therapy for high-grade serous ovarian cancer will depend on optimal molecular stratification and this is just as true for bevacizumab and olaparib as it will be for future agents. While this will help rise to the challenge of optimizing therapy for inter-patient molecular heterogeneity, monotherapy may never overcome intra-patient heterogeneity. If we want to improve the durability of responses, that pool of resistant clones may need to be narrowed by using combination therapies. Indeed, recent clinical data for the addition of the VEGFR inhibitor, cedirinib, to olaparib have shown a significant increase in response rate and a near-doubling of progression free survival [bib_ref] Combination cediranib and olaparib versus olaparib alone for women with recurrent platinum-sensitive..., Liu [/bib_ref]. The majority of this benefit was in the BRCA1/BRCA2 wild-type (or unknown) group, perhaps demonstrating that combinations can overcome monotherapy dependencies but also highlighting that there is still a lot to learn about biomarkers for anti-angiogenic and PARP inhibitor agents in ovarian cancer.
[table] TABLE 1 |: pivotal phase 2 and phase 3 bevacizumab studies in ovarian cancer. [/table]
|
Variation in the estimated prevalence of multimorbidity: systematic review and meta-analysis of 193 international studies
## Table of contents
1. (multimorbidit$ or multi-morbidit$ or comorbidit$ or co-morbidit$ or polymorbidit$ or poly-morbidit$ or multicondition$ or multicondition$ or "multiple chronic condition$" or "morbidity burden" or ((multiple or coexisting [bib_ref] Prevalence and patterns of multimorbidity in australia, Britt [/bib_ref] 275706 0.08 Moderate Contributing to high levels of heterogeneity of effect sizes (Leave-one-out analysis) and the studentized residual of this study is more than 2 standard deviations away from its expected value. [bib_ref] Burden of multimorbidity in relation to age, gender and immigrant status: A..., Lenzi [/bib_ref] Moderate Contributing to high levels of heterogeneity of effect sizes (Leave-one-out analysis) and the studentized residual of this study is more than 2 standard deviations away from its expected value. [bib_ref] Epidemiologic characteristics of multimorbidity and sociodemographic factors associated with multimorbidity in a..., Low [/bib_ref] Low et al
Singapore Asia High Community All adults 39.6 1181024 Self-report 48 309428 0. Contributing to high levels of heterogeneity of effect sizes (Leave-one-out analysis) [bib_ref] Epidemiology of multimorbidity in china and implications for the healthcare system: Cross-sectional..., Wang [/bib_ref] Studies that were carried out in hospital settings
## Data source
Self-report Studies that collected data using self-report or interviews
Medical records and administrative databases Studies that collected data using electronic medical records, medical chart reviews, insurance claims databases, pharmacy databases, or research databases
## Study population
All
[fig] 40, Figure S1 41, Figure S2 42, Figure S3 43, Figure S4: S1: Search strategy....................................................................................................... 3Table S2: Summary of the characteristics of outlying studies ............................................... 4 Table S3: Characteristics of 24 outlying studies.................................................................... 5 Table S4: Characteristics of 193 included studies ................................................................. 9 Table S5: Associations between predictors ......................................................................... 25 Table S6: Risk of bias assessment of included studies ........................................................ 26 Table S7: Output of adjusted meta-analytic model based on 217 studies ........................... 39 Table S8: Definition of variables ......................................................................................... Graphical display of study effect sizes and heterogeneity ................................. Process of examining and identifying outlying studies in meta-analysis .......... Summary of risk of bias assessment .................................................................. Meta-regression trees for predicting the pooled estimated prevalence of multimorbidity (based on 'mean age' and 'number of conditions' predictors. unit: log(odds)) ............................................................................................................................. 44 [/fig]
[fig] Figure S2, Figure S4: adults Studies with a sample of population aged 18 and older (n=45), aged 20 and older (n=8), aged 21 and older (n=3), aged 25 and older (n=2), or others (n=27) (e.g. aged 16 and older, or aged 17 and older) Middle-aged and older Studies with a sample of population aged 50 and older (n=25), aged 40 and older (n=5), aged 40 and older (n=10), or others (n=6) (e.g. aged 57 and older, or aged 45 and older) Only older people Studies with a sample of population aged 65 and older (n=22), aged 60 and older (n=25), aged 70 and older (n=5) or others (n=11) (e.g. aged 68 and older, aged 77 and older, aged 78 and older, Process of examining and identifying outlying studies in metaMeta-regression trees for predicting the pooled estimated prevalence of multimorbidity (based on 'mean age' and 'number of conditions' predictors. unit: log(odds)) , et al. Ho IS-S [/fig]
[table] Table S1: Search strategy [/table]
[table] Table S2: Summary of the characteristics of outlying studies [/table]
[table] Table S4: Characteristics of 193 included studies [/table]
[table] Table S5: Associations between predictors [/table]
[table] Table S6: Risk of bias assessment of included studiesBMJ Publishing Group Limited (BMJ) disclaims all liability and responsibility arising from any reliance Supplemental material placed on this supplemental material which has been supplied by the author(s) [/table]
[table] Table S7: Output of adjusted meta-analytic model based on 217 studies [/table]
[table] Table S8: Definition of variablesCommunityStudies that used population surveys, insurance claims databases, or research databases Primary care Studies that were carried out in primary care settings [/table]
|
Genome-wide identification and characterization of Solanum tuberosum BiP genes reveal the role of the promoter architecture in BiP gene diversity
[table] Table S1: Molecular characterization of additional genes used for phylogeny a The templates here are the Human BiP (P11021) and Chinese Hamster (G3I8R9) c The Cluster density signifies the model quality. A Higher cluster density value means a better model quality. Technically, the cluster density is the density of decoys per unit space. Decoys are the replicas of peptide chain fragments generated during simulation to generate full atomic models of the protein. [/table]
[table] Table S2: I-TASSER modeling results of yeast, Arabidopsis and potato BiPs [/table]
[table] Table S3: Key CREs in plant BiP promoters that engage in cytoprotective responses to environmental challenges [/table]
[table] Table S4: Key CREs in StBiP and AtBiP promoters that engage in cytoprotective responses to environmental challenges [/table]
|
ARHGAP24 represses β-catenin transactivation-induced invasiveness in hepatocellular carcinoma mainly by acting as a GTPase-independent scaffold
Rationale: Accumulating evidence shows that Rho-GTPase-activating proteins (RhoGAPs) exert suppressive roles in cancer cell proliferation and metastasis. However, no study has systematically investigated the clinical significance of RhoGAPs and analyzed the functions of ARHGAP24 in hepatocellular carcinoma (HCC). Methods: The relationship between RhoGAP expression and HCC prognosis was investigated via using The Cancer Genome Atlas and Gene Expression Omnibus databases. ARHGAP24 expression was detected by reverse transcription-polymerase chain reaction, western blot and immunohistochemistry staining assays. Moreover, in vitro assays including cell counting kit-8, colony formation, wound healing and Transwell assays, and in vivo tumor growth and pulmonary metastases evaluations were conducted to evaluate the biological function of ARHGAP24 in HCC. Liquid chromatography-tandem mass spectrometry, co-immunoprecipitation, GTPase activation, ubiquitination, and luciferase reporter assays and bioinformatics analysis were carried out to gain insights into the mechanisms underlying the tumor-suppressive function of ARHGAP24. Results: ARHGAP24 expression was dramatically decreased in HCC tissues, and low ARHGAP24 expression was an independent poor prognostic indicator for progression-free survival in HCC patients. ARHGAP24 overexpression significantly inhibited cell proliferation, migration and invasion, while knockdown of ARHGAP24 exerted the opposite effects. Through Gene Set Enrichment Analysis (GSEA), we found ARHGAP24 mainly suppressed HCC cell proliferation and invasion by attenuating β-catenin transactivation and blocking β-catenin signaling could effectively abolish the promotional effects of ARHGAP24 knockdown in HCC cells. Notably, GAP-deficient mutant of ARHGAP24 exerted similar inhibitory effects as the wild-type did, indicating suppressive function of ARHGAP24 was independent of its RhoGAP activity. Moreover, we identified pyruvate kinase M2 (PKM2) as a new binding partner of ARHGAP24, which recruited a novel E3 ligase (WWP1) and subsequently promoted PKM2 degradation. WWP1 knockdown significantly reduced the inhibitory function of ARHGAP24, and the C-terminal fragments of ARHGAP24 (amino acids 329 -430 and 631 -748) bound directly to WWP1 and PKM2 (amino acids 388 -531), respectively. Conclusions: Our data indicate that ARHGAP24 may be an independent prognostic indicator for HCC. It is a critical suppressor of HCC that recruits WWP1 for PKM2 degradation. Targeting the ARHGAP24/WWP1/PKM2/β-catenin axis may provide new insights into HCC prevention and treatment.
# Introduction
Hepatocellular carcinoma (HCC) is the main type of liver cancer, accounting for approximately 85% -90% of all cases. It is the fifth most common malignant tumor in China with the third highest mortality among all cancer types [bib_ref] Cancer Statistics, 2021, Siegel [/bib_ref] [bib_ref] Epidemiology and surveillance for hepatocellular carcinoma: New trends, Singal [/bib_ref] , and thus presents a serious threat to health and quality of life. At present, radical resection is the only cure for liver cancer, but patients with early-stage HCC have the opportunity to undergo surgery. Although the clinical application of molecular targeted drugs and immune drugs has greatly improved the survival of patients with HCC [bib_ref] Challenges of combination therapy with immune checkpoint inhibitors for hepatocellular carcinoma, Cheng [/bib_ref] [bib_ref] Molecular targeted and immune checkpoint therapy for advanced hepatocellular carcinoma, Liu [/bib_ref] , the 5-year metastasis and recurrence rates of HCC remain high, reaching 50% -70% [bib_ref] Guidelines for the Diagnosis and Treatment of Hepatocellular Carcinoma, Zhou [/bib_ref]. These important factors therefore restrict the prognosis of HCC patients. Further in-depth exploration of the molecular basis of HCC progression is thus needed to improve patient outcomes.
The identification of new tumor suppressor genes is of great significance for understanding the molecular mechanism of liver cancer progression and for exploring new intervention strategies. The Rho-GTPase-activating protein (RhoGAP) family is a class of emerging tumor suppressors with more than 60 members [bib_ref] Rho GTPase regulatory proteins in podocytes, Matsuda [/bib_ref] [bib_ref] Targeting the Small GTPase Superfamily through Their Regulatory Proteins, Gray [/bib_ref] [bib_ref] Systems analysis of RhoGEF and RhoGAP regulatory proteins reveals spatially organized RAC1..., Müller [/bib_ref]. RhoGAPs have been reported to be involved in regulating Rho GTPase (i.e., RhoA, Rac1 and CDC42) activity. Rho GTPases activate a diverse array of downstream effectors, while the GDP-bound states have the opposite effects. RhoGAPs suppress the formation of the active GTP-bound state of Rho GTPases by catalyzing the exchange of GTP for GDP. The aberrant activation or overexpression of RhoGAPs may thus inhibit tumor growth [bib_ref] Rho GTPase regulatory proteins in podocytes, Matsuda [/bib_ref] [bib_ref] Targeting the Small GTPase Superfamily through Their Regulatory Proteins, Gray [/bib_ref] [bib_ref] Systems analysis of RhoGEF and RhoGAP regulatory proteins reveals spatially organized RAC1..., Müller [/bib_ref]. However, less than half of all RhoGAPs currently have clear biological functions and the roles of most members of this family are still unclear, especially in highly heterogeneous tumors such as HCC. Further systematic research is therefore urgently needed.
RhoGAP 24 (ARHGAP24) is a member of the RhoGAP protein family [bib_ref] FilGAP and its close relatives: a mediator of Rho-Rac antagonism that regulates..., Nakamura [/bib_ref] [bib_ref] FilGAP, a Rho-and ROCK-regulated GAP for Rac binds filamin A to control..., Ohta [/bib_ref] with strong tumor suppressor potential. Zhang et al. reported that it induced G0/G1 phase arrest of colorectal cancer cells by regulating the expression of p53 and p21 and promoted tumor cell apoptosis [bib_ref] ARHGAP24 regulates cell ability and apoptosis of colorectal cancer cells via the..., Zhang [/bib_ref]. ARHGAP24 also inhibited the activation of signal transducer and activator of transcription 6 signaling in lung cancer cells and induced tumor cell apoptosis and inhibited cell proliferation through the WWP2/p27 pathway [bib_ref] ARHGAP24 inhibits cell proliferation and cell cycle progression and induces apoptosis of..., Wang [/bib_ref]. ARHGAP24 also inhibited the growth of kidney cancer, breast cancer, and astrocytoma, and its low expression can be used as a predictor of poor prognosis in patients with these tumors [bib_ref] ARHGAP24 inhibits cell cycle progression, induces apoptosis and suppresses invasion in renal..., Xu [/bib_ref] [bib_ref] RASAL2 activates RAC1 to promote triple-negative breast cancer progression, Feng [/bib_ref] [bib_ref] The role of FilGAP, a Rac-specific Rho-GTPase-activating protein, in tumor progression and..., Hara [/bib_ref]. Previous studies found that the single nucleotide polymorphism locus rs346473 of ARHGAP24 was closely related to susceptibility to hepatitis B virus and the progression of related diseases in the Chinese population [bib_ref] Effects of variant rs346473 in ARHGAP24 gene on disease progression of HBV..., Liu [/bib_ref]. However, the biological role of ARHGAP24 in HCC has not yet been explored. In-depth analysis of its specific molecular mechanisms will provide a new theoretical basis for preventing metastasis and recurrence of HCC and for exploring therapeutic targets.
In this study, we investigated the relationship between RhoGAP expression and the prognosis of HCC using data from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) databases. We found that ARHGAP24 expression was reduced in HCC tissues and its expression was significantly related to a poor prognosis compared with other RhoGAPs. Our clinical data further confirmed that ARHGAP24 was an independent indicator predicting time to tumor recurrence (TTR). In addition, in vivo and in vitro experiments showed that high ARHGAP24 levels could inhibit cancer cell proliferation, migration and invasion. Co-immunoprecipitation combined with mass spectrometry showed that ARHGAP24 could serve as a scaffolding protein to promote the binding of the E3 ubiquitin ligase WWP1 to pyruvate kinase M2 (PKM2), and then degrade PKM2 through the ubiquitinproteasome pathway to inhibit liver cancer invasion and metastasis.
# Materials and methods
## Hcc patients and follow-up
HCC tissues and adjacent tissues were obtained from 131 adult patients who underwent surgery at Zhongshan Hospital, Fudan University (Shanghai, China) between April 2018 and July 2019. All tumors were histologically confirmed according to the American Association for the Study of Liver Diseases guidelines. None of the patients received preoperative chemotherapy or radiotherapy. In addition, 20 pairs of frozen HCC and non-tumor tissues, eight recurrent tumors and seven non-recurrent tumors were collected after surgical resection. This study was approved by the ethics committee of Zhongshan Hospital, Fudan University. Written informed consent was obtained from all subjects. PFS was set as the endpoint of follow-up in our study. PFS was defined as the interval between resection and intrahepatic recurrence or extrahepatic metastasis. Follow-up ended on 31 st July 2021.
## Liquid chromatography-tandem mass spectrometry (lc-ms/ms)
## Hcclm3 cells transfected with flag-arhgap24
were immunoprecipitated with anti-FLAG antibody. Proteins interacting with ARHGAP24 were identified as the experimental group and proteins interacting with IgG were identified as the non-specific binding. All MS experiments were performed on a Thermo Fusion Lumos mass spectrometer connected to an Easy-nLC 1200 via an Easy Spray (Thermo Fisher Scientific, MA, USA). The resulting sequences were searched against the UniProt Human Proteome database (downloaded 5 May 2018). The candidate proteins are listed in .
## Rhogtpase activity detection
Levels of GTP-bound Rac1, GTP-bound CDC42 and GTP-bound RhoA were detected using an Active Rho Detection Kit (Active Rac1 Detection Kit; Active CDC42 Detection Kit; Cell Signaling Technology) according to the manufacturer's protocol. Briefly, GST-Rhotekin-RBD fusion protein or GST-PAK-PBD was used to bind activated forms of GTP-bound Rho and GTP-bound Rac1/CDC42, which were then immunoprecipitated with glutathione resin. The level of Rho activation or Rac1/CDC42 activation was then determined by western blotting using Rho/Rac1/ CDC42 rabbit antibody, respectively.
## Statistical analyses
Statistical analyses were performed using SPSS 20.0 software. Continuous variables were presented as mean ± standard deviation. Differences between groups were analyzed by two-tailed unpaired Student's t-test, Pearson's χ 2 test, Mann-Whitney U test, two-way ANOVA, or log-rank test. Differences were considered significant at P < 0.05.
Further details of the methods are presented in the Supplementary Material.
# Results
## Identification of arhgap24 as a novel prognostic biomarker for hcc
The prognostic values of the 64 RhoGAP members were investigated systematically by K-M plotter analysis. Seventeen members were significantly associated with all four clinical outcomes, including overall survival (OS), progression-free survival (PFS), relapse-free survival (RFS) and disease-specific survival (DSS) (all P < 0.05; [fig_ref] Figure 1: ARHGAP24 is downregulated in Hepatocellular carcinoma [/fig_ref]. Among above 17 RhoGAP members, 8 members were classified as hazard indicator for HCC prognosis (hazard ratio (HR) > 1), while 9 members including ARHGAP24 protein were considered as protective factors (HR < 1, [fig_ref] Figure 1: ARHGAP24 is downregulated in Hepatocellular carcinoma [/fig_ref] , [fig_ref] Figure 1: ARHGAP24 is downregulated in Hepatocellular carcinoma [/fig_ref]. Because RhoGAPs were conventionally considered to have capacities on inhibiting tumor growth, we selected members with HR < 1 for further analysis. Only ARHGAP24 expression was significantly reduced in HCC tissues compared with normal liver tissues in all GEO (GSE164760, GSE76427, GSE101728 and GSE101685), TCGA and CPTAC databases enrolled (all P < 0.05, [fig_ref] Figure 1: ARHGAP24 is downregulated in Hepatocellular carcinoma [/fig_ref]. We further confirmed experimentally that ARHGAP24 mRNA levels were significantly decreased in patients with recurrent tumors compared with patients with non-recurrent tumors [fig_ref] Figure 1: ARHGAP24 is downregulated in Hepatocellular carcinoma [/fig_ref]. Moreover, ARHGAP24 mRNA levels were also dramatically reduced in HCC tissues when compared to paired non-cancerous tissues [fig_ref] Figure 1: ARHGAP24 is downregulated in Hepatocellular carcinoma [/fig_ref]. Western blotting assays consistently showed that ARHGAP24 protein levels were downregulated in recurrent tumors and HCC tissues [fig_ref] Figure 1: ARHGAP24 is downregulated in Hepatocellular carcinoma [/fig_ref]. We further validated the prognostic value of ARHGAP24 in HCC by immunohistochemistry staining in 131 HCC patients. Representative images were shown in [fig_ref] Figure 1: ARHGAP24 is downregulated in Hepatocellular carcinoma [/fig_ref]. We compared different clinicopathological features of HCC patients and found that ARHGAP24 downregulation was significantly correlated with satellite lesions (P = 0.031), CNLC stage (P = 0.020), microvascular invasion (P = 0.001) and tumor recurrence (P = 0.002) [fig_ref] Figure 1: ARHGAP24 is downregulated in Hepatocellular carcinoma [/fig_ref] , . Besides, patients with low ARHGAP24 expression had significantly shorter PFS compared with those who with high-ARHGAP24 expression (P < 0.01; [fig_ref] Figure 1: ARHGAP24 is downregulated in Hepatocellular carcinoma [/fig_ref] , left). Additionally, patients with low ARHGAP24 expression also had significantly higher early-relapse rate (within 2 years; P < 0.01, [fig_ref] Figure 1: ARHGAP24 is downregulated in Hepatocellular carcinoma [/fig_ref] , right). Notably, patients with low ARHGAP24 in early tumor stage (BCLC: 0 + A; P < 0.01) or low alpha-fetoprotein subgroups (≤ 400 ng/mL; P < 0.05) also had higher probabilities of tumor progression when compared to the patients with high ARHGAP24 [fig_ref] Figure 1: ARHGAP24 is downregulated in Hepatocellular carcinoma [/fig_ref]. Similar results were observed in patients with small tumors (< 5 cm), single tumors, early tumor differentiation and stage, without satellite lesions and without microvascular invasion [fig_ref] Figure 2: ARHGAP24 inhibits proliferation potentials in HCC [/fig_ref]. Univariate Cox regression analysis showed that ARHGAP24, tumor size, BCLC stage and other clinical parameters were associated with tumor progression in HCC patients [fig_ref] Figure 2: ARHGAP24 inhibits proliferation potentials in HCC [/fig_ref]. These factors were further analyzed by multivariate analysis, which revealed that high ARHGAP24 expression in HCC cells was an independent predictive indicator for tumor progression (HR 0.46 (0.23 -0.94), P = 0.034; [fig_ref] Figure 1: ARHGAP24 is downregulated in Hepatocellular carcinoma [/fig_ref].
Interestingly, low ARHGAP24 was also associated with shorter OS in patients with other tumors, such as renal cell carcinoma (P < 0.001), lung cancer (P < 0.05) and pancreatic ductal adenocarcinoma (P = 0.045) [fig_ref] Figure 3: ARHGAP24 restrains invasiveness and metastasis in HCC [/fig_ref] , suggesting its common inhibitory role in tumors. Alterations of proliferation related molecules after ARHGAP24 expression modulations were detected by WB assays. C. Effects of ARHGAP24 overexpression or downregulation on the short-term proliferation potentials were assessed by CCK-8 assays. D. Effects of ARHGAP24 overexpression or downregulation on the long-term proliferation potentials were assessed by colony-formation assays. E. Effects of ARHGAP24 overexpression or downregulation on cell cycle were determined via flow cytometry assays. F. Effects of ARHGAP24 overexpression or downregulation on cell apoptosis were evaluated by flow cytometry assays. G. In vivo inhibitory efficiency of ARHGAP24 on HCC growth was evaluated by constructing orthotopic xenograft models, and representative images, tumor volumes and tumor weights were illustrated.
## Arhgap24 inhibited cell proliferation and induced g0/g1 arrest in hcc
ARHGAP24 expression was determined in six HCC cell lines and one normal liver cells. Quantitative reverse transcription-polymerase chain reaction (qRT-PCR) and western blotting (WB) assays showed that ARHGAP24 was downregulated in HCC cell lines with strong metastatic ability (HCCLM3 and MHCC97H), but was highly expressed in cells with weak metastatic potential (MHCC97L and Li-7) [fig_ref] Figure 2: ARHGAP24 inhibits proliferation potentials in HCC [/fig_ref]. To investigate the functional role of ARHGAP24 in HCC proliferation, we induced stable overexpression of ARHGAP24 in HCCLM3 cells and Huh7 cells (ARH-OE) and stable silencing of ARHGAP24 in Li-7 cells (sh1 and sh2). Western blot assays showed successful overexpression and knockdown of ARHGAP24 expression, respectively [fig_ref] Figure 2: ARHGAP24 inhibits proliferation potentials in HCC [/fig_ref]. Notably, protein levels of cell proliferation and anti-apoptosis markers (PCNA, Bcl2, cyclin D1 and CDK2) were significantly reduced in ARHGAP24-overexpressing cells while increased in ARHGAP24-silenced cells, while the pro-apoptosis marker, caspase-3, exhibited the opposite results [fig_ref] Figure 2: ARHGAP24 inhibits proliferation potentials in HCC [/fig_ref]. Cell Counting Kit-8 and colony-formation assays showed that knockdown of ARHGAP24 in Li-7 cells significantly increased proliferation, while overexpression of ARHGAP24 in HCCLM3 cells had the opposite effect [fig_ref] Figure 2: ARHGAP24 inhibits proliferation potentials in HCC [/fig_ref]. Similarly, proliferation was reduced in ARHGAP24overexpressing Huh7 cells [fig_ref] Figure 4: ARHGAP24 represses the transcriptional activity of β-catenin [/fig_ref]. Additionally, cell cycle assays demonstrated that ARHGAP24 knockdown accelerated the cell cycle in Li-7 cells, while ARHGAP24 overexpression in HCCLM3 and Huh7 cells resulted in G0/G1 arrest [fig_ref] Figure 2: ARHGAP24 inhibits proliferation potentials in HCC [/fig_ref] , [fig_ref] Figure 4: ARHGAP24 represses the transcriptional activity of β-catenin [/fig_ref]. Apoptosis assays further showed that silencing ARHGAP24 prevented apoptosis in HCC cells under serum-free culture for 24 h, while ARHGAP24 overexpression induced apoptosis in HCCLM3 and Huh7 cells [fig_ref] Figure 2: ARHGAP24 inhibits proliferation potentials in HCC [/fig_ref] , [fig_ref] Figure 4: ARHGAP24 represses the transcriptional activity of β-catenin [/fig_ref]. Moreover, analysis of liver orthotopic xenograft tumors further confirmed that ARHGAP24 knockdown promoted tumor growth in vivo, as evidenced by increased tumor volume and weight (P < 0.01). Conversely, the volumes of tumor xenografts derived from HCCLM3 vector control and HCCLM3 ARHGAP24 overexpressing cells were 1845.22 ± 152.17 and 398.25 ± 73.89 mm 3 , respectively (P < 0.01; [fig_ref] Figure 2: ARHGAP24 inhibits proliferation potentials in HCC [/fig_ref] , suggesting that ARHGAP24 overexpression markedly inhibited tumor growth.
## Arhgap24 attenuated cell invasion and tumor metastasis in hcc
Given the correlation between ARHGAP24 expression and microvascular invasion, we hypothesized that ARHGAP24 might play an important role in HCC metastasis. To verify this, we investigated HCC cell migration and invasion in vitro using transwell and wound-healing assays, respectively. The results revealed that knockdown of ARHGAP24 increased the numbers of migrating and invading cancer cells, while ARHGAP24 overexpression decreased these cells [fig_ref] Figure 3: ARHGAP24 restrains invasiveness and metastasis in HCC [/fig_ref] -B, [fig_ref] Figure 5: ARHGAP24 inhibits cell invasiveness mainly through an RhoGAP activity independent manner [/fig_ref]. In addition, we detected epithelial-mesenchymal transition (EMT)related mRNA and protein expression in HCC cell lines. ARHGAP24 knockdown resulted in increased mesenchymal expressions (N-cadherin, vimentin, and MMP-9), while overexpression of ARHGAP24 resulted in an epithelial-like molecular phenotype [fig_ref] Figure 3: ARHGAP24 restrains invasiveness and metastasis in HCC [/fig_ref]. Consistently, overexpression of ARHGAP24 in Huh7 cells also inhibited mesenchymal-like molecular phenotype [fig_ref] Figure 5: ARHGAP24 inhibits cell invasiveness mainly through an RhoGAP activity independent manner [/fig_ref]. Immunofluorescence analysis confirmed that downregulation of ARHGAP24 decreased E-cadherin expression but increased N-cadherin expression in Li-7 cells, while overexpression of ARHGAP24 in HCCLM3 cells produced the opposite effects [fig_ref] Figure 3: ARHGAP24 restrains invasiveness and metastasis in HCC [/fig_ref]. Moreover, phalloidin staining showed that HCCLM3 cells evolved from a mesenchymal to an epithelial morphology following ARHGAP24 overexpression, while Li-7 cells with ARHGAP24 knockdown changed from their original round shape to a shuttle-like shape [fig_ref] Figure 3: ARHGAP24 restrains invasiveness and metastasis in HCC [/fig_ref].
To verify the in vitro experimental results, mice were injected with 5 × 10 6 HCCLM3 cells via the tail vein to observe the typical lung metastasis sites. The incidence of lung metastasis was reduced (50.00% vs. [bib_ref] Effects of variant rs346473 in ARHGAP24 gene on disease progression of HBV..., Liu [/bib_ref].67%) after ARHGAP24 overexpression, while the incidence of lung metastasis of Li-7 cells was increased (0.00% vs. 33.33%) after ARHGAP24 knockdown [fig_ref] Figure 3: ARHGAP24 restrains invasiveness and metastasis in HCC [/fig_ref]. In addition, we detected the expression levels of E-cadherin (epithelial marker), N-cadherin (mesenchymal marker) and Ki67 (cell proliferation marker) in an orthotopic liver xenograft mouse model by immunohistochemistry. E-cadherin expression was downregulated and N-cadherin and Ki67 expression levels were upregulated in liver tumor tissues with ARHGAP24 knockdown compared with the control group. However, we observed the opposite result when ARHGAP24 was overexpressed [fig_ref] Figure 3: ARHGAP24 restrains invasiveness and metastasis in HCC [/fig_ref]. These results indicated that ARHGAP24 inhibited the migration and invasion of liver cancer cells.
## Arhgap24 regulated hcc progression mainly via inhibiting the wnt/β-catenin signaling pathway
To further identify the underlying signaling of ARHGAP24 in HCC, HCC patients from TCGA dataset were divided into high-and low-ARHGAP24 expression groups, according to the upper and lower quartiles of ARHGAP24 expression. Differentially expressed genes (DEGs) were identified as follows: |log2 fold change (FC)| > 1 (ARHGAP24 high verse ARHGAP24 low) and P < 0.05 [fig_ref] Figure 4: ARHGAP24 represses the transcriptional activity of β-catenin [/fig_ref]. Reactome pathway analysis revealed that Wnt/β-catenin signaling pathways were conformably enriched in the low-ARHGAP24 expression group [fig_ref] Figure 4: ARHGAP24 represses the transcriptional activity of β-catenin [/fig_ref]. As Wnt/β-catenin signaling pathway played vital roles in HCC physiological processes, including cell proliferation, migration and invasion, we selected this pathway as the potential candidate for the following investigation. Gene set enrichment analysis (GSEA) showed that enriched pathways related to Wnt/β-catenin pathways were highly activated in the low-ARHGAP24 expression group [fig_ref] Figure 4: ARHGAP24 represses the transcriptional activity of β-catenin [/fig_ref]. Furthermore, ARHGAP24 knockdown significantly augmented β-catenin transcriptional activity as demonstrated by TOP/FOP Flash reporter assay in Li-7 cells, while its overexpression suppressed β-catenin transcription in HCCLM3 cells [fig_ref] Figure 4: ARHGAP24 represses the transcriptional activity of β-catenin [/fig_ref]. qRT-PCR and western blot assays revealed that silencing ARHGAP24 enhanced the expression of the downstream target genes, such as MYC and CCND1, of the Wnt/β-catenin pathway, while ARHGAP24 overexpression inhibited their expressions [fig_ref] Figure 4: ARHGAP24 represses the transcriptional activity of β-catenin [/fig_ref] , [fig_ref] Figure 5: ARHGAP24 inhibits cell invasiveness mainly through an RhoGAP activity independent manner [/fig_ref]. To validate the critical role of β-catenin transactivation in ARHGAP24-regulated process, we further treated HCCLM3 cells (low-ARHGAP24 expression) and ARHGAP24-knockdown Li-7 cells with ICG-001, a high specific inhibitor of β-catenin transcriptional activity. Expression levels of downstream target genes of β-catenin were reduced in HCCLM3 cells [fig_ref] Figure 4: ARHGAP24 represses the transcriptional activity of β-catenin [/fig_ref] , as shown by qRT-PCR and western blotting assays. Proliferation, migration and invasion capacities were significantly restrained after treatment with ICG-001 in HCCLM3 cells [fig_ref] Figure 4: ARHGAP24 represses the transcriptional activity of β-catenin [/fig_ref]. Notably, the addition of ICG-001 to Li-7 cells alleviated the pro-HCC effects, including the increases in cell proliferation, migration, invasion and β-catenin activity caused by ARHGAP24 knockdown [fig_ref] Figure 4: ARHGAP24 represses the transcriptional activity of β-catenin [/fig_ref]. Collectively, these data suggested that ARHGAP24 suppressed HCC cell proliferation and invasion mainly by inhibiting the transcriptional activity of β-catenin.
## Arhgap24 inhibited the transcriptional activity of β-catenin mainly by an enzyme-independent manner
β-catenin expression and localization were critical for the activation of Wnt/β-catenin signaling [bib_ref] Wnt-β-catenin signalling in liver development, health and disease, Perugorria [/bib_ref]. We further investigated how ARHGAP24 regulated β-catenin signaling by detecting the expression and subcellular distribution of β-catenin. qRT-PCR and western blot assays revealed that knockdown of ARHGAP24 in Li-7 cells and overexpression of ARHGAP24 in HCCLM3 cells did not affect β-catenin expression [fig_ref] Figure 5: ARHGAP24 inhibits cell invasiveness mainly through an RhoGAP activity independent manner [/fig_ref]. Overexpression of ARHGAP24 also had no effect on the intracellular distribution of β-catenin in HCCLM3 cells, while knockdown of ARHGAP24 in Li-7 cells promoted the nuclear accumulation of β-catenin [fig_ref] Figure 5: ARHGAP24 inhibits cell invasiveness mainly through an RhoGAP activity independent manner [/fig_ref]. Immunofluorescence assay showed similar results [fig_ref] Figure 5: ARHGAP24 inhibits cell invasiveness mainly through an RhoGAP activity independent manner [/fig_ref]. Moreover, overexpression of ARHGAP24 in Huh7 cells slightly reduced β-catenin protein expression but not mRNA expression, and also reduced the nuclear distribution of β-catenin [fig_ref] Figure 5: ARHGAP24 inhibits cell invasiveness mainly through an RhoGAP activity independent manner [/fig_ref]. Previous studies reported that ARHGAP24 was a Rac-specific RhoGAP, inactivating Rac1 and thereby inhibiting tumor progression [bib_ref] FilGAP, a Rho-and ROCK-regulated GAP for Rac binds filamin A to control..., Ohta [/bib_ref]. We therefore examined the Rho GTPase activity of ARHGAP24 and observed the effects of Rac1 activation on β-catenin transcriptional activity. We pulled-down GTP-Rac1, GTP-RhoA and GTP-CDC42 using GST-tagged fusion protein beads, as a well-recognized approach for evaluating Rho GTPase activity, followed by immunoblotting assays to determine the levels of the GTP-bound fractions of Rac1, RHOA and CDC42 after modulation of ARHGAP24 expression. The results showed that GTP-RAC1 and GTP-CDC42 levels were decreased in ARHGAP24-overexpressing Huh7 cells, whereas knockdown of ARHGAP24 in Li-7 cells resulted in slight increases in GTP-Rac1 and GTP-CDC42 levels. However, ARHGAP24 overexpression in HCCLM3 cells had no effects on levels of GTP-RAC1, GTP-CDC42 and GTP-RHOA [fig_ref] Figure 5: ARHGAP24 inhibits cell invasiveness mainly through an RhoGAP activity independent manner [/fig_ref]. Strangely, we observed a phenomenon that ARHGAP24 exhibited common inhibitory effects in HCC which was unparalleled with its inhibitory effects on RAC1. We therefore raised a hypothesis that there might be an unknown but enzyme-independent mechanism in addition to inhibition of canonical RAC1 pathway.
To further verify our hypothesis, we constructed a GAP-deficient mutant (Q158R) of the ARHGAP24 gene, encoding a protein lacking RhoGAP activity (ARH-MUT, [fig_ref] Figure 6: ARHGAP24 restrains β-catenin signaling via downregulating PKM2 protein abundance [/fig_ref] , which failed to inhibit RAC1 activation [fig_ref] Figure 5: ARHGAP24 inhibits cell invasiveness mainly through an RhoGAP activity independent manner [/fig_ref]. Biological function assays showed that ARH-MUT protein could also effectively restrain the growth and migration of HCCLM3 cells, and the inhibitory effects were consistent with wild-type ARHGAP24 (ARH-WT) protein [fig_ref] Figure 5: ARHGAP24 inhibits cell invasiveness mainly through an RhoGAP activity independent manner [/fig_ref]. Importantly, despite the loss of function in terms of suppressing Rac1 activity and β-catenin nuclear accumulation [fig_ref] Figure 5: ARHGAP24 inhibits cell invasiveness mainly through an RhoGAP activity independent manner [/fig_ref] , ARH-MUT also exerted inhibitory effects on the proliferation, mesenchymal-like phenotype and invasiveness potential in Huh7 cells as the ARH-WT did [fig_ref] Figure 5: ARHGAP24 inhibits cell invasiveness mainly through an RhoGAP activity independent manner [/fig_ref]. Similarly, ARH-MUT expression still successfully restrained β-catenin transactivation and transcriptional activity [fig_ref] Figure 5: ARHGAP24 inhibits cell invasiveness mainly through an RhoGAP activity independent manner [/fig_ref] , [fig_ref] Figure 6: ARHGAP24 restrains β-catenin signaling via downregulating PKM2 protein abundance [/fig_ref]. Overall, above findings indicated that, despite the cell-specific function of ARHGAP24 in inhibiting Rho activity in HCC cell lines, ARHGAP24 mainly inhibited the transcriptional activity of β-catenin in a Rho-GTPase-independent manner.
## Arhgap24 interacted and reduced pkm2 to retrain the transcriptional activity of β-catenin
RhoGAPs shed their enzyme-independent functions mainly by working as a scaffold to facilitate interactions between other proteins to sustain or restrain tumor progression [bib_ref] ArhGAP30 promotes p53 acetylation and function in colorectal cancer, Wang [/bib_ref] [bib_ref] Arhgap36-dependent activation of Gli transcription factors, Rack [/bib_ref]. We therefore performed LC-MS/MS to identify proteins interacting with ARHGAP24 in ARHGAP24 overexpressed HCCLM3 cells [fig_ref] Figure 6: ARHGAP24 restrains β-catenin signaling via downregulating PKM2 protein abundance [/fig_ref]. The results identified pyruvate kinase (PKM) as the top-ranked protein in addition to ARHGAP24 . Previous studies reported that PKM2 was highly expressed in liver cancer tissues and promoted cancer cell proliferation and invasion. Importantly, PKM2 expression was reported to be the vital enhancer for β-catenin transactivation [bib_ref] ERK1/2-dependent phosphorylation and nuclear translocation of PKM2 promotes the Warburg effect, Yang [/bib_ref] [bib_ref] Nuclear PKM2 regulates β-catenin transactivation upon EGFR activation, Yang [/bib_ref]. We carried out further immunoprecipitation (IP) assays to confirm the interaction between PKM2 and ARHGAP24 [fig_ref] Figure 6: ARHGAP24 restrains β-catenin signaling via downregulating PKM2 protein abundance [/fig_ref]. Interestingly, PKM2 mRNA expression was not affected by ARHGAP24 modulation, as shown by qRT-PCR assays [fig_ref] Figure 6: ARHGAP24 restrains β-catenin signaling via downregulating PKM2 protein abundance [/fig_ref] ; however, PKM2 protein expression was significantly increased after ARHGAP24 knockdwon or decreased after ARHGAP24 overexpression, according to western blotting [fig_ref] Figure 6: ARHGAP24 restrains β-catenin signaling via downregulating PKM2 protein abundance [/fig_ref] and confirmed by immunofluoresence staining [fig_ref] Figure 6: ARHGAP24 restrains β-catenin signaling via downregulating PKM2 protein abundance [/fig_ref]. To clarify if the ARHGAP24 modulation of cell proliferation and invasion was dependent on PKM2, we silenced PKM2 in HCCLM3 and Li-7-shARH cells [fig_ref] Figure 7: ARHGAP24 enhances PKM2 ubiquitination by promoting the interaction of PKM2 and WWP1 [/fig_ref]. PKM2 knockdown suppressed the mRNA and protein expression levels of EMT-related markers and downstream target genes of β-catenin in HCCLM3, which mimicked the inhibitory effects of ARHGAP24 overexpression. Notably, interfering PKM2 in Li-7-shARH cells almost abrogated the enhanced expressions of EMT-related markers and downstream target genes of β-catenin caused by ARHGAP24 knockdown, as shown by qRT-PCR and western blot assays [fig_ref] Figure 6: ARHGAP24 restrains β-catenin signaling via downregulating PKM2 protein abundance [/fig_ref]. Moreover, silencing PKM2 greatly abolished the increased β-catenin transcriptional activity resulted from ARHGAP24 knockdown, without affecting RAC1 activity [fig_ref] Figure 6: ARHGAP24 restrains β-catenin signaling via downregulating PKM2 protein abundance [/fig_ref]. Functional experiments further revealed that silencing PKM2 decreased the proliferation and invasion of HCCLM3 as the ARHGAP24 knockdown did, and significantly abolished the promotional effects of ARHGAP24 silence on the invasiveness and growth in Li-7 cells [fig_ref] Figure 6: ARHGAP24 restrains β-catenin signaling via downregulating PKM2 protein abundance [/fig_ref]. Collectively, these findings revealed that ARHGAP24 suppressed β-catenin transactivation by interacting with PKM2 to decrease its protein expression in HCC.
## Arhgap24 degraded pkm2 via the ubiquitinproteasome pathway by recruiting wwp1
Since expression of PKM2 mRNA was not affected by ARHGAP24 expression modulation, we speculated that ARHGAP24 might exert its function by inhibiting the stability of PKM2 protein. We tested this hypothesis by cycloheximide chase assays to investigate the stability of PKM2. Overexpression of ARHGAP24 in HCCLM3 cells substantially decreased the half-life of PKM2 protein [fig_ref] Figure 7: ARHGAP24 enhances PKM2 ubiquitination by promoting the interaction of PKM2 and WWP1 [/fig_ref] , while knockdown of ARHGAP24 in Li-7 cells remarkably increased the half-life of PKM2 [fig_ref] Figure 7: ARHGAP24 enhances PKM2 ubiquitination by promoting the interaction of PKM2 and WWP1 [/fig_ref]. We further explored the mechanism underlying ARHGAP24-induced PKM2 protein degradation by adding the proteasome inhibitor MG132 or the lysosome inhibitor NH4CI to the culture medium of Li-7 (high ARHGAP24 expression) and HCCLM3 cells (ARHGAP24 overexpression). The results showed that MG132 could effectively rescue the decrease in PKM2 induced by high ARHGAP24 expression, while NH 4 CI showed weaker effects [fig_ref] Figure 7: ARHGAP24 enhances PKM2 ubiquitination by promoting the interaction of PKM2 and WWP1 [/fig_ref]. Furthermore, ARHGAP24 knockdown inhibited but ARHGAP24 overexpression promoted the ubiquitination of PKM2 [fig_ref] Figure 7: ARHGAP24 enhances PKM2 ubiquitination by promoting the interaction of PKM2 and WWP1 [/fig_ref]. These results suggest that ARHGAP24-mediated PKM2 protein decrease mainly via E3 ligase-induced ubiquitination, followed by proteasomal degradation pathway.
We further analyzed the potential ARHGAP24binding candidate E3 ligases from by Co-IP combined with MS in ARHGAP24-FLAG-overexpressing cells. WWP1 was identified as the only potential E3 ligase candidate.
We therefore hypothesized that ARHGAP24-enhanced PKM2 ubiquitination was dependent on WWP1, via formation of a regulator complex. Co-IP analysis confirmed that these three proteins formed a complex in HCCLM3 cells transfected with Flag-tagged ARHGAP24 [fig_ref] Figure 7: ARHGAP24 enhances PKM2 ubiquitination by promoting the interaction of PKM2 and WWP1 [/fig_ref] , and exogenous, Flag-tagged PKM2 could successfully interact with endogenous WWP1 in ARHGAP24-high Li-7 cells [fig_ref] Figure 7: ARHGAP24 enhances PKM2 ubiquitination by promoting the interaction of PKM2 and WWP1 [/fig_ref]. The interactions between ARHGAP24 and PKM2, WWP1 and ARHGAP24 were also confirmed by IP analysis in HEK293T cells [fig_ref] Figure 7: ARHGAP24 enhances PKM2 ubiquitination by promoting the interaction of PKM2 and WWP1 [/fig_ref]. Notably, knockdown of WWP1 abrogated the decrease in PKM2 induced by high ARHGAP24 expression, but expression levels of ARHGAP24 and WWP1 were not affected by each other [fig_ref] Figure 7: ARHGAP24 enhances PKM2 ubiquitination by promoting the interaction of PKM2 and WWP1 [/fig_ref]. ARHGAP24-mediated PKM2 ubiquitination was markedly weakened when WWP1 was silenced [fig_ref] Figure 7: ARHGAP24 enhances PKM2 ubiquitination by promoting the interaction of PKM2 and WWP1 [/fig_ref]. Furthermore, co-transfection of His-WWP1 plasmids with FLAG-PKM2 and HA-ARH in HEK293 cells promoted WWP1-PKM2 interactions [fig_ref] Figure 7: ARHGAP24 enhances PKM2 ubiquitination by promoting the interaction of PKM2 and WWP1 [/fig_ref]. Consistently, more ubiquitinated PKM2 proteins were immunoprecipitated from cells co-expressing the three plasmids compared with cells co-transfected with WWP1 and PKM2 [fig_ref] Figure 7: ARHGAP24 enhances PKM2 ubiquitination by promoting the interaction of PKM2 and WWP1 [/fig_ref]. In contrast, WWP1-PKM2 interactions and WWP1-mediated PKM2 ubiquitination were markedly weakened by ARHGAP24 knockdown in Li-7 cells [fig_ref] Figure 7: ARHGAP24 enhances PKM2 ubiquitination by promoting the interaction of PKM2 and WWP1 [/fig_ref]. These data collectively indicate that ARHGAP24 acts as a scaffolding protein to potentiate WWP1-mediated PKM2 ubiquitination and degradation.
## C-terminal region of arhgap24 was responsible for the scaffolding function
To further verify the role of ARHGAP24 as a scaffold, we transfected different concentrations of ARHGAP24 plasmids into HCCLM3 and MHCC97H cells, and showed that interaction of PKM2 and WWP1 was increased in ARHGAP24-overexpressing cells compared with control cells, as shown by western blot and IP assays [fig_ref] Figure 8: The C-terminal region of ARHGAP24 for the interaction of WWP1 and PKM2 [/fig_ref] , validating the critical role of ARHGAP24 as a scaffold. Deletion mutants of HA-tagged ARHGAP24 and FLAG-tagged PKM2 were constructed to further identify the binding motifs for the ARHGAP24-PKM2 interaction. Representative images were shown in [fig_ref] Figure 8: The C-terminal region of ARHGAP24 for the interaction of WWP1 and PKM2 [/fig_ref]. We mapped the domain that accounted for ARHGAP24 binding to PKM2 by generating expression constructs for full-length HA-tagged ARHGAP24 (ARH-FL-HA) and a series of ARHGAP24 mutants lacking different domains, including ARH-PHD, ARH-GAPD, ARH-CD and ARH-ΔCD, and co-expressed these constructs along with full-length FLAG-tagged PKM2. The results indicated that the C-terminal domain of ARH (ARH-CD-HA) was responsible for the interaction with PKM2 [fig_ref] Figure 8: The C-terminal region of ARHGAP24 for the interaction of WWP1 and PKM2 [/fig_ref]. Additionally, we transfected expression vectors encoding ARH-FL-HA or deletion mutants (PKM2-ΔCD, PKM2-ABD, PKM2-CD) and FLAG-tagged PKM2 into HEK293T cells, followed by IP and western blot assays with anti-HA or anti-PKM2 antibody, and showed that the C-terminal of PKM2 interacted with ARHGAP24 [fig_ref] Figure 8: The C-terminal region of ARHGAP24 for the interaction of WWP1 and PKM2 [/fig_ref]. Furthermore, bioinformatics analysis of ARHGAP24-PKM2 interactions revealed that the interaction domains of both proteins were at their respective C-terminal, with a probability > 70% [fig_ref] Figure 8: The C-terminal region of ARHGAP24 for the interaction of WWP1 and PKM2 [/fig_ref]. Representative protein structures of PKM2 and ARHGAP24 and their interaction domains are shown in [fig_ref] Figure 8: The C-terminal region of ARHGAP24 for the interaction of WWP1 and PKM2 [/fig_ref]. mRNA expression of downstream target genes of β-catenin was detected by qRT-PCR assay in HCCLM3 cells. J. Effects of wild-type ARHGAP24 (ARH-WT) and Q158R mutant ARHGAPP24 (ARH-MUT) overexpression on proliferation potentials of HCCLM3 were determined by CCK8 assays. K. The transcriptional activity of β-catenin in ARH-WT and ARH-MUT overexpressed HCCLM3 cells was analyzed by TOP/FOP Flash assay. L. Effects of ARH-WT and ARH-MUT overexpression on migration and invasion capacities of HCCLM3 cells were validated by Transwell assays. M. mRNA expression of downstream target genes of β-catenin was detected by qRT-PCR assay in Huh7 cells. N. Effects of wild-type ARHGAP24 (ARH-WT) and Q158R mutant ARHGAPP24 (ARH-MUT) overexpression on proliferation potentials of Huh7 cells were determined by CCK8 assays. O. Effects of ARH-WT and ARH-MUT overexpression on migration and invasion capacities of Huh7 cells were validated by Transwell assays. P. The transcriptional activity of β-catenin in ARH-WT and ARH-MUT overexpressed Huh7 cells was analyzed by TOP/FOP Flash assay. To confirm the role of the functional domain of ARHGAP24 in PKM2 protein stability and ubiquitination, we transfected expression vectors encoding FL-ARH-HA or deletion mutant (ARH-ΔCD-HA) into HCCLM3 cells. Results showed that mutant ARHGAP24 failed to induce ubiquitination of PKM2 as the WT-ARHGAP24 did [fig_ref] Figure 8: The C-terminal region of ARHGAP24 for the interaction of WWP1 and PKM2 [/fig_ref]. Notably, PKM2 and WWP1 proteins in HCCLM3 cells could not be immunoprecipitated with ARHGAP24 when the C-terminal domain of ARHGAP24 was deleted [fig_ref] Figure 8: The C-terminal region of ARHGAP24 for the interaction of WWP1 and PKM2 [/fig_ref]. Furthermore, C-terminal domain-deleted ARHGAP24 also failed to induce ubiquitination of PKM2 [fig_ref] Figure 8: The C-terminal region of ARHGAP24 for the interaction of WWP1 and PKM2 [/fig_ref]. These findings consistently suggested that ARHGAP24 serves as a scaffolding protein that recruits WWP1 to PKM2 via its C-terminal. We further transfected expression vectors encoding PKM2-FL-FLAG, WWP1-FL-His and different deletion mutants of ARHGAP24, including ARH-FL-HA, M1 (amino acids (aa) 1 -630), M2 (aa 1 -530), M3 (aa 1 -430) and M4 (aa 1 -330) into HEK293T cells, followed by IP and western blot assays. The results showed that the C-terminal fragments 329 -430 aa and 631 -748 aa of ARHGAP24 bound directly to WWP1 and PKM2, respectively [fig_ref] Figure 8: The C-terminal region of ARHGAP24 for the interaction of WWP1 and PKM2 [/fig_ref]. A schematic diagram of the mechanism of ARHGAP24 in HCC progression [fig_ref] Figure 8: The C-terminal region of ARHGAP24 for the interaction of WWP1 and PKM2 [/fig_ref] and the structural domains of ARHGAP24 interacting with WWP1 and PKM2 are shown in [fig_ref] Figure 8: The C-terminal region of ARHGAP24 for the interaction of WWP1 and PKM2 [/fig_ref].
# Discussion
Abnormalities in Rho GTPase activation have major consequences for cancer cell proliferation and metastasis [bib_ref] Rho GTPase signaling complexes in cell migration and invasion, Lawson [/bib_ref] [bib_ref] RhoE is frequently downregulated in hepatocellular carcinoma (HCC) and suppresses HCC invasion..., Ma [/bib_ref]. RhoGAPs have been shown to inhibit the activation of Rho GTPase with an inactive GDP-bound state [bib_ref] Rho GTPase regulatory proteins in podocytes, Matsuda [/bib_ref] [bib_ref] Targeting the Small GTPase Superfamily through Their Regulatory Proteins, Gray [/bib_ref]. RhoGAPs thus inactivate Rho GTPases (Rac1, CDC42) and have generally been presumed to act as tumor suppressors. However, the tumor suppressor roles of RhoGAPs have generally been reported in tumors other than HCC. For example, low ARHGAP30 expression promoted the proliferation and migration of colorectal carcinoma cells [bib_ref] ArhGAP30 promotes p53 acetylation and function in colorectal cancer, Wang [/bib_ref]. Yagi found that ARAP3 expression inhibited cancer invasiveness by modulating cell adhesion and motility [bib_ref] ARAP3 inhibits peritoneal dissemination of scirrhous gastric carcinoma cells by regulating cell..., Yagi [/bib_ref]. Knockdown of ARHGAP15 resulted in activation of PAK1/2 and indirectly promoted Rac activation, thereby enhancing cancer-promoting signal transduction [bib_ref] ArhGAP15, a Rac-specific GTPase-activating protein, plays a dual role in inhibiting small..., Radu [/bib_ref]. Other RhoGAPs, such as ARHGAP4, ARHGAP6, ARHGAP9, ARHGAP12 and ARHGAP25, have also demonstrated tumor suppressor roles in different types of tumors [bib_ref] ARHGAP4 regulates the cell migration and invasion of pancreatic cancer by the..., Shen [/bib_ref] [bib_ref] ARHGAP6 regulates the proliferation, migration and invasion of lung cancer cells, Wu [/bib_ref] [bib_ref] ARHGAP9 suppresses the migration and invasion of hepatocellular carcinoma cells through up-regulating..., Zhang [/bib_ref] [bib_ref] Met-driven invasive growth involves transcriptional regulation of Arhgap12, Gentile [/bib_ref] [bib_ref] ARHGAP25 Inhibits Pancreatic Adenocarcinoma Growth by Suppressing Glycolysis via AKT/ mTOR Pathway, Huang [/bib_ref]. However, no study has systematically investigated the clinical significance of RhoGAPs and analyzed the functions of significant molecules in HCC. DLC1 and ARHGAP9 were reported to be tumor suppressors inhibiting HCC progression [bib_ref] ARHGAP9 suppresses the migration and invasion of hepatocellular carcinoma cells through up-regulating..., Zhang [/bib_ref] [bib_ref] Integrative genomic identification of genes on 8p associated with hepatocellular carcinoma progression..., Roessler [/bib_ref]. In the current study, we investigated the prognostic value of 64 RhoGAPs using TCGA and GEO databases, and showed that ARHGAP24 was significantly correlated with tumor progression. We also confirmed that ARHGAP24 was an independent indicator for TTR and OS in HCC patients. Importantly, low ARHGAP24 expression could help clinicians to identify HCC patients at high risk of recurrence, such as patients with alphafetoprotein < 400 µg/µL and early-stage tumors.
Previous studies reported that ARHGAP24 was a Rac-specific Rho-GTPase-activating protein that could inhibit cell morphology, migration and invasion [bib_ref] ARHGAP24 ameliorates inflammatory response through inactivating Rac1/Akt/NF-κB pathway in acute pneumonia model..., Liu [/bib_ref] [bib_ref] Arhgap24 inactivates Rac1 in mouse podocytes, and a mutant form is associated..., Akilesh [/bib_ref] [bib_ref] Systems analysis of RhoGEF and RhoGAP regulatory proteins reveals spatially organized RAC1..., Müller [/bib_ref] , in accordance with the current results showing dramatic differences in the inhibitory function of ARHGAP24 on Rac1 signaling among different cell lines. We speculated that this phenomenon might be attributed to the distinct intracellular environment. Previous studies showed that Rac1 signaling was typically regulated by RhoGEFs, RhoGAPs and guanine nucleotide dissociation inhibitors (RhoGDIs), and balancing the Rho signaling responses required coordination among all these factors [bib_ref] Targeting the Small GTPase Superfamily through Their Regulatory Proteins, Gray [/bib_ref] [bib_ref] Systems analysis of RhoGEF and RhoGAP regulatory proteins reveals spatially organized RAC1..., Müller [/bib_ref]. For instance, Feng et al. found that RASAL2 promoted small GTPase Rac1 signaling, which could bind and antagonize the Rac1-GAP protein ARHGAP24 in breast cancer [bib_ref] RASAL2 activates RAC1 to promote triple-negative breast cancer progression, Feng [/bib_ref]. Interestingly, the highly metastatic HCCLM3 HCC cell line was reported to show high RhoGEF expression (DOCK1, IQGAP1) [bib_ref] Genome-wide CRISPR screen identifies synthetic lethality between DOCK1 inhibition and metformin in..., Feng [/bib_ref] [bib_ref] AMD1 upregulates hepatocellular carcinoma cells stemness by FTO mediated mRNA demethylation, Bian [/bib_ref] , which might activate Rac1 and maintain GTP-bound Rac1 at a high level regardless of ARHGAP24 overexpression. On the other hand, modulation of ARHGAP24 expression might break the intracellular balance among RhoGEF, RhoGAP and RhoGDIs in Huh7 and Li-7 cells, resulting in alteration of Rac1 activity.
In fact, overexpression of ARHGAP24 could still result in dramatic inhibition of cell proliferation and invasion of HCCLM3 cells, regardless of the alteration in Rac1 activity (i.e., overexpression of ARHGAP24 could led to decreased proliferation and invasion but have no influence on GTP-Rac1 or GTP-CDC42 levels). Rac1 inactivation resulting from ARHGAP24 overexpression led to less nuclear accumulation of β-catenin in Huh7 cells, which might also contribute to the suppression of invasiveness of HCC cells. However, the present study found that forced expression of a mutant ARHGAP (Q158R) in Huh7 cells, which failed to inactivate Rac1 activity, had similar inhibitory effects on the proliferation, migration and invasion capacities of HCC cells without impairing Rac1 activity, as well as on the nuclear accumulation of β-catenin, a process regulated by Rac1 and considered to be crucial for β-catenin signaling activation. This lack of an association between the biological function and inhibition potentials of Rac1 signaling suggest that the main function of ARHGAP24 was not dependent on its enzymatic activity, and thus did not rely on its role in regulating Rac1 signaling. RhoGAPs have multiple domains. One RhoGAP domain contains catalytic arginine and thus maintains the GDP activation of Rho proteins, as the major mechanism for regulating cancer cell migration and invasion [bib_ref] Systems analysis of RhoGEF and RhoGAP regulatory proteins reveals spatially organized RAC1..., Müller [/bib_ref]. However, other domains with unknown structures and functions may also be responsible for cell biological functions. Members of RhoGAPs could work as a scaffold to facilitate interactions between other proteins during tumor development, in an enzyme-independent manner. Wang et al. showed that ARHGAP30 promoted p53 acetylation and function independent of RhoGAP activity [bib_ref] ArhGAP30 promotes p53 acetylation and function in colorectal cancer, Wang [/bib_ref] , while Yang et al. found that DLC1 interactions with S100A10 did not affect its RhoGAP activity [bib_ref] DLC1 interaction with S100A10 mediates inhibition of in vitro cell invasion and..., Yang [/bib_ref]. We therefore reasoned that ARHGAP24 might also exert its suppressive function as a scaffold.
In our study, LC/MS revealed PKM2 as a new binding partner of ARHGAP24 and showed that it could recruit WWP1, an E3 ligase, to promote ubiquitination as well as proteasomal degradation of PKM2 in HCC cells, even in Rac1 signaling-activated HCC cell lines. Notably, our results further revealed that the C-terminal of ARHGAP24 (fragments 329 -430 and 631 -748 aa), rather than the RhoGAP domain (135 -330 aa), bound directly to WWP1 and PKM2. These findings consistently suggest that ARHGAP24 serves as a scaffolding protein that recruits WWP1 to PKM2 via its C-terminal. As reported previously, Rac1 activated β-catenin signaling mainly by facilitating its nuclear accumulation [bib_ref] Rac1 activation controls nuclear localization of beta-catenin during canonical Wnt signaling, Wu [/bib_ref] [bib_ref] Trio cooperates with Myh9 to regulate neural crest-derived craniofacial development, Guo [/bib_ref]. However, the transcription activity of β-catenin was controlled by PKM2, indicating that Rac1 worked as an upstream regulator for β-catenin signaling, while PKM2 played a more elemental role and acted as a more crucial hub for β-catenin activation than Rac1. Knockdown of PKM2 accordingly resulted in a significant decrease in invasiveness among HCCLM3 cells in which Rac1 was highly activated. Critically, downregulation of PKM2 also abolished the promotional effects of ARHGAP24 knockdown on β-catenin transcription activities and invasive potential in Rac1-activated HCC cells without suppression of Rac1 activity (ARHGAP24-knockdown Li-7 cells). Yang et al. accordingly revealed that PKM2 regulated β-catenin transactivation upon epidermal growth factor receptor activation, and PKM2 depletion significantly inhibited the binding of β-catenin to the promoter region of CCND1 and MYC [bib_ref] ERK1/2-dependent phosphorylation and nuclear translocation of PKM2 promotes the Warburg effect, Yang [/bib_ref] [bib_ref] Nuclear PKM2 regulates β-catenin transactivation upon EGFR activation, Yang [/bib_ref]. Our data thus identified a novel role for ARHGAP24 in restraining PKM2 abundance by serving as a scaffold in HCC, independent of Rac1 activation.
PKM2 has been found to be highly expressed in various cancers [bib_ref] Pyruvate kinase M2 (PKM2) in cancer and cancer therapeutics, Zhu [/bib_ref]. It can serve as a rate-limiting enzyme of cellular glycolysis or a transcriptional coactivator to promote cancer cell proliferation and invasion [bib_ref] PKM2 coordinates glycolysis with mitochondrial fusion and oxidative phosphorylation, Li [/bib_ref] [bib_ref] Pyruvate kinase M2 is a PHD3-stimulated coactivator for hypoxia-inducible factor 1, Luo [/bib_ref]. Exploration of the underlying mechanisms responsible for the high expression of PKM2 will therefore provide new insights into HCC therapy. Ubiquitination modification was recently determined to be the core mechanism regulating intracellular protein stability, which is closely related to the expression of PKM2 [bib_ref] LINC01554-Mediated Glucose Metabolism Reprogramming Suppresses Tumorigenicity in Hepatocellular Carcinoma via Downregulating PKM2..., Zheng [/bib_ref] [bib_ref] E3 ligase ZFP91 inhibits Hepatocellular Carcinoma Metabolism Reprogramming by regulating PKM splicing, Chen [/bib_ref]. Additionally, phosphorylation of PKM2 at Thr328, Thr454 and Tyr105 also relies on ubiquitination modification to maintain the stability of PKM2 [bib_ref] HSP90 promotes cell glycolysis, proliferation and inhibits apoptosis by regulating PKM2 abundance..., Xu [/bib_ref] [bib_ref] Proviral insertion in murine lymphomas 2 (PIM2) oncogene phosphorylates pyruvate kinase M2..., Yu [/bib_ref] [bib_ref] 2'-hydroxycinnamaldehyde inhibits cancer cell proliferation and tumor growth by targeting the pyruvate..., Yoon [/bib_ref]. However, the critical E3 ligases that directly mediate the degradation of PKM2 are rarely reported. Chen et al. found that E3 ligase ZFP91 promoted the ubiquitination of hnRNPA1 and proteasomal degradation, thereby resulting in PKM2 splicing [bib_ref] E3 ligase ZFP91 inhibits Hepatocellular Carcinoma Metabolism Reprogramming by regulating PKM splicing, Chen [/bib_ref]. It has also been shown that PKM2 protein stability is regulated by Parkin, TRIM58 and CHIP E3 ligases [bib_ref] TRIM58 Interacts with Pyruvate Kinase M2 to Inhibit Tumorigenicity in Human Osteosarcoma..., Yuan [/bib_ref] [bib_ref] CHIP/Stub1 regulates the Warburg effect by promoting degradation of PKM2 in ovarian..., Shang [/bib_ref]. In our study, co-IP together with MS revealed that WWP1 was a novel E3 ligase that directly degraded PKM2. These results suggest that WWP1 could identify a novel substrate, PKM2, in high ARHGAP24-expressing cells and degrade it to regulate cell proliferation and invasion.
Several oncogenes related to the progression and poor prognosis of tumors were recently identified as suppressor genes, including MYH9, USP9X and PHKB [bib_ref] Myosin Heavy Chain 9: Oncogene or Tumor Suppressor Gene?, Wang [/bib_ref] [bib_ref] PTENα and PTENβ promote carcinogenesis through WDR5 and H3K4 trimethylation, Shen [/bib_ref] [bib_ref] Phosphorylase Kinase β Represents a Novel Prognostic Biomarker and Inhibits Malignant Phenotypes..., Yang [/bib_ref]. Additionally, canonical tumor suppressors, such as TP53 and PTEN, have been found to promote carcinogenesis [bib_ref] PTENα and PTENβ promote carcinogenesis through WDR5 and H3K4 trimethylation, Shen [/bib_ref] [bib_ref] p53: 800 million years of evolution and 40 years of discovery, Levine [/bib_ref]. WWP1 has been implicated as an oncogene in breast, prostate and liver cancer [bib_ref] CircWAC induces chemotherapeutic resistance in triple-negative breast cancer by targeting miR-142, upregulating..., Wang [/bib_ref] [bib_ref] Ubiquitin E3 ligase WWP1 as an oncogenic factor in human prostate cancer, Chen [/bib_ref] [bib_ref] Knockdown of WWP1 inhibits growth and induces apoptosis in hepatoma carcinoma cells..., Cheng [/bib_ref] , and has been identified as a physical PTEN interactor, inducing polyubiquitination of PTEN to suppress its dimerization and membrane recruitment and unleash its tumor suppressive activity [bib_ref] WWP1 Gain-of-Function Inactivation of PTEN in Cancer Predisposition, Lee [/bib_ref]. ARID5a was also found to be a substrate of WWP1, and degradation of ARID5a resulted in the amplification of interleukin-6 expression, thereby inducing further inflammation [bib_ref] TLR4-induced NF-κB and MAPK signaling regulate the IL-6 mRNA stabilizing protein Arid5a, Nyati [/bib_ref]. However, we unexpectedly found that WWP1 had a tumor suppressor role. When ARHGAP24 is highly expressed in cancer cells, it may recruit WWP1 to form protein complexes and then promote PKM2 degradation. Taken together, our findings suggest that WWP1 can serve as an oncogene or a tumor suppressor, depending on its interactions with different substrates. Notably, the tumor suppressor role of WWP1 in HCC was mediated by ARHGAP24 expression.
# Conclusions
This study was the first to identify ARHGAP24 as an independent prognostic indicator for HCC. Functional experiments revealed that it inhibits HCC progression and metastasis independent of RhoGAP activity, but attenuates β-catenin transactivation upon PKM2 degradation. Importantly, we identified a novel E3 ligase, WWP1, that can be recruited by the C-terminal of ARHGAP24 and subsequently induce proteasomal degradation of PKM2. These findings establish a novel function of RhoGAPs in HCC and provide a promising therapeutic target for HCC.
[fig] Figure 1: ARHGAP24 is downregulated in Hepatocellular carcinoma (HCC) and predicts poor prognosis. A. Prognostic significance of RhoGAP family members in predicting progression-free survival (PFS), overall survival (OS), relapse-free survival (RFS) and disease-specific survival (DSS) according to K-M plotter database. B. Lists of RhoGAPs as significant hazard factor (all HR > 1, left) or protective factor (all HR < 1, right) for prognosis in HCC according to K-M plotter database. C. Comparisons of indicated RhoGAP family members expressions between tumoral and normal liver tissues according to GEO databases, TCGA and CPTAC datasets. D. Detection of ARHGAP24 mRNA expression in primary tumoral tissues from patients with recurrent and non-recurrent HCC by qRT-PCR. E. The mRNA expression of ARHGAP24 in paired HCC tissues and adjacent non-tumor tissues. F. Detection of protein expression of ARHGAP24 in tumoral tissues from patients with recurrent and non-recurrent HCC by immunoblotting assays. G. The protein expression of ARHGAP24 in paired HCC tissues (T) and adjacent non-tumor tissues (P). H. Representative immunohistochemistry staining of ARHGAP24 with different intensity. I. Heatmap demonstrating the association between ARHGAP24 expression and clinicopathologic features. Chi-square test. J. Kaplan-Meier analysis of the progression-free survival (PFS, left) and early-relapse (right) in HCC patients stratified by ARHGAP24 expression. K. Kaplan-Meier analysis of the progression-free survival (PFS) in HCC patients with tumor early stage (BCLC: 0 + A, right) and with low AFP levels (≤ 400, left), respectively. L. Multivariate Cox analysis of prognostic factors associated with tumor progression in HCC. [/fig]
[fig] Figure 2: ARHGAP24 inhibits proliferation potentials in HCC. A. mRNA (upper panel) and protein (lower panel) expression of ARHGAP24 in indicated cell lines were detected by qRT-PCR and WB assays, respectively. B. [/fig]
[fig] Figure 3: ARHGAP24 restrains invasiveness and metastasis in HCC. A. Effects of ARHGAP24 on migrative and invasive capacities of HCC cells were measured by Transwell assays. B. Effects of ARHGAP24 on cell migration were evaluated by scratch wound healing assays. C. The mRNA expressions of indicated EMT-related markers were detected by qRT-PCR assays. D. The protein expressions of EMT-related markers were determined by WB assays. E. Expressions of E-Cadherin and N-Cadherin in HCC cell lines after ARHGAP24 expression modulation were evaluated by immunofluorescence staining. F. Cell morphology alterations after ARHGAP24 expression modulations were assessed by FITC-phalloidin staining. G. Representative images of H&E-stained lung tissues (left), and incidence of lung metastasis in indicated groups (right). H. Representative IHC images of ARHGAP24 and indicated markers in tumor tissues derived from orthotopic xenograft models. [/fig]
[fig] Figure 4: ARHGAP24 represses the transcriptional activity of β-catenin. A. Differential genes (DEGs) were identified when ARHGAP24 high group (upper quarter) was compared to ARHGAP24 low group (lower quarter) in TCGA dataset according to follow criteria: p value < 0.05 and |log2 fold change (FC)| > 1. B. Bubble plot revealed significantly enriched Wnt/β-catenin-related signaling according to Reactome pathway analysis. C. GSEA analysis demonstrated highly activated Wnt/β-catenin signaling especially the transcriptional activity were enriched in ARHGAP24-low subgroup. D. The transcriptional activity of β-catenin in indicated cells received distinct treatment was analyzed by TOP/FOP Flash assays. E. The mRNA expressions of downstream target genes of β-catenin were detected by qRT-PCR assay. F. The mRNA (upper panel) and protein (lower panel) expression of downstream target genes of β-catenin were determined in HCCLM3 cells in the presence of ICG-001, a specific antagonist of Wnt/β-catenin pathway. G. Effects of ICG-001 on the long-term proliferation potentials of HCCLM3 cells assessed by colony formation assays. H. Effects of ICG-001 treatment on the migration and invasion capacities of HCCLM3 cells were determined by Transwell and scratch wound healing assays. I. Effects of ICG-001 on the migration capacities of ARHGAP24-knockdown Li-7 cells were assessed by Transwell and scratch wound healing assays. J. mRNA (upper panel) and protein (lower panel) expressions of downstream target of β-catenin in Li-7 cells received indicated treatment were detected by qRT-PCR and WB assays. K. Effects of ICG-001 on the proliferation of ARHGAP24-knockdown Li-7 cells were evaluated using colony formation assays. [/fig]
[fig] Figure 5: ARHGAP24 inhibits cell invasiveness mainly through an RhoGAP activity independent manner. A. The mRNA expression of β-catenin was detected in HCC cells received indicated treatments. B. The distribution of β-catenin in nucleus and cytoplasm was detected nucleocytoplasmic separation experiment followed by WB assays. C. The distribution of intracellular β-catenin detected by immunofluorescence assay. D. The mRNA expressions of ARHGAP24 and β-catenin were detected in ARHGAP24-overexpressed Huh7 cells received indicated treatments. E. The intracellular distribution of β-catenin in Huh7 cells was detected by WB assays. F. The levels of GTP-bound RAC1, CDC42 and RhoA were determined by PKA-GST pull-down followed by WB assays. G. Levels of GTP-bound RAC1 and GTP-bound CDC42 were detected in the ARH-WT overexpressed-cells and ARH-MUT overexpressed cells. H. The subcellular distributions of β-catenin in the indicated cells were determined by WB assays. I. [/fig]
[fig] Figure 6: ARHGAP24 restrains β-catenin signaling via downregulating PKM2 protein abundance. A. Schematic diagram showing IP-MS strategies for identification of the interactor partner protein of ARHGAP24. B. CO-IP analysis revealed strong interaction between ARHGAP24 and PKM2 in HCC cells. C. mRNA expression of ARHGAP24 and PKM2 were detected by qRT-PCR. D. Protein expression of PKM2 after ARHGAP24 knockdown (upper panel) or overexpression (lower panel) were detected by WB assay. E. Effects of ARHGAP24 modulation on PKM2 expression were assessed via using immunofluorescence staining. F and G. The effects of silencing PKM2 on mRNA (F) and protein (G) expressions of the downstream targeted genes of β-catenin in ARHGAP24-low (upper panel) or ARHGAP24 knockdown cells (lower panel). H. The transcriptional activity of β-catenin in the indicated cells was analyzed by TOP/FOP Flash assay. I. The effects of silenced PKM2 on the abilities of cell migration and invasion in Li-7-shARH cells. J. The effects of silenced PKM2 on the abilities of cell migration and invasion in HCCLM3 cells. [/fig]
[fig] Figure 7: ARHGAP24 enhances PKM2 ubiquitination by promoting the interaction of PKM2 and WWP1. A. WB analysis of PKM2 and ARHGAP24 protein in HCCLM3 cells transfected with ARHGAP24 overexpressing plasmid and corresponding control in the presence of CHX. B. WB analysis of PKM2 and ARHGAP24 protein in Li-7 cells transfected with ARHGAP24 knocking down plasmid and corresponding control in the presence of CHX. C. WB analysis of PKM2 and ARHGAP24 protein in Li-7 cells and ARHGAP24 overexpressed-HCCLM3 cells in the presence of 10 µM MG132 or 10 µM NH4Cl for 8 h. D. Ubiquitination assay in HCC cells transfected with the indicated plasmids, followed by WB analysis of indicated proteins. E. Co-IP analysis of the interaction of ARHGAP24, WWP1 and PKM2 in HCLCM3 cells in the presence of MG132. F. IP and WB analyses of the interaction of PKM2 and WWP1 in Li-7 cells transfected with the indicated plasmids. G. IP and WB analyses of the interaction of ARHGAP24 and WWP1 in HEK293T cells transfected with the indicated plasmids. H. IP and WB analyses of the interaction of ARHGAP24 and PKM2 in HEK293T cells transfected with the indicated plasmids. I. The effects of silenced WWP1 on the protein expression of WWP1, ARHGAP24 and PKM2 by WB analysis. J. IP analyses of the PKM2 ubiquitination in ARHGAP24 overexpressed-HCCLM3 cells transfected with the indicated plasmids. K. IP analyses of the interaction of Flag-PKM2 and His-WWP1 in the presence and absence of ARHGAP24. L. IP analyses of the PKM2 ubiquitination in the presence and absence of ARHGAP24. M. The interaction of WWP1 and PKM2 was detected by IP assays in ARHGAP24 knockdown cells or control cells. N. IP analyses of the PKM2 ubiquitination in ARHGAP24 knockdown cells or control cells. [/fig]
[fig] Figure 8: The C-terminal region of ARHGAP24 for the interaction of WWP1 and PKM2. A. Schematic representation of Flag-tagged full-length PKM2 and HA-tagged full-length ARHGAP24, along with their various deletion mutants. B. HEK293T cells were co-transfected with the full-length PKM2 along with the indicated HA-tagged ARHGAP24 constructs. C. HEK293T cells were co-transfected with the full-length ARHGAP24 along with the indicated Flag-tagged PKM2 constructs. The interaction between ARHGAP24 and PKM2 was detected by IP and WB assays. D. Representative pictures of the binding domain of the ARHGAP24 and PKM2 structure. E. Ubiquitination assays of PKM2 in the presence or absence of ARHGAP24 C-terminal domain in HCCLM3 cells. F. IP and WB analyses of the interaction of ARHGAP24, WWP1 and PKM2 in the presence and absence of ARHGAP24 C-terminal domain. G. HEK293T cells were co-transfected with the Flag-tagged full-length PKM2 and His-tagged full-length WWP1, along with the indicated ARHGAP24 mutants. H. The schematic diagram of the mechanisms of ARHGAP24 on HCC progression. I. The structural domains of ARHGAP24 interacted with WWP1 and PKM2. [/fig]
|
Differential innate immune signalling via Ca2+ sensor protein kinases
SupplementaryTable 1| Expression level of CPK genes relative to TUB4 in mesophyll cell protoplasts. The expression level of CPK genes was monitored by qRT-PCR and was assigned the value « 0 » when no PCR product was obtained from mesophyll cell protoplasts. The quality of the primers was checked on genomic DNA or cDNA from other organs(Supplementary Fig. 3). [fig_ref] Table 8: as described below. [/fig_ref] is visualized in and available as a Microsoft Excel file.
## Supplementary
## Microarray quality assessment. for microarrays processed at the partners
HealthCare Center for Personalized Genetic Medicine Microarray Facility, the array data quality was initially assessed in R using the BioConductor package arrayQualityMetrics [bib_ref] arrayQualityMetrics--a bioconductor package for quality assessment of microarray data, Kauffmann [/bib_ref]. In addition, the BioConductor package Harshlight 13 was used to provide assessment of possible large scale artefacts, and R code for SmudgeMiner 14 was used for assessment of possible spatial bias. Affymetrix ATH1 GeneChip data were generated from duplicate or triplicate biological samples.
## Comparison of various data processing methods. to define cpk5ac and cpk11ac
target genes with high stringency and statistical significance, we processed and compared the ATH1 GeneChip data using various microarray analysis and statistic algorithms [bib_ref] False discovery rate paradigms for statistical analyses of microarray gene expression data, Cheng [/bib_ref] [bib_ref] Comparison of small n statistical tests of differential expression applied to microarrays, Murie [/bib_ref] [bib_ref] Significance analysis of microarrays applied to the ionizing radiation response, Tusher [/bib_ref] , and a Bayesian version of the t-test, cyber-T 24 . Cyber-T is a version of the t-test that uses a Bayesian estimate of the treatment variance (window size of 101 and confidence ratio of 10 was used). In SAM, the statistic test computed for each gene is slightly different from the t-test in that it includes a factor in the denominator selected to minimize the coefficient of variation across arrays, and is also intended to guard against unrealistically small estimates of variability occasionally observed due to low sample size. Also, SAM does not use the hypothetical t-distribution to assess significance of observed effects, but permutations of the data set generate an empirical null distribution. Significance is evaluated using this alternative null as a reference (100 permutations and a random number seed of 0 was used).
Affymetrix probe level data for control, CPK5ac and CPK11ac were also analyzed using Gene Chip Operating Software (GCOS)Regulated genes were then determined by removing genes consistent with the following criteria: a change call of 'no change'; a signal log 2 ratio between -0.99 and 0.99 (less than two-fold change in expression); a detection call of absent in both control and treatment arrays. Because three separate algorithms are used in GCOS to determine detection, change, and signal value, there can be occasional discrepancies between the results of these algorithms. Genes showing these discrepancies were eliminated under the two additional criteria: a change call of 'increase' or 'marginal increase' combined with either a signal log 2 ratio less than 0, an absent call in the CPK5ac or CPK11ac array, or a signal value in the control RNA array that was greater than the signal value in the CPK5ac or CPK11ac array; a change call of 'decrease' or 'marginal decrease' combined with either a signal log 2 ratio greater than 0; an absent call in the control RNA array, or a signal value in the control that was less than the signal value in the CPK5ac or CPK11ac array.
After comparing the results from the extensive analyses of control, CPK5ac
and CPK11ac arrays, it was found that RMA [bib_ref] Summaries of Affymetrix GeneChip probe level data, Irizarry [/bib_ref] [bib_ref] Exploration, normalization, and summaries of high density oligonucleotide array probe level data, Irizarry [/bib_ref] with cyber-T 24 and a p-value cutoff of 0.05 was most in agreement with our qRT-PCR data for the validation of eight marker genes (n=6) in triplicate biological samples . The combination of RMA 19,20 followed by cyber-T 24 (a p-value cutoff of 0.05) was chosen to analyze arrays for 30 min and 60 min flg22 treatment in protoplasts and seedlings as well as arrays for MAMP signals and microbial infection downloaded from the public domain.
Identification of CPK5ac-specific, CPK11ac-specific, and common CPK5ac and CPK11ac target genes. Arrays for CPK5ac and CPK11ac along with respective control arrays were analyzed in FlexArray 18 using RMA 19,20 followed by cyber-T 24 as well as dChip 21 followed by cyber-T 24 . Candidate target genes were chosen from each of the two analysis based on a p-value of 0.05 or less .
Significant target genes separately showed a p-value of 0.05 and log 2 ratio >1 or <-1 in both RMA and dChip processing (Supplementary . The p-values of majority (93%) of these target genes were below 0.03. The p-value of 0.05 was accepted to cover the validated marker gene CYP82C2 , which had relatively low expression level, thus high variance. They were categorized as CPK5ac-specific (72 genes), CPK11ac-specific (59 genes), or common CPK5ac and CPK11ac (113 genes) target genes (Supplementary After processing, all data were imported into Excel. VBA scripts were then used to extract CDPK target genes from the protoplast, leaf, and seedling data for early flg22 responsive genes. The resulting list of candidate flg22-CDPK early target genes (Supplementary was then selected manually by eliminating genes that met the following criteria: genes not regulated in any of the flg22 conditions; genes with a CDPK signal log 2 ratio of between -1 and 1; genes that were reverse regulated between CDPK and flg22. Three gene lists resulted; common CPK5ac and CPK11ac target genes that were also early flg22 responsive genes (81 genes) Supplementary ; CPK5ac-specific genes that were also early flg22 responsive genes (59 genes) , CPK11ac-specific genes that were also early flg22 responsive genes (31 genes) Early MAMP responsive genes were selected by the criteria of p-value of 0.01 or less.
After processing, all data were imported into Excel. VBA scripts were then used to extract genes from the RMA-cyber-T processed MAMP data using the list of 171 and processed with RMA 19,20 followed by cyber-T [bib_ref] A Bayesian framework for the analysis of microarray expression data: regularized t..., Baldi [/bib_ref]. Regulated genes were selected by the criteria of p-value of 0.01 or less. After processing, data were imported into Excel.
VBA scripts were then used to extract genes from the RMA-cyber-T processed microbial infection data using the list of 171 flg22-CDPK early target genes. The resulting genes lists were presented using Hierarchical Clustering analysis in Cluster 3.0 [bib_ref] Open source clustering software, De Hoon [/bib_ref] and Java TreeView 1.1.3 [bib_ref] Java Treeview--extensible visualization of microarray data, Saldanha [/bib_ref] ,
[formula] Supplementary [/formula]
## Clustering of data
Gene lists with signal log ratio values, log 2 (fold-change), were imported into Cluster 3.0 27 for Hierarchical Clustering analysis using binary, agglomerative, hierarchical centroid linkage clustering with a Pearson correlation similarity metric for genes.
More specifically, Correlation (uncentered), a modified Pearson correlation in which the mean is assumed to be 0, was the similarity metric chosen. The resulting clustering was viewed with Java TreeView 1.1.3 28 .
## Ath1 genechip data from public domains
All the original data for 18 ATH1 GeneChips generated for this project were submitted to Gene Expression Omnibus (GEO) microarray database at National
Centre for Biotechnology Information (NCBI) for public access.
## Flg22
Leaf data for flg22 (1 µM, 60 min) were obtained from the AtGenExpress project:
B20-ATGEN, B34-ATGEN, and B6-ATGEN (F. Brunner and T. Nürnberger).
## Elf26
Seedling data for elf26 (1 µM, 60 min) obtained in A.
## Npp1
Leaf data for NPP1 (4 h) were downloaded from AtGenExpress at TAIR website,
[fig] plants Col- 0: The constructs for inactive kinase mutant (KM) forms of CDPK were generated using site-directed mutagenesis to replace the conserved lysine residue of the ATP-binding domain with a methionine. The effector constructs FLS2-HA, MKK4a-MYC, MPK3-HA and MPK6-HA have been previously described 7 . For VIGS, the CPK4 construct was generated using a gene specific 500-In-gel protein kinase assays. For in-gel kinase assay of endogenous MAPKs, 10day-old seedlings were ground in liquid nitrogen and homogenized in 150 µl extraction buffer (50 mM HEPES-KOH pH 7.5, 5 mM EDTA, 5 mM EGTA, 2 mM DTT, 25 mM b-glycerophosphate, 2 mM Na 3 VO 4 , 10 mM NaF, 1 mM PMSF, 5 µg/ml antipain, 5 µg/ml leupeptin). The protein extract was recovered after centrifugation at 17600 g for 15 min at 4°C. Equal amounts of proteins were loaded on 10% SDS-polyacrylamide gel embedded with 0.25 mg/ml myelin basic protein (MBP, Invitrogen). The gel was washed 3 times with washing buffer (25 mM Tris-HCl pH=7.5, 0.5 mM DTT, 5 mM NaF, 0.1 mM Na 3 VO 4 , 0.5 mg/ml BSA, 0.1% Triton X-100), then incubated for 16 h with 3 changes in renaturation buffer (25 mM Tris-HCl pH 7.5, 0.5 mM DTT, 5 mM NaF, 0.1 mM Na 3 VO 4 ) at 4°C. After a 30 min preincubation at room temperature in the reaction buffer (25 mM Tris-HCl pH 7.5, 1 mM DTT, 0.1 mM Na 3 VO 4 , 12 mM MgCl 2 , 1 mM EGTA), the kinase reaction was performed for 1 h in the reaction buffer supplemented with 25 µM cold ATP and 50µCi [g-32 P] or [g-33 P] ATP. The reaction was stopped by extensive washes in 5% TCA and 1% sodium pyrophosphate for 6 h. The protein kinase activity was detected on dried gel by the Typhoon imaging system (GE Healthcare). [/fig]
[fig] For: in-gel kinase assay of endogenous CDPKs, protoplasts were directly resuspended in SDS sample buffer. Proteins were loaded on 10% SDS-polyacrylamide gel embedded with 0.25 mg/ml histone III-S (Sigma). The gel was treated for washing, renaturating and reaction steps as described above, with the following modifications : the reaction buffer contains 25 mM Tris-HCl pH 7.5, 1 mM DTT, 0.1 mM Na 3 VO 4 , 12 mM MgCl 2 , 1 mM CaCl 2 and no cold ATP was added during the kinase reaction.In vitro protein kinase assays. in vitro kinase assay, protoplasts were lysed in 200 µl immunoprecipitation (IP) buffer (50 mM Tris-HCl pH 7.5, 150 mM NaCl, 5 mM EDTA, 1 mM DTT, 2 mM NaF, 2 mM Na 3 VO 4 , 1X cocktail inhibitor [Roche], 1% Triton X-100). To immunoprecipitate CPK-FLAG, protein extracts were incubated with anti-FLAG antibody at 4°C for 2 h, and additional 1 h after adding 15 µl protein G sepharose beads (GE Healthcare). The immunoprecipitated kinases were washed twice with IP buffer and once with kinase buffer (20 mM Tris-HCl pH=7.30 min in 15 µl kinase buffer containing 1 µg histone (Sigma), 50 µM cold ATP and 2 µCi [g-32 P] or [g-33 P] ATP. The reaction was stopped by adding SDS-PAGE loading buffer. After separation on 10% SDS-PAGE, radiolabelled histone was detected on dried gel by the Typhoon imaging system (GE Healthcare). MAPK activity, immunoprecipitation was carried out with anti-HA antibody, and the kinase reaction was performed with 5 mM EGTA and 1 µg MBP. Bacterial growth assays. The bacterial growth assays were performed as previously described in leaf 8 and seedlings 11 with some modifications. leaf assay, an overnight culture of Pseudomonas syringae pv tomato DC3000 was washed and resuspended to 5x10 4 cfu/ml in H 2 O. Arabidopsis leaves of 4-week-old plants were infiltrated with bacteria using a needleless syringe. Two and four days after inoculation, leaf disks were ground in 100 µl H 2 O and serial dilutions were plated on KB medium. Bacterial cfu were counted 2 days after incubation at 28°C. seedling assay, seedlings were grown in 6-well plates containing 1 ml of liquid medium (0.5 x MS, 0.5% sucrose) for 5 days under continuous light at 22°C. Seedlings were treated with 100 nM flg22 for 1 day prior to inoculation with Pst DC3000 at a final concentration of 1x10 7 cfu/ml. After 3 days of co-cultivation, seedlings were surface sterilized with 70% ethanol and rinsed with H 2 O before grinding tissues for bacterial counting as above. The experiments were repeated three times with similar results. [/fig]
[table] Table 9 |: Sequences of oligonucleotide primers for CPKac SupplementaryTable 12 | Sequences of oligonucleotide primers for qRT-PCR. Supplementary Table 13 | Sequences of oligonucleotide primers for semiquantitative RT-PCR. F: forward; R: reverse. Supplementary Table 14 | Sequences of oligonucleotide primers for genotyping cpk mutants. LP: left primer; RP: right primer; LB: left border primer; RB: right border primer.DNA constructs. All primers used for cloning and mutagenesis are listed inSupplementary Tables 9-11and PCR products were checked by sequencing. Effector constructs of full-length (FL) or constitutively active CDPKs (CPKac) were generated by inserting the cDNA (including the first ATG) in a plant expression vector containing a FLAG epitope tag or GFP at the C-terminus, between a 35S-derived promoter and NOS terminator 4,7-10 . The sequence between the promoter and the first ATG of each CDPK contains 8 nucleotides, including the BamHI restriction site for cloning and no additional ATG, which supports the translation to start at the first ATG of CDPKs. Sequencing validated that no additional amino acids were added to the N-terminus of CDPKs by the vector. The cDNA was amplified from Arabidopsis [/table]
[table] Table 5: The resulting gene lists were presented using Hierarchical Clustering analysis in Cluster 3.0 27 and Java TreeView 1.1.3 28 (Fig. 2a, Supplementary Table 6) as described below. of flg22-CDPK early target genes with early MAMP responsive genes. Duplicate or triplicate microarray data from various MAMP treatments (elf26 29 , chitin 30 , harpin 31 , NPP1 32,33 and LPS 34 ) for 30 min, 60 min or 4 h in seedlings or leaves were downloaded from various public domain sources while data for PGN 35treatment was kindly provided by A. Gust and T. Nürnberger (see below). ABA early (60 min) genes in seedlings were used as a microarray control. Affymetrix .CEL files were imported into FlexArray 18 and processed with RMA 19,20 followed by cyber-T 24 . [/table]
[table] Table 8: as described below. [/table]
|
Anaphylactoid hypersensitivity reaction from intra-arterial cetuximab: Clinical considerations and management
Intra-arterial infusion of drugs shows promising results in terms of safety and efficacy. Intra-arterial cetuximab, a monoclonal antibody treatment, is currently being tested for its use in head and neck cancers. We present the case of a 45-year-old Asian male who developed an anaphylactoid hypersensitivity reaction, manifesting itself in the form of bronchospasm, tachycardia, and hypotension, during intra-arterial infusion of cetuximab. The symptoms were quickly diagnosed, and the patient was treated accordingly. Despite the safety profile of cetuximab and the decreased risk of systemic effects with intra-arterial infusion versus intravenous infusion, severe hypersensitivity reactions are still a risk in intra-arterial cetuximab infusions. Consequently, proper planning and care must be taken to prophylactically prevent and in the case of a reaction, treat the reaction accordingly. The case presented herein is, to the best of our knowledge, the first recorded moderate-to-severe infusion reaction in a patient receiving intra-arterial cetuximab treatment for head and neck cancer.
# Introduction
Intra-arterial (IA) infusion is a unique modality of drug administration. Clinically, the main utility of this technique is seen in chemotherapy administration, in which IA infusion of the chemotherapy agent allows for high doses of drugs to be targeted directly to the cancerous tissue, while reducing exposure of healthy tissues to the effect of the drugs. In the literature, IA infusions are noted for having lower risk of systemic effects. [bib_ref] Regional intra-arterial vs. systemic chemotherapy for advanced pancreatic cancer: a systematic review..., Liu [/bib_ref] [bib_ref] A critical evaluation of the principles governing the advantages of intra-arterial infusions, Eckman [/bib_ref] The current explanation for this is based on two working assumptions, one direct and one indirect: the direct effect is by targeting the cancer tissue, reducing healthy tissue exposure to chemotherapy agents, and the indirect effect of the reduction in overall dosing needed to treat, whereby the patient is exposed to an appreciable reduction in the overall amount of chemotherapy agent throughout treatment.
Cetuximab is a monoclonal antibody agent that has been established as a first-line treatment for recurrent head and neck cancer. [bib_ref] Intra-arterial cetuximab for the treatment of recurrent unresectable head and neck squamous..., Tham [/bib_ref] Monoclonal antibodies, such as cetuximab, have proven to be quite effective chemotherapeutic agents but they have also been associated with hypersensitivity infusion reactions (IR). [bib_ref] Radiotherapy plus cetuximab for squamous-cell carcinoma of the head and neck, Bonner [/bib_ref] With respect to intravenous infusions of cetuximab, mild to moderate reactions (grades 1 and 2 IRs), including flushing, rash, fever, chills, dyspnea, and mild hypotension, occur in 16%-19% of patients after the first infusion. [bib_ref] Management and preparedness for infusion and hypersensitivity reactions, Lenz [/bib_ref] Severe reactions (grades 3 and 4 IRs), such as bronchospasms and anaphylaxis, are much less common and occur roughly 3% of the time. Despite previous reports to the safety and the decreased risk of systemic side effects associated with IA, we present the first reported case of a patient who developed a hypersensitivity reaction to IA cetuximab.
## Case
A 45-year-old Asian male with a history of nasopharyngeal carcinoma (NPC) was referred to our clinic in the spring of 2016 after he was found to have NPC recurrence. His past medical history was insignificant except for a four pack-year smoking history . He had no personal or family history of allergies, asthma, chronic obstructive pulmonary disease (COPD), or atopy. The patient was diagnosed via fine needle aspiration of a cervical lymph node in 2013 and underwent definitive chemoradiation therapy (CRT) from late 2013 to early 2014 (cisplatin and conventional radiation therapy (XRT) with a total dose of 56 Gy over 28 fractions). However, 9 months after completion of CRT, a positron emission tomography (PET) scan was suspicious for recurrence. Nasal biopsy was performed, but was found to be negative. Roughly 19 months after completion of his CRT, the patient's surveillance PET scan was again suggestive of recurrence. In January 2016, the patient underwent an endoscopic left maxillary antrostomy, left total ethmoidectomy, and sphenoidotomy with biopsies at another institution. Pathology confirmed recurrence of poorly differentiated NPC. At that time, the patient's disease was determined to be endoscopically unresectable so he was referred to our institution for consideration for open resection. Upon referral, a magnetic resonance imaging (MRI) was obtained, which demonstrated extensive left skull base recurrence of the primary NPC with cavernous sinus, Meckel's cave, and perineural involvement along V3 to the left mandible [fig_ref] Figure 1: Coronal MRI. [/fig_ref]. There was also involvement of the orbital apex and temporal lobe dura. Symptomatically, the patient reported numbness along the left side of his face and decreased taste along the left hemi-tongue. He denied any difficulty swallowing, vision loss, hearing loss, or weakness. After presenting the patient at our tumor board, it was decided that the patient's volume of disease was unresectable, but his excellent performance status and minimal symptoms made him an appropriate candidate for our phase I IA cetuximab clinical trial with concurrent re-irradiation.In April 2016, the patient was enrolled and underwent the first dose of his two scheduled IA cetuximab doses of 100 mg/ m 2 in a Phase I dose-escalation clinical trial (NCT02438995). The morning of the infusion the patient's vitals were stable (T: 99, heart rate (HR): 78 bpm, blood pressure (BP): 119/65, respiratory rate (RR): 14, SaO 2 : 98%) and consistent with patient's baseline vitals. As per the study protocol, the patient was prophylactically treated with 50 mg of diphenhydramine 1 hour prior to the start of the IA cetuximab infusion. After groin puncture and demonstrating a safe intravascular positioning of the microcatheter for IA infusion, a total of 81.5 mL of cetuximab was infused over 12 min [fig_ref] Figure 2: Selected angiographic images [/fig_ref] and b). After infusion of the first 50 mL, the patient developed a cough with concurrent elevation of his HR up to 110 bpm. He was treated with 100 mg intravenous hydrocortisone sodium succinate and his symptoms resolved, allowing for the remainder of the infusion to be given to completion without adverse effects. The patient did not experience any severe vascular pain or vascular spasms as a result of the procedure. The catheter and femoral sheath were removed and hemostasis was achieved with a perclose device. Subsequently, while still in the cath lab, the patient's HR and RR suddenly increased to 110 bpm and over 20, respectively. His BP dropped to 50/30, his oxygen saturation dropped to 90, and he experienced bronchospasms. The anesthesiologist treated the patient with 20 mg of epinephrine, 180 mcg of phenylephrine, and 4 mg of Zofran, along with positive ventilation and a total of 900cc of Lactated Ringers. After intervention, the patient's vitals stabilized (HR: 91 bpm, BP: 111/66, RR: 16, SaO 2 : 99), and he was transferred to the holding area for observation and recovery. The patient never demonstrated signs of flushing, rash, fever, chills, and dyspnea and his vitals remained continuously stable for the following 24 h of observation. The patient was then discharged home and reported feeling well with no complaints during a follow-up phone call 3 days post infusion.
Per protocol, the patient was removed from the IA cetuximab trial due to severe IR, so he never received the second scheduled dose. However, he proceeded to receive intensitymodulated radiation therapy (IMRT) to his left nasopharynx, skull base, cavernous sinus, and Meckel's cave. He received a total of 7000 cGy over 35 fractions with some minor overlapping of his prior radiation site; 16 months out from completion of his one dose of cetuximab and re-irradiation, the patient's surveillance imaging still shows no evidence of disease.
# Discussion
Due to IA cetuximab's purported specificity and safety, it is viewed as a potential preferred method of administration for the treatment of head and neck cancer. IA infusion has shown to be associated with fewer complications and side effects compared to intravenous (IV) infusion. [bib_ref] Regional intra-arterial vs. systemic chemotherapy for advanced pancreatic cancer: a systematic review..., Liu [/bib_ref] However, based on the case we have reported, IA cetuximab still maintains a risk of inducing adverse reactions, albeit at presumably lower rates than IV infusion. The patient experienced an anaphylactoid hypersensitivity reaction, despite prophylactically being treated with diphenhydramine, an H 1 antagonist recommended for the purpose of preventing IRs with cetuximab. During the infusion, the patient experienced a cough and elevated HR, requiring hydrocortisone sodium succinate. Furthermore, within 30 min after the infusion, the patient displayed unstable vitals, including decreased BP, increased RR and HR, decreased SaO 2 , and the development of bronchospasms. The patient was immediately treated for the IR accordingly and his vitals normalized. Epinephrine, a sympathomimetic catecholamine often used in IRs, was used to stabilize the patient. Epinephrine functions as an α and β-adrenergic agonist, reversing peripheral vasodilation, alleviating hypotension, reducing erythema and angioedema, and resulting in bronchodilation and increased myocardial output and contractility. [bib_ref] Epinephrine: the drug of choice for anaphylaxis-a statement of the world allergy..., Kemp [/bib_ref] Although the exact reason for why our patient experienced an IR is unknown, there are several theories as for why adverse reactions may occur. Monoclonal antibodies, such as cetuximab, have been postulated to elicit a response from human anti-chimeric antibodies (HACAs) and human antihuman antibodies (HAHAs). [bib_ref] Anti-EGFR mechanism of action: antitumor effect and underlying cause of adverse events, Lenz [/bib_ref] Other theories include IgEmediated responses as leading contributors to IRs, and in particular, anaphylaxis. 9,10 Geographic differences result in different natural exposures to galactose-α-1,3-galactose which has been implicated in the production of IgE antibodies. Cetuximab contains galactose-α-1,3-galactose, or alphagal and therefore puts those with increased IgE production at risk for IRs. Although information on our patient's IgE sensitization to alpha-gal was unavailable, we suggest using an enzyme-linked immunosorbent assay (ELISA) assay to screen for anti-cetuximab IgE antibodies prior to infusion, as this appears to be a highly specific and sensitive method of determining whether a severe IR may occur. [bib_ref] Cetuximab-induced anaphylaxis and IgE specific for galactose-alpha-1,3-galactose, Chung [/bib_ref] [bib_ref] Management of infusion reactions associated with cetuximab treatment: a case report, Ohshita [/bib_ref] [bib_ref] Case report about fatal or near-fatal hypersensitivity reactions to cetuximab: anticetuximab IgE..., Dupont [/bib_ref] In order to properly monitor for such reactions, vitals must be carefully recorded before, during, and after infusion, and although IRs normally occur within the first hour after infusion, delayed reactions may still occur. [bib_ref] Management and preparedness for infusion and hypersensitivity reactions, Lenz [/bib_ref] [bib_ref] Severe infusion reactions to cetuximab occur within 1 h in patients with..., Yamaguchi [/bib_ref] Among patients who experience IRs, 90% experience adverse reactions after the first infusion. A "crash cart" with epinephrine, ephedrine, aerosolized bronchodilator, and other equipment, such as an oxygen tank, should be available as a preventive measure. Even with pre-medication, up to 19% of patients treated with cetuximab report an IR. Hence, diphenhydramine may not be enough to prevent reactions. [bib_ref] Management and preparedness for infusion and hypersensitivity reactions, Lenz [/bib_ref] Literature suggests that the administration of albuterol, famotidine, and corticosteroids, along with diphenhydramine, may significantly decrease the risk of severe reactions and should be considered in future prevention of IRs. [bib_ref] Risk factors for and pre-medications to prevent cetuximab-induced infusion reactions in patients..., Touma [/bib_ref] There is also literature that suggests that patients who have developed a mild-moderate IR due to cetuximab can be successfully re-challenged, but with a reduced infusion rate. [bib_ref] Management and preparedness for infusion and hypersensitivity reactions, Lenz [/bib_ref] [bib_ref] Cetuximab-induced anaphylaxis and IgE specific for galactose-alpha-1,3-galactose, Chung [/bib_ref] However, re-challenging is not suggested in patients with severe IR. Due to the protocol of the study, our patient was removed from the clinical trial due to the allergic reaction and was not re-challenged. 6
# Conclusion
To our knowledge, this is the first report of an IR arising from IA cetuximab for head and neck cancer. It is important to note that even with evidence of increased safety using IA infusion, our patient experienced a severe IR despite diphenhydramine prophylaxis. Prompt intervention and cognizance of this potentially life-threatening complication is vital during the administration of IA cetuximab.
[fig] Figure 1: Coronal MRI. [/fig]
[fig] Figure 2: Selected angiographic images: (a) Digital subtraction angiography (DSA) examination in lateral view of the left internal maxillary artery at the site of SSIA cetuximab injection and (b) DSA examination in anterior-posterior (AP) view of the left internal maxillary artery at the site of the SSIA cetuximab. [/fig]
|
Positively Selected Effector Genes and Their Contribution to Virulence in the Smut Fungus Sporisorium reilianum
Plants and fungi display a broad range of interactions in natural and agricultural ecosystems ranging from symbiosis to parasitism. These ecological interactions result in coevolution between genes belonging to different partners. A wellunderstood example is secreted fungal effector proteins and their host targets, which play an important role in pathogenic interactions. Biotrophic smut fungi (Basidiomycota) are well-suited to investigate the evolution of plant pathogens, because several reference genomes and genetic tools are available for these species. Here, we used the genomes of Sporisorium reilianum f. sp. zeae and S. reilianum f. sp. reilianum, two closely related formae speciales infecting maize and sorghum, respectively, together with the genomes of Ustilago hordei, Ustilago maydis, and Sporisorium scitamineum to identify and characterize genes displaying signatures of positive selection. We identified 154 gene families having undergone positive selection during species divergence in at least one lineage, among which 77% were identified in the two investigated formae speciales of S. reilianum. Remarkably, only 29% of positively selected genes encode predicted secreted proteins. We assessed the contribution to virulence of nine of these candidate effector genes in S. reilianum f. sp. zeae by deleting individual genes, including a homologue of the effector gene pit2 previously characterized in U. maydis. Only the pit2 deletion mutant was found to be strongly reduced in virulence. Additional experiments are required to understand the molecular mechanisms underlying the selection forces acting on the other candidate effector genes, as well as the large fraction of positively selected genes encoding predicted cytoplasmic proteins.
# Introduction
Plants and fungi have a long history of coevolution since the emergence of pioneering land plants $400 Ma. The development of early plants was likely supported by associations with symbiotic fungi, as suggested by analyses of ribosomal RNAs and fossil records [bib_ref] Four hundred-million-year-old vesicular arbuscular mycorrhizae, Remy [/bib_ref] [bib_ref] Geosiphon pyriforme, a fungus forming endocytobiosis with Nostoc (cyanobacteria), is an ancestral..., Gehrig [/bib_ref] [bib_ref] Ancestral alliances: plant mutualistic symbioses with fungi and bacteria, Martin [/bib_ref]. Different forms of plant-fungus interactions have evolved, including mutualistic symbiosis where both plant and fungus benefit [bib_ref] Arbuscular mycorrhiza: the mother of plant root endosymbioses, Parniske [/bib_ref] , and pathogenic interactions where fungal colonization greatly reduces plant fitness [bib_ref] The top 10 fungal pathogens in molecular plant pathology, Dean [/bib_ref]. Pathogenic interactions play critical roles in natural and agricultural ecosystems, and understanding the evolutionary mechanisms shaping them is of great importance to plant production, food security, and protection of biodiversity in natural ecosystems [bib_ref] Emerging fungal threats to animal, plant and ecosystem health, Fisher [/bib_ref].
Secreted fungal effector proteins are key players in pathogenic interactions as they are involved in protecting and shielding growing hyphae, suppressing plant defense responses and changing plant physiology to support growth of the pathogen [bib_ref] Fungal effector proteins, Stergiopoulos [/bib_ref] [bib_ref] How filamentous pathogens co-opt plants: the ins and outs of fungal effectors, De Jonge [/bib_ref] [bib_ref] Filamentous plant pathogen effectors in action, Giraldo [/bib_ref]. Many effector proteins lack known functional domains, and expression of a subset of effectors is linked to plant colonization [bib_ref] Fungal effectors and plant susceptibility, Presti [/bib_ref] [bib_ref] Plant-pathogen effectors: cellular probes interfering with plant defenses in spatial and temporal..., Toruño [/bib_ref] [bib_ref] Effectors of filamentous plant pathogens: commonalities amid diversity, Franceschetti [/bib_ref] [bib_ref] Ustilago maydis effectors and their impact on virulence, Lanver [/bib_ref]. Effector proteins with a strong effect on virulence phenotype are thought to coevolve with their plant targets either in an arms race or a trench-warfare scenario [bib_ref] Plant-parasite coevolution: bridging the gap between genetics and ecology, Brown [/bib_ref]. In the former, fungal effectors manipulating the host are under positive directional selection, and plant targets evolve in response to changes in effector proteins [bib_ref] Filamentous pathogen effector functions: of pathogens, hosts and microbiomes, Rovenich [/bib_ref]. In the latter scenario, sets of alleles are maintained by balancing selection in both host and pathogen populations [bib_ref] Plant-parasite coevolution: bridging the gap between genetics and ecology, Brown [/bib_ref] [bib_ref] Speed of adaptation and genomic footprints of host-parasite coevolution under arms race..., Tellier [/bib_ref]. Several methods are available for identifying genomic regions under selection [bib_ref] Molecular signatures of natural selection, Nielsen [/bib_ref]. It has been proposed that genes with signatures of positive selection have important functions during host pathogen interaction or have contributed to host specialization [bib_ref] Molecular evolution of plant immune system genes, Tiffin [/bib_ref]. It is therefore expected that the deletion of such genes reduces virulence when tested on a susceptible host.
Depending on the aim of the investigation, studies identifying genes with signatures of positive selection are carried out within or between species. Although studies on the population level focus on recent and ongoing selective processes and are instrumental in the understanding of adaptation, comparative genomic studies employing different species encompass a broader time span and provide insight into the underlying genetic basis of host specialization [bib_ref] Using population and comparative genomics to understand the genetic basis of effector-driven..., Plissonneau [/bib_ref]. The signature of positive selection in such case typically takes the form of an excess of divergence between species due to increased fixation of mutations by selective sweeps compared with a neutral expectation . This is commonly measured by the ratio of nonsynonymous (d N ) over synonymous (d S ) divergence and a d N /d S ratio > 1 is taken as evidence for positive selection under the assumption that synonymous substitutions are neutral while nonsynonymous are not. Positive selection studies in a number of plant pathogen systems revealed that genes encoding secreted effector proteins are enriched in signatures of positive selection (Mö ller [bib_ref] Evolution and genome architecture in fungal plant pathogens, Mö Ller [/bib_ref]. Such studies include investigations in diverse plant pathogens like Microbotryum species causing anther-smut disease of Caryophyllaceae species , the wheat pathogen Zymoseptoria tritici, the rust fungus Melampsora larici-populina [bib_ref] A Comprehensive Analysis of Genes Encoding Small Secreted Proteins Identifies Candidate Effectors..., Hacquard [/bib_ref] , the rice BLAST fungus Magnaporthe oryzae [bib_ref] Rapid evolution of avirulence genes in rice blast fungus Magnaporthe oryzae, Huang [/bib_ref] , the wheat stem rust fungus Puccinia graminis f. sp. tritici [bib_ref] Diversifying selection in the wheat stem rust fungus acts predominantly on pathogen-associated..., Sperschneider [/bib_ref] , the Irish potato famine pathogen Phytophthora infestans [bib_ref] Effector specialization in a lineage of the Irish potato famine pathogen, Dong [/bib_ref] , a group of Fusarium species [bib_ref] Genome-wide analysis in three Fusarium pathogens identifies rapidly evolving chromosomes and genes..., Sperschneider [/bib_ref] , and a group of smut fungi parasitizing different grasses and a dicot host [bib_ref] Gene loss rather than gene gain is associated with a host jump..., Sharma [/bib_ref] [bib_ref] Comparative genomics including the early-diverging smut fungus Ceraceosorus bombacis reveals signatures of..., Sharma [/bib_ref]. Yet, as only a few genes under positive selection have been functionally studied, the link between the selected genotypes and their corresponding phenotypes are only beginning to be understood and only a few studies used evolutionary predictions to unravel the molecular mechanisms of host adaptation. For example, the population genomics study in the wheat pathogen Z. tritici which identified candidate effector genes under positive selectionwas followed up experimentally, and in this case it was shown that the deletion of three of four candidate genes reduced virulence. In the gray mold fungus Botrytis cinerea, four positively selected genes were deleted without affecting virulence, and this finding was attributed to functional redundancy, the limited number of tested host plants, or experimental conditions different from natural infections (Aguileta 2012). A study of the oomycete effector protein EpiC1 showed that a single amino acid substitution at a site under positive selection affected the binding affinity of different host proteases determining host specificity [bib_ref] Effector specialization in a lineage of the Irish potato famine pathogen, Dong [/bib_ref].
Smut fungi, belonging to the division of Basidiomycota, are a group of about 550 species parasitizing mostly grasses, including important crops like maize, sorghum, oat, barley, and sugarcane . In smut fungi, sexual reproduction is linked to pathogenic development and smut fungi therefore depend on successful plant colonization to complete their life cycle. As biotrophic pathogens, they require living plant tissue for establishing a successful interaction [bib_ref] The Ustilaginales as plant pests and model systems, Martinez-Espinoza [/bib_ref]. With few exceptions like Ustilago maydis, smut fungi usually develop symptoms only in the female or male inflorescence of their respective host plants. During the last 10 years, quality draft genome sequences of prominent species were obtained, including U. maydis, the causative agent of smut disease on maize and teosinte , Sporisorium reilianum causing head smut of maize and sorghum [bib_ref] Pathogenicity determinants in smut fungi revealed by genome comparison, Schirawski [/bib_ref] , Ustilago hordei infecting barley [bib_ref] Genome comparison of barley and maize smut fungi reveals targeted loss of..., Laurie [/bib_ref] , and Sporisorium scitamineum parasitizing sugarcane [bib_ref] Genome sequencing of Sporisorium scitamineum provides insights into the pathogenic mechanisms of..., Que [/bib_ref] [bib_ref] Complete genome sequence of Sporisorium scitamineum and biotrophic interaction transcriptome with sugarcane, Taniguti [/bib_ref] [bib_ref] A tale of genome compartmentalization: the evolution of virulence clusters in smut..., Dutheil [/bib_ref]. The head smut fungus S. reilianum occurs in two formae speciales that infect maize (S. reilianum f. sp. zeae) or sorghum (S. reilianum f. sp. reilianum) [bib_ref] Host specificity of Sporisorium reilianum is tightly linked to generation of the..., Zuther [/bib_ref]. The concept of formae speciales is used in phytopathology to distinguish members of the same species based on their ability to colonize a certain host plant (in this example maize or sorghum) . The divergence of U. hordei, U. maydis, S. scitamineum, and S. reilianum was inferred to have occurred in the interval of seven and 50 Ma [bib_ref] Domestication of maize, sorghum, and sugarcane did not drive the divergence of..., Munkacsi [/bib_ref]. The availability of genome sequences of several species with different host ranges, together with established tools for genetic manipulations [bib_ref] A reverse genetic approach for generating gene replacement mutants in Ustilago maydis, Brachmann [/bib_ref] [bib_ref] A PCR-based system for highly efficient generation of gene replacement mutants in..., K€ [/bib_ref] [bib_ref] The use of FLP-mediated recombination for the functional analysis of an effector..., Khrunyk [/bib_ref] [bib_ref] Genome editing in Ustilago maydis using the CRISPR-Cas system, Schuster [/bib_ref] make this group of smut fungi particularly interesting to study the evolution of effector genes as well as their contributions to virulence, speciation, and host specificity.
Here, we employed the genome of the recently sequenced strain S. reilianum f. sp. reilianum SRS1_H2-8 (http://www.ebi. ac.uk/ena/data/view/LT795054-LT795076) [bib_ref] Host specificity of Sporisorium reilianum is tightly linked to generation of the..., Zuther [/bib_ref] together with the genomes of U. maydis, U. hordei, S. scitamineum, and S. reilianum f. sp. zeae to identify potential effector genes with signatures of positive selection. Candidate genes were individually deleted in S. reilianum f. sp. zeae and the phenotype of the deletion strains was assessed after infection of maize in order to understand their function with respect to virulence. We report that the deletion of one candidate gene, pit2, led to a strong reduction in virulence and we further discuss hypotheses on the origin of positive selection for the other candidate genes.
# Materials and methods
## Construction of homologous protein families
Fungal species used in this study, their number of gene models, number of predicted secreted proteins, and sources of genome data are listed in supplementary table S1, Supplementary Material online. The predicted proteome of the five smut fungi U. hordei, U maydis, S. scitamineum, S. reilianum f. sp. zeae, and S. reilianum f. sp. reilianum were used to perform an all-against-all BLASTp search . The SiLiX algorithm was subsequently used to infer homology relationships based on the BLAST hits [bib_ref] Ultra-fast sequence clustering from similarity networks with SiLiX, Miele [/bib_ref]. Two parameters are considered to decide whether a BLAST hit can be taken as evidence for homology: the percent identity between two sequences and the relative length of the hit compared with the total length of the two sequences, hereby referred to as "coverage." In order to maximize the number of families comprising 1:1 orthologues (that is families that have an equal number of members in each species), SiLiX [bib_ref] Ultra-fast sequence clustering from similarity networks with SiLiX, Miele [/bib_ref] was run with a range for coverage and identity thresholds between 5% and 95% in 5% steps. An identity of 40% and coverage between 5% and 45% lead to the maximum number of families with 1:1 orthologues (5,394; supplementary fig. S1, Supplementary Material online) while settings with 40% identity and 80% coverage lead to 5,326 families with 1:1 orthologues (supplementary [fig_ref] FIG. 1: -Phylogeny, divergence estimates, and number of genes under positive selection in five... [/fig_ref] , Supplementary Material online). Since using a higher coverage had only a cost of 68 core families, the stricter criteria were applied for family clustering. Families with at least two members were aligned on the codon level using MACSE 1.01 b [bib_ref] MACSE: multiple Alignment of Coding SEquences accounting for frameshifts and stop codons, Ranwez [/bib_ref] and on the protein level using PRANK v.100802 (Lö ytynoja and Goldman 2008). The resulting alignments were subsequently compared and column scores (CS) computed for each position in the alignment [bib_ref] A comprehensive comparison of multiple sequence alignment programs, Thompson [/bib_ref]. Only positions with CS of 100% (that is, alignment columns identically found by both methods) and a maximum of 30% gaps were retained for further analysis.
## Estimation of genome-wide divergence values and divergence times
The five genomes of U. hordei, U. maydis, S. scitamineum, S. reilianum f. sp. zeae, and S. reilianum f. sp. reilianum were aligned using the Multiz genome aligner from the TBA package and projected on the U. maydis genome as reference. The resulting multiple genome alignment had a total size of 21 Mb and was further restricted to regions with homologous sequences in the five species (total length after this step: 14.3 Mb) and processed to remove coding regions. The final noncoding sequence alignment had a total length of 2.2 Mb, for which pairwise nucleotide similarities were computed in nonoverlapping windows of 10 kb.
Gene families with exactly one member in each species were concatenated and pairwise protein sequence similarities computed using the seqinr package for R [bib_ref] SeqinR 1.0-2: a contributed package to the R project for statistical computing..., Charif [/bib_ref]. Protein alignments were also used to infer dates of divergence, using a relaxed clock model. The PhyloBayes version 4.1 software [bib_ref] PhyloBayes 3: a Bayesian software package for phylogenetic reconstruction and molecular dating, Lartillot [/bib_ref] was used with the auto-correlated model of [bib_ref] Estimating the rate of evolution of the rate of molecular evolution, Thorne [/bib_ref] under a GTR þ CAT model. A unique calibration point was used, based on the divergence time of the most divergent lineage U. hordei, previously estimated to have occurred between 27 and 21 Myr (Bakkeren and Kronstad 2007). A uniform prior was used on this interval for the Monte-Carlo Markov Chain. As convergence issues arise when large alignments (>20,000 positions) are used, we followed the PhyloBayes authors' recommendation to conduct a jackknife procedure. We generated three data sets of circa 20,000 amino acids by randomly sampling families and concatenating the corresponding alignments. Two chains were run in each case and convergence was assessed. Sampling was performed after a burning of 10,000 iterations, and every ten subsequent iterations. Chains were run to ensure that the minimum effective sample size was >50 and maximum relative difference <0.3 in at least one sample. Results are summarized in supplementary table S2, Supplementary Material online, and , Supplementary Material online, shows the six chains for the three samples. In addition to the convergence of the two chains for each sample, our results reveal extremely consistent results between samples. [fig_ref] FIG. 1: -Phylogeny, divergence estimates, and number of genes under positive selection in five... [/fig_ref] shows estimates from one chain of the third data set, which shows a minimum effective sample size >300.
## Detection of positive selection
For gene families with at least three members, translated sequences were employed to create maximum likelihood phylogenetic trees using PhyML 3.0 [bib_ref] New algorithms and methods to estimate maximum-likelihood phylogenies: assessing the performance of..., Guindon [/bib_ref] with a minimum parsimony starting tree and the LG amino acid substitution model with a four-classes gamma distribution of sitespecific substitution rate [bib_ref] An improved general amino acid replacement matrix, Le [/bib_ref]. The best tree topology obtained from nearest neighbor interchange (NNI) and subtree pruning recrafting (SPR) searches was kept [bib_ref] New algorithms and methods to estimate maximum-likelihood phylogenies: assessing the performance of..., Guindon [/bib_ref]. BppML [bib_ref] Non-homogeneous models of sequence evolution in the Bioþþ suite of libraries and..., Dutheil [/bib_ref] was then used to re-estimate branch lengths from the codon alignment using the YN98 substitution model . We next aimed at inferring the occurrence of positive selection for each gene family. This is typically achieved by measuring the ratio of nonsynonymous versus synonymous substitutions (d N /d S ratio) using models of codon sequence evolution [bib_ref] Computational molecular evolution, Yang [/bib_ref]. In particular, nonhomogeneous models of sequence evolution estimate the d N /d S ratio independently in different lineages, yet at the cost of potential overparametrization issues. In the manual of the PAML package, the authors state that such models should only be used for hypothesis testing and advise against using them for scans of positive selection. [bib_ref] Efficient selection of branch-specific models of sequence evolution, Dutheil [/bib_ref] proposed a model selection approach (implemented in the TestNH package) allowing to select for the best nonhomogeneous model supported by the data. They start by fitting the simplest (homogeneous) model and sequentially add parameters to model variation of selective regime among lineages. Because the number of possible models is large even for small data sets, two heuristic approaches have been introduced: the "free" heuristic permits unconnected branches from the tree to evolve under the same regime, whereas the "joint" heuristic restricts model sharing to connected branches (see [bib_ref] Efficient selection of branch-specific models of sequence evolution, Dutheil [/bib_ref] for details). The choice of models to test is guided by statistics on the patterns of substitutions on the phylogenetic tree, an approach named substitution mapping [bib_ref] Fast and robust characterization of timeheterogeneous sequence evolutionary processes using substitution mapping, Romiguier [/bib_ref]. Apart from the model selection approach, the underlying models of codon sequence evolution are identical to the one originally described by Yang [bib_ref] Likelihood ratio tests for detecting positive selection and application to primate lysozyme..., Yang [/bib_ref] [bib_ref] Synonymous and nonsynonymous rate variation in nuclear genes of mammals, Yang [/bib_ref]. Model selection was performed with the TestNH software, which contains two programs: 1) MapNH [bib_ref] Fast and robust characterization of timeheterogeneous sequence evolutionary processes using substitution mapping, Romiguier [/bib_ref] was used for mapping substitutions on the previously inferred phylogenetic tree and 2) PartNH [bib_ref] Efficient selection of branch-specific models of sequence evolution, Dutheil [/bib_ref] was subsequently employed to fit time nonhomogeneous models of codon substitutions. PartNH uses the previously inferred substitution maps in order to perform model comparisons and select a nonhomogeneous model with minimal number of parameters. Both methods "free" and "join" were compared in order to scan for positive selection. Finally, putative secreted effector proteins were identified by predicting secretion using SignalP 4.0 [bib_ref] SignalP 4.0: discriminating signal peptides from transmembrane regions, Petersen [/bib_ref] and proteins were considered as secreted if the program indicated the presence of a signal peptide but no transmembrane domain.
To detect residues under positive selection in homologues of pit2, the branch-site model with Bayes Empirical Bayes (BEB) analysis as implemented in PAML4 [bib_ref] PAML 4: phylogenetic analysis by maximum likelihood, Yang [/bib_ref] was applied. We employed information about family composition, alignment, and phylogeny as outlined above and defined sr10529 and srs_10529 as foreground branches. A posterior probability threshold of > 95% was used for the BEB analysis.
## Association of positively selected genes with repeats in u. hordei
We tested whether genes under positive selection are located significantly closer to repetitive elements than average genes in the genome of U. hordei, which shows the highest content of repetitive elements in the group of smut fungi investigated here. For this analysis, only a group of "uncharacterized interspersed repeats" was investigated, because it was shown previously that this is the only category showing a strong association with candidate effector genes [bib_ref] A tale of genome compartmentalization: the evolution of virulence clusters in smut..., Dutheil [/bib_ref]. Binary logistic regressions were conducted in R using the rms package. The "robcov" function of the rms package was used in order to get robust estimates of each effect. The variable "distance to the closest interspersed repeat" was transformed by log(x þ 1) because of its extreme distribution.
## Comparing d n /d s ratios of genes residing in virulence clusters
Previous work has identified several virulence gene clusters in U. maydis and some of them play important roles during pathogenic development [bib_ref] Insights from the genome of the biotrophic fungal plant pathogen Ustilago maydis, K€ Amper [/bib_ref] [bib_ref] Pathogenicity determinants in smut fungi revealed by genome comparison, Schirawski [/bib_ref]. In total, these clusters contain 163 genes, where 100 reside in clusters without virulence phenotype and 63 reside in clusters with virulence phenotype upon deletion. Both types of clusters contain each 32 genes for which a d N /d S ratio could be determined (the missing genes are part of families that do not have at least three members and were therefore not analyzed). The d N /d S ratios of all genes in clusters were compared between clusters with and without virulence phenotype (Wilcoxon rank-sum test).
## Gene ontology terms enrichment analysis
All proteins in S. reilianum f. sp. zeae, in S. reilianum f. sp. reilianum, and in U. hordei were considered for Gene Ontology (GO) term enrichment analyses. GO terms were assigned using iprscan 1.1.0 (http://fgblab.org/runiprscan; developed by Michael R. Thon) which links GO information provided by Interpro to each protein. In this way, 1,759 unique GO terms could be assigned to 4,130 proteins in S. reilianum f. sp. zeae, 1,744 unique GO terms could be assigned to 4,124 proteins in S. reilianum f. sp. reilianum, and 1,757 unique GO terms could be assigned to 3,922 proteins in U. hordei (supplementary table S3, Supplementary Material online). The Bioconductor package topGO was then used to link each GO term to the three major categories "Cellular Component," "Biological Process," or "Molecular Function." Enrichment analysis was performed by computing P values for each GO term using Fisher's classic test with parent-child correction [bib_ref] Improved detection of overrepresentation of Gene-Ontology annotations with parent child analysis, Grossmann [/bib_ref]. Cytoplasmic proteins with and without signatures of positive selection were compared for the three species separately, and differences were considered to be significant at the 5% level.
## Strains and growth conditions
The Escherichia coli derivative Top10 (Invitrogen, Karlsruhe, Germany) and the Saccharomyces cerevisiae strain BY4741 (MATa his3D1 leu2D met15D ura3D; Euroscarf, Frankfurt, Germany; kindly provided by M. Bö lker, Marburg) were used for cloning purposes. Sporisorium reilianum f. sp. zeae strains used in this study are listed in supplementary table S4, Supplementary Material online. They are derivatives of the haploid solopathogenic strain JS161 which is capable of plant colonization without the need of a mating partner, because it expresses a compatible pheromone/receptor pair [bib_ref] Pathogenicity determinants in smut fungi revealed by genome comparison, Schirawski [/bib_ref]
## Construction of s. reilianum strains
Polymerase chain reactions were performed using the Phusion High-Fidelity DNA Polymerase (New England Biolabs). Templates were either JS161 genomic or indicated plasmid DNA. Restriction enzymes were obtained from New England Biolabs. Protoplast-mediated transformation was used to transform S. reilianum following a method established for U. maydis [bib_ref] The b alleles of U. maydis, whose combinations program pathogenic development, code..., Schulz [/bib_ref]. Transformants were selected on RegAgar plates (
[formula] 1.0% [w/v] yeast extract, 0.4% [w/v] Bacto(R)-Pepton, 0.4% [w/v] Sucrose, 1 M Sorbitol, 1.5% [w/v] Bacto(R)-Agar) supplemented with 200 mg/mL [/formula]
Geneticin and true resistance was tested by growing single colonies on PD plates (3.9% [w/v] Potato-Dextrose Agar, 1%
[v/v] Tris-HCl [1 M, pH 8.0]) supplemented with 50 mg/mL Geneticin. Gene replacements with resistance markers were generated with a PCR-based method employing the previously described SfiI insertion cassette system [bib_ref] A reverse genetic approach for generating gene replacement mutants in Ustilago maydis, Brachmann [/bib_ref] [bib_ref] A PCR-based system for highly efficient generation of gene replacement mutants in..., K€ [/bib_ref] and were confirmed by Southern blot analysis. Genomic regions residing $1-kb upstream (left border) or downstream (right border) adjacent to open reading frames of candidate genes were PCR-amplified using the listed primer pairs (supplementary table S5, Supplementary Material online) and genomic DNA of JS161 as template. The resulting fragments were used for cloning plasmids containing the respective deletion constructs.
To obtain deletion constructs for the genes sr10529 and sr14347, PCR fragments containing the left and right borders of each gene were ligated to the hygromycin resistance cassette of pBS-hhn via SfiI restriction sites and cloned into pCRII-TOPO (Life Technologies) to generate pTOPO Dsr10529 #1 and pTOPO Dsr14347 #1, respectively. Since the use of Geneticin as selection marker resulted in much less false positive transformants compared with the use of Hygromycin B, the hygromycin resistance cassettes in these plasmids were replaced by the Geneticin resistance cassette of pUMA 1057 [bib_ref] A reverse genetic approach for generating gene replacement mutants in Ustilago maydis, Brachmann [/bib_ref] by ligation via SfiI restriction sites, yielding plasmids pTOPO Dsr10529 G418 and pTOPO Dsr14347 Gen #1, respectively. Deletion constructs were PCR-amplified from plasmids pTOPO Dsr10529 G418 and pTOPO Dsr14347 Gen #1 using the listed primers (supplementary table S5, Supplementary Material online) and used to transform the S. reilianum strain JS161 to generate the gene deletion strains JS161Dsr10529 and JS161Dsr14347, respectively.
The drag and drop cloning method in yeast [bib_ref] Drag&Drop cloning in yeast, Jansen [/bib_ref] was used to generate plasmids pRS426 Dsr12968 Hyg #1, pRS426 Dsr14944 Hyg #2, pRS426 Dsr10059 Hyg #1, pRS426 Dsr10182 Hyg #1, pRS426 Dsr14558 Hyg #1, and pRS426 Dsr12897 Hyg #1 which contain deletion constructs for deleting the candidate genes sr12968, sr14944, sr10059, sr10182, sr14558, or sr12897. These plasmids are a derivate of plasmid pRS426, which can be maintained in E. coli and S. cerevisiae [bib_ref] A system of shuttle vectors and yeast host strains designed for efficient..., Sikorski [/bib_ref]. PCR-amplified left and right borders of each candidate gene and the hygromycin resistance cassette were integrated in pRS426 by homologous recombination in S. cerevisiae. Subsequently, the hygromycin resistance cassette was replaced with the Geneticin resistance cassette by ligation via SfiI restriction sites, yielding plasmids pRS426 Dsr12968 Gen #1, pRS426 Dsr14944 Gen #3, pRS426 Dsr10059 Gen #1, pRS426 Dsr10182 Gen #1, pRS426 Dsr14558 Gen #1, and pRS426 Dsr12897 Gen #5, respectively. Gene deletion constructs were PCR-amplified from the respective plasmid using listed primers (supplementary table S5, Supplementary Material online). The obtained deletion constructs were used to transform the S. reilianum strain JS161 to generate the gene deletion strains JS161Dsr12968, JS161Dsr14944, JS161Dsr10059, JS161Dsr10182, JS161Dsr14558, and JS161Dsr12897, respectively.
The drag and drop cloning method was also used to generate plasmid pRS426 Dsr12084 Gen #1. PCR-amplified left and right borders of sr12084 and the Geniticin resistance cassette were integrated in pRS426 by homologous recombination in S. cerevisiae. The gene deletion construct for deleting the candidate gene sr12084 was PCR-amplified from plasmid pRS426 Dsr12084 Gen #1 using primers sr12084_lb_fw/sr12084_rb_rv and transformed into the S. reilianum strain JS161 to generate the gene deletion strain JS161Dsr12084.
## Virulence assays
The solopathogenic strain JS161 and deletion mutants thereof were grown in YEPS light liquid medium to an optical density at 600 nm (OD 600 ) of 0.8-1.0 and cell cultures were adjusted to an OD 600 of 1.0 with sterile water prior to injection into 1-week-old maize (Zea mays) seedlings of the dwarf cultivar "Gaspe Flint" (kindly provided by B. Burr, Brookhaven National Laboratories and maintained by self-pollination). Plants were sowed in T-type soil of "Fruhstorfer Pikiererde" (HAWITA, Vechta, Germany) and grown in a temperature-controlled greenhouse (14 h-/10 h-light/dark cycle, with 28/20 C and 25,000-90,000 lux during the light period). Virulence symptoms were scored 9-to 10-week postinfection according to previously described symptoms [bib_ref] Sporisorium reilianum infection changes inflorescence and branching architectures of maize, Ghareeb [/bib_ref] and the following categories were distinguished: the plant did not develop ears, the plant developed healthy ears shorter or equal to 1 cm or the plant developed healthy ears longer than 1 cm, the plant developed spiky ears, phyllody in ears or phyllody in tassels. Spore formation was only observed occasionally, and rarely the plant died due to the infection. Three independent infections were carried out per strain, and at least three independent deletion strains were tested for virulence.
# Results
## Candidate effector genes are less conserved between species compared with other genes
We reconstructed families of homologous genes for the five smut fungi U. hordei, U. maydis, S. scitamineum, S. reilianum f. sp. zeae, and S. reilianum f. sp. reilianum using the SiLiX clustering algorithm [bib_ref] Ultra-fast sequence clustering from similarity networks with SiLiX, Miele [/bib_ref]. We optimized the clustering parameters to maximize the occurrence of orthologues and minimize the number of paralogues within each family. In this way, we were able to reconstruct 8,761 families, among which 5,266 had at least one gene in each species (supplementary table S6, Supplementary Material online). As a consequence, we found at least one homologous sequence in four species for 78% of all genes. About 5,254 gene families are found to have exactly one member in each species and were therefore taken as true orthologues (referred to as "core orthologous set" in the following). Considering that secreted proteins are putative effectors, we used SignalP [bib_ref] SignalP 4.0: discriminating signal peptides from transmembrane regions, Petersen [/bib_ref] to predict secretion of the encoded protein for each gene (supplementary table S1, Supplementary Material online). We report that 920 (11%) families contained only genes encoding a predicted secreted protein, whereas 7,657 (87%) contained only genes encoding a protein not predicted to be secreted. The remaining 184 (2%) families contained both predicted secreted and cytoplasmic proteins (supplementary table S6, Supplementary Material online). The occurrence of families with both secreted and cytoplasmic proteins can be explained by 1) false negative predictions for secretion, as truncated C-terminal sequences were not removed from the data set, 2) wrong gene annotations, or 3) gain or loss of a secretion signal peptide during effector evolution. Among all predicted secreted proteins, 52% have at least one orthologue in all other species, which is significantly less than the global 78% proportion for all proteins (v 2 test, P value < 2.2Â10 À16 ). Genes encoding putative effector proteins are therefore less conserved across species than other genes, either because their sequence is evolving faster, preventing the recovery of homologous relationships, or because effector genes are created or lost at a higher rate. In U. hordei, we observe several species-specific family expansions. There were 17 families which encompassed 5-25 members, but no orthologue in other species (supplementary table S6, Supplementary Material online). Moreover, we identified three families with up to 62 members in U. hordei, but only one member in up to three of the other species , Supplementary Material online). Gene duplications in U. hordei have been hypothesized to be driven by mobile elements [bib_ref] Genome comparison of barley and maize smut fungi reveals targeted loss of..., Laurie [/bib_ref].
The Genomes of S. reilianum f. sp. zeae and S. reilianum f. sp. reilianum Diverged around 1 Ma
To establish a frame for our comparative analysis, we first calculated sequence similarity of the five smut fungi for noncoding intergenic and protein sequences. In addition, we estimated divergence times by performing a molecular dating analysis based on the core orthologues set of the five pathogens, using advanced models of protein sequence evolution and Bayesian inference as implemented in the PhyloBayes package [bib_ref] PhyloBayes 3: a Bayesian software package for phylogenetic reconstruction and molecular dating, Lartillot [/bib_ref]. As calibration point, we used the divergence time of U. hordei and U. maydis, previously estimated to be between 27 and 21 Myr (Bakkeren and Kronstad 2007). In alignable intergenic regions U. hordei shares 57% identity with S. reilianum f. sp. reilianum and 77% identity in protein sequences [fig_ref] FIG. 1: -Phylogeny, divergence estimates, and number of genes under positive selection in five... [/fig_ref]. Monte-Carlo Markov chains were run for three independent gene samples totaling >20,000 amino acid positions each, and two chains were run in each case to assess convergence. The resulting posterior distribution of divergence times were used to infer 95% posterior intervals. The split between U. maydis and the Sporisorium species was estimated to have occurred $20 Myr ago (95% posterior interval 25-12 Myr; [fig_ref] FIG. 1: -Phylogeny, divergence estimates, and number of genes under positive selection in five... [/fig_ref]. Sporisorium reilianum f. sp. reilianum shares 61% nucleotide identity in alignable intergenic regions with U. maydis, and 79% sequence identity at the protein level [fig_ref] FIG. 1: -Phylogeny, divergence estimates, and number of genes under positive selection in five... [/fig_ref]. The divergence times of S. scitamineum and the two formae speciales of S. reilianum were calculated to be 13 Myr ago (95% posterior interval 19-7 Myr; [fig_ref] FIG. 1: -Phylogeny, divergence estimates, and number of genes under positive selection in five... [/fig_ref] and supplementary table S2, Supplementary Material online), which is consistent with the mean divergence estimated between the hosts sorghum and sugarcane (10 Myr with a posterior interval of 8-13 Myr, average over eight studies, source: timetree.org; [bib_ref] TimeTree: a resource for timelines, timetrees, and divergence times, Kumar [/bib_ref]. Sporisorium reilianum f. sp. reilianum and S. scitamineum share 74% noncoding nucleotide identity and 88% identity at the protein level [fig_ref] FIG. 1: -Phylogeny, divergence estimates, and number of genes under positive selection in five... [/fig_ref]. Finally, the two S. reilianum strains diverged 1.1 Myr ago (95% posterior interval 2.4-0.4 Myr; [fig_ref] FIG. 1: -Phylogeny, divergence estimates, and number of genes under positive selection in five... [/fig_ref] and share 98% noncoding nucleotide identity and 99% protein identity [fig_ref] FIG. 1: -Phylogeny, divergence estimates, and number of genes under positive selection in five... [/fig_ref]. We note that the estimation of this divergence date varied with the gene set used, and was in some cases found to be older (1.7 Myr, with a 95% posterior interval of 4.7-0.6 Myr, see supplementary table S2, Supplementary Material online). The comparison of the five smut genomes therefore encompasses a broad evolutionary time, and the divergence times obtained are compatible with previous estimates from smaller data sets [bib_ref] Domestication of maize, sorghum, and sugarcane did not drive the divergence of..., Munkacsi [/bib_ref]. The speciation times of the investigated smut species largely predate the 10,000 years of crop plant domestication, which implies that adaptation to the agricultural host, if any, will be negligible when interpreting the interspecific patterns of sequence divergence, as it represents a marginal proportion of the time since the divergence from the ancestral species.
## Sporisorium reilianum contains the largest number of positively selected genes
To detect positive selection, 6,205 families with at least three members (orthologues and/or paralogues, see supplementary table S6, Supplementary Material online) were, regardless of their species composition, aligned on the codon and amino acid level and a phylogenetic tree was inferred. Obtaining accurate alignments is critical for detecting positive selection since alignment errors frequently inflate the false discovery rate [bib_ref] The effects of alignment error and alignment filtering on the sitewise detection..., Jordan [/bib_ref]. We therefore developed a stringent bioinformatics pipeline for the filtering of sequence alignments by masking ambiguous alignment positions for further analysis (see Materials and Methods). To scan for positive selection, we employed a nonhomogeneous model of sequence evolution allowing d N /d S ratios to vary along the phylogeny, in combination with two heuristic model selection procedures to avoid overparametrization issues [bib_ref] Efficient selection of branch-specific models of sequence evolution, Dutheil [/bib_ref]. Model parameters could not be fitted by either one of the two methods in 1.7% of branches. The two model selection procedures led to highly consistent estimates of branch-specific d N /d S ratios (Spearman's rank correlation coefficient equal to 0.85, P value < 2.2Â10 À16 ). The distribution of d N /d S was highly skewed with a median value of 0.06, demonstrating the strong predominance of purifying selection throughout lineages and genes. The mean value of d N / d S ratios for lineages undergoing positive selection (d N /d S >1) was 4.1 (median 1.9). Although a d N /d S ratio >1 is indicative of positive selection, the absolute value of the ratio is a poor indicator of the strength of undergoing selection. In particular, high ratio values can be obtained because of low d S values, and the d N rate might include neutral substitutions, such as nonsynonymous substitution that are conservative regarding certain biochemical properties of the amino acids involved [bib_ref] Detecting site-specific physicochemical selective pressures: applications to the Class I HLA of..., Sainudiin [/bib_ref]. The largest number of genes with signs of positive selection was found in S. reilianum f. sp. zeae (84 genes, of which 25 encode predicted secreted proteins) and S. reilianum f. sp. reilianum (111 genes of which 27 encode predicted secreted proteins) [fig_ref] FIG. 1: -Phylogeny, divergence estimates, and number of genes under positive selection in five... [/fig_ref]. In addition, a substantial number of positively selected candidate genes was also found in U. hordei (49, and of these, 22 genes are predicted to code for secreted proteins), but only very few in U. maydis (2 genes) and S. scitamineum (7 genes) [fig_ref] FIG. 1: -Phylogeny, divergence estimates, and number of genes under positive selection in five... [/fig_ref]. A list of all proteins with their associated d N /d S ratios in each species is provided in supplementary table S3, Supplementary Material online. Predicted secreted proteins were significantly enriched in the group of proteins under positive selection in U. hordei and in the two investigated formae speciales of S. reilianum (P values < 10 À5 ; Fisher's exact test). This corroborates results of earlier studies in other pathosystems that showed that predicted secreted proteins are often under positive selection, which can be attributed to their direct interaction with host proteins [bib_ref] Comparative analysis of secreted protein evolution using expressed sequence tags from four..., Joly [/bib_ref] [bib_ref] The wheat powdery mildew genome shows the unique evolution of an obligate..., Wicker [/bib_ref]. Notably, all genes found under positive selection in the two strains of S. reilianum, in S. scitamineum and in U.
## Genes under positive selection in u. hordei are associated with uncharacterized interspersed repeats
Among the species compared here, the genome of U. hordei shows the highest fraction of repetitive elements [bib_ref] Genome comparison of barley and maize smut fungi reveals targeted loss of..., Laurie [/bib_ref] [bib_ref] A tale of genome compartmentalization: the evolution of virulence clusters in smut..., Dutheil [/bib_ref]. Such elements are known to contribute to gene family expansions [bib_ref] Mobile elements: drivers of genome evolution, Kazazian [/bib_ref] , and have been suggested to contribute to adaptation by providing advantageous mutations, for instance by repeat-induced point mutations (RIP) leakage which was revealed in a species complex of Leptosphaeria [bib_ref] Effector diversification within compartments of the Leptosphaeria maculans genome affected by Repeat-Induced..., Rouxel [/bib_ref] [bib_ref] Transposable element-assisted evolution and adaptation to host plant within the Leptosphaeria maculans-Leptosphaeria..., Grandaubert [/bib_ref]. As sequence signatures of RIP were found in LTR elements of U. hordei [bib_ref] Genome comparison of barley and maize smut fungi reveals targeted loss of..., Laurie [/bib_ref] , we tested whether genes under positive selection in U. hordei are physically associated with repetitive elements. We performed a binary logistic regression with the prediction of positive selection as a response variable (that is, whether the underlying branch has a d N /d S ratio >1) and we considered three putative explanatory variables for each analyzed gene: 1) whether the gene is predicted to encode a secreted protein, 2) whether the gene is duplicated, and 3) the distance of the gene to the closest interspersed repeat. The complete linear model explains 50% of the observed variance, and the three explanatory variables are all significant at the 0.1% level (supplementary table S7, Supplementary Material online). These results suggest that positively selected genes in U. hordei are associated with duplication events, and positive selection is more likely to occur at genes encoding putative effectors. In addition, the proximity of interspersed repeats increases the odds of positive selection, independently of the two other effects, and is confirmed by a stratification approach: the effect still holds when only duplicated genes are considered, or only genes encoding a secreted protein, or the combination of the two (supplementary table S7, Supplementary Material online). This finding corroborates previous results obtained in other microbial plant pathogens where it was described that effector genes tend to localize in repeat rich regions and where it was suggested that such regions contribute to the rapid evolution of effector genes [bib_ref] Genome evolution in filamentous plant pathogens: why bigger can be better, Raffaele [/bib_ref].
## Positively selected genes encoding cytoplasmic proteins in s. reilianum and u. hordei
Although we expect effector genes to be under positive selection, we find that the majority of positively selected genes in S. reilianum encodes cytoplasmic proteins [fig_ref] FIG. 1: -Phylogeny, divergence estimates, and number of genes under positive selection in five... [/fig_ref]. To assess the putative functional role of these genes, we performed a Gene Ontology term enrichment analysis, comparing cytoplasmic proteins under positive selection to cytoplasmic proteins not under positive selection (table 1). This analysis revealed that genes with a potential role in metabolic processes, like sulfur compound metabolism, molybdopterin cofactor metabolic process, RNA metabolic process, organic cyclic compound metabolic process, and oxidoreductase activity, as well as responses to starvation and extracellular stimuli are significantly overrepresented at the 5% level (Fisher's classic test with Parent-Child correction; see table 1). This could indicate that cytoplasmic proteins under positive selection contribute to metabolic changes which might be needed to survive with the limited nutrients available on the surface or in the biotrophic interface of different host plants. A similar analysis for cytoplasmic proteins under positive selection in U. hordei was conducted, but only led to top-level categories (DNA integration, DNA metabolic process, and isomerase activity; see table 1).
## Virulence contribution of effector genes showing signs of positive selection in s. reilianum
Candidate effector genes inferred to be under positive selection in a particular species could play a critical role in pathogenicity. Therefore, we sought to assess the contribution to virulence of such candidate genes by creating individual deletion mutants. In total, we tested nine candidate genes with high d N /d S ratios and predicted to encode secreted proteins: three with signatures of positive selection only in S. reilianum f. sp. zeae, three with signatures of positive selection in S. reilianum f. sp. zeae as well as in S. reilianum f. sp. reilianum and three with signatures of positive selection only in S. reilianum f. sp. reilianum. All nine chosen candidate genes together with their characteristics are summarized in table 2. Deletion mutants were generated in the haploid solopathogenic strain JS161 of S. reilianum f. sp. zeae. This strain is capable of colonizing maize plants and cause disease without a compatible mating partner [bib_ref] Pathogenicity determinants in smut fungi revealed by genome comparison, Schirawski [/bib_ref] but virulence is much reduced relative to infection with mating-compatible wild-type strains and spores are only rarely produced. Deletion mutants were also generated in strain JS161 in cases where positive selection was only detected in S. reilianum f. sp. reilianum [fig_ref] Table 2: Positively Selected Genes That Were Deleted in Sporisorium reilianum f [/fig_ref] , because no solopathogenic strain is presently available for S. reilianum f. sp. reilianum. For each gene, at least three independent deletion mutants were generated and tested for virulence. To determine virulence, Gaspe Flint, a dwarf variety of corn, was infected and symptoms were scored in male and female flowers . Only the deletion of sr10529, a gene showing positive selection in both formae speciales of S. reilianum, showed a strong reduction in virulence [fig_ref] Table 2: Positively Selected Genes That Were Deleted in Sporisorium reilianum f [/fig_ref].
The gene sr10529 in S. reilianum f. sp. zeae is orthologous to the previously identified and characterized gene pit2 (UMAG_01375) in U. maydis. Pit2 plays an essential role in virulence as inhibitor of a group of maize papain-like cysteine proteases that are secreted to the apoplast [bib_ref] Two linked genes encoding a secreted effector and a membrane protein are..., Doehlemann [/bib_ref] [bib_ref] Compatibility in the Ustilago maydis-maize interaction requires inhibition of host cysteine proteases..., Mueller [/bib_ref]. Previous work identified a conserved domain of 14 amino acids (PID14) in Pit2 as required and sufficient for the inhibition of maize cysteine proteases [bib_ref] Compatibility in the Ustilago maydis-maize interaction requires inhibition of host cysteine proteases..., Mueller [/bib_ref]. When the branch-site model of PAML 4 [bib_ref] PAML 4: phylogenetic analysis by maximum likelihood, Yang [/bib_ref] was used to identify amino acid residues under positive selection in the Pit2 orthologues of the two S. reilianum species, only two residues residing in the PID14 domain were found under positive selection. However, 24 positively selected residues were detected outside this domain in the 57 amino acid long C-terminal part [fig_ref] FIG. 3: -Distribution of positively selected amino acids in the cysteine protease inhibitor Pit2 [/fig_ref]. The two paralogs include: sr12085 and sr12086.
# Discussion
We used evolutionary comparative genomics of five related smut fungi infecting four different host plants to identify genes with a signature of positive selection during species divergence, with a focus on genes encoding predicted secreted proteins, as such genes were suggested to contribute to virulence in various plant pathogenic microbes [bib_ref] A Comprehensive Analysis of Genes Encoding Small Secreted Proteins Identifies Candidate Effectors..., Hacquard [/bib_ref] [bib_ref] Effector specialization in a lineage of the Irish potato famine pathogen, Dong [/bib_ref] [bib_ref] Rapid evolution of avirulence genes in rice blast fungus Magnaporthe oryzae, Huang [/bib_ref] [bib_ref] Gene loss rather than gene gain is associated with a host jump..., Sharma [/bib_ref]. Our analysis revealed that positive selection is found between paralogous genes in U. hordei, where they belong to families with species-specific expansions. In contrast, genes under positive selection in the other four species belong to families of orthologous sequences. Although we find evidence for a large set of genes under positive selection in the S. reilianum species, signatures for positive selection are hardly detectable in the more distant relatives U. hordei, U. maydis, and S. scitamineum that diverged earlier. Finding evidence for positive selection over time spans of several millions of years is notoriously difficultbecause of two main reasons: 1) periods where genes are evolving under positive selection may occur episodically and may be followed by long episodes of purifying selection, leading to an average d N /d S <1 on long periods of time and 2) fast evolving genes may diverge to an extent where their homology is difficult to infer and where they can no longer be aligned reliably. To overcome this problem, more genome information of species with intermediate branching points is needed.
Predicted secreted proteins were about three times overrepresented in the set of positively selected genes, which illustrates the importance of secreted proteins in adaptation processes of smut fungi. This also corroborates results in other plant pathogenic microbes like Melampsora sp., Z. tritici and the wheat powdery mildew Blumeria graminis [bib_ref] Comparative analysis of secreted protein evolution using expressed sequence tags from four..., Joly [/bib_ref] [bib_ref] The wheat powdery mildew genome shows the unique evolution of an obligate..., Wicker [/bib_ref]. However, the majority of positively selected genes encodes FIG. 2. Continued of sr10529 led to a strong reduction in virulence (A). In contrast, deletion of the candidate genes sr12968 (B), sr14944 (C), sr10059 (D), sr10182 (E), sr14558 (F), sr14347 (G), sr12897 (H), and sr12084 (I) did not alter virulence. Symptoms were scored $9-week postinfection and categorized according to severeness as illustrated in the legend below the bar plots. Results are shown as mean of three independent experiments in relation to the total number of infected plants, which is indicated above each bar (n). Note that strains JS161DSr10529 #G4 and #G5 (A) were only infected in one replicate. cytoplasmic proteins [fig_ref] FIG. 1: -Phylogeny, divergence estimates, and number of genes under positive selection in five... [/fig_ref] , suggesting that both secreted and nonsecreted proteins are important targets of adaptation. A Gene Ontology analysis in S. reilianum showed that mainly processes related to metabolism and its regulation as well as responses to starvation and external stimuli are enriched in cytoplasmic proteins under positive selection. This points at a role of these proteins in adaptation to differences in nutrient availability in the respective host plants maize and sorghum as well as responses to cues originating from the respective host [bib_ref] Life cycle specialization of filamentous pathogens -colonization and reproduction in plant tissues, Haueisen [/bib_ref]. A study conducted in U. maydis has shown that the fungus induces major metabolic changes in the host plant upon infection during establishment of biotrophy and undergoes a series of developmental transitions during host colonization that are likely influenced by the host environment [bib_ref] Reprogramming a maize plant: transcriptional and metabolic changes induced by the fungal..., Doehlemann [/bib_ref]. It is thus conceivable that the two S. reilianum accessions have adapted to their different hosts that differ significantly for example in their amino acid and vitamin composition [bib_ref] Nutrient composition and feeding value of Sorghum for livestock and poultry: a..., Etuk [/bib_ref]. Furthermore, recent studies in U. maydis suggested that intracellular changes of metabolism influence virulence, and therefore the underlying proteins could be targets of positive selection [bib_ref] Defects in mitochondrial and peroxisomal b-oxidation influcence virulence in the maize pathogen..., Kretschmer [/bib_ref] [bib_ref] Carbon acquisition and metabolic changes during fungal biotrophic plant pathogenesis: insights from..., Goulet [/bib_ref].
Out of nine deletions of positively selected genes, only one mutant, lacking sr10529, was affected in virulence. Although six of the deleted genes are single genes in S. reilianum f. sp. zeae for which we failed to identify paralogs, sr12084 has two paralogs, sr14347 has five paralogs, and sr10182 has ten paralogs. We restricted our analyses to generating deletion mutants in some of the genes under positive selection. This leaves open the possibility that the paralogous genes have redundant functions in virulence. Adapting the CRISPR-Cas9 technology allowing multiplexing [bib_ref] Comparative analyses of secreted proteins in plant pathogenic smut fungi and related..., Schuster [/bib_ref] to S. reilianum will be instrumental in testing this hypothesis in future studies. Alternatively, the candidate effectors we investigated may be needed under conditions which differ from those tested here. For example, S. reilianum f. sp. zeae can also systemically colonize maize plants via root infection [bib_ref] Assessment of Ustilago maydis as a fungal model for root infection studies, Mazaheri-Naeini [/bib_ref] , a colonization route we have not assessed in our experimental setup. Moreover, we employed only one maize cultivar for infection assays. Results from other pathosystems suggest that virulence effects can strongly depend on the host and pathogen genotypes, in particular in the presence of R and avr genes (Petit-Houdenot and Fudal 2017). No avr-R gene interaction was described so far in the S. reilianum f. sp. zeae-maize pathosystem. Instead, quantitative virulence differences are observed when different host cultivars are infected [bib_ref] QTL mapping of resistance to Sporisorium reiliana in maize, Lü Bberstedt [/bib_ref]. Knowing the expression profile of effector genes may assist the identification of differences in development of the mutants compared with wild-type strains. Since we lack this information, we scored disease symptoms only in the inflorescences $9 weeks after infection. Additionally, it may be possible that small differences in virulence between JS161 and deletion mutants of candidate genes remain undetected due to the weak infection behavior of JS161. For example, spore formation is only rarely observed after infecting maize plants with the solopathogenic strain. In contrast, infections resulting from infections with two compatible haploid strains show spores in $40% of the infected plants [bib_ref] Host specificity of Sporisorium reilianum is tightly linked to generation of the..., Zuther [/bib_ref]. This means that defects related to spore formation will not be evident in mutants of JS161. In three cases, positive selection was detected in orthologous genes in S. reilianum f. sp. reilianum while candidate effector genes were for experimental reasons deleted in S. reilianum f. sp. zeae. Therefore, it cannot be excluded that these effectors might have a virulence function in S. reilianum f. sp. reilianum. In this case, the positively selected effector genes might have evolved during adaptation to the sorghum host and present host specificity genes. In summary, our virulence assays leave open the possibility that the eight candidate genes which did not show a contribution to virulence could play a role in pathogenicity under conditions not tested here. Alternatively, candidate effector proteins might also be positively selected for traits that are not directly linked to pathogenicity. Such traits could for instance involve competition with large numbers of other plant colonizing microbes [bib_ref] Experimental measures of pathogen competition and relative fitness, Zhan [/bib_ref] [bib_ref] Filamentous pathogen effector functions: of pathogens, hosts and microbiomes, Rovenich [/bib_ref]. Secreted proteins of S. reilianum could act for example as toxin or could efficiently utilize resources from the environment and thereby limit the growth of other microbes. In these cases, a contribution to virulence is not expected to be observed in the employed infection assay with the effector gene mutants. Moreover, our molecular dating analysis showed that the common ancestors of the investigated smut species originated before the beginning of crop domestication. Therefore, positive selection, whose signs we detect by our approach, has most likely occurred on ancestral host plants and not on the domesticated host maize. Consequently, some of the candidate effector genes under positive selection might not be important for the colonization of crop plants, but for infection of related wild species.
In U. maydis, we note that effector genes residing in clusters whose deletion affected virulence [bib_ref] Insights from the genome of the biotrophic fungal plant pathogen Ustilago maydis, K€ Amper [/bib_ref] [bib_ref] Pathogenicity determinants in smut fungi revealed by genome comparison, Schirawski [/bib_ref] have similar d N /d S ratios as effector genes in clusters where the deletion had no effect on virulence (median d N /d S ratio 0.0619 vs. 0.1094; Wilcoxon rank test with P value ¼ 0.1848). Furthermore, orthologues of the effectors Pep1, Stp1, and Cmu1, which were shown to have important roles in pathogenicity of U. maydis [bib_ref] Pep1, a secreted effector protein of Ustilago maydis, is required for successful..., Doehlemann [/bib_ref] [bib_ref] Metabolic priming by a secreted fungal effector, Djamei [/bib_ref] showed no signature of positive selection. These observations could suggest that certain fungal effector proteins are under evolutionary constraint and are therefore not free to accumulate nonsynonymous mutations. Such effectors are conserved over long time spans [bib_ref] Pathogenicity determinants in smut fungi revealed by genome comparison, Schirawski [/bib_ref] [bib_ref] The fungal core effector Pep1 is conserved across smuts of dicots and..., Hemetsberger [/bib_ref] [bib_ref] Comparative genomics including the early-diverging smut fungus Ceraceosorus bombacis reveals signatures of..., Sharma [/bib_ref] and this illustrates that they are instrumental for successful infections in a large group of smut fungi. They probably target molecules shared by several host plants, for example, housekeeping functions that cannot easily evolve in response to the binding of an effector.
One candidate gene (sr10529) under positive selection in both formae speciales of S. reilianum showed a strong contribution to virulence upon deletion. It is orthologous to the previously described protease inhibitor Pit2 in U. maydis, where the deletion also abolished virulence [bib_ref] Two linked genes encoding a secreted effector and a membrane protein are..., Doehlemann [/bib_ref] [bib_ref] Compatibility in the Ustilago maydis-maize interaction requires inhibition of host cysteine proteases..., Mueller [/bib_ref]. Positively selected residues in the PID14 domain of Pit2 might reflect that different proteases need to be inhibited in maize and sorghum. Pit2 might thus contribute to determining the host range of the respective species. A role of cysteine protease inhibitors in host specificity was demonstrated in Phytophthora infestans, a pathogen of potato and its sister species P. mirabilis, which infects the ornamental plant Mirabilis jalapa. Positively selected orthologous protease inhibitors were shown to inhibit proteases specific to the respective host plants and this specificity could be traced back to a single amino acid substitution [bib_ref] Effector specialization in a lineage of the Irish potato famine pathogen, Dong [/bib_ref]. Surprisingly, 24 positively selected sites in Pit2 were detected outside the PID14 domain in the 57 amino acid long C-terminal part in both S. reilianum f. sp. zeae and S. reilianum f. sp. reilianum. This finding raises the intriguing possibility that the C-terminus of Pit2 might possess a second function that is independent of protease inhibition. Earlier work has shown that the pit1 gene encoding a transmembrane protein is located next to the pit2 effector gene and both genes contribute similarly to virulence [bib_ref] Two linked genes encoding a secreted effector and a membrane protein are..., Doehlemann [/bib_ref]. Furthermore, pit1 and pit2 are divergently transcribed, which makes it likely that the expression of pit1 and pit2 is coregulated. In addition, this gene arrangement of pit1 and pit2 is conserved in U. hordei, U. maydis, S. scitamineum, and S. reilianum [bib_ref] Comparative genomics including the early-diverging smut fungus Ceraceosorus bombacis reveals signatures of..., Sharma [/bib_ref]. This finding has led to the speculation that Pit1 and Pit2 somehow act together to govern virulence of U. maydis and related smut fungi. It was hypothesized that that Pit2 shuttles apoplastic maize proteins toward Pit1, thereby scavenging damage-associated molecules [bib_ref] Two linked genes encoding a secreted effector and a membrane protein are..., Doehlemann [/bib_ref]. In this scenario, the positively selected amino acids in the C-terminus of Pit2 could have been selected for scavenging such molecules as adaptation to the two hosts. In future studies, it will be highly interesting to complement the pit2 mutant of S. reilianum f. sp. zeae with the pit2 orthologue of S. reilianum f. sp. reilianum to see if this promotes virulence on sorghum.
# Conclusions
Screens for genes with signs of positive selection are commonly used to identify candidate effector genes in various plant pathogenic microbes. However, it is currently largely open whether positively selected effector genes play indeed a role in virulence. Here, we used comparative genomics of five smut fungi and showed that only one out of nine genes under positive selection contributes to virulence of S. reilianum. Moreover, the majority of positively selected genes did not encode predicted secreted proteins. Our results leave open the possibility that many genes with signatures of positive selection contribute to virulence under conditions not tested in this study or are selected in traits that are not directly related to pathogenicity.
# Supplementary material
Supplementary data are available at Genome Biology and Evolution online.
[fig] FIG. 1: -Phylogeny, divergence estimates, and number of genes under positive selection in five related smut fungal species parasitizing different host plants. (A) Chronogram of the five fungal pathogens as estimated under a relaxed molecular clock. Boxes represent 95% posterior intervals, with corresponding values indicated below. (B) Pairwise sequence differences, for both the noncoding genome and the proteome (nonsynonymous differences). (C) Number of positively selected genes on each terminal branch (total number of genes and genes predicted to encode a secreted protein). [/fig]
[fig] FIG. 3: -Distribution of positively selected amino acids in the cysteine protease inhibitor Pit2. The alignment shows the protein sequences of orthologues in Ustilago hordei (UHOR_02064), U. maydis (UMAG_01375), Sporisorium scitamineum (SPSC_03677), S. reilianum f. sp. zeae (sr10529), and S. reilianum f. sp. reilianum (srs_10529). Sites under positive selection detected by a branch-site model are indicated by colored bold letters. Residues colored in red indicate positive selection detected in the respective species and purple residues indicate sites found under positive selection in both species. The yellow shaded area is orthologous to the previously identified conserved PID14 domain, which is required and sufficient for inhibition of a group of papain-like cysteine proteases. Green sequences indicate secretion signal peptides and bold numbers above the alignment indicate positions in UHOR_02064. [/fig]
[table] Table 1: Gene Ontology Terms Significantly Overrepresented in Positively Selected Genes Encoding Cytoplasmic Proteins in Sporisorium reilianum and Ustilago hordei [/table]
[table] Table 2: Positively Selected Genes That Were Deleted in Sporisorium reilianum f. sp. zeae and Their Selection Criteria Species are S. reilianum f. sp. zeae (Srz) and S. reilianum f. sp. reilianum (Srr). [/table]
|
Overexpression of Acetyl CoA Carboxylase 1 and 3 (ACCase1 and ACCase3), and CYP81A21 were related to cyhalofop resistance in a barnyardgrass accession from Arkansas
Barnyardgrass [Echinochloa crus-galli (L.) P. Beauv.] is the most difficult-to-control weed species of rice production systems worldwide. It has evolved resistance to different herbicide sites of action, including the acetyl-CoA carboxylase (ACCase)-inhibiting herbicides. Target-site mutations conferring resistance to ACCase-inhibiting herbicides are well documented; however, the role of the different ACCase genes in conferring resistance to cyhalofop-p-butyl (cyhalofop), an ACCase-inhibiting herbicide, remains poorly understood. This research assessed the contribution of gene amplification and expression of ACCase genes in a cyhalofop-resistant barnyardgrass accession. Additionally, the expression of glutathione-S-transferases (GSTs) and cytochrome P450 monooxygenases (P450s) genes as possible contributors to resistance to cyhalofop were investigated. Results demonstrated that ACCase gene amplification does not contribute to cyhalofop resistance. However, ACCase1 and ACCase3 were found to be overexpressed in the cyhalofop-resistant barnyardgrass accession. At 24 h after cyhalofop treatment, an overexpression of 2.0and 2.8-fold was detected in ACCase1 and ACCase3, respectively. In addition, CYP81A21 (a P450 gene) was found to be 2.5-fold overexpressed compared to the susceptible accession in the same time period. These results suggest that ACCase1, ACCase3, and CYP81A21 are crucial genes in contributing cyhalofop resistance in this barnyardgrass accession.ARTICLE HISTORY
# Introduction
Barnyardgrass [Echinochloa crus-galli (L.) P. Beauv.] is a troublesome grass weed species in different cropping systems, but its presence is frequently related to rice (Oryza sativa L.) cultivation. In rice, the most prevalent weed species reported worldwide include barnyardgrass, weedy rice (Oryza sativa L.), and Cyperus weed species. [bib_ref] Global distribution of rice weeds -a review, Kraehmer [/bib_ref] Barnyardgrass is a hexaploid (2 n = 6x = 54) weed species. [bib_ref] Genetic relationship between Echinochloa crus-galli and Echinochloa oryzicola accessions inferred from internal..., Aoki [/bib_ref] [bib_ref] Echinochloa crus-galli genome analysis provides insight into its adaptation and invasiveness as..., Guo [/bib_ref] It possesses a remarkable ability to adapt to new environments and produces an exorbitant number of seeds (ranging from 34,600 to 39,000 seeds plant −1 ) that are easily dispersed by different means. [bib_ref] Seed production of barnyardgrass (Echinochloa crus-galli) in response to time of emergence..., Bagavathiannan [/bib_ref] In addition, barnyardgrass has allelopathic characteristics. [bib_ref] Eco-biology and management of Echinochloa crus-galli, Bajwa [/bib_ref] Yield reductions in rice due to barnyardgrass vary depending on the conditions studied; however, yield reductions can be up to 70% [reviewed by [bib_ref] Eco-biology and management of Echinochloa crus-galli, Bajwa [/bib_ref].
Control of barnyardgrass relies on the use of herbicides, and unfortunately, many accessions with resistance to different molecules have been reported worldwide.In the US, herbicide-resistant barnyardgrass accessions exist in various states, including Arkansas, the largest rice-producing state in the US. [bib_ref] Resistance of Echinochloa crus-galli populations to acetolactate synthase-inhibiting herbicides, Riar [/bib_ref] [bib_ref] Echinochloa resistance to herbicides continues to increase in Arkansas rice fields, Rouse [/bib_ref] Globally, barnyardgrass has evolved resistance to many herbicide sites of action, with some specific herbicide examples being propanil (photosystem II-inhibiting herbicide), [bib_ref] Verification and distribution of propanil-resistant barnyardgrass (Echinochloa crus-galli) in Arkansas, Carey [/bib_ref] pendimethalin (microtubule assembly-inhibiting herbicide), 7 mefenacet (very long-chain fatty acid synthesisinhibiting herbicide), [bib_ref] Mefenacet resistance in multiple herbicide-resistant Echinochloa crus-galli L, Cai [/bib_ref] quinclorac (auxin mimic-inhibiting herbicide), [bib_ref] Quinclorac absorption and translocation characteristics in quinclorac-and propanil-resistant and -susceptible barnyardgrass (Echinochloa..., Lovelace [/bib_ref] cyhalofop, [bib_ref] Cross-resistance of barnyardgrass [Echinochloa crus-galli (L.) P. Beauv.] to aryloxyphenoxypropionate herbicides, Hwang [/bib_ref] penoxsulam (acetolactate synthase inhibiting-herbicide), [bib_ref] A novel Phe 206 Leu mutation in acetolactate synthase confers resistance to..., Fang [/bib_ref] glyphosate [5-enolpyruvylshikimate-3-phosphate synthase-inhibiting herbicide (EPSPS)], [bib_ref] Distribution of glyphosate-resistance in Echinochloa crus-galli across agriculture areas in the Iberian..., Vázquez-García [/bib_ref] and clomazone (carotenoid biosynthesis-inhibiting herbicide).The acetyl CoA carboxylase (ACCase; EC. 6.4.1.2) is a crucial enzyme that catalyzes the carboxylation of acetyl-CoA to produce malonyl-CoA. [bib_ref] Molecular bases for sensitivity to acetyl-Coenzyme A carboxylase inhibitors in black-grass, Délye [/bib_ref] In plants, ACCase is essential for the formation of primary fatty acids together with the buildup of long-chain fatty acids and flavonoids. ACCaseinhibiting herbicides work by blocking the formation of malonyl CoA, and, as a consequence, susceptible plants die from not having the ability to produce fatty acids. [bib_ref] Molecular bases for sensitivity to acetyl-Coenzyme A carboxylase inhibitors in black-grass, Délye [/bib_ref] [bib_ref] Resistance to acetyl-CoA carboxylase-inhibiting herbicides, Kaundun [/bib_ref] Necrosis of meristematic tissues, chlorosis and necrosis of apical meristem (flag leaf), and a purple leaf coloration due to the accumulation of anthocyanins are typical injury symptoms in plants treated with ACCase-inhibiting herbicides. [bib_ref] Molecular bases for sensitivity to acetyl-Coenzyme A carboxylase inhibitors in black-grass, Délye [/bib_ref] [bib_ref] Herbicidal properties of sethoxydim for the control of gramineous weeds, Ishikawa [/bib_ref] Rice can safely tolerate cyhalofop due to a lack of esterase function, reduced absorption of the herbicide, and a rapid metabolism of cyhalofop acid. [bib_ref] Basis of selectivity of cyhalofop-butyl in Oryza sativa L, Ruiz-Santaella [/bib_ref] In barnyardgrass, a severe cessation of growth, chlorosis, and desiccation may be observed together with new leaf growth-inhibition. [bib_ref] Physiological basis of differential phytotoxic activity between fenoxaprop-P-ethyl and cyhalofop-butyl-treated barnyardgrass, Kim [/bib_ref] Nonetheless, resistant plants have deployed different strategies to avoid herbicide damage. Among those, plants have evolved a target-site mutation in the ACCase gene together with metabolic resistance, resulting in the herbicide being converted to nontoxic forms to surpass the herbicide toxicity. [bib_ref] Molecular bases for sensitivity to acetyl-Coenzyme A carboxylase inhibitors in black-grass, Délye [/bib_ref] [bib_ref] Resistance to acetyl-CoA carboxylase-inhibiting herbicides, Kaundun [/bib_ref] [bib_ref] Mechanisms of evolved herbicide resistance, Gaines [/bib_ref] Metabolic resistance of herbicides is convoluted and is normally governed by superfamilies of GSTs and P450s. However, it also includes glucosyltransferases (GTs) and aldo-keto reductase enzymes. Those superfamilies have been correlated to herbicide metabolism in different plant species. [bib_ref] Mechanisms of evolved herbicide resistance, Gaines [/bib_ref] [bib_ref] Quizalofop-p-ethyl resistance in Polypogon fugax involves glutathione S-transferases, Chen [/bib_ref] [bib_ref] Key role for a glutathione transferase in multiple-herbicide resistance in grass weeds, Cummins [/bib_ref] [bib_ref] Target-site and metabolic resistance mechanisms to penoxsulam in barnyardgrass (Echinochloa crus-galli (L.)..., Fang [/bib_ref] [bib_ref] Cytochrome P450 CYP81A10v7 in Lolium rigidum confers metabolic resistance to herbicides across..., Han [/bib_ref] [bib_ref] CYP 81A P450s are involved in concomitant cross-resistance to acetolactate synthase and..., Iwakami [/bib_ref] [bib_ref] CYP81A68 confers metabolic resistance to ALS and ACCase-inhibiting herbicides and its epigenetic..., Pan [/bib_ref] [bib_ref] Aldo-keto reductase metabolizes glyphosate and confers glyphosate resistance in Echinochloa colona, Pan [/bib_ref] [bib_ref] Cyhalofopbutyl resistance conferred by a novel Trp-2027-Leu mutation of acetyl-CoA carboxylase and..., Zhao [/bib_ref] In previous research, the metabolic resistance of barnyardgrass to cyhalofop herbicide was reported. The resistance in this population was attributed to a reduction in cyhalofop-acid compounds in the resistant accession compared to the susceptible standard. [bib_ref] Cross-resistance of barnyardgrass [Echinochloa crus-galli (L.) P. Beauv.] to aryloxyphenoxypropionate herbicides, Hwang [/bib_ref] Here, the contribution of ACCase genes to cyhalofop resistance has been explored. Thus, the main objective of this research was to describe the target-site resistance mechanisms involved in a cyhalofop-resistant barnyardgrass accession collected in eastern Arkansas. Gene-specific ACCase primers were designed to study their role in gene expression, gene amplification, and assess their involvement in the observed resistance to cyhalofop herbicide. Additionally, the role of GSTs and P450s genes as possible contributors to resistance to cyhalofop herbicide in barnyardgrass was examined.
# Materials and methods
# Plant material
A barnyardgrass accession previously characterized for resistance to cyhalofop and other aryloxyphenoxypropionate herbicides, together with a known susceptible accession were used in this research. [bib_ref] Cross-resistance of barnyardgrass [Echinochloa crus-galli (L.) P. Beauv.] to aryloxyphenoxypropionate herbicides, Hwang [/bib_ref] Under greenhouse conditions, the resistant barnyardgrass accession displayed a resistance factor of ≈ 15 compared to the susceptible standard when treated with cyhalofop. [bib_ref] Cross-resistance of barnyardgrass [Echinochloa crus-galli (L.) P. Beauv.] to aryloxyphenoxypropionate herbicides, Hwang [/bib_ref] To decrease biological variation among plants, the resistant accession followed an additional selection by treating plants with 3× cyhalofop field rate (1× = 313 g ai ha −1 ). Treatments were carried out using a lab track sprayer calibrated to deliver 187 L ha −1 using 1100067 nozzles. Further selected seeds were germinated using growing medium (Promix, LP15, Premier Horticulture Inc., PA, USA). When seedlings reached the one-leaf stage, they were transplanted to a 5 cm-diameter pot containing silt loam field soil (34% sand, 53% silt, 13% clay, and 1.5% organic matter). Plants were maintained under greenhouse conditions as described elsewhere. 13
## Accase-specific gene sequencing
Complementary DNA (cDNA) was extracted as described before. [bib_ref] Understanding resistance mechanisms to trifluralin in an Arkansas Palmer amaranth population, González-Torralva [/bib_ref] To describe the presence of target-site mutations, six specific primer sets, as described in [bib_ref] Multiple-herbicide resistance in Echinochloa crus-galli var. formosensis, an allohexaploid weed species, in..., Iwakami [/bib_ref] , were used to partially amplify the putative six ACCase gene copies in the cyhalofopresistant and -susceptible barnyardgrass accessions [fig_ref] Table 1: Primer sets used to partially amplify the ACCase gene copies in cyhalofop-resistant... [/fig_ref]. Primer sets covered the region where target-site mutations have been previously associated with resistance to ACCaseinhibiting herbicides. Polymerase Chain Reactions (PCRs) were performed using OneTaq Hot Start 2× Master Mix with Standard Buffer (New England Biolabs, Ipswich, MA, USA) according to manufacturer directions and using 5 µL cDNA (five-fold dilution) as template in 25 µL reaction. To assess the best PCR conditions for ACCase amplification, gradient PCRs were carried out at different temperatures to select the optimum annealing temperature for each ACCase gene. Final PCR conditions were as follows: 94°C for 30s, 30 cycles of 94°C for 30s, 59-60°C for 30s, and 68°C for 135 s, a final extension of 68°C for 5 min, and reactions held at 10°C. After cycling, a 5 µL PCR product aliquot was loaded onto a 1.2% agarose gel to corroborate a single presence band. The remaining PCR products were cleaned using the Wizard SV Gel and PCR Clean up System (Promega Corp., Madison, WI, USA) and sent for bidirectionally Sanger sequencing (Eurofins Genomics, Louisville, KY, USA). Raw sequences were processed using BioEdit software. [bib_ref] BioEdit: a user-friendly biological sequence alignment editor and analysis program for Windows..., Hall [/bib_ref] At least three biological replicates for each resistant and susceptible barnyardgrass accession were sequenced.
## Gene expression under metabolic inhibition
Plants of resistant and susceptible accessions were managed as described in "Plant material" section. Plants were treated at the three-to four-leaf growth stage with either malathion (a P450 inhibitor) or NBD-Cl (a GST inhibitor) to assess ACCase, GSTs, and P450s gene expression under metabolic inhibition. Treatments consisted of a non-treated control (T1); cyhalofop at 313 g ai ha −1 (T2); malathion at 2 kg ai ha −1 + cyhalofop at 313 g ai ha −1 (T3); and NBD-Cl at 80 g ha −1 followed by cyhalofop at 313 g ai ha −1 (T4) 48 h later. NBD-Cl was dissolved on 1% acetone + 0.1% Tween 20. [bib_ref] Key role for a glutathione transferase in multiple-herbicide resistance in grass weeds, Cummins [/bib_ref] Cyhalofop treatments contained 1% v/v crop oil concentrate (Agri-dex, Helena Chemical Company, Collierville, TN, USA). Each pot had two seedlings which was considered one replicate, and each treatment had three replicates (n = 6).
## Rna extraction and complementary dna (cdna) synthesis
RNA extraction and complementary DNA (cDNA) synthesis were performed as described before. [bib_ref] Understanding resistance mechanisms to trifluralin in an Arkansas Palmer amaranth population, González-Torralva [/bib_ref] Fresh tissue of resistant and susceptible barnyardgrass accessions was collected, placed
## Accase gene expression
Transcripts were measured in non-treated and treated plants of resistant and susceptible barnyardgrass accessions. Thus, tissue of non-treated and treated plants was collected 24 h after treatments (HAT) for total RNA and cDNA synthesis as described earlier. Quantitative PCR (qPCR) was utilized to measure the transcript levels using cDNA according to MIQE guidelines. [bib_ref] The MIQE guidelines: minimum information for publication of quantitative real-time PCR experiments, Bustin [/bib_ref] Homologs of β-actin and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) genes in barnyardgrass were used as reference genes to estimate the basal transcript levels in both treated and non-treated plants of resistant and susceptible accessions [bib_ref] CYP81A68 confers metabolic resistance to ALS and ACCase-inhibiting herbicides and its epigenetic..., Pan [/bib_ref] [bib_ref] A novel mutation Asp-2078-Glu in ACCase confers resistance to ACCase herbicides in..., Fang [/bib_ref] [bib_ref] Resistance to quinclorac caused by the enhanced ability to detoxify cyanide and..., Gao [/bib_ref] [fig_ref] Table 2: µL total volume ran on 96-well plates [/fig_ref]. Previously research in barnyardgrass (E. crus-galli var. formosensis) resistant to cyhalofop reported the presence of six ACCase gene copies. [bib_ref] Multiple-herbicide resistance in Echinochloa crus-galli var. formosensis, an allohexaploid weed species, in..., Iwakami [/bib_ref] These six copies of ACCase genes were used in this study as target genes. Nucleotide sequences of ACCase gene copies deposited at the National Center for Biotechnology Information (NCBI) were retrieved to design a set of gene-specific primers. Additionally, a set of primers common to all six ACCase gene copies was designed based on a highly conserved region. Primers were designed using the Primer3Plus software 37 ( template controls (deionized water instead of cDNA) on each primer set were included in every qPCR run. Amplicon efficiency for each primer set and sample reactions was obtained using the LinRegPCR software version 2021.2. [bib_ref] Amplification efficiency: linking baseline and bias in the analysis of quantitative PCR..., Ruijter [/bib_ref] Fold-change expression values were calculated using the formula:
[formula] Relative gene expression ¼ ðE TG Þ ΔCqTG GeoMean½ E RG Þ ΔCqRG � i [/formula]
where E corresponds to base of exponential amplification, ΔCq TG is the difference in the quantification cycles (Cq) of the target gene (TG), GeoMean represents the geometric mean of both reference genes (RG) (β-actin and GAPDH), and ΔCq RG is the difference in quantification cycles of reference genes. [bib_ref] Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple..., Vandesompele [/bib_ref] To reduce biological variation among samples, equal amounts of leaf tissue from two plants were pooled to generate a single biological replicate. Subsequently, three biological replicates (n = 6) were obtained and further processed. A Student's t-test was used to detect significant differences in fold-change expression between resistant and susceptible barnyardgrass accessions on each treatment.
## P450s and gsts gene expression
Homologs of the CYP81A21, CYP81A68, GST1, GST1a, GST2, GST2c, and GSTL3 genes in barnyardgrass were retrieved from the NCBI database. Primer design followed the same methodology as described earlier. Such genes have been reported in conferring resistance to different ACCase-inhibiting herbicides in Asia minor bluegrass (Polypogon fugax Nees ex Steud.), barnyardgrass, late watergrass [Echinochloa phyllopogon (Stapf) Koso-Pol.], and rigid ryegrass (Lolium rigidum Gaudin). [bib_ref] Quizalofop-p-ethyl resistance in Polypogon fugax involves glutathione S-transferases, Chen [/bib_ref] [bib_ref] Cytochrome P450 CYP81A10v7 in Lolium rigidum confers metabolic resistance to herbicides across..., Han [/bib_ref] [bib_ref] CYP 81A P450s are involved in concomitant cross-resistance to acetolactate synthase and..., Iwakami [/bib_ref] [bib_ref] CYP81A68 confers metabolic resistance to ALS and ACCase-inhibiting herbicides and its epigenetic..., Pan [/bib_ref] ACCase gene copy number Genomic DNA (gDNA) extracted from resistant and susceptible barnyardgrass was used to estimate the gene copy number of the ACCase genes relative to two reference genes using a qPCR approach and adhering to MIQE guidelines 34 as described in the "ACCase-specific gene sequencing" section. Homologs of the cinnamoyl-CoA reductase (CCR) and peter Pan-like (PPAN) genes in barnyardgrass were used as reference genes [fig_ref] Table 2: µL total volume ran on 96-well plates [/fig_ref]. Such genes have been reported as a single gene copy in different plant genomes. [bib_ref] Isolation and characterization of a cinnamoyl-CoA reductase gene from perennial ryegrass (Lolium..., Mcinnes [/bib_ref] Gene copy number estimation was carried out as described before. [bib_ref] Understanding resistance mechanisms to trifluralin in an Arkansas Palmer amaranth population, González-Torralva [/bib_ref] The gDNA was extracted from resistant and susceptible barnyardgrass accessions using the E.Z.N.A. Plant DNA kit (Omega Bio-Tek Inc., Norcross, GA, USA) and diluted to 10 ng µL −1 to use as template in subsequent qPCR reactions. Standard curves were created using 5-fold gDNA dilutions, and the efficiency (E) for each primer set was calculated as E = [10 (−1/slope) -1] × 100.Each qPCR reaction included the same reagents as described earlier but 15 ng gDNA and run under the same conditions. The amplification protocol included 3 min at 98°C followed by 40 cycles of 10 sec at 98°C and 30 sec at 61°C and finalized with a standard melt curve analysis to corroborate correct amplification. Relative ACCase gene copy number was expressed as 2 ΔCq as described elsewhere. [bib_ref] Gene amplification confers glyphosate resistance in Amaranthus palmeri, Gaines [/bib_ref] [bib_ref] Analysis of relative gene expression data using real-time quantitative PCR and the..., Livak [/bib_ref] The experiment included four plants (n = 4) per accession, and each run included two technical replicates for each gene. A Student t-test was used to identify differences between resistant and susceptible barnyardgrass accessions.
# Results and discussion
## Accase-specific gene sequencing
Target-site mutations in the ACCase gene have been previously reported as responsible for conferring resistance to ACCaseinhibiting herbicides. [bib_ref] Molecular bases for sensitivity to acetyl-Coenzyme A carboxylase inhibitors in black-grass, Délye [/bib_ref] Those target-site mutations can be found commonly at positions 1781, 1999, 2027, 2041, 2078, 2088, and 2096 (numbered respect to that of Alopecurus myosuroides: GenBank accession AJ310767.1), all in the carboxyl transferase domain of ACCase enzyme. [bib_ref] Resistance to acetyl-CoA carboxylase-inhibiting herbicides, Kaundun [/bib_ref] This research attempted to sequence all six ACCase gene copies previously described in barnyardgrass. [bib_ref] Multiple-herbicide resistance in Echinochloa crus-galli var. formosensis, an allohexaploid weed species, in..., Iwakami [/bib_ref] Under the lab conditions, primers to amplify ACCase2, ACCase4, and ACCase6 did not yield any product amplification compared to ACCase1, ACCase3, and ACCase5 while using annealing temperatures ranging from 53 to 59°C [fig_ref] Figure 1: Quantitative PCR analysis of acetyl CoA carboxylase [/fig_ref] Supplementary) in the barnyardgrass accessions investigated. Consequently, those were discarded for furthersequencing investigation. The lack of product amplification using the same primers could be related to genetic diversity. Echinochloa crus-galli var. crus-galli is native to Central America, East Asia, and Europe, whereas E. crus-galli var. formosensis is native to East Asia.It is important to note that the number of ACCase genes in barnyardgrass remains unclear since some researchers have described different numbers of copies in the ACCase gene of this plant species. For instance, ref. [bib_ref] A novel mutation Asp-2078-Glu in ACCase confers resistance to ACCase herbicides in..., Fang [/bib_ref] sequenced four copies of the ACCase gene from barnyardgrass.
In this research, analysis of the ACCase1, ACCase3, and ACCase5 between resistant and susceptible accessions did not show any nucleotide change that could lead to a target-site mutation (no further details). Similar results have been published in other barnyardgrass accessions and other ACCaseresistant weed species. A target-site mutation was shown as an unlikely resistance mechanism in cyhalofop-resistant barnyardgrass. [bib_ref] Cross-resistance of barnyardgrass [Echinochloa crus-galli (L.) P. Beauv.] to aryloxyphenoxypropionate herbicides, Hwang [/bib_ref] In another barnyardgrass population resistant to cyhalofop, no target-site mutation was reported, 32 and similar findings were reported by [bib_ref] Investigating the resistance levels and mechanisms to penoxsulam and cyhalofop-butyl in barnyardgrass..., Yang [/bib_ref] , where a target-site mutation was rejected as a possible resistance mechanism in a multiple herbicide-resistant barnyardgrass accession from Ningxia province in China. Asia minor bluegrass resistant to quizalofopp-ethyl (quizalofop) had no target-site mutation in any of the four ACCase genes reported. [bib_ref] Quizalofop-p-ethyl resistance in Polypogon fugax involves glutathione S-transferases, Chen [/bib_ref] In large crabgrass [Digitaria sanguinalis (L.) Scop.] resistant to clethodim, sethoxydim, fenoxaprop-p-ethyl (fenoxaprop), fluazifop-p-butyl, and quizalofop herbicides, no target-site mutation was found in the ACCase gene. 47
## Gene expression under metabolic inhibition
In this study, resistant and susceptible barnyardgrass plants were treated with malathion and NBD-Cl to assess ACCase, GSTs, and P450s gene expression under metabolic inhibition. Treatments included a non-treated control (T1), cyhalofop at 313 g ai ha −1 (T2); malathion at 2 kg ai ha −1 + cyhalofop at 313 g ai ha −1 (T3); and NBD-Cl at 80 g ha −1 followed by cyhalofop at 313 g ai ha −1 48 h later (T4). ACCase genespecific primers were designed to describe their role in the observed resistance of barnyardgrass to cyhalofop. Additionally, genes from the literature reported to contribute to herbicide resistance in other weed species were chosen. At 24 HAT, which was the time set up for sample collection, there were differences between resistant and susceptible accessions (P ≤ .05) [fig_ref] Figure 1: Quantitative PCR analysis of acetyl CoA carboxylase [/fig_ref]. The rationale in having selected 24 HAT for qPCR experiments is based on the production of cyhalofopacid, thus differences were found at this time period between resistant and susceptible accessions. Resistant accession produced lower amounts of cyhalofop-acid compared to the susceptible accession and no differences were detected after 24 h time period. [bib_ref] Cross-resistance of barnyardgrass [Echinochloa crus-galli (L.) P. Beauv.] to aryloxyphenoxypropionate herbicides, Hwang [/bib_ref] Gene expression of ACCase1-6 differed between resistant and susceptible accessions in treatments T2, T3, and T4 (P ≤ .05). Fold-change expression in the resistant accession was approximately 2-times higher than the gene expression found in the susceptible one [fig_ref] Figure 1: Quantitative PCR analysis of acetyl CoA carboxylase [/fig_ref]. Ref. [bib_ref] A novel mutation Asp-2078-Glu in ACCase confers resistance to ACCase herbicides in..., Fang [/bib_ref] , reported similar results to those described here while using a common ACCase primer. At 6 h after cyhalofop treatment, there were differences in the expression level of resistant and susceptible accessions. Resistant accession was found to be 3.3-times overexpressed compared to susceptible, and the same trend was also observed in fenoxaprop-and pinoxadentreated plants during the same time period. [bib_ref] A novel mutation Asp-2078-Glu in ACCase confers resistance to ACCase herbicides in..., Fang [/bib_ref] However, the specific ACCase gene causing such overexpression remains unspecified.
In a quizalofop-resistant barnyardgrass accession treated with quizalofop at 60 g ai ha −1 , the basal gene expression in the resistant accession remained similar (≈ 1-fold) to that found in the non-treated control during the time of the study (8 days). However, in the susceptible accession, the gene expression decreased after herbicide application, and by 8 d after treatment, it was practically null. [bib_ref] Determination of ACCase sensitivity and gene expression in quizalofop-ethyl-resistant and -susceptible barnyardgrass..., Huan [/bib_ref] In that research, it was concluded that gene overexpression was not involved in the resistance mechanism, but the number of ACCase genes present in such accessions was unclear. [bib_ref] Determination of ACCase sensitivity and gene expression in quizalofop-ethyl-resistant and -susceptible barnyardgrass..., Huan [/bib_ref] Gene overexpression has been reported as well in johnsongrass [Sorghum halepense (L.) Pers.] resistant to sethoxydim and quizalofop, whereby the resistant accession displayed approximately 2.5-fold higher ACCase enzyme activity compared to susceptible accession. [bib_ref] The mechanism of resistance to aryloxyphenoxypropionate and cyclohexanedione herbicides in a johnsongrass..., Bradley [/bib_ref] ACCase1 also displayed differences in T2 and T3 treatments (P ≤ .05), where the resistant accession had approximately 2-fold more expression than the susceptible used for comparison [fig_ref] Figure 1: Quantitative PCR analysis of acetyl CoA carboxylase [/fig_ref]. Similarly, ACCase3 in the resistant barnyardgrass showed 2.8-, 2.7-, and 1.8-fold more expression (P ≤ .05) than the susceptible for T2, T3, and T4, respectively [fig_ref] Figure 1: Quantitative PCR analysis of acetyl CoA carboxylase [/fig_ref]. Expression levels in ACCase2, ACCase4, and ACCase5 did not differ among treatments [fig_ref] Figure 1: Quantitative PCR analysis of acetyl CoA carboxylase [/fig_ref]. Surprisingly, in ACCase6, susceptible barnyardgrass had a 2.8-fold increase in expression than the resistant accession (P ≤ .05) in T2 treatment. Such overexpression would result from ACCase inhibition and stress caused by cyhalofop herbicide in this susceptible accession. [bib_ref] Comparative proteomic analysis of horseweed (Conyza canadensis) biotypes identifies candidate proteins for..., González-Torralva [/bib_ref] Nonetheless, it is interesting to note that the expression levels in both susceptible and resistant accessions in T3 and T4 treatments remained mostly below that of non-treated controls (T1) [fig_ref] Figure 1: Quantitative PCR analysis of acetyl CoA carboxylase [/fig_ref]. Interestingly, expression levels observed in the susceptible accession in ACCase1, ACCase3, and ACCase6 mainly in T3 and T4 treatments were lower than those found in the resistant accession. That could be explained by the presence of the metabolic inhibitors in those treatments that "enhanced" the efficacy of cyhalofop. Additionally, results suggest the involvement of metabolic resistance in the cyhalofop-resistant barnyardgrass. Expression levels found using cyhalofop alone versus in combination with the metabolic inhibitors decreased mainly in ACCase1 and ACCase3. Perhaps follow up gene expression experiments should explore longer harvest time points to gain more details about ACCase gene profiles. On the other hand, the detection of transcripts in ACCase2, ACCase4, and ACCase6 can be explained by the high sensitivity and specificity by qPCR technique along with the detection limits compared to a conventional PCR. [bib_ref] Real-time PCR detection chemistry, Navarro [/bib_ref] Expression levels using the GST1 and GST1a had no differences between accessions in the treatments evaluated [fig_ref] Figure 1: Quantitative PCR analysis of acetyl CoA carboxylase [/fig_ref]. Interestingly, GST2 was overexpressed by 2.7-, 3.0-, and 2.5-fold (P ≤ .05) in the susceptible accession in T2, T3, and T4 treatments, respectively [fig_ref] Figure 2: Quantitative PCR analysis of glutathione-S-transferase [/fig_ref]. A similar response was obtained with GST2c and GSTL3 genes. In GST2c, treatment T2 caused a higher expression (1.3-fold) level in the susceptible accession compared to that of the resistant (P ≤ .05) [fig_ref] Figure 2: Quantitative PCR analysis of glutathione-S-transferase [/fig_ref]. In GSTL3, susceptible accession displayed 3.4-, 3.3-, and 2.5-fold higher expression levels in T2, T3, and T4 treatments, respectively compared to that found in the resistant accession (P ≤ .05) [fig_ref] Figure 2: Quantitative PCR analysis of glutathione-S-transferase [/fig_ref] , GSTL3). Contrary to the results found in this research, in a quizalofop-resistant accession of Asia minor bluegrass with no herbicide treatment, authors reported that GST2c and GSTL3 were constitutively expressed in the resistant accession. The authors concluded that quizalofop resistance in Asia minor bluegrass is related to non-target-site resistance mechanisms and that GST2c and GSTL3 are involved in resistance. [bib_ref] Quizalofop-p-ethyl resistance in Polypogon fugax involves glutathione S-transferases, Chen [/bib_ref] Regarding the P450s genes analyzed, gene expression using the CYP81A68 did not differ between susceptible and resistant accessions in the treatments evaluated [fig_ref] Figure 1: Quantitative PCR analysis of acetyl CoA carboxylase [/fig_ref]. Nonetheless, in other barnyardgrass accessions resistant to ALS and ACCase-inhibiting-herbicides, it has been reported that the overexpression of CYP81A68 confers metabolic resistance to penoxsulam (an ALS-inhibiting herbicide) but also can confer resistance to cyhalofop and metamifop (ACCase-inhibiting herbicides). [bib_ref] CYP81A68 confers metabolic resistance to ALS and ACCase-inhibiting herbicides and its epigenetic..., Pan [/bib_ref] Using the CYP81A21, the gene expression in the resistant accession showed a 2.5-fold increase in T2 treatment compared to that found in the susceptible accession (P ≤ .05) [fig_ref] Figure 1: Quantitative PCR analysis of acetyl CoA carboxylase [/fig_ref]. This result suggests that CYP81A21, a P450 gene previously reported in contributing to resistance in late watergrass, is also contributing to resistance in those barnyardgrass accessions. Overexpression of CYP81A21 gene has resulted in an increase in cross-resistance to ALS and ACCase-inhibiting herbicides. [bib_ref] CYP 81A P450s are involved in concomitant cross-resistance to acetolactate synthase and..., Iwakami [/bib_ref] The accessions used in this research were both susceptible to the ALS-inhibiting herbicide imazethapyr-ammonium. [bib_ref] Cross-resistance of barnyardgrass [Echinochloa crus-galli (L.) P. Beauv.] to aryloxyphenoxypropionate herbicides, Hwang [/bib_ref] These results suggest that genes involved in metabolic resistance in one plant species not precisely should be involved in another one. For instance, GST2c and GSTL3 were discarded in this research as contributing to ACCase resistance. However, in Asia minor bluegrass, both genes were involved in the resistance to quizalofop. [bib_ref] Quizalofop-p-ethyl resistance in Polypogon fugax involves glutathione S-transferases, Chen [/bib_ref] Additionally, CYP81A68 was not associated to cyhalofop resistance in this research but it was in other barnyardgrass accessions. [bib_ref] CYP81A68 confers metabolic resistance to ALS and ACCase-inhibiting herbicides and its epigenetic..., Pan [/bib_ref] Conversely, CYP81A21 was linked to cyhalofop resistance in this study but also in late watergrass. [bib_ref] CYP 81A P450s are involved in concomitant cross-resistance to acetolactate synthase and..., Iwakami [/bib_ref] Metabolism of ACCase-inhibiting herbicides has been reported as the non-target-site resistance mechanism in several resistant weed species, including Asia minor bluegrass resistant to quizalofop, [bib_ref] Quizalofop-p-ethyl resistance in Polypogon fugax involves glutathione S-transferases, Chen [/bib_ref] [bib_ref] Target-site and metabolic resistance mechanisms to penoxsulam in barnyardgrass (Echinochloa crus-galli (L.)..., Fang [/bib_ref] [bib_ref] CYP81A68 confers metabolic resistance to ALS and ACCase-inhibiting herbicides and its epigenetic..., Pan [/bib_ref] and rigid ryegrass resistant to multiple herbicide sites of action. [bib_ref] Cytochrome P450 CYP81A10v7 in Lolium rigidum confers metabolic resistance to herbicides across..., Han [/bib_ref] However, in ALS-resistant accessions of barnyardgrass and late watergrass, both target and non-target-site resistance mechanisms were reported. Resistance was attributed to an Ala122Asn and Trp574Leu target-site mutations in the ALS1 gene and overexpression of the ALS1 and no ALS2 or ALS3 copies. [bib_ref] Target-site mutations and expression of ALS gene copies vary according to Echinochloa..., Panozzo [/bib_ref] Gene overexpression has been correlated to herbicide resistance in different herbicide-resistant weed species, which can result in an increased amount of enzyme produced by the plant to avoid herbicide damage. Thus, gene overexpression can be due to regulatory changes that increase gene transcription and/ or a difference in the genomic copy number of the target gene, which also causes an increased gene transcription. [bib_ref] Mechanisms of evolved herbicide resistance, Gaines [/bib_ref] Results obtained in this research support the idea that gene overexpression of specific ACCase genes contributes to cyhalofop resistance in those barnyardgrass accessions. Rapid detection of herbicideresistant weed species is crucial to implement the best possible weed management strategies.
## Accase gene copy number
Results displayed a low number of copies in all gene-specific ACCase compared to CCR and PPAN genes. Relative to CCR, the resistant accession showed a higher copy number (P ≤ .05) in the ACCase4 gene. However, biologically, this difference could not be related to ACCase resistance. Using PPAN, the relative gene copy number in susceptible and resistant accessions among ACCase genes, had no differences (P ≥ .05) [fig_ref] Figure 4: Relative gene copy number estimation of acetyl CoA carboxylase [/fig_ref].
Gene amplification of the target-site gene has conferred herbicide resistance in a diverse range of herbicide-resistant weed accessions. [bib_ref] Mechanisms of evolved herbicide resistance, Gaines [/bib_ref] Gene amplification of the EPSPS gene (target-site of glyphosate herbicide) in a diverse range of CCR and PPAN were used as reference genes to measure the relative gene copy number. Bars ± standard deviation of the mean. A significant difference between susceptible and resistant barnyardgrass accession on each gene is indicated by an asterisk (P ≤ .05).
glyphosate-resistant weed species has been reported as a common target-site resistant mechanism. [bib_ref] Mechanisms of evolved herbicide resistance, Gaines [/bib_ref] Other than glyphosate, gene amplification has been found in more herbicideresistant accessions with different herbicide sites of action. For instance, in Palmer amaranth (Amaranthus palmeri S. Watson) resistant to glufosinate, the resistant mechanism deployed was attributed to the amplification of the chloroplastic glutamine synthetase (GS) gene. [bib_ref] Unraveling the mechanism of resistance in a glufosinate-resistant Palmer amaranth (Amaranthus palmeri)..., Carvalho-Moore [/bib_ref] In large crabgrass accessions resistant to ACCase-inhibiting herbicides, gene amplification of the target ACCase gene was also reported. [bib_ref] Acetyl-CoA carboxylase overexpression in herbicide-resistant large crabgrass (Digitaria sanguinalis), Laforest [/bib_ref] In both cases, overexpression of the target-site gene was due to gene amplification.
In summary, this research focused on deciphering the role of gene expression and gene amplification of specific ACCase genes on cyhalofop resistance in barnyardgrass. Previously, the presence of a non-target-site resistance mechanism was documented, where a difference in cyhalofop-acid formation was detected between resistant and susceptible accessions at 24 h after cyhalofop treatment. [bib_ref] Cross-resistance of barnyardgrass [Echinochloa crus-galli (L.) P. Beauv.] to aryloxyphenoxypropionate herbicides, Hwang [/bib_ref] In this research, it has been demonstrated that ACCase1 and ACCase3 genes contribute to cyhalofop resistance. Furthermore, the overexpression observed in CYP81A21 suggests the involvement of this gene in cyhalofop resistance. Additionally, results indicate that gene amplification of ACCase genes is not contributing to resistance in this cyhalofop-resistant barnyardgrass accession.
[fig] Figure 1: Quantitative PCR analysis of acetyl CoA carboxylase (ACCase) genes in susceptible (S) and resistant (R) barnyardgrass accessions under different treatments. Bars ± standard deviation of the mean. A significant difference between susceptible and resistant barnyardgrass accession on each treatment is indicated by an asterisk (P ≤ .05). [/fig]
[fig] Figure 2: Quantitative PCR analysis of glutathione-S-transferase (GSTs) genes in susceptible (S) and resistant (R) barnyardgrass accessions under different treatments. Bars ± standard deviation of the mean. A significant difference between susceptible and resistant barnyardgrass accession on each treatment is indicated by an asterisk (P ≤ .05). [/fig]
[fig] Figure 3: Quantitative PCR analysis of CYP P450 genes in susceptible (S) and resistant (R) barnyardgrass accessions under different treatments. Bars ± standard deviation of the mean. A significant difference between susceptible and resistant barnyardgrass accession on each treatment is indicated by an asterisk (P ≤ .05). [/fig]
[fig] Figure 4: Relative gene copy number estimation of acetyl CoA carboxylase (ACCase) genes in susceptible (S) and resistant (R) barnyardgrass accessions. [/fig]
[table] Table 1: Primer sets used to partially amplify the ACCase gene copies in cyhalofop-resistant and -susceptible barnyardgrass accessions.32 [/table]
[table] Table 2: µL total volume ran on 96-well plates. The amplification protocol consisted of 30 sec at 98°C followed by 40 cycles of 10 sec at 98°C and 30 sec at 61°C and finalized with a standard melt curve analysis to corroborate correct amplification. No-Table 2. Primer sequences of genes used to measure either transcript abundance or gene copy number in resistant and susceptible barnyardgrass accessions by qPCR. ACCase, acetyl CoA carboxylase; CYP, cytochrome P450 monooxygenase; GST, glutathione-S-transferase; GSTL, glutathione S-transferase lambda; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; CCR, cinnamoyl-CoA reductase; PPAN, peter Panlike. b F, forward primer; R, reverse primer. c Primer efficiency in gene copy number experiments was determined using E = [10 (−1/slope) -1] × 100; for gene expression experiments efficiency was obtained using LinRegPCR software. d Common primer to ACCase1-ACCase6. [/table]
|
Advances in the Treatment of Mycoses in Pediatric Patients
The main indications for antifungal drug administration in pediatrics are reviewed as well as an update of the data of antifungal agents and antifungal policies performed. Specifically, antifungal therapy in three main areas is updated as follows: (a) Prophylaxis of premature neonates against invasive candidiasis; (b) management of candidemia and meningoencephalitis in neonates; and (c) prophylaxis, empiric therapy, and targeted antifungal therapy in children with primary or secondary immunodeficiencies. Fluconazole remains the most frequent antifungal prophylactic agent given to high-risk neonates and children. However, the emergence of fluconazole resistance, particularly in non-albicans Candida species, should be considered during preventive or empiric therapy. In very-low birth-weight neonates, although fluconazole is used as antifungal prophylaxis in neonatal intensive care units (NICU's) with relatively high incidence of invasive candidiasis (IC), its role is under continuous debate. Amphotericin B, primarily in its liposomal formulation, remains the mainstay of therapy for treating neonatal and pediatric yeast and mold infections. Voriconazole is indicated for mold infections except for mucormycosis in children >2 years. Newer triazoles-such as posaconazole and isavuconazole-as well as echinocandins, are either licensed or under study for first-line or salvage therapy, whereas combination therapy is kept for refractory cases.
# Introduction
Invasive fungal infections (IFIs) are important causes of excessive morbidity and mortality in pediatrics. Candida spp. and Aspergillus spp. are the most common fungi causing IFIs in neonates, infants, children, and adolescents. Mucorales, Cryptococcus neoformans, and other fungi are less frequent causes [bib_ref] Epidemiology of Invasive Fungal Disease in Children, Pana [/bib_ref]. The incidence of IFIs in various countries and hospitals depends on many factors; however, the degree of exposure to fungal elements and the predisposition of the host to develop IFI are the two main determinants of the fungal epidemiology in particular patient populations.
There are various factors that can increase the risk for developing an IFI in pediatrics. Premature birth and very-low-birth weight (VLBW) have been recognized as an important factor for increased IFIs, especially for invasive candidiasis (IC) in neonatal age. For older infants and children, primary or especially secondary immunodeficiencies are main underlying conditions predisposing to IFIs. Secondary immunodeficiencies are those resulted from cancer-related chemotherapy, stem cell or organ transplantation, and administration of various immunosuppressive agents. Other risk factors are the existence of foreign bodies, and most frequently central catheters predisposing to biofilm development, and overuse of broad-spectrum antibiotics, resulting in antibiotic-induced modified microbiota.
Specific characteristics of neonates and young children are extremely relevant to antifungal drug therapy-such as pharmacokinetics, drug metabolism, age-dependent adverse effects, route of drug administration, and limited clinical data. The impact of these pediatric factors has been recognized as a major issue during development of pediatric guidelines of IFI management [bib_ref] Therapeutic strategies for invasive fungal infections in neonatal and pediatric patients: An..., Pana [/bib_ref].
In this article, the main indications for antifungal drug administration in pediatrics will be reviewed as well as an update of the data of antifungal agents as well as antifungal policies will be performed. Specifically, antifungal therapy in three main areas will be updated as follows:
(a) Prophylaxis of premature neonates against invasive candidiasis; (b) management of candidemia and meningoencephalitis in neonates; and (c) prophylaxis, empiric therapy, and targeted antifungal therapy in children with primary or secondary immunodeficiencies.
Doses and indications of antifungal agents in pediatrics are summarized in [fig_ref] Table 1: Antifungal drugs in pediatric patients [/fig_ref]. - Treatment of IFIs in NICU [bib_ref] ESCMID* guideline for the diagnosis and management of Candida diseases 2012: Prevention..., Hope [/bib_ref] [bib_ref] Clinical Practice Guideline for the Management of Candidiasis: 2016 Update by the..., Pappas [/bib_ref] [bib_ref] Amphotericin B in Neonates: Deoxycholate or Lipid Formulation-What is the Right Choice?, Turkova [/bib_ref] - CNS and disseminated cryptococcal disease (combined with 5FC) [bib_ref] Clinical practice guidelines for the management of cryptococcal disease: 2010 update by..., Perfect [/bib_ref] [bib_ref] Cryptococcosis diagnosis and treatment: What do we know now, Perfect [/bib_ref] LAMB IV - 2.5-7 or 3-5 mg/kg/d - Treatment of IC in neonates [bib_ref] ESCMID* guideline for the diagnosis and management of Candida diseases 2012: Prevention..., Hope [/bib_ref] [bib_ref] Clinical Practice Guideline for the Management of Candidiasis: 2016 Update by the..., Pappas [/bib_ref] - No dosage established - Treatment of HCME in neonates [bib_ref] Liposomal amphotericin B treatment for neonatal fungal infections, Scarcella [/bib_ref] [bib_ref] Comprehensive drug utilization review in neonates: Liposomal amphotericin B, Silver [/bib_ref] - 1-3 mg/kg/d - Empiric fever-driven therapy in hemato-oncological patients at high risk for invasive fungal disease with neutropenia and refractory or new fever of at least 4 days, despite broad-spectrum antibacterial therapy [bib_ref] Guidelines for diagnosis, prevention, and treatment of invasive fungal diseases in paediatric..., Groll [/bib_ref] [bib_ref] Guideline for the Management of Fever and Neutropenia in Children With Cancer..., Lehrnbecher [/bib_ref] - 3 mg/kg/d - Empiric treatment in patients with refractory or new fever episode in the PICU, despite broad-spectrum empirical antibacterial therapy, who are at high risk for Candida infection with moderate-to-severe disease, hemodynamic instability, recent azole exposure, or at high risk for C. glabrata or C. krusei infections [bib_ref] Clinical Practice Guideline for the Management of Candidiasis: 2016 Update by the..., Pappas [/bib_ref] [bib_ref] Antifungal therapy in children: An update, Cecinati [/bib_ref] - 3-5 mg/kg/d - First-line treatment of IFIs (IA, IC) in pediatrics [bib_ref] ESCMID* guideline for the diagnosis and management of Candida diseases 2012: Prevention..., Hope [/bib_ref] [bib_ref] Clinical Practice Guideline for the Management of Candidiasis: 2016 Update by the..., Pappas [/bib_ref] [bib_ref] Guidelines for diagnosis, prevention, and treatment of invasive fungal diseases in paediatric..., Groll [/bib_ref] - ≥5 mg/kg - First-line treatment of mucormycosis [bib_ref] Update on antifungal agents for paediatric patients, Groll [/bib_ref] [bib_ref] Rare fungal infections in children: An updated review of the literature, Pana [/bib_ref] - 5 mg/kg/d - Candida meningitis and endocarditis (combined with 5FC) [bib_ref] Evidence of excessive concentrations of 5-flucytosine in children aged below 12 years:..., Soltani [/bib_ref] - 5 mg/kg/d - Second-line treatment of CNS and disseminated cryptococcal disease (AMB-intolerant patients) [bib_ref] Clinical practice guidelines for the management of cryptococcal disease: 2010 update by..., Perfect [/bib_ref] [bib_ref] Cryptococcosis diagnosis and treatment: What do we know now, Perfect [/bib_ref] ABLC IV
- No dosage established - Treatment of IC in neonates [bib_ref] Antifungal agents in neonates: Issues and recommendations, Almirante [/bib_ref] - 5 mg/kg/d - Treatment of IFIs (IA, IC) in pediatrics [bib_ref] Update on antifungal agents for paediatric patients, Groll [/bib_ref] - 5 mg/kg/d - Treatment for: Blastomycosis, coccidioidomycosis, histoplasmosis, endemic mycoses [bib_ref] Update on antifungal agents for paediatric patients, Groll [/bib_ref] [bib_ref] Rare fungal infections in children: An updated review of the literature, Pana [/bib_ref] - 5 mg/kg/d - Second-line treatment of CNS and disseminated cryptococcal disease (AMB-intolerant patients) [bib_ref] Antifungal prophylaxis in pediatric hematology/oncology: New choices & new data. Pediatr, Dvorak [/bib_ref] [bib_ref] A controlled trial of fluconazole to prevent fungal infections in patients undergoing..., Goodman [/bib_ref] [bib_ref] Efficacy and safety of fluconazole prophylaxis for fungal infections after marrow transplantation-A..., Slavin [/bib_ref] [bib_ref] Clinical practice guidelines for the management of candidiasis: 2009 update by the..., Pappas [/bib_ref] [bib_ref] Guideline for primary antifungal prophylaxis for pediatric patients with cancer or hematopoietic..., Science [/bib_ref] [bib_ref] Safety and tolerability of fluconazole in children, Novelli [/bib_ref] - 6-12 mg/kg/d - Anti-Candida prophylaxis in patients with primary immunodeficiencies at high risk for IFIs or presenting as chronic fungal infections [bib_ref] Prevention of infections during primary immunodeficiency, Aguilar [/bib_ref] [bib_ref] Invasive fungal infections in congenital immunodeficiencies, Antachopoulos [/bib_ref] [bib_ref] Diagnosis and management of Aspergillus diseases: Executive summary of the 2017 ESCMID-ECMM-ERS..., Ullmann [/bib_ref] [bib_ref] International Pediatric Lung Transplant Collaborative. Antifungal prophylaxis in pediatric lung transplantation: An..., Mead [/bib_ref] IV, PO
- 10-12 mg/kg/d - Maintenance treatment of CNS and disseminated cryptococcal disease [bib_ref] Clinical practice guidelines for the management of cryptococcal disease: 2010 update by..., Perfect [/bib_ref] [bib_ref] Cryptococcosis diagnosis and treatment: What do we know now, Perfect [/bib_ref] - 6-12 mg/kg - Treatment of Cryptococcal pneumonia [bib_ref] Clinical practice guidelines for the management of cryptococcal disease: 2010 update by..., Perfect [/bib_ref] [bib_ref] Cryptococcosis diagnosis and treatment: What do we know now, Perfect [/bib_ref] - 12 mg/kg/d - Treatment of IC provided that: It is caused by fluconazole-susceptible organisms, the patient is in stable condition, and has not received prior azole therapy [bib_ref] ESCMID* guideline for the diagnosis and management of Candida diseases 2012: Prevention..., Hope [/bib_ref] [bib_ref] Guidelines for diagnosis, prevention, and treatment of invasive fungal diseases in paediatric..., Groll [/bib_ref] [bib_ref] Antifungal therapy in children: An update, Cecinati [/bib_ref] ITC PO
[formula] - 200 mg b.i.d. [/formula]
- Antifungal prophylaxis in high-risk, immunocompromised pediatric patients (anti-mold activity) [bib_ref] Fluconazole versus itraconazole for antifungal prophylaxis in neutropenic patients with haematological malignancies:..., Vardakas [/bib_ref] [bib_ref] Itraconazole versus fluconazole for prevention of fungal infections in patients receiving allogeneic..., Marr [/bib_ref] - 200 mg b.i.d.
- Anti-Aspergillus prophylaxis in patients with primary immunodeficiencies at high risk for IFIs or presenting as chronic fungal infections [bib_ref] Prevention of infections during primary immunodeficiency, Aguilar [/bib_ref] [bib_ref] Invasive fungal infections in congenital immunodeficiencies, Antachopoulos [/bib_ref] [bib_ref] Diagnosis and management of Aspergillus diseases: Executive summary of the 2017 ESCMID-ECMM-ERS..., Ullmann [/bib_ref] [bib_ref] International Pediatric Lung Transplant Collaborative. Antifungal prophylaxis in pediatric lung transplantation: An..., Mead [/bib_ref] VRC IV PO
[formula] - 8 mg/kg b.i.d. - 9 mg/kg b.i.d. [/formula]
- Antifungal prophylaxis in pediatric patients with allogeneic HSCT (anti-mold activity) [bib_ref] Guidelines for diagnosis, prevention, and treatment of invasive fungal diseases in paediatric..., Groll [/bib_ref] [bib_ref] Randomized, double-blind trial of fluconazole versus voriconazole for prevention of invasive fungal..., Wingard [/bib_ref] IV, PO - Same as above - Anti-Aspergillus prophylaxis in patients with primary immunodeficiencies at high risk for IFIs or presenting as chronic fungal infections - 600 mg/d, (given in 3 doses)
- Antifungal prophylaxis for hematological/oncological patients with acute myeloid leukemia, myelodysplastic syndromes, GVHD or in patients undergoing HSCT, in whom a long neutropenic period due to chemotherapy is expected [bib_ref] Antifungal therapy in children: An update, Cecinati [/bib_ref] [bib_ref] Update on antifungal agents for paediatric patients, Groll [/bib_ref] [bib_ref] Retrospective survey on the off-label use of posaconazole in pediatric hematology patients, Cesaro [/bib_ref] - 600 mg/d, (given in 3 doses) - Antifungal prophylaxis in primary immunodeficiencies (including CGD) [bib_ref] Therapeutic drug monitoring of posaconazole oral suspension in paediatric patients younger than..., Jancel [/bib_ref] [bib_ref] Spriet, I. Pharmacokinetics of Posaconazole Oral Suspension in Children Dosed According to..., Vanstraelen [/bib_ref] - 800 mg/d in 2-4 divided doses - Treatment for: Scedosporiosis, fusariosis [bib_ref] Guidelines for diagnosis, prevention, and treatment of invasive fungal diseases in paediatric..., Groll [/bib_ref] - Second line treatment for mucormycosis [bib_ref] Guidelines for diagnosis, prevention, and treatment of invasive fungal diseases in paediatric..., Groll [/bib_ref] PO (tabl.)
- Children ≥13 years old: 300 mg/d, q.d. Children <13 years: dose not established
- Antifungal prophylaxis in HSCT pediatric patients [bib_ref] Efficacy, safety and feasibility of antifungal prophylaxis with posaconazole tablet in paediatric..., Doring [/bib_ref] - Antifungal prophylaxis in primary immunodeficiencies (including CGD) [bib_ref] Posaconazole in immunocompromised pediatric patients, Vicenzi [/bib_ref] - - Treatment of IC in neonates and infants <3 months (limited data) [bib_ref] Caspofungin therapy in immunocompromised children and neonates, Somer [/bib_ref] [bib_ref] Successful caspofungin treatment of persistent candidemia in extreme prematurity at 23 and..., Jeon [/bib_ref] [bib_ref] Pharmacokinetics and safety of caspofungin in neonates and infants less than 3..., Saez-Llorens [/bib_ref] - 50 mg/m 2 /d - Antifungal prophylaxis in HSCT pediatric patients [bib_ref] Open-label, randomized comparison of itraconazole versus caspofungin for prophylaxis in patients with..., Mattiuzzi [/bib_ref] [bib_ref] Caspofungin as antifungal prophylaxis in pediatric patients undergoing allogeneic hematopoietic stem cell..., Doring [/bib_ref] [bib_ref] Comparison of Efficacy and Safety of Caspofungin Versus Micafungin in Pediatric Allogeneic..., Maximova [/bib_ref] - 70 mg/m 2 /d loading dose, followed by 50 mg/m 2 /d - Empiric fever-driven therapy in hemato-oncological patients at high risk for invasive fungal disease with neutropenia and refractory or new fever of at least 4 days, despite broad-spectrum antibacterial therapy [bib_ref] Guidelines for diagnosis, prevention, and treatment of invasive fungal diseases in paediatric..., Groll [/bib_ref] [bib_ref] Guideline for the Management of Fever and Neutropenia in Children With Cancer..., Lehrnbecher [/bib_ref] -
## Prophylaxis of preterm babies against candidiasis
Measures to prevent colonization of infants, especially of VLBW neonates, by Candida spp. are the cornerstone of prevention of IFIs among them [bib_ref] ESCMID* guideline for the diagnosis and management of Candida diseases 2012: Prevention..., Hope [/bib_ref] [bib_ref] Invasive fungal infections in newborns and current management strategies, Hundalani [/bib_ref] [bib_ref] Neonatal candidiasis: Prophylaxis and treatment, Tiffany [/bib_ref]. Avoidance of horizontal transmission by rigorous infection control measures, rational use of broad-spectrum antibacterial agents (especially third-generation cephalosporins and carbapenems), and proper handling of central venous catheters appear to be extremely important [bib_ref] ESCMID* guideline for the diagnosis and management of Candida diseases 2012: Prevention..., Hope [/bib_ref] [bib_ref] Challenging issues in neonatal candidiasis, Kaufman [/bib_ref].
Early administration of probiotics is associated with prevention of Candida colonization in neonates; however, their efficacy in reducing IFIs and mortality, their safety, risk-benefit potential, optimal dosage, and duration of administration is still unclear [bib_ref] Evaluation of efficacy of probiotics in prevention of candida colonization in a..., Kumar [/bib_ref] [bib_ref] Role of probiotics in prevention of Candida infection in critically ill children, Kumar [/bib_ref] [bib_ref] Bovine lactoferrin prevents invasive fungal infections in very low birth weight infants:..., Manzoni [/bib_ref] [bib_ref] Prophylactic Saccharomyces boulardii versus nystatin for the prevention of fungal colonization and..., Demirel [/bib_ref]. A subsequent meta-analysis did not show any significant impact of probiotic administration in reducing incidence of late onset sepsis [bib_ref] Prebiotic supplementation in preterm neonates: Updated systematic review and meta-analysis of randomised..., Srinivasjois [/bib_ref]. Thus, the impact of probiotics in reducing IFIs in neonates needs further study.
Several neonatal units around the world use fluconazole as means of IFI prevention depending on their IFI incidence [bib_ref] Fluconazole prophylaxis against fungal colonization and infection in preterm infants, Kaufman [/bib_ref] [bib_ref] Antifungal prophylaxis in neonates, Manzoni [/bib_ref] [bib_ref] Fluconazole prophylaxis in extremely low birth weight neonates reduces invasive candidiasis mortality..., Healy [/bib_ref] [bib_ref] Effect of fluconazole prophylaxis on candidiasis and mortality in premature infants: A..., Benjamin [/bib_ref]. A more recent network meta-analysis, including 11 randomized controlled studies of fluconazole prophylaxis, confirmed that fluconazole at 3, 4, or 6 mg/kg twice weekly was superior to placebo preventing invasive candidiasis (IC) and IC-related mortality in preterm neonates, but failed to find any difference between these three dosage regimens [bib_ref] Fluconazole Doses Used for Prophylaxis of Invasive Fungal Infection in Neonatal Intensive..., Leonart [/bib_ref]. Although the study could not demonstrate any effect on long-term toxicity or the development of resistance due to the design of the studies, it recommended the smallest dose of 3 mg/kg twice weekly as sufficient for prophylaxis as well as being better tolerated, minimizing exposure, and reducing cost. Ericson et al. collected patient-level data from four US-based randomized studies and conducted an aggregated analysis of these data. They found that fluconazole prophylaxis significantly reduced IC and Candida colonization in preterm infants, whereas it had no impact on the fluconazole resistance rate during the period of the study [bib_ref] Fluconazole Prophylaxis Study Team on behalf of the Best Pharmaceuticals for Children..., Ericson [/bib_ref]. Nevertheless, neonatal intensive care units (NICUs) should avoid overuse of prophylactic fluconazole and try to use infection control bundles and antibiotic stewardship to avoid increased use of fluconazole with subsequent increases of Candida minimal inhibitory concentrations to fluconazole. Until more definitive results on drug exposure exist, fluconazole 3-6 mg/kg twice weekly for 42 days is recommended for <1000 gr preterm neonates in NICUs with a relatively high incidence of IC (>10%) [bib_ref] Clinical Practice Guideline for the Management of Candidiasis: 2016 Update by the..., Pappas [/bib_ref]. The rational of this twice weekly regimen is to reduce drug exposure (leading to adverse effects, especially long-term neurodevelopmental consequences) as well as the development of resistant fungal isolates showing, nevertheless, sufficient efficacy [bib_ref] Fluconazole Doses Used for Prophylaxis of Invasive Fungal Infection in Neonatal Intensive..., Leonart [/bib_ref]. In addition, one study found that the twice-weekly regimen at a dose of 3 mg/kg twice weekly was not significantly associated with neurodevelopmental sequelae after 8-10 years or impaired quality of life [bib_ref] Fluconazole prophylaxis in extremely low birth weight infants and neurodevelopmental outcomes and..., Kaufman [/bib_ref]. As the long-term complications of neonatal period are frequently multifactorial, further studies are needed to identify the impact of antifungal prophylaxis.
Nystatin has been also proposed as an alternative antifungal prophylactic regimen in neonates, because it is safe, well tolerated, effective, and cheap, although data are limited and especially for neonates weighing less than 750 g. It may be used in an alternative antifungal prophylaxis setting with fluconazole shortages or resistance [bib_ref] Clinical Practice Guideline for the Management of Candidiasis: 2016 Update by the..., Pappas [/bib_ref]. In a non-inferiority randomized study, prophylactic use of nystatin or fluconazole reduced both the incidence of colonisation and IFIs in VLBW neonates [bib_ref] Randomised controlled trial of prophylactic fluconazole versus nystatin for the prevention of..., Aydemir [/bib_ref]. Of interest, in another study of fluconazole or nystatin oral prophylactic use in VLBW infants (birth weight <1500 g), the nystatin group showed an unexplained proportion of deaths [bib_ref] Comparison of fluconazole and nystatin oral suspensions for prophylaxis of systemic fungal..., Violaris [/bib_ref]. The role of oral non-absorbable antifungals to reduce Candida gut colonization and indirect risk of IFIs in the neonates has not been adequately studied [bib_ref] Prophylactic oral/topical non-absorbed antifungal agents to prevent invasive fungal infection in very..., Austin [/bib_ref].
## Management of invasive candidiasis in neonates
Significant differences in IFI pathophysiology and antifungal drug pharmacokinetics between premature neonates and older patients emphasize the need to further evaluate the effectiveness and safety of different antifungal drugs and dosing in the neonatal period [bib_ref] Antifungal agents and therapy for infants and children with invasive fungal infections:..., Lestner [/bib_ref]. In a recent multicenter population-based study of antifungal treatment in premature neonates in Germany, 5.4% of VLBW and 10.7% of extremely-low-birth weight (ELBW) neonates received any empirical or targeted antifungal therapy. Among them, 95.5% of them received empirical instead of targeted therapy. Fluconazole was the most frequently used agent followed by liposomal amphotericin B and caspofungin [bib_ref] Antifungal Treatment and Outcome in Very-Low-Birth-Weight-Infants-A Population-Based Observational Study of the German..., Fortmann [/bib_ref].
Prompt initiation of empiric antifungal treatment (before knowing the positive culture), especially in high-risk neonates, is a prerequisite to efficiently combat candidemia, leading to increased survival without neurodevelopmental impairment [bib_ref] Neonatal candidiasis: Diagnosis, prevention, and treatment, Greenberg [/bib_ref] [bib_ref] Empiric antifungal therapy and outcomes in extremely low birth weight infants with..., Greenberg [/bib_ref]. The empiric treatment should be based on epidemiology of the specific NICU; any premature infant with clinical or microbiological evidence of IC should be treated as having disseminated disease and hematogenous Candida meningoencephalitis (HCME) [bib_ref] ESCMID* guideline for the diagnosis and management of Candida diseases 2012: Prevention..., Hope [/bib_ref]. In all neonates with suspected IC, a lumbar puncture and a retinal examination should be performed and in case of persistent positive blood cultures, ultrasonography of head, heart, and kidneys for end-organ dissemination is strongly recommended [bib_ref] Clinical Practice Guideline for the Management of Candidiasis: 2016 Update by the..., Pappas [/bib_ref]. Removal of a central venous catheter, if possible, is very important for IC treatment to diminish the niche of a persistent Candida biofilm [bib_ref] Clinical Practice Guideline for the Management of Candidiasis: 2016 Update by the..., Pappas [/bib_ref]. No consensus on the duration of antifungal therapy in the neonatal period has been established; however, most NICUs treat IC for a minimum of 14 days after negative cultures [bib_ref] Clinical Practice Guideline for the Management of Candidiasis: 2016 Update by the..., Pappas [/bib_ref]. The duration of therapy for end-organ infections-such as endocarditis and renal infection-is even less clear, but a duration of six weeks or longer is recommended [bib_ref] Candida endocarditis in neonates: Report of five cases and review of the..., Levy [/bib_ref].
Non-culture assays have been studied for the early diagnosis of IC in neonates and thus initiation of preemptive therapy with the goal to improve the outcome. (1,3)-β-D-glucan (BDG), a reliable biomarker of most clinically relevant fungi in adults, has been studied very little in neonates with variable results [bib_ref] -3)-β-D-glucan levels in candidiasis infections in the critically ill neonate, Goudjil [/bib_ref]. In a more recently published study of premature neonates comparing those with proven or probable invasive yeast infections with controls, BDG was shown to have promising results [bib_ref] Evaluation of the (1,3)-β-D-glucan assay for the diagnosis of neonatal invasive yeast..., Cornu [/bib_ref]. The authors suggested a higher threshold, i.e., 106 or 125 pg/mL, may be used, but interpretation of BDG results in neonates should be made with caution, taking into consideration all limitations. A multicenter study evaluated the role of polymerase chain reaction (PCR) and BDG in the early diagnosis of IC in premature neonates in comparison to blood culture [bib_ref] Performance of a Quantitative PCR-Based Assay and β-D-Glucan Detection for Diagnosis of..., Ramos [/bib_ref]. With a cutoff value of either >80 pg/mL or >120 pg/mL used, the authors found that Candida PCR and serum BDG are useful as complementary diagnostic techniques to detect IC in premature neonates.
T2 magnetic resonance assay is a very promising assay that may give the result of positivity and Candida spp. identification much faster (three to four hours) than culture (which may take up to four to five days) [bib_ref] T2Candida Provides Rapid and Accurate Species Identification in Pediatric Cases of Candidemia, Hamula [/bib_ref]. The use of valid diagnostics with good sensitivity and specificity may make the treatment more target-oriented and much earlier. No published data of T2 magnetic resonance assay exist with IC in neonates.
Deoxycholate amphotericin B (DAMB) continues to be useful in the treatment of IFIs in NICU. While its pharmacokinetics in neonates are considerably variable [bib_ref] Pharmacokinetics, outcome of treatment, and toxic effects of amphotericin B and 5-fluorocytosine..., Baley [/bib_ref] , the dosage regimen recommended is 1 mg/kg/d [bib_ref] ESCMID* guideline for the diagnosis and management of Candida diseases 2012: Prevention..., Hope [/bib_ref] [bib_ref] Clinical Practice Guideline for the Management of Candidiasis: 2016 Update by the..., Pappas [/bib_ref] [bib_ref] Amphotericin B in Neonates: Deoxycholate or Lipid Formulation-What is the Right Choice?, Turkova [/bib_ref]. However, its use is limited by its adverse effects of nephrotoxicity and hypokalemia [bib_ref] Amphotericin B in Neonates: Deoxycholate or Lipid Formulation-What is the Right Choice?, Turkova [/bib_ref]. Therefore, frequent monitoring of electrolytes and renal function is recommended during therapy.
Pharmacokinetic data are available on the liposomal and lipid complex preparations of amphotericin B and they support their use in neonates; however, the optimal dosage and duration of therapy is difficult to establish [bib_ref] Antifungal agents in neonates: Issues and recommendations, Almirante [/bib_ref]. Amphotericin B lipid formulations are administered in increased daily dosages (up to 10-fold), and are characterized by high tissue concentrations in macrophage-rich organs (lungs, liver, spleen), a decrease in infusion-related adverse effects, and a marked decrease in renal toxicity. Dosing of liposomal amphotericin B (LAMB) in neonatal IC ranges from 2.5-7 or 3-5 mg/kg/d according to ESCMID or IDSA guidelines, respectively [bib_ref] ESCMID* guideline for the diagnosis and management of Candida diseases 2012: Prevention..., Hope [/bib_ref] [bib_ref] Clinical Practice Guideline for the Management of Candidiasis: 2016 Update by the..., Pappas [/bib_ref]. When comparing the efficacy and safety of DAMB and lipid amphotericin B (AMB) formulations (liposomal and lipid complex) in neonatal IFIs, there are no significant differences noticed, although the number of studies is limited [bib_ref] Neonatal invasive candidiasis: A prospective multicenter study of 118 cases, Lopez Sastre [/bib_ref] [bib_ref] The efficacy of two different lipid-based amphotericin B in neonatal Candida septicemia, Cetin [/bib_ref] [bib_ref] A comparison of AmBisome to amphotericin B for treatment of systemic candidiasis..., Jeon [/bib_ref]. While there is no specific clinical information for the optimal regimen for HCME, LAMB penetrates the central nervous system (CNS) in a preclinical model of HCME and has antifungal activity in the brain [bib_ref] Liposomal amphotericin B treatment for neonatal fungal infections, Scarcella [/bib_ref] [bib_ref] Comprehensive drug utilization review in neonates: Liposomal amphotericin B, Silver [/bib_ref].
Fluconazole has potent in vitro activity against almost all Candida spp., fair pharmacokinetic properties with good penetration to the CNS, and has been used in neonates with IC at 12 mg/kg/d [bib_ref] ESCMID* guideline for the diagnosis and management of Candida diseases 2012: Prevention..., Hope [/bib_ref] [bib_ref] Clinical Practice Guideline for the Management of Candidiasis: 2016 Update by the..., Pappas [/bib_ref].
A loading dose of 25 mg/kg during the first day is probably needed to achieve steady state levels within 24 h. Fluconazole resistance has emerged, particularly in non-C. albicans species, reaching high levels in some regions and local epidemiology should be considered [bib_ref] ESCMID* guideline for the diagnosis and management of Candida diseases 2012: Prevention..., Hope [/bib_ref]. Nevertheless, in Europe and in America, resistance of main Candida spp. causing IC in neonates (C. albicans and C. parapsilosis) remains low (<5%) [bib_ref] Isavuconazole, micafungin, and 8 comparator antifungal agents' susceptibility profiles for common and..., Pfaller [/bib_ref] [bib_ref] Neonatal and Pediatric Candidemia: Results From Population-Based Active Laboratory Surveillance in Four..., Benedict [/bib_ref] [bib_ref] Candidemia in the Neonatal Intensive Care Unit: A Retrospective, Observational Survey and..., Caggiano [/bib_ref] [bib_ref] Antifungal drug susceptibility of candida spp. in neonatal and paediatric candidaemia: A..., Warris [/bib_ref]. Fluconazole should not be used as empiric therapy in neonates that have prophylactically received fluconazole in the past.
Micafungin is the only echinocandin approved for use in infants aged less than three months in the EU and Japan, but not in the USA [bib_ref] A Review in the Prophylaxis and Treatment of Invasive Candida Infections in..., Scott [/bib_ref]. An analysis of nine clinical trials containing neonates showed that micafungin has a substantial efficacy, with a success rate of IC treatment reaching 73% in both premature and non-premature groups [bib_ref] Micafungin in Premature and Non-Premature Infants: A Systematic Review of Nine Clinical..., Manzoni [/bib_ref] that is somewhat less than the success rate found in adults (approximately 83%) [bib_ref] International, open-label, noncomparative, clinical trial of micafungin alone and in combination for..., Ostrosky-Zeichner [/bib_ref]. Because of increased clearance in neonates, an increased dose of 4-10 mg/kg/d is recommended. The most appropriate doses to achieve levels in the brain parenchyma are 7-10 mg/kg/d. This has been suggested from pharmacokinetics and pharmacodynamics studies of micafungin in experimental HCME [bib_ref] The pharmacokinetics and pharmacodynamics of micafungin in experimental hematogenous Candida meningoencephalitis: Implications..., Hope [/bib_ref]. A population pharmacokinetics study of micafungin in neonates and young infants showed that a higher weight-based dose is required for infants than for adults to achieve effective CNS drug concentrations [bib_ref] Population pharmacokinetics of micafungin in neonates and young infants, Hope [/bib_ref]. This has been validated in a recent clinical trial in neonates, and with the use of a Monte Carlo simulation it was shown that micafungin was efficient and well tolerated at a dose of 10 mg/kg and that there was no need for a loading dose (some authors have used 15 mg/kg) [bib_ref] Pharmacokinetics and safety of fluconazole and micafungin in neonates with systemic candidiasis:..., Leroux [/bib_ref]. In a recently published very small study of micafungin at 10 mg/kg/d (20 patients) vs. deoxycholate amphotericin B at 1 mg/kg/d (10 patients) in infants up to four months of age with IC, there was no difference in fungal-free survival and both agents were well tolerated by the infants [bib_ref] A Phase 3 Study of Micafungin Versus Amphotericin B Deoxycholate in Infants..., Benjamin [/bib_ref].
As neonatal IC is frequently followed by HCME, higher doses of echinocandins are required to achieve adequate drug levels in cerebrospinal fluid (CSF). In a study of 18 neonates with IC, three of whom also had meningitis, micafungin was given at a high dose of 8-15 mg/kg/d. Micafungin concentrations achieved in the CSF were 0.80-1.80 mg/L. While 78% of the subjects treated had clinical resolution of IC, five had neurologic impairments. In three patients treated with 10-15 mg/kg/d, marked γ-GT elevations were observed that were improved after dose reduction [bib_ref] High-Dose Micafungin for Preterm Neonates and Infants with Invasive and Central Nervous..., Auriti [/bib_ref]. Micafungin has been used for shunt lock therapy combined with systemic treatment to treat shunt-associated Candida CNS infections [bib_ref] Shunt lock therapy with micafungin to treat shunt-associated Candida albicans meningitis in..., Auriti [/bib_ref].
Micafungin has few drug-drug interactions and an acceptable safety profile, but transaminase monitoring is recommended during treatment. While echinocandins have a favorable safety profile, the lack of clinical study data inhibits recommendations as first-line agents in neonates [bib_ref] A Phase 3 Study of Micafungin Versus Amphotericin B Deoxycholate in Infants..., Benjamin [/bib_ref] [bib_ref] Echinocandin use in the neonatal intensive care unit, Caudle [/bib_ref].
Caspofungin, despite being the first echinocandin used in adults, has limited data for neonates and infants less than three months [bib_ref] Caspofungin therapy in immunocompromised children and neonates, Somer [/bib_ref] [bib_ref] Successful caspofungin treatment of persistent candidemia in extreme prematurity at 23 and..., Jeon [/bib_ref]. In a clinical trial, a dosage of 25 mg/m 2 given once daily was well tolerated and reached the same pharmacokinetic/pharmacodynamic (PK/PD) parameters as in adults treated with the recommended dose of 50 mg [bib_ref] Pharmacokinetics and safety of caspofungin in neonates and infants less than 3..., Saez-Llorens [/bib_ref]. In a case report of device-associated meningitis in a premature neonate, the use of the above recommended dose (25 mg/m 2 ) resulted in adequate caspofungin concentrations in the CSF and a microbiological and clinical response without the need for device removal [bib_ref] Favorable outcome of neonatal cerebrospinal fluid shunt-associated Candida meningitis with caspofungin, Jans [/bib_ref].
Anidulafungin is not approved for use in neonates and infants less than three months of age. A systematic review of anidulafungin showed no drug-related adverse events, and good pharmacodynamics [bib_ref] Anidulafungin for neonatal hematogenous Candida meningoencephalitis: Identification of candidate regimens for humans..., Warn [/bib_ref] [bib_ref] Invasive fungal infections in infants-focus on anidulafungin, Wilke [/bib_ref]. The efficacy of anidulafungin in neonatal HCME was evaluated using a well-established and proven rabbit model and applying a mathematical model to translate the results to humans. It was concluded that the current dosing regimen of 1.5 mg/kg/d with a loading dose of 3 mg/kg is not sufficient to treat HCME [bib_ref] Anidulafungin for neonatal hematogenous Candida meningoencephalitis: Identification of candidate regimens for humans..., Warn [/bib_ref]. Another study exploring the safety and pharmacokinetics of anidulafungin in neonates and young infants showed that administration of anidulafungin at 1.5 mg/kg/d caused exposure levels that were similar to those of children on the same dose and adults that received the recommended dose of 100 mg/d [bib_ref] Safety and pharmacokinetics of multiple-dose anidulafungin in infants and neonates, Cohen-Wolkowiez [/bib_ref].
## Prophylaxis, empiric therapy, and targeted antifungal therapy in children with primary or secondary immunodeficiencies
## Prophylaxis
Prevention of IFIs in high-risk immunocompromised patients is very important. An appropriate bundle of preventive strategies consists of environmental strategies to lower the chance of fungal acquisition and primary/secondary antifungal drug prophylactic administration. Implementation of high efficiency-particulate air filters, room cleaning, and decontamination, protective clothing, care regarding food, and rational use of antibiotics with an active antibiotic stewardship program require special consideration and are prerequisite preventive measures in high-risk children.
Data on antifungal prophylaxis in children are limited and recommendations are mainly extrapolated from adults. On the other hand, it should be kept in mind that pediatric patients differ from adults in underlying conditions and physiology, in pharmacologic considerations, including dosing, toxicity, and administration, and have a much smaller evidence base, upon which recommendations are relied. Evidence-based recommendations for antifungal prophylaxis have been developed for the following categories of secondary immunodeficiencies: Acute myeloid leukemia, high-risk acute lymphatic leukemia, recurrent acute leukemia, intensive care unit (ICU) admission, and the presence of graft-versus-host disease (GVHD) after allogeneic hematopoietic stem cell transplantation (HSCT) [bib_ref] Guidelines for diagnosis, prevention, and treatment of invasive fungal diseases in paediatric..., Groll [/bib_ref]. In addition, patients with primary immunodeficiencies may need life-long antifungal prophylaxis, the kind of which may depend on the predominant fungi causing IFIs in the particular immunodeficiency.
According to ESCMID and to ECIL's four guidelines, antifungal prophylaxis should be administered until engraftment after HSCT [bib_ref] ESCMID* guideline for the diagnosis and management of Candida diseases 2012: Prevention..., Hope [/bib_ref] [bib_ref] Guidelines for diagnosis, prevention, and treatment of invasive fungal diseases in paediatric..., Groll [/bib_ref]. Antifungal prophylaxis should be continued after engraftment in case of GVHD (with antifungal agents active against both yeasts and molds). In children with HSCT, but without GVHD, prolongation of antifungal therapy after engraftment may be considered. In this case, the duration of antifungal prophylaxis should be determined by the immune recovery and the end of immunosuppression [bib_ref] Guidelines for diagnosis, prevention, and treatment of invasive fungal diseases in paediatric..., Groll [/bib_ref].
The most frequent antifungal agent prophylactically administered to high-risk pediatric patients is fluconazole, based on trials that have shown that patients with allogeneic HSCT receiving fluconazole presented a significant better long-term follow-up probably due to less severe gut GVHD [bib_ref] Antifungal prophylaxis in pediatric hematology/oncology: New choices & new data. Pediatr, Dvorak [/bib_ref] [bib_ref] A controlled trial of fluconazole to prevent fungal infections in patients undergoing..., Goodman [/bib_ref] [bib_ref] Efficacy and safety of fluconazole prophylaxis for fungal infections after marrow transplantation-A..., Slavin [/bib_ref]. Fluconazole is safe and well tolerated up to 12 mg/kg/d and doses up to 10 mg/kg/d were successfully used in a prophylaxis trial [bib_ref] Clinical practice guidelines for the management of candidiasis: 2009 update by the..., Pappas [/bib_ref] [bib_ref] Guideline for primary antifungal prophylaxis for pediatric patients with cancer or hematopoietic..., Science [/bib_ref] [bib_ref] Safety and tolerability of fluconazole in children, Novelli [/bib_ref]. The pediatric fluconazole dose recommended for prophylaxis ranges between 6 and 12 mg/kg/d depending on the patient's age and weight [bib_ref] Antifungal prophylaxis in pediatric hematology/oncology: New choices & new data. Pediatr, Dvorak [/bib_ref] [bib_ref] Guideline for primary antifungal prophylaxis for pediatric patients with cancer or hematopoietic..., Science [/bib_ref].
Since fluconazole does not possess anti-mold activity, studies have compared other antifungal agents with fluconazole to further reduce the incidence of mold infections in high-risk children. In a meta-analysis, antifungal prophylaxis with fluconazole was compared to a mold-active azole, or amphotericin B or echinocandin in adult and pediatric cancer patients receiving chemotherapy or HSCT, but no difference in overall mortality between the two arms was observed (RR 1.0; 95% CI 0.88-1.13) [bib_ref] Mould-active compared with fluconazole prophylaxis to prevent invasive fungal diseases in cancer..., Ethier [/bib_ref]. However, mold active agents significantly reduced both IFI episodes and IFI-related deaths, but with the cost of a higher prevalence of adverse events [bib_ref] Mould-active compared with fluconazole prophylaxis to prevent invasive fungal diseases in cancer..., Ethier [/bib_ref]. Other mold-active azoles were evaluated in high-risk patients for their preventive activity against mold infections. In particular, itraconazole reduced the incidence of IFIs, but due to its adverse effects and frequent drug-drug interactions it is not favorite to the clinicians as prophylaxis in high-risk pediatric patients [bib_ref] Fluconazole versus itraconazole for antifungal prophylaxis in neutropenic patients with haematological malignancies:..., Vardakas [/bib_ref] [bib_ref] Itraconazole versus fluconazole for prevention of fungal infections in patients receiving allogeneic..., Marr [/bib_ref]. In general, azoles are associated with many drug-drug interactions, leading to changes in concentrations of azoles or of interacting medications [bib_ref] Frequency of potential azole drug-drug interactions and consequences of potential fluconazole drug..., Yu [/bib_ref]. The list of drugs that have been implicated is long, including, but not limited to, rifampin, phenytoin, carbamazepine, vinca alkaloids, and calcineurin inhibitors [bib_ref] Frequency of potential azole drug-drug interactions and consequences of potential fluconazole drug..., Yu [/bib_ref].
Voriconazole is a broad-spectrum azole with anti-mold activity, but its use in pediatrics is complicated by inadequate drug exposure and by significant drug-drug interactions [bib_ref] Pharmacokinetic Modeling of Voriconazole to Develop an Alternative Dosing Regimen in Children, Gastine [/bib_ref] [bib_ref] Risk of azole-enhanced vincristine neurotoxicity in pediatric patients with hematological malignancies: Old..., Pana [/bib_ref]. A large randomized controlled trial (RCT) has compared voriconazole to fluconazole in adult and pediatric patients with allogeneic HSCT [bib_ref] Randomized, double-blind trial of fluconazole versus voriconazole for prevention of invasive fungal..., Wingard [/bib_ref]. The results have shown no difference in fungal-free or overall survival at 180 days. The proposed maintenance dose is 8 mg/kg twice daily for intravenous administration and 9 mg/kg twice daily for oral administration for all children less than 12 years of age, and for those 12-14 years of age weighing <50 kg [bib_ref] Guidelines for diagnosis, prevention, and treatment of invasive fungal diseases in paediatric..., Groll [/bib_ref] [bib_ref] Randomized, double-blind trial of fluconazole versus voriconazole for prevention of invasive fungal..., Wingard [/bib_ref]. In a recent nationwide, multicenter observational study regarding antifungal prophylaxis in children with cancer, the use of voriconazole was safe and well tolerated [bib_ref] Voriconazole Antifungal Prophylaxis in Children With Malignancies: A Nationwide Study, Pana [/bib_ref]. Therapeutic drug monitoring should be applied to optimize its serum levels between 1-1.5 and 5 mg/L [bib_ref] Guidelines for diagnosis, prevention, and treatment of invasive fungal diseases in paediatric..., Groll [/bib_ref] [bib_ref] Therapeutic drug monitoring of voriconazole for treatment and prophylaxis of invasive fungal..., Allegra [/bib_ref]. It is not recommended for patients <2 years [bib_ref] Guidelines for diagnosis, prevention, and treatment of invasive fungal diseases in paediatric..., Groll [/bib_ref].
Posaconazole is another alternative choice for hematological/oncological patients, including those with GVHD. Factors limiting its use in children include variable plasma concentrations in pediatric patients, especially those less than 13 years of age, lack of an intravenous formulation (since these children typically have nausea, mucositis, and poor oral intake during the at-risk period), and unreliable absorption of oral suspension. One of the first clinical trials was specifically conducted in patients with GVHD [bib_ref] Posaconazole or fluconazole for prophylaxis in severe graft-versus-host disease, Ullmann [/bib_ref]. This double-blind trial compared fluconazole to posaconazole in patients 13 years and older with acute Grade II-IV or chronic extensive GVHD. Patients receiving posaconazole prophylaxis had lower rates of proven and probable IFIs and a lower fungal-related mortality. Currently, posaconazole is recommended for use in patients >13 years of age with HSCT.
Posaconazole was approved for antifungal prophylaxis (for Aspergillus and Candida infections) in patients with acute myeloid leukemia, myelodysplastic syndromes, GVHD, or in patients undergoing HSCT, in whom a long neutropenic period due to chemotherapy is expected [bib_ref] Antifungal therapy in children: An update, Cecinati [/bib_ref] [bib_ref] Update on antifungal agents for paediatric patients, Groll [/bib_ref] [bib_ref] Retrospective survey on the off-label use of posaconazole in pediatric hematology patients, Cesaro [/bib_ref]. Accumulated data from studies in children, including all ages, with neutropenia have shown that posaconazole is safe and efficacious when used for prophylaxis in these patients [bib_ref] Posaconazole in immunocompromised pediatric patients, Vicenzi [/bib_ref] [bib_ref] Analysis of posaconazole as oral antifungal prophylaxis in pediatric patients under 12..., Doring [/bib_ref]. In comparison with other azoles (voriconazole and itraconazole), posaconazole was well-tolerated, safe, and effective oral antifungal prophylaxis in pediatric patients who underwent high-dose chemotherapy and HSCT [bib_ref] Analysis of posaconazole as oral antifungal prophylaxis in pediatric patients under 12..., Doring [/bib_ref] [bib_ref] Comparison of itraconazole, voriconazole, and posaconazole as oral antifungal prophylaxis in pediatric..., Doring [/bib_ref] [bib_ref] Antifungal prophylaxis with posaconazole vs. fluconazole or itraconazole in pediatric patients with..., Doring [/bib_ref]. Recently, the use of a new posaconazole formulation (gastro-resistant tablets) has been studied in HSCT pediatric patients for antifungal prophylaxis [bib_ref] Efficacy, safety and feasibility of antifungal prophylaxis with posaconazole tablet in paediatric..., Doring [/bib_ref]. In this study, the use of posaconazole as a suspension or tablet had similar efficacy, but patients who received a posaconazole tablet achieved much earlier the targeted plasma concentration than those who received the suspension formulation [bib_ref] Efficacy, safety and feasibility of antifungal prophylaxis with posaconazole tablet in paediatric..., Doring [/bib_ref]. Posaconazole has also been used in patients with primary immunodeficiencies, including chronic granulomatous disease (CGD) [bib_ref] Therapeutic drug monitoring of posaconazole oral suspension in paediatric patients younger than..., Jancel [/bib_ref]. Although a daily dose of 120 mg/m 2 body surface given in three doses has been proposed for oral posaconazole suspension [bib_ref] Spriet, I. Pharmacokinetics of Posaconazole Oral Suspension in Children Dosed According to..., Vanstraelen [/bib_ref] , the optimal dosage needs to be further studied in children, including the new formulation [bib_ref] Posaconazole in immunocompromised pediatric patients, Vicenzi [/bib_ref]. For children ≥13 years, a dose of 300 mg per day in one dose of gastro-resistant tablets (preferred formulation) is recommended. Alternatively, 600 mg per day of oral suspension in three divided doses for prophylaxis (not approved in the EU in patients <18 years). The oral solution administration needs therapeutic drug monitoring (TDM).
Data on echinocandin prophylaxis in pediatrics are limited, with only one study revealing that a higher proportion of patients receiving micafungin than fluconazole had no proven/probable/possible IFI at four weeks following HSCT [bib_ref] Micafungin versus fluconazole for prophylaxis against invasive fungal infections during neutropenia in..., Van Burik [/bib_ref]. There was no difference in rates of proven or probable IFI or overall and fungal-related mortality. However, the number of pediatric subjects enrolled was small (n = 84) and a reduction in the incidence of proven or probable IFI was not demonstrated. Micafungin was also tested as antifungal prophylaxis at 2 mg/kg/d in pediatric patients with allogeneic HSCT [bib_ref] Safety, tolerability, and feasibility of antifungal prophylaxis with micafungin at 2 mg/kg..., Yoshikawa [/bib_ref]. The main disadvantages for widespread echinocandin use are a lack of oral formulation and cost.
Caspofungin has been shown to be at least equivalent to itraconazole as antifungal prophylaxis with little caspofungin-related adverse events [bib_ref] Open-label, randomized comparison of itraconazole versus caspofungin for prophylaxis in patients with..., Mattiuzzi [/bib_ref]. Two retrospective studies conducted in children undergoing HSCT caspofungin was safe and with a similar efficacy with a comparator antifungal agent, which was liposomal amphotericin B in one study and micafungain in the other [bib_ref] Caspofungin as antifungal prophylaxis in pediatric patients undergoing allogeneic hematopoietic stem cell..., Doring [/bib_ref] [bib_ref] Comparison of Efficacy and Safety of Caspofungin Versus Micafungin in Pediatric Allogeneic..., Maximova [/bib_ref].
Patients with primary immunodeficiencies being at high risk for IFIs or presenting as chronic fungal infections (i.e., chronic mucocutaneous candidiasis, CGD, severe combined immunodeficiency, and others) may require life-long antifungal prophylaxis [bib_ref] Prevention of infections during primary immunodeficiency, Aguilar [/bib_ref] [bib_ref] Preventing fungal infections in chronic granulomatous disease, Verweij [/bib_ref] [bib_ref] A twice daily posaconazole dosing algorithm for children with chronic granulomatous disease, Welzen [/bib_ref]. Fluconazole for anti-Candida prophylaxis or maintenance therapy and a mold-active azole-such as itraconazole, voriconazole, or posaconazole-for anti-Aspergillus prophylaxis are appropriate and are thoroughly reviewed elsewhere [bib_ref] Prevention of infections during primary immunodeficiency, Aguilar [/bib_ref] [bib_ref] Invasive fungal infections in congenital immunodeficiencies, Antachopoulos [/bib_ref] [bib_ref] Diagnosis and management of Aspergillus diseases: Executive summary of the 2017 ESCMID-ECMM-ERS..., Ullmann [/bib_ref] [bib_ref] International Pediatric Lung Transplant Collaborative. Antifungal prophylaxis in pediatric lung transplantation: An..., Mead [/bib_ref].
## Empiric therapy
Empiric fever-driven antifungal therapy has been considered as standard of care in hemato-oncological patients at high risk for invasive fungal disease with neutropenia, who present with refractory or a new fever of at least four days, despite broad-spectrum antibacterial therapy [bib_ref] Guideline for the Management of Fever and Neutropenia in Children With Cancer..., Lehrnbecher [/bib_ref]. Both ECIL's four guidelines in 2014 and a recent update of clinical practice guidelines in children with cancer and HSC recipients recommend the following options in pediatric patients of all age groups: LAMB (1-3 mg/kg/d) or caspofungin (loading dose 70 mg/m 2 /d, followed by 50 mg/m 2 /d) [bib_ref] Guidelines for diagnosis, prevention, and treatment of invasive fungal diseases in paediatric..., Groll [/bib_ref] [bib_ref] Guideline for the Management of Fever and Neutropenia in Children With Cancer..., Lehrnbecher [/bib_ref]. Children with febrile and neutropenia with low risk for IFI have no benefit with starting empiric antifungal treatment. Nowadays, a more diagnostics-driven empiric (or pre-emptive) antifungal therapy is desirable based on findings of biomarkers, including galactomannan, and BDG, as well as imaging of chest and other organs [bib_ref] Guideline for the Management of Fever and Neutropenia in Children With Cancer..., Lehrnbecher [/bib_ref] [bib_ref] Critical review of current clinical practice guidelines for antifungal therapy in paediatric..., Morgan [/bib_ref]. This eliminates the need for administration of empiric antifungal therapy in neutropenic patients, who present with fever of non-fungal etiology. In a very recent randomized, multicenter clinical trial conducted in Chile, pre-emptive antifungal therapy was non-inferior to standard empiric driven antifungal therapy in pediatric patients aged less than 18 years with cancer and high-risk febrile neutropenia [bib_ref] Efficacy of pre-emptive versus empirical antifungal therapy in children with cancer and..., Santolaya [/bib_ref]. In addition, pre-emptive therapy was associated with statistically significant less antifungal exposure.
A randomized, double-blinded, multicenter study compared caspofungin and LAMB for empiric antifungal therapy in pediatric patients with persistent febrile neutropenia (FN). It revealed that caspofungin (50 mg/m 2 /d) and LAMB (3 mg/kg/d) were comparable in safety, tolerability, and efficacy [bib_ref] A randomized, double-blind, multicenter study of caspofungin versus liposomal amphotericin B for..., Maertens [/bib_ref]. Caspofungin doses can be increased to 70 mg/m 2 /d if clinically indicated (maximum dose of 70 mg/d). Another multicenter study evaluated the efficacy and safety of micafungin at a median dose of 3.0 mg/kg/d and a duration of 13.5 days for FN in pediatric patients with hematological malignancies, and revealed that micafungin may be an alternative safe and effective agent to treat pediatric patients with FN [bib_ref] Efficacy and safety of micafungin for febrile neutropenia in pediatric patients with..., Kobayashi [/bib_ref].
Empirical therapy in adult ICU patients has been shown to be of no benefit when using a fever criterion [bib_ref] ESCMID* guideline for the diagnosis and management of Candida diseases 2012: Prevention..., Hope [/bib_ref] [bib_ref] Clinical Practice Guideline for the Management of Candidiasis: 2016 Update by the..., Pappas [/bib_ref]. Similarly, a recent meta-analysis in non-neutropenic pediatric and adult ICU patients found that non-targeted antifungal therapy (including prophylactic, pre-emptive, and empiric antifungal treatment) compared to placebo or no therapy had no significant reduction in overall mortality [bib_ref] Antifungal agents for preventing fungal infections in non-neutropenic critically ill patients, Cortegiani [/bib_ref]. Diagnostic-driven empiric antifungal therapy has been developed both for children and adults, but with significant limitations [bib_ref] Role of Molecular Biomarkers in the Diagnosis of Invasive Fungal Diseases in..., Huppler [/bib_ref]. Currently, no conclusive evidence exists for the use of galactomannan, BDG, mannan antigen/antibody, and fungal PCR in children with high or low risk for IFD [bib_ref] Role of Molecular Biomarkers in the Diagnosis of Invasive Fungal Diseases in..., Huppler [/bib_ref] [bib_ref] β-D-Glucan, and Polymerase Chain Reaction-Based Assays for the Diagnosis of Invasive Fungal..., Lehrnbecher [/bib_ref]. Improvement of current biomarkers alone or in combination [bib_ref] Comparative evaluation of pan-fungal real-time PCR, galactomannan and (1-3)-β-D-glucan assay for invasive..., Gupta [/bib_ref] is warranted. A new and promising biomarker, T2 magnetic resonance assay, can detect the five most common isolated Candida species within a few hours in whole blood [bib_ref] Use of T2MR in invasive candidiasis with and without candidemia, Zacharioudakis [/bib_ref]. This T2 magnetic resonance biomarker has already been applied in pediatric children with and without candidemia and although there were few patients, a high accuracy was found [bib_ref] T2Candida Provides Rapid and Accurate Species Identification in Pediatric Cases of Candidemia, Hamula [/bib_ref].
In general, patients with a refractory or new fever episode in the pediatric intensive care unit (PICU), despite broad-spectrum empirical antibacterial therapy, who are at high risk for Candida infection with moderate-to-severe disease, hemodynamic instability recent azole exposure, or at high risk for C. glabrata or C. krusei infections require empiric treatment with an echinocandin or LAMB [bib_ref] Clinical Practice Guideline for the Management of Candidiasis: 2016 Update by the..., Pappas [/bib_ref] [bib_ref] Antifungal therapy in children: An update, Cecinati [/bib_ref].
## Targeted therapy
LAMB at 3-5 mg/kg/d has been approved as a first-line treatment of IFIs in pediatrics, including invasive aspergillosis (IA) and IC [bib_ref] ESCMID* guideline for the diagnosis and management of Candida diseases 2012: Prevention..., Hope [/bib_ref] [bib_ref] Clinical Practice Guideline for the Management of Candidiasis: 2016 Update by the..., Pappas [/bib_ref] [bib_ref] Guidelines for diagnosis, prevention, and treatment of invasive fungal diseases in paediatric..., Groll [/bib_ref]. Amphotericin B lipid complex at 5 mg/kg/d can be also used [bib_ref] Update on antifungal agents for paediatric patients, Groll [/bib_ref]. Other indications for AMB administration in children include blastomycosis, coccidioidomycosis, histoplasmosis, and endemic mycoses, whereas LAMB at a higher dose (≥5 mg/kg) is the first-line treatment of mucormycosis [bib_ref] Update on antifungal agents for paediatric patients, Groll [/bib_ref] [bib_ref] Rare fungal infections in children: An updated review of the literature, Pana [/bib_ref]. The combination of LAMB with flucytosine is proposed to treat difficult fungal infections, such as Candida meningitis and endocarditis [bib_ref] Evidence of excessive concentrations of 5-flucytosine in children aged below 12 years:..., Soltani [/bib_ref]. For the management of cryptococcosis in children, data are limited. The main principle is to start induction therapy for meningoencephalitis using LAMB combined with flucytosine for at least two weeks, followed by fluconazole for a long period [bib_ref] Clinical practice guidelines for the management of cryptococcal disease: 2010 update by..., Perfect [/bib_ref]. According to 2010 guidelines, the proposed dose regimen in children for CNS and disseminated cryptococcal disease is DAMB (1 mg/kg/d iv) in combination with flucytosine (100 mg/kg/d orally in four divided doses) for two weeks, followed by fluconazole (10-12 mg/kg/d orally) for eight weeks; for AmB-intolerant patients, either liposomal AmB (5 mg/kg per day) or amphotericin B lipid complex (ABLC) (5 mg/kg per day). Fluconazole is also proposed as a maintenance treatment at a dose of 6 mg/kg per day orally, while for cryptococcal pneumonia, fluconazole use is recommended (6-12 mg/ kg) for 6-12 months [bib_ref] Clinical practice guidelines for the management of cryptococcal disease: 2010 update by..., Perfect [/bib_ref] [bib_ref] Cryptococcosis diagnosis and treatment: What do we know now, Perfect [/bib_ref].
Among echinocandins, caspofungin is most commonly used to treat IFIs in pediatric patients [bib_ref] Systemic antifungal prescribing in neonates and children: Outcomes from the Antibiotic Resistance..., Lestner [/bib_ref]. An open-label prospective study of caspofungin for the treatment of IC and IA in patients aged three months to 17 years reported success (defined as complete or partial response) at the end of therapy in 5/10 (50%) patients with IA and 30/48 (62.5%) patients with IC without presenting any major adverse effects during treatment [bib_ref] A prospective, multicenter study of caspofungin for the treatment of documented Candida..., Zaoutis [/bib_ref]. Another prospective study evaluated caspofungin in 83 patients with IA, most of which were refractory to other therapies. An overall favorable response to caspofungin was noted in 45% of patients [bib_ref] Efficacy and safety of caspofungin for treatment of invasive aspergillosis in patients..., Maertens [/bib_ref]. A recent systematic review and meta-analysis of three randomized controlled trials of a high quality, but with small sample size, in children and neonates found a favorable outcome for patients treated with caspofungin in comparison to amphotericin B [bib_ref] Efficacy and safety of caspofungin in children: Systematic review and meta-analysis, Rosanova [/bib_ref]. Based on these randomized trials as well as observational studies, caspofungin was approved for pediatric patients as second-line therapy of IA and as primary targeted therapy of IC in Europe [bib_ref] ESCMID* guideline for the diagnosis and management of Candida diseases 2012: Prevention..., Hope [/bib_ref] [bib_ref] Guidelines for diagnosis, prevention, and treatment of invasive fungal diseases in paediatric..., Groll [/bib_ref].
Micafungin pediatric data are also limited. A retrospective study of patients <2 years of age found nine clinical studies having enrolled patients of this age [bib_ref] Micafungin in Premature and Non-Premature Infants: A Systematic Review of Nine Clinical..., Manzoni [/bib_ref]. Treatment-related adverse events were recorded in 23% of patients, with no difference between prematurely and non-prematurely born infants. For a subgroup of 30 patients with IC, treatment success was achieved in 73% in both premature and non-premature groups. The safety and pharmacokinetics of micafungin at 3 mg/kg and 4.5 mg/kg once-daily were evaluated in children with proven, probable, or suspected IC [bib_ref] Safety and pharmacokinetic profiles of repeated-dose micafungin in children and adolescents treated..., Benjamin [/bib_ref] ; both dosages of micafungin were well tolerated. In a randomized double-blind trial of micafungin (2 mg/kg/d) versus LAMB (3 mg/kg/d) in 98 children aged <16 years, both drugs achieved similar success rates for treating candidemia/IC [bib_ref] Micafungin versus liposomal amphotericin B for pediatric patients with invasive candidiasis: Substudy..., Queiroz-Telles [/bib_ref]. In a recent meta-analysis, the use of micafungin for the prevention or treatment of IFIs in neutropenic cancer pediatric patients had significantly more success rates than comparators, but similar mortality rates [bib_ref] Efficacy and safety of micafungin versus extensive azoles in the prevention and..., Lee [/bib_ref]. According to ESCMID and ECIL's four guidelines, micafungin is licensed for targeted therapy of IC at a dose of 2-4 mg/kg [bib_ref] ESCMID* guideline for the diagnosis and management of Candida diseases 2012: Prevention..., Hope [/bib_ref] [bib_ref] Guidelines for diagnosis, prevention, and treatment of invasive fungal diseases in paediatric..., Groll [/bib_ref].
A meta-analysis was recently conducted to evaluate the limited clinical data of echinocandins vs. other antifungal therapy for safety and efficacy in the treatment of IC in children [bib_ref] Efficacy and Safety of Echinocandins for the Treatment of Invasive Candidiasis in..., Tsekoura [/bib_ref]. In particular, from four randomized clinical trials (324 patients), two confirmed IC (micafungin vs. LAMB and caspofungin vs. LAMB) and two empirical therapy trials (caspofungin vs. DAMB and caspofungin vs. LAMB) were included. There was no significant difference between echinocandins and the comparator in terms of treatment success (OR = 1.61, 95% CI 0.74-3.50) and incidence of treatment-related adverse events (OR = 0.70, 95% CI 0.39-1.26). However, fewer children treated with echinocandins discontinued treatment due to adverse events than amphotericin B formulations (OR = 0.26, 95% CI 0.08-0.82, p = 0.02).
Pediatric data on anidulafungin efficacy and safety are restricted in only one published pediatric study to date [bib_ref] Safety and pharmacokinetics of intravenous anidulafungin in children with neutropenia at high..., Benjamin [/bib_ref]. Anidulafungin was well tolerated. Pediatric patients receiving 0.75 mg/kg/d or 1.5 mg/kg/d had anidulafungin concentration profiles similar to those of adult patients receiving 50 or 100 mg/d, respectively. In a prospective study conducted in a single center in Argentina, a total of 55 patients received anidulafungin for treatment or prophylaxis due to a temporary shortage of amphotericin B [bib_ref] Anidulafungin in children: Experience in a tertiary care children's hospital in Argentina, Rosanova [/bib_ref]. The majority of the children had bone marrow transplantation and anidulafungin was well tolerated and efficacious when given at a loading dose of 3 mg/kg/d and a maintenance dose of 1.5 mg/kg/d [bib_ref] Anidulafungin in children: Experience in a tertiary care children's hospital in Argentina, Rosanova [/bib_ref]. Preliminary data of an open study of anidulafungin to patients two years to 17 years were recently presented, showing that anidulafungin at doses of 3 mg/kg as a loading dose followed by 1.5 mg/kg/d was effective, with a global response success rate of 70.8% at end-of-iv treatment and an acceptable tolerability and safety profile in children diagnosed with IC [bib_ref] on behalf of the anidulafungin A8851008 pediatric study group. A Prospective, Open-label..., Roilides [/bib_ref]. Similar data are currently being analyzed for patients one month to two years of age. ESCMID/ECIL/IDSA guidelines recommend a dose of 3 mg/kg as a loading dose followed by 1.5 mg/kg/d [bib_ref] ESCMID* guideline for the diagnosis and management of Candida diseases 2012: Prevention..., Hope [/bib_ref] [bib_ref] Clinical Practice Guideline for the Management of Candidiasis: 2016 Update by the..., Pappas [/bib_ref] [bib_ref] Guidelines for diagnosis, prevention, and treatment of invasive fungal diseases in paediatric..., Groll [/bib_ref]. Anidulafungin is not licensed yet for patients younger than 18 years.
Fluconazole can be safely used in IC caused by fluconazole-susceptible organisms in pediatric patients of all ages who are in a stable condition and have not received prior azole therapy [bib_ref] ESCMID* guideline for the diagnosis and management of Candida diseases 2012: Prevention..., Hope [/bib_ref] [bib_ref] Guidelines for diagnosis, prevention, and treatment of invasive fungal diseases in paediatric..., Groll [/bib_ref] [bib_ref] Antifungal therapy in children: An update, Cecinati [/bib_ref]. Fluconazole or echinocandin can also be used as an initial therapy for candidemia in non-neutropenic patients, depending on the disease severity and possibility of azole resistance. According to IDSA guidelines, an echinocandin should be started for patients with moderately severe to severe illness or for patients with recent azole exposure, while fluconazole is recommended for patients with a less critical condition and without recent azole exposure or risk for a fluconazole resistant candida [bib_ref] Clinical Practice Guideline for the Management of Candidiasis: 2016 Update by the..., Pappas [/bib_ref].
Voriconazole is indicated in pediatric patients as first line therapy for IA as well as for scedosporiosis, fusariosis, and other IFIs in pediatric patients, who are intolerant of or refractory to conventional antifungal therapy [bib_ref] Guidelines for diagnosis, prevention, and treatment of invasive fungal diseases in paediatric..., Groll [/bib_ref] [bib_ref] Voriconazole in the treatment of aspergillosis, scedosporiosis and other invasive fungal infections..., Walsh [/bib_ref] [bib_ref] Practice Guidelines for the Diagnosis and Management of Aspergillosis: 2016 Update by..., Patterson [/bib_ref]. The recommended voriconazole dose for the treatment of IA is based on the age and weight of the child and TDM is strongly recommended [bib_ref] Guidelines for diagnosis, prevention, and treatment of invasive fungal diseases in paediatric..., Groll [/bib_ref]. As voriconazole has linear PK and high variability in children 2-11 years old, the dose for children is a loading of 9 mg/kg/dose × 2 followed by maintenance of 8 mg/kg/dose × 2 with pos 9 mg/kg/dose × 2. Higher voriconazole dosages or even given every eight hours have been suggested for children and especially in younger children aged less than two years of age [bib_ref] An Optimized Voriconazole Dosing Strategy to Achieve Therapeutic Serum Concentrations in Children..., Zembles [/bib_ref].
Mucormycosis is a life-threatening IFI particularly affecting children with hematological malignancies [bib_ref] Collaborators of Zygomyco.net and/or FungiScope Registries. Invasive mucormycosis in children: An epidemiologic..., Pana [/bib_ref]. Posaconazole is active against mucormycosis, and when combined with other antifungal drugs, can be effective in immunocompromised children and be used as a salvage therapy [bib_ref] Posaconazole salvage treatment in paediatric patients: A multicentre survey, Lehrnbecher [/bib_ref] [bib_ref] Mucormycosis treated with posaconazole: Review of 96 case reports, Vehreschild [/bib_ref] [bib_ref] ESCMID and ECMM joint clinical guidelines for the diagnosis and management of..., Cornely [/bib_ref]. According to ECIL's four recommendations, posaconazole can be used in children >13 years for scedosporiosis and fusariosis and as second line treatment for mucormycosis [bib_ref] Guidelines for diagnosis, prevention, and treatment of invasive fungal diseases in paediatric..., Groll [/bib_ref]. In a multicenter, phase three, open-label study in juvenile (age range 8-17 years old) and adult (18-64 years) patients who were intolerant of or had IFIs refractory to standard antifungal therapies, posaconazole at 800 mg/d as an oral suspension in divided doses was administered to almost all patients [bib_ref] Posaconazole plasma concentrations in juvenile patients with invasive fungal infection, Krishna [/bib_ref]. The overall success rates and adverse event profiles were comparable. Posaconazole plasma concentrations were similar for juvenile and adult patients, suggesting that clinical outcomes are expected to be similar in adults and children with refractory IFIs. Nowadays, the gastro-resistant tablets of posaconazole have eliminated the need for TDM. For the treatment of children ≥13 years, a dose of 300 mg per day in one dose (day 1, two doses of 300 mg) of gastro-resistant tablets (preferred formulation) is recommended. Alternatively, a dose of 800 mg per day in two or four divided doses of oral solution is recommended (not approved in the EU in patients <18 years). The oral solution administration needs therapeutic drug monitoring (TDM).
There are limited data for the use of posaconazole for the treatment of IFIs in children aged less than 13 years of age. In a recent study, 13 pediatric cancer patients were treated with posaconazole at a median dose of 12.5-16.5 mg/kg/d and 77% of them achieved the targeted concentration of 1 µg/mL within the first week [bib_ref] Posaconazole oral dose and plasma levels in pediatric hematology-oncology patients, Vicenzi [/bib_ref]. Similarly, a retrospective analysis of the administration of oral posaconazole suspension for the treatment of fungal infections was conducted in 12 patients aged less than 13 years [bib_ref] Therapeutic drug monitoring of posaconazole oral suspension in paediatric patients younger than..., Jancel [/bib_ref]. In this study, which included mostly male patients with CGD, posaconazole dose ranged from 18.5-47.9 mg/kg/d, the daily frequency varied from two to four doses, and plasma concentrations ranged from 0.23-2.16 µg/mL. TDM is the same in children and adults, 0.7 for prophylaxis and 1 mg/L for treatment.
Isavuconazole, a new triazole with activity against both Aspergillus and Mucorales, and both oral and intravenous formulation, has been recently approved for use in adults above 18 years [bib_ref] Isavuconazole: A New Broad-Spectrum Triazole Antifungal Agent, Miceli [/bib_ref]. Currently, its pharmacokinetics are being studied in children from 1-18 years for the intravenous regimen and from 6-18 years for the oral regimen (ClinicalTrials.gov NCT03241550). It has been used for treating a few children with mucormycosis successfully [bib_ref] Successful treatment of invasive mucormycosis with isavuconazole in pediatric patients, Barg [/bib_ref].
A number of rarer IFIs may be encountered in immunocompromised children-such as invasive fusariosis, scedosporiosis, or trichosporonosis. A review of these infections in children has been published [bib_ref] Rare Fungal Infections in Children: An Updated Review of the Literature, Pana [/bib_ref]. Antifungal therapy for them in infants or children does not differ from that in adults even though some of the active agents (such as voriconazole) cannot be given in young patients. Surgical intervention and adjunctive immunotherapy have been used in some cases [bib_ref] Immunotherapy of infections caused by rare filamentous fungi, Katragkou [/bib_ref].
# Conclusions
Fluconazole remains the most frequent antifungal prophylactic agent given to high-risk neonates and children. However, the emergence of fluconazole resistance, particularly in non-albicans Candida species, should be considered during preventive or empiric therapy. In VLBW neonates, although fluconazole is used as antifungal prophylaxis in NICU's with a relatively high incidence of IC, its role is under continuous debate. The mainstay of therapy for treating neonatal and pediatric yeast and mold infections remains amphotericin B, primarily in its liposomal formulation. Voriconazole is indicated for mold infections except mucormycosis in children >2 years. Newer triazoles-such as posaconazole and isavuconazole-and echinocandins are either licensed or under study for first-line or salvage therapy, whereas combination therapy is kept for refractory cases.
Funding: The authors of this review received no external funding.
Conflicts of Interest: E Roilides has received research grants from Astellas, Gilead, and Pfizer Inc.; and is a scientific advisor and member of speaker bureaux for Astellas, Gilead, Merck, and Pfizer Inc. The other authors declare no conflict of interest.
[table] Table 1: Antifungal drugs in pediatric patients. [/table]
|
XCluSim: a visual analytics tool for interactively comparing multiple clustering results of bioinformatics data
Background: Though cluster analysis has become a routine analytic task for bioinformatics research, it is still arduous for researchers to assess the quality of a clustering result. To select the best clustering method and its parameters for a dataset, researchers have to run multiple clustering algorithms and compare them. However, such a comparison task with multiple clustering results is cognitively demanding and laborious.Results: In this paper, we present XCluSim, a visual analytics tool that enables users to interactively compare multiple clustering results based on the Visual Information Seeking Mantra. We build a taxonomy for categorizing existing techniques of clustering results visualization in terms of the Gestalt principles of grouping. Using the taxonomy, we choose the most appropriate interactive visualizations for presenting individual clustering results from different types of clustering algorithms. The efficacy of XCluSim is shown through case studies with a bioinformatician.Conclusions: Compared to other relevant tools, XCluSim enables users to compare multiple clustering results in a more scalable manner. Moreover, XCluSim supports diverse clustering algorithms and dedicated visualizations and interactions for different types of clustering results, allowing more effective exploration of details on demand. Through case studies with a bioinformatics researcher, we received positive feedback on the functionalities of XCluSim, including its ability to help identify stably clustered items across multiple clustering results.
# Background
Since Eisen lab's Cluster and TreeView [bib_ref] Cluster analysis and display of genome-wide expression patterns, Eisen [/bib_ref] popularized cluster analyses and visualizations of microarray data, cluster analysis has been widely used in the bioinformatics community. As genetic probing technologies rapidly improve in capacity and accuracy (e.g. Next Generation Sequencing), cluster analysis is playing an even more important role in the descriptive modeling (segmentation or partitioning) of the large data produced by high-throughput probing technologies. Though cluster analysis has become a routine analytic task for bioinformatics research, it is still arduous for a researcher to quantify the quality of a clustering method's clustering results.
There have been a few attempts to develop objective measures for clustering quality assessment; however, in most practical research projects, determining the quality of a clustering result is subjective and application specific [bib_ref] Interactively exploring hierarchical clustering results, Seo [/bib_ref]. To make things even more challenging, there are a large number of clustering methods, which could generate diverse clustering results. Moreover, even an individual clustering algorithm could end up with different results depending on the clustering parameters.
Since there is no generally accepted objective metric for selecting the best clustering method and its parameters for a given dataset, researchers often have to run multiple clustering algorithms and compare different results while examining the concordance/discordance among them. Such a comparison task with multiple clustering results for a large dataset is cognitively demanding and laborious.
In this paper, we present XCluSim, a visual analytics tool that enables users to interactively compare multiple clustering results and explore individual clustering results using dedicated visualizations.
This paper is structured as follows. In the next section we discuss some of the most relevant visualization tools and techniques, focusing on a comparative analysis of multiple clustering results. Each visualization component and its interactions in XCluSim are described in the Methods section. The Results and discussion section contains case studies and discussions followed by a conclusion.
## Related work visual comparison using visualizations for multidimensional categorical data
Since multiple clustering results can be treated as multidimensional categorical datasets, they can be visualized using various visualization techniques corresponding to the specific data types. These techniques include Parallel Sets [bib_ref] Parallel sets: visual analysis of categorical data, Kosara [/bib_ref] and Parallel Coordinate Plot [bib_ref] Parallel coordinates: a tool for visualizing multidimensional geometry. Visualization, 1990. Visualization '90, Inselberg [/bib_ref]. Lots of prior work on the visual comparison of multiple clustering results employed these techniques [bib_ref] Interactively exploring hierarchical clustering results, Seo [/bib_ref] [bib_ref] iGPSe: A visual analytic system for integrative genomic based cancer patient stratification, Ding [/bib_ref] [bib_ref] Visually comparing multiple partitions of data with applications to clustering, Zhou [/bib_ref] [bib_ref] Diverse information integration and visualization, Havre [/bib_ref] [bib_ref] Comparative analysis of multidimensional, quantitative data, Lex [/bib_ref] [bib_ref] Comparing clusterings using Bertin's idea, Pilhofer [/bib_ref] [bib_ref] StratomeX: visual analysis of large-scale heterogeneous Genomics data for cancer subtype characterization, Lex [/bib_ref] , but we focus our discussion on the ones that are most relevant to us in terms of utilizing ribbon-like bands to represent concordance/discordance among multiple clustering results.
In iGPSe [bib_ref] iGPSe: A visual analytic system for integrative genomic based cancer patient stratification, Ding [/bib_ref] , to visually compare clustering results of two different expression data types (i.e. gene expression and micro-RNAs expression), two dimensional axes were juxtaposed, allowing for the use of parallel sets. By observing the flow of ribbon-like bands, users were easily able to see which items were shared between a pair of clusters from two different clustering results. HCE [bib_ref] Interactively exploring hierarchical clustering results, Seo [/bib_ref] also juxtaposed a pair of hierarchical clustering results in parallel to enable comparison tasks with the two results. In contrast to iGPSe, HCE used a partitioned heatmap instead of a simple node to show the details of each data item. To reveal the relations between items in a pair of heatmaps, matching items were connected with straight lines. However, these two visual analytics tools only supported the comparison of a pair of clustering results. Moreover, because they used connectivity between related items, it was often the case that there were too many crossing lines with a large dataset.
CComViz [bib_ref] Visually comparing multiple partitions of data with applications to clustering, Zhou [/bib_ref] alleviated the line crossing problem while focusing on the comparison tasks of more than two clustering results. In their work, multiple clustering results were visualized with a parallel coordinate plot: clustering results as dimensions, clusters as vertical positions in each dimension, and items as lines. Users could grasp the overall distribution of items across multiple clustering results by tracking the flow of lines crossing multiple dimensions. Similar representations were used in [bib_ref] Diverse information integration and visualization, Havre [/bib_ref] , but CComViz devised an algorithm for rearranging clusters and their members to minimize visual clutter between each dimension. Matchmaker [bib_ref] Comparative analysis of multidimensional, quantitative data, Lex [/bib_ref] also utilized the parallel coordinate plot, but to show raw data simultaneously, partitioned heatmaps were shown in dimensional axes. The items in each dimension were rearranged by their average values so that heatmaps clearly showed the patterns of their raw data. Unlike the case of CComViz, in this case, partitioned heatmaps used a bundling strategy to maintain the position of each item in a dimension. This reduced line crossings between adjacent dimensions. Although this method generated a clearer overview of the distributions of items, it had some drawbacks. First of all, the flows of inner lines were invisible unless users explicitly highlighted the lines. Secondly, since the lines were bundled, the width of a band may not have accurately conveyed the number of the items belonging to the band.
CComViz and Matchmaker were probably most relevant to XCluSim. They depended on a linear ordering of dimensions (or clustering results), which made it difficult to do all-pairs comparison with a large number of clustering results at once. For example, as the authors said, Matchmaker only enabled users to compare, at most, six clustering results simultaneously, even with the limited linear ordering of dimensions. Since the same dataset can yield a large number of different clustering results, it is necessary to provide a more scalable way of comparing them. In XCluSim, we present diverse overviews to help in comparison tasks with many clustering results.
## Visualization using similarity measures
There are a few approaches to visualize measured similarity values between clusters (or items) in different clustering results instead of explicitly visualizing shared items among multiple clustering results. Sharko et al. [bib_ref] Heat map visualizations allow comparison of multiple clustering results and evaluation of..., Sharko [/bib_ref] utilized a color-coded similarity matrix view to show the stability between items or clusters across different clustering results. Similarities were measured by counting how many times each pair of items was clustered together or how many items each pair of clusters shared. Kothur et al. [bib_ref] Visual Analytics for Comparison of Ocean Model Output with Reference Data: Detecting..., Kothur [/bib_ref] used bar charts arranged in a matrix layout to show similarity values between a pair of clusters. However, these two works were restricted to comparing a pair of clustering results since they both used a matrix layout.
iGPSe [bib_ref] iGPSe: A visual analytic system for integrative genomic based cancer patient stratification, Ding [/bib_ref] used Silhouette Plot [bib_ref] Silhouettes: a graphical aid to the interpretation and validation of cluster analysis, Rousseeuw [/bib_ref] to help compare a pair of clustering results. Each item got a standardized dissimilarity value ranging from -1 to 1. This value represented dissimilarity in such a way that, when a value was close to 1, its average dissimilarity from all other items in the same cluster was much smaller than the maximum average dissimilarity from all items in another cluster. When the value was close to -1, the meaning of the value was reversed. By representing these similarity values between clustering results using a bar chart, users were able to assess the relative quality of clustering results.
These previous works using similarity measures allowed for comparisons of only a small number of clustering results. However, it is clear that, by abstracting detailed differences to simpler similarity measures, the visual comparison could be rendered more scalable. In our work, we used a graph layout and a dendrogram to show similarity overviews in a more scalable way.
## Color encoding for clusters
Color is a powerful visual cue for representing a cluster membership. It is used in many visualization techniques, including parallel coordinate plot [bib_ref] Visually comparing multiple partitions of data with applications to clustering, Zhou [/bib_ref] [bib_ref] Heat map visualizations allow comparison of multiple clustering results and evaluation of..., Sharko [/bib_ref] and scatterplot [bib_ref] Just-in-time annotation of clusters, outliers, and trends in point-based data visualizations, Kandogan [/bib_ref] [bib_ref] Interactive visual clustering of large collections of trajectories, Andrienko [/bib_ref] [bib_ref] Scatter/gather clustering: Flexibly incorporating user feedback to steer clustering results, Hossain [/bib_ref] , to discriminate clusters while revealing trends in raw data. Similar efforts exist in the visualizations of multiple clustering results. For example, when using the parallel sets view, a few distinct colors are used to encode each cluster to discriminate it from others [bib_ref] iGPSe: A visual analytic system for integrative genomic based cancer patient stratification, Ding [/bib_ref] [bib_ref] Visually comparing multiple partitions of data with applications to clustering, Zhou [/bib_ref].
However, if there are clusters from different clustering results that share the same members, it is not desirable to encode them in distinct colors since it may mislead a user into thinking that those clusters are different. Moreover, when the number of clusters increases, it is hard to color-code clusters differently, because it is hard to discriminate between more than 10 colors.
A useful color encoding strategy is Tree Colors [bib_ref] Tree Colors: Color Schemes for Tree-Structured Data, Tennekes [/bib_ref] , which was devised for tree-structured data to represent similarities between nodes. A part of the parent's hue range is recursively assigned to its child nodes. As a result, nodes with the same parent have similar colors, while those that are less similar have different colors. Moreover, this color scheme reflects the level of a node by using differentially encoded chroma and luminance in each level. If the similarities between clusters from multiple clustering results can be represented as a tree structure, Tree Colors may be well-suited to represent similarity among them. In XCluSim, we used this color scheme to color-code clusters after building a hierarchical structure by running a hierarchical agglomerative clustering (HAC) [bib_ref] Cluster analysis and display of genome-wide expression patterns, Eisen [/bib_ref] with all clusters.
# Methods
## Task analysis and design goals
When performing a cluster analysis with a gene expression dataset, bioinformaticians typically follow an iterative analytics process: 1) they filter out unnecessary genes from the dataset for more focused analysis; 2) they run a clustering algorithm with the selected genes; and 3) they validate clusters in the clustering result to determine whether genes are clustered properly in the biological context. When the quality of the clustering result is not satisfactory at the validation stage, they often have to return to previous steps and run the same clustering algorithm with different parameters or run a different clustering algorithm.
Years of close collaboration with bioinfomaticians have revealed to us that they often faced challenges in this iterative analytics process. First of all, there is no flexible analytics environment that supports them through the iterative process while providing diverse clustering algorithms and keeping track of their exploration history (i.e., the sequence of the clustering algorithms and parameter settings). Moreover, it is challenging for them to effectively compare different clustering results generated during multiple iterations while investigating the quality of the results at diverse levels (i.e. clustering results level, cluster level, and gene level).
To address these challenges in the iterative process of cluster analysis, we set the following design goals for our visual analytics tool:
- To facilitate scalable visual comparison of many clustering results at diverse levels;
- To support generation of diverse clustering results;
- To promote understanding of the characteristics of each clustering algorithm and its parameters in results;
- To provide dedicated visualizations effective for different types of individual clustering results.
We designed XCluSim based on the visual information seeking mantra (i.e. overview first, zoom and filter, and details-on-demand)to better support scalable visual comparison. Since each combination of different clustering algorithms and their parameters may yield different clustering results, it is inevitable from those many clustering results to 1) see their overall similarity first, 2) choose a subset of them, and then 3) perform detail comparisons and explore individual clustering results.
XCluSim provides as many clustering options as possible by implementing famous clustering algorithms and linking the clustering algorithms available in Weka [bib_ref] The WEKA data mining software: an update, Hall [/bib_ref]. It also keeps track of clustering options that users try during the analysis process.
In the following subsections, we introduce visualization techniques and user interactions for comparison tasks. They include overview, filtering/selection, and detail view. Then we present visualization techniques that help users to explore individual clustering results. For better comprehension of the visualization components in XCluSim, we first describe a color encoding strategy for clusters, which we consistently apply to every visualization component of XCluSim prior to explaining each visualization.
## Color encoding of clusters using tree colors
To help users identify similarities among multiple clustering results, we color-code each cluster based on Tree Colors [bib_ref] Tree Colors: Color Schemes for Tree-Structured Data, Tennekes [/bib_ref] , which provides a color-coding scheme for tree-structured data. We first hierarchically cluster all clusters from every clustering result using HAC. The correlation coefficient is used as the similarity measure between a pair of clusters as in [bib_ref] Visually comparing multiple partitions of data with applications to clustering, Zhou [/bib_ref]. This maintains consistency in the use of the cluster similarity measure in XCluSim, which is also used for rearranging bands (i.e. clusters) in the enhanced parallel sets view (see the Enhanced parallel sets view section). In the resulting tree-structured cluster hierarchy, we assign an appropriate color to each cluster based on the Tree Colors colorcoding scheme so that similar clusters have similar colors.
This color encoding helps users intuitively assess the similarity of clusters. For example, in (the enhanced parallel sets view), ① and ② have very similar colors while ① and ③ do not, which means that ① and ② share most items while ① and ③ barely share any items. This color-coding scheme is consistently applied to overviews, detail views, and every visualization for individual clustering results.
## Overview of all clustering results
## Parameter information view
XCluSim provides an overview of parameters for all clustering results in the parameter information view , 2A). This view is vertically divided into subsections, each of which corresponds to an individual clustering algorithm (e.g. "K-means clustering"). Inside each subsection, there are multiple bar charts arranged in a matrix layout. Each bar chart shows the number of clustering results generated by the corresponding algorithm with the corresponding parameter setting. For example, in , the parameter information view is divided into more than four subsections (some subsections are hidden under the scroll view) since a user made clustering results using algorithms such as HAC, self-organizing map (SOM) clustering, K-means clustering, and expectation-maximization (EM) clustering. As shown in , the bar in the left bottom cell of K-means clustering is taller than any bars shown in any clustering algorithms, indicating that the K-means clustering algorithm with a distance measure of Euclidean distance and with 9 as the number of clusters is the one mostly used . We note here that bioinformaticians often run a clustering algorithm multiple times Visualization techniques for comparing multiple clustering results in XCluSim. There are three types of overviews: (A) parameter information view, (B) force-directed layout overview, and (C) dendrogram overview. They enable users to simultaneously compare multiple clustering results in a scalable way. When some clustering results are selected in the overviews, they are added to (D) enhanced parallel sets view for more in-depth comparison tasks. Users can access the detailed information of the selected clustering results with each result in each tab of (E) the tabular list view. even with the same parameter setting when the algorithm (e.g. K-means) works non-deterministically. For more details on clustering parameters, the user can also look into the visualization of individual clustering results.
To help users determine which results to select for detailed analysis, XCluSim provides scalable similarity overviews both at the cluster level and at the clustering result level using a force-directed layout (FDL) and a dendrogram view. In the next two sections, we present details of these two overviews.
## Force-directed layout (fdl) overview
In the FDL overview, overall similarity relations among multiple clustering results are visualized in a force-directed layout, where more similar results are placed closer together and connected with thicker edges . 2B). The similarity metric for calculating distances between nodes is F-measure [bib_ref] Foundation of evaluation, Van Rijsbergen [/bib_ref] , which is the harmonic mean of the precision and recall measure. Each of the precision and recall measures for the two clustering results is calculated by dividing the number of agreed pairs of items by the number of all pairs of items belonging to a clustering result. An agreed pair refers to two items that "agree" to be clustered together in both clustering results.
Since the FDL overview uses physical distance to visually encode similarity between clusters, it has a perceptual advantage in revealing similarity relations among them. In addition, a pie chart is embedded in each node to enable users to visually estimate the number of clusters and their sizes. Since the global color encoding scheme also helps users to grasp similarities among clusters, users can estimate which clusters remain stable across different clustering results. For the scalability of the FDL overview, nodes become smaller as more results are added to the view. Moreover, an edge between two clusters is displayed only when similarity between the clusters exceeds a predetermined similarity threshold.
## Dendrogram overview
The overall similarity relations are also visualized in the dendrogram overview , 2C) after running an HAC with all clustering results (i.e. each row or node represents a result). As in the FDL overview, we use the F-measure as the distance measure between a pair of results. However, the visual representation and its purpose are different from the FDL overview. While the FDL overview intuitively shows similarities using physical distance, the dendrogram overview uses a more familiar clustering visualization component (i.e. a dendrogram) to represent similarities between clustering results. Moreover, the dendrogram overview is more space efficient so that users can see clustering results and cluster distributions more clearly without occlusion.
## Visualization for comparing select clustering results
When users identify clustering results of their interests in the overview of all results, they want to select them and perform more in-depth comparison with them. In the next two subsections, we introduce visualizations for comparing the selected clustering results: the enhanced parallel sets view and the tabular list view. When a user selects a result either in the FDL or dendrogram overviews, the selected result is added to the enhanced parallel sets view for more in-depth comparison. The tabular list view, located on the rightmost side of XCluSim, enables users to access detailed information of the selected clustering results with each result in a separate tab.
## Enhanced parallel sets view
To visualize the concordance and discordance of multiple clustering results in more detail, we utilized parallel sets [bib_ref] Parallel sets: visual analysis of categorical data, Kosara [/bib_ref]. We enhanced the parallel sets for effective clustering result comparison by designing more appropriate interactions and revealing more relevant information, i.e., stable group (explained in detail later in this section). In the parallel sets view , each horizontal row of stacked bars represents a clustering result. A tiny gap is placed between each bar to assist users to correctly perceive a single cluster since adjacent bars can occasionally have similar colors when the Tree Colors scheme is used. Rows are arranged in such a way that the distance between adjacent rows encodes the dissimilarity between the corresponding clustering results. Each horizontal bar in a row represents a cluster in the corresponding result. We define a stable group of items as a set of items that are clustered together through all selected clustering results. A stable group is represented as a ribbon-like band across all rows. Since the parallel sets view only enables comparisons based on a linear ordering of results, users can interactively switch any two rows by dragging one over the other. When the vertical order of the rows is changed, all rows are replaced accordingly to reflect the similarity between new adjacent clustering results.
The aggregated band representation for links connecting items in a stable group significantly reduces visual clutter compared to the use of a single line representation to connect individual items. The width of a band is an important visual cue that encodes important information about a stable group (i.e. its size) in XCluSim. Users can easily recognize the largest groups of items that are clustered together across multiple clustering results as they spot thick bands. Moreover, users can visually estimate the stability of a cluster by looking at the width of each stable group in it. For example, since the average width of stable groups in ① is bigger than ② in , a user can infer that ① is a more stable cluster than ②. Cluster-similarity based on the color-coding of bars (i.e. clusters) helps to facilitate the comparison of multiple clustering results.
However, the aggregation method could still suffer from clutter due to band-crossings. We applied a rearrangement algorithm [bib_ref] Visually comparing multiple partitions of data with applications to clustering, Zhou [/bib_ref] to address this issue. To provide more flexible user interaction depending on a user's need, we divided the algorithm into two rearrangement features: rearranging clusters (i.e. bar rearrangement) and rearranging their members (i.e., band rearrangement). These features can be evoked by pressing on the button at the bottom of the enhanced parallel sets view . When a user uses any of these two features, smooth animated transition is supported to reduce the cognitive burden that accompanies users' attempts to trace the movement of bands or bars.
XCluSim provides more user interactions to overcome the cluttering problem. First of all, users can alleviate the visual clutter in the region of interest by rearranging the bars in a row. This involves dragging them horizontally. After manually rearranging bars (i.e. clusters), users can The enhanced parallel sets view with various user interactions for in-depth comparison. (a) The parallel sets view provides rearranging algorithms that minimize line crossings. (b) When users hover a mouse pointer over the node of a cluster, the stable groups contained in it are highlighted while other stable groups fade out to reveal flows more clearly. By using a filtering feature on the stable group histogram at the bottom of the parallel sets view, users can hide less interesting bands. (c) Moreover, by using common angle plot [bib_ref] Common angle plots as perception-true visualizations of categorical associations, Hofmann [/bib_ref] , users can compare the sizes of different bands more accurately. employ the band rearrangement feature to reduce the visual clutter of bands across multiple rows due to the current manual arrangement of bars in the row. Secondly, there is a band filtering feature similar to that in. The stable group histogram at the bottom of shows the distribution of bands by size. There are two blue filtering bars on both sides. Users can filter out bands that are too small or too big from the parallel sets view by adjusting the position of the filtering bars. Finally, when the mouse pointer hovers over a cluster, it highlights the bands, allowing the clusters to show their flows across other clustering results clearly . This can be helpful when a user is especially interested in stable groups that belong to a specific cluster.
The perception of a stable group's size could be distorted by a line width illusion [bib_ref] Common angle plots as perception-true visualizations of categorical associations, Hofmann [/bib_ref]. Such an illusion causes humans to perceive line width incorrectly at slanted angles. This distortion may disrupt the task of band size comparison. In order to prevent it, we adopt the common angle plot [bib_ref] Common angle plots as perception-true visualizations of categorical associations, Hofmann [/bib_ref] idea . By comparing the straight, vertical parts of bands, users can compare the sizes of the stable groups more accurately. However, since the common angle plot represents a single line as three connected straight lines, it may generate more clutter and occlusions. Thus, it is better to use this feature when only a small number of bands are displayed in the parallel sets view.
## Tabular list view
Users can access detailed information concerning the selected clustering results with each result in a separate tab in the tabular list view . The tabular view provides detailed information in two different modes: the group-by mode and the heatmap mode. In the group-by mode, users can see the data grouped by stable groups or by clusters. A group is represented by a representative item in a single row with the number of group members between parentheses. Moreover, there is a line graph glyph in each row to show the overall average pattern of the corresponding group. In the heatmap mode, the tabular list view shows numerical details with each cell color-coded according to its value. There is a text search field on top of the tabular list view so that users can directly access specific items. A user can export a selected subset of data (e.g. a specific stable group) as a CSV text file for further analysis.
XCluSim provides brushing and linking among all visualization components. Thus, the tabular list view is coordinated with all visualization components in XClu-Sim. Thus, whenever a user selects a group of items in any visualization, they are highlighted in the tabular list view to help the user access detailed information about them. In addition, when the mouse pointer hovers over an item in a component, it highlights the item in whiteblue color, and all related items on the other components are also highlighted. This could lead to additional meaningful insights. For example, hovering a mouse pointer over the title of a specific algorithm in the parameter information view results in the highlighting of all related clustering results in overviews and detail views . As a consequence, users are able to understand that K-means clustering can produce totally different clustering results depending on the clustering parameters chosen (e.g. compare "K-means clustering (10)" to "K-means clustering(11)" in the dendrogram overview in .
## Figure 4
The tabular list view enables users to access numerical details. (a) Users can see detailed information for each item grouped by cluster or stable group. (b) Users also can see raw data in a heatmap form. When a user wants to access an item or a group directly, he/she can use the search box provided on top of the tabular list view.
## Interactive data manipulation
Simple file formats such as comma separated values (CSV) and tab-delimited text are used for XCluSim. XCluSim enables researchers to interactively manipulate the input dataset when loading it, prior to clustering it . Users can generate a ratio value by selecting two columns from the original dataset. XCluSim provides filters such as a range filter and RPKM threshold adjustment. It also provides features for calculating fold changes.
## Visualization for individual clustering results
To make XCluSim a more general visual analytics tool for comparing clustering results, we try to provide a wide variety of clustering algorithms. First of all, we implement frequently-used clustering algorithms in Interactive manipulation of input data supported by XCluSim: derive a new column (ratio, fold change), change color mapping, filter items using a range filter and RPKM adjustment.
XCluSim. These include Hierarchical Agglomerative Clustering [bib_ref] Cluster analysis and display of genome-wide expression patterns, Eisen [/bib_ref] , SOM clustering, K-means clustering, and OPTICS clustering [bib_ref] OPTICS: ordering points to identify the clustering structure, Ankerst [/bib_ref]. Moreover, all clustering algorithms from Weka [bib_ref] The WEKA data mining software: an update, Hall [/bib_ref] are also available in XClu-Sim. Users can also import any clustering results made by any other clustering algorithms that are not available in XCluSim.
## Taxonomy of visualization techniques for visualizing clustering results
Different clustering algorithms work on different principles. For example, there are three major categories of clustering algorithms: hierarchical, partitional, and density-based. Clustering algorithms in different categories need different visualization techniques to effectively visualize their clustering results.
To suggest effective visualizations for each category of clustering algorithms, we first surveyed visual encoding techniques for visualizing the clustering results of various algorithms [fig_ref] Table 1: Taxonomy of visualization techniques for visualizing clustering resultsVisualization components for visualizing clustering... [/fig_ref]. Sedlmair et al. presented a related taxonomy of factors in visual cluster separation [bib_ref] A taxonomy of visual cluster separation factors, Sedlmair [/bib_ref]. They evaluated the effect of each factor on visual cluster separation in scatterplots. Building upon this work, we consider the appropriateness of visual encoding techniques in representing the characteristics of each type of clustering algorithm. To broaden the perspective of our taxonomy, we further categorize the visual encoding techniques in terms of Gestalt principles of grouping: similarity, proximity, connectedness, and enclosure.
Similarity: The similarity principle is the one most commonly used in cluster visualization. It helps users to perceive cluster membership by employing similar colors, shapes, or sizes. Among them, color is the most frequently used visual cue. However, using color as the main visual cue may not scale well because the use of human color perception to discriminate between classes is limited to a number of colors. Thus, it is often used in conjunction with visual cues such as in reachability plot [bib_ref] The WEKA data mining software: an update, Hall [/bib_ref] and silhouettes plot [bib_ref] iGPSe: A visual analytic system for integrative genomic based cancer patient stratification, Ding [/bib_ref].
Proximity: This principle facilitates the perception of cluster membership by placing related items closer together. For example, in the silhouettes plot [bib_ref] Silhouettes: a graphical aid to the interpretation and validation of cluster analysis, Rousseeuw [/bib_ref] , bars belonging to the same cluster are placed next to each other. However, this principle is not used alone. It is typically used together with other visual cues. For example, the partitioned heatmap sometimes puts gaps between clusters to show their boundaries clearly [bib_ref] Interactively exploring hierarchical clustering results, Seo [/bib_ref] [bib_ref] Comparative analysis of multidimensional, quantitative data, Lex [/bib_ref] [bib_ref] StratomeX: visual analysis of large-scale heterogeneous Genomics data for cancer subtype characterization, Lex [/bib_ref].
Connectedness: The connectedness principle helps users to identify groups by connecting related items using a visual artifact such as a line. Line connection is one of the most powerful visual cues among the Gestalt principles of grouping. However, it can confuse users when there are too many lines in a single view. The connectedness principle is especially used with hierarchical clustering results since hierarchy structures can best be demonstrated with line connections. For example, HCE [bib_ref] Interactively exploring hierarchical clustering results, Seo [/bib_ref] , Matchmaker [bib_ref] Comparative analysis of multidimensional, quantitative data, Lex [/bib_ref] , and others use this principle to represent clusters in dendrograms.
Enclosure: The enclosure principle is adopted particularly when drawing a closed boundary containing items belonging to a cluster. For example, when a dataset contains spatial information, all items of a cluster are shown on a color-coded region with a solid boundary [bib_ref] Visual Analytics for Comparison of Ocean Model Output with Reference Data: Detecting..., Kothur [/bib_ref] [bib_ref] Visual analytics for spatial clustering: Using a heuristic approach for guided exploration, Packer [/bib_ref]. Another typical technique based on this principle is the partitioned heatmap [bib_ref] StratomeX: visual analysis of large-scale heterogeneous Genomics data for cancer subtype characterization, Lex [/bib_ref]. It is a powerful way to display raw data while clearly specifying the boundary surrounding the members of each cluster. In addition to these four Gestalt principles of grouping, there are some attempts to use abstract representations (such as glyphs or special shapes) for clusters without showing any individual items in clusters. The cluster graph [bib_ref] Visualization of large hierarchical data by circle packing, Wang [/bib_ref] uses an abstract representation of a circular node for a cluster. Clusters derived from SOM clustering results are visualized in a hive-shaped grid view while each item is abstracted as a node. As these attempts do not allow for the visualization of individual items, they are not a good fit for the classification based on Gestalt principles.
After reviewing and categorizing visual encoding techniques for visualizing clustering results, we selected visualization techniques appropriate for visualizing each of three main kinds of clustering algorithms (namely, hierarchical clustering, partitional clustering, and density based clustering). In the next three subsections, we describe the visualization techniques in detail.
## Visualization technique for hierarchical clustering
We visualized HAC results with the combination of a dendrogram and heatmap visualization [fig_ref] Figure 6: Visualization techniques for individual clustering results in XCluSim [/fig_ref] , where users could interactively compress/expand, flip, and swap sub-trees. The batch compression of sub-trees using the minimum similarity bar [bib_ref] Interactively exploring hierarchical clustering results, Seo [/bib_ref] is also possible. By adjusting the position of the similarity bar, users can dynamically determine the clusters. There is a compact bird's-eye overview using heatmap [bib_ref] Caleydo: Design and Evaluation of a Visual Analysis Framework for Gene Expression..., Lex [/bib_ref] in the left most part which is tightly coupled with the dendrogram. By dragging a black-bordered rectangle that represents the current viewport (see the black rectangle in the top left of [fig_ref] Figure 6: Visualization techniques for individual clustering results in XCluSim [/fig_ref] in the heatmap overview, users can efficiently navigate through the dendrogram+heatmap view.
## Visualization technique for partitional method
Partitional clustering results other than SOM clustering (e.g. K-means clustering, EM clustering, farthest first clustering, etc.), and all imported results are visualized in a force-directed layout [fig_ref] Figure 6: Visualization techniques for individual clustering results in XCluSim [/fig_ref] , where each cluster is represented as a rectangle whose size is proportional to the cluster size. The force between nodes is determined by the similarity between members of each cluster so that similar clusters are closely positioned and have thicker links between them. To show an overview of a cluster, XCluSim also visualizes the average pattern of all members of the cluster in a line chart, which is shown as a glyph in the cluster's node. XCluSim also supports semantic zooming to enable users to explore clusters in more detail. When a cluster is zoomed into, more details of its members are dynamically visualized in a parallel coordinate plot.
SOM clustering results are visualized using the typical hive-shaped visualization [fig_ref] Figure 6: Visualization techniques for individual clustering results in XCluSim [/fig_ref] , where each hexagonal cell represents a cluster. In XCluSim, the background intensity of each cell represents the size of the corresponding cluster. As a visual summary of each cluster, XCluSim presents the average pattern of the cluster members in a line chart within each hexagonal cell. XCluSim also supports semantic zooming. Users can zoom into a cluster by double-clicking on the corresponding cell and look at the details of their members in a parallel coordinate plot in the same way they would in a force-directed layout.
## Visualization technique for density-based method
Density-based clustering algorithms calculate a kind of density-related information for each item during the clustering process. For example, OPTICS [bib_ref] OPTICS: ordering points to identify the clustering structure, Ankerst [/bib_ref] calculates the reachability distance for each item. We believe that users can more intuitively understand a density-based clustering result when the density-related information is revealed. Therefore, a bar-chart-like visualization, with each item arranged on the horizontal axis and the density-related information on the vertical axis, can effectively visualize density-based clustering results. The conventional reachability plot for OPTICS is a typical example. In XCluSim, we enhance the plot for better cluster identification and for improved examination of details [fig_ref] Figure 6: Visualization techniques for individual clustering results in XCluSim [/fig_ref]. To clearly show the position of each cluster, XCluSim places a horizontal bar from the start to the end positions of the cluster right below the reachability plot. The parallel coordinate plot at the bottom shows more details of cluster members. These two plots support brushing and linking between the cluster members. For example, when a mouse pointer hovers over a cluster in the reachability plot, the lines for the members of the cluster are highlighted in the parallel coordinate plot.
## Implementation
XCluSim was developed using Java Standard Edition 7 (Java SE 7), which enables it to run on any platform with JRE version 1.7 or higher. We used the Piccolo 2D framework to implement visualization components and interactions. Weka's clustering algorithms were integrated into XCluSim using Weka SDK 3.6 [bib_ref] The WEKA data mining software: an update, Hall [/bib_ref].
# Results and discussion
## Case studies
To evaluate the efficacy of XCluSim, we conducted two case studies with our collaborator in a major bioinformatics research laboratory. He is a senior research engineer and has years of experience in genome and transcriptome analyses.
Elucidating the role of ferroxidase in cryptococcus neoformans var. grubii H99 (case study 1)
This study was carried out in his laboratory for 80 minutes. Pre-and post-study interviews were conducted for 10 minutes each. The participant used XCluSim for 50 minutes after a 10-minute tutorial. We used a dataset containing normalized expression levels of 6,980 genes belonging to the Cryptococcus neoformans var. grubii H99 strain. The dataset had been prepared for his previous work [bib_ref] A defect in iron uptake enhances the susceptibility of Cryptococcus neoformans to..., Kim [/bib_ref].
His task was to elucidate the role of ferroxidase (cfo1) by knocking it out. He was interested in finding a meaningful set of genes whose expression would be influenced and in identifying the affected pathways. For the task, he tried to see the effect of fluconazole on two different strains: the wild type of Cryptococcus neoformans var. grubii H99 and the cfo1 mutant of the same strain. In the dataset, each gene has four expression levels: two different strains, each cultured in two conditions (i.e. wild-type strain and cfo1 mutant with and without fluconazole treatment).
When he loaded the data, he made four new data columns of ratio values, including the wild-type strain with fluconazole versus the wild-type strain without fluconazole treatment (WT+F/WT-F) and the cfo1 mutant with fluconazole versus the cfo1 mutant without fluconazole treatment (MT+F/MT-F) . Subsequently, he adjusted the RPKM threshold and used log fold changes to filter out less interesting genes for more efficient analysis.
After data pre-processing, XCluSim showed the results of three clustering algorithms (i.e. HAC, SOM clustering, K-means clustering) in three independent views. Since he was most familiar with dendrogram and heatmap visualization, he examined the HAC results first. He was interested in genes that were highly expressed with fluconazole treatment. Among them he found the gene named Erg11 (CNAG_00040). He said that this gene was reported to be associated with azole resistance.
Next, he tried to see which genes were stably grouped together across different clustering results. He tried to load as many clustering results as possible to see the differences between them. The parameter information view provided him with a good overview of all clustering results (clustering algorithms and their parameters). He was able to make diverse clustering results without generating any duplicate results.
After generating 15 different clustering results, he selected four diverse results from the FDL overview to find out which genes were clustered together with Erg11. However, he recognized that the stable groups were excessively thin because of the result named "FarthestFirst [bib_ref] Visually comparing multiple partitions of data with applications to clustering, Zhou [/bib_ref]." This had to do with the fact that it was the most dissimilar result to other selected clustering results . So he removed that result from the parallel sets view. Then he selected a more similar one named "KMeans Clustering(4)" . He subsequently accessed the stable group with Erg11 directly, utilizing the search feature in the tabular list view. He was able to confirm that 17 other genes belonged to the stable group. After validating the members of the stable group with an enrichment analysis, he found that most of them (10 out of 18) belonged to the ergosterol biosynthetic pathway.
Once he had selected the stable group in the tabular list view, he was able to efficiently inspect the flow of the group across different clustering results in the enhanced parallel sets view . While he looked into the flow of the stable group across all rows (the rightmost highlighted-band in , he also noticed that the clustering result from "KMeans Clustering(4)" had the tightest cluster, which included the stable group. However, there were no more genes outside the stable group in the cluster that belonged to the ergosterol pathway.
Then he tried to find the best algorithm and those of its parameters that gave the tightest cluster containing genes belonging to the ergosterol pathway. Since "KMeans Clustering(4)" had previously been the best clustering result among the selected results, he ran K-means clustering algorithms with different parameters to arrive at similar results. He then inserted three of the most similar results in the parallel sets view . Again, he highlighted a stable group with Erg1 (the band indicated with a red arrow in . By checking the flow of the stable group crossing each result, he recognized that "KMeans Clustering [bib_ref] Silhouettes: a graphical aid to the interpretation and validation of cluster analysis, Rousseeuw [/bib_ref] " gave the tightest cluster. This led to the conclusion that K-means clustering with the corresponding parameter configurations (i.e. Euclidean distance as the distance metric and 9 as the number of clusters) was the best result for the given dataset among all the results.
Finding a clustering result that clearly represents biological relations (case study 2)
A second case study was subsequently carried out with the same participant in his laboratory. The study was conducted for 150 minutes on a different day. Since the participant was already familiar with XCluSim, we skipped the tutorial. In the study, he relied on the gene expression profiles of 169 genes in Escherichia coli, which used a DNA microarray [bib_ref] DNA microarray analysis of gene expression in response to physiological and genetic..., Khodursky [/bib_ref]. In the dataset, each gene contained 19 expression levels in order to investigate the effects of the perturbations on tryptophan metabolism. The expressions were measured under the following conditions: wild type growth with and without tryptophan (five conditions), wild type growth with and without tryptophan starvation (nine conditions), and the growth of wild type and a trp repressor mutant (five conditions).
Through the case study, the participant wanted to find a clustering result that clearly reflected biological relations in tryptophan metabolism. In the original paper [bib_ref] DNA microarray analysis of gene expression in response to physiological and genetic..., Khodursky [/bib_ref] , the authors used HAC to cluster the 169 gene expression profiles measured in the 19 conditions. It was indicated in the paper that genes showing similar expression responses did not necessarily fall into the same cluster. One example included the genes associated with aromatic amino acid metabolism.
He first wanted to see if the optimal algorithm and its parameters in the previous case study would work for another dataset. To determine this, he produced 11 clustering results in XCluSim, including the result produced using previous optimal settings: K-means clustering with Euclidean distance as the distance metric and 9 as the number of clusters. He validated each cluster in the result ("KMeans Clustering(6)" in through an enrichment analysis using the DAVID website (http://david.abcc.ncifcrf.gov/). After validating each cluster, he concluded that most of the clusters were grouped well in the sense that they represented biological relations in pathways. However, he recognized two problems in the result. First of all, a cluster that had both Arg and Art regulons also contained a gene named tnaA that was considered to be noise. This was because tnaA showed a different expression pattern and was not highly related to other cluster members in biological terms. Secondly, one gene from the fli operon, fliS, fell into a different cluster from the other genes in the same operon while they had homogeneous expression patterns.
By utilizing visualizations in XCluSim, he wanted to find the clustering result that properly represented biological relations as "KMeans Clustering(6)" while the two problems were revisited. For this intended task, he selected all the similar results from the FDL overview: "KMeans Clustering(5)", "KMeans Clustering(8)", and "KMeans Clustering". Then he accessed the stable groups that contained tnaA and the Arg/Art regulon. He easily recognized that genes in both the Arg and Art regulons fell into same stable group while tnaA was not stably clustered with them. The results, which separately clustered tnaA from the Arg and Art regulons, were "KMeans Clustering(5)" and "KMeans Clustering(8)". Similarly, by checking the flow of stable groups in each horizontal row, he easily recognized that two clustering results that used the correlation coefficient as a distance metric clustered two stable groups together: one with the fli operon and the other with fliS. The two results were "KMeans Clustering(5)" and "KMeans Clustering". As a consequence, "KMeans Clustering(5)", using the correlation coefficient as the distance metric and 13 as the number of clusters, was the most satisfying result for the dataset.
Additionally, our participant gained insight by seeing a stable group in XCluSim. Genes in the trp operon (i.e. trpE, trpD, trpC, trpB, and trpA) were stably clustered together with yciF through the four different results (see the highlighted stable group in . Since yciF was assigned to a putative function, he said that the gene might be closely related to tryptophan synthase as a trp operon.
After he found the best result, he compared it with a clustering result provided in the original work [bib_ref] DNA microarray analysis of gene expression in response to physiological and genetic..., Khodursky [/bib_ref] to see if his result better represented biological relations . The clustering result presented in the paper had been prepared prior to the study and was imported to XCluSim for visual comparisons. After comparing two results, he found that some of the genes involved in aromatic amino acid metabolism, aroF, tyrA, aroL, and aroP, were clustered together in our best result while only three of them fell into the same cluster in their original result. Moreover, their result did not cluster fliS with the other fli operon. These results suggested that the authors of the original work [bib_ref] DNA microarray analysis of gene expression in response to physiological and genetic..., Khodursky [/bib_ref] could have generated more biologically meaningful results if they had used XCluSim in the first place.
# Discussion
During the case studies, we received positive subjective feedback on XCluSim from the participant. He especially liked the ability to identify stable groups across multiple clustering results. Moreover, he was satisfied that he could select and run diverse clustering algorithms and interactively compare them by adding/removing a clustering result to/from the enhanced parallel sets view. He could quickly shift his attention to a more interesting set of results for more in-depth comparison. However, he also pointed out the limitations of XCluSim. Since filtering sets of items was only available at the data manipulation step, he said it would be helpful to allow users to interactively filter raw data in the visualization components as well.
We color-coded each cluster consistently across the whole system using the Tree Colors scheme after building a hierarchical structure of all clusters from multiple clustering results. With the help of this color coding, Results of the second case study are visualized in the enhanced parallel sets view. (a) The highlighted stable group contained the trp operon with yciF. (b) Visual comparison of two results: the best clustering result ("KMeans Clustering(5)") derived from the case study and a result ("A Result from Original Paper") presented in the original research paper [bib_ref] DNA microarray analysis of gene expression in response to physiological and genetic..., Khodursky [/bib_ref].
overviews became even more useful in XCluSim. While the color encoding was applied for a specific purpose in this work (i.e. for the visualization of clusters), we think it can also be applied to parallel sets applications in a more general and scalable way. For example, instead of distinguishing only a small number of categories while visualizing a categorical dataset, it might be possible to distinguish many more nodes in the parallel sets once a hierarchical structure of the nodes has been built in a similar manner to the one we employed in XCluSim.
We provided a taxonomy of visualization techniques for visualizing clustering results based on the Gestalt principle of grouping and the types of clustering algorithms. The design space defined by this taxonomy can help researchers to make design decisions for clustering results visualization. By thinking about visualization techniques in terms of the Gestalt principle, researchers can come up with better visual encoding without overlooking important features. For example, since the graph layout is used to visualize cluster memberships by colorcoding each item [bib_ref] iGPSe: A visual analytic system for integrative genomic based cancer patient stratification, Ding [/bib_ref] [bib_ref] Interactive visual clustering of large collections of trajectories, Andrienko [/bib_ref] , one can also utilize the enclosure principle (like GMap [bib_ref] GMap: Visualizing graphs and clusters as maps, Gansner [/bib_ref] and BubbleSets [bib_ref] Bubble sets: Revealing set relations with isocontours over existing visualizations, Collins [/bib_ref] to represent their membership more clearly.
## Future work
At present, when a clustering algorithm does not assign all items to clusters, all un-clustered items are treated as a single cluster in XCluSim. OPTICS and DBSCAN clustering algorithms can give rise to results of this kind. XClu-Sim treats un-clustered items as a group of less interesting items as if it were a special cluster. Otherwise, it could make a huge number of stable groups since each un-clustered item will become a single stable group. This would make it hard for users to gain insight from visualizations. In the future, we plan to improve XCluSim to resolve this problem. For example, we can represent these kinds of groups with different textures in the parallel sets view to distinguish them from other normal clusters.
In this paper, we concentrated mostly on supporting comparison tasks based on the concordance/discordance of multiple clustering results. However, since bioinformaticians' cluster analysis is highly integrated with the validation stage, it would also be valuable to provide a visual representation of cluster validity measures (e.g. internal cluster validity indices). For example, the gray scale intensity of each band (i.e. stable group) in the parallel sets view, which currently represents the size of a stable group, can be utilized to represent its internal validity measures. In such a case, stable group provided by XCluSim will become more reliable information.
# Conclusion
In this paper, we presented XCluSim, a visual analytics tool that enables users to compare multiple clustering results. XCluSim provides three different overviews to help users grasp their overall similarity relationships in a more scalable and flexible way. Moreover, the enhanced parallel sets view enables users to detect differences among select clustering results even more clearly by using improved user interactions. To help users not only compare but also explore individual clustering results more effectively, we proposed dedicated visualizations for each of the three distinctive classes of clustering algorithms. To design them, we defined a design space for clustering results visualization by building a visualization taxonomy based on the Gestalt principles of grouping. This taxonomy could be useful for other researchers when they design new visualizations for a clustering result. We conducted case studies to evaluate the usefulness of XCluSim, and the participants gave positive feedback.
List of abbreviations HAC: Hierarchical agglomerative clustering; SOM: Self-organizing map; EM: Expectation-maximization; FDL: Force-directed layout.
[fig] Figure 2: Three overviews supported in XCluSim. (a) The parameter information view provides the parameter settings used for the clustering results produced. The table in the parameter information view is for a clustering algorithm, and it shows a bar in each cell to represent the number of clustering results using the corresponding parameter setting. (b) The force-directed layout overview intuitively shows similarity among multiple clustering results with the distance between nodes representing similarity. (c) The dendrogram overview shows similarities between clustering results in a familiar dendrogram layout with a clustering result visualized at a terminal node. [/fig]
[fig] Figure 6: Visualization techniques for individual clustering results in XCluSim. (a) Dendrogram+heatmap visualization for hierarchical agglomerative clustering results. (b) Force directed layout for every partitional clustering result and imported clustering results. (c) Common hive-shaped visualization for SOM clustering results. (d) Reachability plot together with parallel coordinate plot for OPTICS. [/fig]
[table] Table 1: Taxonomy of visualization techniques for visualizing clustering resultsVisualization components for visualizing clustering results use visual cues based on Gestalt principles of grouping[27] to represent cluster membership. We categorize the visualization components by principle and indicate how appropriate each visualization component is for showing clustering results by different types of clustering algorithms. (i.e. "O" for most appropriate, "Δ" for moderately appropriate, and "X" for not applicable). [/table]
|
The duality between particle methods and artificial neural networks
the algorithm behind particle methods is extremely versatile and used in a variety of applications that range from molecular dynamics to astrophysics. for continuum mechanics applications, the concept of 'particle' can be generalized to include discrete portions of solid and liquid matter. this study shows that it is possible to further extend the concept of 'particle' to include artificial neurons used in Artificial Intelligence. This produces a new class of computational methods based on 'particleneuron duals' that combines the ability of computational particles to model physical systems and the ability of artificial neurons to learn from data. The method is validated with a multiphysics model of the intestine that autonomously learns how to coordinate its contractions to propel the luminal content forward (peristalsis). Training is achieved with Deep Reinforcement Learning. The particleneuron duality has the advantage of extending particle methods to systems where the underlying physics is only partially known, but we have observations that allow us to empirically describe the missing features in terms of reward function. During the simulation, the model evolves autonomously adapting its response to the available observations, while remaining consistent with the known physics of the system. open
The family of particle-methods comprise of a variety of computational techniques that range from molecular dynamics to computational astrophysics. Despite the different fields of application, however, almost all particle methods follow the same algorithm [fig_ref] Figure 1: Functioning of particle methods and link with Artificial Neural Networks [/fig_ref]. Firstly, the computational domain is subdivided in discrete entities (i.e. computational particles) that, according to the circumstances, represent atoms, portions of matter, solid objects, or even entire stars. Secondly, the forces exchanged among the particles are calculated. Finally, based on these forces, the particles' positions are updated by solving Newton's equation of motion. The difference between one particle method and another lies essentially on how these forces are calculated and what type of physical phenomena are represented. For instance, in Molecular Dynamics (MD), they represent the attractive and repulsive forces between atoms 1,2 ; in the Lattice Spring Model (LSM) or in Peridynamics (PD), the elastic forces occurring in solids [bib_ref] Peridynamic bond-associated correspondence model: stability and convergence properties, Chen [/bib_ref] [bib_ref] Higher toughness of metal-nanoparticle-implanted sodalime silicate glass with increased ductility, Ono [/bib_ref] ; in Smooth Particle Hydrodynamics (SPH), the pressure and viscous forces occurring in fluids [bib_ref] Mesoscale modelling of miscible and immiscible multicomponent fluids, Zhao [/bib_ref] [bib_ref] Smoothed particle hydrodynamics method for fluid flows, towards industrial applications: motivations, current..., Shadloo [/bib_ref] ; and, in direct gravitational N-body simulations, gravity between stars 7 . Because of their common algorithm, particle-methods can also be coupled together in Discrete Multiphysics (DMP) simulations [bib_ref] Numerical simulations of MitraClip placement: clinical implications, Kamakoti [/bib_ref] [bib_ref] Discrete multi-physics simulations of diffusive and convective mass transfer in boundary layers..., Ariane [/bib_ref] [bib_ref] Discrete multiphysics: a mesh-free approach to model biological valves including the formation..., Ariane [/bib_ref] [bib_ref] Using discrete multi-physics for detailed exploration of hydrodynamics in an in vitro..., Alexiadis [/bib_ref]. Fluid structure interactions, for instance, combine solid and liquid particles: the LSM calculates the forces occurring in the solid, SPH the forces in the fluid; and additional forces (e.g. repulsive forces to prevent compenetration between phases) model the solid-liquid interface. Moreover, DMP simulations can include heat or mass transfer [bib_ref] Solidification using smoothed particle hydrodynamics, Monaghan [/bib_ref] [bib_ref] Assessment of smoothed particle hydrodynamics (SPH) models for predicting wall heat transfer..., Ng [/bib_ref] : the algorithm in [fig_ref] Figure 1: Functioning of particle methods and link with Artificial Neural Networks [/fig_ref] still applies, but updates properties other than position, and calculates physical interactions other than forces. In the case of heat transfer, for instance, the new property is temperature, and, besides forces, particles also exchange heat.
What makes particle methods so versatile is their inherently simple concept: discrete objects interact with each other and, by exchanging some type of information (e.g. forces or heat), update their properties (e.g. position or temperature). Given its fundamental nature, it should not come as a surprise that, sometimes in disguise, this concept is often found in other algorithms. This study shows that Artificial Neural Networks (ANNs) are one of those algorithms and it is possible to unify ANNs and particle methods under the same framework. Within this Scientific RepoRtS | (2020) 10:16247 | https://doi.org/10.1038/s41598-020-73329-0 www.nature.com/scientificreports/ framework, it is possible to define a particle-neuron dual that combines the ability of computational particles to correctly reproduce the physics with the ability of artificial neurons to autonomously learn how to improve the (physical) model during the simulation. A previous study of ours [bib_ref] Deep multiphysics: coupling discrete multiphysics with machine learning to attain self-learning in-silico..., Alexiadis [/bib_ref] coupled DMP with Reinforcement Learning. That article did not make use of ANNs. However, in the conclusions, it mentioned the particle-neuron duality as a theoretical possibility that could arise if Reinforcement Learning were replaced by Deep Reinforcement Learning. DMP was linked with ANNs in another study [bib_ref] Deep multiphysics and particle-neuron duality: a computational framework coupling (discrete) multiphysics and..., Alexiadis [/bib_ref] , where the duality was introduced in the context of a classification problem requiring Supervised Learning (i.e. cell-sorting in a microfluidic device). To the best of our knowledge this is the first time that the particle-neuron duality is established in the context of Deep Reinforcement Learning.
## The particle-neuron duality
As mentioned, particle methods function by exchanging forces among neighbouring particles. This can occur in different ways. Non-bonded particles (e.g. fluid particles) move during the simulation, and the neighbours, with which they exchange forces, are not always the same [fig_ref] Figure 1: Functioning of particle methods and link with Artificial Neural Networks [/fig_ref]. On the contrary, bonded particles (e.g. solid particles) only exchange forces with predetermined particles [fig_ref] Figure 1: Functioning of particle methods and link with Artificial Neural Networks [/fig_ref]. Additionally, particles can have different types of interactions; in heat transfer problems, for instance, they exchange heat instead forces [fig_ref] Figure 1: Functioning of particle methods and link with Artificial Neural Networks [/fig_ref].
We can look at ANNs from a similar angle. Similarly, to computational particles, artificial neurons are discrete computational elements: they do not exchange forces or heat, but general (i.e. not necessarily physical related) information [fig_ref] Figure 1: Functioning of particle methods and link with Artificial Neural Networks [/fig_ref]. They do not have position or temperature that varies under the effect of forces or heat, but more general outputs that varies under the effect of inputs. The algorithm for forward propagation in ANNs [fig_ref] Figure 1: Functioning of particle methods and link with Artificial Neural Networks [/fig_ref] is also surprisingly similar to the particle methods algorithm [fig_ref] Figure 1: Functioning of particle methods and link with Artificial Neural Networks [/fig_ref]. We can take advantage of these similarities and design a mathematical object that is, at the same time, a computational particle and an artificial neuron. When these objects exchange forces (or heat), they behave like computational particles; when they exchange non-physical related information, they behave like artificial neurons. We call these new mathematical objects particle-neuron duals and a mathematical model based on particle-neuron duals a Deep Multiphysics (DeepMP) model. In the next section, we discuss how this can be implemented in practice.
# Discussion and conclusions
Consider a fluid in a flexible pipe. The pipe can contract any of its parts squeezing the fluid completely or partially out of the contracted section [fig_ref] Figure 2: The DeepMP model combines a DMP model with an ANN [/fig_ref]. We can loosely think of this system as a section of the intestine and at the liquid as the luminal content. The objective of the DeepMP model is to learn, without being explicitly programmed to do so, how to simulate the process of peristalsis by coordinating its contractions and move the luminal fluid from left to right. Peristalsis is only one of the many motility patterns occurring in the intestine; the goal of the model is not to reproduce the complexity of the real intestine, but only peristaltic coordination. We choose peristalsis because this type of motility has the specific purpose of effectively propelling the contents distally. Thus, the expected outcome of the training is somehow predictable providing a visual baseline for evaluating the performance of the model.
The DeepMP model combines a DMP model that calculates the physics, and an ANN that learns peristalsis [fig_ref] Figure 2: The DeepMP model combines a DMP model with an ANN [/fig_ref]. The DMP model accounts for two types of particles, SPH particles that model the fluid and LSM particles that model the solid: details of the DMP model are given in the Methods section. The DMP model can contract specific sections, but it is not specifically programmed to move the fluid in any given direction. Random contractions would simply move the fluid back and forth; and only a well-coordinated sequence will result in the overall motion of the fluid's centre of mass.
In the intestine, coordination is controlled by the Enteric Nervous System (ENS), but the DMP model alone cannot replicate the effect of the ENS. To achieve this goal in our simulations, we provide the model with an 'artificial brain' [fig_ref] Figure 2: The DeepMP model combines a DMP model with an ANN [/fig_ref] that, based on the current status of the pipe, determines which section should be contracted www.nature.com/scientificreports/ next (and for how long). This artificial brain is a fully connected ANN with four layers that takes as input the particles contracting at time T (current state of the membrane) and output the particles that will contract at time T + ΔT (next action based on the current state of the membrane). In [fig_ref] Figure 2: The DeepMP model combines a DMP model with an ANN [/fig_ref] , the membrane particles subjected to contraction are 'activated' and given a value of 1; the non-contracting particles are 'inactive' and given a value of 0. This information is gathered into a vector of ones and zeros that represents the input layer of the ANN. The input layer passes the information to two hidden layers with 50 neurons each. After they process the information, the ANN outputs a vector of the same size of the input: each element of the vector corresponds to a particle of the membrane. If the value of the vector is 1, the membrane is contracted at that location. The ANN updates its state and decides a new course of action every ΔT = 1 s. It is also possible it continues squeezing the same section for longer; in this case, the input and the output vectors are the same. To take advantage the axial symmetry, we group membrane particles in concentric 'slices' (see "Methods" section).
The SPH particles that model the fluid are pure computational particles, since they only exchange forces (with each other and with LSM particles). The LSM particles that model the membrane are particle-neuron duals. They exchange forces (with each other and with SPH particles), but, since they also constitute the input and output layers of the ANN, they also exchange information with the hidden layers. The hidden neurons are pure computational neurons, since they only exchange information (with each other and with LSM particles). All particles, neurons and particle-neuron duals belong to the same computational domain, but the neurons are 'invisible' because do not have a position property that is updated by forces.
At this stage, the ANN does not yet perform correctly as it needs to be trained and learn an effective strategy for coordinating the contractions. This is different from traditional ANNs that are trained by adjusting the weights to maximize the ability of the network to classify a series of observations. In this case, the ANN must learn a policy that produces the best strategy to achieve a given goal. In order to train the network, we use Deep Reinforcement Learning (DRL) and, specifically, Q-learning (QL), which is the algorithm behind AlphaGo, the computer program that defeated the human word champion at the board game Go [bib_ref] Mastering the game of Go without human knowledge, Silver [/bib_ref]. The reader can refer to Neftci and Averbeck 17 and Lapan 18 for details on DRL and QL. In the Methods section, we discuss how DRL is applied to our specific case.
The DRL algorithm is designed to maximize the 'reward' during the training of the ANN. In our case, we define the reward as the variation of the centre of mass of the fluid (more details in the "Methods" section). [fig_ref] Figure 3: Training of the DeepMP model [/fig_ref] shows how the dimensionless cumulative reward changes during the training of the network. The dimensionless reward is defined as the ratio between the total change of volume of the contracting sections and the volume of liquid displaced forward. If the value of the dimensionless reward is close to one, the algorithm is very efficient in pushing the luminal content forward. If it is zero, the liquid is equally split in the two directions. If it is negative, the liquid overall flows backwards.
Initially, the ANN is not trained; hence this produces random, non-coordinated, contraction (see Video 1). After 2,000 episodes, the model develops an effective, but not yet optimal, strategy (see . Finally, after 10,000 episodes the model fully masters the task at hand (see Video 3). The dimensionless reward never reaches unity since, as Video 3 shows, a small amount of the liquid leaks back during the episode. The amplitude of the contraction is such that the tube closes almost completely, but, as it also occurs in human peristalsis [bib_ref] Insights into the mechanisms underlying colonic motor patterns, Spencer [/bib_ref] , this does not prevent that, from time to time, a small amount of liquid leaks back during the contraction. After training, the propulsion mode of the model is similar to that observed in the intestine of humans and experimental animals [bib_ref] Insights into the mechanisms underlying colonic motor patterns, Spencer [/bib_ref]. Moreover, in the human intestine [bib_ref] Insights into the mechanisms underlying colonic motor patterns, Spencer [/bib_ref] , the peristaltic wave travels at speeds between 0.1 and 1 cm s −1 ; in www.nature.com/scientificreports/ the model, the speed is around 0.6 cm s −1 . Another feature of [fig_ref] Figure 3: Training of the DeepMP model [/fig_ref] is the periodic drop of performance affecting the learning phase. This is a relatively common phenomenon in DRL known as catastrophic forgetting. There are various techniques to mitigate it, but they are not employed here since the performance after 10,000 episodes is considered satisfactory. This example shows how DeepMP provides a new paradigm for coupling physics-based models with Artificial Intelligence (AI). Commonly, if we know the physics behind a given phenomenon, we model it based on first-principles (FPs); if we do not know the physics, but we have (big) data, we use Machine Learning (ML). In practice, most engineering problems are in between: we know some of the physics, but not everything; we have some data, but not enough for ML. DeepMP is designed for this type of 'middle-ground' problems. It does not require perfect knowledge of the FPs because the AI component will improve the physical model during the simulation. It does not require big data, because the physical model will generate new data during the simulation. DeepMP only requires observations that help identifying the nature of the reward in the DRL algorithm. In this study, for instance, we can model the mechanics of both the membrane and the luminal content by FPs, but we cannot model the complexity of the Enteric Nervous System by FPs. DeepMP bypasses this issue. We know the goal of peristalsis is to efficiently move the luminal content forwards; this observation is enough to identify a reward for the DRL algorithm (i.e. variation of the centre of mass of the fluid). Once the reward is set, the ANN learns peristalsis based on the data generated by the FP model during the simulation.
Because of its features, DeepMP differs from any other technique that couples FP modelling with ML. These techniques can be divided in two groups. In the first group, the FP model generates data that are fed into the ML algorithm, e.g. for dimensionality-reduction [bib_ref] Deep learning approach based on dimensionality reduction for designing electromagnetic nanostructures, Kiarashinejad [/bib_ref] [bib_ref] Deep learning reveals underlying physics of light-matter interactions in nanophotonic devices, Kiarashinejad [/bib_ref]. In the second group, the ML algorithm generates a data-driven model that is fed into the FP model [bib_ref] Machine learning for active matter, Cichos [/bib_ref] [bib_ref] Knowledge discovery in nanophotonics using geometric deep learning, Kiarashinejad [/bib_ref]. These approaches have their benefits, but, because they link the FP model and the ML algorithm in series, they still require either a complete knowledge of the FPs or a large quantity of data. If the first element of the series is the FP model, complete knowledge of the physics is required; if it is the ML algorithm, big data are required. This is not the case of DeepMP that is designed for scenarios where we have partial knowledge of the FPs and limited amount of data.
This study also anticipates a question that is likely to become increasingly important in the next years: i.e. what is the most effective way to couple Multiphysics models with Reinforcement Learning? DRL is rapidly becoming the new frontier of ML. However, the physical models currently used by the Reinforcement Learning community to simulate the interaction between the agent and the physical environment are very simple, almost trivial, if compared with the tools developed in the last twenty years by the Multiphysics community. Often, physical models with less than 100 degrees of freedom are considered 'complex' and 'highdimensional' in DRL. The DMP model used in this study, which is small for multiphysics standards, accounts for 14,578 three-dimensional particles, which implies 43,734 degrees of freedom. So far, this question has remained somehow in the background because both Reinforcement Learning and Multiphysics are computationally intensive, and it is currently impractical to couple multiphysics models of millions of degrees of freedom with extensive DRL based on ANNs with millions of artificial neurons. However, things are changing rapidly: computational power is increasing at great strides and quantum computers are around the corner. Therefore, it is not too soon to ask the question on how complex multiphysics models and Reinforcement Learning should be best coupled together; Deep Multiphysics, based on the underlying duality between computational particles and artificial neurons, provides a first answer to this question. where h = 9 × 10 -3 m is the smoothing length, v ij the relative velocity between the two particles, and ρ ij = ρ j + ρ i . The parameter a depends on the kinematic viscosity, while b = 0.01 ensures the stability of the simulation. Periodic boundary conditions are used, i.e. particles that exit the computational domain from one side of the tube, re-enter from the other side. The flexible pipe is divided 2500 particles with mass 2.5 × 10 -4 kg modelled with the LSM [bib_ref] Dynamic patterns of compaction in brittle porous media, Guillard [/bib_ref]. Elastic forces between two particles i and j are calculated using Hooke's law where k = 9 × 10 -2 N m −1 is the Hookean constant, r 0 = 6 × 10 -3 m the equilibrium distance between the particles, and r ij their instantaneous distance. An additional Hookean spring with k′ = 5 × 10 -4 N m −1 is used to tether each particle of the membrane to its initial position and return the particles to their initial position after the contraction. Contractions are achieved by adding an axisymmetric force f 0 = 4 × 10 -4 N pointing towards the axis of the pipe to all particles of a given section. This section can be located everywhere along the tube, but its thickness is always one tenth of the pipe length. Fluid-solid interaction is achieved with a repulsive potential of the type with A = 2 × 10 -6 J, to avoid compenetration, while no-slip boundary conditions are approximated by exchange viscous forces like those of Eq. (2) between liquid and solid particles. The time step used in the simulations is ∆t = 10 -3 s. The DMP model was derived from a previous study [bib_ref] Modelling and simulation of the hydrodynamics and mixing profiles in the human..., Schütt [/bib_ref] , where more details and additional numerical considerations (e.g. convergence of the results with respect to the size and the number of particles used in the model) can be found.
# Methods
## Deep reinforcement learning (drl).
When the tube contracts at a certain location, we perform an 'action' a. As a result of this action, the flexible membrane changes shape and, therefore, its 'state' s. If we contract the tube on a different location, we perform a new action a′ and the system acquires a new state s′. In our case, every state is a 10-dimensional vector of zeros and ones that constitutes the input of the ANN (i.e. which particles are contracting at time T). Actions are also represented by 10-dimensional vectors of zeros and ones that constitute the output of the ANN (i.e. which particles are going to be contracting at time T + ∆T). Every action produces a movement ∆x of the centre of mass of the fluid. If the change of centre of mass of the fluid is in the positive axial direction, ∆x is positive and we call it the 'reward' r of the action a. If the change is in the opposite direction, ∆x is negative and produces a negative reward. During a simulation, we perform N actions a that bring to N corresponding states s and N rewards r: the series of N sets (a, s, r) is called an 'episode' . The objective of DRL is to train the ANN in such a way that given an initial state, the ANN determines a series of actions that maximize the cumulative reward of the whole episode. To achieve this goal, RL introduces the concept of 'quality' Q(s, a) of an action a given the state s where max Q(s′, a′) returns the maximum Q value for the best possible action a' in the next state s′, and the parameter γ (< 1) is the discounting factor that weights the importance of future rewards on Q. In DRL, an ANN is used to approximate Q(s, a); the multiphysics model will then execute the action with the highest quality, i.e. max Q(s, a). To achieve this goal, Eq. (5) must be the training target of the ANN, which implies a loss function of the type where Q * (s, a) is the ANN approximation of . At the beginning of the training, the values of Q * (s, a) are meaningless because the ANN is initialized with random weights. If the action is selected only based on max Q * (s, a), the model would always repeat the same actions: it will learn from these actions, but they will be far from optimal. Therefore, we need a way to force the model to explore alternatives to max Q * (s, a). To achieve this, from time to time, actions are selected randomly. The probability α to choose a random action over max Q * (s, a) is higher at the beginning of the training and decays as it progresses. In our case, initially α = 1 (fully random) and decays linearly to α = 0.1 during 30,000 episodes.
The simulations of the physical model are carried out with the open-source software LAMMPS 28 compiled as a Python library, while the Python library Keras is used for training the network. Details are given in [fig_ref] Table 1: Architecture of the ANN and Hhyperparameters used for training [/fig_ref] :
[formula] (1) P = c 0 ρ 0 7 ρ ρ 0 7 − 1 (2) � ij = −ah c 0 ρ ij v ij r ij r 2 ij + bh 2 (3) F ij = k(r ij − r 0 ) (4) U ij = A 1 + cos [/formula]
πr ij r 0 , r < r 0 (5) Q(s, a) = r + γ max a ′ Q(s ′ , a ′ ) [fig_ref] Table 1: Architecture of the ANN and Hhyperparameters used for training [/fig_ref] not explicitly discussed in the text. In [fig_ref] Table 1: Architecture of the ANN and Hhyperparameters used for training [/fig_ref] , the size of the output layer is 10. However, in the most general case, the size of the output layer should be 2 [bib_ref] Discrete multiphysics: a mesh-free approach to model biological valves including the formation..., Ariane [/bib_ref]. In fact, if a is a 10-dimensional vector (of ones and zeros indicating which sections are contracting or relaxing at time T + ∆T), the number of all possible a (e.g. all possible contraction patterns) is N a = 2 [bib_ref] Discrete multiphysics: a mesh-free approach to model biological valves including the formation..., Ariane [/bib_ref]. The ANN, therefore, takes s (size 10) as input and gives Q * (s, a) (size 2 10 ) as output. However, if we assume that only one section at the time can be contracted, the size of the output layer reduces to 10, which is consistent with the description of the ANN given in [fig_ref] Figure 2: The DeepMP model combines a DMP model with an ANN [/fig_ref].
The optimal model has a dimensionless reward close to 1 (which is close to theoretically maximal efficiency). The model was also trained with a settings (e.g. architecture of the ANN, hyperparameters) different from those of [fig_ref] Table 1: Architecture of the ANN and Hhyperparameters used for training [/fig_ref] , and some of the runs produced suboptimal models, which are not shown here. Visualization [fig_ref] Figure 2: The DeepMP model combines a DMP model with an ANN [/fig_ref] and Videos were processed with the software Ovito 30 .
## Code availability
The code used for the simulations is freely available under the GNU General Public License v3 and can be downloaded from the Cranfield repository https ://publi c.cranfi eld.ac.uk/e1020 81/DeepM P/.
Received: 21 July 2020; Accepted: 14 September 2020
[fig] Figure 1: Functioning of particle methods and link with Artificial Neural Networks. (a) Typical flow of particle methods algorithms. (b) Particle methods like molecular dynamics, smoothed particle hydrodynamics or the discrete element method exchange forces among non-bonded particles. (c) Methods like the lattice spring model or peridynamics exchange forces among bonded particles. (d) In discrete multiphysics, heat transfer occurs by exchanging heat among neighbouring particles. (e) Artificial neural networks exchange information among interconnected neurons. (f) The algorithm for forward propagation in ANNs has the same flow as the particle methods algorithm. Scientific RepoRtS | (2020) 10:16247 | https://doi.org/10.1038/s41598-020-73329-0 [/fig]
[fig] Figure 2: The DeepMP model combines a DMP model with an ANN. (a) DMP model: SPH particles model the fluid, LSM particles model the membrane. (b) Liquid particles are computational particles that only exchange forces; hidden neurons are computational neurons that only exchange (non-physical) information; solid particles are particle-neuron duals that exchange forces with the other computational particles, and information with the hidden neurons. Given the state of the membrane at time T, the ANN calculates which section of the membrane (and for how long) should be contracted next to maximize the amount of fluid moved from left to right. Scientific RepoRtS | (2020) 10:16247 | https://doi.org/10.1038/s41598-020-73329-0 [/fig]
[fig] Figure 3: Training of the DeepMP model. Evolution of dimensionless cumulative reward during training showing catastrophic forgetting. 0 and ρ 0 are reference velocities and densities. Viscous forces between particles i and j are calculated with [/fig]
[table] Table 1: Architecture of the ANN and Hhyperparameters used for training. [/table]
|
A case of oncogenic osteomalacia owing to inguinal tumor
The oncogenic hypophosphatemic osteomalacia is a very incapacitating disease and the mortality rate, mainly due to metabolic disorder, depends on the early diagnosis, since the surgery is curative. The major difficulty is to consider this kind of disease in patients with complex clinical presentation. Moreover, medical centers have to provide a good diagnostic infrastructure because these tumors, in most cases, are small and do not have an obvious site. This case report is about a man with a rapid loss of strength and muscle mass, which had his diagnosis in a late, culminating in significant deformities and organic dysfunctions with clinical repercussions. However, the fast diagnosis with appropriate tests determined the stop point of the evolution of disease and marked the beginning of metabolic recovery. This case reinforces the global problem health care infrastructure and the access to diagnostic equipment, demonstrating the impact on the patient's health of our service.
# Introduction
Oncogenic hypophosphatemic osteomalacia (OHO) is a rare paraneoplastic syndrome, in most cases induced by small benign mesenchymal tumors in any location [bib_ref] Tumor induced osteomalacia, Chong [/bib_ref] [bib_ref] Most osteomalacia-associated mesenchymal tumors are a single histopathologic entity, Folpe [/bib_ref]. The structural deformations are induced by bone demineralization, presenting hypophosphatemia and 1,25-dihydroxy-vitamin D, hyperphosphaturia and initial normal serum parathyroid hormone (PTH), calcium and 25-hydroxy-vitamin-D levels.
These structural and laboratory changes are consequences of excessive fibroblast growth factor 23 (FGF 23) secreted by the tumor, that causes low tubular phosphorus absorption and modify the renal hydroxylation of vitamin D [bib_ref] Oncogenic osteomalacia: is there a new phosphate regulating hormone?, Nelson [/bib_ref] [bib_ref] Osteomalacia oncogénica-presentación de dos casos, Jerkovich [/bib_ref]. Gradually, the patient develop pain on walking and during routine activities, associated to body moves limitations, multiple fractures, renal and cardiac dysfunction and death. Nevertheless, a simply resection of this tumor stops the evolution [bib_ref] A case of oncogenic osteomalacia due to occult nasal sinus tumor. Clinical..., Ray [/bib_ref].
## Case report
A 40-year-old patient, male, was referred to the General Surgery Unit of Complexo Hospitalar Santa Casa with several structural deformities (fractures with poor bone healing, vertebral and limb shortening) that have evolved over the last three years. Bone destruction, fractures, salt and pepper skull effect was demonstrated. These fractures generated a thoracic architecture deformity by spine and ribs alterations. This condition started with pain to walk that progressed to walking disability without aid.
In one year, he suffered an acute myocardial infarction without known primary risk factors. The heart evaluation by a transthoracic echocardiogram has shown a left ventricular dysfunction-grade 2, diffuse hypokinesia, diastolic dysfunctiongrade 2, minimal mitral regurgitation and a 35% ejection fraction. Due to the volumetric decrease of the thorax, the pulmonary function deteriorated.
The investigation started with some laboratory exams in an attempt to elucidate the cause of the structural deformities (osteomalacia) [fig_ref] Table 1: Preoperative laboratory exams [/fig_ref].
He had some radiographic exams that showed multiple fractures in limbs, decreased axial skeleton and loss of bone mass. The computer tomography (CT) revealed a diffuse reduction of bone attenuation coefficient and vertebral depression, suggesting osteoporosis. However, no visceral abnormality was found. At the clinical examination, no additional anatomical changes were found to detect the initial site of the disease. A wholebody magnetic resonance imaging (MRI) [fig_ref] Figure 1: MRI (inguinal view). [/fig_ref] found a nodular image in left inguinoscrotal area (2.5 × 2.1 × 30 cm 3 ), with heterogeneous signal intensity in T1 and T2 and high contrast impregnation near the left Long Adductor Muscle, without invasion, followed by lymph nodes. It was also found fracture signs in inferior ischiopubic area, bilateral, pubic and femur, suggesting pathologic fractures.
Due to the high suspicion of a secretive tumor, the surgical team decided to investigate with a whole-body positron emission tomography (PET/CT) [fig_ref] Figure 2: PET-CT. [/fig_ref] that found several bone deformities by metabolic disorder and a nodular lesion (soft part aspect on acral site) located on the subcutaneous tissue, left inguinal region at the base of the scrotum, prior to the medial margin of the pectineus muscle, measuring 2.5 × 1.6 cm 2 (SUV 3.8).
The bone scintigraphy with technetium [fig_ref] Figure 3: bone scintigraphy with technetium [/fig_ref] , that can be performed to study the degree of loss of bone mass and other structural abnormalities, found a diffuse capitation in multiple articulations (arms, legs, ribs, spine, head and pelvis) compatible with bone wear and the same scrotal injury.
The final diagnosis was confirmed after a surgical resection with free margins, that presented to histopathological and immunohistochemical analysis, a mesenchymal phosphaturic tumor with 3.5 × 3.0 × 2.0 cm 3 without invasion, known as OHO [fig_ref] Table 2: Immunohistochemical analysis [/fig_ref]. Our service could not afford the necessary kit for FGF-23 testing.
After seventh postoperative day and continuous multivitamin and mineral supplementation (indicated by our Nutrition Team), the patient underwent a second transesophageal echocardiogram, presenting a great improve of cardiac parameters, with normal chamber function and a 66% ejection function (much more than we expected). No cardiovascular intervention was performed, except respiratory physiotherapy. Postoperative laboratory exams improved drastically [fig_ref] Table 3: Postoperative laboratory exams [/fig_ref]. After three months, the pain almost disappeared with low levels of analgesics, physiotherapy two times a week, walking without aid and normal lung function.
# Discussion
Disabling diseases are common, like trauma consequences, some kind of cancers or congenital malformations. Nevertheless, some diseases that causes such damage to the patient, society and costs to government could be treated since the first signals and have a better upshot. In case of OHO, this kind of tumor, when simply resected with free margin, determinate the cure of this disease and stop the evaluation to limited structural and organic dysfunction [bib_ref] Oncogenic osteomalacia: loss of hypophosphatemia might be the key to avoid misdiagnosis, Chang [/bib_ref] [bib_ref] Fibroblast growth factor 23 in oncogenic osteomalacia and Xlinked hypophosphatemia, Jonsson [/bib_ref]. There is no large review about OHO tumoral location (maybe because there is no specific site to this cancer grow up), but probably this is not the first one with inguinal location. One important aspect is to be secure that all tumoral lesion was removed (not just to prevent an uncommon recidive, but to cure the disease) and a perioperative analysis by pathologist can solve that question. The early search for tertiary centers with expertise in OHO are the basis to have good results with no or minimum sequelae. The physician that takes into account some nonspecific points, like pain and difficult to walk, that many times could be confused with orthopedic diseases, is the key point to this patient life [bib_ref] Resolution of severe oncogenic hypophosphatemic osteomalacia after resection of a deeply located..., Radaideh [/bib_ref]. Simple laboratory tests (serum and urinary phosphorus, calcium, vitamin D varieties, PTH) and some radiographs could help the diagnosis without so much cost [bib_ref] Fibroblast growth factor 23 in oncogenic osteomalacia and Xlinked hypophosphatemia, Jonsson [/bib_ref].
Several genetic and acquired diseases must be differentiated to OHO, once can be found elevated FGF-23 and clinical characteristics (OHO-like).
Mesenchymal tumors can have recurrences, but OHO does not need regular screening tests because of the low incidence of recurrences [bib_ref] Fibroblast growth factor 23 in oncogenic osteomalacia and Xlinked hypophosphatemia, Jonsson [/bib_ref]. Our team proposes just a regular clinical review, following the recovery of motor function and clinical implications, only requesting some tests if there are complaints. In some services, radiological work up functional imaging is done before anatomical imaging, but that's not a reality of all medical centers because this exams are expensive and restrict to few centers. The great challenge is how to locate exactly this mesenchymal tumor site. Nowadays, the main exams include CT, MRI and scintigraphy with some radioisotope (technetium is the most commonly used) [bib_ref] Osteomalacia inducida por tumor: hemangiopericitoma rinosinusal, Serafini [/bib_ref]. These kinds of exams are very expensive and normally not easy to be found in the routine of the medical center radiology units around the world. There is no consensus on literature about the gold standard exam, but most services use MRI to screen the tumoral site.
The close follow up in rehabilitation to reintegrate this patient to family and society is crucial, once this disease starts, mostly, when this individual is economically active and has a family [bib_ref] Phosphaturic mesenchymal tumor: 2 new oral cases and review of 53 cases..., Qari [/bib_ref].
[fig] Figure 2: PET-CT. [/fig]
[fig] Figure 1: MRI (inguinal view). [/fig]
[fig] Figure 3: bone scintigraphy with technetium. [/fig]
[fig] Figure 4: bone scintigraphy with technetium (focus). [/fig]
[table] Table 1: Preoperative laboratory exams. [/table]
[table] Table 2: Immunohistochemical analysis. [/table]
[table] Table 3: Postoperative laboratory exams. Anti-nuclear antibody (NR: negative) Negative Calcium (serum) (NR: 4.4-5.4 mg/dl) 5.2 mg/dl Inorganic phosphorus (serum) (NR: 3.4-4.5 mg/dl) 3.8 mg/dl Inorganic phosphorus (urinary) (NR: 400-1.300 mg/24 h) [/table]
|
Algal Autophagy Is Necessary for the Regulation of Carbon Metabolism Under Nutrient Deficiency
Autophagy is a mechanism to recycle intracellular constituents such as amino acids and other carbon-and nitrogen (N)-containing compounds. Although autophagy-related (ATG) genes required for autophagy are encoded by many algal genomes, their functional importance in microalgae in nutrient-deficiency has not been appraised using ATG-defective mutants. Recently, by characterization of an insertional mutant of the ATG8 encoding a ubiquitin-like protein indispensable for autophagosome formation in a green alga Chlamydomonas reinhardtii, we have provided evidence that supports the following notions. ATG8 protein is required for the degradation of lipid droplets and triacylglycerol (TAG) triggered by resupply of N to cell culture in N-deficient conditions. ATG8 protein is also necessary for starch accumulation under phosphorus-deficient conditions. Algal autophagy is not necessary for inheritance of chloroplast and mitochondrial genomes. In this review, we discuss the physiological roles of algal autophagy associated with nutrient deficiency revealed by the genetic and biochemical analyses using disruption mutants and reagents that inhibit the fatty acid biosynthesis and vacuolar H + -ATPase.
# Introduction
Macroautophagy (hereafter autophagy) is a recycling system for degradation of cytoplasmic constituents in the vacuole or lysosomes. This pathway is conserved among eukaryotes. In order to understand the physiological functions of plant autophagy, the following strategies has been conducted, because the processes of autophagy, from autophagosome formation to degradation of target cellular components in the vacuole are very rapid: 1) analyzing the phenotype of autophagy-defective core ATG gene mutants; 2) visualizing subcellular localization of ATG8 protein with a fluorescent protein; and 3) artificially halting the autophagic flux in the vacuole by treatment with concanamycin A, which inhibits H + -ATPase and thus inactivating the vacuolar acid hydrolases, leading to the autophagic body accumulating inside the vacuole.
The following findings have been reported regarding the regulation of the accumulation of photosynthetic assimilation products, such as lipids and starch, by autophagy. Firstly, starch synthesized in leaves by photosynthesis during the day is degraded at night, and phenotypic analysis of atg mutants in Arabidopsis thaliana suggested that autophagy facilitates this process. In addition, autophagy was shown to promote triacylglycerol (TAG) degradation under C-deficient conditions using the seedlings of A. thaliana atg mutants. Recently,reported that basal autophagy is required for TAG biosynthesis by lipid turnover providing fatty acids from organellar membrane lipids in A. thaliana. Similarly, studies on atg mutants in rice which showed male sterility concluded that autophagy is necessary for TAG and starch accumulation and lipid droplet formation as well as normal reproductive postmeiotic anther development during pollen maturation. However, studies of plant autophagy have been limited to a few model species for which atg mutants are available, and the role of autophagy in the metabolism of carbon (C)-assimilation products throughout the plant kingdom needs to be further investigated. Because in algae the role of autophagy in the accumulation and metabolism of photosynthetic assimilation products remained unclear, studies using autophagy-deficient mutant strains have been considered necessary in addition to those using wild type cells treated with autophagy-inhibiting chemicals. In this review, the physiological functions of algal autophagy in response to nutrient deficiency will be discussed, based on recent reports of autophagy-defective mutants in Chlamydomonas reinhardtii (hereafter Chlamydomonas) and the effects of treatment of the wild-type Chlamydomonas cells with cerulenin and concanamycin A for inhibition of fatty acid synthesis and vacuolar lysosomal function, respectively.
## Algal autophagy
Unicellular algae known as microalgae are photosynthetic eukaryotes, classified as one of protists. For the purpose of biofuel production, many microalgae have been nominated by screening, which accumulate high levels of C-storage compounds such as lipid and starch. Especially understanding the physiological culture conditions and the molecular mechanisms for accumulation of their lipid and starch is necessary to achieve realistic biofuel production. So far, it is reported that many algae accumulate these C-storage compounds in cells when exposed to nutrient-deficient stress conditions such as nitrogen (N)-deficiency after stopping their growth in these stress conditions but they do not die immediately and that instead, they maintain cell viability for a period of time. This cell survival under nutrient-deplete conditions suggests that autophagy is involved in the system for keeping the cell viability. Although autophagy is involved in the regulation of C metabolism in yeast, animals, and terrestrial plants, contribution of autophagy to lipid and starch metabolism in algae has not been fully understood. In several species of algae from Chlorophyta, Rhodophyta, and Chromalveolata, orthologs for known ATG genes have been found in genome databases, except for red algae whose genomes lack the ATG gene. However, no autophagy-defective mutant has been reported in any species of algae so far.
In a model photosynthetic single-cell eukaryote, Chlamydomonas, a series of studies have been reported to provide molecular evidence of autophagy and autophagic activities responding to rapamycin, an inhibitor of the target of rapamycin (TOR) protein kinase, abiotic stresses such as endoplasmic reticulum (ER) stress, oxidative stress, and metal toxicity, and the following nutrient deficiency; N-deficiency, C-deficiency, and carotenoid deficiency. In the all mentioned stress conditions, the abundance of ATG8 protein and its PE-conjugated/lipidated form increases. Accumulation of ATG8 and ATG3 proteins increased in a conditional repression line of a chloroplast protease ClpP1, implying that chloroplast proteolysis systems and autophagy in Chlamydomonas partially complement each other. Furthermore, in vitro assays using recombinant ATG8 and ATG4 proteins and complementation tests with the corresponding yeast mutants revealed that correspond to the orthologs for ATG8 and ATG4 genes in Chlamydomonas, respectively, encode proteins with functions similar to those in other organisms.
Recently, we reported the functional importance of ATG8 and ATG3 proteins by using insertion mutants defect in corresponding genes in Chlamydomonas . Using these autophagy-defective strains, contribution of autophagy on cell processes including photosynthesis, metabolism, and reproduction under various nutrient-deficient conditions was clarified.
## Algal autophagy in tor signaling
The TOR kinase complex 1 (TORC1) is reported to be an important regulatory factor making balance between growth and autophagy in the eukaryotes. When nutrients are abundantly supplied to the cell, TOR is activated and promotes protein synthesis and cell growth, while negatively regulating autophagy by inhibiting the formation of the ATG1 complex. Land plants have TORdependent and TOR-independent pathways for regulation of autophagy. Activation of autophagy in nutrient deficiency, salt, and osmotic stress requires suppression of TOR. On the other hand, induction of autophagy under oxidative stress and ER stress conditions is suggested to be controlled independently by TOR. Although there are still many unknown regarding algal TOR systems, it is reported that TOR, LST8 and Raptor genes, which are components of TORC1 protein complex, are widely conserved in algal genomes.
Chlamydomonas has all the components of the TORCl complex. The arrest of cell cycle, bleaching, and vacuolization have been shown to be proceeded by the addition of rapamycin to the Chlamydomonas cells, indicating that Chlamydomonas has a rapamycin-sensitive TOR signaling network. Addition of rapamycin also increases the expression of ATG8, led to detection of lipidated ATG8, and altered the subcellular localization of ATG8, suggesting that the rapamycin-sensitive TOR signaling network regulates autophagy in Chlamydomonas. Recently,demonstrated that rapamycin-induced TOR inhibition increases de novo amino acid synthesis due to upregulated ammonium assimilation in Chlamydomonas. It has been also reported that TOR inhibition by addition of TOR inhibitors, such as rapamycin, AZD8055, or Torin1, activates expression of genes encoding TAG biosynthesis-related enzymes and increased TAG accumulation in Chlamydomonas and a red alga Cyanidioschyzon merolae. Furthermore, it has been suggested that TOR in Chlamydomonas regulates the biosynthesis of the signaling molecules, inositol polyphosphates (InsPs). A mutant, vip1-1, of the inositol polyphosphate kinase (VIP1) gene involved in the biosynthesis of the InsPs, showed hypersensitivity to rapamycin and increased TAG accumulation even under nutrient replete conditions. In addition, these reports indicate an interaction between TOR signaling network and lipid metabolism in algae. Most recently, through the characterization of a hypomorphic mutant of the LST8 gene, lst8-1 under P deficiency,demonstrated that P deficiency reduces the abundance of LST8 proteins and downregulates activity of TORC1 to activate autophagy and TAG synthesis in Chlamydomonas. In the N-deficient conditions, any significant phenotype was not observed in the lst8-1 mutant.
Despite these energetic studies in this field, any direct molecular evidence that algal autophagy is involved in interaction between TOR signaling network and lipid metabolism has not been provided and needs to be examined experimentally in future. By quantitative phosphoproteomic approach, total 258 of phosphosites from 219 unique phosphopeptides were detected in Chlamydomonas wild-type cells as significantly modulated by inhibition of TOR. The key target factors downstream of TOR that are involved in autophagy induction and regulation of lipid turnover and link each other might be discovered from proteins composed of such phosphopeptides.
## Role of algal autophagy in carbon metabolism
Chlamydomonas accumulates high levels of neutral lipids and starch in the absence of nutrients such as N, P, and sulfur (S). In addition, the Chlamydomonas genome contains ATG genes related to autophagy. Therefore, we attempted to isolate mutants that lack the ATG8 and ATG3 genes, which are required for the formation of autophagosomal membranes-an early process of autophagy . For both genes, only one copy was present in the Chlamydomonas genome, and we predicted that their deletion would lead to an autophagy defect. DNA tag-insertion mutants of each gene were selected, one by one, from approximately 4,600 mutant strain libraries, and antibodies were used to show that these were functionally deficient mutants. In the atg8 and atg3 mutants, the cell viability under N-deficient conditions was significantly reduced. After two weeks, the survival rate of the wild-type cells was 80%, whereas the survival rate in both mutant strains decreased to less than 5%, indicating that autophagy is necessary for algal cell survival.
The role of autophagy in the control of C metabolic processes was determined by examining the accumulation of neutral lipids (TAG) and starch under nutrient-deficient conditions, using atg8 mutants as a representative strain. Under N-and S-deficient conditions, the atg8 mutant accumulated starch in the same way as did the wild-type strain during the early stress phase. However, since the accumulated starch in the atg8 mutant was broken down after the second day of stress, this suggests that autophagy was required to maintain the accumulated starch content. In addition, when the medium was replaced with Ndeficient conditions before being returned to the N-replete condition, TAG and starch degraded rapidly in the wild-type strain, while degradation of TAG and oil droplets in the atg8 cells occurred approximately 6 h later than in the wild strain. However, the rate of starch degradation in the atg8 mutant was unaffected, suggesting that autophagy is required for rapid degradation of TAG.
The green algae accumulated starch and TAG under Pdeficient conditions, as well as under N-and S-deficient conditions. However, although the atg8 mutants accumulated TAG at rates similar to that of the wild-type strain, starch accumulation did not occur from the beginning of the deficiency period, suggesting that autophagy was required for starch accumulation in response to P deficiency.
Cellular organelles, such as chloroplasts and mitochondria, have their own genomes, and only the genomes from one parent are inherited during mating-a process known as maternal inheritance. The Chlamydomonas cell undergoes a shift from vegetative growth to reproductive growth, and gametogenesis occurs meiotically under N-deficient conditions. Because autophagy is involved in the maternal inheritance of mitochondria in mice (Aland nematode, experiments were conducted using the atg8 mutant to determine whether autophagy was involved in the maternal inheritance of algae. The results showed that maternal inheritance continued as normal during mating of both mitochondrial genomes, and that algal autophagy is not involved in maternal inheritance.
## Revealing the relationship between algal autophagy and lipid metabolism by adding inhibitors
Heredia-Martínez et al. reported that chloroplast damage in Chlamydomonas cells induced by the addition of the fatty acid biosynthesis inhibitor, cerulenin, triggers autophagy. In addition,reported that changes in cells were suppressed when the protein degradation system was suppressed by treating the cells with concanamycin A to stop vacuole flux. The advantage of using such inhibitors is to be able to avoid downstream effects and to examine the effects on cells when a particular in vivo pathway is inhibited by time-limited conditions. In both cases, the addition of the inhibitor increased the expression of the ATG8 protein, which suggested autophagy was activated. When cerulenin was added to the cells, chloroplast shrinkage and thylakoid membrane aggregation was observed along with a decrease in photosynthetic proteins and ribosomal proteins, and a decrease in maximum quantum yield. Furthermore, the amount of ROS increased, while the thylakoid membrane lipid, MGDG, decreased. Both qRT-PCR and RNAseq analyses showed that the expression of genes involved in protein degradation, stress response, and signal transduction were altered.have proposed a model demonstrating that when fatty acid synthesis is inhibited by cerulenin, the amount of MGDG is reduced, thereby impairing chloroplast function and causing thylakoid membrane aggregation and generation of ROS. This chloroplast damage propagates up to the nucleus as a retrograde signal and causes activation of autophagy with changes in gene expression, as well as the activation of proteolytic systems. Future research will be necessary to further verify the time-based vertical relationship and causality of these multiple-response reactions. In addition, by comparing the effect of treatment with cerulenin between the wildtype strain and the autophagy-deficient strain, it is possible to verify to what extent autophagy affects these chloroplast injuries.reported that development of lipid droplets and TAG accumulation under N-or P-deficient conditions were repressed when the lytic digestion in the vacuolar or lysosomal lumen was inhibited in Chlamydomonas WT cells after treatment with concanamycin A, an inhibitor of vacuolar H + -ATPase activity. Moreover, treatment with concanamycin A during these nutrient-deficient periods suppressed the degradation of ribosomal proteins such as RPS6 and RPL37. This suggested that concanamycin A inhibits autophagic flux. By the live-cell imaging using the Chlamydomonas transgenic line expressing red fluorescent protein (mCherry)tagged ATG8, the interactions and fusion between mCherrylabeled structures and lipid droplets were observed in the N deficiency. Based on this observation,suggests that autophagy-related pathway might be involved in lipid droplet turnover in the alga. On the other hand, genetic inhibition of autophagy by knockout of the core ATG genes compromises mobilization of various cytoplasmic components into the vacuolar or lysosomal lumen . The atg mutants formed lipid droplets and accumulated TAG under both N-and P-deficient conditions as the WT cells did. In the red alga C. merolae which do not have any core ATG gene, mRNA abundance of genes encoding glycerol-3phosphate acyltransferase (GPAT) and acyl-CoA:diacylglycerol acyltransferase (DGAT)in TAG biosynthesis and the level of TAG content increase under TOR-inhibition condition. These reports suggest that TOR signaling regulates development of lipid droplets and TAG accumulation without though activation of autophagic pathway in the algae.
Interestingly, a dual-specificity tyrosine phosphorylationregulated kinase, TAG accumulation regulator 1 (TAR1) has been reported to be necessary for the degradation of chlorophyll and photosynthesis-related proteins and acetate-dependent TAG accumulation in photomixotrophic N-deficient conditions. In addition, in photoautotrophic Ndeficient conditions, the tar1-defective mutant maintained higher levels of cell viability, TAG, and starch with lower accumulation of hydrogen peroxide compared with those of wild type (WT) cells with bubbling of air containing 5% carbon dioxide . Recently,andreported that TOR represses an A. thaliana TAR1 ortholog, YAK1 to promote meristem activity and plant growth. YAK1 interacts with RAPTOR and is directly phosphorylated by TOR complex in A. thaliana. These findings open a possibility that the Chlamydomonas TAR1 is also directly regulated by TORmediated phosphorylation. Relationship of the Chlamydomonas TAR1 kinase and regulation of algal autophagy and TOR signaling network could further deepen the understanding of the regulatory mechanism of cellular responses to nutrient-deplete stress conditions. Especially analyses of protein-protein interactions among regulatory kinases and autophagy components could contribute to better understanding of autophagy-related processes in photosynthetic organisms.
## Concluding remarks and future perspectives
The research into autophagy-defective mutants and the effects of inhibitor addition have clarified the role of algal autophagy in the accumulation and degradation of neutral lipids and starch under various nutritional conditions. In the future, by combining these approaches-i.e. by adding rapamycin, cerulenin, concanamycin A, and other inhibitors of a specific pathway to the autophagydefective mutants, and examining the resultant time-dependent phenotypic changes-we can verify the mechanism that works in parallel with autophagy, and which plays a role in complementing autophagy in algae.
# Author contributions
MK and HF wrote this manuscript.
# Funding
This work was supported by JSPS KAKENHI grant number 16H04805 (to HF), 17K07753 (to MK), JST ALCA program JPMJAL1105 (to HF), and NEDO project number P10010 (to HF). |
Will Sirtuins Be Promising Therapeutic Targets for TBI and Associated Neurodegenerative Diseases?
Traumatic brain injury (TBI), a leading cause of morbidity worldwide, induces mechanical, persistent structural, and metabolic abnormalities in neurons and other brain-resident cells. The key pathological features of TBI include neuroinflammation, oxidative stress, excitotoxicity, and mitochondrial dysfunction. These pathological processes persist for a period of time after TBIs. Sirtuins are evolutionarily conserved nicotinamide-adenine dinucleotide (NAD+)-dependent deacetylases and mono-ADP-ribosyl transferases. The mammalian sirtuin family has seven members, referred to as Sirtuin (SIRT) 1-7. Accumulating evidence suggests that SIRT1 and SIRT3 play a neuroprotective role in TBI. Although the evidence is scant, considering the involvement of SIRT2, 4-7 in other brain injury models, they may also intervene in similar pathophysiology in TBI. Neurodegenerative diseases are generally accepted sequelae of TBI. It was found that TBI and neurodegenerative diseases have many similarities and overlaps in pathological features. Besides, sirtuins play some unique roles in some neurodegenerative diseases. Therefore, we propose that sirtuins might be a promising therapeutic target for both TBI and associated neurodegenerative diseases. In this paper, we review the neuroprotective effects of sirtuins on TBI as well as related neurodegeneration and discuss the therapeutic potential of sirtuin modulators.
# Introduction
Traumatic brain injury (TBI) is a leading cause of morbidity worldwide. It is associated with long-term disability and significant healthcare expenditures and has become a priority for public health policy. Although most of the neurological injuries are limited and temporary, TBI still has multiple sequelae including sleep disturbance, hypopituitarism, seizures, epilepsy, and neurodegenerative diseases. Indeed, TBI is widely recognized as a risk factor for neurodegenerative diseases, which are characterized by a slow progressive loss of neurons or myelin sheaths and increasing disability and include Alzheimer's disease (AD), Parkinson's disease (PD), Huntington's disease, and multiple sclerosis. Notably, patients with different types of TBI tend to have an increased risk of different neurodegenerative diseases. For example, moderate or severe TBI contributes to the development of late-onset neurodegenerative diseases, in particular AD and PD. Chronic traumatic encephalopathy (CTE), a condition that shares features with neurodegenerative diseases, is linked to mild TBI. Numerous random-effects meta-analysis, epidemiologic study, and retrospective cohort studyhave lent support to these associations. Besides raising the risk, TBI lowers the age at onset of TBI-related neurodegenerative diseases. Mechanically, it has previously been observed that pathological processes, including oxidative stress, neuroinflammation, excitotoxicity, proteinopathies, and mitochondrial dysfunction, may persist for months or years post-TBI, and these pathophysiological substrates serve as triggers yielding the progression of neurodegenerative diseases. Accordingly, targeting these post-TBI pathological processes holds the promise for the intervention of TBI and late-onset neurodegenerative diseases.
Homologs of Silent information regulator 2 (Sir2) are collectively known as sirtuins. The mammalian sirtuin family has seven members, which are classified into four groups: class I (SIRT1-3), class II (SIRT4), class III (SIRT5), and class IV (SIRT6 and 7) . All sirtuins have an evolutionarily conserved nicotinamideadenine dinucleotide (NAD + )-dependent catalytic core domain, and each sirtuin has unique N-terminal and C-terminal sequences. SIRT1 and SIRT5 tend to elicit NAD + -dependent deacetylase activity, whereas SIRT4 and SIRT6 are likely to express mono-ADP-ribosyl transferase activity. Sirtuins occupy different cellular locations including mitochondria (SIRT3-5), the cell nucleus (SIRT6), the nucleolus (SIRT7), and between the cytoplasm and nucleus (SIRT1, 2). In addition, sirtuins are widespread in brain cells, such as neurons, astrocyte , and microglia .
Sirtuins are of vital interest in oxidative stress, neuroinflammation, blood-brain barrier (BBB) permeability, astrocyte activation, and neural apoptosis in brain injuries. To date, studies regarding SIRT1 and SIRT3 have provided direct evidence about the functions of sirtuins in TBI. Despite the lack of evidence, other sirtuins may also intervene in similar pathophysiology in TBI, as have been shown in other models such as subarachnoid hemorrhage (SAH), ischemic stroke, and ischemia/reperfusion (I/R) injury. In addition, sirtuins also have some considerable impact on neurodegenerative diseases, such as AD and PD. Based on the epidemiological association and pathological similarities of TBI and neurodegenerative diseases, we propose that sirtuins may be a promising therapeutic target for TBI and related neurodegenerative diseases. In this review, we will dissect the similarity on the pathophysiological processes between TBI and neurodegenerative diseases and summarize the neuroprotective effects of sirtuins in these pathophysiological processes. We will also outline sirtuin modulators for pharmacotherapeutics and the probable difficulties in the development of pharmacological agents.
## Pathophysiology of tbi and related neurodegenerative diseases tbi pathophysiology
TBI is classified into mild, moderate, and severe type, which can lead to different degrees of primary and secondary brain injuries. Primary injuries are attributed to the direct result of physical injuries, including hemorrhage and tissue damage. Secondary injuries include cellular hyperexcitability, vasogenic and cytotoxic edema, hypoxia-ischemia, microglia polarization, and astrocyte activation. Reducing primary injuries depends on preventive measures whereas second injuries are sensitive to therapeutic measures. Evidence suggests that second injuries considerably determine the outcomes of TBI patients. Therefore, in this review, we mainly focus on the second injuries and corresponding therapeutic measures.
Second injuries can further induce metabolic disorders, vascular abnormalities, extensive neuroinflammation, oxidative stress, microglial activation, and excitotoxicity. TBI-induced mitochondrial dysfunction causes cellular metabolic alterations, excitotoxicity, and endogenous antioxidant system exhaustion. The excessive production of reactive oxygen species (ROS) leads to lipid peroxidation, cytotoxicity, necrotic cell apoptosis, and oxidative stress, which in turn, exacerbate mitochondrial dysfunction. Besides, mitochondrial dysfunction can affect cell FIGURE 1 | Mammalian sirtuins. Mammals have seven sirtuins. They are classified into four groups. SIRT1, 2, 3 are class I, SIRT4 is class II, SIRT5 is class III, and SIRT6, 7 are class IV. All of them have a conserved NAD + -dependent catalytic core domain. Each of the sirtuins has unique additional N-terminal and C-terminal sequences. SIRT1, 5 tend to show NAD + -dependent deacetylase (DAC) activity first whereas SIRT4, 6 are likely to express its mono-ADP-ribosyl transferase (ART) activity. Sirtuins have distributed cellular localization, mitochondria (SIRT3, 4, 5), cell nucleus (SIRT6), nucleolus (SIRT7), and both in cytoplasm and nucleus (SIRT1, 2). Reprinted and modified from International Journal of Biological Sciences, 7 (5), Xiaoling membrane permeability, resulting in the extensive release of apoptotic proteins. Neuroinflammation, the physiological response to the injury, triggers the release of cytokines and chemokines and activates microglia and astrocytes. A recent study of brain injury identified a significant number of rod-shaped microglia arranged like the carriages of a train. Microglia are the primary immune sentinels of the central nervous system responding to inflammatory eventsand the main source of ROS. Studies provided evidence that microglial activity persists for a long-term period in the brain of TBI survivors. Moreover, vascular abnormalities may disrupt the BBB and cause encephaledema. Furthermore, TBI is a common cause of olfactory dysfunction, which is mediated by excitotoxicity, an intercellular cascade initiated by excessive release of glutamate and hyperactivation of the glutamatergic receptors; .
## Mechanisms underlying tbi-induced neurodegenerative diseases
TBI is one of the well-established contributing factors for lateonset neurodegenerative diseases. TBI may alter the normal trajectory of the aging processes of the affected individual. Multiple injuries produce synergistic effects and accelerate neurodegeneration (DeKosky and Asken, 2017). The common causes of neurodegenerative diseases are neuroinflammation, excitotoxicity, oxidative stress, mitochondrial dysfunction, and proteinopathies, all of which are involved in the pathophysiology of TBI. Inflammation is a normal immune response coping with injuries or harmful stimuli, yet overactive inflammation damages the brain. Post-mortem and animal model data indicate that oxidative stress and neuroinflammation can persist for years after the acute phase of injury as a result of chronic microglial activation. According to research, reactive microglia were still present in 28% brains over a year after a single TBI. Excitotoxicity refers to neuronal cell death caused by the toxic action of excitatory amino acids. Glutamate is the primary excitatory amino acid neurotransmitter in the central nervous system and is associated with neuronal degeneration, aging, and death. Oxidative stress triggers excessive free radical production, leading to neuronal death and stimulates inflammation. TBI-induced mitochondrial dysfunction decreases the metabolic energy supply and interferes with synaptic function. The resulting depletion of energy stores stimulates the release of ROS, which in turn, aggravates oxidative stress. Besides,indicated that the detrimental role of iron deposits in the TBI pathogenesis may constitute one of the reasons for boosting late-onset neurodegenerative diseases.
AD, one of the most common neurodegenerative diseases, is characterized by the aggregation of β-amyloid protein and FIGURE 2 | TBI induces inflammation, mitochondrial damage, microglial activation, excitotoxicity, and BBB disruption. TBI results in HMFB1 translocation and secretion. Extracellular HMGB1activates TLR4 and CXCR4 to trigger the NF-kB pathway which induces microglia secreting inflammation cytokines to further activate other microglia. TBI induces mitochondrial damage. Damaged mitochondria produce apoptosis protein and ROS to aggravate oxidative stress. Besides, TBI leads to excitotoxicity and BBB disruption which makes inflammation cytokines leakage more severe. TBI, traumatic brain injury; TLR4, toll-like receptor 4; CXCR4, chemokine (C-X-C motif) receptor 4; BBB, blood-brain barrier; HMGB1, high-mobility group box 1; NF-κB, nuclear factor-κB; ROS, reactive oxygen species. tau protein. PD, which is caused by the selective degeneration of dopamine and is characterized by bradykinesia and rigidity, is the second most common neurodegenerative disease. Studies showed that TBI is one of the strong environmental contributing factors for AD and PD. One of the possible reasons is that TBI causes diffuse axonal injury and leads to proteinopathies. TBI causes mechanical trauma to axons and induces the excessive release of glutamate, which may substantially alter neuronal metabolism. Dysregulated axonal transport systems lead to axonal swelling, which may disrupt the delivery of necessary proteins and ultimately lead to focal nerve disconnection. In damaged axons, accumulated amyloid precursor proteins and amyloid β-peptide are markers of axon injury . These markers are frequently found existing in the brains of TBI survivorsand post-mortem brains . Although the mechanism underlying TBI-induced PD is not clear, it may involve mitochondrial dysfunction and oxidative stress. Amyloid β-peptide accelerates tau protein pathology, but the underlying mechanism remains unknown. TBI causes hyperphosphorylation, misfolding, and aggregation of tau proteins, which then form neurofibrillary tangles (NFTs). Hyperphosphorylated tau proteins are released into the extracellular space and stimulate microglia and astrocytes to produce pro-inflammatory cytokines (e.g., IL-1β and TNFα), which further activate tau kinases and aggravate tau hyperphosphorylation (Fesharaki-Zadeh, 2019). Besides phosphorylation, tau acetylation at Lys174 is a crucial change during the post-translational modification of tau proteins, determining the toxicity in mice, tau homeostasis, and the initial step of AD.
CTE, a neurodegenerative disease caused by repeated mild or concussive head injuries, is characterized by an abnormal accumulation of hyperphosphorylated tau protein. Clinically, patients with CTE present with two distinct phenotypes: affective changes and cognitive impairment. Several large epidemiological studies suggest that a single severe TBI is a significant risk factor for the later development of AD. The effects of TBI are long-term and extend beyond the immediate acute phase. Previous findings suggest that neuronal densities in the hippocampus and thalamus continue to decrease 1 year or more after a severe TBI.
Taken collectively, it appears likely that targeting the aforementioned pathological processes post-TBI holds the promise for the treatment of TBI as well as the treatment and even prevention of TBI-induced neurodegenerative diseases.
## Potential roles of sirtuins in tbi and related neurodegeneration
After TBI, the expressions of sirtuins increase . Previous research has established that sirtuins exert a neuroprotective effect after TBI. Nevertheless, related research concerning the functions of sirtuins, particularly SIRT2, 4, 5, 6, 7, in TBI remains insufficient. Notably, sirtuins are also reported to be involved in the pathophysiology of other acute brain injury models, such as SAH, ischemic stroke, and I/R injury, whose pathophysiologies have some similarities with that of TBI. Altogether, sirtuins regulate inflammation, oxidative stress, astrocyte activation, BBB permeability, axon elongation, and synaptic plasticity during acute brain injuries, all of which are characteristic pathological features of TBI. Furthermore, some researchers indicated that sirtuins could reduce the accumulation of amyloid β peptideand tau protein, suggesting their preventive potential for TBI-induced neurodegenerative diseases. In this section, we outline the potential underlying mechanisms of sirtuins in TBI and related neurodegeneration, particularly with regards to AD. Of note, some of these candidate mechanisms are deduced from the models of other acute brain injuries, which are highlighted in corresponding sentences.
## Sirtuins reduce inflammation
In TBI models, existing research has shown the critical role played by SIRT1 to reduce neuroinflammation . Moreover, SIRT2 was observed to alleviate inflammation in LPS-induced neuroinflammation models. SIRT1 reduces inflammation by modulating high mobility group box 1 (HMGB1). SIRT1 maintains HMGB1 in a deacetylated state to control neuroinflammation resulting from microglial activation after TBI. Acetylated HMGB1 is required for HMGB1 transcription and extracellular secretion. HMGB1 is a non-histone-binding protein. Under normal conditions, HMGB1 binds to DNA and regulates transcription and translation in the nucleus. After TBI damage, HMGB1 translocates from the nucleus to the extracellular space. Extracellular HMGB1 induces microglial activation, which leads to further HMGB1 translocation . Extracellular HMGB1 binds to receptors on the cell surface, including toll-like receptor 4 and chemokine receptor 4 . HMGB1 activates receptors via medullary differentiation factor (MyD88) and non-MyD88-dependent pathways, triggering a signaling cascade in the microglia via the nuclear factor-κB (NF-κB) pathway. Nuclear translocation of NF-κB induces microglial polarization and cascade amplification of inflammatory cytokines. Furthermore, SIRT1 inhibits activation of the NOD-like receptor (NLR) family pyrin domain containing 3 (NLRP3) inflammasome to attenuate inflammation after TBI. NLRP3 serves as a platform to produce interleukin-1β (IL-1β), an inflammatory initiating cytokine in a caspase reaction. Inflammation flooding further activates microglia, which in turn, aggravates inflammation . Previous studies have shown that microglial activation up to 1 year after TBI may cause chronic inflammation. Activated microglia have been identified around lesions in various neurodegenerative diseases . Excessive activation of microglia further damages neurons by releasing cytotoxic factors and accelerating the progression of neurodegenerative diseases. Moreover, SIRT1 was shown to mitigate astrocyte activation by inhibiting the p38 mitogen-activated protein kinase (MAPK) signaling pathway in an experimental model of TBI . Astrocyte activation results in the overproduction of glial scar and inflammatory cytokines, which further activates the astrocytes, creating a positive inflammatory feedback loop that may injure the remaining neurons.
## Sirtuins reduce excitotoxicity
Excitotoxicity is implicated in the pathogenesis of several neurodegenerative diseases. Excessive release of glutamate results in glutamate receptor hyperactivation, calcium buffering impairment, free radical overproduction, and mitochondrial hyperpermeability, consequently leading to further neuronal degeneration and programmed cell death. Studies have shown that SIRT1, SIRT3, and SIRT4 are key players in reducing excitotoxic lesion in cortical neurons models.
An in vitro study in cortical neurons found that SIRT1 deacetylated and decreased peroxisome proliferator-activated receptor-coactivator 1α (PGC1α) to reduce glutamate excitotoxicity. PGC1α is a key factor in the maintenance of mitochondrial biogenesis and respiratory function. Previous studies showed that SIRT4 promotes glutamate transport capacity by inhibiting excitotoxicityand interacts with glutamine catabolism by hampering glutamate dehydrogenase (GDH). Glutamine catabolism is crucial during the DNA repair response. Excessive glutamine catabolism leads to neuronal dysfunction and cell death. Furthermore, mitochondrial SIRT3 expression increases after poly(ADP-ribose) polymerase-1-mediated NAD depletion in cortical neurons.concluded that SIRT3 is essential for neuroprotection against excitotoxicity. The authors showed that NAD depletion further increased ROS production, aggravated oxidative stress, and upregulated the expression of SIRT3 in mitochondria.
## Sirtuins improve mitochondrial function and reduce oxidative stress
Mitochondrial dysfunction is a fundamental pathogenic event in most neurodegenerative diseases. Findings from rat models of TBI have revealed that ceramide is the first lipid to accumulate in the mitochondria after TBI, suggesting it may serve as a therapeutic indicator. Ceramide accumulation causes mitochondrial dysfunction, which is associated with increased mitochondrial fission and excessive ROS production. According to existing research, SIRT1 and SIRT3 have been detected to alleviate oxidative stress in TBI models. The study byshowed that SIRT1 reduces the mitochondrial damage to play its neuroprotective roles in TBI. SIRT3, SIRT6, and SIRT7 showed their functions of reducing oxidative stress in other brain models. SIRT1 has been shown to improve mitochondrial function by inhibiting the p38 MAPK pathway in animal models of TBI. Given that MAPK protein is associated with multiple phosphorylation/dephosphorylation signaling cascades, SIRT1 may inhibit this pathway by dephosphorylating p38 MAPK . A previous study found that SIRT1 activity was downregulated in the brains of patients with neurodegenerative diseases including PD. Furthermore, the authors reported that SIRT1 reduced oxidative stress-induced neural cell death and concluded that SIRT1 is a pro-survival protein that is downregulated under cellular stress. These findings suggest that SIRT1 has the potential to reduce oxidative stress in TBI and associated neurodegenerative diseases.
Investigations of the effects of SIRT3 on mitochondrial dysfunction and oxidative stress have yielded contradictory findings. Results from animal models provide evidence supporting a promoting role for SIRT3 in the accumulation of ceramide within the mitochondria, which serves as a causal factor of mitochondrial dysfunction by inhibiting the activity of respiratory chain complex III. Indeed, SIRT3 deacetylates ceramide synthase 1, 2, and 6 to keep them in hyperacetylated state (primed state), propelling the ceramide accumulation. However, some study showed that the inhibition of SIRT3 had no beneficial effects on brain injury in a model of stroke. Inversely, inhibition of SIRT3 in microglia leads to excessive ROS production. SIRT3 deacetylates Forkhead box O 3a (Foxo3a), a transcription factor that transactivates antioxidant genes, to enable it to translocate into the nucleus. Thus, SIRT3 increases the production of antioxidants, including catalase and manganese superoxide dismutase (mnSOD), to protect against oxidative stress. Another study pointed out that it is a compensatory increase in SIRT1 rather than the inhibition of SIRT3 that exerts neuroprotective effect. Not only that, recent work by researchers proved that SIRT3 exerted neuroprotective effect against oxidative stress not only in models of TBIbut also in SAH, I/R, and oxygen-glucose deprivation (OGD) models. Moreover, SIRT3 protects dopaminergic neurons against oxidative stress. The degeneration of dopaminergic neurons is the primary cause of PD. SIRT3 regulation of mnSOD activity via deacetylation at the lysine 68 site attenuates oxidative stress and restores mitochondrial membrane potential in dopaminergic neurons. Collectively, these findings support SIRT3-induced attenuation of oxidative stress after TBI and related neurodegeneration.
SIRT6 locates in the nucleus, where it promotes resistance to DNA damage and oxidative stress by supporting DNA doublestrand break repair. SIRT6 levels decrease after I/R injury. Nuclear factor erythroid-derived 2like 2 (NRF2) is activated via the deacetylation by SIRT6. NRF2 is a basic leucine zipper transcription factor that regulates the expression of antioxidant proteins, such as heme oxygenase-1 and SOD, that resist oxidative stress and have a neuroprotective function. Although SIRT7 is expressed ubiquitously in the cortex, striatum, hippocampus, and thalamus, its function in the brain is unclear. A recent study in an I/R model found that SIRT7 regulated apoptosis and resisted oxidative stress via deacetylation of the p53 protein. This finding suggests that SIRT6 and 7 may have similar effects on oxidative stress after TBI.
## Sirtuins mitigate proteinopathies
Sirtuins are a kind of longevity assurance factor and involved in the promotion of healthy aging mechanisms. Sirtuins have some special functions on lowering the risk of neurodegenerative diseases such as reducing the accumulation of tau proteins and amyloid β-peptide.
Tau proteins and amyloid β-peptide accumulations serve as indispensable parts in the development of AD. Cytological studies have shown that SIRT1 in growth cones promotes axon genesis and elongation in embryonic hippocampal neurons. The accumulation of amyloid precursor protein and amyloid β-peptide in injured axons after TBI has been demonstrated in vivo and in vitrostudies. These pathologies are associated with AD . SIRT1 locates in the distal regions of the axon, the axonal growth cone in particular. SIRT1 regulates the deacetylation and activation of protein kinase B (Akt). Akt is the upstream inhibitory kinase of glycogen synthase kinase3 (GSK3). Akt activation inhibits GSK3 activity. GSK3 is a multifunctional serine/threonine kinase and a key regulator of neurogenesis, polarization, neurite outgrowth, and plasticity in the nervous system. Intracellular mechanisms involved in axon genesis include actin filaments and the reorganization of microtubules. In addition to reducing amyloid β-peptide accumulation, SIRT1 upregulates the amount of lysosome in primary astrocytes to promote the degradation of amyloid β-peptide by its deacetylase activity . Recent studies showed that SIRT1 may exert its antiamyloidogenic effects by direct activating gene ADAM10, which encodes disintegrin protein. Furthermore, an in vitro study in Chinese hamster ovarian cells found that SIRT3 reduced ROS production and lipid peroxidation and improved mitochondrial function, resulting in the reduction of amyloid β-peptide production.
Hyperphosphorylated tau protein accumulation is widespread after a single TBI. Hyperphosphorylated tau proteins misfold and aggregate to form NFTs, consequently resulting in impaired tau function. Hyperphosphorylated tau proteins are liberated into the extracellular space, stimulating microglia and astrocytes to release pro-inflammatory cytokines (e.g., IL-1β and TNFα), which further activate tau kinases and aggravate tau hyperphosphorylation. In addition to phosphorylation, acetylation is a crucial step in the posttranslational modification of tau proteins. The acetylation of tau protein has been demonstrated in all stages of AD. An in vitro study in cortical neurons argued that SIRT3 deacetylated tau proteins, which attenuated the pathological accumulation of tau proteins.research has shown that SIRT1 suppressed tau protein accumulation in hippocampus neurons in a rat model of brain insulin resistance.
## Sirtuins and bbb integrity
The BBB is an interface between the brain parenchyma and cerebral circulation. BBB permeability is regulated by pericytes, astrocytes, and cerebral microvascular endothelial cells through the expression of tight and adherent junctions. BBB disruption serves as one of the critical parts in the development of AD and other neurodegenerative disorders. Disruption of the BBB allows neurotoxic cells, pathogens, and inflammatory cytokines to leak from the cerebral circulation to the brain parenchyma, where they activate inflammatory and immune responses associated with severe TBI and the development of neurodegenerative diseases.
SIRT2 serves as a pivotal factor in BBB integrity. SIRT2 inhibition has been shown to increase BBB permeability and exacerbate inflammation and brain edema after TBI. Matrix metalloproteinase (MMP)-9 degrades tight junction proteins of BBB. SIRT2 inhibition stimulates MMP-9 expression and disturbs the integrity of BBB, allowing plasma components (e.g., macrophage and inflammatory factors) crossing the BBB more easily, consequently resulting in the exacerbation of neuroinflammation. Moreover, SIRT2 inhibition promotes microglia and macrophage activation, which intensifies the inflammatory processes. SIRT2 inhibition promotes nuclear translocation of NF-kB p65 via acetylation of the p65 subunit to increase NF-kB activity and upregulate its targets including aquaporin 4, MMP-9, and proinflammatory cytokines. By contrast, SIRT5 adversely affects the maintenance of BBB integrity. SIRT5 has previously been observed to exacerbate inflammation in an experimental model of I/R injury, while SIRT5 inhibition decreased endothelial permeability of the BBB and increased the expression of occludin and claudin-5 via the phosphatidylinositol 3-kinase/Akt pathway.
In experimental models of SAH and aging, SIRT1 prevents BBB hyperpermeability. This effect can be reversed by inhibiting SIRT1. Tight junction proteins, such as claudin-5 and occludin, are the main components of the BBB structure. SIRT1 upregulates tight junction protein expression in response to brain edema to protect BBB integrity. In addition, SIRT1 attenuates damage caused by brain edema via the deacetylation of p53. P53 is a protein upregulating endothelial MMP-9 via the NF-kB pathway to reduce occlusion in animal models of SAH. Furthermore, a study of ischemic brain tissue from mice found that SIRT1 was involved in the repair of BBB vascular destruction by upregulating vascular endothelial growth factor expression in astrocytes. SIRT1 was proved to accelerate endothelial cell apoptosis in an animal model of SAH. Furthermore, a recent study suggested that endothelial SIRT6 protected BBB integrity after an I/R injury in a stroke model; however, the mechanism is not fully understood.
## Sirtuins and neuronal apoptosis
Sirtuins reduce the apoptosis of normal neurons and promote apoptosis in damaged neurons. TBI induces SIRT1 expression and activates the MAPK/extracellular signal-related kinase (ERK) pathway. Inhibition of SIRT1 and the MAPK/ERK pathway decreases neuronal apoptosis. Moreover, inhibition of the MAPK/ERK pathway further decreases SIRT1 levels. The dynamic relationship between SIRT1 and apoptosis has been demonstrated in in vivo and in vitro models of TBI. Moreover, sirtuins play an essential role in reducing apoptosis in other models of brain injury. SIRT1 is a nicotinamide-adenine dinucleotide-dependent p53 deacetylase. P53 induces apoptosis in the acetylated state. SIRT1 promotes p53 proteolysis via deacetylation to inhibit p53 transactivation activity and suppress apoptosis in response to oxidative stress in an experimental model of SAH. SIRT3 was found to promote autophagy of damaged neurons to protect ischemic neurons in a model of OGD. Cortical neurons overexpress SIRT3 under ischemic and hypoxia conditions, further increasing the phosphorylation of adenosine 5 -monophosphate (AMP)-activated protein kinase (AMPK), which inhibits the phosphorylation of mTOR and induces autophagy of ischemia neurons.
## Sirtuins as promising therapeutic targets
Given that primary TBIs are irreversible, interventional strategies to prevent secondary injuries and promote plasticity and the recovery process are of primary concern. As mentioned previously, the sirtuin family are vividly involved in several TBIinduced pathophysiological processes including microglial and astrocyte activation, inflammation, proteinopathies, oxidative stress, and mitochondrial damage, all of which may facilitate the onset of neurodegenerative diseases. Moreover, sirtuins have a wide range of effects on BBB integrity, apoptosis, and excitotoxicity in other in vivo and in vitro models of brain injury. As a result of these pleiotropic effects, sirtuins have been acknowledged to be emerging targets for TBI and TBI-related neurodegeneration.
There is a growing body of literature that recognizes the therapeutic potential of the modulators of sirtuins. Although these modulators have not been tested clinically, there is an abundance of preclinical data in cortical neuron cultures and animal models that describe specific mechanisms of modulator actions. Identification of a modulator that allows precise control of sirtuin expression will enable clinicians to treat TBI and prevent pathological processes that trigger late-onset neurodegenerative diseases.
Resveratrol is one of the most effective natural SIRT1-activating compounds. Resveratrol is a natural polyphenolic substance extracted from grapes that has been found to increase SIRT1 activity about 10-fold. By modifying at the B ring 4 position, scientists developed a variety of synthetic resveratrol derivatives to lower toxicity to cells or improve potency. Resveratrol and its derivatives play various roles in apoptosis , oxidative stress, and recovery of motor functions. Moreover, resveratrol modulates the activity of SIRT3 and SIRT5. Several other natural extracts also activate SIRT1, including magnolol, oxymatrine, curcumin, salvianolic acid B, arctigenin, astaxanthin , salidroside , ginsenoside, icariin, low molecular weight fucoidan , and saponin from Aralia taibaiensis. These findings suggest that Chinese traditional medicine may be a source of sirtuin modulators.
Moreover, some hormones modulate SIRT1. Melatonin, a kind of hormone mainly secreted by the pineal gland, regulates circadian rhythms. Melatonin is a powerful antioxidant that attenuates early brain injury-induced oxidative stress and neuronal apoptosis by activating the SIRT1/NF-kB pathway . Furthermore, melatonin upregulates the expression of SIRT3 to reduce the production of ROS and resist oxidative stress. Estrogen also regulates SIRT1. Estrogen restores resistance to cell apoptosis and attenuates ischemic brain injury via the SIRT1/AMPK pathway . Activation of the SIRT1/AMPK pathway may underlie the neuroprotective effect of estrogen.
Some drugs that are not originally developed to regulate sirtuins have currently been found to have neuroprotective properties via a SIRT-involved pathway. For instance, the longacting glucagon-like peptide-1 analogs, liraglutide, agomelatine, and dexmedetomidine, exert a SIRT1-dependent neuroprotective effect and critically regulate the metabolism. Furthermore, the antibiotic, minocycline, modulates SIRT3 to protect BBB integrity , and bexarotene attenuates neuroinflammation by modulating SIRT6.
Intermittent fasting, and daily caloricor protein restrictionshave been shown to reduce inflammation and oxidative stress by upregulating SIRT1. Specifically, reduced energy intake without affecting nutritional requirements has a neuroprotective effect after brain injury. Besides, docosahexaenoic acid, vitamin E, and hydrogen-rich salinereduce oxidative stress. The antioxidant, alpha-lipoic acid, reduces oxidative stress by upregulating SIRT1.
Besides, NAD + nicotinamide mononucleotide (NMN; precursor of NAD + ) provides neuroprotection by directly activating sirtuins . Dysregulation of NAD + metabolism may cause neurodegenerative diseases and acute brain injury. NMN is a multi-targeted modulator: it activates sirtuins, inhibits mitochondrial fission, and is an NAD+ supplement. Therefore, NMN is a promising therapeutic agent for TBI and associated neurodegenerative diseases . NMN is an endogenous cellular metabolic compound that may have lower toxicity and fewer side effects at high doses.
Furthermore, some enzymes modulate sirtuin. Nicotinamide phosphoribosyltransferase (Nampt) is the rate-limiting enzyme in mammalian NAD + biosynthesis. Nampt, which is located primarily in neurons, regulates autophagy and apoptosis. Rolipram, a phosphodiesterase-4 inhibitor, modulates inflammation and neuronal apoptosis via different SIRT1-dependent pathways.
However, a better understanding of sirtuins and their molecular structures is still needed to facilitate the discovery of additional modulators. Referring to, we can find that sirtuins seem to play contradictory roles under certain circumstances, which may bring difficulties to the development of TBI pharmacological agents.
# Conclusion and limitations
TBI is an accepted risk factor for neurodegenerative diseases. The key pathological features of TBI and neurodegenerative diseases include neuroinflammation, oxidative stress, excitotoxicity, proteinopathies, and mitochondrial dysfunction, with many similarities and overlaps. Therefore, targeting these processes has the therapeutic potential for TBI and associated neurodegenerative diseases. The sirtuin family have seven members, namely SIRT1-7 and are strictly implicated in the pathological mechanisms of acute brain injuries. Among these sirtuins, SIRT1 and SIRT3 have been demonstrated to reduce the post-TBI damage by alleviating inflammation, oxidative stress and proteinopathies, neuronal apoptosis, and BBB compromise. The rest sirtuins intervene in similar pathophysiology in other brain injury models and merit further exploration in TBI models. Recent progress on sirtuin modulators has made sirtuins promising therapeutic targets. However, sirtuins may show contradictory effects under certain conditions, and certain off-target effects should be taken into consideration. This review has some limitations. First, studies about sirtuins' effects on TBI are not enough yet, especially SIRT2, 4, 5, 6, and 7, which lack support from direct evidence in TBI models. We speculate that functions of SIRT2, 4, 5, 6, and 7 in other kinds of brain injury models may be also valid in TBI models. Second, the demonstration of the functions of sirtuins in this article is all based on animal test results. Hence, further exploration of human and clinical evidence is warranted. Third, the hypothetical linking between TBI and neurodegenerative diseases is supported by similarities of pathological processes. We should explore more supporting evidence to offer this association a more solid foundation.
# Author contributions
YZ and AS conceptualized the research project. QY, YZ, YuS, and YL wrote the manuscript and made the original figures. AS, YeS, and QY critically revised the texts and figures. AS, YeS, and YZ supervised the research and led the discussion. All authors contributed to the article and approved the submitted version.
# Funding
This work was funded by the National Natural Science Foundation of China (81701144). |
Coping with Oxidative Stress in Reproductive Pathophysiology and Assisted Reproduction: Melatonin as an Emerging Therapeutical Tool
Citation: Cosme, P.; Rodríguez, A.B.; Garrido, M.; Espino, J. Coping with Oxidative Stress in Reproductive Pathophysiology and Assisted Reproduction: Melatonin as an Emerging Therapeutical Tool.Antioxidants 2023, 12, 86. https:// Abstract: Infertility is an increasing global public health concern with socio-psychological implications for affected couples. Remarkable advances in reproductive medicine have led to successful treatments such as assisted reproductive techniques (ART). However, the search for new therapeutic tools to improve ART success rates has become a research hotspot. In the last few years, pineal indolamine melatonin has been investigated for its powerful antioxidant properties and its role in reproductive physiology. It is considered a promising therapeutical agent to counteract the detrimental effects associated with oxidative stress in fertility treatments. The aim of the present narrative review was to summarize the current state of the art on the importance of melatonin in reproductive physiology and to provide a critical evaluation of the data available encompassing basic, translational and clinical studies on its potential use in ART to improve fertility success rates.
# Introduction
Infertility is an increasing global public health issue that affects up to 16% of couples of reproductive age. According to the World Health Organization (WHO), infertility is a disease of the reproductive system defined by the failure to achieve a clinical pregnancy after 12 months of regular, unprotected sexual intercourse [bib_ref] The International Glossary on Infertility and Fertility Care, Zegers-Hochschild [/bib_ref]. Half of infertility cases are due to female factors [bib_ref] Some of the Factors Involved in Male Infertility: A Prospective Review, Babakhanzadeh [/bib_ref] , which are generally attributed to hormonal, functional or anatomical dysfunction of the organs of the reproductive tract [bib_ref] Investigation of Infertility Using Endometrial Organoids, Nikolakopoulou [/bib_ref]. Moreover, the age of the female is of great importance because of its physiological and genetic influence on conception, which includes a reduced ovarian follicular pool, perturbations in ovulation and increased meiotic errors within the oocyte [bib_ref] Physiological Aspects of Female Fertility: Role of the Environment, Modern Lifestyle, and..., Hart [/bib_ref]. As for male factors, these are responsible for 20-30% of infertility cases. The most important factors are hormonal deficits, physical causes, sexually transmitted problems, and genetic factors. Nevertheless, the origin of about 40% of male infertility is unknown [bib_ref] Some of the Factors Involved in Male Infertility: A Prospective Review, Babakhanzadeh [/bib_ref].
Melatonin is well-known for its potent antioxidant properties, as it can remove free radicals and convert reactive oxygen species (ROS) into less harmful species. In this way, melatonin protects lipids, proteins and DNA from oxidative damage [bib_ref] Melatonin as a Hormone: New Physiological and Clinical Insights, Cipolla-Neto [/bib_ref]. As a potent free radical scavenger, the antioxidant cascade of melatonin, i.e., melatonin and its secondary and tertiary metabolites, distinguishes it from other classic antioxidants. Given that the metabolites of melatonin, following its interaction with ROS, retain their ability to scavenge free radicals, one melatonin molecule has the capacity to scavenge up to ten ROS, this in contrast to the classic antioxidants that detoxify only one or less ROS. In this way, melatonin can be hydroxylated by interaction with free radicals, something which causes an immediate intramolecular rearrangement when then leads to the formation of a third ring, thereby giving rise to a metabolite denominated as cyclic 3-hydroxymelatonin (c3OHM) [fig_ref] Figure 1: Cascade reaction of melatonin interaction with free radicals and its main metabolites [/fig_ref]. causes an immediate intramolecular rearrangement when then leads to the formation of a third ring, thereby giving rise to a metabolite denominated as cyclic 3-hydroxymelatonin (c3OHM) [fig_ref] Figure 1: Cascade reaction of melatonin interaction with free radicals and its main metabolites [/fig_ref]. Likewise, c3OHM is also a potent free radical scavenger that can be converted by two - OH to another key metabolite of melatonin, N 1 -acetyl-N 2 -formyl-5methoxykynuramine (AFMK). At the same time, AFMK can be generated by direct conversion of melatonin, which comprises various enzymatic, pseudoenzymatic, free radicalmediated and photochemical mechanisms. Among these, the most relevant is the enzymatic conversion of melatonin to AFMK by indoleamine 2,3-dioxygenase (IDO) or myeloperoxidase [fig_ref] Figure 1: Cascade reaction of melatonin interaction with free radicals and its main metabolites [/fig_ref]. Finally, AFMK is easily deformylated to N 1 -acetyl-5-methoxykynuramine (AMK). To date, two enzymes capable of catalysing this reaction have been identified, arylamine formamidase and hemoperoxidase. Moreover, the formation of AMK from AFMK by - OH has also been suggested. The deformylated product AMK appears to be a radical scavenger of considerably higher reactivity than AFMK because it easily undergoes single-electron transfer reactions, which can in turn generate further metabolites [fig_ref] Figure 1: Cascade reaction of melatonin interaction with free radicals and its main metabolites [/fig_ref] (for an extensive review, see. On the other hand, melatonin is a pleiotropic molecule that is not only involved in the control of circadian rhythms but also participates in the regulation of the immune response, inhibition of carcinogenesis, proliferation of stem cells and modulation of aging, among other things [bib_ref] Physiological and Pharmacological Perspectives of Melatonin, Samanta [/bib_ref]. Interestingly, this indolamine has been suggested as a promising agent for the management of reproductive disorders over the last two decades.
The interest in human reproduction has progressively grown over the years with the aim of generating new diagnostic and therapeutic tools to improve human fertility. However, it is a challenging subject of study due to population heterogeneity, ethical limits when experimenting with humans, high research costs and appropriate technologies. The purpose of this article review is to summarize the current state of knowledge of the importance of melatonin in reproductive physiology and to provide a critical evaluation of the data available from existing studies on its potential uses in assisted reproductive techniques (ART) to improve fertility success rates.
## Functions of melatonin in reproductive physiology
## Role of melatonin in sperm physiology
Spermatogenesis is a testosterone-dependent event driving male gamete differentiation and maturation in the testis. However, mature spermatozoa are maintained in a quiescent state within the testis and must be therefore activated for a successful fertilization. The phenomenon of sperm activation is referred to as sperm capacitation, which is a multistep process involving changes in sperm form and function that are induced by extracellular structures of the oocyte. Such changes include initiation of motility, chemotaxis (i.e., swimming toward the oocyte in response to chemical concentration gradients), binding to the oocyte coat, the acrosome reaction (i.e., the release of hydrolytic enzymes from the acrosome), oocyte matrix penetration, and fusion of the two plasma membranes. Parallelly, reciprocal sperm-induced oocyte activation occurs with structural and metabolic changes in the oocyte that result in fertilization and which ultimately trigger embryo development (for a detailed review, see [bib_ref] Gamete Activation: Basic Knowledge and Clinical Applications, Tosti [/bib_ref]. In this context, melatonin can pass through the blood-testis barrier and enter testicular cells. Moreover, the indoleamine acts via membrane melatonin receptors 1 (MT1) and 2 (MT2), which are G-protein-coupled receptors that control testosterone synthesis, and hence spermatogenesis, by regulating cyclic adenosine monophosphate (cAMP) transduction cascades [bib_ref] Melatonin Regulates the Synthesis of Steroid Hormones on Male Reproduction: A Review, Yu [/bib_ref].
Normal sperm function depends on low levels of ROS generation in order to promote the signal transduction pathways associated with capacitation, binding to the zona pellucida (i.e., the oocyte extracellular coat) and sperm chromatin condensation. When generated in excess, however, ROS can induce lipid peroxidation that, in turn, disrupts the membrane characteristics that are critical for the maintenance of sperm function, including the capacity to fertilize an egg [bib_ref] Reactive Oxygen Species and Sperm Function-in Sickness and in Health, Aitken [/bib_ref]. Taking advantage of its outstanding antioxidant and free radical scavenger properties, melatonin reduces oxidative damage in mitochondria, DNA fragmentation, lipid peroxidation of plasma membrane and apoptotic markers in spermatozoa [bib_ref] Role of Melatonin in Male Reproduction, Bhattacharya [/bib_ref]. Nonetheless, this protection is not only due to its free radical scavenger properties but also to its action on the MT1 membrane receptor [bib_ref] Melatonin Protects Human Spermatozoa from Apoptosis via Melatonin Receptor-and Extracellular Signal-Regulated Kinase-Mediated..., Espino [/bib_ref]. Furthermore, melatonin regulates sperm capacitation through the modulation of bicarbonate secretion and mobilization of intracellular calcium, which is dependent on the MT2 receptor.
## Influence of melatonin on ovarian follicle development and ovulation
Melatonin functions in female reproduction are based on its direct actions in the ovary [bib_ref] The Role of Melatonin as an Antioxidant in the Follicle, Tamura [/bib_ref]. The expression of MT1 and MT2 receptors in granulosa cells, luteal cells, antral follicles, and corpus luteum of rats indicates that it has essential roles in the regulation of mammalian reproductive processes. In fact, melatonin modulates granulosa cells steroidogenesis and follicular function in rodents and humans [bib_ref] Modulation of Human Ovarian Function by Melatonin, Rai [/bib_ref].
Folliculogenesis is the intricated process of ovarian follicle formation and is fundamentally dependent on circulating levels of FSH. The formation of ovarian follicles relies on low levels of ROS as well, as they act as second messengers modulating the expression of genes involved in oocyte maturation. Nevertheless, an excess of ROS may produce oxidative stress that can cause damage to granulosa and oocyte cells in the follicle [bib_ref] Management of Ovarian Functions by Melatonin, Haldar [/bib_ref]. For this reason, melatonin is essential for the maintenance of the pro-oxidant-antioxidant balance of the oocyte, thus protecting the female gamete from oxidative damage and regulating a healthy folliculogenesis [bib_ref] Modulation of Human Ovarian Function by Melatonin, Rai [/bib_ref] [bib_ref] Management of Ovarian Functions by Melatonin, Haldar [/bib_ref]. Melatonin is found in follicular fluid at high concentrations (three-times higher than blood levels), where it stimulates granulosa cell proliferation through activation of mitogen-activated protein kinases (MAPK). Melatonin increases proportionally with the growth of the follicle [bib_ref] Antioxidative Action of Melatonin and Reproduction, Tamura [/bib_ref] so that the larger the follicles the greater the concentration of melatonin. This high concentration of the indolamine during the preovulatory phase is involved in the production of progesterone, which leads to luteinization and therefore to a successful ovulation [bib_ref] Modulation of Human Ovarian Function by Melatonin, Rai [/bib_ref].
## Melatonin and luteal phase
Melatonin levels increase during the luteal phase as compared with the follicular phase, thus suggesting that the indolamine has a direct action in the modulation of this phase. Furthermore, melatonin binding sites as well as MT1 and MT2 receptors have been found to be expressed in human granulosa lutein cells, which is in line with the fact that melatonin stimulates progesterone release in human granulosa lutein cells [bib_ref] Modulation of Human Ovarian Function by Melatonin, Rai [/bib_ref] [bib_ref] Hormone of Darkness" and Human Reproductive Process: Direct Regulatory Role of Melatonin..., Scarinci [/bib_ref]. On the other hand, melatonin also acts in the balance between luteotrophic and luteolytic regulators by inducing an increase of luteotrophic prostaglandin E 2 (PgE2) and a reduction of the luteolytic modulator PgF2α [bib_ref] Hormone of Darkness" and Human Reproductive Process: Direct Regulatory Role of Melatonin..., Scarinci [/bib_ref].
In women suffering luteal phase defect, which is characterized by low blood flow and ROS-evoked oxidative stress, it has been observed that melatonin provides protection to granulosa lutein cells and increases progesterone production by corpus luteum via a reduction in oxidative stress [bib_ref] Melatonin as a Free Radical Scavenger in the Ovarian Follicle, Tamura [/bib_ref].
## Effects of melatonin in the placenta
The placenta synthesizes melatonin de novo and expresses melatonin MT1 and MT2 receptors, through which melatonin promotes placental cell survival [bib_ref] Placental Melatonin Production and Melatonin Receptor Expression Are Altered in Preeclampsia: New..., Lanoix [/bib_ref]. Melatonin performs anti-apoptotic effects in the villous cytotrophoblasts, avoiding their excessive cell death, and acts as an antioxidant in the syncytiotrophoblasts. Furthermore, melatonin helps to maintain homeostatic processes in the placenta, which reduces the probability of pathologies such as preeclampsia. Preeclampsia is a systemic maternal-foetal disorder characterized by hypertension after twenty weeks of gestation and is associated with maternal organ dysfunction and/or foetal growth restriction. In preeclampsia, circulating levels of melatonin as well as its synthesis and receptor abundance are decreased [bib_ref] Melatonin Enhances Antioxidant Molecules in the Placenta, Reduces Secretion of Soluble Fms-like..., Hannan [/bib_ref]. For this reason, melatonin treatment could be potentially useful in this disorder, as suggested by [bib_ref] Melatonin Enhances Antioxidant Molecules in the Placenta, Reduces Secretion of Soluble Fms-like..., Hannan [/bib_ref] , who have demonstrated that melatonin increased the expression of several antioxidant enzymes, including thioredoxin (TXN) in primary trophoblasts, placental explants and human umbilical vein endothelial cells (HUVEC), glutamate-cysteine ligase (GCLC) in placental explant tissue and HUVEC cells, and quinone acceptor oxidoreductase 1 (NQO1) in placental explant tissue [bib_ref] Melatonin Enhances Antioxidant Molecules in the Placenta, Reduces Secretion of Soluble Fms-like..., Hannan [/bib_ref].
## Actions of melatonin during parturition
Melatonin levels increase with advancing gestation and reach their peak during labour, which suggests that the indolamine helps promote uterine contractions [bib_ref] A Double-Blind Randomised Placebo-Controlled Trial of Melatonin as an Adjuvant Agent in..., Swarnamani [/bib_ref]. In fact, at the end of pregnancy, uterine contractions are stronger at night, when the concentration of circulating melatonin is the highest [bib_ref] Riding the Rhythm of Melatonin through Pregnancy to Deliver on, Mccarthy [/bib_ref]. Interestingly, the myometrium expresses both oxytocin and melatonin MT2 receptors, whose expression are also increased during parturition. Melatonin and oxytocin trigger the same signalling pathway involving phospholipase C (PLC) and protein kinase C (PKC), which leads to the activation of myosin light chain kinase (MYLK) and ultimately promotes uterine muscle contractions. It has been demonstrated that melatonin acts synergistically with, and sensitizes uterine muscle to, oxytocin, thus producing its maximal contraction [bib_ref] Actions and Implication for Successful Pregnancies, Sagrillo-Fagundes [/bib_ref]. Moreover, melatonin increases the expression of protein connexin 43, which is necessary for myometrial cell communication and uterine contractions synchronization [bib_ref] A Double-Blind Randomised Placebo-Controlled Trial of Melatonin as an Adjuvant Agent in..., Swarnamani [/bib_ref]. After childbirth, serum melatonin levels decrease rapidly [bib_ref] A Double-Blind Randomised Placebo-Controlled Trial of Melatonin as an Adjuvant Agent in..., Swarnamani [/bib_ref].
## Influence of melatonin on seasonal reproduction
Melatonin is an important rhythmic hormone that regulates the circadian and seasonal rhythms of the body, thereby affecting the reproductive physiology of seasonally reproducing animals [bib_ref] Seasonal Reproduction in Vertebrates: Melatonin Synthesis, Binding, and Functionality Using Tinbergen's Four..., Vivid [/bib_ref]. On one hand, seasonal reproduction assures that the offspring are delivered at the time of the year that maximizes their survival, which is mostly spring or early summer. Therefore, these species are reproductively mature when the day lengths are long with a brief nocturnal melatonin peak (e.g., hamster, horse). On the other hand, short day breeders (e.g., buffalo, sheep, deer) successfully breed during the winter when the duration of the nocturnal melatonin rise is prolonged. Importantly, in both long-and short-day breeding mammals, the annual cycle of reproductive competence and mating is regulated by the changing duration of the nocturnal melatonin peak [bib_ref] Historical Perspective and Evaluation of the Mechanisms by Which Melatonin Mediates Seasonal..., Reiter [/bib_ref]. Due to the important role of melatonin in the reproductive function of these animals, multiple studies have investigated the effect of exogenous melatonin out of the breeding season. Generally, findings of these studies have demonstrated that melatonin successfully improved sperm quality, as well as conception and fertilization rates in different animal species [bib_ref] Melatonin Treatment in Winter and Spring and Reproductive Recovery in Sarda Breed..., Mura [/bib_ref] [bib_ref] Exogenous and Endogenous Factors in Seasonality of Reproduction in Buffalo: A Review, D'occhio [/bib_ref] [bib_ref] Effect of Melatonin Treatment on Semen Parameters and Endocrine Function in Black..., Egerszegi [/bib_ref] [bib_ref] The Effect of Exogenous Melatonin during the Non-Reproductive Season on the Seminal..., Casao [/bib_ref] [bib_ref] Effectiveness of Melatonin and Controlled Internal Drug Release Device Treatment on Reproductive..., Ramadan [/bib_ref] [bib_ref] Melatonin-Improved Buffalo Semen Quality during Nonbreeding Season under Tropical Condition, Ramadan [/bib_ref] [bib_ref] Manipulation of Reproductive Seasonality Using Melatonin Implantation in Anglo-Nubian Does Treated with..., El-Mokadem [/bib_ref] [bib_ref] Effects of Melatonin Administration on Seminal Plasma Metabolites and Sperm Fertilization Competence..., Satta [/bib_ref]. However, more investigations are necessary to give a definite recommendation on the use of melatonin as treatment to increase reproductive efficiency, thus allowing for increased productivity and flexibility of reproductive management. For an extensive review on this topic, the reader is encouraged to check the recent article [bib_ref] Historical Perspective and Evaluation of the Mechanisms by Which Melatonin Mediates Seasonal..., Reiter [/bib_ref] and the references therein.
## Melatonin application in assisted reproductive techniques (art)
In recent years, many studies have investigated the potential use of melatonin in ART to improve success rates. These techniques include oocyte manipulation, artificial insemination, in vitro fertilization (IVF), and embryo culture and transfer [bib_ref] Importance of Melatonin in Assisted Reproductive Technology and Ovarian Aging, Tamura [/bib_ref].
## Sources of reactive oxygen species (ros) in art
Despite the physiological role of ROS on gamete structure and function, an exacerbated production of ROS could be detrimental for gamete physiology [bib_ref] Reactive Oxygen Species and Sperm Function-in Sickness and in Health, Aitken [/bib_ref] [bib_ref] Management of Ovarian Functions by Melatonin, Haldar [/bib_ref]. In this sense, there is a higher risk of oxidative stress during ART procedures compared with in vivo physiological conditions. The reason is the lack of physiological defence mechanisms and the presence of intrinsic sources of ROS such as oocytes, cumulus mass cells, spermatozoa and leukocytes [bib_ref] Utility of Antioxidants during Assisted Reproductive Techniques: An Evidence Based Review, Agarwal [/bib_ref]. Likewise, there are extrinsic factors responsible for ROS generation such as culture media, pH, temperature, oxygen concentration, centrifugation and cryopreservation [bib_ref] Utility of Antioxidants during Assisted Reproductive Techniques: An Evidence Based Review, Agarwal [/bib_ref] [bib_ref] Radicals Generation in an in Vitro Fertilization Setting and How to Minimize..., Lampiao [/bib_ref] [fig_ref] Figure 2: Sources of reactive oxygen species [/fig_ref]. Culture media used during ART have an important impact on embryo quality and, consequently, on treatment success [bib_ref] Utility of Antioxidants during Assisted Reproductive Techniques: An Evidence Based Review, Agarwal [/bib_ref]. Some culture media contain metallic ions such as iron and copper. These ions lead to ROS generation, which implies that supplementation of the culture media with antioxidants could be beneficial to reduce ROS formation [bib_ref] Sources of ROS in ART, Agarwal [/bib_ref]. Culture media used during ART have an important impact on embryo quality and, consequently, on treatment success [bib_ref] Utility of Antioxidants during Assisted Reproductive Techniques: An Evidence Based Review, Agarwal [/bib_ref]. Some culture media contain metallic ions such as iron and copper. These ions lead to ROS generation, which implies that supplementation of the culture media with antioxidants could be beneficial to reduce ROS formation [bib_ref] Sources of ROS in ART, Agarwal [/bib_ref]. The maintenance of pH is also an important variable in culture media, as it influences sperm motility and its binding to oocyte, oocyte maturation and embryo development [bib_ref] Biological PH Buffers in IVF: Help or Hindrance to Success, Will [/bib_ref]. The maintenance of culture media pH is highly dependent on levels of CO 2 and temperature, which should remain constant at 5% and 37 - C, respectively. Nevertheless, in the case of those procedures carried out outside an incubator, the pH is maintained by using culture media with reduced levels of bicarbonate or by including a pH buffer [bib_ref] Sources of ROS in ART, Agarwal [/bib_ref]. Furthermore, high atmospheric oxygen concentrations can influence embryo quality due to oxidative stress induction. This is the rationale behind the use of low atmospheric oxygen concentrations (5%) in some ART laboratories to mimic in vivo conditions-where others may use an oxygen concentration of 20% [bib_ref] Low Oxygen Concentrations for Embryo Culture in Assisted Reproductive Technologies, Mantikou [/bib_ref]. In fact, it has been demonstrated that an atmospheric oxygen concentration of 5% increases embryo quality, implantation and pregnancy rates, as well as live birth rates compared with an oxygen concentration of 20% [bib_ref] The Effect of Two Distinct Levels of Oxygen Concentration on Embryo Development..., Kasterstein [/bib_ref].
Regarding spermatozoa preparation, centrifugation is a common step to separate spermatozoa from the seminal plasma and other components such as death cells, immature spermatozoa and leukocytes [bib_ref] Sources of ROS in ART, Agarwal [/bib_ref]. However, spinning sperm cells for more than 10 min leads to increased levels of ROS [bib_ref] Radicals Generation in an in Vitro Fertilization Setting and How to Minimize..., Lampiao [/bib_ref]. Furthermore, prolonged centrifugation times increase temperature of the sample, which also affects sperm motility [bib_ref] Sources of ROS in ART, Agarwal [/bib_ref]. Finally, cryopreservation is an ultra-low-temperature technique to maintain cells and tissues (from −80 - C to −196 - C) [bib_ref] Human Sperm Cryopreservation: Update on Techniques, Effect on DNA Integrity, and Implications..., Santo [/bib_ref]. This method is the best option to preserve human gametes; however, freeze-thaw cycles dramatically increase ROS production and reduce antioxidant defences of spermatozoa, thus rendering them more sensitive to oxidative stress. At the same time, this oxidative stress leads to lipid peroxidation of the sperm membrane [bib_ref] Antioxidants in Sperm Cryopreservation, Majzoub [/bib_ref].
The field of reproductive medicine has achieved remarkable advances in the last few years. However, ongoing research is nowadays focused on enhancing the success rates of infertility treatments. For this purpose, in vitro and in vivo studies have concentrated their efforts on the application of antioxidants (e.g., vitamins C, D, E, resveratrol, and quercetin) [bib_ref] Effect of Vitamins on the Quality of Insemination Doses of Bulls, Hashim [/bib_ref] [bib_ref] Role of Vitamin E and D3 Supplementation in Intra-Cytoplasmic Sperm Injection Outcomes..., Fatemi [/bib_ref] [bib_ref] Resveratrol Improves the Mitochondrial Function and Fertilization Outcome of Bovine Oocytes, Takeo [/bib_ref] [bib_ref] Resveratrol and Epigallocatechin-3-Gallate Addition to Thawed Boar Sperm Improves in Vitro Fertilization, Gadani [/bib_ref] [bib_ref] Resveratrol Promotes the Embryonic Development of Vitrified Mouse Oocytes after in Vitro..., Wang [/bib_ref] [bib_ref] The Effect of Quercetin on Fertility of Frozen-Thawed Ram Epididymal Spermatozoa, Ardeshirnia [/bib_ref] , and particularly melatonin, in ART to counteract the negative effects of oxidative stress due to its free radical scavenging properties [fig_ref] Figure 3: Applications and benefits of melatonin in assisted reproductive techniques [/fig_ref] [bib_ref] Antioxidative Action of Melatonin and Reproduction, Tamura [/bib_ref].
Antioxidants 2023, 11, x FOR PEER REVIEW 7 of 22 [bib_ref] Effect of Vitamins on the Quality of Insemination Doses of Bulls, Hashim [/bib_ref] [bib_ref] Role of Vitamin E and D3 Supplementation in Intra-Cytoplasmic Sperm Injection Outcomes..., Fatemi [/bib_ref] [bib_ref] Resveratrol Improves the Mitochondrial Function and Fertilization Outcome of Bovine Oocytes, Takeo [/bib_ref] [bib_ref] Resveratrol and Epigallocatechin-3-Gallate Addition to Thawed Boar Sperm Improves in Vitro Fertilization, Gadani [/bib_ref] [bib_ref] Resveratrol Promotes the Embryonic Development of Vitrified Mouse Oocytes after in Vitro..., Wang [/bib_ref] [bib_ref] The Effect of Quercetin on Fertility of Frozen-Thawed Ram Epididymal Spermatozoa, Ardeshirnia [/bib_ref] , and particularly melatonin, in ART to counteract the negative effects of oxidative stress due to its free radical scavenging properties [fig_ref] Figure 3: Applications and benefits of melatonin in assisted reproductive techniques [/fig_ref] [bib_ref] Antioxidative Action of Melatonin and Reproduction, Tamura [/bib_ref].
## Effect of melatonin in oocyte quality and embryo quality
Numerous in vitro studies have supplemented culture media with melatonin so as to enhance oocyte maturation, oocyte fertilization and embryo development [bib_ref] Antioxidative Action of Melatonin and Reproduction, Tamura [/bib_ref]. This approach assumes that oxidative stress accelerates apoptosis in oocytes and hence influences their capacity for fertilization. In fact, animal studies have shown that oxidative stress oc-
## Effect of melatonin in oocyte quality and embryo quality
Numerous in vitro studies have supplemented culture media with melatonin so as to enhance oocyte maturation, oocyte fertilization and embryo development [bib_ref] Antioxidative Action of Melatonin and Reproduction, Tamura [/bib_ref]. This approach assumes that oxidative stress accelerates apoptosis in oocytes and hence influences their capacity for fertilization. In fact, animal studies have shown that oxidative stress occurs after oocyte in vitro incubation for only 8 h; however, supplementation of oocyte culture media with 1 mM melatonin markedly relieved such a stress in a time-dependent manner in mouse oocytes, thereby delaying the onset of apoptosis. Moreover, melatonin supplementation also significant improved embryo quality [bib_ref] Melatonin Prevents Postovulatory Oocyte Aging in the Mouse and Extends the Window..., Lord [/bib_ref]. The addition of 1 µM melatonin in oocyte culture media has also been proved with a prolonged in vitro incubation time of 52 h. The results demonstrate that the indoleamine was able to enhance the quality and development of porcine oocytes by decreasing ROS generation, apoptosis and DNA damage [bib_ref] Il Jin, D. Melatonin Supplementation during Prolonged in Vitro Maturation Improves the..., Lin [/bib_ref]. Similarly, low melatonin doses, i.e., 1 nM and 0.1 µM, improved the production and quality of bovine blastocysts, substantially increased the expression of important genes related to embryo development such as DNA methyltransferase 3a (DNMT3A), occludin (OCC) and cadherin (CDH1), and decreased the expression of aquaporin 3 (AQP3), which leads to an enhanced resistance to apoptosis [bib_ref] Melatonin Reduces Apoptotic Cells, SOD2 and HSPB1 and Improves the in Vitro..., Marques [/bib_ref] [bib_ref] Melatonin Improves the Quality of in Vitro Produced (IVP) Bovine Embryos: Implications..., Wang [/bib_ref].
Ovarian aging is characterized by a gradually depleted number of primordial follicles and a diminished quality of oocytes, thus causing a progressive reduction of fertility. An investigation with 10-week-old female mice has demonstrated that the administration of water containing 100 µg/mL of melatonin delays ovarian aging. This supplementation, kept until mice were 43-weeks old, resulted in a higher number of primordial, primary and antral follicles, as well as better fertilization and blastocysts rates in treated mice in comparison with littermate control mice [bib_ref] Long-Term Melatonin Treatment Delays Ovarian Aging, Tamura [/bib_ref]. Additionally, melatonin significantly enhanced telomere length in old mice, and improved the expression of aging-related genes, such as sirtuins (SIRT1, SIRT3) and the autophagy-related gene microtubule-associated protein light chain 3 (LC3). Melatonin has also been shown to be able to upregulate 40 ribosome-related genes that are commonly downregulated during aging, these results demonstrating the capacity of melatonin to delay ovarian aging [bib_ref] Long-Term Melatonin Treatment Delays Ovarian Aging, Tamura [/bib_ref].
As for in vivo human studies, several trials have investigated the efficacy of melatonin administration to patients who underwent in vitro fertilization and embryo transfer (IVF-ET) procedure with the idea of rising follicular melatonin concentrations and hence improving oocyte quality [bib_ref] Importance of Melatonin in Assisted Reproductive Technology and Ovarian Aging, Tamura [/bib_ref]. In this sense, oral supplementation with 3 mg/day melatonin in women undergoing IVF-ET increased the percentage of mature oocytes and the number of top-quality embryos, although no significant differences were observed in fertilization rates and clinical pregnancy rates compared to a control group [bib_ref] The Efficacy of Melatonin Administration on Oocyte Quality, Batiioǧlu [/bib_ref]. The same melatonin dosage was also tested in patients with poor oocyte and embryo quality and resulted in a better fertilization rate in the second cycle of IVF-ET in comparison with the first cycle without melatonin supplementation [bib_ref] Oral Melatonin Supplementation Improves Oocyte and Embryo Quality in Women Undergoing in..., Nishihara [/bib_ref]. Other studies have been carried out in women with diminished ovarian reserve who received 3 mg/day melatonin commencing the fifth day of their menstrual cycle till the day of follicular puncture. The number of mature oocytes and top-quality embryos were higher in melatonin treated women than in the control group; however, no statistically significant differences were found in clinical pregnancy and spontaneous miscarriage rates between both groups [bib_ref] Effect of Melatonin on the Outcome of Assisted Reproductive Technique Cycles in..., Jahromi [/bib_ref]. In relation to unexplained infertility, oral supplementation with 3 mg/day or 6 mg/day of melatonin for 40 days rebalanced intrafollicular oxidative state, and enhanced oocyte quality and IVF success rates [bib_ref] Impact of Melatonin Supplementation in Women with Unexplained Infertility Undergoing Fertility Treatment, Espino [/bib_ref]. Furthermore, melatonin supplementation (5 mg/day) in IVF cycles was also effective in women aged over 40, and raised intrafollicular levels of indolamine, the number of mature oocytes and embryo quality [bib_ref] Beneficial Effects of Melatonin on Oocytes and Embryo Quality in Aged IVF..., Valeri [/bib_ref]. Nevertheless, all these positive findings differ from other studies in which melatonin was unable to ameliorate oocyte quality. For example, in a clinical trial with oral administration of different antioxidants, including 0.975 mg of melatonin, there was an observed improvement in embryo quality upon melatonin treatment, but with no significant differences in terms of the number of follicles, mature oocytes and clinical pregnancy rates [bib_ref] Randomized Prospective Study to Evaluate the Efficacy of Previous Therapy with Melatonin,..., Jiménez-Tuñón [/bib_ref]. Similarly, doses of 2, 4 and 8 mg of melatonin administered twice a day enhanced neither the number of mature oocytes and embryos nor clinical pregnancy rates, even though the dose of 8 mg resulted in higher concentrations of intrafollicular melatonin compared with the placebo group.
As in female patients, melatonin administration was also studied in infertile men to investigate its effects on sperm quality and the quality of the embryos retrieved from their couples when undergoing an IVF cycle. The results demonstrate that supplementation for 45 days of 6 mg melatonin/day promoted a remarkable increase of seminal total antioxidant activity and a reduction in sperm DNA oxidative damage. Moreover, embryos obtained from women whose male couple was taking melatonin experienced significant increment in the percentage of grade A (top quality), B (good quality) and C (impaired quality) embryos, but a decrease in grade D (poor quality, not recommended for ET) embryos, according to the Spanish Association for the Study of Reproductive Biology (ASEBIR) criteria [bib_ref] Exogenous Melatonin Supplementation Prevents Oxidative Stress-Evoked DNA Damage in Human Spermatozoa, Bejarano [/bib_ref].
## Application of melatonin in sperm preparation for art
Sperm preparation for ART aim at the selection and enrichment of motile and functionally competent spermatozoa from the ejaculate [bib_ref] Sperm Preparation for ART, Henkel [/bib_ref]. Starting from simple washing of spermatozoa, conventional techniques for the separation of spermatozoa from seminal plasma are based on different principles such as migration (which relies on the presence of motile spermatozoa within the semen sample, e.g., swim-up procedure), filtration (relies on sperm motility and the propensity of sperm to adhere to filtration matrices, e.g., glass wool filtration) and density gradient centrifugation (relies on sperm motility and the property of sperm to collect at the border between liquid phases, e.g., continuous density gradient with different media) [bib_ref] Identification and Preparation of Sperm for Art, Mehta [/bib_ref]. Different studies carried out in diverse animal models have proved the effects of melatonin supplementation, at different doses, during sperm preparation [bib_ref] Effect of Addition of Trehalose on the Liquid Storage (5 • C)..., Perumal [/bib_ref] [bib_ref] Nourmohammadi, Z. Melatonin Modulates the Expression of BCL-Xl and Improve the Development..., Dehghani-Mohammadabadi [/bib_ref] [bib_ref] Supplementation of IVF Medium with Melatonin: Effect on Sperm Functionality and in..., Cheuquemán [/bib_ref] [bib_ref] Effect of Melatonin on Mobility and Velocity Parameters of Mithun (Bos Frontalis)..., Perumal [/bib_ref] [bib_ref] Effects of Adding Melatonin on the Quality Offrozen-Thawed Boar Semen, Thongrueang [/bib_ref] [bib_ref] Melatonin Improves Rate of Monospermic Fertilization and Early Embryo Development in a..., Gutiérrez-Añez [/bib_ref] [bib_ref] Melatonin Improves the Quality of Frozen Bull Semen and Influences Gene Expression..., Su [/bib_ref]. In this regard, thawed bovine sperm samples were treated with 1 mM melatonin. The results demonstrate that the indolamine decreased the expression of proapoptotic genes such as caspase-3 and BAX and caused a dramatic rise in the expression of both the anti-apoptotic genes Bcl-2 and X-linked inhibitor of apoptosis protein (XIAP), and the antioxidant enzyme catalase (CAT). Likewise, in the same study, a concentration of 10 µM melatonin enhanced plasma membrane integrity and acrosome integrity, along with a reduction of the intracellular ROS levels [bib_ref] Protective Effects of Melatonin on Bovine Sperm Characteristics and Subsequent in Vitro..., Pang [/bib_ref]. The very same dose (10 µM) also proved to be efficient in sex-sorted bull semen as it protected semen samples against oxidative stress by increasing the activity of endogenous antioxidants such as Gpx, superoxide dismutase (SOD) and CAT, while inhibiting phosphatidylserine externalization and lipid peroxidation (measured as MDA levels), which are events related to apoptosis and acrosomal membrane integrity, respectively. Moreover, it was found that the dose of 10 µM melatonin led to an increment of the fertilization capacity and an enhancement in embryo development with respect to both the untreated control group and the different doses of the indolamine (1 nM, 0.1 µM and 1 mM) [bib_ref] Melatonin Improves the Fertilization Capacity of Sex-Sorted Bull Sperm by Inhibiting Apoptosis..., Li [/bib_ref].
The role of melatonin in sperm capacitation has also been studied, this process being necessary for the sperm cells to acquire fertilizing capacity [bib_ref] Sperm Capacitation and Acrosome Reaction in Mammalian Sperm, Stival [/bib_ref]. In this sense, low melatonin concentrations (100 pM) have been shown to modulate sperm capacitation by rising motile spermatozoa subpopulation in ram samples, thereby leading to better oocyte fertilization rates after IVF [bib_ref] Melatonin Reduces CAMP-Stimulated Capacitation of Ram Spermatozoa, Gimeno-Martos [/bib_ref] [bib_ref] Melatonin Prevents Capacitation and Apoptotic-like Changes of Ram Spermatozoa and Increases Fertility..., Casao [/bib_ref]. In fact, one of the first investigations reporting the involvement of melatonin in sperm motility modulation revealed that melatonin concentrations ranging from 1 pM to 1 µM, when added to supernatant after swim-up sperm selection, enhanced sperm hyperactivation, this action being dependent on MT1 receptor [bib_ref] Melatonin-Enhanced Hyperactivation of Hamster Sperm, Fujinoki [/bib_ref].
In relation to human studies, it has been observed that preincubation with 6 mM melatonin during sperm capacitation readily improved progressive motility and membrane integrity of asthenoteratozoospermic men samples [bib_ref] Ameliorative Effect of Melatonin versus the Passage of Time and Lipid Peroxidation..., Sohrabi [/bib_ref]. Moreover, it has been reported that the use of lower doses, i.e., 1 mM melatonin, also produced good results since it provoked a significant increment of spermatozoa suitable for oocyte fertilization. Additionally, melatonin (1 mM) enhanced migration of spermatozoa with compacted DNA in oligozoospermic human samples and also avoided DNA fragmentation in normozoosper-mic human samples [bib_ref] Melatonin Diminishes Oxidative Damage in Sperm Cells, Improving Assisted Reproductive Techniques, Monllor [/bib_ref] [bib_ref] High Endogenous Melatonin Concentrations Enhance Sperm Quality and Short-Term in Vitro Exposure..., Ortiz [/bib_ref]. Likewise, a concentration of 2 mM melatonin also displayed beneficial effects in human sperm motility. Thus, the addition of melatonin after swim-up resulted in an increase in the number of fast and progressively motile spermatozoa, along with an improved sperm viability [bib_ref] The in Vitro Effects of Melatonin on Human Sperm Function and Its..., Du Plessis [/bib_ref]. On the other hand, it has been demonstrated that a concentration of 1 mM melatonin exerted an anti-apoptotic effect in human spermatozoa treated with H 2 O 2 or progesterone due to the inhibition of caspase-3 and the activation of caspase-9, and also prevented phosphatidylserine externalization, which is one of the main hallmarks of apoptosis [bib_ref] Melatonin as a Potential Tool against Oxidative Damage and Apoptosis in Ejaculated..., Espino [/bib_ref]. Subsequent experiments have demonstrated that 1 mM melatonin managed to revert H 2 O 2 -induced DNA fragmentation and suggested that the protective effect of melatonin on sperm apoptosis is dependent on MT1 receptor and ERK signalling [bib_ref] Melatonin Protects Human Spermatozoa from Apoptosis via Melatonin Receptor-and Extracellular Signal-Regulated Kinase-Mediated..., Espino [/bib_ref]. Interestingly, the fact that the indoleamine prevents DNA oxidative damage is not only dependent on its antioxidant effect but also relies on its ability to regulate diverse DNA repair pathways, including, but not limited to, base excision repair, homologous recombination and mismatch mediated repair, by different mechanisms (for a detailed review, see [bib_ref] Melatonin: A Smart Molecule in the DNA Repair System, Mir [/bib_ref].
## Melatonin as protective agent in gametes cryopreservation
Sperm cryopreservation is the most commonly used method in cancer patients as they undergo aggressive treatments that can affect sperm quality and ultimately lead to azoospermia [bib_ref] Approaches and Technologies in Male Fertility Preservation, Huleihel [/bib_ref]. The main drawback of cryopreservation is that sperm quality can be negatively affected due to the freeze-thaw cycles, which may result in oxidative stress, lipid peroxidation increase and loss of plasma membrane integrity, hence disturbing the capacity of sperm-oocyte fertilization [bib_ref] Protective Effects of Melatonin on Male Fertility Preservation and Reproductive System, Sun [/bib_ref] [bib_ref] Targeted Antioxidant Delivery Modulates Mitochondrial Functions, Ameliorates Oxidative Stress and Preserve Sperm..., Tiwari [/bib_ref] [bib_ref] The Effect of Cryopreservation on the Genome of Gametes and Embryos: Principles..., Kopeika [/bib_ref] [bib_ref] Implication of Apoptosis in Sperm Cryoinjury, Said [/bib_ref] [bib_ref] Effect of Cryopreservation on Sperm Quality and Fertility, Lemma [/bib_ref] [bib_ref] Recent Advances in Bovine Sperm Cryopreservation Techniques with a Focus on Sperm..., Grötter [/bib_ref]. For this reason, the scientific community has focused its attention on the possible protective effects of melatonin on spermatozoa during cryopreservation given its powerful antioxidant action. This protection is associated with a reduction of lipid peroxidation events in sperm cells, which is related to the melatoninevoked increase in both total antioxidant capacity and activity of antioxidant enzymes [bib_ref] Role of Melatonin on Production and Preservation of Gametes and Embryos: A..., Cruz [/bib_ref].
Several studies have proved the role of melatonin as an effective cryoprotectant in sperm cryopreservation. For instance, it has been reported that the supplementation with 2 mM or 3 mM melatonin in the semen extender counteracted the adverse effects of freeze-thaw cycles in bull sperm, as it lessened lipid peroxidation and boosted total antioxidant capacity and activity of antioxidant enzymes [bib_ref] Antioxidative Effects of Melatonin on Kinetics, Microscopic and Oxidative Parameters of Cryopreserved..., Ashrafi [/bib_ref]. Similarly, the use of melatonin at a concentration of 100 µM or 1 mM improved both motility and viability parameters in cryopreserved buffalo semen samples, which was positively reflected in their in vitro fertilization capacity and the percentage of embryos obtained [bib_ref] Evidences for the Role of Melatonin as a Protective Additive during Buffalo..., El-Raey [/bib_ref] [bib_ref] Melatonin Supplementation Improved Cryopreserved Thai Swamp Buffalo Semen, Inyawilert [/bib_ref]. Moreover, it has also been documented that the addition of 1 mM melatonin in the cryoprotectant was efficient in various animal models such as rabbit, ram, horse, and dog. These studies proved that melatonin enhanced sperm DNA and acrosome integrity [bib_ref] Melatonin Can Improve Viability and Functional Integrity of Cooled and Frozen/Thawed Rabbit..., Fadl [/bib_ref] [bib_ref] Melatonin Added to Cryopreservation Extenders Improves the Mitochondrial Membrane Potential of Postthawed..., Lançoni [/bib_ref] [bib_ref] Supplementation of Melatonin to Cooling and Freezing Extenders Improves Canine Spermatozoa Quality..., Divar [/bib_ref] , which led to increased total cleavage rates and, hence, to higher fertilization and birth rates [bib_ref] Melatonin Can Improve Viability and Functional Integrity of Cooled and Frozen/Thawed Rabbit..., Fadl [/bib_ref] [bib_ref] Melatonin Protects Ram Spermatozoa from Cryopreservation Injuries in a Dose-Dependent Manner, Succu [/bib_ref]. Furthermore, the indolamine managed to decrease oxidative stress by ameliorating antioxidant enzymes activation and, therefore, reducing ROS concentration during cryopreservation process [bib_ref] Antioxidative Effect of Melatonin on Cryopreserved Chicken Semen, Appiah [/bib_ref]. Furthermore, the supplementation of the extender with 500 µM melatonin has been shown to improve the viability of post-thaw mouse sperm samples because of an increased expression of the anti-apoptotic gene B-cell lymphoma-extra-large Bcl-xL and a reduction in the percentage of viable spermatozoa with ROS overproduction [bib_ref] Effect of Melatonin Supplementation on Cryopreserved Sperm Quality in Mouse, Chen [/bib_ref].
On the other hand, studies carried out with human sperm samples have demonstrated that the use of 100 µM melatonin added as cryoprotectant was able to significantly raise sperm viability and membrane integrity, while diminishing intracellular ROS levels and lipid peroxidation. Of note, this supplementation did not have any detrimental effect on human sperm during cryopreservation [bib_ref] Melatonin Regulates the Synthesis of Steroid Hormones on Male Reproduction: A Review, Yu [/bib_ref] [bib_ref] Protective Effects of Melatonin on Male Fertility Preservation and Reproductive System, Sun [/bib_ref]. Likewise, other authors have also reported that different melatonin concentrations (10 µM and 3 mM) resulted in higher viability and motility of cryopreserved spermatozoa, and lower intracellular ROS levels [bib_ref] The Protective Effects of Melatonin against Cryopreservation-Induced Oxidative Stress in Human Sperm, Karimfar [/bib_ref] [bib_ref] Melatonin Affects Membrane Integrity, Intracellular Reactive Oxygen Species, Caspase3 Activity and AKT..., Najafi [/bib_ref]. Finally, a recent study investigated the effect of 2 mM caffeine added before cryopreservation in normozoospermic semen samples previously treated with 2 mM melatonin. The findings showed that the combination of caffeine and melatonin ameliorated sperm motility and mitochondrial activity compared with samples treated only with melatonin [bib_ref] Melatonin and Caffeine Supplementation Used, Respectively, as Protective and Stimulating Agents in..., Pariz [/bib_ref].
Regarding mature oocyte cryopreservation, this technique can be used for women facing anticipated fertility decline for various reasons, including gonadotoxic cancer therapies, surgeries with risk of damage to ovary or oophorectomies, and women with increased risk of primary ovarian insufficiency [bib_ref] Mature Oocyte Cryopreservation for Fertility Preservation, Liang [/bib_ref] [bib_ref] Cryopreservation of Oocytes, Schattman [/bib_ref] [bib_ref] Management of Women with Endometriosis: Guideline of the European Society of Human..., Dunselman [/bib_ref]. With the idea of minimizing cellular osmotic and/or oxidative stresses during this procedure, the use of antioxidants such as melatonin has been implemented in the last few years. In this sense, it has been reported that loading porcine cumulus-oocyte complexes with melatonin plus glycine (1 µM and 6 mM, respectively) during vitrification (an ultra-rapid method of cryopreservation) enhanced the developmental competency of vitrified porcine oocytes, while lessening levels of ROS and apoptotic occurrence in mature oocytes [bib_ref] Glycine and Melatonin Improve Preimplantation Development of Porcine Oocytes Vitrified at the..., Tang [/bib_ref]. Similarly, it has been demonstrated that the addition into vitrification media of melatonin and resveratrol (1 pM and 0.5 µM, respectively) co-encapsulated by solid lipid nanocarriers synergistically improved maturation, fertilization, and embryo development rates and decreased extra/intracellular ROS levels in mature oocytes [bib_ref] Enhanced Cryoprotective Effect of Melatonin and Resveratrol by Coencapsulation: Improved in Vitro..., Aghaz [/bib_ref]. Importantly, the indoleamine can also improve the effect of cryopreservation in human oocytes, as has been recently demonstrated [bib_ref] Melatonin Improves the Effect of Cryopreservation on Human Oocytes by Suppressing Oxidative..., Zhang [/bib_ref]. Apart from protecting oocytes during cryopreservation, melatonin has also been proven to foster in vitro maturation of vitrified mouse [bib_ref] Effects of Melatonin and Human Follicular Fluid Supplementation of in Vitro Maturation..., Doroudi [/bib_ref] [bib_ref] Melatonin Promotes in Vitro Maturation of Vitrified-Warmed Mouse GV Oocytes Potentially by..., Yang [/bib_ref] [bib_ref] Melatonin Improves the First Cleavage of Parthenogenetic Embryos from Vitrified-Warmed Mouse Oocytes..., Pan [/bib_ref] [bib_ref] Melatonin Improves Parthenogenetic Development of Vitrified-Warmed Mouse Oocytes Potentially by Promoting G1/S..., Pan [/bib_ref] [bib_ref] Melatonin Rescues the Aneuploidy in Mice Vitrified Oocytes by Regulating Mitochondrial Heat..., Gao [/bib_ref] [bib_ref] Anti-Apoptotic Regulation Contributes to the Successful Nuclear Reprogramming Using Cryopreserved Oocytes, Lee [/bib_ref] [bib_ref] Improved Development by Melatonin Treatment after Vitrification of Mouse Metaphase II Oocytes, Zhang [/bib_ref] and equine [bib_ref] Mitochondrial Function, Blastocyst Development and Live Foals Born after ICSI of Immature..., Clérico [/bib_ref] oocytes. Nevertheless, other studies found no effect of exogenous melatonin on development of cryopreserved oocytes in mouse [bib_ref] No Effect of Exogenous Melatonin on Development of Cryopreserved Metaphase II Oocytes..., Li [/bib_ref].
Altogether, these studies indicate that melatonin can be used as an effective cryoprotectant, which would acquire special clinical relevance in cryopreserved samples from oncological patients that choose this method for preserving their fertility.
## Impact of melatonin on reproductive organs pathophysiology
## Endometriosis
Endometriosis is associated with an exacerbated production of ROS due to an imbalance of oxidants and antioxidants and, therefore, the search for new treatments is focused on antioxidant therapy with the use of scavenging molecules such as melatonin [bib_ref] Potential Markers of Endometriosis: Latest Update, Gupta [/bib_ref].
In vitro studies with endometriotic epithelial cells derived from patients with endometriosis have shown that 1 mM of melatonin blocked 17β-estradiol-induced migration, invasion and epithelial-mesenchymal transition (EMT) through the upregulation of the Numb endocytic adaptor protein (Numb) and the low activity of the neurogenic locus notch homolog protein 1 (Notch1) [bib_ref] Melatonin Inhibits 17β-Estradiol-Induced Migration, Invasion and Epithelial-Mesenchymal Transition in Normal and Endometriotic..., Qi [/bib_ref].
On the other hand, a large number of animal studies have investigated the potential use of melatonin in endometriosis. Thus, an intraperitoneal injection of 10 mg melatonin/kg/day for 28 days in a rat model of surgically induced endometriosis caused volume and weight reduction of the implants via modulation of the expression of vascular endothelial growth factor (VEGF), which is involved in angiogenesis, and tissue inhibitor of metalloproteinase-2 (TIMP-2), which is significantly decreased in women with endometriosis [bib_ref] Melatonin Causes Regression of Endometriotic Implants in Rats by Modulating Angiogenesis, Tissue..., Yilmaz [/bib_ref] [bib_ref] Effects of Pinealectomy and Melatonin Supplementation on Endometrial Explants in a Rat..., Koc [/bib_ref]. Similar results have been observed with different doses of melatonin. For instance, the administration of 20 mg melatonin/kg/day for two weeks in surgically induced endometriotic rats produced a greater regression of endometriotic foci than that observed with letrozole, which is used to treat endometriosis, and the recurrence rate was also lower after the cessation of treatment compared with letrozole [bib_ref] The Effects of Letrozole and Melatonin on Surgically Induced Endometriosis in a..., Yildirim [/bib_ref]. Other studies have shown that the dose of 20 mg melatonin/kg/day was more efficient than 10 mg/kg/day, causing a higher regression of endometriotic lesions and a significant decrease of MDA in an oophorectomized rat endometriosis model and in a severe combined immunodeficient (SCID) mice endometriosis model [bib_ref] Melatonin Treatment Results in Regression of Endometriotic Lesions in an Ooferectomized Rat..., Kocadal [/bib_ref] [bib_ref] The Effects of Different Doses of Melatonin Treatment on Endometrial Implants in..., Cetinkaya [/bib_ref] [bib_ref] The Effects of Melatonin on Endometriotic Lesions Induced by Implanting Human Endometriotic..., Yesildaglar [/bib_ref]. Additionally, a higher dose of melatonin (48 mg/kg/day) administered for 10-20 days in ovariectomized mice promoted apoptosis in the endometriotic tissue mediated by a reduction of the anti-apoptotic B-cell lymphoma 2 gene (Bcl-2) and an increase of the proapoptotic Bcl-2-associated X protein (BAX) and caspase-9. Moreover, melatonin suppressed metalloproteinase-3 (MMP-3) expression, whose activity is associated with the formation of endometriotic lesions at an early stage [bib_ref] Melatonin Protects against Endometriosis via Regulation of Matrix Metalloproteinase-3 and an Apoptotic..., Paul [/bib_ref].
The potential use of melatonin for endometriosis has also been investigated in a phase II, randomized, double-blind, placebo-controlled trial. Interestingly, daily oral supplementation with 10 mg of melatonin for eight weeks reduced daily chronic pelvic pain scores (39.80%) and dysmenorrhea (38.01%) in patients with endometriosis. Melatonin also improved sleep quality and diminished the levels of brain-derived neurotrophic factor (BDNF), which is related to the pathogenesis of chronic pain, independently of pain levels [fig_ref] Table 1: Melatonin clinical trials on ovarian pathologies [/fig_ref] [bib_ref] Efficacy of Melatonin in the Treatment of Endometriosis: A Phase II, Randomized,..., Schwertner [/bib_ref]. Nevertheless, oral administration of 10 mg/day of melatonin was ineffective as an analgesic in women with dysmenorrhea but without signs of endometriosis [fig_ref] Table 1: Melatonin clinical trials on ovarian pathologies [/fig_ref] [bib_ref] Adjuvant Use of Melatonin for Pain Management in Dysmenorrhea-A Randomized Double-Blinded, Placebo-Controlled..., Söderman [/bib_ref] , contrary to the results obtained after administering 3 mg melatonin /day, which reduced pain and enhanced subjective sleep in women with primary dysmenorrhea [fig_ref] Table 1: Melatonin clinical trials on ovarian pathologies [/fig_ref] [bib_ref] Both Melatonin and Meloxicam Improved Sleep and Pain in Females with Primary..., Keshavarzi [/bib_ref]. While some previous findings lend support to the use of indoleamine [bib_ref] Melatonin Causes Regression of Endometriotic Implants in Rats by Modulating Angiogenesis, Tissue..., Yilmaz [/bib_ref] [bib_ref] Effects of Pinealectomy and Melatonin Supplementation on Endometrial Explants in a Rat..., Koc [/bib_ref] [bib_ref] The Effects of Letrozole and Melatonin on Surgically Induced Endometriosis in a..., Yildirim [/bib_ref] [bib_ref] Melatonin Treatment Results in Regression of Endometriotic Lesions in an Ooferectomized Rat..., Kocadal [/bib_ref] [bib_ref] The Effects of Different Doses of Melatonin Treatment on Endometrial Implants in..., Cetinkaya [/bib_ref] [bib_ref] The Effects of Melatonin on Endometriotic Lesions Induced by Implanting Human Endometriotic..., Yesildaglar [/bib_ref] [bib_ref] Melatonin Protects against Endometriosis via Regulation of Matrix Metalloproteinase-3 and an Apoptotic..., Paul [/bib_ref] [bib_ref] Efficacy of Melatonin in the Treatment of Endometriosis: A Phase II, Randomized,..., Schwertner [/bib_ref] , new clinical trials are warranted to convincingly demonstrate the effectiveness of melatonin as a possible pharmacological and/or adjuvant treatment in the management of endometriosis.
## Polycystic ovary syndrome (pcos)
Polycystic ovary syndrome (PCOS) is a common endocrine disorder that causes hyperandrogenism and infertility due to a dysfunctional follicular maturation and anovulation. Numerous animal studies have focused on the use of melatonin to investigate its putative effects on oocyte quality in PCOS. In this sense, the addition of 10 µM melatonin in the culture media of oocyte cumulus obtained from a PCOS female mouse model enhanced oocyte quality due to the lowered presence of free radicals, which resulted in increased fertilization rates [bib_ref] Effects of Melatonin on Oocyte Maturation in PCOS Mouse Model, Nikmard [/bib_ref]. A recent study using the same animal model and melatonin concentration showed that the indoleamine promoted an increment in both oocyte maturation-related genes, such as growth differentiation factor-9 (Gdf9) and bone morphogenetic protein 5 (Bmp5), and antioxidant enzymes, such as glutathione peroxidase 1 (Gpx1) and superoxide dismutase 1 (Sod1), compared to untreated group [bib_ref] The Boosting Effects of Melatonin on the Expression of Related Genes to..., Nikmard [/bib_ref]. Furthermore, there was an observed anti-apoptotic effect of melatonin mediated by a decline of the expression of BAX and a rise of anti-apoptotic Bcl-2. Finally, this study demonstrated an inverse correlation between levels of ROS and concentration of the indolamine in the culture media in the melatonin-treated groups [bib_ref] The Boosting Effects of Melatonin on the Expression of Related Genes to..., Nikmard [/bib_ref].
As for human studies, melatonin supplementation of in vitro culture medium has been evaluated in PCOS patients undergoing IVF-ET. The results demonstrate that the addition of 10 µM melatonin to in vitro maturation media enhanced embryo implantation and pregnancy rates with respect to non-supplemented control group [bib_ref] Does Supplementation of In-Vitro Culture Medium with Melatonin Improve IVF Outcome in..., Kim [/bib_ref]. In relation with the utilization of the indoleamine as oral pharmacological treatment in PCOS patients, the findings of some trials are encouraging. First, the administration of 2 mg melatonin/day for six months to PCOS patients enhanced menstrual irregularities in 95% of patients and waned levels of androgens and anti-Mullerian hormone, whose basal levels are above the normal range in PCOS patients. Concurrently, FSH levels, which are shrunk in this syndrome, were significantly enlarged compared with a control group [fig_ref] Table 1: Melatonin clinical trials on ovarian pathologies [/fig_ref] [bib_ref] Melatonin Treatment May Be Able to Restore Menstrual Cyclicity in Women with..., Tagliaferri [/bib_ref]. The effects of melatonin oral supplementation were also investigated in PCOS patients undergoing intrauterine insemination [bib_ref] Effects of Melatonin Administration on Chemical Pregnancy Rates of Polycystic Ovary Syndrome..., Mokhtari [/bib_ref] and in women with PCOS undergoing an IVF-ET cycle [bib_ref] Effect of Myo-Inositol and Melatonin versus Myo-Inositol, in a Randomized Controlled Trial,..., Pacchiarotti [/bib_ref]. In these trials, the synergistic effect of 3 mg melatonin/day and 4000 mg myoinositol/day remarkably elevated the number of mature oocytes and grade I embryos with respect to an untreated control group and myoinositol-treated group, although no significant differences were found in pregnancy rates [fig_ref] Table 1: Melatonin clinical trials on ovarian pathologies [/fig_ref].
On the other hand, the combination of 3 mg of melatonin with 250 mg of magnesium also proved to be effective in PCOS patients since it enhanced sleep quality and decreased serum testosterone and insulin levels, which are usually high in women with PCOS [fig_ref] Table 1: Melatonin clinical trials on ovarian pathologies [/fig_ref] [bib_ref] Metabolic and Hormonal Effects of Melatonin and/or Magnesium Supplementation in Women with..., Alizadeh [/bib_ref].
Given the safety profile of the indoleamine, and despite the necessity of larger clinical studies to confirm melatonin effectiveness, some fertility clinics have already included melatonin as a supplement in IVF protocols, especially, in the case of PCOS patients.
## Varicocele
Varicocele is a common clinical disease in andrology that has detrimental effects on semen quality, sperm function, and pregnancy outcomes in some men [bib_ref] Varicocele and Male Infertility, Jensen [/bib_ref]. Most importantly, oxidative stress seems to have a central role in the pathogenesis of varicocele-induced infertility [bib_ref] Varicocele and Male Infertility, Jensen [/bib_ref]. In this context, the role of melatonin in the prevention of testicular damage was first investigated in animal model of experimentally induced varicocele [bib_ref] The Effects of Melatonin and the Antioxidant Defence System on Apoptosis Regulator..., Onur [/bib_ref] [bib_ref] Effects of Melatonin on Testicular Tissue Nitric Oxide Level and Antioxidant Enzyme..., Semercioz [/bib_ref]. Such reports found that melatonin administration at a dose of 10 mg/kg/day antagonized the activation of germ cell apoptosis evoked by varicocele, which was attributed to the prevention of oxidative lipid damage. In the same line, more recent investigations have revealed that the antioxidant/anti-apoptotic efficacy of melatonin treatment of varicocele rats is mediated by microRNA-34a/SIRT1/forkhead transcription factors-class O (type1) (FOXO1) signal transduction pathway [bib_ref] Melatonin Epigenetic Potential on Testicular Functions and Fertility Profile in Varicocele Rat..., El Gheit [/bib_ref].
Though there are no studies on the therapeutical potential of melatonin on varicocele patients, low levels of melatonin in semen have been observed in infertile men with varicocele [bib_ref] Melatonin Hormone Profile in Infertile Males, Awad [/bib_ref]. Likewise, one study has recently investigated the impact of melatonin supplementation on semen quality and total antioxidant capacity after varicocelectomy of infertile male patients [bib_ref] Melatonin Therapy Adds Extra Benefit to Varicecelectomy in Terms of Sperm Parameters,..., Lu [/bib_ref] , thus concluding that melatonin therapy adds an extra benefit to varicocelectomy by improving sperm concentration, motility, and morphology as well as total antioxidant capacity.
## Ovarian cancer
A plethora of new drugs has been proposed as adjuvant therapeutic strategies for ovarian cancer management, including melatonin. This is due to its antiproliferative, antiangiogenic, pro-apoptotic and immunomodulatory properties [bib_ref] Melatonin as a Promising Agent to Treat Ovarian Cancer: Molecular Mechanisms, Chuffa [/bib_ref]. In fact, in vitro studies have demonstrated the anti-tumour effect of 800 µM melatonin for 72 h, as it is able to inhibit tumour growth through the downregulation of cyclin-dependent kinases 2 and 4 (CDK2 and CDK4) in ovarian cancer cell lines PA-1 and OVCAR-429 [bib_ref] Melatonin Suppresses the Growth of Ovarian Cancer Cell Lines (OVCAR-429 and PA-1)..., Shen [/bib_ref]. Similarly, a concentration of 3.4 mM melatonin has been shown to hinder proliferation and migration by 23% in cancer stem cells derived from ovarian cancer cell line SK-OV-3. Moreover, melatonin (3.4 mM) also decreased the expression of EMT-related proteins such as zinc finger Ebox-binding homeobox 1 and 2 (ZEB1 and ZEB2), snail, and vimentin, and increased the levels of E-cadherin, a negative regulator of EMT. On the other hand, melatonin may enhance the therapeutic effect of cisplatin in ovarian cancer as it has been proven that, in cisplatin-treated SK-OV-3 cells, the indolamine further promoted cytotoxicity and apoptosis through caspase-3 activation and through suppression of the extracellular signal-regulated kinase (ERK)/90-kDa ribosomal S6 kinase (p90RSK)/heat shock protein 27 (HSP27) cascade, thereby improving cisplatin-induced apoptosis [bib_ref] Melatonin Synergistically Enhances Cisplatin-Induced Apoptosis via the Dephosphorylation of ERK/P90 Ribosomal S6..., Kim [/bib_ref].
Regarding animal studies, the anti-tumour effect of melatonin in an ovarian cancerinduced rat model has been investigated. Melatonin, injected at a dose of 200 µg/100 g body weight/day for 60 days, reduced ovarian tumour mass by 20%, synchronized the oestrous cycle and forestalled the incidence of sarcomas, endometrioid carcinomas and cystic teratomas [bib_ref] Characterization of Chemically Induced Ovarian Carcinomas in an Ethanol-Preferring Rat Model: Influence..., Chuffa [/bib_ref]. Furthermore, the same dose of melatonin may modulate different molecular events associated with ovarian cancer in rat models. Thus, melatonin has been shown to be involved in apoptosis induction as it upregulates of pro-apoptotic proteins (i.e., p53, BAX and caspase-3), downregulates the anti-apoptotic protein Bcl-2 and strengthens DNA fragmentation in an in vivo model of ovarian cancer [bib_ref] Apoptosis Is Triggered by Melatonin in an in Vivo Model of Ovarian..., Chuffa [/bib_ref]. Likewise, melatonin treatment has also been effective in modulating the angiogenic signalling pathway by lowering the expression of several angiogenic factors such as transforming growth factor β-1 (TGFβ1), VGEF and vascular endothelial growth factor receptor 2 (VEGFR2) [bib_ref] Melatonin Reduces Angiogenesis in Serous Papillary Ovarian Carcinoma of Ethanol-Preferring Rats, Zonta [/bib_ref]. Additionally, melatonin has been shown to be able to suppress the increased expression levels of proteins involved in ovarian cancer signalling, including epidermal growth factor receptor 2 (Her-2), p38 mitogen-activated protein kinase (p38 MAPK), protein kinase B (AKT), mammalian target of rapamycin (mTOR), and toll-like receptor 4 (TLR4) [bib_ref] Melatonin Attenuates the TLR4-Mediated Inflammatory Response through MyD88-and TRIF-Dependent Signaling Pathways in..., Chuffa [/bib_ref].
As for human studies, a phase II study has investigated the effectiveness of a combination of melatonin with tamoxifen as therapy in patients with untreatable metastatic solid tumours, wherein two ovarian cancer patients were enrolled. Drugs were given orally at a daily dose of 20 mg of tamoxifen in the midday and 20 mg of melatonin at bedtime. The first ovarian cancer patient, who had lung metastasis, did not respond to the treatment, and survived for only four months. However, in the second ovarian cancer patient, who had lung and lymph nodes metastasis, the disease was stabilized, and a 10-month survival was achieved [fig_ref] Table 1: Melatonin clinical trials on ovarian pathologies [/fig_ref] [bib_ref] A Phase II Study of Tamoxifen plus Melatonin in Metastatic Solid Tumour..., Lissoni [/bib_ref]. This trial demonstrates the potential of melatonin as a therapeutic agent in ovarian cancer, although further clinical assays are necessary to draw definitive, unequivocal conclusions. BDNF: brain-derived neurotrophic factor; FSH: follicle stimulating hormone; hCG: human chorionic gonadotropin; IVF-ET: in vitro fertilization-embryo transfer; MEL: melatonin; PCOS: polycystic ovary syndrome.
# Conclusions
As impaired oxidative balance appears to explain fertility treatment failure, research on reproductive medicine has focused on developing efficient antioxidant therapies to bypass such an issue. In this scenario, melatonin has emerged in the last few years as a valuable tool. Thus, the addition of melatonin to culture media helps gametes to efficiently fight against oxidative stress in so far as the indoleamine reportedly improves oocyte and sperm quality during gametes preparation and/or in vitro culture. Furthermore, melatonin also acts as an effective cryoprotectant by counteracting cellular osmotic and/or oxidative stresses that occur during cryopreservation of spermatozoa and oocytes. More importantly, both animal research and clinical trials have provided evidence that oral supplementation with melatonin prevents oxidative stress-induced damage in vivo, which enhances gametes quality and may eventually lead to higher fertility success rates. This intervention is particularly relevant in the management of diverse ovarian pathologies since melatonin supplementation has been informed to be beneficial in animal models of, and patients with endometriosis, PCOS and/or ovarian cancer.
From a medical perspective, melatonin has increasingly gained attention in recent years due to scientific evidence supporting its potential therapeutical use in reproductive medicine and its safety profile. On one hand, most fertility clinics recommend the use of dietary supplements to aid fertility as they may exert positive effects on different fertility targets, such as hormonal balance, ovulation, gametes quality or embryo quality, and hopefully on the likelihood of achieving pregnancy. In this sense, several dietary supplements for male and female fertility that contain melatonin are currently being marketed in Europe and the USA (e.g., Gestagyn men, Seidivid, FertyBiotic, Ovosicare Fertility, or Theratonin). Additionally, some fertility clinics have already included melatonin as antioxidant therapy in IVF protocols, especially, in the case of PCOS patients. Nonetheless, further clinical studies are warranted to better understand the proper dose, timing, and application of melatonin to enhance fertility.
## Conflicts of interest:
The authors declare no conflict of interest.
[fig] Figure 1: Cascade reaction of melatonin interaction with free radicals and its main metabolites. R: free radical; RH: reduced agent. [/fig]
[fig] Figure 2: Sources of reactive oxygen species (ROS) during assisted reproductive techniques (ART). Created in BioRender.com. [/fig]
[fig] Figure 3: Applications and benefits of melatonin in assisted reproductive techniques (ART). Created in BioRender.com. [/fig]
[fig] Author: Contributions: Conceptualization, M.G. and J.E.; data curation, P.C.; writing-original draft preparation, P.C., M.G. and J.E.; writing-review and editing, P.C., A.B.R., M.G. and J.E.; visualization, P.C., A.B.R., M.G. and J.E.; supervision, A.B.R., M.G. and J.E.; project administration, M.G. and J.E.; funding acquisition, A.B.R., M.G. and J.E. All authors have read and agreed to the published version of the manuscript.Funding: This study was supported by Junta de Extremadura (GR21042). J. Espino and M. Garrido hold research post-doctoral fellowship from Junta de Extremadura (ref. TA18002 and TA18029, respectively). Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. Data Availability Statement: Not applicable. [/fig]
[table] Table 1: Melatonin clinical trials on ovarian pathologies. [/table]
|
Identifiers.org and MIRIAM Registry: community resources to provide persistent identification
The Minimum Information Required in the Annotation of Models Registry (http://www.ebi.ac .uk/miriam) provides unique, perennial and location-independent identifiers for data used in the biomedical domain. At its core is a shared catalogue of data collections, for each of which an individual namespace is created, and extensive metadata recorded. This namespace allows the generation of Uniform Resource Identifiers (URIs) to uniquely identify any record in a collection. Moreover, various services are provided to facilitate the creation and resolution of the identifiers. Since its launch in 2005, the system has evolved in terms of the structure of the identifiers provided, the software infrastructure, the number of data collections recorded, as well as the scope of the Registry itself. We describe here the new parallel identification scheme and the updated supporting software infrastructure. We also introduce the new Identifiers.org service (http://identifiers.org) that is built upon the information stored in the Registry and which provides directly resolvable identifiers, in the form of Uniform Resource Locators (URLs). The flexibility of the identification scheme and resolving system allows its use in many different fields, where unambiguous and perennial identification of data entities are necessary.
# Introduction
The size and complexity of data produced in biology has made it increasingly important to provide metadata alongside the core data itself. This metadata may comprise domain-specific information as described by minimal information 'checklists' meant to enable accurate data reuse or may be ontological in nature, specifying more precisely the kind of entities under consideration. Community-level collaborative bodies such as Minimum Information for Biological and Biomedical Investigations (MIBBI) [bib_ref] Promoting coherent minimum reporting guidelines for biological and biomedical investigations: the MIBBI..., Taylor [/bib_ref] and Open Biomedical Ontologies (OBO) (2) exist to formalize and coordinate such efforts across the Life Sciences. The computational systems biology community developed one such checklist, entitled the Minimum Information Required in the Annotation of Models (MIRIAM) [bib_ref] Minimum information requested in the annotation of biochemical models (MIRIAM), Nove`re [/bib_ref] , in order to define the meta-information needed to ensure the re-usability of computational models of biological processes. These guidelines describe the need to unambiguously and perennially identify model components, as well as other information regarding model origin and development. When used in conjunction with standard computerreadable formats such as Systems Biology Markup Language (SBML) (4), controlled annotations facilitate not only model reuse, but also permit efficient search strategies, accurate model comparison and meaningful model conversion between different formats. Furthermore, the relevant linking of models to biological knowledge transforms them into repositories of information.
In order to provide globally unique, perennial and location-independent identifiers for data used in the biomedical domain, we developed MIRIAM Identifiers and the MIRIAM Registry [bib_ref] MIRIAM Resources: tools to generate and resolve robust cross-references in Systems Biology, Laibe [/bib_ref]. MIRIAM Identifiers are Uniform Resource Identifiers (URIs), which unambiguously identify a record in a data collection, independently of the specific resources distributing instances of those records. The MIRIAM Registry provides information about the different data collections and how to access instances of their records. Definitions of the terms used in subsequent descriptions are detailed in the (definition).
## Miriam identifiers
To completely fulfil its roles, an identifier must be: (i) unique (two identifiers should not be associated with the same entity); (ii) unambiguous (an identifier must only be associated with a single entity); (iii) perennial (the same identifier should remain associated with an entity for the whole duration of its existence). In addition, an identifier should preferably also be (iv) standard compliant (for easier software support); (v) resolvable (convertible into *To whom correspondence should be addressed. Tel: +44 1223 494 403; Fax: +44 1223 494 468; Email: [email protected] a physical address on the World Wide Web); and (vi) free to use. MIRIAM identifiers were designed to satisfy all these criteria [bib_ref] MIRIAM Resources: tools to generate and resolve robust cross-references in Systems Biology, Laibe [/bib_ref].
In order to provide a unique identifier for a record, regardless of the physical location(s) where that information can be retrieved, MIRIAM identifiers are composed of three parts. The first part is a prefix, dependent on the scheme used (see below) and that specifies 'this is a MIRIAM URI'. The second part is the namespace that identifies the data collection. The third and final part is the internal identifier of a specific record in a data collection (this identifier is created and provided by the data collection itself). MIRIAM URIs were initially only provided as Uniform Resource names (URNs). The prefix of the URN scheme is urn:miriam. For example, the enzyme alcohol dehydrogenase in the enzyme classification collection is identified by urn:miriam:ec-code:1.1.1.1 and the species Homo sapiens in the taxonomy of living species by urn:miriam:taxonomy:9606.
In order to access the data for a record, one must rely on a URI resolving system, such as the MIRIAM Registry described below. The URNs can be resolved into URLs, for instance, using Web Services or by processing the XML export of the MIRIAM Registry. But they are not directly resolvable, so one cannot successfully copy/paste them into a browser and get an informative page. In order to provide directly dereferencable URIs and comply with the second rule of Linked Data (6), we recently introduced a new URL-based identification scheme. This scheme provides directly resolvable identifiers, based on the information stored in the MIRIAM Registry. The prefix of the URL scheme is http://identifiers.org/. This identification scheme runs entirely in parallel with the URN form of MIRIAM identifiers. Both forms essentially share the same structure, are based on the same shared list of namespaces and are fully inter-convertible. For example, the enzyme alcohol dehydrogenase in the enzyme classification collection is identified by http://identifiers.org/eccode/1.1.1.1 and the species Homo sapiens in the taxonomy of living species by http://identifiers.org/ taxonomy/9606.
The key difference between the two parallel schemes, which offer exactly the same access to data, is that the Identifiers.org URLs can be resolved directly and do not require special software tools on the user side for their handling; the dereferencing is performed by the Identifiers.org service (see below). The inter-relationships between the information captured by this service are illustrated in [fig_ref] Figure 1: Concepts and component information captured in the MIRIAM Registry [/fig_ref].
## Miriam registry
While the location of information on the Web is a convenient endpoint for cross-references, it is fraught with issues which can result in 'dead links'. These can be caused by changes in the underlying infrastructure of a resource or in modification of the access URL or identification scheme used by that resource. In addition, data of a given collection can often be accessed via several providers on the Web, for example, when it is 'mirrored' but also when associated with different metadata. In these cases, the record is identical (the data relevant to the collection describing the entity), though the physical instances of the record differ. MIRIAM Registry tackles this problem through the recording of resources, which are physical locations (associated with URLs) where one can access the entity or information about the entity. The concept of resource effectively allows the decoupling of the identification of an entity from its location on the Web, enabling the association of a single entity identifier with multiple locations. Resources and data collections are themselves identified as records in the MIRIAM Registry and have their own namespace. For instance, the enzyme nomenclature is identified by http://identifiers.org/miriam.collection/MIR:00000004 while the enzyme nomenclature distributed by the Swiss Institute of Bioinformatics is http://identifiers.org/miriam.resource/MIR:00100003. For a detailed list of all the information stored for each data collection, refer to Table 2 (information).
Initial population of the collections in the MIRIAM Registry came largely from the Systems Biology community, and more specifically from those collections that were commonly used in the annotation of models stored in BioModels Database [bib_ref] BioModels Database: An enhanced, curated and annotated resource for published quantitative kinetic..., Li [/bib_ref]. Further collections have been incorporated based on the requests from individual users, from collaborative bodies, such as the Protein Standards Initiative (http://www.psidev.info/) and from publicly available listings of Life Sciences databases, such as those used in cross-referencing (for example, http://www. geneontology.org/cgi-bin/xrefs.cgi). The public-facing web application provides easy access to the catalogue of data collections. Any visitor to the site can suggest modification . Definitions
## Data collection
A data collection gathers data of the same type (e.g. DNA, RNA or protein) and stores information regarding the same sets of 'properties' (e.g. sequence, references). It should make use of a well-defined internal identifier scheme. For example, the namespace 'uniprot' identifies a data collection whose subject is proteins, whose representation is protein sequence-centric, where each entry stores protein domain information and where the identifier scheme can be described using a specific regular expression. Similarly, 'ec-code' identifies a data collection that provides access to enzyme records and 'chebi' to an ontological representation of chemicals.
## Resource
A resource is the physical location on the Web where information about a data record can be accessed. A resource provides the instances of all the records belonging to a collection. Since the record identifier is independent of physical location (URL), it may be resolved using any of the resources listed for that data collection. Namespace
The namespace is the unique syntactic string which defines a data collection. For example, given the identifier 'urn:miriam:ec-code:1.1.1.1', the namespace is defined as 'ec-code'. This precise lexical string is used in both URN and URL forms of the identifiers.
of the information recorded (through a link at the bottom of each page) and submit new data collections (through a dedicated form linked from the left menu panel). More detailed information on making submissions and ascertaining the suitability of a collection for inclusion can be found here in the FAQ (http://www.ebi.ac. uk/miriam/main/mdb?section=faq). All records are manually curated in order to ensure accuracy and consistency. When the information provided by the submitter, usually obtained from the relevant resources on the web, is incomplete, we liaise with the developers and administrators of those collections. This is particularly important when issues arise, for example, regarding the identifier schemes used by the collection or the specific details of the license under which the information is made available.
The MIRIAM Registry infrastructure is written in Java (http://java.sun.com/javaee/) and makes use of the Model-View-Controller design pattern (http://en .wikipedia.org/wiki/Model-view-controller). The application runs inside an Apache Tomcat Web container (http://tomcat.apache.org/). All the information in the Registry is stored in a MySQL database (http://www .mysql.com/). . Information
## Data collection information
Identifier A stable MIRIAM Registry identifier of the data collection. Name
The name usually used to refer to the data collection. Synonym(s) Alternative name(s) of the data collection. Namespace
The part of the URIs which identifies the data collection. For example 'ec-code' for enzymes. Deprecated root URI(s) MIRIAM URNs or URLs that have become obsolete over time. Deprecated identifiers are stored in the Registry, allowing conversion to current forms. Definition
Short description of the data collection, indicating the focus of its content. Identifier pattern A regular expression pattern that describes the identifiers used within the data collection. Reference(s) Link(s) to documentation about the data collection and relevant publication(s).
## Resource information
Identifier Each resource associated with a collection is given a unique identifier in the MIRIAM Registry. Access URL URL used to retrieve a given data entry, where the token ($id) is replaced with a specified identifier for a record. Website
Root URL of the resource, usually its home page. Description
Brief description about the resource, used to distinguish the current resource from all the others recorded for the same data collection. Institution
The institution responsible for hosting the resource.
## Health status
Though not a textual field, the resource health status is displayed by the colour-coded text area. Deprecated Physical Location(s)
A list of deprecated resource(s) which are no longer usable to resolve information for this data collection. Each collection, which itself can be referenced via a URI, is assigned a namespace. This namespace can be combined with a suitable identifier in order to form a URI identifying the specific data record, independently of any physical locations holding that information. Each of these resolvable physical locations are regarded as an instance of the data record, and can themselves be identified using a URI.
Since its original introduction, there has been significant growth and development of the MIRIAM Registry. It currently contains information on over 250 collections, with a further 64 undergoing the curation process. The Registry also provides supporting facilities to enable the convenient usage of both URN and URL identification schemes.
## Finding collections and resources
With the plethora of available data collections potentially suitable to annotate biomedical information, it can be somewhat problematic to locate the most appropriate one. To aid in this task, each collection is associated with one or more tags. For instance, should a user require a data collection with which to annotate a protein sequence, it is possible to search using the two tags 'protein' and 'sequence' to find suitable data collections. This 'Tags' search function is linked from the left menu panel and displays a selectable list of the tags used. There are currently 39 tags available with which collections can be associated. These were created in ad hoc fashion when the system was implemented and currently are sufficient in number to allow the existing collections to be associated with two to three tags each. Additional tags can be created as needed by curators or as requested by users. The tags are of a coarse granularity, describing the type of data recorded ('sequence' 'expression' 'phenotype'), the subject of those data ('gene' 'protein' 'drug'), the domain area to which they relate ('disease' 'pharmacogenomics' 'neuroscience') or taxonomic associations of the data ('mammalian' 'human'). The purpose of the tagging system is to allow users to identify appropriate collections based on a gross level query. Plans to improve this tagging mechanism, for instance by incorporating ontological information at the level of resources and the data they provide, are discussed below.
The MIRIAM Registry provides access to several resources serving the same data collection. Those resources are not necessarily identical and there may be reasons to prefer one over another. To allow users to make an informed choice when selecting such a resource, we provide additional information, including the uptime of the servers running the service. For this purpose a 'health status' has been implemented and a daily health check is automatically performed for each resource listed in the Registry [fig_ref] Figure 2: An illustration of the variety of information captured for each data collection... [/fig_ref]. A summary of the health status of a specific resource is depicted by colour coding where green indicates an uptime in excess of 90% and graduated colour coding with downtime to below 20% being represented in red. More detailed information is available (by clicking on a resource's identifier), such as a calendar view of the uptime and details of the last check made. The system is also used by the Registry curators as a warning system to highlight otherwise unnoticed changes in the way data is accessed from a particular resource.
## Programmatic use of the registry
Simple Object Access Protocol (SOAP) and REpresentational State Transfer (REST) access methods are available to query the Registry. These Application Programming Interfaces (APIs) can be used to generate and resolve the identifiers, as well as to extract information about individual resources providing access to the data records.
To facilitate the usage of the web services by third party tools, a Java library is provided. It allows querying of the Registry in a quick and convenient way. It is available for download from the SourceForge.net project (http:// sourceforge.net/projects/miriam/).
The entire content of the Registry is also made available as an XML file export. This file is auto-generated daily and additionally can be created on demand.
## Identifiers.org resolving system
Identifiers.org is the resolving system of MIRIAM identifiers' URL form. For more information, readers may refer to http://identifiers.org/. Access to a given collection in the Registry, such as the enzyme nomenclature is given by appending the namespace, as in http://identifiers.org/eccode/. Due to the decoupling between the data collections and the resources that provide record information, the resolution of an Identifiers.org URL, such as http:// identifiers.org/ec-code/1.1.1.1, directs the user to an intermediate page listing all recorded physical locations where a record may be accessed, allowing the user to choose the most suitable one. This process is illustrated in .
One can directly access the instance of a record in a given resource by appending a parameter as a suffix. For instance, the following URL provides access to the record for alcohol dehydrogenase of the enzyme nomenclature collection provided by the IntEnz resource: http://identi fiers.org/ec-code/1.1.1.1?resource=MIR:00100001. Alter natively, the concept of 'profile' allows one to customize the behaviour of the resolving system: it allows the pre-selection of resources to be used in the dereferencing, for a whole range of data collections. For example, the pre-defined profile 'most_reliable' as in http://identi fiers.org/ec-code/1.1.1.1?profile=most_reliable, always returns the instance of a record in the resource with the best uptime. The 'most_reliable' profile is currently based on the health check history over the whole lifetime of the resource since its inclusion in the Registry. The valid parameters available for use are illustrated at: http://identi
[formula] fiers.org/examples/. [/formula]
The information about all the instances of a record is presented by default as HTML, but may also be retrieved in RDF/XML format. Either format can be recovered through content negotiation or using the 'format' parameter within the URL (for example: http://identifiers .org/ec-code/1.1.1.1?format=rdfxml). It is possible to accommodate further output formats as requested by the user community. The information represented in an RDF form allows the additional incorporation of semantic information. These semantics are captured using standard vocabularies such as SIO (Semanticscience Integrated Ontology; http://semanticscience.org/ontology/ sio.owl) and EDAM (EMBRACE Data and Methods; http://edamontology.sourceforge.net/), using terms such as 'has_identifier' and 'accession' to describe relationships and data concepts, respectively.
Parameter specification should be used only in conjunction with user interface instantiation (for example when used in a browser) and should be avoided when a URL is to be used as unique identifier for a collection record. There are a number of potential parameter name-value combinations possible for URIs, making a direct comparison difficult if provided with accompanying parametrization. Hence, in the unambiguous and perennial identification of data, the 'atomic' identifier should be considered as the minimal string that specifies the record.
The resolving system also provides direct information for malformed queries, for example, where the identifier is not properly encoded, or when a deprecated URL is used. In both cases a clear message is given to inform users of the situation. In addition to the human-readable description of the error, the system also returns the appropriate HTTP status code. , are highlighted: (1) Namespace; (2) Identifier pattern, which allows automated checking of identifier validity with respect to the expected expression pattern; and (3) Resource health status, which provides information on resource up-and down-time. A notification of the health status is given through colour coding, while more details are presented on a separate page, via a link on the resource identifier (see inset).
## Current status and future developments
MIRIAM URNs are already widely used, particularly within the Computational Systems Biology community. For example, in the 20th release (September 2011) of BioModels Database, the 764 models contain over 25 000 MIRIAM identifiers.
An overview of the widespread use of MIRIAM Registry information and its identifiers is detailed on the website (http://www.ebi.ac.uk/miriam/main/mdb? section=use). URI identifiers are supported by a variety of file formats; software libraries and tools have been developed to generate, resolve and leverage upon MIRIAM URIs in novel scientific research. Registry namespaces are being used as controlled vocabularies in databases and standardization efforts. Moreover, the Life Science Registry Name identification scheme, which will imminently no longer be developed and supported, has decided upon Identifiers.org as its replacement and successor.
The launch of the Identifiers.org URL as an alternative to the URN form of MIRIAM identifiers answers the ever-growing need to provide directly resolvable identifiers, especially for Semantic Web applications, such as the Linking Open Data initiative from the W3C. Collections and resources used in those efforts, such as those from the Life Science Dataset Registry which supports the Bio2RDF project [bib_ref] Bio2RDF: towards a mashup to build bioinformatics knowledge systems, Belleau [/bib_ref] , are currently being integrated in the Registry.
To enable the incorporation of a wider array of data collections, the Registry is being updated to store additional information, such as restrictions on the access to the data entries (for example the need to register and login), on its use (utilization of specific license and/or copyright), etc. In addition, resource and collection descriptions will be further enhanced by the incorporation of information from ontologies such as the Biomedical Resource Ontology (http://bioportal.bioontology.org/ ontologies/1104). These modifications will allow users to ascertain the appropriateness of use of particular data collections and will improve existing search facilities.
Profiles predefine specific resolving locations for each selected data collection and may be shared using the 'profile' parameter described above. Moreover, there is currently ongoing work to allow users to create profiles in the Registry through a dedicated user interface. This interface will also list the publicly available profiles that have been created by users, stating the collection and preferred resources associated.
The MIRIAM Registry is a collaborator of the BioDBCore effort [bib_ref] Towards BioDBcore: a community-defined information specification for biological databases, Gaudet [/bib_ref]. This effort focuses on a community approved minimum information checklist to which database providers should comply. It recommends that the standards that are implemented by a data provider (such as which standard formats it accepts and provides, which terminologies it uses, etc.) should be recorded with reference to those standards listed by BioSharing (http:// biosharing.org). This information, ideally, would be provided by database administrators as an RDF file. The BioDBCore database will rely on dedicated resources for the storage of some information. Identifiers.org URIs will be used for identification and data access information.
All the code used to develop the MIRIAM Registry and the associated helper utilities are released under the terms of the GNU, General Public License and are available at: http://sourceforge.net/projects/miriam/.
# Conclusions
The MIRIAM Registry is a stable resource, which provides both an identifier scheme and resolution system. While it originates from within the Computational Systems Biology community, it is certainly not limited to that domain, submissions are encouraged for new data collections from any biological community that has a desire to create and/or use unambiguous perennial identifiers or to generate and resolve physical locations from which data can be accessed. The system is of particular interest to tool and database developers who need to manage annotations and cross-references. Identifiers.org helps to ensure that data entities are resolvable, thereby avoiding the creation of 'dead ends' in the network of linked data.
[fig] Figure 1: Concepts and component information captured in the MIRIAM Registry. The MIRIAM Registry collects information about data collections and resources, allowing them to be referenced using URIs. Red-bounded boxes represent concepts, while green ones depict specific instances. [/fig]
[fig] Figure 2: An illustration of the variety of information captured for each data collection in the MIRIAM Registry. Some fields, described in the Information [/fig]
|
Brain-related proteins as serum biomarkers of acute, subconcussive blast overpressure exposure: A cohort study of military personnel
Repeated exposure to blast overpressure remains a major cause of adverse health for military personnel who, as a consequence, are at a higher risk for neurodegenerative disease and suicide. Acute, early tracking of blast related effects holds the promise of rapid health assessment prior to onset of chronic problems. Current techniques used to determine blastrelated effects rely upon reporting of symptomology similar to that of concussion and neurocognitive assessment relevant to operational decrement. Here, we describe the results of a cross sectional study with pared observations. The concentration of multiple TBI-related proteins was tested in serum collected within one hour of blast exposure as a quantitative and minimally invasive strategy to augment assessment of blast-exposure effects that are associated with concussion-like symptomology and reaction time decrements. We determined that median simple reaction time (SRT) was slowed in accordance with serum Nf-L, tau, Aβ-40, and Aβ-42 elevation after overpressure exposure. In contrast, median levels of serum GFAP decreased. Individual, inter-subject analysis revealed positive correlations between changes in Nf-L and GFAP, and in Aβ-40 compared to Aβ-42. The change in Nf-L was negatively associated with tau, Aβ-40, and Aβ-42. Participants reported experiencing headaches, dizziness and taking longer to think. Dizziness was associated with reaction time decrements, GFAP or NfL suppression, as well as Aβ peptide elevation. UCH-L1 elevation had a weak association with mTBI/concussion history. Multiplexed serum biomarker quantitation, coupled with reaction time assessment and symptomology determined before and after blast exposure, may serve as a platform for tracking adverse effects in the absence of a head wound or diagnosed concussion. We propose further evaluation of serum biomarkers, which are often associated with TBI, in the context of acute operational blast exposures.Overpressure (OP) is defined as the pressure caused by a shock wave that exceeds normal atmospheric pressure. OP exposure (Exp) may be caused by a variety of explosive devices or charges, as well as munitions. A subset of military personnel, "breachers", use a tactical technique to force entry into a closed area and within structures experience and are regularly exposure to repeated exposure. Low levels of exposure are linked to acute reduction in operational performance indicated, in part, by a decrement in reaction time (RT)[1]. Repetitive exposure has been linked to a complex array of symptoms including headaches, tinnitus, fatigue, and dizziness. This symptom complex has been termed "breacher's brain"[2][3][4]. Symptoms are similar to those observed among persons who have a clinically diagnosed mild traumatic brain injury (mTBI) or concussion, one of the most common injuries sustained by military personnel, particularly those who engage in training and combat roles (http://dvbic.dcoe.mil/tbimilitary). These effects are often transient, underreported, and challenging to identify due to symptom variability; which makes classification of an objective "injury response" difficult to achieve.Assessment of exposure mediated effects as they relate to performance, resilience, or mTBI, are often achieved through neurocognitive testing. The Defense Automated Neurobehavioral Assessment (DANA) is a field deployable neurocognitive and psychological assessment tool developed and extensively tested by the Department of Defense[5]. The DANA was commissioned to assist with detecting performance change over a variety of issues, such as concussion, occurring in combat deployment settings. The DANA has been tested in several operational environments and being used in research contexts[1,6]. More recently, the use of blood based biomarkers have been suggested to further augment stratification and, potentially, health status relevant directly to exposure.Central nervous system-enriched proteins, such as neurofilament light chain (Nf-L), tau, and glial fibrillary protein (GFAP) have been used as objective measurements to identify TBI[7,8]. Similarly, ubiquitin carboxy-terminal hydrolase L1 (UCH-L1) and amyloid precursor protein (APP), the precursor of amyloid beta (Aβ) peptides, have been identified in peripheral blood collected weeks-months after exposure[9]. These biomarkers have been robust identifiers within the context of chronic paradigms relevant to overpressure as well as TBI. Yet, despite the high incidence of exposure and the likelihood of symptomology similar to mTBI, an objective and quantifiable evidence of an "injury effect" per biomarker assessment during an acute, or near immediate time frame, remains elusive. Therefore, this preliminary study was conducted to determine the changes in reaction time, self-reported symptoms, and quantitation of TBI-associated serum biomarkers among military personnel exposed to overpressure exposure caused by blast. Early, sensitive quantitation of exposure-mediated peripheral biomarkers may be capable of identifying biological effects of overpressure and augment in-field care, even in the absence of a fully diagnosed concussion or visible traumatic brain injury.Materials and methodsStudy participantsActive duty United States Army personnel (n = 29) within a single site, Fort Leonard Wood, MO engaged in a two-week breacher training course. Heavy wall breaching exercises occurred within one training day during which neurocognitive testing, blood sampling, and symptomology assessments were conducted. Serum biomarkers of acute, low-level blast overpressure PLOS ONE | https://doi.org/10.
a1111111111 a1111111111 a1111111111 a1111111111 a1111111111
# Introduction
## Overpressure measurements
All participants were exposed to two overpressure events (back-to-back heavy wall breaches to defeat concrete walls) in a single training session. Exposure levels were measured as psi (pound per square inch) using the B3-H pressure sensor mounted on the left shoulder of each participant to approximate incident pressure. The B3-H is a small, lightweight, accurate, disposable, and off the shelf device that records and collects data peak pressure, acceleration (rate at which speed changes) and impulse (time exposed to certain levels of overpressure exposure). Peak pressure (psi) and impulse (psi X milliseconds are displayed for each incident and as cumulative values for the training session. Participants are in a static position during exposure. Therefore, acceleration does not occur and does not meet the threshold to be automatically recorded by the B3-H sensor.
## Assessment of neurocognitive performance
The Defense Automatized Neurocognitive Assessment (DANA) tool was administered prior to (pre: -8h) and after (post: +1h) Exp in accordance with blood-draw time. The DANA consists of three subtasks conducted with a hand-held device and monitor screen. (1) Simple reaction time (SRT) measures pure reaction time. The participant was required to tap on the location of the yellow asterisk symbol as quickly as possible each time it appeared; (2) Procedural reaction time (PRT) is a choice reaction time that measures accuracy, reaction time, and impulsivity. The screen displays one of four numbers for 3 seconds (sec). The participant was required to press a left button ("2" or "3") or right button ("4" or "5"). This choice reaction time task targets simple executive functioning and working memory; and (3) Go-No-Go (GNG) is a forced choice reaction-time task. A picture of a house is presented on the screen. Either a "friend" (green) or "foe" (white) appeared in a window. The respondent must push a "fire" button only when a "foe" appears. The choice reaction time measures sustained attention and impulsivity. The test quantifies speed and accuracy of target omissions and commissions.
## Symptom and tbi/concussion history reporting
Participants completed a 32-item, paper-and-pencil health symptom inventory before (pre-) and after (post-) Exp, in conjunction with each blood draw. The symptoms on the inventory are similar to that of the Rivermead instrument, but with additional items and responses relevant to the breaching exercise context rather than exclusively to concussion. Participants were instructed to use a 5-point Likert scale (0 "not experienced at all," 1 "no more of a problem than before training," 2 "mild problem-present but don't really notice and doesn't concern me," 3 "moderate problem-I can continue what I am doing but I notice the problem," 4 "severe problem-constantly present, feels like it could affect my performance"). Participants noted prior history of concussion or mTBI which is reported as a binary metric (No = 0, Yes = 1). Clinical data was not available.
## Serum preparation and quantitative biomarker measurements
Venous blood was collected directly into BD Vacutainer SST Serum Separation Tubes (Fisher Scientific, Waltham, MA) and processed within 30 minutes according to the manufacturer's instructions. Samples were centrifuged at 1,000 x g for 10 minutes, at room temperature. Samples were stored in 1mL aliquots, supplemented with HALT protease/phosphatase inhibitors, and then stored at -80 C until use. GFAP, UCH-L1, Nf-L, tau, Aβ-40, and Aβ-42 were measured using digital immunoassays performed using the Simoa HD-1 according to manufacturer's instructions (Quanterix Corporation, Lexington, MA). All assays were performed based on manufacturer's recommendations. Briefly, serum was thawed on ice then centrifuged at 10,200 x g for 10 minutes at 4˚C. Thereafter, 120μL of serum supernatant was directly loaded onto a 96 well plate and diluted 1/4 during the assay. Curve fitting analysis was conducted using pre-set programs designed by the manufacturer.
## Data management and statistical analysis
The full dataset containing age (years), sampling time (hours), peak pressure (psi and kPa), impulse (psi X ms, DANA values (ms), biomarker concentrations [pg/mL], and dichotomized symptomology or mTBI/concussion history for each participant is shown (S1 . Nondichotomized, Likert scale symptomology reporting is provided (S2 . All data was analyzed using Prism version 7 (GraphPad, La Jolla, CA). Biomarker concentrations [pg/mL] were compared using the Wilcoxon signed rank test, � p � 0.05. Data is shown as the median concentration [pg/mL] +/-IQR. Symptomology was transformed into binary variables (No change or a decrease = 0 or "-"; an increase = 1, or "+"). Biomarker and DANA values were transformed into a delta (d = post-Exp minus pre-Exp) and outliers were removed (ROUT = 1%) prior to correlative analysis using 1-tailed, Spearman rank correlation coefficient, � p � 0.05, or to comparisons against dichotomized (-, no change or decrease vs. +, increased) symptom data. Distribution of delta DANA or delta biomarker values were tested for normality using the D'Agostino & Pearson test prior to comparison using 1-tailed Mann-Whitney U-Test or Welch's t-Test as appropriate, � p � 0.05, and are displayed as the median +/-5-95%-ile range. Delta DANA or biomarker values compared to self-reported symptomology or to mTBI/concussion history are displayed (S3 .
# Results
Participants (n = 29) within this study were males aged 21-43 (mean: 29.5 yrs.), with variable duration of service (mean +/ SD: 8.5+/-4.6 yrs.; range: 2-20 yrs.). Acute blood sampling occurred within one hour following Exp (mean: 0.99h, range: 0.52-1.68h). The levels (mean +/-SD) of peak pressure (4.35+/-0.49 psi or 30.0+/-3.37 kPa / incident); cumulative peak pressure (8.71+/-psi or 60.0+/-6.83 kPa / session); impulse (11.7+/-1.13 psi X ms / incident); and cumulative impulse (23.5+/-2.26 psi X ms/session) derived from B3-H sensors mounted on each participant's left shoulder are indicated. Participants experienced symptomology similar to concussion . Headaches (15/29, 52%) and taking longer to think (12/29, 41%) were the most frequently reported, followed by dizziness (9/29, 31%), slowed thinking (8/29, 29%), and poor concentration .
DANA administration and serum biomarker testing revealed several changes relevant to OPChanges in biomarker levels derived from the delta (d = post-pre) were evaluated among of individual study participants using 1-tailed Spearman rank correlations after outlier removal . Delta Nf-L had a positive relationship with dGFAP (r = +0.40, p = 0.020). Interestingly, suppression of dNf-L was associated with higher dAβ-40 (r = -0.63, p < 0.001), dAβ-42 (r = -0.76, p < 0.001), and dTau (r = -0.37, p = 0.028); thus, each comparison has a negative correlation. As expected, dAβ-40 had a high degree of concordance with dAβ-42 (r = +0.93, p < 0.001). Overall, evaluation of inter-subject changes show that suppressed GFAP and NfL . Demographic characteristics of study participants and biosample collection timelines.
## Number of subjects (n) 29
## Age (years)
Mean (SD)Taking longer to think 12 Feelings of dizziness 9 20 31
Slowed thinking 8 Poor concentration 8 Ringing in ears 5 Feeling anxious or tense 5 Blurred vision 5 Easily upset by loud noise 4 Being irritable or easily angered 4 The table indicating the number of participants and self-reported symptoms is shown. The top ten of 32 symptoms are shown with the number of participants reporting an increase, a decrease or no change. The percent (%) of participants who reported an increase is indicated.
https://doi.org/10.1371/journal.pone.0221036.t002 are generally associated with increased levels of tau, Aβ-40, and Aβ -42 as a consequence of Exp. The remaining comparisons were not significant. Next, relationships between the top three symptoms (headaches, dizziness, and taking longer to think), the changes in SRT, and that of biomarkers that showed effects after Exp were determined. Correlation analysis indicated that there was no relationship between dSRT, dPRT, or dGNG compared to the changes in biomarker levels.
The top three symptoms (headaches, dizziness, and/ or taking longer to think) as well as mTBI/concussion history based on self-reporting, were dichotomized into two groups exemplifying participants who reported either a decrease or no change (-) compared to those who reported an increase (+) in symptomology or prior mTBI/concussion (S3 . This dichotomized data was compared to changes in DANA metrics and biomarker levels after outlier removal.
As expected, DANA metrics were generally associated with symptomology. Delta SRT was slower (exemplified by an increased value) in participants who reported dizziness 1 to 90.9, n = 12, p = 0.05) were also significant. Next, delta biomarker values were compare to symptomology and indicated that changes in biomarker levels were not linked to headaches or taking longer to think experience after Exp. In contrast, the decreased value of dGFAP was associated with dizziness (decreased or no change-median: -1.57, range: -37.7 to 33.6, n = 20; increased-median: -8.25, range: -22.0 to -4.00, n = 8, p = 0.02), whereas dTau was not related to any symptoms. Higher dUCH-L1 (decreased or no change-median: 0, range: -1.90 to 1.20, n = 17; increased-median: 0, range: -1.16 to 4.27, n = 9, p = 0.05). However, this result may be skewed by two study participants who had high dUCH-L1 levels even after outlier removal (S2 . Suppressed dNfL (decreased or no change-median: 0.30, range: -0.31 to 1.55, n = 19; increased-median: 0.24, range: -0.83 to 0.22, n = 8, p = 0.05)and elevated dAβ-40 (decreased or no change-median: -13.5, range: -91.1 to 88.2, n = 20; increased-median: 48.2, range: -44.4 to 90.3, n = 9, p < 0.001)values were also associated with post-Exp dizziness. Delta Aβ-42 held a similar trend (decreased or no changemedian: -0.09, range: -4.04 to 5.08, n = 20; increased-median: 2.73, range: -0.72 to 8.81, n = 9, p = 0.06), but was not significant. Changes in DANA metrics were, as expected, aligned with the top three symptoms reported. Interestingly, biomarker levels, particularly GFAP or NfL suppression and AB elevation, occurred in participants who reported post-Exp dizziness. DANA and biomarker changes were compared to dichotomized mTBI/concussion history reporting. Although the range among participants who reported former mTBI/concussion was greater, median values of dUCH-L1 were equivalent (No-median: 0.00, range: -2.54 to 1.20, n = 19, Yes-median: 0.00, range: -0.18 to 4.27, n = 9, p = 0.05) (S3 . As with comparisons to symptomology, these results are skewed by a few participants. No other trends were observed.
# Discussion
Quantitation of peripheral biomarkers offers a means to objectively monitor the effects of overpressure Exp within groups and among individuals involved in military training operations. Measurement of acute biomarkers remains sparse, particularly if the individual does not have outwardly obvious, clinically defined mTBI or concussion marked by well-known symptoms, such as loss of consciousness or changes in gait. Therefore, this study compared serum biomarker levels, neurocognitive deficits, and reported symptoms caused before and within one hour after mild-moderate Exp among military breachers within a single training session. The main findings show that median elevation of Nf-L, tau, Aβ-40 or -42, but a suppression of GFAP was evident in serum collected one hour post-Exp compared to pre-Exp sampling. Changes in DANA metrics were aligned with the top three symptoms reported and serum GFAP, NfL, and Aβ peptide changes were largely associated post-Exp dizziness.
## A subset of tbi-related proteins are potential biomarkers of acute overpressure exposure based on evaluation of group effects
The utility of blood-based biomarkers are increasingly investigated for concussion or subconcussion. Breacher's brain symptomology caused by overpressure exposure is similar to that of mTBI or concussion. Therefore, we hypothesized that biomarkers would also have utility for symptomatic overpressure exposure.
GFAP and UCH-L1 are, perhaps, the most thoroughly studied as biomarkers for moderatesevere TBI or closed head hemorrhagic injury. Levels dramatically increase in the serum or plasma within 12-24h. GFAP and UCH-L1 are not typically evident in patients with an mTBI unless hemorrhage or intracranial lesions are presented. However, composite assessment of GFAP and UCH-L1 (in addition to spectrin break down product (SBDP-150) revealed that these proteins were elevated in blood collected from study participants who had the most striking decrements in neurocognitive performance, including simple reaction time, as well as symptom reportingin the absence of a clinically defined concussion. UCH-L1 was also increased in serum collected two weeks after training from a subset of participants, but these results were not linked to neurocognitive performance or symptomology. Participants in the present study do not suffer from complicated mTBI within this context. The lack of a robust post-exposure UCH-L1 response among the cohort is not surprising. The moderate drop in median GFAP levels is not known in the context of physiology. However, this observation presents a novel observation that may be deserving of further investigation, specifically in the context of overpressure exposure and its systemic outcomes.
Nf-L and tau are two of the most abundant cytoskeletal proteins in both the peripheral (PNS) and central nervous system (CNS). Both have recently become more prevalent as potential biomarkers of brain trauma, neurological disease, and repeated concussion. In the context of low level exposure without a direct impact to the head, this study indicated that median levels of serum Nf-L and tau were elevated (although tau did not meet statistical thresholds). Exposure is reported to impact brain tissues in a way that may mirror a sub-concussive event, causing cytoskeletal abnormalities and demyelination in rodent models. In animal models, serum tau is elevated within 6h-1d of and the heavy chain isoform of neurofilament is increased within 2h after mild exposure. Serum Nf-L is reported to increase acutely, after sub-concussive head impacts when viewed in the context of TBI status. Both proteins are elevated in blood 1-6h hour after play among athletes who have prolonged return to play status.
This work indicated that median Aβ-42 was elevated one hour post-exposure. Aβ peptides are toxic monomers shown to be crucial to pathogenesis of chronic neurodegenerative diseases, such as Alzheimer's disease (AD)or chronic traumatic encephalopathy. Aβ is increased in the brains derived from veterans with a history of between chronic exposures caused by blast, which may offer associative or causative relationship to neurodegenerative diseases. Assessment of Aβ levels in blood is primary viewed in the context of cognitive decline or advanced age wherein Aβ levels in the blood typically decrease in accordance with increased plaque burden in the brain. The effect of Aβ in blood collected from cognitively normal, yet acutely injured and symptomatic, subjects is not fully understood. Recently, increased levels Aβ peptides have been detected in the blood of active duty and veteran populations. Plasma Aβ-40 is elevated among service members who sustained a clinically diagnosed TBI or experienced chronic symptoms associated with PTSD. Interestingly, increased serum Aβ is also associated with hypoxia or hypoxemia, which is proposed to occur as a consequence of altered cerebral blood flow after blast overpressure exposure.
When study participants are viewed collectively, assessment of median biomarker values indicate that serum GFAP and NfL decrease while Aβ-42 increases after overpressure exposure. These biomarker changes occur within the same time frame as SRT decrement, although there is no overall correlation to the DANA metrics within the cohort of participants. Dichotomizing biomarker changes according to breacher's brain symptomology may offer additional insight regarding individual and group post-exposure responses.
## Acute shifts in biomarker levels and neurocognitive decrements are associated with post-exposure symptomology, not mtbi or concussion history
Breacher's brain symptomology, such as post-exposure headaches, dizziness and taking longer to think, are well established, yet typically studied in relation to mTBI/concussion diagnosis and chronic blast exposure. The current study indicated that low levels of overpressure exposure consistently have this effect, even at early time frames. Neurocognitive decrements have been proven to be useful during acute timeframes. Therefore, association of SRT, PRT, and GNG decrements with the top three reported symptoms among participants is fitting for blast exposure.
Interestingly, decreased GFAP or NfL, elevated UCH-L1, and, to a greater extent, increased Aβ peptide levels were detected in serum of participants who reported increased dizziness after exposure. Acute GFAP and NfL suppression among symptomatic participants was surprising, and may appear to be contrary to the effects shown for concussed athletes, wherein biomarkers generally increase among cohorts with a clinically defined concussion. However, it is notable the pre-game levels were not determined and that the temporal dynamics indicate a decrease 1-12 hours after play. Suppressed GFAP in relation to symptomology remains a novel observation.
The change in serum UCH-L1 was positive (e.g. increased) among the subset of participants who reported post-exposure dizziness. Previously, UCH-L1 levels were shown to be unchanged in serum derived from symptomatic concussion patients who were negative for CT abnormalities. However, UCH-L1 was elevated in serum one hour after game-play among subconcussive football players, although the relationship to specific symptoms were not determined.
Direct relevance of acute Aβ-peptide elevation is not well known in the context of symptomology caused by overpressure exposure or subconcussive paradigms. Rather, fluctuation in blood Aβ are largely understood in the context of subacute-chronic symptomatic mTBI, after activities such as boxing, without stratification of specific symptoms. However, mass spectrometry-based proteomics of serum revealed peptides (proteins) that were specifically associated with the decree of PTSD or post-concussive syndrome symptomology among veterans who suffered a mTBI. It is possible that acute Aβ elevation among symptomatic participants in this study are aligned with these observations. Overall, this study is the first to show dysregulation of blood biomarkers, specifically Aβ, are aligned with symptomology that is common among participants exposed to blast overpressure.
This study is not without a few potential caveats. First, symptomology and mTBI/concussion history may be under-reported. Changes in UCH-L1 levels were higher in serum of participants who reported prior mTBI/concussion based on self-report. However, definitive medical evidence is not available. The change in UCH-L1 after low levels of overpressure exposure is considered a small effect that will remain under consideration for future studies. Second, amyloid precursor protein expression and release of Aβ peptides may occur outside of the CNS, including the epidermis and muscle and leak into the blood stream. However, there were no reports of tissue injury among participants. Lastly, circadian variation of Aβ levels among healthy controls (< 5%)has been reported, but the changes induced by exposure within this study eclipsed those found associated with circadian patterns.
To our knowledge, this work is the first to explicitly report changes in biomarker levels in the context of acute overpressure exposure compared to pre-exposure values. The strict definition of a clinical mTBI/concussion was not met within this paradigm. However, acute measurement of proteins in serum, particularly Aβ peptides, coupled with symptomology and neurocognitive assessment may provide a novel biomarker relevant to subconcussive effects of blast overpressure exposure. Acute serum biomarker dynamics among overpressure-exposed persons, who have subconcussive breacher's brain symptoms, may be worth further collective evaluation particularly in the absence of a clinically defined mTBI/concussion.
# Conclusions
Mild-moderate blast exposure is associated with acute elevation of serum Aβ peptides, a slight increase in tau, but a reduction in GFAP and NfL. Reaction time decrements and these biomarker profiles were collectively associated with post-exposure dizziness. Acute evaluation of serum protein levels, well-known neurocognitive tests, and symptoms before and after exposure have the potential to serve as a multiplexed surrogate biomarkers of exposure in absence of a direct impact to the head. These metrics may be adaptable to field-ready tools and aid return to duty decisions independent of TBI status.
Supporting information S1 Field metrics, biomarker levels, and self-reports of study participants before and after overpressure exposure. The age (years), peak pressure (psi, kPa), impulse (psi-ms), DANA metrics (milliseconds), quantitative biomarker values [pg/mL], changes in symptomology (0 = Decrease or No Change), and mTBI/concussion history (0 = No, 1 = Yes) are shown for each study participant prior to outlier removal or statistical analysis. Missing data for responses is indicated as no data (ND). (XLSX) S2 Table. Self-reported symptomology of participants. Self-reported symptoms from the 32-item survey for each study participant are shown. (A) Pre-exposure and (B) Post-exposure data is displayed in the Likert scale format from none (0) to severe (4). (XLSX) S3 Relationship between changes in DANA or biomarker levels with symptomology or mTBI/concussion history based on self-reporting. Dichotomized symptoms (headaches, dizziness, and taking longer to think) or mTBI/concussion reporting is shown in relation to changes (post-pre delta) in DANA metrics or biomarker levels. |
Posterior Reversible Encephalopathy Syndrome after a Variety of Combined Chemotherapies Containing Bevacizumab for Metastatic Colon Cancer
A 44-year-old woman with advanced metastatic colon cancer received chemotherapies comprising oxaliplatin and capecitabine (XELOX), irinotecan hydrochloride, leucovorin calcium and fluorouracil irinotecan (FOLFIRI)/panitumumab and mFOLFOX6/bevacizumab. Fifteen months later, she presented with the acute onset of a headache, drowsiness and seizure with a fever and hypertension. Brain magnetic resonance imaging (MRI) indicated bilateral regions of signal hyperintensity in the white matter with spasms of bilateral cerebral arteries apparent on magnetic resonance angiography. Posterior reversible encephalopathy syndrome (PRES) was diagnosed, and treatments resulted in improvement of the MRI findings, but the patient experienced cerebral infarction and ultimately died of deterioration of cancer on day 26 after the onset of PRES.
# Introduction
Posterior reversible encephalopathy syndrome (PRES) is a neuro-radiological syndrome characterized by cortical blindness, an altered level of consciousness and seizures. PRES is associated with hyper-intense lesions on magnetic resonance imaging (MRI), typically in the posterior region [bib_ref] Posterior reversible encephalopathy syndrome: clinical and radiological manifestations, pathophysiolology, and outstanding questions, Fugate [/bib_ref]. In most cases, the symptoms and radiological lesions of PRES are reversible. However, the term PRES is not precisely suitable, as the syndrome is not always reversible (2) and is often not confined to either the white matter or posterior regions of the brain [bib_ref] Posterior reversible encephalopathy syndrome with extensive deep white matter lesions including the..., Ohira [/bib_ref].
The mechanisms underlying PRES have been postulated to be severe hypertension leading to failed auto-regulation and endothelial injury/vasogenic edema, or vasoconstriction leading to brain ischemic and subsequent vasogenic edema [bib_ref] Permeability change and brain tissue damage after intracarotid administration of cisplatin studied..., Sugimoto [/bib_ref]. The risk factors for this syndrome include malignant hypertension, eclampsia, renal failure and treatment with anti-neoplastic agents.
Certain combination regimens have been associated with PRES, including a XELOX regimen comprising oxaliplatin, capecitabine and folinic acid; a panitumumab plus FOLFIRI regimen comprising irinotecan hydrochloride, leucovorin calcium and fluorouracil; and a bevacizumab plus mFOLFOX6 regimen comprising oxaliplatin, 5-fluorouracil and Lleucovorin [bib_ref] Posterior reversible encephalopathy syndrome following chemotherapy with oxaliplatin and a fluoropyrimidine: a..., Giuseppe [/bib_ref] [bib_ref] Reversible posterior leukoencephalopathy syndrome associated with oxaliplatin, Pinedo [/bib_ref] [bib_ref] Delayed posterior encephalopathy syndrome following chemotherapy with oxaliplatin and gemcitabine, Moris [/bib_ref] [bib_ref] Posterior reversible encephalopathy syndrome with bevacizumab, Lau [/bib_ref] [bib_ref] Reversible posterior leukoencephalopathy syndrome and bevacizumab, Glusker [/bib_ref] [bib_ref] Reversible posterior leukoencephalopathy syndrome and bevacizumab, Ozcan [/bib_ref] [bib_ref] Reversible posterior leukoencephalopathy syndrome after bevacizumab/ FOLFIRI regimen for metastatic colon cancer, Allen [/bib_ref] [bib_ref] Late-onset leukoencephalopathy induced by long-term chemotherapy with capecitabine and cyclophosphamide for liver..., Yasaki [/bib_ref] [bib_ref] Bevacizumab plus irinotecan, fluorouracil, and leucovorin for the metastatic colorectal cancer, Hurwitz [/bib_ref] [bib_ref] Reversible posterior leukoencephalopathy syndrome after bevacizumab/FOLFIRI regimen for metastatic colon cancer, Allen [/bib_ref].
We herein report the case of a patient with advanced metastatic colon cancer who received chemotherapies of a XELOX regimen, a FOLFIRI regimen plus panitumumab and a mFOLFOX 6 regimen plus bevacizumab. The diagnosis of PRES was supported by the findings on MRI and MR angiography (MRA), and the patient died of deterioration of cancer. A link between PRES and bevacizumab, which was ultimately administered with mFOLFOX6 after the XELOX and panitumumab/FOLFIRI regimens, was therefore suggested.
## Case report
A 44-year-old woman underwent rectosigmoid colectomy and left mesorectal resection for rectosigmoid carcinoma in April 2015 with stage b NSCLS (tumor-node-metastasis staging score, T3N2M0). A histopathological examination revealed poorly differentiated adenocarcinoma that had infiltrated the subserosa and regional metastases to the left supraclavicular lymph nodes. The patient was started on a XELOX regimen comprising oxaliplatin, capecitabine and folinic acid in May 2015. Thereafter, additional metastases to the para-aortic lymph node at the level of the left renal artery were detected on abdominal computed tomography. After radiation therapy to the lesion, a 7-cycle course of panitumumab plus FOLFIRI regimen consisting of irinotecan hydrochloride, leucovorin calcium and fluorouracil was administered from January 2016, followed by an 8-cycle course of bevacizumab plus the mFOLFOX6 regimen, comprising oxaliplatin, 5-fluorouracil and L-leucovorin, from June 2016.
Treatment was uneventful until September 2016, when she presented with the acute onset of a headache, drowsi-ness and seizure. Vital signs indicated a fever and hypertension (188/112 mmHg) with no history of hypertension. A neurological examination indicated a limited attention span, disorientation, generalized hyperreflexia, bilateral Babinski sign and no focal neurological signs in the limbs. A laboratory investigation revealed an elevated white blood cell count (15,540/μL; normal range, 3,800-8,500/μL), mildly elevated C-reactive protein (1.5 mg/dL; normal range, 0.0-0.3 mg/dL), a normal ammonia level, a normal renal function, mild hyponatremia (128 mmol/L; normal range, 138-146 mmol/L) and a mildly elevated glucose level (122 mg/ dL; normal range, 70-109 mg/dL). Brain MRI was performed on the day of the onset of neurological symptoms. Fluid-attenuated inversion recovery (FLAIR) imaging revealed bilateral regions of signal hyperintensity in the occipital, parietal and periventricular white matter [fig_ref] Figure 1: Axial-section fluid-attenuation inversion recovery [/fig_ref] , but T1-weighted imaging indicated regions of signal isointensity in the same areas. MRA indicated spasms of the bilateral anterior, middle and posterior cerebral arteries [fig_ref] Figure 2: Magnetic resonance angiography [/fig_ref]. The patient was ultimately diagnosed with PRES associated with reversible cerebral vasoconstriction syndrome (RCVS).
Her blood pressure was controlled within a normal range through the administration of nimodipine given intravenously to control her hypertension and the spasms of cerebral arteries, followed by the administration of anticonvulsants, anti-platelet drugs and anti-coagulants. To reduce the cerebral edema, edaravone, glycerol and steroid were administrated. She recovered gradually from the symptoms and regained full consciousness, with a partial improvement of her findings on brain MRI 11 days later [fig_ref] Figure 1: Axial-section fluid-attenuation inversion recovery [/fig_ref]. In addition, she showed progression of spasms in the bilateral anterior, middle and posterior cerebral arteries [fig_ref] Figure 2: Magnetic resonance angiography [/fig_ref]. Treatment with the above drugs continued, but she developed cerebral infarction presenting as moderate right hemiparesis. Brain MRI revealed new lesions in the left fronto-parietal area [fig_ref] Figure 1: Axial-section fluid-attenuation inversion recovery [/fig_ref] , and a tendency toward the improvement of spasms in the bilateral anterior, middle and posterior cere-bral arteries [fig_ref] Figure 2: Magnetic resonance angiography [/fig_ref]. The patient ultimately died of deterioration of cancer on day 26 after the onset of PRES.
# Discussion
PRES is a rare neurological syndrome with presenting symptoms ranging from headache, altered mental status, seizures and visual loss to loss of consciousness. The term describes a potentially reversible imaging appearance and symptomatology with a variety of causes (1). The term PRES is not necessarily accurate, as brain edema is often not isolated to the posterior region (3), and the syndrome is not uniformly reversible. As discussed previously, cerebral hemorrhaging and infarction are the most common reasons for incomplete recovery, and PRES can prove fatal in severe cases [bib_ref] Fatal outcome of posterior "reversible encephalopathy syndrome in metastatic colorectal carcinoma after..., Plaveti [/bib_ref].
PRES is primarily associated with hypertension, eclampsia, renal impairment, cytotoxic drugs, immunosuppressants and molecular-targeted agents. Chemotherapy-induced complications show a wide spectrum of clinical manifestations, including PRES. Certain combination regimens have also been associated with PRES, including the combination of capecitabine and oxaliplatin (XELOX), the combination of irinotecan hydrochloride, leucovorin calcium and fluorouracil (FOLFIRI)/panitumumab, and oxaliplatin, 5fluorouracil and L-leucovorin (mFOLFOX6)/ bevacizumab [bib_ref] Posterior reversible encephalopathy syndrome following chemotherapy with oxaliplatin and a fluoropyrimidine: a..., Giuseppe [/bib_ref] [bib_ref] Reversible posterior leukoencephalopathy syndrome associated with oxaliplatin, Pinedo [/bib_ref] [bib_ref] Delayed posterior encephalopathy syndrome following chemotherapy with oxaliplatin and gemcitabine, Moris [/bib_ref] [bib_ref] Posterior reversible encephalopathy syndrome with bevacizumab, Lau [/bib_ref] [bib_ref] Reversible posterior leukoencephalopathy syndrome and bevacizumab, Glusker [/bib_ref] [bib_ref] Reversible posterior leukoencephalopathy syndrome and bevacizumab, Ozcan [/bib_ref] [bib_ref] Reversible posterior leukoencephalopathy syndrome after bevacizumab/ FOLFIRI regimen for metastatic colon cancer, Allen [/bib_ref] [bib_ref] Late-onset leukoencephalopathy induced by long-term chemotherapy with capecitabine and cyclophosphamide for liver..., Yasaki [/bib_ref] [bib_ref] Bevacizumab plus irinotecan, fluorouracil, and leucovorin for the metastatic colorectal cancer, Hurwitz [/bib_ref] [bib_ref] Reversible posterior leukoencephalopathy syndrome after bevacizumab/FOLFIRI regimen for metastatic colon cancer, Allen [/bib_ref]. The development of PRES associated with RCVS in our patient may therefore have been secondary to bevacizumab, given her clinical course.
Bevacizumab, a monoclonal antibody against vascular endothelial growth factor (VEGF), was recently approved as a first-line treatment for patients with metastatic colorectal cancer when administered in combination with FOLFILI [bib_ref] Bevacizumab plus irinotecan, fluorouracil, and leucovorin for the metastatic colorectal cancer, Hurwitz [/bib_ref]. Bevacizumab may have contributed to a hypertensive state in an otherwise non-hypertensive patient, as hypertension is a known adverse effect of this medication. Global endothelial cell dysfunction induced by bevacizumab may then have led to PRES.
The mechanisms underlying PRES are not entirely understood. The syndrome has been postulated to represent either severe hypertension leading to failed auto-regulation and endothelial injury/vasogenic edema, or vasoconstriction leading to brain ischemia and subsequent vasogenic edema. Sudden elevations in the systemic blood pressure disrupt the bloodbrain barrier, causing a local exchange of fluids. The cerebral white matter is composed of myelinated fiber tracts in an extracellular matrix of glial cells, arterioles and capillaries and is susceptible to vasogenic edema [bib_ref] A reversible posterior leukoencephalopathy syndrome, Hinchey [/bib_ref]. The hypothesis that pronounced hypertension can lead to the breakdown of the blood-brain barrier, hyperperfusion and increased interstitial edema in PRES was initially proposed [bib_ref] Autoregulation of brain circulation in severe arterial hypertension, Strandgaard [/bib_ref]. However, many recent radiological perfusion studies have suggested an opposing hypothesis, in which hypoperfusion is central to the pathophysiological changes and brain imaging findings associated with PRES, and vasospasm may precipitate vasogenic edema, leading to cyto-toxic edema if left untreated [bib_ref] Reversible cerebral vasoconstriction syndrome, Ducros [/bib_ref].
PRES and RCVS share many clinicoradiographic features, suggesting overlapping or similar pathophysiological mechanisms. PRES and RCVS are frequently associated, and reversible brain edema occurs in 8-38% of all cases of RCVS [bib_ref] Reversible cerebral vasoconstriction syndromes: analysis of 139 cases, Singhal [/bib_ref] [bib_ref] The clinical and radiological spectrum of reversible cerebral vasoconstriction syndrome: a prospective..., Ducros [/bib_ref]. Multifocal cerebral vasoconstriction has been noted in more than 85% of patients with PRES, and this vasoconstriction was shown to be reversible on followup MRA [bib_ref] Catheter angiography, MR angiography, and MR perfusion in posterior reversible encephalopathy syndrome, Bartynski [/bib_ref]. PRES is associated with endothelial cell dysfunction and was initially thought to be caused by severe hypertension, leading to altered cerebral autoregulation with hyperperfusion and vasogenic edema [bib_ref] Posterior reversible encephalopathy syndrome, part 2: controversies surrounding pathophysiology of vasogenic edema, Bartynski [/bib_ref]. However, a quarter of patients with PRES are normotensive, and these patients have more extensive edema than do hypertensive patients, suggesting that hypertension may be a protective reaction [bib_ref] Posterior reversible encephalopathy syndrome, part 2: controversies surrounding pathophysiology of vasogenic edema, Bartynski [/bib_ref]. Recently, endothelial dysfunction of any cause has been shown able to affect the regulation of the cerebral arterial tone and trigger vasoconstriction with subsequent hypoperfusion, breakdown of the blood-brain barrier, and vasogenic edema [bib_ref] Posterior reversible encephalopathy syndrome, Staykov [/bib_ref].
In the present case, follow-up MRI of the brain revealed radiographic resolution of parieto-occipital lobe edema, which correlated with clinical improvement. However, brain MRA performed at the same time indicated that spasms of the bilateral cerebral arteries had progressed compared to the situation prior to treatment. Furthermore, instead of continuing treatment for PRES, when the patient developed cerebral infarction presenting as moderate right hemiparesis, brain MRI revealed new lesions in the left fronto-parietal area, but MRA indicated an improvement in the spasms of bilateral cerebral arteries. Cerebral infarctions induced by PRES occur mainly in the arterial watershed regions of the cerebral hemispheres, often between the posterior circulation and carotid territories [bib_ref] Reversible cerebral vasoconstriction syndromes: analysis of 139 cases, Singhal [/bib_ref] [bib_ref] Magnetic resonance angiography in reversible cerebral vasoconstriction syndromes, Chen [/bib_ref]. Edema resulting from PRES is often seen on MRI with symmetrical FLAIR hyperintensities, usually showing total reversal within one month after the clinical onset, which is much earlier than the reversal of vasoconstriction [bib_ref] Reversible cerebral vasoconstriction syndrome, Ducros [/bib_ref]. The maximum vasoconstriction of the branches of the middle cerebral arteries is reportedly reached a mean of 16 days after the clinical onset [bib_ref] Magnetic resonance angiography in reversible cerebral vasoconstriction syndromes, Chen [/bib_ref]. The MRA findings in the present case might support this report.
We suspect that PRES in our patient resulted from systemic endothelial cell dysfunction induced by bevacizumab. Clinical suspicion toward patients presenting with symptoms characteristic of PRES while receiving bevacizumab is encouraged, particularly when administered in combination with the FOLFILI or mFOLFOX6 regimens. In conclusion, we should consider PRES when cancer patients are treated with new molecular-targeted agents, as PRES is an oncological emergency.
The authors state that they have no Conflict of Interest (COI).
[fig] Figure 1: Axial-section fluid-attenuation inversion recovery (FLAIR) magnetic resonance imaging (MRI) (1.5 T; TR, 9,000 ms; TE, 105ms) performed on the day of the onset of PRES (a, b), on day 11 after the onset (c, d) and on day 19 after the onset, when the patient developed moderate right hemiparesis (e, f), and diffusion-weighted imaging (DWI) (1.5 T; TR, 3,000 ms; TE, 88 ms) performed on the day that moderate right hemiparesis developed (g, h). FLAIR images indicate bilateral increases in the signal intensity in the occipital, parietal and periventricular white matter (arrows) (a, b), partial improvement of the brain MRI findings 11 days later (c, d), new lesions in the left fronto-parietal area (e, f) and new hyperintense lesions in the left fronto-parietal area on DWI (arrowheads) (g, h). [/fig]
[fig] Figure 2: Magnetic resonance angiography (MRA) performed on the day of the onset of PRES (a), on day 11 after the onset (b) and on day 19 after the onset, when the patient developed moderate right hemiparesis (c). MRA reveals spasms of the bilateral anterior, middle and posterior cerebral arteries (arrows) (a); progression of the spasms of the bilateral anterior, middle and posterior cerebral arteries (b); and a tendency toward improvement of the spasms of the bilateral anterior, middle and posterior cerebral arteries (c). [/fig]
|
Meesmann Corneal Dystrophy; a Clinico-Pathologic, Ultrastructural and Confocal Scan Report
Purpose: To report the microstructural features of Meesmann corneal dystrophy (MCD) in two patients. Case Report: The first patient was a 10-year-old boy who presented with bilateral visual loss, diffuse corneal epithelial microcystic changes, high myopia and amblyopia. With a clinical impression of MCD, automated lamellar therapeutic keratoplasty was performed in his left eye. Histopathologic examination of the corneal button disclosed epithelial cell swelling and cyst-like intracytoplasmic inclusions. The cells contained moderate amounts of periodic acid-Schiff-positive and diastase-sensitive material (glycogen). Transmission electron microscopy revealed numerous vacuoles and moderate numbers of electron-dense membrane-bound bodies in the cytoplasm, similar to lysosomes, some engulfed by the vacuoles. The second patient was a 17-year-old female with a clinical diagnosis of MCD and episodes of recurrent corneal erosion. On confocal scan examination of both corneas, hyporeflective round-shaped areas measuring 6.8 to 41.4 µm were seen within the superficial epithelium together with irregular and ill-defined high-contrast areas in the sub-basal epithelial region. The subepithelial nervous plexus was not visible due to regional hyperreflectivity. Conclusion: This case report further adds to the microstructural features of Meesmann corneal dystrophy and suggests confocal scan as a non-invasive method for delineating the microstructural appearance of this rare dystrophy.
# Introduction
Meesmann corneal dystrophy (MCD) is a rare bilateral corneal epithelial disorder which appears in the first or second year of life and was first described by Pameijer in 1935.The pattern of inheritance is autosomal dominant but an autosomal recessive form has also been reported.On slitlamp biomicroscopy, the lesions appear as punctate, bubble-like, round to oval opacities in the corneal epithelium. On histopathology, the dystrophic epithelium is characterized by cellular swelling, cyst-like inclusions, and cytoplasmic vacuoles. The cysts appear to contain degenerated cell debris which is periodic acid-Schiff (PAS) positive. [bib_ref] Degenerations and dystrophies, Spencer [/bib_ref] Although the cells contain PAS-positive material, this may not be excessive glycogen as was previously believed; the material has been reported to be a dense intracellular substance of unknown composition.On electron microscopic examination, an electron-dense and amorphous "peculiar substance" has been reported in the cytoplasm of epithelial cells. Deposition of the substance in the epithelium leads to cyst formation and cell death followed by rapid regrowth of the epithelium. [bib_ref] Degenerations and dystrophies, Spencer [/bib_ref] Confocal scan is a non-invasive diagnostic tool for rapid evaluation of all corneal layers and in vivo diagnosis of corneal disorders. [bib_ref] Clinical confocal microscopy, Petroll [/bib_ref] [bib_ref] In vivo confocal microscopy of the human cornea, Jalbert [/bib_ref] The reported confocal microscopic features of MCD 6,7 include well delineated cystic lesions containing hyperreflective points, [bib_ref] Confocal microscopy of cystic disorders of the corneal epithelium, Hernández-Quintela [/bib_ref] hyporeflective areas in the basal epithelial layer, large elongated intraepithelial clefts and reflective spots within the hyporeflective areas. [bib_ref] Imaging the microstructural abnormalities of Meesman corneal dystrophy by in vivo confocal..., Patel [/bib_ref] Reports on the microstructural features of MCD are limited in the literature and there are only two reports 6,7 on confocal scan findings of this rare dystrophy. We believe that this report adds to the microstructural information available for MCD.
## Case reports
## Case 1
A 10-year-old boy presented with bilateral decreased vision since the age of three. His parents were consanguineous but normal on ophthalmic examinations. Best-corrected visual acuity (BCVA) was 20/160 and 20/800 in his right and left eyes with correction of -12.00 and -13.00 sphere, respectively. On slitlamp biomicroscopy, diffuse intra-epithelial microcystic changes were present within the entire corneal epithelium [fig_ref] Figure 1: The cells contained moderate amounts of PASpositive [/fig_ref]. Intraocular pressure (IOP) was normal and funduscopic examination disclosed pathologic myopic changes in both eyes. The clinical diagnosis was MCD associated with amblyopia due to high myopia. To improve visual acuity and anterior corneal clarity, [bib_ref] Recurrent Meesmann's corneal dystrophy: treated with keratectomy and mitomycin C, Yeung [/bib_ref] the patient underwent automated lamellar therapeutic keratoplasty (ALTK) with a thickness of 250 µm in his left eye.
The corneal button was sent in 10% formalin to the pathology laboratory. After bisecting the specimen, one half was processed and embedded in paraffin wax. Sections were prepared and stained with Hematoxylin and Eosin (H&E) for studying the general morphology and by PAS sequence with and without diastase to identify glycogen. [bib_ref] Endocervical adenocarcinoma. Clinico-pathologic and histochemical study of 29 cases, Resta [/bib_ref] The histopathological sections were examined by light microscopy (Olympus BX43, Olympus Co., Tokyo, Japan). The other half of the specimen was sent in 2.5% glutaraldehyde to the electron microscopy laboratory for transmission electron microscopy (EM 900, Zeiss, Germany). Histopathological examination disclosed a partial-thickness cornea with abnormal-appearing epithelium consisting of numerous intracytoplasmic cyst-like inclusions together with cellular swelling . lysosomes were also noted, some within the vacuoles . No abnormal findings were noted elsewhere. The histopathologic and electron microscopic findings confirmed the clinical diagnosis of MCD.
## Case 2
A 17-year-old female presented with foreign body sensation and pain in both eyes since 3 years ago without any significant medical or family history. BCVA in her right and left eyes was 20/40 and 20/50, respectively with refractive error of -0.5 sphere in both eyes. On retroillumunation by slitlamp biomicroscopy, there were diffuse (limbus to limbus) intraepithelial microcystic lesions together with regional haze in both corneas [fig_ref] Figure 1: The cells contained moderate amounts of PASpositive [/fig_ref] but other corneal layers were unremarkable. IOP and funduscopic examinations were within normal limits. The clinical features were characteristic for MCD. After topical anesthesia, confocal scan 3.0 (Nidek Technology, Padova, Italy) was performed on both eyes using methylcellulose as a coupling agent between the front lens (40×, 0.75 objective lens) and the surface of A B the cornea. The automatic full thickness and epithelial modes were used to capture images from all corneal layers with particular attention to the anterior parts of the involved cornea. The manual analytic software of the confocal scan was utilized to measure the abnormal confocal findings. Confocal scan examination of both corneas disclosed scattered and well-defined round to oval hyporeflective intracytoplasmic areas measuring 6.8 to 41.4 µm in their largest diameter within the superficial corneal epithelium [fig_ref] Figure 4: Round to oval, hyporeflective intracytoplasmic cystic structures [/fig_ref] , diffuse hyperreflective spots in the basal epithelium , irregular and poorly-defined high contrast areas in the subbasal epithelial region and foci of sub-epithelial fibrosis . A few hyperreflective lesions containing high contrast spots, corresponding to cell nuclei, were also present. The subepithelial nerve plexus was not visible because of the regional hyperreflectivity. No abnormal finding was noted in the rest of corneal stroma and the endothelium. The confocal microscopic features were consistent with the clinical diagnosis of MCD.
# Discussion
The histopathologic features of the excised corneal button in our first case were similar to those described by Chiou et al: [bib_ref] Recurrent Meesmann's corneal epithelial dystrophy after penetrating keratoplasty, Chiou [/bib_ref] an increase in corneal epithelial thickness, presence of intraepithelial microcysts, and increased a m o u n t s o f i n t r a c e l l u l a r g l y c o g e n . T h e ultrastructural appearance was also similar to that reported by Nakanishi et al 11 in terms of presence of intense intracytoplasmic vacuolation and formation of lysosome-like, electron-dense bodies within the cytoplasm of epithelial cells. These features are distinctly different from the electron-dense "peculiar substance" 1 or the electron-dense fibrillogranular material 10 reported earlier.
We observed hyporeflective cystic structures of various sizes in the superficial corneal epithelium on confocal scan examination of the second case, which is similar to findings previously reported for MCD. [bib_ref] Imaging the microstructural abnormalities of Meesman corneal dystrophy by in vivo confocal..., Patel [/bib_ref] [bib_ref] Confocal microscopy of cystic disorders of the corneal epithelium, Hernández-Quintela [/bib_ref] The presence of diffuse hyperreflective spots in the basal epithelium, high contrast sub-basal epithelial areas and subepithelial fibrosis were new findings in our study not previously described. It has been suggested that hyperreflective spots within cystic lesions may correspond to cell nuclei, 7 however they may be due to accumulation of intracytoplasmic lysosomes containing degenerated cellular material. We assume that the irregular high contrast areas in the sub-basal epithelial region correspond A B to nonspecific irregular thickening of the epithelial basement membrane which may be seen in most cases of MCD. [bib_ref] Degenerations and dystrophies, Spencer [/bib_ref] The diagnosis of MCD is based on clinical findings such as bilateral limbus to limbus microcystic intraepithelial changes on high power slitlamp biomicroscopy, [bib_ref] Imaging of microstructural abnormalities of Meesmann corneal dystrophy by in vivo confocal..., Tuft [/bib_ref] and may be further confirmed microstructurally with light and electron microscopy 1 or through confocal microscopy. [bib_ref] Imaging the microstructural abnormalities of Meesman corneal dystrophy by in vivo confocal..., Patel [/bib_ref] Since surgical intervention such as lamellar or penetrating keratoplasty is not indicated in the majority of patients with MCD, [bib_ref] Imaging the microstructural abnormalities of Meesman corneal dystrophy by in vivo confocal..., Patel [/bib_ref] [bib_ref] Recurrent Meesmann's corneal dystrophy: treated with keratectomy and mitomycin C, Yeung [/bib_ref] in vivo confocal microscopy can provide a noninvasive method for confirming the diagnosis.
In conclusion, this report could further add to the microstructural information available on Meesmann corneal dystrophy and present new confocal microscopic features of this rare dystrophy.
[fig] Figure 1: The cells contained moderate amounts of PASpositive (Fig. 2C), diastase-sensitive (Fig. 2D) material consistent with glycogen. Other corneal layers were unremarkable. Transmission electron microscopic examination disclosed numerous and variable-sized vacuoles within the cytoplasm in all epithelial layers (Fig. 3A). Moderate numbers of electron-dense and membranebound intracytoplasmic bodies similar to A B Diffuse intraepithelial microcystic lesions on slitlamp biomicroscopy visible by retroillumination in the first (A) and second (B) patient. [/fig]
[fig] Figure 3, Figure 2: Transmission electron microscopy: A) Numerous vacuoles (V) of variable size and clusters of electron-dense bodies (arrows) within the cytoplasm of an epithelial cell (magnification ×4400). B) Electrondense, membrane-bound bodies similar to lysosomes, some within the vacuoles in the cytoplasm of epithelial cells (magnification ×20,000). Abnormal corneal epithelium with cellular swelling and intracytoplasmic cyst-like inclusions (A & B) on Hematoxylin & Eosin staining (A: magnification × 400, B: magnification ×1000). Note the presence of moderate amounts of periodic acid-Schiff-positive (C) and diastase-sensitive (D) material within the abnormal epithelial cells (magnification ×1000). [/fig]
[fig] Figure 4: Round to oval, hyporeflective intracytoplasmic cystic structures (arrows) in the superficial corneal epithelium on confocal scan examination of the second case (A&B). [/fig]
[fig] Figure 5, Figure 6: Note the presence of diffuse hyperreflective spots in the basal corneal epithelium on confocal microscopy. Note the presence of irregular and poorlydefined high contrast areas in the sub-basal epithelial region and foci of sub-epithelial fibrosis (A&B) on confocal microscopy. [/fig]
|
Adult-onset cyclic neutropenia is a benign neoplasm associated with clonal proliferation of large granular lymphocytes
[bib_ref] Cyclic hematopoiesis: human cyclic neutropenia, Lange [/bib_ref]
# Materials and methods
Patients . Clinical details concerning patients 1-5 have been published previously (3). Patients 1-3 had adult-onset disease; patients 4 and 5 had childhood-onset cyclic neutropenia. The diagnosis of cyclic neutropenia was established by performing blood counts a minimum of three times per week for at least 6 wk, as previously described [bib_ref] Cyclic hematopoiesis: human cyclic neutropenia, Lange [/bib_ref] [bib_ref] Lithium is an ineffective therapy for human cyclic hematopoiesis, Hammond [/bib_ref]. Cycle lengths of patients 1, 3, 4, and 5 were within the 19-22-d cycle period seen in 85% of patients with cyclic neutropenia (1), whereas the cycle length in patient 2 was a bit longer at 27 d.
Patients 1-3 with adult-onset cyclic neutropenia had markedly increased LGL counts at time of diagnosis, ranging from 2,170-6,143/mms (normals in our laboratory : 223 ± 99, n = 10), whereas the two patients with childhood-onset disease had normal LGL counts. At the time of this study, patients 2 and 3 were in clinical remission from neutrophil cycling on alternate-day steroid therapy. Such therapy resulted in reduction of numbers of LGL, although LGL counts remained greater than normal (3).
Blot Hybridization Analysis. Genomic DNA was extracted from PBMC as previously described. 75-90% of these PBMC were LGL in the patients with adult-onset cyclic neutropenia . The DNA samples were then digested with restriction enzymes Bam HI, Eco RI, or Hind III . Digested DNA was separated on 1 .1 % agarose gels and transferred onto nitrocellulose filter by the method of Southern [bib_ref] Detection of specific sequences among DNA fragments separated by gel electrophoresis, Southern [/bib_ref]. Filters were then hybridized to DNA probes that had been "P-labeled by nick translation and visualized by autoradiography as previously described. The cDNA clone Jurkat B2 containing the C and J regions of the To gene (7) was kindly provided by Dr. Tak Mak (Ontario Cancer Institute, Toronto, Canada). A fragment representing nucleotides 100-870 (7) was isolated on agarose gels and used as the hybridization probe .
# Results
The human To gene locus has two constant region genes designated C,31 and CF2 [bib_ref] Organization and sequences of the diversity, joining, and constant region genes of..., Toyonaga [/bib_ref]. Digestion of non-T cell DNA with restriction enzyme Bam HI produces a 23-kb germline fragment containing both constant region genes. Therefore, rearrangements at either C a gene locus may be detected after digestion with this enzyme by the appearance of a smaller Bam HI fragment containing the C a gene . Eco RI cleaves within the T a gene locus and produces two germline fragments of 11 and 4 kb, containing CFI and CF2, respectively . Since the CF2 rearrangements are not detected by constant region probe when DNA is digested with this enzyme (9), any nongermline band observed represents rearrangement of the CFI gene . Digestion of non-T cell DNA with restriction enzyme Hind III produces two germline CF2 containing fragments of 8 and 6.5 kb, and one 3 .5kb fragment containing the CF1 gene . These patterns of somatic rearrangement of T cell receptor gene can be detected in clonal populations of T cells, and therefore can be used to demonstrate clonality of various T cell malignancies [bib_ref] Rearrangement s of T-cell receptor gene YT 35 in human DNA from..., Toyonaga [/bib_ref].
Results of Southern blot hybridization analyses using the T F gene probe are shown in [fig_ref] FIGURE 1: Results of Southern blot hybridization analyses using To gene probe showing rearrangement... [/fig_ref]. Analysis of DNA after digestion with Barn HI or Eco RI showed clonal rearrangement of T F gene in patients 1-3. In contrast, we saw a germline pattern in DNA from patients 4 and 5 after digestion with these enzymes. Hind III digestions showed germline pattern in all five patients (data not shown) . As indicated by Eco RI analysis, rearrangement in patients 1 and 3 involved the CFI gene . In patient 2, it was not possible to determine by Eco RI or Hind III digestion whether the rearrangement indicated by Bam HI analysis involved the CFI or CF2 gene . The intensity of signal of the rearranged bands indicates that the vast majority of PBMC were clonal . Therefore, it is not possible that T cells with a non-LGL morphology represented the clonal proliferation, since they were only a minor population of the PBMC .
# Discussion
These results show that all three patients with adult-onset cyclic neutropenia had a clonal proliferation of LGL, as indicated by somatic rearrangement of the TF gene . In contrast, both patients with childhood-onset cyclic neutropenia had no evidence for a clonal lymphocyte proliferation . These data suggest that adultonset cyclic neutropenia can be distinguished from the childhood-onset form of the disease by the presence of a clonal proliferation of LGL .
Having shown clonality in these patients with acquired cyclic neutropenia, the question remains whether this disease is actually malignant. The clinical course of these patients has been remarkably stable, with cyclical illnesses having occurred over 8-14 yr of duration . These patients with adult-onset cyclic neutropenia share several features of patients with LGL leukemia, who also have a clonal proliferation of LGL [bib_ref] Leukemi a of large granular lymphocytes : Association with clonal chromosomal abnormalities..., Loughran [/bib_ref] [bib_ref] Rearrangement of the gene for the beta chain of the T cell..., Aisenberg [/bib_ref] [bib_ref] Rearrangements of genes for the antigen receptor on T-cells as markers of..., Waldmann [/bib_ref] [bib_ref] Clonal rearrangement of T-cell receptor genes in LGL leukemia, Loughran [/bib_ref]. Both groups have excess LGL that express some NK cell surface antigens. These LGL have little NK cell activity in vitro [bib_ref] Rearrangement of the gene for the beta chain of the T cell..., Aisenberg [/bib_ref] ; however, cytotoxicity can be induced by treatment with anti-CD3 mAb or IL-2. Furthermore, lymphocytic infiltration of splenic red pulp cords and bone marrow has been documented in both groups [bib_ref] Rearrangement of the gene for the beta chain of the T cell..., Aisenberg [/bib_ref]. Most patients with LGL leukemia also have a chronic clinical course, with morbidity and mortality generally resulting from infections acquired during severe neutropenia rather than from tissue infiltration by abnormal lymphocytes [bib_ref] T y-lymphoproliferative disorders in man and experimental animals : a review of..., Reynolds [/bib_ref]. Thus it would appear that in most instances an abnormal clone remains under partial immunoregulatory control, as occurs, for example, in benign monoclonal gammopathy . The etiology of cyclic neutropenia is not certain, although marrow transplantation studies have shown that the defect originates at the stem cell level in both dogs and man [bib_ref] Transplantation of allogeneic bone marrow in canine cyclic neutropenia, Dale [/bib_ref] [bib_ref] Canine cyclic neutropenia : a stem cell defect, Weiden [/bib_ref] [bib_ref] Canine cyclic hematopoiesis : marrow transplantation between littermates, Jones [/bib_ref] [bib_ref] Human cyclic neutropenia transferred by allogeneic bone marrow grafting, Krance [/bib_ref]. In both childhood-onset and adult-onset disease in man, a final common pathway for the neutropenia, namely a periodic failure of production, has been demonstrated by kinetic studies [bib_ref] Periodic hematopoiesis in human cyclic neutropenia, Guerry [/bib_ref]. Furthermore, mathematical models have stressed that stable oscillations almost certainly result from an abnormality in feedback regulation of hematopoiesis [bib_ref] Cyclic neutropenia : a clue to the control of granulopoiesis, Vonschulthess [/bib_ref] [bib_ref] An analogue model of granulopoiesis for the analysis of isotopic and other..., Reeve [/bib_ref] [bib_ref] Unified hypothesis for the origin of aplastic anemia and periodic hematopoiesis, Mackey [/bib_ref] [bib_ref] Cyclic hematopoiesis and feedback control, Morley [/bib_ref] [bib_ref] Cycli c hematopoiesis: the biomathematics, Dunn [/bib_ref]. Although several investigators [bib_ref] Cell kinetics in human cyclic hematopoiesis, Dresch [/bib_ref] [bib_ref] Some immunological and haematological aspects of human cyclic neutropenia, Andrews [/bib_ref] [bib_ref] Alteratio n of colonystimulating factor output, endotoxemia, and granulopoiesis in cyclic neutropenia, Greenberg [/bib_ref] [bib_ref] Cyclic neutropenia and T lymphocyte suppression of granulopoiesis : abrogation of the..., Verma [/bib_ref] [bib_ref] Cyclica l neutropenia and T S lymphocyte mediated stimulation of granulopoiesis, Smith [/bib_ref] have attempted to document a causative role for specific feedback abnormalities in patients with this disease, considerable differences in interpretation of the data exist and no consensus for the mechanism has come forth. The data we report here strongly implicate a role for a population of LGL in the etiology of cyclic hematopoiesis.
LGL have been reported to LOUGHRAN AND HAMMOND BRIEF DEFINITIVE REPORT produce multiple regulatory factors including colony-stimulating factor [bib_ref] Human large granular lymphocytes are potent producers of interleukin-1, Scala [/bib_ref] , as well as cause inhibition ofgranulocyte/macrophage colony formation [bib_ref] Inhibition of in vitro granulopoiesis by autologous allogeneic human, Hansson [/bib_ref] [bib_ref] Inhibition of bone marrow colony formation by human natural killer cells and..., Degliantoni [/bib_ref] [bib_ref] Natural killer (NK) cell-derived hematopoietic colony-inhibiting activity and NK cytotoxic factor ...., Degliantoni [/bib_ref]. The clonal expansion of LGL could also conceivably decrease the number or function of cells (such as other lymphocytes or monocytes) crucial to the regulatory feedback loop. The demonstration of a clonal expansion of LGL in this subset of patients with cyclic hematopoiesis provides an opportunity to examine one component of the homeostatic control mechanisms for granulopoiesis.
Summary Human cyclic neutropenia occurs in children and adults. Adult-onset cyclic neutropenia is an acquired disease characterized by increased numbers of large granular lymphocytes (LGL), in contrast to childhood-onset cyclic neutropenia in which LGL counts are normal . We investigated the clonality of lymphocytes in these two groups of patients by assessing the rearrangement status of the T cell receptor ,B chain gene. Patients with adult-onset cyclic neutropenia showed clonal rearrangement of the Ts gene whereas the children did not . Since LGL are known to have multiple regulatory effects on normal hematopoiesis, the finding of a clonal proliferation of this lymphocyte population implicates these cells in the pathogenesis of cyclic neutropenia .
We thank Dr. Paul Neiman for helpful discussions .
Received for publication 14 August 1986 and in revised form
[fig] FIGURE 1: Results of Southern blot hybridization analyses using To gene probe showing rearrangement pattern after digestion with Bam HI or Eco RI . Lanes 1-3 represent patients 1-3 with adult-onset cyclic neutropenia, lanes 4-5 represent patients 4-5 with childhood-onset cyclic neutropenia. Lane 6 represents the germline pattern obtained when analyzing DNA extracted from neutrophils from a normal volunteer. Positions of rearranged bands are indicated by arrowheads . [/fig]
|
Prolonged oral sildenafil use-induced Mondor disease: a case report
[fig_ref] Figure 1: Photo of the penis [/fig_ref] [fig_ref] Figure 2: Penile ultrasonography and color Doppler ultrasound imaging [/fig_ref]
# Discussion
MD is a rare, painful, superficial venous thrombosis that has been reported to predominantly occur in the chest wall or penis [bib_ref] Mondor's disease: what's new since 1939?, Laroche [/bib_ref]. Penile MD comprises venous thrombosis of the dorsal superficial penile vein and is typically benign and self-limiting. However, its pathophysiology remains unclear [bib_ref] Mondor's disease: a review of the literature, Amano [/bib_ref]. Virchow's triad, involving vascular wall injury, venous stasis, and hypercoagulation, has been reported to cause venous thrombosis [bib_ref] Mondor's disease: a review of the literature, Amano [/bib_ref]. The most likely cause of penile MD is vascular wall injury, which is one factor in Virchow's triad, due to lengthening or splitting of the vessel from prolonged vigorous sexual activity, typically within 24 to 48 hours [bib_ref] A common presentation to an uncommon disease: Penile Mondor's disease: a case..., Walsh [/bib_ref]. Hypercoagulation due to protein S, protein C, and antithrombin III deficiency has also been reported to cause penile thrombosis [bib_ref] Subcutaneous penile vein thrombosis (penile Mondor's disease): pathogenesis, diagnosis, and therapy, Al-Mwalad [/bib_ref]. Di-
## A b
agnosis is possible through a detailed history, physical examination, and color Doppler ultrasound [bib_ref] Mondor's disease: a review of the literature, Amano [/bib_ref]. Penile MD is most frequently treated conservatively, and patients are advised to avoid sexual activity and to take nonsteroidal anti-inflammatory medication. However, in rare cases, thrombectomy can be proposed in chronic forms of penile MD or in forms resistant to treatment (persistence of the symptomatology and absence of permeabilization of the vein at 6 weeks) [bib_ref] Mondor's disease: a review of the literature, Amano [/bib_ref] [bib_ref] Penile Mondors' disease: an underestimated pathology, Sasso [/bib_ref]. Antibiotics are prescribed if there is evidence of infection, such as superficial cellulitis or a sexually transmitted disease [bib_ref] A common presentation to an uncommon disease: Penile Mondor's disease: a case..., Walsh [/bib_ref]. Sildenafil acts by blocking phosphodiesterase-5, an enzyme that promotes the breakdown of cyclic guanosine monophosphate (cGMP), causing relaxation of the corpora cavernosa and penile arteriolar smooth muscle. Owing to the localized vasodilatory effect of the nitric oxide-guanosine monophosphate pathway, the U.S. Food and Drug Administration has approved its use in the treatment of pulmonary hypertension. However, sildenafil has been reported to have various adverse effects, including venous thrombosis. shows previous reports on the association between venous thrombosis and oral sildenafil [bib_ref] Bilateral cerebral hemispheric infarction associated with sildenafil citrate (Viagra) use, Kim [/bib_ref] [bib_ref] Sildenafil-related cerebral venous sinus thrombosis and papilledema: a case report of a..., Karti [/bib_ref]. According to previous reports, the mechanisms of vascular thrombosis may be due to vascular stasis and/or platelet aggregation resulting from long-term sildenafil use. cGMP can inhibit platelet aggregation. Initially, cGMP causes platelets to clump to seal a wound and later reverses this action to stop an excessive build-up of cells that might then block a blood vessel. Sildenafil-enhanced intracellular cGMP activates cGMP-dependent protein kinase (PKG) [bib_ref] Cyclic nucleotides and phosphodiesterases in platelets, Haslam [/bib_ref]. PKG promotes von Willebrand factor-induced activation of human platelets [bib_ref] A stimulatory role for cGMP-dependent protein kinase in platelet activation, Li [/bib_ref]. Rufa et al.suggested that the chronic use of sildenafil causes high cGMP levels, which might interfere with normal endothelial function. Moreover, sildenafil lowers systolic blood pressure by 8 to 10 mmHg owing to its vascular relaxation effect, which could be the cause of vascular insufficiency [bib_ref] Overall cardiovascular profile of sildenafil citrate, Zusman [/bib_ref]. Decreased blood flow owing to blood vessel relaxation may be an additional cause of thrombus formation.
Penile MD is not life-threatening, but it can negatively affect a patient's quality of life. Considering the adverse effects of sildenafil, prolonged oral administration could cause abnormal venous thrombosis of the superficial dorsal penile vein. It is also reason-able to consider that sildenafil can cause penile MD in men with other risk factors, such as penile vascular wall injury due to vigorous sexual activity. We believe that long-term use of sildenafil causes genital vascular thrombosis, and further studies are needed to support the development of optimal guidelines for longterm use.
## Notes
## Conflicts of interest
No potential conflict of interest relevant to this article was reported.
# Funding
None.
# Author contributions
## Orcid
Han Sol Chung, https://orcid.org/0000-0003-2193-9060 You Ho Mun, https://orcid.org/0000-0003-0585-8674
[fig] Figure 1: Photo of the penis. Arrows indicate a slight protrusion of the superficial dorsal vein with mild swelling of the penis. No erythema is present. [/fig]
[fig] Figure 2: Penile ultrasonography and color Doppler ultrasound imaging. (A) Ultrasonography shows thrombosis (arrows) in the penile superficial dorsal vein. (B) Color Doppler untrasound indicates the absence of a flow signal (blue or red signal) in the thrombotic lesion (arrow). Imaging results support the diagnosis of penile Mondor disease. [/fig]
|
A systematic review of person-centered care interventions to improve quality of facility-based delivery
Introduction:We conducted a systematic review to summarize the global evidence on person-centered care (PCC) interventions in delivery facilities in order to: (1) map the PCC objectives of past interventions (2) to explore the impact of PCC objectives on PCC and clinical outcomes. Methods: We developed a search strategy based on a current definition of PCC. We searched for English-language, peer-reviewed and original research articles in multiple databases from 1990 to 2016 and conducted hand searches of the Cochrane library and gray literature. We used systematic review methodology that enabled us to extract and synthesize quantitative and qualitative data. We categorized interventions according to their primary and secondary PCC objectives. We categorized outcomes into person-centered and clinical (labor and delivery, perinatal, maternal mental health). Results: Our initial search strategy yielded 9378 abstracts; we conducted full-text reviews of 32 quantitative, 6 qualitative, 2 mixed-methods studies, and 7 systematic reviews (N = 47). Past interventions pursued these primary PCC objectives: autonomy, supportive care, social support, the health facility environment, and dignity. An intervention's primary and secondary PCC objectives frequently did not align with the measured person-centered outcomes. Generally, PCC interventions either improved or made no difference to person-centered outcomes. There was no clear relationship between PCC objectives and clinical outcomes. Conclusions: This systematic review presents a comprehensive analysis of facility-based delivery interventions using a current definition of person-centered care. Current definitions of PCC propose new domains of inquiry but may leave out previous domains.
## Plain english summary
When births are conducted in health facilities, it is challenging to balance life-saving surgical interventions, physiologic birth, and humanized care for all women. Person-centered care has recently been proposed as a promising approach to provide evidence-based and equitable birth care that is tailored to a woman's unique medical and social needs. However, there is no consensus on how to define and implement all or only some aspects of PCC into facility-based delivery settings. Luckily, many past interventions have been designed to improve the level of PCC in birth facilities. We undertook this review of person-centered delivery interventions from 1990 to 2016 in order to understand the full range of past interventions, their person-centered goals, and their impact on clinical and PCC outcomes. We knew that these interventions would be diverse in their designs and goals, so we used a method that allowed us to integrate diverse sources of data.
We used a current definition of person-centered care to systematically search the English-language literature. We explored the relationship between an intervention's stated person-centered goals and outcomes. We found close to 10,000 abstracts in our original search and narrowed this list to 47 interventions. We found that past interventions principally had the goals to improve the levels of autonomy, supportive care, social support, dignity, as well as the quality of the health facility environment. Past interventions were frequently inconsistent in their stated goals and measured outcomes; in other words, while many interventions intended to impact autonomy, they either did not measure autonomy and/or measured person-centered outcomes unrelated to autonomy. Generally, when researchers measured the level of PCC it either improved or stayed the same. We found no clear relationship between the level of PCC and clinical outcomes.
Our review presents a comprehensive picture of personcentered care interventions conducted in birth facilities. Current definitions of PCC propose new elements that have not been explored well in the past literature. At the same time, past interventions could prove informative around how to enhance PCC through decision-making, continuity midwifery care, and centering in pregnancy.
# Background
In her 1723 impassioned argument against the encroaching class of men-midwives, English midwife Elizabeth Nihell took particular issue with their use of forceps, titling her text "A treatise on the art of midwifery. Setting forth abuses therein, especially in the practices of instruments." Possibly the first mention of "abuses" in relation to childbirth in the English-language literature, Nihell foresaw the ongoing debates surrounding obstetric interventions and person-centered care in birth facilities. A spectrum of inappropriate obstetric interventions can be found in today's birth facilities, from "too much too soon" to "too little too late" [bib_ref] Beyond too little, too late and too much, too soon: a pathway..., Miller [/bib_ref] , with women on both ends of this spectrum experiencing mistreatment [bib_ref] Defining disrespect and abuse of women in childbirth: a research, policy and..., Freedman [/bib_ref]. At this time prevalence estimates of mistreatment are challenged by systematic errors in measurement, but nonetheless high percentages of women in many places experience multiple forms of mistreatment during childbirth [bib_ref] Methods used in prevalence studies of disrespect and abuse during facility based..., Sando [/bib_ref]. Over-medicalization and mistreatment can both lead to excess morbidity and mortality and both represent a violation of women's fundamental human rights [bib_ref] Health care experiences of pregnant, birthing and postnatal women of color at..., Mclemore [/bib_ref] [bib_ref] The effect of cesarean delivery rates on the future incidence of placenta..., Solheim [/bib_ref].
In developed settings many potential solutions to the problem of over-medicalization and mistreatment of women in birth facilities have been proposed [bib_ref] Discussions of findings from a Cochrane review of midwife-led versus other models..., Sandall [/bib_ref] [bib_ref] Safe prevention of the primary cesarean delivery, Caughey [/bib_ref] [bib_ref] Invisible wounds: obstetric violence in the United States, Diaz-Tello [/bib_ref]. "Person-centered care" (PCC), a concept grounded in strong provider-patient relationships, effective communication and shared-decision making, has figured large in these discussions [bib_ref] Association of the quality of interpersonal care during family planning counseling with..., Dehlendorf [/bib_ref] [bib_ref] Pain and women's satisfaction with the experience of childbirth: a systematic review, Hodnett [/bib_ref]. Lack of PCC in less developed settings may contribute to delays in care and avoidable maternal mortality [bib_ref] Manifestations and drivers of mistreatment of women during childbirth in Kenya: implications..., Warren [/bib_ref] [bib_ref] Facilitators and barriers to facility-based delivery in low-and middle-income countries: a qualitative..., Bohren [/bib_ref]. Thus, a personcentered approach holds promise in both developed and less-developed settings to improve quality of maternity care.
Why this review was necessary PCC frameworks are complex and feature multiple domains, a fact that may hinder intervention design and result in slow translation of PCC objectives into practice. PCC frameworks span anywhere from 7 to 9 domains of experience [bib_ref] The mistreatment of women during childbirth in health facilities globally: a mixed-methods..., Bohren [/bib_ref] and are made up of challenging concepts to operationalize, such as "humanization" and "dignity" [bib_ref] Methods used in prevalence studies of disrespect and abuse during facility based..., Sando [/bib_ref]. PCC domains could either extensively overlap or be at odds with each other depending on the legal, clinical, or cultural contexts [bib_ref] Methods used in prevalence studies of disrespect and abuse during facility based..., Sando [/bib_ref]. Finally, there is little to no guidance as to how a given PCC objective, or combination of objectives, might plausibly impact outcomes.
The current complexity of concepts, contexts, and impact when designing PCC interventions could be clarified by using a consistent logic, or theoretical rationale, to inform intervention design. [bib_ref] Models, strategies, and tools. Theory in implementing evidence-based findings into health care..., Sales [/bib_ref] argue that developing a rationale can "provide a foundation for designing and planning strategies for intervention and selecting tools with a better than random probability of success in implementation [bib_ref] Models, strategies, and tools. Theory in implementing evidence-based findings into health care..., Sales [/bib_ref]." We conducted this systematic reviewto map the PCC objectives of past interventions using a current definition of PCC delivery care and [bib_ref] Beyond too little, too late and too much, too soon: a pathway..., Miller [/bib_ref] to explore the impact of PCC objectives on PCC and clinical outcomes.
Refinement of instruments and quantification of the maternity care experience are critical steps prior to development of interventions with a better than random chance of impact. In one recent systematic review Nilver et al. found a wide-ranging set of 36 existing instruments that measure the maternity care experience [bib_ref] Measuring women's childbirth experiences: a systematic review for identification and analysis of..., Nilver [/bib_ref]. Sando et al. found extensive heterogeneity in the sampling techniques, eligibility criteria, and operational definitions of mistreatment during childbirth and advised caution in interpreting prevalence measures of mistreatment [bib_ref] Methods used in prevalence studies of disrespect and abuse during facility based..., Sando [/bib_ref]. Adding to the complexity, existing studies span countries and regions, and there may be cultural differences in how women define positive person-centered experiences, not to mention the diversity of both biomedical and midwifery models in health systems around the world. Similar to these reviews, we expected to find an expansive literature on the subject of person-centered care in birth facilities. Our paper uses a review method that is particularly suited to synthesizing disparate literatures and is a timely contribution to those seeking to design impactful interventions to improve the quality of maternity care around the world.
## Person-centered care frameworks in the global context
Person-centered delivery care in developed settings emerged almost exclusively in response to the impersonal and excessive medicalization of childbirth [bib_ref] Beyond too little, too late and too much, too soon: a pathway..., Miller [/bib_ref]. Only recently has PCC become an area of inquiry as a potential deterrent to facility-based childbirth. Modern objections to the over-medicalization of childbirth are rooted in events of the early twentieth century when white women in Europe and America began to pursue 'twilight sleep' [bib_ref] The curse of civilised woman: race, gender and the pain of childbirth..., Rich [/bib_ref]. The woman's movement of the 1960s and 1970s redefined birth into an event of social and personal significance, ideally controlled by an awake and empowered mother. Around the same time midwifery care re-emerged and became solidified to different extents in the official health systems of Europe and North America.
By the early 1990s efforts to reform the experience of maternity care and to shore up the profession of midwifery in developed nations were fully underway. A direct line between current definitions of PCC delivery care can be drawn to the United Kingdom's National Health Service's 1993 Changing Childbirth expert report. The report included both a strategy for midwifery care and a "Patient's Charter" that laid out the rights of maternity patients. Meanwhile, in the international context the 1994 International Conference on Population and Development in Cairo became a turning point for a rights-based approach to sexual and reproductive health [bib_ref] The role and limitations of the Cairo International Conference on Population and..., Dejong [/bib_ref]. Twenty years later a synthesis of international reproductive rights declarations produced the first publications on "respectful maternity care" (RMC). The RMC frameworks were soon followed by statements from all the major international health organizations denouncing the mistreatment of women in childbirth [bib_ref] Defining disrespect and abuse of women in childbirth: a research, policy and..., Freedman [/bib_ref]. The WHO, leveraging its power as a norm-setting organization, then published a framework to establish the experience of care as a pillar of quality maternity care [bib_ref] Quality of care for pregnant women and newborns-the WHO vision, Tuncalp [/bib_ref]. Now with close to 30 years of discussion around person-centered care, several overlapping strains of PCC exist, but with gaps between the different approaches. We described one strain above that aligns with the re-emergence of midwifery care and includes different approaches to the provision of care (continuity midwifery models, centering in pregnancy, doula-supported childbirth). More recently, the Lancet Midwifery series examined the contributions of midwives in the global context, including resource-poor settings which have not been exposed to excessive medicalization [bib_ref] Health-care professionals in midwifery care, Stones [/bib_ref]. A second strain of PCC in the global literature has emerged relatively separate from midwifery, namely the framework around mistreatment of pregnant women, as advanced by Bowser and Hill. Bohren et al. revised the Bowser and Hill typology to develop the most comprehensive set of PCC categories in the maternity context [bib_ref] The mistreatment of women during childbirth in health facilities globally: a mixed-methods..., Bohren [/bib_ref].
However, between the midwifery approach and the mistreatment typologies there existed a persistent gap as to how these two areas might be inter-related.
Thus, a broader approach was necessary if the model of care and mistreatment categories were going to be useful in a range of resource settings and in health systems with different proportions of technological obstetric care and primary midwifery care. Given this tension between local contexts and universal frameworks for PCC and for models of care, Sudhinaraset et al. [bib_ref] Advancing a conceptual model to improve maternal health quality: the person-centered care..., Sudhinaraset [/bib_ref] conducted a trans-disciplinary review to create the most comprehensive and adaptable PCC framework to date, which they call The Person-Centered Framework for Reproductive Health Equity. In this framework Sudhinaraset et al. link the provision of care to the experience of person-centered care, using a similar typology to Bohren et al. However, they go beyond the typology approach and link the provision of care to PCC. The provision of care encompasses evidence-based care, both the over and underuse of technology, information and referral systems, infrastructure, human resources, and the medical supply chain. Sudhinaraset and colleagues also situate PCC within the context of a community's experiences with care, as a community's specific history with discrimination can determine care-seeking behaviors. Ongoing experiences with mistreatment in the facility can in turn influence a community's care-seeking behaviors. Finally, the authors link the facility and care-seeking behaviors to societal and community determinants of health equity, including gender and violence norms [bib_ref] Moving beyond disrespect and abuse: addressing the structural dimensions of obstetric violence, Sadler [/bib_ref] [bib_ref] Neither medicine nor health care staff members are violent by nature: obstetric..., Morales [/bib_ref]. In this review, we chose to use Sudhinaraset and colleagues' framework as we see theirs as the most comprehensive and flexible PCC framework. Importantly, for a global review of PCC delivery interventions, the Sudhinaraset et al. framework points to the ways in which PCC components need to be contextualized within specific health systems, gender and violence norms, and community behaviors [bib_ref] Advancing a conceptual model to improve maternal health quality: the person-centered care..., Sudhinaraset [/bib_ref]. Finally, this systematic review protocol and the PCC Framework for Reproductive Health Equity were developed contemporaneously within a cooperating research group.
# Methodology
We knew that the interventions and outcomes included under the framework of person-centered care would not be amenable to meta-analysis. As a result, this systematic review applied the qualitative method of framework analysis in order to define concepts, map the range of the phenomena, create typologies, find associations, seek explanations, and develop new ideas [bib_ref] Carrying out qualitative analysis, Ritchie [/bib_ref]. The initial step consisted of a systematic approach to problem identification, which we identified to be the complexity of concepts, contexts, and potential impacts that result from PCC frameworks. We followed reporting standards for systematic reviews of social interventions set forth by the Campbell Collaboration, including the development and publication of a protocol with pre-determined inclusion criteria and analysis plan which was registered with the PROSPERO International prospective register of systematic reviews.
## Inclusion and exclusion criteria
In order to be included, an article had to: (1) contain original data (quantitative or qualitative), (2) consist of an evaluation, (3) have at least one PCC objective designed into the intervention and (4) be facility-based. We defined quantitative data as using inferential statistics and qualitative data as primary narratives from participants. We defined an "evaluation" as any quantitative study that utilized a control group (experimental, quasiexperimental). Quasi-experimental quantitative studies needed to collect longitudinal and/or cross-sectional data from treatment and comparison groups. A qualitative evaluation had to be associated with a new personcentered delivery intervention, but we did not require a control group. We defined an objective as the primary goal that the intervention sought to impact. We defined "person-centered" objectives using a current definition grounded in the literature, encompassing: dignity, autonomy, privacy/confidentiality, communication, social support, supportive care, trust, and the health facility environment (See Additional file 1: [fig_ref] Table 1: Characteristics of all articles included in the review [/fig_ref] for definitions) [bib_ref] Advancing a conceptual model to improve maternal health quality: the person-centered care..., Sudhinaraset [/bib_ref]. We defined "facility-based" as having some linkage to a hospital or birth center and, required that outcomes be measured at the level of the facility. We defined person-centered outcomes according to the same criteria as the objectives. We defined clinical outcomes to include: labor and delivery, perinatal, and maternal mental health. If a PCC intervention was conducted in the prenatal setting, the measured outcomes had to cross over into the delivery setting.
We excluded quantitative studies that lacked a valid control or comparison group, exclusively examined prenatal outcomes (e.g., ambulatory prenatal diabetes care) or postpartum outcomes (e.g., breastfeeding). We recorded the number of excluded studies and the reason for exclusion at each stage.
## Search strategy
We designed a search strategy to maximize the number of primary sources. We searched the English language literature from 1990 until 2016. We conducted a first search in October 2015 and a second search in April 2016 in order to identify any new publications. We systematically searched peer-reviewed literature in PubMed, CINAHL, EconLit, and EMBASE using controlled search terms and free-text terms combining three main components: (a) pregnancy and delivery care (b) person-centered care and (c) interventions (See Additional file 2: Search Strategy). The final keyword chain from April 2016 differed slightly in that we added terms for group prenatal care and birth plans. Otherwise, the 2016 search was only adjusted for the date of publication. The same keywords were used in CINAHL, EconLit, and EMBASE according to their respective search engine requirements.
We hand searched the Cochrane database for all studies related to maternity care. We extensively searched the gray literature, including reports from relevant governmental and non-governmental organizations' websites by using Google Scholar keyword searches (See Additional file 2: Search Strategy). We searched dissertations and theses in the ProQuest database. Finally, we used bibliographic back referencing to identify additional studies not captured by any of the above searches. We maintained a search diary describing the search methods, keywords used, and search results.
## Screening and data extraction
We excluded duplicate references. Next, we independently reviewed titles, abstracts, and executive summaries; we excluded all references that were clearly not relevant. Following this, two team members independently applied the pre-specified inclusion/exclusion criteria to the remaining abstracts. Disagreements regarding the inclusion status of any article were presented to a third team member for a final decision. When the abstract did not contain sufficient information for inclusion, the full text was retrieved.
Three researchers then independently performed fulltext reviews and extracted quantitative or qualitative data. Data extractions were checked by one other researcher. Descriptions of interventions were assembled including the study setting; sample characteristics; objectives; design; data collection and analysis methods. We extracted person-centered and clinical outcomes that had significance testing of p < 0.05. We summarized these outcomes in tables in a qualitative manner by the direction of their effect (positive, negative, no differences). Two data extractors worked independently, followed by an independent third checker. Themes, findings, and participant quotations were extracted from qualitative studies.
# Analysis
We first categorized interventions according to their PCC objectives. All interventions were assigned a primary PCC objective, such as "autonomy" or "supportive care," through a consensus process. An intervention frequently was assigned multiple secondary PCC objectives. Then, undergoing a data reduction process, interventions that shared a primary objective were grouped together and then sub-categorized into conceptual groupings. Following Sudhinaraset et al.'s framework, we formed the conceptual groupings into established models of care.
We categorized outcomes using the same categories as the PCC objectives. We summarized overarching themes and directions of PCC and clinical outcomes. We conducted a thematic analysis of qualitative data [bib_ref] Using thematic analysis in psychology, Braun [/bib_ref] , a process that included reading repeatedly to extract concepts, categories, and metaphors used to describe or interpret PCC as experienced by the women interviewed. We integrated the quantitative and qualitative data into a final summary. We restricted our search terms by intervention type, rather than outcomes and thus included studies that measured a range from outcomes, from clinical maternal and perinatal outcomes to satisfaction and mental health. Given the diversity of quantitative outcomes, we are not able to combine results to make statements about effect sizes.
## Data quality assessment and risk of bias
Two researchers independently assessed the risk of selection, confounding, performance, and reporting bias;
we coded the studies as low, medium, or high risk for each of the four types of bias. Qualitative data were independently appraised by two researchers using the 9-item Critical Appraisal Skills Programme Qualitative Research Checklist. Risk of bias and quality assessment summaries can be found in the Additional file 3: [fig_ref] Figure 1: Study search flow diagram [/fig_ref] and [fig_ref] Table 2: Detailed descriptions of included quantitative studies, organized by primary PCC objective PCC [/fig_ref].
# Results
## General overview
The initial database searches yielded 11,409 articles with N = 9378 after duplicates were removed. After title-abstract screening was performed, N = 947 remained, and after full text reviews N = 100 potentially eligible studies were found. Of the final included studies (N = 47), 34 resulted from database searches and were supplemented by hand-searches (N = 7), bibliographic back referencing (n = 5), and theses and dissertations (n = 1). The review process and descriptive characteristics of the studies are reviewed in [fig_ref] Figure 1: Study search flow diagram [/fig_ref] and [fig_ref] Table 1: Characteristics of all articles included in the review [/fig_ref]. We categorized interventions (both quantitative and qualitative findings) into five primary PCC objectives: autonomy (N = 20), supportive care (N = 17), social support (N = 11), the health facility environment (N = 2), and dignity (N = 1). We did not find any intervention designed with these primary objectives: trust, privacy/confidentiality, or communication. Nonetheless, these three domains were addressed as secondary objectives. Below brief summaries of the results are organized from most common primary PCC objective to least common (for definitions of PCC objectives/outcomes and more detailed descriptions of the interventions, see Additional file 1: [fig_ref] Table 1: Characteristics of all articles included in the review [/fig_ref]. In [fig_ref] Table 2: Detailed descriptions of included quantitative studies, organized by primary PCC objective PCC [/fig_ref] we present detailed information about each quantitative study, PCC domain, intervention type and description, and the effect size of the outcomes; [fig_ref] Table 3: Detailed descriptions of included qualitative studies, organized by primary PCC objective [/fig_ref] presents the comparable information for the qualitative studies. In Tables 5, 6, 7 and 8 we briefly summarize the quantitative interventions and in the qualitative interventions. Each [fig_ref] Table 5: Primary PCC Objective [/fig_ref] , 7, 8 and 9) represents a single primary PCC objective; most tables contain a sub-grouping of interventions according to a model of care (for instance, "continuity midwifery" under the PCC objective of "autonomy"). Each intervention has two rows: the first row contains the PCC objectives (marked with an "X"); the second row contains PCC and clinical outcomes. The key below explains how to interpret the direction of the outcomes.
# Results
## Primary pcc objective#1: autonomy (20 papers)
Autonomy: Interventions The most common primary PCC objective was autonomy with 16 total interventions. Autonomy interventions fell broadly into two categories: labor and birth decisions and continuity midwifery. Specific labor and birth decision interventions implemented birth plans, open access to case notes, rigorous informed consent processes, and decision support. Of the 10 labor and birth decision interventions the largest proportion (4) tested some type of birth plan (Kuo [bib_ref] Evaluation of the effects of a birth plan on Taiwanese women's childbirth..., Kuo [/bib_ref] , Lundgren [bib_ref] Is the childbirth experience improved by a birth plan?, Lundgren [/bib_ref] , Martinez, Mehdizadeh [bib_ref] Evaluation of the impact of birth preparation courses on the health of..., Mehdizadeh [/bib_ref]. The Horey [bib_ref] Interventions for supporting pregnant women's decision-making about mode of birth after a..., Horey [/bib_ref] systematic review included three Randomized Control Trials (RCTs) of decision support for women who desired a vaginal birth after cesarean (VBAC), and Fraser [bib_ref] Randomized controlled trial of a prenatal vaginal birth after cesarean section education..., Fraser [/bib_ref] and Martinexamined specialty support clinics for women who desired a VBAC. The majority of interventions (8) connected women to specialized research staff (lay or clinical) to assist with individualized plans. Only two studies gave women access to information without additional assistance (Brown PCC: More likely to have named midwife as primary caregiver (97% vs 74%), to say they knew their primary provider "very well" (16% vs 4%), preferred to see their primary caregiver (86% vs 50%), to state they were "very well prepared" for birth (18% vs 12%), to feel confident about labor (51% vs 39%), to rate the birth as "hard work but wonderful" (51% vs 39%), have continuous support from midwife (90% vs 53%), and more likely to be "very satisfied" (79% vs 71% [bib_ref] Giving women their own case notes to carry during pregnancy, Brown [/bib_ref] , O'Cathain [bib_ref] Use of evidence based leaflets to promote informed choice in maternity care:..., O'cathain [/bib_ref]. Six manuscripts examined continuity midwifery care. Even though autonomy was the primary aim of these interventions, investigators stated multiple secondary person-centered objectives including supportive care, trust, dignity, privacy, and social support.
Autonomy: Outcomes Eight of the ten labor and birth decision interventions measured autonomy, 2 out of 10 measured social support and supportive care, 1 of 10 measured the health facility environment, dignity, and communication. Generally, interventions either improved or made no difference to PCC outcomes, with the exception of Lundgren's [bib_ref] Is the childbirth experience improved by a birth plan?, Lundgren [/bib_ref] birth plan experiment that showed a negative impact on dignity, communication, and supportive care. Seven of the studies looked at a labor and delivery outcomes, and found positive, negative, and null results. Regarding the negative outcomes, Brown [bib_ref] Giving women their own case notes to carry during pregnancy, Brown [/bib_ref] found women with access to case notes had more operative deliveries and cesarean sections, and Mehdizadeh [bib_ref] Evaluation of the impact of birth preparation courses on the health of..., Mehdizadeh [/bib_ref] found more use of oxytocin. None of the studies that measured perinatal outcomes or maternal mental health found any impact. Five of the six continuity midwifery care interventions measured aspects of autonomy and all reported beneficial effects. Five of 6 measured trust by inquiring about the nature and extent of continuity with the known midwife. Three of 6 measured supportive care, communication, and dignity, while 1 of 6 measured privacy and the health facility environment. Continuous care with a midwife decreased obstetric interventions almost across the board. Only Gu [bib_ref] The effectiveness of a Chinese midwives' antenatal clinic service on childbirth outcomes..., Gu [/bib_ref] measured mental health and found no difference in levels of maternal anxiety.
Autonomy: Qualitative evaluations (4 papers) Many of the qualitative evaluations concerning the PCC objective of autonomy confirmed the quantitative findings in that women generally gave positive reviews to decision support. Brown [bib_ref] Giving women their own case notes to carry during pregnancy, Brown [/bib_ref] found that women supported carrying their own pregnancy records to facilitate shared communication with providers. Horey [bib_ref] Interventions for supporting pregnant women's decision-making about mode of birth after a..., Horey [/bib_ref] found four qualitative studies conducted concurrently with VBAC decision support trials; women with a prior cesarean perceived a sense of choice and gave positive evaluations to the information provided. However, some women reported that information about their options (VBAC or repeat cesarean) raised anxiety levels if the likelihood of certain risks was not included. In terms of feasibility, many women had to seek out additional support from research staff to use the decision tools. Walsh [bib_ref] An ethnographic study of women's experience of partnership caseload midwifery practice: the..., Walsh [/bib_ref] conducted ethnographic interviews with women (N = 10) in a continuity midwifery practice. Women placed great importance on their relationships with known midwives, and as a result felt more comfortable asking questions and felt that their concerns were validated. Home antenatal visits were well-reviewed because partners and children could be involved. While in labor at home, women appreciated having a known midwife and used expressions of "delight" to describe their labor experiences. This contrasted to women's first hospital labors that felt de-humanized and lacked privacy. De Koninck [bib_ref] Comparing women's assessment of midwifery and medical care in Québec, De Koninck [/bib_ref] conducted a mixed-methods assessment of midwifery and medical care in Quebec and used open ended responses (N = 182) and interviews (N = 10) to contextualize the quantitative findings. Women described visits with doctors as "rushed" and "austere". As a result, women reported feeling undervalued and held back questions. Clients of midwives "felt respected" and didn't "feel like a number". During labor women felt they had to be in a position suitable to the obstetrician, but with midwives, "I did not have to move for the midwife to be comfortable." Women extended similar positive reviews to labor and delivery nurses, who they often found to be responsive to their needs [bib_ref] Comparing women's assessment of midwifery and medical care in Québec, De Koninck [/bib_ref] [fig_ref] Table 4: How to interpret direction of outcomes in summary tables [/fig_ref].
## Primary pcc objective#2: supportive care (17 papers)
Supportive care: Interventions The second most common set of interventions identified in this review concerned supportive care (N = 15). These interventions fell broadly into two categories: enhanced prenatal care for at-risk women and psychological support. A total of 9 studies targeted at-risk women, including adolescents (Grassley [bib_ref] Evaluation of the supportive needs of adolescents during childbirth Intrapartum nursing intervention..., Grassley [/bib_ref] , low-income and/or ethnic minority women (El-Mohandes [bib_ref] Very preterm birth is reduced in women receiving an integrated behavioral intervention:..., El-Mohandes [/bib_ref] , Harris [bib_ref] Effect of a collaborative interdisciplinary maternity care program on perinatal outcomes, Harris [/bib_ref] , Kildea [bib_ref] The Murri clinic: a comparative retrospective study of an antenatal clinic developed..., Kildea [/bib_ref] , and women at risk for pre-term birth , Mason [bib_ref] Effects of a pregnancy management program on birth outcomes in managed Medicaid, Mason [/bib_ref] , Newman [bib_ref] South Carolina Partners for Preterm Birth Prevention: a regional perinatal initiative for..., Newman [/bib_ref] , Panaretto [bib_ref] Impact of a collaborative shared antenatal care program for urban indigenous women:..., Panaretto [/bib_ref]. Generally, this group of interventions aimed to optimize access to quality prenatal care, in order to decrease the effects of socio-medical risk factors that influence adverse pregnancy outcomes. Another 6 studies in the supportive care category sought to intervene upon the woman's psychological state, especially as it related to anxiety or fear of childbirth, her level of mindfulness, and her coping skills.
Supportive care: Outcomes While the stated objective was to provide supportive care to at-risk women, no study actually measured supportive care. Of the three studies focused on low-income or ethnic minority [bib_ref] The Murri clinic: a comparative retrospective study of an antenatal clinic developed..., Kildea [/bib_ref]. Two of 6 studies inquired about various aspects of at-risk women's autonomy. The majority of trials focused on at-risk women sought to decrease pre-term birth and found positive or null results. Those that measured clinical outcomes (epidural, cesarean, planned VBAC) generally also had positive or null results. Compared to the at-risk women, psychological interventions consistently measured person-centered outcomes, including support (3/6) and autonomy (3/6). Many also explored the impact on mental health (5/6) using validated scales for depression, anxiety, and post-traumatic stress disorder (PTSD). Results on PCC measures were mixed, with Ryding [bib_ref] An evaluation of midwives' counseling of pregnant women in fear of childbirth, Ryding [/bib_ref] finding that women in the intervention group reported a more frightening experience in childbirth and more post-traumatic stress. Clinical outcomes measured were generally positive (more spontaneous vaginal births, shorter labor lengths). Perinatal outcomes were not frequently explored, and those that did found no impact.
Supportive care: Qualitative evaluations (2 papers) Kildea [bib_ref] The Murri clinic: a comparative retrospective study of an antenatal clinic developed..., Kildea [/bib_ref] conducted a mixed methods analysis of a specialized clinic for ethnic minority women. In faceto-face interviews women gave positive reviews to the continuity of care model, as women liked not having to repeat information with providers. Women reported that it was more important that the antenatal provider listened to them rather than share the same cultural identity. Women also appreciated the flexible drop-in appointment system and the proximity of the clinic to the labor ward, which helped partners/family know where to take the women when labor started. Some women reported that the waiting room could be crowded and lacked privacy. While in labor, women wanted to maintain physical modesty and privacy by limiting the number of hospital staff. Finally, they feared that people entering the room indicated a problem with the labor or baby. Stapleton [bib_ref] Women from refugee backgrounds and their experiences of attending a specialist antenatal..., Stapleton [/bib_ref] conducted 4 focus groups with refugee women about their experiences with a specialized antenatal clinic. Refugee women also appreciated the continuity model of care because they did not have to repeat traumatic histories. Continuity allowed for more efficient use of interpreters; however, a challenge was clients using interpreters as sources of clinical information. Geographic distance created barriers to access, either because language difficulties arose on public transit or women because women had to rely on husbands for transportation. Refugee women were accustomed to having female kin support them in labor and reported feeling isolated in the new Australian context .
## Primary pcc objective#3: social support (11 papers)
Social support: Interventions The third group of interventions sought to increase support for women during pregnancy and birth care through involvement of male partners and/or continuous labor support, or group prenatal care. Through these interventions male partners were invited to be more involved in care, or women were provided with the continuous support of a doula during labor and birth. Group prenatal care interventions encouraged women to connect and support each other outside the confines of individual prenatal visits.
Social support: Outcomes While the expressed objective of male partner and continuous labor support interventions was social support, only Gungor [bib_ref] Effects of fathers' attendance to labor and delivery on the experience of..., Gungor [/bib_ref] and Kuneneactually measured support, both with positive outcomes. Three of the male partner/continuous support studies measured autonomy (Gungor [bib_ref] Effects of fathers' attendance to labor and delivery on the experience of..., Gungor [/bib_ref] , Hodnett [bib_ref] Continuous support for women during childbirth, Hodnett [/bib_ref] , Mullany [bib_ref] The impact of including husbands in antenatal health education services on maternal..., Mullany [/bib_ref] , all with positive effects. Results were mixed for clinical outcomes, with some evidence of positive impact on labor and delivery (Hodnett [bib_ref] Continuous support for women during childbirth, Hodnett [/bib_ref]. However, other studies found no impact on pain meds, obstetric interventions, or type of delivery (Mullany [bib_ref] The impact of including husbands in antenatal health education services on maternal..., Mullany [/bib_ref] , Gungor [bib_ref] Effects of fathers' attendance to labor and delivery on the experience of..., Gungor [/bib_ref] , Gruber [bib_ref] Impact of doulas on healthy birth outcomes, Gruber [/bib_ref]. Two papers looked at perinatal outcomes (Hodnett [bib_ref] Continuous support for women during childbirth, Hodnett [/bib_ref] ; Gruber [bib_ref] Impact of doulas on healthy birth outcomes, Gruber [/bib_ref] with positive or null results. None of the group prenatal care studies measured social support. Two of the three group prenatal care studies measured autonomy with positive or null results. Results were mixed for clinical outcomes, with some significant (Barr [bib_ref] Evaluation of a group prenatal care-based curriculum in a family medicine residency...., Barr [/bib_ref] and some null findings . Similarly, results were mixed for perinatal outcomes (lower preterm birth for Barr [bib_ref] Evaluation of a group prenatal care-based curriculum in a family medicine residency...., Barr [/bib_ref] and no difference for Catling [bib_ref] Group versus conventional antenatal care for women, Catling [/bib_ref]. Only Catling measured maternal mental health and found no differences in depression with group prenatal care.
Social support: Qualitative results (3 papers) Herrman [bib_ref] Womenʼs perceptions of CenteringPregnancy: a focus group study, Herrman [/bib_ref] explored the strengths and weaknesses of group antenatal care by conducting 5 focus groups. Women felt respected in the group environment, drew on the knowledge of the other mothers in the room, and reported a greater sense of capability to become mothers. Risisky [bib_ref] Women's perceptions using the CenteringPregnancy model of group prenatal care, Risisky [/bib_ref] conducted 3 focus groups of group antenatal care during which women reported appreciating the rich conversations created by women sharing experiences together. This helped women feel more empowered as decision makers. Women appreciated having partners attend the group prenatal sessions, as this helped partners become more effective sources of support. Even though few quantitative studies measured social support, social support was a prominent theme in the qualitative evaluations.
Hazard [bib_ref] Hispanic labor friends initiative: supporting vulnerable women, Hazard [/bib_ref] interviewed Spanish-speaking Hispanic women to evaluate a culturally-sensitive program of labor support compared to women who received standard care. Women who received the intervention appreciated having the cultural and social support from trained Hispanic doulas, demonstrated increased use of healthcare resources, reported enhanced quality of informed consent, and fewer language barriers with providers. As a whole the qualitative data did not address clinical outcomes with the exception of Hazard [bib_ref] Hispanic labor friends initiative: supporting vulnerable women, Hazard [/bib_ref] , who reported that intervention women exhibited more care-seeking behaviors .
## Primary pcc objective#4: the health facility environment (2 papers)
The health facility environment: Interventions Two studies examined the health facility environment through alternative birth sites (Hodnett [bib_ref] Alternative versus conventional institutional settings for birth, Hodnett [/bib_ref] and a new physical organization for a labor room (Janssen [bib_ref] Single room maternity care and client satisfaction, Janssen [/bib_ref]. These interventions also intended to impact trust and clinical outcomes.
The health facility environment: Outcomes Janssen [bib_ref] Single room maternity care and client satisfaction, Janssen [/bib_ref] measured women's perceptions of the new physical Primary PCC objective: Supportive Care space, which were positive; Hodnett [bib_ref] Alternative versus conventional institutional settings for birth, Hodnett [/bib_ref] did not measure women's perceptions of the alternative health environment. Regarding other PCC outcomes, Hodnett only measured autonomy, with a positive impact. Janssen found positive impacts on dignity, autonomy, privacy, social support, supportive care, and the health facility environment. Both studies found positive impacts on clinical outcomes, with lower rates of epidurals, labor augmentation, and episiotomy and higher rates of vaginal birth and women using more comfort measures for pain (Janssen). Hodnett found a reduction in low Apgar scores. Neither study measured maternal mental health outcomes.
Primary PCC objective#5: Dignity (1 paper) Dignity: Interventions One intervention (Abuya [bib_ref] The effect of a multi-component intervention on disrespect and abuse during childbirth..., Abuya [/bib_ref] included in this review utilized the current framework of respectful maternity care. We categorized the primary objective of Abuya as increasing dignity by decreasing mistreatment of women, although the intervention took a multi-pronged approach and addressed several secondary PCC objectives.
Dignity: Outcomes Abuya found a positive impact on dignity (decreased disrespect and abuse of women), an increase in women's autonomy, and no difference in supportive care and privacy. No clinical outcomes were measured .
# Discussion
We conducted the first systematic review of person-centered care interventions in birth facilities using a current and comprehensive framework for PCC. We found that since the 1990s the absolute number of PCC delivery interventions has increased. We found that applying a current PCC framework was feasible and applicable to multiple prior interventions, covering five primary PCC objectives (autonomy, supportive care, social support, the health facility environment, and dignity). Past PCC interventions attempted to empower and support pregnant women to a variety of ends, usually to decrease inappropriate obstetric interventions, improve perinatal outcomes, to directly impact maternal mental health, or decrease disrespect and abuse. We found very few examples of harm caused by a person-centered intervention, and many examples of either null or positive effects.
Given the contextual nature of person-centered objectives, using the mixed-methods systematic review allowed us to examine the contour and correlations between PCC objectives and PCC outcomes. Within our own consistent use of PCC categories, we found that Primary PCC Objective: Social Support PCC objectives frequently did not overlap with PCC outcomes. For instance, while supportive care was the explicit goal for at-risk women, none in this group of interventions measured supportive care. At the other end of the spectrum, some researchers found improvements in PCC outcomes that were not stated anywhere in the intervention's objectives.
Building on Sudhinaraset et al.'s framework [bib_ref] Advancing a conceptual model to improve maternal health quality: the person-centered care..., Sudhinaraset [/bib_ref] , our review highlights several gaps in the PCC intervention literature. While we did find numerous studies that explored the relationship between PCC and the provision of care, we found only one intervention that linked the health facility to changes in the health system, gender and violence norms, or community care-seeking behaviors. Abuya et al linked the mistreatment of women in the facility to "community accountability and governance" [bib_ref] The effect of a multi-component intervention on disrespect and abuse during childbirth..., Abuya [/bib_ref]. Also, while we did find a number of interventions designed to improve perinatal outcomes for at-risk Primary PCC Objectives: Qualitative Evaluations Primary PCC Objective: The Health Facility Environment and Dignity women, none of these interventions measured individual perceptions of discrimination, nor did these interventions attempt to link the facility to external accountability mechanisms nor to structural interventions to transform systemic discrimination [bib_ref] Racism and health II: a needed research agenda for effective interventions, Williams [/bib_ref].
This review raises questions about the theoretical coherence of PCC interventions, in terms of how faithfully researchers should match objectives and outcomes, which combinations of domains might result in the greatest benefit (or rarely, harm), and the relationship between PCC objectives and clinical outcomes. Thus, our review builds on a growing body of evidence as to the heterogeneous approaches to measuring and intervening upon women's experiences of evidence-based, quality maternity care [bib_ref] Methods used in prevalence studies of disrespect and abuse during facility based..., Sando [/bib_ref] [bib_ref] Measuring women's childbirth experiences: a systematic review for identification and analysis of..., Nilver [/bib_ref]. By tying together past interventions with the newer frameworks around respectful maternity care, we have demonstrated a longer tradition of personcentered objectives in maternity care. Future interventions can and should draw on the rich literature related to decision-support, continuity midwifery, group prenatal care, and alternative birth sites.
# Limitations
There are several limitations to this systematic review. First, the nature of our search terms and exclusion strategy lead to many papers from low-resource settings being excluded and thus the majority of papers are from high-resource settings. While there is a longer history of interventions on person-centered care from the developed world and subsequently more literature, it is unclear if these findings are as relevant to a developing world setting. As mentioned above, women in different settings may desire different aspects of person-centeredness, so interventions in one setting may not be relevant in another. Relatedly, while our definition of PCC was based on findings from around the globe, our domains might not appropriately represent all women's expectations. Furthermore, we limited our search to interventions in facilities and to outcomes measured post-delivery. There are many exciting person-centered interventions that were exclusively based in the community, prenatal or post-natal settings. These diverse PCC interventions deserve attention and critical review. Finally, we found that the qualitative literature was very narrow based on our search terms; we recommend that a future mixed-methods systematic review use broader search terms, especially to identify relevant studies that were not directly attached to an intervention.
[fig] Figure 1: Study search flow diagram [/fig]
[table] Table 1: Characteristics of all articles included in the review [/table]
[table] Table 2: Detailed descriptions of included quantitative studies, organized by primary PCC objective PCC: Longer visits (78 vs. 33 min, p < 0.001), had the opportunity to ask questions "very often" (84.6% vs. 64.1%, p < 0.001), rated their care as "very personalized" (87.9% vs. 33% p < 0.001). Delivered by a continuity provider (70.5% vs. 38.8%), able to choose labor position (84% vs. 25%, p < 0.001). Feeling of control over delivery (mean 4.33 vs. 3.95, p p < 0.001). PCC: No difference in perception of control on the Birth Experience Rating Scale. Labor and delivery: No difference in vaginal delivery. Perinatal: No differences in perinatal mortality or maternal morbidity. [/table]
[table] Table 3: Detailed descriptions of included qualitative studies, organized by primary PCC objective [/table]
[table] Table 4: How to interpret direction of outcomes in summary tables [/table]
[table] Table 5: Primary PCC Objective: Autonomy [/table]
|
Efficient generation of twin photons at telecom wavelengths with 2.5 GHz repetition-rate-tunable comb laser
## I n t e n s i t y ( v )
T i m e ( n s ) (a-f) The spectra, autocorrelation, and temporal sequences of the comb laser at 10 GHz and 2.5 GHz repetition rates. The full-width-at-half-maximum (FWHM) of the autocorrelation data are round 3.6 ps, corresponding to FWHM of 2.6 ps for the fundamental lasers.
## Supplementary-ii
In this part we investigate the relationship between signal to noise ratio (SNR) and the main/side peaks in Time of Arrival (ToA) data. The output state from the spontaneous parametric down conversion (SPDC) can be expressed as |ψ = 1 − λ 2 ∞ n=0 λ n |n, n = 1 − λ 2 (|0, 0 + λ |1, 1 + λ 2 |2, 2 + ...),
where λ is the squeezing parameter. The probability for the n-pair photons per pulse is
[formula] P r n = (1 − λ 2 )λ 2n .(2) [/formula]
The average number of photon pairs per pulse is
[formula] p = (1 − λ 2 ) ∞ n=0 nλ 2n = λ 2 1 − λ 2 .(3) [/formula]
The 1-pair photons (|1, 1 components in Eq. (1)) are the "signal", e.g., for constituting a single-photon source. All the other photons are contamination, and should be viewed as the "noise". Therefore, the signal to noise ratio (SNR) can be naturally defined as the ratio of single pair emission rate over all the other n-pair emission rates,
[formula] SN R ≡ P r 1 P r 2 + P r 3 + P r 4 + ... = (1 − λ 2 )λ 2 (1 − λ 2 )(λ 4 + λ 6 + λ 8 + ...) = 1 ∞ n=1 λ 2n = 1 − λ 2 λ 2 = 1 p .(4) [/formula]
In Eq. (4), it can be noticed that the SNR is the inverse of the averaged photon pair number per pulse, p. For a low pump power, where λ 2 1,
[formula] SN R = 1 − λ 2 λ 2 ≈ 1 λ 2 .(5) [/formula]
In the realistic situation, it may be difficult to estimate the SNR by using the experimental data. In contrast, here we provide a new method that enables us to simply evaluate the SNR from the main/side peaks in the experimental ToA data. In the following calculation, we assume the pump power is very small so that λ 2 1 is satisfied. We can omit all the higher order terms in Eq. (1), and only consider the 0-, 1-, and 2-pair emissions. In the ToA data, the probability of the main peak (P main ) is proportional to the 1,1 click probability from total emission.
[formula] P main ≈ (1 − λ 2 )λ 2 η 2 + (1 − λ 2 )λ 4 (1 − (1 − η) 2 ) 2 = (1 − λ 2 )λ 2 η 2 [1 + λ 2 (2 − η) 2 ],(6) [/formula]
where we only consider the 1-and 2-pair emission. η is the overall detection efficiency for the signal and idler photons. Assuming the dead time of the detector (∼40 ns in our experiment) is longer than the peak-to-peak interval of the pump laser (∼13 ns in our experiment), the probability of the side peak (P side ) is proportional to the 1, 0 click probability at the main peak position (P 1,0m , 1-click in start channel, and 0-click in stop channel), multiplied by the 1 click probability at the side peak position (P 1s , 1-click in stop channel).
[formula] P 1,0m ≈ (1 − λ 2 )λ 2 η(1 − η) + (1 − λ 2 )λ 4 (1 − (1 − η) 2 )(1 − η) 2 = (1 − λ 2 )λ 2 η(1 − η)[1 + λ 2 (2 − η)(1 − η)].(7)P 1s ≈ (1 − λ 2 )λ 2 η + (1 − λ 2 )λ 4 (1 − (1 − η) 2 ) = (1 − λ 2 )λ 2 η[1 + λ 2 (2 − η)].(8)P side = P 1,0m × P 1s ≈ (1 − λ 2 ) 2 λ 4 η 2 (1 − η) × [1 + λ 2 (2 − η)][1 + λ 2 (2 − η)(1 − η)].(9) [/formula]
Therefore, we can evaluate [(main peak − side peak)/side peak] as main peak−side peak side peak
[formula] = Pmain−P side P side ≈ (1−λ 2 )λ 2 η 2 [1+λ 2 (2−η) 2 ]−(1−λ 2 ) 2 λ 4 η 2 (1−η)[1+λ 2 (2−η)][1+λ 2 (2−η)(1−η)] (1−λ 2 ) 2 λ 4 η 2 (1−η)[1+λ 2 (2−η)][1+λ 2 (2−η)(1−η)] ≈ (1−λ 2 )λ 2 η 2 −(1−λ 2 ) 2 λ 4 η 2 (1−η) (1−λ 2 ) 2 λ 4 η 2 (1−η) ≈ λ 2 η 2 −λ 4 η 2 (1−η) λ 4 η 2 (1−η) ≈ λ 2 η 2 λ 4 η 2 (1−η) = 1 λ 2 (1−η) ∝ 1 λ 2 .(10) [/formula]
In Eq. (10), the approximations were achieved by assuming λ 2 1. By comparing Eq. (5) and Eq. (10), we can learn that Eq.(5) can be used to approximate the SNR. Therefore, it is reasonable to calculate the SNR in a logarithmic scale as
[formula] SN R ≈ 10log 10 [(main peak − side peak)/side peak] ≈ 10log 10 [ 1 λ 2 (1 − η) ] ≈ 10log 10 [ 1 λ 2 ].(11) [/formula]
The last approximation is valid if η is sufficiently low.
In the SNR test in this experiment, at 30 mW pump power, the overall detection efficiencies (η) were estimated as 0.31 for 76 MHz laser and 0.30 for 2.5 GHz laser; the average photon numbers per pulse (p) were estimated as 0.0079 for 76 MHz laser and 0.00021 for 2.5 GHz laser; λ 2 = p/(1 + p) were estimated as 0.0078 for 76 MHz laser and 0.00021 for 2.5 GHz laser. Therefore, the condition of λ 2 1 is fully satisfied in the experiment. In Eq.(11), with η = 0.31,
[formula] 10log 10 [ 1 λ 2 (1−η) ] = 10log 10 [ 1 λ 2 ] + 10log 10 [ 1 (1−η) ] ≈ 10log 10 [ 1 λ 2 ] + 1.61. While 10log 10 [ 1 λ 2 ] ≈ 21 [/formula]
.08 for λ 2 = 0.0078.
## Supplementary-iii
In this part, we numerically analyze the relationship between photon-pair generation rate (i.e., average photon pair per pulse) and HOM interference visibility.
## The model
Here, we describe a numerical model of the HOM experiment. The model is described in [fig_ref] FIG. 3: Models [/fig_ref] (without delay) and (b) (with delay) where η A,B represent transmittances of mode A and B (losses are effectively described by beam splitters), respectively, and η D1 and η D2 are the detector efficiencies. The mode mismatch between the signal and idler pulses is directly reflected to the HOM interference visibility. In general, the signal and idler pulses occupy slightly different modes in frequency, time, or spatial degrees of freedom. This is phenomenologically modeled by introducing two virtual beam splitters with transmittance η M (which directly corresponds to the mode matching efficiency) that split the signal and idler into three modes, overlapped part (A and B) and unoverlapped parts occupied by the signal (E) and the idler (F) (see [fig_ref] FIG. 3: Models [/fig_ref]. The HOM interference visibility is defined as
[formula] V = CC mean − CC min CC mean ,(12) [/formula]
where CC min and CC mean are the coincidence count rates with zero-delay and large delay (i.e., with and without interference between the signal and idler), respectively. In the following we derive CC min and CC mean separately from our model. The initial state from the SPDC source is given by a two-mode squeezed-vacuum state
[formula] |ψ AB = 1 − λ 2 ∞ n=0 λ n |n A |n B ,(13) [/formula]
where λ is the squeezing parameter, and λ 2 /(1 − λ 2 ) = p is the average photon pairs per pulse. LetV η A(B) be a beam splitting operator on mode A(B) with transmittance η which transforms the photon number states |n 1 |n 2 aŝ
[formula] V η AB |n 1 |n 2 = 1 √ n 1 !n 2 ! n1 k1=0 n2 k2=0 n 1 k 1 n 2 k 2 (−1) k2 × √ η n2+k1−k2 1 − η n1−k1+k2 × (k 1 + k 2 )!(n 1 + n 2 − k 1 − k 2 )! × |k 1 + k 2 |n 1 + n 2 − k 1 − k 2 .(14) [/formula]
Applying the beam splittersV
[formula] η A AC ,V η B BD ,V η M AE ,V η M BF , andV 1/2 [/formula]
AB onto the two-mode squeezed vacuum (state at X in [fig_ref] FIG. 3: Models [/fig_ref] , we obtain
[formula] V 1/2 ABV η M BFV η M AEV η B BDV η A AC |ψ AB |0 C |0 D |0 E |0 F = 1 − λ 2 ∞ n=0 λ n n k1=0 n k 1 1/2 η k1/2 A (1 − η A ) n−k 1 2 n k2=0 n k 2 1/2 η k2/2 B (1 − η B ) n−k 2 2 × k1 k3=0 k 1 k 3 1/2 η k3/2 M (1 − η M ) k 1 −k 3 2 k2 k4=0 k 2 k 4 1/2 η k4/2 M (1 − η M ) k 2 −k 4 2 × k3 k5=0 k4 k6=0 k 3 k 5 k 4 k 6 1 2 k 3 +k 4 2 (−1) k6 (k 5 + k 6 )! (k 3 + k 4 − k 5 − k 6 )! k 3 ! k 4 ! 1/2 ×|k 5 + k 6 A |k 3 + k 4 − k 5 − k 6 B |n − k 1 C |n − k 2 D |k 1 − k 3 E |k 2 − k 4 F = 1 − λ 2 ∞ n=0 λ n n k1=0 n k 1 1/2 η k1/2 A (1 − η A ) n−k 1 2 n k2=0 n k 2 1/2 η k2/2 B (1 − η B ) n−k 2 2 × k1 k3=0 k 1 k 3 1/2 η k3/2 M (1 − η M ) k 1 −k 3 2 k2 k4=0 k 2 k 4 1/2 η k4/2 M (1 − η M ) k 2 −k 4 2 1 2 k3+k4 × k3+k4 l=0 min{l,k3} k5=max{0,l−k4} (−1) l−k5 k 3 k 5 k 4 l − k 5 l k 5 k 3 + k 4 − l k 3 − k 5 1/2 ×|l A |k 3 + k 4 − l B |n − k 1 C |n − k 2 D |k 1 − k 3 E |k 2 − k 4 F ,(15) [/formula]
where l = k 5 + k 6 and we have used the relation
[formula] k 3 k 5 k4 k 6 (k 5 + k 6 )! (k 3 + k 4 − k 5 − k 6 )! k 3 ! k 4 ! 1/2 = k 3 k 5 k4 k 6 k 5 + k 6 k 5 k 3 + k 4 − k 5 − k 6 k 3 − k 5 1/2 .(16) [/formula]
Note thatV 1/2 should be applied to mode E and F , which will be discussed later. From Eq. (15) we find the joint probability of having l, k 3 + k 4 − l, n − k 1 , n − k 2 , k 1 − k 3 , k 2 − k 4 photons in mode A-F at X:
[formula] P X ABCDEF (l, k 3 + k 4 − l, n − k 1 , n − k 2 , k 1 − k 3 , k 2 − k 4 ) = (1 − λ) 2 λ 2n η k1 A (1 − η A ) n−k1 η k2 B (1 − η B ) n−k2 η k3+k4 M (1 − η M ) k1+k2−k3−k4 1 2 k3+k4 n k 1 n k 2 k 1 k 3 k 2 k 4 × min{l,k3} k5=max{0,l−k4} (−1) l−k5 k 3 k 5 k 4 l − k 5 l k 5 k 3 + k 4 − l k 3 − k 5 1/2 2 .(17) [/formula]
The 50/50 beam splitting of mode E (F ) into E A and E B (F A and F B ) adds extra binomial distribution terms k1−k3 k7 k2−k4 k8 1 2 k1+k2−k3−k4 to Eq. (17). The joint probability distribution for the state at the detectors is thus given by
[formula] P ABCDE A F A E B F B (l, k 3 + k 4 − l, n − k 1 , n − k 2 , k 7 , k 2 − k 4 − k 8 , k 1 − k 3 − k 7 , k 8 ) = (1 − λ) 2 λ 2n η k1 A (1 − η A ) n−k1 η k2 B (1 − η B ) n−k2 η k3+k4 M (1 − η M ) k1+k2−k3−k4 1 2 k1+k2 n k 1 n k 2 k 1 k 3 k 2 k 4 × k 1 − k 3 k 7 k 2 − k 4 k 8 min{l,k3} k5=max{0,l−k4} (−1) l−k5 k 3 k 5 k 4 l − k 5 l k 5 k 3 + k 4 − l k 3 − k 5 1/2 2 .(18) [/formula]
The coincidence rate CC min is then obtained by the sum of the joint probability:
[formula] CC min = ∞ n=0 n k1=0 n k2=0 k1 k3=0 k2 k4=0 k1−k3 k7=0 k2−k4 k8=0 k3+k4 l=0 1 − (1 − η D1 ) l+k2−k4+k7−k8 1 − (1 − η D2 ) −l+k1+k4−k7+k8 ×P ABCDE A F A E B F B (l, k 3 + k 4 − l, n − k 1 , n − k 2 , k 7 , k 2 − k 4 − k 8 , k 1 − k 3 − k 7 , k 8 ).(19) [/formula]
The derivation of CC mean is rather simple since there is no interference at the 50/50 beam splitter due to the delay. This is illustrated in [fig_ref] FIG. 3: Models [/fig_ref]. Note that we do not need η M . The two-mode squeezed vacuum from the SPDC source has a joint photon distribution:
[formula] P AB (n, n) = (1 − λ 2 )λ 2n .(20) [/formula]
The beam splitting operation simply spread this distribution in a binomial manner. For example, after the beam splitter η A , the joint distribution is given by
[formula] P ABC (n, k 1 , n − k 1 ) = (1 − λ 2 )λ 2n n k 1 η k1 A (1 − η A ) n−k1 .(21) [/formula]
Applying the η B and 50/50 beam splitters in a similar way, we have
[formula] P AA BB CD (k 3 , k 2 − k 4 , k 4 , k 1 − k 3 , n − k 1 , n − k 2 ) = (1 − λ 2 )λ 2n n k 1 n k 2 k 1 k 3 k 2 k 4 η k1 A (1 − η A ) n−k1 η k2 B (1 − η B ) n−k2 1 2 k1+k2 ,(22) [/formula]
before the detectors. The coincidence count CC mean is then given by
[formula] CC mean = ∞ n=0 n k1=0 n k2=0 k1 k3=0 k2 k4=0 1 − (1 − η D1 ) k2+k3−k4 1 − (1 − η D2 ) k1−k3+k4 ×P AA BB CD (k 3 , k 2 − k 4 , k 4 , k 1 − k 3 , n − k 1 , n − k 2 ).(23) [/formula]
The HOM visibility in Eq. (12) is thus calculable from Eqs. (19) and (23).
# Numerical result
The transmittances (efficiencies) of each components in the experiment are summarized in [fig_ref] TABLE I: Transmittance and efficiency of the components in the experiment [/fig_ref] 3 for the theoretical model and the corresponding experimental setup in Main text. In fact, the HOM visibility is extremely sensitive to the mode matching factor η M . It is however not easy to estimate the mode matching factor η M experimentally with enough accuracy.
In , we plot the numerical results with various η M , and the experimental data with the 76 MHz laser. The experimental average photon-pair p is estimated from the experimental count rates. The experimental data fit the theoretical lines well within 0.9848 ≤ η M ≤ 0.9888. With the parameters in [fig_ref] TABLE I: Transmittance and efficiency of the components in the experiment [/fig_ref] , we also calculated the performance of our scheme at high photon-pair generation rate, as shown in and [fig_ref] TABLE I: Transmittance and efficiency of the components in the experiment [/fig_ref]. From this simulation, we find several interesting relationship. (1). The visibility is directly determined by the average photon-pairs. (2). The slope of this line is very sensitive to the unbalanced loss in the delay arm and non-delay arm. (3). The Y-intercept of this line very sensitive to the mode matching efficiency.
[fig] FIG. 2: Hong-Ou-Mandel dip for comb laser at 10 GHz repetition rate, fitted with triangle shape function. The pump power is 2 mW. [/fig]
[fig] FIG. 3: Models. (a) No delay (HOM dip). The mode mismatch at the 50/50 beam splitter is represented by the beam splitters ηM . (b) With delay. The delay is represnted by the spatial difference at the 50/50 beam splitter. [/fig]
[fig] FIG. 4 FIG. 5: The HOM visibility versus p. The solid lines represent theoretical curves with ηM = 0.9888, 0.9878, 0.9868, 0.9858, 0.9848 from the top to the bottom. The plots are the experimental results with the 76 MHz laser. The HOM visibilities at different p values, with ηM = 0.9878. [/fig]
[table] TABLE I: Transmittance and efficiency of the components in the experiment. SMFC: single mode fiber coupler. FC: fiber connector. SNSPD: superconducting nanowier single photon detector. [/table]
|
The Cumulative Risk of Chemical and Nonchemical Exposures on Birth Outcomes in Healthy Women: The Fetal Growth Study
Metals, stress, and sociodemographics are commonly studied separately for their effects on birth outcomes, yet often jointly contribute to adverse outcomes. This study analyzes two methods for measuring cumulative risk to understand how maternal chemical and nonchemical stressors may contribute to small for gestational age (SGA). SGA was calculated using sex-specific fetal growth curves for infants of pregnant mothers (n = 2562) enrolled in the National Institute of Child Health and Human Development (NICHD) Fetal Growth Study. The exposures (maternal lead, mercury, cadmium, Cohen's perceived stress, Edinburgh depression scores, race/ethnicity, income, and education) were grouped into three domains: metals, psychosocial stress, and sociodemographics. In Method 1 we created cumulative risk scores using tertiles. Method 2 employed weighted quantile sum (WQS) regression. For each method, logistic models were built with three exposure domains individually and race/ethnicity, adjusting for age, parity, pregnancy weight gain, and marital status. The adjusted effect of overall cumulative risk with three domains, was also modeled using each method. Sociodemographics was the only exposure associated with SGA in unadjusted models ((odds ratio) OR: 1.35, 95% (confidence interval) CI: 1.08, 1.68). The three cumulative variables in adjusted models were not significant individually, but the overall index was associated with SGA (OR: 1.17, 95% CI: 1.02, 1.35). In the WQS model, only the sociodemographics domain was significantly associated with SGA. Sociodemographics tended to be the strongest risk factor for SGA in both risk score and WQS models.
# Introduction
The Environmental Protection Agency (EPA) has defined cumulative risk as "the combined risks from aggregate exposures to multiple agents or stressors". While much of the scientific community, including the EPA, has advocated for the study of cumulative risk models , this is rarely done in epidemiological studies. When it is studied, researchers typically do not combine social and chemical exposures. Several studies have used various methodologies to study chemical mixtures, and some scientists have combined psychosocial stressors, but most researchers tend to construct separate models for chemical and nonchemical stressors. Recently, researchers have advocated for combining these exposures in order to have a more complete understanding of how exposures interact to affect outcomes. Although researchers have attempted to model the combined effects of chemical and nonchemical exposures to study lead and stress on cognition; benzene and stress on birthweight; and lead, cadmium, and polychlorinated biphenyls on blood pressure, methods to study combined chemical and nonchemical are still lacking.
Studying multiple chemical and nonchemical exposures is of special interest to researchers studying perinatal health. One area of concern are indicators for fetal growth restriction, such as small for gestational age (SGA), which have shown to have both short-term and long-term effects on infants as they grow into adulthood. Despite extensive research into the exposures contributing to SGA, understanding the exact etiology has proven difficult. Several studies reveal possible associations with SGA, including environmental toxicants, race, ethnicity, income, maternal educational attainment, and maternal mental health, but findings are often inconsistent and the biological rationale is not conclusively known. Studies are consistent in indicating that women who are non-Hispanic-black, low income, and have not attained degree past high schoolare more at risk for most adverse perinatal outcomes, including SGA.
Chemical exposures are also risk factors for reduced fetal growth. Heavy metals, which are common neurotoxins, have been linked to impeded fetal development at high levels of occupational exposure. In utero exposure to certain heavy metals, like lead (Pb), cadmium (Cd), and mercury (Hg), have been associated with low birthweight (LBW) and preterm birth (PTB), although there are many studies that do not find an increased risk between chemical exposure and adverse birth outcomes. This is similar to studies researching other chemicals. For example, researchers studying the effects of phthalates and pesticides on fetal development have reached inconclusive results. However, there is evidence that, while at low doses there are no detectable risks when studied individually, there could be a cumulative risk when these low doses are studied in combination. Animal models have indicated that if modeled cumulatively, these chemicals do in fact impact fetal development.
Beyond the possible link between environmental toxicants and birth outcomes, other researchers have focused on social stressors for their risk of adverse birth outcomes. Many point to the stress experienced during preconception and pregnancy, with evidence of an association between many indicators of stress and fetal growth. Women who report having depressive symptoms tend to have worse birth outcomes, as well as a higher risk for SGA. Although women who experience these psychosocial issues trend towards having adverse birth outcomes, like chemical exposures, literature reviews have reported some ambiguity; the evidence is far from conclusive that stress disorders directly cause adverse birth outcomes. Inconclusive findings could be the result of inadequate methods for studying stress, but it could also be a result of the type of populations that are studied. When studying low risk, healthy pregnant women, Voegtline et al. found that stress indicators, such as depressive symptoms, trait anxiety, and emotional well-being, tended to be non-significantly correlated with cortisol measured throughout pregnancy.
Further complicating perinatal research is the possible interaction between chemical and nonchemical exposures. Common socioeconomic exposures, like race/ethnicity, income, and education, are correlated with each other. African Americans, for example, have higher incidences of PTB, impaired fetal growth, and maternal mortality, with studies identifying stress induced by racism as an underlying factor. Levels of cortisol are higher among women of lower socioeconomic status (SES), with black women tending to report higher exposure to stress compared to other races. Race and ethnicity are commonly multi-correlated with other social factorsand those that are disadvantaged are at additional risk to be 'vulnerable' to environmental exposures and psychosocial stress. DeFur et al. theorized that vulnerability, which is typically a result of low SES, interacts with physical and social environments and results in disparities in health outcomes. This theory is supported by numerous studies finding that lower-SES populations most commonly live in areas with higher environmental exposures, including proximity to pollution, industrial areas, and other areas containing large amounts of environmental toxicants, and are thus more vulnerable to be exposed to psychosocial stress as a result of these circumstances. SES may also be an indicator of different dietary intakes or occupational exposures that may put people of lower SES at greater risk for environmental toxicants.
Due to the multi-correlated nature of risk factors contributing to fetal growth, perinatal research could benefit from researching exposures cumulatively. This is especially true when studying relatively healthy women, as healthy women tend to be exposed to toxicants at levels lower than the regulatory concerned and associations of individual chemical and nonchemical exposures with birth outcomes are typically only evident at high levels of exposure. However, just as in chemical exposures that may only show risk at low doses with multiple exposures, there could be associations between the nonchemical and chemical risk factors and perinatal outcomes, if a combination of lower levels of exposure are measured cumulatively. These risks also may interact with each other, causing a high degree of multicollinearity. We have included a conceptual model to explain the hypothesized interrelationships of possible chemical and nonchemical stressors. The objective of this study is to understand whether cumulative risk, indicated by the combination of exposure to heavy metals, psychosocial stress, and sociodemographic risk, contributes to incidence of SGA infants in a healthy cohort of pregnant women. We compare two possible ways of assessing this cumulative risk, one a relatively simple risk score that weights exposures equally, and one which allows for varying weights, weighted quantile sums. toxicants, and are thus more vulnerable to be exposed to psychosocial stress as a result of these circumstances. SES may also be an indicator of different dietary intakes or occupational exposures that may put people of lower SES at greater risk for environmental toxicants. Due to the multi-correlated nature of risk factors contributing to fetal growth, perinatal research could benefit from researching exposures cumulatively. This is especially true when studying relatively healthy women, as healthy women tend to be exposed to toxicants at levels lower than the regulatory concerned and associations of individual chemical and nonchemical exposures with birth outcomes are typically only evident at high levels of exposure. However, just as in chemical exposures that may only show risk at low doses with multiple exposures, there could be associations between the nonchemical and chemical risk factors and perinatal outcomes, if a combination of lower levels of exposure are measured cumulatively. These risks also may interact with each other, causing a high degree of multicollinearity. We have included a conceptual model to explain the hypothesized interrelationships of possible chemical and nonchemical stressors. The objective of this study is to understand whether cumulative risk, indicated by the combination of exposure to heavy metals, psychosocial stress, and sociodemographic risk, contributes to incidence of SGA infants in a healthy cohort of pregnant women. We compare two possible ways of assessing this cumulative risk, one a relatively simple risk score that weights exposures equally, and one which allows for varying weights, weighted quantile sums.
# Material and methods
## Study participants
We conducted a secondary analysis on the data collected through the Fetal Growth Study, a cohort study conducted on pregnant women through the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) with the primary goal of calculating a national standard for fetal growth and size for gestational age. Secondary goals were to study fetal growth in women with gestational diabetes mellitus (GDM) among the n = 468 obese participants, as well as measuring fetal growth in twin gestations. The study was conducted between July 2009 and January
# Material and methods
## Study participants
We conducted a secondary analysis on the data collected through the Fetal Growth Study, a cohort study conducted on pregnant women through the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) with the primary goal of calculating a national standard for fetal growth and size for gestational age. Secondary goals were to study fetal growth in women with gestational diabetes mellitus (GDM) among the n = 468 obese participants, as well as measuring fetal growth in twin gestations. The study was conducted between . The data are comprised of 3270 low-risk pregnant women with singleton gestation, with 468 of them classified as obese. Women were recruited within their first trimester, had a singleton gestation, did not consume alcohol, were nonsmokers, and were not suspected to have a fetus with a fetal abnormality. Alcohol and smoking habits were assessed upon recruitment in a questionnaire which asked participants the amount of alcohol they consumed the amount they smoked within 3 months prior to pregnancy and current consumption and smoking status. The dataset includes six visits: one initial recruitment visit and five additional follow-up visits. Participants were randomized into four groups (A, B, C, or D) with each group having varying schedules of follow-up visits, depending on gestational age. The first follow-up visit for group A was scheduled for 15 to 17 weeks gestation; in B, for 17 to 18 weeks gestation; in C, for 19 to 21 weeks gestation; and in D, for 21 to 23 weeks gestation. A detailed account of the cohort recruitment, including exclusion criteria, has been previously published. For the purposes of this analysis, only women with a singleton gestation and who had a live birth were included in this study. After accounting for missing data, the complete dataset consisted of n = 1569 women. Women were recruited within their first trimester, had a singleton gestation, did not consume alcohol, were nonsmokers, and were not suspected to have a fetus with a fetal abnormality. Alcohol and smoking habits were assessed upon recruitment in a questionnaire which asked participants the amount of alcohol they consumed the amount they smoked within 3 months prior to pregnancy and current consumption and smoking status. The dataset includes six visits: one initial recruitment visit and five additional follow-up visits. Participants were randomized into four groups (A, B, C, or D) with each group having varying schedules of follow-up visits, depending on gestational age. The first followup visit for group A was scheduled for 15 to 17 weeks gestation; in B, for 17 to 18 weeks gestation; in C, for 19 to 21 weeks gestation; and in D, for 21 to 23 weeks gestation. A detailed account of the cohort recruitment, including exclusion criteria, has been previously published. For the purposes of this analysis, only women with a singleton gestation and who had a live birth were included in this study. After accounting for missing data, the complete dataset consisted of n = 1569 women.
## Variables
Education, race/ethnicity, and income were assessed upon recruitment. Women were asked their highest degree obtained and whether they were white, black, American Indian, an Alaskan Native, or Asian. Women were additionally asked whether they were Hispanic. Women reported their family income from the previous year in 10 categorical groupings, which we collapsed into 4 categories for better power: (1) ≤$29,999, (2) $30-$49,999, (3) $50-$74,999, (4) ≥$75,000. We categorized education into high school or less, some college or an associate's degree, and completed college. Race/ethnicity was categorized as black, white, Hispanic, or Asian. We collapsed the categories as black or non-black
## Variables
Education, race/ethnicity, and income were assessed upon recruitment. Women were asked their highest degree obtained and whether they were white, black, American Indian, an Alaskan Native, or Asian. Women were additionally asked whether they were Hispanic. Women reported their family income from the previous year in 10 categorical groupings, which we collapsed into 4 categories for better power: (1) ≤$29,999, (2) $30-$49,999, (3) $50-$74,999, (4) ≥$75,000. We categorized education into high school or less, some college or an associate's degree, and completed college. Race/ethnicity was categorized as black, white, Hispanic, or Asian. We collapsed the categories as black or non-black (non-black participants include white, Hispanic, and Asian) in the final analysis, because black women tended to be more at risk for adverse outcomes, while risk was similar in the other groups.
Information on depression and perceived stress was collected during the first and second trimesters across a total of seven study visits (recruitment and six follow-up visits); however, we only included information from the first 5 visits due to missing psychosocial information after five study visits. Each depression score and perceived stress score was averaged across study visits. Perceived life stress was assessed using the Cohen Perceived Stress scale. The 10-item questionnaire captures perceived stress in daily life, where the top tertile used as an indicator of high stress. The Edinburgh Postnatal Depression Index (EPDS) was used to assess depression, with a cutoff value of 12 used to indicate depression. Birthweight, gestational age, and sex of the infant were assessed through birth record abstraction at the participating hospitals.
The outcome variable was SGA. Infants were categorized as being SGA if their weight fell below the 10th percentile for their recorded gender and gestational age, in accordance with published national standards.
Women were dichotomized as being married/cohabitating with a partner or single. Parity was categorized into nulliparous, one previous birth, or two or more births. Age was categorized to 24 years old or less, 25-34 years old, and 35 or over based on the risk distribution within the age categories. We used the guidelines published by the Institute of Medicine (IOM)to determine whether women's weight gain during the entire pregnancy was adequate, under the recommended amount, or over the recommended amount according to reported pre-pregnancy BMI.
## Specimen collection and analysis
For this study, we analyzed data from blood samples collected during recruitment (8-12 weeks gestation). Obese women provided 30 mL non-fasting blood samples while women from other BMI groups provided 20 mL samples, which were sent to the Trace Elements Section of the Laboratory of Inorganic and Nuclear Chemistry at the Wadsworth Center, New York State Department of Health (Albany, NY, USA) where they were stored at −80 - C. Whole blood Pb, Cd, and Hg were determined by coupled plasma-mass spectrometry (ICP-MS). Methods to quantitate heavy metals were previously validated. In total, 2063 participants provided blood samples upon recruitment, which were then analyzed for Pb, Cd, and Hg. Seven percent of participants who provided blood samples (n = 148) had amounts less than the limit of detection (LOD) for Pb; this amount was 1% (n = 23) for Hg and 8% (n = 171) for Cd. Values that were below the LOD were included during analysis in order to prevent introducing biasusing the values that were originally detected, consistent with other studies using the same laboratory analysis methods.
# Statistical analysis
Cumulative models were analyzed in two ways: (1) by creating a cumulative risk score, using tertiles of the exposure and (2) by constructing weighted quantile sum (WQS) regression models. Cumulative risk scores were created by first dividing each predictor into tertiles; the highest tertile was considered exposed (x = 1) while the 2 lowest tertiles were unexposed (x = 0). We used tertiles given their precedent in studies of chemical exposures. In a sensitivity analysis using quartiles and quintiles we found that results did not vary widely from the tertile method; however, using tertiles avoided issues with model convergence. In the first method, we divided each of the continuous exposures (Pb, Cd, Hg, EPDS scores, and perceived stress scores) into tertiles, with the highest tertile considered exposed. The ordinal predictors (income levels, and educational attainment) were considered exposed if income was <$30,000 and if women had attained less than a high school education. We created indices for each of the three domains: (1) metals, (2) psychosocial stress, and (3) sociodemographics. Each domain's index was equal to the sum of the variables within that domain. The total cumulative domain was also created, consisting of the sum of all three domains. We created an unadjusted regression model for each of the four domains individually (metals, psychosocial, sociodemographics and total cumulative), as well as an additional unadjusted logistic regression model that assessed metals, psychosocial stress, and sociodemographic in a single model. We then created adjusted models looking at the three domains models separately in three different models. Lastly, we created two fully adjusted models. The first included Metals Psychosocial stress, and sociodemographics adjusted for parity, marital status, weight gain during pregnancy, and age. The last model included the total cumulative domain adjusted for parity, marital status, weight gain during pregnancy, and age.
We compared the above models to those developed using the weighted quantile sum (WQS) regression method developed by Carrico et al.. WQS regression allows researchers to create an index of correlated predictors that are weighted according to their strength of association with the outcome. Weaker variables are zeroed out in the WQS index. The method accounts for highly correlated exposures and was developed in the context of multiple chemical exposures, such as phthalates.
Specifically, the equation presented by Carrico et al. seeks to calculate the weights of c set of correlated variables:
[formula] g(µ) = β 0 + β 1 c i=0 w i q i + z ϕ [/formula]
The sum term is the index for the c items, and weights are represented by the sum of w i . Each w i is constrained to a value between 0 and 1. All confounders are represented by z ϕ. Prior to analysis, the data is split into two datasets at random: a training dataset and the validation dataset. Using the training dataset, bootstrap samples are selected and the strength of the associations for each c item is determined by the beta coefficient. The index is calculated based on the mean w i s across all bootstrap samples. Weights are estimated based on optimization algorithms employed to maximize the likelihood in a nonlinear model. Variables with more influence within the quantile were assigned higher weights. Variables are selected based on a previously set significance threshold (we used p = 0.10).
Four models were created, one for each domain (metals, psychosocial, and sociodemographic) and one for the total cumulative index that included all domains. For each domain, we created weights adjusting for race/ethnicity and confounders (weight gain, parity, marital status, and age). The test dataset consisted of a random sample of 40% of the observations with the results validated in the remaining 60% of the sample population. Weights in the unadjusted and adjusted model (each of the domains modeled separately with race/ethnicity and confounders) were noted, as well as the final betas (β) for each of the models. The following equations represent the WQS models used:
[formula] logit (SGA) = β 0 + β Metals x Metals + ε logit (SGA) = β 0 + β Psychosocial x Psychosocial + ε logit (SGA) = β 0 + β Demographics x Demographics + ε logit (SGA) = β 0 + β Total x Total + ε. [/formula]
logit (SGA) = β 0 + β Total x Total + β Race x Race + β Age x Age + ε SAS software 9.4 (SAS Institute Inc., Cary, NC, USA) was used to generate the descriptive and cumulative risk scores. R software was used to find weights and estimates in WQS regression analysis. All participating sites received human subjects' approval and participants gave their informed consent before data was collected. As a secondary analysis of de-identified data, this analysis was not considered human-subjects research.
# Results
There were 2038 participants that met the inclusion criteria and provided a blood specimen. Women were divided evenly between non-Hispanic white, non-Hispanic black, and Hispanic (27.3%, 26.1%, and 27.4% respectively). The majority of women either earned <$30,000 (n = 489, 27.8%) or ≥$100,000 (n = 521, 29.6%). Only 3% of women (n = 58) scored high enough on the EPDS to be categorized as depressed. Seven percent of infants (n = 150) were SGA. A description of continuous variables is presented in. About 7% (n = 148) of women had Pb levels below the limit of detection. Only 1% (n = 23) of women had Hg levels below the limit of detection and 8% (n = 171) of women had Cd levels below the limit of detection. The average Pb blood level was 0.51 micrograms per deciliter (µg/dL) and only 1% (n = 21) of the study population had a Pb blood level of greater than 5 µg/dL. Only one participant had an excessive level of Cd, measured at 5 µg/L, and one participant had blood Hg content (>5 µg/L).
Most variables were individually associated with SGA, with the exception of depression (OR: 0.32, 95% CI: 0.08, 1.33) and Hg (OR: 0.91, 95% CI: 0.78, 1.45). Non-Hispanic black women were at a higher risk of having SGA infants compared to non-black women (OR: 2.46, 95% CI: 1.59, 3.81). Higher levels of Cd and Pb also were associated with having an SGA infant (OR: 1.18, 95% CI: 1.04, 1.35 and OR: 1.22, 95% CI: 1.03, 1.35). Higher income was protective against SGA when compared to lower income (less than $30,000) and attaining higher levels of education was also protective against SGA. Women with previous pregnancies, married women, and women who were older than 25 were less likely to have SGA infants compared to nulliparous, nonmarried, and younger women respectively. Variables within domains tended to be correlated with each other; however, Hg and Pb were not significantly correlated. Though many of the variables were statistically significantly correlated, education and income appeared to be the only variables highly correlated to each other (r = 0.68, p < 0.05). Depression and perceived stress were moderately correlated (r = 0.31, p < 0.05).
Only the two sociodemographic variables (education and income) were correlated with SGA (not included in table). In unadjusted cumulative models, the demographic index and the total cumulative index were the only variables associated with SGA. The total cumulative indexwas associated with SGA even with race/ethnicity included in the model (OR: 1.17, 95% CI: 1.02, 1.35). While not significantly associated with SGA, the unadjusted metals domain and the psychosocial domain appeared to have an association, given that the lower confidence level was close to the null (OR 1.16, 95% CI 0.97, 1.38 and OR: 1.20, 95% CI 0.96, 1.52 respectively). In the adjusted models, inputting the domains separately did not yield any associations between the indices or race/ethnicity and SGA. The weights for the WQS regression models and the betas are described in. The sociodemographic domain was the only domain associated with SGA (β = 0.41, SE = 0.20, p = 0.04). According to the weights assigned in the WQS regression models, income appears to have the most influence on the index, as the weight for income was 0.69, while it was only 0.31 for education. Weights for adjusted and unadjusted WQS models, betas (β) and standard errors (SE) for final adjusted WQS regression models n WQS models, Fetal Growth Study.
# Discussion
This analysis indicates that a cumulative measure of multiple exposures could be associated with SGA. The index that included all three domains in the cumulative risk score analysis (heavy metals, psychosocial stress, and sociodemographics) was consistently associated with SGA, suggesting that modeling exposures cumulatively could help explain how lower levels of exposures jointly increase women's risk of these adverse birth outcomes. However, in WQS regression models, the total cumulative model was not associated with SGA, and the only domain to be associated with the outcome was the sociodemographic index. Sociodemographics and race/ethnicity were so strongly associated with the outcome they outweighed effects of other exposures. Previous studies have indicated that controlling for indicators of socio-economic levels could control for other effects. This results in only the strongest SES terms being associated with the outcome, while producing null results among other indicators. In Lefmann et al's study, for example, they believed that the reason why Medicaid status was the only predictor associated with SGA, LBW, and PTB was because Medicaid controlled for other correlated indicators, such as race and experiencing stressful events. They postulated that Medicaid status could be an indicator for poverty, thereby negating associations between other sociodemographic and stress variables in the models. A similar process could be operating with respect to exposures to toxicants; certain lower-SES populations are more vulnerable to environmental exposures. Several studies have indicated disadvantaged populations are at higher risk of exposure to both chemical and nonchemical exposures. DeFur et al., described the vulnerability in lower SES groups, meaning not just a susceptibility of lower SES groups to harmful exposures, but also an inability of these groups to recover. That results in disparities in outcomes between SES groups. DeFur et al.'s conclusion is plausible given that lower SES people tend to live in closer proximity to industrial areas and other areas with higher chemical exposures. However, depending on the population, other potential sources of pollutant exposure (diet, consumer products, and occupation) may not have the same correlation with SES: for instance, jobs with relatively high levels of exposure may also be high-paying, or women with higher incomes may be able to afford seafood that contains mercury. While the study does not allow us to determine why lower SES women had a higher risk of having SGA infants, the analyses do provide more evidence on the lasting and overwhelming impacts of lower income and educational attainment on perinatal outcomes. WQS regression's ability to weight predictors may be of special interest to researchers who study multiple correlated predictors, as it avoids a simplistic assumption of equal weighting provided by a simple cumulative risk score, such as our initial analysis provided. It should be noted, however, that WQS still assumes an equal increase/decrease in risk with every one-quantile change, which may be incorrect if an association is nonlinear. We found no other studies that used the WQS method to estimate the joint effect of both chemical and nonchemical exposures. Since its development, the approach has been primarily used to combine multiple chemical exposures, although more recently, the approach has been used to create SES indices, nutrient indices from self-reported food frequency questionnaires, and an index for psychosocial stress. Though the authors of these studies chose to build models that did not combine chemical and nonchemical exposures, some authors did attempt to build WQS coefficients that combined several domains. Yorita, Christensen et al. created three WQS indices (heavy metals, co-planar polychlorinated biphenyls , and non-dioxin-like PCBs) to predict alanine amino transferase (ALT) levels (which indicate liver damage). Czarnota et al. used one WQS index to estimate the combined effect of PCBs, polycyclic aromatic hydrocarbons (PAH), and pesticides. Risk factors across different domains are commonly correlated with each other, so using an approach that accounts for this correlation could be helpful for researchers in many fields of epidemiology. However, if exposure variables are strongly correlated with each other and weakly correlated with the outcome WQS will perform poorly.
# Limitations
We conducted this analysis on a cohort of pregnant women with no previous history of very PTB or LBW infants, as well as no other pregnancy complications. Results cannot be generalized to other populations of women and are not representative of the entire population of the United States. A review of literature reveals that scientists are also currently unaware of the amounts of heavy metals that may cause adverse birth outcomes. While most women did not have blood samples that were higher than the CDC recommended limits, there is still evidence that lower levels of exposure could have adverse effects on perinatal outcomes (which was one of the objectives of this study). As no clear cut-off point was available from the literature, we used the highest tertile to classify exposure.
The primary intention of this study was to understand whether cumulative risk, of chemical and nonchemical exposures, was associated with SGA in a healthy population of women, which included women who were obese but otherwise healthy. Investigating women who do not present prior health issues results in the exclusion of higher-risk women. If the examined risk factors are associated with intermediates, selection bias could exist; regardless, this fact limits generalizability. Studying the effects in more-exposed and/or less-healthy populations is an important next step in this research. While we chose variables that have effects on fetal growth, there may have been other variables that could be more appropriate for this analysis that were not measured for this study. Other chemical exposures, such as PCBs, air pollution, or other environmental pollutants, may interact with heavy metals, but cannot be studied in this secondary data analysis. We were similarly limited to the psychosocial stress measures studied in this analysis; we were unable to study other indicators of stress, such as childhood traumatic events or post-traumatic stress disorder (PTSD), which have also been associated with adverse birth outcomes. Each factor studied is affected by a degree of measurement error; some of them, such as stress and metal levels, may change over time as well. (Age is likely to be a susceptibility factor, but thoroughly examining this interaction would require multiple measures over time.) Combining multiple imperfectly measured variables can attenuate effects under some circumstances. As methods for combined exposure analysis develop, the degree to which errors are amplified will need to be assessed. The study also suffers from missing data, as illustrated in. One limitation of WQS, or any assessment studying cumulative risk, is that the many exposure variables included in the study will inevitably lead to more missing data compared to traditional studies, which only assess one exposure variable with one outcome. To the extent that missing data is informative or associated with exposure and outcome, this may bias the sample; most likely, women with complete data will be at lower risk than the overall cohort.
# Conclusions
This study demonstrates that risk factors may affect SGA cumulatively and introduces WQS as a possible tool to measure cumulative risk. WQS has been presented as a possible tool to measure risks in perinatal research, as well as presenting the possibility of using the method in studies involving chemical and nonchemical risk factors. The present analysis also supports previous research on the importance of socioeconomic factors on reproductive health outcomes. When researchers study cumulative risk, special attention should be placed on acknowledging the important role of sociodemographics on maternal health. |
The impact of BCG vaccination on tuberculin skin test responses in children is age dependent: evidence to be considered when screening children for tuberculosis infection
Background Following exposure to TB, contacts are screened to target preventive treatment at those at high risk of developing TB. The UK has recently revised its recommendations for screening and now advises a 5 mm tuberculin skin test (TST) cut-off irrespective of age or BCG status. We sought to evaluate the impact of BCG on TST responses in UK children exposed to TB and the performance of different TST cut-offs to predict interferon γ release assay (IGRA) positivity. Methods Children <15 years old were recruited from 11 sites in the UK between January 2011 and December 2014 if exposed in their home to a source case with sputum smear or culture positive TB. Demographic details were collected and TST and IGRA undertaken. The impact of BCG vaccination on TST positivity was evaluated in IGRA-negative children, as was the performance of different TST cut-offs to predict IGRA positivity. Results Of 422 children recruited (median age 69 months; IQR: 32-113 months), 300 (71%) had been vaccinated with BCG. BCG vaccination affected the TST response in IGRA-negative children less than 5 years old but not in older children. A 5 mm TST cut-off demonstrated good sensitivity and specificity in BCG-unvaccinated children, and an excellent negative predictive value but was associated with low specificity (62.7%; 95% CI 56.1% to 69.0%) in BCG-vaccinated children. For BCG-vaccinated children, a 10 mm cut-off provided a high negative predictive value (97.7%; 95% CI 94.2% to 99.4%) with the positive predictive value increasing with increasing age of the child. Discussion BCG vaccination had little impact on TST size in children over 5 years of age. The revised TST cut-off recommended in the recent revision to the UK TB guidelines demonstrates good sensitivity but is associated with impaired specificity in BCG-vaccinated children.BACKGROUND
ABSTRACT Background Following exposure to TB, contacts are screened to target preventive treatment at those at high risk of developing TB. The UK has recently revised its recommendations for screening and now advises a 5 mm tuberculin skin test (TST) cut-off irrespective of age or BCG status. We sought to evaluate the impact of BCG on TST responses in UK children exposed to TB and the performance of different TST cut-offs to predict interferon γ release assay (IGRA) positivity. Methods Children <15 years old were recruited from 11 sites in the UK between January 2011 and December 2014 if exposed in their home to a source case with sputum smear or culture positive TB. Demographic details were collected and TST and IGRA undertaken. The impact of BCG vaccination on TST positivity was evaluated in IGRA-negative children, as was the performance of different TST cut-offs to predict IGRA positivity. Results Of 422 children recruited (median age 69 months; IQR: 32-113 months), 300 (71%) had been vaccinated with BCG. BCG vaccination affected the TST response in IGRA-negative children less than 5 years old but not in older children. A 5 mm TST cut-off demonstrated good sensitivity and specificity in BCG-unvaccinated children, and an excellent negative predictive value but was associated with low specificity (62.7%; 95% CI 56.1% to 69.0%) in BCG-vaccinated children. For BCG-vaccinated children, a 10 mm cut-off provided a high negative predictive value (97.7%; 95% CI 94.2% to 99.4%) with the positive predictive value increasing with increasing age of the child. Discussion BCG vaccination had little impact on TST size in children over 5 years of age. The revised TST cut-off recommended in the recent revision to the UK TB guidelines demonstrates good sensitivity but is associated with impaired specificity in BCG-vaccinated children.
# Background
Following household exposure to an infectious adult with pulmonary TB, up to half of all children living with this source case will themselves have evidence of sensitisation to Mycobacterium tuberculosis. This is usually manifest as a positive immune test in either the skin or blood. 1 TB infection implies that organisms are dormant within the body and are not causing disease; the child is clinically well with a normal chest radiograph. In the majority of children, disease never develops but over time, in a proportion, the mycobacteria will overcome the immune-mediated containment and the child will become ill with TB disease. Certain risk factors increase the likelihood of progressing from asymptomatic TB infection to TB disease. These include young age and suppression of the immune system. Over 50 years ago it was demonstrated that by providing drug therapy to individuals with evidence of TB infection, the risk of progression to TB disease could be significantly reduced. [bib_ref] Controlled chemoprophylaxis trials in tuberculosis. A general review, Ferebee [/bib_ref] The use of daily isoniazid, given for at least 6 months, has subsequently been demonstrated in a large number of studies to be effective in reducing the risk of disease progression in those with TB infection for HIV-positive 7 and HIV-negative contacts. [bib_ref] Isoniazid for preventing tuberculosis in non-HIV infected persons, Smieja [/bib_ref] The Why read on?
▸ The revised UK national guidance of using a 5 mm tuberculin skin test cut-off results in impaired specificity in BCG-vaccinated children.
diagnostic tests rely on demonstrating evidence of the presence (or absence) of prior immunological sensitisation. Traditionally, the tuberculin skin test (TST) has been used for this purpose. In the TST, a crude mixture of mycobacterial antigens is injected into the dermis of the skin in the forearm and the presence of a skin reaction at the site of injection 'read' at 48 h for evidence of a type IV delayed hypersensitivity reaction, signifying immunological memory to antigens within the tuberculin. These antigens, however, are not unique to M. tuberculosis; some are also found in BCG and environmental, non-tuberculous mycobacteria. [bib_ref] Use of the tuberculin skin test in children, Enarson [/bib_ref] Newer tests, such as the interferon γ release assays (IGRAs), measure either the quantity of interferon γ released by T cells or the number of T cells which release interferon γ, after stimulation by M. tuberculosis-specific antigens. Consequently, they have been found to have superior specificity. [bib_ref] The utility of an interferon gamma release assay for diagnosis of latent..., Machingaidze [/bib_ref] Until 2005, it was advised that all children in the UK be given BCG between the ages of 10 and 14 years. In addition, targeted vaccination, to be given soon after birth, was advised for children deemed at increased risk of TB exposure. This included children living in areas of high TB prevalence and those born within families from high incidence countries.From 2005, universal BCG vaccination was no longer recommended; targeted neonatal vaccination continued. In 2006 the National Institute for Health and Care Excellence (NICE) published guidance on the management of TB contacts.NICE suggested a stepwise screening strategy for TB infection in childhood contacts, using combinations of TST and IGRA with different algorithms for children of different ages, resulting in significant complexities in management. The 2006 NICE guidance also suggested using different TST cut-offs for children who had previously been vaccinated with BCG (15 mm induration) and those who had not been (6 mm induration). The rationale for this was that a prior BCG immunisation might lead to falsepositive results and impaired specificity of the TST. However, it was felt that the size of the induration would differentiate between BCG and TB infection-the larger the induration, the greater the probability that the response was due to TB infection. [bib_ref] Tuberculin reactivity and the risk of tuberculosis: a review, Watkins [/bib_ref] It may be that part of this recommendation is a legacy from a time when BCG was given to adolescents, where BCG would have a significant impact on TST response. Many clinicians were, however, concerned that a response of less than 15 mm in a BCG-vaccinated child might still indicate TB infection, and should not be ignored. International guidelines recommend a cut-off of 10 mm induration for a positive TST result regardless of BCG status. 14 In early 2016, NICE updated its guidance and suggested that a TST cut-off of 5 mm be used irrespective of BCG vaccination status in all ages of children.Prior to the advent of IGRA tests, it was not possible to evaluate the impact of a previous BCG vaccination on the TST response in TB contacts as it was not possible to determine whether a TST response was due to the prior BCG vaccination rather than TB infection. However, by evaluating TST responses in child contacts of TB cases shown to be IGRA-negative, the contribution of BCG can now be more precisely documented. While fully acknowledging that no gold standard for TB infection exists, we felt that a negative IGRA test result would act as a proxy measure for lack of sensitisation to M. tuberculosis. We therefore set out to answer two questions. First, to determine the influence of prior BCG vaccination on TST responses in IGRA-negative children living in the UK who were being screened because of household exposure to an infectious TB case. Second, to evaluate the sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of different TST cut-offs to predict IGRA positivity in BCG vaccinated and unvaccinated children of different ages.
# Methods
## Study setting
This investigation was carried out within a larger study, the NIHR-funded IGRA Kids Study (NIKS), which sought to evaluate the NPV of IGRA to predict incident TB disease in children exposed to TB in the UK. Sites included five paediatric TB clinics in London, together with paediatric TB clinics in Southampton, Bristol, Birmingham, Manchester, Glasgow and Newcastle. All children (<15 years) presenting to one of the 11 participating centres between 1 January 2011 and 31 December 2014 were eligible for recruitment if they had a history of household exposure to a source case with pulmonary sputum smear or culture positive TB.
## Study procedures and definitions
Following the diagnosis of a source case of infectious pulmonary TB, household contacts were identified and screened. Screening aimed to identify contacts already unwell with TB disease and asymptomatic individuals eligible for treatment of TB infection. Families were invited to participate in the study and written informed consent was obtained. Screening for TB disease in children included history, examination, chest radiography, TST and IGRA tests, and microbiology if indicated. The evaluation of contacts to decide on the provision of treatment for TB infection was based on NICE guidelines, in conjunction with clinical context. Children with TB infection were treated with isoniazid and rifampicin daily for 3 months.
TST and IGRA were evaluated at baseline. TST was placed by experienced members of the local TB nursing teams, by injecting two tuberculin units intradermally ( purified protein derivative RT23, Statens Serum Institute) with results read at 48-72 h. IGRA tests (either QuantiFERON-TB Gold In-Tube (Cellestis Ltd) or T-SPOT.TB (Oxford Immunotec Ltd) depending on the practice of the recruiting clinic) were carried out by the clinically available laboratory services, following the manufacturer's specifications. The IGRA test was repeated in all children after 2 months. If the TST was negative at baseline, it was also repeated after 2 months. TST was defined as being positive if the transverse diameter of the induration was ≥6 mm in BCG-unvaccinated children and ≥15 mm in vaccinated children, in line with the 2006 NICE recommendations. The largest TST measurement was used for analysis. The child was classified as IGRA positive (and hence assumed to be TB infected for the purposes of this analysis) if either the baseline or the 2-month IGRA was positive. Repeatedly indeterminate results were considered negative. Children were examined for the presence of a BCG scar; if present, children were classified as BCG vaccinated. If no scar was seen, but there was clear documentation in paper or electronic records, or if the parents gave a clear history of vaccination, the child was classified as BCG vaccinated. Otherwise the child was classified as BCG unvaccinated.
# Statistical analysis
Data were entered into a tailor-made online database in real time and later checked centrally for entry errors. Data were analysed using STATA software (V.11; Stata Corp, College Station, Texas, USA). For analysis, children were divided into four age groups: <2 years, 2 to <5 years, 5 to <10 years and 10 to <15 years. The Mann-Whitney test was used to assess differences between TST measurements in BCG vaccinated and unvaccinated children due to the non-normal distribution of the data. ORs and 95% CIs were calculated to determine the impact of BCG vaccination on TST positivity, in IGRA-negative children, using three TST cut-offs (5, 10 and 15 mm). Sensitivity, specificity, PPV, NPV, together with receiver operator characteristic (ROC) curves and their respective area under the curves were determined for different TST cut-offs to predict IGRA positivity. 95% CIs were determined for each of the above.
# Results
Of 422 children recruited, 216 (51%) were boys; median age was 69 months (IQR: 32-113). Of 370 children tested for HIV, none tested positive and no children in the study were known to be HIV positive. The majority of children (361; 86%) were born in the UK. Three hundred (71%) had been vaccinated with BCG [fig_ref] Table 1: Demographic and clinical characteristics of children in the study [/fig_ref].
In IGRA-negative children <2 years, there was a difference in TST size between those who were BCG vaccinated and those who were unvaccinated (median: 4 mm (IQR: 0-12) vs median: 0 mm (IQR: 0-0); p<0.001; table 2). There was also a difference for children aged 2 years to <5 years ( p=0.001) but no significant difference was seen in children aged 5 to <10 years ( p=0.12) or children aged 10 to <15 years ( p=0.09) [fig_ref] Table 2: Median TST induration in BCG-vaccinated and BCG-unvaccinated children at different ages who... [/fig_ref]. For IGRA-negative children aged <2 years and 2 to <5 years, those who were BCG vaccinated were more likely to have a positive TST result than those who were unvaccinated, at all three TST cut-offs. For children aged 5 to <10 years and 10 to <15 years there was a smaller and less significant difference in TST positivity between children who were BCG vaccinated and those who were unvaccinated, at all TST cut-offs [fig_ref] Table 3: Odds of having a positive TST response of 5 mm, 10 mm... [/fig_ref].
In BCG-unvaccinated children, the use of a 5 mm or a 10 mm TST cut-off to predict IGRA positivity resulted in similar sensitivities and specificities and provided excellent NPVs at all ages [fig_ref] Table 4: Sensitivity, specificity, PPVs and NPVs for different TST cut-offs in correctly identifying... [/fig_ref]. For all ages of unvaccinated children, the area under the ROC curve was above 0.94 (figure 1). For children under 5 years who had been vaccinated with BCG, the sensitivity and specificity of TST to predict IGRA positivity was low using all three cut-offs [fig_ref] Table 5: Sensitivity, specificity, PPVs and NPVs for different TST cut-offs in correctly identifying... [/fig_ref]. In older children, good sensitivities were seen using 5 or 10 mm, but at 5 mm specificity was poor. By using a 15 mm cut-off the specificity improved, but at the expense of sensitivity. For BCG-vaccinated children under 5 years, the area under the ROC curve was below 0.8. However, in children over 5 years, the area under the curve was greater than 0.9.
# Discussion
This large UK cohort study of paediatric household TB contacts shows that BCG vaccination had little effect on TST size in children over 5 years of age. In younger children, BCG did affect TST measurements, but as age increased this effect was less pronounced. In BCG-unvaccinated children, 5 and 10 mm TST cut-offs showed good sensitivity and specificity to detect TB infection, defined as IGRA positivity, while 15 mm was associated with impaired sensitivity. In BCG-vaccinated children, TST performed poorly in children under 5 years in terms of the sensitivity and specificity to predict IGRA positivity, but appeared to be reliable in children over this age, with 10 and 15 mm cut-offs providing reasonable sensitivity and specificity.
Although this topic has been evaluated in other settings, we have not been able to identify any studies carried out previously in the UK. A systematic review of studies that evaluated the impact of BCG on TST found that if vaccination was given at birth, the impact on TST size waned fairly rapidly and by the age of 10 years less than 1% of children had a TST of 10 mm or greater. [bib_ref] False-positive tuberculin skin tests: what is the absolute effect of BCG and..., Farhat [/bib_ref] If BCG was given after the first year of life, however, a much higher proportion of children remained TST positive. In line with our own data, a recent study from Spain found that the risk of a false-positive TST result, due to BCG given at birth, disappeared by 3 years post vaccination. [bib_ref] Tuberculin skin test in bacille Calmette-Guérin-vaccinated children: how should we interpret the..., Piñeiro [/bib_ref] Similar findings were also reported from a study conducted in Canada, where the data showed that when using a 5 mm cut-off, TST positivity was higher in children of all ages who had been vaccinated at birth compared with unvaccinated children. However, after 1 year, when using a cut-off of 15 mm, the rate of TST positivity was no different in vaccinated children. When using a cut-off of 10 mm, more vaccinated children had a positive TST at the age of 1 year, but by 4 years of age, no differences were seen. [bib_ref] The effect of neonatal bacille Calmette-Guérin vaccination on purified protein derivative skin..., Reid [/bib_ref] Using different TST size cut-offs for BCG vaccinated and unvaccinated children is not a universal policy. A recent review of national and international childhood TB guidelines found that most countries and agencies use a 10 mm induration cut-off for all children irrespective of BCG vaccination status, [bib_ref] Tuberculosis in childhood: a systematic review of national and international guidelines, Berti [/bib_ref] in line with the advice of WHO.To inform policy it is important to obtain evidence from the population for whom the policy is intended. To this end, this large study of children living in the UK, evaluated post exposure to infectious cases of pulmonary TB, is informative. This is especially so in the context of the recent revision to the NICE guidelines for the management of TB in the UK where a TST cut-off of 5 mm has been recommended for all ages of children irrespective of BCG vaccination history.This study has some acknowledged limitations. The first limitation was that we used the UK NICE 2006 definitions of TST positivity to decide if a second TST was required. Therefore, for BCG-unvaccinated children, those with an initial TST of ≥6 mm did not have a repeat test after 2 months. Although this pragmatic approach reflected UK policy, it may have resulted in TST measurements for BCG-vaccinated children being lower than if we had repeated the TST in all children and used the largest TST response in analysis. A second limitation was that as this study was carried out under programmatic conditions, some data were missing. Although, the absolute numbers were small, the most common missing data were repeat TST testing for those children with initially negative tests (five children). Third, as we were using routine, clinical data for IGRA testing, some centres used QuantiFERON-TB Gold In-Tube and some T-SPOT.TB. Although it would have been ideal to have used the same test for all children, we, and other research groups, have previously shown these two IGRA tests to be highly concordant and it was a deliberate decision of the principal investigator and the collaborators to use their locally available IGRA. Fourth, we used evidence of a BCG scar, documentation of vaccination, or a clear history of vaccination from the parents as evidence of BCG vaccination. It is possible that some children were incorrectly classified using this approach. Fifth, although the majority of children were vaccinated soon after birth, we were unable to document exactly when each child received the BCG. This makes it more difficult to determine how quickly the effects of BCG on TST responses wane. Sixth, this study was carried out in the UK and caution should be exercised when generalising results to other contexts. We anticipate that similar findings would be seen in countries with a similar epidemiological TB pattern but it should be acknowledged that the PPVs and NPVs would likely change with the background TB prevalence and consequent pre-test probability that the child was IGRA positive. Seventh, although all children were exposed to TB in their households, it is theoretically possible that children living in areas that recommended BCG vaccination were, in some way, more exposed to TB than children from other regions in the UK. This could have led to some degree of selection bias that might have impacted on TST responses. Eighth, it is possible that in some of the age subgroups, the sample size was not large enough to detect modest variations in TST response due to BCG, especially in older children. Finally, and most importantly, we used the IGRA result as the 'gold standard' to define who was genuinely infected and compared different TST cut-offs in BCG vaccinated and unvaccinated children against this result. It is acknowledged that IGRA tests have imperfect sensitivity and specificity, especially in younger children. It is therefore possible that some IGRA-negative children who had a reactive TST were actually infected and their positive TST response may have been due to M. tuberculosis. These would have been classified as false positives. Conversely some IGRA-positive children with unreactive TST responses may not have been infected with M. tuberculosis. These children would have been classified as false negatives. As we sought to determine the impact of BCG on TST responses in children, we used IGRA as a tool to screen out the majority of the TST responses that were caused by genuine TB infection and not BCG. Although there may have been one or two misclassifications, we are confident that the overall result is valid.
Our study suggests that in the youngest children BCG does affect TST measurements. However, these children are the most vulnerable to progression from infection to disease and are also the most susceptible to severe disseminated forms of TB disease, such as TB meningitis and miliary TB. Because of this, greater screening test sensitivity and a high NPV is more important than specificity and PPV for many clinicians when evaluating children in this age group. WHO recommends treating all children aged under 5 years for TB infection following significant exposure to an infectious TB case, irrespective of the result of the test of TB infection and irrespective of BCG vaccination status.For those managing TB in children, there is a need to consider the interpretation of screening tests. As no tests are perfect there will always be a trade-off between sensitivity and specificity of the test chosen as well as their respective PPVs and NPVs. The question of whether it is better to treat some children for TB infection unnecessarily (because of greater test sensitivity) or to fail to treat some children with TB infection who may progress to disease (because of greater test specificity) remains challenging. Although 6 months of daily isoniazid is a long course of treatment for a child who is well, it is safe [bib_ref] Toxicity of first-line drugs for treatment of tuberculosis in children: review, Frydenberg [/bib_ref] and effective. Three months of daily isoniazid and rifampicin has been shown to be an appropriate alternative strategy. [bib_ref] Effectiveness of 3 months of rifampicin and isoniazid chemoprophylaxis for the treatment..., Bright-Thomas [/bib_ref] Ideally, tests with a high NPV are used for screening, while tests for diagnosis should have a high PPV. However, both false negatives [fig_ref] Figure 1: Receiver operating characteristic curves for tuberculin skin test results of different cut-offs... [/fig_ref] Continued. and false positives have clinical and cost implications. Modelling exercises suggest that not only is the treatment of TB infection in children likely to be a highly cost-effective strategy, but in children under 2 years old, screening for exposure and treating without testing for evidence of infection is the most costeffective intervention. [bib_ref] Modelling the cost-effectiveness of strategies to prevent tuberculosis in child contacts in..., Mandalakas [/bib_ref] As newer, simpler treatments become available, such as the 3-month duration once-a-week regimen with rifapentine and isoniazid, the benefits of treating may increasingly outweigh the risks and costs and of not treating younger children.
# Conclusion
Our data suggest that the impact of infant BCG vaccination on TST responses in children exposed to TB wanes with age. In BCG-vaccinated children a TST cut-off of 5 mm is associated with poor specificity.
Funding This work was funded by a NIHR Senior Research Fellowship to BK (NIHR/ SRF-2009-02-07). JAS was supported by an NIHR Academic Clinical Lectureship and also through grants from the Academy of Medical Sciences and the BRC. The research and DK were supported by the National Institute for Health Research (NIHR) Biomedical Research Centre based at Imperial College Healthcare NHS Trust and Imperial College London.
Competing interests None declared.
Ethics approval The study was approved by the UK National Research Ethics Service (REC: 11/11/11). Parents of all children included in the study provided written informed consent, with additional consent provided in older children.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement Presented data are fully anonymised. No consent for data sharing with other parties was obtained, but the corresponding author may be contacted to forward requests for data sharing. Open Access This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/ by/4.0/
[fig] Figure 1: Receiver operating characteristic curves for tuberculin skin test results of different cut-offs to predict interferon γ release assay positivity. (A) BCG-unvaccinated children <2 years (n=20; AUC=1.00; 95% CI not calculable). (B) BCG-vaccinated children <2 years (n=64; AUC=0.79; 95% CI 0.65 to 0.93). (C) BCG-unvaccinated children aged 2 to <5 years (n=32; AUC=0.97; 95% CI 0.91 to 1.00). (D) BCG-vaccinated children aged 2 to <5 years (n=65; AUC=0.78; 95% CI 0.60 to 0.95). (E) BCG-unvaccinated children aged 5 to <10 years (n=41; AUC=0.95; 95% CI 0.88 to 1.00). (F) BCG-vaccinated children aged 5 to <10 (n=108; AUC=0.92; 95% CI 0.86 to 0.97). (G) BCG-unvaccinated children aged 10 to <15 (n=29; AUC=0.99; 95% CI 0.96 to 1.00). (H) BCG-vaccinated children aged 10 to <15 (n=63; AUC=0.95; 95% CI 0.90 to 1.00). [/fig]
[table] Table 1: Demographic and clinical characteristics of children in the study [/table]
[table] Table 3: Odds of having a positive TST response of 5 mm, 10 mm and 15 mm between IGRA negative children who had been vaccinated with BCG at birth versus children who had not been vaccinated with BCG Unable to calculate ORs where a zero exists in one of the four cells needed to generate the OR. No BCG-unvaccinated children under 2 years with negative IGRA results had a TST induration of greater than 15 mm. IGRA, interferon γ release assay; TST, tuberculin skin test. [/table]
[table] Table 2: Median TST induration in BCG-vaccinated and BCG-unvaccinated children at different ages who are IGRA negative [/table]
[table] Table 4: Sensitivity, specificity, PPVs and NPVs for different TST cut-offs in correctly identifying IGRA positivity in BCG-unvaccinated children [/table]
[table] Table 5: Sensitivity, specificity, PPVs and NPVs for different TST cut-offs in correctly identifying IGRA positivity in BCG-vaccinated children 5 mm (with 95% CIs) 10 mm (with 95% CIs) 15 mm (with 95% CIs)IGRA, interferon γ release assay; NPV, negative predictive value; PPV, positive predictive value; TST, tuberculin skin test. [/table]
|
Variability in phase and amplitude of diurnal rhythms is related to variation of mood in bipolar and borderline personality disorder
Method for calculating variability parametersThe mean of the difference in timings between the daily and total minimums were calculated as:where D is the total number of days. The standard deviation of the difference in timings between the daily and total minimums is defined as:Similarly, the mean (µ A and µ M ) and standard deviation (σ A and σ A ) of the difference between the amplitudes and MESORs of the daily and total sinusoids can be calculated from Ad and Aw and Md and Mw respectively. The residual sum of squares between the daily and total sinusoids is calculated as a total measure of regularity for the period of recording:where SIN represents the fitted sinusoids.In addition to regularity measures coming from the comparison of the daily sinusoids to the weekly sinusoids, means and standard deviations of successive differences of the daily sinusoids were found. The mean was found for the successive 1
# Supplementary materials
## Method for calculating variability parameters
The mean of the difference in timings between the daily and total minimums were calculated as:
[formula] µ T = 1 D D ∑ i=1 (T d i − Tw i )(1) [/formula]
where D is the total number of days. The standard deviation of the difference in timings between the daily and total minimums is defined as:
[formula] σ T = 1 D D ∑ i=1 (T d i − Tw i ) − µ T 2(2) [/formula]
Similarly, the mean (µ A and µ M ) and standard deviation (σ A and σ A ) of the difference between the amplitudes and MESORs of the daily and total sinusoids can be calculated from Ad and Aw and Md and Mw respectively. The residual sum of squares between the daily and total sinusoids is calculated as a total measure of regularity for the period of recording:
[formula] RSS = N ∑ i=1 SIN day i − SIN week i 2 (3) [/formula]
where SIN represents the fitted sinusoids.
In addition to regularity measures coming from the comparison of the daily sinusoids to the weekly sinusoids, means and standard deviations of successive differences of the daily sinusoids were found. The mean was found for the successive differences of timings of the minimum values:
[formula] µDIF T = 1 (D − 1) D−1 ∑ i=1 (T d i+1 − T d i )(4) [/formula]
with the standard deviation defined as:
[formula] σ DIF T = 1 (D − 1) D−1 ∑ i=1 (T d i+1 − T d i ) − µDIF T 2 (5) [/formula]
All these features were calculated for HR and acceleration data to give measures of the range of: HR and activity levels through the amplitude, the average HR and activity levels through the MESOR and the timings of the maximum or minimum HR and activity levels through the phase. As the vertical acceleration is a measure of sleep, the timings of the sleep can be indirectly measured through the phase and the amount of sleep or rest-activity, can be measured through the MESOR. |
Quantification of one Prenylated Flavanone from Eysenhardtia platycarpa and four derivatives in Ex Vivo Human Skin Permeation Samples Applying a Validated HPLC Method
Prenylated flavanones are polyphenols that have diverse biological properties. The present paper focuses on a HPLC method validation for the quantification of prenylated flavanones (2S)-5,7-dihydroxy-6-(3-methyl-2-buten-1-yl)-2-phenyl-2,3-dihydro-4H-1Benzopyran-4-one 1 and derivatives (2S)-5,7-bis(acetyloxy)-6-(3-methyl-2-buten-1-yl)-2-phenyl-2,3-dihydro-4H-1-Benzopyran-4-one A; (2S)-5-hydroxy-7-methoxy-6-(3-methyl-2-buten-1-yl)-2-phenyl-2,3-dihydro-4H-1-Benzopyran-4-one B; (8S)-5-hydroxy-2,2-dimethyl-8-phenyl-3,4,7,8-tetrahydro-2H,6H-Benzo [1,2-b:5,4-b ] dipyran-6-one C; and (8S)-5-hydroxy-2,2-dimethyl-8-phenyl-7,8-dihydro-2H,6H-Benzo [1,2-b:5,4-b ] dipyran-6-one D applied in biopharmaceutic studies. The linear relationships are proven with significant correlation coefficients (R 2 > 0.999) in the range of 1.56 to 200 µg/mL with low limits of detection and quantification, on average of 0.4 µg/mL and 1.2 µg/mL, respectively. The validation method used in this work is highly accurate and precise, with values lower than 15%. The relative standard deviation values of repeatability of the instrumental system are demonstrated with less than 0.6% for all studied flavanones. Therefore, the applicability method of the quantification of the prenylated flavanones was established using the permeation of human skin in the Franz cell system. During the method previously described, there was no interference observed from human skin components in ex vivo permeation studies.
# Introduction
Prenylated flavanones are part of a diverse class of natural flavonoids conformed of oxygen-containing heterocycles and prenyl substituents. They have proven to have antibacterial and anti-fungal properties conferring protection to plants against diseases [bib_ref] Biocatalytic Access to Diverse Prenyl Flavonoids by Combining a Regiospecific C-Prenyltransferase and..., Li [/bib_ref] [bib_ref] Prenylated Flavonoids from Desmodium caudatum and Evaluation of Their Anti-MRSA Activity, Sasaki [/bib_ref]. Nowadays, several studies have confirmed their biological effects against oxidation, obesity, and inflammation, which may be useful in the prevention of several diseases, including cancer [bib_ref] Natural Antioxidants: Sources, Compounds, Mechanisms of Action, and Potential Applications, Brewer [/bib_ref] [bib_ref] Prenylated Chalcones and Flavonoids for the Prevention and Treatment of Cancer, Venturelli [/bib_ref] [bib_ref] Prenylated Flavonoids, Promising Nutraceuticals with Impressive Biological Activities, Yang [/bib_ref]. The genus Eysenhardtia comprises 14 species, and some of them, including E. platycarpa, which contains prenylated flavanones, are widely used in traditional Mexican medicine for the treatment of kidney and bladder infections [bib_ref] Introduction of Nanotechnology in Herbal Drugs and Nutraceutical: A Review, Gopi [/bib_ref]. E. platycarpa is a tree found throughout southern Mexico and is popularly known as "taray", "palo dulce" and "palo azul". Previous studies on Eysenhardtia species have highlighted hypoglycemic and antidiabetic properties [bib_ref] Antihyperglycemic Activity and Chemical Constituents of Eysenhardtia platycarpa, Narváez-Mastache [/bib_ref]. The antioxidant activity of the alcoholic extract of E. platycarpa was also confirmed [bib_ref] Antioxidant Evaluation of Eysenhardtia Species (Fabaceae): Relay Synthesis of 3-O -Acetyl-11 α,..., Narváez Mastache [/bib_ref] [bib_ref] Evaluation of Antidiabetic, Antioxidant and Antiglycating Activities of the Eysenhardtia polystachya, Gutierrez [/bib_ref]. Domínguez-Villegas et al. isolated five flavanones from a methanolic extract obtained from the aerial parts of E. platycarpa, one of them were (2S)-5,7-dihydroxy-6-(3-methyl-2-buten-1-yl)-2-phenyl-2,3-dihydro-4H-1Benzopyran-4-one 1 [fig_ref] Figure 1: Chemical structures of studied compounds [/fig_ref]. The formerly mentioned compounds proved to have anti-inflammatory activity. Furthermore, flavanones revealed a high percentage reduction of free radical DPPH (2,2-Diphenyl-1-picrylhydrazyl), and thus exhibited strong cytotoxic activity on brine shrimp [bib_ref] Antioxidant and Cytotoxicity Activities of Methanolic Extract and Prenylated Flavanones Isolated from..., Domínguez-Villegas [/bib_ref].
Biomolecules 2020, 10, x FOR PEER REVIEW 6 of 10 kidney and bladder infections [bib_ref] Introduction of Nanotechnology in Herbal Drugs and Nutraceutical: A Review, Gopi [/bib_ref]. E. platycarpa is a tree found throughout southern Mexico and is popularly known as "taray", "palo dulce" and "palo azul". Previous studies on Eysenhardtia species have highlighted hypoglycemic and antidiabetic properties [bib_ref] Antihyperglycemic Activity and Chemical Constituents of Eysenhardtia platycarpa, Narváez-Mastache [/bib_ref]. The antioxidant activity of the alcoholic extract of E. platycarpa was also confirmed [bib_ref] Antioxidant Evaluation of Eysenhardtia Species (Fabaceae): Relay Synthesis of 3-O -Acetyl-11 α,..., Narváez Mastache [/bib_ref] [bib_ref] Evaluation of Antidiabetic, Antioxidant and Antiglycating Activities of the Eysenhardtia polystachya, Gutierrez [/bib_ref]. Domínguez-Villegas et al. isolated five flavanones from a methanolic extract obtained from the aerial parts of E. platycarpa, one of them were (2S)-5,7-dihydroxy-6-(3-methyl-2-buten-1-yl)-2-phenyl-2,3-dihydro-4H-1Benzopyran-4-one 1 [fig_ref] Figure 1: Chemical structures of studied compounds [/fig_ref]. The formerly mentioned compounds proved to have anti-inflammatory activity. Furthermore, flavanones revealed a high percentage reduction of free radical DPPH (2,2-Diphenyl-1-picrylhydrazyl), and thus exhibited strong cytotoxic activity on brine shrimp [bib_ref] Antioxidant and Cytotoxicity Activities of Methanolic Extract and Prenylated Flavanones Isolated from..., Domínguez-Villegas [/bib_ref].
In addition, the prenylated flavanone 1, as well as its derivatives obtained from structural modifications:
(2S)-5,7-bis(acetyloxy)-6-(3-methyl-2-buten-1-yl)-2-phenyl-2,3-dihydro-4H-1- [fig_ref] Figure 1: Chemical structures of studied compounds [/fig_ref] ; were loaded into polymeric nanoparticles exhibiting cytotoxic potential against pancreatic cell line MiaPaCa-2 [bib_ref] Cytotoxic Evaluation of (2S)-5,7-Dihydroxy-6-Prenylflavanone Derivatives Loaded PLGA Nanoparticles against MiaPaCa-2 Cells, Andrade-Carrera [/bib_ref].
[formula] Benzopyran-4-one A; (2S)-5-hydroxy-7-methoxy-6-(3-methyl-2-buten-1-yl)-2-phenyl-2,3-dihydro- 4H-1-Benzopyran-4-one B; (8S)-5-hydroxy-2,2-dimethyl-8-phenyl-3,4,7,8-tetrahydro-2H,6H- Benzo[1,2-b:5,4-b ]dipyran-6-one C; and (8S)-5-hydroxy-2,2-dimethyl-8-phenyl-7,8-dihydro-2H,6H- Benzo[1,2-b:5,4-b ]dipyran-6-one D (O O H OH O 1 O O O O O O A O O OH O B O O OH O C D O O OH O Figure 1. Chemical structures of studied compounds. (2S)-5,7-dihydroxy-6-(3-methyl-2-buten-1-yl)- 2-phenyl-2,3-dihydro-4H-1Benzopyran-4-one (1); (2S)-5,7-bis(acetyloxy)-6-(3-methyl-2-buten-1-yl)-2- phenyl-2,3-dihydro-4H-1-Benzopyran-4-one (A); (2S)-5-hydroxy-7-methoxy-6-(3-methyl-2-buten-1- yl)-2-phenyl-2,3-dihydro-4H-1-Benzopyran-4-one (B); (8S)-5-hydroxy-2,2-dimethyl-8-phenyl-3,4,7,8- tetrahydro-2H,6H-Benzo[1,2-b:5,4-b ]dipyran-6-one (C); and (8S)-5-hydroxy-2,2-dimethyl-8-phenyl- 7,8-dihydro-2H,6H-Benzo[1,2-b:5,4-b ]dipyran-6-one (D). [/formula]
The increasing interest in the medicinal properties of flavanones, with specific antiinflammatory activity on skin diseases [bib_ref] Development of Fl Avanone and Its Derivatives as Topical Agents against Psoriasis:..., Alalaiwe [/bib_ref] , has led to a demand for accurate, reproducible, and sensitive analytical methods to quantify new compounds that have not been validated yet. Highperformance liquid chromatography (HPLC) is the most widely used separation method for quantifying phenolic compounds [bib_ref] Recent Developments in the HPLC Separation of Phenolic Food Compounds, Pyrzynska [/bib_ref] [bib_ref] Recent Advances and Trends in the Liquid-Chromatography-Mass Spectrometry Analysis of Flavonoids, Villiers [/bib_ref]. The conditions include mainly the use of C18 reverse phase columns and a diode array and a fluorescence detector. Aqueous solutions and acetonitrile or In addition, the prenylated flavanone 1, as well as its derivatives obtained from structural modifications: (2S)-5,7-bis(acetyloxy)-6-(3-methyl-2-buten-1-yl)-2-phenyl-2,3-dihydro-4H-1-Benzopyran -4-one A; (2S)-5-hydroxy-7-methoxy-6-(3-methyl-2-buten-1-yl)-2-phenyl-2,3-dihydro-4H-1-Benzopyran dipyran-6-one D [fig_ref] Figure 1: Chemical structures of studied compounds [/fig_ref] ; were loaded into polymeric nanoparticles exhibiting cytotoxic potential against pancreatic cell line MiaPaCa-2 [bib_ref] Cytotoxic Evaluation of (2S)-5,7-Dihydroxy-6-Prenylflavanone Derivatives Loaded PLGA Nanoparticles against MiaPaCa-2 Cells, Andrade-Carrera [/bib_ref]. The increasing interest in the medicinal properties of flavanones, with specific anti-inflammatory activity on skin diseases [bib_ref] Development of Fl Avanone and Its Derivatives as Topical Agents against Psoriasis:..., Alalaiwe [/bib_ref] , has led to a demand for accurate, reproducible, and sensitive analytical methods to quantify new compounds that have not been validated yet. High-performance liquid chromatography (HPLC) is the most widely used separation method for quantifying phenolic compounds [bib_ref] Recent Developments in the HPLC Separation of Phenolic Food Compounds, Pyrzynska [/bib_ref] [bib_ref] Recent Advances and Trends in the Liquid-Chromatography-Mass Spectrometry Analysis of Flavonoids, Villiers [/bib_ref]. The conditions include mainly the use of C18 reverse phase columns and a diode array and a fluorescence detector. Aqueous solutions and acetonitrile or methanol are usually the mobile phases. Notwithstanding, there are HPLC validated methods to quantify compounds similar to flavanones assayed [bib_ref] Validation of a Rapid and Sensitive Reversed-Phase Liquid Chromatographic Method for the..., Sus [/bib_ref]. The present job focused on the validation method of unpublished molecules prenylated flavanone 1 extracted from E. platycarpa and its derivatives A-D; considering the human skin as the principal biologic matrix. Moreover, this method was used to determine the concentration of prenylated flavanones in permeation and retention samples of ex vivo diffusional studies, using human skin and following bioanalytical guidelines to evaluate their intrinsic permeation, before they were analyzed in vivo as potential anti-inflammatory drugs candidates.
# Materials and methods
## Chemicals and reagents
The purified water used in all experiments was obtained from the MilliQ ® Plus System lab supply. All other chemical reagents used in this study were purchased from Fisher Scientific (Leicestershire, UK). The solvents were appropriately filtered through a 0.45 µm Millipore membrane filter (Merck, Darmstadt, Germany) and degassed in an ultrasonic bath for 20 min.
## Extraction and isolation of plant material
E. platycarpa leaves were collected in Tetipac, Guerrero, Mexico, and identified by Prof. Ramiro Cruz (Registration Number: Ramiro Cruz 1325 from the Sciences Faculty Herbarium Facilities of the Autonomous University of the State of Morelos). The leaves were dried at room temperature, then pulverized and extracted by three consecutive macerations with methanol at room temperature (100g of dried vegetal material per 1000 mL methanol). The extraction solvent was removed under reduced pressure. Next, the prenylated flavanone 1 was isolated by column chromatography at reduced pressure. Finally, it was purified and characterized by direct thin-layer chromatography (TLC) comparison with original samples available in the laboratory. The product was a yellow powder precipitate with a melting point of 200.2 - C. The compound obtained was characterized by comparison with previously published melting point data and with 1 H-NMR results [bib_ref] Antihyperglycemic Activity and Chemical Constituents of Eysenhardtia platycarpa, Narváez-Mastache [/bib_ref].
## Semi-synthesis from natural prenylated flavanone
Each prenylated flavanone was prepared following the method as previously reported [bib_ref] Cytotoxic Evaluation of (2S)-5,7-Dihydroxy-6-Prenylflavanone Derivatives Loaded PLGA Nanoparticles against MiaPaCa-2 Cells, Andrade-Carrera [/bib_ref] getting (2S)-5,7-bis(acetyloxy)-6-(3-methyl-2-buten-1-yl)-2-phenyl-2,3-dihydro-4H-1-Benzopyran-4-one A;
[formula] (2S)-5-hydroxy-7-methoxy-6-(3-methyl-2-buten-1-yl)-2-phenyl-2,3-dihydro-4H-1-Benzopyran-4-one B; (8S)-5-hydroxy-2,2-dimethyl-8-phenyl-3,4,7,8-tetrahydro-2H,6H-Benzo[1,2-b:5,4-b ] dipyran-6-one C; and (8S)-5-hydroxy-2,2-dimethyl-8-phenyl-7,8-dihydro-2H,6H-Benzo[1,2-b:5,4-b ] dipyran-6-one D. [/formula]
## Chromatographic operating conditions
The HPLC system consisted of a Waters 515 HPLC pump, a 717 Plus autosampler, and a dual λ absorbance UV-vis 2487 detector (Waters, Milford, MA, USA). The analytical column was Atlantis ® C18 5 µm 250 mm × 4.6 mm, Waters. The analyte separation was performed with 10 µL sample injection volume. The separations were done in isocratic mode at room temperature. The mobile phase with a flow rate of 1 mL/min comprised of W-water and AcN-acetonitrile (%W: %AcN) with a different composition for each prenylated flavanone studied: 1 (30:70), A (20:80), B (40:60), C (20:80) and D (10:90). The detection wavelengths determined by spectrum scan were 300 nm for 1, B, C, D, and 320 nm for A. The Peak area was used to quantify each analyte.
## Ex vivo human skin permeation
A blank sample (ethanol: water; 70:30; v/v) was used as a negative control and the samples of prenylated flavanones (1, A, B, C, and D) with a concentration of 200 µg/mL were permeated through human skin membrane in the receptor compartment of the Franz diffusion cells (FDC 400, Crown Glass, Somerville, NY, USA), with a diffusion area of 2.54 cm 2 . Human skin from abdominal plastic surgery of healthy patients was used as a permeation membrane. The skin was cut into 400 µm thickness and placed between the donor, and the receptor compartment of the Franz diffusion cells, avoiding the formation of bubbles [bib_ref] Biopharmaceutical pro Fi Le of Hydrogels Containing Pranoprofen-Loaded PLGA Nanoparticles for Skin..., Abrego [/bib_ref]. [fig_ref] Figure 1: Chemical structures of studied compounds [/fig_ref] were applied to the donor compartment and the receptor compartment was filled with ethanol: water (70:30) solution. The receptor compartment was kept at 32 ± 1 - C. Twenty-four h after the application of the tests, 300 µL aliquots were collected from the receptor side. Sink conditions were always followed. The flavanones amount permeated (Q) through human skin were determined by HPLC analysis described in Section 2.4.
## Prenylated flavanone extraction
At the end of the ex vivo human skin permeation study, the flavanones amounts remaining in the skin were quantified by calculating the flavanone amount extracted from the skin to the flavanone amount added. For this purpose, the skin was removed from the Franz cells, cleaned with gauze soaked in a 0.05% solution of dodecyl sulfate and washed with distilled water. The permeated areas of the skin were then excised and weighed. The flavanone contained in the skin was extracted with ethanol: water (70:30) mixture under sonication (20 min) in an ultrasonic bath. The resulting solutions were measured with HPLC, quantifying the flavanone amount retained in the skin in micrograms of prenylated flavanone per grams of skin and per area unit µg/g skin .cm 2 ).
## Recovery from human skin tissues and prenylated flavanone retained
The accuracy of the extraction was evaluated by adding 1 mL of each prenylated flavanone solution (200 µg/mL) to their corresponding vials containing approximately 100 mg of human skin. These vials remained for 24 h at 32 - C to simulate the permeation conditions experiments. This experiment was conducted in triplicate. After the time of the study, the skin was submitted to drug extraction, as described in Section 2.6. The initial solutions and the eluates from each assay were collected and analyzed with HPLC. The differences obtained between the initial flavanone amount in the solution and the final flavanone amount in the collected solutions after 24 h were considered to be the value of the respective flavanone amount bound to tissue. Recovery percentage was calculated comparing the corresponding drug extraction results with the flavanone amount bound to the tissue [bib_ref] Validation of Chromatographic Methods in Biomedical Analysis Viewpoint and Discussion, Causon [/bib_ref]. A comparison of the amount of prenylated flavanone extracted and the recovery percentage was made in order to find out the real amount of flavanone retained in the skin.
# Analytical method validation
The method was validated according to the International Conference on Harmonization guidelines (ICH) [bib_ref] Validation of High-Performance Liquid Chromatography Methods for Pharmaceutical Analysis, Shabir [/bib_ref] for linearity, the limit of detection (LOD), the limit of quantification (LOQ), accuracy, and precision. Calibration curves were analyzed in two ranges; from 200 to 12.5 µg/mL in a high concentration level, and from 12.5 to 1.56 µg/mL in a low concentration level.
## Standard solutions for calibration curves
Standard stock solutions of each compound [fig_ref] Figure 1: Chemical structures of studied compounds [/fig_ref] were prepared daily by dissolving the appropriate amount of each analyte in ethanol: water (7:3; v/v) to obtain a final concentration of (200 µg/mL). The working solutions were elaborated by the dilution of appropriate aliquots of the stock solutions with the diluting solvent to reach the concentration ranges 1.56, 3.12, 6.25, 12.5, 25, 50, 100 µg/mL.
## Linearity
The linearity was evaluated by a one-way analysis of variance (ANOVA) test to compare peak areas versus nominal concentrations of each standard [bib_ref] Development and Validation of a High-Performance Liquid Chromatography Method for the Quantification..., Alvarado [/bib_ref]. Differences were considered statistically significant when p < 0.05. The least-square linear regression analysis and mathematical determinations were performed by Prism, V 5.0 software (Graph Pad Software Inc., San Diego, CA, USA).
## Limit of detection and limit of quantification
The limit of detection (LOD) and the limit of quantitation (LOQ) for each analyte [fig_ref] Figure 1: Chemical structures of studied compounds [/fig_ref] were calculated based on the standard deviation of the response and the slope of the calibration curve, generated from six replicate analysis applying the formula 1 [bib_ref] Development and Validation of Reverse-Phase High-Performance Liquid Chromatographic (RP-HPLC) Method for Quantification..., Surve [/bib_ref] :
[formula] LOD or LOQ = k SD Sa S b (1) [/formula]
where k is the factor related to the level of confidence (k = 3.3 and 10 for LOD and LOQ respectively), SD Sa is the standard deviation of the intercept, and S b is the slope.
## Repeatability, accuracy, and precision
The instrumental repeatability was assayed by analyzing the concentration sample of 200 µg/mL for each flavanone [fig_ref] Figure 1: Chemical structures of studied compounds [/fig_ref] repeatedly seven times, consecutively. The accuracy and precision were investigated by measuring samples in three concentrations 1.56, 12.5, and 200 µg/mL [bib_ref] Analysis Pranoprofen Quantification in Ex Vivo Corneal and Scleral Permeation Samples: Analytical..., Abrego [/bib_ref]. The inter-day test was conducted by analyzing each analyte [fig_ref] Figure 1: Chemical structures of studied compounds [/fig_ref] with each of the three concentration levels mentioned before, once a day for six consecutive days. The accuracy was expressed as a relative error (RE%). The precision was defined as the relative standard deviation (RSD%) of the measurement. The method is considered accurate and precise if RE% and RSD%, respectively, are within ±15%.
## Specificity
Specificity is defined as the ability of the method to distinguish the analyte from all other substances present in the sample. This can be proven by comparing the analyte chromatographic retention time in extracted matrix samples and with its retention time in at least one reference solution [bib_ref] Validation of Chromatographic Methods in Biomedical Analysis Viewpoint and Discussion, Causon [/bib_ref] [bib_ref] Advances in Validation, Risk and Uncertainty Assessment of Bioanalytical Methods, Rozet [/bib_ref]. To test the specificity of the analytical method, the ex vivo human permeation procedure described in Section 2.5 was followed. The blank sample peaks should not appear at the same retention times of the prenylated flavanones.
# Results and discussions
Due to the fact of the biological properties of prenylated flavanones, it is of utmost importance to count on analytical method validation in order to promote future studies. HPLC is highly sensitive in the determination of small quantities of natural molecules in biopharmaceutical studies based on previous studies [bib_ref] Development and Validation of a High-Performance Liquid Chromatography Method for the Quantification..., Alvarado [/bib_ref].
# Analytical method validation
## Linearity
The linearity of the analytical method is the capability over a range of data to obtain proportional results. The applied HPLC method for the flavanone's quantification [fig_ref] Figure 1: Chemical structures of studied compounds [/fig_ref] showed satisfactory linearity in the tested concentrations. In order to have a better mathematical analysis, the linearity was evaluated in two concentration ranges: from 200 to 12.5 µg/mL and from 12.5 to 1.56 µg/mL. Two linear calibration curves were fitted for each flavanone. The R 2 value for each analyte was found above 0.999 for all studied flavanones, indicating the linear relationship between the analyte concentration and the peak area. No statistical differences were found (p > 0.05) after the ANOVA test of the calibration curves of each flavanone 1, A, B, C, and D with p values for each two levels of 0.12 and 0.08, 0.93 and 0.08, 0.38 and 0.47, 0.63 and 0.56, and 0.53 and 0.46, respectively (see . . Linearity (expressed in R 2 and p with two ranges, one by row), Precision, Accuracy (calculated at maximum, medium, and minimum concentration values), and Repeatability of the HPLC Method for the determination of flavanones.
## Compound
## Limit of detection and limit of quantification
LODs and LOQs for all the investigated flavanones were calculated using the response standard deviation and the calibration curve slope of 12.5 to 1.56 µg/mL for each flavanone, described in Section 2.8.3. The values of LODs and LOQs for each flavanone are listed in . These results indicate that the method is sensitive enough for flavanones determination in the range of 1.56 to 200 µg/mL.
## Repeatability, accuracy, and precision
Precision and accuracy values were obtained from sample analyses of the 1.56, 12.5, and 200 µg/mL flavanones concentrations [fig_ref] Figure 1: Chemical structures of studied compounds [/fig_ref]. The inter-day precision and accuracy were calculated after analyzing the samples on six different days. The results are reported in . Both parameters were lower than the 15 % limit value in EMA (European Medicines Agency) guidelines. These results suggest that the proposed method has satisfactory accuracy and precision. Repeatability studies of the instrumental system showed RSD % not greater than 0.6% for all flavanones.
## Specificity
The analytical methodology was implemented for the flavanone's quantitation [fig_ref] Figure 1: Chemical structures of studied compounds [/fig_ref] in skin permeation studies. In order to show the specificity, 300 µL of the flavanones at 200 µg/mL concentration was permeated into human skin using Franz cells. The amount of each flavanone normalized by the surface (Q) in the receptor compartment during percutaneous permeation experiments (n = 3) is indicated in [fig_ref] Table 2: Results of the permeation studies expressed by mean and SD [/fig_ref]. Percentages of permeation were calculated, accounting for the experimental flavanone content of each assay. At the end of each experiment, skin samples were removed from the diffusion cell, flavanones amount retained (Q ret ) were quantified as previously described and expressed in micrograms of prenylated flavanone per grams of skin and per area unit (µg/g skin .cm 2 ) in [fig_ref] Table 2: Results of the permeation studies expressed by mean and SD [/fig_ref]. As shown in [fig_ref] Table 2: Results of the permeation studies expressed by mean and SD [/fig_ref] , the prenylated flavanone 1 is the one which most permeates, followed by D and C. In the case of A and B, Q was detectable but not quantifiable because their values were below LOQ determinate before (see . On the contrary, A and B are found retained in the skin in a greater amount, so we can infer that it is possible that these molecules have greater interactions with the different skin components due to their physicochemical characteristics. Therefore, they have prevented permeation as 1, C, and D did too. The chromatograms showed the absence of interference of any other peak corresponding to each flavanone [fig_ref] Figure 1: Chemical structures of studied compounds [/fig_ref]
## Recovery
Flavanone extraction was done, as described in Section 2.6. Recovery was calculated comparing the corresponding extraction result with the amount of flavanone bound to the skin. The results were reported as the mean value of the percentage between the amount of flavanone in each sample and the weight of the skin sample (see in [fig_ref] Table 2: Results of the permeation studies expressed by mean and SD [/fig_ref]. The aforementioned results show the real quantity for each prenylated flavanone that can be recovered using the extraction method previously described. Thus, the exact quantity of flavanone retained can be known, and this quantity is responsible for the exercising of the biological effect.
# Conclusions
The results in the present research describe a liquid chromatographic method validation for the analysis of prenylated flavanones 1, A-D using UV-VIS detection. According to the data obtained, the method developed is linear, accurate, and precise. In addition, the method can be used in the quantification of prenylated flavanones samples from permeation and retention studies in human skin. Finally, the method is selective since the chromatograms allowed the observer to identify the signal for each prenylated flavanone without interference. We can conclude that this method is suitable for further analysis of prenylated flavanones 1, A-D in biological systems, and for other biopharmaceutical studies.
[fig] Figure 1: Chemical structures of studied compounds. (2S)-5,7-dihydroxy-6-(3-methyl -2-buten-1-yl)-2-phenyl-2,3-dihydro-4H-1Benzopyran-4-one (1); (2S)-5,7-bis(acetyloxy)-6-(3-methyl -2-buten-1-yl)-2-phenyl-2, 3-dihydro-4H-1-Benzopyran-4-one (A); (2S)-5-hydroxy-7-methoxy- [/fig]
[fig] Figure 2: HPLC chromatograms of prenylated flavanones 1, A-D classified with sub-index a, b, c, and d. a correspond to: Blank sample (ethanol: water, 70:30) permeated at chromatographic conditions of each prenylated flavanone; b correspond to prenylated flavanone standard; c correspond to prenylated flavanone from receptor compartment Franz diffusion cells; and, d correspond to prenylated flavanone extracted from human skin after permeation study. [/fig]
[fig] Author: Contributions: Conceptualization, P.B.-S.; validation, P.B.-S., B.A.-C. and H.A.; formal analysis, P.B.-S.; investigation, P.B.-S.; resources, A.C.-C. and M.L.G.-R.; writing-original draft preparation, P.B.-S., H.A., M.L.G.-R. and A.C.-C.; writing-review and editing, P.B.-S., B.A.-C., H.A., M.L.G.-R. and A.C.-C.; supervision, M.L.G.-R. and A.C.-C.; project administration, A.C.-C.; funding acquisition, P.B.-S. All authors have read and agreed to the published version of the manuscript. [/fig]
[table] Table 2: Results of the permeation studies expressed by mean and SD (n = 3). NQ = non quantifiable; value below LOQ, * skin extracted not corrected by percentage recovery. [/table]
|
Proteomic and biological profiling of extracellular vesicles from Alzheimer's disease human brain tissues
Introduction: Extracellular vesicles (EVs) from human Alzheimer's disease (AD) biospecimens contain amyloid beta (A ) peptide and tau. While AD EVs are known to affect brain disease pathobiology, their biochemical and molecular characterizations remain ill defined.Methods: EVs were isolated from the cortical gray matter of 20 AD and 18 control brains. Tau and A levels were measured by immunoassay. Differentially expressed EV proteins were assessed by quantitative proteomics and machine learning.Results: Levels of pS396 tau and A 1-42 were significantly elevated in AD EVs. High levels of neuron-and glia-specific factors are detected in control and AD EVs, respectively. Machine learning identified ANXA5, VGF, GPM6A, and ACTZ in AD EV compared to controls. They distinguished AD EVs from controls in the test sets with 88% accuracy.Discussion: In addition to A and tau, ANXA5, VGF, GPM6A, and ACTZ are new signature proteins in AD EVs.
# Background
Alzheimer's disease (AD) is the most common form of dementia affecting nearly 50 million people worldwide. Neuropathologically, disease is characterized by amyloid plaques formed by extracellular aggregation of amyloid beta (A ) peptide and intracellular accumulation of neurofibrillary tangles (NFTs). These are formed in brain tissue by the hyperphosphorylated and misfolded microtuble-associated protein tau.As AD progresses, A and tau aggregates spread throughout the brain in a spatiotemporal manner.among others), lipids and proteins that are transferred from cell to cell, and are found in blood, urine, and cerebrosprinal fluid (CSF).In the central nervous system (CNS), it has been reported that ADassociated pathogenic proteins in brain EVs including tau and A oligomers play important roles in AD pathogenesis.Moreover, it has been reported that inhibition of EV synthesis reduced amyloid plaque deposition in the mouse model of AD, and stimulation of EV secretion increased intracellular transfer of prion protein in vitro.EVs are involved in the extracellular enzymatic degradation of A and promote both A aggregation and clearance by microglia.Moreover, models of neuron-to-neuron transfer of tau seeds by EVs were reported.Our own prior work demonstrated that microglia spread tau by EV secretion and that reducing EV synthesis significantly reduces tau propagation.One mechanism centers around bridging integrator 1 (BIN1), which is associated with the progression of tau pathology and observed by its abilities to alter tau clearance and by promoting the release of tau-enriched microglial EVs.While EVs recovered from human and mouse brain tissues were examined by morphology, proteomics, and RNA analyses, no comprehensive and quantitative proteomics databases have yet been acquired for AD human brain tissues. Herein, we provide the first proteomic profiling of EVs isolated from post mortem AD and cognitively impaired control brain tissues. The analyses were combined with machine learning and quantitation of A and tau species by epitopespecific enzyme-linked immunosorbent assay (ELISA). Machine learning identified and distinguished protein signatures of AD brain-derived EVs from controls with high degrees of accuracy.
# Methods
## Brain sample acquisitions
Two cohorts of brains were used in this study. The first cohort was obtained from the University of Nebraska Medical School (11 AD and 9 control) and the Greater Los Angeles Veteran's Affairs Hospital (9 AD and 9 control) as part of NIH NeuroBioBank, which were matched for age and sex. The second cohort was obtained from the NIH NeuroBioBank
## Purification of evs from human brain samples
0.5 g of largely gray matter tissue from the frontal cortex of deceased AD or control cases were processed for EV extraction based on reported method with some modifications (see supporting information for detailed methods). 21
## Protein concentrations
The bicinchoninic acid (BCA) assay was used to determine protein concentration for each sample using Pierce BCA protein assay kit (# 23225
Pierce) (see supporting information for detailed methods).
## Enzyme-linked immunosorbent assay
ELISAs were performed to assess levels of t-tau, p-tau, A 1-40 and A 1-42, and ANXA5 (see supporting information for detailed methods).
# Nanoparticle tracking analysis
All samples were diluted in dfPBS at least 1:1000 or more to get particles within the target reading range for the Nanosight 300 machine (Malvern Panalytical Inc) (see supporting information for detailed methods).
## Transmission electron microscopy
The EV isolated from AD or control brain tissue were analyzed by transmission electron microscopy (TEM; see supporting information for detailed methods).
## Mass spectrometry
The EV samples were subjected to chemical treatment, tryptic digestion, and liquid chromatography (LC)-electrospray ionization (ESI) tandem mass-spectroscopy (MS/MS) analysis (see supporting information for detailed methods).
## Sequence database
The raw LC-MS/MS data were converted into mZML format using Pro-teoWizard msConvert and analyzed using using PeaksDB and Peak-sPTM using Peaks Studio version 8.0 (Bioinformatics Solutions, Inc.,
## Waterloo, on, canada) against the uniprot/swissprot database for
Homo sapiens with a 0.1% false discovery rate (FDR) and at least two unique peptides (see supporting information for detailed methods). 24
## Statistical analyses
Statistical analysis was conducted using IBM SPSS software ver-
## Machine learning
The protein biomarkers to distinguish patients with AD from controls were selected using least absolute shrinkage and selection operator (LASSO) on the proteomics data from the training set (n = 21), in which each patient's true state is labeled. An ensemble machine learning classifier to evaluate the performance of the selected proteins was developed as described (see supporting information for detail methods).The machine learning generated model's performance was evaluated on a separate, user-blinded test set (n = 17).
# Results
A . There were no statistical differences in the demographics between AD and controls with the exception of post mortem intervals (PMI) of the validation set, which will be discussed in the validation study (see.
## Biochemical characterization of brain-derived evs
The experimental workflow is summarized in . The EV samples were isolated from 20 AD and 18 sex-and age-matched cognitively unimpaired controls using the discontinuous sucrose gradient ultracentrifugation as previously described with modifications F I G U R E 1 Biophysical and biochemical characterization of extracellular vesicles (EVs) isolated from Alzheimer's disease (AD) and control (CTRL)brain tissues: A, Schematic of extracellular vesicle isolation protocol from human frozen brain tissue (see supporting information for detailed methods). B, Left: particle numbers of brain-derived EV fraction from control (CTRL) or AD by nanoparticle tracking analysis. P = .6075 by Mann-Whitney test. Right: Particle size of brain-derived EV fraction. P = .0095 by Mann-Whitney test. C, Transmission electron microscopy image of frozen human brain-derived EVs. Scale bar = 100 nm. Left: CTRL, Right: AD. D, Comparison of total tau and tau phosphorylated at threonine 181, serine 199, and serine 396 in EVs. pS 396 tau; P = .0375 by Mann-Whitney test. E, Comparison of amyloid beta 1-40 or 1-42 in EVs. A 1-42; P < .0001 by Mann-Whitney test. F, Scattered plot of brain tissue homogenates and brain-derived EVs. Left: pS396 tau (r = 0.4897, P = .005 using two-tailed t-test), Right: A 1-42 (r = 0.5632, P = .0005 using two-tailed t-test) (see Materials and Methods). This technique of EV isolation has been successfully used to isolate EVs from frozen mouse brain tissues. In addition, we performed the quantitative proteomics analysis of human brain tissue homogenates and purified EV samples. The tetraspanins and ESCRT proteins were enriched in the EV samples, and contamination of non-EV molecules such as nucleus, mitochondria, ER, and Golgi-related proteins as indicated in MISEV2018 guidelineswere diminished in the EV samples The mode size distribution for EVs was significantly different and peaked at 122 nm for AD and 131 nm for controls (P = .0095) . The EVs isolated from frozen brain tissue showed cap-shaped morphology by transmission electron microscopy (TEM; . We next measured the concentration of total tau (t-tau) and
p-tau at threonine 181 (pT181 tau), serine 199 (pS199 tau), and serine 396 (pS396 tau) in lysed EVs by ELISA. The levels of t-tau, pT181 tau, and pS199 tau showed no significant differences between AD and controls (t-tau: P = .398, pT181 tau: P = .7235, and pS199 tau: P = .4384; and . Conversely, pS396 tau was significantly increased in AD-brain derived EVs over controls (pS396 tau: P = .0375; and S1). Moreover, we observed a significant increase in A 1-42 in AD-derived EVs over controls (P < .0001), but not in A 1-40 (P = .119; and . The amount of pS396 and A 1-42
in the brain homogenate tissue were quantified by ELISA, and we performed the bivariate correlation analysis between the homogenate samples and EV samples purified from the same AD brain samples.
There is a significant difference in pS396 tau and A 1-42 levels in the brain tissue homogenate between AD and controls (pS396 tau; P < .0001, A 1-42; P = .0001). In addition, there is a significant positive correlation between the brain tissue homogenate and the EVs. The results suggest increased pS396 and A 1-42 level in the brain tissue might be related with elevation of pS396 and A 1-42 in the EVs (pS396; r = 0.4987, P = .0050, A 1-42; r = 0.5632, P = .0005; .
## Proteomic profiling of brain-derived evs
We performed a label-free Nano-LC-MS/MS analysis of 38 EV samples for proteomic profiling. Across both cohorts, a total of 1088 proteins were identified with at least two unique peptidesand Tables to the AD group were linked to mitochondria metabolism known to be dysfunctional in AD brain 28. Interestingly, in pathway analysis by DAVID, neurodegenerative disorders, including AD and Parkinson's and Huntington's diseases were enriched in common and unique proteins.shows the peptides identified in AD. Notably, A sequence was identified in APP fragments.shows the AD pathway from KEGG pathway analysis based on 68 proteins identified in the AD group, which are designated with red stars. The list of AD pathwayrelated proteins is provided in in supporting information. Proteins known to play an important role in AD pathogenesis, such as APP, APOE, tau, and NADH-ubiquinone oxidoreductase (Cx I-V), were all identified in the AD group, although they were not unique to this group.
## Analysis of label-free quantitative proteomics comparison of ad and control brain-derived evs
Label-free quantitative proteomics analysis was performed using PEAKS studio software. A total of 949 proteins were quantified and in supporting information). The 934 quantified proteins were common between AD and control groups. Between these groups, three proteins were uniquely expressed in the controls, while 12 proteins were uniquely expressed in the AD group. The principal component analysis (PCA) showed a marginal separation of the two groups . shows the volcano plot of the common 289 proteins that were detected in > 50% of the group (AD: n > 10 and controls: n > 9). Among these proteins, 15 proteins were significantly upregulated and three proteins were significantly downregulated in AD compared to the control group (as determined by P < 0.05 and log 2 fold change threshold of > 1 or < −1; and . The expression levels of 18 molecules identified in the AD group relative to the control group are displayed in a heatmap . We next searched for brain cell-type-specific molecules within the EV proteomics dataset using the mouse proteomics dataset as a reference.The top 100 ranked cell type-specific molecules, which have at least two-fold change in concentration in the cell type of interest over the other cell types, were tested with our EV proteomics dataset . The distribution of these markers indicates that in the human brain, 49% of the identified molecules were related to neuronal origin, whereas the other 50% of EV proteins are related to glial origin, including microglia, astrocytes, and oligodendrocytes. Moreover, using label-free quantitative value, differences in the expression of cell type-specific marker molecules between AD and controls were seen . Interestingly, neuron-specific molecules were enriched in the control group , blue), while glia-specific marker molecules were enriched in the AD group . These results suggest that glia may proliferate upon inflammatory response or accelerate
## Machine learning to identify distinctive ad brain-derived ev proteins
To discover a combination of protein molecules that can accurately distinguish AD EVs from controls, the label-free quantitative proteomics dataset was analyzed using a machine learning method. For this purpose, we split the proteomics dataset into an unblinded training subset (AD: n = 11; control: n = 10) and a blinded testing subset (AD: n = 9; control: n = 8). The ensemble machine learning model was built using only the data from the training subset, and then the accuracy of the diagnosis was determined using only the blinded testing set.
We found that a panel that included annexin-A5 (ANXA5), Neurosecretory protein VGF (VGF), neuronal membrane glycoprotein M6-a (GPM6A), and alpha-centractin (ACTZ), selected by the LASSO algorithm, resulted in an area under the ROC curve (AUC) of 0.95 within the training setand in supporting information).
We then examined the accuracy in the independent blinded test set using the four proteins in the dotted green box. Using this model, we achieved an 88% accuracy (AUC = 0.97) in identifying AD patients using the panel that consisted of the four proteins.
Further, we ran two control experiments that randomly selected four proteins from a total possible 949 proteins to form the diagnosis panel (repeated 20 times, AUC = 0.58, accuracy = 55%) and shuffling the true labels of the subjects within the training set (AUC = 0.47, accuracy = 48%). The control's AUC was significantly worse than using the four-protein panel's AUC (P < .001).shows the scatter plot of four selected proteins, which were significantly differentially expressed between AD and control groups The statistical significance of the differences were calculated using student's t test.. Although there was statistical difference in PMI between AD and control cases in this cohort , Pearson's correlation analysis of PMI and ANXA5 levels showed no significant correlation (r = −0.165, P = .149). Thus, this is not due to the difference in PMI between the two groups. Interestingly, ANXA5 expression level shows a tendency to increase along with Braak stages in an AD-dependent manner. Therefore, ANXA5 is a potential EV molecule for both distinguishing AD and control EVs and as a surrogate marker for Braak stage.
# Discussion
The biophysical properties of EVs isolated from unfixed post mortem human brain tissues, quantitative analysis of tau and A species, and label-free quantitative proteomic profiling by Nano-LC-MS/MS analyses were performed. pS396 tau and A 1-42 levels were significantly increased in AD brain-derived EVs compared to controls. A total of 1088 unique proteins from brain-derived EVs were found to be enriched as extracellular exosomes molecules. We also quantified 949 proteins by label-free quantitative proteomic analysis, which were enriched in neuron-specific molecules in controls and glial cell typespecific molecules in the AD group. We used the feature selection algorithm LASSO to select a panel of protein biomarkers that could accurately diagnose AD, including ANXA5, VGF, GPM6A, and ACTZ. It is important to note that the feature selection algorithm we used identifies the best panel of biomarkers for diagnosis, but is not necessarily a list of the most informative individual biomarkers. LASSO has the property that if there were, for example, two excellent biomarkers for AD, but which correlated highly to one another, only one of these biomarkers would be included in the panel because including both would bring only redundant information.Using the validation cohort with the larger sample size, the increased protein level of ANXA5 in the AD group, compared to controls, was confirmed by ELISA.
Previous reports indicate pT181 tau to be an early PTM associated with AD, pS199 tau modification is thought to promote tau accumulation, and pS396 tau modification is associated with tau seeding activity and aggregation.The PTM of tau at either pT181, pS199, or/and pS396 may facilitate the recruitment of tau in EVs, but neither pT181 nor pS199 tau is enriched in AD brain-derived EVs. This was unexpected because both pT181 and pS396 tau are elevated in neuronderived exosomes in plasma samples obtained from AD and prodromal AD cases.One potential explanation is that phosphorylation of specific sites on tau may be more preferentially incorporated into EV presumably via their ubiquitination, which is necessary for their multivesicular body (MVB) incorporation. In general, protein phosphorylation induces ubiquitination of lysine residues proximal to the phosphorylation sites, and it is necessary for ESCRT-mediated incorporation of ubiquitinated molecules into MVBs. Ubiquitination sites known for PHF-tau are K254, K311, and K353,and phosphorylation of tau at S396 may facilitate its MVB sorting through ubiquitination of those lysine residues. In addition, we also observed tau fragments from the mid-region (156 to 406), which is inclusive of proline-rich domains (151 to 240), and microtubule binding repeat domains (24 to 369). Tau can be cleaved by various proteases including calpain-1 and -2 (at R230), thrombin, cathepsins, asparagine endopeptidase (at D255 and N368), caspase-2 (at D314), and caspase-3 (at D421).A further investigation is needed to determine how tau is truncated, phosphorylated, and enriched in AD brain-derived EVs.
In the present study, we observed that the size of EVs derived from AD brain samples was smaller than from control EVs. It is possible that cholesterol level might be related to EV size, as small EVs contain sig- The gene ontology of proteins identified in AD group showed mitochondria metabolism category. It is well known that mitochondrial dysfunction occurs in AD and by A stimulation.Recombinant tau mutant mimicking S396/S404 phosphorylation can enhance A -induced mitochondrial damage and neuronal dysfunction,and mitochondrial dysfunction can be an upstream inducer of tau phosphorylation in AD.In addition, mitochondrial dysfunction leads to their sorting to endolysosomal system and MVBs, which can release mitochondria-derived vesicles into the extracellular space as exosomes.A number of previous studies in CSF and brain tissue have reported markedly lowered concentration of VGF in AD cases compared to controls. 51-54 GPM6A expression level was reported to be downregulated in the hippocampus of human AD brain tissues and transgenic AD mouse brains.Because GPM6A homolog is located in neuropil in Drosophila melanogaster, 57 elevated GPM6A in AD brain EVs may be an indicator of the neuropil loss. Indeed, GPM6A is reported to cluster in lipid rafts upon palmitoylation, which are also enriched in sphingolipids and cholesterol.Thus, reduction in GPM6A in brain tissue may be negatively correlated with GPM6A enrichment in EVs. In our study, ANXA5 was detected in brain-derived EVs from validation cohort, but VGF, GPM6A, and ACTZ were undetectable by commercial ELISA kits.
Further study is necessary to validate these molecules with higher sensitivity ELISA. Finally, the combination of A , tau, and cell type-specific molecules from brain cells, including ANAX5, VGF, GPM6A, or ACTZ may serve as potential biomarker candidate molecules in AD patient body fluid samples.
# Acknowledgments
The author thanks Maria Ericsson (Electron Microscopy Facility, Harvard Medical School) for electron microscopic imaging services, participating laboratories of NIH NeuroBioBank for providing frozen human brain tissue specimens, and Li Wu (University of Nebraska Medical
Center) for providing human brain tissue specimens.
# Funding information
This work is in part funded by Alzheimer's Association AARF- The t test was caluculated by Mann-Whitney test. E, Scattered plot of candidate molecules and AD pathogenic molecule in brain-derived EVs. Left: GPM6A and pS396 tau (r = 0.380, P = .019 using two-tailed t-test). Center: GPM6A and A 1-42 (r = 0.387, P = .016). Right: VGF and A 1-42 (r = −0.538, P = .002). F, Left: Significant difference in total ANXA5 to total EV protein between AD and CTRL group by ELISA (P = .0042 by Mann-Whitney test) (AD = 42, CTRL = 36). Right: Braak stage-dependent increase in the ANXA5 expression level to total EV protein in AD-dependent manner Issadore, Weiming Xia, Joseph Zaia, and Tsuneya Ikezu edited the paper.
## Orcid
Tsuneya Ikezu https://orcid.org/0000-0002-3979-8596 |
Extraosseous Ewing's sarcoma / primitive neuroectodermal tumor of the sacral nerve plexus
[bib_ref] Ewing's sarcoma family of tumors: Ewing's sarcoma of bone and soft tissue..., Horowitz [/bib_ref] [bib_ref] Primary spinal epidural extraosseous Ewing's sarcoma: Report of five cases and literature..., Mukhopadhyay [/bib_ref] [bib_ref] Primary spinal epidural extraosseous Ewing's sarcoma, Kasper [/bib_ref] [bib_ref] Extraosseous Ewing's sarcoma: A study of 42 cases, Rud [/bib_ref] [bib_ref] Ewing's sarcoma in the spinal nerve root: A case report and review..., Isefuku [/bib_ref] [bib_ref] Primary spinal epidural extraosseous Ewing's sarcoma: Report of five cases and literature..., Mukhopadhyay [/bib_ref]
## Case report
A 9-year-old apparently healthy boy presented with a gradually increasing painful swelling of 4 months' duration over his lower back. The pain was localized to the lower back, severe in intensity, and piercing in nature; it did not radiate and increased with movements. It was associated with urinary incontinence for the same duration. There was no history of trauma. There was no history of any significant medical or surgical illness in the child or his family.
Local examination of the site revealed a small tender swelling in the lumbosacral region, with no signs of inflammation, no visible scar / sinus, and no abnormal pulsations. The swelling was fixed to the underlying structures and the overlying skin was freely mobile. The child was anemic, with hemoglobin of 10.2 gm%; the ESR was 62 mm in the first hour (Westergren's method). The other routine investigations were within normal limits.
## Radiographs of the pelvis and lumbosacral spine revealed
# Abstract
We report an unusual case of Ewing's sarcoma / primitive neuroectodermal tumor (PNET) of the sacral nerve plexus in a 9-yearold boy who presented with a soft tissue swelling and severe piercing pain in the lower back region. MRI of the lumbosacral spine showed a lobulated soft tissue mass with clubbed finger-like projections along the path of the sacral nerves, which had caused widening of the spinal canal and the sacral foramina (S2-S4 level). There was presacral extension and posterior scalloping of the sacral vertebrae. Histopathology of the lesion confirmed Ewing's sarcoma / PNET of the sacral spinal nerve plexus. The patient responded favorably to chemotherapy and radiotherapy, showing clinical and radiological improvement. An open biopsy of the lesion revealed a proliferative growth. Microscopically, there were monomorphic-appearing small round cells arranged in sheets, with dispersed congested vessels. The cytoplasmic borders were indistinct, producing a syncytial appearance. Mitoses were seen. The nuclei appeared to be round to cleaved. These features are suggestive of Ewing's sarcoma / PNET [ [fig_ref] Figure 3: Histopathology slide [/fig_ref] ]. Based on the clinicopathological and radiological findings, we diagnosed Ewing's sarcoma / PNET of the sacral nerve plexus.
The patient was started on chemotherapy (VAIA regimen: vincristine, adriamycin, ifosfamide, and actinomycin D), which was followed by radiotherapy to the tumor bed. Radionuclide scan of the whole body (Tc-99m-MDP) after six cycles of chemotherapy revealed no evidence of metastases. The patient responded well to the chemotherapy and radiotherapy and became asymptomatic. Repeat scans showed almost complete resolution of the mass.
# Discussion
EES / PNET is a rare malignant small round cell neoplasm of undifferentiated mesenchymal origin. [bib_ref] Primary spinal epidural extraosseous Ewing's sarcoma, Kasper [/bib_ref] It was first described by in four patients who had paravertebral soft tissue tumors with a histologic appearance resembling Ewing's sarcoma. EES / PNET is usually seen in the second or third decades, with the reported incidence in children below 10 years being 0.5%.EES / PNET has equal frequency in both males and females, as contrasted with Ewing's sarcoma of the bone (ESB), where there is a male to female ratio of 2:1. [bib_ref] Primary spinal epidural extraosseous Ewing's sarcoma, Kasper [/bib_ref] The sites most commonly involved by EES / PNET are the extremities, mainly the lower limbs. Other common sites are the head and neck region, the paravertebral region, and the pelvis. [bib_ref] Primary spinal epidural extraosseous Ewing's sarcoma, Kasper [/bib_ref] In the present case, the tumor arose from the sacral spinal nerve roots; this is infrequently reported, [bib_ref] Ewing's sarcoma in the spinal nerve root: A case report and review..., Isefuku [/bib_ref] although there are reports of cases occurring in the cervical and lumbar paravertebral and epidural regions. [bib_ref] Primary spinal epidural extraosseous Ewing's sarcoma: Report of five cases and literature..., Mukhopadhyay [/bib_ref] [bib_ref] Primary spinal epidural extraosseous Ewing's sarcoma, Kasper [/bib_ref] [bib_ref] Extraosseous Ewing's sarcoma: A study of 42 cases, Rud [/bib_ref] On MRI, the tumor is usually isointense to the muscles on T1W and hyperintense on T2W images, with enhancement on postcontrast scans, as in our case. [bib_ref] Spinal epidural Extraskeletal Ewing sarcoma: MR findings in two cases, Shin [/bib_ref] The tumor in our patient, appeared to extend along and follow the path of the sacral nerves.
The classical histopathological features of ESFT consist of uniform round cells, with irregularly shaped chromatic nuclei surrounded by scanty cytoplasm. Mitotic figures may be seen. Special cellular arrangements, like rosettes or differentiations, are not often seen. The cells often show immunohistochemical positivity for various neurofilaments, CD99, and S-100. [bib_ref] Primary spinal epidural extraosseous Ewing's sarcoma, Kasper [/bib_ref] [bib_ref] Ewing's sarcoma in the spinal nerve root: A case report and review..., Isefuku [/bib_ref] Ewing's sarcoma / PNET arising from the sacral nerve plexus is only rarely reported and there is a need to differentiate it from other tumors arising in this region. The benign nerve sheath tumors such as neurofibromas are well-defined lesions, with a 'target' appearance due to a hyperintense rim and a hypointense center on T2W and contrast-enhanced T1W images (due to collagen and condensed Schwann cells); they are also often associated with neurofibromatosis-1. Schwannomas (neurilemmoma / neurinoma) are well-encapsulated, slow-growing benign nerve sheath neoplasms seen most commonly in the lumbar region, pushing the cord, conus, or hilum to the contralateral side; these tumors have a cystic component (40% cases) and are seen as central low-signal foci with an enhancing periphery on postcontrast T1W images.Spinal ependymomas (myxopapillary) usually involve the filum terminale / conus medullaris and there is symmetric cord expansion and cavitation; these are hypointense on T1W images and hyperintense on T2W images. Hypointensity of the tumor margin on T2W images, is a sign suggestive of, but not pathognomonic of, ependymomas.Non-neuraxis neoplasms such as lymphomas (particularly non-Hodgkin's lymphoma) present in older age-groups (peak incidence 40-60 years) and spread via the subarachnoid space, causing diffuse rope-like thickening of the spinal nerve roots. These lesions are often missed on MRI without contrast enhancement, as they appear isointense to the spinal cord on precontrast images. They, however show marked enhancement on postcontrast fat-suppressed MRI images.Differentiation of these lesions requires imaging studies along with histopathological confirmation.MRI imaging is considered the imaging modality of choice, being better than CT scan for demonstrating the pattern of tumor extension and for delineating the soft tissues. [bib_ref] Spinal epidural Extraskeletal Ewing sarcoma: MR findings in two cases, Shin [/bib_ref] Narula et al.: EES/PNET of the sacral nerve plexus Depending on the site of the tumor and its extension, treatment can be with surgery, chemotherapy, and radiotherapy, used separately or in various combinations. [bib_ref] Primary spinal epidural extraosseous Ewing's sarcoma, Kasper [/bib_ref] [bib_ref] Extraosseous Ewing's sarcoma: A study of 42 cases, Rud [/bib_ref] Our patient responded favorably to chemotherapy and radiotherapy.
In conclusion, the present case highlights a rare case of EES / PNET presenting in the first decade of life and arising from the sacral nerve plexus. The mass was seen to grow along the sacral nerves and the diagnosis could he established only after a biopsy. The complaints resolved following therapy.
[fig] Figure 1, Figure 2: (A,B): Anteroposterior (A) and lateral (B) plain radiographs of the lumbosacral (LS) spine reveal posterior scalloping of the sacral vertebrae (S2-S4) (arrows), widened and ill-defined sacral foramina (arrowheads), and an overlying right-sided soft tissue haze (curved arrow) (A-D): T1W axial (A) and T2W mid-sagittal (B) MRI images of the LS spine show a lobulated soft tissue mass (arrows), isointense to pelvic muscles on T1W and hyperintense on T2W images with areas of necrosis; both intraspinal and extraspinal components cause widening of the sacral spinal canal and sacral foramina. T2W coronal image (C) shows clubbed, finger-like extensions along the path of the sacral nerves.Axial T1W contrast-enhanced image (D) shows intense enhancement with areas of low intensity within, corresponding to necrotic areas. Involvement of the pyriformis (arrow) and right erector spinae muscles (arrowhead) can be seen foramina; there was an overlying right-sided soft tissue haze. MRI of the lumbosacral spine revealed a lobulated soft tissue mass with both intraspinal and extraspinal components in the sacral region (S2-S4). It was isointense to pelvic muscles on T1W and hyperintenseenhancement on postcontrast scans [Figure 2D].It had caused widening of the spinal canal and the sacral foramina and showed clubbed finger-like projections along the path of the sacral spinal nerves. There was presacral extension and infiltration of the pyriformis and right erector spinae muscles bilaterally. Based on the MRI findings, we made the diagnosis of a complex nerve sheath tumor / soft tissue sarcoma of the sacral nerve roots. [/fig]
[fig] Figure 3: Histopathology slide (hematoxylin and eosin) shows proliferative growth of monomorphic-appearing small round cells arranged in sheets, with dispersed congested vessels. The nuclei appear round to cleaved [/fig]
|
Robot-assisted posterior retroperitoneoscopic adrenalectomy: single port access
# Introduction
Recent advances in the technologic and surgical instrument fields have resulted in the developments of several novel surgical techniques, such as, Natural Orifice Transluminal Endoscopic Surgery and Laparoendoscopic Single-site Surgery, which minimize perioperative co-morbidities and increase cosmetic benefits [bib_ref] Nomenclature of natural orifice translumenal endoscopic surgery (NOTES) and laparoendoscopic single-site surgery..., Box [/bib_ref]. However, these new techniques have several limitations due to restrictive instrumentation movement because of the small access ports used and relatively low-quality images produced.
In 1992, Gagner et al. [bib_ref] Laparoscopic adrenalectomy in Cushing's syndrome and pheochromocytoma, Gagner [/bib_ref] first introduced the use of laparoscopic adrenalectomy for the treatment of small adrenal tumors, and since, laparoscopic adrenalectomy has become a gold standard in adrenal gland surgery. More recently, some minimally invasive trials have been conducted on single access surgery on the adrenal gland [bib_ref] Single access retroperitoneoscopic adrenalectomy (SARA)--one step beyond in endocrine surgery, Walz [/bib_ref].
In 2001, Horgan and Vanuno [bib_ref] Robots in laparoscopic surgery, Horgan [/bib_ref] described a robot-assisted laparoscopic adrenalectomy technique, and since, several institutions have reported their experiences of this technique using the da Vinci surgical robot system (Intuitive Surgical, Inc., Sunnyvale, CA, USA). The multi-articulated instruments of this robotic system and the 3-D, magnified image provided by its stable camera platform greatly benefit endoscopic adrenalectomy.
In this study, we introduce our first experiences of robot-assisted posterior retroperitoneoscopic adrenalectomy using single-port access and the da Vinci system. To the best our knowledge, this is the first case report issued on robotic single port adrenalectomy using a posterior retroperitoneal approach.
## Case report
A 56-year-old woman was referred to our endocrine surgery division due to the presence of a 1.5 cm sized mass in her left adrenal gland [fig_ref] Figure 1: Computed tomography scan of the patient showing a 1 [/fig_ref] , which was incidentally found on an abdominal computed tomography scan during health screening. She had been taking 25 mg atenolol, 4 mg lacidipine, and 80 mg telmisartan daily under a diagnosis of hypertension for 12 years. Her initial blood pressure was 130/80 mmHg, and her routine laboratory tests, which included renal function tests, electrolyte, and 24-hour urinary catecholamines were normal. In the standing position, her morning plasma aldosterone level was 398 pmoL/L (standard range, 38.9 to 306.8 pmoL/L), her concurrently measured plasma renin activity was 0.01 μg/L/hr (standard range, 1.31 to 3.95 μg/L/hr), and her plasma aldosterone to renin activity ratio was 3,980 pmoL/L per μg/L/hr. In view of the hypertension and a raised plasma aldosterone to renin activity ratio, a working diagnosis of hyperaldosteronism-induced hypertension was made. We discussed surgical options with the patient and decided to perform robot-assisted single port retroperitoneoscopic adrenalectomy.
The route used to approach the adrenal gland was slightly modified version of that described by Walz et al. [bib_ref] Posterior retroperitoneoscopy as a new minimally invasive approach for adrenalectomy: results of..., Walz [/bib_ref].
Under general anesthesia, the patient lies in prone, jack-knife position with hip joints bending at a right angle.
Robotic adrenalectomy and posterior retroperitoneal single port access thesurgery.or.kr S23 Soft pillows were applied at the ventral portion of the chest and pelvis [fig_ref] Figure 2: Operative views of robotic posterior retroperitoneoscopic adrenalectomy using single access [/fig_ref]. Initially, a 3 cm-sized transverse skin incision was made just below the lowest tip of the 12th rib [fig_ref] Figure 2: Operative views of robotic posterior retroperitoneoscopic adrenalectomy using single access [/fig_ref] , and then subcutaneous and muscle layers were split and opened. After exposing the retroperitoneal space, the Glove port (NELIS, Bucheon, Korea) was applied to the skin incision and maintained pneumoretroperitoneum. CO2 was then insufflated to a pressure of 18 mmHg to create adequate working space. In 2001, Horgan and Vanuno [bib_ref] Robots in laparoscopic surgery, Horgan [/bib_ref] reported the first robot-assisted laparoscopic adrenalectomy, and since, more than 100 cases of robotic adrenalectomy using the transperitoneal approaches have been performed. However, no report on robot-assisted posterior retroperitoneoscopic adrenalectomy or on single port robotic adrenalectomy has been issued. Elias and colleagues reviewed the literature on robotic adrenalectomy and comparisons made between laparoscopic and robotic adrenalectomy. Operative times for robotic and laparoscopic cases were 99 to 188 and 82 to 131 minutes, respectively, and times to discharge were 4 to 6.7 and 3.4 to 6.9 days, respectively [bib_ref] The role of robotics for adrenal pathology, Hyams [/bib_ref]. Wu et al. [bib_ref] Comparison of robot-assisted laparoscopic adrenalectomy with traditional laparoscopic adrenalectomy -1 year follow-up, Wu [/bib_ref] demonstrated that robotic adrenalectomy is well tolerated and effective, has short-term outcomes, which are comparable with those of laparoscopic adrenalectomy, and has subjective advantages regarding fine motion, dexterity, and surgeon comfort. As compared with robotic adrenalectomy data previously reported [bib_ref] The role of robotics for adrenal pathology, Hyams [/bib_ref] , our results indicate that operative time was somewhat longer, which is probably because it was our first experience and more time was required to create an adequate working space and to effectively utilize robotic system instruments via the single access port.
Compared to robot-assisted transperitoneal adrenalectomy, robot-assisted single port retroperitoneoscopic adrenalectomy could not only reduce postoperative ileus, bacterial contamination, and intestinal complications because of not opening the peritoneal cavity, but also reduce postoperative pain because of using minimally invasive single-access approaches.
Our initial experiences of robot-assisted single port retroperitoneoscopic adrenalectomy did much to assure us of its safety and feasibility. We suggest that robot-assisted single port retroperitoneoscopic adrenalectomy be view-Jae Hyun Park, et al.
## S24
thesurgery.or.kr ed as a useful surgical option for the treatment of adrenal pathologies.
## Conflicts of interest
No potential conflict of interest relevant to this article was reported.
[fig] Figure 1: Computed tomography scan of the patient showing a 1.5 × 1.1 cm mass in the left adrenal gland. [/fig]
[fig] Figure 2: Operative views of robotic posterior retroperitoneoscopic adrenalectomy using single access. (A) Anatomic landmark, (B) skin incision, (C) instrumentation of each robotic arms, and (D) operative position and postoperative wound. [/fig]
[fig] A 10: mm robotic camera with 30 degree up view and three 5 mm robotic ports were inserted through 4 port of the Glove port. The camera was located at the center of incision and placed in its most cephalic portion. The Maryland dissector (Intuitive Surgical Inc.) was placed on the lateral side of the incision, and a harmonic curved shears (Intuitive Surgical Inc.) was placed on the medial side of the incision. The assistant port was placed on the most caudal portion and endoscopic metal suction tip (WISP, Munchen, Germany) was introduced. Using the Maryland dissector and the harmonic curved shears, retroperitoneal fatty tissue in the vessel free layer was detached from Gerota's fascia and pushed up for the identification of upper margin of the kidney. The adrenal gland and its tumor were then visualized and mobilized. Small vessels around the adrenal gland were then coagulated using the harmonic shears. The adrenal vein was found caudally, medial to the upper kidney pole, and ligated using the harmonic curved shears. The adrenal gland was then fully mobilized, and the adrenal gland and its tumor were delivered using a retrieval bag. The total operation time was 188 minutes, and the console time was 57 minutes. On the day of surgery, the patient started taking sips of water after full awakening, and on the 1st postoperative day, the patient gradually took a soft, general diet. Postoperatively, the patient received only an intravenous narcotic agent on the operative day. nalgesics were not required. The patient recovered well uneventfully and was discharged on the 3rd postoperative day. The permanent pathologic diagnosis was of adrenal cortical adenoma. Six month postoperatively, her blood pressure was well controlled without an antihypertensive agent. DISCUSSION Laparoscopic adrenalectomy can be performed in four ways, that is, by using transperitoneal approaches in the supine or lateral position or retroperitoneal approaches in the lateral or prone position. Walz et al. [6] demonstrated the feasibility, safety, wide applicability, and speed of the posterior retroperitoneal approaches to the adrenal glands. [/fig]
|
Cortical regulation of cell size by a sizer cdr2p
Cells can, in principle, control their size by growing to a specified size before commencing cell division. How any cell actually senses its own size remains poorly understood. The fission yeast Schizosaccharomyces pombe are rod-shaped cells that grow to ∼14 µm in length before entering mitosis. In this study, we provide evidence that these cells sense their surface area as part of this size control mechanism. We show that cells enter mitosis at a certain surface area, as opposed to a certain volume or length. A peripheral membrane protein kinase cdr2p has properties of a dose-dependent 'sizer' that controls mitotic entry. As cells grow, the local cdr2p concentration in nodes at the medial cortex accumulates as a measure of cell surface area. Our findings, which challenge a previously proposed pom1p gradient model, lead to a new model in which cells sense their size by using cdr2p to probe the surface area over the whole cell and relay this information to the medial cortex.
# Introduction
The fundamental process by which a cell controls its own size is not understood for any cell type. In actively dividing cells, growth, and size need to be coordinated for cells to maintain their size. In several cell types, cells have been shown to have a size threshold, in which they need to grow to a minimal cell size before committing to cell division [bib_ref] Cell size control in yeast, Turner [/bib_ref]. This mechanism however requires that cells somehow monitor their own size. The molecular mechanism for how size is sensed, and what aspect of size-surface area, volume, mass, linear dimensions etc-is monitored remains unknown.
The fission yeast Schizosaccharomyces pombe is an attractive eukaryotic model for cell size studies because of its highly regular dimensions, simple rod-shape, and growth patterns. During interphase, these cells grow from the cell tips at a nearly constant rate to approximately 14 µm in length before entering mitosis, when cell growth ceases until the next cell cycle [bib_ref] Growth in cell length in the fission yeast Schizosaccharomyces pombe, Mitchison [/bib_ref]. Genetic analyses in fission yeast have identified a pathway of conserved protein kinases for cell size control: the DYRK kinase pom1p is an inhibitor of the SAD family kinase cdr2p, which inhibits wee1p, which in turn inhibits the cell division kinase cdk1p [bib_ref] Negative regulation of mitosis by wee1+, a gene encoding a protein kinase..., Russell [/bib_ref] [bib_ref] The cdr2(+) gene encodes a regulator of G2/M progression and cytokinesis in..., Breeding [/bib_ref] [bib_ref] Polar gradients of the DYRK-family kinase Pom1 couple cell length with the..., Martin [/bib_ref] [bib_ref] A spatial gradient coordinates cell size and mitotic entry in fission yeast, Moseley [/bib_ref]. Loss of function of pom1 and wee1 leads to abnormally short cells, whereas loss of function of cdr2 leads to abnormally long ones. Interestingly, these factors largely localize to different sites in the cell. Pom1p localizes in cortical gradients emanating from cell tips [bib_ref] The cell-end factor pom1p inhibits mid1p in specification of the cell division..., Padte [/bib_ref] [bib_ref] A phosphorylation cycle shapes gradients of the DYRK family kinase Pom1 at..., Hachet [/bib_ref] [bib_ref] Noise reduction in the intracellular pom1p gradient by a dynamic clustering mechanism, Saunders [/bib_ref]. Cdr2p localizes to a medial band of plasma membrane protein complexes termed 'nodes', which overlie the medial nucleus [bib_ref] The GIN4 family kinase, Cdr2p, acts independently of septins in fission yeast, Morrell [/bib_ref]
# Results
Testing the pom1p gradient model for size control
To test the pom1p gradient model, we quantitatively analyzed pom1p in living cells expressing a functional pom1-tomato-dimer fusion protein at near-endogenous levels [bib_ref] The cell-end factor pom1p inhibits mid1p in specification of the cell division..., Padte [/bib_ref] [bib_ref] A phosphorylation cycle shapes gradients of the DYRK family kinase Pom1 at..., Hachet [/bib_ref] [bib_ref] Noise reduction in the intracellular pom1p gradient by a dynamic clustering mechanism, Saunders [/bib_ref]. Pom1p cortical gradients exhibit large cell-to-cell variability in intensity and distribution, fluctuate over time in individual cells, and show little consistent change with cell length [bib_ref] Noise reduction in the intracellular pom1p gradient by a dynamic clustering mechanism, Saunders [/bib_ref]. This variability, plus a short decay length relative to cell length, eLife digest Although different types of cells come in a variety of shapes and sizes, most cells are able to maintain a fairly consistent size and shape as they grow and divide. For example, the rod-shaped cells of the fission yeast S. pombe grow to be 14 microns long before dividing in the middle to form two new cells. This prevents any single cell becoming too large or small.
A similar phenomenon has been observed in other types of cells, so it is clear that cells must be able to measure their own size, and then use that information to trigger cell division. A number of proteins that regulate cell size and cell division in fission yeast have now been identified. These proteins form a pathway in which a protein called pom1p inhibits another protein, cdr2p, which in turn causes a third protein, cdk1p, to start the process of cell division. However, the details of the measurement process and the property that the cells are actually measuring-surface area, volume, mass or something else-remain mysterious.
Pan et al. have now used imaging techniques and mathematical modeling to probe the distribution and movements of proteins in fission yeast cells. Their results do not support a previous model in which the cell uses the gradient of pom1p as a ruler to measure cell length. Rather, Pan et al. propose a new model in which the level of cdr2p is used to sense the size of the cell. Individual molecules of cdr2p come together to from clusters called nodes on the cell membrane. As the cell grows larger, more and more cdr2p proteins accumulate in these nodes, which are found in a band around the middle of the cell. When the cells reaches a critical cell size, the increased concentration of cdr2p at these nodes may help to trigger the start of cell division.
By examining cells that grow at different rates, Pan et al. show that the rate of accumulation of cdr2p in the nodes depends on how big the cells are, rather than on the length of time that has elapsed. Analysis of fission yeast cells of different shapes shows that cell division starts when the surface area of the cell grows to a certain value, as opposed to starting when the volume or length reach a given value.
Pan et al. also show that cdr2p binds to all parts of the cell membrane, not just to the nodes near the middle, and go on to provide a simple mathematical model showing how this property can allow cells to measure their surface area. However, as Pan et al. point out, this is probably just one component of a larger mechanism that tells cells when they need to divide. led us to question whether these gradients can function reliably as 'rulers'. One of the key predictions of the gradient-based model is that pom1p levels decrease on the medial cortex as cells grow. We measured pom1p concentration in a 3-µm region along the medial cortex, where cdr2p nodes are located. Using time-averaged data (reducing fluctuations in the gradient over time [bib_ref] Noise reduction in the intracellular pom1p gradient by a dynamic clustering mechanism, Saunders [/bib_ref] , we detected low but measurable intensities [fig_ref] Figure 1: Gradient distribution of pom1p is not the basis for cell size control [/fig_ref] , [fig_ref] Figure 1: Gradient distribution of pom1p is not the basis for cell size control [/fig_ref]. Importantly, measurements of pom1-tomato at the medial cortex showed no detectable decrease with cell length in a population of cells [fig_ref] Figure 1: Gradient distribution of pom1p is not the basis for cell size control [/fig_ref] , or in individual cells imaged over time [fig_ref] Figure 1: Gradient distribution of pom1p is not the basis for cell size control [/fig_ref]. These cortical measurements improve on previously reported pom1p measurements that integrate intensities over the whole cell [bib_ref] Polar gradients of the DYRK-family kinase Pom1 couple cell length with the..., Martin [/bib_ref] [bib_ref] A spatial gradient coordinates cell size and mitotic entry in fission yeast, Moseley [/bib_ref] , which have artifacts stemming from the normal exclusion of pom1p from the nucleus [bib_ref] Noise reduction in the intracellular pom1p gradient by a dynamic clustering mechanism, Saunders [/bib_ref] ; [fig_ref] Figure 1: Gradient distribution of pom1p is not the basis for cell size control [/fig_ref] , [fig_ref] Figure 2 -: figure supplement 4 [/fig_ref].
To further test the gradient model, we examined the effect of altering the gradient profile. Pom1-3GFP (pom1p fused to three tandem GFPs) produced a consistently steeper gradient profile than pom1-GFP or pom1-tomato fusions [fig_ref] Figure 1: Gradient distribution of pom1p is not the basis for cell size control [/fig_ref] ,F, [fig_ref] Figure 1: Gradient distribution of pom1p is not the basis for cell size control [/fig_ref]. The reason for this change was not clear, as these fusion proteins displayed similar dynamics (our unpublished observations). The gradient model predicts that a change in gradient distribution would lead to a significant change in cell size at division. However, we detected no differences in cell length at division between these pom1-tagged strains [fig_ref] Figure 1: Gradient distribution of pom1p is not the basis for cell size control [/fig_ref]. Consistent with this result, there were no significant differences in the intensities or number of cdr2p nodes (Figure 1-figure supplements 4C, 5). Overall, these data are inconsistent with the gradient model.
## Cdr2p at cortical nodes scales with cell size
To further investigate how this regulatory pathway may sense cell size, we focused on how cell size affects cdr2p and its behavior at these medial cortical nodes. Pom1p may exert its cell size effects in part by ensuring the proper localization of cdr2p nodes to this region [bib_ref] Pom1 kinase links division plane position to cell polarity by regulating Mid1p..., Celton-Morizur [/bib_ref] [bib_ref] The cell-end factor pom1p inhibits mid1p in specification of the cell division..., Padte [/bib_ref] [bib_ref] Polar gradients of the DYRK-family kinase Pom1 couple cell length with the..., Martin [/bib_ref] [bib_ref] A spatial gradient coordinates cell size and mitotic entry in fission yeast, Moseley [/bib_ref] (see below). We quantitated cdr2p levels using a functional cdr2-GFP construct [bib_ref] The GIN4 family kinase, Cdr2p, acts independently of septins in fission yeast, Morrell [/bib_ref]. Cdr2-GFP concentration in the whole cell remained approximately constant in interphase cells of various lengths [fig_ref] Figure 2 -: figure supplement 4 [/fig_ref]. Interestingly, the intensity of cdr2-GFP at the medial cortex increased with cell length [fig_ref] Figure 2 -: figure supplement 4 [/fig_ref] , [fig_ref] Figure 2 -: figure supplement 4 [/fig_ref]. The cortical area containing the nodes also increased slightly with cell length, but the relative change was less than for the cdr2-GFP intensity [fig_ref] Figure 2 -: figure supplement 4 [/fig_ref] ; [bib_ref] The GIN4 family kinase, Cdr2p, acts independently of septins in fission yeast, Morrell [/bib_ref]. Measurement of cdr2-GFP intensity within a 3-μm wide region of the medial cortex showed directly that the local cdr2p concentration in this region rises approximately twofold as cells grow through interphase Time-lapse imaging also revealed dynamics of cdr2p nodes. Mature cdr2p nodes, estimated to each contain an average of ∼90 cdr2-GFP molecules (Figure 2-figure supplement 1), moved very slowly and exhibited little change over hours (Video 1). FRAP studies, however, revealed that cdr2-GFP turned over with a t 1/2 of about 3 min within each node ( ). Thus, cell growth is accompanied by the formation of new nodes, leading to an increase in local cdr2p density. Imaging also revealed a subpopulation of less intense and more motile cortical nodes that may be newly assembling ones (Video 1).
## Cdr2p is a dose-dependent regulator of cell size
To determine if the cdr2p concentration is important in cell size control, we tested the effects of varying its expression level [fig_ref] Figure 3: Cdr2p is a dose-dependent regulator of cell size [/fig_ref]. Cdr2p was expressed from an nmt81 promoter, regulated by thiamine in the media. Mild cdr2p overexpression in the absence of thiamine (estimated 1.6-fold) caused cells to divide at abnormally short cell lengths. Consistent with previous studies [bib_ref] The cdr2(+) gene encodes a regulator of G2/M progression and cytokinesis in..., Breeding [/bib_ref] , higher levels of overexpression caused cytokinesis defects and accumulation of longer cells. Conversely, decreased cdr2p expression led cells to divide at much longer lengths, similar to a cdr2 null strain [bib_ref] The GIN4 family kinase, Cdr2p, acts independently of septins in fission yeast, Morrell [/bib_ref] [bib_ref] Polar gradients of the DYRK-family kinase Pom1 couple cell length with the..., Martin [/bib_ref] [bib_ref] A spatial gradient coordinates cell size and mitotic entry in fission yeast, Moseley [/bib_ref]. Thus, cdr2p is a dose-dependent regulator of cell size and mitotic entry [fig_ref] Figure 3: Cdr2p is a dose-dependent regulator of cell size [/fig_ref]. The persistence of cdr2-GFP in cells treated with the protein synthesis inhibitor cyclohexamide showed that the 4 of 24 Research article [fig_ref] Figure 1: Gradient distribution of pom1p is not the basis for cell size control [/fig_ref]. Continued on next page majority of cdr2p is highly stable in interphase cells [fig_ref] Figure 3: Cdr2p is a dose-dependent regulator of cell size [/fig_ref]. Together, these findings suggest that as the cell grows, cdr2p is a stable protein that is synthesized to maintain a constant concentration in the whole cell, and accumulates at the medial cortex, where it promotes mitotic entry in a concentration-dependent manner [fig_ref] Figure 3: Cdr2p is a dose-dependent regulator of cell size [/fig_ref].
## Cdr2p monitors cell size not time
A critical issue in cell size regulation is whether cdr2p levels at nodes report cell size or passage of time [bib_ref] Cell size control in yeast, Turner [/bib_ref] : is cdr2p a 'sizer' or a 'timer'? To test these possibilities, we examined cdr2p behavior in cells arrested for cell growth upon treatment with an actin inhibitor Latrunculin A [bib_ref] High rates of actin filament turnover in budding yeast and roles for..., Ayscough [/bib_ref] [bib_ref] Movement of a cytokinesis factor cdc12p to the site of cell division, Chang [/bib_ref]. Levels of a simple timer should continue to increase over time, even without cell growth, while a sizer would not increase without cell growth. Latrunculin A-treated cells exhibited no growth and no increase in cdr2-GFP levels at nodes . Next, we compared cdr2p in cells growing at different rates. We used for3Δ (formin) mutants, which are defective in cell polarity regulation and exhibit highly variable growth rates [bib_ref] Roles of the fission yeast formin for3p in cell polarity, actin cable..., Feierbach [/bib_ref]. This mutant allowed us to measure cells in the same microscope field with identical genotype and growth conditions, but with over twofold varying growth rates
## Cdr2p binds all over the cortex
Our findings raise the key question of how nodal cdr2p concentration is able to scale with cell size. In further characterizing the dynamic behavior of cdr2p, we found that in addition to cdr2p in nodes and a diffuse cytoplasmic haze, it also localized to dim, dynamic dots all around the cortex [fig_ref] Figure 5 -: figure supplement 2B, 'Materials and methods' [/fig_ref] , Video 2). This dim cortical population has not been described previously. Interestingly, the distribution of these dim cortical cdr2-GFP dots did not vary over the cell tip, and thus did not correlate with levels of pom1p at cell tips. Thus, cdr2p is able to bind to the whole surface of the cell.
## Mathematical models for size-dependent accumulation of cdr2p at nodes
Because it is not intuitively clear how these dynamic behaviors of cdr2p might cause it to concentrate in the nodal region in a cell-size-dependent manner, we developed a mathematical model to probe the mechanism responsible [fig_ref] Figure 5 -: figure supplement 2B, 'Materials and methods' [/fig_ref]. Based on our experiments, this model postulates that:
(1) the concentration of cdr2p in the cytoplasm is homogeneous and changes only slightly with cell length [fig_ref] Figure 1: Gradient distribution of pom1p is not the basis for cell size control [/fig_ref] ; (2) cytoplasmic cdr2p molecules can bind all over the plasma membrane [fig_ref] Figure 5 -: figure supplement 2B, 'Materials and methods' [/fig_ref] ; Video 2), and subsequently move rapidly by diffusion on the cortex; (3) cortical cdr2p molecules can transition to associate with a nodal region on the medial cortex. Note that the details of the formation and growth of individual nodes are beyond the scope of the model. Rather, we simply model the overall number of cdr2p molecules in the nodal region.
(4) Both cortical and nodal cdr2p can then unbind and return cdr2p to the cytoplasm; (5) cytoplasmic cdr2p can then diffuse rapidly before rebinding to the membrane. As the timescale of cell growth (hours) is much slower than the timescale of the cdr2p dynamics (minutes, [fig_ref] Figure 2 -: figure supplement 4 [/fig_ref] -figure supplement 4), we assumed that the molecular system is, at any given time, effectively in steady state. This steady-state assumption is also consistent with experimental findings that cdr2p levels at nodes are stable over time when cells are not growing . This model [fig_ref] Figure 5 -: figure supplement 2B, 'Materials and methods' [/fig_ref] was implemented by two mass-action equations for cortical and nodal cdr2p, and solved analytically [fig_ref] Figure 5 -: figure supplement 2B, 'Materials and methods' [/fig_ref]. Importantly, the model reveals how the cdr2p dynamics ensure a nodal cdr2p density that scales with cell size, or more specifically, with the surface area of the plasma membrane [fig_ref] Figure 5 -: figure supplement 2B, 'Materials and methods' [/fig_ref]. The simplicity of the model allowed us to clarify the two key elements important for this area sensing. The first is that the area of the nodal region must not scale proportionally with the total cell membrane area as the cell size increases [fig_ref] Figure 2 -: figure supplement 4 [/fig_ref]. We then have one process (cdr2p membrane association) that scales proportionally with cell area, with a second process (uptake of cortical cdr2p into the nodes), which does not. The second key element is that the nodal region receives information via cdr2p about the entire surface area of the cell. In this model, cdr2p needs to be able bind the membrane long enough to move on the membrane to reach the nodal region. The outcome is then a rising cdr2p nodal density with increasing cell area. Using the experimentally determined nodal/cortical areas, and with other parameters measured/constrained from our experiments [fig_ref] Figure 5 -: figure supplement 2B, 'Materials and methods' [/fig_ref] , 'Materials and methods'), we fitted the cdr2p density in the medial nodes as a function of cell length to that measured experimentally, with good results [fig_ref] Figure 5 -: figure supplement 2B, 'Materials and methods' [/fig_ref]. Note that wild-type S. pombe cells are rodshaped and have an approximately constant width, so that surface area and cell length are proportional to one another. A more sophisticated version of the same underlying model, including spatially varying cdr2p on the cortex, generated similar results [fig_ref] Figure 5 -: figure supplement 2B, 'Materials and methods' [/fig_ref] -figure supplement 1B-F, 'Materials and methods').
In addition, alternative models in which cdr2p does not need to diffuse long distances on the plasma membrane to the nodes are also consistent with the current findings. We analyzed a model in which cdr2p was now modified (e.g., phosphorylated) at the cortex and remains modified for a period even if it returns into the cytoplasm, from where it then can diffuse to and accumulate at the nodal region [fig_ref] Figure 2A ,: Figure 1-figure supplement 1A, Figure 2-figure supplement 1 [/fig_ref]. The underlying area-sensing mechanism was nevertheless conserved in this alternative model, with similar key elements as discussed above ( [bib_ref] Effects of heat shock and cycloheximide on growth and division of the..., Polanshek [/bib_ref]. Strain used: FC2688. Scale bar = 5 μm. (E) Total cell and nodal cdr2-GFP intensities in individual cells treated with cycloheximide over time. Total cell intensity was measured as in [fig_ref] Figure 2 -: figure supplement 4 [/fig_ref] n = 9 cells. Error bars = SD. DOI: 10.7554/eLife.02040.016
## Research article
Cdr2p scales with cell surface area not volume A central prediction of the modeling is that cdr2p is sensing the surface area of the cell: cdr2p at nodes should scale with surface area, and not, for instance, cell volume. To experimentally test if cdr2p scales with surface area or volume, we analyzed cdr2p levels in S. pombe mutants with different widths, so that surface area and volume are uncoupled. Rga2p and rga4p are Rho-GAPs involved in regulation of cell polarity and width [bib_ref] Regulation of cell diameter, For3p localization, and cell symmetry by fission yeast..., Das [/bib_ref] [bib_ref] Rga2 is a Rho2 GAP that regulates morphogenesis and cell integrity in..., Villar-Tajadura [/bib_ref] [bib_ref] Spatial control of Cdc42 activation determines cell width in fission yeast, Kelly [/bib_ref] ; rga2Δ mutants are thinner while rga4Δ mutants are fatter than wild type [fig_ref] Figure 6: Cdr2p and cell size at division scale with cell surface area [/fig_ref]. We measured surface areas and volumes in these cells ('Materials and methods'; [fig_ref] Figure 6: Cdr2p and cell size at division scale with cell surface area [/fig_ref] -figure supplement 1). In a group of interphase cells of similar surface area but of different volumes, nodal cdr2-GFP intensities correlated with surface area [fig_ref] Figure 6: Cdr2p and cell size at division scale with cell surface area [/fig_ref]. Conversely, in considering cells of similar volume but with a range of different surface areas, cdr2-GFP nodal intensity correlated with surface area and not volume [fig_ref] Figure 6: Cdr2p and cell size at division scale with cell surface area [/fig_ref]. These results thus suggest that nodal cdr2p scales with cell surface area, in agreement with the predictions of the mathematical models.
## Cells enter mitosis at a given surface area
These findings lead to another key prediction that cells enter mitosis at a specific cell surface area. We measured cell length, surface area and volume in wild-type, rga2Δ and rga4Δ strains in dividing cells; these dimensions are indicative of the size of the cells at entry into mitosis. These cells with different shapes entered mitosis with more similar cell surface areas but differing cell volumes and lengths [fig_ref] Figure 6: Cdr2p and cell size at division scale with cell surface area [/fig_ref]. All three strains exhibited average surface areas of 150 µm 2 ± 8 µm 2 , while the average volumes varied from 120 µm 3 to 150 µm 3 and average lengths from 11 µm to 15 µm. A more rigorous analysis based on Jensen-Shannon distances ('Materials and methods') showed quantitatively that the distributions of surface area were more similar than those for volume or length. These findings suggest that cells monitor their size at the G2/M transition by measuring their surface area.
## Regulation of cdr2p nodes by pom1p
We next examined how pom1p quantitatively affects cdr2p. In pom1Δ mutants, cdr2p is thought to be somehow more 'active' and promotes division at slightly shorter cell lengths than wild type [bib_ref] Polar gradients of the DYRK-family kinase Pom1 couple cell length with the..., Martin [/bib_ref] [bib_ref] A spatial gradient coordinates cell size and mitotic entry in fission yeast, Moseley [/bib_ref]. In pom1Δ cells, cdr2-GFP is spread in dots throughout much of the cortex, except for the growing cell tip [fig_ref] Figure 7: Cdr2p behavior in pom1Δ mutants [/fig_ref] ; [bib_ref] Pom1 kinase links division plane position to cell polarity by regulating Mid1p..., Celton-Morizur [/bib_ref] [bib_ref] The cell-end factor pom1p inhibits mid1p in specification of the cell division..., Padte [/bib_ref] [bib_ref] Polar gradients of the DYRK-family kinase Pom1 couple cell length with the..., Martin [/bib_ref] [bib_ref] A spatial gradient coordinates cell size and mitotic entry in fission yeast, Moseley [/bib_ref]. The total amount of cdr2p in the cell was similar in wild-type and pom1Δ mutant cells over a range of cell lengths [fig_ref] Figure 7: Cdr2p behavior in pom1Δ mutants [/fig_ref]. Cortical profiles showed that in pom1Δ cells, cdr2p was still enriched over the medial cortex and that the non-growing end had levels roughly half that of the medial region [fig_ref] Figure 7: Cdr2p behavior in pom1Δ mutants [/fig_ref]. The fraction of cdr2p that is cortical and the area of nodal cdr2p were both substantially increased in pom1Δ cells [fig_ref] Figure 7: Cdr2p behavior in pom1Δ mutants [/fig_ref]. Interestingly, the increase of cortical cdr2p with cell length was similar in pom1Δ vs wild-type cells, but the curve was shifted slightly upwards [fig_ref] Figure 7: Cdr2p behavior in pom1Δ mutants [/fig_ref]. In contrast, in the medial cortical region, cdr2p levels were lower than wildtype [fig_ref] Figure 7: Cdr2p behavior in pom1Δ mutants [/fig_ref]. A simple interpretation is that cdr2p is able to signal to promote entry into mitosis from nodes on non-medial sites in this mutant. However, another factor to consider is that cdr2p kinase activity may also be altered in these cells. Time-lapse imaging showed that cdr2p nodes are more motile in pom1Δ cells than in WT [fig_ref] Figure 1: Gradient distribution of pom1p is not the basis for cell size control [/fig_ref] , suggesting a defect in the anchoring of these nodes in the membrane. At the growing end, there are also the dim cortical motile cdr2p dots, similar to those present at cell ends in WT [fig_ref] Figure 5 -: figure supplement 2B, 'Materials and methods' [/fig_ref]. Thus, pom1p affects the distribution and mobility of cdr2p nodes.
We also examined the effect of disrupting pom1p localization on the cdr2p distribution. A construct in which pom1p is targeted all over the plasma membrane has been described (PMT-Pom1C fusion, [fig_ref] Figure 7: Cdr2p behavior in pom1Δ mutants [/fig_ref] -figure supplement 1A) . Although cdr2p was expressed at denote local peaks in the cdr2-GFP signal that are clearly distinct from the mean cdr2-GFP cortical signal. Strain: FC2678. (B) Outline of mathematical model for cdr2p dynamics. (C) Equations and analytic solutions describing cortical and nodal cdr2p number. (D) Model parameters. 'Measured': deduced directly from experiment, 'constrained': limited by nodal cdr2p density scaling with cell length, 'not important': plays no role in nodal cdr2p density scaling. (E) Model fit to nodal cdr2-GFP density as function of cell length (data from maximum intensity projection as described in . A subset of cells whose surface area was within 10-20% of the mean surface area was selected for each cell type ('Materials and methods'). The graphs show the surface area, volume, and nodal cdr2-GFP intensity (cdr2-GFP intensity [fig_ref] Figure 6: Cdr2p and cell size at division scale with cell surface area [/fig_ref]. Continued on next page normal levels in the whole cell, it was evenly distributed all over the cortex at a low level, and did not increase at the medial cortex with increasing cell length [fig_ref] Figure 1: Gradient distribution of pom1p is not the basis for cell size control [/fig_ref]. As shown previously , these cells divided at abnormally long cell lengths, similar to cdr2Δ mutants. These data show that pom1p has an inhibitory effect on cdr2p localization to nodes. These results further provide support that cdr2p needs to be present at these medial nodes in order to function effectively in cell size control.
# Discussion
Here we propose a mechanism for cell size sensing based on a cortical sizer protein cdr2p. We provide evidence that cells sense a critical cell size by measuring cell surface area rather than, for example, cell volume or absolute length, a mechanism that could function regardless of the cell shape. As the cell grows, the concentration of cdr2p at the medial cortex increases. We have developed models explaining how cdr2p probes the surface area of the cell, and conveys this information to the medial cortex. There, cdr2p may signal to cell cycle regulators located on the nearby spindle pole body and nucleus (see below). When the cell reaches a critical size, cdr2p at the nodes may reach a critical local concentration that promotes mitotic entry.
Our quantitative models show how cdr2p can convey information about global cell area and deliver it in the form of a local (nodal) concentration. This size-sensing model shares elements with a proposed microtubule length control mechanism termed the 'antenna model'. In the microtubule model, longer microtubules bind more motor proteins, which then accumulate at the microtubule end in a lengthdependent manner [bib_ref] Yeast kinesin-8 depolymerizes microtubules in a length-dependent manner, Varga [/bib_ref]. In the cell size sensing case, the whole surface area of the plasma membrane may be regarded as an 'antenna'. Similar to the microtubule model, the property of cdr2p to first bind to the plasma membrane 'antenna' (as opposed to merely binding the nodes directly) is critical for this mechanism to read out cell size. This membrane cdr2p must then transit to the nodal region, where the cdr2p nodal density serves as a read-out of cell area. Although cdr2p may not exhibit directed motor-driven movements, this movement can still occur by diffusion along the membrane. We also considered an alternative model, where cdr2p is modified on the membrane, but after unbinding is able to diffuse through the cytoplasm to the nodes. The modification allows information about membrane area to be preserved in the cytoplasm, from where it can be relayed to the nodes [fig_ref] Figure 2A ,: Figure 1-figure supplement 1A, Figure 2-figure supplement 1 [/fig_ref]. Furthermore, as the amount of nodal cdr2p reflects cell size rather than time, we postulate that the system is effectively in a dynamic steady state at a given cell size, with fast cdr2p dynamics compared to the timescales of cell growth.
The localization of a cdr2p sizer at cortical nodes provides several key advantages over other locales. First, it allows the local concentration of nodal cdr2p to increase as the cell grows. Previously proposed mechanisms have been based upon nuclear concentration or the nuclear/cytoplasmic ratio of a sizer, but in many cell types (including fission yeast), nuclear volume also increases as cells grow [bib_ref] Nuclear size control in fission yeast, Neumann [/bib_ref] [fig_ref] Figure 1: Gradient distribution of pom1p is not the basis for cell size control [/fig_ref]. Second, we speculate that measured as defined in [fig_ref] Figure 2 -: figure supplement 4 [/fig_ref] in these selected cells. For each data type, normalization is by mean value for rga4Δ cells. Error bars = Error on the mean. n = 24 (wt) cells, 27 (rga4Δ), 32 (rga2Δ). Strains used in B and C: FC1441, FC2792, FC2793. See [fig_ref] Figure 6: Cdr2p and cell size at division scale with cell surface area [/fig_ref] -figure supplement 1. (C) As in B, except groups of cells were selected with similar volumes (mean measured volume ± 10-20%). n = 24 (wt) cells, 27 (rga4Δ), 27 (rga2Δ). These data show cdr2-GFP scaling with surface area. The difference in surface area and cdr2-GFP intensity between the rga2Δ and rga4Δ cells is statistically significant (**p<10 −3 , ***p<10 −4 ). See [fig_ref] Figure 1: Gradient distribution of pom1p is not the basis for cell size control [/fig_ref] Comparison of cell lengths, surface areas and volumes in rga4Δ, wild type and rga2Δ at time of septation ('Materials and methods'). The septum is not included in these measurements. Data for each set is normalized by the appropriate value for the rga4Δ cells. Error bars = SD. Strains used: FC2554, FC2555, FC2556. n = 76 (wt), 64 (rga4Δ), 60 (rga2Δ). (E) Quantitating differences between rga4Δ, wt and rga2Δ at time of septation. Left: probability density distributions for measured surface area (top) and volume (bottom) for wild type (red), rga2Δ (green) and rga4Δ (blue) cells in (D). Gray area marks the overlap region between the distributions. Error bars not shown for clarity. Right: to quantitatively compare these distributions, we calculated the Jensen-Shannon distance [bib_ref] Divergence measures based on the Shannon entropy, Lin [/bib_ref] between the length, surface area and volume distributions for the different cell types (where 1 corresponds to the distributions having no shared information and 0 to identical distributions, see 'Materials and methods'). This analysis shows that these cells with different shapes divide with similar surface area. DOI: 10.7554/eLife.02040.023
The following figure supplements are available for figure 6: [fig_ref] Figure 1: Gradient distribution of pom1p is not the basis for cell size control [/fig_ref]. pom1Δ cells are orientated such that the cell end with the higher cdr2-GFP level is defined to be at d = 0 µm. n = 52 (wt) cells, 72 (pom1Δ). Error bars not shown for clarity. See 'Materials and methods' for further details. (D) Fraction of cdr2-GFP signal observed on the cortex compared with total measured cdr2-GFP in the medial plane. The cortical signal is calculated as the sum of measured intensity along a mask around the cortex (see 'Materials and methods' for mask definition). The total signal is defined as the total measured cdr2-GFP intensity on and inside the mask. Error bars = SD. n = 52 cells (wt), 72 (pom1Δ). (E) Measured area of nodal cdr2-GFP region from maximum intensity projection images. Regions were measured manually for individual cells. n = 46 (wt) cells, 77 (pom1Δ). Error bars = SD. (F) Accumulation of total membrane [fig_ref] Figure 7: Cdr2p behavior in pom1Δ mutants [/fig_ref]. Continued on next page medial cortical placement of nodes surrounding the medial nucleus may allow cdr2p to communicate its local concentration to presumed targets such as wee1p and cdk1p on the nucleus. Although wee1p can be observed at some nodes upon overexpression , its localization in late G2-phase is clear in the nucleus, and at the spindle pole body (SPB), a structure on the nuclear envelope situated close (often <0.5 μm) to the nodes. Cdk1/ cyclin B and polo kinase are also located at the SPB and nucleus [bib_ref] Distinct nuclear and spindle pole body populations of cyclin-cdc2 in fission yeast, Alfa [/bib_ref] [bib_ref] Centrosomal MPF triggers the mitotic and morphogenetic switches of fission yeast, Grallert [/bib_ref]. Potentially, the SPB could detect local gradients of cdr2p (or other molecules) emanating from nearby cortical nodes. However, as a simple cdr2p concentration gradient in the cytoplasm is expected to be very shallow (due to rapid diffusion), it is likely that additional layers of regulation such as through phosphorylation states or diffusion barriers would be needed to generate suitably steep gradients. The potential importance of the geometric relationship between the nodes and SPB/nucleus remains to be tested.
The localization of these nodes to the medial cortical region involves multiple inputs. One important contributor is pom1p. Although pom1p clearly regulates cdr2p function and phosphorylation, our data indicate that the pom1p gradient distribution may not be the primary size sensing mechanism as previously proposed. Indeed, our data are consistent with a recent report that size correction still occurs in pom1Δ mutants [bib_ref] Pom1 and cell size homeostasis in fission yeast, Wood [/bib_ref]. Rather, a primary role of pom1p may be to ensure the medial localization of nodes. Thus, pom1p may affect cdr2p nodes in part by affecting distribution and general properties (such as its mobility in the membrane) of the nodes. Recent studies (published while this work was in press) suggest that cdr2p activity is also regulated by phosphorylation by pom1p and ssp1p protein kinases [bib_ref] Distinct levels in Pom1 gradients limit Cdr2 activity and localization to time..., Bhatia [/bib_ref] [bib_ref] Dueling Kinases Regulate Cell Size at Division through the SAD Kinase Cdr2, Deng [/bib_ref]. Another important factor in cdr2p localization is likely to be the nucleus that is situated in the cell interior with roughly the same width as the nodal region. Studies on mid1p, another component of the nodes, suggest that the nucleus governs dynamic nodal localization, in a mechanism that may involve nuclear shuttling [bib_ref] Analysis of mid1p, a protein required for placement of the cell division..., Paoletti [/bib_ref] [bib_ref] Dynamic positioning of the fission yeast cell division plane, Daga [/bib_ref] [bib_ref] Spatial control of cytokinesis by Cdr2 kinase and Mid1/anillin nuclear export, Almonacid [/bib_ref]. Furthermore, the organization of the cortical endoplasmic reticulum also influences nodal stability and localization [bib_ref] The cortical ER network limits the permissive zone for actomyosin ring assembly, Zhang [/bib_ref]. There are also likely to be additional (or alternative) inputs into size control [bib_ref] Driving the cell cycle with a minimal CDK control network, Coudreuse [/bib_ref] [bib_ref] A systematic screen reveals new elements acting at the G2/M cell cycle..., Navarro [/bib_ref] [bib_ref] Pom1 and cell size homeostasis in fission yeast, Wood [/bib_ref]. Additional cell size regulators include the cell tip protein nif1p [bib_ref] Polar gradients of the DYRK-family kinase Pom1 couple cell length with the..., Martin [/bib_ref] [bib_ref] Pom1 and cell size homeostasis in fission yeast, Wood [/bib_ref] , and skb1p, which localizes to cortical patches distinct from the nodes [bib_ref] Compartmentalized nodes control mitotic entry signaling in fission yeast, Deng [/bib_ref]. Cells expressing a cdk1p-cyclinB fusion still exhibit apparently near-normal size control in the absence of wee1p/mik1p or cdk1p-tyrosine phosphorylation control [bib_ref] Driving the cell cycle with a minimal CDK control network, Coudreuse [/bib_ref] [bib_ref] A systematic screen reveals new elements acting at the G2/M cell cycle..., Navarro [/bib_ref] [bib_ref] Pom1 and cell size homeostasis in fission yeast, Wood [/bib_ref] , suggesting the existence of controls that are entirely independent of cdk1-tyrosine phosphorylation. Thus, this simple cdr2p-based mechanism is likely to be a core component of a larger network responsible for cell size control.
# Materials and methods
## S. pombe strain construction
Standard methods for S. pombe growth and genetics were used [bib_ref] Molecular genetic analysis of fission yeast Schizosaccharomyces pombe, Moreno [/bib_ref]. In general, strains were constructed using a PCR-based homologous recombination method to insert markers in the yeast chromosome . Pom1-mGFP (= pom1-GFP) and pom1-3GFP strains were constructed by inserting mGFP and 3GFP constructs into the pom1 + chromosomal locus from fragments amplified from pFA6a-mGFP-kanMX6 (monomeric GFP A206K [bib_ref] Partitioning of lipid-modified monomeric GFPs into membrane microdomains of live cells, Zacharias [/bib_ref] and pFA6a-3GFP-kanMX6 (triple tandem GFP) [bib_ref] Counting cytokinesis proteins globally and locally in fission yeast, Wu [/bib_ref] [bib_ref] Dynamics of the formin for3p in actin cable assembly, Martin [/bib_ref] , (from JQ Wu). In general, constructs were checked by PCR and sequencing, and strains were outcrossed multiple times.
## Imaging s. pombe cells
For live cell imaging, S. pombe cells were typically grown in exponential phase in liquid YE5S media at 25°C with shaking for 18-24 hr. In some experiments, the cells were mounted in liquid YE5S media directly on glass. For long term imaging experiments, the cells were placed in open 35-mm glass bottom dishes (MatTek Corp, Ashland, MA). To stick cells to the glass, dishes were coated with lectin by drying 5 μl of 1 μg/μl lectin on the dishes; the cells in media were applied and incubated for 5 min and then 2 ml YE5S were added [fig_ref] Figure 2 -: figure supplement 4 [/fig_ref]. The cells were also imaged on 1% agarose YE5S pads under a glass coverslip.
## Modulating cdr2p expression levels
For experiments to alter cdr2 levels [fig_ref] Figure 3: Cdr2p is a dose-dependent regulator of cell size [/fig_ref] , the nmt81 promoter [bib_ref] TATA box mutations in the Schizosaccharomyces pombe nmt1 promoter affect transcription efficiency..., Basi [/bib_ref] was inserted upstream of the cdr2 chromosomal locus by homologous recombination using a PCR-generated DNA fragment derived from pFA6a-kanMX6-P81nmt1 using the following primers:
nmt-cdr2-F: (5′-TATGCTGTTCTATGAATGGGGTTTGGATTTGGCCATCACCACTTCACCGATTT ACTGGTTCTTTTGAATAGTTGAAGTGTGAATTCGAGCTCGTTTAAAC-3′) nmt-cdr2-R: (5′-TTGGCTAAACGTGATGAATTTGGTCCTCCTGATCCTAAGGAAAGACCAAGC TCCCAAGGTCCAACTTCTGAAATTGTACTCATGATTTAACAAAGCGACTATA-3′).
Correct insertion was verified by PCR of both sides of the construction using specific primers for the endogenous and inserted DNA. Multiple transformants showed the same cell size phenotypes. nmt81-cdr2 cells and the parental wild-type strain (FC15) were grown in EMM +5 µg/ml thiamine at 25°C for 2 days, keeping the OD 600 of the culture below 0.5 over the entire period. The cells were then washed three times by centrifugation at 2000 rpm with EMM, innoculated into EMM with or without 5 µg/ml thiamine, and then grown with shaking at 25°C for 20 hr, and then samples were collected for microscopy for cell length measurements and for RNA preparation. Relative RNA expression levels were assayed by RT-PCR. RNA was isolated from cells using the RNeasy Mini Kit (Qiagen, Germantown, MD). 50 ng of RNA was used for real-time PCR using iScript One-Step RT-PCR kit with SYBR green (Bio-Rad) on a Bio-Rad Real-Time PCR system. The actin act1 mRNA was used as standard. Amplicons for act1 or cdr2 were generated with the following primers: act1-F (5′-GAAGAAGAAATCGCAGCGTTGG-3′), act1-R (5′-CGCTTGCTTTGAGCTTCATCAC-3′) cdr2-F (5′-TGGGAGCTTGGTCTTTCCTTAG-3′), cdr2-R (5′-TAGCCTGTTGGCTCGAAGTAAG-3′). Expression levels of cdr2 in nmt81-cdr2 cells were measured as fold change relative to levels in a wild-type (FC15) strain grown under the same conditions. The changes in cell lengths were consistent in multiple experiments.
## Pharmacological inhibitors
Cycloheximide (Sigma, St Louis, MO) was used at a final concentration of 100 µg/ml from a stock of 10 mg/ml stock solution in ethanol and added to exponential phase cultures in YE at 25°C [bib_ref] Effects of heat shock and cycloheximide on growth and division of the..., Polanshek [/bib_ref]. Latrunculin A (LatA) was used at a final concentration of 200 µM from a 100X stock in DMSO [bib_ref] Movement of a cytokinesis factor cdc12p to the site of cell division, Chang [/bib_ref]. LatA or cycloheximide were added to cells in a 35-mm glass bottom dish (described above) and imaged over time.
## Research article
## Microscopy
Images were generally acquired using a spinning-disc confocal fluorescence NikonTI-based microscope system (Nikon Instruments, Melville, NY, Yokogawa, Tokyo, Japan, Solamere Technology, Salt Lake City, UT) with an EM CCD camera (Hamamatsu Corp, Boston, MA) and a 100X 1.4 N.A. objective with a 1.5X magnifier [bib_ref] Noise reduction in the intracellular pom1p gradient by a dynamic clustering mechanism, Saunders [/bib_ref]. A wide-field Nikon Eclipse 800 microscope and a 60X 1.4 N.A. objective was also used for some studies. FRAP studies were performed with a Zeiss 710 scanning confocal microscope.
# Image analysis
ImageJ (NIH) and custom MatLab (Mathworks, Natick, MA) software were used for analysis.
## Pom1p gradient analyses
Fluorescence intensity values around the cortex of cells were measured from images of cells in a medial focal plane, using custom MatLab software for the automated generation of a one-pixel wide mask around the cell cortex, followed by manual correction [bib_ref] Noise reduction in the intracellular pom1p gradient by a dynamic clustering mechanism, Saunders [/bib_ref]. Timeaveraged images of pom1-fusions used average projections of 50 0.5 s frames over 25 s. The average pom1-tomato intensity at the medial cortex was measured in a 3-pixel wide by 3 µm long rectangle over the medial cortex, and the mean background value outside of the cells was subtracted. To measure the pom1p gradient decay lengths, cells expressing the appropriate pom1-fusion were imaged for 3 s in a single confocal section through the middle of the cell. Cells were segmented as described in [bib_ref] Noise reduction in the intracellular pom1p gradient by a dynamic clustering mechanism, Saunders [/bib_ref] Intensities were normalized to one at the cell tip and background subtraction performed such that the different fusions had zero intensity 5 µm from the tip. Curves were then fitted to exp (−x/λ), where λ is the decay length of the profile, with λ shown in [fig_ref] Figure 1: Gradient distribution of pom1p is not the basis for cell size control [/fig_ref] -figure supplement 4B.
## Cdr2p node analyses
Cdr2-GFP intensity was quantified using six different methods [fig_ref] Figure 2 -: figure supplement 4 [/fig_ref]. (A) Maximum projections were made of 13 slices of confocal sections taken 0.4 μm apart. A region of interest (ROI) was selected in ImageJ by hand around the cdr2p nodes, excluding as much background as possible. The area and total intensity of the ROI was recorded, and the ROI width was determined by the spread of cdr2p nodes along the long axis of the cell. (B) Similar to (A) but the maximum projection was taken from the top three slices consisting of the 'top' cortical section of the cell. (C) Similar to (A), except the ROI was selected by an image analysis program in Matlab (custom-written) which selected only pixels over a predetermined threshold (approximately two times the mean background intensity). In this case, the width was not determined. (D) Maximum projections were taken similar to (A). We used the Find Maxima macro function in ImageJ to find the brightest pixel from a local intensity source (likely nodes), counting their number and totaling their intensity to estimate total intensity levels. In this case, width was also not determined. (E) We used a single confocal section through the middle of the cell and acquired images over 30 s. A region was then chosen for each time-averaged data set that overlapped the nodal region (now seen as a line on the perimeter of the cell) in a single pixel wide line that was 3 µm long. The intensity was measured from that line and summed. (F) Maximum projections were made of 13 slices of confocal sections taken 0.4 μm apart. Individual cells were then taken and rotated so their long axis was horizontal. A rectangular ROI was fixed at 3 μm wide and 3.72 μm tall for all cells and placed at the center of the nodal region. The mean intensity was then recorded in this fixed area.
In all these instances, the mean background intensity from an area outside of the cells was subtracted for each pixel. These different methods all resulted in the same linear increase of cdr2-GFP intensity levels in the nodal region as a function of increasing cell length. However, due to the fact that each method measured cdr2-GFP levels in different ways, the exact slope and variance of the correlation differed from method to method.
The single cell analyses of cdr2-GFP in the wildtype [fig_ref] Figure 2 -: figure supplement 4 [/fig_ref] used 13 confocal sections 0.4 µm apart. Intensities were measured in a hand drawn ROI that contained the majority of cdr2p nodes and the mean background outside the cells was subtracted.
In [fig_ref] Figure 3: Cdr2p is a dose-dependent regulator of cell size [/fig_ref] , in analysis of LatA and cycloheximide-treated cells, and for3Δ cells, maximum projections of Z-stacks comprising 13 confocal sections 0.4 µm apart were used. Intensities were measured in a hand drawn ROI that contained the majority of cdr2p nodes and the mean cytoplasmic value inside the cells at each time point was subtracted.
In the measurements of rates of growth and cdr2-GFP accumulation in for3 mutants , growth rates were calculated by a least squares linear fit to the cell length as a function of time (images every 30 min), over 60-120 min. The rate of change in nodal cdr2p intensity with time was also calculated by a least squares linear fitting. In both the cases, the error on the fit was found for each cell. To test whether a positive correlation between growth and cdr2p accumulation rates was robust, we performed numerical simulations using the distributions of the measured rates, and their errors, to create in silico data. Corresponding to each pair of values in the measured data set, we created a new in silico pair by drawing from Gaussian distributions with widths given by the measured errors in growth rate and cdr2p accumulation rate. We then performed a linear least squares fitting on each in silico data set to find the level of correlation and test whether it was greater than zero-that is whether a positive correlation existed between cdr2p accumulation rate and cell growth rate. Repeating this process 10 6 times, we found a probability of ∼0.0005 that a positive correlation would be absent. Hence, our conclusion of a positive correlation between the cdr2p accumulation rate and cell growth rate is robust. Results shown in are for a single experiment (n = 21 cells); similar results were found in multiple additional experiments (data not shown).
Protein counts were estimated by quantitative fluorescence intensity in ratios with standard proteins that had been quantitated previously [bib_ref] Counting cytokinesis proteins globally and locally in fission yeast, Wu [/bib_ref] [bib_ref] Counting protein molecules using quantitative fluorescence microscopy, Coffman [/bib_ref]. GFP-MotB complexes in live bacteria were used as a standard at 22 GFP molecules/dot [bib_ref] Stoichiometry and turnover in single, functioning membrane protein complexes, Leake [/bib_ref] [bib_ref] CENP-A exceeds microtubule attachment sites in centromere clusters of both budding and..., Coffman [/bib_ref] [bib_ref] Assembly and architecture of precursor nodes during fission yeast cytokinesis, Laporte [/bib_ref].
To calculate the width of the nodal cdr2p region, we fitted the function ( ) σ − − + 2 2 0 ( ) / 2 x x ae b to the cdr2p profile from a time-averaged (90 s) confocal section through the middle of each cell (385 cells). We only analyzed cells with a good quality of fit (so that the measured σ is meaningful) and with σ >0.5 µm (thereby excluding cells with distorted fits due to one very bright nodal region). This process left 237 cdr2p intensity profiles for analysis. Each cell was binned according to length (8-9 µm, 9-10 µm, …) and the mean and standard deviation calculated within each bin, see [fig_ref] Figure 2 -: figure supplement 4 [/fig_ref].
For the cortical Cdr2-GFP profiles shown in [fig_ref] Figure 7: Cdr2p behavior in pom1Δ mutants [/fig_ref] , [fig_ref] Figure 7: Cdr2p behavior in pom1Δ mutants [/fig_ref] -figure supplement 1C, a cortical mask was extracted as described in [bib_ref] Noise reduction in the intracellular pom1p gradient by a dynamic clustering mechanism, Saunders [/bib_ref]. The center of each cell was located and the angles between a chosen tip and each pixel on the mask were calculated (so a pixel at the opposing tip would have angle π). Angles were then binned into 100 sectors from 0 to 2π and the mean cdr2-GFP intensity at a given angle around the cell was calculated. Angles were converted into the mean distance from the tip by assuming that in the mid plane the cell can be approximated as two semicircles connected by straight lines, using the mean cell length and radius for each cell type. For pom1Δ cells, the tip with the highest cdr2-GFP intensity was defined to be at d = 0 µm.
To calculate total cortical signal the sum of the cdr2-GFP signal on the mask was used [fig_ref] Figure 7: Cdr2p behavior in pom1Δ mutants [/fig_ref]. For analysis of cdr2-GFP intensity in a 3-µm cortical region around the cell middle [fig_ref] Figure 7: Cdr2p behavior in pom1Δ mutants [/fig_ref] , [fig_ref] Figure 7: Cdr2p behavior in pom1Δ mutants [/fig_ref] -figure supplement 1D) each pixel in the cell cortical mask within ± 1.5 µm of the cell centre was identified and then the cdr2-GFP was summed over only these pixels.
## Measuring cell surface area and volume
For [fig_ref] Figure 6: Cdr2p and cell size at division scale with cell surface area [/fig_ref] , cells were grown in liquid YE5S media at 25°C, and imaged on agarose pads. Cell surface area and volumes were measured using manual segmentation. In [fig_ref] Figure 6: Cdr2p and cell size at division scale with cell surface area [/fig_ref] ,C, we used a single mid focal plane brightfield image, whereas in [fig_ref] Figure 6: Cdr2p and cell size at division scale with cell surface area [/fig_ref] ,E, we used a single mid focal plane of a fluorescent image of blankofluor-stained septated cells. Cell perimeters were manually traced, with the mean surface area, A cor and mean volume, V, calculated in Matlab assuming radial symmetry around the long axis of the cell (as the cross-section of fission yeast cells are nearly circular). To compare cells of similar surface area, we selected all cells with surface areas in the range A cor ± 0.10-0.20 A cor . The range of ± 10-20% was taken to ensure we had enough cells included for statistical significance (between 24 and 32 cells), but that the range was reasonably constrained. The specific range was adjusted for each subset of cells for the different cell lines such that the mean surface area [fig_ref] Figure 6: Cdr2p and cell size at division scale with cell surface area [/fig_ref] or mean volume [fig_ref] Figure 6: Cdr2p and cell size at division scale with cell surface area [/fig_ref] were equal to within ± 1%. The unbinned data for each cell type is shown in [fig_ref] Figure 6: Cdr2p and cell size at division scale with cell surface area [/fig_ref] -figure supplement 1. Likewise, for comparing cells of similar volume, we included all cells with volume in the range V ± 0.10-0.20 V. For measuring cell size at septation [fig_ref] Figure 6: Cdr2p and cell size at division scale with cell surface area [/fig_ref] , cells without cdr2-GFP were analyzed. The septum was not included in this analysis, as we wanted to extract cellular dimensions at entry into mitosis prior to septum formation. We also analyzed a separate data set with the cdr2-GFP strains (n > 45 cells for each genotype), and a data set using brightfield images, which all showed the same behavior.
## Research article
The similarity of the distributions for cell length, surface area and volume at mitosis were compared in the wild-type, rga2Δ and rga4Δ mutants using the Jensen-Shannon distance [fig_ref] Figure 6: Cdr2p and cell size at division scale with cell surface area [/fig_ref]. The Jensen-Shannon distance is a statistical measure that quantitatively compares the overlap of two or more distributions, with a distance of 1 corresponding to the distributions having no shared information and a distance of 0 to identical distributions. The Jensen-Shannon distance is the square root of the Jensen-Shannon divergence, which is defined in terms of the Shannon entropy function of the probability distributions (see [bib_ref] Divergence measures based on the Shannon entropy, Lin [/bib_ref].
## Description of mathematical modeling cell morphology
Wild-type fission yeast geometry is approximated as a cylindrical body with hemispherical caps at either end. The radius of the cell is approximately constant at about R = 1.5 μm, while over the cell cycle the cell length grows from about L = 7 μm to L = 14 μm. In this case, the surface area A cor and length L are strictly proportional and related by A cor = 2πRL. This approximation of the cell morphology predicted surface areas and volumes consistent with experimentally measured values (data not shown).
## Timescales
From cdr2p FRAP experiments, the lifetime of the nodal cdr2p is on the order of 3 min [fig_ref] Figure 2 -: figure supplement 4 [/fig_ref]. From live imaging of cortical cdr2p [fig_ref] Figure 5 -: figure supplement 2B, 'Materials and methods' [/fig_ref] , cortical cdr2p dynamics are rapid, with a timescale on the order of seconds. However, cell growth is considerably slower (with doubling times on the order of hours) and hence we solve the subsequent equations for cdr2p with each cell size considered to be in quasi-steady-state.
## Model i different forms of cdr2p
In our first model cdr2p is taken to have three forms: cytoplasmic, cortical, and nodal. (1) Cytoplasmic cdr2p has a homogeneous concentration, ρ cyt = N cyt /V, which does not change significantly with cell size, as found experimentally [fig_ref] Figure 1: Gradient distribution of pom1p is not the basis for cell size control [/fig_ref].
(2) Cytoplasmic cdr2p can associate with the membrane. For this cortical cdr2p population, N cor denotes the copy number and ρ cor = N cor /A cor is the corresponding concentration. (3) The cortical cdr2p can cluster in nodes at the midcell cortex. For this nodal cdr2p population, N nod denotes the copy number, with corresponding concentration ρ nod = N nod /A nod , where A nod is the area of the cell membrane occupied by the nodes.
We employ two approaches in our analysis. First, we take the cortical cdr2p population as diffusing rapidly and hence having an approximately uniform distribution around the cell membrane. The nodal cdr2p is taken to be uniformly distributed within the nodal region, though below we also consider a model variant where we explicitly consider diffusion of the cortical cdr2p population.
## Uniform cdr2p populations
Here, the uniformly distributed cdr2p populations in quasi-steady-state are described by the following equations:
[formula] 0 cor nod cyt cor cor cor A A N N N V A β ν α = − − 0 , nod cor nod cor A N N A α η = − [/formula]
where, β is the association parameter of cytoplasmic to cortical cdr2p, ν is the disassociation rate of cortical cdr2p back into the cytoplasm, α is the rate of uptake of cortical to nodal cdr2p and η is the disassociation rate of nodal cdr2p back into the cytoplasm. These equations can be solved exactly:
-1 A = + and = .
nod nod cyt cor nod cor
[formula] A β ν η ρ ρ ρ ρ η α α [/formula]
The value of β is not important as it only enters our solutions as a constant prefactor. The rate of cdr2p disassociation from the nodes back into the cytoplasm, η, can be estimated from our FRAP experiments [fig_ref] Figure 2 -: figure supplement 4 [/fig_ref]. We find that the cdr2p has a nodal occupancy time of around 3 min. We can therefore estimate η = 5 × 10 −3 s −1 .
Much higher concentrations of cdr2p in the nodes are experimentally observed than elsewhere on the cortex. From above, since ρ nod /ρ cor = α/η ≫ 1, we therefore require that the rate α of uptake of cortical cdr2p into the nodes be considerably greater than the rate η of nodal cdr2p disassociation back into the cytoplasm. This constraint places a lower bound on α, and consistently we choose α = 1.0 s −1 .
Experimentally, we observe significant scaling of ρ nod with increasing A cor (or equivalently with cell length in the wild type). For this to occur, our model requires that two key criteria be met. First, the cortical cdr2p must be much more likely to be taken up into the nodes than disassociate from the cortex, that is from above ν/α ≪ 1. Since α is already constrained, we have a further restriction on ν. Accordingly, we choose ν = 5 × 10 −3 s −1 , meaning that the cdr2p disassociation rates from the nodes and cortex are the same. Second, A nod , the area of the nodal region, must not scale proportionally with A cor , the total cell area, as the cell size increases. Importantly, this model requirement was verified experimentally, see [fig_ref] Figure 2 -: figure supplement 4 [/fig_ref]. Although the cdr2p does spread to an extent during growth [fig_ref] Figure 2 -: figure supplement 4 [/fig_ref] , the majority of nodal cdr2p is localized to the center of the cortex [fig_ref] Figure 1: Gradient distribution of pom1p is not the basis for cell size control [/fig_ref]. Therefore, we take A nod to be constant in the [fig_ref] Figure 5 -: figure supplement 2B, 'Materials and methods' [/fig_ref] fitting. In [fig_ref] Figure 5 -: figure supplement 2B, 'Materials and methods' [/fig_ref] -figure supplement 1, we also include the effect of A nod increasing with cell length (see below). Finally, the model also included the experimentally observed (slight) decrease in cyt ρ as a function of cell length, [fig_ref] Figure 5 -: figure supplement 2B, 'Materials and methods' [/fig_ref] -figure supplement 1A black line (gradient = −0.01 µm −1 after normalization to the average cytoplasmic intensity). However, there was little difference in our results between this case and when assuming a strictly constant cyt ρ (data not shown).
As shown in [fig_ref] Figure 5 -: figure supplement 2B, 'Materials and methods' [/fig_ref] , the above model can recapitulate the observed cdr2p scaling. The larger the value of α, the stronger the scaling effect will be, as more cortical cdr2p-which effectively 'measures' the cell area-is taken up into the nodes-which effectively 'read-out' the area measurement.
## Cdr2p membrane localization does not necessarily imply cell size control
If cytoplasmic cdr2p can only associate to the membrane by being directly taken up by nodes at midcell then, we can simply leave out the cortical cdr2p form. By balancing the cdr2p coming onto the nodes (parameter β) with that disassociating (rate η) we find ρ nod = (β/η) ρ cyt . The concentration of nodal cdr2p is now independent of cell size, assuming ρ cyt is constant. A similar conclusion is reached if the nodes can form anywhere on the cortex by direct association of cdr2p from the cytoplasm. In both cases, the system cannot sense cell area because both association and disassociation occur over the same region. To sense cell area we require one process, which here is the association of cdr2p anywhere onto the membrane, to scale proportionally with cell area. However, the second process, which here is the uptake of cortical cdr2p into the nodes, must be localized over a region whose size does not scale proportionally with the total cell area as the cell grows. The outcome is then a density (for both ρ cor and ρ nod ) that scales with total cell membrane area.
## Incorporating cortical cdr2p diffusion
Incorporating cortical diffusion into the model is straightforward, though solutions now need to be found numerically. We assume that the cytoplasmic cdr2p is still homogeneous and at an almost constant concentration, decreasing only slightly with cell length as described above. Assuming that on the membrane the cdr2p densities only depend on the long-axis coordinate, x, the equations become, at quasi-steady-state: with β, η and ν taking the same values as before [fig_ref] Figure 5 -: figure supplement 2B, 'Materials and methods' [/fig_ref]. We now incorporate the width of the nodal region by using an association function α(x) for the uptake rate of cortical to nodal cdr2p. This scheme is, of course, a simplification of the true uptake dynamics, which presumably involve cdr2p aggregation and clustering. Nevertheless, this simplification is sufficient for understanding the mechanistic basis of size scaling. We take α(x) = α 0 exp(−x 2 /2ω 2 ), with α 0 = 0.5 s −1 . Here, ω is the fitted width of the nodal region (fitted to the appropriate data set in [fig_ref] Figure 2 -: figure supplement 4 [/fig_ref] : ω = a (1−e −L/s ) where s = 7 µm and a = 2.2 µm. In [fig_ref] Figure 5 -: figure supplement 2B, 'Materials and methods' [/fig_ref] , we see that the nodal cdr2p does not move significantly over an extended period, suggesting D nod /D cor ≪ 1. Therefore, we set D nod = 0.
For the nodal cdr2p density to serve as a read-out of the entire cell membrane area, the typical cortical cdr2p diffusional displacement along the long cell axis must be greater than 5 µm. This requirement ensures that cortical cdr2p can diffuse along the long axis from the cell tips to the nodal region without first disassociating. Hence √(2D cor τ) > 5 µm, where τ is the lifetime of cortical cdr2p. Given a cortical cdr2p lifetime of around 3 min (see above), this implies that the diffusion constant should be greater than about D cor = 0.1 µm 2 s −1 . In our simulations, we use D cor = 0.2 µm 2 s −1 , but if the lifetime of the cortical cdr2p is shorter, then the cortical diffusion constant will need to be larger.
We solve the equations numerically in one-dimension with length L and hard wall boundary conditions, using Matlab. This model can reproduce the profile of cortical/nodal cdr2p for different cell lengths (or equivalently with cell surface areas in the wildtype) [fig_ref] Figure 1: Gradient distribution of pom1p is not the basis for cell size control [/fig_ref]. The increase in nodal cdr2p concentration as the cell grows [fig_ref] Figure 2 -: figure supplement 4 [/fig_ref] is also captured [fig_ref] Figure 1: Gradient distribution of pom1p is not the basis for cell size control [/fig_ref]. In conclusion, within reasonable parameter ranges, the model prediction-that the concentration of cdr2p in the nodes increases with cell area-is robust to the inclusion of cortical cdr2p diffusion.
## Model ii
A simple alternative model for area scaling involves cdr2p becoming modified (e.g., phosphorylated). We assume that unmodified cdr2p diffuses rapidly in the cytoplasm, with homogeneous density ρ cyt = N cyt /V. Unmodified cdr2p can then bind to the membrane with a binding constant β. Once present on the membrane, with a correspondingly homogeneous density ρ cor = N cor /A cor , cdr2p can unbind back into the cytoplasm at a rate ν, while at the same time becoming modified (e.g., phosphorylated). This cytoplasmic, modified form of cdr2p, with homogeneous density ρ * cyt = N * cyt /V can rapidly diffuse, and then bind to the nodal region on the cortex, with a binding constant α, or spontaneously become unmodified at a rate µ. Finally, modified cdr2p in the nodal region, with density ρ nod = N nod /A nod , can unbind and become unmodified cytoplasmic cdr2p at a rate η In the case without spontaneous reversion of cytoplasmic, modified cdr2p back to its unmodified form (i.e., µ = 0), we can solve these equations to find = .
cor nod cyt nod
[formula] A A β ρ ρ η [/formula]
Experimentally, we observe significant scaling of ρ nod with increasing A cor (or equivalently with cell length in the wildtype). This is in agreement with Model II, again provided that A nod , the area of the nodal region, does not scale proportionally with A cor , the total cell membrane area, as the cell size increases. If there is spontaneous reversion of cytoplasmic, modified cdr2p back to its unmodified form, then the above solution for ρ nod becomes 1 = . +1 cor nod cyt nod
[formula] nod A V A A β ρ ρ μ η α Provided nod A V α µ ≫ [/formula]
, then the reversion process can be neglected and our cell area scaling results are unchanged. As in Model I, the larger the value of α, the stronger the scaling effect will be, as more modified, cytoplasmic cdr2p-which effectively 'measures' the cell area-is taken up into the nodeswhich effectively 'read-out' the area measurement.
As above with Model I, the value of β is not important as it only enters our solutions as a constant prefactor. The value of ν is also not important for the behavior of ρ nod (see above), though we use a value of ν = 0.5 s −1 to ensure low levels of cortical cdr2p, as observed experimentally. We again use the FRAP data to estimate η = 5 × 10 −3 s −1 . Since ρ nod = (α/η) ρ * cyt , we take α = 0.5 µms −1 so that the concentration of modified cdr2p in the nodal region is relatively large compared to that in the cytoplasm (required since there is no observed scaling of cytoplasmic cdr2p concentration with cell length). Further, we use a low rate of spontaneous modification loss µ = 0.03 s −1 , so that the lifetime of modified, cytoplasmic cdr2p is relatively long. The fitting of this model to the experimental scaling of cdr2p is shown in [fig_ref] Figure 2 -: figure supplement 4 [/fig_ref].
In this model, the mechanism of size scaling is similar to that of model I. One process, the unbinding of cortical cdr2p can occur from anywhere on the membrane, and so scales proportionally with cell area. However, a second process, in this case the uptake of modified, cytoplasmic cdr2p into the nodes, occurs over a region whose area does not scale proportionally with the total cell area as the cell grows. The outcome is again a density (for both ρ * cyt and ρ nod ) that scales with total cell membrane area. The presence of the modified form of cdr2p is vital, so that information about membrane area can be protected and relayed to the nodes without being lost into the general cytoplasmic cdr2p population, which does not show size scaling characteristics [fig_ref] Figure 1: Gradient distribution of pom1p is not the basis for cell size control [/fig_ref].
## Role of pom1p in nodal cdr2p scaling
The model does not explicitly include pom1p. However, this does not mean that pom1p is unimportant in the regulation of nodal cdr2p. As the pom1Δ experiments demonstrate, without pom1p acting as a tip inhibitor for cdr2p, nodal cdr2p can form in an extended part of the cell, with such cells observed to divide at shorter lengths. Conversely, in pom1p mutants where pom1p is mistargeted all over the plasma membrane, cdr2p does not form localized regions of high nodal concentration, with such cells dividing at longer lengths. Therefore, pom1p appears to play an important role in defining the region of nodal cdr2p accumulation, without which the cdr2p-dependent control of cell length is perturbed. Accordingly, pom1p is implicitly included in the model by defining a spatially limited region that can be occupied by the cdr2p nodes.
[fig] Figure 2A ,: Figure 1-figure supplement 1A, Figure 2-figure supplement 1; [/fig]
[fig] Figure 2 -: figure supplement 4). With increasing cell length, the number of nodes in each cell increased (Figure 2G, Figure 2-figure supplement 5A), whereas the intensities of individual nodes remained unchanged (Figure 2H, Figure 2-figure supplement 5B [/fig]
[fig] Figure 1: Gradient distribution of pom1p is not the basis for cell size control. (A) Time-averaged spinning disc confocal images of fission yeast cells expressing pom1-tomato in a medial focal plane (60 frames over 3 min). Scale bar = 3 μm. Strain used: FC2054. (B) Total fluorescence intensities of pom1-tomato in a medial 3-μm segment along cortical edge of interphase cells, from images like A (n > 100). See Figure 1-figure supplements 1-3. [/fig]
[fig] Figure supplement 1 004, Figure supplement 2 005, Figure supplement 3 006, Figure supplement 4 007, Figure supplement 5: C) Time-lapse images of pom1-tomato in individual cell. Images are time averaged (5 frames over 25 s) in medial focal plane. Scale bar = 3 μm. (D) Pom1-tomato intensities at medial cortex (as in B) of individual growing interphase cells. (E) Cells expressing pom1-GFP or the pom1-3GFP. Imaging as in A. Strains used: FC1162, FC2685. Scale bar = 3 μm. (F) Gradient profiles of pom1-3GFP, pom1-GFP and pom1-tomato (n > 30 each strain). Peak absolute protein numbers in pom1-3GFP and pom1-GFP gradients were similar. Error bars not shown for clarity. See Figure 1-figure supplement 4. (G) Effect of pom1p fusions on cell size as measured by length of septated cells (n > 100). Error bars: SDs. Strains used: FC420, FC1162, FC2685, FC2054. See Figure 1-figure supplement 4C, 5. (H) Distribution of cell lengths at division in indicated strains. DOI: 10.7554/eLife.02040.003 The following figure supplements are available for figure 1: Pom1p concentration at the medial cortex does not vary with cell length. DOI: 10.7554/eLife.02040.Pom1p concentration at the medial cortex: Comparison with previous data. DOI: 10.7554/eLife.02040.Nuclear width as a function of cell length. DOI: 10.7554/eLife.02040.Pom1p gradients with different decay lengths do not affect cdr2p distribution. DOI: 10.7554/eLife.02040.Pom1p gradients with different decay lengths do not affect cdr2p node intensity or number. DOI: 10.7554/eLife.02040.008 [/fig]
[fig] Figure supplement 1 010, Figure supplement 2 011, Figure supplement 3 012, Figure supplement 4 013, Figure supplement 5: -lapse maximum projection images of a cell expressing cdr2-GFP. Scale bar = 3 μm. (F) Total normalized intensities of nodal cdr2-GFP in 5 cells tracked over time (measured from images like E, using method of Figure 2-figure supplement 3A). See Figure 2-figure supplement 5. (G) Number of cdr2-GFP nodes as function of cell length (n = 51). Black line: linear fit with r 2 = 0.67. Nodes identified by thresholding, using method of Figure 2-figure supplement 3C, which provides a lower-bound estimate. (H) Distributions of cdr2-GFP node intensities in short vs long cells. n = 89 nodes in 9 cells, n = 286 nodes in 7 cells, respectively (nodes as determined in G). DOI: 10.7554/eLife.02040.009 The following figure supplements are available for figure 2: Measurement of cdr2p protein number. DOI: 10.7554/eLife.02040.Cdr2p and pom1p intensity measurements as a function of cell length. DOI: 10.7554/eLife.02040.Comparison of image analysis methods for quantitating cdr2p fluorescence in the nodes. DOI: 10.7554/eLife.02040.FRAP analysis of cdr2-GFP. DOI: 10.7554/eLife.02040.Cdr2p node number but not maximal intensity in each node increases with cell length. DOI: 10.7554/eLife.02040.014 [/fig]
[fig] Figure 5 -: figure supplement 2B, 'Materials and methods'). [/fig]
[fig] Figure 3: Cdr2p is a dose-dependent regulator of cell size. (A) Effect of cdr2p expression level on cell size. cdr2 + was expressed at different levels using a thiamine-regulatable promoter (nmt81-cdr2). Inverted images of cells stained with cell wall dye blankofluor. Cells express cdr2p at levels on average of 1.6, 1.0 and 0.3-fold relative to wild type (top to bottom). Strains used: FC15, FC2691. Scale bar = 5 μm. (B) Length of cells at septation. n = 87, 47, 121, 123 cells. T = thiamine. Error bars = SD, ***p<0.0001 as determined by Kolmogorov-Smirnov statistical tests. (C) Model that the local concentration of cdr2p increases in the region of the medial cortical nodes as cells grow, and when it reaches a critical level, promotes entry into mitosis. (D) Stability of cdr2-GFP protein. Time-lapse images of cells expressing cdr2-GFP treated with 100 μg/ml cycloheximide [/fig]
[fig] Figure 6: Cdr2p and cell size at division scale with cell surface area. (A) Fission yeast cells expressing cdr2-GFP and pom1-tomato in wt, rga4Δ (fat morphology) and rga2Δ (thin morphology) backgrounds. Maximum Z-projection images. Cells lacking nodes are in mitosis. Strains used: FC2678, FC2794, FC2795. Scale bar = 5 μm. (B) Comparison of measured nodal Cdr2-GFP intensity in cells of different volumes. For each cell, the surface area and volume were measured by segmentation (' [/fig]
[fig] Figure supplement 1: Scaling of nodal cdr2-GFP intensity with surface area and volume. DOI: 10.7554/eLife.02040.024 [/fig]
[fig] Figure 7: Cdr2p behavior in pom1Δ mutants. (A) Fission yeast cells expressing Cdr2-GFP in wt and pom1Δ background. Brightfield, maximum projection, and mid-focal plane images are shown. Strains used: FC1441 and FC2057. Scale bar = 3 μm. (B) Comparison of total measured cdr2-GFP intensity (from sum projection after background subtraction) with cell length. n = 52 (wt), 72 (pom1Δ). (C) Average cdr2-GFP intensity profile around cortex of cell (spatial direction as defined in cartoon in [/fig]
[fig] Figure 5 ,: Figure 5-figure supplement 2A). The corresponding steady-state equations are: [/fig]
|
Influence of Third Particle on the Tribological Behaviors of Diamond-like Carbon Films
Tribological mechanisms of diamond-like carbon (DLC) films in a sand-dust environment are commonly unclear due to the complicated three-body abrasion caused by sand particles. This study investigates the three-body abrasion of the DLC film via molecular dynamics simulations. The influence factors such as the load, velocity, shape of the particle and its size are considered. It has been found that the friction and wear of the DLC film are determined by adhesion at a small load but dominated by both adhesion and plowing at a large load. A high velocity can increase the friction of the DLC film but decrease its wear, due to the response of its networks to a high strain rate indicated by such velocity. The shape of the particle highly affects its movement mode and thus changes the friction and wear of the DLC film. It is found that a small-sized particle can increase the friction and wear of the DLC film by enhancing plowing. These unique tribological mechanisms of the DLC film can help to promote its wide applications in a sand-dust environment.
Diamond-like carbon (DLC) films that consist of sp 2 and sp 3 hybridized carbon atoms exhibit excellent tribological behaviors [bib_ref] Tribological mechanism of hydrogenated amorphous carbon film against pairs: a physical description, Bai [/bib_ref]. The DLC film can highly reduce the friction and wear of workpieces in various environments such as water, desert and outer-space 2 . The sand or dust particles in the desert and outer-space can cause a three-body abrasion of the DLC film that is different from its commonly tribological mechanisms. Since the world is becoming desertization and more explorations are conducted in the outer-space this century, it is important to understand the tribological mechanisms of the DLC film to improve the stability of machines with the presence of sand or dust particles.
The mechanism of the three-body abrasion is complicated, because it is influenced by many factors such as the load, velocity, size of the particle, its shape and number density, hardness ratio of particle to substrates and hardness ratio between the substrates 3 . Moreover, the mechanical properties of DLC films vary largely according to their deposition parameters and compositions [bib_ref] Effect of different ion beam energy on properties of amorphous carbon film..., Bai [/bib_ref] , and thus further complicate their abrasion mechanisms.
Tribological behaviors of DLC films in a sand dust environment have been only studied in several experiments. Previous studies mainly located a large number of sand particles at the contact interface between the DLC films [bib_ref] Ultra-high tribological performance of magnetron sputtered a C: H films in sand-dust..., Qi [/bib_ref] [bib_ref] The tribological performance of selected solid lubricant films in sand-dust environments, Qi [/bib_ref] [bib_ref] A comparative investigation of the wear behavior of PTFE and PI under..., Li [/bib_ref] [bib_ref] Influences of added sand-dust particles on the tribological performance of graphite-like coating..., Qi [/bib_ref]. In fact, in common situations the surface damages of the DLC films are caused by few sand particles instead of a plenty of them. This is because the few particles can cause a huge contact stress and severely damage the DLC film by inducing its plastic deformation, thus degrading its surface morphology and structure and highly influencing its tribologcial behaviors. Influence of such few particles is hardly investigated in experiment. This is because these particles and the thickness of the DLC films are commonly at the microscale or nanoscale and the corresponding friction and wear phenomenon are hardly observed in experiment [bib_ref] Effect of different ion beam energy on properties of amorphous carbon film..., Bai [/bib_ref].
In this case, simulations are commonly employed to understand the three-body abrasion. reported that the substrate deformation in a three-body contact condition follows the regimes of no-wear, condensing, adhering and ploughing [bib_ref] Atomic scale deformation in silicon monocrystals induced by two-body and three-body contact..., Zhang [/bib_ref]. Sun et al. found that the nanoparticle purely rolls during the friction process and the wear of materials is dominated by plowing [bib_ref] Abrasive wear of nanoscale single crystal silicon, Sun [/bib_ref]. In the study by Si et al., wear caused by the particle rolling is high and cannot be neglected compared with that by sliding 11 . These previous simulations provide a fundamental for the understanding of the three-body abrasion at the nanoscale. This study investigates the tribological behaviors of DLC films in a three-body contact condition via molecular dynamics (MD) simulations. The influence factors such as the load, velocity, shape of the particle and its size are considered. It is believed that this study can improve the understanding of the three-body abrasive mechanisms of the DLC film and promote its applications in a sand-dust environment.
# Results and discussions
Load effect. The load effect is considered in the cases with a spherical particle. The sliding configuration in [fig_ref] Figure 1: Atomic configuration of the friction process with a spherical particle [/fig_ref] shows that many covalent bonds are formed at the interface between the particle and the DLC films, indicating that the sticky DLC networks attach to the spherical particle. These bonds cause a strong interfacial adhesion and thus influence the movement of the particle that can be evaluated by calculating its mass-center velocity. It is found that the spherical particles undergo pure rolling in the friction process. This is consistent with observations in the literature [bib_ref] Abrasive wear of nanoscale single crystal silicon, Sun [/bib_ref]. The sliding configuration demonstrates that the friction and wear of the DLC films with a rolling third particle may closely relate with adhesion.
When the load F n increases, the friction force F f of the DLC films increases [fig_ref] Figure 2: Effect of load F n of the DLC film on [/fig_ref]. The increase of F f is because a large F n can penetrate the particle into the DLC films and thus increase the interfacial adhesion strength. This can be verified by the increase of the number of bonds n b at the contact interface with the F n [fig_ref] Figure 2: Effect of load F n of the DLC film on [/fig_ref]. The n b is the total number of bonds formed between the particle and the DLC films, due to the fact that all these bonds contribute to the interfacial adhesion.
Recent theoretical studies showed that at the nanoscale F f is linearly proportional to n b which represents the real contact area, indicating the validation of the macroscale Bowden-Tabor model at the nanoscale 12,13 . The configuration in [fig_ref] Figure 1: Atomic configuration of the friction process with a spherical particle [/fig_ref] has shown that the friction and wear highly depend on the interfacial adhesion due to the pure rolling of the particle. Therefore, the F f in the three-body contact condition should also be closely related with n b . In this case the relation between the F f and n b in the three-body contact condition is firstly studied, as shown in [fig_ref] Figure 2: Effect of load F n of the DLC film on [/fig_ref]. It can be seen that this relation can be regarded as linear with a small n b but becomes nonlinear when the n b is large. Such relation evolution indicates that at a small load F n the friction is simply determined by the interfacial adhesion strength while at a large F n the friction is dominated by both the adhesion and other factors.
At a large F n , the spherical particle highly penetrates into DLC films, which can be verified by the increase of the sliding depth of the spherical particle [fig_ref] Figure 3: Effect of load F n of the DLC film on [/fig_ref]. As a result, the large deformations of the DLC films can be caused by the spherical particle and highly resist their sliding, thus increasing the F f . This keeps consistent with results in the literature [bib_ref] Molecular dynamics simulation of rolling friction using nanosize spheres, Lee [/bib_ref]. It has been reported that the rolling friction can be highly generated by the energy dissipation involved in the deformation of materials [bib_ref] C 60 molecular bearings, Miura [/bib_ref]. Sun et al. further found that the plowing is a significant factor in determining the three-body friction and wear within the elastic-plastic regime [bib_ref] Abrasive wear of nanoscale single crystal silicon, Sun [/bib_ref]. Bhushan et al. also reported that the three-body friction is caused by both the adhesion and plastic deformation which represents the plowing [bib_ref] Comprehensive model for scale effects in friction due to adhesion and two-and..., Bhushan [/bib_ref]. Therefore, it is evident that the nonlinear relation between the F f and n b at a large F n is caused by the plastic deformation and plowing of the DLC films.
Moreover, it is noticed that the F f is still high when F n = 0. The high F f is due to the presence of interfacial adhesion at the zero F n and indicates that the F f in the three-body contact condition has a direct proportion with the n b instead of F n . This keeps consistent with observations in the previous studies under the two-body contact condition [bib_ref] Friction laws at the nanoscale, Mo [/bib_ref] [bib_ref] Applicability of macroscopic wear and friction laws on the atomic length scale, Eder [/bib_ref].
The load F n also significantly influences the wear performance of DLC films, as shown in [fig_ref] Figure 3: Effect of load F n of the DLC film on [/fig_ref]. The wear rate k increases with the F n . Since wear induced by adhesion is proportional to the real contact area that can be represented by n b , the relation between the k and n b is meaningful in the investigation of adhesive wear. [fig_ref] Figure 3: Effect of load F n of the DLC film on [/fig_ref] shows that the relation can be regarded as linear at a small n b but becomes nonlinear when the n b is large. The linear relation is due to the fact that the wear at a small F n is determined by the interfacial adhesion while the nonlinear relation is attributed to the plastic deformation of the DLC films when the F n is large.
It is noticed that k is nonzero when F n = 0. Such nonzero k is different from results in the literature. Zhang et al. reported that a neglected wear can be obtained at a small load that only induces elastic deformation of material surface [bib_ref] Atomic scale deformation in silicon monocrystals induced by two-body and three-body contact..., Zhang [/bib_ref]. This neglected wear should be due to the von der Waals interactions at the sliding interface [bib_ref] Atomic scale deformation in silicon monocrystals induced by two-body and three-body contact..., Zhang [/bib_ref]. In the present study, however, the interfacial forces caused by the strong covalent C-C bonds are quite high and thus can induce worn atoms even when F n = 0.
The contribution of the deformation of DLC films to their wear can be examined by analyzing the sliding depth h of the particle. [fig_ref] Figure 3: Effect of load F n of the DLC film on [/fig_ref] shows that at the maximum F n the h approaches to 5 Å which is higher than the displacement criteria for the definition of worn atoms. Because the surface of the DLC films deforms locally, many atoms are worn by such huge deformation even without the sliding. This keeps consistent with the worn atoms caused by plastic deformation in the literature [bib_ref] Molecular dynamics simulation of rolling friction using nanosize spheres, Lee [/bib_ref] [bib_ref] Comprehensive model for scale effects in friction due to adhesion and two-and..., Bhushan [/bib_ref] [bib_ref] Atomistic simulation of the effect of roughness on nanoscale wear, Hu [/bib_ref].
It is noticed that the wear of DLC films caused by their deformations is neglected when the F n is small. In this case, the wear is determined by adhesion and can be predicted according to the proportion between k and n b . Since this proportion is commonly obtained in the two-body contact conditions 18 , its validation in this study indicates it represents the essence of the adhesion wear regardless of in the two-body or three-body contact conditions. Velocity effect. Besides the load, the velocity v x also highly influences the friction and wear of DLC films, as shown in [fig_ref] Figure 4: Effect of velocities v x of the DLC film on [/fig_ref]. The friction force F f increases with the v x . This is different from the observations in the two-body contact conditions. It has been observed that the F f commonly decreases with the v x due to the fact that a high v x can largely increase the friction temperature and reduce the n b [bib_ref] Effect of environmental hydrogen atoms on the tribological behaviors of diamond-like carbon..., Bai [/bib_ref]. However, in the present study the n b almost keeps constant when the v x varies [fig_ref] Figure 4: Effect of velocities v x of the DLC film on [/fig_ref]. This constant n b seems to conflict with the increase of F f , indicating the existence of a unique friction mechanism.
The friction mechanism can be understood by analyzing the friction configurations, as shown in [fig_ref] Figure 5: Atomic configuration with different velocities v x [/fig_ref]. When the v x increases, although the n b is constant, many networks of DLC films attach to the particle. These networks can highly resist the sliding of the DLC film and thus largely increase the F f . The attaching of these networks to the particle is due to the structural response of DLC films to v x . With a high v x which indicates a high strain rate, the micro-cracks in the DLC films have insufficient time to be initiated and propagate, thus increasing their yield strains. This keeps consistent with the mechanical theory of solids [bib_ref] Influence of strain rate on superelastic properties of TiNi shape memory alloy, Tobushi [/bib_ref] [bib_ref] Influence of strain-rate sensitivity on necking under uniaxial tension, Hutchinson [/bib_ref] [bib_ref] Influence of Strain Rate and Yarn Number on Tensile Test Results, Vangheluwe [/bib_ref] and has also been proved by our tensile simulations of the DLC films. For these films in the present study, such high strain is also attributed to the high flexibility of DLC networks at a high v x , which will be discussed later. As a result, more networks can attach to the third particle at a high v x , leading to the increase of F f .
The wear performances of DLC films are also influenced by the v x , as shown in [fig_ref] Figure 6: Effect of velocities v x of the DLC film on [/fig_ref]. The wear rate k decreases with the v x . This trend agrees well with the decrease of the sliding depth h at a high v x [fig_ref] Figure 6: Effect of velocities v x of the DLC film on [/fig_ref]. The decrease of the h indicates the reduced deformation of the DLC films and thus contributes to the decrease of k by highly reducing plowing.
Effect of v x on the adhesive wear also contributes to the reduction of k. [fig_ref] Figure 7: Formation process of worn atoms at a high velocity v x of... [/fig_ref] shows that at a high v x many DLC networks attach to the spherical particle. The relative sliding between the DLC films largely deform their networks. When the strain energy is high enough to break the bonds between these networks and the spherical particle, majority of them return to the DLC films while only few atoms are worn and still attach to the particle.
The return of the networks is determined by their flexibility and the instability of their atoms attached to the spherical particle. The flexibility of the DLC networks can be characterized by the fraction of sp 2 atoms. This fraction can be represented by the total number of new sp 2 atoms Δ N sp2 in the friction process. A higher Δ N sp2 indicates a larger fraction of sp 2 atoms in the DLC film. [fig_ref] Figure 8: Effect of velocities v x of the DLC film on its [/fig_ref] shows that the Δ N sp2 increases with the v x , i.e., the DLC film exhibit more flexible behavior at a high v x . Such flexibility makes the DLC networks undergo high strain before their yielding.
On the other hand, the instability of the network atoms attached to the spherical particle can be simply characterized by their temperature which can be regarded as the friction temperature T f . The atoms with a high T f are active and easily influenced by external forces. [fig_ref] Figure 8: Effect of velocities v x of the DLC film on its [/fig_ref] shows that the T f highly increases with the v x . The combination of the flexibility of the DLC networks and the high T f can raise a new wear mechanism. When the DLC films relatively slide, their flexible networks can largely attach to the spherical particle and deform during the sliding [fig_ref] Figure 7: Formation process of worn atoms at a high velocity v x of... [/fig_ref]. As a result, such networks can highly draw the atoms bonded to the particle and make them tend to return to the DLC films. Moreover, the tendency is further enhanced by the high T f , since it can highly improve the possibility of the breaks for the bonds formed between these atoms and the particle. As a result, majority of these atoms return to the DLC film and only few of them are worn, resulting in the decrease of k.
The tribological mechanisms of DLC films with different v x in this study are quite different from results under the two-body contact condition in the literature [bib_ref] Effect of environmental hydrogen atoms on the tribological behaviors of diamond-like carbon..., Bai [/bib_ref] [bib_ref] Tribology of diamond-like carbon films: recent progress and future prospects, Erdemir [/bib_ref] [bib_ref] A study of the wear mechanism of diamond-like carbon films, Liu [/bib_ref] [bib_ref] An investigation of the relationship between graphitization and frictional behavior of DLC..., Liu [/bib_ref]. It has been reported that the high v x can reduce the friction and wear of DLC films by improving their level of graphitization and promoting the formation of a transfer film which is easy to shear and capable of isolating the DLC films from their counterparts.
The present study shows that although DLC networks attaching to the third particle are actually the transfer layer, the friction reduction by such layer is neglected. This is attributed to that the friction reduction by the transfer layers is caused by their shear deformation due to their easy-shear properties. However, in this study their shear deformations hardly occur because the pure rolling of the spherical particle mainly causes tensile deformation of the DLC networks.
Therefore, the present study generally shows that the wear reduction at a high v x in the three-body contact conditions is caused not by the formation of transfer layer but by suppressing the plowing and improving the flexibility of the DLC networks and the high friction temperature.
Particle effect. The effect of the particle size is considered by changing the radius of the spherical particle, as shown in . The friction force F f increases with a small-sized particle. Such increase is attributed to that under the same load the small-sized particle causes a high contact stress and thus largely penetrates into the DLC films. Such penetration highly deforms the DLC films and thus increases the F f . This keeps consistent with results in the literature which stated that the plowing is significant in determining the three-body friction at the nanoscale [bib_ref] Abrasive wear of nanoscale single crystal silicon, Sun [/bib_ref] [bib_ref] Atomistic scale tribological behaviors in nano-grained and single crystal copper systems, Sun [/bib_ref]. The large deformation of the DLC films with the small-sized particle also increases their wear rate.
The large F f and k of DLC films with a small-sized particle is different from observations in the previous studies 2,5,6,8 . Qi et al. reported that small-sized sand particles can reduce the friction and wear of DLC films mainly by reducing the contact stress 5,6 . This is because in their studies a large quantity of sand particles are located at the interfaces between DLC films. The small size of such particles can make them form a relatively flat layer between the DLC films and thus reduce the contact stress. However, the present study demonstrates that a small number of particles surely exhibit a different abrasive behavior and may severely damage the DLC film and change their tribological behaviors. The particle shape can also influence the tribological behaviors of DLC films. [fig_ref] Figure 9: Atomic configuration with [/fig_ref] shows that the cubic particle also purely rolls when the DLC films relatively slide. However, the cuboid particle hardly rolls and is initially attached to the lower DLC film. Such cuboid particle highly ploughs the upper DLC film. This can be proved by the chip formation in front of the cuboid particle. Moreover, the chip exerts a high force to the cuboid particle. When this chip becomes large, this force plus the adhesion force from the upper DLC film can cause the movement of the cuboid particle. As a result, such particle is in turn attached to the upper DLC film and ploughs the lower DLC film. It can be seen that the shape of particles highly influences their movement modes which affect tribological mechanisms of DLC films. This keeps consistent with observations in the previous study which stated that the nanoparticle exhibits an optimum shape to realize its rolling [bib_ref] Dynamical evolution of wear particles in nanocontacts, Anantheshwara [/bib_ref]. [fig_ref] Figure 1: Atomic configuration of the friction process with a spherical particle [/fig_ref] shows that the cuboid particle causes a high F f while the spherical and cubic particles cause a low F f . The dependence of F f on the particle shape is due to the transition of the friction mechanisms. For the cuboid particle, the ploughing is present and highly increases the F f . For the spherical and cubic particles, the friction is determined by rolling and the F f is caused by the adhesion instead of the ploughing. The variance of the F f for cubic particle is due to the varied interfacial adhesion strength when the particle rolls. Therefore, it can be seen that the presence of ploughing is a significant reason for the high friction in the three-body contact condition.
The shape of the third particle also highly influences the wear performance of DLC films [fig_ref] Figure 1: Atomic configuration of the friction process with a spherical particle [/fig_ref]. The k with the cuboid particle is much higher than those with the cubic and spherical particles and is attributed to the occurrence of ploughing. Meanwhile, the low k with cubic and spherical particles is due to the rolling adhesive wear which is highly suppressed by the flexibility of the DLC networks. The influence of the particle shape indicates that the presence of ploughing can highly increase the friction and wear of DLC films. Moreover, such influence demonstrates that the rolling wear rate of the DLC films in the third-body contact condition is lower than their sliding wear rate. This is consistent with results in the previous studies 10, [bib_ref] Movement patterns of ellipsoidal particle in abrasive flow machining, Fang [/bib_ref]. It has been reported that plowing or ploughing dominates the three-body wear at the nanoscale [bib_ref] Abrasive wear of nanoscale single crystal silicon, Sun [/bib_ref]. Experimental results also found that grooving movement of particles shows a higher contribution to wear volume of workpiece than their rolling mode [bib_ref] Movement patterns of ellipsoidal particle in abrasive flow machining, Fang [/bib_ref]. The ploughing caused by the cuboidal particle actually reflects the properties of the two-body abrasion which cannot be well explained by the mechanisms for spherical and cubic particles. Therefore, it is evident that the shape of the particle can directly determine its movement mode and the friction and wear mechanisms of the DLC films 27 .
# Conclusions
The tribological behaviors of DLC films with a third particle at their contact interface are investigated via molecular dynamics simulations. The influence factors such as the load F n , velocity v x , shape of the particle and its size are considered. It has been found that the friction force F f and wear rate k of the DLC film are determined by adhesion at a small F n but dominated by both adhesion and plowing at a large F n . This can be verified by examining the relation of the F f and k with the number of bonds n b at the contact interface. With the increase of v x , the F f increases and k decreases while the n b almost keeps constant. This is because with a large v x the DLC networks exhibit a large yield strain and thus largely attach to the third particle to resist the relative sliding of the DLC films. These attached networks highly increase the F f . The decrease of k at a large v x is caused by the flexibility of the DLC networks and the decrease of the sliding depth. The small-sized particle can increase the F f and k by enhancing plowing. The shape of the third particle can highly influence its movement mode and change the friction and wear of the DLC films. It is found that the spherical and cubic particle purely roll without sliding. However, the cuboidal particle highly increases the F f and k by purely sliding and ploughing the DLC films, indicating that the cuboidal particle can induce the transition of tribological mechanisms from a three-body rolling to a two-body sliding. Note that in this study the third particles are set to be rigid and thus their wear are not considered. However, such
# Methods
The simulation system consists of two relatively sliding DLC films and a particle located at the interface between them [fig_ref] Figure 1: Atomic configuration of the friction process with a spherical particle [/fig_ref]. The relative sliding is realized by setting the upper and lower DLC film with a velocity of 0.5v x and − 0.5v x along the x-direction, respectively. During the sliding, a load is applied by maintaining a force F n to the upper DLC film along the y-direction. The periodic boundary conditions of the system are set along its x and z-directions.
The particle has a diamond crystalline structure and is set as a rigid body to avoid its wear during the simulation. The spherical, cubic and cuboidal particles are generated to study the effect of their shapes. The spherical particle has a radius of 12 Å. The cube and cuboid particle have dimensions of 20 × 20 × 20 Å 3 and 40 × 15 × 20 Å 3 , respectively. A small spherical particle with a radius of 10 Å is also generated to study its size effect.
The DLC films have dimensions of 120 × 25 × 60 Å 3 . Each of the DLC films is defined into three different layers from the contact interface to subsurface along the y-direction: Newtonian, thermostatic and rigid layers. The Newtonian layer with a thickness of 18 Å is in contact with the diamond ball and contains atoms that are free to move under the forces of their neighbors. The thermostat layer with a thickness of 3 Å is employed keep a constant temperature of 300 K by rescaling the velocities of atoms. The rigid layer with a thickness of 4 Å can help to maintain velocity. Moreover, the rigid layer in the upper DLC film is set to take F n along the y-direction while that in the lower DLC film is prohibited to move along the same direction. More details of the simulation model can refer to our previous studies [bib_ref] Effect of environmental hydrogen atoms on the tribological behaviors of diamond-like carbon..., Bai [/bib_ref] [bib_ref] Friction between silicon and diamond at the nanoscale, Bai [/bib_ref].
The DLC films in the simulations are obtained by a melt-quenching procedure [bib_ref] Investigation on tensile behaviors of diamond-like carbon films, Bai [/bib_ref]. A block of crystalline diamond is firstly generated. In a canonical NVT ensemble, the temperature of such block is raised to be above the melting point of crystalline diamond, thus leading to the formation of carbon liquid. After a period of equilibration by keeping the block thermostatic, its temperature decreases to 300 K with a high rate of about 1000 K/ps which allows proper structural relaxations in amorphous structures. The dimension of the block is finally adjusted to release its residual stress in an isothermal-isobaric NPT ensemble at 300 K. More details of the melt-quenching procedure can refer to previous studies [bib_ref] Investigation on tensile behaviors of diamond-like carbon films, Bai [/bib_ref] [bib_ref] A modified Tersoff potential for pure and hydrogenated diamond-like carbon, Sha [/bib_ref].
The simulations are conducted via the large-scale atomic/molecular massively parallel simulator (LAMMPS) [bib_ref] Fast parallel algorithms for short-range molecular dynamics, Plimpton [/bib_ref]. The atomic interactions are described by the Tersoff potential which is capable of studying the structures and energetics of carbon-based materials [bib_ref] Empirical interatomic potential for carbon, with applications to amorphous carbon, Tersoff [/bib_ref] [bib_ref] New empirical approach for the structure and energy of covalent systems, Tersoff [/bib_ref]. Moreover, the time step of the simulations is set as 1 fs, and their molecular visualizations are conducted via the software OVITO [bib_ref] Visualization and analysis of atomistic simulation data with OVITO-the Open Visualization Tool...., Stukowski [/bib_ref].
Prior to the friction simulation, a load F n (7.8, 39.2, 78.4, 117.6 and 196 nN) is applied to the upper DLC film to cause contact between the particle and the DLC films. The contact is equilibrated within 30 ps. After the equilibration period, the DLC films start to relatively slide and the relative velocity v x is set as 2, 5, 7 and 10 Å/ps, respectively. The total sliding distance is always kept as 300 Å. In the cases with different F n , v x is kept as 2 Å/ps. The F n is set as 196 nN for the cases with various v x . In the cases with different shapes and sizes of particles, the F n and v x are set as 196 nN and 2 Å/ps, respectively.
The friction force F f is calculated by summing up the forces of atoms in the rigid layer of the upper DLC film along the x-direction. The number of worn atoms N is simply calculated by evaluating their displacements [bib_ref] Atomistic simulation of the effect of roughness on nanoscale wear, Hu [/bib_ref]. For a mild wear which is determined by atom-by-atom attritions [bib_ref] Ultralow nanoscale wear through atom-by-atom attrition in silicon-containing diamond-like carbon, Bhaskaran [/bib_ref] , a worn atom can be conveniently defined as the one whose bonds with its nearest neighbors break during the friction process, Such breaks can be caused when the displacement of the atom is larger than two-bond length. Since the maximum length of a C-C bond in the DLC films is about 2 Å which corresponds to the first minimum in their radial distribution functions, the length of 4 Å can be chosen as the displacement criteria to estimate the worn atoms. This criteria is also useful to estimate the N for the severe wear of materials that is determined by their plastic deformations. Previous studies calculated this N as those removed from the wear track [bib_ref] Molecular dynamics simulation of severe adhesive wear on a rough aluminum substrate, Zhong [/bib_ref] [bib_ref] Molecular dynamics simulations of nano-indentation and wear of the γ Ti-Al alloy, Xu [/bib_ref]. It is evident that these atoms have displacements larger than 4 Å. Therefore, the present criteria can provide more information of wear than the method employed in the literature [bib_ref] Molecular dynamics simulation of severe adhesive wear on a rough aluminum substrate, Zhong [/bib_ref] [bib_ref] Molecular dynamics simulations of nano-indentation and wear of the γ Ti-Al alloy, Xu [/bib_ref]. It should be noticed that the large displacement of atoms in the DLC films can also be induced by [fig_ref] Figure 1: Atomic configuration of the friction process with a spherical particle [/fig_ref]. Atomic configuration of the simulation model with (a) a front view and (b) a side view. The model consists of two relative sliding DLC films with a third particle located between them.
Scientific RepoRts | 6:38279 | DOI: 10.1038/srep38279 their elastic deformation. These elasticity-induced displacements can cause an error in the wear calculation. In order to eliminate such error, the wear calculation is conducted after the friction sliding. In this case, the elastic deformation has recovered, and thus all the large displacements of atoms are caused by the plastic deformation of the DLC films.
The wear rate k is calculated as k = N/L, where N is the number of worn atoms and L is the sliding distance. The average k of two DLC films is employed to indicate their wear performance. It is noticed that the sliding tracks overlap in each simulation due to the periodic condition in the x-direction. The overlapping hardly changes the wear rate and the friction force, thus indicating that this study actually investigates the phenomenon in the running-in period of wear tests in experiments. The hybridization states of C atoms are also evaluated by calculating the number of their nearest neighbor atoms within the cutoff of the maximum bond length. The fourfold, threefold and twofold atoms are regarded as sp 3 , sp 2 and sp bonded, respectively [bib_ref] Investigation on tensile behaviors of diamond-like carbon films, Bai [/bib_ref]. The temperature of atoms is calculated based on its relation with their kinetic energies.
[fig] Figure 1: Atomic configuration of the friction process with a spherical particle.Scientific RepoRts | 6:38279 | DOI: 10.1038/srep38279 [/fig]
[fig] Figure 2: Effect of load F n of the DLC film on (a) its friction force F f and (b) the number of contact bonds n b ; (c) F f vs n b . Scientific RepoRts | 6:38279 | DOI: 10.1038/srep38279 [/fig]
[fig] Figure 3: Effect of load F n of the DLC film on (a) its wear rate k and (b) the sliding depth of the spherical particle h; (c) k vs h. Scientific RepoRts | 6:38279 | DOI: 10.1038/srep38279 [/fig]
[fig] Figure 4: Effect of velocities v x of the DLC film on (a) its friction force F f and (b) the number of contact bonds n b . [/fig]
[fig] Figure 5: Atomic configuration with different velocities v x : (a) 2 Å/ps and (b) 10 Å/ps. Scientific RepoRts | 6:38279 | DOI: 10.1038/srep38279 [/fig]
[fig] Figure 6: Effect of velocities v x of the DLC film on (a) its wear rate k and (b) the sliding depth h of the spherical particle. [/fig]
[fig] Figure 7: Formation process of worn atoms at a high velocity v x of 10 Å/ps. The dark-red color in the figure highlights the networks attached to the third particle. [/fig]
[fig] Figure 8: Effect of velocities v x of the DLC film on its (a) number of new sp 2 atoms Δ N sp2 and (b) friction temperature T f . [/fig]
[fig] F f nN k Å −Table 1: Effect of particle radius r on the friction and wear of DLC films. [/fig]
[fig] Figure 9: Atomic configuration with (a) cubic and (b) cuboidal particles during the friction process. Scientific RepoRts | 6:38279 | DOI: 10.1038/srep38279 [/fig]
|
A Low FODMAP Gluten-Free Diet Improves Functional Gastrointestinal Disorders and Overall Mental Health of Celiac Disease Patients: A Randomized Controlled Trial
A subset of patients with celiac disease (CD) on a gluten-free diet (GFD) reported the persistence of functional gastrointestinal disorders. Foods containing fermentable, oligosaccharides, disaccharides, monosaccharides, and polyols (FODMAP) can trigger a broad range of gastrointestinal symptoms in sensitive individuals. We evaluated the effects of a low FODMAP diet (LFD) on gastrointestinal and psychological symptomatology in CD patients. A total of 50 celiac patients on GFDs and with persistence of gastrointestinal symptoms were included. The patients were randomly allocated to one of two dietary groups-one on a low FODMAP GFD (LF-GFD, n = 25) and the other on a regular GFD (R-GFD, n = 25)-for 21 days. Psychological symptomatology and quality of life were evaluated by the Symptom Checklist-90-R (SCL-90) and the Short Form (36) Health Survey (SF-36) questionnaires, respectively. Gastrointestinal symptomatology and general well-being were evaluated by visual analogue scale (VAS) scores. After 21 days, 21 and 23 patients completed the dietary treatment on LF-GFD and R-GFD, respectively. A reduced global SCL-90 index (p < 0.0003) was found in the LF-GFD group but not in the R-GFD one. However, the SF-36 scores did not differ between groups after treatment. The VAS for abdominal pain was much lower, and the VAS for fecal consistency enhanced after treatment in the LF-GFD group. General well-being increased in both groups but with a much higher improvement in the LF-GFD (p = 0.03). A short-term LFD regimen helps to improve the psychological health and gastrointestinal symptomatology with enhanced well-being of CD patients with persisting functional gastrointestinal symptomatology. The long-term clinical effects of LFD in particular subgroups of CD patients need further evaluation.
# Introduction
Celiac disease (CD) is an autoimmune multisystem disorder triggered by gluten ingestion. CD affects genetically susceptible individuals who are known to possess the Human Leukocyte Antigen HLA DQ2 (90%-95%) or the HLA DQ8 (5%-10%) haplotypes [bib_ref] HLA types in celiac disease patients not carrying the DQA1*05-DQB1*02 (DQ2) heterodimer:..., Karell [/bib_ref]. CD symptomatology is mainly gastrointestinal with patients usually reporting diarrhea, bloating, abdominal pain, and weight loss. Extra-intestinal symptoms can be also frequent [bib_ref] Extraintestinal manifestations of coeliac disease, Leffler [/bib_ref].
A gluten-free diet (GFD) is the current treatment for CD [bib_ref] Celiac disease: Understanding the gluten-free diet, Bascuñán [/bib_ref]. In this dietary treatment, foods containing gluten, which is a protein found in grains, such as wheat, barley, rye, and triticale, are excluded. Gluten induces small intestine inflammation, and a GFD helps to counteract the clinical signs/symptoms and to prevent complications [bib_ref] Treatment of celiac disease: From gluten-free diet to novel therapies, Francavilla [/bib_ref]. Although this treatment is highly successful, following a strict GFD poses great difficulty to patients in their family, social, and working contexts, thus deteriorating their quality of life [bib_ref] Predictors of reduced health-related quality of life in adults with coeliac disease, Hauser [/bib_ref] and causing psychological distress [bib_ref] Prevalence of eating disorders in adults with celiac disease, Passananti [/bib_ref].
It is not uncommon that patients on GFDs report symptoms resembling those of irritable bowel syndrome (IBS), which is a frequent condition in clinical practice. Reportedly, around 20-23% of treated CD patients fulfill the Rome III criteria for IBS and also suffer from various functional gastrointestinal symptoms, further affecting their quality of life [bib_ref] Celiac disease and irritable bowel-type symptoms, O'leary [/bib_ref]. In fact, a meta-analysis has shown that IBS-like symptoms are common in CD patients (the pooled prevalence of IBS symptoms in treated CD patients was 38%), concluding that higher levels of adherence to a GFD are possibly associated with some reduction in symptomatology [bib_ref] Prevalence of irritable bowel syndrome-type symptoms in patients with celiac disease: A..., Sainsbury [/bib_ref] ; however, the authors also highlighted that in some patients, IBS-like symptoms persist even after following a strict GFD.
Functional gastrointestinal disorders are characterized by recurrent or current gastrointestinal symptoms that have no identifiable structural or biochemical basis. The most common functional gastrointestinal disorder is IBS [bib_ref] Prokinetics in the Management of Functional Gastrointestinal Disorders, Quigley [/bib_ref]. The variety of clinical manifestations has limited the effective treatment of these syndromes, and most treatments to date only alleviate the primary manifestation. A novel option for IBS treatment, which is currently generating great excitement, is the dietary regimen with reduced amounts of fermentable, oligosaccharides, disaccharides, monosaccharides, and polyols (FODMAP) [bib_ref] Does a diet low in FODMAPs reduce symptoms associated with functional gastrointestinal..., Marsh [/bib_ref]. FODMAP are short-chain carbohydrates that are poorly absorbed in the small intestine and increase gas production and intestinal osmolarity because of their rapid fermentation and osmotic action [bib_ref] Evidence-based dietary management of functional gastrointestinal symptoms: The FODMAP approach, Gibson [/bib_ref] [bib_ref] The low FODMAP diet improves gastrointestinal symptoms in patients with irritable bowel..., De Roest [/bib_ref] [bib_ref] Dietary Triggers of Abdominal Symptoms in Patients with Irritable Bowel Syndrome: Randomized..., Shepherd [/bib_ref]. Foods containing FODMAP can trigger gastrointestinal symptoms in sensitive individuals [bib_ref] Evidence-based dietary management of functional gastrointestinal symptoms: The FODMAP approach, Gibson [/bib_ref] [bib_ref] Low fermentable oligosaccharides, disaccharides, monosaccharides and polyols (FODMAP) diet improves symptoms in..., Varjú [/bib_ref].
A low FODMAP diet (LFD) appears to be associated with the reduction of IBS symptoms [bib_ref] A diet low in FODMAPs reduces symptoms of irritable bowel syndrome, Halmos [/bib_ref]. A very recent meta-analysis found evidence for the short-term efficacy and safety of LFD for patients with IBS, but the long-term effects are still under investigation [bib_ref] Low fermentable, oligo-, di-, mono-saccharides and polyol diet in the treatment of..., Schumann [/bib_ref]. Recently, LFD has been evaluated in patients with inflammatory bowel disease, showing improved symptomatology after treatment [bib_ref] Low-FODMAP diet reduces irritable bowel symptoms in patients with inflammatory bowel disease, Pedersen [/bib_ref]. From a clinical point of view, CD and IBS may coexist [bib_ref] Is gluten sensitivity a no man's land or a fertile crescent for..., Ball [/bib_ref]. However, it is more likely that the inflammatory process occurring in CD does not revert completely in some patients on GFD, and similar low-grade inflammation can be present both in patients with CD and IBS [bib_ref] Celiac disease and irritable bowel-type symptoms, O'leary [/bib_ref] [bib_ref] Irritable bowel syndrome: Diagnosis and pathogenesis, El-Salhy [/bib_ref]. To date there are no reports showing the potential effect of LFD on gastrointestinal symptomatology for patients with CD; thus, we have evaluated the role of LFD on treated CD patients with the persistence of functional gastrointestinal disorders. In addition, given the frequent manifestation of psychopathological and behavioral abnormalities in CD patients undergoing dietary changes, their overall psychological distress and disability were also assessed. In particular, we hypothesized that patients being administered LFD would show improved conditions in terms of gastrointestinal and psychopathological symptoms compared with patients on regular-GFD (R-GFD).
# Materials and methods
This was a randomized double-blind intervention-controlled study, previously registered at ClinicalTrials.gov (ref. no. IDNCT02946827). All the authors had access to the study data and reviewed and approved the final version of the manuscript. The patients were recruited at the Center for Prevention and Diagnosis of Celiac Disease of Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico in Milan (Italy). The University of Milan's Institutional Review Board reviewed and approved the study protocol according to the Helsinki Declaration, the Project Identification Code of the Ethics Committee Approval of our study: 744_2015bis, the protocol was approved by the Ethics Committee of Milano Area B (date . All the patients gave and signed their informed consent prior to participation in this study.
Between December 2015 and December 2017, we studied patients with CD fulfilling the following inclusion criteria: adult age (between 18 and 60 years), treated with a GFD for at least one year, with negative plasma tissue transglutaminase values, with IBS-like symptoms and functional gastrointestinal disorders according to the Rome III criteria [bib_ref] New standard for functional gastrointestinal disorders, Drossman [/bib_ref] , and with a global well-being score assessed by a visual analogue scale (VAS) of <4. As exclusion criteria, we considered the following: low adherence to the GFD (as evaluated by the Celiac Dietary Adherence Test [bib_ref] A Simple Validated Gluten-Free Diet Adherence Survey for Adults with Celiac Disease, Leffler [/bib_ref] , refractory CD (as evaluated through biopsy to assess the persistence of intestinal atrophy while on a GFD and by means of interview carried out by a trained nutritionist, who assessed patients' adherence to the diet), individual intolerance to disaccharides (as evaluated by hydrogen test (lactose and fructose), history of previous nutritionist evaluation or nutritional treatment for the dietary management of IBS, taking IBS pharmacological therapy, abdominal surgery, and type 2-diabetes.
CD was diagnosed according to positivity to the serological tests of endomysial antibodies and tissue transglutaminase antibodies and on the basis of histological abnormalities at duodenal biopsy according to the modified Marsh classification (following the European Society for Pediatric Gastroenterology and Nutrition criteria) [bib_ref] Nutrition European Society for Pediatric Gastroenterology, Hepatology, and Nutrition guidelines for the..., Husby [/bib_ref]. The allocation ratio was 1:1, and the recruited patients (n = 50) were randomly allocated to either of two dietary treatments-a low FODMAP GFD (LF-GFD, n = 25) or a R-GFD (R-GFD, n = 25)-for a time length of 21 days. Before randomization (baseline) and at day 21, the patients underwent a physical examination, and biochemical and nutritional parameters were assessed. After the intervention period, 21 patients in the LF-GFD and 23 in the R-GFD group completed the protocol and were included for the analyses reported in this study. The primary outcome was change in the VAS score for general well-being after 21 days of intervention. Secondary outcomes were changes in the VAS score for gastrointestinal symptomatology and the Short Form (36) Health Survey (SF-36) and the Symptom Checklist-90-R (SCL-90) scores for quality of life and psychological symptomatology, respectively.
The sample size was calculated using G*Power v. 3.1.9.2 for Windows (Düsseldorf, Germany [bib_ref] A flexible statistical power analysis program for the social, behavioral, and biomedical..., Faul [/bib_ref] based on the difference in the reduction of overall IBS-like symptomatology of at least 50% after 21 days following a LFD or high-FODMAP diet in IBS patients, as reported by McIntosh et al.. Considering an α-error 1% (two-tailed test) and a power of 90%, with a response of 72% after the LFD and 21% after the high-FODMAP diet, the estimated sample size was estimated in 22 patients per group, including an additional 20% of patients for potential losses during the follow-up. The random allocation sequence was planned by one of the researchers (L.R.), and the participants' enrollments were carried out by F.B., F.F., and L.E. Both the researchers (gastroenterologists and trained nutritionists) and the patients were blind after the assignment to each intervention group.
## Clinical evaluation
At the baseline and at the end of the intervention period, each patient underwent a clinical and nutritional evaluation (by two gastroenterologists and two trained nutritionists). Overall health and gastrointestinal and extra-gastrointestinal symptoms were assessed, and the gastrointestinal symptomatology was further classified.
## Diets
The nutritional evaluation aimed at assessing anthropometrical parameters, nutritional status, and usual dietary patterns. After clinical evaluation, a personalized GFD adjusted to match energy and macronutrients and micronutrients daily requirements was indicated to each patient. This task was carried out by a trained nutritionist who was only in charge of performing this task, without involvement in patient management. In each dietary treatment, a structured 21-day dietary plan excluding all food gluten sources was indicated. The dietary plan included structured daily meals and specific foods/beverages and was explained in detail to each patient at the beginning of the study. After the initial explanation, the nutritionist was available to answer any doubts or issues strictly related to the dietary plan via e-mail or telephone. As all the patients received a structured dietary plan, both the R-GFD, as well as the LF-GFD, received a review of their current dietary habits, in addition to the change in FODMAP content in the LF-GFD group. The FODMAP content of the R-GFD and LF-GFD was a median (interquartile range) of 21.8 and 3.7 (3.0-4.12) g/day, respectively, as previously described [bib_ref] Measurement of short-chain carbohydrates in common Australian vegetables and fruits by high-performance..., Muir [/bib_ref] [bib_ref] Quantification of fructans, galacto-oligosacharides and other short-chain carbohydrates in processed grains and..., Biesiekierski [/bib_ref]. An example of the meals and foods used in both types of diets is shown in [fig_ref] Table 1: Examples of the two different prescribed diets for a typical day 1 [/fig_ref]. The group with the LF-GFD received an in-depth GFD review, food education regarding GFD and LFD, and dietary counseling to initiate the modification of the FODMAP content towards the LFD. The patients in the R-GFD group received an in-depth GFD review together with food education regarding their diet. Compliance and doubts about the diet were checked 10 days later by means of a telephone call by the same nutritionist.
## Psychological symptoms and quality of life
The Symptom Checklist-90-R (SCL-90) questionnaire was used to evaluate a broad range of psychological problems and symptomatology [bib_ref] Symptom Checklist-90-R (SCL-90-R): Administration, Scoring, and Procedures Manual, Derogatis [/bib_ref]. The scale is composed of 90 questions and assesses the presence and severity of symptoms of mental distress regarding different symptomatic domains (somatization, obsessive-compulsive, interpersonal sensitivity, depression, anxiety, hostility, phobic anxiety, paranoid ideation, and psychotic). Each question is awarded a score on a five-point Likert scale with extremes from "not at all" (0 points) to "extremely" (4 points). In addition to the scores rating specific symptoms' intensities, a global severity index was calculated to estimate the assessment of the patient's psychopathological state and as an indicator of symptomatic severity and psychic distress.
The Short Form (36) Health Survey (SF-36) questionnaire evaluated the patients' quality of life. This is a 36-question-long instrument conceptually referring to eight health domains: physical activity (10 questions), role limitations due to physical health (4 questions), role limitations due to emotional state (3 questions), physical pain (2 questions), general health perception (5 questions), vitality (4 questions), social activities (2 questions), mental health (5 questions), and a single question on changes in the state of health [bib_ref] The MOS 36-item short-form health survey (SF-36): I. Conceptual framework and item..., Ware [/bib_ref]. The scores for each domain ranged between 0 and 100, where 100 represents the best possible perception of quality of life.
## Gastrointestinal symptoms
We used a series of 10-cm long visual analogue scales (VASs) referring to the level of satisfaction with their health status and the severity of specific symptoms (abdominal pain, satisfaction with stool consistency, bloating, postprandial fullness, early satiety, epigastric pain, and other symptoms). A further VAS evaluated satisfaction with general well-being (0 being extremely poor satisfaction and 10 very high satisfaction). These VASs were previously used by our group in a population with Non Celiac Gluten Sensitivity to evaluate gastrointestinal manifestations and general well-being [bib_ref] Evidence for the presence of non-celiac gluten sensitivity in patients with functional..., Elli [/bib_ref]. The magnitude of change in gastrointestinal symptoms between the baseline and at the end of the 21-day-long intervention was assessed as follows: (a) comparing each VAS score at both time points, and (b) estimating the number of patients achieving a change in VAS score for general well-being higher than or equal to 50% from the baseline.
# Statistical analysis
The data were described as median ± Standard Deviation (SD) or median (inter-quartile range), depending on the parametric or non-parametric distribution of variables as assessed by graphical inspection and the Shapiro-Wilk test. All the patients who had fully completed the intervention period, were included in the analysis (per-protocol analysis). For SCL-90 and SF-36 scores, a within-group comparison at both time points (the baseline and day 21) and a between-group comparison at day 21 was conducted using an independent Student's t-test or the non-parametric Wilcoxon rank sum test, depending on the distribution of variables. For the main outcome, the VAS score for general well-being, with two-way Analysis of Variance (ANOVA) (factors 'treatment' and 'time') with one repeated measure ('time'), was used. The magnitude of change in the VAS score for the general well-being comparison between groups at day 21 was evaluated by Fisher's exact two-tailed test. A 5% significance level was used, and the software packages STATA ® v. 13.1 (StataCorp LLC, College Station, TX, USA) and GraphPad Prism v. 6 (GraphPad Software, La Jolla, CA, USA) were used for analysis and figures processing.
# Results
## Patients
The participant flow is shown in [fig_ref] Figure 1: Change in the SF-36 questionnaire scores between the baseline and day 21... [/fig_ref]. The patients were middle-aged, mainly women, and within the normal weight range, according to mean body-mass index [fig_ref] Table 2: Background and gastrointestinal symptoms at baseline 1 [/fig_ref]. Regarding their clinical symptomatology at the baseline, 64% reported the presence of IBS-like symptoms, whereas 34% reported functional symptomatology [fig_ref] Table 2: Background and gastrointestinal symptoms at baseline 1 [/fig_ref]. Among the specific symptoms, diarrhea and constipation were the most frequently reported (34% and 32%, respectively) with lower frequencies of mixed and non-specified gastrointestinal symptoms (12% and 8%, respectively). At the baseline the LF-GFD and R-GFD groups were similar in relation to the presence of evaluated symptoms.
## Psychological symptoms and quality of life
A consistent reduction in most SCL-90 scores was found in the LF-GFD group but not in the R-GFD group, with the global SCL-90 score being significantly reduced (p < 0.0003) compared with the R-GFD group at day 21 (p < 0.04, [fig_ref] Table 3: SCL-90 scores according to studied groups 1 [/fig_ref] , and [fig_ref] Figure 1: Change in the SF-36 questionnaire scores between the baseline and day 21... [/fig_ref] in the Supplementary Materials). With respect to specific sub-items of the SCL-90, there were no differences in the R-GFD group. However, in the LF-GFD group some significant changes were observed between the baseline and day 21 in relation to the majority (7 out of 9) of the psychopathological dimensions [fig_ref] Table 3: SCL-90 scores according to studied groups 1 [/fig_ref].Data shown as mean ± SD or median (interquartile range) for non-parametrical variables. † p-value for comparison within groups using a non-parametric Wilcoxon rank sum test unless otherwise is indicated; ‡ independent t-test. R-GFD: regular gluten-free diet; LF-GFD: low-FODMAP gluten-free diet.
The results of the SF-36 scores are shown in [fig_ref] Table 4: Short Form [/fig_ref]. Overall, there were no differences in the SF-36 sub-scores both within and between groups through the intervention [fig_ref] Table 4: Short Form [/fig_ref] and [fig_ref] Figure 2: Cont. [/fig_ref] in the Supplementary Materials). However, when evaluating the change percentage at day 21, a statistically significant improvement in health perception, as well as in the physical functioning scores was found in the LF-GFD compared with the R-GFD group [fig_ref] Figure 1: Change in the SF-36 questionnaire scores between the baseline and day 21... [/fig_ref]. [fig_ref] Figure 1: Change in the SF-36 questionnaire scores between the baseline and day 21... [/fig_ref]. Change in the SF-36 questionnaire scores between the baseline and day 21 from intervention. Data shown as mean (symbol) ± SEM (upper and lower whiskers). For each sub-item, the difference between the values after intervention and the baseline was calculated and divided by the respective baseline value, expressed as a percentage, * p < 0.05; ‡ p = 0.06, for comparison between groups for each sub-item. R-GFD: regular gluten-free diet; and LF-GFD: low-FODMAP gluten-free diet.
## Gastrointestinal symptoms
A significant interaction was found with regard to the VAS score of abdominal pain with a significant decrease in the LF-GFD group versus the R-GFD group at day 21 (p < 0.01, [fig_ref] Figure 2: Cont. [/fig_ref] , and [fig_ref] Figure 3: Change in the VAS of well-being between the baseline and day 21... [/fig_ref] in the Supplementary Materials). The VAS score for satisfaction about fecal consistency showed a tendency for a higher increase in the LF-GFD group at day 21 (p < 0.09). Post-prandial fullness severity was lower in the LF-GFD group and decreased in both groups at day 21 (p < 0.006) but without significant interaction [fig_ref] Figure 2: Cont. [/fig_ref] and in the Supplementary Materials). No differences were found for non-specific functional gastrointestinal symptoms. For each sub-item, the difference between the values after intervention and the baseline was calculated and divided by the respective baseline value, expressed as a percentage, * p < 0.05; ‡ p = 0.06, for comparison between groups for each sub-item. R-GFD: regular gluten-free diet; and LF-GFD: low-FODMAP gluten-free diet.
## Gastrointestinal symptoms
A significant interaction was found with regard to the VAS score of abdominal pain with a significant decrease in the LF-GFD group versus the R-GFD group at day 21 (p < 0.01, [fig_ref] Figure 2: Cont. [/fig_ref] , and [fig_ref] Figure 3: Change in the VAS of well-being between the baseline and day 21... [/fig_ref] in the Supplementary Materials). The VAS score for satisfaction about fecal consistency showed a tendency for a higher increase in the LF-GFD group at day 21 (p < 0.09). Post-prandial fullness severity was lower in the LF-GFD group and decreased in both groups at day 21 (p < 0.006) but without significant interaction [fig_ref] Figure 2: Cont. [/fig_ref] and in the Supplementary Materials). No differences were found for non-specific functional gastrointestinal symptoms. For each sub-item, the difference between the values after intervention and the baseline was calculated and divided by the respective baseline value, expressed as a percentage, * p < 0.05; ‡ p = 0.06, for comparison between groups for each sub-item. R-GFD: regular gluten-free diet; and LF-GFD: low-FODMAP gluten-free diet.
## Gastrointestinal symptoms
A significant interaction was found with regard to the VAS score of abdominal pain with a significant decrease in the LF-GFD group versus the R-GFD group at day 21 (p < 0.01, [fig_ref] Figure 2: Cont. [/fig_ref] , and [fig_ref] Figure 3: Change in the VAS of well-being between the baseline and day 21... [/fig_ref] in the Supplementary Materials). The VAS score for satisfaction about fecal consistency showed a tendency for a higher increase in the LF-GFD group at day 21 (p < 0.09). Post-prandial fullness severity was lower in the LF-GFD group and decreased in both groups at day 21 (p < 0.006) but without significant interaction [fig_ref] Figure 2: Cont. [/fig_ref] and in the Supplementary Materials). No differences were found for non-specific functional gastrointestinal symptoms. The VAS score for satisfaction about general well-being was significantly enhanced in both groups. However, the improvement in well-being was greater in the LF-GFD group (p < 0.01, [fig_ref] Figure 3: Change in the VAS of well-being between the baseline and day 21... [/fig_ref]. Consistently, the evaluation of the change in this VAS score of general well-being from the baseline revealed a greater change at day 21 in the LF-GFD group as compared with the R-GFD group [fig_ref] Figure 3: Change in the VAS of well-being between the baseline and day 21... [/fig_ref]. The VAS score for satisfaction about general well-being was significantly enhanced in both groups. However, the improvement in well-being was greater in the LF-GFD group (p < 0.01, [fig_ref] Figure 3: Change in the VAS of well-being between the baseline and day 21... [/fig_ref]. Consistently, the evaluation of the change in this VAS score of general well-being from the baseline revealed a greater change at day 21 in the LF-GFD group as compared with the R-GFD group [fig_ref] Figure 3: Change in the VAS of well-being between the baseline and day 21... [/fig_ref].
# Discussion
To our knowledge this is the first randomized double-blind intervention-controlled study that has investigated the effects of a LFD on patients with CD following GFDs but with persisting functional gastrointestinal symptoms. Our results showed a positive response to LFD, an improvement in psychological health scores-but only a limited change in quality of life-and a significant improvement in gastrointestinal symptoms with improved perception of well-being by the patients on the LFD.
Foods containing FODMAP can trigger IBS-like symptoms. These dietary compounds can trigger an increase in flatulence, diarrhea, and bloating that may lead to abdominal pain [bib_ref] Evidence-based dietary management of functional gastrointestinal symptoms: The FODMAP approach, Gibson [/bib_ref]. Our results suggest that LFD may improve persistent gastrointestinal symptomatology in those patients who undergo GFD and also successfully improve the psychological aspects already described in this group of patients [bib_ref] Anxiety but not depression decreases in coeliac patients after one-year gluten-free diet:..., Addolorato [/bib_ref]. After our intervention, the severity of gastrointestinal symptoms, such as abdominal pain and stool consistency, decreased when compared with the situation at the baseline and with the R-GFD group, along with improvement in the general well-being VAS. These results are in agreement with those reported by Halmos and colleagues, who showed that a low-FODMAP diet effectively reduced functional gastrointestinal symptoms in patients with IBS [bib_ref] A diet low in FODMAPs reduces symptoms of irritable bowel syndrome, Halmos [/bib_ref]. On the other hand, a study conducted on patients with inflammatory bowel disease has showed a positive response to LFD, thus suggesting that a reduction in FODMAP intake offers an efficacious strategy for those patients who present with concurrent functional gastrointestinal symptoms [bib_ref] Reduction of dietary poorly absorbed short-chain carbohydrates (FODMAPs) improves abdominal symptoms in..., Gearry [/bib_ref]. Although we were able to show an effect on symptomatology as measured by VAS in the LF-GFD group, we also observed that the patients receiving GFD reinforcement (our comparison group) did also show a decrease in symptomatology. This issue has already been addressed in Sainsbury and co-workers, [bib_ref] Prevalence of irritable bowel syndrome-type symptoms in patients with celiac disease: A..., Sainsbury [/bib_ref] , who emphasized that in-depth dietary revision as carried out by a trained nutritionist can improve any persistent symptomatology in CD patients. Nevertheless, our results have showed that such improvement was to a greater extent in the LF-GFD group.
Together with the improvements in IBS-like symptoms, we have also found that LF-GFD can overall improve the psychopathological symptoms as measured by a well-validated instrument. The relationship between psychological and psychiatric disturbances and CD is already well-established,
# Discussion
To our knowledge this is the first randomized double-blind intervention-controlled study that has investigated the effects of a LFD on patients with CD following GFDs but with persisting functional gastrointestinal symptoms. Our results showed a positive response to LFD, an improvement in psychological health scores-but only a limited change in quality of life-and a significant improvement in gastrointestinal symptoms with improved perception of well-being by the patients on the LFD.
Foods containing FODMAP can trigger IBS-like symptoms. These dietary compounds can trigger an increase in flatulence, diarrhea, and bloating that may lead to abdominal pain [bib_ref] Evidence-based dietary management of functional gastrointestinal symptoms: The FODMAP approach, Gibson [/bib_ref]. Our results suggest that LFD may improve persistent gastrointestinal symptomatology in those patients who undergo GFD and also successfully improve the psychological aspects already described in this group of patients [bib_ref] Anxiety but not depression decreases in coeliac patients after one-year gluten-free diet:..., Addolorato [/bib_ref]. After our intervention, the severity of gastrointestinal symptoms, such as abdominal pain and stool consistency, decreased when compared with the situation at the baseline and with the R-GFD group, along with improvement in the general well-being VAS. These results are in agreement with those reported by Halmos and colleagues, who showed that a low-FODMAP diet effectively reduced functional gastrointestinal symptoms in patients with IBS [bib_ref] A diet low in FODMAPs reduces symptoms of irritable bowel syndrome, Halmos [/bib_ref]. On the other hand, a study conducted on patients with inflammatory bowel disease has showed a positive response to LFD, thus suggesting that a reduction in FODMAP intake offers an efficacious strategy for those patients who present with concurrent functional gastrointestinal symptoms [bib_ref] Reduction of dietary poorly absorbed short-chain carbohydrates (FODMAPs) improves abdominal symptoms in..., Gearry [/bib_ref]. Although we were able to show an effect on symptomatology as measured by VAS in the LF-GFD group, we also observed that the patients receiving GFD reinforcement (our comparison group) did also show a decrease in symptomatology. This issue has already been addressed in Sainsbury and co-workers, [bib_ref] Prevalence of irritable bowel syndrome-type symptoms in patients with celiac disease: A..., Sainsbury [/bib_ref] , who emphasized that in-depth dietary revision as carried out by a trained nutritionist can improve any persistent symptomatology in CD patients. Nevertheless, our results have showed that such improvement was to a greater extent in the LF-GFD group.
Together with the improvements in IBS-like symptoms, we have also found that LF-GFD can overall improve the psychopathological symptoms as measured by a well-validated instrument. The relationship between psychological and psychiatric disturbances and CD is already well-established, significantly influencing the reduction of the quality of life and worsening the symptoms of affected patients [bib_ref] Coeliac disease and psychiatric comorbidity: Epidemiology, pathophysiological mechanisms, quality-of-life, and gluten-free diet..., Cossu [/bib_ref]. The results related to the SCL-90 questionnaire have shown post-intervention differences in the LF-GFD group but none for the patients on the R-GFD. In the former group, in fact, we have observed a change in the vast majority of the dimensions as compared with the baseline, suggesting that the decrease in gastrointestinal symptoms may positively influence the improvement of a patient's overall psychopathological burden. These results are consistent with those reported by a previous study, in which patients with CD undertook GFDs and for whom a post-intervention change was observed, which determined a decrease in the score for anxiety but not for depression [bib_ref] Anxiety but not depression decreases in coeliac patients after one-year gluten-free diet:..., Addolorato [/bib_ref].
In another report some patients newly diagnosed with CD were evaluated in relation to the dimensions of SCL-90 and compared with a healthy control group: the scores for somatization, obsessive-compulsive, interpersonal sensitivity, depression, anxiety, and sleep were found to be higher in CD patients [bib_ref] Prevalence of eating disorders in adults with celiac disease, Passananti [/bib_ref]. The use of SCL-90 in a population that suffers from gastrointestinal symptoms has been already evaluated in IBS patients, who exhibited significantly more distress compared with other groups. In addition, the patients with gastrointestinal symptoms as a group, compared with the healthy controls, were characterized by high levels of irritable depression and somatization [bib_ref] Psychopathological symptom dimensions in patients with gastrointestinal disorders, Kovács [/bib_ref]. From this perspective, the results from the present study, which shows a significant psychopathological improvement in CD patients on the LF-GFD versus the R-GFD, provide an important confirmation about the relationship between gastrointestinal and mental health in CD patients who undergo different types of diet. In particular, to our knowledge this present study offers the first report demonstrating a significant amelioration for CD patients of most SCL-90 items in the short-term (i.e., after 21 days) as a result of following a LF-GFD.
The quality-of-life perception, as assessed by means of the SF-36 questionnaire, has shown only minor differences between the studied groups: we could not establish for this group of celiac patients any significant improvement in this regard after the LFD. This finding is in discordance with what was previously reported. Previous data about our group of patients with non-celiac gluten sensitivity treated with GFDs, showed an improvement in the majority of the SF-36 scores after 7 days of treatment in a cross-over study, with both mental and physical components of the SF-36 questionnaire being significantly lower for patients positive to a gluten challenge [bib_ref] Evidence for the presence of non-celiac gluten sensitivity in patients with functional..., Elli [/bib_ref]. Other authors have shown a quality-of-life improvement in patients with atypical and typical CD, compared with healthy controls after a one-year-long treatment, but only with differences in two items (general health and vitality) for subjects with typical CD [bib_ref] Quality of life in screen-detected and typical coeliac disease and the effect..., Johnston [/bib_ref].
Of note, even though we were not able to demonstrate any improvement in quality of life when comparing our study groups, when the percent change was evaluated for each of the items consulted, both health perception and physical functioning turned out higher in the LF-GFD group compared with the R-GFD group. Therefore, we could not rule out that the aforementioned mixed results possibly depended on the limited period of observation, with a longer follow-up period to be required in order to better assess the changes in quality of life for CD patients following the LF-GFD.
Among the strengths of our report there is the fact that it comes from the first randomized double-blind study performed on patients with CD and to evaluate the potential effects of a reduction in dietary FODMAP on overall health and gastrointestinal symptomatology. As a limitation, even if we could show significant improvement in clinical symptoms, our results were obtained only after a short period of time (i.e., limited to three weeks) and only on patients who had fully completed the intervention (92% and 84% in the R-GFD and LF-GFD groups, respectively).
In conclusion, our results show that nutritional intervention by a LFD can have beneficial effects for CD patients who are on a GFD but present with persisting functional gastrointestinal disorders, even without major changes in their quality-of-life perception. The same results also suggest that, for those patients with CD being treated with a GFD and experiencing IBS-like symptoms, a LFD can be indicated by a trained nutritionist, but its beneficial effects and long-term clinical effects for this group of selected CD patients need further investigation.
ClinicalTrials.gov, ref. no. DNCT02946827.
Supplementary Materials: The following are available online at http://www.mdpi.com/2072-6643/10/8/1023/s1, [fig_ref] Figure 1: Change in the SF-36 questionnaire scores between the baseline and day 21... [/fig_ref] : SCL-90 global index, [fig_ref] Figure 2: Cont. [/fig_ref] : SF-36 score for general health, [fig_ref] Figure 3: Change in the VAS of well-being between the baseline and day 21... [/fig_ref] : VAS score for abdominal pain, : VAS score for fecal consistency, : VAS score for postprandial fullness severity.
## Conflicts of interest:
The authors declare no conflict of interest.
# Appendix a
Nutrients 2018, 10, x FOR PEER REVIEW 12 of 14
Supplementary Materials: The following are available online at www.mdpi.com/xxx/s1, [fig_ref] Figure 1: Change in the SF-36 questionnaire scores between the baseline and day 21... [/fig_ref] : SCL-90 global index, [fig_ref] Figure 2: Cont. [/fig_ref] : SF-36 score for general health, [fig_ref] Figure 3: Change in the VAS of well-being between the baseline and day 21... [/fig_ref] : VAS score for abdominal pain, : VAS score for fecal consistency, : VAS score for postprandial fullness severity.
[fig] Figure 1: Change in the SF-36 questionnaire scores between the baseline and day 21 from intervention. Data shown as mean (symbol) ± SEM (upper and lower whiskers). [/fig]
[fig] Figure 2: Cont. [/fig]
[fig] Figure 3: Change in the VAS of well-being between the baseline and day 21 from intervention. VAS score for overall well-being was evaluated at the baseline and at the end of the intervention period (day 21) for the R-GFD (A) and LF-GFD (B) groups; the magnitude of change in well-being perception (C) was calculated by estimating the number of patients (shown in percentage for each group) with a change in VAS score greater than or equal to 50% of the baseline value. Data are individual values at both time points. R-GFD: regular gluten-free diet; LF-GFD: low-FODMAP gluten-free diet; VAS: visual analogue scale. * p = 0.03 Fisher´s exact test. [/fig]
[fig] Author: Contributions: Conceptualization, L.R., K.A.B., and L.E.; Methodology, L.R., K.A.B., and L.E.; Investigation, L.R., A.S., V.L., F.B., F.F., and L.E.; Data Curation, L.R.; Formal analysis, L.R., K.A.B., and L.E.; Original Draft Preparation, L.R. and K.A.B.; Review and Editing of Manuscript, L.R., K.A.B., B.D.O., L.D., M.T.B., and L.E.; and Funding Acquisition, L.E. Funding: This research was funded by Fondazione IRCCS Ca' Granda and received grants from Italy's Ministry of Health and Lombardy's Regional Government ity (Ministero della Salute e Regione Lombardia, grant number 2011-02348234. The Article Processing Charge was funded by Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico and Università degli Studi di Milano, Milan, Italy. [/fig]
[fig] Figure A1: CONSORT Flow diagram. [/fig]
[table] Table 1: Examples of the two different prescribed diets for a typical day 1 . [/table]
[table] Table 2: Background and gastrointestinal symptoms at baseline 1 . [/table]
[table] Table 3: SCL-90 scores according to studied groups 1 . [/table]
[table] Table 4: Short Form (36) Health Survey (SF-36) subscales and global score 1 . [/table]
|
Anthranilate Acts as a Signal to Modulate Biofilm Formation, Virulence, and Antibiotic Tolerance of Pseudomonas aeruginosa and Surrounding Bacteria
## Name
Genotype References
## Pseudomonas aeruginosa
|
Quantitative and pattern recognition analyses of magnoflorine, spinosin, 6′′′-feruloyl spinosin and jujuboside A by HPLC in Zizyphi Semen
Two rapid and simple HPLC methods with UV detector to determine three main compounds (magnoflorine, spinosin and 6 000 -feruloyl spinosin) and evaporative light scattering detector (ELSD) to determine jujuboside A were developed for the chemical analyses of Zizyphi Semen. Magnoflorine, spinosin, and 6 000 -feruloyl spinosin were separated with an YMC J'sphere ODS-H80 column (250 mm 9 4.6 mm, 4 lm) by the gradient elution followed by the isocratic elution using methanol with 0.1 % formic acid and water with 0.1 % formic acid as the mobile phase. The flow rate was 1.0 mL/min. Jujuboside A was separated by HPLC-ELSD with YoungJinBioChrom Aegispak C18-L column (250 mm 9 4.6 mm, 5 lm) column in a gradient elution using methanol with 0.1 % formic acid (A) and water with 0.1 % formic acid as the mobile phase. These two methods were fully validated with respect to linearity, precision, accuracy, stability, and robustness. These HPLC methods were applied successfully to quantify four compounds in a Zizyphi Semen extract. The HPLC analytical methods were validated for pattern recognition analysis by repeated analysis of 91 seed samples corresponding to 48 Zizyphus jujuba var. spinosa (J01-J48) and 43 Zizyphus mauritiana (M01-M43). The results indicate that these methods are suitable for a quality evaluation of Zizyphi Semen.
# Introduction
Zizyphi Semen is the dried seeds of Zizyphus jujuba Miller var. spinosa Hu ex H. F. Chou (Z. jujuba var. spinosa) in the Korean Pharmacopoeia (K.P.) and the Chinese Pharmacopoeia (C.P.), and belongs to the Rhamnaceae family [bib_ref] Three triterpene esters from Zizyphus jujube, Lee [/bib_ref]. It is distributed mainly in tropical and subtropical regions of the world. Zizyphi Semen has been used as an analgesic, a tranquilizer, and an anticonvulsant in oriental countries such as Korea and China for over 2,500 years [bib_ref] Anxiolytic-like effects of sanjoinine A isolated from Zizyphi Spinosi Semen: possible involvement..., Han [/bib_ref]. It has been used as an anticonvulsant and for treating anxiety and insomnia in folk medicine in India [bib_ref] Interaction profile of Zizyphus jujuba with phenytoin, phenobarbitone, and carbamazepine in maximal..., Pahuja [/bib_ref] , and for treating depression, insomnia, and anxiety in other Asian countries [bib_ref] Antidepressant-like effect of the ethanolic extract from Suanzaorenhehuan formula in mice models..., Liu [/bib_ref].
Studies have found that Zizyphi Semen possesses beneficial effects on the cardiovascular system such as antiarrhythmia and anti-hypertension [bib_ref] Betulinic acid ameliorates endothelium-dependent relaxation in L-NAME-induced hypertensive rats by reducing oxidative..., Fu [/bib_ref] , antianxiety [bib_ref] Anxiolytic effect of seed of Zizyphus jujuba in mouse models of anxiety, Peng [/bib_ref] , amelioration of seizures and oxidative stress [bib_ref] Hydroalcoholic extract of Zizyphus jujuba ameliorates seizures, oxidative stress, and cognitive impairment..., Pahuja [/bib_ref] , enhancement of pentobarbital-induced sleep [bib_ref] Cyclopeptide alkaloid fraction from Zizyphi Spinosi Semen enhances pentobarbital-induced sleeping behaviors, Ma [/bib_ref] , protection of N-methyl-D-aspartate-induced neuronal cell damage [bib_ref] Protection of NMDA-induced neuronal cell damage by methanol extract of Zizyphi Spinosi..., Park [/bib_ref] , inhibition of histamine release [bib_ref] CML-1 inhibits TNF-a-induced NF-jB activation and adhesion molecule expression in endothelial cells..., Mao [/bib_ref] , reduction of atherosclerosis by inhibiting foam cell formation and prevention of foodborne pathogens [bib_ref] Potential roles of essential oil and organic extracts of Zizyphus jujuba in..., Al-Reza [/bib_ref].
Magnoflorine, one of main alkaloid components in Zizyphi Semen [bib_ref] Flavonoids from the seeds of Zizyphus jujuba var. spinosa, Lee [/bib_ref] , has anti-glycemic [bib_ref] Magnoflorine from Tinospora cordifolia stem inhibits a-glucosidase and is antiglycemic in rats, Patel [/bib_ref] and antioxidant [bib_ref] Antiradical and antioxidant activities of alkaloids isolated from Mahonia aquifolium, Rackova [/bib_ref] activities. Spinosin, another major flavonoid compound [bib_ref] Flavonoids from the seeds of Zizyphus jujuba var. spinosa, Lee [/bib_ref] , potentiates pentobarbital-induced sleep via a serotonergic mechanism [bib_ref] Spinosin, a C-glycoside flavonoid from Semen Zizyphi spinosae, potentiated pentobarbital-induced sleep via..., Wang [/bib_ref] [bib_ref] Potentiating effect of spinosin, a C-glycoside flavonoid of Semen Zizyphi spinosae, on..., Wang [/bib_ref]. Jujuboside A, the other main component [bib_ref] Flavonoids from Zizyphus jujuba Mill var. spinosa, Cheng [/bib_ref] , has been studied for its effect on hippocampal neurons of rat [bib_ref] Effects on the expression of GABA A receptor subunits by jujuboside A..., You [/bib_ref] and for insomnia .
The regulation of Zizyphi Semen content in C.P. (2010) has been already stipulated in 2010; it is prescribed to contain no less than 0.08 % spinosin and 0.03 % jujuboside A from Z. jujuba var. spinosa. However, the K.P. has no stipulation on the main compounds contained in Zizyphi Semen. The purpose of this study was to establish a reliable high-performance liquid chromatographic (HPLC) method to quantitatively analyze the major compounds in Zizyphi Semen, and to provide analytical method which would be used as the official analytical method in K.P. revision. The dried seeds of Z. mauritiana, which are normally distributed and cropped in low-latitudes of Asia, Africa, and Australia [bib_ref] Zizymauritic acids A-C, three novel nortriterpenes from Zizyphus mauritiana, Ji [/bib_ref] , is mislabeled as Zizyphi Semen in Korean herbal markets. Therefore, we also suggest analytical marker compounds to distinguish the seeds of Z. jujuba var. spinosa from those of Z. mauritiana.
In previous studies, several analytical methods such as ultraviolet spectrophotometry [bib_ref] Quantitative determination of total flavonoids in Semen Ziziphi Spinosae by UV spectrophotometry, Li [/bib_ref] , liquid chromatography/mass spectroscopy [bib_ref] Simultaneous analysis and identification of main bioactive constituents in extract of Zizyphus..., Liu [/bib_ref] [bib_ref] Quantitative determination of spinosin in rat plasma by liquid chromatography-tandem mass spectrometry..., Li [/bib_ref] , ultra-high-performance liquid chromatography coupled with diode-array detector (UPLC-DAD) [bib_ref] Quantitative analysis and chromatographic fingerprinting of the Semen Zizyphi Spinosae by ultrahigh-performance..., Niu [/bib_ref] and HPLC-UV [bib_ref] Determination of flavonoids in seeds of Zizyphus vulgaris var. spinosus by high..., Shin [/bib_ref] have been established to quantify or identify the components in Zizyphi Semen. Ultraviolet spectrophotometry, targeting spinosin and other flavonoids, is a simple method but does not provide detailed chemical information like retention times of magnoflorine and jujuoboside A [bib_ref] Quantitative determination of total flavonoids in Semen Ziziphi Spinosae by UV spectrophotometry, Li [/bib_ref]. HPLC-photo diode array detection and HPLC-DADelectrospray ionization-mass spectroscopy (HPLC-DAD-ESI-MS) method had been developed to identify 11 compounds including spinosin, 6 000 -feruloyl spinosin, and jujuboside A in Zizyphi Semen, in which a complicated elution method more than 6 steps and long running time (65 min) were used [bib_ref] Simultaneous analysis and identification of main bioactive constituents in extract of Zizyphus..., Liu [/bib_ref]. UPLC-DAD has been applied for chromatographic fingerprint analysis and quantitative analysis of six flavonoids to classify and discriminate 23 Zizyphi Semen samples, but had complicated elution conditions like poor elution times of 15.45 or 22.95 min. HPLC chromatogram also exhibited some overlapped peaks of marker compounds [bib_ref] Quantitative analysis and chromatographic fingerprinting of the Semen Zizyphi Spinosae by ultrahigh-performance..., Niu [/bib_ref]. In C.P., two methods, such as HPLC-UV to determine spinosin and HPLC-ELSD to determine jujuboside A, have been used to assay marker compounds in Zyziphi Semen. However, complicated elution conditions were used for both methods. Magnoflorine, a major marker compound with different content between Z. jujuba var. spinosa and Z. mauritiana resulted from this study, was not adopted as a marker compound for Zyziphi Semen.
In this study, newly developed method not only has short analytical time but also shows good resolution. Magnoflorine, one of maker compounds, was not adopted in conventional experiments, even though it was regarded as an important marker compound in this study.
We suggest a suitable analytical method for quantitative and pattern recognition analyses of Zyziphi Semen together with the establishment of appropriate marker compounds to distinguish between Z. jujuba var. spinosa and Z. mauritiana.
# Materials and methods
# Reagents and materials
The magnoflorine (1), spinosin (2), 6 000 -feruloyl spinosin (3), and jujuboside A (4) standards were kindly provided by the Zizyphi Semen separation team of Korean National Center for Standardization of Herbal Medicines, which were separated from Z. jujuba var. spinosa. The internal standards (I.S.), naringin (5) and nargingenin (6), were purchased from Sigma-Aldrich (St. Louis, MO, USA). The compound structures are shown in . The purities of these compounds were determined to be [98 % by normalizing the peak areas detected by HPLC analyses. Methanol was purchased from Merck K GaA (Darmstadt, Germany). All other chemicals used were analytical grade. Deionized water was prepared using the Milli-Q purification system (Millipore, Bedford, MA, USA). This study adopted the seed samples of 48 Z. jujuba var. spinosa (J01-J48), and 43 Z. mauritiana (M01-M43). All Z. jujuba var.
## Sample preparation
Each standard stock solution was prepared by adding 1.0 mg magnoflorine, spinosin and 6 000 -feruloyl spinosin to 1.0 mL of methanol containing 80 ppm naringin, respectively. A standard stock solution was prepared by adding 1.0 mg jujuboside A to 1.0 mL of methanol containing 50 ppm naringenin.
A powdered sample of Zyziphi Semen (1.0 g) for HPLC-UV was mixed with 50 mL of 50 % methanol containing 80 ppm I.S. (naringin) in a vial and the mixture was refluxed for 30 min. A powdered sample of Zyziphi Semen (5.0 g) for HPLC-ELSD was mixed with 50 mL of 50 % methanol containing 50 ppm I.S. (naringenin) in a vial. Each mixture was sonicated for 30 min. The solution was weighed again, and the loss in weight was made up with methanol. The solution was filtered through a 0.45-lm membrane filter (Whatman), and the filtrate was used as the test solution. A 10 lL aliquot of the test solution was injected into the HPLC system.
## Hplc-uv conditions
The HPLC equipment was a Waters HPLC system (Empower pro) with a Waters 600 pump, a Waters 486 tunable absorbance detector and Waters 717 autosampler (Waters Inc., Milford, MA, USA). Three different columns were used and compared: YMC J'sphere ODS-H80 (250 mm 9 4.6 mm, 4 lm), YoungJinBioChrom Aegispak C18-L (250 mm 9 4.6 mm, 5 lm) and Phenomenex Gemini ODS C18 (250 mm 9 4.6 mm, 5 lm). The mobile phase consisted of water containing 0.1 % formic acid (A) and methanol containing 0.1 % formic acid (B). Elution was performed at a flow rate of 1 mL/min in gradient and isocratic modes. The solvent gradient was changed
## Hplc-elsd conditions
The HPLC equipment was a Gilson HPLC system (Unipoint 2.0) with a Gilson 321 pump, a Gilson Prep TM II ELSD detector and Gilson 321 XL auto-sampler (Gilson Inc. Middleton, WI, USA). The above three different columns were compared in HPLC-ELSD and two mobile phases, A and B, were also same with HPLC-UV. Elution was performed at a flow rate of 0.1 mL/min in a gradient mode. The solvent gradient was changed according to the following program: from 45 % (A):55 % (B) to 25 % (A):75 % (B) at 0-30 min. The column was washed by 100 % of (B) for 20 min and re-equilibrated by 45 % (A):55 % (B) for 20 min. The mobile phase was filtered under vacuum through a 0.21-lm membrane filter and was degassed prior to use. The ELSD parameters of the spray chamber and drift tube temperatures, and gas pressure were optimized at 30, 60°C and 50 psi, respectively.
Magnoflorine (1) Spinosin (2) 6 '''-Feruloyl spinosin (3) Jujuboside A (4) N a r i n gin (5) Naringenin (6) Fig.
## Limits of detection and quantification
The lowest concentration of working solution was diluted with appropriate concentrations, and LOD and LOQ under the chromatographic conditions were separately determined at signal-to-noise ratios (S/N) of about 3 and 10, respectively.
## Accuracy and precision
Precision and accuracy were determined in HPLC-UV by spiking three concentration levels of the magnoflorine, spinosin, and 6 000 -feruloyl spinosin standards, which were mixed with a Zyziphi Semen (J14) sample for subsequent extraction and filtration. Three concentrations of 0.9, 90.0, and 135.0 lg/mL for magnoflorine and spinosin, and 1.0, 100.0, and 150.0 lg/mL for 6 000 -feruloyl spinosin were evaluated. Precision and accuracy in HPLC-ELSD were determined as the same way except three concentrations of 40.0, 100.0, and 200.0 lg/mL were used with the jujuboside A standard. The HPLC-UV and HPLC-ELSD analytical experiments were performed in triplicate for each control level. Data from the standard solution and the extracted sample were compared. Precision and accuracy were determined by multiple analyses (n = 5) of quality control samples prepared at low, medium and high concentrations spanning the calibration range.
## Robustness
The robustness of the method was studied by introducing changes in the column (i.e., J'sphere, Aegispak, Gemini), column temperature (i.e., 25, 30, 35, and 40°C) and flow rates (i.e., 0.8, 1.0, and 1.2 mL/min).
# Pattern recognition analysis
A pattern recognition analysis was conducted to evaluate the phytochemical equivalency among the 91 samples (48 Z. jujuba var. spinosa (J01-J48), 43 Z. mauritiana (M01-M43) samples). We used two major marker compound HPLC-UV peaks of magnoflorine and spinosin, and one major marker compound HPLC-ELSD peak of jujuboside A for the pattern recognition analysis using IBM SPSS Statistics Version 19 software (SPSS, Inc., Chicago, IL, USA).
# Results
## Optimization of chromatographic conditions
HPLC conditions were selected to obtain good resolution on the chromatograms within a short retention time. We investigated YMC J'sphere ODS-H80, YoungJinBioChrom Aegispak C18-L, and Phenomenex Gemini ODS C18 columns to optimize the HPLC-UV chromatographic conditions. These three columns showed similar results, but ODS-H80 showed better resolution and theoretical plate of each peak in robustness. Above three columns also showed similar results for HPLC-ELSD, but Aegispak C18-L showed better resolution and theoretical plate for jujuboside A. UV detector was used for magnoflorine, spinosin and 6 000 -feruloyl spinosin because these compounds have good absorption in UV wavelengths. We used 270 nm because this was the maximum absorption of the three compounds. Mobile phase of water-methanol was adequate for good resolution of compounds during UV and ELSD. Adding 0.1 % formic acid to both water and methanol significantly improved the separation. Furthermore, we set the ELSD parameters for a spray chamber and drift tube temperatures, and gas pressure, with the purpose of generating a reproducible jujuboside A peak. Ultimately, the optimal mobile phase was a 0.1 % formic acid in methanol and a 0.1 % formic acid in deionized water in the gradient elution followed by the isocratic elution mode. Typical chromatograms of the sample and standard mixtures are shown in Figs. 2 and 3; the target compounds including I.S. were completely separated within 40 min by UV, and 30 min by ELSD. Naringin was selected as the I.S. for UV, and naringenin for ELSD .
## Optimization of the sample preparation conditions
Four extracting solvents, including 70 % ethanol, 50 % ethanol, 70 % methanol, and 50 % methanol containing 80 ppm of naringin (I.S.) for HPLC-UV and containing 50 ppm of naringenin (I.S.) for HPLC-ELSD, were compared in sample assays after extraction by sonication for 30 min at room temperature. When the samples were extracted with 50 % methanol, the sample assays were higher than the other solvent samples in both methods. Therefore, we employed 50 % methanol as the extracting solvent throughout this work . Ultra-sonication and reflux using each 50 % methanol extraction solvent Extraction by reflux showed better results than extraction by sonication for HPLC-UV. However, extraction by sonication showed better results than extraction by reflux for HPLC-ELSD . To determine the time needed to complete the extraction, samples were extracted for 30, 45, 60, 90 and 120 min. When the extraction time was set to 30 min, the sample assay results were similar to those of the others in both methods. Therefore, all of the compounds were sufficiently extracted when the extraction time was 30 min . The stability of naringin was compared between standing at room temperature and reflux at 80°C for 30 min in 50 % methanol.
Linearity, calibration range, and limits of detection and quantification
The calibration curves showed good linearity (r 2 [ 0.999) within the test ranges, as shown in [fig_ref] Table 1: Linearity, linear ranges, LOD and LOQ [/fig_ref]. The stock solution containing the reference compound was diluted with methanol to give a series of appropriate concentrations and the aliquots of the diluted solutions were injected. The LOD (S/N = 3) and LOQ (S/N = 10) values for magnoflorine, spinosin, 6 000 -feruloyl spinosin, and jujuboside A are presented in [fig_ref] Table 1: Linearity, linear ranges, LOD and LOQ [/fig_ref]. The values for both LOD and LOQ for these four standards were low enough to detect traces of these compounds in either a crude extract or its preparation.
## Precision and accuracy
The extraction precision and accuracy were assessed by extracting a known amount of compounds from Zizyphi Semen powdered samples. Known amounts of each standard compound at three levels were mixed with the sample powder and then extracted with 50 % methanol. Average recovery was calculated by the formula: R (%) = [(amount from the sample spiked standard -amount from the sample)/amount from the spiked standard] 9 100. Intraassay precision and accuracy were determined from the variability obtained from multiple analyses (n = 5) of quality control samples analyzed within the same analytical run. The quality control samples had intra-assay precision B4.82 % and accuracy of 95.18-101.37 %. Inter-assay precision and accuracy were evaluated from the differences in multiple analyses (n = 3) of quality control samples analyzed for 3 consecutive days. The quality control samples had an inter-assay precision of B3.17 % and accuracy of 97.61-101.87 %. Thus, the methods were highly reproducible. The precision and accuracy data are presented in [fig_ref] Table 2: Precision and accuracy of analytical results [/fig_ref].
## Robustness
Robustness was determined to evaluate the reliability of the established HPLC method. The experimental conditions, such as column temperature, column species and flow rates, were purposely altered, and the theoretical plate (N), retention factor (k), separation factor (a) and resolution (Rs) were evaluated. The four analytical factors showed that the experimental conditions were sufficiently robust (data not shown).
# Sample analysis
The HPLC method was applied to analyze 91 samples corresponding to the seeds of 48 Z. jujuba var. spinosa (J01-J48) and 43 Z. mauritiana (M01-M43) samples. The average contents (wt%) of magnoflorine, spinosin, 6 000feruloyl spinosin, and jujuboside A are presented in [fig_ref] Table 3: Average contents [/fig_ref]. The average content of magnoflorine (0.156 %) in the Z. jujuba var. spinosa samples was higher than that of Z. mauritiana (0.055 %). In contrast, the average contents of spinosin (0.104 %) and 6 000 -feruloyl spinosin (0.040 %) in the Z. jujuba var. spinosa samples was lower than those of spinosin (0.142 %) and 6 000 -feruloyl spinosin (0.052 %) in Z. mauritiana. Interestingly, the average content of jujuboside A in the Z. jujuba var. spinosa samples was 0.058 %, whereas there was no jujuboside A in Z. mauritiana. This quantitative analysis results of magnoflorine, spinosin, 6 000 -feruloyl spinosin, and jujuboside A will be reflected in the contents regulation of these four compounds for Zizyphi Semen in the next revision of the K.P.
# Pattern recognition analysis
To evaluate the phytochemical equivalency among the seeds of the 48 Z. jujuba var. spinosa (J01-J48) and 43 Z. mauritiana (M01-M43) samples, pattern recognition analysis was conducted using the contents of three (magnoflorine, spinosin, and jujuboside A) and four (magnoflorine, spinosin, 6 000 -feruloyl spinosin, and jujuboside A) marker compounds. The content of 6 000 -feruloyl spinosin did not affect the result of the pattern recognition, because the average 6 000 -feruloyl spinosin content between Z. jujuba var. spinosa and Z. mauritiana was not much different compared to that of the other three marker compounds. Therefore pattern recognition analysis was conducted using the magnoflorine, spinosin and jujuboside A contents. Consequently, considering the concatenation of the three compounds which was significantly different between two species of Z. jujuba var. spinosa and Z. mauritiana, all of the samples were divided into two groups, Z. jujuba var. spinosa (A) and Z. mauritiana (B), by the pattern analysis .
# Discussion
We have provided a fully validated HPLC method for quality control of Zizyphi Semen and pattern recognition analysis resulted in distinguishing between Z. jujuba var. spinosa and Z. mauritiana. The analytical conditions using a simple gradient elution system with UV and ELSD detectors allowed for a concise experiment and enhanced the analytical conditions. Our results suggest that magnoflorine, spinosin and jujuboside A are marker compounds for quality evaluations of Zizyphi Semen. Magnoflorine was not adopted as a marker compound of Zizyphi Semen in the C.P., even though the magnoflorine content was higher than that of spinosin from our assay results. Consequently, we suggest that including magnoflorine together with spinosin and jujuboside A as marker compounds is more reasonable compared with the marker compounds (spinosin and jujuboside A) currently in the C.P.
[fig] Figure 2: HPLC-UV chromatograms of standard mixture (a), the sample of Z. jujuba var. spinosa (J01, b) and the sample of Z. mauritiana (M01, c). 1 Magnoflorine, 2 Spinosin, 3 6 000 -Feruloyl spinosin, 5 Naringin [/fig]
[fig] Figure 3, Figure 4: HPLC-ELSD chromatograms of standard mixture (a), the sample of Z. jujuba var. spinosa (J01; b) and the sample of Z. mauritiana (M01; c). 4 Jujuboside A, 6 Naringenin Comparison of the extraction solvents for extraction efficiencies of marker compounds (n = 3, ppm of naringin (I.S.) for HPLC-UV and containing 50 ppm of naringenin (I.S.) for HPLC-ELSD were compared as extraction methods in sample assays. [/fig]
[fig] Figure 5, Figure 6: Comparison of the extraction methods (sonication and reflux) for extraction efficiencies of marker compounds (n = 3, w/w %) Comparison of the extraction time for extraction efficiencies of marker compounds (1 = 3, w/w %) [/fig]
[table] 1: Structures of standards and an internal standardsThe developed HPLC method was validated according to Korea Food and Drug Administration (KFDA) guidelines for the following parameters: linearity, limits of detection (LOD), limits of quantification (LOQ), accuracy, precision, and robustness. [/table]
[table] Table 2: Precision and accuracy of analytical results [/table]
[table] Table 3: Average contents (wt%) of magnoflorine, spinosin, 6 000 -feruloyl spinosin, and jujuboside A in Zizyphi Semen Mean ± SD (wt%) Each value represents the mean ± SD (n = 3) [/table]
|
Breast Cancer-Derived Microparticles Reduce Cancer Cell Adhesion, an Effect Augmented by Chemotherapy
Tumor cell heterogeneity is primarily dictated by mutational changes, sometimes leading to clones that undergo a metastatic switch. However, little is known about tumor heterogeneity following chemotherapy perturbation. Here we studied the possible involvement of tumor-derived extracellular vesicles, often referred to as tumor-derived microparticles (TMPs), as mediators of the metastatic switch in the tumor microenvironment by hindering cell adhesion properties. Specifically, we show that highly metastatic or chemotherapy-treated breast cancer cells shed an increased number of TMPs compared to their respective controls. We found that these TMPs substantially reduce cell adhesion and disrupt actin filament structure, therefore increasing their biomechanical force pace, further implicating tumor cell dissemination as part of the metastatic cascade. Our results demonstrate that these pro-metastatic effects are mediated in part by CD44 which is highly expressed in TMPs obtained from highly metastatic cells or cells exposed to chemotherapy when compared to cells with low metastatic potential. Consequently, when we inhibited CD44 expression on TMPs by a pharmacological or a genetic approach, increased tumor cell adhesion and re-organized actin filament structure were observed. We also demonstrated that breast cancer patients treated with paclitaxel chemotherapy exhibited increased CD44-expressing TMPs. Overall, our study provides further insights into the role of TMPs in promoting metastasis, an effect which is augmented when tumor cells are exposed to chemotherapy. Cells 2020, 9, 2269 2 of 20 still incurable, with metastasis being the main cause of death. In breast cancer, the most frequently diagnosed cancer in women, approximately half a million cancer deaths due to metastasis are reported per year [1]. Although a minority of breast cancer patients are diagnosed with stage IV advanced metastatic incurable disease, approximately 30% of all breast cancer will develop metastasis within months and years after diagnosis[2]. Metastasis is a multi-step process that includes cancer cell dissemination from the primary tumor, intravasation to the blood or lymphatic system, survival in the circulation, extravasation to a target organ, and seeding and proliferation at a distant site[3,4]. These effects require a tight regulation of the cellular machinery that supports tumor cell detachment from the primary tumor and their binding to the metastatic site.Microparticles have recently emerged as having a potentially significant role in tumor progression and metastasis. Microparticles belong to heterogeneous double layered membrane-coated particles exerted from cells, called extracellular vesicles (EVs)[5]. EVs encompass lipids, proteins, mRNA, non-coding RNA, and DNA. EVs have been mostly studied in the context of intercellular communication, whereby they transfer cargo between cells, including signaling proteins and RNA, as well as stimulate cells by membrane binding[5,6]. There are three main family members of EVs: apoptotic bodies (1000-4000 nm in diameter), microparticles (100-1000 nm in diameter), and exosomes (20-100 nm in diameter)[5,7]. EVs have been shown to affect cancer progression in various ways. For example, at remote secondary sites, the EVs are taken up by organ-derived cells, thereby preparing the microenvironment for future tumor cell seeding. These effects are associated with integrins expressed by EVs, which direct them to specific organs[8]. In melanoma, EVs derived from tumors contribute to the mobilization of myeloid cells to the pre-metastatic site and support metastasis[9]. Such pre-metastatic niche formation has been recently reported to be enhanced in response to chemotherapy[10]. At the primary tumor site, EVs promote tumor progression by transferring oncogenic proteins such as EGFRvIII between glioblastoma cells[11]. Thus, EVs play a significant role in tumor growth and metastasis.When focusing on microparticles, it has been shown that tumor-derived microparticles (TMPs) obtained from multidrug-resistant breast cancer cells support immune evasion by polarizing macrophages into an inactive state. This process contributes to the colonization of breast cancer cells at distant sites[12]. In another study, TMPs were shown to selectively transfer p-glycoprotein to breast cancer cells, supporting their multidrug resistance capacity and therefore contributing to tumor growth[13]. We have recently demonstrated that breast cancer cells exposed to chemotherapy shed an increased number of TMPs expressing osteopontin. These TMPs support the mobilization and tumor homing of pro-angiogenic bone marrow-derived cells (BMDCs), ultimately enhancing tumor growth and metastasis[14]. However, little is known about the effect of TMPs on the biomechanical pro-metastatic forces of cells within the primary tumor site, which support their dissemination, and about the contribution of chemotherapy to this process.Metastasis is regulated by an abundant number of proteins, which participate in each step of the metastatic process. Cell adhesion is one of the key factors contributing to the early stage of metastasis, when tumor cells disseminate from the primary tumor site. To do this, tumor cells reduce their adhesion ability and increase their biomechanical forces in order to allow their dissemination from the surrounding microenvironment. CD44 is a transmembrane glycoprotein which is expressed on the cell surface of many cancer cells[15]. CD44 confers several variant isoforms, some of which are found to be upregulated in tumor progression[16]. It has been demonstrated that the high expression of CD44 on cancer cells is associated with increased metastasis[17]. The interaction of CD44 with ECM ligands such as hyaluronic acid, osteopontin, collagens, and matrix metalloproteinases contributes to cell migration and invasion and supports metastasis[18]. Furthermore, the transfer of CD44/CD44v6 expressed by EVs supports tumor initiation and progression[19]. Similarly, CD44-expressing EVs also contribute to metastasis by remodeling the ECM within the primary tumor, supporting tumor cell invasion[20]. Thus, CD44 is an important regulator for the metastatic process.Here we show that TMPs from highly metastatic tumor cells or cells exposed to chemotherapy enhance the metastatic characteristics of low-metastatic recipient cells. These effects include decreased Cells 2020, 9, 2269 3 of 20 focal adhesions, actin cytoskeleton rearrangement, and an increased number of catch/release cycles in the process of biomechanical force production, all driven by TMPs expressing CD44. Thus, this study describes a biomechanical mechanism that drives the metastatic switch in chemotherapy-treated tumors.Materials and MethodsCell CultureMDA-MB-231 human breast carcinoma cells, 4T1 murine breast carcinoma cells, and HEK293T human embryonic kidney cells were purchased from the American Type Culture Collection (ATCC). The LM2-4 cell line, a lung metastatic variant of the MDA-MB-231 cell line, was kindly provided by Prof. Robert Kerbel (Sunnybrook Health Sciences Centre, Toronto, Ontario, Canada). 67NR, a low metastatic variant of the 4T1 cell line, was kindly provided by Prof.
# Introduction
Although significant progress has been made in the last decades towards the development of novel anti-cancer therapies for the treatment of advanced metastatic disease, most cancer types are
## Extraction and quantification of tmp
The extraction and quantification of TMPs were carried out as previously demonstrated and validated. Briefly, cells were grown to 80% confluency, at which point medium was replaced with serum free (SF) medium or SF medium supplemented with 200 nM paclitaxel (PTX), as previously described. After 48 h, conditioned medium (CM) was collected and centrifuged at 4000× g for 20 min at room temperature (RT) to remove floating cells and apoptotic bodies. Supernatants were collected and centrifuged at 20,000 g for 1 h at 4 - C. The TMP-containing pellet was resuspended in phosphate buffered saline (PBS) and stored at −80 - C until further use. It should be noted that the dosage of 200 nM PTX for a duration of 48 h in serum-free conditions did not result in cell death, excluding the possibility that samples were contaminated with apoptotic bodies, as previously described.
TMP quantification was performed using flow cytometry by calculating the ratio between 7.35-µm counting beads (Calbiochem, Burlington, MA) and the number of events collected in the TMP gate (approximately 0.6-0.9 µm), as previously described. Additional information can be found in the Supplemental Online Materials.
Quantification and measurement of TMPs by Nanosight NS300 (NanoSight LTD., Malvern, UK) was performed as previously described. Additional information can be found in the Supplemental Online Materials.
## Modified boyden chamber assay
The invasion properties of MDA-MB-231 or 67NR cells pre-exposed to different TMP conditions were evaluated in Matrigel-coated Boyden chambers as previously described. Additional information can be found in the Supplemental Online Materials.
## Cell spreading assay
TMPs were pre-exposed to 1 µg/mL IgG or anti-CD44 (BioXcell, West Lebanon, NH) for 1 h, followed by extensive washing, in which a volume of 100 times PBS was used. The TMPs were then added to MDA-MB-231 GFP+ cultures and incubated for 24 h. Subsequently, the cells were trypsinized and re-plated on fibronectin-coated glass plates. Time-lapse movies were generated from images acquired every 5 min for a total of 4 h using an ImageXpress Micro Confocal system (Molecular Devices, San Jose, CA). The percentage of cells spreading over fibronectin was analyzed using FIJI, as previously described.
## Cell viability alamarblue tm assay
The metabolic indicator dye AlamarBlue TM (Serotec Ltd., Oxford, UK) was used to determine cell viability, as previously described. Additional information can be found in the Supplemental Online Materials.
## Pillar fabrication
Pillar fabrication was performed as previously described. Briefly, PDMS (Sylgard 184, Dow Corning; 10:1 base to curing agent ratio) was poured over silicon molds with wells that were 1.3 µm deep, 0.5 µm wide, and spaced 1 µm apart (center-to-center distance). The molds were then flipped over onto glass-bottom dishes (no. 0 glass coverslip, Cellvis), which were then placed at 60 - C for 12 h to cure the PDMS. The molds were peeled off from the plates while immersed in pure ethanol, which was then replaced by PBS. The pillars were coated with fibronectin (10 µg/mL, 1 hr, 37 - C). To measure cellular forces on the fibronectin-coated pillars, serum-starved MDA-MB-231 cells were cultured for 24 h in the presence of 100,000 TMPs obtained from untreated or PTX-treated MDA-MB-231 cells, or from untreated or PTX-treated LM2-4 cells. The cells were then spread on the pillar arrays and imaged at 37 - C using a Leica DMIRE2 microscope (Leica Microsystems, Wetzlar, Germany), 100× 1.4 NA oil objective, and a CCD camera (QImaging Retiga EXi). Analyses of pillar movements was performed with ImageJ (V.1.51, National Institutes of Health, Bethesda, MD, USA) using the Nano Tracking plugin as previously described. Pillar displacement curves were generated using Matlab (V.9.4, MathWorks, Natick, MA, USA). Analyses of pillar releases by the cells was performed on the pillar displacement curves using the 'findpeaks' function in Matlab. All peaks identified were categorized into bins ranging from 1 to 12 nm, and the number of releases per second was calculated for each of these bins. For each condition >30 pillars from >3 cells were analyzed.
## Cd44 short hairpin rna
Several different short hairpin RNA (shRNA) sequences specific to human CD44 or a scrambled control sequence were cloned into the GIPZ Lentiviral Human shRNA plasmid (Horizon Discovery, Lafayette, CO, USA), as previously described. Subconfluent LM2-4 cells were transfected with shRNA plasmids using lipofectamine reagent (Thermo Fisher Scientific, MA, USA) according to the manufacturer's instructions. Forty-eight hours post-transfection, cells were incubated in growth medium containing puromycin (1 µL/mL) for the selection of stable transfectants. After 3 weeks of selection, the protein level of CD44 was evaluated by flow cytometry using appropriate controls. Two independent CD44-depleted clones generated from the pooled cells were selected for this study.
## Western blot
Paxillin and phospho-paxillin analysis was performed as follows. MDA-MB-231 cells were pre-exposed to 100,000 TMPs for 24 h, and then seeded on a fibronectin (20 µg/mL) coated plate overnight. Subsequently, the cells were lysed in Hepes 50 mM PH-7.5, EDTA 4 mM, Triton 1%, 0.5 mg/mL Na 3 VO 4 , and 4.5 mg/mL Na 2 P 2 O 7 and lysates were subjected to SDS-PAGE. Proteins were electro-transferred to nitrocellulose membranes, which were then probed with polyclonal rabbit anti-paxillin (1:1000; catalog number 2542, Cell Signaling Technologies, Danvers, MA, USA), anti-phospho-paxillin (1:1000; catalog number 2541, Cell Signaling Technologies), or anti-GAPDH (1:5000, catalog number sc-25778, Santa Cruz Biotechnology Inc., Dallas, TX, USA). GAPDH was used for loading controls. All experiments were performed in triplicate.
## Tumor-derived microparticle proteomics
TMP proteomics were assessed and analyzed in accordance with the previously described method. Briefly, TMPs were isolated by multiple centrifugations of conditioned media from MDA-MB-231 and LM2-4 cells cultured in serum-free media in the presence or absence of PTX, as described above. Pellets were immediately resuspended in 200 µL of lysis buffer (6 M urea, 2 M thiourea in 0.1 M Tris pH 7.6). Protein concentration was calculated based on the Bradford assay, and 10 µg of each sample was used for the analysis. The proteins were reduced using 1 mM DTT and alkylated using 5 mM iodoacetamide, followed by overnight in-solution trypsin/LysC digestion. The peptides were acidified with 0.1% TFA and purified on C18 stageTips. A total of three biological replicates were prepared for each system, some of which were analyzed with two technical replicates, and were later combined into a single data point.
Resulting peptides were analyzed by LC-MS/MS on the EASY-nLC1000 UHPLC system (Thermo Fisher Scientific) coupled to the Q-Exactive Plus mass spectrometer (Thermo Fisher Scientific). The peptides were separated using a gradient of 140 min (technical replicate I) or 240 min (technical replicate II) of water-acetonitrile on a 50 cm EASY-spray column. Data were acquired using a top-10 method in a data-dependent mode. Resolution values were 70,000 and 17,500 in the MS and MS/MS scans, respectively.
MS raw files were processed using MaxQuant version 1.5.6.9and the Andromeda search engine. MS/MS spectra were searched against the human uniprot database on a forward and decoy database with 1% false discovery rate for both the protein and the peptide levels. The 'match between runs' option was enabled to transfer identifications between runs. The label-free algorithm was used for protein quantification. Technical replicates were averaged into a single biological replicate during the MaxQuant data processing. The proteinGroups output was further analyzed using Perseus software. Label-free quantification (LFQ) intensity valueswere log2 transformed. Each comparison (MDA-MB-231 vs. MDA-MB-231 PTX or MDA-MB-231 vs. LM2-4) was processed separately following data filtration that kept only those proteins that were quantified in at least two of the three replicates in at least one of the sample groups. Student's t-test was performed with permutation-based false discovery rate (FDR) correction with a cutoff of 0.1 and S0 correction of 0.2. Enrichment analysis was done using the Fisher exact test with an FDR cutoff of 0.02. Principal component analysis (PCA) was performed following data imputation, by replacing the missing values with values that formed a normal distribution with a downshift of 1.6 standard deviations and a width of 0.4 of the original data distribution.
## Ex vivo pulmonary metastasis assay
The ex vivo pulmonary metastasis assay (PuMA) was performed as previously described. Additional information can be found in the Supplemental Online Materials.
## Blood samples from cancer patients
The study was approved by the hospital's Ethics Committee (Ha'Emek medical center), and written informed consent was obtained from all involved patients, in compliance with the Helsinki declaration for human studies. Patients with localized breast carcinoma were treated by neoadjuvant chemotherapy (before tumor removal), consisting of adriamycin and cyclophosphamide followed by PTX chemotherapy, at the Ha'Emek Medical Center, Afula, Israel. Blood was collected in 5-mL sodium citrate (3.2%) tubes before the first cycle of PTX chemotherapy and 24 h post-treatment (n = 15). Plasma was obtained by centrifugation at 1000 g for 10 min at RT. Subsequently, platelet-poor plasma (PPP) was extracted by plasma centrifugation at 3000 g for 5 min at RT, as previously described. This procedure eliminated most of the platelets (<10,000 per µL). To ensure purified PPP, another centrifugation (3000 g for 5 min) was performed on separated PPP at RT. Subsequently PPP was stored at −80 - C. Thawed PPP was centrifuged at 20,000 g for 1 h, and the pellet was resuspended in PBS. Anti-MUC-1 Cells 2020, 9, 2269 6 of 20 antibody and anti-CD44 antibody were added according to the manufacturer's instructions and were subsequently analyzed by flow cytometry as previously described. A fluorescence-minus-one control (FMO) was used to determine nonspecific binding of antibodies, as previously described.
## Immunostaining and imaging
Serum-starved cells were cultured for 24 h in the presence of 100,000 TMPs obtained from untreated or PTX-treated cells, as indicated in the text. In some experiments, TMPs were cultured with cells for 24 h in serum-free medium in the presence of anti-CD44 antibodies (1 µg/mL). Cells were rigorously washed in PBS (100-fold volume), and then seeded on fibronectin-coated plates (20 µg/mL fibronectin, Biological Industries). After 4 h, cells were fixed using 4% paraformaldehyde (PFA) and immunostained with a primary anti-vinculin antibody (1:100, Sigma-Aldrich) and Cy2-conjugated secondary antibody. Actin was stained with Alexa 488 conjugated phalloidin (1:100, Invitrogen, Carlsbad, CA, USA). Images were acquired with a LSM 700 Zeiss confocal microscope (Zeiss Ltd. Oberkochen, Germany).
# Statistical analysis
Data are presented as mean ± standard error (SE). All in vitro studies were performed at least in three biological replicates. The in vivo studies were performed twice. Statistically significant differences were assessed by one-way ANOVA, followed by the Tukey post-hoc test (when comparing between more than two groups) using GraphPad Prism (V.6, La Jolla, CA, USA). When applicable, an estimate of variance was performed and statistical significance comparing only two sets of data was determined by the two-tailed Student's t-test. Significance was set at values of p < 0.05, and designated as follows: * p < 0.05; ** p < 0.01; *** p < 0.001.
# Results
## Tmps from highly metastatic or chemotherapy-treated tumor cells induce tumor cell invasion
To study the effect of TMPs on the metastatic potential and the adhesion properties of tumor cells, we compared TMPs from MDA-MB-231 cells with their highly metastatic variant, LM2-4 cells, or in the presence or absence of PTX chemotherapy. TMPs were extracted from the medium and quantified by flow cytometry and NanoSight. Although the number of TMPs extracted from highly metastatic cells or from cells exposed to chemotherapy was higher than control cells, as previously reported, their size was not significantly different as evaluated by NanoSight. Based on their size, we ruled out the possibility that some of the TMPs detected are apoptotic bodies, as also previously reported.
We next postulated that TMPs from highly metastatic cells or from cells exposed to chemotherapy could augment the metastatic potential of low-metastatic cells. To this end, TMPs were extracted from MDA-MB-231 and LM2-4 cells which had been treated with PTX or the vehicle control. Subsequently, these TMPs were cultured with low-metastatic MDA-MB-231 cells for 24 h and further analyzed for invasion properties using the Boyden chamber assay. TMPs from PTX-treated MDA-MB-231 cells enhanced the invasive properties of MDA-MB-231 cells in comparison to control cultures, including TMPs from non-treated MDA-MB-231 cells or serum-free conditions. Furthermore, MDA-MB-231 cells cultured with TMPs from control or PTX-treated LM2-4 cells exhibited the highest invasive ability. Notably, TMPs from PTX-treated cells displayed induced invasion properties regardless of the metastatic potential of the tumor cells, implicating a potent effect of chemotherapy on TMP-induced cell invasion. Similar results were observed when using 4T1 highly metastatic cells and their low metastatic 67NR cells. Of note, TMPs had no effect on cell proliferation. Taken together, these results suggest that TMPs affect tumor cell invasion properties.
## Figure 1. tumor-derived microparticles (tmps) from highly metastatic cells or cells exposed to chemotherapy increase cell invasion. (a,b) mda-mb-231 cells cultured with serum-free medium or
TMPs from MDA-MB-231 or LM2-4 cells exposed to paclitaxel (PTX) or vehicle control were assessed for invasion properties using the Boyden chamber assay. Representative images of invading cells are shown in (A). Scale bar, 200 µm. Quantifications of invading cells are shown in (B) (n = 5 repeats and 7 images/repeat). *, differences compared to MDA-MB-231 control. ***, p < 0.001, as assessed by oneway ANOVA followed by Tukey post-hoc test.
## Tmps from highly metastatic or chemotherapy-treated tumor cells inhibit tumor cell seeding
We next sought to test if TMPs affect the seeding of low metastatic cells in distant tissues. We therefore performed an ex vivo pulmonary metastasis assay (PuMA), which primarily evaluates the seeding potential of tumor cells in the lungs. GFP-expressing MDA-MB-231 cells pre-cultured for 24 h with TMPs from vehicle-or PTX-treated MDA-MB-231 or LM2-4 cells were injected through the tail vein of naïve mice to perform the PuMA. The PuMA can be used a method to test the ability of cells to adhere to a tissue, therefore it is suitable to test the early stage of metastasis, when tumor cells detach from the primary tumor microenvironment. We found that tumor cell seeding represented by the number of GFP+ foci in the lungs was reduced when the MDA-MB-231 cells were pre-cultured with TMPs from PTX-treated MDA-MB-231 cells or with TMPs from vehicle-or PTXtreated LM2-4 cells. These results were also supported by flow cytometry analysis. Specifically, a reduced percentage of GFP+ cells in the whole lung tissue was observed in mice injected with MDA-MB-231 cells pre-cultured with TMPs from PTX-treated MDA-MB-231 cells or with TMPs from vehicle-or PTX-treated LM2-4 cells compared to control MDA-MB-231 cells. Importantly, the short time-interval between cell injection and animal sacrifice suggests that the
## Tmps from highly metastatic or chemotherapy-treated tumor cells inhibit tumor cell seeding
We next sought to test if TMPs affect the seeding of low metastatic cells in distant tissues. We therefore performed an ex vivo pulmonary metastasis assay (PuMA), which primarily evaluates the seeding potential of tumor cells in the lungs. GFP-expressing MDA-MB-231 cells pre-cultured for 24 h with TMPs from vehicle-or PTX-treated MDA-MB-231 or LM2-4 cells were injected through the tail vein of naïve mice to perform the PuMA. The PuMA can be used a method to test the ability of cells to adhere to a tissue, therefore it is suitable to test the early stage of metastasis, when tumor cells detach from the primary tumor microenvironment. We found that tumor cell seeding represented by the number of GFP+ foci in the lungs was reduced when the MDA-MB-231 cells were pre-cultured with TMPs from PTX-treated MDA-MB-231 cells or with TMPs from vehicle-or PTX-treated LM2-4 cells. These results were also supported by flow cytometry analysis. Specifically, a reduced percentage of GFP+ cells in the whole lung tissue was observed in mice injected with MDA-MB-231 cells pre-cultured with TMPs from PTX-treated MDA-MB-231 cells or with TMPs from vehicle-or PTX-treated LM2-4 cells compared to control MDA-MB-231 cells. Importantly, the short time-interval between cell injection and animal sacrifice suggests that the TMPs lowered the cells' adhesive ability. These results implicate an early process of metastasis at the primary tumor site where Cells 2020, 8 of 20 the cells detached from their surrounding microenvironment. To support these results, we found that cells exposed to TMPs from PTX-treated MDA-MB-231 cells or to TMPs from vehicle-or PTX-treated LM2-4 cells spread to a much lower extent when seeded on fibronectin-coated plates, compared to control conditions. This property directly relates to the inability to form strong cell-matrix adhesions. Once again, TMPs from PTX-treated cells displayed similar reduced tumor cell adhesion properties, regardless of the metastatic potential of the tumor cells, therefore suggesting a strong effect of chemotherapy on TMPs which hinder cell adhesion.
Cells 2020, TMPs lowered the cells' adhesive ability. These results implicate an early process of metastasis at the primary tumor site where the cells detached from their surrounding microenvironment. To support these results, we found that cells exposed to TMPs from PTX-treated MDA-MB-231 cells or to TMPs from vehicle-or PTX-treated LM2-4 cells spread to a much lower extent when seeded on fibronectincoated plates, compared to control conditions. This property directly relates to the inability to form strong cell-matrix adhesions. Once again, TMPs from PTX-treated cells displayed similar reduced tumor cell adhesion properties, regardless of the metastatic potential of the tumor cells, therefore suggesting a strong effect of chemotherapy on TMPs which hinder cell adhesion.
## Tmps from highly metastatic or chemotherapy-treated tumor cells inhibit cellular adhesion and destroy actin filaments
Metastatic cells acquire the ability to detach from the primary tumor by inhibiting adhesion molecules and rearranging the actin cytoskeleton to promote the migratory properties. To study the effect of TMPs on the cell adhesion, MDA-MB-231 cells were cultured for 24 h in the presence. In addition, cell area was reduced in the presence of TMPs from all groups in comparison to control cells cultured in the absence of TMPs. Lastly, cells cultured with TMPs from PTX-treated MDA-MB-231 cells or with TMPs from vehicle-or PTX-treated LM2-4 cells demonstrated less organized and elongated actin stress fibers. These fibers were displayed with shorter actin filaments positioned in multiple directions, compared to those observed in control cells. Thus, TMPs support metastatic cell characteristics by means of downregulating the formation of mature adhesion plaques and less organized actin cytoskeleton. Cells 2020, 9, x 10 of 20
## Tmps increase the pace of metastatic cell biomechanical forces
Since there is reciprocal regulation between adhesions and their associated force-producing actin cytoskeleton, we sought to determine which of these structures was affected by TMPs, thereby decreasing adhesion and promoting cell motility. To test this, we evaluated pillar displacement when using MDA-MB-231 cells in different conditions, as demonstrated in. Specifically, when MDA-MB-231 cells were cultured with TMPs from LM2-4 or from PTX-exposed cells, they pulled on the pillars at a similar pace compared to control cells. However, they could not maintain their hold, leading to repeated pull/release cycles that occurred at a higher rate in comparison to control cells, thus demonstrating a significant increase in the average number of pillar releases per second. Since defects in cytoskeletal assembly lead to changes in the rate of force application, whereas defects in adhesions lead to their breakage, we concluded that exposure to TMPs from LM2-4 or from PTX-exposed cells directly decreases cell adhesion properties, rendering the adhesions too weak to withstand cytoskeletal forces and prevent the formation of stress fibers. Furthermore, Western blot analysis of cell lysates revealed significantly lower adhesion-associated paxillin signaling in cells cultured with TMPs from PTX-treated MDA-MB-231 cells or with TMPs from vehicle-or PTX-treated LM2-4 cells. These results suggest that adhesion is reduced in cells exposed to TMPs from PTX-treated cells or highly metastatic cells. Overall, our results demonstrate that TMPs from PTX-treated low metastatic cells or from highly metastatic cells (regardless of chemotherapy treatment) disrupt cell adhesion, which is consistent with cell dissemination and an increased ability to invade. Each of the control conditions was significantly different than the conditions in which the cells were exposed to TMPs from LM2-4 cells or from cells exposed to PTX (α < 0.01; Kolmogorov-Smirnov test). n = 8-10 repeats/group. (D-E) Cells treated as above were seeded on fibronectin-coated plates to evaluate phospho-paxillin (P-Pax) and total-paxillin (Pax) expression by Western blot. GAPDH was used as a loading control (D). The ratio between P-Pax over Pax was calculated based on densitometry, and presented as fold change, for three biological repeats (E). n = 3 repeats/group. * differences compared to MDA-MB-231 control group, * p < 0.05, as assessed by oneway ANOVA followed by Tukey post-hoc test.
## Proteomic analysis of tmps
We next undertook a proteomic approach to identify factors within TMPs that affect cell adhesion, when the major focus was placed on adhesion-related proteins. To this end, TMPs were (C) Histograms presenting the average number of pillar releases per second binned according to the size of the release. Each of the control conditions was significantly different than the conditions in which the cells were exposed to TMPs from LM2-4 cells or from cells exposed to PTX (α < 0.01; Kolmogorov-Smirnov test). n = 8-10 repeats/group. (D-E) Cells treated as above were seeded on fibronectin-coated plates to evaluate phospho-paxillin (P-Pax) and total-paxillin (Pax) expression by Western blot. GAPDH was used as a loading control (D). The ratio between P-Pax over Pax was calculated based on densitometry, and presented as fold change, for three biological repeats (E). n = 3 repeats/group. * differences compared to MDA-MB-231 control group, * p < 0.05, as assessed by one-way ANOVA followed by Tukey post-hoc test.
## Proteomic analysis of tmps
We next undertook a proteomic approach to identify factors within TMPs that affect cell adhesion, when the major focus was placed on adhesion-related proteins. To this end, TMPs were isolated from MDA-MB-231 and LM2-4 cells (untreated or treated with PTX) and processed for high-resolution mass-spectrometry-based proteomic analysis. Overall, 5209 proteins were identified, with control MDA-MB-231 TMPs exhibiting the lowest number of proteins . PCA further reflected the large variation between control MDA-MB-231 TMPs and all other TMP groups. We then performed two comparisons. In the first, we compared the proteomes of TMPs derived from control and PTX-treated MDA-MB-231 cells. In the second, we compared the proteomes of TMPs derived from control MDA-MB-231 cells and untreated LM2-4 cells. Over 300 significantly altered proteins were detected in each comparison. When specifically focusing on adhesion membrane proteins of which the levels were significantly changed, 15 proteins were common to both comparisons. Fourteen of these proteins were higher in MDA-MB-231 control TMPs. Interestingly, out of the 15 cell adhesion proteins common to both comparisons, only one, CD44, was significantly increased in TMPs from PTX-treated MDA-MB-231 or untreated LM2-4 cells. Its intensity in these TMPs was over two-fold higher than in TMPs derived from control MDA-MB-231 cells. Altogether, these results suggest that CD44, probably among other proteins that are significantly differentially expressed between the various TMP groups, is a possible co-effector that can potentially account for the disruption of cell adhesion induced by TMPs. . PCA further reflected the large variation between control MDA-MB-231 TMPs and all other TMP groups. We then performed two comparisons. In the first, we compared the proteomes of TMPs derived from control and PTX-treated MDA-MB-231 cells. In the second, we compared the proteomes of TMPs derived from control MDA-MB-231 cells and untreated LM2-4 cells. Over 300 significantly altered proteins were detected in each comparison. When specifically focusing on adhesion membrane proteins of which the levels were significantly changed, 15 proteins were common to both comparisons. Fourteen of these proteins were higher in MDA-MB-231 control TMPs. Interestingly, out of the 15 cell adhesion proteins common to both comparisons, only one, CD44, was significantly increased in TMPs from PTX-treated MDA-MB-231 or untreated LM2-4 cells. Its intensity in these TMPs was over two-fold higher than in TMPs derived from control MDA-MB-231 cells. Altogether, these results suggest that CD44, probably among other proteins that are significantly differentially expressed between the various TMP groups, is a possible co-effector that can potentially account for the disruption of cell adhesion induced by TMPs.
## Cd44-expressing tmps decrease cell adhesion and support tumor cell metastatic properties
CD44 is a glycoprotein known to participate in cell adhesion and migration. Therefore, increased levels of CD44 in TMPs may explain the TMP-induced increased adhesive and invasive properties of tumor cells. To test this, TMPs derived from untreated or PTX-treated MDA-MB-231 or LM2-4 cells were incubated with CD44 blocking antibodies or IgG control antibodies for 24 h. Excess antibodies were removed by rigorous washing when adding a 100-fold volume of PBS to the sample, followed by centrifugation to obtain TMPs. Subsequently, the TMPs were added to MDA-MB-231 cultures, and invasive and adhesive properties were assessed. Pre-treating TMPs with anti-CD44 antibodies restored the invasive properties of TMP-exposed tumor cells to control levels. Similar results were obtained when TMPs from 4T1 cells were pre-treated with anti-CD44 antibodies and subsequently cultured with 67NR (low metastatic) cells.
Next, TMPs pre-treated with anti-CD44 were added to MDA-MB-231 cultures, and the number of focal adhesions, assessed by vinculin, cell area, efficiency of cell spreading, and the phenotype of actin stress fibers, was evaluated. Blocking CD44 on TMPs from highly metastatic cells or from cells exposed to chemotherapy restored the conditions found in control MDA-MB-231 cells. Specifically, CD44 inhibition on TMPs increased focal adhesion plaques assessed by vinculin, cell area, and percentage of cell spreading. Notably, the cell size reduction noticed in MDA-MB-231 cells exposed to TMPs from PTX-treated or LM2-4 cells is in line with cell spreading, demonstrated in. Furthermore, TMPs in which CD44 was inhibited resulted in rearrangement of actin stress fibers, mimicking those found in control serum-free conditions. Similar results were obtained when TMPs from highly metastatic cells were knocked down for CD44 and subsequently were added to MDA-MB-231 cultures. Specifically, clones 1 and 2 demonstrated an increased number of focal adhesion plaques and better organization of actin filaments when compared to scrambled control cells. Taken together, these results further indicate that CD44-expressing TMPs contribute to the inhibition of cell adhesion and cytoskeleton re-arrangement, and thus enhance the metastatic properties of tumor cells. Cells 2020, 9, x 15 of 20
## Tmps expressing cd44 were increased in breast cancer patients treated with chemotherapy
To further investigate the possibility that TMPs expressing CD44 could be found in breast cancer patients following chemotherapy, we collected plasma samples from 15 breast cancer patients at baseline or 24 h after PTX chemotherapy administration. We used MUC-1 as an epithelial breast carcinogenic marker, known also to be expressed on TMPs. The number of MUC-1+/CD44+ TMPs detected in plasma samples from 12 out of 15 breast cancer patients post-chemotherapy was 3-4-fold higher than those at baseline. Taken together, the data suggest that chemotherapy induces CD44 expression on TMPs and, as a result, may contribute to tumor cell aggressiveness.
## Tmps expressing cd44 were increased in breast cancer patients treated with chemotherapy
To further investigate the possibility that TMPs expressing CD44 could be found in breast cancer patients following chemotherapy, we collected plasma samples from 15 breast cancer patients at baseline or 24 h after PTX chemotherapy administration. We used MUC-1 as an epithelial breast carcinogenic marker, known also to be expressed on TMPs. The number of MUC-1+/CD44+ TMPs detected in plasma samples from 12 out of 15 breast cancer patients post-chemotherapy was 3-4-fold higher than those at baseline. Taken together, the data suggest that chemotherapy induces CD44 expression on TMPs and, as a result, may contribute to tumor cell aggressiveness.
# Discussion
Metastasis is the main cause of death in cancer patients and is still a major obstacle for the success of therapy. Treatment protocols involve chemotherapy, radiation, surgery, or their combination. These protocols have shown clinical benefits, yet resistance is common and recurrence following chemotherapy is usually more aggressive. This study provides a suitable explanation for tumor cell aggressiveness in response to chemotherapy. Specifically, we demonstrate that TMPs originating from highly metastatic or chemotherapy-exposed tumor cells mediate a paracrine effect that decreases cell adhesion and rearranges the actin cytoskeleton. These effects promote early-stage metastatic properties in recipient low metastatic tumor cells. To strengthen the fact that cell adhesion is decreased, we used the ex vivo pulmonary metastasis assay, which specifically tests the seeding properties of tumor cells, namely, their ability to adhere to the lung tissue. Although one would expect an increased number of metastatic foci in the lung sections of the PuMA, the fact that we observed a reduced number of metastasisstrengthens our assumption that the
# Discussion
Metastasis is the main cause of death in cancer patients and is still a major obstacle for the success of therapy. Treatment protocols involve chemotherapy, radiation, surgery, or their combination. These protocols have shown clinical benefits, yet resistance is common and recurrence following chemotherapy is usually more aggressive. This study provides a suitable explanation for tumor cell aggressiveness in response to chemotherapy. Specifically, we demonstrate that TMPs originating from highly metastatic or chemotherapy-exposed tumor cells mediate a paracrine effect that decreases cell adhesion and rearranges the actin cytoskeleton. These effects promote early-stage metastatic properties in recipient low metastatic tumor cells. To strengthen the fact that cell adhesion is decreased, we used the ex vivo pulmonary metastasis assay, which specifically tests the seeding properties of tumor cells, namely, their ability to adhere to the lung tissue. Although one would expect an increased number of metastatic foci in the lung sections of the PuMA, the fact that we observed a reduced number of metastasisstrengthens our assumption that the major effect of TMPs is on reduced cell adhesion properties. Reduced cell adhesion is the early process required for the dissemination of tumor cells from the primary tumor site. These results further demonstrate that TMPs support different stages of the metastatic cascade. Furthermore, we should note that our study focused on the in vitro changes in tumor cell characteristics in the presence of TMPs, yet there are several studies which reported that reduced adhesion of tumor cells supports early stage of metastatic properties. Additional in vivo studies can strengthen these conclusions by testing whether TMPs originating from highly metastatic clones or following chemotherapy within the primary tumor microenvironment, are directly associated with increased metastasis in distant sites. Taken together, our results provide further support for the role of TMPs in promoting metastasis.
To further identify the possible mechanism for TMP-induced tumor cell metastatic properties, we have undertaken an unbiased proteomic approach to analyze the expression of various proteins associated with adhesion. We demonstrate that the pro-metastatic effects of TMPs are associated with the expression of CD44, a non-oncogenic protein. EVs expressing CD44 have been shown to promote tumor cell aggressiveness and metastasis mainly by contributing to the pre-metastatic niche. Here we found that TMPs expressing CD44 contribute to metastatic tumor cell characteristics in part by disrupting the actin cytoskeleton filament structure, inhibiting focal adhesion plaques and cell spreading, reducing contractility, and downregulating adhesion-related signaling measured by phospho-paxillin. While these effects clearly contribute to the understanding of the extracellular function of CD44, further investigation is needed regarding the intracellular domain of CD44. Specifically, it has been shown that the intracellular domain of CD44 recruits actin binding complexes such as ezrin, radixin, moesin (ERM), and ankyrin. These complexes are associated with intracellular signaling via Ras-MAPK, Wnt, and PI3K, and can thus contribute to the metastatic characteristics of tumor cells by other mechanisms. Notably, although our studies have focused on specific adhesion molecules, based on the proteomic analysis it is plausible that other molecules expressed by TMPs may affect cell adhesion properties, including specific extracellular matrix-associated enzymes. Thus, additional studies in this direction will provide further insights into the overall effects of TMPs on tumor cell metastatic characteristics.
Elevated CD44 expression levels were also found in TMPs from breast cancer patients treated with PTX. When analyzing their plasma samples, we focused on MUC1+ EVs that are known to originate from tumor cells in various malignancies, suggesting that they are TMPs. We have previously demonstrated that the number of TMPs from breast cancer patients increased following chemotherapy. These results are supported by our current study, in which we demonstrate that tumor cells exposed to chemotherapy shed an increased number of TMPs. Clinically, we found that the percentage of TMPs expressing CD44 was increased in plasma samples from breast cancer patients following chemotherapy when compared to baseline levels. Taken together, our in vitro findings support the observations found in breast cancer patients.
In summary, we describe a mechanism whereby CD44-expressing TMPs affect tumor cell metastatic characteristics by inhibiting cell adhesion properties. This effect is augmented in response to chemotherapy, which can explain the potential increase in metastasis. Indeed, breast cancer patients treated with neoadjuvant chemotherapy may exhibit an increased tumor microenvironment of metastasis (TMEM). TMEMs promote the dissemination of tumor cells from the primary tumor site. It is plausible that combining chemotherapy with agents that inhibit CD44 on TMPs may reduce the potential for metastasis, and therefore can serve as a strategy to inhibit the possible risks of chemotherapy-induced metastasis in breast cancer.
Supplementary Materials: The following are available online at http://www.mdpi.com/2073-4409/9/10/2269/s1, Figures S1-S5 . |
A higher BMI is not associated with a different immune response and disease course in critically ill COVID-19 patients
Background/objectives Obesity appears to be an independent risk factor for ICU admission and a severe disease course in COVID-19 patients. An aberrant inflammatory response and impaired respiratory function have been suggested as underlying mechanisms. We investigated whether obesity is associated with differences in inflammatory, respiratory, and clinical outcome parameters in critically ill COVID-19 patients. Subjects/methods Sixty-seven COVID-19 ICU patients were divided into obese (BMI ≥ 30 kg/m 2 , n = 18, 72% class I obesity, 28% class II obesity) and non-obese (BMI < 30 kg/m 2 , n = 49) groups. Concentrations of circulating interleukin (IL)-6, IL-8, IL-10, tumor necrosis factor alpha (TNF-α), interferon gamma (IFN-γ), interferon gamma-induced protein (IP)-10, monocyte chemoattractant protein (MCP)-1, and IL-1 receptor antagonist (RA) were determined from ICU admission until 10 days afterward, and routine laboratory and clinical parameters were collected. Results BMI was 32.6 [31.2-34.5] and 26.0 [24.4-27.7] kg/m 2 in the obese and non-obese group, respectively. Apart from temperature, which was significantly lower in obese patients (38.1 [36.9-38.9] vs. 38.7 [38.0 −39.5]°C, p = 0.02), there were no between-group differences on ICU admission. Plasma cytokine concentrations declined over time (p < 0.05 for all), but no differences between obese and non-obese patients were observed. Also, BMI did not correlate with the cytokine response (IL-6 r = 0.09, p = 0.61, TNF-α r = 0.03, p = 0.99, IP-10 r = 0.28, p = 0.11). The kinetics of clinical inflammatory parameters and respiratory mechanics were also similar in both groups. Finally, no differences in time on ventilator, ICU length of stay or 40-day mortality between obese and non-obese patients were apparent. Conclusions In COVID-19 patients requiring mechanical ventilation in the ICU, a higher BMI is not related to a different immunological response, unfavorable respiratory mechanics, or impaired outcome.
# Background
The Coronavirus Disease 2019 (COVID-19) pandemic currently sweeps across the globe, leading to high morbidity and mortality. This infection, caused by Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2), can cause acute respiratory distress syndrome (ARDS) requiring mechanical ventilation in the intensive care unit (ICU). Age, male sex, and multiple comorbidities, including type 2 diabetes, hypertension, and coronary artery disease, were identified as risk factors related to severity of COVID-19 disease [bib_ref] Clinical course and risk factors for mortality of adult inpatients with COVID-19..., Zhou [/bib_ref] [bib_ref] Risk factors for severity and mortality in adult COVID-19 inpatients in Wuhan, Li [/bib_ref] [bib_ref] Clinical predictors of mortality due to COVID-19 based on an analysis of..., Ruan [/bib_ref] [bib_ref] Factors associated with hospital admission and critical illness among 5279 people with..., Petrilli [/bib_ref]. These comorbidities are more prevalent in the obese population than in patients with a normal body weight [bib_ref] Obesity and cardiovascular disease: pathophysiology, evaluation, and effect of weight loss: an..., Poirier [/bib_ref]. Also, the prevalence of obesity among COVID-19 patients requiring mechanical ventilation was shown to be higher than expected when compared to controls with Members of the RCI-COVID-19 study group are listed above Author contributions.
non COVID-19-related acute pulmonary diseases, and the need for mechanical ventilation increased with increasing body mass index (BMI) [bib_ref] High prevalence of obesity in severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) requiring..., Simonnet [/bib_ref] [bib_ref] Obesity and mortality of COVID-19. Meta-analysis, Hussain [/bib_ref]. Finally, mortality was shown to be significantly higher in overweight/obese COVID-19 patients compared to normal-weight patients [bib_ref] Obesity and mortality of COVID-19. Meta-analysis, Hussain [/bib_ref]. Of interest, obesity itself is also recognized as an independent risk factor of severe pulmonary H1N1 influenza infection and mortality [bib_ref] Populations at risk for severe or complicated influenza illness: systematic review and..., Mertz [/bib_ref].
The underlying mechanisms of obesity as a factor contributing to COVID-19 disease severity are currently poorly understood, but various possible explanations have been proposed. For instance, in obese patients, pulmonary function may be restricted because of decreased functional residual capacity, resulting in lower blood oxygen levels [bib_ref] The effects of obesity on lung volumes and oxygenation, Littleton [/bib_ref]. In addition, the angiotensin converting enzyme 2 receptor, which is utilized by SARS-CoV-2 to enter cells, is highly expressed by adipocytes and expression levels were shown to be higher in adipocytes of patients with obesity and type 2 diabetes [bib_ref] The role of adipocytes and adipocytelike cells in the severity of COVID-19..., Kruglikov [/bib_ref]. Therefore, it appears plausible that adipose tissue may serve as a reservoir for the virus, leading to more pronounced and sustained viral shedding, resulting in a perpetual inflammatory response and impaired outcome.
To date, it is not known whether obesity is only a risk factor for COVID-19 susceptibility and for the need of intensive care, or whether it also influences the outcome of the patients once admitted on ICUs. Previous studies in critically ill COVID-19 patients have yielded ambiguous results, with some reporting associations between lower BMI and mortality [bib_ref] ICU and Ventilator Mortality Among Critically Ill Adults With Coronavirus Disease, Auld [/bib_ref] , whereas others showed increased mortality in (severely) obese patients [bib_ref] Obesity, overweight and survival in critically ill patients with SARS-CoV-2 pneumonia: is..., Halasz [/bib_ref] [bib_ref] Factors associated with death in critically ill patients with coronavirus disease 2019..., Gupta [/bib_ref]. In the present study, we investigated associations between obesity and the immune response as well as the clinical course of SARS-CoV-2 infection in critically ill patients on the ICU of a tertiary university hospital. Elucidating these mechanisms may pave the way for the development of new treatment strategies for obese COVID-19 patients.
# Methods
## Study design and participants
In this prospective observational cohort study, all consecutive COVID-19 patients admitted to the ICU in the Radboud University Medical Center (Nijmegen, The Netherlands) between March 11 and April 27 were included. COVID-19 was diagnosed by a positive SARS-CoV-2 RT-PCR test in nasopharyngeal and throat swabs and/or by typical chest CT-scan findings. Patients with a pre-existing immunosuppressed status or other comorbidities that strongly influence prognosis were excluded. Patients were divided into an obese (BMI ≥ 30 kg/m 2 ) and non-obese (BMI < 30 kg/m 2 ) group, according to the classification of obesity by the World Health Organisation. Also, individual BMI values were correlated with circulating cytokine concentrations. The study was carried out in the Netherlands in accordance with the applicable rules concerning the review of research ethics committees and informed consent. All patients or legal representatives were informed about the study details and could decline to participate.
## Data collection
Data were collected from the electronic patient files (EPIC, EPIC Systems Corporation, Verona, Wisconsin, USA) and recorded in the good clinical practice-compliant data management system Castor (Castor EDC, Amsterdam, the Netherlands). ICU admission day was designated as day 1, and serial data were obtained for 10 consecutive days. Clinical outcomes (time on mechanical ventilation, ICU length of stay (LOS), and mortality) were recorded for 40 consecutive days. Serial values of mean arterial pressure, body temperature, and leukocyte differentiation as well as circulating levels of C-reactive protein (CRP), procalcitonin (PCT), D-dimer, ferritin, and cytokines (see below) were used to assess the inflammatory response to SARS-CoV-2 infection. Patients were actively cooled when they had a fever of >40°C. In these cases, body temperature was imputed as 40°C. Positive end expiratory pressure (PEEP), tidal volume per kilograms ideal body weight (TV IBW ) and PaO 2 /FiO 2 ratio were used to assess the level of respiratory support required. Tidal volume was only recorded in patients on volume-controlled mechanical ventilation. Because the vast majority of patients were switched to pressure support ventilation after 4 days of ICU admission, TV IBW data were analyzed until this timepoint. Ideal body weight was calculated based on sex and height, using the formulas provided by the ARDS Network [bib_ref] Ventilation with lower tidal volumes as compared with traditional tidal volumes for..., Brower [/bib_ref].
## Cytokine concentration measurements
A baseline blood sample was obtained within the first 48 h following ICU admission and serial samples were collected every other day. Ethylenediaminetetraacetic acidanticoagulated blood was centrifuged (2000 × g, 10 min, 4°C), after which plasma was stored at −80°C until analysis. Concentrations of interleukin (IL)-6, IL-8, IL-10, tumor necrosis factor (TNF)-α, interferon gamma (IFN-γ), IFN-γ-induced protein (IP)-10, monocyte chemoattractant protein (MCP)-1, and IL-1 receptor antagonist (IL-1RA) were determined in one batch using a Luminex assay (Milliplex, Millipore, Billerica, USA). The lower detection limit was 3.2 pg/mL for all cytokines.
# Statistical analysis
Statistical analysis was performed using SPSS 25 (IBM) and Graphpad Prism 8 software (GraphPad Software). Several variables were not measured daily; therefore, depending on the frequency of measurements, data were binned into 2 or 3 days using a custom script made in Rstudio v3.6.2 (www.r-project.org). Because of the relatively small group size, normality was not assumed and all data are displayed as median with interquartile range or geometric mean with 95% confidence interval. Baseline characteristics were analyzed using chi-square tests and Mann-Whitney U tests. Between-group differences over time were analyzed using linear mixed effects model analysis on log-transformed data followed by posthoc analyses using Sidak's multiple comparisons tests in all serially measured variables. Relationships between BMI and cytokines were analyzed using Spearman's correlation. Time on mechanical ventilation, ICU LOS, and mortality were analyzed using log-rank and chisquare tests. Patients who died in the hospital or those who were still in the ICU and/or receiving mechanical ventilation on day 40 were censored at day 41 for the analysis of time on mechanical ventilation and ICU LOS. For the mortality analysis, patients who were discharged alive from the hospital or were still in the ICU or hospital on day 40 were censored at day 41.
# Results
## Patient characteristics
All 77 patients with proven COVID-19 consecutively admitted to the ICU of Radboud University Medical Center were assessed for study inclusion. Patients with a preexisting immunosuppressed status were excluded (n = 9, see [fig_ref] Figure 1: Body mass index distribution and plasma cytokine levels [/fig_ref] , and one patient refused participation. The remaining study population (n = 67) was divided in obese (n = 18, of which 72% with class I obesity and 28% with class II obesity) and non-obese (n = 49) groups [fig_ref] Figure 1: Body mass index distribution and plasma cytokine levels [/fig_ref]. BMI of the entire study population was 27.7 [24.9-30.8] kg/m 2 [fig_ref] Figure 1: Body mass index distribution and plasma cytokine levels [/fig_ref]
## Inflammation and respiration
Plasma concentrations of all measured pro-inflammatory cytokines were highest at ICU admission and decreased over time both in the obese and non-obese groups (p < 0.05 for all cytokines, [fig_ref] Figure 1: Body mass index distribution and plasma cytokine levels [/fig_ref]. Except for slightly higher IP-10 levels in obese vs. non-obese patients at days 9-10, no between-group differences in any of the measured cytokines on any timepoints were observed [fig_ref] Figure 1: Body mass index distribution and plasma cytokine levels [/fig_ref]. Furthermore, BMI did not correlate with concentrations of IL-6, TNF-α, and IP-10 at ICU admission (r = −0.09 p = 0.61, r = 0.03 p = 0.99 and r = 0.28 p = 0.11, respectively, [fig_ref] Figure 2: Relationship between BMI and concentrations of circulating cytokines on day of admission... [/fig_ref] or any other cytokine measured (r values ranging from −0.11 to 0.06, p values all >0.50, see [fig_ref] Figure 2: Relationship between BMI and concentrations of circulating cytokines on day of admission... [/fig_ref]. Despite a trend toward less pronounced leucocytosis in obese patients, no significant differences were present between both groups at any of the timepoints [fig_ref] Figure 3: Routine laboratory and clinical inflammatory parameters [/fig_ref]. The baseline difference in temperature (i.e., higher in non-obese patients) persisted at later timepoints, but did not attain statistical significance [fig_ref] Figure 3: Routine laboratory and clinical inflammatory parameters [/fig_ref]. No differences were observed in CRP levels between the obese and non-obese groups [fig_ref] Figure 3: Routine laboratory and clinical inflammatory parameters [/fig_ref]. Also, no significant differences in plasma concentrations of PCT, ferritin and d-dimer were observed between both groups [fig_ref] Figure 3: Routine laboratory and clinical inflammatory parameters [/fig_ref]. Finally, no between-group differences in any of the ventilation parameters were observed .
## Clinical outcomes
On day 40 of ICU admission, 11% (n = 2) of the patients of the obese group was still mechanically ventilated and 17% (n = 3) was still in the ICU, while 16% (n = 8) of patients of the non-obese group were still mechanically ventilated and in the ICU (p = 0.60 and p = 0.97, respectively). Time on ventilator was 22 (16-40) days in the obese group and 27 (14-40) days in the non-obese group (p = 0.41). ICU length of stay was 25 (17-40) and 29 (15-40) days in the obese and non-obese group, respectively (p = 0.53). Forty-day ICU mortality was 17% (n = 3) in the obese group and 24% (n = 12) in the non-obese group (p = 0.50). The Kaplan-Meier curves for time on mechanical ventilation, ICU LOS and mortality are presented in .
# Discussion
In this observational study in critically ill COVID-19 patients, we observed that, while disease severity and other baseline covariates were similar, temperature at ICU admission was lower in obese compared to non-obese patients. This observation may suggest that the endogenous immune-hypothalamus axis responsible for the induction of fever (IL-1-IL-6-PGE2 pathway) is less pronounced in obese compared to non-obese patients. However, no significant differences in plasma concentrations of various inflammatory markers and cytokines (most importantly IL-6) were observed between both groups. To gain sensitivity to detect a possible relationship between BMI and the cytokine response we performed correlation analyses that confirmed lack of such a relationship. Also, no significant difference in duration of mechanical ventilation, ICU length of stay, or mortality was observed between the obese and non-obese group. Taken together, these data indicate that once a COVID-19 patient becomes critically ill, the immune response and clinical course of the disease is not relevantly influenced by the patients' BMI. Obesity has been identified as a risk factor for severe disease and ICU admission in COVID-19 patients admitted to non-intensive care hospital wards [bib_ref] COVID-19 and the role of chronic inflammation in patients with obesity, Chiappetta [/bib_ref] [bib_ref] Obesity is a risk factor for greater COVID-19 severity, Gao [/bib_ref]. This is not necessarily in contrast with our findings, as obese individuals may be more susceptible to infection, might have a different immune response prior to becoming critically ill, and/or may require more respiratory support [bib_ref] High prevalence of obesity in severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) requiring..., Simonnet [/bib_ref]. Although several studies have assessed associations between BMI and clinical outcomes/mortality in critically ill COVID-19 patients, conflicting findings have been reported [bib_ref] ICU and Ventilator Mortality Among Critically Ill Adults With Coronavirus Disease, Auld [/bib_ref] [bib_ref] Obesity, overweight and survival in critically ill patients with SARS-CoV-2 pneumonia: is..., Halasz [/bib_ref] [bib_ref] Factors associated with death in critically ill patients with coronavirus disease 2019..., Gupta [/bib_ref]. As such, the relationship between obesity and mortality in these critically ill patients remains unclear. What our study adds pertaining to the role of BMI in critically ill COVID-19 patients are data of inflammatory/immunological markers. Of interest, a paradoxical relationship between obesity and mortality has been described in the general ICU population, as well as in critically ill patients suffering from different infectious diseases, such as pneumonia and bacterial sepsis [bib_ref] Obesity survival paradox in pneumonia: a meta-analysis, Nie [/bib_ref] [bib_ref] Increased body mass index and adjusted mortality in ICU patients with sepsis..., Pepper [/bib_ref]. Overall, obese patients demonstrate a better survival compared to non-obese individuals, even when corrected for several covariates. This phenomenon has been coined the obesity paradox [bib_ref] Obesity in the critically ill: a narrative review, Schetz [/bib_ref]. Our study was underpowered to demonstrate statistically significant differences in mortality between obese and non-obese patients, and therefore we are unable to confirm or reject the presence of the obesity paradox in COVID-19 patients. The present study has several limitations. First, because of the observational nature of the study, no direct link between cause and effect can be deduced. Second, because of the relatively small group of (especially obese) patients, a type 2 error is possible, and our findings should be regarded as indicative, not conclusive. Nevertheless, there were also no nonsignificant trends in inflammatory parameters or correlations between BMI and any of the circulating cytokines, which would be expected in case BMI would play an crucial role in the immune response in critically ill COVID-19 patients. Third, the present study was conducted in a single center in the Netherlands, where (morbid) obesity Clinical outcome data. a Time on mechanical ventilation, b ICU length-of-stay, c ICU mortality. p values depicted in the panels were calculated using log-rank tests. Dotted lines indicate 95% confidence intervals. may be less widespread compared to other countries. Accordingly, the obese group consisted largely of mildly obese patients and no patients with a BMI above 40 kg/m 2 were present in our cohort. Fourth, because patients in our cohort were all included after ICU admission and received mechanical ventilation, no statements can be made about the relationship between obesity and the risk of ICU admission or requirement of mechanical ventilation. Finally, we focused on innate immunity, so possible differences between obese and non-obese COVID-19 patients in adaptive immune response remain to be explored.
In conclusion, the results of this study indicate that, in critically ill COVID-19 patients requiring mechanical ventilation, the patients' BMI is not related to a different innate immune response, unfavorable respiratory mechanics, or impaired outcome. A larger multi-center study with a more expansive BMI distribution is warranted to confirm our findings.
## Data availability
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.
[fig] Figure 1: Body mass index distribution and plasma cytokine levels. a Histogram depicting body mass index (BMI) frequencies. Kinetics of concentrations of circulating b interleukin (IL)-6, c IL-8, d IL-10, e tumor necrosis factor alpha (TNF-α), f interferon gamma (IFN-γ), g interferon gammainduced protein (IP)-10, h monocyte chemoattractant protein (MCP)-1 and i IL-1 receptor antagonist (RA). Data are presented as geometric mean with 95% confidence interval. *p < 0.05, calculated using Sidak's post-hoc multiple comparisons tests on individual timepoints. [/fig]
[fig] Figure 2: Relationship between BMI and concentrations of circulating cytokines on day of admission to the intensive care unit. a Interleukin (IL)-6, b tumor necrosis factor alpha (TNF-α) and c interferon gamma-induced protein (IP)-10. r-and p values were calculated using Spearman's correlation. See supplementary for the relationships between BMI and the other measured circulating cytokine concentrations. [/fig]
[fig] Figure 3: Routine laboratory and clinical inflammatory parameters. a White blood cell counts (WBC), b temperature, c circulating C-reactive protein (CRP). Data are represented as geometric mean (a) or geometric mean with 95% confidence interval (b, c). No significant differences on any of the individual timepoints were observed according to Sidak's post-hoc multiple comparisons tests. [/fig]
[fig] Figure 4: Clinical outcome data. a Time on mechanical ventilation, b ICU length-of-stay, c ICU mortality. p values depicted in the panels were calculated using log-rank tests. Dotted lines indicate 95% confidence intervals. [/fig]
[table] Table 1: Data are presented as n (%) or median[IQR]. [/table]
|
Clinical characteristics, outcomes and prognosticators in adult patients hospitalized with COVID-19
A B S T R A C TBackground: COVID-19 is a novel disease caused by SARS-CoV-2. Methods: We conducted a retrospective evaluation of patients admitted with COVID-19 to one site in March 2020. Patients were stratified into 3 groups: survivors who did not receive mechanical ventilation (MV), survivors who received MV, and those who received MV and died during hospitalization. Results: There were 140 hospitalizations; 22 deaths (mortality rate 15.7%), 83 (59%) survived and did not receive MV, 35 (25%) received MV and survived; 18 (12.9%) received MV and died. Thee mean age of each group was 57.8, 55.8 and 72.7 years, respectively (P = .0001). Of those who received MV and died, 61% were male (P = .01). More than half the patients (n = 90, 64%) were African American. First measured d-dimer >575.5 ng/mL, procalcitonin > 0.24 ng/mL, lactate dehydrogenase >445.6 units/L, and brain natriuretic peptide (BNP) >104.75 pg/mL had odds ratios of 10.5, 5, 4.5 and 2.9, respectively for MV (P < .05 for all). Peak BNP >167.5 pg/mL had an odds ratio of 6.7 for inpatient mortality when mechanically ventilated (P = .02). Conclusions: Age and gender may impact outcomes in COVID-19. D-dimer, procalcitonin, lactate dehydrogenase and BNP may serve as early indicators of disease trajectory.
# Methods
The protocol was reviewed by the local Institutional Review Board and deemed exempt.
The study was conducted at a large, academic, Midwestern institution which serves as a referral center for the State of Indiana. Surge planning allowed our institution to accommodate up to 278 patients with COVID-19 who may require intensive care and up to 400 patients who may require medical-surgical or progressive levels of care. The first patient with testing confirmed COVID-19 was admitted to our hospital on Initially, polymerase chain reaction based testing was available through coordination with the State Department of Health for all patients. Polymerase chain reaction testing became available within the institution on March 18, 2020. A list of all patients who present to the hospital and have testing performed is maintained by our Infection Preventionist (K.K.). We used this list to identify all patients admitted to the hospital between March 1 and March 31, 2020 whose test results indicated infection by the novel coronavirus.
A data collection form was created in REDCap, a secure web-based tool to facilitate research (data collection form available as supplementary material). [bib_ref] Research electronic data capture (REDCap)-A metadata-driven methodology and workflow process for providing..., Harris [/bib_ref] Items included demographics, comorbid conditions, clinical presentations, time stamped laboratory values, and hospital course. Data collection was discussed and operationalized between 3 authors who reviewed the electronic medical record of each patient (W.G., E.C., A.K.). The form was pilot tested and edited to enhance ease of use and consistency. Each patient's admission history and physical was reviewed and corroborating diagnostic information was retrieved when relevant (eg. hemoglobin A1c for patients with diabetes mellitus). Presenting symptoms were categorized as (1) respiratory complaints (cough, shortness of breath, and chest pain), (2) gastrointestinal (GI) complaints (nausea, vomiting, diarrhea, and abdominal pain), (3) fever, (4) syncope and altered mental status and, [bib_ref] Clinical predictors of mortality due to COVID-19 based on an analysis of..., Ruan [/bib_ref] constitutional symptoms (myalgias, anosmia, dysgeusia, anorexia, night sweats, fatigue, and weakness). The first recorded set of vital signs, imaging and laboratory data was captured. Vital signs and respiratory care notes were reviewed to assess the timing and magnitude of increasing oxygen needs. Laboratory data was reviewed for the entire hospital stay. The timing and values of first drawn possible prognosticators were recorded and peak values and timing were captured when these laboratories were checked more than once. Medications prior to admission were determined by reviewing the pharmacist's admission medication history. The medication administration record was accessed to confirm the receipt of COVID-19 specific therapies, steroids and vasopressors. The discharge summary was reviewed for complications. All patients admitted for confirmed or suspected COVID-19 during the study period at our institution had a complete blood count (CBC) and basic chemistries drawn on admission. Timing of prognostic laboratories and clinical trajectories were reported relative to the timing of this admission CBC. Data were collected for all patients until June 5, 2020.
Data were then downloaded from REDCap and described using descriptive statistics. We stratified the sample into three distinct groups of worsening severity based on outcomes: patients who did not receive mechanical ventilation (MV) and survived, those who received MV and survived, and those who received MV and died during the hospitalization.
Laboratory values that may serve as markers of disease severity were compared among these three groups. Based on prior research, laboratory values that were tracked included alanine transaminase, brain natriuretic peptide (BNP), creatine kinase (CK), c-reactive protein (CRP), d-dimer, ferritin, lactate dehydrogenase (LDH), procalcitonin and troponin. First and peak recorded values were retrieved and compared using analysis of variance testing. If testing revealed statistically significant differences in values between groups at P < .05, odds ratios were then calculated. At the time of presentation, concerns about the trajectory of patients with COVID-19 often revolve around whether the patient will require MV; and shift towards survival once MV is necessary. To parallel these clinical questions, the first measured values were used to calculate the odds of receiving MV, while peak values were used to calculate the odds of inpatient mortality in those who received MV. The 75th percentile value in the group that did not receive MV and survived was used as the cut-off to calculate the odds for receiving MV, while the 75th percentile value in those who received MV and survived was used as the cut-off to calculate the odds of mortality in those who received MV. The timing of both the first and peak prognostic laboratory values relative to the admission CBC time were also compared between the three groups.
Data was analyzed using the pandas package for Python, with Fisher exact testing used for categorical variables and contingency tables. Analysis of variance testing was performed for continuous variables, using the SciPy STATS package. [bib_ref] Author correction: SciPy 1.0: fundamental algorithms for scientific computing in Python, Virtanen [/bib_ref] Characteristics of the sample were compared between the 3 outcome groups.
# Results
Between March 1 and March 31, 2020, there were 140 admissions to the hospital with testing confirmed COVID-19. More than half (59.3%) of the sample did not receive MV.
There were a total of 22 deaths (15.7% mortality rate) however, 4 patients had goals of care that were focused on comfort and did not receive MV. Of those who received MV, 35 (66%) survived.
## Demographics and comorbidities
Overall the sample had roughly equal numbers of males and females; however, 68% of those who received MV were male. The mean age of those who received MV and died was 73 years while the mean age of those who received MV and survived was 55.8 years.
More than half the sample (64%) was African American and 67% of those who did not survive MV were African American. Gender and age distribution were statistically significantly different between the three outcome groups [fig_ref] Table 1: Demographics and medical comorbidities [/fig_ref].
Hypertension and diabetes mellitus were the most frequently noted comorbidities with the overall sample having a mean of 2.9 comorbidities per patient. The mean number of outpatient medications at the time of admission per patient was 7.6. Patients had few inpatient stays in our hospital system before the current encounter with a mean of 0.4 hospitalizations per patient over the prior 12 months.
The group that received MV and survived had the highest mean body mass index (BMI) (36) and the lowest proportion of individuals with normal BMIs. The mean BMI for those who received MV and did not survive was 27.2 which was lower than the mean BMI for the group who did not receive MV (32.3) (P <.05 for all pair-wise comparisons of BMI).
## Presenting symptoms and initial evaluation
Most patients presented with multiple symptoms, however patients who received MV and died reported fewer symptoms on presentation. Symptoms related to the respiratory system were the most frequent (93.6%) followed by reports of fever (65%) and GI complaints (51%). Fewer patients who received MV and died reported fever on admission and reported shorter duration of symptoms at the time of presentation [fig_ref] Table 2: Details of initial clinical presentation for adult admissions for COVID-19 between March... [/fig_ref].
While fever was a common complaint, the first mean recorded temperature for all groups was <38°C. More than half the sample (52%) met systemic inflammatory response (SIRS) criteria on admission; however, the distribution of the mean quick sequential organ failure assessment (Q SOFA) score differed between the three groups with the highest mean score noted in those who received MV and died. [bib_ref] Assessment of clinical criteria for Sepsis: for the third international consensus definitions..., Seymour [/bib_ref] Initial chest X-ray imaging was normal in 13.6% of presentations and bilateral infiltrates were the most commonly noted abnormality (69%). Bilateral infiltrates were less frequently observed in those who did not receive MV and survived. Mean presenting white counts, absolute neutrophil counts and blood urea nitrogen values were noted to be highest amongst those who received MV and died.
Only 3% of patients had a coinfection detected by respiratory viral panel testing.
## Clinical trajectory
Upon initial presentation, 79 patients (56%) did not receive supplemental oxygen while 26 (18.6%) did not receive any supplemental oxygen throughout the stay.
In patients who did not receive MV support and survived, the mean peak oxygen requirement by nasal cannula was 5.1 L. If needs were not met by nasal cannula, the mean peak FiO2 was 42.5%. These peak needs were reached in a mean of 32.4 hours following admission.
Patients who received MV and survived had a mean peak preintubation requirement of 12.1 liters by nasal cannula or mean FiO2 80% when nasal cannula did not suffice, reaching this peak in a mean of 35.9 hours following admission. In those who received MV and died, mean peak preintubation requirements by nasal cannula were 10.25 liters or FiO2 72.5%. These peak needs were reached in a mean of 36.1 hours following admission (Supplementary [fig_ref] Table 1: Demographics and medical comorbidities [/fig_ref].
The mean time to MV from admission in those who survived was 30 hours and for those who died was 52.5 hours (P = .1).
## Laboratory values and prognostication
Laboratory values that showed differences in distribution between groups with P-value <.05 in the first reported values included BNP, d-dimer, LDH, and procalcitonin. Odds ratio using the specified cut-offs for receiving MV were statistically significant for each of these laboratory values with the highest odds ratio (10.5) for (1) BP, blood pressure; IQR, interquartile range. *P values comparing the three outcome groups. Total column includes 4 patients whose goals of care focused on comfort. y SIRS= systemic inflammatory response syndrome (met ≥ 2 of the following criteria: Temp > 38°C, HR> 90, RR> 20, white count > 12k or < 4k). **q SOFA =quick sequential organ failure assessment (1 point each for Glasgow coma scale < 15, respiratory rate >= 22, systolic BP < = 100). receiving MV noted for d-dimer values elevated above 575.5 ng/mL [fig_ref] Table 3: Initial and peak laboratory values in all patients [/fig_ref].
Alanine transaminase, BNP, CK, CRP, d-dimer IL-6, and LDH values were statistically different between the groups at peak and were used to calculate the odds for mortality in patients receiving MV. At the thresholds used, only peak BNP achieved statistical significance with levels elevated above 167.5 pg/mL associated with a 6.8-fold increased risk for mortality in patients receiving MV [fig_ref] Table 3: Initial and peak laboratory values in all patients [/fig_ref].
Ferritin and troponin did not achieve statistically significant differences between groups at either first or peak measured values.
The timing of first reported laboratory values relative to admission CBC were similar between the 3 groups, however there were statistically significant differences in when peak values of CK,CRP, ferritin, procalcitonin, and troponin were achieved between the groups [fig_ref] Table 3: Initial and peak laboratory values in all patients [/fig_ref].
## Treatment
Most (n = 109, 78%) patients received at least one dose of hydroxychloroquine during the course of the hospitalization. At least one dose of azithromycin was given in 68% (n = 95) of cases. Six (4%) patients received tociluzimab and 1 (0.7%) received remdesivir. Systemic steroids were administered in 44 (31%) patients.
## Outcomes and complications
The mean length of stay was 6.5 days for patients who did not receive MV and 21.3 days for those who received MV and survived (P < .00001).
Shock requiring vasopressors was noted in 54.3% of those who survived MV and in 77.8% of those who received MV and died. Secondary bacterial infection which included pneumonia and bacteremia was common amongst those who received MV (noted in 42.9% of survivors who received MV and in 44.4% of those who received MV and died). Venous thromboembolism occurred in 3.6% of patients who did not receive MV and survived, in 20% of those who received MV and survived and in 27.8% of those who received MV and died. The 14-day readmission rate was 8.4% for those who did not receive MV and 11.4% for those who did. Renal failure necessitating the initiation of renal replacement therapy was noted in a third of those who received MV and died [fig_ref] Table 5: Outcomes and complications for adult patients admitted with COVID-19 between March 1-March... [/fig_ref].
# Discussion
We present in rich detail the clinical characteristics, laboratory evaluation, trajectories, and outcomes of all patients admitted with COVID-19 to our institution in March 2020 and explore prognostic implications of certain clinical and laboratory markers.
Our data corroborates the increased risk of severe disease and mortality in COVID-19 conferred by increasing age and male gender noted in previous studies. An early US report found a mortality rate of 67% in patients admitted to the intensive care unit where the mean age of the population was 70 years, and more serious illnesses in the US have been noted in older adults. 9,10 In our sample, while more males received ventilatory support than females, these differences were less marked than the initial data from China where 85% of those who required ICU care were male. [bib_ref] Clinical features of patients infected with 2019 novel coronavirus in Wuhan, Huang [/bib_ref] Data from hospitalized patients in the New York City area also noted worse outcomes in older and male patients. [bib_ref] Presenting characteristics, comorbidities, and outcomes among 5700 patients hospitalized with COVID-19 in..., Richardson [/bib_ref] Strategies to protect vulnerable, older adults should continue to be prioritized and further data adjusted for potential confounders will be needed to explore gender related disparities in COVID-19 outcomes.
We also note racial differences in the epidemiology of COVID-19. African Americans represented 64% of all hospitalized COVID-19 patients in our sample. To place this in perspective, in the five months preceding March 2020, the proportion of patients admitted to our facility who were African American was 21%. Our work does not allow us to explain the root causes of these differences however urgent attention is needed to understand and mitigate this trend.
Diabetes mellitus was the most frequently reported comorbid condition in patients hospitalized with COVID-19 in China. [bib_ref] Clinical characteristics of coronavirus disease 2019 in China, Guan [/bib_ref] [bib_ref] Clinical features of patients infected with 2019 novel coronavirus in Wuhan, Huang [/bib_ref] However, both our sample and a large series from New York City found hypertension to be the most frequent comorbidity noted in patients hospitalized with COVID-19. [bib_ref] Clinical characteristics of coronavirus disease 2019 in China, Guan [/bib_ref] [bib_ref] Presenting characteristics, comorbidities, and outcomes among 5700 patients hospitalized with COVID-19 in..., Richardson [/bib_ref] Interestingly, patients admitted for COVID-19 in our sample did not appear to be high utilizers of health care with few prior hospitalizations within our system.
There is data emerging linking obesity and increased disease severity in COVID-19. [bib_ref] Obesity in patients younger than 60 years is a risk factor for..., Lighter [/bib_ref] We found similar higher mean BMIs in patients who received MV compared to those who did not. However, in those who received MV but died, mean BMIs fell in the overweight category with more than a third of those who died having normal BMIs. The interaction between weight, need for MV and outcomes when ventilated requires further exploration.
In terms of presenting symptoms, half the patients in our sample had GI complaints, higher than reports from China where GI symptoms were recorded for only 13% of patients. Fewer patients who received MV and died reported fever as a presenting complaint and presented with a shorter duration of symptoms. These findings may offer clues to the differences in presentation that may signal different trajectories.
Presenting vital signs between patients in the 3 outcome groups appeared comparable and the high prevalence of meeting SIRS criteria on admission limits its utility as a predictor of clinical trajectory. However, q SOFA scoring on admission may prove to be a useful prognosticator on admission. Imaging and laboratory evaluations may also be helpful in guiding initial clinical concern as the presence of bilateral infiltrates, higher white counts, higher absolute neutrophil and blood urea nitrogen values on admission were noted more frequently in those who received MV and died.
Patients appear to "declare" themselves in the first 48 hours of admission with the mean time to reaching maximum oxygen requirements ranging from 32 to 36 hours following admission. Statistically nonsignificant differences were noted in the times to receiving MV. While the time from admission to MV in those who subsequently survived appears to be shorter than the time from admission to MV in those who died, we cannot determine whether this observation represents a difference in the rate of decline (with slower rates of decline portending worse prognosis), an impact on outcomes by early versus late MV or whether the decision to intubate was impacted by the team's awareness of prognosis.
Several laboratory values are being investigated as prognostic markers for severe disease. We analyzed the predictive ability of laboratory evaluations in two critical clinical periods posing two different decisions. We used the first measured values to predict the need for mechanical ventilation and the peak values to predict inpatient mortality in those who were mechanically ventilated. Multiple studies and our own data have demonstrated the marked derangements seen in patients with COVID-19. 14 Accordingly, we used cutoff values derived from the distribution of our own data set rather than reference ranges as thresholds to calculate odds ratios to present more meaningful and discriminatory interpretations for clinicians. D-dimer values have been reported to be abnormal in more than a third of patients presenting in China and with values >1,000 ng/mL associated with mortality. Our findings indicate that initial d-dimer values may also be used to predict the need for mechanical ventilation with values >575.5 ng/mL conferring a 10.5-fold increase in risk. Increasing procalcitonin and LDH values have also been associated with increased odds of mortality in COVID-19. [bib_ref] Clinical features of patients infected with 2019 novel coronavirus in Wuhan, Huang [/bib_ref] [bib_ref] Risk factors of fatal outcome in hospitalized subjects with coronavirus disease 2019..., Chen [/bib_ref] Our findings indicate that their first reported values may also predict the need for mechanical ventilation. We additionally identified the potential of first and peak BNP values to predict the need for mechanical ventilation and mortality respectively. Previous reports have found ferritin and troponin values to be predictive of mortality. [bib_ref] Clinical predictors of mortality due to COVID-19 based on an analysis of..., Ruan [/bib_ref] [bib_ref] Clinical course and risk factors for mortality of adult inpatients with COVID-19..., Zhou [/bib_ref] In our sample, however, these values were not statistically significantly different between groups at either first or peak measurement. Our findings may form the basis of future work to create scoring systems to improve our ability to predict the trajectory of patients presenting with COVID-19, stratify risk and guide subsequent management. The differences in the timing of peak values noted for certain laboratory ALT, alanine transaminase; BNP, brain natriuretic peptide; CK, creatine kinase; CRP, c-reactive protein; IL-6, interleukin 6; LDH, lactate dehydrogenase. Notes: ***'Other' complications included diabetic ketoacidosis (2), prolonged encephalopathy/ delirium (6) thrombocytopenia (2) stridor following extubation (3). IQR, interquartile range; NA, not applicable; O2, oxygen. *P values comparing the three outcome groups. Total column includes 4 patients whose goals of care focused on comfort.
values (CK, CRP, ferritin, procalcitonin, and troponin) raise additional areas for future research. The medical management of these patients appears to be complex and resource intensive with multiple complications, long lengths of stay and need for placement upon discharge. In addition to shock, venous thromboembolism and secondary bacterial infections, we also noted patients with prolonged encephalopathy and sequelae of prolonged intubation. Long-term monitoring is needed to identify delayed or prolonged deficits arising from the initial illness.
Our study has limitations. It is a single center's experience with a novel illness over the first month of its appearance at our institution. We relied on discharge summary documentation of complications with targeted review of corroborating diagnostics and may therefore be under reporting adverse outcomes. While our health system shares an electronic medical record platform across all 18 hospitals, we did not access city wide data to confirm readmissions or prior hospitalizations. Importantly, we present unadjusted odds ratios and analysis of differences between the three outcome groups. The differences found should be considered exploratory and hypothesis generating requiring confirmation in larger, multivariate analysis.
As the burden of this novel disease grows, sharing clinical information about patients will help us generate hypotheses and adapt our management and prevention strategies. Continued research on presentations, outcomes and complications in different settings and over longer periods will help improve the care we provide our patients.
[table] Table 1: Demographics and medical comorbidities: adult admissions for COVID-19 between March 1-March 31, 2020 at an academic health center BMI, body mass index; HIV, human immunodeficiency virus; IQR, interquartile range; SD, standard deviation; TIA, transient ischemic attack. *P values comparing the three outcome groups. Total column includes 4 patients whose goals of care focused on comfort. y Chronic lung disease = COPD, sarcoidosis, interstitial lung disease, pulmonary hypertension, cystic fibrosis, restrictive lung disease. z Immune suppressed = chronic steroid use, biologic agents for rheumatologic disorders or inflammatory bowel disease, recent chemotherapy, hematologic malignancy, history of bone marrow or solid organ transplant. [/table]
[table] Table 2: Details of initial clinical presentation for adult admissions for COVID-19 between March 1-March 31, 2020 at an academic health center [/table]
[table] Table 3: Initial and peak laboratory values in all patients [/table]
[table] Table 4: Unadjusted odds ratios for receiving mechanical ventilation and mortality if mechanically ventilated based on laboratory evaluation [/table]
[table] Table 5: Outcomes and complications for adult patients admitted with COVID-19 between March 1-March 31, 2020 at an academic health center [/table]
|
Comparison of breath‐hold and free‐breathing positions of an external fiducial by analysis of respiratory traces
An internal target volume (ITV) accounting for respiratory-induced tumor motion is best obtained using 4DCT. However, when 4DCT is not available, inspiratory/ expiratory breath-hold (BH insp , BH exp ) CT images have been suggested as an alternative. In such cases, an external fiducial on the abdomen can be used as a substitute for tumor motion and CT images are acquired when the marker position matchesas judged by the therapist/physicist -its positions at previously determined freebreathing (FB) respiratory extrema (FB insp , FB exp ). In this study we retrospectively determined the accuracy of these matches. Free breathing 4DCT images were acquired, followed by BH insp and BH exp CT images for 25 patients with non-small-cell lung cancer. Respiration was monitored using a commercial external fiducial system, which generates positional information while CT studies are conducted. Software was written for statistically analyzing the displacement of the external fiducial during BH insp and BH exp CT acquisition and comparing these displacements with corresponding mean FB extrema positions (FB insp and FB exp , respectively) using a Student's t-test. In 72% of patients, mean positions at BH insp differed significantly from mean positions at FB insp (p < 0.05: 0.13 -1.40 cm). In 92% of patients, mean positions at BH exp differed significantly from mean positions at FB exp (p < 0.05: 0.03 -0.70 cm), although this difference was smaller than 0.5 cm in many cases (median = 0.34 cm). Our findings indicate that relying solely on abdominal external markers for accurate BH CT imaging in order to accurately estimate FB extrema positions may be subject to significant error.
# I. introduction
Because of respiratory motion, lung tumors may move up to 2 cm during a single fraction during radiation therapy. [bib_ref] Respiratory kinematics of the upper abdominal organs: a quantitative study, Korin [/bib_ref] [bib_ref] Ultrasound quantitation of respiratory organ motion in the upper abdomen, Davies [/bib_ref] [bib_ref] Respiration gated radiotherapy treatment: a technical study, Kubo [/bib_ref] To account for respiratory motion in designing radiation therapy plans, the International Commission on Radiation Units and Measurements (ICRU) has identified the internal target volume (ITV) as encompassing the entire range of motion of a tumor, both demonstrably and microscopically, during treatment delivery.Initially, researchers defined population-based margins for expanding the clinical target volume to generate the ITV. State-of-the-art radiation therapy simulation can define the ITV based on the extent of tumor motion explicitly measured using four- dimensional computed tomography (4DCT). [bib_ref] Four Dimensional CT scans for treatment planning in stereotactic radiotherapy for stage..., Underberg [/bib_ref] Consequently, margins that account for respiratory motion can now be made patient-specific, in both magnitude and direction, resulting in better coverage for tumors with a great deal of motion and a significant decrease in the amount of irradiated uninvolved lung for tumors with little motion. [bib_ref] Dosimetric benefits of respiratory gating: a preliminary study, Butler [/bib_ref] The technology that enables explicit ITV determination based on 4DCT requires a high-end multislice CT scanner, as well as 4DCT reconstruction software. This exclusive combination of hardware and software may be prohibitively expensive for smaller institutions. Furthermore, because 4DCT images are taken while the patient is freely breathing, they can still contain residual motion blurring and artifacts. Therefore, an alternative procedure that has been suggested to determine the ITV is breath-hold (BH) CT imaging, in which the patient holds his or her breath during imaging at both end-inspiration (BH insp ) and end-expiration (BH exp ). [bib_ref] Internal target volume determined with expansion margins beyond composite gross tumor volume..., Shih [/bib_ref] The suitability of BH imaging for ITV determination in patients who undergo treatment while freely breathing is unknown. Even if one accounts for hysteresis and the non-rectilinear tumor motion between extrema positions, [bib_ref] Precise and real-time measurement of 3D tumor motion in lung due to..., Seppenwoolde [/bib_ref] how well the tumor positions at BH respiratory extrema reflect the tumor positions at free-breathing (FB) respiratory extrema remains unclear. Moreover, a common method of monitoring the respiratory cycle is the use of external fiducials placed on the abdomen, whose correlation with internal tumor positions is still questionable. [bib_ref] The correlation between internal and external markers for abdominal tumors: implications for..., Gierga [/bib_ref] Yet it is assumed during BH imaging that the positions of the external fiducial matches its corresponding positions at FB respiratory extrema. The simplest approach to determine appropriate BH positions is to rely on the patient's estimation of FB extrema positions with no external marker. However, this method results in no quantitative measurement of the BH position accuracy.
We performed the present retrospective study to quantitatively determine how accurately BH positions -as defined using an external marker on the abdomen -reflected free-breathing (FB) respiratory extrema positions in patients who had routinely undergone BH CT imaging for ITV construction. During institutional conversion from multiple BH CT studies to FB 4DCT for lung tumor ITV-generation, we investigated the suitability of FB 4DCT for deriving tumor ITVs. During this period, combined FB 4DCT and BH scans were performed at our institute under an Institutional Review Board (IRB) approved protocol. Although this combined, redundant study would not be routinely performed by institutions -and is no longer done at our own institute -the scans from this investigative period provided us with the opportunity for this retrospective study. Also, although the use an external fiducial system might be economically unfeasible for a small institution without 4DCT capabilities, we were in the position to exploit the Varian RPM system for quantitative assessment of BH position in this study.
## Ii. methods and materials
The study population consisted of 25 patients who underwent radiation therapy for non-small-cell lung cancer from to , at The University of Texas M. D. Anderson Cancer Center. During this period, free-breathing (FB) 4DCT and breath-hold (BH) CT scans were routinely performed for treatment-planning purposes. The patient data were retrospectively acquired under an IRB-approved retrospective chart-review protocol.
Patients underwent scanning using one of two multislice helical CT scanners: Mx8000 IDT (Philips Medical Systems, Cleveland, OH) and Discovery ST (GE Healthcare, Waukesha, WI). Immobilzation was achieved in the standard supine position using a wing board with a T-bar handgrip in conjunction with a vacuum immobilization device (BlueBAG Vacuum Cushions; Medical Intelligence, Schwabmünchen, Germany).
Real-time monitoring of patient respiration during FB 4DCT and BH CT scanning was accomplished using an external fiducial device (RPM; Varian Medical Systems, Palo Alto, CA) that measured the vertical displacement of the abdomen during respiration. The position of a reflector on the surface of a block positioned on each patient's abdomen was tracked using an infrared light source and a charge-coupled device (CCD) detector. Patients were allowed to observe their respiratory motion by viewing a liquid crystal display (LCD) flat-panel video monitor (Optiview, Rancho Dominguez, CA) mounted vertically at the end of the treatment couch through a mirror assembly. [bib_ref] Respiration-correlated treatment delivery using feedback-guided breath hold: a technical study, Nelson [/bib_ref] Some of the patients later used virtual reality goggles (i-O Display Systems, Sacramento, CA) rather than the monitor/mirror assembly for visual feedback, as the latter interfered less with patient setup and immobilization. Neither visual feedback method was deemed to be more accurate or superior to the other. The choice of visual feedback method depended on patient preference. Some patients could not tolerate wearing goggles, whereas the positioning of others did not allow the use of the monitor/mirror assembly method. Either way, by having a method to observe their own breathing traces, patients were able to exert finer control over the position of breath-hold.
Prior to acquiring the CT images during breath hold, the patients' respiration was monitored to estimate the average FB amplitude of fiducial motion during normal respiration. For BH imaging, patients were shown a horizontal bar that moved vertically, tracking the motion of the fiducials in real-time within a color-coded region that corresponded to the previously determined average FB amplitude. The patients thus obtained visual feedback that displayed the real-time position of the fiducial during CT image acquisition as well as target positions for their breath-holds corresponding to the upper and lower limits of the color-coded region. Patients were instructed to observe a trace of their respiratory cycle and perform a BH at end-inspiration (BH insp ) and end-expiration (BH exp ), corresponding to the upper and lower limits of the colored region.
After positioning the patient on the CT couch, the fiducial marker was placed on the patient approximately at midline and midway between the xiphoid process and umbilicus. This position usually corresponds with the maximum abdominal motion. The FB 4DCT scan was taken first, followed by the two BH scans (at end-inspiration and end-expiration) with no change in the patient or fiducial marker position.
Respiratory traces consisting of fiducial displacement as a function of both time and phase were obtained during the FB 4DCT and BH scans. For this study, the traces were exported as text files. A custom software program was developed to analyze these traces (MATLAB; The MathWorks, Natick, MA). This program displayed a respiratory trace on a time axis with indicators identifying the period of CT data acquisition (beam-on). For each FB 4DCT trace, the user selected the portion of the trace corresponding to the beam-on time, and the program automatically extracted the displacements corresponding to the end-inspiration and end-expiration periods of the breathing cycles within the selected time intervals. From the FB 4DCT traces, the mean position of the external fiducial at end-inspiration (FB insp ) and its standard deviation (SD-FB insp ) and the mean position of the external fiducial at end-expiration (FB exp ) and its standard deviation (SD-FB exp ) were calculated. Typically, FB statistics were extracted from the 12-15 breathing cycles required for the FB 4DCT study acquired after patients had achieved steady-state breathing. From BH traces, the mean position of the external fiducial at end-inspiration (BH insp ) and its standard deviation (SD-BH insp ) and the mean position of the external fiducial at end-expiration (BH exp ) and its standard deviation (SD-BH exp ) were calculated.
The differences between the displacements measured using the two methods and their uncertainty were computed for each patient. Because there was no alteration in the position of the fiducial marker between the FB 4DCT and BH scans, the positions measured during the scans were relative to the same coordinate system and could be compared directly. The differences were analyzed and tested for statistical significance using Student's t-test (p ≤ 0.05 for statistical significance). [fig_ref] FIG. 1: Data from Patient no [/fig_ref] summarizes the process of statistical analysis of respiratory traces. lists the absolute differences between average FB and BH positions at end-inspiration (Δ insp ) and end-expiration (Δ exp ), respectively. For end-inspiration, the absolute difference between the FB and BH fiducial positions ranged from 0 -1.4 cm (mean = 0.4 cm, SD = 0.4 cm). For endexpiration, the absolute difference between mean FB and BH fiducial positions ranged from 0 -0.7 cm (mean = 0.3 cm, SD = 0.2 cm). [fig_ref] FIG. 2: Histograms of the absolute differences between average FB and BH positions at [/fig_ref] represents a histogram of the absolute differences between the FB and BH end-inspiration positions. [fig_ref] FIG. 2: Histograms of the absolute differences between average FB and BH positions at [/fig_ref] is a histogram of the absolute differences between the FB and BH end-expiration positions. lists the standard deviations of the fiducial positions at end-inspiration and end-expiration and their mean values as a measure of stability of these positions.
# Iii. results
Finally, [fig_ref] TABLE 3: The statistical significance of absolute differences between free-breathing and breath-hold extrema positions... [/fig_ref] shows the p values for the differences in the FB and BH fiducial positions at inspiration (insp) and expiration (exp) extrema. In most cases, the p value was < 0.001, indicating highly significant differences between the FB and BH extrema positions.
# Iv. discussion
We performed the present study to determine how accurately BH positions -as defined using an external marker on the abdomen -reflect free-breathing (FB) respiratory extrema positions. shows that at end-inspiration the magnitude of the differences between the mean FB and BH fiducial positions ranged from 0 -1.4 cm. In the extreme case of patient number 15, a 1.4 cm difference in fiducial positions at inspiration could result in a tumor position at BH, which is very different to that during FB at the corresponding phase of the breathing cycle. At end-expiration, the magnitude of the differences between FB and BH fiducial positions ranged from 0 -0.7 cm. These data are summarized in the histograms of [fig_ref] FIG. 2: Histograms of the absolute differences between average FB and BH positions at [/fig_ref] , which show that there is less variation between BH and FB positions at end-expiration than at end-inspiration. Reproducibility of freebreathing extrema positions while the patient is freely breathing, is crucial for passively-gated respiratory-correlated therapy. The data in show that the average FB insp SD is greater than the average FB exp SD, suggesting that end-expiration is the best phase for passive respiratorygated treatment where the patient is freely breathing, which is in agreement with previous observations. [bib_ref] Potential radiotherapy improvements with respiratory gating, Keall [/bib_ref] [bib_ref] Patient training in respiratory-gated radiotherapy, Kini [/bib_ref] [bib_ref] Breathing-synchronized radiotherapy program at the University of California Davis Cancer Center, Kubo [/bib_ref] [bib_ref] Evaluation of internal lung motion for respiratory-gated radiation therapy using MRI: Part..., Liu [/bib_ref] [bib_ref] Deep inspiration breath hold and respiratory gating strategies for reducing organ motion..., Mageras [/bib_ref] [bib_ref] Correlation of gross tumor volume excursion with the desirability of respiratory gating, Starkschall [/bib_ref] [bib_ref] Respiratory gating for liver tumors: use in dose escalation, Wagman [/bib_ref] For BH studies, we found that the mean BH insp position was significantly different (≤ 0.05: 0.13 -1.40 cm) from the mean FB insp position in 18 (72%) patients and that the mean BH exp position was significantly different (p < 0.05: 0.03 -0.70 cm) from the mean FB exp position in 23 (92%) patients [fig_ref] TABLE 3: The statistical significance of absolute differences between free-breathing and breath-hold extrema positions... [/fig_ref]. The larger number of cases which reach statistically significant difference for end-expiration positions is due to the low variability (smaller standard deviations) of the endexpiration FB positions as seen in . While these differences for end-expiration positions are statistically significant, in practice the large magnitude differences are of most importance. For example, five patients had differences at end-expiration greater than 4 mm and six additional patients One reason for the differences between the BH and FB extrema positions is that during BH scanning, patients can contract or relax their abdominal muscles, causing the external fiducial to move significantly without any corresponding changes in lung volume or tumor position. These differences are not necessarily reflective of differences in internal tumor positions during BH compared to FB extrema. However, during a typical BH CT examination the therapist qualitatively judges the integrity of patient BH solely by comparing the external fiducial BH positions to FB extrema positions. This study quantitatively addresses the accuracy of this method by statistically comparing FB and BH respiratory traces derived from an external fiducial.
Correlation between the external fiducial and internal tumor position is an important, separate topic actively being investigated in our institute as well as others. Balter et al. compared positions of tumor GTVs at FB extrema (from the 0 and 50% phases of 4DCT studies) with BH CT derived GTVs at BH insp and Bh exp . [bib_ref] Comparison of 4DCT with breath-hold CT for determination of tumor motion with..., Balter [/bib_ref] This study showed that GTV positions during BH studies do not accurately represent the limits of the FB GTV, especially during inspiration. Similarly, with respect to the external marker, we found that the position of the fiducial during BH does not always accurately represent the limit of the fiducial position during FB. Therefore, care must be taken during BH imaging, interpretation of respiratory traces, and creation of treatment-planning margins based solely upon external markers. Alternatively, if available, an approach other than BH CT such as multiple slow CT scans, [bib_ref] Multiple 'slow' CT scans for incorporating lung tumor mobility in radiotherapy planning, Lagerwaard [/bib_ref] [bib_ref] Are multiple CT scans required for planning curative radiotherapy in lung tumors..., Van Sörnsen De Koste [/bib_ref] positron emission tomography, [bib_ref] Can PET provide the 3D extent of tumor motion for individualized internal..., Caldwell [/bib_ref] extended-time CT, or 4DCT, [bib_ref] Acquiring a four-dimensional computed tomography dataset using an external respiratory signal, Vedam [/bib_ref] [bib_ref] 4-dimensional computed tomography imaging and treatment planning, Keall [/bib_ref] may more accurately elucidate the true extent of tumor motion during FB and provide more accurate ITVs for treatment planning.
# V. conclusion
External fiducial positions during end-inspiration end-expiration breath-holds do not accurately correspond to positions of the fiducial during corresponding end-inspiration end-expiration phases of free breathing. Our findings indicate that relying solely on abdominal external markers for accurate BH CT imaging in order to accurately estimate FB extrema positions may be subject to significant error. Consequently, an internal target volume (ITV) generated using CT scans acquired with the fiducial at these breath hold positions could be erroneous and care must be taken using such an approach. If 4DCT, which explicitly accounts for tumor motion during the breathing cycle, is not available then additional margins may be warranted in the ITV generated using surrogate breath-hold techniques.
[fig] FIG. 1: Data from Patient no. 1 summarizing the process of respiration trace analysis. Vertical direction of traces represents vertical displacement of reflective marker (cm) and the horizontal direction represents time (t) (axes omitted for clarity). Free breathing (FB) extrema positions (FB insp and FB exp ) are extracted and compared to their breath-hold (BH) counterparts (BH insp and BH exp , respectively) using a Student's t-test. Time points at which CT data are acquired are highlighted in green. (a) End-inspiration FB (FB insp ) points (red squares) are extracted from which the mean and standard deviation (SD) are calculated. (b) The mean and SD of inspiration BH (BH insp ) are calculated and a Student's t-test was used to compare FB insp and BH insp . (c) Summary of the statistical analysis (scale expanded for visual clarity) showing no significant difference between the average displacement of the two endinspiration datasets. The box plot has lines at the lower quartile, median, and upper quartile values. The whiskers are lines extending from each end of the box to show the extent of the rest of the data. Outliers are data with values beyond the ends of the whiskers (red crosses). A similar analysis is applied at end-expiration. (d) Mean and SD extracted from end-expiration FB (FB exp ) points (blue circles) and compared to (e) expiration BH (BH exp ). (f) Summary of the analysis (scale expanded for visual clarity) of end-expiration data showing a significant difference, Δ, between the FB exp and BH exp data (p < 0.05) TABLE 2. Summary of the standard deviations of mean positions of the fiducials during end inspiration and end expiration (FB -free-breathing; BH -breath-hold; units in cm) Patient SD insp position SD exp position FB BH FB BH [/fig]
[fig] FIG. 2: Histograms of the absolute differences between average FB and BH positions at (a) end-expiration and (b) end-inspiration. [/fig]
[table] TABLE 3: The statistical significance of absolute differences between free-breathing and breath-hold extrema positions at end-inspiration (insp) and end-expiration (exp) Examination of the magnitude of the difference between FB and BH fiducial positions at end-inspiration |FB-BH| insp versus end-expiration |FB-BH| exp reveals that 14 patients show larger deviation for |FB-BH| insp than for |FB-BH| exp . Although this is a small study, these results suggest that BH exp is the more appropriate position for BH radiation therapy. Practically, however, there may be significant normal tissue sparing with the BH insp protocol as is evident during deep-inspiration breath hold (DIBH) radiation therapy.(16,(19)(20)(21)(22)(23)(24) [/table]
|
Mathematical Modeling and Data Analysis of NMR Experiments using Hyperpolarized 13C Metabolites
Rapid-dissolution dynamic nuclear polarization (DNP) has made significant impact in the characterization and understanding of metabolism that occurs on the sub-minute timescale in several diseases. While significant efforts have been made in developing applications, and in designing rapid-imaging radiofrequency (RF) and magnetic field gradient pulse sequences, very few groups have worked on implementing realistic mathematical/kinetic/relaxation models to fit the emergent data. The critical aspects to consider when modeling DNP experiments depend on both nuclear magnetic resonance (NMR) and (bio)chemical kinetics. The former constraints are due to the relaxation of the NMR signal and the application of 'read' RF pulses, while the kinetic constraints include the total amount of each molecular species present. We describe the model-design strategy we have used to fit and interpret our DNP results. To our knowledge, this is the first report on a systematic analysis of DNP data.
# Introduction
Dynamic nuclear polarization (DNP) involves saturating the magnetization of unpaired electrons in a radical-molecule (or metal ion) to lead to cross polarization of the magnetization of neighboring nuclei. Even though the theoretical description of the phenomenon in metal ions (and later on for electron spins) was made by Overhauser in 1953, [bib_ref] Polarization of nuclei in metals, Overhauser [/bib_ref] and its experimental demonstration was in the same year by Carver and Slichter, 2 experimental applications of the phenomenon did not evolve greatly over the next 50 years.
With new technological developments (highfrequency microwave sources, cryo-technology), a rapid-dissolution DNP system was created [bib_ref] Increase in signal-to-noise ratio of .10,000 times in liquid-state NMR, Ardenkjaer-Larsen [/bib_ref] and commercialized (HyperSense ® , Oxford Instrument). Typically, the sample, a 13 C-labelled compound dissolved in aqueous solution with a free radical, is cooled to ∼1.4 K inside a superconducting magnet (3.35 T). The electrons of the radical are excited by microwave irradiation (∼94.1 GHz) and their magnetization is transferred to the nuclei via electron-nuclear dipolar coupling. This operation takes from a few tens of minutes to several hours. At the end of the hyperpolarization process, the sample is rapidly returned to its physiological temperature (typically 37 °C) by automatically dissolving it in a buffer that is at high pressure and temperature, and the buffer-dissolved sample is shuttled via a delivery pipe (typically Teflon) to the magnet to be injected either in vitro or in vivo. The first reported signal enhancement in [bib_ref] IDEAL spiral CSI for dynamic metabolic MR imaging of hyperpolarized [1-13 C]pyruvate, Wiesinger [/bib_ref] C NMR (urea sample) was greater than 10,000fold. [bib_ref] Increase in signal-to-noise ratio of .10,000 times in liquid-state NMR, Ardenkjaer-Larsen [/bib_ref] Over the last decade, this technique has mostly been developed for in vivo applications to detect and characterize diseases like cancer. [bib_ref] Molecular imaging with endogenous substances, Golman [/bib_ref] [bib_ref] Real-time metabolic imaging, Golman [/bib_ref] [bib_ref] Detecting tumor response to treatment using hyperpolarized 13 C magnetic resonance imaging..., Day [/bib_ref] [bib_ref] Magnetic resonance imaging of pH in vivo using hyperpolarized 13 C-labelled bicarbonate, Gallagher [/bib_ref] [bib_ref] Kinetics of hyperpolarized 13 C 1 -pyruvate transport and metabolism in living..., Harris [/bib_ref] [bib_ref] Hyperpolarized 13 C spectroscopic imaging informs on hypoxia-inducible factor-1 and Myc activity..., Dafni [/bib_ref] [bib_ref] Tumor imaging using hyperpolarized 13 C magnetic resonance spectroscopy, Brindle [/bib_ref] The molecule most used to investigate metabolism has been pyruvate as it hyperpolarizes well, and rapidly enters into the tricarboxylic acid cycle (TCA cycle; or Krebs cycle) or is converted into alanine, lactate and bicarbonate. It has a reasonably long observable life-time (related to the spin-lattice relaxation T 1 ). 11 As dissolution-DNP requires rapid recording, fast image-acquisition procedures have been developed (based on EPSI, and SPIRAL). [bib_ref] A method for simultaneous echo planar imaging of hyperpolarized 13 C pyruvate..., Reed [/bib_ref] [bib_ref] IDEAL spiral CSI for dynamic metabolic MR imaging of hyperpolarized [1-13 C]pyruvate, Wiesinger [/bib_ref] In vivo studies have focused mostly on the 13 C nucleus, but for completeness we note that other nuclei have been hyperpolarized, some having direct biological interest ( 15 N, [bib_ref] Kinetic modeling of hyperpolarized 13 C 1 -pyruvate metabolism in normal rats..., Zierhut [/bib_ref] F, 31 P) and others being more interesting from a chemistry perspective, or as indirect biological probes (eg, 89 Y, 107,109 Ag). [bib_ref] Monitoring the solid-state polarization of 13 C, 15 N, 2 H, 29..., Reynolds [/bib_ref] [bib_ref] Feasibility of in vivo 15 N MRS detection of hyperpolarized 15 N..., Cudalbu [/bib_ref] [bib_ref] DNP by thermal mixing under optimized conditions yields .60 000-fold enhancement of..., Lumata [/bib_ref] [bib_ref] Production and NMR characterization of hyperpolarized 107,109 Ag complexes, Lumata [/bib_ref] Huge efforts have been made to refine rapid-dissolution DNP to demonstrate its applicability in medical diagnosis. While the results are convincing, they are nevertheless preliminary. The production rate of particular metabolites is significantly different between the healthy and diseased tissues, but quantitative flux analysis is as yet naive. Most of the data fitting has been done by considering a simple one-way reaction without taking into account the enzyme concentrations or possible Michaelis-Mententype saturation kinetics, or the possible inhibition of the reaction(s) due to the injection of a large amount of an enzyme's substrate. From this perspective, data analyzed using a simple linear relationship between enzyme concentration/activity and the apparent reaction flux are unconvincing. During the conversion of pyruvate to lactate, some researchers have noted that increases in the lactate signal not only are due to biochemical reactions, but also to magnetization transfer between the substrate and product of the reaction. [bib_ref] Detecting tumor response to treatment using hyperpolarized 13 C magnetic resonance imaging..., Day [/bib_ref] [bib_ref] Magnetization transfer measurements of exchange between hyperpolarized [1-13 C]pyruvate and [1-13 C]lactate..., Kettunen [/bib_ref] To the best of our knowledge, very few studies have used the more realistic models of enzymic reactions or investigated the use of the mathematical models of the relaxation-reaction scheme in their data analysis. One of the first studies to use a kinetic model including the Michaelis-Menten mechanism involved an animal model of prostate cancer. [bib_ref] Kinetic modeling of hyperpolarized 13 C 1 -pyruvate metabolism in normal rats..., Zierhut [/bib_ref] The authors showed an effect of the substrate dose on the estimates of the kinetic parameters. By combining 13 C and 31 P data and a "realistic" relaxation-kinetic model, Harris et al claimed that the conversion of hyperpolarized pyruvate to lactate in breast cancer cells (T47D) was limited by the rate at which the pyruvate crossed the cell membrane. [bib_ref] Kinetics of hyperpolarized 13 C 1 -pyruvate transport and metabolism in living..., Harris [/bib_ref] This realization was later discussed by Witney et al 20 and a recent study compared the data fitting using different models based on first-order kinetics with either two-or three-pool, uni-or bi-directional chemical reactions. [bib_ref] Comparison of kinetic models for analysis of pyruvate-to-lactate exchange by hyperpolarized 13..., Harrison [/bib_ref] While data from hyperpolarized substrates by themselves could be fitted by a simple 2-pool model, the inclusion of mass spectrometry data necessitated the use of a 3-pool model to achieve realistic fits to the data. Most of the developments in modeling the experiments were performed on in vitro systems. Unfortunately, none of these papers extensively explain the strategy used to design their model.
In this article, we describe the physical phenomena that must be considered when constructing a mathematical model that describes data from a rapiddissolution DNP experiment. The strategy used to create the model and fit the experimental data is explained; it derives its concepts from the older literature used to describe 'tracer exchange' enzyme kinetics and its later application to magnetization transfer in NMR spectroscopy. [bib_ref] NMR spin exchange kinetics at equilibrium in membrane transport and enzyme systems, Kuchel [/bib_ref] To our knowledge, this is the first paper to explicitly explain in extenso the necessary constraints when modeling rapiddissolution DNP experiments.
## Hyperpolarization considerations
The mathematical modeling of an experiment involving hyperpolarized metabolites requires account to be taken of more parameters than those in classical (bio)chemical kinetics. Some of these refinements are well described in the literature (for example, [bib_ref] Magnetization transfer measurements of exchange between hyperpolarized [1-13 C]pyruvate and [1-13 C]lactate..., Kettunen [/bib_ref] [bib_ref] Kinetic modeling of hyperpolarized 13 C 1 -pyruvate metabolism in normal rats..., Zierhut [/bib_ref] [bib_ref] Comparison of kinetic models for analysis of pyruvate-to-lactate exchange by hyperpolarized 13..., Harrison [/bib_ref] while others are not. In the following sections we explain the different kinetics aspects that arise as a result of the excited (hyperpolarized) state.
## Spin-lattice relaxation t 1
The first key parameter that must be included in a mathematical model of the kinetics/relaxation scheme is the spin-lattice relaxation time, T 1 . Indeed, after hyperpolarization, the nuclear spins relax back to a Boltzmann-distribution of states, ie, to equilibrium (at the same rate as 'normal' polarized spins). The magnetization at any time, t, can be expressed as a function of the initial magnetization M 0 and the relaxation time:
[formula] M t M e z t/T ( ) = − 0 1 (1) [/formula]
where M z (t) denotes the magnetization of the hyperpolarized species along the direction of the magnetic field B 0 in the NMR spectrometer magnet. According to this equation the magnetization relaxes without giving rise to an observable signal. To observe the NMR signal, the magnetization must be transferred into the x,y-plane by applying an RF pulse. This point is more broadly described in the following section.
Because the relaxation rate is dependent on both the chemical and physical environments, it is important to estimate its value for each of the species studied for all physical and biological systems. [fig_ref] Figure 1: Magnetization [/fig_ref] shows simulations of magnetization decay for a range of relaxation times; the figure reveals how critical it is to consider differences in T 1 values between different chemical species and/or media. Furthermore, in conventional NMR spectroscopy, long T 1 values are usually considered a drawback. This is because they require the use of a long recycle time (or relaxation delay) between acquisition of free induction decays (FIDs), thus ensuring so called 'fully relaxed' spectra that quantitatively reflect the number of spins in the sample. However, as shown in [fig_ref] Figure 1: Magnetization [/fig_ref] , for hyperpolarized spins in which the signal can only be observed before the spins return to equilibrium, molecules having long T 1 values are preferred in order to observe the signal over a longer timescale.
## 'read' rf pulses
As noted in the previous section, the magnetization of hyperpolarized spins is aligned along B 0 , so it is not directly detected. Therefore at least one RF pulse must be applied in order to nutate ("sample") the magnetization into the x,y-plane; the magnetization in this latter plane is what is detected by the spectrometer. The detected magnetization is directly removed from the longitudinal magnetization, thus diminishing its intensity. [fig_ref] Figure 2: Schematic representation of the influence of RF pulse angle on the magnetization... [/fig_ref] is a schematic illustration of the effect of an RF pulse on the z-magnetization. The application of an RF pulse having a nutation angle α allows the detection of a signal that is proportional to the sine of the RF pulse angle (sin α), while the remaining z-magnetization is proportional to cos α. Meanwhile, the magnetization that was just detected (and so is no longer part of the hyperpolarized population) can be considered to have become "invisible". In other words, an amount that is proportional to (1 − cos α) has been added to this non-hyperpolarized pool.
The importance of choosing an appropriate z-magnetization-sampling RF nutation angle ("flip" angle) should be obvious: if the nutation angle is too small, no or low signal (and hence poor signalto-noise) will be obtained in the emergent spectrum. On the other hand, a large flip angle will very rapidly deplete the z-magnetization. A 90° RF pulse allows the recording of the maximum signal intensity within a single transient (FID) (sin 90° = 1) but subtracts all of the hyperpolarized magnetization (cos 90° = 0). Typically, the NMR spectroscopist is interested in following a (bio)chemical reaction over time, so they will use a series of small-angle RF pulses (usually between 2° and 20°) with a short intertransient (between FIDs) repetition time (of the order of 1 s).
To demonstrate the effect of the magnitude of the flip RF pulse angle on the signal attenuation we show the case of magnetization decaying according to a T 1 value of 40 s. The signal decays faster when RF pulses are applied; this is due to the transfer of z-magnetization into the x,y-plane in order to record the FID (NMR signal). The inset shows increased magnification ("zoom") of the beginning of the signal attenuation curve; it reveals the discretization of the magnetization that is lost on application of each RF pulse. , we generated data points for signal that was sampled with a read RF pulse applied every 1 s and fitted these data by using the 'classical' T 1 equation, (see . While the error in the estimate of T 1 was small for a small train of read RF pulses, it became much larger for a larger flip angle. The error is quite pronounced for a pulse angle as small as 5°. We recently proposed a method based on non-regular time intervals to accurately fit both T 1 and the RF pulse angle. 23
## Kinetic model
Once the MR considerations have been made, the second key aspect for modeling the experiment is the (bio)chemical kinetics. Under the experimental conditions used (ie, low RF flip angle, short repetition hyperpolarization decreases during the experiments due to both longitudinal (T 1 ) relaxation and the read RF pulses. Indeed, there is no difference in terms of kinetics between the "hot" (hyperpolarized) and "cold" (non-hyperpolarized) pools of a solute; the difference only exists in terms of the NMR signal. is a representation of the simplest chemical exchange reaction: the reversible interconversion of A and B. Considering that the nuclear spins in the molecules are (partially) hyperpolarized (denoted by * in , we must consider both the hot and cold pools when writing the kinetic equations. The chemical fluxes between A and B, characterized by the unitary rate constants k 1 and k −1 , are independent of the hyperpolarized state of the molecule and apply to the total amounts of each chemical species, namely A+A* and B+B*. However, the detected signal only depends on A* and B* and it decreases according to the longitudinal relaxation time of each species. It also decreases as a result of the z-magnetization-sampling RF pulses. While the relaxation effect is relatively easy to incorporate into the model (eg, as represented in , the strategy for considering the effect of the z-magnetization-sampling RF pulses is more difficult and is described next.
## Building a mathematical model and fitting it to experimental data building the mathematical model
The construction of a realistic mathematical model of the experimental system requires the use of dedicated software such as Matlab or Mathematica; in our research group, we use the latter. The model is built as a 'function' inside a module, thus enabling the use of locally defined variables. All the parameters to be evaluated are combined into a vector which becomes the 'argument' (input) of the model. This approach allows the ready numerical evaluation of the model. As an illustration, consider the system described in with the notable parameters whose values must be estimated: T 1 , k 1 and k −1 . Suppose that the function is given the name MODEL, the variables are input in the argument of the function as follows:
[formula] vector = {T 1,A , T 1,B , k 1 , k −1 }; MODEL[vector]; [/formula]
The most important feature in MODEL is the differential rate equations that describe chemical flux in the (bio)chemical reactions. As mentioned above, two sets of equations must be written: the first set for the hyperpolarized (detected) molecules, and the second set for the cold (non-hyperpolarized) molecules. Then the pools of hot (hyperpolarized) and cold molecules are interconnected by the relaxation of the hot molecules. For the system described in , the relevant differential equations are (these take the form that has been used in tracer exchange enzyme kinetics, and also in earlier studies using magnetization-transfer NMR spectroscopy to study membrane transport processes in cell
## B t k a t (2)
As described in the section on the effect of the read RF pulses, magnetization is decreased in discrete steps by these pulses. Since the system of differential equations may be nonlinear and hence not have a simple analytical solution, we choose to calculate the magnetization after each RF pulse by numerical integration of the differential equations, ie, to carry out the integration over the discrete time intervals between each z-magnetization-sampling RF pulse. Thus, these calculations generate a series of values of all signals for the series of time points (see below for a more detailed explanation). To simulate signal evolution for each reactant during a time course, we employ a loop that has the same number of increments as we have experimental time points (1 NMR spectrum = 1 time point); thus it is necessary to define the time during which the signal freely evolves (no RF pulse). Typically, this time is the inter-FID repetition time (TR). We then solve the set of differential equations over TR using as the initial conditions the magnetizations at the beginning of the new TR interval. The amount of each reactant is cumulatively added to the output matrix. The new initial conditions must be determined, and these depend on the flip angle, α, of the z-magnetization-sampling RF pulse. The hyperpolarized reactants will experience diminution by a factor (1 − cos α) while the non-hyperpolarized molecules will have their amounts increased by this factor. Typically this outcome is expressed in the program as follows:
[formula] A A t B B t A A [/formula]
where t end denotes the signal at the end of the time interval, and the subscript 0 denotes the initial value for the next numerical-integration period. At the end of the simulation, the discrete time values and the amounts of the different reactants, including the initial quantities, are output.
## Further comments on the model
The readily apparent parameters whose values we seek to estimate are concerned with relaxation and (bio)chemical interconversions. However, any of the parameters appearing in the above description can be estimated in the statistical fitting analysis. Uncertainties exist regarding the value of the z-magnetizationsampling RF pulse as this one is strongly dependent on the electronics of the NMR spectrometer, as well as the ionic composition (electrical conductivity) of the sample-medium. Therefore, the pulse angle can be added as a floating (to be fitted) parameter to the input vector.
For both in vitro and especially for in vivo experiments the model must account for the "arrival time" (delay after injection into the sample or animal) of the hyperpolarized reactant. For the former case, the hyperpolarized molecules are usually injected directly into the sample in NMR magnet, thus constituting a stopped-flow type of experiment. However, because the injection of a solution creates distortion of the magnetic field, the first transients cannot be used in the subsequent data analysis (or alternatively the start of data acquisition can be judiciously delayed). To take this effect into account, a delay parameter can be added to the model as a floated (to be estimated/fitted) parameter. In the case of in vivo experiments, the hyperpolarized solute is typically injected into the tail vein of the animal (rodent) and it travels throughout the circulatory system to reach the organ of interest. This injection is usually not as fast as is achievable in an in vitro experiment. Thus the signal built-up can readily be observed and the arrival time of the chemical at the particular organ must be factored into the model. This is done by modifying the differential equations for the injected hyperpolarized reactant during the time that its signal increases. A kinetic constant k inj , to simulate the time course of delivery of the compound, is included. Using the previous example and considering that A* is injected, its differential equation during its delivery phase (the first few seconds) is expressed as:
Finally, to check that there are no obvious mistakes in the model, the output data, including all molecules for both hot and cold states are scrutinized for their conservation of mass throughout the simulated time courses.
## Fitting the data
Once the model has been composed, it is facile to generate apparent signal evolutions for a set of values of relaxation and kinetic parameters and to compare the predictions with real experimental data. To fit the experimental dataset, we have used an approach based on Markov-chain Monte-Carlo (MCMC) analysis. [bib_ref] Relaxation times of spin states of all ranks and orders of quadrupolar..., Kuchel [/bib_ref] [bib_ref] Quantitative model of NMR chemical shifts of 23 Na + induced by..., Puckeridge [/bib_ref] Any constraints (on T 1 and k values such as only positive numbers being allowed) are specified in the Mathematica program. Because a detailed description of the fitting method used in our recent work on DNP is beyond the scope of this article, the interested reader is referred to the following references. [bib_ref] Relaxation times of spin states of all ranks and orders of quadrupolar..., Kuchel [/bib_ref] [bib_ref] An introduction to MCMC for machine learning, Andrieu [/bib_ref] conclusions We have shown which key parameters need to be taken into account while modeling data from DNP experiments. On one hand, NMR parameters (viz., T 1 or the 'magnetization-sampling' RF pulse angle α) must be built into the model to properly describe the intrinsic process of signal attenuation. On the other hand, the total quantity of metabolites must be accounted for when describing the (bio)chemical kinetic processes. The consequence of this is to double the number of differential equations to account for the pools of both hyperpolarized and non-hyperpolarized reactants. We have explained our model-design strategy and how to incorporate all parameters into the model. Extra parameters like an initial transient time (delay) before the start of the reaction(s), or progressive delivery over a finite time interval of hyperpolarized reactant(s), can readily be added to the model by using the basic strategy described here.
[fig] Figure 1: Magnetization (signal) decrease due to spin-lattice relaxation predicted by equation (1). notes: The chosen T 1 values were (s): 40 (red); 35 (blue); 30 (purple); 20 (brown), and 10 (black). [/fig]
[fig] Figure 2: Schematic representation of the influence of RF pulse angle on the magnetization due to the hyperpolarized state. notes: To "sample" the magnetization a small-angle RF pulse (angle α) is applied partially converting the magnetization into a detectable one in the x,yplane and thus decreasing the residual magnetization. [/fig]
[fig] Figure 3, Figure 4: (A) Comparison of the attenuation of magnetization in the absence (red) and presence (blue) of a 4° RF pulse. The inset shows the "sloping step-wise" signal attenuation due to the RF pulses applied every 1 s. (B) Data points were generated for nuclear spins having a longitudinal relaxation time, T 1 , of 40 s and different read pulse angles. notes: These data were fitted using the 'classical' Bloch T 1 relaxation equation (equation (1), full lines); we report the error in such T 1 estimations. time TR), the only detected signal emerges from the hyperpolarized molecules. However, the (bio) chemical kinetics of a system depends on the mole amounts of the molecules and not only on the observed MR signal intensity (which is proportional to the number of molecules present in the sensitive volume of the coil). In the case of hyperpolarized molecules, it is important to consider the undetected (nonhyperpolarized) pool of molecules when writing down the rate equations for the reaction scheme. This is because (1) the initial polarization level does not reach 100% (up to 10s of percent), and (2) the Representation of a simple kinetic model in which molecular species A is converted to B and vice-versa with unitary rate constants k 1 and k −1 , respectively. notes: The * denotes the hyperpolarized molecules that relax to their equilibrium Boltzmann state according to the relaxation time T 1 (subscript A and B denote the corresponding species). [/fig]
|
Structure of the drug target ClpC1 unfoldase in action provides insights on antibiotic mechanism of action
[bib_ref] The structure of ClpP at 2.3 A resolution suggests a model for..., Wang [/bib_ref] [bib_ref] Mycobacterium tuberculosis ClpP1 and ClpP2 function together in protein degradation and are..., Raju [/bib_ref] [bib_ref] Mutation analysis of the interactions between Mycobacterium tuberculosis caseinolytic protease C1 (ClpC1)..., Jung [/bib_ref] [bib_ref] The natural product cyclomarin kills Mycobacterium tuberculosis by targeting the ClpC1 subunit..., Schmitt [/bib_ref] [bib_ref] The antibiotic cyclomarin blocks arginine-phosphate-induced millisecond dynamics in the N-terminal domain of..., Weinhäupl [/bib_ref] [bib_ref] A rufomycin analogue is an anti-tuberculosis drug lead targeting CLPC1 with no..., Choules [/bib_ref] [bib_ref] Lassomycin, a ribosomally synthesized cyclic peptide, kills Mycobacterium tuberculosis by targeting the..., Gavrish [/bib_ref] [bib_ref] Mutation analysis of the interactions between Mycobacterium tuberculosis caseinolytic protease C1 (ClpC1)..., Jung [/bib_ref] [bib_ref] Structural basis of mycobacterial inhibition by cyclomarin A, Vasudevan [/bib_ref] [bib_ref] High-resolution structure of ClpC1-rufomycin and ligand binding studies provide a framework to..., Wolf [/bib_ref] [bib_ref] Structural basis of mycobacterial inhibition by cyclomarin A, Vasudevan [/bib_ref] [bib_ref] High-resolution structure of ClpC1-rufomycin and ligand binding studies provide a framework to..., Wolf [/bib_ref] [bib_ref] The antibiotic cyclomarin blocks arginine-phosphate-induced millisecond dynamics in the N-terminal domain of..., Weinhäupl [/bib_ref] [bib_ref] Mycobacterium tuberculosis ClpC1 N-terminal domain is dispensable for adaptor protein-dependent allosteric regulation, Marsee [/bib_ref]
# Results
Mutation of phenylalanine 444 into alanine results in a fully functional MtbClpC1 with a stabilized hexameric state
Although the ClpC1-NTD has been extensively studied by X-ray crystallography and solution NMR [bib_ref] Structural basis of mycobacterial inhibition by cyclomarin A, Vasudevan [/bib_ref] [bib_ref] High-resolution structure of ClpC1-rufomycin and ligand binding studies provide a framework to..., Wolf [/bib_ref] [bib_ref] Structure of the N-terminal domain of ClpC1 in complex with the antituberculosis..., Wolf [/bib_ref] , no structural information on the full-length ClpC1 has so far been reported. This can be rationalized by several key factors. Firstly, we have shown that ClpC1 exists in an equilibrium between a resting state and the active hexameric state [bib_ref] The antibiotic cyclomarin blocks arginine-phosphate-induced millisecond dynamics in the N-terminal domain of..., Weinhäupl [/bib_ref]. Second, the active hexamer is only formed in the presence of ATP, which is rapidly converted into ADP. Finally, the ClpC1P1P2 Mtb system is, compared to homologs from Staphylococcus aureus and Bacillus subtilis, more insoluble and therefore harder to work with in vitro [bib_ref] Toxic activation of an AAA+ protease by the antibacterial drug cyclomarin A, Maurer [/bib_ref]. Indeed, these limitations and the fact that so far all the NPAs targeting ClpC1 bind to the ClpC1 NTD have led others to employ chimeras of ClpC1 and the D1D2 domains of S. aureus to study the function and mode of action of cyclomarin [bib_ref] Toxic activation of an AAA+ protease by the antibacterial drug cyclomarin A, Maurer [/bib_ref]. While this is an interesting approach, it has the limitation of inferring results from nonphysiological proteins and therefore not providing any direct and useful structural data for drug development. This is particularly evident, considering the existing mechanistic differences between the S. aureus and Mtb proteins. For example, whereas SaClpC depends on MecA to catalyze the degradation of GFPssra by ClpP, ClpC1 is fully capable of doing so, independently of the adapter [bib_ref] Structure and functional properties of the active form of the proteolytic complex,..., Li [/bib_ref] [bib_ref] Development of high throughput screening methods for inhibitors of ClpC1P1P2 from Mycobacteria..., Fraga [/bib_ref] [bib_ref] Cleavage specificity of Mycobacterium tuberculosis ClpP1P2 protease and identification of novel peptide..., Akopian [/bib_ref]. In fact, MecA does not exist in the Mtb genome. It is therefore fundamental to obtain structural information on ClpC1 in its functional state.
Given the presence of an equilibrium between a decameric resting state and the active hexameric ClpC1 states in solution, we rationalized that the stabilization of the ClpC1 hexamer could allow us to obtain structural information on this important drug target. Previously, stabilization was achieved by removing intrinsically disordered loops, allowing the elucidation of the only ClpC X-ray structure described to date [bib_ref] Structure and mechanism of the hexameric MecA-ClpC molecular machine, Wang [/bib_ref] , but at the expense of enzymatic activity and therefore mechanistic relevance. Carroni et al.have demonstrated that point mutations in the middle domain (MD) of SaClpC shift the equilibrium towards the hexameric state. We hypothesized that similar modifications could also stabilize the active hexameric state of ClpC1, allowing subsequent structural characterization. In particular, residue F436 in SaClpC (F444 in Mtb) has been shown to be important for the resting state-hexamer equilibrium [fig_ref] Figure 1: ClpC1 resting statehexamer equilibrium [/fig_ref].
By means of site-directed mutagenesis, we mutated the homologous residue in ClpC1 to an alanine (F444A), to test if it could shift the ClpC1 equilibrium. Using a Superose 6 10/300 GL column, we observed that, as previously shown (7), WT ClpC1 (ClpC1 WT ) migrates as a species larger than the canonical hexamer. By contrast, in the presence of ATP, the F444A main species (ClpC1 F444A ) eluted in a volume consistent with the size expected for a ClpC1 hexamer [fig_ref] Figure 1: ClpC1 resting statehexamer equilibrium [/fig_ref]. Furthermore, the presence of hexameric rings in the mutant was also confirmed by negative stain EM [fig_ref] Figure 1: ClpC1 resting statehexamer equilibrium [/fig_ref].
After successful stabilization of the hexameric state, we sought to verify that biological activity was maintained in mutant ClpC1 F444A . ClpC1 F444A ATPase activity was checked using an assay monitoring the fluorescence decrease at 340 nm, associated with NADH to NAD + conversion as the ADP formed by ClpC1 ATPase activity is reconverted into ATP by pyruvate kinase and phosphoenolpyruvate dehydrogenase [bib_ref] Development of high throughput screening methods for inhibitors of ClpC1P1P2 from Mycobacteria..., Fraga [/bib_ref]. As shown in [fig_ref] Figure 2: Effect of NPAs on ClpC1 activity [/fig_ref] , ClpC1 F444A displayed an increase in the ATPase activity in comparison to ClpC1 WT . An increase in ATPase activity was also observed for the SaClpC1 F436A mutant, which was explained by a shift in the resting state-active hexamer equilibrium towards the latter, which is consistent with the described size-exclusion chromatography experiments.
Although cyclomarin and ecumicin binding sites are located at the ClpC1 NTD and are therefore distant from the F444A mutation, we tested the functional consequences of cyclomarin and ecumicin binding on the mutant. Using saturating concentrations (10 μM), ecumicin binding to the WT results in a strong increase in ClpC1 ATPase activity, while cyclomarin does not. Curiously, a similar increase in ATPase activity is observed using the F444A mutant, showing that ecumicin effects do not exclusively result from the modulation of the ClpC1 resting state-hexamer equilibrium [fig_ref] Figure 2: Effect of NPAs on ClpC1 activity [/fig_ref]. An important in vivo function of ClpC1, if not the most relevant, is to associate with ClpP1P2 and target proteins for degradation by the protease. The roles of ClpC1 in this process are multiple, as it is responsible for substrate recognition, unfolding, and translocation but also for association with ClpP1P2 and its concomitant allosteric activation [bib_ref] Development of high throughput screening methods for inhibitors of ClpC1P1P2 from Mycobacteria..., Fraga [/bib_ref]. Because these processes are very hard to study independently, ClpC1 function is often evaluated indirectly by measuring protein target degradation, that is, the rate at which a protein is degraded by ClpP1P2 in association with ClpC1 [bib_ref] The active ClpP protease from M. tuberculosis is a complex composed of..., Akopian [/bib_ref]. Two substrates that are commonly used are FITC-casein (casein with a fluorescein fluorophore attached to lysine residues) and GFP, representing two classes of proteins. FITC-Casein, which lacks a welldefined tertiary structure, is usually used as a model for unfolded substrates, while the stable, beta-barrel containing GFP is used as a model for a structured protein. While in the Structure of the drug target Mycobacterium tuberculosis ClpC1 case of FITC-casein, protein degradation results in a net fluorescence increase as fluorescein quenching is reduced with degradation; in the case of GFPssra, degradation of the protein fluorophore results in a decrease in fluorescence [bib_ref] Development of high throughput screening methods for inhibitors of ClpC1P1P2 from Mycobacteria..., Fraga [/bib_ref]. As can be seen in [fig_ref] Figure 2: Effect of NPAs on ClpC1 activity [/fig_ref] , B and C, the F444A mutant is fully functional to catalyze the degradation of both FITC-casein and GFPssra by ClpP1P2, showing that this mutation does not impair ClpC1 enzymatic activity.
Reflecting different modes of action, the effects of cyclomarin and ecumicin in protein degradation by the ClpC1P1P2 complex are distinct. At the concentration used, cyclomarin is a moderate activator of FITC-casein degradation and a weak inhibitor of GFPssra degradation, whereas for ecumicin, we observe a mild activation of FITC-casein degradation and a strong inhibition of GFPssra degradation [fig_ref] Figure 2: Effect of NPAs on ClpC1 activity [/fig_ref] , B and C). The different effects of these two NPAs are also clear when they are used together. Indeed, cyclomarin is able to prevent both ATPase activation [fig_ref] Figure 2: Effect of NPAs on ClpC1 activity [/fig_ref] as well as GFP degradation inhibition induced by ecumicin [fig_ref] Figure 1: ClpC1 resting statehexamer equilibrium [/fig_ref] , showing that they are competing for a similar pocket but they induce distinct biochemical effects. The distinct structural consequences resulting from cyclomarin and ecumicin binding are also clear from other biophysical data. Differential scanning fluorimetry is a very useful method to monitor ligand binding to a given target. We have previously shown that arginine phosphate (ArgP) is able to stabilize the NTD, and we aimed to test the effects of cyclomarin and ecumicin on the NTD (7). Using Sypro orange as a fluorescence reporter, we obtained a Tm of 77.7 (±0.3) C and 83.6 (±0.3) C, for the apo and ArgP-bound NTD. These values are higher than the ones we previously reported using intrinsic tryptophan fluorescence as a reporter that were 69 C and 79 C for the WT and ArgP-bound NTD [bib_ref] The antibiotic cyclomarin blocks arginine-phosphate-induced millisecond dynamics in the N-terminal domain of..., Weinhäupl [/bib_ref]. Quite striking was the difference observed between cyclomarin and ecumicin. While cyclomarin binding resulted in a very strong stabilization of NTD, with a calculated Tm of 91.5 (±0.9) C, ecumicin, on the contrary, did not stabilize the domain and in fact lead to a decrease in the Tm [fig_ref] Figure 2: Effect of NPAs on ClpC1 activity [/fig_ref].
## Structural characterization of mtbclpc1 by cryo-em
Having shown that the MD mutant (ClpC1 F444A ) is active, we attempted unsuccessfully to crystallize the resulting hexamer. We therefore set out to obtain structural information on ClpC1 via cryo-EM. To further stabilize the sample, considering that ATP is being continuously degraded even at low temperatures and that some AAA+ proteins have been shown to hydrolyze ATP analogs, we introduced two additional mutations in the double walker B motif (E288A and E626A). We hypothesized that these mutations allow nucleotide binding but prevent hydrolysis and therefore further stabilize the hexameric assembly.
After initial cryo-EM grid screening, a dataset was collected of the apo form of MtbClpC1 E288A/F444A/E626A (see Experimental procedures). Processing of the data resulted in a structure for apo ClpC1 determined to a resolution of 3.6 Å (see Experimental procedures, Figs. S3 and S4, and . As expected, the overall structure of apo MtbClpC1 E288A/F444A/ E626A consists of a hexamer composed of six ClpC1 subunits (A-F), with the D1 and D2 domains arranged in a ring that together forms a pore through which the substrate can be Structure of the drug target Mycobacterium tuberculosis ClpC1 actively translocated. In line with recent structures, including the recent cryo-EM structure of B. subtilis ClpC (21), the observed ClpC1 hexamer is not symmetric, as observed originally in the crystal structure of ClpC [bib_ref] Structure and mechanism of the hexameric MecA-ClpC molecular machine, Wang [/bib_ref] , and instead adopts an asymmetric spiral structure [fig_ref] Figure 3: The MtbClpC1 active hexameric structure [/fig_ref]. An important characteristic of our structure is the observation of a 23-residue long peptide visible in its central pore. The presence of substrates in the pore formed by D1 and D2 has been previously reported for other members of the family, but as no substrate has been added to our cryo-EM sample preparation, it was probably taken up and trapped inside of the inactive protein during purification [fig_ref] Figure 3: The MtbClpC1 active hexameric structure [/fig_ref]. This unexpected finding is nevertheless useful as it provides important details on the ClpC1 mechanism of action. Indeed, the substrate is bound by each of the two pore loops from subunits A to E in both D1 and D2, with the exception of pore loop 2 in D1 subunit E, forming a spiral along the substrate. Subunit E is positioned at the highest point and subunit A at the lowest. Subunit F is detached from the substrate in both D1 and D2, and the pore loops are not visible. Furthermore, all nucleotide-binding sites in D1 are occupied by ADP, while for D2, four sites are occupied by ADP (subunits B, C, D, and E), while two are in their apo state (subunits A and F) [fig_ref] Figure 3: The MtbClpC1 active hexameric structure [/fig_ref]. This organization is in line with the asymmetric disposition we referred to above and has been reported for other members of the family, particularly for structures obtained using cryo-EM [bib_ref] BacPROTACs mediate targeted protein degradation in bacteria, Morreale [/bib_ref] [bib_ref] Conformational plasticity of the ClpAP AAA+ protease couples protein unfolding and proteolysis, Lopez [/bib_ref] [bib_ref] Cryo-EM structures of the Hsp104 protein disaggregase captured in the ATP conformation, Lee [/bib_ref].
Unfortunately, several structural features are not visible in the ClpC1 apo structure, presumably due to flexibility or intrinsic disorder. One of these missing features are the LGF loops which, in line with other AAA+ ATPase structures, will likely only become structured upon interaction with the ClpP protease [bib_ref] Conformational plasticity of the ClpAP AAA+ protease couples protein unfolding and proteolysis, Lopez [/bib_ref]. Another unresolved feature is the MD, which connects D1 and D2 and carries the F444A mutation. The MD is presumably a key point for interactions with adapters or binding partners and is usually only visible in the resting state or when it is engaged with a binding partner (see MecA in ClpC1 or DnaK in the homologous ClpB). Although there is no high-resolution structure of the ClpC1 resting state, we presume that the overall organization of the resting state is similar to SaClpC for two reasons: (1) our previously reported small-angle X-ray scattering structure of the ClpC1 resting state overlays perfectly with the SaClpC cryo-EM structure of the resting state (Protein Data Bank, PDB: 6EM9/6EM8) and (2) the same residues appear to be involved in the stabilization of the resting state as shown by our mutation experiments involving F444. It is thus likely that the MD of ClpC1 in the resting state is properly folded to provide contacts between the different MDs, thereby stabilizing the resting state, and only becomes unfolded upon formation of the hexamer. Finally, the NTD and its adjacent 26 residue linker are not visible in our apo ClpC1 structure. The NTD is thought to have an important role in substrate recognition and targeting to the ClpC1 pore, and it is the binding site of several recently identified NPAs against Mtb. Similar to the MD, the NTD can only be seen in the resting state or in structures where an adapter or the substrate itself stabilizes its position. The isolated NTD is a well-folded globular domain and the fact that it is not visible in our apo ClpC1 structure presumably does not reflect on an intrinsic disorder of the domain but rather on the high flexibility of the linker and the resulting multiple orientations with respect to the D1/D2 domains.
## Cryo-em structures of mtbclpc1 bound to npas
Following successful structure determination of apo ClpC1, two additional datasets were collected on cryo-EM grids prepared after the addition of 30 μM cyclomarin or ecumicin to purified MtbClpC1 E288A/F444A/E626A (see Experimental procedures). Processing of the two datasets resulted in a structure for cyclomarin-bound ClpC1 determined to a resolution of 3.3 Å and two 3D classes of ecumicin-bound ClpC1 determined to resolutions of 4.3 Å and 8.6 Å, respectively (see Experimental procedures, Figs. S3 and S4, and .
The structure of cyclomarin-bound ClpC1 [fig_ref] Figure 2: Effect of NPAs on ClpC1 activity [/fig_ref] is virtually identical to the apo ClpC1 structure, with a calculated rmsd of 0.39 Å (over 3239 aligned C α -atoms). No differences in nucleotide binding or substrate interaction were observed. Similar to the apo ClpC1 structure, a 23-residue substrate is trapped in the ClpC1 pore and the NTDs, MDs, and LGF loops are invisible.
Processing of the MtbClpC1 dataset with added ecumicin resulted in 2D classes that included side-views revealing hints of the NTDs, present as a blurry sphere on top of the ClpC1 hexamer [fig_ref] Figure 3: The MtbClpC1 active hexameric structure [/fig_ref]. Ensuing hetero refinement using two 3D classes followed by final non-uniform (NU) refinements in cryoSPARC (https://cryosparc.com) (25) resulted in two distinct maps at resolutions of 4.3 Å and 8.6 Å. The first map, populated by roughly 60% of the particles, represents a structure identical to the ones observed for apo MtbClpC1 and cyclomarin-bound MtbClpC1. The second map, containing 40% of particles, has a significantly lower resolution but reveals three globular domains positioned on top of the D1 domains of the hexameric ClpC1 ring in an organization reminiscent of an NTD trimer observed in a hyperactive ClpB mutant bound to casein [bib_ref] Structural basis for substrate gripping and translocation by the ClpB AAA+ disaggregase, Rizo [/bib_ref]. Rigid-body fitting of the NTD-trimer taken from PDB 6OG3 [bib_ref] Structural basis for substrate gripping and translocation by the ClpB AAA+ disaggregase, Rizo [/bib_ref] in the low-resolution map results in a good fit [fig_ref] Figure 4: Structural basis of the ecumicin mechanism of action [/fig_ref] and provides a model where the antibiotic-binding sites in helix 1 and 5 on the NTD and the ArgP/substratebinding sites are facing towards the center of the ClpC1 hexameric ring [fig_ref] Figure 4: Structural basis of the ecumicin mechanism of action [/fig_ref]. The three remaining NTD domains as well as all flexible linkers connecting the NTD domains with the adjacent D1 domain are not visible on the map, most probably due to the inherent flexibility of the linker region. Nonetheless, the observed trimerization of the MtbClpC1 NTDs upon addition of ecumicin suggests that ecumicin can function by mimicking substrate binding.
# Discussion
It is well established that minor differences in protein structure may result in important functional differences. Therefore, even when structures of homologs are available, it is still important to obtain accurate structural information, particularly for drug development. For instance, while displaying high primary sequence homology to other ClpCs, MtbClpC1 displays unique mechanistic features. Indeed, contrary to its homologs, ClpC1 does not depend on activators as MecA and is fully competent to catalyze, together with ClpP1P2, the degradation of both unfolded (as casein) and folded (as GFPssra) substrates. In addition, MtbClpC1 is the sole target of the NPAs cyclomarin, lassomycin, ecumicin, or rufomycin.
In this study, we took advantage of a MD mutation (F444A) to stabilize the ClpC1 hexameric state. This mutation was chosen based on previous data obtained for SaClpC and our own results using MtbClpC1 that proved the existence of an equilibrium between a resting state and a hexameric form. Indeed, the change of an aromatic phenylalanine to an alanine resulted in a large shift in the size exclusion profile towards the size expected for a hexameric complex.
While the introduction of the F444A mutation shifts the equilibrium towards the hexameric state, it may potentially modify MtbClpC1's mechanism of action. For example, the first structure of a ClpC complex, the only one obtained using X-ray crystallography, was from an inactive mutant, unable to catalyze ATP hydrolysis and therefore protein degradation [bib_ref] Structure and mechanism of the hexameric MecA-ClpC molecular machine, Wang [/bib_ref]. We therefore biochemically characterized the F444A mutant with respect to ATPase activity and ClpP1P2-mediated protein degradation of unfolded and folded substrates and show that MtbClpC1 F444A is fully active, with increased specific activities versus the WT protein. This activation was not unexpected, considering that the equilibrium is shifted towards the active state, but it is worth mentioning that it is more moderate than the activation described for ClpC1 homologs with similar modifications in the MD. This fact likely reflects the functional differences between these homologs but nonetheless shows that the F444A mutant is fully able to catalyze unfolded and folded protein degradation in association with ClpP1P2.
One of our goals was to understand the mechanism of action of antibiotics targeting ClpC1. Therefore, an important question was whether the F444A mutation might influence the effects of cyclomarin and ecumicin. It is known that even though cyclomarin and ecumicin largely share their binding surfaces in the NTD, their binding results in different mechanistic consequences. In fact, as we have shown here and has been reported elsewhere [bib_ref] Mutation analysis of the interactions between Mycobacterium tuberculosis caseinolytic protease C1 (ClpC1)..., Jung [/bib_ref] , ecumicin results in a marked increase in ClpC1 ATPase activity, whereas cyclomarin fails to do so. Two other striking differences concern FITC-casein and GFPssra degradation in association with ClpP1P2. Although both cyclomarin and ecumicin lead to a mild activation of FITC-casein degradation, ecumicin is a much better inhibitor of GFPssra degradation.
Introduction of the F444A mutation and the shift towards a hexameric state does not abolish cyclomarin and ecumicin effects, and a similar profile of activation and inhibition is observed for all enzymatic activities tested. This does not support the recent proposition, based on protein chimeras, that cyclomarin mechanism of action depends only on a shift of the resting state-hexamer equilibrium [bib_ref] Toxic activation of an AAA+ protease by the antibacterial drug cyclomarin A, Maurer [/bib_ref]. If that was the case, the effect of cyclomarin should be abolished with the introduction of the mutant. Indeed, the same conclusion could be simply derived from the fact that the two antibiotic effects are dissimilar. If ecumicin and cyclomarin effects were exclusively dependent on resting state-hexamer equilibrium, we would expect their action to be the same in all activities measured-which is not the case. The cryo-EM structures we determined further corroborate our assumption that the F444A mutant is a valid model for an active ClpC1 hexamer. Supporting the validity of our approach, multiple structural features demonstrate that we have in fact a functional enzyme. First, the asymmetric hexameric structure we observe here has been observed for other members of the AAA+ family, including a very recent structure of SaClpC as well as ClpA [bib_ref] BacPROTACs mediate targeted protein degradation in bacteria, Morreale [/bib_ref] [bib_ref] Conformational plasticity of the ClpAP AAA+ protease couples protein unfolding and proteolysis, Lopez [/bib_ref]. This shows that the mutation we introduced does not impair hexamer formation, domain architecture, or nucleotide binding. Even more importantly, the unexpected presence of a peptide bound in the pore formed by the D1 and D2 domains and in close contact with the canonical tyrosine loops demonstrates that F444A ClpC1 is fully competent to bind and translocate protein substrates.
In addition to the apo MtbClpC1 cryo-EM structure, we determined additional structures after the addition of either cyclomarin or ecumicin antibiotics to purified MtbClpC1. Both cyclomarin-bound MtbClpC1 as well as one of the two 3D classes of ecumicin-bound MtbClpC1 are strikingly similar to the apo MtbClpC1 structure and have missing NTDs, which contain the binding sites for both cyclomarin and ecumicin. The role of the NTD in ClpC1's mechanism of action remains largely unknown, but data obtained with other AAA+ ATPases suggest that it may transfer the polypeptide substrate to D1 and the D1 pore. Curiously, truncating part of the ClpC1 NTD, residues 1 to 78, has been shown to paradoxically activate the refolding capability of ClpC1 [bib_ref] Mycobacterium tuberculosis ClpC1 N-terminal domain is dispensable for adaptor protein-dependent allosteric regulation, Marsee [/bib_ref]. Considering our data, it seems, however, unlikely that cyclomarin would induce a stable conformational change in the NTD, as for example, occurs with MecA binding in other species, where an additional ring adjacent to the D1 ring is clearly observed both in X-ray and cryo-EM structures.
The second 3D class of ecumicin-bound MtbClpC1, containing roughly 40% of the imaged particles, reveals three globular domains above the D1 N-termini, most likely corresponding to three NTDs. Additional densities corresponding to the other NTDs are not observed, strongly suggesting that they remain flexible. Interestingly, the position of the putative NTDs in our map of ecumicin-bound MtbClpC1 is identical to the structure of an NTD trimer described for a hyperactive Escherichia coli ClpB MD mutant engaged with casein [bib_ref] Structural basis for substrate gripping and translocation by the ClpB AAA+ disaggregase, Rizo [/bib_ref]. In this structure, obtained at much higher resolution (2.9 Å), the NTD trimer forms a substrate entrance channel, positioning the casein polypeptide above the translocation pore. Additionally, trimerization of the NTD of MtbClpC1 upon addition of ecumicin was recently demonstrated in a crystal structure of the ecumicin-ClpC1 NTD complex [bib_ref] Structure of the N-terminal domain of ClpC1 in complex with the antituberculosis..., Wolf [/bib_ref]. The relative position of the NTDs in the ecumicin: ClpC1 NTD trimer is similar to the NTD trimer in casein-bound E. coli ClpB and therefore fits equally well in our low-resolution map. Interestingly, contrary to the casein: ClpB NTD trimer, the orientation of the individual NPA-binding sites of the NTDs in the ecumicin: ClpC1 NTD trimer is facing outwards rather than inwards. However, the lower resolution of the ecumicin-bound MtbClpC1 map we obtained here (8.6 Å) does not allow us to accurately assess the ecumicin-binding site or the relative conformations of the NTD domains.
How can the structures we obtained help in understanding the mechanism of action of cyclomarin and ecumicin? As described previously, albeit binding to similar regions of the NTD, cyclomarin and ecumicin result in different mechanistic consequences. Isothermal titration calorimetry results as well as X-ray crystallographic studies have shown that only one molecule of cyclomarin binds to the NTD (7, 11), forming a bridge over a hydrophobic ridge dominated by the aligned phenyl rings of phenylalanines F2 and F80. It is still unclear if cyclomarin, like rufomycin, forms a covalent adduct with the NTD N-terminal methionine. Quite striking is the fact that cyclomarin binding results in a dramatic increase in the stability of the NTD, and this is reflected in an increase of circa 15 C in thermal stability as well as an important restriction of the domain dynamics induced by ArgP binding (7).
Ecumicin is a tridecamer depsipeptide with a larger scaffold than cyclomarin and an extended tail of three amino acids important for binding to the protein NTD (see [fig_ref] Figure 1: ClpC1 resting statehexamer equilibrium [/fig_ref]. Furthermore, it has a different binding stoichiometry where two ecumicin molecules are accommodated by one NTD [bib_ref] Structure of the N-terminal domain of ClpC1 in complex with the antituberculosis..., Wolf [/bib_ref]. Interestingly, in the crystal packing of the ecumicin:ClpC1 NTD complex, clusters of three dimers arrange themselves in a looser hexameric arrangement. Wolf et al. [bib_ref] Structure of the N-terminal domain of ClpC1 in complex with the antituberculosis..., Wolf [/bib_ref] suggested that the presence of two ecumicin molecules may help to stabilize this aggregate via hydrophobic binding. It is speculative to consider that a similar NTD trimer occurs in solution. Yet, the structure of ecumicin-bound ClpC1 shown here could represent such a complex, where ecumicin stabilizes the interaction between the three NTD subunits. Still, several mechanistical questions can be addressed here, particularly how the formation of an NTD trimer upon ecumicin binding can result in ATPase activation, inhibition of GFPssra degradation, and a mild activation of FITC-casein degradation. As described above, due to its high mobility, very little is known about NTD function. Nevertheless, two roles for the NTDs in ClpA have been proposed [bib_ref] The N-terminal substrate-binding domain of ClpA unfoldase is highly mobile and extends..., Ishikawa [/bib_ref]. One of the proposed roles is to work as an antenna to capture substrates via an initial weak interaction, transferring them to the second, stronger binding site for unfolding, and subsequent translocation to the D1-D2 pore. In this way, the NTDs would not be fundamental for catalysis but would make the processing of some challenging (folded) substrates more productive. This hypothesis is consistent with the observation that removal of half of the ClpC1 NTD appears not to affect firefly luciferase refolding by ClpC1 [bib_ref] Mycobacterium tuberculosis ClpC1: characterization and role of the Nterminal domain in its..., Kar [/bib_ref] or that casein degradation by ClpA [bib_ref] Functional domains of the ClpA and ClpX molecular chaperones identified by limited..., Singh [/bib_ref] is independent of the NTD. In contrast to these unfolded substrates, impairment of the NTD can drastically affect the degradation of (folded) GFPssra by ClpA [bib_ref] Conserved residues in the N-domain of the AAA+ chaperone ClpA regulate substrate..., Erbse [/bib_ref]. Another putative role for the NTD would be to function as an "entropic brush" in order to prevent unspecific interactions with the pore-binding sites. Resulting from these two proposed roles, impairment of NTD function should likely affect the degradation of folded proteins like GFP more than unfolded proteins such as FITC-casein. The results presented here seem to corroborate this hypothesis, since a strong inhibition of GFPssra degradation and a mild activation of FITC-casein degradation by MtbClpC1 are observed after the addition of ecumicin. However, it is still not clear how the formation of NTD oligomers may explain the strong ATPase activation observed after the addition of ecumicin.
In summary, we provide here the first high-resolution structure of MtbClpC1, an important drug target, in its apo, cyclomarin-, and ecumicin-bound states. The ecumicin-bound ClpC1 structure presented here allows us to suggest a model for antibiotic action based on the formation of stable ecumicin-bound NTD oligomers, which would block folded substrate degradation by the ClpC1P1P2 complex.
## Experimental procedures biochemistry
Mutations F444A, E288A, and E626A were introduced into a pet20-ClpC1 plasmid using an NZYMutagenesis kit. The primers used are listed in . Mutations were confirmed by DNA sequencing (Eurofins Scientific) using the T7 forward primer. ClpC1 and mutants were expressed and purified as previously described (7). ClpC1 ATPase activity and FITCcasein and GFPssra degradation by ClpC1P1P2 complex were measured as described previously [bib_ref] Development of high throughput screening methods for inhibitors of ClpC1P1P2 from Mycobacteria..., Fraga [/bib_ref]. Differential scanning fluorimetry measurements were executed in a Pierce Light Cycler 96 using a 34 to 95 C temperature ramp using 5 μM NTD, 12.5 μM cyclomarin and ecumicin, and 166 μM ArgP in Hepes pH 7.4 50 mM NaCl 100 mM. The T m was calculated using the first derivate of the respective curves.
## Cryo-em sample preparation
Cryo-EM grids of Apo Mtb ClpC1 (MtClpC1), MtClpC1 in complex with cyclomarin, and MtClpC1 in complex with ecumicin were vitrified using a Vitrobot Mark IV (FEI). Quantifoil Cu/Rh 1.2/1.3 300 mesh grids were previously glow-discharged for 30 s at 15 mA. Aliquots of 3 μl of the different samples were added onto the grids, blotted for 3 s at 4 C and 95% humidity, and plunged into liquid ethane.
## Cryo-em data collection
Screening and data acquisition of all samples were performed using a 200 kV FEI Talos Arctica equipped with a Falcon III direct electron detector at the Centro Nacional de Biotecnología cryo-EM facility. A total of 944 movies of Apo MtClpC1, 2840 movies of MtClpC1 + cyclomarin, and 1520 movies of MtClpC1 + ecumicin were acquired at a nominal magnification of 120,000× (corresponding to a pixel size of 0.855 Å/pixel), with a defocus range of −1.2 to −3.1 μm. Movies were fractionated to 60 frames with a total exposure of 40 s (Apo MtClpC1 and MtClpC1 + ecumicin) or 30 s (MtClpC1 + cyclomarin), with a total dose per movie of 34.3 e − /Å 2 (Apo MtClpC1), 36.9 e − /Å 2 (MtClpC1 + cyclomarin), and 32.2 e − /Å 2 (MtClpC1 + ecumicin).
## Cryo-em data processing
Collected movies (Apo dataset: 944, cyclomarin-bound dataset: 2840, ecumicin-bound dataset: 1520) were motion corrected and dose weighted using MotionCor2 [bib_ref] MotionCor2: anisotropic correction of beam-induced motion for improved cryo-electron microscopy, Zheng [/bib_ref]. Further data processing was performed using cryoSPARC 3.3.1 [bib_ref] The N-terminal substrate-binding domain of ClpA unfoldase is highly mobile and extends..., Ishikawa [/bib_ref]. Initial CTF estimation on the imported aligned and dose weighted micrographs was performed using Patch CTF estimation. For each dataset, particle picking was performed using crYOLO [bib_ref] SPHIRE-crYOLO is a fast and accurate fully automated particle picker for cryo-EM, Wagner [/bib_ref]. Imported particle stacks were extracted using a box size of 500 pixels downsampled to 250 pixels (2× binned), corresponding to a pixel size of 1.71 Å/pixel. Extracted particle stacks were cleaned using several rounds of iterative 2D classification and 2D class selection. Cleaned particle stacks were used as an input for Ab Initio model generation and subsequent NU 3D refinement. For Apo and cyclomarin-bound datasets, a final NU 3D refinement was performed on unbinned particles from the prefinal 3D refinement, extracted using a box size of 500 pixels, corresponding to a pixel size of 0.885 Å/pixel. For the Apo dataset, the final NU refinement was performed using optimized pergroup CTF parameters, whereas for the cyclomarin-bound dataset, both per-group CTF parameters as well as perparticle defocus were optimized. For the ecumicin-bound dataset, particles selected after 2D classification were used as an input for Ab Initio model generation and hetero refinement using two classes, and each class was separately refined using a final NU 3D refinement. Maps for the apo and cyclomarin-bound datasets as well as class 1 of the ecumicinbound dataset were postprocessed using DeepEMhancer (https://github.com/rsanchezgarc/deepEMhancer) for model building purposes [bib_ref] DeepEMhancer: a deep learning solution for cryo-EM volume post-processing, Sanchez-Garcia [/bib_ref].
## Model building and refinement
An initial model was generated based on the AlphaFold2 (34) prediction for monomeric ClpC1 (https://alphafold.ebi.ac. uk/entry/P9WPC8). First, parts with a low per-residue confidence score were removed from the AlphaFold2 ClpC1 monomer model. Next, six copies of the trimmed ClpC1 model were manually placed in the DeepEMhancer-sharpened Apo ClpC1 map using rigid-body fitting in USCF Chimera [bib_ref] UCSF Chimera -a visualization system for exploratory research and analysis, Pettersen [/bib_ref] , followed by automatic molecular dynamics-based flexible fitting using NAMDINATOR [bib_ref] Namdinator -automatic molecular dynamics flexible fitting of structural models into cryo-EM and..., Kidmose [/bib_ref] and subsequent manual building in Coot [bib_ref] Features and development of Coot, Emsley [/bib_ref]. The resulting structure was further refined in the cryoSPARC-sharpened Apo ClpC1 map in
## Mutation
Forward primer Reverse primer Phenix, using global minimization, local grid search, ADP refinement, secondary structure and Ramachandran restraints, noncrystallographic symmetry constraints, and a nonbonded weight parameter of 300. The apo structure of Mtb ClpC1 was subsequently placed in the DeepEMhancer-sharpened ClpC1 + cyclomarin and ClpC1 + ecumicin, class 1 maps using rigid-body fitting in UCSF Chimera, followed by manual building in Coot. Next, several cycles of refinement using the cryoSPARC-sharpened maps were performed in Phenix, using global minimization, local grid search, ADP refinement, secondary structure and Ramachandran restraints, noncrystallographic symmetry constraints, and a nonbonded weight parameter of 300.
For ClpC1 + ecumicin class 2, the ClpC1 + ecumicin class 1 structure was rigid-body fitted in the map using UCSF Chimera, together with three copies of an available structure of the Mtb ClpC1 NTD in complex with Ecumicin PDB-ID: 6PBS (15), aligned with the E. coli ClpB NTD trimer found in PDB-ID 6OG3 [bib_ref] Structural basis for substrate gripping and translocation by the ClpB AAA+ disaggregase, Rizo [/bib_ref].
## Data availability
Atomic coordinates and associated structure factors have been deposited in the PDBs 8AUA, 8AUV, and 8AUW.
Supporting information-This article contains supporting information.
[fig] Figure 1: ClpC1 resting statehexamer equilibrium. A, model of the MtbClpC1 resting state created by Swiss-model using the Staphylococcus aureus ClpC resting state as a template (PDB: 6EM9) and of the MtbClpC1 theoretical active hexameric state using the crystal structure of Bacillus subtilis ClpC (PDB 3PXI). The protomers P1 to P6 are shown in rainbow colors. P1, pink; P2, dark blue; P3, light blue; P4, green; P5, yellow; and P6, dark red. Residue F444 is marked in light red. B, extract of the pairwise sequence alignment of MtbClpC1 and S. aureus ClpC. The conserved F444 residue is marked in red. C, size exclusion profile of WT MtbClpC1 (black) and the two mutants MtbClpC1 F444A (blue) and MtbClpC1 F444A E288A E626A (red) in the absence and presence of ATP. D, negative stain EM images of the hexameric ClpC1 F444A mutant. Mtb, Mycobacterium tuberculosis; NPAs, natural product antibiotics; PDB, Protein Data Bank. [/fig]
[fig] Figure 2: Effect of NPAs on ClpC1 activity. A, ATPase activity of WT ClpC1 and the F444A mutant in the apo state and bound to natural product antibiotics. Degradation of FITC-casein (B) and GFPssra (C) by the ClpC1P1P2 complex of apo ClpC1 and the F444A mutant in the presence of natural product antibiotics. D, cyclomarin binding prevents the activation of ClpC1 ATPase activity promoted by ecumicin binding. NPAs, natural product antibiotics. [/fig]
[fig] Figure 3: The MtbClpC1 active hexameric structure. A, cryo-EM map of the apo MtbClpC1 hexamer bound to a substrate peptide in top, side view and with the bound substrate visible (without the protomers P1 and P6). Individual protomers are colored independently and labeled P1 to P6. B, occupation of the nucleotide-binding pockets in the MtbClpC1 hexamer with ADP (in the D1 domain orange, in the D2 domain red). From left to right hexamer in top view (substrate entry pore), side view, and bottom view (interface with ClpP1P2). C, cartoon image of the nucleotide occupation in the D1 and D2 domains and the attachment of the pore loops in D1 and D2 to the substrate. P6 is detached from the substrate in both D1 and D2. D, top view of the ClpC1-bound substrate with pore loops attached in the typical spiral arrangement and the pore loop of P6 detached. Mtb, Mycobacterium tuberculosis. [/fig]
[fig] Figure 4: Structural basis of the ecumicin mechanism of action. A, class 1 (rainbow) and 2 (white) maps of ecumicin-bound MtbClpC1 with the percentage of particles found in each state. B, class 2 ecumicin-bound ClpC1 map with the fitted model of the MtbClpC1 NTD (created by alpha fold) and the hexameric high-resolution structure of ecumicin-bound MtbClpC1 (rainbow cartoon). C, model of the ecumicin-bound hexameric ClpC1 and three visible NTD domains in top, side, and substrate visible view (without P1 and P6). Protomers P1 to P6 are colored individually, because of the low-resolution map and the invisibility of the linker region, NTDs cannot be assigned to specific protomers and are therefore colored in grayscale. Mtb, Mycobacterium tuberculosis; NTD, N-terminal domain. [/fig]
|
Effect of Financial Stress and Positive Financial Behaviors on Cost-Related Nonadherence to Health Regimens Among Adults in a Community-Based Setting
IntroductionLittle is known about the role of positive financial behaviors (behaviors that allow maintenance of financial stability with financial resources) in mitigating cost-related nonadherence (CRN) to health regimens. This study examined the relationships between positive financial behaviors, financial stress, and CRN.MethodsData came from the 2011 Speak to Your Health! Community Survey (n = 1,234). Descriptive statistics were computed to examine financial stress and CRN, by chronic condition and health insurance status. We used multivariate logistic regression models to examine the relationship between positive financial behaviors and financial stress and their interaction on a composite score of CRN, controlling for health insurance status, educational level, age, marital status, number of chronic conditions, and employment status.ResultsThirty percent of the sample engaged in CRN. Participants reported moderate financial stress (mean, 13.85; standard deviation [SD] = 6.97), and moderate positive financial behavior (mean, 8.84; SD = 3.24). Participants with employer-sponsored insurance, Medicaid, Medicare, the Genesee Health Plan, high blood pressure, asthma, and diabetes had the highest proportion of CRN. The relationship between financial stress and CRN was not significantly different between those who reported lower versus higher levels of positive financial behavior (P = .32). Greater financial stress was associated with a greater likelihood of CRN (odds ratio [OR] = 2.49; 95% confidence interval [CI], 2.08-2.99). Higher level of positive financial behavior was associated with a lower likelihood of CRN (OR = 0.80; 95% CI, 0.67-0.94).ConclusionFinancial literacy as a means of promoting positive financial behavior may help reduce CRN. An intervention strategy focused on improving financial literacy may be relevant for high-risk groups who report high levels of financial stress. Facilitating the development of a county health coverage plan with data from a community-based health survey.
# Introduction
One of 4 Americans reports financial difficulty in paying medical bills [bib_ref] Adherence influencing factors -a systematic review of systematic reviews, Mathes [/bib_ref] ; this difficulty has significant public health implications, especially for the 50% of the population that is managing chronic illness. Seven systematic reviews concluded that several factors influence adherence to treatment, but cost to the patient is one that demonstrates a consistent negative effect [bib_ref] Cost-related medication underuse among chronically ill adults: the treatments people forgo, how..., Piette [/bib_ref]. Nearly 18% of chronically ill Americans report underusing medications and delaying or not fulfilling therapeutic recommendations because of cost [bib_ref] Comparing adherence and persistence across 6 chronic medication classes, Yeaw [/bib_ref] , which is referred to as cost-related nonadherence (CRN) [bib_ref] Comparing adherence and persistence across 6 chronic medication classes, Yeaw [/bib_ref] and varies by therapeutic class across chronic therapies [bib_ref] Frequency of and reasons for medication non-fulfillment and non-persistence among American adults..., Mchorney [/bib_ref]. Fifty-six percent of American adults with common chronic dis-eases self-report nonfulfillment of medication as a result of financial hardship [bib_ref] Patient perceptions of asthma-related financial burden: public vs. private health insurance in..., Patel [/bib_ref].
Health insurance coverage is a strong predictor of financial burden [bib_ref] Why health insurance literacy matters, Levitt [/bib_ref]. Nearly half of Americans have literacy challenges with health insurance and pay more for health care out of pocket because of these challenges, despite improvements as a result of the Affordable Care Act (ACA) [bib_ref] Poorer financial and health literacy among community-dwelling older adults with mild cognitive..., Han [/bib_ref].
Although health literacy and health insurance literacy are commonly discussed as integral for individuals to have the capacity to obtain, process, and understand basic health information or services and health insurance, financial literacy in the context of health has received little attention. Financial literacy is a set of skills and knowledge that allows individuals to make informed decisions with their financial resources [bib_ref] Low-income consumer's lack of financial literacy could impede use of ACA coverage..., Bauhoff [/bib_ref] , and it is associated with more frequent engagement in health-promoting behaviors [bib_ref] The impact of health and financial literacy on decision making in communitybased..., James [/bib_ref] [bib_ref] Correlates of health and financial literacy in older adults without dementia, Bennett [/bib_ref] [bib_ref] Assessing the relationship between neighborhood factors and diabetes related health outcomes and..., Smalls [/bib_ref]. Studies show that social determinants of health that contribute to financial burden correlate with CRN [bib_ref] Treat or eat: food insecurity, cost-related medication underuse, and unmet needs, Berkowitz [/bib_ref] [bib_ref] Clinical outcomes and incremental costs from a medication adherence pilot intervention targeting..., Ryan [/bib_ref]. Therefore, financial burden may be experienced in the context of a growing concern for financial insecurity and may not be exclusively health-related. Given the role that cost to the patient plays in adherence to therapeutic regimens, improving financial literacy to influence positive financial behaviors (behaviors that allow individuals to maintain financial stability with their financial resources) may have implications for CRN, and may be a necessary adjunct to policy reforms.
Few interventions have aimed to mitigate CRN beyond reducing out-of-pocket costs, which have shown modest improvements in health status [bib_ref] Using a community-based health survey as a tool for informing local health..., Kruger [/bib_ref]. Whether positive financial behavior is protective of CRN has not been explored and may have implications for behavioral interventions to promote financial literacy, especially among people who have chronic illnesses.
We examined financial stress and CRN by type of chronic condition and health insurance status in a community-dwelling sample in Michigan. We also examined the relationship between financial stress, positive financial behaviors, and CRN, testing the hypothesis that the relationship between financial stress and CRN is different between people who report lower numbers versus higher numbers of positive financial behaviors.
# Methods
## Study sample
Data came from the 2011 wave of the Speak to Your Health! Community Survey administered to residents of Genesee County, Michigan. Genesee County is among the most economically disadvantaged counties in Michigan [bib_ref] Challenges and lessons learned in developing a community-based health survey, Shirey [/bib_ref]. We collected data from October 2011 through March 2012. The survey's purpose was to monitor and understand local health concerns to inform efforts to improve the health of Genesee County communities and was developed by the team through a community-based participatory research process [bib_ref] Neighborhood social conditions mediate the association between physical deterioration and mental health, Kruger [/bib_ref]. All study procedures were approved by the University of Michigan Institutional Review Board.
Random samples of households were drawn from 129 Genesee County residential census tracts in an attempt to include at least 20 residents from each tract in Flint, Michigan, and 10 residents in suburban and rural census tracts. Inclusion criteria included being at least aged 18 and a resident of Genesee County. More details on the Speak to Your Health! Survey are available elsewhere [bib_ref] Neighborhood social conditions mediate the association between physical deterioration and mental health, Kruger [/bib_ref]. From 9,944 telephone listings, 1,234 respondents provided complete data. Inability to contact participants because of telephone issues totaled 10.9% (n = 1,088) of all attempted records. The response rate was 25% (1,234 of 4,936). Telephone numbers that were not in service or were never answered after 10 attempts (not including answering machines) were not included in response rate calculations. The complete and analytic samples were similar in age, years of education, sex, and body mass index (data not shown). All participants provided informed consent before data collection. Professional survey staff conducted a 25minute computer-aided telephone interview (CATI) with randomly selected respondents.
## Measures
The outcome of interest in this study was CRN, a composite binary measure (yes/no) of a positive response to at least 1 of 2 items. The first item was not seeing a doctor because of cost, which was measured by asking participants, "Was there a time during the last 12 months when you needed to see a doctor, nurse, or other health professional but could not because of cost?" The second item was not filling a prescription because of cost, which was measured by asking participants, "Was there a time during the last 12 months when you needed to fill a prescription but could not because of cost?".
Participants were asked a series of 9 questions to determine level of financial stress and use of positive financial behaviors. Items included whether participants set aside money from each pay period for savings, have trouble sleeping because of financial problems, are concerned because they cannot afford health insurance, and only spend what they can afford. All items were assessed on a 5point Likert scale (1 being strongly disagree to 5 being strongly agree) [bib_ref] An increase in economic adversity is associated with poorer self-reported physical and..., Kruger [/bib_ref] [bib_ref] Out-of-pocket health care expenditures, by insurance status, Catlin [/bib_ref].
To determine whether participants had any of 8 prevalent chronic conditions that require routine medical management (high blood pressure, heart disease, cancer, diabetes, asthma, sarcoidosis, sickle cell anemia, and lupus), participants were asked if they had ever been diagnosed by a doctor as having any of the conditions (yes/no). We developed this list on the basis of conditions with the highest overall health burdens and the interests of communitybased organization partners. To assess the type of health insurance participants had, they were asked if they had a health insurance policy through the following mechanisms: employersponsored, nonemployer/private, Medicaid, Medicare, Genesee Health Plan, other, and no insurance coverage (yes/no).
Analyses were conducted using SAS version 9.4 (SAS Institute, Inc). We used a principal-axis factor analysis with a varimax rotation for the 9 survey items related to finances to produce 2 orthogonal factors of interest (eg, financial stress and positive financial behaviors). Question items with factor loadings greater than 0.50 were included in a factor. Internal consistency was evaluated with the Cronbach α statistic for each of the final 2 orthogonal factors produced. Health-related financial stress and positive financial behavior items were both summed to create a score. The range of scores for financial stress was 1 to 30, with 1 indicating low stress and 30 indicating high stress. The range of scores for positive financial behaviors was 1 to 15, with 1 indicating low level of positive financial behaviors and 15 indicating high level of positive financial behaviors.
Descriptive statistics were computed for all demographic characteristics to examine the frequency of CRN and financial stress by chronic condition and type of health insurance. Multiple variable logistic regression analyses were used to examine the independent association between financial stress and positive financial behaviors (main independent variables) and CRN (dependent variable). Both of the variables for financial stress and positive financial behaviors were centered at the mean to standardize values and were included in the model as an interaction term to test the hypothesis that the relationship between financial stress and CRN is different between those who report lower levels of positive financial behaviors and those who report higher levels of positive financial behaviors. All models were adjusted for health insurance status, educational level, age, marital status, number of chronic conditions, and employment status.
# Results
The mean age of the sample (N = 1,234) was 53.5 years (standard deviation [SD] = 15.2 y), 73% were female, 47% were married, and 70% reported their race/ethnicity as white [fig_ref] Table 1: Demographic and Clinical Characteristics of the Sample [/fig_ref]. Ninety percent of the sample was employed, and 30% had an educational attainment of college or more. Nearly half of participants reported having employer-sponsored health insurance coverage and high blood pressure. Twenty percent of participants did not see a doctor, and 24% did not fill a prescription because of cost, resulting in a 30% prevalence of CRN.
Inter-item correlations among the 9 items that determined levels of financial stress and positive financial behaviors ranged from 0.07 to 0.83 [fig_ref] Table 2: Characteristics of Financial Items From Exploratory Factor Analysis, Speak to Your Health!... [/fig_ref]. Financial stress had strong reliability (6 items; Cronbach α = 0.89), and positive financial behaviors had adequate reliability (3 items; Cronbach α = 0.63). Participants reported a mean financial stress score of 13.85 (SD = 6.97) indicating moderate financial stress, and a mean score of 8.84 (SD = 3.24) for positive financial behaviors, indicating moderate positive financial behaviors. Financial stress and positive financial behaviors were not significantly correlated (r = 0.31). [fig_ref] Figure 1: Mean level of financial stress, by chronic condition and health insurance status [/fig_ref] shows mean financial stress by number and type of chronic conditions and health insurance status. Moderate financial stress was evident among participants reporting diabetes, asthma, sarcoidosis, sickle cell anemia, lupus, 2 or more chronic conditions, coverage by the Genesee Health Plan, and no insurance coverage. Multiple variable logistic regression analyses indicated that the interaction of financial stress and positive financial behaviors was not significant (P = .32), so it was dropped from the final model. Greater financial stress was associated with a greater likelihood of CRN behavior (odds ratio [OR] = 2.49; 95% confidence interval [CI], 2.08-2.99). Higher levels of positive financial behavior were associated with a lower likelihood of CRN (OR = 0.80; 95% CI, 0.67-0.94) [fig_ref] Table 3: Association Between Positive Financial Behaviors and Financial Stress on a Composite Score... [/fig_ref].
## Preventing chronic disease
## Preventing chronic disease
# Discussion
Many factors motivate adherence to a therapeutic health regimen, and this analysis focused on the role of cost as a factor. To our knowledge, this is the first study to examine the association between positive financial behaviors and CRN. Positive financial behaviors were moderate in this study of community-dwelling adults in an economically disadvantaged county in Michigan. The relationship between financial stress and CRN was not significantly different between those who reported lower versus higher levels of positive financial behavior in our sample; therefore, our hypothesis was not supported. This relationship requires further investigation in other samples. We found, however, that regardless of health insurance, higher levels of positive financial behaviors were associated with a lower likelihood of CRN, suggesting that positive financial behaviors may influence CRN. Our findings show an association between behavioral practices related to finances that influence CRN, whereas other work has shown that higher levels of health and financial literacy can promote health through improved health care decision making and behavior [bib_ref] Correlates of health and financial literacy in older adults without dementia, Bennett [/bib_ref] [bib_ref] Assessing the relationship between neighborhood factors and diabetes related health outcomes and..., Smalls [/bib_ref].
Researchers that found effects of financial literacy on health and health care decision making investigated financial literacy through measures that assess comprehension of financial concepts and numeracy skills based on mathematical computation [bib_ref] The impact of health and financial literacy on decision making in communitybased..., James [/bib_ref] [bib_ref] Correlates of health and financial literacy in older adults without dementia, Bennett [/bib_ref] [bib_ref] Assessing the relationship between neighborhood factors and diabetes related health outcomes and..., Smalls [/bib_ref]. In our study, we examined a set of behaviors for effective financial management in the interest of long-term financial well-being that include saving, spending within means, and paying off debt. Our findings expand our understanding of the influences of financial literacy on health by considering functional financial literacy and a set of skills and behaviors related to financial management that may also be protective against the negative consequences of health-related financial burdens on health behaviors. This is different from health literacy or health insurance literacy, which may encompass knowledge, behaviors, and self-efficacy around general or specific areas of health or health insurance.
Improving financial literacy as an intervention strategy as a means to promote positive financial behavior to prevent CRN may be especially relevant for certain high-risk groups. Although we found an association between financial stress and CRN regardless of health insurance status and number of chronic conditions, our descriptive findings and other work (3) suggest that attention to financial literacy may be relevant for certain conditions in which levels of both financial stress and CRN are high: high blood pressure, asthma, and diabetes. The adequate prevention of adverse events for high blood pressure, asthma, and diabetes through a medication regimen can have implications for the development of severe comorbidity and poor long-term health, as well as in-creased financial burden. We are not aware of any evidence-based self-management programs for any of these conditions that incorporate skills training in financial literacy to promote positive financial behaviors or have an emphasis on reducing cost-related barriers to prevent CRN. Although the chronic conditions included in this study may differ by typical cost of treatment, this variance my affect individuals in terms of perceived financial stress based on the type of health insurance they have and what services are covered in their plan.
Although individuals with employer-sponsored health insurance in our sample reported the lowest levels of financial stress, rates of CRN were similar for those with local insurance options for lowincome families (Genesee Health Plan) and those with no insurance, both of whom reported high levels of financial stress. Between 2007 and 2010, per-person out-of-pocket spending grew most rapidly for people primarily covered by employer-sponsored insurance because of a greater shift to high-deductible plan offerings by employers [bib_ref] An early look at changes in employer-sponsored insurance under the Affordable Care..., Blavin [/bib_ref] , which may encourage CRN. To date, no significant changes as a result of the ACA have been noted on consumer behavior among those with employer-sponsored insurance, suggesting that strategies to improve financial literacy may also be relevant for this group to prevent CRN.
Our study has several limitations. This was a secondary analysis of available survey data. Lack of household income data limited our interpretation of results to perceived financial burden. Median household income in Genesee County is $42,089, which is lower than state and national averages, suggesting economic burden in the region (25). Items used to generate constructs of financial stress and positive financial behaviors were based on available data. The items used may not capture all aspects of either construct, which may increase their content validity. Nevertheless, the measures have adequate reliability and capture key aspects of the construct, and the results were in the hypothesized direction. Adherence is also influenced by factors beyond cost; however, we did not have measures in our survey to examine other factors that influence adherence. We also did not have access to data to verify the CRN reported in this study against pharmacy claims or medical records, which would strengthen confidence in self-reported data. However, the rates of CRN reported in this study align with national averages and those reported elsewhere [bib_ref] Patient perceptions of asthma-related financial burden: public vs. private health insurance in..., Patel [/bib_ref] [bib_ref] Why health insurance literacy matters, Levitt [/bib_ref] , so the selfreported nature of the data was not a major concern for our confidence in the results. Another limitation was that participants were from one economically disadvantaged county in Michigan. Furthermore, women and older adults were overrepresented in the sample; as a result, findings may not be generalizable nationally or to Michigan as a whole, but they may be generalizable to other economically disadvantaged communities. Future work that considers replicating these analyses with nationally representative PREVENTING CHRONIC DISEASE www.cdc.gov/pcd/issues/2016/16_0005.htm - Centers for Disease Control and Prevention data would be useful. Finally, the data for the study were collected when the ACA was first passed. Rates of the provision and type of health insurance may have shifted since the enactment of ACA reforms in the region. However, the Genesee Health Plan was available before the ACA to provide basic health insurance for uninsured, low-income adults (26); therefore, the differences in the rate of the insured versus uninsured before and after the ACA may not differ. Although we did not have information for types of health insurance, our results reflect current trends of financial stress based on cost sharing that are consistent with the data reported here.
Despite these limitations, this study has implications for behavioral interventions. Policy interventions may intend to reduce financial stress from medical bills, but they do not provide the requisite skills for individuals and families to understand and therefore effectively manage the financial elements of their health care. Literacy and numeracy may have significant influence on how people choose cost-effective health plans and manage out-of-pocket expenses. These skills may be especially salient for those who manage chronic diseases, which require lifetime management with a therapeutic regimen and routine interface with the health care system. Building health literacy around the importance of adherence is imperative. Because financial literacy may protect against CRN, behavioral interventions for chronic diseases that teach and facilitate self-management skills may also consider skills training in health-related financial literacy. This may be a vital direction for intervention design, because financial stress was predictive of CRN despite health insurance status and number of chronic conditions. Improving financial literacy to promote positive financial behavior may be relevant for certain high-risk groups who report high levels of financial stress, such as those managing high blood pressure, asthma, or diabetes, or who have health insurance plans with high or variable cost sharing.
# Tables
[fig] Figure 1: Mean level of financial stress, by chronic condition and health insurance status. Mean level of financial stress is a composite score based on 6 items; scores ranged from 1 to 30, with 1 indicating low levels of stress and 30 indicating levels of high stress. Speak to Your Health! Community Survey, 2011. [/fig]
[fig] Figure 2 shows: CRN by chronic condition and health insurance. The highest proportion of individuals who reported CRN were those who had employer-sponsored health insurance, Medicaid, Medicare, the Genesee Health Plan, high blood pressure, asthma, or diabetes. [/fig]
[fig] Figure 2: Cost-related nonadherence behaviors, by chronic condition and health insurance status. Cost-related nonadherence is a composite binary measure of a positive response to 1 of 2 cost-cutting behaviors with the treatment regimen. Speak to Your Health! Community Survey, Michigan, 2011. [/fig]
[table] Table 1: Demographic and Clinical Characteristics of the Sample (N = 1,234), Speak to Your Health! Community Survey, Michigan, 2011 a The opinions expressed by authors contributing to this journal do not necessarily reflect the opinions of the U.S. Department of Health and Human Services, the Public Health Service, the Centers for Disease Control and Prevention, or the authors' affiliated institutions.Table 1. Demographic and Clinical Characteristics of the Sample (N = 1,234), Speak to Your Health! Community Survey, Michigan, 2011 a Values expressed as no. (%), unless otherwise indicated. Not all categories add to total, because some survey respondents did not answer all questions (percentages based on number of respondents who answered the question). b Respondents could choose more than one answer. c Twenty percent of participants did not see a doctor and 24% did not fill a prescription because of cost, resulting in a 30% prevalence of cost-related nonadherence in this sample. The opinions expressed by authors contributing to this journal do not necessarily reflect the opinions of the U.S. Department of Health and Human Services, the Public Health Service, the Centers for Disease Control and Prevention, or the authors' affiliated institutions. www.cdc.gov/pcd/issues/2016/16_0005.htm • Centers for Disease Control and Prevention [/table]
[table] Table 2: Characteristics of Financial Items From Exploratory Factor Analysis, Speak to Your Health! Community Survey, Michigan, 2011My financial situation is much worse this year than it was in the previous year.I do not know how I will be able to support myself in the next 12 months.How difficult is it for you to live on your total household income right now? (% high difficulty) Abbreviation: NA, not applicable.The opinions expressed by authors contributing to this journal do not necessarily reflect the opinions of the U.S. Department of Health and Human Services, the Public Health Service, the Centers for Disease Control and Prevention, or the authors' affiliated institutions. [/table]
[table] Table 3: Association Between Positive Financial Behaviors and Financial Stress on a Composite Score of Cost-Related Nonadherence, Speak to Your Health! Community Survey, Michigan, 2011 a Models adjusted for the provision of health insurance, education, age, marital status, number of chronic conditions, and employment. The opinions expressed by authors contributing to this journal do not necessarily reflect the opinions of the U.S. Department of Health and Human Services, the Public Health Service, the Centers for Disease Control and Prevention, or the authors' affiliated institutions. www.cdc.gov/pcd/issues/2016/16_0005.htm • Centers for Disease Control and Prevention [/table]
|
Research on Hysteresis of Piezoceramic Actuator Based on the Duhem Model
To improve the modeling accuracy of piezoceramic actuator in the precision positioning system, the Duhem hysteretic model of the piezoceramic actuator was proposed. The paper used the polynomial function to approach the piecewise continuous function and (V) and (V) in the Duhem model, adopted recursive least squares algorithm and gradient correction algorithm to identify parameter , polynomial coefficients of and in the Duhem model, and established the nonlinear parametric model of the piezoceramic actuator. Contrasting the simulation results of recursive least squares algorithm and gradient correction algorithm, the modeling accuracy is 0.24% when adopting the recursive least squares algorithm, and the modeling accuracy is 0.11% when adopting the gradient correction method. The result showed that the gradient correction algorithm could meet the modeling accuracy better, and the structure of the algorithm is simple, adaptable, and easy to implement.
# Introduction
The piezoceramic actuator is a kind of ideal drive elements in the microdisplacement technology currently and has the advantage of high positioning accuracy, large driving force, and fast response speed. Since the hysteresis, nonlinear, and creep resistance, inherent the piezoceramic actuator, the repeatability and accuracy of microdisplacement mechanism decreased, and the transient response speed becomes slower. [bib_ref] Modeling and control of hysteresis in magnetostrictive actuators, Tan [/bib_ref] [bib_ref] Identification and control of dynamic modeling for piezoceramic actuator, Chen [/bib_ref] [bib_ref] Identification of dynamic hysteresis based on Duhem model, Chen [/bib_ref] [bib_ref] Tracking control of a biaxial piezoactuated positioning stage using generalized Duhem model, Lin [/bib_ref] [bib_ref] An efficient vector Preisach hysteresis model based on a novel rotational operator, Sutor [/bib_ref] [bib_ref] A modified Preisach hysteresis operator for the modeling of temperature dependent magnetic..., Sutor [/bib_ref] To decrease the impact of the nonlinearity and obtain better performance of the piezoceramic actuator, many researchers carried on an investigation in modeling and controlling of hysteresis nonlinear system. [bib_ref] New kind of generalized Preisach hysteresis model and its identification based on..., Liu [/bib_ref] [bib_ref] A neural networks based model for rate-dependent hysteresis for piezoceramic actuators, Dong [/bib_ref] [bib_ref] A modified Prandtl-Ishlinskii modeling method for hysteresis, Dong [/bib_ref] [bib_ref] Robust adaptive control of systems with hysteretic nonlinearities: a Duhem hysteresis modelling..., Feng [/bib_ref] [bib_ref] Hysteresis modeling of the grain-oriented laminations with inclusion of crystalline and textured..., Baghel [/bib_ref] [bib_ref] Elastic, piezoelectric, and dielectric properties of Ba(Zr 0.2 Ti 0.8 )O 3..., Xue [/bib_ref] [bib_ref] Iterative compensation for hysteresis effects in positioning and tracking problems, Tseng [/bib_ref] Al Janaideh et al. [bib_ref] An analytical generalized Prandtl-Ishlinskii model inversion for hysteresis compensation in micropositioning control, Janaideh [/bib_ref] analyzed the inverse model of generalized Prandtl-Ishlinskii (PI) model to compensate the hysteresis nonlinearities of smart actuators. The experimental results verified that the inverse model of generalized PI model could be conveniently applied as a feedforward compensator and saturated hysteresis in magnetostrictive and SMA actuators. Dong and Tan [bib_ref] A modified Prandtl-Ishlinskii modeling method for hysteresis, Dong [/bib_ref] proposed a modified PI modeling method for rate-independent hysteresis in piezoelectric actuators and introduced a generalized backlash operator as the elementary operator into the model. By applying the proposed method, the hysteresis in the piezoelectric actuators and ultrasonic motor are illustrated, respectively. Hamimid et al. [bib_ref] Minor hysteresis loops model based on exponential parameters scaling of the modified..., Hamimid [/bib_ref] proposed the minor hysteresis loops model based on parameters scaling of the modified Jiles-Atherton model by using judicious expressions. The proposed model had been applied for 3.2% Fe-Si nonoriented magnetic sheet, and the result showed the expected behavior when the flux density level was low. Chwastek [bib_ref] Modelling hysteresis loops in thick steel sheet with the dynamic Takács model, Chwastek [/bib_ref] used the dynamic Takacs model to describe hysteresis loops in a thick nonoriented steel sheet and achieved the dynamic extension by using an additional component of the effective field. The experiment obtained a satisfactory agreement between the measured and the modeled hysteresis loops. Baghel and Kulkarni [bib_ref] Parameter identification of the Jiles-Atherton hysteresis model using a hybrid technique, Baghel [/bib_ref] proposed a hybrid technique to solve the parameter identification problem based on the Jiles-Atherton hysteresis model. And the proposed technique is flexible enough to incorporate improved GA and LM algorithms. Zirka et al. [bib_ref] On physical aspects of the Jiles-Atherton hysteresis models, Zirka [/bib_ref] analyzed the physical assumptions under the static and dynamic Jiles-Atherton (JA) hysteresis models. This led to the using in the model of a misleading entity resembling the coenergy instead of the actual energy. It is necessary to take measures to avoid this nonphysical feature; therefore the JA static model is reserved for applying in circuit simulators.
The Duhem model is a kind of differential hysteresis model, proposed by Duhem and Stefanini in 1897 [bib_ref] Traite elementaire de mecanique chimique, fondee sur la thermodynamique, Duhem [/bib_ref]. The Scientific World Journal
The Duhem model has the explicit function expression and is the function of input signal derivative; the output of model is related to the rate of input signal. The model is a kind of dynamic models [bib_ref] A neural networks based model for rate-dependent hysteresis for piezoceramic actuators, Dong [/bib_ref] , conforms the dynamic characteristic of hysteresis nonlinear in the actual intelligent materials, and could describe the hysteresis nonlinear precisely. However it is difficult to obtain the parameter , coefficients of and in the Duhem model, and it would be the obstacles for application of Duhem model.
## Parameters identification of the hysteresis model
The Duhem has an explicit differential expression, through adjusting the parameter , coefficients of and of the Duhem model; different hysteresis characteristics could be reflected; while identifying the parameters of the Duhem model accurately, the hysteresis model of the piezoceramic actuator could be obtained [bib_ref] Identification and control of dynamic modeling for piezoceramic actuator, Chen [/bib_ref]. The differential function of the Duhem model is
[formula] = V [ (V) − ] + V (V) ,(1) [/formula]
where is constant, V is the input voltage, is output displacement, (V) and (V) are piecewise continuous functions.
[ , ] represents the set of all continuous functions defined in the closed interval , to the arbitrary
[formula] (V) and (V) in the [ , ]; ‖ − ℎ‖ ∞ = sup ≤V≤ | (V) − ℎ(V)| [/formula]
represents the distance between them [bib_ref] Identification and control of dynamic modeling for piezoceramic actuator, Chen [/bib_ref]. Letting ∈ [ , ], to the arbitrary given > 0, the polynomial existed and the following equation holds:
[formula] − ℎ ∞ = sup ≤V≤ (V) − ℎ (V) ≤ .(2) [/formula]
To the arbitrary given (V) ∈ [ , ] and approximation precision, there has an algebraic polynomial
[formula] ℎ (V) = 0 + 1 V + 2 V 2 + ⋅ ⋅ ⋅ + V ,(3) [/formula]
where is natural number and ‖ − ℎ‖ ∞ ≤ . When accuracy > 0, the order of (V) and (V) is and , respectively; the polynomials are as follows:
[formula] (V) = 0 + 1 V + 2 V 2 + ⋅ ⋅ ⋅ + V = ∑ =0 V , (V) = 0 + 1 V + 2 V 2 + ⋅ ⋅ ⋅ + V = ∑ =0 V . (4) Substituting (4) into (1), = V [ (V) − ] + V (V) .(5) [/formula]
And (5) could be transformed into
[formula] = V [ ∑ =0 V − ] + V ∑ =0 V .(6) [/formula]
As the input voltage V, output displacement , and V/ , / are measurable, while identifying the parameters , , and accurately, the parameterized model of the Duhem model could be obtained.
Letting
[formula] ( ) = |V( ) − V( − 1)|, ( ) = V( ) − V( − 1), ( ) = ( )− ( −1), = 2, 3, . . ., the dynamic discretization Duhem model of the system is ( ) = ( ) ⋅ [ ∑ =0 V( ) − ( )] + ( ) ⋅ ∑ =0 V( ) ,(7) [/formula]
where V( ) is the input voltage of the system at time , ( ) is the output displacement at time . Letting ( ) = ( ) × , ( ) is the data vectors of the input voltage, is the identified parameter vector. That is,
[formula] ( ) = [ ( ) , ( ) V ( ) , . . . , ( ) V( ) , − ( ) ( ) , ( ) , ( ) V ( ) , . . . , ( ) V( ) ] ( ) ∈ 1×( + +3) , = [ 0 , 1 , 2 , . . . , , , 0 , 1 , 2 , . . . , ] ∈ 1×( + +3) .(8) [/formula]
Let
[formula] ( ) = ∞ ∑ =1 [ ( )] 2 = ∞ ∑ =1 {[ ( ) − ( ) × ] 2 } ,(9) [/formula]
that is,
[formula] ( ) = ( ) − ( ) × .(10) [/formula]
The target of the parameter identification is to obtain the value of the parameter when the function is the minimum.
Applying the recursive least squares algorithm to recursive equations (11), [bib_ref] Robust adaptive control of systems with hysteretic nonlinearities: a Duhem hysteresis modelling..., Feng [/bib_ref] , and (13), the identification parameters are
[formula] ∧ ( ) = ∧ ( − 1) + ( ) [ ( ) − ( ) ∧ ( − 1)] , (11) ( ) = ( ) ( + 1) 1 + ( ) ( − 1) ( ) ,(12)( ) = [1 − ( ) ( ) ] ( − 1) .(13) [/formula]
Equation (11)
where ( ) is the weight matrix, the effect of the weight is to control the influence of the input component.
If the element of the weight matrix meets the following conditions:
(1) 0 < Λ ≤ Λ ( ) ≤ Λ ( = 1, 2, . . . , ), Λ and Λ are the determined upper and lower bound values; (2) at least one Λ ( ) existed for , that
[formula] Λ ( ) − Λ ( + 1) Λ ( ) ≥ Λ ( ) − Λ ( + 1) Λ ( )(16) [/formula]
or
[formula] Λ ( + 1) Λ ( ) ≤ Λ ( + 1) Λ ( ) ;(17) [/formula]
(3) 0 < ( ) < 2/ ∑ =1 Λ ( ) 2 ( );
[formula] (4)̃( ) and ( ) are disjoint;̃( ) = 0 − ∧ ( ), [/formula]
Then regardless of the initial value of the parameter estimate, the parameter estimated value is always wide range asymptotic convergence, that is
[formula] lim → 0 ∧ ( ) = 0 .(18) [/formula]
## Parameters identification simulation and modeling of the duhem model
To verify the accuracy of the algorithms for the Duhem model identification, the paper applied recursive least squares algorithm and gradient correction algorithm to identify the parameters of the Duhem model based on MATLAB simulation software, respectively, and contrasted the influence of the two identification algorithms to the model modeling accuracy.
## Identification of recursive least squares algorithm.
In the experiment, the order of the polynomial (V), (V) is = 3, = 2, respectively; that is, Utilizing the parameter identification data, the hysteresis curve of the model is shown in [fig_ref] Figure 2: Input-output hysteresis curves of Duhem model [/fig_ref]. The red and blue curves represent input-output hysteresis curve of the Duhem model and actual input-output hysteresis curve, respectively. [fig_ref] Figure 2: Input-output hysteresis curves of Duhem model [/fig_ref] showed that the output of the Duhem model and the actual output data are basically consistent. [fig_ref] Figure 3: Error curve between the actual output and model output [/fig_ref] is the relative error curve between the model output and actual output. It can be seen that the maximum error is 0.066 m. The result of the experiment verified the validity of the recursive least squares algorithm.
[formula] (V) = 0 + 1 V ( ) + 2 V( ) 2 + 3 V( ) 3 , (V) = 0 + 1 V ( ) + 2 V( ) 2 .(19) [/formula]
## Gradient correction algorithm.
The data of the input and output is shown in [fig_ref] Figure 1: is the given input-output curves [/fig_ref] ; under the gradient correction algorithm, Λ ( ) is as follows:
[formula] diag Λ ( ) = [1, 1 3 , 1 3 2 , 1 3 3 , 1, 1, 1 3 , 1 3 2 ] .(21) [/formula]
4
The Scientific World Journal The parameter identification result is shown in [fig_ref] Figure 4: Parameter identification curves of the gradient correction algorithm [/fig_ref]. It can be seen form [fig_ref] Figure 4: Parameter identification curves of the gradient correction algorithm [/fig_ref] that the identification parameters tend to be stable when recursiving to = 6; the parameter identification results are as follows: Utilizing the gradient correction parameter identification results, the model hysteresis curve is shown in , The red and blue curves represent input-output hysteresis curve of the Duhem model and real input-output hysteresis curve, respectively. The error curve between the system output and model output is shown in . It can be seen from that the maximum is 0.048 m. The result also verified the validity of the gradient correction algorithm.
The identification parameters of the two algorithms are shown in [fig_ref] Table 1: Identification parameters of two algorithms [/fig_ref]. And we show part of the relative errors contrast results under the two algorithms in [fig_ref] Table 2: The relative errors of two algorithms [/fig_ref]. It can be seen from [fig_ref] Table 1: Identification parameters of two algorithms [/fig_ref] that the relative errors between the real output and model output under the recursive least squares algorithm could reach 0.24%, the mean square deviation of the error is 0.0263, and the maximum error is 0.066 m; in contrast, the relative errors between the actual output and model output under the gradient correction algorithm could reach 0.11%, the mean square deviation of the error is 0.0222, and the maximum error is 0.048 m.
The Scientific World Journal 5
# Conclusion
The paper utilized the polynomial to approach the piecewise continuous functions and of the Duhem model, adopted the recursive least squares and gradient correction algorithm, respectively, to identify the parameter , coefficients of and of the Duhem model, and applied the identified parameters to model the Duhem model. The experiment results showed that the modeling accuracy of the recursive least squares algorithm could reach 0.24%, the mean square deviation of the error is 0.0263; the modeling accuracy of the gradient correction algorithm could reach 0.11%, the mean square deviation of the error is 0.0222. The results of the experiment certified validity of the recursive least squares algorithm and gradient correction algorithm. Contrasting with the least squares algorithm, the gradient correction algorithm is adaptable and suitable for engineering. Applying the gradient correction algorithm, the Duhem model could be established more precisely and lay the foundation for the further control research of the piezoceramic.
[fig] Figure 1: is the given input-output curves; there are 21 sets of data totally. The red line represents the voltage input signal, and the blue line represents the displacement output signal. Under recursive least squares algorithm, the identification result is = 0.0874, (V) = − 0.015 + 0.006V ( ) + 5.57 × 10 −6 V( ) 2 − 8.6 × 10 −8 V( ) 3 , (V) = 0.053 − 8.316 × 10 −4 V ( ) + 8.195 × 10 −6 V( ) 2 . [/fig]
[fig] Figure 2: Input-output hysteresis curves of Duhem model. [/fig]
[fig] Figure 3: Error curve between the actual output and model output. [/fig]
[fig] Figure 4: Parameter identification curves of the gradient correction algorithm. [/fig]
[fig] Figure 5, Figure 6: 6096 × 10 −8 V( ) 3 , (V) = 0.051 − 8.3166 × 10 −4 V ( ) + 8.1960 × 10 −6 V( ) 2 . Input-output hysteresis curves of Duhem model. Error curve between the actual output and model output. [/fig]
[table] Table 1: Identification parameters of two algorithms. [/table]
[table] Table 2: The relative errors of two algorithms. [/table]
|
Full-Length Enriched cDNA Libraries and ORFeome Analysis of Sugarcane Hybrid and Ancestor Genotypes
Sugarcane is a major crop used for food and bioenergy production. Modern cultivars are hybrids derived from crosses between Saccharum officinarum and Saccharum spontaneum. Hybrid cultivars combine favorable characteristics from ancestral species and contain a genome that is highly polyploid and aneuploid, containing 100-130 chromosomes. These complex genomes represent a huge challenge for molecular studies and for the development of biotechnological tools that can facilitate sugarcane improvement. Here, we describe full-length enriched cDNA libraries for Saccharum officinarum, Saccharum spontaneum, and one hybrid genotype (SP803280) and analyze the set of open reading frames (ORFs) in their genomes (i.e., their ORFeomes). We found 38,195 (19%) sugarcane-specific transcripts that did not match transcripts from other databases. Less than 1.6% of all transcripts were ancestor-specific (i.e., not expressed in SP803280). We also found 78,008 putative new sugarcane transcripts that were absent in the largest sugarcane expressed sequence tag database (SUCEST). Functional annotation showed a high frequency of protein kinases and stress-related proteins. We also detected natural antisense transcript expression, which mapped to 94% of all plant KEGG pathways; however, each genotype showed different pathways enriched in antisense transcripts. Our data appeared to cover 53.2% (17,563 genes) and 46.8% (937 transcription factors) of all sugarcane full-length genes and transcription factors, respectively. This work represents a significant advancement in defining the sugarcane ORFeome and will be useful for protein characterization, single nucleotide polymorphism and splicing variant identification, evolutionary and comparative studies, and sugarcane genome assembly and annotation.
# Introduction
Sugarcane (Saccharum spp.) is a C4 grass that stores large amounts of sucrose in its stems, which can account for as much as 40%-50% of the culm dry weight [bib_ref] Temporal and spatial regulation of sucrose accumulation in the sugarcane stem, Moore [/bib_ref]. Sucrose from sugarcane has been used for human consumption for centuries and has recently been used for bioethanol production. Biomass from sugarcane can also be used for bioenergy production; the bagasse can be burned to generate electricity [bib_ref] Scientific challenges of bioethanol production in Brazil, Amorim [/bib_ref] and can be hydrolyzed to yield simple sugars from the complex plant cell wall, which can be fermented to produce bioethanol [bib_ref] Plant genetic engineering for biofuel production: towards affordable cellulosic ethanol, Sticklen [/bib_ref]. These features place sugarcane among the best feedstock options for future bioenergy production.
Early breeding programs used two main species for sugarcane improvement: Saccharum officinarum (2n = 80, basic chromo-some number x = 10) and Saccharum spontaneum (2n = 36-128, basic chromosome number x = 8) [bib_ref] Saccharum species as horticultural classes, Irvine [/bib_ref]. The domesticated species, Saccharum officinarum, has thick and juicy culms with high sucrose content, whereas the wild species, Saccharum spontaneum, has thin, fibrous, low-sucrose culms and higher stress tolerance. As a result of interspecies hybridization and breeding strategies, sugarcane cultivars are now highly polyploid and aneuploid, with 10-12 sets of chromosomes and a monoploid genome size of 760-930 Mb. Among the 100-130 chromosomes present in modern varieties [bib_ref] Sugarcane Improvement through Breeding and Biotechnology, Ming [/bib_ref] , 70%-80% are derived from Saccharum officinarum, 10%-20% are derived from Saccharum spontaneum, and 10% are recombinants of these two species [bib_ref] Unraveling the genome structure of polyploids using FISH and GISH; examples of..., D'hont [/bib_ref] [bib_ref] Characterisation of the double genome structure of modern sugarcane cultivars (Saccharum spp.)..., D'hont [/bib_ref] [bib_ref] Sugarcane: a Major Source of Sweetness, Alcohol, and Bio-energy, D'hont [/bib_ref]. Recently, breeding efforts are aiming to increase energy content (GJ/ha) using Saccharum spontaneum genotypes to produce an Energycane, a higher yield cane, with increased fiber content and higher tolerance to drought.
The sugarcane complex genome represents a great challenge for molecular studies. Despite recent advances [bib_ref] Haplotype structure around Bru1 reveals a narrow genetic basis for brown rust..., Costet [/bib_ref] [bib_ref] SNP genotyping allows an in-depth characterisation of the genome of sugarcane and..., Garcia [/bib_ref] [bib_ref] A mixed model QTL analysis for sugarcane multiple-harvest-location trial data, Pastina [/bib_ref] [bib_ref] A novel linkage map of sugarcane with evidence for clustering of retrotransposon-based..., Palhares [/bib_ref] [bib_ref] Functional markers for gene mapping and genetic diversity studies in sugarcane, Marconi [/bib_ref] [bib_ref] Sugarcane for water-limited environments. Genetic variation in cane yield and sugar content..., Basnayake [/bib_ref] [bib_ref] Simultaneously accounting for population structure, genotype by environment interaction, and spatial variation..., Wei [/bib_ref] , biotechnology for sugarcane analysis is less advanced than that for other economically important crops. For example, compared with sorghum, maize, and rice, sugarcane does not have a draft genome sequence available or a well-annotated transcriptome. A multinational effort to produce a reference sugarcane genome (sugarcanegenome.org) is in progress, but many obstacles must be overcome, particularly because a high number of closely related homologs, paralogs, and alleles collapse into single contigs during the assembly.
To date, functional genomics and transcriptomics studies in sugarcane have relied primarily on expressed sequence tag (EST) and microarray datasets [bib_ref] Biofuel and energy crops: high-yield Saccharinae take center stage in the post-genomics..., De Siqueira Ferreira [/bib_ref]. Despite their limitations, these studies produced important information on sugarcane gene functions [bib_ref] Biofuel and energy crops: high-yield Saccharinae take center stage in the post-genomics..., De Siqueira Ferreira [/bib_ref] [bib_ref] Role of Bioinformatics as a Tool for Sugarcane Research, Casu [/bib_ref] [bib_ref] Functional Genomics: Transcriptomics of Sugarcane -Current Status and Future Prospects, Casu [/bib_ref] [bib_ref] Transcriptome Analysis and Functional Genomics of Sugarcane, Manners [/bib_ref] [bib_ref] Sugarcane Functional Genomics: gene discovery for agronomic trait development, Menossi [/bib_ref]. Experiments addressing sucrose accumulation [bib_ref] Identification of a novel sugar transporter homologue strongly expressed in maturing stem..., Casu [/bib_ref] [bib_ref] Sugarcane genes associated with sucrose content, Papini-Terzi [/bib_ref] , stem and cell wall development [bib_ref] Identification of differentially expressed transcripts from maturing stem of sugarcane by in..., Casu [/bib_ref] [bib_ref] Identification of transcripts associated with cell wall metabolism and development in the..., Casu [/bib_ref] , and responses to hormones and stresses [bib_ref] Identification of sense and antisense transcripts regulated by drought in sugarcane, Lembke [/bib_ref] [bib_ref] Signal transduction-related responses to phytohormones and environmental challenges in sugarcane, Rocha [/bib_ref] have improved our understanding of mechanisms mediating sugarcane growth; however, there is still much to learn. Thus, to obtain insights into the functions of putative trait-related genes and unknown proteins, we must first obtain full-length transcripts.
In this paper, we present the development of a protocol to produce full-length cDNA libraries for cloning and next generation sequencing (NGS), using a commercial hybrid (SP803280) and two ancestor genotypes (Saccharum officinarum and Saccharum spontaneum). We also present high-throughput sequencing data from the set of open reading frames (ORFs) in the sugarcane genome (i.e., the sugarcane ORFeome). Defining this full-length cDNA dataset will be critical for large-scale protein characterization and genome assembly and annotation, facilitating sugarcane improvement. The use of the two ancestor genotypes will help define genes and regulatory networks that may underlie the distinct phenotypes that can be useful for improving sugarcane yield and for designing new varieties for bioethanol generation.
# Materials and methods
# Plant materials
Leaves, immature and intermediate internodes samples from field-grown SP803280 plants were sampled after plants were cultivated for 9 months. The immature internode was considered the first internode at the apex of the culm. The intermediate internode was considered the fifth internode from the immature internode. Saccharum officinarum (Caiana listrada) and Saccharum spontaneum (IN8458) were grown in a greenhouse, and leaf samples were harvested after plants were cultivated for 11 months. Each sample was a pool from five individual plants. Field samples were cultivated at Centro de Ciências Agrárias -Universidade Federal de São Carlos, city of Araras, state of São Paulo, Brazil (http://www.cca.ufscar.br/). GPS coordinates: 22u18946.40S 47u23909.60W. No specific permissions were required for these locations/activities and field studies did not involve endangered or protected species.
RNA isolation, full-length first-strand cDNA (FLFS cDNA) synthesis, and enrichment Total RNA was isolated using TRIzol (Life Technologies) followed by polyA mRNA isolation using a FastTrack MG mRNA Isolation kit (Life Technologies). The polyA mRNA (10 mg) from each sample was used for first-strand cDNA synthesis using SuperScript III (Life Technologies) and oligo-dT 21 VN primers (1 mg) for direct sequencing or an attB2-dT [bib_ref] Functional Genomics: Transcriptomics of Sugarcane -Current Status and Future Prospects, Casu [/bib_ref] VN primer (GGGGACAACTTTGTACAAGAAAGTTGGGT (T) [bib_ref] Functional Genomics: Transcriptomics of Sugarcane -Current Status and Future Prospects, Casu [/bib_ref] VN) for library construction. Reactions were carried out in 16 reaction buffer and incubated at 45uC for 40 min, 50uC for 40 min, and 55uC for 40 min. To enrich full-length cDNA, the mRNA:cDNA hybrid sample was treated with RNAse One (Ambion) in the same buffer and incubated at 37uC for 30 min. The sample was extracted using phenol:chloroform:isoamyl alcohol (Life Technologies) and precipitated by ethanol. The cDNA:mRNA sample was resuspended in 600 mL of 16 TE buffer containing 100 mM NaCl 2 , followed by addition of 3 mg Cap antibody-conjugated magnetic beads (160 mg antibody/mg beads, Life Technologies) and gentle rotation at room temperature for 1 h. The beads were washed three times with 16TE buffer containing 100 mM NaCl 2 . For library construction and 454 sequencing (Roche), the fulllength enriched cDNA:mRNA sample was eluted from beads in 300 mL buffer containing 1.5 M guanidine isothiocyanate, 20 mMTris-HCl (pH 7.5), and 10 mM EDTA and incubated at room temperature for 1 h. For direct sequencing, the full-length enriched single-stranded cDNA was eluted from cDNA:mRNA beads in 16 RNase H buffer containing 2 units of RNAse H (Life Technologies) and 12 mg of RNAse A (Life Technologies) for 30 min at 37uC, followed by incubation at 95-100uC for 5 min. The eluted FLFS cDNA was precipitated by ethanol and used for subsequent experiments.
## Cdna library construction and cloning
The full-length enriched cDNA:RNA was used for secondstrand cDNA synthesis in a 150-mL reaction containing 16 second-strand cDNA synthesis buffer (Life Technologies), 1.3 mM DTT, 266 mM dNTP, 10 units E.coli DNA ligase, 40 units E.coli DNA polymerase, and 2 units E.coli RNAse H. The samples were incubated at 16uC for 2 h, and the double-stranded cDNA was polished by incubation with T4 DNA polymerase at 16uC for 5 min. A double-stranded 59 adaptor containing an attB1 site (top strand: TCGTCGGGGACAACTTTGTACAAAAAAGTTGG-A; bottom strand: PHO-TCCAACTTTTTTGTACAAAGTTG-TCCCC) was ligated to the double-stranded cDNA. The cDNA sample containing an attB1 site at the 59end and an attB2 site at the 39end was cloned into the pDNOR222 vector (Gateway, Life Technologies) in a 20-mL recombination reaction containing 16 TE buffer and 4 mL BP enzyme mix (Life Technologies) and incubated at room temperature overnight.
## Titanium sequencing
Double-stranded cDNA was fragmented using the Covaris S220 system (Covaris, Woburn, MA, USA) to target DNA fragments ranging in size from 400 to 1000 bp following the manufacturer's protocol using MID tags for each sample.
## Ion pgm sequencing
To maintain transcriptome complexity, we used 75% FLFS cDNA and 25% original polyA mRNA from the sample preparation for Ion PGM sequencing (Life Technologies). The FLFS cDNA sample was fragmented using the Covaris S220 system (Covaris) to target DNA fragments ranging from 300 to 500 bp, following the manufacturer's protocol. A DNA Ionadaptor mixture (containing the standard Ion A and P1 sequences) was ligated to the FLFS cDNA fragments. For the polyA mRNA, RNAse III was used for fragmentation, followed by size selection (300-500 bp), adaptor ligation, and reverse transcription using the ''Ion-Total RNA-Seq'' protocol (version 2) for preparation of 300base-read RNA libraries (Life Technologies). The adapted cDNA fragments derived from FLFS cDNA and mRNA were equally mixed for emulsion PCR using the One Touch 2 system following the Ion PGM Template OT2 400 base protocol (Life Technologies). The enriched template-positive Ion sphere particles were loaded onto the Ion 318 chips following manufacturer's protocol. The raw data were trimmed, and FASTQ files were generated using Ion plug-in software.
## Raw reads processing
Data from 454 and Ion PGM were processed and filtered individually. The Ion PGM raw data were trimmed in Torrent Suite (version 3.6.2) to remove primer, adaptor, and low-quality base calling sequences. The FASTQ files containing Q20 reads were generated using Ion plug-in FastqCreator software (3.6.0). Reads from the 454 platform were processed and trimmed with Prinseq [bib_ref] Quality control and preprocessing of metagenomic datasets, Schmieder [/bib_ref]. First, we applied a quality filter by trimming reads using mean Phred quality scores below 20, reads with at least 5 N's or 5 poly-A/T's in the ends, and removing reads with minimum length of 70 after trimming. Next, we applied the 'dust' method to remove low complexity sequences. All reads were deposited in the Sequence Read Archive (SRA-NCBI) database under accession number SRP042605.
## De novo transcriptome assembly
Ion PGM and 454 processed reads were combined and assembled using Trinity 2013-02-25 software [bib_ref] Fulllength transcriptome assembly from RNA-Seq data without a reference genome, Grabherr [/bib_ref]. Reads from all five samples were combined, and the minimum sequence length in the assembly was set to 200 bp. Trinity software processed data sequentially through three modules, which allows for recovery of transcript isoforms using the de Bruijn graph algorithm [bib_ref] De novo transcript sequence reconstruction from RNA-seq using the Trinity platform for..., Haas [/bib_ref].
## Gene prediction
We used three ESTScan scripts to create a sugarcane codonusage matrix and to predict ORFs: (i) extract_mrna, using 832 manually annotated genes from 303 Sugarcane BAC sequences [bib_ref] Building the sugarcane genome for biotechnology and identifying evolutionary trends, De Setta [/bib_ref] with a mean size of 1,820 bp, composed of at least two exons and 100 bp in 59 and 39 untranslated regions (UTRs); (ii) prepare_data, to separate genes in the training and test groups according to GC-content, to mask redundancies, and to classify coding and noncoding parts of genes; and (iii) build_model, to create the matrix using the previous data. This matrix was used for protein prediction with ESTScan [bib_ref] ESTScan: a program for detecting, evaluating, and reconstructing potential coding regions in..., Iseli [/bib_ref].
## Functional annotation of sugarcane transcripts
All contigs were used for comparative and functional genomic analyses. Putative genes with functional annotations were compared in the following order, for species with more specific transcript hits, against each grass genome dataset: Sorghum bicolor, Oryza sativa, Zea mays, Brachypodium distachyon, Panicum virgatum, and Setaria italica (Phytozome v9.0). Comparisons were performed by BLAST search (e-value 1e-05) and were processed with an in-house script based on two methods: single-directional best hit (SBH) information (blastx) and bidirectional best hit information, in both forward (BLASTX) and reverse directions (TBLASTN), taking each gene in genome A as a query compared against all genes in genome B, and vice versa [bib_ref] KAAS: an automatic genome annotation and pathway reconstruction server, Moriya [/bib_ref]. The mapping step allowed the cross-species annotation (i.e. to take the annotation from a grass gene and assign it to the matching sugarcane contig), for BLAST hits with bit scores higher than 30, following the above species order. Since the bit scores of a gene pair a and b from two genomes A and B, respectively, can be different in forward and reverse orientations, and because the top scores do not necessarily reflect the order of the rigorous Smith-Waterman scores, we used the bidirectional hit rate (BHR) and selected genes with BHRs greater than 0.95 [bib_ref] KAAS: an automatic genome annotation and pathway reconstruction server, Moriya [/bib_ref]. To annotate the assembled contigs according to GO terms [bib_ref] Gene ontology: tool for the unification of biology. The Gene Ontology Consortium, Ashburner [/bib_ref] , we used the BHR method as described above to determine gene ontologies. All predicted proteins were categorized according to PFAM domains [bib_ref] The Pfam protein families database, Punta [/bib_ref] using HMMPFAM from InterProScan software [bib_ref] InterProScan 5: genome-scale protein function classification, Jones [/bib_ref] with an e-value threshold of 1e-03.
Next, we identified sugarcane metabolic pathways for predicted proteins using the KEGG Automatic Annotation Server (www. genome.jp/kegg/kaas) with the eukaryote and plants dataset, which assigns KO identifiers as controlled vocabularies to sets of new sequences based on BHRs $30, excluding non-plant pathways.
## Functional activity scores
Following the functional class scoring (FCS) approach [bib_ref] A systems biology approach for pathway level analysis, Draghici [/bib_ref] , we developed an algorithm that integrated transcript expression profiles and metabolic pathways to estimate the activities of metabolic pathways using in-house scripts and R-statistics modules. To define and compare the functional activities of NATs for the metabolic pathways in each sample, we initially selected the significantly expressed antisense transcripts and then estimated the pathway activity based on the antisense transcripts identified in each pathway and their relative abundances from each sample. The pathway activity comparisons represented the likelihood that a metabolic pathway was active and the degree of activation. To calculate the relative pathway activity (PA), the pathway(s) associated with each antisense expression profile was retrieved. Then, we summed the gene intensities for each pathway to obtain the total pathway score and normalized it dividing by the proportion of detected antisense transcripts for its respective pathway. The score for PA was calculated as:
[formula] PA(Pathway)~g 1 zg 2 zg 3 :::zg N Ã N M [/formula]
where: g j is the intensity of gene j identified from pathway P, N is the total number of genes detected in pathway P, and M is the total number of genes part of pathway P. Then, the scores were log2-transformed and centered by the median value. Subsequently, we performed a hierarchical clustering using the average method (for pathways) and the Spearman correlation method (for samples). The cut in the tree to decide the number of clusters in the hierarchical clustering was determined using the R package NbClustwith the ''Euclidean distance'' to create the dissimilarity matrix.
## Reads mapping and identification of antisense transcripts
The six grass genomes cited above, SUCEST transcripts [bib_ref] Analysis and functional annotation of an expressed sequence tag collection for tropical..., Vettore [/bib_ref] , and Trinity assembled contigs were used to map reads from each cDNA library to identify the expression levels of contigs in each genotype or sample and the presence of antisense transcripts using the Bowtie aligner [bib_ref] Ultrafast and memoryefficient alignment of short DNA sequences to the human genome, Langmead [/bib_ref] with the following parameters: verysensitive-local: (-D 20 -R 3 -N 0 -L 20 -i S,1,0.50); p: 32 (number of threads); and hits: $100 bp (minimum alignment length). Only reads from Ion PGM sequencing were used for antisense identification since the protocol used allowed to kept the strand orientation, what was not possible for 454 sequencing.
## Tfs
TF families were obtained from predicted proteins based on specific rules as established in GRASSIUS [bib_ref] GRASSIUS: a platform for comparative regulatory genomics across the grasses, Yilmaz [/bib_ref]. The domain identification involved a HMMPFAM search from InterProScan against all available PFAM hidden Markov models, keeping only significant hits with an e-value threshold of 0.001. After that, we applied the script with the GRASSIUS rules for the identification and classification of putative TFs.
## Prediction of full-length transcripts
To estimate the number of full-length sugarcane transcripts, we used the full-length CDS genes from six grasses as references: Sorghum bicolor, Zea mays, Panicum virgatum, Setaria italica, Oryza sativa, and Brachypodium distachyon. These genes contained the complete mRNA (59UTR+CDS+39UTR) [fig_ref] Table 2: Data from Ion PGM and 454 sequencing [/fig_ref]. Identification of full-length sugarcane transcripts was conducted in three analyses. The first two were based on the alignment between sugarcane contigs and CDSs of full-length species genes by BLASTN, with parameters as follows: e-value threshold: 1e-15, max_target_seqs: 1, outfmt: 6, dust: NO, num_threads: 20, only contigs with coverage $50%, respective gene hit coverage $95%, and identity $80%. In Analysis I, we considered transcripts to be full-length only if they completely covered the first gene hit, according to the above criteria, and such genes and contigs were disregarded in Analysis II. In Analysis II, we selected full-length genes of each species if two or more contigs covered the given gene, based on the above criteria. In both analysis, the genes were analyzed in a specific species order (Sorghum bicolor, Zea mays, Panicum virgatum, Setaria italica, Oryza sativa, and Brachypodium distachyon), based on the species with more specific transcript hits. Transcripts identified as full-length in comparison to Sorghum bicolor genes were not employed in the analysis in Zea mays, and those identified as full-length in comparison to Zea mays were not employed in the analysis in Panicum virgatum and so on. In Analysis III, we calculated the mean CDS size (1,030.85 bp, [fig_ref] Table 2: Data from Ion PGM and 454 sequencing [/fig_ref] for full-length genes from the six grasses, and contigs not classified in the two previous analyses were selected as putative full-length genes if their CDS sizes were greater than 1,030.85 bp and complete CDS (including the start and stop codons) similar to [bib_ref] Efficient assembly and annotation of the transcriptome of catfish by RNA-Seq analysis..., Liu [/bib_ref]. Prediction of paralogs and gene families (see below) allowed the identification of uniquely represented full-length contigs by clustering alleles, variants and paralogs of a given gene. Full-length TFs were identified by integration between full-length transcript analysis and the TF identification as described above.
# Gene family analysis
Paralogs and orthologs were identified from the sugarcane, Sorghum, maize, Oryza, Setaria, Brachypodium, and Panicum putative full-length predicted proteins [fig_ref] Table 2: Data from Ion PGM and 454 sequencing [/fig_ref] with Hieranoid software (Schreiber et al. 2013) by inferring pair-wise homology relationships between the translated transcripts and combining them into groups of orthologous sequences, with specific parameters (bootstrap: 1, matrix: BLOSUM62, group_overlap_cutoff: 0.5).
# Results and discussion
Library construction and cDNA sequencing
The genotypes Saccharum officinarum (Caiana listrada) and Saccharum spontaneum (IN8458) were chosen since they are representatives of the genotypes used in early breeding programs to develop sugarcane hybrids. Variety SP803280 was chosen because it represents a common sugarcane variety model used in studies in all areas and is currently being sequenced by the sugarcane genome project (sugarcanegenome.org). Five samples were used to produce individual cDNA libraries, three from SP803280 (leaves, immature and intermediate internodes) and one from each of the ancestor genotypes (leaves only). Using these samples, we implemented a protocol for construction of full-length enriched cDNA libraries for cloning and NGS [fig_ref] Figure 1: Full-length enrichment for library cloning and next generation sequencing [/fig_ref]. Briefly, oligo(dT) primers were used to synthesize cDNA and were then treated with RNAse I to digest single-stranded RNAs (non-reversetranscribed into cDNA) and by full-length cDNA selection using Cap-antibody magnetic beads. A fraction of each cDNA sample was ligated to barcoded double-stranded DNA adaptors for cloning. The remaining fraction was fragmented, ligated to double-stranded adaptors, mixed with original polyA mRNA that was also reverse transcribed into cDNA, and sequenced by NGS.
For cloning, all five barcoded libraries were mixed, cloned into Gateway vectors (pDNOR222, Life Technologies), and divided into two fractions based on their average insert size, i.e., approximately 1.1 and 1.4 kb. Thus, the sugarcane cDNA library was consistent with previously reported full-length libraries for other plants. The cDNA average insert lengths of 1.1 and 1.4 kb were similar to those of tomato (1,418 bp) [bib_ref] Large-scale analysis of full-length cDNAs from the tomato (Solanum lycopersicum) cultivar Micro-Tom,..., Aoki [/bib_ref] , Arabidopsis (1,445 bp) [bib_ref] Features of Arabidopsis genes and genome discovered using full-length cDNAs, Alexandrov [/bib_ref] , and poplar (1,045 bp) [bib_ref] Analysis of 4,664 high-quality sequence-finished poplar full-length cDNA clones and their utility..., Ralph [/bib_ref] and longer than that of Brachypodium (808 bp) [bib_ref] Large-scale collection and analysis of full-length cDNAs from Brachypodium distachyon and integration..., Mochida [/bib_ref]. Twenty-six randomly selected clones were sequenced by the Sanger method from both cDNA ends, achieving full-length inserts from 90% of clones after excluding low-quality sequences and repeated clones [fig_ref] Table 2: Data from Ion PGM and 454 sequencing [/fig_ref]. All trimmed reads were assembled with Trinity software [bib_ref] Fulllength transcriptome assembly from RNA-Seq data without a reference genome, Grabherr [/bib_ref] , yielding 195,765 contigs with a mean contig size of 684 bp and N50 of 963 bp [fig_ref] Table 2: Data from Ion PGM and 454 sequencing [/fig_ref]. Around 20% of contigs were 1 kb or more in length [fig_ref] Table 2: Data from Ion PGM and 454 sequencing [/fig_ref].
A recent work by Cardoso-Silva and colleagues reported a de novo assembly of the sugarcane transcriptome using GAIIx Illumina 2636 bp (paired ends) of six sugarcane commercial varieties from three different breeding programs [bib_ref] De novo assembly and transcriptome analysis of contrasting sugarcane varieties, Cardoso-Silva [/bib_ref] , which included the genotype SP803280 used in this present work. However, the experiments carried out in this previous analysis [bib_ref] De novo assembly and transcriptome analysis of contrasting sugarcane varieties, Cardoso-Silva [/bib_ref] did not focus on full-length enrichment as described here. Instead, they aimed to identify molecular markers in the transcriptome that could be correlated to different phenotypes among the six varieties sampled, particularly resistance to rust and sucrose accumulation. Besides the GAIIx Illumina platform gives shorter reads than PGM and 454, it fits better to the expression profile analysis and the study of molecular markers, such as SSRs and SNPs. The authors produced over 445 million reads in six runs, which yielded almost five times more sequence (32 Gb) than reported here (6.6 Gb, [fig_ref] Table 2: Data from Ion PGM and 454 sequencing [/fig_ref]. They reported 119,768 assembled contigs with an average length of 921 bp, N50 1,367 bp, and 18,624 unigenes with 1 kb or more in length. The results regarding average length and N50 are higher than those of our results (see [fig_ref] Table 2: Data from Ion PGM and 454 sequencing [/fig_ref]. However, the work by Cardoso-Silva and colleagues excluded all contigs shorter than 300 bp, which certainly helped to increase these two values. If we assumed the same criteria used by Cardoso-Silva (contig minimum length = 300 bp and excluding isoforms), the N50 increased from 963 bp [fig_ref] Table 2: Data from Ion PGM and 454 sequencing [/fig_ref] to 1,071 (data not shown). These results suggest that sequencing depth may be more important than read length to produce assemblies with longer transcripts on average, since better results of average length and N50 were achieved by shorter reads from the GAIIx platform [bib_ref] De novo assembly and transcriptome analysis of contrasting sugarcane varieties, Cardoso-Silva [/bib_ref] than from similar technologies and longer reads from the 454 platform (this present study). However, recent publications have shown that longer reads could be useful for the identification of splicing isoforms, determination of the exon-intron structure, allele discrimination, and single nucleotide variations [bib_ref] Long-Read Sequencing of Chicken Transcripts and Identification of New Transcript Isoforms, Thomas [/bib_ref] [bib_ref] Defining a personal, allelespecific, and single-molecule long-read transcriptome, Tilgner [/bib_ref] , as in the case of Iso-Seq (Pacific Bioscience, Menlo Park, CA, USA), which generates longer reads than the 454 platform, using amplification-free long-read sequencing, originated from a single RNA molecule. Despite the higher error rate of the Iso-seq method compared to other technologies, these very long reads and their circular consensus reads could be used as a scaffold for de novo assembling, and the error could be corrected by short reads from other platforms [bib_ref] Hybrid error correction and de novo assembly of single-molecule sequencing reads, Koren [/bib_ref]. The mRNA:cDNA hybrid was treated with RNase I (scissor) to remove the single-stranded RNA that was not fully extended by the first-strand cDNA, followed by selection for full-length transcripts using Capantibody magnetic beads to enrich the full-length mRNA:cDNA. The full-length single-stranded DNA (FLssDNA) was eluted from beads and used for both cDNA library cloning (lower left) and NGS (lower right). For full-length library cloning, a double-stranded adaptor (green) was linked to the 59 end of ssDNA. Second-strand cDNA synthesis was then carried out, followed by cloning into a vector. For NGS, the full-length enriched ssDNA was fragmented by sonication to target fragments in the range of 200-400 bp, followed by ligation of the double-stranded DNA sequencing adaptor mixture (purple) to 39 and 59 ends of ssDNA. To maintain the complexity of the library while enriching the full-length cDNA for NGS, the original polyA mRNA was also fragmented using RNAse III, followed by ligation of the double-stranded RNA sequencing adaptor mixture (brown) to 39 and 59 ends of mRNA. After first-and second-strand synthesis, the polyA and capped mRNA and polyA and non-capped mRNA samples were mixed in a 3:1 ratio and applied to the downstream NGS procedure. doi:10.1371/journal.pone.0107351.g001. Description of cloned full-length cDNA libraries. Comparisons with other species and among genotypes revealed a high number of sugarcane-specific transcripts, but a low number of genotype-specific transcripts Sugarcane contigs were mapped against the predicted genes of six grasses plus the main sugarcane EST database, SUCEST [bib_ref] Analysis and functional annotation of an expressed sequence tag collection for tropical..., Vettore [/bib_ref] [bib_ref] The libraries that made SUCEST, Vettore [/bib_ref] (http://sucest-fun.org). At least 70% of all grass genes showed a matching contig, with higher percentages for closely related species, such as sorghum [fig_ref] Figure 2: Number of grass transcripts mapping to sugarcane contigs [/fig_ref]. About 7% of SUCEST sequences did not match any contig [fig_ref] Figure 2: Number of grass transcripts mapping to sugarcane contigs [/fig_ref] ; this may be explained by tissue-specific genes that are present in SUCEST but were not sampled in this study, such as those found in roots, lateral buds, and flowers [bib_ref] The libraries that made SUCEST, Vettore [/bib_ref]. Interestingly, 78,008 contigs did not match any SUCEST sequence [fig_ref] Figure 3: Venn diagram comparing sugarcane transcripts as obtained by RNAseq [/fig_ref] , and only 3,086 contigs were specific to Saccharum officinarum and/or Saccharum spontaneum ; since only 15% of SUCEST reads were generated from leaf and internode tissues [bib_ref] The libraries that made SUCEST, Vettore [/bib_ref] , these may be novel and specific sugarcane transcripts in leaf and internode tissues. This represented an increase of 1.8-fold compared to the SUCEST database, which contains around 43,000 sequences. Moreover, 38,195 contigs (19.5%) may represent sugarcane-specific transcripts since they did not match any analyzed public database [fig_ref] Figure 2: Number of grass transcripts mapping to sugarcane contigs [/fig_ref]. This percentage was similar to that in SUCEST (20%-25%), but much higher than reported for the tomato full-length library [bib_ref] Large-scale analysis of full-length cDNAs from the tomato (Solanum lycopersicum) cultivar Micro-Tom,..., Aoki [/bib_ref] , in which around 6% of all sequenced transcripts have no match.
After mapping all reads from each sample against the assembly, we identified only a few genotype-specific contigs, and less than 1.6% were not present in SP803280, i.e., were expressed only in Saccharum officinarum (0.16%), Saccharum spontaneum (0.69%), and both ancestors (0.73%), while 153,251 (78.28%) contigs were expressed in all three genotypes [fig_ref] Figure 4: Number of contigs expressed by genotype [/fig_ref] and. A note should be made here that tissues were sampled in parallel under similar conditions from green house grown plants in the case of ancestor genotypes and field-grown plants in the case of SP803280, therefore expression profiles should be noted with care. The divergence between Saccharum officinarum and Saccharum spontaneum took place 1.5-2 million years ago [bib_ref] Orthologous comparison in a gene-rich region among grasses reveals stability in the..., Jannoo [/bib_ref] , and this low number of genotype-specific transcripts (contigs) suggests that few major changes occurred in expressed regions of the genome. The genetic differences that account for major phenotypic differences (e.g., sucrose accumulation, yield, and stress tolerance) may be related to differentially expressed genes and alleles, single-nucleotide polymorphisms (SNPs), splicing variants, and so on. This hypothesis is in agreement with recent work on the sucrose synthase gene family in ancestor genotypes [bib_ref] Haplotype analysis of sucrose synthase gene family in three Saccharum species, Zhang [/bib_ref] , which exhibits significant differences in SNP frequency and resulting minor alterations in protein sequences. In addition, polymorphisms are more frequent in Saccharum spontaneum than in Saccharum officinarum and hybrid genotypes [bib_ref] A Survey Sequence Comparison of Saccharum Genotypes Reveals Allelic Diversity Differences, Berkman [/bib_ref]. Compared to sorghum, which diverged from sugarcane 8-9 million years ago [bib_ref] Orthologous comparison in a gene-rich region among grasses reveals stability in the..., Jannoo [/bib_ref] , we observed greater differences in the presence/absence of transcripts; 121,562 contigs matched sorghum genes [fig_ref] Figure 2: Number of grass transcripts mapping to sugarcane contigs [/fig_ref] , indicating that 37.9% of sugarcane contigs differed from sorghum transcripts. Further analysis on these results will improve our understanding of genome evolution and grass species divergence.
BLASTN comparisons between the two sugarcane de novo assemblies described in this present work (195,765 contigs) and by a previous study [bib_ref] De novo assembly and transcriptome analysis of contrasting sugarcane varieties, Cardoso-Silva [/bib_ref] (72,269 unigenes) revealed 15% and 23% of contigs and unigenes, respectively, that did not match (data not shown). This high number of unmatched transcripts between these two assemblies suggests that sugarcane might present a very complex transcriptome, with possible variety-specific transcripts. However, the environmental conditions and tissues harvested were not the same in these two studies, and therefore, sampling conditions may partly explain these differences.
## Functional annotation of full-length sugarcane cdna libraries
Functional annotation was carried out for sugarcane contigs based on Gene Ontology (GO) [bib_ref] Gene ontology: tool for the unification of biology. The Gene Ontology Consortium, Ashburner [/bib_ref] , Kyoto Encyclopedia of Genes and Genomes (KEGG) [bib_ref] KEGG for integration and interpretation of large-scale molecular data sets, Kanehisa [/bib_ref] , PFAM [bib_ref] The Pfam protein families database, Punta [/bib_ref] , and Phytozome (v9.0) (www.phytozome.net) databases . We found 78,273 (40%) contigs with categories in GO terms. The three main categories, i.e., ''cellular component'', ''molecular function'', and ''biological process'', showed ''nucleus'', ''protein binding'', and ''regulation of transcription, DNA dependent'', respectively, as the most frequent categories [fig_ref] Table 4: Protein prediction of sugarcane contigs [/fig_ref]. Compared to the sugarcane transcriptome annotation by Cardoso-Silva [bib_ref] De novo assembly and transcriptome analysis of contrasting sugarcane varieties, Cardoso-Silva [/bib_ref] (14,983 unigenes annotated in GO terms), a higher number of transcripts was annotated by the GO database, even when using the same criteria as used in the previous study [bib_ref] De novo assembly and transcriptome analysis of contrasting sugarcane varieties, Cardoso-Silva [/bib_ref] (i.e., minimum contig length and excluding isoforms; 24,085 transcripts, data not shown). KEGG pathway annotation yielded 14,298 contigs assigned into 132 pathways; 7,385 and 2,761 of these contigs were assigned into ''metabolic pathways'' and ''biosynthesis of secondary metabo-lites'', respectively [fig_ref] Table 4: Protein prediction of sugarcane contigs [/fig_ref]. Using ESTScan [bib_ref] ESTScan: a program for detecting, evaluating, and reconstructing potential coding regions in..., Iseli [/bib_ref] we found protein prediction for 187,935 (96%) contigs [fig_ref] Table 4: Protein prediction of sugarcane contigs [/fig_ref] and these proteins showed a mean length of 188 amino acids. These predicted protein sequences were compared with PFAM domain database, obtaining 57,526 matching proteins, with protein kinase (Pkinase) as the most frequent domain [fig_ref] Table 4: Protein prediction of sugarcane contigs [/fig_ref]. Next, using the results from mapping sugarcane contigs against grasses [fig_ref] Figure 2: Number of grass transcripts mapping to sugarcane contigs [/fig_ref] , categories from grass best hits were retrieved and assigned to the respective matching sugarcane contig, resulting in 139,980 contigs with 6,274 different categories . ''Unknown'', ''NB-ARC domain'', and ''protein kinase superfamily'' were the three most frequent categories [fig_ref] Table 4: Protein prediction of sugarcane contigs [/fig_ref]. Interestingly, although the plants were not noticeably stressed at the time of sampling, we identified enriched stress-related categories, such as ''response to salt stress'' and ''defense response'' in GO and NB-ARC domain [bib_ref] Structure-function analysis of the NB-ARC domain of plant disease resistance proteins, Van Ooijen [/bib_ref] in Phytozome and PFAM databases [fig_ref] Table 4: Protein prediction of sugarcane contigs [/fig_ref]. We selected the categories of genotype-specific contigs [fig_ref] Table 5: Number of full-length transcripts identified by each analysis [/fig_ref] , focusing in contigs expressed only in leaves, to identify differences among genotypes. [fig_ref] Figure 5: Top four categories of genotype-specific contigs based on Phytozome annotations [/fig_ref] shows the four most frequent categories, which were the same for all three genotypes. Saccharum spontaneum is the genotype responsible for introducing stress tolerance into hybrid sugarcane. The presence of several individual contigs containing the stress-related NB-ARC domain [bib_ref] Structure-function analysis of the NB-ARC domain of plant disease resistance proteins, Van Ooijen [/bib_ref] among genotype-specific transcripts, especially in Saccharum spontaneum (36 contigs), suggested their involvement in this phenotypic difference. Moreover, the presence of a high number of kinases and transcription factors reinforced the importance of signal transduction for phenotypic differences among Saccharum genotypes. All of these genotype-specific transcription factors were actually small contigs, ranging from 200 to 450 bp (data not shown), and were therefore unlikely to be complete genes. Instead, we consider that these may represent different variants of the same gene.
## Natural antisense transcripts (nats) appeared to affect several biological processes
Another important feature in sugarcane is the presence of NATs. NATs modulate the expression of their sense counterparts through several mechanisms [bib_ref] Genome-wide natural antisense transcription: coupling its regulation to its different regulatory mechanisms, Lapidot [/bib_ref]. NAT expression has been studied in several plants, such as maize [bib_ref] Comparative profiling of the sense and antisense transcriptome of maize lines, Ma [/bib_ref] , rice [bib_ref] Genome-wide identification and analysis of small RNAs originated from natural antisense transcripts..., Zhou [/bib_ref] , Arabidopsis [bib_ref] Distinct expression patterns of natural antisense transcripts in Arabidopsis, Henz [/bib_ref] [bib_ref] Genome-wide identification of long noncoding natural antisense transcripts and their responses to..., Wang [/bib_ref] , and Brassica rapa [bib_ref] Global analysis of cis-natural antisense transcripts and their heat-responsive nat-siRNAs in Brassica..., Yu [/bib_ref]. Studies have also suggested that [bib_ref] The libraries that made SUCEST, Vettore [/bib_ref] , and those that have been studied using oligoarrays in customized Agilent sugarcane chip (red) [bib_ref] Identification of sense and antisense transcripts regulated by drought in sugarcane, Lembke [/bib_ref] [bib_ref] Circadian rhythms of sense and antisense transcription in sugarcane, a highly polyploid..., Hotta [/bib_ref]. Green and red boxes show the number of transcripts present in the SUCEST data (green) and the Agilent Chip (red) but not in the sugarcane ORFeome (this work). doi:10.1371/journal.pone.0107351.g003 they may be important for regulating gene expression under stress conditions [bib_ref] Identification of sense and antisense transcripts regulated by drought in sugarcane, Lembke [/bib_ref] [bib_ref] Global analysis of cis-natural antisense transcripts and their heat-responsive nat-siRNAs in Brassica..., Yu [/bib_ref]. In sugarcane, NAT expression has been found to be responsive to drought [bib_ref] Identification of sense and antisense transcripts regulated by drought in sugarcane, Lembke [/bib_ref] and is modulated in a circadian manner [bib_ref] Circadian rhythms of sense and antisense transcription in sugarcane, a highly polyploid..., Hotta [/bib_ref]. Since sugarcane does not have a draft genome with predicted gene models, antisense identification was performed here by mapping the reads against gene models of other grasses using only Ion PGM reads as it is the only protocol used that maintained strand orientation. We detected NATs in all samples [fig_ref] Figure 6: Number of reads identified as natural antisense transcripts [/fig_ref] ; their expression was higher in leaf samples, which was particularly evident when the analysis was carried out in Brachypodium distachyon, Setaria italica, Sorghum bicolor, and SUCEST [fig_ref] Figure 6: Number of reads identified as natural antisense transcripts [/fig_ref]. These data suggested that some processes in leaves might be more dependent on or affected by antisense mechanisms than in internodes. However, this is different from data reported for Arabidopsis, where different tissues (roots, leaves, and inflorescence) have comparable expression levels of NATs [bib_ref] Genome-wide identification of long noncoding natural antisense transcripts and their responses to..., Wang [/bib_ref]. Furthermore, we identified no significant differences in total NAT expression among ancestral genotypes and the commercial hybrid [fig_ref] Figure 6: Number of reads identified as natural antisense transcripts [/fig_ref] , suggesting that this feature was conserved among Saccharum species, regardless of domestication and genome hybridization. Using SUCEST as reference, adding up all antisense reads in all five libraries, there are close to 1,700,000 antisense reads, which means ,5.8% of antisense reads (1,700,000/29,260,184).
Recently, NATs were predicted to represent approximately 70% of all Arabidopsis protein-coding loci, suggesting that NATs are much more widespread than previously thought [bib_ref] Genome-wide identification of long noncoding natural antisense transcripts and their responses to..., Wang [/bib_ref]. Here, antisense mapping yielded 28,884 contigs (14.7%) with antisense reads [fig_ref] Figure 7: Number of contigs and percentage of coverage by antisense reads [/fig_ref] , a much lower percentage than estimated for Arabidopsis. To obtain insights into the roles of these NATs, contigs showing NAT expression were assigned into biological processes by mapping them against the KEGG pathway database to cluster different proteins or enzymes of a given pathway into a single ''expression'' value for this KEGG pathway [fig_ref] Figure 1: Full-length enrichment for library cloning and next generation sequencing [/fig_ref]. A wide range of pathways is represented, with 94% (126/134) of plant KEGG pathways showing antisense expression, suggesting that NATs were widely present in the sugarcane transcriptome. Leaf-specific pathways, such as ''carbon fixation'' and ''photosynthesis'', showed NAT expression [fig_ref] Figure 1: Full-length enrichment for library cloning and next generation sequencing [/fig_ref] , correlating with the higher number of NATs in leaves [fig_ref] Figure 6: Number of reads identified as natural antisense transcripts [/fig_ref]. Pathways with higher ''expression'' of NATs were different in each genotype. Saccharum spontaneum showed ''alanine, aspartate, and glutamate metabolism'' as NAT-enriched pathway, whereas SP803280 and Saccharum officinarum showed ''indole alkaloid biosynthesis'' and ''lipoic acid metabolism'', respectively.
## A dataset of 17,563 unique sugarcane full-length genes
We conducted a three-step analysis to estimate the number of full-length transcripts in our ORFeome. Analysis I [fig_ref] Table 5: Number of full-length transcripts identified by each analysis [/fig_ref] considered a contig as full-length if a single contig covered more than 95% of a grass CDS with an identity of $80%, yielding 9,960 (9,410 unique) sugarcane full-length contigs. After manual analysis of alignment results, we observed that several full-length grass genes were covered by two or more contigs [fig_ref] Figure 2: Number of grass transcripts mapping to sugarcane contigs [/fig_ref]. These contigs represent a single transcript that was not assembled . Functional annotation of sugarcane full-length cDNA contigs. together, probably due to the high stringency of the assembler, problems in de novo assembly, or lack of sequencing depth. Therefore, we carried out a second analysis (Analysis II, [fig_ref] Table 5: Number of full-length transcripts identified by each analysis [/fig_ref] using the same criteria of coverage and identity cited above, but considering the non-overlapping coverage of different contigs in a given gene. This analysis did not take into account the contigs identified in Analysis I and resulted in 26,384 contigs representing 4,027 genes (3,952 unique genes), with an average of 6.55 contigs per full-length gene. Finally, using the remaining contigs, Analysis III searched for contigs with a complete predicted ORF (start and stop codons) and took into account the average CDS length of fulllength grass genes [fig_ref] Table 2: Data from Ion PGM and 454 sequencing [/fig_ref] , comparing it with the length of the predicted ORF of the sugarcane contig. If the CDS size was bigger than the average CDS length, the contig was considered fulllength; 4,243 contigs fit these criteria. As a result of these three analyses, we identified 40,587 contigs representing 17,563 unique, full-length genes [fig_ref] Table 5: Number of full-length transcripts identified by each analysis [/fig_ref]. Assuming 33,000 as the approximate total number of genes in sugarcane [bib_ref] GRASSIUS: a platform for comparative regulatory genomics across the grasses, Yilmaz [/bib_ref] , we identified 53.2% of all sugarcane full-length sequences. This number of unique full-length transcripts is much higher than reported for other plant full-length libraries, such as poplar [bib_ref] Analysis of 4,664 high-quality sequence-finished poplar full-length cDNA clones and their utility..., Ralph [/bib_ref] , tomato [bib_ref] SNP genotyping allows an in-depth characterisation of the genome of sugarcane and..., Garcia [/bib_ref] [bib_ref] Large-scale analysis of full-length cDNAs from the tomato (Solanum lycopersicum) cultivar Micro-Tom,..., Aoki [/bib_ref] , and Brachypodium (10,513) [bib_ref] Large-scale collection and analysis of full-length cDNAs from Brachypodium distachyon and integration..., Mochida [/bib_ref] , even though the estimated number of total genes in the genome is similar among these plants (34,000 for tomato, 31,000 for Brachypodium, and 33,000 for sugarcane).
Notably, identification of full-length sugarcane genes was carried out by de novo assembly in this study, in contrast to the full-length libraries cited above, which used an available draft genome to facilitate the assembly. Among all 40,587 contigs identified in the full-length analysis [fig_ref] Table 5: Number of full-length transcripts identified by each analysis [/fig_ref] , 40,407 showed a match against the Phytozome database [fig_ref] Table 4: Protein prediction of sugarcane contigs [/fig_ref]. Unknown proteins, NB-ARC domains, and protein kinases were the most frequent categories [fig_ref] Figure 8: Functional annotation of full-length contigs [/fig_ref] , similar to the results for all annotated contigs. The length of both sugarcane ORFs and transcripts follows the length distribution of other grasses full-length genes, especially when considered lengths over 1,500 nt for ORFs [fig_ref] Figure 9: Length distribution of sugarcane ORFs, full-length transcripts [/fig_ref] and 2,000 nt for transcripts [fig_ref] Figure 9: Length distribution of sugarcane ORFs, full-length transcripts [/fig_ref]. Almost 40% of sugarcane ORFs presents length ranging from 1,000 to 1,500 nt [fig_ref] Figure 9: Length distribution of sugarcane ORFs, full-length transcripts [/fig_ref] , whereas close to 60% of sugarcane full-length transcripts has 1,000-2,000 nt in length [fig_ref] Figure 9: Length distribution of sugarcane ORFs, full-length transcripts [/fig_ref]. Besides our analysis for fulllength identification primarily considered only CDS coverage as mentioned above, we can see that sugarcane full-length transcripts possess 59 and 39 UTRs [fig_ref] Figure 9: Length distribution of sugarcane ORFs, full-length transcripts [/fig_ref] , which reinforces that our transcripts are indeed full-length sequences or, at least, comprises the entire CDS plus part of the UTRs.
A recent work with NATs in the Arabidopsis transcriptome showed that NATs are widespread throughout the genome and that 60% of all sense-antisense pairs overlap each other entirely [bib_ref] Genome-wide identification of long noncoding natural antisense transcripts and their responses to..., Wang [/bib_ref]. Therefore, we further analyzed our dataset to identify, using the full-length analysis I and III, ''full-length'' NATs in sugarcane. However, only 46 (0.16% of contigs showing NATs) full-length sugarcane transcripts were fully overlapped by their antisense counterpart [fig_ref] Table S6: List of ''full-length'' NATs and their respective annotation [/fig_ref]. This difference in the number of NATs (28,844) and full-length NATs [bib_ref] Large-scale collection and analysis of full-length cDNAs from Brachypodium distachyon and integration..., Mochida [/bib_ref] should be interpreted carefully, since our methodology enriched for polyadenylated RNAs and there is evidence that NATs tend to be non-polyadenylated, at least in mammals [bib_ref] Disclosing hidden transcripts: mouse natural sense-antisense transcripts tend to be poly(A) negative..., Kiyosawa [/bib_ref] [bib_ref] Epigenetic regulation of human cis-natural antisense transcripts, Conley [/bib_ref] and Arabidopsis thaliana [bib_ref] Disclosing hidden transcripts: mouse natural sense-antisense transcripts tend to be poly(A) negative..., Kiyosawa [/bib_ref]. Therefore, this result may suggest that sugarcane NATs, particularly full-length NATs, tend to be non-polyadenylated or rarely show polyA tail.
Our sugarcane ORFeome comprised 937 unique, full-length transcription factors (TFs)
TFs are key regulators that allow plants to differentially regulate gene expression profiles in response to internal or external stimuli. Sugarcane TFs have been identified by microarray experiments as differentially expressed in several conditions, indicating that different TFs function in a wide range of traits and responses. Differentially expressed TFs are involved in tissue specificity [bib_ref] Identification of transcripts associated with cell wall metabolism and development in the..., Casu [/bib_ref] [bib_ref] Transcription profiling of signal transduction-related genes in sugarcane tissues, Papini-Terzi [/bib_ref] , sucrose content [bib_ref] Sugarcane genes associated with sucrose content, Papini-Terzi [/bib_ref] , and responses to hormones [bib_ref] Signal transduction-related responses to phytohormones and environmental challenges in sugarcane, Rocha [/bib_ref] , cold [bib_ref] RNA expression profiles and data mining of sugarcane response to low temperature, Nogueira [/bib_ref] , elevated CO 2 [bib_ref] Elevated CO2 increases photosynthesis, biomass and productivity, and modifies gene expression in..., Souza [/bib_ref] , and drought [bib_ref] Identification of sense and antisense transcripts regulated by drought in sugarcane, Lembke [/bib_ref]. Here, we identified 3,399 contigs belonging to 49 TF families [fig_ref] Table S7: Number of transcription factors [/fig_ref]. Since the estimated gene models indicate that around 2,000 TFs are encoded in the sugarcane genome [bib_ref] GRASSIUS: a platform for comparative regulatory genomics across the grasses, Yilmaz [/bib_ref] , our ORFeome identified a good portion of TF variants and alleles. The most abundant TF families were ''far red-impaired response'' (FAR1like), ''MYB-related'', and ''DNA binding protein phosphatase'' (DBP). Among the 17,563 unique full-length transcripts, we identified 937 unique TFs [fig_ref] Table 5: Number of full-length transcripts identified by each analysis [/fig_ref] , therefore, we detected 46.8% (937 out of 2,000) of all unique sugarcane TFs as being full-length. These results are expected to facilitate the functional characterization of sugarcane TFs. Further analysis of these TFs would improve our knowledge of the relationship between TF expression and specific responses or traits of interest, particularly because we sampled Saccharum genotypes exhibiting a wide range of differing characteristics including stress tolerance, sugar content, yield and fiber content (Saccharum spontaneum and Saccharum officinarum are highly different for these traits). [bib_ref] Ultrafast and memoryefficient alignment of short DNA sequences to the human genome, Langmead [/bib_ref]. The graph shows the 20 most frequent categories from Phytozome annotation. In total, 5,038 categories were observed [fig_ref] Table 4: Protein prediction of sugarcane contigs [/fig_ref]. doi:10.1371/journal.pone.0107351.g008 Sugarcane transcripts were first mapped against full-length grass transcripts and were then assigned as full-length if the CDS coverage was $95% and the identity was $80% (Analysis I). In this analysis, we found 9,960 full-length contigs, which aligned against 29,914 grass genes. Analysis II took into account more than one contig that mapped to the same grass gene [fig_ref] Figure 2: Number of grass transcripts mapping to sugarcane contigs [/fig_ref] , and non-overlapped coverage was calculated. Those that fit in the criteria of coverage $95% and identity $80% were considered full-length. This analysis yielded 26,384 contigs representing 3,952 unique grass genes and therefore 3,952 full-length sugarcane transcripts. For Analysis III, we calculated the average CDS size based on six grasses [fig_ref] Table 2: Data from Ion PGM and 454 sequencing [/fig_ref] , and all contigs with a predicted protein larger than this average CDS size were considered to be fulllength. doi:10.1371/journal.pone.0107351.t005
# Conclusion
In summary, we added 78,008 new transcripts to the SUCEST-FUN Database, identified 38,195 sugarcane-specific transcripts, 17,563 unique full-length transcripts (including 937 full-length TFs), putative genotype-specific transcripts, and numerous NATs affecting 126 KEGG pathways. [fig_ref] Figure 1: Full-length enrichment for library cloning and next generation sequencing [/fig_ref] summarizes our results from this initial analysis of the sugarcane ORFeome and presents the next steps that should be carried out. The use of ancestor This ORFeome can be used for gene discovery related to a range of traits since over 5,000 different categories [fig_ref] Table 4: Protein prediction of sugarcane contigs [/fig_ref] have a full-length representative. Differentially expressed alleles in each sample and in the hybrid and their origin from each ancestral genotype can also be analyzed. Several types of polymorphisms and genetic variability can be further investigated. Both genome assembly and annotation can make use of this sugarcane ORFeome dataset to validate and improve results. TF, transcription factor; NAT, natural antisense transcript. doi:10.1371/journal.pone.0107351.g010 genotypes is important for the development of the Energycane, identification of genes of interest (gene targeting), and investigation of polymorphisms and will improve our knowledge on Saccharum genetic variability. Furthermore, the ORFeome described in this report will provide information regarding allelic expression and ancestry, especially those related to stress responses, since the ORFeome was enriched in genes from this category. The ORFeome will facilitate assembly of the sugarcane genome since transcriptome data can be used as guides to join, order, and orient genomic fragments [bib_ref] Scaffolding a Caenorhabditis nematode genome with RNA-seq, Mortazavi [/bib_ref] [bib_ref] L_RNA_scaffolder: scaffolding genomes with transcripts, Xue [/bib_ref]. Moreover, a high percentage of gene models (in silico prediction) are not experimentally verified or might be different from expressed genes [bib_ref] Large-scale collection and analysis of full-length cDNAs from Brachypodium distachyon and integration..., Mochida [/bib_ref] [bib_ref] Large-scale RACE approach for proactive experimental definition of C. elegans ORFeome, Salehi-Ashtiani [/bib_ref] [bib_ref] Annotation and expression profile analysis of 2073 full-length cDNAs from stress-induced maize..., Jia [/bib_ref] ; thus, genome annotation could rely on the sugarcane ORFeome to create a better-suited matrix for gene prediction.
In conclusion, this sugarcane ORFeome represents an important advancement in sugarcane biotechnology and will be of great importance for the improvement of different areas of sugarcane research, providing a suitable reference for genomic and transcriptomic comparative analyses. This database has been made available to the scientific community at the Sequence Read Archive (SRA-NCBI) under accession number SRP042605. [fig_ref] Figure 1: Full-length enrichment for library cloning and next generation sequencing [/fig_ref] Heat map of 'pathway activity' of natural antisense transcripts based on KEGG pathway annotation. (PDF) [fig_ref] Figure 2: Number of grass transcripts mapping to sugarcane contigs [/fig_ref] Three examples of sorghum genes covered by two or more sugarcane contigs, i.e., a full-length sugarcane gene formed by two or more contigs. Blue thick arrows denote sugarcane contigs. Lines with numbers at the top of each figure denote the position in sorghum chromosome. Gray and orange thick arrows denote UTRs and CDS part, respectively, of sorghum genes. All identities fit in the previous criteria of . = 80%. Alignments were carried out at phytozome.net against sorghum genome v2.1.Contigs with minor alignments were excluded from figure. A, alignment of two sugarcane contigs against sorghum 4-coumarate CoA:ligase gene; B, alignment of three sugarcane contigs against sorghum sucrose synthase gene; C, alignment of two sugarcane contigs against sorghum caffeic acid 3-O-methyltransferase gene. (PDF) Table S3 Number of contigs grouped by Samples. Contigs in each genotype were analyzed by mapping the reads of each sample against the assembly (contigs). Colored lines highlight: officinarum specific contigs (green), spontaneum specific contigs (red); SP803280 specific contigs (Gray); SP803280 leaf (dark gray); ancestral specific (spontaneum + officinarum) contigs (Blue). (XLSX)
## Supporting information
[fig] Figure 1: Full-length enrichment for library cloning and next generation sequencing (NGS). Full-length (blue line with 59 cap) or truncated (short blue line without 59 cap) mRNAs were reverse transcribed into first-strand cDNA using oligo-dT primers (red arrow). [/fig]
[fig] Figure 2: Number of grass transcripts mapping to sugarcane contigs. A, Percentage of total grass transcripts mapping to sugarcane contigs. B, Total grass transcripts mapping to sugarcane contigs (white bars) and total sugarcane contigs mapping to each grass database (black bars). C, Total sugarcane contigs mapping to grasses, Uniprot, and NR databases and total unmatched sugarcane contigs (putative sugarcane-specific transcripts). doi:10.1371/journal.pone.0107351.g002 [/fig]
[fig] Figure 3: Venn diagram comparing sugarcane transcripts as obtained by RNAseq (blue, this work), SUCEST (green) [/fig]
[fig] Figure 4: Number of contigs expressed by genotype. doi:10.1371/journal.pone.0107351.g004 [/fig]
[fig] Figure 5: Top four categories of genotype-specific contigs based on Phytozome annotations. Only contigs from leaf samples in each genotype were considered. doi:10.1371/journal.pone.0107351.g005 [/fig]
[fig] Figure 6: Number of reads identified as natural antisense transcripts (NATs), based on their orientation on alignment to grass genes. Tissues are indicated as follows: In1, immature internodes; In5, intermediate internodes; L, leaves. doi:10.1371/journal.pone.0107351.g006 [/fig]
[fig] Figure 7: Number of contigs and percentage of coverage by antisense reads. Non-overlapping contig coverage by antisense reads was calculated for all contigs showing antisense expression (28,844 contigs). doi:10.1371/journal.pone.0107351.g007 [/fig]
[fig] Figure 8: Functional annotation of full-length contigs [/fig]
[fig] Figure 9: Length distribution of sugarcane ORFs, full-length transcripts (FL), and UTRs. Graphs denote the comparison of sugarcane length distribution (gray bars) of ORFs (A) and full-length transcripts (B) to other grasses (colored lines). Length distribution of 59 and 39 UTRs (C, black and white bars, respectively) of sugarcane full-length transcripts is shown as well. doi:10.1371/journal.pone.0107351.g009 [/fig]
[table] Table 1: The libraries were sequenced by two different NGS platforms. Eight runs were carried out (five on an Ion PGM instrument [Life Technologies, Carlsbad, CA, USA] and three on a 454 Titanium instrument [Roche, Branford, CT, USA]), totaling 31,663,934 trimmed reads and 6.6 Gb of sequence [/table]
[table] Table 2: Data from Ion PGM and 454 sequencing (after trimming) and Trinity assembly output. [/table]
[table] Table 4: Protein prediction of sugarcane contigs. [/table]
[table] Table 5: Number of full-length transcripts identified by each analysis. Total unique full-length sugarcane transcription factors (TFs) 937 [/table]
[table] Table S1: Number of Unique Contigs based on 1st Hit vs Full-Length grasses Species with Coverage .95% and Identity .80%. Species: Sb: S. bicolor; Zm: Z. mays; Pv: P. virgatum; Si: S. italica; Os: O. sativa; Bd: B. distachyum. (XLSX) [/table]
[table] Table S2: Number of genes, full-length genes and CDS mean size in each species from Phytozome v9.0. (XLSX) [/table]
[table] Table S4: Annotation of sugarcane contigs in Gene ontology, KEGG, PFAM and Phytozome. Annotation of full-length contigs in Phytozome is shown separately. (XLSX) [/table]
[table] Table S5: Annotation of genotype-specific contigs in Phytozome. (XLSX) [/table]
[table] Table S6: List of ''full-length'' NATs and their respective annotation. (XLSX) [/table]
[table] Table S7: Number of transcription factors (TFs) and full-length TFs by family. (XLSX) [/table]
|
Sustained accuracy improvement in intraocular lens power calculation with the application of quality control circle
Accurate intraocular lens (IOL) power calculation is always a challenge in ophthalmology, and unoptimized process may lead to inaccurate refractive outcomes. Quality control circle (QCC) hasshown its success in many fields as a process management tool. However, its efficacy in ophthalmology remains unclear. Here we utilized the QCC method to optimize the process and evaluate its efficacy in improving the accuracy of IOL power calculation. After the QCC application, the percentage of eyes with achieved refractive outcomes within 0.5 diopter significantly increased from 63.2% to 80.8% calculated by Haigis formula and 59.2% to 75.8% by SRK/T formula in patients with normal axial length (AL) (22 mm ≤ AL < 26 mm). Although there were no statistically significant differences in patients with long AL by the two formulas (p = 0.886 and 0.726), we achieved an accuracy of 75% with the application of the PhacoOptics software, which was significantly higher than that using the other two formulas (p < 0.001). Our findings indicated that QCC optimized and standardized the process of IOL power calculation, thus improved the accuracy of IOL power calculation in patients who underwent cataract surgery.With the remarkable update of surgical technology and equipment, phacoemulsification combined with intraocular lens (IOL) implantation has become the most prevalent treatment for cataract. Although precision medicine is becoming an appealing concept, accurate IOL power calculation is always a challenge in ophthalmology, while the involvement of various departments, doctors and technicians made it more complicated 1-9 .In many hospitals, the technicians performed the preoperative examinations on cataract patients including measurement of axial length, anterior chamber depth and corneal curvature. Then according to the results of biometric instrument IOL-Master or Lenstar (A-scan results were utilized in patients who were unable to be measured by biometric instrument), the surgeon selected the formula, calculated the IOL power and postoperative refractive prediction. The final refractive results were finished by optometrists. Hence, every simple step during the process may cause errors, and unoptimized process will inevitably result in undesirable postoperative refractive outcomes 1,5,10,11 .As a process management and problem-solving technique, quality control circle (QCC) was firstly employed in business management and company operation in Japan 12-15 . With the joint effort by all members and referring to certain procedures, the group members can make the most of their advantages and cooperate with other related departments, thus finally improving the complex work flow and solve the problems in work and management13,14,16. Some studies evaluated the role of QCC in medical improvement and highlighted the success of its application in medical and hospital management[17][18][19]. However, to our knowledge, little literature focused on the unoptimized process of IOL power calculation and the application of QCC in the field of cataract. Therefore, the aim of this study was to apply the novel QCC method to optimize the IOL power calculation process under the QCC principle and evaluate its accuracy and efficacy in accuracy improvement. at our hospital was included (Without QCC group). After applying the optimized process assisted by QCC, 104 patients (151 eyes) that underwent the surgery by the same experienced surgeon (Yun-e Zhao) between March 2015 and January 2016 were regarded as the QCC group. We excluded eyes with pterygium, strabimus, epiretinal membrane, retinal detachment, uveitis, or a history of corneal or intraocular surgery. Patients with complications or a best corrected visual acuity less than 20/40 after surgery were also excluded. According to the axial length (AL), patients were divided into the normal AL (22 mm ≤ AL < 26 mm) and long AL (AL ≥ 26 mm) subgroups.
Process optimization and standardization. The QCC technique was applied to optimize and standardize the process of IOL power calculation. The retrospective data of 117 patients were reviewed to analyze the potential reasons based on fishbone analytical diagram method [fig_ref] Figure 1: Fishbone analytical diagram [/fig_ref]. Then every QCC members voted to choose the main target. Finally, three leading causes were selected as the main target of this QCC , which were respectively inaccurate biometric measurement, IOL and formula selection and optimization. Taking the feasibility, economy and circle capacity into consideration, two experienced doctors from the cataract department and 1 technician participated in the strategy assessment and graded each strategy based on a 5-point system. The strategy with the lowest score would not be selected in this study. According to the main reasons, sufficient circle meetings were hold to optimize the positive strategies as follows.
Firstly, normalize the biometric measure protocol for IOLMaster: (1) make sure that the examined eye of patients focused on the fixation lamp and use penlight to guide the fellow eye if the fixation is poor; [bib_ref] Benchmark standards for refractive outcomes after NHS cataract surgery, Gale [/bib_ref] delete the data with a deviation of more than 0.02 mm, and reexamine the data less than 5 groups; while the most important tip for curvature measurement is nice focus; then the notes for the A scan: (a) adjust the sonic velocity related to lens opacity; (b) ensure the favorable fixation of the patients with the red dot; (c) determine and record the final data based on the adjusted waveform in A scan, delete the data with an abnormal waveform or a deviation of more than 0.05 mm from the mean value, reexamine the data less than 10 groups, and assign another technician to finish the examination if the measurement results could not meet the standard mentioned above.
Secondly, optimize the IOL constant in the light of User Group for Laser Interference Biometry (ULIB, website: http://ocusoft.de/ulib/). Thirdly, select the ray-tracing assisted IOL calculation software PhacoOptics which was reported by Olsen when it comes to patients with long AL [bib_ref] C constant: new concept for ray tracing-assisted intraocular lens power calculation, Olsen [/bib_ref] [bib_ref] Ray-tracing analysis of intraocular lens power in situ, Olsen [/bib_ref]. During the QCC activity, the quality management cycle, namely PDCA (plan, do, check and action), should be followed throughly to achieve the sustained enhancement of strategy execution.
Main outcome measurement. The results of the subjective refraction at 3 months postoperatively were recorded as achieved refraction and expressed as the spherical equivalent (SE). The predicted refraction error was defined as the achieved refraction minus the predicted refraction result by each formula. The mean difference and absolute difference of the predicted refraction were calculated as the mean prediction error (ME) and mean absolute prediction error (MAE), respectively. Additionally, the percentage of eyes with a final MAE within 0.5D was regarded as an evaluation index of the accuracy.
Statistical analyses. The sample size in the QCC group was determined by PASS 11.0 software (NCSS Statistical Software, Kaysville, UT) based on 80% power to detect the effect of QCC. Then the data was analyzed by SPSS 18.0 software for Windows (SPSS Inc. Chicago, IL, U.S.). Normal distribution of the data was analyzed by the Kolmogorov-Smirnov test. Independent sample t test or non-parametric Wilcoxon was used to compare the differences of age, axial length and refractive results between the two groups. Pearson's chi-square test was used for the investigation of the patient's demographics and the accuracy between the two groups after the QCC application. P values lower than 0.05 were regarded as statistically significant. summarized the demographics of patients who underwent surgery before and after the QCC application. There were no significant differences in the sex ratio, age and axial length between the two groups. [fig_ref] Table 3: Refractive outcomes of Haigis and SRK/T formulas in patients with different axial... [/fig_ref] summarized the refractive outcomes of Haigis and SRK/T formulas in patients with different AL in the two groups. As shown in , there was significantly higher accuracy in patients with normal axial length after QCC when calculated by Haigis or SRK/T formulas, while there was no statistically significant difference in the patients with long axial length by the two regular formulas. However, we achieved a remarkable accuracy of IOL power prediction as high as 75.0% if the PhacoOptics software was used in patients with long AL, which was significantly higher than that using the other two formulas (p < 0.001).
# Results
## Patient demographics.
## Accuracy of iol power estimation.
QCC protocol. We established the QCC group and selected the IOL power calculation accuracy improvement as our goal. Then after analyzing the current status and related reasons, we formulated several strategies and executed them. Finally, the results were checked and the process was standardized. During the QCC activity, the PDCA circulation would continue until the results were effective. The ten steps protocol of QCC was shown in [fig_ref] Figure 2: Protocol of the quality control circle [/fig_ref] , and the badge of our QCC could be found in [fig_ref] Figure 1: Fishbone analytical diagram [/fig_ref].
Standardization. The standard operation flow [fig_ref] Figure 3: Standard flow of the intraocular lens [/fig_ref] was established and optimized by the continuous improvement of the QCC activity.
# Discussion
Cataract surgery is the only treatment for cataract patients at present, while accurate IOL power calculation continues to be a big challenge in ophthalmology, especially in primary hospitals [bib_ref] Calculation of intraocular lens power: a review, Olsen [/bib_ref] [bib_ref] Benchmark standards for refractive outcomes after NHS cataract surgery, Gale [/bib_ref] [bib_ref] Accuracy of 3 new methods for intraocular lens power selection, Kane [/bib_ref] [bib_ref] Comparison of 9 intraocular lens power calculation formulas, Cooke [/bib_ref] [bib_ref] Formula choice: Hoffer Q, Holladay 1, or SRK/T and refractive outcomes in..., Aristodemou [/bib_ref] [bib_ref] Optimizing intraocular lens power calculations in eyes with axial lengths above 25.0..., Wang [/bib_ref] [bib_ref] Precision medicine for metastatic breast cancer-limitations and solutions, Arnedos [/bib_ref] [bib_ref] A new initiative on precision medicine, Collins [/bib_ref] [bib_ref] Intraocular lens power calculation for eyes with an axial length greater than..., Abulafia [/bib_ref] [bib_ref] Intraocular lens formula constant optimization and partial coherence interferometry biometry: Refractive outcomes..., Aristodemou [/bib_ref] [bib_ref] Intraocular lens power calculation using the IOLMaster and various formulas in eyes..., Wang [/bib_ref]. However, unoptimized process of IOL power calculation may lead to inaccurate refractive outcomes. Process optimization has drawn great attention worldwide for its potential in accuracy improvement [bib_ref] Intraocular lens formula constant optimization and partial coherence interferometry biometry: Refractive outcomes..., Aristodemou [/bib_ref] [bib_ref] Intraocular lens power calculation using the IOLMaster and various formulas in eyes..., Wang [/bib_ref] [bib_ref] Optimizing glaucoma screening in high risk population: design and 1-year findings of..., Zhao [/bib_ref] [bib_ref] Intraocular lens power calculation and optimized constants for highly myopic eyes, Petermeier [/bib_ref]. Nevertheless, QCC, as a process optimization tool, has never been used in the field of ophthalmology. Therefore, we firstly performed this study to optimize the IOL power calculation process by utilizing QCC technique and evaluated its efficacy in improving the accuracy of IOL power calculation.
Through our retrospective data analyses, we found that before QCC, the percentages of eyes within 0.5D of target refraction were respectively 63.2%/59.2% (Haigis/SRK/T formula), which were consistent with the benchmark standards [bib_ref] A new initiative on precision medicine, Collins [/bib_ref]. Although they were lower than other literature 3-5 , we thought they were acceptable as the baseline since they were obtained under conditions with unoptimized process including biometric measurement, A constant and formula selection. Based on the QCC technique, our study analyzed the potential reasons related to the low accuracy by data review and brainstorm. In order to optimize the process and achieve a higher level of accuracy of IOL power estimation, we laid down relevant solutions to realize this goal. Given that feasibility, economy and circle capacity, three leading causes were selected as the main target (see . In China or other third world countries, patients are more prone to undergo the . Accuracy of intraocular lens power prediction before and after the quality control circle activity. QCC = Quality control circle, AL = Axial length. *The P value comes from comparing the results of PhacoOptics with those of the Haigis and SRK/T formulas after the QCC activity. cataract surgery until the cataract is mature, thus A scan or other enhanced optical measure methods will be required. Therefore, accurate biometric measurement and standardized process are essential and critical. For the purpose of more precise biometric measurement, we formulated the standard operation procedure such as using the average of repeated measurement and/or checking the results by different operators, especially in patients performed by A scan. In fact, no matter how well we controlled the protocol, the precision of the A scan is inferior to the optical measurement, and the axial length performed by A scan will influence the IOL power calculation more or less. In terms of IOL 23 , we optimized our routine IOL type, A constant and formula by refer to the ULIB. With respect to formula selection in patients with abnormal axial length especially long axial length, we found that researchers have introduced the C constant as a new concept for the ray-tracing assisted IOL calculation software PhacoOptics in the latest literature, which indeed bettered the prediction of the IOL power [bib_ref] C constant: new concept for ray tracing-assisted intraocular lens power calculation, Olsen [/bib_ref] [bib_ref] Ray-tracing analysis of intraocular lens power in situ, Olsen [/bib_ref]. Therefore, we applied this method in IOL power calculation in patients with long axial length. It turned out that the accuracy was 75%, which was consistent with other literature [bib_ref] Accuracy of 3 new methods for intraocular lens power selection, Kane [/bib_ref] [bib_ref] Formula choice: Hoffer Q, Holladay 1, or SRK/T and refractive outcomes in..., Aristodemou [/bib_ref] [bib_ref] Optimizing intraocular lens power calculations in eyes with axial lengths above 25.0..., Wang [/bib_ref] [bib_ref] Intraocular lens power calculation for eyes with an axial length greater than..., Abulafia [/bib_ref] [bib_ref] Intraocular lens formula constant optimization and partial coherence interferometry biometry: Refractive outcomes..., Aristodemou [/bib_ref] [bib_ref] Intraocular lens power calculation and optimized constants for highly myopic eyes, Petermeier [/bib_ref].
In accordance with the PDCA cycle of the QCC activity, the accuracy of predicted IOL power increased from 63.2%/59.2% (Haigis/SRK/T formula) to 80.8%/75.8% respectively in patients with normal axial length and the results were similar with other studies [bib_ref] Accuracy of 3 new methods for intraocular lens power selection, Kane [/bib_ref] [bib_ref] Optimizing intraocular lens power calculations in eyes with axial lengths above 25.0..., Wang [/bib_ref] [bib_ref] Intraocular lens power calculation for eyes with an axial length greater than..., Abulafia [/bib_ref]. Although there was no difference between the two formulas after QCC in patients with long axial length, the utilization of PhacoOptics software led to a dramatic growth in accuracy. In patients with normal axial length, we attributed the enhancement of accuracy to the standardization of axial length measurement, IOL constant and formula optimization, indicating that the whole process optimization could lead to accuracy improvement. Regarding to the IOL power calculation in patients with long axial length, both the Haigis and SRK/T formulas would take the axial length and anterior chamber depth into consideration, while the PhacoOptics software was independent of axial length as it was based on the preoperative lens thickness and anterior chamber depth to predict the postoperative IOL position [bib_ref] C constant: new concept for ray tracing-assisted intraocular lens power calculation, Olsen [/bib_ref] [bib_ref] Ray-tracing analysis of intraocular lens power in situ, Olsen [/bib_ref]. In addition, as an optical formula, PhacoOptics software attached great importance to the true net corneal power, realistic geometric position and corneal asphericity. Furthermore, previous study has demonstrated that C constant was an unbiased concept to predict the effective lens position 20 . Hence, we owed the success in patients with long axial length to the ray tracing method. Although the ray tracing method demonstrated its potential and efficacy in accuracy improvement, we did not apply it in patients with normal AL because of the desirable accuracy by Haigis/SRK/T formula, high manpower costs and cost-benefit analysis. In this study, our department utilized the scientific management approach as a tool, established the QCC group, executed the strategies under the PDCA principle, and obtained desirable results, which was a pioneering attempt and validated the importance of QCC in ophthalmology. It not only gave rise to accuracy, but also promoted the formation of standardization and a smooth process flow. Furthermore, it increased the satisfaction of patients as well. As we know, the worsening of doctor-patient relationship is becoming an increasing threat in China due to misunderstanding and distrust [bib_ref] Facing up to the threat in China, Huang [/bib_ref] [bib_ref] A gloomy future for medical students in China, Zeng [/bib_ref]. The dramatic growth in accuracy could bring relief in the deteriorated condition and facilitate the harmonious coexistence between patients and doctors.
Promoting the QCC activity can not only obtain the tangible achievements such as accuracy improvement, but also produce some intangible achievements. First, the QCC activity could draw out the capacity and potential ability of the group. Circle members are willing to spend more time, strength and creativity to realize the self-management and harmonious development of the group. In the modern society, the pattern of solving problems by a single department is no longer an ideal method. Additionally, allowing for the involvement of the QCC activity with many different departments, a sense of responsibility and coordination will be promoted. Overall, the more active the QCC is, the better results we can achieve, and everyone can benefit from the process. Moreover, the QCC activity exerted some certain effect to encourage the work morale and attitudes. Although this investigation only performed in cataract patients, we hope that our novel discovery will bring more studies to further explore the function of QCC in ophthalmology.
In spite of the success in this study, several limitations should also be noted. First, a large and multi-center study was definitively needed to verify the efficacy and accuracy of QCC. In addition, as limited by economy and circle capacity, abnormal anterior chamber depth modification was not performed in the present study. Thus, further investigations were still required for the benefits of complex cases.
In summary, this study showed the potentially important role that optimization of process has in the IOL power calculation in patients who underwent cataract surgery. The results proved the efficacy and accuracy of QCC in IOL power calculation process optimization and accuracy improvement. The application of QCC really made a difference from the conventional methods, indicating the feasibility of extended utilization of QCC in more fields of ophthalmology, clinical practice and even the scientific research. Furthermore, the successful experience of QCC could be beneficial to the establishment of primary eye hospitals. With the current medical climate of precision medicine and personalized medicine [bib_ref] Intraoperative aberrometry versus preoperative biometry for intraocular lens power selection in axial..., Hill [/bib_ref] [bib_ref] Protocols for studies of intraocular lens formula accuracy, Hoffer [/bib_ref] [bib_ref] A gloomy future for medical students in China, Zeng [/bib_ref] , there is no doubt that QCC should be considered as a promising approach to achieve a higher level of efficacy and precision in medical and research field.
Meeting. This study was presented at the Annual Meeting of the Association for Research in Vision and Ophthalmology (ARVO) during May 7 to May 11, 2017 in Baltimore, Maryland, and selected as a hot topic.
[fig] Figure 1: Fishbone analytical diagram. This fishbone diagram includes main reasons that lead to the low accuracy of intraocular lens (IOL) power prediction. [/fig]
[fig] Figure 2: Protocol of the quality control circle. The ten steps and PDCA (plan, do, check and act) flow of the quality control circle. [/fig]
[fig] Figure 3: Standard flow of the intraocular lens (IOL) power calculation process. After quality control circle, the process of the IOL power calculation was standardized for cataract patients with different axial lengths (AL). [/fig]
[table] Table 3: Refractive outcomes of Haigis and SRK/T formulas in patients with different axial lengths in the two groups. QCC = Quality control circle, AL = Axial length, ME = Mean prediction error, MAE = Mean absolute prediction error. *The P value comes from comparing the results of PhacoOptics software with those of the Haigis and SRK/T formulas after the QCC activity. [/table]
|
2-hydroxylated sphingomyelin profiles in cells from patients with mutated fatty acid 2-hydroxylase
Fatty acid 2-hydroxylase (FA2H) is the enzyme responsible for the hydroxylation of free fatty acids prior to their incorporation into 2-hydroxylated sphingolipids, which are the major constituents of the myelin leaflet. Mutated FA2H has been associated with neurodegenerative diseases. Decreased FA2H activity was demonstrated only in vitro, but not in patient tissues. In this study we characterized the 2-hydroxylated sphingomyelin (SM) profiles in blood and fibroblasts from patients harboring a deleterious FA2H mutatation, and found that hydroxylated fatty acid sphingomyelin is present in normal amounts in patient lymphocytes, but decreased to a different extent in fibroblasts and erythrocytes.
# Background
A compact myelin sheet surrounding the axon is essential for correct nerve conduction. Myelin is composed of over 70% lipids, the most abundant of which are the galactolipids galactosylceramide (GalC) and its sulfated form, sulfatide. As more than fifty percent of GalC and sulfatide are hydroxylated at the C2 position on the fatty acid (FA) moiety, it has been estimated that approximately twenty five percent of the outer leaflet lipids in myelin are hydroxylated [bib_ref] Occurrence of 2-hydroxy fatty acids in animal tissues, Kishimoto [/bib_ref] [bib_ref] Ceramide galactoside of enriched neuronal and glial fractions from rat brain, Raghavan [/bib_ref]. Based on studies of model membranes, hydroxylation likely contributes to the stability of myelin by virtue of the hydrogen bonding between the hydroxy groups, the galactose head group and the polar part of the ceramide backbone [bib_ref] Influence of structural modifications on the phase behavior of semi-synthetic cerebroside sulfate, Boggs [/bib_ref].
The 2-hydroxylation of sphingolipids [for review, see [bib_ref] Fatty acid 2-Hydroxylation in mammalian sphingolipid biology, Hama [/bib_ref] ] occurs during de novo ceramide synthesis and is catalyzed by the enzyme fatty acid 2-hydroxylase (FA2H), a membrane-bound protein containing a cytochrome b5-like heme-binding domain responsible for the redox activity, and a sterol desaturase domain [bib_ref] The human FA2H gene encodes a fatty acid 2-hydroxylase, Alderson [/bib_ref]. The products of FA2H are free 2-hydroxy fatty acids (OHFA), which are subsequently incorporated into ceramide, the precursor of galactosylceramide [bib_ref] Fatty acid 2-hydroxylase, encoded by FA2H, accounts for differentiation-associated increase in 2-OH..., Uchida [/bib_ref]. FA2Hdeficient mice lacking hydroxylated fatty acids in the central and peripheral nervous system had normal neuronal development, but showed late onset axon and myelin sheet degeneration [bib_ref] Absence of 2-hydroxylated sphingolipids is compatible with normal neural development but causes..., Zöller [/bib_ref] and exhibited CNS dysfunction.
Mutations in the human FA2H gene were identified in patients with autosomal recessive leukodystrophy characterized by childhood (4-5 y) onset spasticity, dystonia and white matter degeneration. In seven patients, the desaturase domain was disrupted by a splice site mutation causing the skipping of exons 5 and 6 while a missense mutation in a conserved residue was detected in the other two patients [bib_ref] Mutations in the fatty acid 2-hydroxylase gene are associated with leukodystrophy with..., Edvardson [/bib_ref]. Recently, mutated FA2H was also found to be the underlying cause of a complicated hereditary spastic paraplegia (SPG35) [bib_ref] Mutation of FA2H underlies a complicated form of hereditary spastic paraplegia (SPG35), Dick [/bib_ref] , and neurodegeneration with brain iron accumulation. Although in vitro transfection studies of mutated FA2H disclosed reduced hydroxy fatty acid synthesis, decreased enzymatic activity was not demonstrated in patients. In fact, tetracosanoic acid hydroxylating activity in patient fibroblasts was indistinguishable from that of normal controls [bib_ref] Mutations in the fatty acid 2-hydroxylase gene are associated with leukodystrophy with..., Edvardson [/bib_ref]. Moreover, the impact of defective FA2H on fatty acid composition in patients is (to our knowledge) unknown. The aim of this study was to characterize the functional impact of FA2H splice site mutation on the fatty acid and hydroxy fatty acid sphingomyelin profiles in patient's blood and fibroblasts.
# Methods
Cell culture reagents were obtained from Biological Industries, Beit HaEmek, Israel. Hisopaque and all other reagents were from Sigma-Aldrich, Israel.
## Subjects
Blood from two patients and fibroblasts from one patient harboring the FA2H c.786+1G A mutation (family 1, described by Edvardsson et al [bib_ref] Mutations in the fatty acid 2-hydroxylase gene are associated with leukodystrophy with..., Edvardson [/bib_ref] and from three controls were obtained with informed consent and approval from the local IRB.
## Lymhocyte isolation
Lymphocytes were isolated from whole blood (EDTA anti-coagulant) using Histopaque-1077 according to the manufacturer's instructions. The resulting pellet was lyophilized prior to lipid analysis.
## Erythrocyte membrane preparation
Erythrocyte membranes were prepared from whole blood collected with heparin as anti-coagulant. The blood was spun at 1000 × g for 10 min, the plasma was removed and the erythrocytes were washed three times with twice their volume of saline. Subsequently, cells were hemolyzed in 5 ml of water then spun at 12000 × g for 10 minutes. The membranes were washed twice with water and lyophilized prior to lipid analysis.
## Cell culture
Primary fibroblasts were grown in DMEM (high glucose) supplemented with 15% fetal calf serum in the presence of penicillin and streptomycin. Confluent cells were harvested by trypsinization, washed with phosphate buffered saline (PBS) and lyophilized prior to lipid analysis.
# Lipid analysis
Lipid analysis was performed at the Lipidomics Core of the Medical University of South Carolina using HPLC/ MS-MS as previously described [bib_ref] Simultaneous quantitative analysis of bioactive sphingolipids by high-performance liquid chromatography-tandem mass spectrometry, Bielawski [/bib_ref] [bib_ref] Comprehensive quantitative analysis of bioactive sphingolipids by highperformance liquid chromatography-tandem mass spectrometry, Bielawski [/bib_ref]. All values are reported normalized to mg protein as determined using the Lowry method [bib_ref] Protein measurement with the Folin phenol reagent, Lowry [/bib_ref].
## Rt-pcr
Total RNA from lymphoblasts was prepared using TRI reagent (Sigma Aldrich) and equal amounts were reverse transcribed using ImProm-II (Promega, Wisconsin, USA) reverse transcriptase kit with a hexamer mixture as the template primer according to the manufacturer's instructions. Primer sequences for PCR analysis available upon request.
# Results and discussion
We have previously shown that the c.786+1G A mutation causes mis-splicing leading to skipping of exons 5 and 6 in fibroblasts. RT-PCR was perfomed in order to verify the exon skipping in lymphocytes. The expected 426bp shorter transcript, obtained in the patient, corroborated the exon skipping as shown in [fig_ref] Figure 1: Lymphocyte FA2H patient transcript and total FA-SM/OH-SM profiles [/fig_ref].
Subsequently we proceeded to investigate the sphingomyelin (SM) profiles in these cells as well as in erythrocyte membranes and fibroblasts [fig_ref] Figure 1: Lymphocyte FA2H patient transcript and total FA-SM/OH-SM profiles [/fig_ref].
The total SM-fatty acid content in patient fibroblasts and lymphocytes was not significantly different from that of control subjects, while patient erythrocytes had a relatively higher SM-fatty acid content [fig_ref] Figure 1: Lymphocyte FA2H patient transcript and total FA-SM/OH-SM profiles [/fig_ref]. SMhydroxy fatty acid content in patient lymphocytes was not significantly different from that of the controls, however it was reduced by 50% in patient fibroblasts and increased in patient erythrocytes [fig_ref] Figure 1: Lymphocyte FA2H patient transcript and total FA-SM/OH-SM profiles [/fig_ref]. The decrease in SM-hydroxy fatty acid content in fibroblasts was especially evident when the ratio of SM-hydroxy fatty acid to SM-fatty acid was calculated. In patient erythrocytes this ratio was also significantly decreased as SM-hydroxy fatty acids were decreased relative to SMfatty acids [fig_ref] Figure 1: Lymphocyte FA2H patient transcript and total FA-SM/OH-SM profiles [/fig_ref].
The distribution of the SM-fatty acid [fig_ref] Figure 2: HO-C14-SM a-HO-C16-SM a-HO-C18-SM a-HO-C20-SM a-HO-C22-SM a-HO-Levels of saturated and unsaturated FA-SM and... [/fig_ref] , B) and SM-hydroxy fatty acids [fig_ref] Figure 2: HO-C14-SM a-HO-C16-SM a-HO-C18-SM a-HO-C20-SM a-HO-C22-SM a-HO-Levels of saturated and unsaturated FA-SM and... [/fig_ref] by chain length was investigated. C16 is by far the most abundant fatty acid in all cell types. No differences were seen between patient and control lymphocytes. SM-C16 fatty acid content was slightly increased in patient fibroblasts, and the levels of most SM-fatty acids were increased in patient erythrocytes. In all cell types, in control as well as patient, the distribution of SM-hydroxy fatty acids does not mirror that of their fatty acid counterpart. In lymphocytes, SM-OHC18 is the most abundant species, and the distribution of SM-hydroxy fatty acids is the same in the patient and the control. However, patient fibroblasts contain approximately 50% of SM-OHC16 and SM-OHC18 of the control values, and all the other SM-hydroxy fatty acid levels were significantly lower. The patient erythrocytes had an elevated content of all SM-hydroxy fatty acids, but with a significant increase of SM-OHC18 vs SM-OHC16.
The fraction of hydroxylated fatty acid relative to fatty acid (OHFA-SM/FA-SM) of the same chain length was calculated [fig_ref] Figure 2: HO-C14-SM a-HO-C16-SM a-HO-C18-SM a-HO-C20-SM a-HO-C22-SM a-HO-Levels of saturated and unsaturated FA-SM and... [/fig_ref]. This ratio is the same for both control and patient lymphocytes, where the ratio is highest for C18. In contrast, in patient fibroblasts, the fraction of hydroxylated C20:1 is only 1/10 that of the control value, with reductions for C18 and C20 as well. C20 is the most hydroxylated saturated fatty acid in both control and patient, whereas C20:1 is the most hydroxylated monounsaturated acid in the controls and C24:1 the most hydroxylated monounsaturated acid in the patient. In erthrocytes, the ratio is decreased for the patient vs the control for all chain lengths, most noticeably for C20 and C22, with C20 and C24:1 the most heavily hydroxylated acids.
As SM is the major sphingolipid in human cultured fibroblasts [bib_ref] Ceramide and sphingomyelin species of fibroblasts and neurons in culture, Valsecchi [/bib_ref] , and taking into account the limited amount of sample, we focused our studies on this sphingolipid in cells obtained from patients using minimally invasive sampling. The lack of difference in the SM fatty acid and hydroxy fatty acid profile between patient and control lymphocytes was unexpected. Apparently the lack of FA2H as demonstrated by RT-PCR had no effect. On the other hand, SM hydroxy fatty acid content in fibroblasts was clearly decreased, with respect to both the total content and the fraction of hydroxylated SM. The high OHFA content but low OHFA-SM/FA-SM ratio in erythrocytes suggests an imbalance, rather than a quantitative change. The distribution of OHFA according to chain length is different in the patient and control fibroblasts, with C20:1 being the most hydroxylated acid in the controls and C24:1 the most hydroxylated in the patient.
The results we obtained for FA content and distribution confirm previous results [bib_ref] Ceramide and sphingomyelin species of fibroblasts and neurons in culture, Valsecchi [/bib_ref] however in that paper the authors did not find any hydroxy fatty acids. They conclude that if present, they must be in amounts below their detection limit of a few pmoles. This is certainly not in agreement with the results presented here. The discrepancy could be due to differences in the sample preparation as well as analytical conditions.
The characterization of the FA and OHFA profiles was carried out also to identify differences between controls and patients which could possibly form the basis for a biochemical diagnostic test. The most significant difference was found in the relative decrease in OHFA-SM relative to FA-SM content in patient erythrocytes vs control. In cases where a skin biopsy is performed and fibroblasts are available, the decrease in saturated OHFA content coupled with the favoring of OHC24:1 in patients vs OHC20:1 in controls could also serve as a marker.
Although genetic analysis is more feasible, FA-SM analysis could complement the molecular data and provide valuable information where genetic analysis is uninformative. It is somewhat surprising that patient cells still contain a measurable amount of hydroxy fatty acids at all, considering the deleterious nature of the mutation in the FA2H gene. This would suggest presence of other enzyme/s with overlapping substrate specificity with FA2H. In fibroblasts and erythrocytes, which show a difference between the patient and the control, the 2-hydroxylating activity may be due to FA2H and to an additional enzyme with a different chain length preference. In lymphocytes, FA2H may be absent altogether while the proposed other enzyme performs the FA2-hydroxylation. Notably we were unable to detect any FA2H enzymatic activity in normal lymphocytes (results not shown).
# Conclusions
We conclude that OHFA-SM is decreased in patient fibroblasts and erythrocytes. Differences in FA-SM and OH-SM between patients and controls could serve as the basis for a diagnostic test.
[fig] Figure 1: Lymphocyte FA2H patient transcript and total FA-SM/OH-SM profiles. A, Lane C shows the transcript obtained in control lymphocytes using the FA2H primer and lane P shows the shorter transcript obtained in patient lymphocytes. B, Total level of FA-SM in control and patient fibroblasts, lymphocytes and erythrocytes was quantitated by HPLC/MS-MS. C, Total level of OHFA-SM in control and patient fibroblasts, lymphocytes and erythrocytes was quantitated by HPLC/MS-MS. D, The ratio of the total OHFA-SM content divided by the total FA-SM content was calculated (mean of 2 experiments +/-SD). [/fig]
[fig] Figure 2: HO-C14-SM a-HO-C16-SM a-HO-C18-SM a-HO-C20-SM a-HO-C22-SM a-HO-Levels of saturated and unsaturated FA-SM and OHFA-SM in the range C14-C24 in fibroblasts, lymphocytes and erythrocytes of control subjects and in patients were quantitated by HPLC/MS-MS. A, Saturated FA-SM. B, Unsaturated FA-SM.C, Saturated OHFA-SM. D, Unsaturated OHFA-SM. E, Ratio saturated OHFA -SM/FA-SM. F. Ratio unsaturated OHFA -SM/FA-SM (mean of 2 experiments +/-SD). [/fig]
|
Oxidative Stress and Mitochondrial Activation as the Main Mechanisms Underlying Graphene Toxicity against Human Cancer Cells
Due to the development of nanotechnology graphene and graphene-based nanomaterials have attracted the most attention owing to their unique physical, chemical, and mechanical properties. Graphene can be applied in many fields among which biomedical applications especially diagnostics, cancer therapy, and drug delivery have been arousing a lot of interest. Therefore it is essential to understand better the graphene-cell interactions, especially toxicity and underlying mechanisms for proper use and development. This review presents the recent knowledge concerning graphene cytotoxicity and influence on different cancer cell lines.
## Graphene: properties and applications
Novoselov et al. first described graphene in 2004 as monocrystalline graphitic film and received Nobel Prize in 2010 for the exploration of its exceptional properties [bib_ref] Electric field in atomically thin carbon films, Novoselov [/bib_ref]. The discovery of graphene became a new driving force in the development of nanoindustry [bib_ref] Toxic response of graphene nanoplatelets in vivo and in vitro, Park [/bib_ref] [bib_ref] Pharmaceutical applications of graphene-based nanosheets, Kim [/bib_ref]. Graphene is a single-atom-thick, two-dimensional sheet of sp 2 -hybridized carbon atoms arranged in a regular hexagonal pattern like in honeycomb structure [bib_ref] A systems toxicology approach to the surface functionality control of graphene-cell interactions, Chatterjee [/bib_ref] [bib_ref] An in vitro evaluation of graphene oxide reduced by Ganoderma spp. in..., Gurunathan [/bib_ref] [bib_ref] Cellular distribution and cytotoxicity of graphene quantum dots with different functional groups, Yuan [/bib_ref] [bib_ref] Carbon-based drug delivery carriers for cancer therapy, Lim [/bib_ref] [bib_ref] Neuroblastoma cells grown on fluorine or oxygen treated graphene sheets, Oh [/bib_ref] [bib_ref] Graphene and graphene oxide: biofunctionalization and applications in biotechnology, Wang [/bib_ref]. Graphene conducts heat and electricity extremely well [bib_ref] Toxic response of graphene nanoplatelets in vivo and in vitro, Park [/bib_ref] and as one of the carbon allotropes it is considered the thinnest and strongest known material [bib_ref] In vitro evaluation of the effects of graphene platelets on glioblastoma multiforme..., Jaworski [/bib_ref]. The ratio of thickness of graphene sheet to the size of its surface differentiates this material from all other known nanomaterials [bib_ref] In vitro evaluation of the effects of graphene platelets on glioblastoma multiforme..., Jaworski [/bib_ref]. The unique physicochemical properties of graphene are large surface area (2630 m 2 /g), extraordinary electrical (mobility of charge carriers, 200,000 cm 2 V −1 s −1 ) and thermal conductivity (∼5000 W/m/K), extremely high mechanical strength (Young's modulus ∼1100 Gpa), and possibility of mass-production at low cost [bib_ref] A systems toxicology approach to the surface functionality control of graphene-cell interactions, Chatterjee [/bib_ref] [bib_ref] Assessment of the toxic potential of graphene family nanomaterials, Guo [/bib_ref] [bib_ref] Biomedical applications of graphene, Shen [/bib_ref] [bib_ref] Graphene oxide can induce in vitro and in vivo mutagenesis, Liu [/bib_ref]. The perfect electronic transport properties and high surface-to-volume ratios are responsible for its exceptional mechanical and rheological properties and resistance to degradation. Graphene has two active sides which are surfaces and edges that improve the attachment of biological molecules to graphene and its adhesion to the cells [bib_ref] Assessment of the toxic potential of graphene family nanomaterials, Guo [/bib_ref]. Graphene has higher ratio of peripheral to central carbon atoms than similar nanomaterials. Consequently atoms at the edge allow better interaction with cell membranes and interference with cell metabolism [bib_ref] In vitro and in vivo effects of graphene oxide and reduced graphene..., Jaworski [/bib_ref]. Unlike other carbon allotropes, that is, fullerenes or carbon nanotubes, graphene exhibits unique chemical and physical properties closely related to the possibility of its surface functionalization which makes it more biocompatible and less toxic [bib_ref] Graphene: promises, facts, opportunities, and challenges in nanomedicine, Mao [/bib_ref].
Graphene and graphene-based nanomaterials are today applied in numerous fields for purposes including nanoelectronics and energy technology (supercapacitors, batteries, composite materials, transistors, solar cells, fuel cells, matrix for mass spectra, and hydrogen storage), energy storage, sensors, catalysis, and biomedicine [bib_ref] Toxic response of graphene nanoplatelets in vivo and in vitro, Park [/bib_ref] [bib_ref] A systems toxicology approach to the surface functionality control of graphene-cell interactions, Chatterjee [/bib_ref] [bib_ref] Assessment of the toxic potential of graphene family nanomaterials, Guo [/bib_ref] [bib_ref] Biomedical applications of graphene, Shen [/bib_ref]. Due to their unique mechanical properties, such as high elasticity, flexibility, and adaptability for tissue engineering graphene family nanomaterials (GFNs) have been investigated in several biomedical applications especially cancer therapy, drug delivery, and diagnosis [bib_ref] An in vitro evaluation of graphene oxide reduced by Ganoderma spp. in..., Gurunathan [/bib_ref] [bib_ref] The in vitro and in vivo toxicity of graphene quantum dots, Chong [/bib_ref] [bib_ref] Systemic administration of polymer-coated nano-graphene to deliver drugs to glioblastoma, Moore [/bib_ref]. Other biomedical applications comprise gene delivery, antibacterial and antiviral materials, tissue engineering, and biocompatible scaffolds for 2 Oxidative Medicine and Cellular Longevity : The graphene structure: single layer of sp 2 -hybridized carbon atoms arranged in 2D crystal honeycomb lattice (adapted from [bib_ref] Graphene and graphene oxide: biofunctionalization and applications in biotechnology, Wang [/bib_ref]. cell cultures. Graphene-based materials are promising in the field of biosensing and bioimaging (optical sensing, fluorescence imaging probes, and electrochemical sensing) [bib_ref] A systems toxicology approach to the surface functionality control of graphene-cell interactions, Chatterjee [/bib_ref] [bib_ref] An in vitro evaluation of graphene oxide reduced by Ganoderma spp. in..., Gurunathan [/bib_ref] [bib_ref] Biomedical applications of graphene, Shen [/bib_ref] [bib_ref] Internalization and cytotoxicity of graphene oxide and carboxyl graphene nanoplatelets in the..., Lammel [/bib_ref]. Furthermore, graphene nanomaterials have been used in advanced therapeutic techniques such as photothermal and photodynamic therapies [bib_ref] Pharmaceutical applications of graphene-based nanosheets, Kim [/bib_ref] [bib_ref] The in vitro and in vivo toxicity of graphene quantum dots, Chong [/bib_ref].
Graphene and its derivatives, referred to as graphene family nanomaterials (GFNs), include graphene oxide (GO), its reduced form (rGO) and single-or few-layer graphene, graphene nanosheets (GNS), and graphene nanoribbons [bib_ref] A systems toxicology approach to the surface functionality control of graphene-cell interactions, Chatterjee [/bib_ref] [bib_ref] Assessment of the toxic potential of graphene family nanomaterials, Guo [/bib_ref] [bib_ref] Nanotoxicity of graphene and graphene oxide, Seabra [/bib_ref]. Graphene nanoparticles, depending on the method of synthesis, can show different morphologies and chemical or physical properties [bib_ref] The effects of graphene nanostructures on mesenchymal stem cells, Talukdar [/bib_ref]. So far various approaches have been developed to synthesize graphene and its derivatives such as mechanical exfoliation, epitaxial growth, or unzipping carbon nanotubes. The mechanical exfoliation, firstly used by [bib_ref] Electric field in atomically thin carbon films, Novoselov [/bib_ref] , resulted in few-layer graphene from highly oriented pyrolytic graphite. Graphene samples with the lateral size up to millimeter-range were obtained after many method modifications but still are too large and cannot be produced on a large scale, hence the inability to be used in most practical applications. Chemical vapor deposition (CVD) based on dissolving carbon atoms into a metal substrate allows producing large scale graphene films. Graphene nanoribbons (GNRs) of precise dimensions and 100% yield can be obtained by the novel strategy based on longitudinal unzipping carbon nanotubes. However, the most developed method for the mass-production of graphene is the exfoliation of graphene oxide (GO). Oxygen functional groups on the graphene surface make GO and rGO sheets strongly hydrophobic although the electrical conductivity is lower than that of pristine graphene. Poor conductivity can be bypassed in the process of liquid phase exfoliation of graphite where high-quality monolayer graphene at significant yield can be produced [bib_ref] Graphene: promises, facts, opportunities, and challenges in nanomedicine, Mao [/bib_ref]. In our previous article we have described numerous methods of graphene synthesis related with the development of various forms of graphene which differ in the quality, number of layers, and the amount of the structure defects [bib_ref] Graphene: one material, many possibilities-application difficulties in biological systems, Skoda [/bib_ref]. Lots of the possible applications of graphene derivatives obtained in different conditions make it problematic to use graphene safely in biomedicine or tissue engineering. In this paper we have focused on the impact of graphene family nanomaterials (GFNs) on the different cancer cells, the possible mechanisms of graphene toxicity, and available applications of graphene in cancer therapy or drug delivery.
## Graphene family nanomaterials (gfns)
Among other members of graphene family nanomaterials (GFNs) graphene oxide (GO) is one of the most important chemical graphene derivatives. GO is a highly oxidized form of graphene [bib_ref] A systems toxicology approach to the surface functionality control of graphene-cell interactions, Chatterjee [/bib_ref] [bib_ref] Flake sizedependent cyto and genotoxic evaluation of graphene oxide on in vitro..., De Marzi [/bib_ref] [bib_ref] Chlorotoxin-conjugated graphene oxide for targeted delivery of an anticancer drug, Wang [/bib_ref] produced mainly by chemical methods through energetic oxidation of graphite using different oxidant agents or known procedures as in Hummers method [bib_ref] In vitro evaluation of the effects of graphene platelets on glioblastoma multiforme..., Jaworski [/bib_ref] [bib_ref] Biomedical applications of graphene, Shen [/bib_ref]. GO nanosheets present hydroxyl and epoxide functional groups on their basal surface and carboxyl functional groups on their plane edges [bib_ref] Pharmaceutical applications of graphene-based nanosheets, Kim [/bib_ref]. GO has usually 1-3 layers (1-2 nm thick), with size ranging from a few to several hundred nanometers [bib_ref] Biomedical applications of graphene, Shen [/bib_ref]. GO is hydrophilic and forms stable suspensions in pure water but in salt and other biological solutions it creates aggregates [bib_ref] Cytotoxicity profile of highly hydrogenated graphene, Chng [/bib_ref] [bib_ref] Nano-graphene oxide for cellular imaging and drug delivery, Sun [/bib_ref]. Reactive COOH and OH groups in GO facilitate connection with various materials, such as polymers, biomolecules, DNA, protein, quantum dots, or Fe 3 O 4 nanoparticles which improve the solubility and prevent aggregation in salt-containing physiological buffers [bib_ref] Pharmaceutical applications of graphene-based nanosheets, Kim [/bib_ref] [bib_ref] Biomedical applications of graphene, Shen [/bib_ref].
Improved properties of graphene oxide make it useful in biological and medical applications, as a surface coating material for implants and also as a stimulator of growth and differentiation of the cells [bib_ref] Biomedical applications of graphene, Shen [/bib_ref] [bib_ref] Systemic administration of polymer-coated nano-graphene to deliver drugs to glioblastoma, Moore [/bib_ref] [bib_ref] Internalization and cytotoxicity of graphene oxide and carboxyl graphene nanoplatelets in the..., Lammel [/bib_ref]. The large aromatic surface of graphene oxide with lots of functional groups allows adsorbing molecules with high affinity and creating stable complexes which make GO an ideal nanocarrier for effective drug and gene delivery [bib_ref] Chlorotoxin-conjugated graphene oxide for targeted delivery of an anticancer drug, Wang [/bib_ref] [bib_ref] Role of surface charge and oxidative stress in cytotoxicity and genotoxicity of..., Wang [/bib_ref]. Different targeting molecules such as folic acid or antibodies can be conveniently immobilized on GO which allows precise and efficient delivery of GO into targeted cells. Solid tumor cells are more acidic (pH ∼ 6.8) than normal cells (pH 7.4) and are ideal candidates for controlled release of anticancer drugs [bib_ref] Controlled release of doxorubicin from graphene oxide based charge-reversal nanocarrier, Zhou [/bib_ref]. Lowered pH in some drug molecules additionally increases their solubility and decreases their tendency to stay adsorbed which eventually leads to the controlled endocytosis and the release in lysosomes [bib_ref] Graphene: promises, facts, opportunities, and challenges in nanomedicine, Mao [/bib_ref]. pH-responsive and integrin v 3 monoclonal antibody functionalized graphene oxide is an example of the nanocarrier for targeted delivery and controlled release of doxorubicin (DOX) into cancer cells [bib_ref] Controlled release of doxorubicin from graphene oxide based charge-reversal nanocarrier, Zhou [/bib_ref].
Reduced graphene oxide (rGO) is the product of thermal or chemical modification of graphene oxide (GO) with reducing agents (e.g., hydrazine) [bib_ref] Pharmaceutical applications of graphene-based nanosheets, Kim [/bib_ref] [bib_ref] A systems toxicology approach to the surface functionality control of graphene-cell interactions, Chatterjee [/bib_ref]. rGO possess lower number of oxygen containing functional groups than GO [bib_ref] Synthesis, characterization and cytotoxicity of europium incorporated ZnO-graphene nanocomposites on human MCF7..., Bera [/bib_ref]. The reducing conditions greatly influence the properties of GO such as electrical conductivity, surface charge, or water dispersibility (increase hydrophobicity) [bib_ref] A systems toxicology approach to the surface functionality control of graphene-cell interactions, Chatterjee [/bib_ref]. rGO possesses high capacity for hydrophobic interactions among various functional molecules but it leads to creation of aggregates with weak stability under physiological conditions. Surface modification of rGO with polymers or biopolymers has been used to stabilize and improve the properties of rGO and use it as a nanocarrier [bib_ref] Reduced graphene oxide nanosheets coated with an anti-angiogenic anticancer lowmolecular-weight heparin derivative..., Shim [/bib_ref].
Graphene platelets (GPs) are produced by physical methods directly by exfoliation of graphite without the initial Oxidative Medicine and Cellular Longevity 3 stage of oxidation. GPs are hydrophobic and form stable hydrocolloids [bib_ref] In vitro evaluation of the effects of graphene platelets on glioblastoma multiforme..., Jaworski [/bib_ref]. Zero-dimensional, single-atom layer graphene quantum dots (GQDs) have lateral dimensions below 100 nm and size of 10 nm or less [bib_ref] The in vitro and in vivo toxicity of graphene quantum dots, Chong [/bib_ref]. GQDs are biocompatible due to their small size and high oxygen content which improves solubility and stability in water or serum [bib_ref] Biomedical applications of graphene, Shen [/bib_ref] [bib_ref] The in vitro and in vivo toxicity of graphene quantum dots, Chong [/bib_ref]. Graphene quantum dots due to their excellent photoluminescent properties are promising agents for optical probes in bioimaging [bib_ref] Cellular distribution and cytotoxicity of graphene quantum dots with different functional groups, Yuan [/bib_ref]. Graphene nanoparticles, referred to as graphene nanoribbons, are formed by the longitudinal unzipping of multiwalled carbon nanotubes [bib_ref] Cell specific cytotoxicity and uptake of graphene nanoribbons, Chowdhury [/bib_ref].
## Graphene and cells
The potential toxic effects of graphene materials on the environment and on the human health have recently attracted considerable attention among researchers. Understanding of the interactions of GFNs with living systems and their adverse effects in vitro and in vivo is essential for further development and safe use of graphene-based nanomaterials [bib_ref] Assessment of the toxic potential of graphene family nanomaterials, Guo [/bib_ref]. Cytotoxicity studies of graphene include the influence on the cell viability and morphology, membrane integrity, ROS generation, DNA damage, gene expression, DNA damage, and mechanism of uptake [fig_ref] Figure 2: Schematic toxicity mechanisms of graphene on human cancer cells [/fig_ref] [bib_ref] A systems toxicology approach to the surface functionality control of graphene-cell interactions, Chatterjee [/bib_ref] [bib_ref] In vitro evaluation of the effects of graphene platelets on glioblastoma multiforme..., Jaworski [/bib_ref] [bib_ref] Graphene oxide can induce in vitro and in vivo mutagenesis, Liu [/bib_ref]. The interactions of graphene nanoparticles with the cells depend on the physicochemical and electrical properties [bib_ref] An in vitro evaluation of graphene oxide reduced by Ganoderma spp. in..., Gurunathan [/bib_ref] [bib_ref] Biomedical applications of graphene, Shen [/bib_ref] [bib_ref] Distribution of graphene oxide and TiO 2 -graphene oxide composite in A549..., Jin [/bib_ref] [bib_ref] Carbon materials for drug delivery & cancer therapy, Liu [/bib_ref] [bib_ref] Cytotoxicity assessment of MDA-MB-231 breast cancer cells on screenprinted graphene-carbon paste substrate, Waiwijit [/bib_ref] [bib_ref] Synthesis, characterization and cytotoxicity of phosphorylcholine oligomer grafted graphene oxide, Liu [/bib_ref]. The reports indicate that morphology (size, shape, and sharp edges), surface charge, surface functionalization, dispersibility, state of aggregation, number of layers, purity, and method of synthesis (e.g., CVD [bib_ref] Graphene: one material, many possibilities-application difficulties in biological systems, Skoda [/bib_ref] , arc-discharge [bib_ref] Cell specific cytotoxicity and uptake of graphene nanoribbons, Chowdhury [/bib_ref] , and biological methods [bib_ref] Green synthesis of graphene and its cytotoxic effects in human breast cancer..., Gurunathan [/bib_ref] are the key factors that influence the mechanism of uptake (passive diffusion and endosomal uptake) and tissue response to graphene-based nanomaterials [bib_ref] Toxic response of graphene nanoplatelets in vivo and in vitro, Park [/bib_ref] [bib_ref] A systems toxicology approach to the surface functionality control of graphene-cell interactions, Chatterjee [/bib_ref] [bib_ref] An in vitro evaluation of graphene oxide reduced by Ganoderma spp. in..., Gurunathan [/bib_ref] [bib_ref] The effects of graphene nanostructures on mesenchymal stem cells, Talukdar [/bib_ref] [bib_ref] Cell specific cytotoxicity and uptake of graphene nanoribbons, Chowdhury [/bib_ref]. Moreover, the toxic effect of graphene highly depends on the conditions of the experiment, which include the time of exposure, dose, type of the cells, and the method used to establish the cell viability [bib_ref] Nanotoxicity of graphene and graphene oxide, Seabra [/bib_ref] [bib_ref] Cytotoxicity profile of highly hydrogenated graphene, Chng [/bib_ref] [bib_ref] Cell specific cytotoxicity and uptake of graphene nanoribbons, Chowdhury [/bib_ref] [bib_ref] Green synthesis of graphene and its cytotoxic effects in human breast cancer..., Gurunathan [/bib_ref] [bib_ref] A comparative study of cellular uptake and cytotoxicity of multi-walled carbon nanotubes,..., Zhang [/bib_ref].
The chemical methods used in the production of graphene nanomaterials including oxidation or reduction of graphene oxide bring harsh conditions and toxic agents, such as hydrazine or its derivatives, which influence the structure of graphene and its safety. One of the approaches used to decrease the toxicity of graphene involves aqueous and environmentally friendly reduction strategy based on bacterial and yeast respiration [bib_ref] Green synthesis of graphene and its cytotoxic effects in human breast cancer..., Gurunathan [/bib_ref]. Recently used microbial biomass for the reduction of GO including Escherichia coli [bib_ref] Microbial reduction of graphene oxide by Escherichia coli: a green chemistry approach, Gurunathan [/bib_ref] , Bacillus marisflavi [bib_ref] Green synthesis of graphene and its cytotoxic effects in human breast cancer..., Gurunathan [/bib_ref] , and Ganoderma extract [bib_ref] An in vitro evaluation of graphene oxide reduced by Ganoderma spp. in..., Gurunathan [/bib_ref] has significantly increased biocompatibility of graphene.
The majority of GFNs have poor solubility and create aggregates in salt-containing physiological buffers due to electrostatic charge and nonspecific binding to proteins [bib_ref] Assessment of the toxic potential of graphene family nanomaterials, Guo [/bib_ref]. Functionalization of pristine graphene via covalent or noncovalent coatings by various materials such as polymers, DNA, proteins, and nanoparticles greatly improves the biocompatibility [bib_ref] Pharmaceutical applications of graphene-based nanosheets, Kim [/bib_ref]. Surface modifications of graphene nanomaterials also improve their solubility and significantly reduce toxic interactions with living systems. Significant changes in biocompatibility have been achieved by producing graphene reinforced composite materials with polyethylene glycol (PEG) or other biopolymers such as chitosan, hyaluronan (HA), or dextran [bib_ref] Pharmaceutical applications of graphene-based nanosheets, Kim [/bib_ref] [bib_ref] Assessment of the toxic potential of graphene family nanomaterials, Guo [/bib_ref] [bib_ref] Graphene oxide can induce in vitro and in vivo mutagenesis, Liu [/bib_ref] [bib_ref] Systemic administration of polymer-coated nano-graphene to deliver drugs to glioblastoma, Moore [/bib_ref] [bib_ref] Nanotoxicity of graphene and graphene oxide, Seabra [/bib_ref].
Most of the members of graphene family nanomaterials easily enter the living cells because of the small size, sharp edges and rough surface [bib_ref] Assessment of the toxic potential of graphene family nanomaterials, Guo [/bib_ref]. Additionally, negatively charged (−30.89 eV) GO can easily accumulate inside the cell [bib_ref] Distribution of graphene oxide and TiO 2 -graphene oxide composite in A549..., Jin [/bib_ref]. The uptake can be also affected by the shape and the aggregation state of GO sheets [bib_ref] In vitro toxicity evaluation of graphene oxide on A549 cells, Chang [/bib_ref]. The presence of carboxyl, epoxy, and hydroxyl groups in GO reduces its cytotoxicity [bib_ref] Distribution of graphene oxide and TiO 2 -graphene oxide composite in A549..., Jin [/bib_ref] and the small size (smaller than 5 nm) and the high content of oxygen improve the solubility and increase biocompatibility [bib_ref] The in vitro and in vivo toxicity of graphene quantum dots, Chong [/bib_ref]. However, the mechanism of cellular uptake and the fate of graphene inside the living cells are still not fully understood. This process may depend on the cell type, on the properties of graphene, or on both of these factors. Some researchers suggest endocytosis as a basic mechanism of cellular uptake for PEG-GO while others combine endocytosis and macropinocytosis depending on the formation of smaller or larger aggregates of PEG-graphene nanoribbons [bib_ref] Pharmaceutical applications of graphene-based nanosheets, Kim [/bib_ref].
The physical interactions of graphene with the cell membranes are one of the major causes of GFNs cytotoxicity [bib_ref] An in vitro evaluation of graphene oxide reduced by Ganoderma spp. in..., Gurunathan [/bib_ref] [bib_ref] Nanotoxicity of graphene and graphene oxide, Seabra [/bib_ref] [bib_ref] Comparative protein profile of human hepatoma HepG2 cells treated with graphene and..., Yuan [/bib_ref]. Hydrophobic forms of graphene interact with the cell membrane lipids [bib_ref] Nanotoxicity of graphene and graphene oxide, Seabra [/bib_ref] while the other forms may bond to the cell receptors and interfere with the cell metabolism, inhibit nutrient supply, and induce stress or cell death [bib_ref] In vitro evaluation of the effects of graphene platelets on glioblastoma multiforme..., Jaworski [/bib_ref]. Moreover, graphene itself can bind the micronutrients and amino acids from the cell culture medium which limits their availability and inhibits cellular growth and viability [bib_ref] Cytotoxicity profile of highly hydrogenated graphene, Chng [/bib_ref]. GO is smaller and less toxic than rGO because of the high oxygen content, smoother edges, and hydrophilic properties. Reduced graphene oxide has high affinity to the cell membranes and the irregular and sharp edges affect their integrity, stimulate receptors, and activate mitochondrial pathways which may cause apoptosis [bib_ref] Nanoparticles containing allotropes of carbon have genotoxic effects on glioblastoma multiforme cells, Hinzmann [/bib_ref].
Oxidative stress and generation of reactive oxygen species (ROS) can be involved in the toxic effects of graphene-based nanomaterials [bib_ref] The in vitro and in vivo toxicity of graphene quantum dots, Chong [/bib_ref] [bib_ref] Nanotoxicity of graphene and graphene oxide, Seabra [/bib_ref] [bib_ref] Comparative protein profile of human hepatoma HepG2 cells treated with graphene and..., Yuan [/bib_ref]. When the cell homeostasis is disrupted and the enzymes responsible for reducing ROS (superoxide dismutase and glutathione peroxidase) fail, the macromolecules, such as proteins, DNA, and lipids, can be damaged, which greatly influence the cell metabolism and signaling [bib_ref] Nanotoxicity of graphene and graphene oxide, Seabra [/bib_ref] [bib_ref] Cytotoxicity assessment of MDA-MB-231 breast cancer cells on screenprinted graphene-carbon paste substrate, Waiwijit [/bib_ref]. The interactions of the GO with the cells can lead to excessive ROS generation, which is the first step in the mechanisms of carcinogenesis, ageing, and mutagenesis [bib_ref] Flake sizedependent cyto and genotoxic evaluation of graphene oxide on in vitro..., De Marzi [/bib_ref].
Except for the plasma membrane damage and oxidative stress induction graphene can cause apoptosis and/or cell necrosis through the direct influence on the cell DNA or mitochondrial activity [bib_ref] Internalization and cytotoxicity of graphene oxide and carboxyl graphene nanoplatelets in the..., Lammel [/bib_ref]. Graphene nanoparticles can induce dissipation of the mitochondrial membrane potential which subsequently increases the generation of intracellular ROS and eventually triggers apoptosis by activating the mitochondrial pathway [bib_ref] Green synthesis of graphene and its cytotoxic effects in human breast cancer..., Gurunathan [/bib_ref]. The interactions of graphene with cell genetic material are based on DNA-intercalation and cleavage mechanisms [bib_ref] Graphene oxide can induce in vitro and in vivo mutagenesis, Liu [/bib_ref]. Difference in the structure of rGO and GO makes rGO more potent to penetrate cell compartments and directly interact with the nuclear DNA resulting in genotoxic effects [bib_ref] Nanoparticles containing allotropes of carbon have genotoxic effects on glioblastoma multiforme cells, Hinzmann [/bib_ref].
Additionally, graphene can directly interact with different genes encoding important proteins and enzymes [bib_ref] A systems toxicology approach to the surface functionality control of graphene-cell interactions, Chatterjee [/bib_ref] [bib_ref] Graphene oxide can induce in vitro and in vivo mutagenesis, Liu [/bib_ref]. Other indirect mechanisms of GO cytotoxicity involve DNA damage caused by ROS [bib_ref] Graphene oxide can induce in vitro and in vivo mutagenesis, Liu [/bib_ref] , inhibition or activation of specific enzymes [bib_ref] Flake sizedependent cyto and genotoxic evaluation of graphene oxide on in vitro..., De Marzi [/bib_ref] , or reaction with other cell components such as proteins and polysaccharides [bib_ref] Graphene oxide can induce in vitro and in vivo mutagenesis, Liu [/bib_ref]. For better understanding of the mechanisms of graphene action inside the cell further studies are required, particularly to explain the cellular interactions of graphene materials with proteins and cell membrane lipids on a molecular level [bib_ref] Nanotoxicity of graphene and graphene oxide, Seabra [/bib_ref].
## Breast cancer cell.
Many of the currently available methods for producing graphene are not environmentally friendly and rGO obtained by these methods is not safe enough to use in biological and medical applications. Therefore researchers developed a novel and simple approach for rGO synthesis using microorganisms which is cost-effective and safe for the environment. Gurunathan et al. compared the cytotoxicity of GO obtained from graphite powder using a modified version of Hummers and Offeman's method with rGO synthesized by Bacillus marisflavi biomass on human breast adenocarcinoma cells (MCF-7) using WST-8 assay. Incubation of MCF-7 cells with both B-rGO (biogenic rGO) and GO at concentrations ranging from 0 to 100 g/mL showed dose-dependent graphene cytotoxicity. In concentrations higher than 60 g/mL graphene markedly decreased the cell viability and increased ROS generation and release of LDH. Surprisingly, bacterial rGO had stronger cytotoxic effect on MCG-7 cells compared to GO [bib_ref] Green synthesis of graphene and its cytotoxic effects in human breast cancer..., Gurunathan [/bib_ref]. In another experiment Gurunathan and colleagues used mushroom extracts (Ganoderma) to reduce graphene oxide. They examined the influence of GO and GE-rGO on MDA-MB-231 human breast cancer cells using WST-8 viability assay, membrane integrity test (LDH assay), and DCFH-DA assay as a quantitative method for oxidative stress assessment. The cytotoxicity of graphene was dose-dependent (0-150 g/mL) especially at the higher concentrations where elevated levels of ROS induced membrane damage and LDH leakage in the presence of GE-rGO [bib_ref] An in vitro evaluation of graphene oxide reduced by Ganoderma spp. in..., Gurunathan [/bib_ref]. These studies indicate that rGO synthesis with the use of bacteria and fungi is easier, less expensive and works better for the development of a potential therapeutic agent that targets breast cancer cells.
In vitro anticancer activity of GO was examined in various concentrations [bib_ref] In vitro evaluation of the effects of graphene platelets on glioblastoma multiforme..., Jaworski [/bib_ref] [bib_ref] The effects of graphene nanostructures on mesenchymal stem cells, Talukdar [/bib_ref] [bib_ref] Distribution of graphene oxide and TiO 2 -graphene oxide composite in A549..., Jin [/bib_ref] , and 80 g/mL) on human breast cancer cells MCF-7 using MTT viability assay. GO showed approximately 13% inhibition of cell viability of MCF-7 cells and the cytotoxicity at dose-dependent manner [bib_ref] Graphene oxide based magnetic nanocomposites for efficient treatment of breast cancer, Chaudhari [/bib_ref]. Other tests concerning cytotoxicity of GO were carried on human adenocarcinoma breast cancer cells (MDA-MB-231) using Cell Counting Kit-8 (CCK-8) assay. 48 h incubation with GO in concentrations ranging from 100 g/mL to 500 g/mL showed increasing cytotoxicity against MDA-MB-231 cells together with the increasing amount of graphene in the medium. Further studies showed that GO reacts directly with genomic DNA and inhibits cell replication with complete blockage of human glyceraldehyde-3-phosphate dehydrogenase (hGAPDH) gene at the concentration of 1 g/mL. MDA-MB-231 cells treated with GO even at low concentration (10 g/mL) after 24 h incubation showed signs of apoptosis. Hence, scientists tested 30,000 genes to examine the impact of GO on the gene expression at the cellular level. The results revealed 101 genes (mainly responsible for DNA-damage control, cell apoptosis, cell cycle, and metabolism) that showed 2-fold or even greater expression changes after GO treatment at the concentrations of 10 g/mL and 100 g/mL. Additionally, GO increased expression of ATM and Rad51 genes (DNA repair proteins) which can explain the influence of graphene on the cell DNA [bib_ref] Graphene oxide can induce in vitro and in vivo mutagenesis, Liu [/bib_ref].
Zhou and coworkers evaluated the cytotoxicity of GO modified with polyethylene glycol (PEG) using three cell lines derived from human breast cancers: MDA-MB-231, MDA-MB-436, and SK-BR-3. PEG-GO had no apparent influence on the cell viability but inhibited cancer cell migration and invasion. PEG-GO disrupted F-actin filaments responsible for cell migration by depleting ATP levels through downregulation of mitochondrial energy metabolism [bib_ref] Energy metabolism analysis reveals the mechanism of inhibition of breast cancer cell..., Zhou [/bib_ref]. Another research on the human breast cancer cells MDA-MB-231 with pristine graphene and graphene oxide also showed no apparent influence on the cell viability at low concentrations but prominent inhibition of migration and invasion [bib_ref] The inhibition of migration and invasion of cancer cells by graphene via..., Zhou [/bib_ref].
Recent findings have proved that the functionalization of the graphene surface makes it less toxic. Mullick Chowdhury et al. investigated the cytotoxicity of oxidized-graphene nanoribbons coated with the amphiphilic polymer PEG-DSPE (O-GNR-PEG-DSPE) at various concentrations (0-400 g/mL) on Sloan Kettering breast cancer (SKBR3) cells and Michigan Cancer Foundation-7 (MCF-7) breast cancer cells using Alamar blue assay. Both cell lines showed the reduction in viability by about 10%-15% at the highest concentrations after 24 h incubation with the copolymer. SKBR3 cells incubated with O-GNR-PEG-DSPE demonstrated slight increase in the LDH release while MCF-7 cells did not show any statistically significant LDH leakage. Additionally, SKBR3 and MCF-7 cells showed small or no uptake of O-GNR-PEG-DSPE. The results indicate that graphene copolymer has no toxic effect on the tested cells up to 10 g/mL and exhibits low cytotoxicity even at the highest concentrations (400 g/mL) [bib_ref] Cell specific cytotoxicity and uptake of graphene nanoribbons, Chowdhury [/bib_ref].
The toxicity of covalently pegylated nano-GO with unmodified rGO was compared using MTS assay and MCF-7 human epithelial breast cancer cells. The half maximal inhibitory concentration (IC50) of nano-rGO was established at the concentration of approximately 80 mg/L, while for pegylated nano-GO it was at about 99 mg/L [bib_ref] Graphene oxide based magnetic nanocomposites for efficient treatment of breast cancer, Chaudhari [/bib_ref]. According to MTT assay fluorinated form of graphene oxide (FGO) even at the concentration of 576 g/mL showed no toxicity to human breast cancer cells (MCF-7) [bib_ref] Fluorinated graphene oxide; a new multimodal material for biological applications, Romero-Aburto [/bib_ref]. Waiwijit and coworkers investigated the toxicity of graphene-carbon paste (GCP) in four different concentrations (1, 2.5, 5, and 10 wt%) on MDA-MB-231 breast cancer cells also using MTT assay. The cell viability decreased after longer incubation periods (48 and 72 h) and at the presence of the highest concentration of GCP in comparison to the cultures with CP alone. Moreover, MDA-MB-231 cancer cells exhibited increased ROS generation with the increasing time of incubation and the amount of GCP in the culture medium [bib_ref] Cytotoxicity assessment of MDA-MB-231 breast cancer cells on screenprinted graphene-carbon paste substrate, Waiwijit [/bib_ref]. Together, these studies demonstrate the different impact of graphene nanomaterials on breast cancer cells including the cell viability and cytotoxicity connected with the generation of ROS, loss of the membrane integrity, and DNA damage which may have potential clinical advantage pertaining to increased therapeutic efficacy and decreased local toxicity of the used nanomaterial.
## Cervical cancer cell.
Remarkably durable and prolific HeLa cells derived from cervical cancer are more sensitive to the graphene than other cell lines. According to Zhang et al. GO showed high cytotoxicity to HeLa cells even at low concentrations. The biological responses induced by GO were evaluated by series of assays, including MTT, malondialdehyde (MDA), superoxide dismutase (SOD), lactate dehydrogenase (LDH), and reactive oxygen species (ROS). HeLa cells were treated with different concentrations of GO ranging from 0 to 80 g/mL and cultured for 3 h and 24 h. MTT test results showed dose-dependent GO cytotoxicity with the cell viability at about 50% at the concentration of 80 g/mL. To evaluate the lipid peroxidation and oxidant stress the levels of MDA and SOD enzyme activity were measured in the cell lysates. The results showed an obvious increase in MDA production after exposure to 80 g/mL of GO and decreased SOD activity. Moreover, the incubation of HeLa cells with 80 g/mL of GO for 24 h increased the levels of ROS 17 times. The researchers suggested that the cytotoxicity of GO is not associated with the cell uptake [bib_ref] A comparative study of cellular uptake and cytotoxicity of multi-walled carbon nanotubes,..., Zhang [/bib_ref].
Instead of the biological assays measuring cell activity and viability there are available more selective, more sensitive, and faster electrochemical approaches to evaluate the toxicity of graphene. Yoon et al. used cell-based electrochemical impedance biosensing with interdigitated indium tin oxide (ITO) electrodes to analyze toxicity of graphene nanoflakes in HeLa cells. Researchers used two different sizes of graphene flakes (80 nm and 30 nm) in the concentration of 400 g/mL and monitored the cytotoxicity for 1 day. The studies showed greater cytotoxic effect of the smaller 30 nm graphene nanoflakes due to their higher uptake, while 80 nm graphene nanoflakes agglomerated on cell membranes causing less harm to the cells [bib_ref] Toxicity of graphene nanoflakes evaluated by cell-based electrochemical impedance biosensing, Yoon [/bib_ref].
Liu's group conjugated graphene oxide with dextran, a widely used surface coating biopolymer. They cultured HeLa cells with different concentrations (10, 50, and 200 mg/L) of GO and GO-DEX and studied in vitro toxicity for 24 h, 48 h, and 72 h. The cell counting data showed dose-dependent decrease in the cell proliferation after incubation with GO and notably smaller influence on the cell count after GO-DEX treatment. The calcein AM/propidium iodide (PI) staining was carried out to further determine graphene toxicity. The results revealed that GO did not induce significant cell death even at high concentrations up to 200 mg/L, while GO-DEX showed no influence on the cell growth and viability. All the evidence demonstrates that dextran coating may improve the biocompatibility of GO [bib_ref] In vitro and in vivo behaviors of dextran functionalized graphene, Zhang [/bib_ref].
In other studies the cytotoxicity of graphene polymer (GQD-PEG) was evaluated on HeLa cells using WST-1 assay. GQD-PEG did not induce apoptosis or necrosis even at the concentration of 160 g/mL. LDH release and ROS level measurements showed no impact of GQD-PEG on the cell membrane integrity and oxidative stress generation probably because of the small size of the particles (smaller than 5 nm) and the presence of PEG polymer [bib_ref] The in vitro and in vivo toxicity of graphene quantum dots, Chong [/bib_ref]. However, HeLa cells showed a reduction of the cell viability by 60% after 24 h incubation with 400 g/mL O-GNR-PEG-DSPE (oxidizedgraphene nanoribbons (O-GNRs) with the amphiphilic polymer). As the dose increased, the survival rate of the cells decreased together with the release of LDH. On the images of the cells lots of swollen intracellular vesicles were observed together with disrupted plasma membranes which are a characteristic feature in necrotic cells [bib_ref] Cell specific cytotoxicity and uptake of graphene nanoribbons, Chowdhury [/bib_ref].
## Lung cancer cell.
The biological effect of GFNs on lung cancer cells depends mainly on the size and concentration of graphene [bib_ref] Flake sizedependent cyto and genotoxic evaluation of graphene oxide on in vitro..., De Marzi [/bib_ref] [bib_ref] In vitro toxicity evaluation of graphene oxide on A549 cells, Chang [/bib_ref]. Hu et al. investigated the cellular effect of different concentrations (0 to 100 g/mL) of GO nanosheets on human alveolar adenocarcinoma cell line (A549). MTT assay showed concentration-dependent cytotoxicity and about 50% decrease in cell viability after incubation with GO at the concentration of 100 g/mL. Interestingly, the cell viability was greatly mitigated after addition of 10% FBS (fetal bovine serum) into the culture medium. TEM imaging demonstrates that precoating of GO with FBS prevents cell membranes from the damage, the outflow of cytoplasm, and eventually cell death. GO nanosheets possess high adsorption capability for proteins in the medium and therefore cytotoxic effect of GO precoated with 10% FBS was largely reduced [bib_ref] Protein corona-mediated mitigation of cytotoxicity of graphene oxide, Hu [/bib_ref].
In other experiments scientists compared the cytotoxicity of GO nanosheets and reduced with hydrazine rGO nanosheets characterized by lower thickness and less surface defects. The metabolic activity assays based on succinate dehydrogenase activity in the mitochondria showed that GO in the concentration of 20 g/mL slightly influenced the viability of A549 cells (20%) but in higher concentration (85 g/mL) reduced the cell viability to 50% within 24 h. rGO nanosheets reduced the A549 cell viability to 47% and 15% with 20 and 85 g/mL, respectively. Therefore, rGO nanosheets are significantly more cytotoxic than GO's which is because of different surface charge and functional groups on the nanosheet surfaces. Transmission electron microscopy (TEM) showed that graphene nanosheets could be internalized within A549 cells via endocytosis. However, flow cytometric analysis demonstrated no apoptosis in A549 cells treated with GO nanosheets (20 and 85 g/mL for 24 h) but cell cycle arrest in the G2 phase (mitosis metaphase). These data suggest that the observed small decrease in the cell viability is not because of the cell death but rather might arise from GO-retarded cell cycle which restrains the proliferation rate [bib_ref] Graphene-based antibacterial paper, Hu [/bib_ref].
The group of scientists investigated also the cytotoxicity of graphene oxide (GO) and highly hydrogenated graphene (HHG) in concentrations that ranged from 3.125 g/mL to 400 g/mL. The results from MTT and WST-8 assays indicated that HHG was more toxic to A549 cells than GO and that the toxicity was dose-dependent. The percentage of viable cells after 24 h treatment with GO and HHG in the concentration of 400 g/mL was 43% and 26%, respectively [bib_ref] Cytotoxicity profile of highly hydrogenated graphene, Chng [/bib_ref].
In contrast, Chang et al. reported that graphene oxide (GO) is a reasonably safe material at the cellular level. Researchers examined the toxicity of GO at the concentration range from 0 to 200 g/mL on human lung carcinoma epithelial cell line A549. In this comprehensive study the morphology, viability, apoptosis, ROS production, and membrane integrity were examined. The CCK-8 assay used to estimate the GO toxicity showed dose-and size-dependent loss of the viability with little influence of the culture period. However, the level of apoptosis was not relevant to the dose or the size of the GO samples and exposure to GO did not induce LDH leakage. The LDH levels of GO-treated cells (for 200 g/mL was 6%) were even slightly lower than those of the control cells (7.5%). GO induced oxidative stress in A549 cells even at low concentrations, but with no obvious toxicity. The results showed that the cells grow on the GO films very well and there is no considerable difference in the morphology and density of the GO-treated and control cells. There is no impact on the ultrastructure of A549 cells and no signs of GO sheets inside the cells. These results indicate that GO is biocompatible and has a great potential for being the substrate for the cell growth [bib_ref] In vitro toxicity evaluation of graphene oxide on A549 cells, Chang [/bib_ref].
de Marzi et al. with the same cell line investigated the impact of graphene oxide on the viability using MTT assay. Graphene oxide was used at various concentrations (10, 50, and 100 g/mL) and in two different flake sizes (1.32 m and Oxidative Medicine and Cellular Longevity 7 130 nm). The results showed slight loss in the viability of the A549 cells after 24 h incubation with both types of GO. The comet assay showed size-dependent genotoxic effect on the cells with high degree of toxicity even at the low concentrations with 130 nm GO flakes [bib_ref] Flake sizedependent cyto and genotoxic evaluation of graphene oxide on in vitro..., De Marzi [/bib_ref]. Cytotoxicity and distribution of GO inside the A549 cells were evaluated by Jin and coworkers using CCK-8 assay and transmission electron microscopy (TEM), respectively. After 4 h incubation with GO in concentrations of 100 and 300 g/mL there was no significant decrease in the cell viability. GO was present inside the cells in the cytoplasm and nucleus but cellular organelles were not affected [bib_ref] Distribution of graphene oxide and TiO 2 -graphene oxide composite in A549..., Jin [/bib_ref].
Yuan et al. examined the cytotoxicity of graphene quantum dots (GQDs) with various surface modifications (NH 2 , COOH, and CO-N (CH 3 ) 2 ) in human lung carcinoma cells (A549 cells) using MTT assay. GQDs with different functional groups had low cytotoxicity even when the concentration reached 200 g/mL. Moreover, the three kinds of GQDs did not induce cell apoptosis and/or necrosis. GQDs (50 g/mL) were localized in the cytoplasm and did not enter into the cell nucleus. GQDs are smaller and provide less damage to cell membranes than GO and therefore are more biocompatible and less cytotoxic to cells even when modified with different chemical groups [bib_ref] Cellular distribution and cytotoxicity of graphene quantum dots with different functional groups, Yuan [/bib_ref]. Other studies showed that pegylated graphene quantum dots (GQDs-PEG) are practically not toxic to A549 cells at all [bib_ref] The in vitro and in vivo toxicity of graphene quantum dots, Chong [/bib_ref]. The inconsistency of these results might come from the different methods of preparation or synthesis of GO and distinct testing models.
## Liver cancer
Cell. The increasing number of possible applications of graphene nanomaterials triggers considerable concerns about the impact on health and environment though further more thorough investigations are vital. Chatterjee et al. investigated toxicity of various concentrations of graphene oxide (GO) and reduced graphene oxide (rGO) on HepG2 cells for 24 h. According to EZ-Cytox assay the cells viability was clearly dose-and time-dependent for both nanomaterials but rGO indicated higher cytotoxicity with unclear converse change after 16 h of exposure. EC20 and EC50 for rGO were 8 mg/L and 46 mg/L, respectively, whereas they were 10 mg/L and 81 mg/L for GO. The microscopic images showed increased internal granularity of the GO-treated cells which indicates that GO was internalized by HepG2 cells through endocytosis. The rGO treated cells showed outsized aggregation and accumulation of rGO on the cell membrane due to its hydrophobic nature. Difference in the uptake efficiency explains various modes of cytotoxicity [bib_ref] A systems toxicology approach to the surface functionality control of graphene-cell interactions, Chatterjee [/bib_ref].
One of the principal mechanisms underlying nanomaterial toxicity involves oxidative stress. In the experiments both GO and rGO induced release of reactive oxygen species (ROS) in HepG2 at dose-dependent manner. However, rGO mediated ROS production was the result of physical interaction while oxidative stress induced by GO involved NADPH oxidase and significant increase in the antioxidative enzyme genes (SOD1, SOD2, CAT, GSTA1, and GSTA4) expression. The toxicity of graphene can also be caused by direct interaction with the cell DNA. GO and rGO induced both single and double stranded DNA damage. rGO did not significantly influence the DNA repair gene expression and DNA damage resulted from physical interactions rather than biological one. Moreover, GO and rGO both caused increase in the apoptosis rate of HepG2 cells. However, apoptosis induced by GO was dose-and time-dependent and involved alterations in expression of the key apoptotic genes whereas rGO elicited apoptosis only at lower dose and early time of exposure. The cytotoxicity of rGO is probably caused by the strong hydrophobic interactions with the cell membranes and eventual destruction by extremely sharp edges and highly depends on their uptake by HepG2 cells [bib_ref] A systems toxicology approach to the surface functionality control of graphene-cell interactions, Chatterjee [/bib_ref].
The objective of the study of Lammel and his coworkers was to evaluate the cytotoxicity and underlying mechanism of two different graphene derivatives: graphene oxide (GO) and carboxyl graphene (CXYG) towards human hepatoma cell line. It was observed that cells exposed to GO and CXYG in concentrations of 16 g/mL for 24 h were completely covered with the nanomaterial and further increase in the concentration caused unspecific cell damage due to mechanical stress. TEM and scanning electron micrographs demonstrated that both GO and CXYG were able to penetrate the plasma membrane and cumulate in the intracellular vesicles resulting in altered cell morphology and an augmented number of apoptotic cells. Exposure of HepG2 to GO (1-16 g/mL) and CXYG (2-32 g/mL) for 72 h caused dosedependent increase in the fluorescence intensity indicating an elevated metabolic activity of the cells which suggests plasma membrane damage. Loss of the membrane integrity was associated with a strong physical interaction of GO with the phospholipid bilayer and increased metabolism was probably associated with energy-dependent process involved in plasma membrane repair. Elevated fluorescence intensity at the high exposure concentrations can be also explained by oxidative stress increase. However, the underlying ROS-generating mechanisms were distinct after GO and CXYG treatment. Exposure to GO and GXVG indicates mitochondrial membrane depolarization and/or a decrease in the amount of mitochondria which leads to increased intracellular ROS. The authors concluded that plasma membrane damage and oxidative stress are the key factors in graphene-induced cytotoxicity of HepG2 cells [bib_ref] Internalization and cytotoxicity of graphene oxide and carboxyl graphene nanoplatelets in the..., Lammel [/bib_ref].
Yuan et al. applied the iTRAQ-coupled 2D LC-MS/MS approach to analyze the protein profile change of HepG2 cells treated with graphene oxide. They observed only a moderate variation of protein levels within the cells [bib_ref] Comparative protein profile of human hepatoma HepG2 cells treated with graphene and..., Yuan [/bib_ref] [bib_ref] Cytotoxicity evaluation of oxidized single-walled carbon nanotubes and graphene oxide on human..., Yuan [/bib_ref]. Moreover, MTT assay resulted in 17% loss of the cell viability in the cells treated with GO [bib_ref] Comparative protein profile of human hepatoma HepG2 cells treated with graphene and..., Yuan [/bib_ref].
## Nerve cell cancer.
Graphene toxicity and biocompatibility were further established by Jaworski et al. who examined the influence of graphene platelets (GPs) on two different human glioma cell lines (U87 and U118) with high degree of malignancy. The GP-treated cells were more oval and denser and in both cases graphene platelets created agglomerates close to the cell bodies but did not enter the cells. GPs caused cell membrane disruption higher in U87 than in U118 cells. Exposure to graphene at the concentration of 100 g/mL for 24 h resulted in 54% and 58% decrease in the cell viability in U87 and U118 cells, respectively. The degree of apoptosis was higher in both glioma cell lines (68% in 8 Oxidative Medicine and Cellular Longevity U87 and 99% in U118) together with necrosis present only in U87 (24%). The results indicate that the high concentration and the direct physical contact with the cells are the main cause of graphene toxicity. Difference in the activity of genes involved in a cell cycle regulation of the U87 and U118 cells is responsible for the susceptibility to programmed cell death indicating the potential applicability of GP in anticancer therapy [bib_ref] In vitro evaluation of the effects of graphene platelets on glioblastoma multiforme..., Jaworski [/bib_ref]. Similar results of nano-rGO were obtained in U87MG glioblastoma cell line using MTS assay where half maximal inhibitory concentration (IC50) reached 85 mg/L [bib_ref] Ultrasmall reduced graphene oxide with high near-infrared absorbance for photothermal therapy, Robinson [/bib_ref]. Jaworski et al. using the same glioma cells (U87 and U118) as previously mentioned investigated cytoand genotoxicity of GO and rGO platelets. In vitro analysis showed that both GO and rGO enter glioma cells and reduce the cell viability and the proliferation with increasing doses. However, the lower cell vitality and the higher degree of apoptosis were observed after rGO treatment which indicates that GO is less toxic to glioma cells than rGO [bib_ref] In vitro and in vivo effects of graphene oxide and reduced graphene..., Jaworski [/bib_ref]. The scope of another experimental in vitro study on glioblastoma cancer cells U87 was to determine the cell viability and DNA fragmentation after exposure to different carbon allotropes. All studied nanoparticles did not alter the cell morphology; however pristine graphene (GN) and reduced graphene oxide (rGO) led to a significant decrease in the cell viability. The comet assay results demonstrated that DNA damage was caused by GN, rGO, graphite, and ultradispersed detonation diamond (UDD) and only GO had no genotoxic effect on U87 cells. These findings indicate the potential use of GO as a drug nanocarrier and GN, rGO, graphite, and UDD in the direct elimination of glioblastoma multiforme cells because of their higher toxicity [bib_ref] Nanoparticles containing allotropes of carbon have genotoxic effects on glioblastoma multiforme cells, Hinzmann [/bib_ref].
Moore and coworkers investigated the impact of nanographene (nGr) in U-138 glioblastoma cells. Cytotoxicity was measured in vitro using PrestoBlue cell viability assay after 24 h incubation. The results showed significant increase in the number of dead cells and the decrease in cell density after graphene treatment in the concentrations higher than 50 g/mL [bib_ref] Systemic administration of polymer-coated nano-graphene to deliver drugs to glioblastoma, Moore [/bib_ref].
Yuan et al. examined the cytotoxicity of graphene quantum dots (GQDs) with different surface modifications (NH 2 , COOH, and CO-N (CH 3 ) 2 ) in human neural glioma cells (C6) using MTT assay. Conversely, data analysis showed low cytotoxicity and good biocompatibility for all tested graphene nanomaterials even at the very high concentrations (200 g/mL) [bib_ref] Cellular distribution and cytotoxicity of graphene quantum dots with different functional groups, Yuan [/bib_ref]. Carboxylated graphene oxide (GO-COOH) and chlorotoxin-conjugated graphene oxide (CTX-GO) both had negligible toxic effects on C6 cells (80% of viability at concentrations of 3.0 g/mL, 7.5 g/mL, and 15.0 g/mL) [bib_ref] Chlorotoxin-conjugated graphene oxide for targeted delivery of an anticancer drug, Wang [/bib_ref]. Coating graphene with the multifunctional PLA-PEG (poly(lactide) and poly(ethylene glycol)) reduced the toxicity of uncoated graphene and did not show signs of dosedependent toxicity up to 250 g/mL [bib_ref] Systemic administration of polymer-coated nano-graphene to deliver drugs to glioblastoma, Moore [/bib_ref].
Interesting results were obtained by Oh et al. Scientists used MTT assay to examine the viability of SH-SY5Y cell line grown on partially functionalized graphene sheets with oxygen or fluorine. SH-SY5Y cells cultured on the oxygenated graphene sheets showed approximately 138% viability but only 50% viability on the fluorinated graphene compared to pristine graphene samples. The increase in cell proliferation can be explained by adhesion of the hydrophilic oxygenated graphene sheets to the cell surface [bib_ref] Neuroblastoma cells grown on fluorine or oxygen treated graphene sheets, Oh [/bib_ref].
3.6. Other Cancer Cells. Except described cancer cell lines where cytotoxic effect was predominant, some reports show only slight decrease in the cell viability with improved influence of graphene on the cell proliferation and survival [bib_ref] Flake sizedependent cyto and genotoxic evaluation of graphene oxide on in vitro..., De Marzi [/bib_ref] [bib_ref] Nano-graphene oxide for cellular imaging and drug delivery, Sun [/bib_ref] [bib_ref] Graphene oxide: a nonspecific enhancer of cellular growth, Ruiz [/bib_ref]. The cytotoxicity of graphene depends on various possible mechanisms including interactions with the cells or culture medium. de Marzi et al. using graphene oxide at growing concentrations (10, 50, and 100 g/mL) and in two different flake sizes (1320 nm and 130 nm) investigated the cytotoxic effect on CaCo2 human colorectal adenocarcinoma cell line. Both micro-and nano-GO exhibited high biocompatibility and increased CaCo2 cell proliferation slightly decreasing with higher concentrations of nano-GO. The 24 h comet assay showed that micro-GO flakes genotoxicity rose together with the used concentration, while nano-GO had no significant genotoxic effect on treated cells [bib_ref] Flake sizedependent cyto and genotoxic evaluation of graphene oxide on in vitro..., De Marzi [/bib_ref].
Beyond exerting little cytotoxic effects on the cells, Ruiz et al. observed morphological changes, cell enlargement, and better attachment to GO-coated slides of HT-29 mammalian colorectal adenocarcinoma cells (control glass slides and glass slides coated with 10 g of GO). The results indicated promotion of mammalian cell proliferation, spreading, and growth after graphene oxide exposure [bib_ref] Graphene oxide: a nonspecific enhancer of cellular growth, Ruiz [/bib_ref].
Wu et al. evaluated the cytotoxicity of graphene oxide (GO) on human multiple myeloma cells (RPMI-8226). Increasing GO concentration from 10 to 100 mg/L after 24 h treatment reduced the cell viability from 95.6% to 79.6%, respectively. Cells treated with GO were round with little cell shrinkage but with no typical apoptotic features. Annexin V-FITC/PI staining by flow cytometry showed no significant differences in the cell apoptotic rate between the untreated and GO-treated cells suggesting only slight cytotoxicity of GO [bib_ref] Cytotoxicity of graphene oxide and graphene oxide loaded with doxorubicin on human..., Wu [/bib_ref].
Sun and his group examined toxicity of single-layer pegylated graphene oxide sheets (NGO-PEG) soluble in buffers and serum. Incubation of Raji cells (Burkitt's lymphoma B lymphocytes) in various concentrations of NGO-PEG for 72 h showed no obvious toxicity except a slight delay of the cell growth at the highest concentration (150 mg/L) [bib_ref] Nano-graphene oxide for cellular imaging and drug delivery, Sun [/bib_ref].
Human prostate cancer cells (PC3) were incubated in the presence of different concentrations (0-180 g/ L) of chemically reduced graphene oxide (CRGO) and chitosan magnetic graphene nanoparticles (CMG) for 72 hours. The cytotoxicity was evaluated using the WST-1 assay and the results revealed dose-dependent increase in graphene oxide cytotoxicity while CMG nanoparticles did not show any toxicity at all the tested concentrations. Chitosan-coated graphene oxide is soluble in both organic and acidic aqueous solutions and less toxic than nonfunctionalized GO and hence has higher therapeutic efficacy [bib_ref] Multifunctional chitosan magnetic-graphene (CMG) nanoparticles: a theranostic platform for tumor-targeted co-delivery of..., Wang [/bib_ref]. New insights into specific cancer treatment were presented in the research on metastasis of prostate cancer cells PC3. With the low influence on the cell viability pristine graphene and GO effectively inhibited migration and invasion of these cancer cells with no apparent effect on the induction of apoptosis [bib_ref] The inhibition of migration and invasion of cancer cells by graphene via..., Zhou [/bib_ref].
Conventional therapeutic approaches to eradicate all cancer cells fail because of the presence of tumor-initiating cells that are resistant to drugs, chemotherapy, and radiation. Cancer stem cells (CSCs) constitute a minority of the overall cancer cell population, although they are highly invasive and tumorigenic and the inability of their efficient elimination results in disease relapse and formation of metastases [bib_ref] Targeting cancer stem cells with nanoparticle-enabled therapies, Burke [/bib_ref]. flakes of GO to inhibit selectively CSCs proliferation in multiple cell lines including breast, lung, ovarian, prostate, and pancreatic cancers. Two different grades of GO were used, small GO (0,2-2 m) and big GO . Both small and big GO flakes inhibited tumorsphere formation in all independent cancer cells. They did not affect the viability of non-CSCs but selectively targeted cancer stem cells. Analysis of these targeted actions showed that GO inhibited a number of several key signal transduction pathways related to cancer stem cells including antioxidant and interferon responses [bib_ref] Graphene oxide selectively targets cancer stem cells, across multiple tumor types: implications..., Fiorillo [/bib_ref].
# Conclusion
Graphene was first isolated in 2004 and since then its properties have been studied widely [bib_ref] Toxic response of graphene nanoplatelets in vivo and in vitro, Park [/bib_ref]. Graphene-based nanomaterials have boosted the development of the interdisciplinary research caused by their unique properties and possible applications in electronics and biotechnology. Single-atom-thick, two-dimensional sheet of sp 2 -hybridized carbon atoms arranged in a regular hexagonal pattern [bib_ref] Pharmaceutical applications of graphene-based nanosheets, Kim [/bib_ref] [bib_ref] A systems toxicology approach to the surface functionality control of graphene-cell interactions, Chatterjee [/bib_ref] owns extraordinary electrical and thermal properties, mechanical strength, and capability of biofunctionalization [bib_ref] Assessment of the toxic potential of graphene family nanomaterials, Guo [/bib_ref] [bib_ref] Biomedical applications of graphene, Shen [/bib_ref] [bib_ref] Graphene oxide can induce in vitro and in vivo mutagenesis, Liu [/bib_ref]. Graphene nanoparticles have been used as drug and gene delivery agents in multimodal imaging and could be useful in biomedicine and cancer therapy [bib_ref] The effects of graphene nanostructures on mesenchymal stem cells, Talukdar [/bib_ref]. Graphene is a nanomaterial whose chemical, physical, or mechanical properties and structure permit the active tissue integration of desirable cell types and tissue components suggesting the potential use in tissue engineering [bib_ref] Biomedical applications of graphene, Shen [/bib_ref] [bib_ref] Systemic administration of polymer-coated nano-graphene to deliver drugs to glioblastoma, Moore [/bib_ref] [bib_ref] Internalization and cytotoxicity of graphene oxide and carboxyl graphene nanoplatelets in the..., Lammel [/bib_ref]. Besides the research confirming graphene biocompatibility there are reports of dose-dependent graphene toxicity against cultured cells. However, most of these reports concentrate mainly on graphene oxide and reduced graphene oxide (rGO) prepared in solutions [bib_ref] Graphenebased materials biocompatibility: a review, Pinto [/bib_ref]. Graphene family nanomaterials include ultrathin graphite, few-layer graphene (FLG), graphene oxide (GO; from monolayer to few layers), reduced graphene oxide (rGO), and graphene nanosheets (GNS) [bib_ref] A systems toxicology approach to the surface functionality control of graphene-cell interactions, Chatterjee [/bib_ref]. Among the most frequently used graphene derivatives in the cytotoxicity study are GO, rGO and graphene quantum dots (GQD) with various surface modifications. Mainly studied cancer cells include lung, breast, cervical, liver, and nerve cancer cell lines [fig_ref] Table 1: Influence of graphene-based nanomaterials on cancer cells [/fig_ref].
Depending on the cell line and type of the nanomaterial, graphene can increase the viability [bib_ref] Flake sizedependent cyto and genotoxic evaluation of graphene oxide on in vitro..., De Marzi [/bib_ref] [bib_ref] Graphene oxide: a nonspecific enhancer of cellular growth, Ruiz [/bib_ref] or cause the cell death [bib_ref] A comparative study of cellular uptake and cytotoxicity of multi-walled carbon nanotubes,..., Zhang [/bib_ref]. In the study of de Marzi et al. GO shows a slight decrease in A549 cells viability while the same concentration and time of exposure result in increased cell viability in CaCo2 colorectal carcinoma cells [bib_ref] Flake sizedependent cyto and genotoxic evaluation of graphene oxide on in vitro..., De Marzi [/bib_ref]. Oxidizedgraphene nanoribbons (O-GNRs) water-solubilized with the amphiphilic polymer PEG-DSPE (O-GNR-PEG-DSPE) show significantly higher toxic effect on cervical cancer cells (HeLa) than on other cancer or normal tested cells [bib_ref] Cell specific cytotoxicity and uptake of graphene nanoribbons, Chowdhury [/bib_ref]. We can assume that reduced graphene oxide (rGO) is more cytotoxic than graphene oxide (GO) to lung, liver, and breast cancer cells [bib_ref] A systems toxicology approach to the surface functionality control of graphene-cell interactions, Chatterjee [/bib_ref] [bib_ref] Green synthesis of graphene and its cytotoxic effects in human breast cancer..., Gurunathan [/bib_ref] [bib_ref] Graphene-based antibacterial paper, Hu [/bib_ref]. However, the influence of rGO is similar among U87 nerve cancer cells and MCF-7 breast cancer cell line (IC50 = 85 mg/L and 80 mg/L, resp.) [bib_ref] Ultrasmall reduced graphene oxide with high near-infrared absorbance for photothermal therapy, Robinson [/bib_ref]. Graphene surface functionalization with different groups of various biomaterials such as PEG or dextran results in better nanomaterial biocompatibility. Pegylated graphene quantum dots (GQDs-PEG) exhibit very low or no toxicity against lung and cervical cancer cells even at very high concentrations (200 g/mL) [bib_ref] Cellular distribution and cytotoxicity of graphene quantum dots with different functional groups, Yuan [/bib_ref] [bib_ref] The in vitro and in vivo toxicity of graphene quantum dots, Chong [/bib_ref]. Pegylated graphene oxide (GO-PEG) [bib_ref] Nano-graphene oxide for cellular imaging and drug delivery, Sun [/bib_ref] [bib_ref] Ultrasmall reduced graphene oxide with high near-infrared absorbance for photothermal therapy, Robinson [/bib_ref] , dextran covered graphene oxide (GO-DEX) [bib_ref] In vitro and in vivo behaviors of dextran functionalized graphene, Zhang [/bib_ref] and fluorinated graphene oxide (FGO) [bib_ref] Fluorinated graphene oxide; a new multimodal material for biological applications, Romero-Aburto [/bib_ref] are more biocompatible than other graphene derivatives such as highly hydrogenated graphene (HHG) which after 24 h incubation reduce the viability of lung cancer cells (A549) to 26% [bib_ref] Cytotoxicity profile of highly hydrogenated graphene, Chng [/bib_ref].
Therefore, each graphene derivative may have diverse effect on the same cell type and the same graphene form can cause different reaction depending on the cell origin. Evaluations of the cytotoxicity and biocompatibility are an essential step in developing of any new biomaterial for in vivo biomedical applications. This review reveals that the toxicity of graphene nanomaterials depends not only on the graphene chemical structure, functionalization, size, concentration, and time of exposure but also on various possible mechanisms including interactions with different types of cells or culture medium components. Moreover the diversity of the samples and methods of the production hinder establishing of the biological impact of graphene [bib_ref] Graphene: one material, many possibilities-application difficulties in biological systems, Skoda [/bib_ref].
One of the proposed mechanisms underlying graphene cytotoxicity involves reactive oxygen species [bib_ref] Nanotoxicity of graphene and graphene oxide, Seabra [/bib_ref] [bib_ref] Comparative protein profile of human hepatoma HepG2 cells treated with graphene and..., Yuan [/bib_ref] while the others include plasma membrane damage, impairment of mitochondrial activity, DNA damage, and interaction with biomolecules which finally lead to apoptotic and/or necrotic cell death [bib_ref] Nanotoxicity of graphene and graphene oxide, Seabra [/bib_ref] [bib_ref] Green synthesis of graphene and its cytotoxic effects in human breast cancer..., Gurunathan [/bib_ref] [bib_ref] Graphenebased materials biocompatibility: a review, Pinto [/bib_ref]. Toxicity of graphene is desirable when used against cancer cells but not in case of surrounding healthy ones. It would be best to use graphene as a delivery agent for water insoluble drugs, antigens, antibodies, or nucleic acids and unload therapeutic molecules selectively inside the cancer cells to impair their activity [bib_ref] Cholesteryl hyaluronic acid-coated, reduced graphene oxide nanosheets for anti-cancer drug delivery, Miao [/bib_ref] [bib_ref] Furin-mediated sequential delivery of anticancer cytokine and small-molecule drug shuttled by graphene, Jiang [/bib_ref]. Use of graphene as a drug delivery agent has been recently the subject of numerous scientific researches [bib_ref] Carbon-based drug delivery carriers for cancer therapy, Lim [/bib_ref] [bib_ref] Biomedical applications of graphene, Shen [/bib_ref] [bib_ref] Systemic administration of polymer-coated nano-graphene to deliver drugs to glioblastoma, Moore [/bib_ref] [bib_ref] Chlorotoxin-conjugated graphene oxide for targeted delivery of an anticancer drug, Wang [/bib_ref] [bib_ref] Nano-graphene oxide for cellular imaging and drug delivery, Sun [/bib_ref] [bib_ref] Controlled release of doxorubicin from graphene oxide based charge-reversal nanocarrier, Zhou [/bib_ref] [bib_ref] Reduced graphene oxide nanosheets coated with an anti-angiogenic anticancer lowmolecular-weight heparin derivative..., Shim [/bib_ref] [bib_ref] Cytotoxicity of graphene oxide and graphene oxide loaded with doxorubicin on human..., Wu [/bib_ref] [bib_ref] Multifunctional chitosan magnetic-graphene (CMG) nanoparticles: a theranostic platform for tumor-targeted co-delivery of..., Wang [/bib_ref] [bib_ref] Carbon materials for drug delivery & cancer therapy, Liu [/bib_ref] [bib_ref] Furin-mediated sequential delivery of anticancer cytokine and small-molecule drug shuttled by graphene, Jiang [/bib_ref] [bib_ref] Targeted delivery system of nanobiomaterials in anticancer therapy: from cells to clinics, Jin [/bib_ref] [bib_ref] Functionalized graphene oxide nanoparticles for cancer cell-specific delivery of antitumor drug, Zhao [/bib_ref] [bib_ref] Delivery of paclitaxel using PEGylated graphene oxide as a nanocarrier, Xu [/bib_ref]. However, the mechanisms of cellular uptake and modes of action are still under investigation. In vitro studies regarding the influence of GFNs on mammalian cells give only a slight overview on the possible interactions with living organisms. The inconsistency of available data and the lack of sufficient information make it impossible to fully assess the suitability of graphene as a biomaterial. To understand better the impact of graphene further studies should be performed especially in vivo on the mechanisms of cell uptake and signaling combined with the results of long term effects of the materials internalization. More thorough research concerning graphene hemo-and biocompatibility together with the impact on immunological system would be essential to establish safe administration or implantation of GFNs. Yet the most important thing in graphene technology is to establish one universal and recurrent method of Chatterjee et al. [bib_ref] A systems toxicology approach to the surface functionality control of graphene-cell interactions, Chatterjee [/bib_ref] U87 nerve [bib_ref] Multifunctional chitosan magnetic-graphene (CMG) nanoparticles: a theranostic platform for tumor-targeted co-delivery of..., Wang [/bib_ref] production which would allow obtaining graphene with the same properties on large scale and in cost-effective manner. Therefore, detailed studies are required to explain the toxicity pathways of GFNs which would allow not only establishing the effect of graphene on cancer cells but also facilitating their proper use in medicine and cancer therapy.
[fig] Figure 2: Schematic toxicity mechanisms of graphene on human cancer cells. Graphene provides the formation of reactive oxygen species (ROS) which are the cause of DNA (fragmentation and condensation) and cell membrane damage (release of LDH, lipid peroxidation, and increase in MDA-malondialdehyde), mitochondrial disorders (reduction of mitochondrial membrane potential ΔΨ, increase in Ca 2+ ), and cell death. [/fig]
[table] Table 1: Influence of graphene-based nanomaterials on cancer cells.160 g/mL = <5% Slight reduction of cell viability Chong et al.[16] O-GNR-PEG-DSPE [/table]
[bib_ref] In vitro and in vivo effects of graphene oxide and reduced graphene..., Jaworski [/bib_ref] |
A structural inventory of native ribosomal ABCE1‐43S pre‐initiation complexes
Cryo-EM maps and resulting molecular models have been deposited in the EMDB and PDB databases and accession codes are listed in the "data availability" section as requested.Cryo-EM maps and resulting molecular models have been deposited in the EMDB and PDB databases and accession codes are listed in the "data availability" section as requested.NA NA NA NA NA
## 6th may 2020 1st editorial decision
Thank you for submit ting your manuscript report ing nat ive ribosomal ABCE1-4 3S pre-init iat ion complex st ruct ures for considerat ion by The EMBO Journal. We have now received three referee report s on your st udy, which are included below for your informat ion.
As you will see, the reviewers are overall posit ive and acknowledge the st udy's cont ribut ion to the field and it s qualit y. Nonet heless they also raise some concerns that would need to be addressed in a revised manuscript . In part icular referee #1's point s 1, 5 and 6, and referee #3's last major concern should be addressed, by adding to the discussion and by providing furt her experiment al dat a if needed. In addit ion, please also carefully respond to the ot her issues raised by the referees and revise the text and figures accordingly. When these concerns are resolved, we will be happy to consider this st udy furt her for publicat ion, and I would therefore now like to invit e you to prepare and submit a revised manuscript .
Please not e that it is our policy to allow only a single round of major revision. We recognize that lab work worldwide is current ly affect ed by the COVID-19/SARS-CoV-2 pandemic and can ext end the revision time when needed. In addit ion, we have ext ended our 'scooping prot ect ion policy' to cover the period required for a full revision. However, it is nonet heless import ant to clarify any quest ions and concerns at this st age.
Please feel free to cont act me should you have any furt her quest ions regarding the revision. Thank you for the opport unit y to consider your work for publicat ion. I look forward to receiving your revised manuscript . First off, I must apologize for the lat e review of this manuscript , as I have been flooded wit h such request s. This in no means reflect s a negat ive opinion of this work; on the cont rary, I found this a careful analysis of both native yeast and human initiation complexes by the Beckmann group. This is a deep and dense manuscript that is rich in novel structural information regarding complexes directly isolated from cells after gradient analysis. Of course, linking this to the biochemistry and mechanism is challenging, but this manuscript is worthy of publication as it provides novel structural perspectives on the complex eukaryotic initiation process. There are several issues that should be rectified in a revised manuscript, outlined below.
1. I had some issues with their subunit splitting assay and analysis. They included Dom/Hbs in their "splitting factors" rather than eRF1/eRF3 for reasons that were unclear in the text. My guess is that they just see poor splitting rates with eRF1/3 -they are 10 fold slower than Dom/Hbs -but they are looking at canonical processes throughout the rest of the paper here (and eRF1 would be present for recycling/re-initiation, not Dom), so it reads pretty strangely to me. Plus, the effect of adding 3j is *really* small, and their signal/noise in that assay is not great. To rectify this,the authors could pull out these results (they can still cite the other paper), and scale back this specific claim in the discussion: "We could corroborate the finding that eIF3j assists in ABCE1-dependent splitting by in vitro dissociation assays and furthermore we established that eIF3j remains bound to the 40S together with ABCE1 after the splitting cycle." 2. The second intro paragraph does not make the point that ABCE1's FeSD seems to push on eRF1/pelota to elicit subunit splitting. This may be conceptually helpful for some readers, especially considering the FeSD is heavily discussed. 3. References are missing for eIF5B statement at the very end of third intro paragraph? 4. The fourth intro paragraph feels like a list of unknown structural details, maybe it would be better to edit this down and call out new findings throughout the text instead? 5. In processes such as recycling and reinitiation, splitting would presumably occur in the presence of eRF1, yet the authors performed splitting assays with Dom34 instead. Although Dom34 elicits faster subunit splitting in vitro, are these results meaningful in the context of recycling? 6. The effects of 3j on splitting with Dom/Hbs in 1C/S1D are quite subtle. Comparing +/-3j, the addition of 3j not only increases the amount of subunits, but also increases the amount of 80S. The magnitude of the increase in 80S is substantial upon addition of 3j, and well outside what would be suggested by error bars for 80S. +3j/-SF and ++3j/-SF would also be good controls to include. The authors state "However, eIF3j alone did not exhibit any activity" so should include these results
## Referee #2:
In this follow-up to their structural study of an in vitro reconstituted, archaeal post-splitting complex with ABCE1 bound (Heuer et al. 2020), Beckmann and co-workers set out to investigate the extent to which ABCE1 remains associated with initiation factor-bound 40S (43S-pre-initiation or 48Sinitiation complexes). In order to do so, they obtained several cryo-EM structures of native small subunits from yeast and human cells, which show that ABCE1 stays bound to the SSU throughout 43S-PIC assembly and, to a lesser extent, in the 48S-IC. Moreover, the conformation of the ABCE1 NBDs differs from that observed in the complex reconstituted in vitro, seemingly as a result of an unknown molecule interacting with its nucleotide binding site, causing NBSI to remain in a semiopen conformation. Overall, this work sheds light on the various steps leading up to translation initiation in eukaryotes by providing a detailed series of cryo-EM snapshots, including a higher resolution view of human eIF3 interacting with the 43S PIC and of initiator tRNA interacting with eIF2 on the SSU.
The structural data are of good quality, are well described and the assignment of the various densities observed is well supported by earlier studies and by the LC/MS data. As a result, I only have a few minor comments: -In the discussion, the authors discuss a putative order of events during 43S PIC and 48S IC assembly. A schematic figure summarizing the discussion would be very helpful for the uninitiated reader.
-On p.7, it would be worth mentioning that complexes were affinity-purified directly from the lysate and not from the gradients. The way it is phrased at the moment is ambiguous.
-The order of the B and C panels inshould be inverted to match the order in which they are discussed in the text.
-Density for eIF3c could be shown in. At the moment it is difficult to make out in panel A.
Referee #3:
In the manuscript entitled "Structural inventory of native ribosomal ABCE1-43S pre-initiation complexes" Beckmann and colleagues investigate the coupling of recycling and initiation steps of eukaryotic translation by cryo-EM. In particular they focus on the role played by the ATPase ABCE1 (Rli1 in yeast). While ABCE1 primarily functions as a recycling factor, the authors of this manuscript find it to remain associated in 43S and 48S stages of re-initiation. In addition, by sorting 43S-PIC particles and employing focused re-classification a near complete picture of the 43S stage of eukaryotic translation initiation could be discerned. As such this manuscript, represents a substantial advance in our understanding of this stage of translation initiation and expands the structural role played by ABCE1 in this process. While this justifies the publication of this manuscript, I feel some re-structuring of the manuscript is in order to extend its accessibility to a broader readership.
Major concerns: -Most of the conclusions reached by the authors is based on the structural analysis of the human samples. However, they intermittantly refer to both the human and yeast samples within the manuscript. I presume the intention of the authors was to highlight that the observations they describe are validated by their simultaneous observation both in the human and yeast system. This is however confusing when reading the manuscript. Therefore, I would suggest to describe all structural findings based on the human samples and then summarise in one chapter the observation made in yeast, which justify the drawn conclusions.
-The authors describe that they were able to redefine existing models for the PCI-MPN core in particular and trace flexible parts of eIF3 in general. For the PCI-MPN core the claim to be able to correct register errors of former models. This is excellent and will be invaluable in future. However, in the main manuscript the authors only present Calpha traces of the final model or side chains in selected areas. Densities are only presented in the Supplemental section. To substantiate the claims (in particular the correction of registry shifts) the authors should provide figures along with densities in the main manuscript, where the density allows the reader to appreciate why registry shifts were corrected.
-regarding the hybrid state of ABCE1, the authors claim that the ominous extra density inis responsible for this. While this is plausible from their description, I feel the cannot exclude a contribution of eIF3j and/or eIF1. Since this is one of the central points of this manuscript, I feel there should be some experimental evidence to backup the authors´ claims. Considering that models are available Xl-MS should be able to reveal the identity of this factor?! Minor concerns: -Please add densities into the main manuscript figures, where interpretation of side chain interactions occur. This includes 5C, 5D, 6F, 7C and 7D.
## Reviewers reports and response
Referee #1:
First off, I must apologize for the late review of this manuscript, as I have been flooded with such requests. This in no means reflects a negative opinion of this work; on the contrary, I found this a careful analysis of both native yeast and human initi ation complexes by the Beckmann group. This is a deep and dense manuscript that is rich in novel structural information regarding complexes directly isolated from cells after gradient analysis. Of course, linking this to the biochemistry and mechanism is c hallenging, but this manuscript is worthy of publication as it provides novel structural perspectives on the complex eukaryotic initiation process. There are several issues that should be rectified in a revised manuscript, outlined below.
1. I had some issues with their subunit splitting assay and analysis. They included Dom/Hbs in their "splitting factors" rather than eRF1/eRF3 for reasons that were unclear in the text. My guess is that they just see poor splitting rates with eRF1/3 -they are 10 fold slower than Dom/Hbs -but they are looking at canonical processes throughout the rest of the paper here (and eRF1 would be present for recycling/re -initiation, not Dom), so it reads pretty strangely to me. Plus, the effect of adding 3j is *really* small, a nd their signal/noise in that assay is not great. To rectify this, the authors could pull out these results (they can still cite the other paper), and scale back this specific claim in the discussion: "We could corroborate the finding that eIF3j assists in ABCE1-dependent splitting by in vitro dissociation assays and furthermore we established that eIF3j remains bound to the 40S together with ABCE1 after the splitting cycle.
We apologize for being unclear on why we used the given set of factors. We agree with the referee that the canonical termination factors eRF1 and eRF3 would in principle be more appropriate than the Dom34/Hbs1 system. We have nevertheless chosen the Dom34/Hbs1 system for several reasons. First, in order to use eRF1 and eRF3 for termination-coupled recycling, stop-codon containing pre-termination complexes need to be prepared in high enough yields from a yeast in vitro translation system. This is in principle possible using for example the stalling sequence encoded by the gp48 uORF2 of cytomegalovirus (coding for the "CMV-stalling" sequence) or by adding a release-deficient eRF1 GGQmutant. However, both approaches would have created rather complicated cases of pre-termination stalled substrates for recycling. Therefore, for practical reasons we chose the use the mechanistically equivalent Dom34/Hbs1-ABCE1 system, since this system works with vacant 80S ribosomes and is -as the referee pointed out -also more efficient. Moreover, as for canonical splitting using termination factors, splitting of vacant ribosomes with Dom34/Hbs1 and ABCE1 results in the same end product, namely 40S subunits to which ABCE1 may remain bound under certain conditions, e.g. using nonhydrolyzable nucleotide analoga. Because of this mechanistic equivalence and the fact that also these subunits eventually enter a new phase of initiation, we consider this system suitable to investigate the effect of Hcr1/eIF3j on ABCE1-mediated splitting.
We also agree that the eIF3j-effect is indeed weak but still significant (see point 6). It may indeed be possible, that this effect would be stronger using canonical termination factors, but for reasons discussed above we haven't performed these assays with "real" termination complexes.
1st Authors' Response to Reviewers 18th Aug 2020
As requested, we actually confirmed the interaction between eIF3j and ABCE1 in native 40S initiation complexes by chemical crosslinking coupled to mass spectrometry (XL-MS) as well as with additional cryo-EM data of such a cross-linked 43S complex (see in detail in the response to reviewer 3). Since these data strengthen our in vitro assays, we kindly ask to not follow the reviewer's suggestion to scale back our claims and pull out these results. We rather added more controls and a clearer explanation of our system. For further justification please see also the response to point 6.
2. The second intro paragraph does not make the point that ABCE1's FeSD seems to push on eRF1/pelota to elicit subunit splitting. This may be conceptually helpful for some readers, especially considering the FeSD is heavily discussed. 4. The fourth intro paragraph feels like a list of unknown structural details, maybe it would be better to edit this down and call out new findings throughout the text instead?
Here, we are not sure, if we understand what the reviewers refers to. In the fourth paragraph we are giving a general overview over the different subunits and modules of eIF3 that we refer to later, but we don't give any structural details here and all this information is commonly known for a long time in the translation field. We thus would ask to not change the text there.
5. In processes such as recycling and reinitiation, splitting would presumably occur in the presence of eRF1, yet the authors performed splitting assays with Dom34 instead. Although Dom34 elicits faster subunit splitting in vitro, are these results meaningful in the context of recycling?
As explained in point 1.) we used the Dom34/Hbs1 system primarily for technical reasons, because we can use vacant 80S ribosomes for splitting. We still think that this is a reasonable approximation to study the effects of eIF3j/Hcr1 in splitting because the mechanism of ABCE1 action in both processes (recycling after canonical termination and after ribosome rescue) is most likely highly similar. While we cannot rule out that eIF3j could have a more prominent effect on splitting after termination, our and other data show, that eIF3j rather affects/assists ABCE1 itself. Thus, we think, when observing a Hcr1 effect on Dom34/Hbs1-mediated splitting, this is also conferrable to canonical termination.
Notably, the origin of the ABCE1-bound 43S/48S complexes found in our native pullouts is not clear and likely represents a mix. While assuming that these complexes originate mostly from recycling after canonical termination, an unknown portion will be the result of Dom34-dependent ribosome quality control pathways or of vacant ribosome splitting. For yeast, it was also reported that as a final step of 6. The effects of 3j on splitting with Dom/Hbs in 1C/S1D are quite subtle. Comparing +/-3j, the addition of 3j not only increases the amount of subunits, but also increases the amount of 80S. The magnitude of the increase in 80S is substantial upon addition of 3j, and well outside what would be suggested by error bars for 80S. +3j/-SF and ++3j/-SF would also be good controls to include. The authors state "However, eIF3j alone did not exhibit any activity" so should include these results
We can only partly agree with the referee at this point. It is correct that the amount of 80S slightly increases in the "+SF +eIF3j" condition, but when then comparing the error bars with the "+SF" condition (the control experiment without eIF3j), the increase in 80S is not "well outside what would be suggested by these error bars", but still within 1.5 standard deviations. Please note, that this control experiment has an overall higher deviation, as also observable in the original triplicates of the gradients inWhile we agree that the addition of only 4-fold molar excess (indicated as "+" in the figures) is not very substantial, the effect upon addition of 20-fold molar excess (indicated as "++" in the figures) shows a significant effect on the splitting rates. As suggested by the reviewer, we added the suggested control experiments ++3j -SF as well as a control with only Dom34-Hbs1 and only ABCE1 to
## Genzentrum, ludwig -maximilians-universität münchen
Referee #2:
In this follow-up to their structural study of an in vitro reconstituted, archaeal post-splitting complex with ABCE1 bound (Heuer et al. 2020), Beckmann and co-workers set out to investigate the extent to which ABCE1 remains associated with initiation factor-bound 40S (43S-pre-initiation or 48S-initiation complexes). In order to do so, they obtained several cryo-EM structures of native small subunits from yeast and human cells, which show that ABCE1 stays bound to the SSU throughout 43S-PIC assembly and, to a lesser extent, in the 48S-IC. Moreover, the conformation of the ABCE1 NBDs differs from that observed in the complex reconstituted in vitro, seemingly as a result of an unknown molecule interacting with its nucleotide binding site, causing NBSI to remain in a semi-open conformation. Overall, this work sheds light on the various steps leading up to translation initiation in eukaryotes by providing a detailed series of cryo-EM snapshots, including a higher resolution view of human eIF3 interacting with the 43S PIC and of initiator tRNA interacting with eIF2 on the SSU.
The structural data are of good quality, are well described and the assignment of the various densities observed is well supported by earlier studies and by the LC/MS data. As a result, I only have a few minor comments:
We are happy about the reviewer's overall very positive comments.
-In the discussion, the authors discuss a putative order of events during 43S PIC and 48S IC assembly. A schematic figure summarizing the discussion would be very helpful for the uninitiated reader.
As suggested by the reviewer we added a cartoon summarizing our findings in-On p.7, it would be worth mentioning that complexes were affinity-purified directly from the lysate and not from the gradients. The way it is phrased at the moment is ambiguous.
The samples used for quantitative mass spectrometry (LC-MS/MS) were indeed affinity purified from the lysates and not from the gradient peaks. We corrected this in the revised manuscript.
-The order of the B and C panels inshould be inverted to match the order in which they are discussed in the text.
We agree that this should be consistent. While we kept the order of the two panels, we adjusted the text accordingly.
-Density for eIF3c could be shown in. At the moment it is difficult to make out in panel A.
We added an additional panel into show the density for the N-terminus of eIF3c.
## Genzentrum, ludwig -maximilians-universität münchen
Referee #3:
In the manuscript entitled "Structural inventory of native ribosomal ABCE1-43S pre-initiation complexes" Beckmann and colleagues investigate the coupling of recycling and initiation steps of eukaryotic translation by cryo-EM. In particular they focus on the role played by the ATPase ABCE1 (Rli1 in yeast). While ABCE1 primarily functions as a recycling factor, the authors of this manuscript find it to remain associated in 43S and 48S stages of re-initiation. In addition, by sorting 43S-PIC particles and employing focused re-classification a near complete picture of the 43S stage of eukaryotic translation initiation could be discerned. As such this manuscript, represents a substantial advance in our understanding of this stage of translation initiation and expands the structural role played by ABCE1 in this process. While this justifies the publication of this manuscript, I feel some re-structuring of the manuscript is in order to extend its accessibility to a broader readership.
Major concerns: -Most of the conclusions reached by the authors is based on the structural analysis of the human samples. However, they intermittantly refer to both the human and yeast samples within the manuscript. I presume the intention of the authors was to highlight that the observations they describe are validated by their simultaneous observation both in the human and yeast system. This is however confusing when reading the manuscript. Therefore, I would suggest to describe all structural findings based on the human samples and then summarise in one chapter the observation made in yeast, which justify the drawn conclusions.
While we agree with the referee that it may be clearer to describe all findings with the human sample, we need to point out that our findings in yeast and human complexes are not always redundant, in a few cases rather complementary. For example, the conformation of eIF3j in ABCE1 containing classes differs between human and yeast, and in the yeast structures eIF1A is present while in the human structure it is absent. Moreover, that ABCE1-containing 48S complex was only found in yeast. For this reason, we decided to arrange the manuscript in a way, that yeast and human complexes from similar stages of initiation can be compared. We thus feel that restructuring the manuscript as suggested by the reviewer would not lead to an improved intelligibility. Nevertheless, we tried to simplify the manuscript wherever possible in order to make it less confusing.
-The authors describe that they were able to redefine existing models for the PCI-MPN core in particular and trace flexible parts of eIF3 in general. For the PCI-MPN core the claim to be able to correct register errors of former models. This is excellent and will be invaluable in future. However, in the main manuscript the authors only present C-alpha traces of the final model or side chains in selected areas. Densities are only presented in the Supplemental section. To substantiate the claims (in particular the correction of registry shifts) the authors should provide figures along with densities in the main manuscript, where the density allows the reader to appreciate why registry shifts were corrected.
As suggested, we added an Expanded View model-to-map fits (including side chains) from PCI-MPN core regions , as well as a new Appendix , where the register shifts are shown. For the sake of clarity, however, we decided to not show this in the main figure. Notably in EMBO J, Expanded View figures are displayed together with the main figure online. This should give the reader an excellent possibility to retrace the claims made in the main figure.
-regarding the hybrid state of ABCE1, the authors claim that the ominous extra density inis GENZENTRUM, LUDWIG -MAXIMILIANS-UNIVERSITÄT MÜNCHEN responsible for this. While this is plausible from their description, I feel the cannot exclude a contribution of eIF3j and/or eIF1. Since this is one of the central points of this manuscript, I feel there should be some experimental evidence to backup the authors´ claims. Considering that models are available Xl-MS should be able to reveal the identity of this factor?!
We thank the referee for his request and we agree that the identification of the extra density at the ABCE1 nucleotide binding site would significantly strengthen the central claim of the manuscript. To that end, we prepared a similar native 40S initiation complex sample as used in the original manuscript: We noted that ABCE1 (Rli1) and eIF3j (Hcr1)-containing (pre)-initiation complexes are enriched in native pullouts using a TAP-tagged Nip1 (eIF3c) as a bait. We prepared native small 40S subunits from this sample and performed XL-MS in collaboration with the Herzog lab in the Gene Center Munich. In addition, we subjected this crosslinked sample to cryo-EM and single particle analysis. Amongst a large number of crosslinks (43 inter-protein and 74 intra-protein crosslinks) largely confirming our previously presented structural data, XL-MS yielded two robust inter-protein crosslinks between a lysine in the eIF3j N-terminal domain and two lysines in the nucleotide-binding cleft of ABCE1-NBD1, unambiguously confirming our claim, that the N-terminus of eIF3j binds into the nucleotide binding cleft of ABCE1-NBD1. Further, with this new sample, we could improve the local resolution of the 40S-bound ABCE1-eIF3j cryo-EM map to around 3 Å. This allowed us to describe the structure of 40S-bound eIF3j for the first time in molecular detail. We now observe a direct density connection between the 6-helix bundle of the eIF3j dimer and the "extra density" in ABCE1. Unfortunately, the local resolution was still too low to build a molecular model for the eIF3j region (1-135) preceding the three C-terminal helices and thus for the peptide that accommodates within ABCE1. Yet, the position determined by the XL-MS experiment, 18 amino acids N-terminal of the eIF3j helix bundle (K118), would perfectly agree with extra density within ABCE1. Taken together, we now can state with high certainty that the eIF3j N-terminal domain binds to the NBSI of ABCE1 and stabilizes this asymmetric ATPase in a novel hybrid conformation.
In addition, the improved resolution allowed us to also identify and de novo build the ultimate Cterminus of eIF3j which locates in the mRNA entry channel (also confirmed by XL-MS). This finding explains previous experiments showing that mRNA-and eIF3j binding are antagonistic, which plays a role during mRNA recycling from the 40S as well as influencing/coordinating mRNA loading and even startsite selection during initiation.
Minor concerns:
-Please add densities into the main manuscript figures, where interpretation of side chain interactions occur. This includes 5C, 5D, 6F, 7C and 7D.
As mentioned above, for better clarity, we added model-to-map fits as new Expanded View 10th Sep 2020 2nd -Editorial Decision Thank you for submit ting your revised manuscript , we have now received the report s from the three init ial referees (see comment s below). I am pleased to say that they overall find that their comment s have been sat isfact orily addressed and now support publicat ion. Referee #3 raises a minor issue regarding how model qualit y is report ed, which can be resolved in the final revised version. In addit ion, I would like to ask you to also address a number of edit orial issues that are list ed in det ail below. Please make any changes to the manuscript text in the at tached document only using the "t rack changes" opt ion. Once these remaining issues are resolved, we will be happy to formally accept the manuscript for publicat ion.
Thank you again for giving us the chance to consider your manuscript for The EMBO Journal. I look forward to receiving your final revision. Please feel free to cont act me if you have furt her quest ions regarding t he revision or any of t he specific point s list ed below. 3. Were any steps taken to minimize the effects of subjective bias when allocating animals/samples to treatment (e.g. randomization procedure)? If yes, please describe.
For animal studies, include a statement about randomization even if no randomization was used.
4.a. Were any steps taken to minimize the effects of subjective bias during group allocation or/and when assessing results (e.g. blinding of the investigator)? If yes please describe. The main method used in this publicatiomn is cryo-EM followed by single particle analysis. Sample size (number of particles) was chosen to enable molecular interpretation of the obtained maps.
NA NA NA definitions of statistical methods and measures:
Any descriptions too long for the figure legend should be included in the methods section and/or with the source data.
In the pink boxes below, please ensure that the answers to the following questions are reported in the manuscript itself. Every question should be answered. If the question is not relevant to your research, please write NA (non applicable). We encourage you to include a specific subsection in the methods section for statistics, reagents, animal models and human subjects.
## B-statistics and general methods
an explicit mention of the biological and chemical entity(ies) that are being measured. an explicit mention of the biological and chemical entity(ies) that are altered/varied/perturbed in a controlled manner.
the exact sample size (n) for each experimental group/condition, given as a number, not a range; a description of the sample collection allowing the reader to understand whether the samples represent technical or biological replicates (including how many animals, litters, cultures, etc.). a statement of how many times the experiment shown was independently replicated in the laboratory.
Source Data should be included to report the data underlying graphs. Please follow the guidelines set out in the author ship guidelines on Data Presentation.
## Captions
Each figure caption should contain the following information, for each panel where they are relevant:
a specification of the experimental system investigated (eg cell line, species name). the assay(s) and method(s) used to carry out the reported observations and measurements
The data shown in figures should satisfy the following conditions:
the data were obtained and processed according to the field's best practice and are presented to reflect the results of the experiments in an accurate and unbiased manner. figure panels include only data points, measurements or observations that can be compared to each other in a scientifically meaningful way. graphs include clearly labeled error bars for independent experiments and sample sizes. Unless justified, error bars should not be shown for technical replicates. if n< 5, the individual data points from each experiment should be plotted and any statistical test employed should be justified |
Management of macular edema due to central retinal vein occlusion – The role of aflibercept
Central retinal vein occlusion (CRVO) can cause vision loss. The pathogenesis of CRVO involves a thrombus formation leading to increased retinal capillary pressure, increased vascular permeability, and possibly retinal neovascularization. Vision loss due to CRVO is commonly caused by macular edema. Multiple treatment modalities have been used to treat macular edema. Currently, the most common therapy used is intravitreal inhibition of vascular endothelial growth factor (VEGF). The three most widely used agents are aflibercept, bevacizumab, and ranibizumab and they are effective at blocking VEGF. In addition, intraocular steroids can be used to treat macular edema. This review will briefly cover the treatment options and discuss in greater detail the efficacy and safety of aflibercept.
# Background
## C entral retinal vein occlusion (crvo) is
a common cause of sight-threatening retinal disease. [bib_ref] Natural history of branch retinal vein occlusion: An evidence-based systematic review, Rogers [/bib_ref] The pathogenesis of CRVO is believed to involve vascular endothelial damage and compression of the retinal vein, leading to thrombus formation. This leads to increased retinal capillary pressure, which causes transudation into the extracellular space and macular edema. [bib_ref] Anti-vascular endothelial growth factor for macular oedema secondary to branch retinal vein..., Mitry [/bib_ref] [bib_ref] Anti-vascular endothelial growth factor for macular oedema secondary to central retinal vein..., Braithwaite [/bib_ref] CRVO may be further divided into ischemic or nonischemic subtypes, with the ischemic subtype having worse visual outcomes. [bib_ref] Anti-vascular endothelial growth factor for macular oedema secondary to branch retinal vein..., Mitry [/bib_ref] [bib_ref] Anti-vascular endothelial growth factor for macular oedema secondary to central retinal vein..., Braithwaite [/bib_ref] In a recent natural history cohort study, the baseline visual acuity was 20/100 or better for 78% of patients with nonischemic CRVO, while only 1% of patients with ischemic CRVO achieved 20/100 or better. Final visual acuity was 20/100 or better in 83% of patients with nonischemic CRVO versus only 12% of patients with ischemic CRVO. [bib_ref] Natural history of visual outcome in central retinal vein occlusion, Hayreh [/bib_ref] Visual loss from CRVO is most commonly caused by macular edema but may also be caused by macular ischemia, neovascular glaucoma, retinal neovascularization, or a combination of these complications. [bib_ref] Anti-vascular endothelial growth factor for macular oedema secondary to branch retinal vein..., Mitry [/bib_ref] [bib_ref] Anti-vascular endothelial growth factor for macular oedema secondary to central retinal vein..., Braithwaite [/bib_ref] Imaging from a patient with CRVO is shown in [fig_ref] Figure 1: Retinal imaging of central retinal vein occlusion [/fig_ref].
Globally, CRVO has a large impact with an estimated 2.5 million affected people. [bib_ref] The prevalence of retinal vein occlusion: Pooled data from population studies from..., Rogers [/bib_ref] [bib_ref] The burden of disease of retinal vein occlusion: Review of the literature, Laouri [/bib_ref] Prevalence rates are similar across the world, ranging between 0.3% and 2.1% in data collected from international epidemiological studies. [bib_ref] The prevalence of retinal vein occlusion: Pooled data from population studies from..., Rogers [/bib_ref] [bib_ref] Prevalence and associations of retinal vein occlusion in Australia. The Blue Mountains..., Mitchell [/bib_ref] [bib_ref] Vein occlusion in Chinese subjects, Liu [/bib_ref] [bib_ref] Prevalence and systemic risk factors for retinal vein occlusion in a general..., Yasuda [/bib_ref] [bib_ref] Prevalence and risk factors of retinal vein occlusion in an Asian population, Lim [/bib_ref] [bib_ref] The epidemiology of retinal vein occlusion: The Beaver Dam Eye Study, Klein [/bib_ref] [bib_ref] Traditional and novel cardiovascular risk factors for retinal vein occlusion: The multiethnic..., Cheung [/bib_ref] CRVO prevalence rates do not significantly vary with regard to race or gender. Increasing age is the biggest risk factor for developing CRVO, which is likely due to increased arteriosclerosis and other systemic and ocular risk factors. [bib_ref] The prevalence of retinal vein occlusion: Pooled data from population studies from..., Rogers [/bib_ref] [bib_ref] The burden of disease of retinal vein occlusion: Review of the literature, Laouri [/bib_ref] Other risk factors for developing CRVO include hypertension, diabetes mellitus, atherosclerosis, high cholesterol, thrombophilia, glaucoma, and other inflammatory and autoimmune conditions. [bib_ref] Risk factors for central and branch retinal vein occlusion: A meta-analysis of..., Kolar [/bib_ref] An extensive laboratory workup for the cause of the CRVO is usually not recommended because the majority of testing will not reveal a systemic coagulopathy. However, in younger patients, especially those with bilateral CRVO disease, a thrombophilic workup is recommended. [bib_ref] Laboratory evaluation of hypercoagulable states in patients with central retinal vein occlusion..., Lahey [/bib_ref] A workup including blood pressure measurement for hypertension screening, intraocular pressure (IOP) measurement to screen for glaucoma, complete blood count with glucose to evaluate if the patient is a diabetic, and a lipid panel to see if the patient is hyperlipidemic is recommended in younger patients under 56-year-old with CRVO. Further work up with tests for thrombophilias such as homocysteinemia, anticardiolipin antibodies, antiphospholipid antibodies, activated protein C resistance, antithrombin III activity, and proteins S and C can be investigated if the initial workup is negative on a case by case basis. [bib_ref] Laboratory evaluation of hypercoagulable states in patients with central retinal vein occlusion..., Lahey [/bib_ref] Additional testing may be performed to detect other systemic disorders associated with CRVO including but not limited to cardiovascular disease, cerebrovascular disease, chronic obstructive pulmonary disease, lupus, and blood dyscrasias.
## Antivascular endothelial growth factor agents in the treatment of central retinal vein occlusion
Vascular endothelial growth factor (VEGF) is a cytokine produced by hypoxic cells to stimulate vascular permeability and proliferation by binding to endothelial cell receptors. This increased vascular permeability leads to the development of macular edema in CRVO. The development of macular edema may cause vision loss. Anti-VEGF agents such as pegaptanib, bevacizumab, ranibizumab, and aflibercept work by binding to VEGF to inhibit endothelial receptor binding. [fig_ref] Figure 2: Ocular coherence tomography changes with antivascular endothelial growth factor treatment [/fig_ref] demonstrates the effectiveness of anti-VEGF agents in decreasing macular edema.
The CRUISE study was a double-masked, randomized, controlled trial (RCT) evaluating the efficacy of ranibizumab injections in treating CRVO compared to sham injections. [bib_ref] Ranibizumab for macular edema following central retinal vein occlusion: Six-month primary end..., Brown [/bib_ref] Participants were randomized to one of three groups to receive monthly sham intraocular injections or intraocular ranibizumab injections of 0.3 mg or 0.5 mg. Significant statistical difference was observed between the ranibizumab groups and the sham group for those achieving 15-letter improvement in BCVA at 7 days, 1 month, and 6 months. At 6 months, 46.2% and 47.7% of participants in the 0.3 mg and 0.5 mg ranibizumab groups, respectively, had achieved 15-letter improvement in BCVA compared to only 16.9% of participants in the sham group. [bib_ref] Ranibizumab for macular edema following central retinal vein occlusion: Six-month primary end..., Brown [/bib_ref] A recent meta-analysis of six RCTs described the positive outcomes associated with anti-VEGF injections compared to sham injections in patients with macular edema secondary to CRVO [ [fig_ref] Table 1: Comparison of antivascular endothelial growth factor agents in the treatment of central... [/fig_ref] ]. [bib_ref] Anti-vascular endothelial growth factor for macular oedema secondary to central retinal vein..., Braithwaite [/bib_ref] These trials include the GALILEO and COPERNICUS trials for aflibercept, CRUISE and ROCC trials for ranibizumab, the Epstein study for bevacizumab, and the Wroblewski study for pegaptanib sodium. [bib_ref] VEGF trap-eye for macular oedema secondary to central retinal vein occlusion: 6-month..., Holz [/bib_ref] [bib_ref] Intravitreal aflibercept injection for macular edema due to central retinal vein occlusion:..., Heier [/bib_ref] [bib_ref] Benefit from bevacizumab for macular edema in central retinal vein occlusion: Twelve-month..., Epstein [/bib_ref] [bib_ref] Pegaptanib sodium for macular edema secondary to central retinal vein occlusion, Wroblewski [/bib_ref] [bib_ref] Efficacy of ranibizumab in patients with macular edema secondary to central retinal..., Kinge [/bib_ref] This meta-analysis showed that patients receiving anti-VEGF treatment were 2.71 times more likely to gain 15 letters or more of visual acuity after 6 months compared to patients who were treated with sham injections. [bib_ref] Anti-vascular endothelial growth factor for macular oedema secondary to central retinal vein..., Braithwaite [/bib_ref] The likelihood of gaining 15 letters of visual acuity at 6 months did not show a statistically significant difference among the different anti-VEGF agents. However, aflibercept had the highest rate of 15-letter improvement in visual acuity. Nearly sixty percent (60.2%) of those receiving aflibercept injections gained 15 letters. [bib_ref] VEGF trap-eye for macular oedema secondary to central retinal vein occlusion: 6-month..., Holz [/bib_ref] Visual acuity was largely maintained at 12 months using anti-VEGF dosing as needed in the applicable studies. [bib_ref] Anti-vascular endothelial growth factor for macular oedema secondary to central retinal vein..., Braithwaite [/bib_ref] Similarly, those who received anti-VEGF agents had a mean visual a acuity 15.23 letters higher than those receiving sham injections at 6 months. [bib_ref] Anti-vascular endothelial growth factor for macular oedema secondary to central retinal vein..., Braithwaite [/bib_ref] These anti-VEGF agents not only improved visual acuity but were also protective in deterioration of visual acuity. The same meta-analysis revealed that receiving intravitreal anti-VEGF treatment was associated with an 80% lower risk of losing 15 letters of visual acuity at 6 months. [bib_ref] Anti-vascular endothelial growth factor for macular oedema secondary to central retinal vein..., Braithwaite [/bib_ref] Anti-VEGF agents also demonstrated the ability to decrease macular edema. Anti-VEGF agents caused an average reduction of 267.4 μm in central retinal thickness (CRT) compared to participants who received sham injections. [bib_ref] Anti-vascular endothelial growth factor for macular oedema secondary to central retinal vein..., Braithwaite [/bib_ref] The positive outcomes from using anti-VEGF agents in the treatment of CRVO comes with no increased risk of IOP elevation or cataract progression. Anti-VEGF therapy was associated with decreased risk of developing iris or retinal neovascularization and neovascular glaucoma compared to sham groups at 6 months. Endophthalmitis, rhegmatogenous retinal detachment, and retinal artery occlusion, which are all known complications of intravitreal injections, occurred at extremely low frequency in all the studies included in the meta-analysis. [bib_ref] VEGF trap-eye for macular oedema secondary to central retinal vein occlusion: 6-month..., Holz [/bib_ref] [bib_ref] Intravitreal aflibercept injection for macular edema due to central retinal vein occlusion:..., Heier [/bib_ref] [bib_ref] Benefit from bevacizumab for macular edema in central retinal vein occlusion: Twelve-month..., Epstein [/bib_ref] [bib_ref] Pegaptanib sodium for macular edema secondary to central retinal vein occlusion, Wroblewski [/bib_ref] [bib_ref] Efficacy of ranibizumab in patients with macular edema secondary to central retinal..., Kinge [/bib_ref] No systemic adverse events were identified in these studies at 6 months. Anti-VEGF therapy for CRVO macular edema is effective at improving and maintaining visual acuity with an excellent safety profile.
## Aflibercept in the treatment of central retinal vein occlusion
Aflibercept is a receptor fusion protein that includes the second domain of human VEGF receptor 1 and the third domain of VEGF receptor 2. These are fused to the Fc domain of human immunoglobulin G1. [bib_ref] Cytokine traps: Multi-component, high-affinity blockers of cytokine action, Economides [/bib_ref] Aflibercept has been designed to have a greater binding affinity for VEGF than bevacizumab and ranibizumab and has shown this ability in vitro. [bib_ref] Binding and neutralization of vascular endothelial growth factor (VEGF) and related ligands..., Papadopoulos [/bib_ref] Aflibercept has been associated with significant visual improvement for patients with CRVO macular edema and the Food and Drug Administration has been approved for this indication since September 2012. Two of the most important RCTs evaluating the use of aflibercept for the treatment of CRVO macular edema were the COPERNICUS and GALILEO studies.
In the COPERNICUS trial, patients with CRVO-related macular edema with a BCVA between 20/40 and 20/320 (Early Treatment Diabetic Retinopathy Study [ETDRS] 73-24 letters) and a CRT of ≥250 μm were randomized into two groups. The first group received a 2.0 mg intravitreal injection of aflibercept every 4 weeks for 24 weeks, while the second group to receive a sham injection at the same time intervals. [bib_ref] Vascular endothelial growth factor trap-eye for macular edema secondary to central retinal..., Boyer [/bib_ref] From weeks 24 to 52, all the patients were evaluated monthly and received an intravitreal aflibercept injection if they had a >50 μm CRT increase from their lowest previous measurement, persistent diffuse edema ≥250 μm, new or persistent cystic retinal changes or subretinal fluid, or a loss of ≥5 letters in BCVA. All the patients were evaluated at a minimum interval of 3 months from weeks 52 to 100 to assess the need for intravitreal aflibercept injections based on the same criteria. [bib_ref] Vascular endothelial growth factor trap-eye for macular edema secondary to central retinal..., Boyer [/bib_ref] The GALILEO trial enrolled patients with CRVO-related macular edema with the same visual acuity and CRT requirements in the COPERNICUS trial. Patients were randomized to the treatment group, which received a 2.0 mg intravitreal injection of aflibercept every 4 weeks for 20 weeks, or the sham group, which received a sham injection at the same interval. [bib_ref] VEGF trap-eye for macular oedema secondary to central retinal vein occlusion: 6-month..., Holz [/bib_ref] From weeks 24 to 48, patients in the aflibercept group received another 2.0 mg injection of aflibercept if they met the same criteria in the COPERNICUS trial, while the sham group participants continued to receive sham injections. From weeks 52 to 68, all patients were evaluated every 8 weeks and received an intravitreal aflibercept injection if they met the specified criteria. [bib_ref] VEGF trap-eye for macular oedema secondary to central retinal vein occlusion: 6-month..., Holz [/bib_ref] After 6 months, 56.1% of patients who received aflibercept injections had a 15-letter increase in visual acuity versus 12.3% receiving sham injections in the COPERNICUS study. [bib_ref] Vascular endothelial growth factor trap-eye for macular edema secondary to central retinal..., Boyer [/bib_ref] About 60.2% of patients receiving aflibercept injections had at 15-letter improvement compared to 22.1% of participants receiving sham injections in the GALILEO study. [bib_ref] VEGF trap-eye for macular oedema secondary to central retinal vein occlusion: 6-month..., Holz [/bib_ref] Combined, 58.1% of participants in these studies had a 15-letter increase in visual acuity after receiving aflibercept injections, and participants receiving aflibercept were 3.37 times more likely to have a 15-letter increase in visual acuity compared to those who receive sham injections. [bib_ref] Anti-vascular endothelial growth factor for macular oedema secondary to central retinal vein..., Braithwaite [/bib_ref] After the 6-month primary endpoint, the GALILEO and COPERNICUS trials continued to show promising results for treatment of CRVO macular edema with aflibercept. At week 52, 55.3% of participants in the original aflibercept group had gained ≥15 letters of visual acuity, and 49.1% of the participants in the aflibercept had achieved this increase in visual acuity at 100 weeks while adhering to study protocol. [bib_ref] Intravitreal aflibercept injection for macular edema due to central retinal vein occlusion:..., Heier [/bib_ref] Similarly, 60.2% of participants in the aflibercept group of the GALILEO study had gained ≥15 letters of visual acuity compared to 32.4% in the original sham group at week 52. By 76 weeks, 57.3% of the aflibercept group participants had achieved this increase in visual acuity compared to 29.4% of participants in the sham group. [bib_ref] VEGF trap-eye for macular oedema secondary to central retinal vein occlusion: 6-month..., Holz [/bib_ref] The findings of the COPERNICUS and GALILEO trials suggest that patients with CRVO-related macular edema may benefit from early aflibercept injections following the initial event. Visual outcomes were best in the groups that initially received aflibercept injections rather than the groups that initially received sham injections and then received aflibercept as needed many weeks later. Intravitreal aflibercept was generally well tolerated, with the most common adverse events being those typically associated with intravitreal injections or events related to the progression of underlying CRVO. When the intervals between treatments were increased in both trials, decreased visual and anatomic gains were observed. This suggests that the treatment interval may be extended after the initiation of treatment, but the monitoring and treatment intervals should be chosen carefully based on clinician discretion. [bib_ref] VEGF trap-eye for macular oedema secondary to central retinal vein occlusion: 6-month..., Holz [/bib_ref] No significant difference in the rate of adverse events was noted between the aflibercept groups and sham groups for both trials. [bib_ref] VEGF trap-eye for macular oedema secondary to central retinal vein occlusion: 6-month..., Holz [/bib_ref] [bib_ref] Intravitreal aflibercept injection for macular edema due to central retinal vein occlusion:..., Heier [/bib_ref] The results of the COPERNICUS and GALILEO trials suggest that early, regular intravitreal
## A possible role for switching antivascular endothelial growth factor agents
There is limited data on how to treat eyes with persistent macular edema and a suboptimal response to a particular anti-VEGF agent. Currently, retina specialists do not have a consensus definition of a suboptimal responder.
Multiple retrospective studies have indicated that aflibercept may be an effective agent in treating CRVO refractory to other anti-VEGF agents. Eadie et al. described a review of six patients who received a minimum of 10 monthly intravitreal ranibizumab or bevacizumab injections and were transitioned to intravitreal aflibercept every 4-6 weeks for refractory macular edema. [bib_ref] Response to aflibercept as secondary therapy in patients with persistent retinal edema..., Eadie [/bib_ref] All of the patients had improvement in their macular edema with a mean decrease in central optical coherence tomography thickness of 316 μm at 1 month and 290 μm at 3 months. Three of the six patients had lasting modest VA improvement, while four of the six had subjective visual improvement. [bib_ref] Response to aflibercept as secondary therapy in patients with persistent retinal edema..., Eadie [/bib_ref] Similarly, Lehmann-Clarke et al. reported a retrospective series of six patients with CRVO who were switched to aflibercept for persistent macular edema after a minimum of 6 monthly ranibizumab injections. These patients had a mean improvement of 9.2 ETDRS letters and a mean decreased central macular thickness of 248.0 μm after switching to monthly intravitreal aflibercept. [bib_ref] The effect of switching ranibizumab to aflibercept in refractory cases of macular..., Lehmann-Clarke [/bib_ref] In a retrospective study by Pfau et al., 13 patients had 48 months of persistent macular edema due to CRVO despite treatment with ranibizumab and/or bevacizumab. These eyes were switched to treatment with aflibercept. [bib_ref] Clinical outcome after switching therapy from ranibizumab and/ or bevacizumab to aflibercept..., Pfau [/bib_ref] After 1 year of aflibercept treatment, the mean injection interval increased by 0.51 months, and the relapse-free interval increased by 3.02 weeks. The mean ETDRS score increased from 66.15 to 76.54 letters, and the mean CRT decreased by 195.84 μm. [bib_ref] Clinical outcome after switching therapy from ranibizumab and/ or bevacizumab to aflibercept..., Pfau [/bib_ref] Another retrospective study was performed by Papakostas et al. who evaluated the efficacy of aflibercept injections in 42 eyes from 42 patients with CRVO-related macular edema refractory to treatment with ranibizumab and/or bevacizumab. Median visual acuity before aflibercept treatment was 20/126. One month after switching to aflibercept, acuity was 20/89, and acuity was 20/100 at the end of the follow-up period, which averaged 14 months. [bib_ref] Intravitreal aflibercept for macular oedema secondary to central retinal vein occlusion in..., Papakostas [/bib_ref] Median CRT before aflibercept was 536 μm, 293.5 μm at 1 month after the first aflibercept injection, and 279 μm at the end of the follow-up period. These results are limited by their retrospective nature and possible recall bias.
It is possible that aflibercept may improve visual acuity and anatomy of patients with CRVO-related macular edema that is not responsive to intravitreal injections of bevacizumab and/or ranibizumab. In addition, some clinicians suggest adding intravitreal steroids in eyes with a suboptimal response to aflibercept, ranibizumab, or bevacizumab. [bib_ref] Outcome of intravitreal dexamethasone implant for the treatment of ranibizumab-resistant macular edema..., Manousaridis [/bib_ref] [bib_ref] Comparison of ozurdex and triamcinolone acetonide for refractory cystoid macular edema in..., Ozkok [/bib_ref] Prospective studies are necessary to determine how to best manage eyes with recalcitrant macular edema secondary to CRVO who have had a suboptimal response to anti-VEGF agents.
## Other treatment options for central retinal vein occlusion
## Laser-induced chorioretinal venous anastomosis
Moderate efficacy was demonstrated in a small, randomized RCT comparing laser-induced chorioretinal venous anastomosis to sham treatment in adults with nonischemic CRVO macular edema. [bib_ref] The central retinal vein bypass study: A trial of laser-induced chorioretinal venous..., Mcallister [/bib_ref] Efficacy was highest in the 76% of patients who developed a functional anastomosis. However, 9% of patients required vitrectomy for macular traction or nonresolving vitreous hemorrhage, and 18% in the treatment arm developed neovascularization at the treatment site. [bib_ref] The central retinal vein bypass study: A trial of laser-induced chorioretinal venous..., Mcallister [/bib_ref] Currently, this type of laser procedure is not commonly used because there are more effective pharmacologic agents available.
## Surgically induced chorioretinal venous anastomosis
A non-RCT comparing surgically induced chorioretinal venous anastomosis to controls for treatment of patients with ischemic CRVO demonstrated improvement of VA in 80% of patients undergoing surgery compared to 28% of the controls after 8 months. [bib_ref] Surgical induction of chorioretinal venous anastomosis in ischaemic central retinal vein occlusion:..., Mirshahi [/bib_ref] The mean improvement of VA at 8 months was 1.5 logMAR units more in the surgical group than the controls, and all the patients had at least one successful shunt formation. Despite these positive results, three of the ten surgical patients developed significant side effects, including cataract formation, retinal detachment, and vitreous cavity hemorrhage. [bib_ref] Surgical induction of chorioretinal venous anastomosis in ischaemic central retinal vein occlusion:..., Mirshahi [/bib_ref] Radial optic neurotomy RON involves a radial incision through the cribriform plate, scleral rim, and adjacent sclera to relax the scleral outlet and was first proposed by Opremcak et al. [bib_ref] Radial optic neurotomy for central retinal vein occlusion: A retrospective pilot study..., Opremcak [/bib_ref] Its efficacy has been demonstrated in multiple studies.
A recent meta-analysis evaluated 200 patients from five studies and found there was no improvement in VA at 6 months for those receiving RON compared to controls, but those receiving RON had significantly improved VA at 12 months compared to controls. Those who had received RON had a pooled risk ratio of 2.27 for improvement of ≥3 lines of logMAR scale after 12 months when compared to controls. [bib_ref] Radial optic neurotomy in treating central retinal vein occlusion: A Meta-analysis, Chen [/bib_ref] RON also demonstrated the ability to decrease the risk of neovascular glaucoma and had similar rates of retinal detachment and vitreous hemorrhage compared to control groups. [bib_ref] Radial optic neurotomy in treating central retinal vein occlusion: A Meta-analysis, Chen [/bib_ref] Corticosteroids Intravitreal corticosteroids are a treatment option for CRVO. The "Standard Care versus Corticosteroid for Retinal Vein Occlusion" trial was a randomized clinical trial which treated individuals with nonischemic CRVO macular edema with 1 mg of 4 mg of intravitreal triamcinolone injections every 4 months compared to observation alone. At 12 months, those eyes treated with either dose of triamcinolone were five times more likely to have a 15-letter gain in visual acuity compared to the control group.About 26.5% of participants who received 1 mg triamcinolone injections achieved a 15-letter improvement in best-corrected visual acuity (BCVA) at 12 months, while 25.6% of participants in the 4 mg injection group achieved this improvement. Only 6.8% of those in the observation group had a 15-letter improvement in BCVA at 12 months.However, the participants who received triamcinolone injections were also more likely to require IOP-lowering treatment and have new development or progression of cataract.In the 1 mg injection group, [bib_ref] VEGF trap-eye for macular oedema secondary to central retinal vein occlusion: 6-month..., Holz [/bib_ref] The GENEVA trial compared a single dexamethasone implant injection at a dose of 0.35 mg or 0.7 mg to a sham implant for participants with nonischemic BRVO or CRVO macular edema. Participants receiving the dexamethasone implant had increased visual acuity at 30 and 60 days, but not at 90 or 180 days. [bib_ref] Randomized, sham-controlled trial of dexamethasone intravitreal implant in patients with macular edema..., Haller [/bib_ref] At day 180, 41% of participants in the 0.7 mg implant group achieved 15-letter improvement in BCVA compared to 40% and 23% in the 0.35 mg implant and sham implant groups, respectively. [bib_ref] Randomized, sham-controlled trial of dexamethasone intravitreal implant in patients with macular edema..., Haller [/bib_ref] There was also a higher incidence of ocular hypertension (4% vs. 0.7%, P < 0.002), requirement for IOP-lowering medication (24% in the steroid group), and procedural treatment for elevated IOP among the patients who received dexamethasone implants. These participants also had increased anterior chamber activity compared to the sham group. [bib_ref] Randomized, sham-controlled trial of dexamethasone intravitreal implant in patients with macular edema..., Haller [/bib_ref] In an additional continuation study for the dexamethasone implant, patients who received a 0.7 mg dexamethasone implant had increased visual acuity at 60 days but no evidence of sustained visual acuity gain at 12 months. [bib_ref] Dexamethasone intravitreal implant in patients with macular edema related to branch or..., Haller [/bib_ref] Participants who received the dexamethasone implant also had increased progression of cataracts (29.8% in the retreated 0.7 mg group), increased incidence of elevated IOP (32.8% of patients in the 0.7 mg group), and more individuals requiring IOP-lowering medications and procedures than those who received a sham implant. [bib_ref] Dexamethasone intravitreal implant in patients with macular edema related to branch or..., Haller [/bib_ref] A recent retrospective study has also demonstrated that dexamethasone intravitreal implants are not only effective as a first-line agent for CRVO-related macular edema but also for refractory cases as well. [bib_ref] Efficacy and safety of dexamethasone intravitreal implant for treatement of refractory macular..., Sheu [/bib_ref] Corticosteroid treatments have demonstrated the ability to improve visual acuity in the short-term and have proven, especially useful in pseudophakic patients or patient who do not experience significant IOP elevation with local steroid use. They are a useful option for the treatment of macular edema due to retinal vein occlusion.
# Conclusion
Intravitreal VEGF blockade is a standard treatment for macular edema due to CRVO. Intravitreal aflibercept is an effective and safe treatment option. Level 1 medical evidence has demonstrated that aflibercept can improve vision and reduce retinal thickening in eyes with macular edema and CRVO.
## Financial support and sponsorship
Dr. Do is a consultant and receives research funding from Allergan, Genentech, and Regeneron.
Dr. Nguyen serves on the Scientific Advisory Board for AbbVie, Genentech, Regeneron, and Santen.
[fig] Figure 2: Ocular coherence tomography changes with antivascular endothelial growth factor treatment. (a) Central retinal thickness before antivascular endothelial growth factor treatment. (b) Central retinal thickness after treatment with aflibercept demonstrating significantly less macular edema [/fig]
[fig] Figure 1: Retinal imaging of central retinal vein occlusion. (a) The ultra-wide fundus photograph shows numerous intraretinal hemorrhages and vascular tortuosity consistent with central retinal vein occlusion. (b) The ocular coherence tomography scan shows intraretinal edema with thickening in the central macula. (c-e) The fluorescein angiogram progresses with time from left to right. The fluorescein angiogram shows leakage to the macula, perivascular leakage in the periphery, and shunt vessels due to old central retinal vein occlusion [/fig]
[table] Table 1: Comparison of antivascular endothelial growth factor agents in the treatment of central retinal vein occlusion a [/table]
|
Traumatic chorioretinitis sclopetaria: Risk factors, management, and prognosis
A B S T R A C TPurpose: To describe new cases of sclopetaria and evaluate the risk factors, management, and visual prognosis of all reported cases in the literature. Observations: We performed a retrospective, observational case series. This study included six cases (median age 23, interquartile range 33) of sclopetaria. Additionally, literature searches were conducted in the PubMed and Cochrane Library databases to uncover risk factors associated with all published cases of sclopetaria. Main outcome measure was best corrected visual acuity (BCVA) worse than 20/20. Sixty-seven cases (71 eyes) of sclopetaria have been reported, of which 59 cases (61 eyes) met inclusion criteria in this study. Most were young (median age 19.5 years) men (51/59, 88.1%). Thirty-seven eyes were observed while 24 underwent immediate surgery including six pars plana vitrectomies and three scleral buckles. Compared to initial presentation, BCVA improved in 31/48 (64.6%) eyes, remained stable in 12/48 eyes (25.0%), and worsened in 5/48 eyes (10.4%). Ten patients (16.4%) achieved a final BCVA of 20/20 with median follow up time of seven months. In a multivariate model, location of sclopetaria in the macula, temporal retina, or immediate orbital foreign body removal predicted poor final BCVA with an area under receiver operating characteristic curve of 0.767. Conclusions and importance: Traumatic chorioretinitis sclopetaria is rare, but reports have increased dramatically over the past two decades. While pars plana vitrectomy may be required for the management of retinal detachments and non-clearing vitreous hemorrhage, close observation is appropriate in most cases. Visual prognosis is poor with most patients attaining 20/200 vision or worse.
# Introduction
Sclopetaria chorioretinitis is the rupture of the choroid and overlying neurosensory retina secondary to a high velocity-projectile adjacent to but not penetrating the globe. The term is thought to originate either from the term sclopetum linked to the name of an old Italian handgun or from the term sclow meaning scratch, pull, or tear. [bib_ref] Retained periorbital and intracranial air-gun pellets causing sclopetaria and visual loss, Al-Amry [/bib_ref] Sclopetaria is a rare condition, yet, an increasing number of cases have been reported in the literature over the last several decadesfrom 11 cases reported between 1980 and 1999 [bib_ref] Chorioretinitis sclopetaria caused by fishing line sinker, Katsumata [/bib_ref] [bib_ref] BB-gun injuries to the eye, Brown [/bib_ref] [bib_ref] Treatment and pathogenesis of traumatic chorioretinal rupture (sclopetaria), Martin [/bib_ref] [bib_ref] Clinicopathologic correlation of chorioretinitis sclopetaria, Dubovy [/bib_ref] to 43 reported between 2000 and 2018. 1, [bib_ref] Chorioretinitis sclopetaria, Beatty [/bib_ref] [bib_ref] Chorioretinitis sclopetaria from BB ex memoria, Otto [/bib_ref] [bib_ref] Radial choroidal ruptures in sclopetaria, Maguluri [/bib_ref] [bib_ref] Retained intraorbital metallic foreign bodies, Ho [/bib_ref] [bib_ref] Concomitant optic nerve transection and chorioretinitis sclopetaria, Mohammadpour [/bib_ref] [bib_ref] Dual cause of blindness: chorioretinitis sclopetaria and homonymous hemianopsia, Perez-Carro [/bib_ref] [bib_ref] Ocular findings following trauma from paintball sports, Taban [/bib_ref] [bib_ref] Management of traumatic macular holes: case report, Brasil [/bib_ref] [bib_ref] Surgical management of sclopetaria associated with macular hole in a young patient:..., Grosso [/bib_ref] [bib_ref] Clinical presentation and outcome of chorioretinitis sclopetaria: a case series study, Ahmadabadi [/bib_ref] [bib_ref] Evolution of retinitis sclopetaria after blunt trauma, Georgalas [/bib_ref] [bib_ref] Bilateral macular hole formation secondary to sclopetaria caused by shockwaves transmitted by..., Kunjukunju [/bib_ref] [bib_ref] Traumatic chorioretinal rupture (sclopetaria), Papakostas [/bib_ref] [bib_ref] Visual outcomes after blunt ocular trauma, Blanch [/bib_ref] [bib_ref] Retinal detachment associated with traumatic chorioretinal rupture, Papakostas [/bib_ref] [bib_ref] Clinical presentation of chorioretinitis sclopetaria, Fraser [/bib_ref] [bib_ref] Spectral-domain optical coherence tomography features of bilateral chorioretinitis sclopetaria, Rayess [/bib_ref] [bib_ref] Traumatic transection of the lateral rectus muscle with chorioretinitis sclopetaria, Mackenzie [/bib_ref] [bib_ref] Intraorbital foreign body: a rifle bullet removed 20 years after the accident, Claros [/bib_ref] The sequelae of sclopetaria were first described in 1872 by Herman Cohn, a German ophthalmologist.He described a 23-year-old male gunshot victim who, following enucleation, was found to have fusion of the retina and choroid in the posterior pole, so he termed the condition "chorioretinitis." Subsequently, in 1901, a Hungarian ophthalmologist, Wilhelm Goldzieher, first used the term "chorioretinitis plastica sclopetaria" to describe the fundus findings following a periorbital bullet wound. [bib_ref] Beitrag zur pathologie der orbitalen Schussverletzungen, Goldzieher [/bib_ref] He noted that the force of the bullet likely caused rupture of the choroid, hemorrhage and ultimately a whitish reactive fibroglial proliferation. The histopathologic findings would later be validated by Dubovy et al. on postmortem examination of the eye of a gentleman shot in 1974. [bib_ref] Clinicopathologic correlation of chorioretinitis sclopetaria, Dubovy [/bib_ref] On histopathology, the macula showed defects in Bruch's membrane and the choroid, extensive photoreceptor loss, and hyperplasia of the retinal pigmented epithelium (RPE). The retina and choroid were replaced with dense and loose fibrous tissue respectively. Most recently, photoreceptor degeneration and decreased cellularity in the retinal ganglion cell layer have been demonstrated in mouse models following blast traumatic brain injury. [bib_ref] Acute vitreoretinal trauma and inflammation after traumatic brain injury in mice, Evans [/bib_ref] Given the low incidence of sclopetaria, consensus on management has not been reached as both observation and immediate surgical management have been described. [bib_ref] Retinal detachment associated with traumatic chorioretinal rupture, Papakostas [/bib_ref] In this study, we report on six new https://doi.org/10.1016/j.ajoc.2019.02.004 Received 10 August 2018; Accepted 7 February 2019 patients with sclopetaria and describe their treatment outcomes. Moreover, to better understand the risk factors, management, and prognosis, we performed a comprehensive review of all reported cases of sclopetaria.
# Methods
An observational, retrospective case series was assembled from a single institution (Stanford University). Clinical exams were performed by vitreoretinal specialists (DD, DM, VM). In addition to the case series, the PubMed and Cochrane Library databases, most recently in August 2018 were filtered using the search terms sclopetaria, traumatic chorioretinal rupture, chorioretinitis proliferans, retinitis proliferans, and traumatic proliferative chorioretinitis of LaGrange. This search was limited to human studies. The search revealed 77 articles and was narrowed to 33 publications on the topic of sclopetaria. Studies were then excluded that noted sclopetaria but did not distinguish management by individual cases. [bib_ref] Retained intraorbital metallic foreign bodies, Ho [/bib_ref] Twenty-three articles (including the present study) were selected for further review based on inclusion of information regarding immediate management and final best corrected visual acuity (BCVA).
Data recorded from the publications included demographic data, injury details, location of sclopetaria, orbital comorbidities, complications, and immediate and delayed management. BCVA was also converted post-hoc to logarithm of the minimum angle of resolution (logMAR) units to allow for comparative analysis.
Data were analyzed using Statistical Analysis Software Enterprise Guide version 7.13 (SAS Institute, Cary, North Carolina). Variables were tested for normality with the Kolmogorov-Smirnov test to determine the appropriate statistical test. Bivariate testing was used to compare subgroups with poor final BCVA (defined a priori as worse than 20/20) to those with final BCVA of 20/20. Exposure variables were excluded if they were missing. Missing variables comprised less than 10% of the sample except for primary location of sclopetaria (18/ 61, or 29.5% missing), initial BCVA (12/61, or 19.7% missing) and follow up time (10/61, or 16.4% missing). There were no missing outcome data. A multivariable logistic regression model was developed to determine factors most predictive of a poor final BCVA. Predictors were included if they showed an association (p < 0.10) with poor final BCVA in crude bivariate analysis. Prediction of poor final BCVA was evaluated using the area under receiver operating characteristic curve (AUROC). Variables were considered to be significant by Bonferroni corrected α-levels of less than 0.003 (0.05/20).
## Findings
## Case 1
A 13-year-old male presented after he was shot near the left eye with a BB gun while playing with friends. His BCVA was 20/25 OD and 20/70 OS. His intraocular pressure (IOP) was 14 mmHg OD and 10 mmHg OS. Both pupils were round and reactive to light without relative afferent pupillary defect (rAPD). On the left, he had a three by three-millimeter partial thickness conjunctival laceration that was siedel negative and 4 + red blood cells in the anterior chamber. His posterior segment exam was notable for commotio retinae throughout the macula and periphery with vitreous hemorrhage inferiorly and nasally. There was a choroidal hemorrhage with overlying sclopetaria nasally [fig_ref] Figure 1: A 13-year-old male was shot near his left eye with a BB... [/fig_ref]. Computed tomography scan (CT) revealed a metallic foreign body in the posterior left ethmoid air cells with mild irregularity of the left lamina papyracea. The patient was managed medically with follow up two days after discharge then again at 1, 3, 5, 7, and 11 weeks. Visual acuity remained stable OD but worsened to counting fingers (CF) at one foot OS at 11 weeks secondary to vitreous hemorrhage and large preretinal hemorrhages inferiorly and nasally. Serial Bscan ultrasonography did not show evidence of retinal breaks or detachments.
## Case 2
A 17-year-old male was assaulted on his way home from school by an unknown male with a BB gun. He presented to the emergency room with BCVA of hand motion OD and 20/20 OS. He had a right rAPD and his visual fields were limited inferiorly. There was an entry wound of the upper eyelid. He had a three-millimeter superonasal conjunctival laceration, microhyphema, and round, non-reactive iris. Posterior segment exam was limited by dense vitreous hemorrhage, but B-scan ultrasonography showed no evidence of tears or detachments [fig_ref] Figure 2: A 17-year-old male was assaulted on his way home from school with... [/fig_ref]. CT showed the BB adjacent to the optic nerve. He underwent urgent orbital exploration with removal of the BB. One week later, the patient was found to have berlin's edema and a fibroglial scar with sharp serrated borders consistent with sclopetaria superiorly. An adjacent localized rhegmatogenous retinal detachment (RRD) was also seen. Optical coherence tomography (OCT) showed rupture through the choroid, RPE and outer retina. Two weeks after presentation, he underwent 25-gauge pars plana vitrectomy (PPV) for RRD repair with endolaser, perfluorocarbon, and gas tamponade. At five months, he was count fingers at two feet in the right eye.
## Case 3
A 63-year-old man suffered a rifle injury to the left eye when the weapon backfired. Presenting BCVA was 20/20 OD and CF at four feet OS. His IOP was 21 mmHg OD and 23 mmHg OS. He had a left rAPD. Examination of the left eye demonstrated a conjunctival laceration without scleral involvement, a corneal epithelial defect, microhyphema and dense vitreous hemorrhage. Plain film showed no metal foreign body and CT showed no orbital or canal fractures. One month later, the patient's visual acuity returned to 20/20 OU. The rAPD resolved and there was clearing of the vitreous hemorrhage. Wide angle photography revealed an area of bare sclera and of subretinal fibrosis consistent with sclopetaria [fig_ref] Figure 3: A 63-year-old male subject following a rifle injury to his left eye [/fig_ref]. Fluorescein angiography showed no leakage on early frames and staining of the fibrogliotic lesions on late frames.
## Case 4
A 33-year-old man presented following a gunshot wound at close range to his right supraorbital region. CT showed fractures of the right frontal bone, medial orbital wall, and ethmoid sinuses with bullet fragments in the extraconal space and bilateral frontal and ethmoid sinuses [fig_ref] Figure 4: A 33-year-old male presented following a gunshot wound to his right supraorbital... [/fig_ref]. BCVA was 20/400 OD and 20/20 OS. Intraocular pressure was 53 mmHg OD and 19 mmHg OS, necessitating lateral canthotomy and cantholysis reducing the IOP to 15 mmHg OD. A rAPD was present on the right. His extraocular movements were limited in all directions OD with complete ptosis. A subcentimeter entry wound was noted in the right medial brow. The posterior segment exam showed inferior subretinal hemorrhages, a raised perifoveal lesion, an infratemporal crescent of commotio retinae, and superior hemorrhages. OCT four days following the accident showed a full-thickness rupture through the retina, RPE, Bruch's, and choroid consistent with sclopetaria. His BCVA at that time was stable. He underwent conservative management and did not return to clinic for follow-up appointments at one or three months.
## Case 5
A 23-year-old male presented after a self-inflicted shotgun wound to the face. CT scan showed bilateral LeFort III fractures with metallic fragments adjacent the right optic nerve. He had no light perception (NLP) vision in both eyes and IOPs were 10 mmHg on the right and unmeasurable on the left. He had a complex ruptured globe of the left eye with complete extrusion of globe contents ultimately requiring enucleation. In the right eye he had dense vitreous hemorrhage, temporal commotio retinae, and elevation of the retina superiorly. There was an extensive area of sclopetaria with subhyaloid and intra-retinal hemorrhages in the posterior pole. The retina was attached, but B-scan ultrasonography showed extensive retinal and choroid thickening. He was followed closely while inpatient. Following discharge from the psychiatry unit, four months after the initial incident, he was seen in Retina clinic and found to have a white macula with areas of hemorrhage and macerated tissue with large areas of atrophy and scleral exposure in both the macula and periphery. B-scan ultrasonography showed a tractional elevation of the retina with extensive necrotic appearing tissue. The decision was made to pursue ophthalmic comfort care and globe salvation.
## Case 6
A 69-year-old male presented to the Retina service for new floaters in the setting of a chorioretinal lesion and prior herpes zoster ophthalmicus. He reported a history of BB gun injury to the right eye in childhood without any subsequent surgeries. His BCVA was 20/25 OD and 20/20 OS. His right cornea had a central opacity and with an otherwise normal anterior segment exam. On fundus examination he had pigmentation and chorioretinal atrophy of the nasal quadrant in an area of previous sclopetaria [fig_ref] Figure 5: A 69-year-old male reported a history of BB gun injury to the... [/fig_ref]. OCT showed a normal foveal contour without cystoid macular edema or subretinal fluid in either eye. No intervention was required. Optical coherence tomography showed disruption and irregularity of the outer macula (Bottom left). Optos ultra wide-field imaging at presentation (Top right), five weeks (Middle right), and eleven weeks (Bottom right) showed nasal choroidal hemorrhage with underlying sclopetaria, commotio retinae throughout the macula and periphery, and a large vitreous hemorrhage inferiorly and nasally that subsequently evolved into dense vitreous hemorrhage. Optos ultra wide-field imaging showed macula-involving commotio retinae, a fibroglial scar with sharp serrated borders superiorly, and an adjacent localized retinal detachment (Top middle). Fundus autofluorescence showed hypoautofluorescence within the area of choroidal rupture (Top right). Optical coherence tomography showed a rupture through the choroid, retinal pigmented epithelium and outer retina consistent with sclopetaria (Bottom right).
## Review of the literature
To date, 71 eyes with sclopetaria in 67 patients (including six in the present study) have been described in the literature. Ten eyes were excluded from this review as there was no description of the management and/or final reported BCVA. Of the 59 patients (61 eyes) meeting inclusion criteria in this review, 52 (88.1%) were male [fig_ref] Table 1: Baseline characteristics of all patients in the literature with sclopetaria [/fig_ref]. The median age at presentation was 19.5-years-old (interquartile range (IQR), 12 years). Where sidedness was reported, 33 were right eyes (60.0%). Patients diagnosed with sclopetaria were most likely to have been injured via indirect trauma to the globe with a BB Location of sclopetaria was significantly related to final BCVA with a higher risk of poor outcomes associated with sclopetaria located in the macula (RR, 1.32; 95% CI, 1.10-1.59), temporal retina (RR, 1.32; 95% CI, 1.10-1.59), and superior retina (RR, 1.30; 95% CI, 1.09-1.54) [fig_ref] Table 2: Primary location of sclopetaria in 43 eyes with sclopetaria [/fig_ref].
The most common comorbidity was intraorbital foreign bodies in 40/61 (65.6%) cases [fig_ref] Table 3: Comorbidities and their relationship to final visual acuity [/fig_ref]. Vitreous hemorrhage occurred in 38/61 (62.3%) cases and typically occurred at the time of injury, though was a delayed complication in three cases -reported at three days, six weeks, and four months. Optic neuropathy (17/61, 27.9%) and hyphemas or microhyphemas (17/61, 27.9%) were also common. Of the seven with subsequent RDs, one was diagnosed at the time of initial injury with the remaining six occurring at one week, one week, two weeks, three weeks, one year, and one and one-half years. Retrobulbar hematomas (RR, 1.20; 95% CI 1.07-1.34), eyelid lacerations (RR, 1.24; 95% CI 1.08-1.41), ptosis (RR, 1.20; 95% CI 1.07-1.34), macular holes (RR, 1.23; 95% CI 1.08-1.39), maculopathy (RR, 1.23; 95% CI 1.08-1.39), and optic nerve avulsion (RR, 1.21; 95% CI 1.08-1.37) were all associated with a higher risk of poor BCVA, while, RD was not (RR, 1.04; 95% CI 0.75-1.44).
Regarding management, 36 eyes were observed, seven underwent globe exploration, six urgent PPV, six urgent foreign body removal, three scleral buckle, two enucleation, one prophylactic scleral buckle, one macular hole repair, one lensectomy, one muscle reattachment surgery, one fracture repair, and one conjunctival closure [fig_ref] Table 4: Immediate and delayed management of complications [/fig_ref]. The six patients who underwent urgent PPV had similar outcomes to those observed, with poor final BCVA in five patients (83.3%). Of those undergoing urgent PPV, five had improvement in final BCVA (83.3%).
## Prediction model
Backward selection was performed on the following variables as possible risk factors for poor BCVA: age, projectile (air-gun, paintball, cork, foam bullet, tree branch), location of sclopetaria (macula, temporal retina, superior retina), comorbidities (retrobulbar hematoma, eyelid laceration, ptosis, cataract, hyphema or microhyphema, macular hole, maculopathy, optic nerve avulsion) and urgent management (foreign body removal, muscle reattachment surgery, fracture repair, lateral canthotomy, lensectomy). Location of sclopetaria in the macula, location in the temporal retina, and urgent foreign body removal were retained in the model. Area under receiver operating characteristic curve was 0.767, and 10-fold cross validation revealed a true AUROC of 0.737. Therefore, the over-optimism was 0.030, indicating a low degree of overfitting.
# Discussion
In 1996, Kuhn et al. presented a standardized classification for ocular trauma with the major delineation between closed globe and open globe injuries. [bib_ref] A standardized classification of ocular trauma, Kuhn [/bib_ref] Under the heading of closed globe injuries posterior manifestations include commotio retinae, choroidal rupture, sclopetaria, macular hole, and retinal detachments. High velocity projectiles can lead to coup injuries at the site of impact (e.g. sclopetaria), while pressure waves opposite the site of impact can lead to countercoup injuries (e.g. commotio retinae, choroidal rupture, and traumatic macular holes). Direct ocular compression, on the other hand, leads to equatorial stretching which can result in vitreous base avulsion and retinal dialysis. [bib_ref] Coup-contrecoup mechanism of ocular injuries*, Wolter [/bib_ref] Sclopetaria is distinguished from commotio retinae and choroidal rupture in that there is more extensive damage including rupture of the choroid, Bruch's membrane, RPE and neurosensory retina adjacent to the projectile. [bib_ref] Neuroretinal cell death in a murine model of closed globe injury: pathological..., Blanch [/bib_ref] On examination, bare sclera can often be visualized, though this will typically evolve into a pigmented scar [fig_ref] Figure 3: A 63-year-old male subject following a rifle injury to his left eye [/fig_ref]. The clinical and pathologic findings following ocular trauma are a result of differences in elasticity of the sclera, Bruch's membrane, RPE, and retina.Bruch's membrane is inelastic and ruptures easily with compressive forces along with its adherent choriocapillaris leading to the acute subretinal hemorrhages often seen in ocular trauma. Similarly, the RPE is relatively inelastic, making rupture more likely and leading to the late pigmentary changes often seen with both choroidal rupture and sclopetaria. Conversely, both the retina and sclera are elastic, requiring high levels of impact energy to disrupt their architecture. Therefore, only a high velocity projectile, such as a bullet, passing adjacent to the globe, could create the shock wave forces that could produce retraction of both the retina and choroid, leaving bare sclera at the site of a break. [bib_ref] Acute vitreoretinal trauma and inflammation after traumatic brain injury in mice, Evans [/bib_ref] Regarding long term complications, commotio retinae can be observed as retinal edema almost universally resolves in a few days. [bib_ref] Retinal opacification after blunt non-perforating concussional injuries to the globe. A clinical..., Hart [/bib_ref] Eyes with choroidal rupture must be observed more closely for the [bib_ref] Retinal detachment associated with traumatic chorioretinal rupture, Papakostas [/bib_ref] It should be noted that these three patients had preceding surgery (globe rupture repair, nail removal, and BB removal from the orbit) prior to retinal detachment. Based on the results of this review, the incidence of retinal detachment following sclopetaria was 11.5% (7/61) though this finding was not associated with a worse visual outcome (p = 0.83). The visual prognosis following sclopetaria is dependent on multiple features. When considering projectile type, those with the lower muzzle energy (e.g. air-gun pellet and paintball) were more likely to result in poor visual outcome than those with high muzzle energy (e.g. bullet and BB) [fig_ref] Table 1: Baseline characteristics of all patients in the literature with sclopetaria [/fig_ref]. [bib_ref] Gunshot wounds, Rhee [/bib_ref] This was a surprising finding as increased muzzle energy is typically associated with increased damage. Therefore, we propose that the visual prognosis is more likely related to the location of sclopetaria and the associated ocular comorbidities than the projectile type. As anticipated, those with temporal and macular sclopetaria were less likely to achieve 20/20 vision [fig_ref] Table 2: Primary location of sclopetaria in 43 eyes with sclopetaria [/fig_ref]. Additionally, the concurrence of a macular hole or maculopathy were risk factors for incomplete visual recovery [fig_ref] Table 3: Comorbidities and their relationship to final visual acuity [/fig_ref]. Those with eyelid lacerations were at higher risk for poor visual prognosis, possibly because these tended to occur in the setting of more extensive orbital damage. The presence of an intraorbital foreign body, vitreous hemorrhage and retinal detachments were not independent risk factors for suboptimal visual recovery. As expected, those with simultaneous optic nerve avulsion had poor visual outcomes.
The most controversial component of sclopetaria is the immediate management. As this review has demonstrated, there is no statistically significant benefit of immediate surgery as compared to observation alone (p = 0.140). However, this finding needs to be taken in context as it is unknown what the visual outcome would have been if surgery was delayed. In cases where the clinical suspicion was high for a ruptured globe, globe exploration did not ultimately worsen the chance for visual recovery (p = 0.865). Additionally, PPV (6 cases) and/or prophylactic scleral buckle (3 cases) were not statistically beneficial in protecting vision, though the sample size was likely too small to detect a difference. Of the nine patients undergoing delayed surgery, only two ultimately achieved a visual acuity of 20/20. Patients undergoing immediate orbital foreign body removal were statistically less likely to achieve 20/20 vision (p = 0.002). The standard of care following inorganic intraorbital foreign body removal includes observation unless they are causing complications or anteriorly located allowing simple removal as loss of vision is almost always associated with the initial trauma. [bib_ref] Clinical features and management of intraorbital foreign bodies, Fulcher [/bib_ref] Immediate surgical intervention (especially orbital foreign body removal) is likely not beneficial for ultimate visual recovery, though this decision needs to be tailored to each individual scenario.
# Conclusions
Ultimately, sclopetaria is a rare sequelae of a high velocity projectile passing tangentially to the eye. The demographic associated with sclopetaria is the same as for trauma elsewhere to the bodyyoung males. Sclopetaria most commonly occurs secondary to BBs, likely due to both the common use of BB guns among this population and the impact velocity enough to rupture the retina, but low not to harm the sclera. [bib_ref] Paintball trauma and mechanisms of optic nerve injury: rotational avulsion and rebound..., Sponsel [/bib_ref] The visual prognosis is most dependent on the status of the macula and the location of the sclopetaria (with superior and temporal disease having worse outcomes). The immediate and delayed management is equivocal and a case by case approach is likely the safest approach. Fortunately, most patients experience improvements in final BCVA though only 16.4% achieve 20/20 vision.
## Patient consent
This report does not contain any information that could lead to the identification of the patients.
[fig] Figure 1: A 13-year-old male was shot near his left eye with a BB gun. Computed tomography scan on presentation showed a metallic BB in the posterior left ethmoid air cells with mild irregularity of the left lamina papyracea (Top left, Middle left). [/fig]
[fig] Figure 2: A 17-year-old male was assaulted on his way home from school with a BB gun. Computed tomography scan demonstrated the BB located within the orbit at the posterior margin of the globe (Top left, Middle left). B-scan ultrasonography shows layered vitreous hemorrhage in posterior pole and hyperechogenicity of the BB (Bottom left). [/fig]
[fig] Figure 3: A 63-year-old male subject following a rifle injury to his left eye. His left eye composite fundus photo at one-month follow-up showed macular and peripheral subretinal fibroglial scars (Left). A nasal fibroglial scar was adjacent to bare sclera consistent with sclopetaria. Additionally, retinal pigmented epithelium hypertrophy was seen throughout the periphery and adjacent to the inferior vascular arcade. Fluorescein angiography (FA) nasally showed a window defect at the area of chorioretinal rupture (Top right), early FA of the macula showed absence of leakage while staining of the fibrogliotic lesion was seen on late frames (Bottom right). [/fig]
[fig] Figure 4: A 33-year-old male presented following a gunshot wound to his right supraorbital region resulting in complete ptosis. The patient's plain film and computed tomography scans demonstrated multiple fractures and bullet fragments in the right superior and inferior extraconal spaces and bilateral frontal and ethmoid sinuses (Top left and right). His Optos ultra wide-field fundus photo showed superior dot blot hemorrhages, inferior subretinal hemorrhages, infratemporal commotio, and a raised perifoveal area (Middle left), while his fundus autofluorescence showed hypoautofluorescence extending beyond the location of the hemorrhages (Middle right). Optical coherence tomography showed a full-thickness rupture through the retina, Bruch's membrane and choroid consistent with sclopetaria (Bottom). [/fig]
[fig] Figure 5: A 69-year-old male reported a history of BB gun injury to the right eye in childhood without any subsequent surgeries. Optical coherence tomography scan showed normal foveal contour and normal retinal layers (Top left, Top right). Optos ultra wide-field imaging showed severe nasal pigmentation and chorioretinal atrophy of the nasal quadrant consistent with sclopetaria (Bottom left). Fundus autofluorescence showed hypoautofluorescence nasally consistent with loss of retinal pigmented epithelium (Bottom right). (23/60, 38.3%) or bullet (16/60, 26.7%). However, direct trauma to the globe with subsequent sclopetaria was also reported with paintballs, a tree branch, foam bullet, and cork. Initial BCVA was NLP to 20/1000 in 35/49 (71.4%), 20/800 to 20/ 200 in 7/49 (14.3%), 20/100 to 20/25 in 5/49 (10.2%), and 20/20 in 2/49 (4.1%). Final BCVA was NLP to 20/1000 in 30/61 (49.2%) patients, 20/800 to 20/200 in 12/61 (19.7%), 20/100 to 20/25 in 9/61 (14.8%) and 20/20 in 10/61 (16.4%). Where both initial and final BCVA were recorded, BCVA improved in 31/49 (63.3%) eyes, remained stable in 13/48 eyes (26.5%), and worsened in 5/48 eyes (10.2%). Median initial BCVA was 2.08 logMAR units (IQR, 2.00; 20/2400 in Snellen equivalent; 12 missing). Median final BCVA was 1.70 logMAR units (IQR, 1.57; 20/1000 in Snellen equivalent) with median follow up time of 7 months (IQR, 23.5 months). Follow up time was not significantly different between those with poor final BCVA (> 20/20) compared to those with final BCVA of 20/20 (p = 0.334). [/fig]
[table] Table 1: Baseline characteristics of all patients in the literature with sclopetaria. [/table]
[table] Table 2: Primary location of sclopetaria in 43 eyes with sclopetaria. [/table]
[table] Table 3: Comorbidities and their relationship to final visual acuity. [/table]
[table] Table 4: Immediate and delayed management of complications. BCVA = best corrected visual acuity. a Variables considered to be significant by Bonferroni corrected a-levels of 0.003 (0.05/20). b Defined as less surgery performed less than one week after the injury.development of secondary choroidal neovascularization membranes.34 Patients with sclopetaria are at higher risk for both delayed vitreous hemorrhage or progression of vitreous hemorrhage (as in Case 1) and retinal detachment, which has given credence to early surgical intervention.21 Prior to 2014, it was felt that retinal detachment was unlikely to occur in the setting of sclopetaria due to adhesions between the retina and choroid causing them to retract as a single unit. Additionally, as most patients with sclopetaria are young, they have formed vitreous and intact posterior hyaloid which tamponades the retinal break. However, in 2014, Papakostas et al. published three cases of retinal detachment in the setting of sclopetaria. [/table]
|
Comparing the Visual Analogue Scale and the Pediatric Quality of Life Inventory for Measuring Health-Related Quality of Life in Children with Oral Clefts
Objectives: To evaluate the performance of the Visual Analogue Scale (VAS), in measuring overall health-related Quality of Life (HRQoL) in children with oral clefts relative to the Pediatric Quality of Life Inventory 4.0 (PedsQL TM ) Generic Core Scales, one of the most validated and commonly used methods to measure pediatric HRQoL.Methods:The study included a population-based sample of 307 children aged 5 to 10 years who were born in Iowa, New York, and Arkansas with non-syndromic oral clefts. Data on HRQoL were obtained using a VAS and PedsQL TM via self-administered interviews with the parents. We evaluated the correlations between the VAS and PedsQL TM total scores, and the correlations of each of these two scales with a series of child health and wellbeing indicators. Results: The VAS and PedsQL TM scores were well-correlated (r = 0.67). ThereOPEN ACCESSConclusions: Our study finds the VAS to perform relatively well in measuring overall HRQoL among children with oral clefts. The VAS may be useful as a screening tool to identify children with oral clefts at risk of low HRQoL for referral into more comprehensive evaluations and for measuring average HRQoL across a sample of children.
were no prominent differences between the correlations of VAS and PedsQL TM with the selected indicators of child health and wellbeing; differences in correlations were less than 0.1. Differences in HRQoL by cleft type were more pronounced on the PedsQL TM .
# Introduction
Oral clefts are one of the most common birth defects worldwide, and have lifelong implications for the wellbeing and health-related quality of life (HRQoL) of affected children and their families [bib_ref] Health professionals' assessment of health-related quality of life values for oral clefting..., Wehby [/bib_ref] [bib_ref] The impact of orofacial clefts on quality of life and healthcare use..., Wehby [/bib_ref] [bib_ref] Oral healthrelated quality of life in children with orofacial clefts, Ward [/bib_ref]. Among the many potential consequences early in life and during childhood are increased risks for fetal growth retardation [bib_ref] Fetal health shocks and early inequalities in health capital accumulation, Wehby [/bib_ref] , hospitalizations [bib_ref] The effects of oral clefts on hospital use throughout the lifespan, Wehby [/bib_ref] and certain behavioral and psychosocial problems such as inattention/hyperactivity and separation anxiety disorder [bib_ref] Oral clefts and behavioral health of young children, Wehby [/bib_ref] [bib_ref] Separation anxiety in children ages 4 through 9 with oral clefts, Tyler [/bib_ref] , which may be partly due to concerns about facial appearance and speech but also other factors such as the multiple needed surgical repairs and healthcare treatments [bib_ref] Oral clefts and behavioral health of young children, Wehby [/bib_ref].
Health-related quality of life (HRQoL) has become a commonly used measure of health and well-being that represents the impact of health on quality of life and captures the desirability of health conditions relative to perfect health. In addition to being a powerful measure of health status, preference-based HRQoL measures can be used to adjust duration of life to generate quality-adjusted life years (QALYs), a standard measure of effectiveness in cost-utility analyses used for assessing the value of health care treatments and resources to society relative to cost. Therefore, measuring the HRQoL of children with oral clefts is of interest for both researchers and clinicians given the greater risk of affected children for a decline in HRQoL compared to the general population due to both physical and psychosocial effects and the increased need for interactions with healthcare professionals including pediatricians.
Several methods have been suggested for measuring HRQoL including instruments that have been developed to measure HRQoL in children. Of these, the 23-item Pediatric Quality of Life Inventory 4.0 (PedsQL TM ) Generic Core Scales is one of the most commonly utilized. The PedsQL TM is a short survey (estimated to take less than 5 min) designed to measure HRQoL for children aged 2 to 12 and adolescents 13-18 years through questions related to the physical, emotional, and social functioning and has high feasibility, reliability and validity [bib_ref] The PedsQL: Measurement model for the pediatric quality of life inventory, Varni [/bib_ref] [bib_ref] The PedsQL 4.0 as a pediatric population health measure: Feasibility, reliability, and..., Varni [/bib_ref] [bib_ref] The PedsQL (TM) 4.0 generic core scales: Sensitivity, responsiveness, and impact on..., Varni [/bib_ref] [bib_ref] The PedsQL (TM)-Reliability and validity of the short-form generic core scales and..., Chan [/bib_ref].
Visual Analogue Scale (VAS) methods that involve direct rating and state comparisons can also be used to measure HRQoL [bib_ref] Health professionals' assessment of health-related quality of life values for oral clefting..., Wehby [/bib_ref]. Unlike instruments that involve answering a series of direct questions about health and well-being, this method asks an individual to rate health status or a particular health state on a scale between two reference states/points (typically death and perfect health) that represents the desirability of that health status or state relative to these states. The simplicity and ease of administering VAS methods allow for their wide use in many different settings [bib_ref] Health professionals' assessment of health-related quality of life values for oral clefting..., Wehby [/bib_ref] [bib_ref] Methodology for measuring health-state preferences-II: Scaling methods, Froberg [/bib_ref]. The reliability, validity, and feasibility of direct rating methods have been demonstrated in the literature [bib_ref] Methodology for measuring health-state preferences-II: Scaling methods, Froberg [/bib_ref] [bib_ref] Visual analogue scales: Measurement of subjective phenomena, Gift [/bib_ref] [bib_ref] A new analogue scale for assessing children's pain: an initial validation study, Mcgrath [/bib_ref].
Given the availability of different methodologies to measure HRQoL and the importance of measuring HRQoL among children with oral clefts, the question arises as to how different methodologies such as the VAS and the PedsQL TM compare to each other in this population. The benefits of VAS include quick adaptability in both research and clinical settings and that VAS-based HRQoL scores can be easily used to obtain quality-adjusted-life-years (QALYs) for cost-effectiveness analysis. There are also limitations with the VAS, including potential bias in measurement due to raters avoiding the ends of the scale or with measuring multiple co-existing health conditions [bib_ref] Visual analog scales: Do they have a role in the measurement of..., Torrance [/bib_ref]. Another limitation is that the VAS is generally designed to obtain a single HRQoL score for overall health status without additional detail on the various domains of health. While this disadvantage may be arguably overcome by constructing VAS for each of the main domains of health including, physical, social, and mental health, this may take away from the practical advantage of the VAS. In general, VAS may be particularly appealing in settings where the primary goal is to screen for low HRQoL or to obtain overall HRQoL across a sample relatively easily and at little cost. Given that the PedsQL TM is one of the most validated and commonly used instruments for pediatric HRQoL, this study compares the performance of the VAS relative to the PedsQL TM in measuring the HRQoL of children with oral clefts and the consistency of the HRQoL scores between the two methods.
# Methods
## Data
The data from this study were obtained from a mail survey conducted with the parents of 307 children age 5 to 10 years who were born with non-syndromic oral clefts in Iowa, New York, and Arkansas. Data were obtained on HRQoL using both the PedsQL TM and VAS measures of HRQoL. The survey was conducted in 2007-2008; the children were required to be currently living in one of the three aforementioned states with a parent. A written survey using a modified Dillman method was mailed asking questions about a wide range of topics. First a survey questionnaire was sent along with a letter discussing the purpose of the study. A week later, a postcard was sent as a means to remind participants to return in their surveys if they wished to participate. If a participant did not respond within 10 days, then another survey questionnaire and letter were sent. All necessary Institutional Review Boards approved the study.
## Comparison of hrql instruments
The mothers were asked to complete the parental version of the PedsQL TM for their children. In addition, they were asked to rate their child's HRQoL on a VAS: 0 100
Worst
## Imagina ble health
Perfect Health Specifically, each mother was asked to draw a vertical line (|) at the point on the scale that she thought represented the status of her child's HRQoL. The VAS score was calculated as the distance between the left anchor of the scale (0 value or worst imaginable health) and the vertical line drawn by the mother. The scale was described as follows:
On the scale below, we ask you to rate your child's health-related quality of life on a scale of 0 to 100. A score of "0" represents the worst health state that you can imagine. A score of "100" represents perfect health. A child with perfect health would be one who has no pain or discomfort, no anxiety or depression, and no problems with usual activities that would be expected for his or her age, such as feeding him or herself, speaking, playing with other children, washing his or her hands, participating in school activities.
We first evaluated the correlations between the VAS and PedsQL TM total scores. Next, we evaluated the correlations of each of these two scales with a series of child health and wellbeing indicators in order to evaluate if any the two was more strongly correlated with these measures. For each of these indicators, we calculated the correlations for the subgroup that had complete (non-missing data) on the indicator and on both the VAS and the PedsQL TM (11 observations had missing data on one or both of these scales). Measures of the child's social and separation anxiety were obtained using the 41-item Screen for Child Anxiety Related Emotional Disorders (SCARED) [bib_ref] Psychometric properties of the screen for child anxiety related emotional disorders (SCARED):..., Birmaher [/bib_ref]. A subscale of 8 SCARED items make up the separation anxiety score which has a maximum of 16, with a score of 5 or greater indicating separation anxiety disorder. Similarly, the social anxiety variable consists of a subscale of 7 items from the SCARED items with a maximum score of 14, with a score of 8 or greater indicating social anxiety disorder. Data were also obtained on the Pediatric Behavior Scale (PBS), a 30 item survey which focuses on four broad areas of depression/anxiety, physical/somatic symptoms, aggression/opposition, and inattention/hyperactivity [bib_ref] Oral clefts and behavioral health of young children, Wehby [/bib_ref]. Other indicators of the child's health and well-being included maternal rating of the child's overall health status on a standard Likert-scale, whether the child suffered from a chronic health condition (under 25 categories such as asthma, vision, dental, hearing, and other problems), and how happy the child was with his or her facial appearance on a four-category scale, a commonly used and particularly relevant measure for this population [bib_ref] Oral clefts and behavioral health of young children, Wehby [/bib_ref]. Additionally, five variables focused on aspects of how the child's condition affected his or her ability to be understood while speaking. [fig_ref] Table 1: Variable Description and Descriptive Statistics [/fig_ref] includes the definitions of the study variables.
# Results
Out of 589 eligible children, questionnaires were received from 307, yielding a response rate of 52.1%. About 62% of the sample were males (which is expected since oral clefts are more common among males) and 92% were Caucasian. The sample was approximately evenly distributed across the ages of four to nine with a range of 42 to 58 children in each year. The rates of cleft type were overall comparable to population rates in the US including 81 children (28%) with cleft lip only (CLO), 95 children (30%) with cleft palate only (CPO), and 131 children (42%) with both cleft lip with palate (CLP). These statistics suggest no response bias over child's gender, race/ethnicity, age, and cleft type. [fig_ref] Figure 1: Scatter Plot and Fitted Line of VAS over PedsQL TM Scores [/fig_ref] shows a scatter plot of the VAS vs. PedsQL TM scores along with their ordinary least squares (OLS) regression line. The two scores had a standardized correlation coefficient (r) of 0.67: a one standard deviation increase in PedsQL TM was associated with a 0.67 standard deviation increase in VAS (and vice versa). This correlation is stronger than those previously reported between the PedsQL TM and HRQoL instruments specific to oral health including the Child Oral Health Impact Profile when used among children and adolescents with oral clefts (r = 0.52) [bib_ref] Evaluation of the similarities and differences in response patterns to the pediatric..., Broder [/bib_ref] and the Early Childhood Oral Health Impact Scale (r = 0.20) [bib_ref] Health-and oral health-related quality of life among preschool children with cerebral palsy, Du [/bib_ref]. The stronger correlation suggests that the VAS is capturing more of the generic HRQoL measured by the PedsQL TM compared to a condition-specific (i.e., oral health) instrument. In [fig_ref] Figure 2: Means of VAS [/fig_ref] , we show the means of the VAS scores across quintiles of the PedsQL TM and vice-versa. The VAS score means increased across the quintiles of the PedsQL TM but the changes became smaller in magnitude with moving to successively higher quintiles. This was also generally the case for changes in the PedsQL TM score means over the VAS quintile groups, with the exception that the PedsQL TM score mean slightly declined between the third and fourth quintiles of the VAS. This indicates that the scores of the PedsQL TM and VAS were overall more consistently related to each other at lower ranges, i.e., for children with lower HRQoL, but were less so for children with high HRQoL. [fig_ref] Table 2: Correlations of VAS and PedsQL TM with selected measures of child health... [/fig_ref] compares the correlations of measured indicators of child health and wellbeing that are thought to be relevant for HRQoL, one at a time, with each of VAS and PedsQL TM scores. The correlations ranged from 0.20 to 0.54 (in absolute values) and were significant at p < 0.001. Overall, there were no prominent and consistent differences in the correlations of VAS and PedsQL TM with these measures. The correlations were generally close with a difference between VAS and PedsQL TM of less than 0.1 in all cases. No instrument clearly dominated the other one in being more strongly correlated with a greater number of the selected child health and wellbeing indicators. The HRQoL of children with oral clefts may vary by cleft type. However, the direction and magnitude of these differences are theoretically ambiguous. For example, it is unclear based on theory whether children with CLO have better or worse HRQoL than those with CPO. Even though speech problems are typically not present among children with CLO unlike those with cleft palate, both cleft types are associated with feeding problems and dental problems and cleft lip is additionally associated with esthetic concerns (and generally more dental issues). In order to evaluate how the two HRQoL scores compare by cleft type, we regressed using OLS the VAS and the PedsQL TM scores on cleft type indicators including an indicator for CLO and another for CPO with CLP as the reference category [fig_ref] Table 3: Mean [/fig_ref]. Children with CLO had significantly higher PedsQL TM scores than those with CLP; however, the difference in VAS scores was smaller and insignificant (p = 0.18). The difference between CPO and CLP was insignificant on both instruments. The difference between the instruments for CLO could suggest that the PedsQL TM is more sensitive to identifying differences in HRQoL by cleft type. However, this difference could also be partly driven by the relatively small number of children in each cleft type and the skewed HRQoL score distributions (especially by cleft type) which could bias mean comparisons. When comparing the medians of the HRQoL scores between children with CLO and those with cleft palate (with or without cleft lip in one group, i.e., combining CPO and CLP together), the VAS indicated higher HRQoL values among children with CLO; the difference in median scores between these two groups was slightly larger with the PedsQL TM than VAS (7 vs. 5 points). Taken as a whole with the other results and considering the theoretical ambiguity about differences in HRQoL by cleft type and the relatively small number of children in each cleft type, differences between the instruments by cleft type do not necessarily suggest a weakness of the VAS in capturing overall HRQoL in this sample.
# Discussion
In this sample of children with oral clefts, the HRQoL scores from the VAS and PedsQL TM were well-correlated and overall similar in their correlation with several health indicators. Given the simplicity of the VAS, it may be an appealing choice for cleft teams and other clinical providers of healthcare for children with clefts who may be primarily interested in screening children with oral clefts to identify those at risk for low HRQoL for more comprehensive evaluations instead of assessing specific HRQoL domains in every child. VAS may also be of interest to researchers of health services and outcomes among children with oral clefts who are mainly interested in measuring average HRQoL across a sample and those who are soliciting HRQoL values to generate QALYs for cost-effectiveness analysis. In contrast, one clear advantage of PedsQL TM is in settings where clinicians or researchers are interested in decomposing total HRQoL across multiple domains to identify areas of functioning most adversely affected by the child's health.
There are other methods besides VAS to obtain HRQoL scores for QALY measurement in cost-effectiveness analysis such as the standard gamble (SG) and time trade-off (TTO). Both of these methods are rooted in economic theory, but they are not necessarily advantageous to VAS on either theoretical or empirical grounds. Among the main theoretical limitations of these methods are their sensitivity and bias to preferences for risk taking (SG) and time/future discounting (TTO). On the practical side, these methods are particularly demanding on the raters' cognitive ability and fairly burdensome (especially the SG), typically requiring an interviewer and illustrative tools to aid the raters' in their task. In contrast, the VAS can be easily self-administered after brief written instructions as done in this study.
Our study has several strengths but some limitations. One strength is that mothers completed the PedsQL TM and VAS at the same time so there is no timing bias due to changes in health status and no interviewer or data collection method bias since mothers self-administered both methods. Also, there was no language in the instructions that would alert the mothers to our objective of comparing the two methods and cause them to compare their own answers between these instruments. Furthermore, the Likert-scale answers to the PedsQL TM questions are not directly comparable to the single VAS score and the total PedsQL TM score derived from the answers is not available to the mothers. Therefore, it is unlikely that there is any bias in the correlation between the two scores due to the mode of administering the instruments. Another strength is that we measured several health indicators that we used to evaluate the sensitivity of the HRQoL scores and their ability to correlate with different aspects of health and wellbeing.
On the limitation side, having other measures of the child's health and well-being, such as measures of pain or cognitive performance, would have been useful to correlate within the quality of life measures. The correlations between the HRQoL scores and certain health/wellbeing indicators such as social anxiety and number of chronic conditions were relatively low. This is not surprising since the total scores are generic measures that capture overall HRQoL and how it is impacted by various physical and psychosocial aspects of health and wellbeing. The differences in correlations across the various health and wellbeing indicators may reflect the relative importance of these indicators for HRQoL. However, these results also highlight the value of domain-specific assessments and analyses which can be done with the PedsQL TM in cases where specific areas of health and wellbeing such as physical or emotional functioning are of interest. We chose to measure overall HRQoL with the VAS in this study and therefore only compared the VAS to the overall PedsQL TM scores instead of the domain-specific scores. However, future studies can evaluate the utility of VAS in specific domains of health and wellbeing. Also, we were unable to evaluate the test-retest and inter-rater reliability of the VAS for our study population and leave this for future research.
It is important to note that our results may not necessarily generalize to other conditions besides oral clefts such other birth defects or chronic health conditions. To the best of our knowledge, very few studies have reported the correlations between the PedsQL TM and global measures of HRQoL using VAS in other pediatric populations, so there are not many previous results with which we can compare our finding. One study reported a correlation of 0.64 between the PedsQL TM and a general VAS-based measure of wellbeing among children with chronic arthritis [bib_ref] Health of children with chronic arthritis: Relationship of different measures and the..., Brunner [/bib_ref] , which is close to the 0.67 correlation coefficient we found. Replicating this study in other pediatric populations may be of interest to further evaluate the value of VAS as a tool for HRQoL screening and assessment.
# Conclusions
Our study finds the VAS, a relatively simple technique, to perform relatively well in measuring overall HRQoL among children with oral clefts. The average VAS score of the sample was very close to the average PedsQL TM , the two scores were well-correlated (r = 0.67), and they were overall comparable in their correlation with several measures of child health and wellbeing. The VAS method may be particularly appealing to cleft teams and other health professionals providing care for children with oral clefts for screening children at risk of low HRQoL and referral into more comprehensive evaluations. This method may also be useful for researchers who are interested in measuring average HRQoL across a sample and generating HRQoL scores to obtain QALYs for cost-effectiveness analysis.
[fig] Figure 1: Scatter Plot and Fitted Line of VAS over PedsQL TM Scores. [/fig]
[fig] Figure 2: Means of VAS (PedsQL TM ) Scores by Quintiles of PedsQL TM (VAS). Notes: The incremental changes in means of scores of one HRQL scale with moving to the next quintile of the other HRQL scale are shown at the top of the bars. [/fig]
[table] Table 1: Variable Description and Descriptive Statistics.Table 1 includes descriptive statistics for the HRQoL measures and other study variables. The average score for VAS was 86.6 on scale from 0-100 (standard deviation of 15.9), while the average of PedsQL TM was 83.8 (standard deviation of 16.4). [/table]
[table] Table 2: Correlations of VAS and PedsQL TM with selected measures of child health and wellbeing. [/table]
[table] Table 3: Mean (OLS) and median regressions of VAS and PedsQL TM scores on cleft type indicators. [/table]
|
Rapid and precise diagnosis of T. marneffei pulmonary infection in a HIV-negative patient with autosomal-dominant STAT3 mutation: a case report
Background: Talaromyces marneffei, also named Penicillium marneffei, is an opportunistic pathogen that can cause systemic or limited infection in human beings. This infection is especially common in human immunodeficiency virus (HIV)-infected hosts; however, it has also been recently reported in HIV-negative hosts. Here, we report a very rarely seen case of T. marneffei pulmonary infection in a non-HIV-infected patient with signal transducer and activator of transcription 3 (STAT3) mutation. Case presentation: A 34-year-old woman was admitted to our hospital for uncontrollable nonproductive cough and dyspnea with exercise. She had been immunocompromised since infancy. Computerized tomography scan showed multiple ground glass opacities with multiple bullae in both lungs. Next generation sequencing (NGS) of the bronchoalveolar lavage fluid identified T. marneffei nucleotide sequences. Culture of bronchoscopy specimens further verified the results. The patient was HIV negative, and blood gene detection indicated STAT3 mutation. To date, following the application of itraconazole, the patient has recovered satisfactorily. Conclusion: In clinical practice, T. marneffei infection among HIV-negative individuals is relatively rare, and we found that patients who are congenitally immunocompromised due to STAT3 mutation may be potential hosts. Early diagnosis and timely treatment are expected to improve the prognosis of T. marneffei infection. NGS is a powerful technique that may play an important role in this progress.The reviews of this paper are available via the supplemental material section.
# Background
Talaromyces marneffei, formerly called Penicillium marneffei, causes mostly opportunistic infections in immunodeficiency individuals, who are particularly susceptible to T. marneffei, especially human immunodeficiency virus (HIV)-positive patients in certain endemic regions such as Southeast Asia. [bib_ref] Penicillium marneffei infection and recent advances in the epidemiology and molecular biology..., Vanittanakom [/bib_ref] In 1973, the first case of T. marneffei infection was reported in an American minister in Southeast Asia. [bib_ref] Infection caused by Penicillium marneffei: description of first natural infection in man, Disalvo [/bib_ref] The incidence rate of T. marneffei infection increased noticeably after the acquired immune deficiency syndrome (AIDS) pandemic in the 1980s. [bib_ref] Penicillium marneffei infection and recent advances in the epidemiology and molecular biology..., Vanittanakom [/bib_ref] Infection by T. marneffei is rarely reported in non-HIV-infected hosts, [bib_ref] Talaromyces (Penicillium) marneffei infection in non-HIVinfected patients, Chan [/bib_ref]
## Rapid and precise diagnosis of t. marneffei pulmonary infection in a hiv-negative patient with autosomal-dominant stat3
mutation: a case report immunosuppressive therapies, and being positive for anti-interferon-gamma autoantibody. Therefore, it is important to increase the diagnostic efficiency of this disease, especially in HIV-negative hosts, with a effective technique. Here, we report a rare case of a HIV-negative patient with lung T. marneffei infection with a STAT3 (signal transducer and activator of transcription 3) mutation.
## Case presentation
A 34-year-old young woman was admitted to our department for "recurrent cough for 6 months, acute exacerbation with dyspnea for 1 month" on 7 January 2019. The female presented with a 6-month history of slight nonproductive cough, shortness of breath after exercise, and complained of mild fever and night sweating with yellowbrown sputum for several days but denied chest pain. After the application of antibiotics (the detail was not clear) in a local hospital, temperature declined to normal, but she still had cough and dyspnea with exercise. A chest computed tomography (CT) scan (1 January 2019,showed multiple ground glass opacities with multiple bullae in both lungs. The patient had been slightly immunocompromised (the detail was not clear) since infancy, and had undergone several surgical treatments for suspected pleurisy and skin infection around the left ear. She had a history of viral hepatitis B, but was not on regular treatment, and no smoking or alcohol history. She was born in Gansu province, northwest of China, and moved to Hangzhou 10 years ago.
On physical examination, vital signs appeared normal, moist rales could be heard in both lungs, and there was no obvious cyanosis. Her HIV test was negative. The serum CA125 and NSE levels were 65.3 (reference 0-33 kU/l), 21.9 (reference 0-16.6 μg/l), respectively. The total T-lymphocyte count and CD4+ were 900 and 380 cells/µl, respectively. Hemoglobin was only 91 g/l. Blood gas analysis was normal. Serum immunoglobulin (Ig)E, IgG, IgA, and IgM were all normal. The plasma galactomannan antigen test and serum cryptococcal antigen agglutination test were normal. The sputum and blood cultures for both fungus and bacteria and sputum acid-fast bacillus test were negative. Other routine laboratory examinations were normal, such as the white blood count, C-reactive protein, glucose, aminotransferases, creatinine, vasculitis antibodies, and autoantibodies. The CT scan showed multiple disseminated ground glass opacities with multiple bullae in both lungs, no pleural effusion and pleural thickening, and no swollen superficial lymph nodes throughout the body, and no abnormal echocardiography, abdominal B-ultrasonic, and brain CT were observed.
Bronchoscopic examinations revealed uneven local membrane surface [fig_ref] Figure 2: Bronchoscopy showed local uneven membrane surface [/fig_ref] , hypoechoic areas in group 7, 4R, and 11Rs mediastinal lymph nodes through convex-probe endobronchial ultrasound [fig_ref] Figure 2: Bronchoscopy showed local uneven membrane surface [/fig_ref] , and hypoechoic shadow in the dorsal segment of the right lower lobe (RLL) through radial-probe endobronchial ultrasound [fig_ref] Figure 2: Bronchoscopy showed local uneven membrane surface [/fig_ref]. Cultures of the bronchoalveolar lavage fluid (BALF) for bacteria and acid-fast bacilli were all negative. The galactomannan test and cryptococcal antigen agglutination test of BALF were also negative. Upon histological examination, chronic granulomatous inflammation was found in the dorsal segment of RLL and the group 7 lymph node. Next generation sequencing (NGS) of the BALF confirmed lung infection with T. marneffei 2 days later [fig_ref] Table 1: NGS of BALF identified 566 T [/fig_ref]. About 1 week later, culture of BALF and the biopsied tissue mass also showed the existence of T. marneffei. Based on the pathogen's temperature-dependent dimorphic growth characteristic and the production of soluble red pigment and PAS-negative cell content, the isolate was identified definitively as T. marneffei [fig_ref] Figure 3: Culture of BALF revealed Talaromyces marneffei, which shows temperature-dependent dimorphic character, growing... [/fig_ref].
On the other hand, considering that the patient's HIV test was negative but the total counts of lymphocytes and CD4+ were slightly decreased, together with her history since infancy, we could not rule out congenital immunodeficiency. A blood gene detection test was taken, which indicated a loss-of-function mutation in the gene STAT3; however, there were no similar mutations in her parents (Table 2,.
# Discussion and conclusion
This is a relatively rare case report of a T. marneffei pulmonary infection in a HIV-negative patient with STAT3 mutation. The application of NGS in BALF greatly assists the rapid diagnosis of T. marneffei infection.
Patients with some immunodeficiencies, such as AIDS, AD-HIES, hyper-IgM syndrome, certain immunosuppressive therapies, and variety of transplants, are susceptible to T. marneffei as an opportunistic fungus. [bib_ref] Disseminated penicilliosis, recurrent bacteremic nontyphoidal salmonellosis, and burkholderiosis associated with acquired immunodeficiency..., Tang [/bib_ref] Affected individuals often suffer quick progression to multiple organ failure and finally death. Infection by T. marneffei is rarely reported in non-HIV-infected hosts, 3 but in clinical practice, the incidence rate of T. marneffei infection in HIV-negative individuals is increasing
year by year. In a report of five Chinese non-HIVinfected children and teenagers with T. marneffei infection, it was shown that four had had chronic mucocutaneous candidiasis since infancy, and one had AD-HIES. [bib_ref] Penicillium marneffei infection and impaired IFN-γ immunity in humans with autosomaldominant gain-of-phosphorylation..., Pamela [/bib_ref] [bib_ref] STAT1 mutations in autosomal dominant chronic mucocutaneous candidiasis, Van De Veerdonk [/bib_ref] Mutation in gene STAT3 was identified by Sanger sequencing in our patient (c.92G>A, p.R31Q); however, neither parent carried a similar mutation. STAT3, as a signal transducer and transcription factor, activates downstream of various of cytokine signals, including interferon-α, interleukin (IL)-6, and IL-10, and others. [bib_ref] JAKs and STATs in immunity, immunodeficiency, and cancer, O'shea [/bib_ref] It is reported that STAT3 mutation is usually related to AD-HIES. AD-HIES is a very rare primary immunodeficiency, characterized by elevated serum IgE and eczema, recurrent skin infections, and sinopulmonary infections. [bib_ref] A novel STAT3 mutation in a Qatari patient with hyper-IgE syndrome, Chaimowitz [/bib_ref] The classic triad of abscesses, pneumonia, and elevated IgE level was identified in 77% of all patients and 85% of those older than 8 years of age. At present, IgE is higher during childhood in some cases phenotypically, which may decrease, or even fall below normal, with age. The diagnostic criteria are not adequate for our patients. A consistent immunophenotype in AD-HIES patients is impaired development of Th17 lymphocytes, due mainly to the key role of cytokine signaling through STAT3 in their generation. In addition, a loss-of-function mutation in the STAT3 gene (STAT3-deficiency) is also frequently associated with susceptibility to fungal infections, including Talaromyces or aspergillosis, although its pathogenesis remains largely unknown. STAT3-deficient patients showed a defective adaptive immune response, with lower production of cytokines, including IFN-γ, IL-17, and IL-22, 9 which could be the reason for their susceptibility to fungal infection in HIV-negative patients. Regrettably, cytokines were not detected and IgE is negative in this patient. From a clinical perspective, the patient fails to meet the recent diagnostic criteria of HIES. To conclude, a lossof-function mutation in STAT3 gene is a rare primary deficiency, and our study was limited in some aspects. We are not sure how big a role it plays in this case, but it indicates that STAT3deficiency indeed increases susceptibility to microbial infections of lungs, which is the direction we are taking for our further research.
T. marneffei can lead to various kinds of infections involving multiple organs or systems, including the lungs, blood, skin, central nervous system, and bone marrow, so if not diagnosed or treated in a timely manner, it can be life-threatening. Whether in HIV-positive or HIV-negative patients, the clinical characteristics are similar. [bib_ref] Retrospective analysis of 15 cases of Penicillium marneffei infection in HIV-positive and..., Li [/bib_ref] The most common symptoms are cough, fever, anemia, weight loss, malaise, hepatosplenomegaly, and cutaneous lesions, but nonspecific and with little significance for differential diagnosis. The lung is the organ most commonly involved, occurring in 64% of HIV-infected patients and 75% of non-HIV-infected patients. [bib_ref] Retrospective analysis of 15 cases of Penicillium marneffei infection in HIV-positive and..., Li [/bib_ref] In this case, dyspnea was the main complaint along with recurrent cough; no other organs and systems seem to be involved.
Similarly in laboratory tests, there is no significant difference between HIV-negative and HIV-positive patients. [bib_ref] Retrospective analysis of 15 cases of Penicillium marneffei infection in HIV-positive and..., Li [/bib_ref] T. marneffei is well known to be the only temperature-dependent dimorphic pathogen in Penicillium, growing as a mycelium at temperatures 25°C-30°C with the generation of a soluble red pigment, and as yeast-like cells at 37°C. Only the yeast phase has pathogenicity. In addition, a mulberry-shaped cell mass, sausage-shaped cells, and a transverse wall are the three main morphological characteristics of T. marneffei growth in tissue, [bib_ref] Talaromyces marneffei infection in a lung cancer patient: a rare case report, Lin [/bib_ref] which were also found in our case.
In former clinical practice, the diagnosis of T. marneffei infection depended highly on tissue culture and histopathologic results, which can be confined to the low positive rate and can be timeconsuming, respectively, especially as fungal cultures usually take about 3-7 days. We performed NGS on the patient's BALF, which detected numerous nucleotide sequence reads corresponding to T. marneffei 2 days later, and thus resulted in the timely diagnosis and treatment of T. marneffei pulmonary infection.
The successful application of NGS assisted the rapid diagnosis of T. marneffei infection, providing a powerful skill in clinical practice and revealing the potential value of this procedure in rapid etiological diagnosis. [bib_ref] Rapid and precise diagnosis of disseminated T.marneffei infection assisted by high-throughput sequencing..., Zhu [/bib_ref] It is well known that amphotericin B, itraconazole, voriconazole, fluconazole, and terbinafine are the antifungal drugs most commonly used for therapy. In addition, itraconazole and amphotericin B are reported to be more effective in clinical practice, whereas the clinical response to fluconazole is relatively poor. [bib_ref] In vitro interactions of calcineurin inhibitors with conventional antifungal agents against the..., Mo [/bib_ref] Current guidelines for the therapy of T. marneffei infection recommend amphotericin B treatment for 2 weeks, then adequate oral itraconazole for 10 weeks, and finally secondary prophylaxis. 14 From our case, we found that, following the application of oral itraconazole (200 mg, every 12 h) for 3 months, the patient recovered satisfactorily and lesions absorbed obviously on CT scans. To date, she continues to receive the application of itraconazole (200 mg per day) and follow up until the present. The general maintenance dose lasts for 1 year, depending on the efficacy or whether the risk factors can be terminated or not. However, despite standard treatment strategy, most infected individuals experience recurrence several months or even years later. Earlier research indicated that mortality in HIV-negative individuals was higher than in HIV-positive individuals, [bib_ref] Retrospective analysis of 15 cases of Penicillium marneffei infection in HIV-positive and..., Li [/bib_ref] which may be related partly to delayed diagnosis because of the lack of an effective and rapid diagnosis technique. 15
In conclusion, the incidence of T. marneffei infection in non-HIV-infected patients has been relatively low in recent years; however, it has shown a significant increase, 16 even in some healthy hosts. Patients who are congenitally immunocompromised by a STAT3 mutation may be among potential hosts. Finally, rapid diagnosis and early stage treatment are critical for improving the prognosis. The successful application of NGS can play an important role in rapid diagnosis, revealing the potential value of this technique in rapid etiological diagnosis. Further studies are required to explore the pathogenesis and mechanisms of infection in HIV-negative patients with STAT3 mutation infected with T. marneffei.
[fig] Figure 1: (A) Past chest CT scan (1 January 2019) showing multiple disseminated ground glass opacities with multiple bullae. (B, C) Chest CT scan during follow up (26 January 2019 and 27 April 2019, respectively) showing distinct resolution of both lungs. CT, computed tomography. journals.sagepub.com/home/tar 3 [/fig]
[fig] Figure 2: Bronchoscopy showed local uneven membrane surface (A), hypoechoic areas in group 7, 4R, and 11Rs mediastinal lymph nodes through CP-EBUS (B), and hypoechoic shadow in the dorsal segment of the right lower lobe through RP-EBUS (C). CP-EBUS, convex-probe endobronchial ultrasound; RP-EBUS, radial-probe endobronchial ultrasound. [/fig]
[fig] Figure 3: Culture of BALF revealed Talaromyces marneffei, which shows temperature-dependent dimorphic character, growing as yeast-like cells at 37°C (A) and as a mycelium at temperatures between 25°C and 30°C (B); the cells produced red pigment at 25°C (C).BALF, bronchoalveolar lavage fluid.So far, the patient was diagnosed as having a STAT3 mutation and lung infection with T. marneffei. Considering her financial condition, she was prescribed oral itraconazole (200 mg, every 12 h) therapy on 13 January 2019 and later discharged. Two repeated chest CT, on 26 January 2019 (Figure 1B) and 27 April 2019 (Figure 1C), revealed distinct resolution of both lungs. The symptoms of shortness of breath and cough were also obviously alleviated, and the patient continues to receive treatment (itraconazole, 200 mg per day) and follow up at present. [/fig]
[fig] Figure 4: (A) Family map and Sanger sequencing. (B) Illustration of the functional structure of STAT3: p.R31Q in red and other known pathogenic, or likely pathogenic mutations, in black. (C) The labeling of p.R31Q on the structure of STAT3 modeled by the I-TASSER algorithm. STAT3, signal transducer and activator of transcription 3. [/fig]
[table] Table 1: NGS of BALF identified 566 T. marneffei nucleotide sequences. BALF, bronchoalveolar lavage fluid; NGS, next generation sequencing. [/table]
[table] Table 2: Heterozygous missense mutation in exome regions of gene STAT3 was identified by Sanger sequencing (c.92G>A, p.R31Q). [/table]
|
Serum IgG titers against periodontal pathogens are associated with cerebral hemorrhage growth and 3-month outcome
[formula] a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 [/formula]
# Materials and methods
## Patients
Consecutive acute cerebral hemorrhage patients, who were admitted to the Hiroshima University Hospital and the Suiseikai Kajikawa Hospital, Japan, from January 2013 to April 2016, were enrolled in this prospective study. The study protocols were approved by the ethics committee of the Hiroshima University Hospital (Epd-614-2) and the Suiseikai Kajikawa Hospital (2015-3) and performed according to the guidelines of the national government based on the Helsinki Declaration of 1964. Written informed consent was obtained from all patients or their relatives. All data analyses were conducted in a blinded manner.
We included patients who were admitted within 7 days from onset, were aged � 20 years, and for whom consent to participate in this study was obtained from the patient or their relatives. We excluded patients who could not undergo head computed tomography (CT) and magnetic resonance imaging (MRI). We also excluded the patients who were diagnosed with hemorrhagic infarction or trauma-induced hemorrhages.
## Data acquisition
Baseline clinical characteristic data, including age, sex, body mass index (BMI), comorbidities (hypertension, diabetes mellitus, dyslipidemia, atrial fibrillation, chronic kidney disease, and stroke), smoking and drinking habits, medication of antithrombotic drugs (anti-platelet and anticoagulant), onset to admission time, blood pressure, and C-reactive protein (CRP) levels were collected from all patients. Two stroke specialists (SA and EI) evaluated the stroke severity and conscious level. Stroke severity on admission was evaluated using the NIH Stroke Scale (NIHSS). Conscious level was evaluated with Glasgow coma scale. Comorbidities were defined according to a previous reportbased on the Japanese hypertension, diabetes mellitus, dyslipidemia, atrial fibrillation, and chronic kidney disease guidelines. Hypertension was defined as the use of anti-hypertensive medication before admission or confirmed blood pressure of �140/90 mmHg at rest measured 2 weeks after onset. Diabetes mellitus was defined as a glycated hemoglobin level of �6.5%, fasting blood glucose level of �126 mg/dL, or use of antidiabetes medication. Dyslipidemia was defined as a total cholesterol level of �220 mg/dL, lowdensity lipoprotein cholesterol level of �140 mg/dL, high-density lipoprotein cholesterol level of <40 mg/dL, triglyceride levels of �150 mg/dL, or use of anti-hyperlipidemia medication. Atrial fibrillation was defined as follows: (1) a history of sustained or paroxysmal atrial fibrillation or (2) atrial fibrillation detection on arrival or during admission. Renal functioning was calculated with the estimated glomerular filtration rate (eGFR) using a revised equation for the Japanese population as follows: eGFR (mL/min/1.73 m 2 ) = 194 × serum creatinine −1.094 × age −-0.287 × 0.739 (for women). Chronic kidney disease was defined as an eGFR <60 mL/min/ 1.73 m 2 . Imaging analysis with head computed tomography (CT) and magnetic resonance imaging (MRI) was performed in all patients for acute cerebral hemorrhage diagnosis. The cerebral hematoma was evaluated using CT. The admission and follow-up CT scans (24 hours after admission) were performed with axial 5-mm section thickness. Two experienced neurologists (MN and YS) measured the cerebral hematomas using the ABC/2 formula. The neuroimaging evaluation remained blinded from the clinical assessment. Based on the criteria used in several large clinical studies, hematoma growth was defined as an increase in hematoma volume of >33% or >12.5 mL at 24-hour follow-up. The other two experienced neurologists (NK and KI) also performed evaluations of MRI findings and hematoma growth neuroimaging predictors, which were detected as the blend sign or black hole sign using plain head CT. Cerebral amyloid angiopathy (CAA) was diagnosed using modified Boston criteria. We collected clinical data regarding the acute phase, including intraventricular hemorrhage extension, pharmacological blood pressure management during the first 24 hours, surgical management approach, tube ventilatory use, and septic complications.
## Plos one
When we evaluated the 3-month outcome, we excluded patients who were disabled prior to stroke (corresponding to premorbid modified Rankin Scale [mRS] score �2). An unfavorable 3-month outcome was defined as a 3 or higher on the mRS.
## Measurement of serum antibody titers of periodontal pathogens
Serum IgG antibody titers of periodontal pathogens were determined using ELISA as previously described. Bacterial antigen-coated wells were washed with phosphate-buffered saline with Tween (PBST); serum samples in PBST were then added to the wells. After incubation at 4˚C overnight, the wells were washed with PBST and filled with alkaline phosphataseconjugated goat anti-human IgG (gamma-chain specific, Abcam, Cambridge, MO) in PBST. After another incubation at 37˚C for 2 hours, the wells were again washed with PBST, an aliquot of p-nitrophenylphosphate at 1 mg/mL (WAKO Pure Chemical Industries Ltd., Osaka, Japan) in 10% diethanolamine buffer was added to each well as a substrate, and incubation was performed at 37˚C for 30 minutes. Optical density at 405 nm was measured using a microplate reader (iMark, Bio-Rad Laboratories Inc., Hercules, CA). Serum samples were collected from patients within 3 days after stroke onset and stored at −80˚C. Sonicated preparations of the following periodontal pathogens were used as bacterial antigens Porphyromonas (P.) gingivalis, Aggregatibacter (A.) actinomycetemcomitans, P. intermedia, Prevotella nigrescens, F. nucleatum, Treponema denticola, Tannerella forsythensis, Campylobacter rectus, and Eikenella corrodens. We selected these representative periodontal pathogens based on the previously reported association of serum antibody titers with stroke outcome. The serum of 5 healthy individuals was pooled and used for calibration. Using serial dilutions of the pooled control serum, the standard reaction was defined based on the ELISA unit (EU) such that 100 EU corresponded to a 1:3200 dilution of the calibrator sample. For statistical analysis, we used the common logarithms of serum IgG antibody titers.
# Statistical analysis
Data are expressed as the mean ± standard deviation or the median (minimum, maximum) for continuous variables, and frequencies and percentages for discrete variables. Statistical analysis was performed using the JMP 14.0 statistical software (SAS Institute Inc., Cary, NC, USA). The statistical significance of intergroup differences was assessed using the unpaired ttest or Mann-Whitney U test for continuous variables or the Fisher exact test or χ 2 test for discrete variables as appropriate. Because there were no reports of hemorrhagic stroke, we calculated the sample size according to the past investigations for the IgG titers of periodontal pathogens in atherothrombotic stroke. Based on an alpha level = 0.05 and power = 0.80, we estimated that we would require a total of n = 99 participants. Baseline data of cerebral hemorrhage patients were analyzed, and two-step strategies were employed to assess the relative importance of variables in their association with hemorrhage growth and poor outcome using least square linear regression analysis. We first performed a univariate analysis, followed by a multi-factorial analysis with selected factors with p < 0.05 in the former analysis. We considered p < 0.05 as statistically significant. We also calculated the statistical power and effect size as post hoc analysis.
# Results
A total of 115 patients (44 females, aged 71.3 ± 13.1 years) with acute cerebral hemorrhage were registered in this study. The baseline clinical characteristics are shown in. Among them, 32 (27.8%) patients had histories of stroke, 16 of whom were cerebral hemorrhage. The number of patients with anti-platelet and anticoagulant drug use was 20 (17.4%) and 15 (13.2%), respectively. The median time from onset to admission was 147 minutes (min-max: 26-7200). CAA was diagnosed 16 (13.9%) patients, of whom 7 patients were probable CAA and 6 patients were possible CAA. The mean serum IgG titers of periodontal disease pathogens from all patients are summarized in.
We found hemorrhage growth in 13 patients (11.3%). The potential factors associated with hemorrhage growth were evaluated using the univariate analysis (listed in Tables 1 and 2). In this analysis, hemorrhage growth was associated with the history of atrial fibrillation, usage of anticoagulant, and the IgG titers of A. actinomycetemcomitans. Among patients with hemorrhage growth, 5 (38.5%) patients had used an anticoagulant, namely warfarin, and were not injected with the prothrombin complex concentrate during the time of admission due to the Japanese medical insurance system at that time. Multivariate logistic regression analysis revealed that usage of anticoagulant (odds ratio 8.36, 95% CI 1.35-51.70, p = 0.02), septic complications (odds ratio 10.20, 95% CI 1.94-53.72, p = 0.01), and the IgG titer of A. actinomycetemcomitans (odds ratio 5.26, 95% CI 1.52-18.25, p = 0.01) were independently associated with hemorrhage growth . Regarding the IgG titer of A. actinomycetemcomitans, the statistical power and effect size of Cohen's d were 0.80 and 0.67, respectively.
When we evaluated the 3-month outcome, 22 patients were excluded based on premorbid mRS scores of �2. The potential factors associated with poor outcome (mRS score �3) (listed in Tables 1-3) were evaluated using the univariate analysis. We found that poor outcome was associated with age, NIHSS score, cerebral hematoma volume, cerebral hematoma growth, and IgG titers of F. nucleatum. Multivariate logistic regression analysis revealed that age (odds ratio 1.09, 95% CI 1.01-1.17, p = 0.02), NIHSS score (odds ratio 1.29, 95% CI 1.09-1.52, p = 0.002), and the IgG titer of F. nucleatum (odds ratio 7.86, 95% CI 1.08-57.08, p = 0.04) were independently associated with poor outcome, but not cerebral hematoma volume or growth. Regarding the IgG titer of F. nucleatum, the statistical power and effect size of Cohen's d were 0.91 and 0.69, respectively.
# Discussion
Our findings suggest that periodontal disease may be associated with cerebral hemorrhage and its clinical course. Specifically, we reveal for the first time that serum IgG titers of a few particular periodontal pathogens may be useful biomarkers for predicting the clinical course of a cerebral hemorrhage.
In the present study, we provide evidence that an elevated serum IgG titer of A. actinomycetemcomitans is an independent factor for predicting hemorrhage growth. A. actinomycetemcomitans is a gram-negative, facultatively anaerobic coccobacillus and is considered the major etiologic agent of localized aggressive periodontitis. It also contributes to chronic periodontitis, and according to previous reports, an elevated serum titer of A. actinomycetemcomitans predicts stroke and coronary heart disease. A potential explanation for these relationships is that individuals infected with A. actinomycetemcomitans harbor T-cells specific to this bacterium in their blood, which express the receptor activator of nuclear factor-κB (RANK) ligand. This ligand stimulates the vascular smooth muscle cells to produce matrix metalloproteinase-9, which may promote the growth of cerebral hemorrhages.
We further demonstrate that the serum IgG titer of F. nucleatum independently predicted an unfavorable outcome in cerebral hemorrhage, which is consistent with a previous study. Fusobacterium nucleatum is strictly an anaerobic gram-negative rod bacterium, normally found in the oral cavity. It is considered to be a periodontal pathogen because it is frequently isolated from periodontitis lesions, produces a high number of tissue irritants, and often aggregates with other periodontal pathogens as a bridge between early and late colonizers. Fusobacterium nucleatum, which can pass the blood-brain barrier, is associated with colon cancer and Alzheimer's disease, and was found to cause brain abscesses. This bacterium also has abilities to adhere to and invade host vascular endothelial cells using the protein, Fusobacterium adhesin A (FadA), which binds to vascular endothelial-cadherin on the cell surface, thereby triggering a breakdown of endothelial cell-to-cell junctions. The virulence of F. nucleatum mediates endotoxin activity, hemagglutination, as well as aggregation and death of immune cells via their outer membrane proteins, such as fibroblast activation protein 2 (FAP-2) and radiation genes (RadD). In addition, F. nucleatum causes an increase in the expression of genes associated with host immune responses, such as There were some limitations to this study. First, there was the issue of sample size and sampling bias. In this study, the sample size was modest. Thus, despite the multicenter study design, sample size calculation, and post hoc analysis, some sampling bias might exist. The average age of patients and the numbers of patients with CAA, previous stroke, and taking antithrombotic drugs were also relatively high, whereas the total frequency of intracerebral hematoma growth was low. Regarding this low intracerebral hematoma growth frequency, the time from onset to admission might have affected the results. In this study, we included patients within 7 days from onset. However, all patients were considered during the acute phase, and the median time from onset to admission was 147 minutes. Time from onset to collection of blood sample also varied because of the study design. However, all patients were tested within a week from onset. Since periodontal disease is a chronic disease, mild variation might not affect the results. Second, this was a cross-sectional observational study, which makes it difficult to adequately assess the biological relationship between periodontal pathogens and cerebral hemorrhage. To provide more conclusive evidence, in vivo animal experiments are required. If periodontal pathogens themselves are risk factors for a malignant outcome of cerebral hemorrhage, daily oral care and regular dental examination could practically improve the clinical course. Third, serum IgG titers were used to investigate the association between periodontal disease and the clinical course of a cerebral hemorrhage, but the severity of periodontal disease and the intraoral environment was not directly evaluated. Therefore, future studies could consider these parameters and assess how they affect cerebral hemorrhage pathology and clinical outcomes. While the influence of serum IgG titers of periodontal pathogen on cerebral hemorrhage has not been fully determined, specific periodontal pathogen infection can be a useful biomarker for predicting the clinical course of cerebral hemorrhage.
Stroke mortality rates have been decreasing owing to the development of advanced medical treatments; however, disability rates of stroke survivors have been increasing. Given our findings of an association between periodontal pathogen titers and the clinical course of a cerebral hemorrhage, it is important to seriously consider each of these pathogens and their potential negative impact on affected individuals.
# Conclusions
Our observations reveal the impact of two periodontal pathogens on cerebral hemorrhage, namely that elevated serum IgG titers of A. actinomycetemcomitans predicted hemorrhage growth and that those of F. nucleatum predicted a poor outcome in patients with this disease. Consequently, periodontal pathogens could play an important role in the management and treatment of stroke patients, for which the assessment of IgG titers has shown to be a valid tool. |
Development of Severe Acute Pancreatitis Following Uncovered Metallic Stent Placement: A Rare Case Report
Self-expandable metallic stents (SEMSs) are widely used for malignant biliary stricture (MBS). Acute pancreatitis is an early complication following SEMS placement. In the present case, the patient developed severe acute pancreatitis after SEMS placement for MBS because of metastatic lymph nodes. Endoscopic retrograde cholangiopancreatography, endoscopic sphincterotomy and an endoscopic nasobiliary drainage tube placement were performed. After seven days, an uncovered SEMS was placed; however, severe acute pancreatitis occurred, and the SEMS was drawn out emergently. In SEMS placement for patients with MBS caused by non-pancreatic cancer, SEMS should be selected carefully while considering each patient's case.
# Introduction
Self-expandable metallic stents (SEMSs) have been widely used for malignant biliary stricture. Acute pancreatitis is an early complication that can follow SEMS placement in between 0% and 24% of cases . However, there are few reports on severe acute pancreatitis onset after SEMS placement (12).
We herein report a rare case of severe acute pancreatitis after SEMS placement for malignant distal bile duct stricture and discuss the development of acute pancreatitis after metallic stent placement.
## Case report
A man in his 70s who was receiving chemotherapy for advanced squamous cell carcinoma of the lung complained of jaundice. Hepatobiliary enzymes and bilirubin levels were markedly elevated, and the patient was referred to our department.
Computed tomography (CT) revealed multiple liver metastases, hepatomegaly, multiple swollen abdominal lymph nodes, and extrahepatic bile duct dilatation [fig_ref] Figure 1: Contrast-enhanced computed tomography showing multiple metastatic liver tumors and metastatic lymph node... [/fig_ref]. Magnetic resonance cholangiopancreatography showed intrahepatic and extrahepatic bile duct dilatation, distal bile duct obstruction, and smooth main pancreatic duct without dilatation [fig_ref] Figure 2: Magnetic resonance cholangiopancreatography showing the intrahepatic and extrahepatic bile duct dilatation, the... [/fig_ref]. Endoscopic retrograde cholangiopancreatography (ERCP) revealed distal bile duct stricture due to metastatic lymph node [fig_ref] Figure 3: ERCP findings [/fig_ref]. After endoscopic sphincterotomy (EST) with a small incision, an endoscopic nasobiliary drainage (ENBD) tube was placed [fig_ref] Figure 3: ERCP findings [/fig_ref] , and the yellowing effect was confirmed. After 7 days, ERCP was performed a second time, and an uncovered SEMS (10 mm in diameter and 8 cm in length; Bonastent , Sewoon Medical, Cheonai, Korea) was placed without any additional procedures, including contrast medium injection and guidewire insertion into the pancreatic duct, a biopsy, and intraductal ultrasonography [fig_ref] Figure 3: ERCP findings [/fig_ref]. These ERCP-related procedures were performed by a well-experienced endoscopist in 11 minutes.
A marked increase in pancreatic enzymes was observed Intern Med 60: 1703-1707, 2021 DOI: 10.2169/internalmedicine.6394-20 (serum amylase level 2,266 U/L after 2 hours, and 4,841 U/ L the following day), and CT showed swelling of the pancreas, peripancreatic inflammation, and the spread of inflammation toward the inferior pole of the left kidney [fig_ref] Figure 4: Computed tomography showing severe acute pancreatitis findings [/fig_ref]. We considered severe acute pancreatitis with CT grade 2 according to the Japan Medical Care Guideline of acute pancreatitis. We believed that the cause of severe acute pancreatitis was SEMS placement and immediately performed ERCP. We smoothly drew out the SEMS via the forceps channel of the endoscope using snare forceps without resistance and placed a tube stent and an ENBD tube; 2 days later, severe acute pancreatitis which was CT grade 2 and fulfilled 3 prognostic factors [C-reactive protein (CRP), 29.3 mg/dL; lactate dehydrogenase, 2,872 U/L; and age, >70 years old] according to the established guidelines [fig_ref] Figure 5: Contrast-enhanced computed tomography two days after the second ERCP procedure showing continuing... [/fig_ref].
The Bedside Index for Severity in Acute Pancreatitis score was 3 points out of 5 (blood urea nitrogen, age, pleural effusion); the Acute Physiology and Chronic Health Evaluation II score was 17 points. The patient received conservative treatment, including 3,800-4,000 mL infusion, 1,500 mg of gabexate mesylate, and 13.5 g of tazobactam piperacillin hydrate intravenous administration daily continuously. Following the removal of the SEMS, the pancreatic enzyme levels dropped dramatically, and the CRP levels declined steadily. The recovery of severe acute pancreatitis was confirmed by laboratory data and CT at 21 days after the first ERCP procedure; a fourth ERCP procedure was subsequently performed, and another tube stent was placed. The patient was discharged 29 days after the first ERCP procedure without pancreatitis onset.
# Discussion
Acute pancreatitis is an early complication that can occur after SEMS placement for malignant biliary stricture. In endoscopic SEMS placement, acute pancreatitis has been reported to occur in 0-14% of cases in the past decade (1-10). Although the frequency varies among reports, acute pancreatitis occurs with a certain probability. However, there are few reports of severe acute pancreatitis following SEMS placement (12).
In our institution, SEMS placement was performed for 559 patients with malignant biliary stricture over the 12-year period between 2008 and 2019. Of the 559 patients, acute pancreatitis occurred in 14 patients (2.5%). These 14 patients underwent EST and biliary drainage using ENBD and/ or tube stent at the first ERCP procedure and SEMS placement at the second or third ERCP procedure. Of these patients, urgent ERCP and SEMS removal were performed in 3 (0.54%), including the present case, due to the rapid increase in pancreatic enzyme levels observed, which was considered to have been caused by SEMS placement. Of these three patients, two had partially covered SEMS, and 1 had an uncovered SEMS; both were braided-type SEMS. Among these 3 patients, severe acute pancreatitis only developed in the current case (0.18%).
Acute pancreatitis after metallic stent placement is considered to occur as a consequence of SEMS obstructing the pancreatic duct orifice of the papilla of Vater, which blocks pancreatic juice outflow and consequently induces acute pancreatitis. The use of EST can help avoid pancreatic duct obstruction to some extent. Indeed, Sugawara et al. reported that 24% of patients with acute pancreatitis were observed after percutaneous transhepatic biliary SEMS placement across the papilla without EST (11). However, EST has been reported to be unrelated to the development of pancreatitis following SEMS placement . In previous reports, the etiology of biliary stricture was limited to pancreatic cancer EST should be performed to reduce acute pancreatitis after SEMS placement, especially for patients with nonpancreatic cancer, as recommended in the European Society of Gastrointestinal Endoscopy Guidelines (13).
Covered SEMS may block the orifice of the pancreatic duct after EST, so these types of stent are thought to carry a higher risk of acute pancreatitis than others. It has been reported that acute pancreatitis occurs more frequently in cases with covered SEMSs than in those with uncovered SEMSs (1, 2); however, there are no marked differences in the rates of acute pancreatitis development between partially covered and uncovered SEMSs (3, 4). This may be related to the underlying disease, the presence or absence of dilatation of the main pancreatic duct, and/or atrophy of the pancreatic parenchyma (2, 5, 11). In patients with pancreatic duct obstruction, such as pancreatic head cancer, we empirically understand that acute pancreatitis is unlikely to occur, even if the pancreatic duct orifice is blocked because of chronic obstructive pancreatitis. In a randomized comparative study of covered versus uncovered SEMSs in cases of malignant bile duct stricture, 76-77% of the total patients had pancreatic cancer; the incidence of acute pancreatitis was reported as 1.5% and 2.0%, respectively (4).
Acute pancreatitis after fully covered SEMS placement has been reported in 9.3% of patients (n=602), while the presence of a moderate to high degree of acute pancreatitis was reported in 1.3% of patients (6). A relatively high rate of acute pancreatitis following fully covered SEMS placement was reported in a small number of patients with pancreatic cancer (n=169; 28%) and in >50% of the patients with a non-dilated pancreatic duct (n=349; 58%).
Acute pancreatitis is likely to occur in patients with no dilation of the main pancreatic duct and no pancreatic atrophy (9). In the current case, we suspected that the large metastatic lymph node had compressed the pancreatic parenchyma over the pancreatic duct due to SEMS placement, resulting in a sudden increase in pancreatic ductal pressure and simultaneous parenchyma damage and the development of acute pancreatitis. Generally, the mechanisms underlying aggravation of acute pancreatitis have been thought to be local inflammation, increased cytokine production that leads to systemic cytokine overflow, and the production of other mediators that induce systemic inflammatory syndrome. Systemic inflammation induces multiple organ failure and/or disseminated intravascular coagulation. However, the exact cause and mechanism of the progression of acute pancreatitis to severe acute pancreatitis were unknown in the present case.
In the context of acute pancreatitis during SEMS placement for malignant bile duct stricture due to non-pancreatic cancer, whether or not there are differences between partially covered/uncovered SEMS and fully covered SEMS is unclear. Further studies are needed to clarify this issue. When the risk of acute pancreatitis is considered high, particularly in patients with bile duct stricture due to nonpancreatic cancer, a stent with a smaller radial force and/or smaller stent diameter may be an option, as previously described (9). In addition, a braided stent that can be endoscopically removed in the unlikely event of acute pancreatitis may represent a viable option.
In conclusion, we herein report a rare case of severe acute pancreatitis after uncovered SEMS placement for malignant bile duct stricture due to metastatic lymph node from lung cancer. A SEMS with a smaller diameter and/or lower axial force may have been ideal. In cases of SEMS placement for patients with malignant biliary stricture due to nonpancreatic cancer, it is necessary to perform EST and select a stent after careful consideration of the patient-specific factors.
The authors state that they have no Conflict of Interest (COI).
[fig] Figure 1: Contrast-enhanced computed tomography showing multiple metastatic liver tumors and metastatic lymph node swelling. a: Lymph node swelling around the extrahepatic bile duct (arrows). b: Lymph node swelling near the pancreatic head (arrow). [/fig]
[fig] Figure 2: Magnetic resonance cholangiopancreatography showing the intrahepatic and extrahepatic bile duct dilatation, the distal bile duct (lower bile duct) obstruction, and the smooth main pancreatic duct in the pancreatic head to body without dilatation. [/fig]
[fig] Figure 3: ERCP findings. a: Endoscopic cholangiography reveals lower-to-middle bile duct stricture. b: The papilla of Vater was of normal size and shape. c: An endoscopic nasobiliary tube was placed. d: Endoscopic sphincterotomy with a middle incision was performed. e, f: An endoscopic metallic stent was placed after 7 days using an uncovered braded type metallic stent (10 mm in diameter and 8 cm in length). ERCP: endoscopic retrograde cholangiopancreatography [/fig]
[fig] Figure 4: Computed tomography showing severe acute pancreatitis findings. a: The swollen pancreas and spread of inflammation to the abdominal cavity and retroperitoneum (arrows). b: The inflammation spread toward the inferior pole of the left kidney (arrow). [/fig]
[fig] Figure 5: Contrast-enhanced computed tomography two days after the second ERCP procedure showing continuing severe acute pancreatitis findings. a: Bilateral pleural effusion and atelectasis in the left lung. b: Peripancreatic inflammation and inflammation to the abdominal cavity (arrows). c: Inflammation spread around the left kidney (arrow). ERCP: endoscopic retrograde cholangiopancreatography (100%) (8), or there were many cases (82%) with dilatation of the main pancreatic duct, including pancreatic cancer (9). [/fig]
|
Randomised controlled trial protocol for the PROTECT-CS Study: PROTein to Enhance outComes of (pre)frail paTients undergoing Cardiac Surgery
# Introduction
By 2031, 25% of Canadians will be older than 65 years.By extension, it is expected that greater numbers of frail (defined as a cumulative decline in multiple physiological systems resulting in a vulnerability to stressor events 1 2 ) older adult individuals will be referred for a cardiac surgery procedure. While frailty is not necessarily synonymous with age, it is more prevalent among an older adult population.More specifically, in the last two decades, the increasing burden of heart disease has resulted in cardiac surgery being offered to older and more frail patients. Previous studies have demonstrated that patients with higher levels of frailty prevalence, ranging from 20% to 53%,Open access used, typically experience higher rates of postoperative morbidity, mortality and prolonged hospital length of stay with associated increased costs to the Canadian healthcare system. It has been previously identified that approximately 50% of older adults undergoing cardiac surgery are frail, resulting in a population who are at a higher risk for poor outcomes versus non-cardiac surgery patients. Furthermore, pre-frail (defined as a transitional physiological state between robustness and frailty, characterised by intermediate accumulation of physiological dysfunction, resulting in a minimal decline in physiological reserve and functional capacity, supported by a Clinical Frailty Scale (CFS) score equal to 3cardiac surgery patients are at an increased risk of adverse events such as increased duration of mechanical ventilation, morbidity and hospital stay, as well as postoperative frailty progression that ultimately results in functional disability and worse healthrelated quality of life (HRQoL). Outcomes associated with frail cardiac surgery patients are impacted further by malnutrition (defined as an unintentional, nutritional intake imbalance (not necessarily a decreased intake)), which is evident in 20% of the patients presenting for cardiac surgery. Chronic malnutrition which results from age-related physiological decline (ie, decreased mobility, sensory functions (olfaction and taste) and cognition), physiological challenges (ie, increased dependency and loneliness) and chronic illness (ie, cardiovascular, polypharmacy) in combination with perioperative starvation, which is the typical perioperative fasting, inflammatory and regenerative processes that induce an exponential nutritional demand resulting in an additional demand-supply imbalance (ie, a state of metabolic stress) can cause a further decline in health status and poor recovery to baseline functional capacity in already frail patients.
Malnutrition and frailty have shared risk factors with overlapping clinical presentation that exacerbate each syndrome.The previous standard preoperative protocol requires preoperative fasting for 6-8 hours (may extend to 8-12 hours). Further, a delay (mean of 2 days) in initiating postoperative nutritional support as well as an inadequacy of caloric intake (approximately 70% less than the recommended intake) has often been reported in cardiac surgery patients. This perioperative underfeeding compound the pre-existing chronic malnutrition in frail older patients undergoing cardiac surgery. Malnutrition can manifest into a catastrophic fat-free muscle mass decline, that intensifies 'senile sarcopenia' (defined as the typical age-related involuntary muscle and organ degradation; typically, 1%-2% per year after 30 years).This altogether results in an exponential loss of lean muscle mass (sarcopenia) and muscle function (dynapenia) with ageing.Importantly, even a minimal (10%) loss of fat-free muscle mass was found to be associated with a progressive functional decline, increased mortality and increased utilisation of healthcare services (including premature institutionalisation) among community-dwelling older adults.Furthermore, the malnutrition-accelerated decline in lean muscle mass and muscle function are consistent features of frailty (coined as the biological substrate of frailty exacerbation. The frail-malnourished patients cannot mobilise enough amino acids to commensurate with the 400% increased demand required to synthesise proteins for wound healing, immune function and acute-phase reactants.To compound this issue, the metabolic and immune response to injury induces insulin resistancefurther aggravating a caloric deficient state. Ultimately, this perioperative demand-supply imbalance results in an extensive insult to a reserve-deficient physiological system (as in a typical frail patient), disproportionately declining the health status and limiting a patient's recovery to baseline functional capacity.
Providing protein-caloric supplementation preoperatively, postoperatively and post-hospital discharge is recommended by many consensus guidelines 39 including: enhanced recovery protocols (ERPs),The European Society for Clinical Nutrition and Metabolism 41 and the Canadian Nutritional Support Clinical practice guidelines.Nutritional supplementation can be viewed as a potentially modifiable factor that can alleviate the preoperative frailty-chronic malnutrition associated with vulnerability. Cardiac surgery patients, in addition to their cardiac disease and associated comorbidities, experience higher levels of frailty and malnutrition versus other non-cardiac surgery patients.There is a significant knowledge gap regarding the efficacy of nutritional prehabilitation in preserving functional capacity (ie, mitigating frailty progression) and promoting enhanced recovery, among (pre)frail older adults undergoing cardiac surgery.
Enhanced recovery protocols, such as prehabilitation, have been associated with a reduction in overall complications and length of stay of up to 50% in vulnerable patients when compared with conventional perioperative patients in non-cardiac surgery populations.Early feasibility studies have shown promise with ERPs in cardiac surgery as well, however, require additional rigorous inquiry. Specifically, the impact of a nutritional supplementation ERP on mitigating long-term frailty progression and improving patient-reported outcomes (patient valued survivorship and recovery) in an increasingly ageing cardiac surgery patient population has been largely understudied. In Canada, when a patient requires elective cardiac surgery, they are placed on a 'waiting list' for as long as 3-4 months. At present, there is no formal process for engaging these patients to enhance their health during this waiting period, which represents a significant opportunity to optimise a (pre)frail older adult patient's clinical condition prior to surgery. Additionally, the consideration of longer term (not just in-hospital) HRQoL is vitally important to the patient-caregiver unit. As a result, it is becoming critically important for the healthcare system to develop strategies to improve clinical outcomes in this high-risk patient population.
The PROTein to Enhance outComes of (pre)frail paTients undergoing Cardiac Surgery (the PROTECT-CS Study) was developed as a practical and sustainable, new perioperative care pathway, informed by patientcaregiver values, to enhance short-term and long-term recovery of vulnerable cardiac surgery patients.
The primary objectives of the trial are: 1. To determine if a leucine-rich protein supplementation, consumed two times per day for a minimum of 7 days pre-procedure for elective patients (minimum of 2 days for non-emergent inpatients), two times per day during in-hospital postoperative recovery and two times per day for 8 weeks after the patient is discharged home can reduce perioperative functional decline among (pre)frail older cardiac surgery patients. 2. To determine if a perioperative leucine-rich protein supplementation will enhance short-term and longer term patient-reported outcomes. We hypothesise that the preoperative and postoperative supplementation of leucine-rich-formulated beverages to a patient's habitual diet will support the physiological processes to counter the impact of cardiac surgery-related stressors to reduce the postoperative functional decline. We also hypothesise that nutrition supplementation using a leucine-rich protein supplementation will facilitate postoperative recovery and improve self-reported HRQoL in (pre)frail cardiac surgery patients.
## Potential impact of study
To date, no high-quality study has prospectively examined the impact of perioperative nutrition supplementation in the higher risk, (pre)frail older adult undergoing cardiac surgery. The impact of the PROTECT-CS Study may provide a much-needed framework to direct future surgical practice and to inform ERP guidelines in the cardiac surgery patient. The PROTECT-CS Study is highly relevant and aims to improve the patient-centred and patient-driven perioperative approach for the optimisation of the (pre)frail older adult undergoing cardiac surgery. We endeavour to ensure patients do not just 'survive but thrive' after their heart surgery.
# Methods and analysis
The PROTECT-CS Study is a patient/researcher collaborative study, involving two Canadian centres. The study is a double-blinded, placebo randomised controlled trial (RCT) with blinded endpoint assessment and intentionto-treat analysis. Randomisation of the placebo and supplement is stratified by site and sex. Additionally, a third party will be randomising the patients for this study to ensure double blinding. At present, there is no formal process for engaging patients undergoing cardiac surgery to enhance their health during their waiting period. There is a significant knowledge gap regarding the efficacy of nutritional supplementation in preserving functional capacity and promoting enhanced recovery, among (pre) frail older adults undergoing cardiac surgery. Improved perioperative processes that address frailty and proper nutrition are of significant value to the patient-caregiver unit. This study is funded by the Heart and Stroke Foundation of Canada. We have used the Standard Protocol Items: Recommendations for Interventions Trials guidelines, the Sex and Gender Equity in Research guidelines and the Consolidated Standards of Reporting Trials guidelines in reporting this clinical trial. A description of this clinical trial is available on http:// ClinicalTrials. gov.
## Patient and public involvement
Patient engagement in research involves meaningful and active collaborations between patients and researchers throughout the different phases of a research project, including planning, study design, data collection, data analysis and knowledge translation (KT).We have recruited 10 patients and caregivers to collaborate as research partners on this study through membership on an advisory panel. Panel members were recruited from a database of previous cardiac surgery patients who agreed to be contacted for future research. We selectively chose panel members to reflect key demographic characteristics of our study's target population. We anticipate that we will collaborate with the advisory panel across the study's entire research cycle-from study planning through to the end of study KT (3 years)-predominantly through face-to-face meetings.outlines the different areas that the advisory panel may contribute to the study. These include providing input on study implementation refinement and procedures (eg, recruitment methods, approaches to maximise adherence to supplement-intake protocols, use of accessible language in recruitment and study materials, and usability of technology), relevance of outcomes to patient concerns, data analysis (eg, interpretation and contextualisation of findings) and KT (eg, alternate dissemination methods, use of accessible language). 50
## Study setting
The PROTECT-CS trial will be conducted and recruited in two sites across Canada: (1) Winnipeg, Manitoba, St Boniface Hospital; (2) Montreal, Quebec, Jewish General Hospital.
RECRUITMENT This study will recruit 150 patients over a 24-month recruitment period, starting January 2020, between both sites (ie, four per month for the primary site and two per month for the second site). We expect a total recruitment rate of approximately 20%-25% of potentially eligible patients based on previous research 30 (figure 2).
## Eligibility criteria screening
Initial screening for the study will include the CFS, the Short Performance Physical Battery (SPPB) and the Short
## Open access
Form-36 Physical Function (SF-36-PF) survey to measure frailty status in potential study patients. We have defined a CFS of greater than or equal to 3 (classified as 'managing well'), but not 7 (severely frail) or higher as an initial indicator of frailty for the purposes of our study's inclusion criteria. An SPPB score of less than or equal to 9 or a score of ≤60 on the SF-36-PF questionnaire for suitable participation. If the screening criteria are met, the research team will obtain signed consent, administer all baseline assessments and randomly allocate each patient to the appropriate research group (ie, protein or placebo). Patients who do not meet the frailty inclusion criteria are excluded based on specific criteria or refuse to be randomised into the main study will be given the opportunity to participate in the registry component of the study (figure 3).
## Outcomes and instrumentation
The primary outcome is a change in the SPPB score and the SF-36-PF score at 2 (short) and 6 (longer term) months compared with baseline. Due to the global pandemic (COVID-19), research initiatives that collected data through in-person meetings were suspended on two occasions (March 2020-August 2020, and November 2020-to be discussed (TBD)). As a result, we developed an alternative primary outcome (ie, SF-36-PF) to implement for the trial in order to accommodate the institutional directive to transition research to use remote/virtual data collection strategies. The SF-36-PF is an optimal outcome measure appropriate for our study cohort, as it is correlated with the SPPB (r=0.5-0.6) and is a surrogate measure of physical function.Secondary outcomes are adherence to supplementation, fat-free muscle mass (measured by a portable bioimpedance device); HRQoL measured by the EQ-5D-3L, EQ-Visual Analogue Scale (EQ-VAS)and the Older American Resource Scale (OARS): Activities of Daily Living (ADL) and Instrumental ADL (IADL) questionnaire 53 ; mood as measured by the Patient Health Questionnaire-9 (PHQ-9); current nutrition as measured by the Mini Nutrition Assessment (MNA) tool 55 ; anxiety as measured by the Cardiac Anxiety Questionnaire (CAQ); self-reported physical activity, exhaustion and nutrition as assessed by the Modified Fried Questionnaire 26 ; physical activity accumulation as measured by actical accelerometers, aerobic fitness as measured by the 6 Min Walk Test (6MWT); and a composite safety endpoint of all-cause mortality, injurious fall, acute kidney injury or readmission Open access for related events at 2 months provided by medical records. The primary and secondary outcomes will be assessed at baseline prior to cardiac surgery procedure and reassessed at 2 and 6 months post-hospital discharge. Additionally, the SPPB, SF-36-PF, fat-free muscle mass (assessed by bioelectrical impedance analysis (BIA)) and the MNA, and an assessment of nausea and vomiting symptoms will be reassessed while in-hospital prior to discharge if appropriate. It should be noted, the portable BIA device uses segmental multifrequency technology to measure muscle mass with an accuracy comparable with dual X-ray absorptiometry; and while the accuracy is inferior to that of CT and MRI, it is endorsed by consensus guidelines as a 'good portable alternative that is inexpensive, easy to use, and readily reproducible'.
## Blinding
Patients will be blinded to the treatment group allocation, as well as the research staff who will assess the fatfree muscle mass, the SPPB and other endpoints at each time point to avoid ascertainment bias.
## Study intervention study participants
All patients consenting to the study will be randomly allocated to either a protein or placebo group and will undergo the standard preoperative evaluation for suitability and planning for their cardiac surgery procedures. Patients will then be placed on the surgical waitlist. At the time of their cardiac surgery, both the placebo and protein group participants will ingest a carbohydrate beverage, PREcovery, (ie, 50 g of complex carbohydrates; aka CHO loading) 2-3 hours before surgery. The carbohydrate beverage consumed before surgery reduces insulin resistance and tissue glycosylation, improves postoperative glucose control, reduces nausea and vomiting, and also enhances the return of gastrointestinal function post-surgery. Protein group Patients randomised to the protein group will receive a leucine-rich protein supplement derived and provided by Enhanced Medical Nutrition (https:// emnhealth. com/). The product contains 25 g protein and 3 g leucine per serving (total caloric value: 160 Kcal) to be reconstituted and consumed two times per day for a minimum of 2 weeks pre-procedure, two times per day during postoperative in-hospital recovery (~5 -10 days) and two times per day for 8 weeks after the patient is discharged home. The supplement will be consumed approximately 1-2 hours after the morning meal and also 1-2 hours after lunch or before bedtime so as to supplement rather than replace meals. The protein supplement (ISOlution) is formulated to minimise appetite suppression. A research assistant, who is blinded to group allocation, will provide the patient's supply of supplements at the time of study consent (supplementation prior to cardiac surgery) and after discharge from the hospital (supplementation after
## Box 1 inclusion and exclusion criteria for the protect-cs trial
Inclusion criteria ► Patients aged 60 years or older, undergoing elective or nonemergent isolated CABG, aortic valve repair or replacement for moderate aortic stenosis or severe regurgitation, mitral valve repair or replacement for moderate mitral stenosis or severe regurgitation, or combined CABG/valve procedures. ► Patient with a Clinical Frailty Scale (CFS) from 3 (managing well) to 6 (moderately frail). 14 67 ► Patients with an estimated wait time of approximately 7 days or longer for elective surgery or 2 days or longer for non-emergent surgery.
## Exclusion criteria
► Decompensated or non-ambulatory class IV symptoms of angina, dyspnoea and claudication. ► Patients with a CFS of 7 or greater (severely frail to terminally ill); this will exclude less than 1% of the population on the elective cardiac surgery waitlist. 52 ► Creatinine clearance <30 mL/min/1.83 m 2 . ► Cirrhosis (Child-Pugh class B or greater). ► Allergy to milk proteins or other ingredients in the supplement. ► Inability to safely ingest beverages by mouth. ► Mild-to-severe cognitive impairment (ie, Montreal Cognitive Assessment <16). 67 ► An inability to speak/read in English or French. ► Emergent patients or non-emergent patients going for surgery within 48 hours of admission to hospital.
CABG, coronary artery bypass grafting; PROTECT-CS, PROTein to Enhance outComes of (pre)frail paTients undergoing Cardiac Surgery.
Open access cardiac surgery). In-hospital supplementation will be provided to the patient directly from research staff on a daily basis until discharged home. Phone calls will also be made to participants once per week to ensure compliance. Patients will be encouraged to perform light-tomoderate-intensity aerobic activity up to 30 min, 5 days per week as tolerated based on the 2016 edition of 'Living well with heart disease', published by the Heart and Stroke Foundation of Canada.This brochure outlines practical tips to safely begin a walking programme and gradually increase walking time to meet the goal of 150 min per week.
## Placebo group
Enrolled patients allocated to the placebo group will receive the same supplementation schedule followed by the protein group, as well as compliance verification; however, they will receive a placebo product with no supplemented protein (no additional nutritional benefit). This placebo product derived and provided by the Enhanced Medical Nutrition (https:// emnhealth. com/) will look exactly like the protein-rich supplement. A research assistant, who is blinded to group allocation, will provide the patient's supply of supplements at the time of study consent (supplementation prior to cardiac surgery) and after discharge from the hospital (supplementation after cardiac surgery). In-hospital supplementation will be provided to the patient directly from research staff on a daily basis until discharged home. Phone calls will also be made to participants once per week to ensure compliance. Additionally, patients will be encouraged to perform light-to-moderate-intensity aerobic activity up to 30 min, 5 days per week as tolerated based on the 2016 edition of 'Living well with heart disease', published by the Heart and Stroke Foundation of Canada.This brochure outlines practical tips to safely begin a walking programme and gradually increase walking time to meet the goal of 150 min per week.
## Registry group
Patients undergoing cardiac surgery who want to participate in the research but refuse to be randomised into the protein group or the placebo group, or those who want to participate in the research but are excluded based on specific study criteria, will be invited to participate in a registry group that will be followed for the purpose of documenting the effects of standard care. Registry participants will not be required to consume study supplements or CHO prior to surgery. Participants will be asked to complete basic demographic and quality of life questionnaires during their consent process and at 2 and 6 months after their cardiac surgery. Additionally, prior to their discharge from the hospital after their cardiac surgery, they will be asked about their postoperative experience specifically related to nausea, vomiting and nutrition while in hospital.
## Data collection and management
Participants will meet with the research staff for data collection at four separate points. This will include after consenting to their surgical procedure (baseline presurgery), during the in-hospital stay (postoperatively prior to discharge), and at the 2-month and 6-month time point post-surgery. The SPPB and SF-36-PF will be assessed to determine frailty status change at discharge, 2-month and 6-month post-surgery. The 6MWT, hand grip assessment and Fried Questionnaires will also be used to assess frailty in study participants at the above time points. Additionally, nutrition will be assessed by the MNA questionnaire at baseline, in-hospital (postoperatively prior to discharge), at 2-month and 6-month post-surgery. During the in-hospital research visit, all participants' nausea and vomiting symptoms will be assessed by research staff. Objective measurements of fat-free muscle mass will be obtained by using a BIA device which will be assessed at baseline, in-hospital (postoperatively prior to discharge), and during 2-month and 6-month post-surgery follow-up visits. Physical activity, sedentary behaviour and compliance to physical activity recommendations will be assessed by an accelerometer that participants will wear for a period of 7 days at baseline and at 2-month and 6-month post-surgery follow-up appointments. As COVID-19-related restrictions continue to persist, the potential of in-person visits is diminished based on institutional protocols. Therefore, assessment of physical parameters may not be possible for each research appointment. However, survey administration will continue for all research time points. The EQ-5D-3L, the EQ-VAS, and the OARS ADL and IADL will assess HRQoL at baseline, 2-month and 6-month post-surgery time points. The PHQ-9 and the CAQ will be administered at baseline, 2-month and 6-month postsurgery time points.
Sex and gender data will be collected at baseline and analysed to identify relationships as they occur within this cohort of cardiac surgery patients. We will capture this information through questionnaires as part of the survey package presented to research participants. Specifically, participants will be asked:what is your assigned sex? (forced choice: male, female or intersex) and (2) what is your gender identity? (man, woman or please specify (openended response)). It has been previously published that gender and sex differences are a consideration regarding risk for and outcomes of cardiovascular disease.Sample size Based on a previously observed distribution of SPPB scores in our elective cardiac surgery population, we expect this effect size will be approximately equivalent to either a 1-point change in the SPPB score or a 10-point change in SF-36-PF score. This magnitude of change in SPPB score is associated with a'substantial' improvement in mobility and quality of life,and survival.Furthermore, the scale of change for the SF-36-PF is sensitive to detect even a small and clinically meaningful change in physical functioning.Due to the global pandemic Open access , research initiatives that collected data through in-person meetings were suspended on two occasions (March 2020-August 2020, and November 2020-TBD). As a result, we developed an alternative primary outcome (ie, SF-36-PF) to implement for the trial in order to accommodate the institutional directive to transition research to use remote/virtual data collection strategies. The SF-36-PF is an optimal outcome measure appropriate for our study cohort, as it is correlated with the SPPB (r=0.5-0.6) and is a surrogate measure of physical function.A sample size of 150 patients (75 in each group) will allow for an overall drop-out rate of 15% (n=22) while maintaining a moderate Cohen's effect size (0.5) to detect the difference in change experienced between the two experimental groups at each follow-up time point with a two-tailed alpha of 0.05 and a power of 80% for the continuous primary outcomes. G*Power V.3.1 was used to calculate sample size for this study.
# Statistical methods
The primary analysis for this project will be intentionto-treat. A secondary a priori per-protocol analysis will be performed to assess the efficacy of the protein-rich nutritional supplement on the primary and secondary outcomes. Continuous outcome variables measured at multiple time points will be analysed using a repeated measure analysis of covariance (ANCOVA) considering site and sex as fixed effects. Additional covariates such as age, comorbidities and baseline cardiac function may be considered in the ANCOVA analysis if other imbalances remain following randomisation. Post hoc analysis that applies a Bonferroni correction to control the familywise error rate will be performed to identify how the primary and secondary outcomes change at each time point. Categorical outcome variables will be compared using a Χ 2 test or Fisher's exact test where appropriate. A sex-based and gender-based analysis will be performed by either stratifying the analysis or by including sex and gender as independent factors in the main analysis. Standardised mean differences will be calculated for several important perioperative characteristics to assess the covariate balance between study groups. Inverse probability of treatment weighting or multivariable regression techniques may also be considered if the ANCOVA assumptions are violated.The study biostatistician, who is experienced in RCT methodology, will oversee all statistical analyses.
## Data monitoring
A research coordinator will work out of the I H Asper Clinical Research Institute at the St Boniface Hospital/Research Centre, where the principal investigator and co-investigator are primarily appointed. Data management services, statistical and methodological support are to be housed within the Cardiac Sciences Program at the St Boniface Hospital. The research coordinator will assume responsibility for the data handling. The REDCap platform, housed on a virtual private network at the University of Manitoba, has been used for data collection to ensure safe storage of electronic data for both study sites.
## Ethics and dissemination informed consent
Patients will first consent to their surgical procedure before being considered for study participation. The patient will be approached and informed about the trial by the research assistant and provided with a copy of the patient information and consent form. Patients will be given adequate amount of time to consider their participation in the trial and will be given an opportunity to ask questions if needed. If the patient decides to participate in the study, they will be asked to provide written consent. All participants are free to withdraw from the study at any time, without any prejudice to future medical treatment.
## Trial monitoring and safety
The study collaborators are responsible for ensuring proposed milestones and deadlines are met. They are also responsible for study design, management, ethical conduct, analysis and dissemination of results. Safety data, including new hospitalisation, worsening heart symptoms and other adverse events will be captured and reportable to the study Data and Safety Monitoring Board (DSMB) as they occur. The DSMB is an independent group of experts that advises study investigators. They are responsible for a periodical evaluation of the study data (ie, every 6 months) for participant safety and study conduct, in addition to making recommendations concerning the modification and/or termination of the trial.
## Dissemination
Results will be distributed to an interdisciplinary team of cardiologists, cardiac surgeons, geriatricians, nurses, dieticians, rehabilitation providers, policymakers and patients by way of printed and electronic educational materials as well as oral presentation at regional meetings. Furthermore, we have formulated an integrated KT (iKT) team with a dedicated patient panel. Our iKT strategy will seek to adapt the intervention to local resources and expertise by including clinicians, dieticians and patient panels to develop a clinical pathway for the institution of treatment strategies with a mean of evaluation of effectiveness using the Plan-Do-Study-Act cycle 65 (a framework for developing, testing and implementing changes leading to improvement). We will validate our results in a larger multicentre trial leveraging ERAS Cardiac Surgery Society and CANCARE Society participating sites. In addition, we will organise focus group sessions with the patient-caregiver unit and our industry partner to seek their input to further refine and improve the leucine-rich protein-caloric supplement. In doing so, patients can transform the research process from one directed by investigators to one driven and informed by the needs of patients and their caregivers. Patient engagement is recognised as a necessary approach to the building of a |
Aberrant Whole-Brain Functional Connectivity and Intelligence Structure in Children with Primary Nocturnal Enuresis
Aim: To assess the potential relationship between intelligence structure abnormalities and whole-brain functional connectivity in children with primary nocturnal enuresis (PNE) with resting-state functional magnetic resonance imaging (fMRI) to provide insights into the association between these two seemingly unrelated conditions.
# Introduction
Primary nocturnal enuresis (PNE) affects up to 20% of young children (,5 years old) and nearly 2% of all young adults. PNE is characterized by involuntary voiding of the bladder during sleep beyond age five, which is the generally accepted age required for complete bladder control development and normal voiding habits during waking hours. PNE can cause significant psychosocial stress, potentially leading to more serious complications later in life.
PNE has been correlated with numerous genetic factors, including deficient arginine vasopressin (AVP) secretion, sleep awareness disorder, and bladder dysfunction. Current research also suggests that behavioral conditions, such as attention deficit hyperactivity disorder (ADHD), increase the risk for persistent PNE in children [bib_ref] Attentiondeficit/hyperactivity disorder (ADHD) as a risk factor for persistent nocturnal enuresis in..., Baeyens [/bib_ref]. Notably, antidiuretic treatments have also been shown to enhance short-term memory; Muller et al.
administered the antidiuretic hormone analog 1-desamino-8-Darginine vasopressin (DDAVP) to PNE patients and observed improvements [bib_ref] The effect of desmopressin on short-term memory in children with primary nocturnal..., Muller [/bib_ref]. Similarly, the incidence rate of enuresis in ADHD patients is higher than in healthy children [bib_ref] Attentiondeficit/hyperactivity disorder (ADHD) as a risk factor for persistent nocturnal enuresis in..., Baeyens [/bib_ref] [bib_ref] The impact of attention deficit hyperactivity disorders on brainstem dysfunction in nocturnal..., Baeyens [/bib_ref] [bib_ref] Behavioural problems and attention-deficit hyperactivity disorder in children with enuresis: a literature..., Baeyens [/bib_ref].
It is thought that PNE children experience cognitive deficits, which manifest as developmental delay and lower gross intelligence level in comparison with healthy control children [bib_ref] Prestimulation-induced startle modulation in attention-deficit hyperactivity disorder and nocturnal enuresis, Ornitz [/bib_ref]. A study by however, indicated that PNE children and agematched healthy children have similar IQ levels, but PNE children with intelligence structure abnormalities exhibit significantly different memory and attention levels.
Current evidence pertaining to the physical mechanism of cognitive deficits in PNE patients remains extremely limited, and the existing study results are not fully consistent. confirmed that the ratio of prepulse inhibition (PPI) signals, assessed using myoelectric tracing technology while simultaneously monitoring eye-wink reaction stimulation signals, was significantly decreased in PNE patients. This result indicated possible defects in the suppressive function of the brain stem associated with PNE [bib_ref] The impact of attention deficit hyperactivity disorders on brainstem dysfunction in nocturnal..., Baeyens [/bib_ref]. Conversely, a recent event-related fMRI study performed by Yu et al. [bib_ref] Evaluation of working memory impairment in children with primary nocturnal enuresis: evidence..., Yu [/bib_ref] indicated that PNE children exhibited an increased number of working memory deficits, a condition potentially associated with dysfunction of the left cerebellum,and their voxelbased morphometry (VBM) study revealed reduced gray matter density in the right dorsolateral prefrontal cortex (DLPFC) and left cerebellum of PNE children [bib_ref] Assessment of memory/ attention impairment in children with primary nocturnal enuresis: A..., Yu [/bib_ref]. In addition, Lei et al. [bib_ref] Altered brain activation during response inhibition in children with primary nocturnal enuresis:..., Lei [/bib_ref] reported that response inhibition in children with PNE is associated with a relative delay in maturation of prefrontal cortex circuitry that is known to suppress inappropriate responses, their resting-state fMRI study revealed abnormal spontaneous blood oxygen leveldependent (BOLD) activities in the left inferior frontal gyrus, medial frontal gyrus (posterior cingulate gyrus, middle temporal gyrus, left parietal lobe/inferior parietal lobule), and left midbrain of PNE children [bib_ref] Spontaneous brain activity changes in children with primary monosymptomatic nocturnal enuresis: a..., Lei [/bib_ref] , and their diffusion tensor imaging study revealed that children with PNE showed both a decrease in fractional anisotropy (FA) and an increase in mean diffusivity (MD) in multiple brain regions, including the frontal lobe, anterior cingulate cortex (ACC), insula, and particularly in the thalamus, compared to healthy children [bib_ref] Changes in the brain microstructure of children with primary monosymptomatic nocturnal enuresis:..., Lei [/bib_ref].
These studies indicate that multiple brain regions and circuits might be associated with the symptoms of enuresis and cognitive disorders in PNE children. However, the functional connectivity among these brain regions in PNE children was previously unknown.
Intrinsic low-frequency functional correlations measured by fMRI have been successfully applied in mapping brain systems in resting state subjects [bib_ref] The restless brain, Raichle [/bib_ref]. Low-frequency (50.08 Hz) fluctuations (LFF) of the BOLD signal in the resting state are considered to be related to spontaneous neuronal activity. These fluctuations have been used to identify functional connectivities between different brain regions, demonstrating that even remotely located regions have functional relationships as indicated by high temporal coherent LFF values [bib_ref] Functional connectivity in the motor cortex of resting human brain using echo-planar..., Biswal [/bib_ref] [bib_ref] Functional disconnectivity of the medial temporal lobe in Asperger's syndrome, Welchew [/bib_ref]. These findings imply the existence of neuronal coordination [bib_ref] Functional connectivity in single and multislice echoplanar imaging using resting-state fluctuations, Lowe [/bib_ref] [bib_ref] Interregional connectivity to primary motor cortex revealed using MRI resting state images, Xiong [/bib_ref]. In addition, several resting state fMRI studies have shown that LFF correlation patterns were altered in some pathological and behavioral conditions [bib_ref] Reductions in interhemispheric motor cortex functional connectivity after muscle fatigue, Peltier [/bib_ref] [bib_ref] Widespread functional disconnectivity in schizophrenia with resting-state functional magnetic resonance imaging, Liang [/bib_ref]. While fMRI techniques are potentially powerful techniques for mapping abnormal neural circuitry, recognition of functional relationships using fMRI is limited by the technique's sensitivity to factors that may not be directly involved in anatomical connectivity [bib_ref] Whole brain functional connectivity in the early blind, Liu [/bib_ref]. Despite these limitations, fMRI is a useful tool for investigating whole-brain functional connectivity with minimal patient discomfort, making the method ideal for assessing altered connectivity in children with PNE.
Whole-brain functional connectivity was explored in order to investigate specific alterations in functional connectivity in children with PNE. Based on previous methods proposed by Tzourio-Mazoyer, the brain was divided into 116 automated anatomical labeling (AAL) regions [bib_ref] Automated anatomical labeling of activations in SPM using a macroscopic anatomical parcellation..., Tzourio-Mazoyer [/bib_ref] , and correlations between each pair of these regions were analyzed in both PNE and normal control subjects. Significant differences in functional connectivity were determined by comparing the correlation coefficients of each pair of regions between the two groups. We also examined the relationships between altered functional connectivity and intelligence tests.
# Materials and methods
A total of 147 right-handed children including 75 PNE children (39 male, 36 female; aged 10.461.3 y) and 72 healthy control children (40 male, 32 female; aged 10.061.2 y) were assessed in the present study. Subjects in each group were matched for age, handedness, and primary school level (grades 4-6).
The study protocol was approved by and performed under the supervision of the ethics committee of the Shengjing Hospital of China Medical University (Shenyang 110004, China, no. 2012PS25K). All participants were informed about the study purposes and protocols, and each participant and their guardian provided written informed consent.
All children in the PNE group met the following inclusion criteria: urination under control during the daytime and involuntary urination during sleep$ twice a week for more than 6 months and normal blood and urine biochemistry, urine culture, and urine flowmetry. In addition, ultrasound examination of the urinary tract revealed no abnormal kidney or urinary tract defects, residual urine, or other urological or neurological disorders or abnormalities. All included subjects underwent routine MRI examination and had normal results.
Prior to inclusion, all children regularly attended the enuresis outpatient clinic, and no children were included that had previously been treated with any typical or atypical psychoactive drug. Children with a current or historical diagnosis of any neurological and psychiatric diseases according to the Diagnostic and Statistical Manual of Mental Disorders published by the American Psychiatric Association (DSM-IV), especially ADHD, were excluded from the present study.
## Intelligence evaluation and data analysis
Intelligence testing was performed at our hospital using the China-Wechsler Intelligence Scale for Children (C-WISC), revised by Gong et al.All children completed 11 individual tests, including 6 speech tests and 5 manipulation tests. The speech tests consisted of tests for Information (I, answering common questions), Comprehension (C, answering the best action under a certain circumstance), Sorting (S, summarize the common aspects of each pair of words in a phrase), Arithmetic (A, solve arithmetic questions with mental calculation), Digit symbol (D, recite numbers in sequential and reverse orders), and Vocabulary (V, sort vocabulary according to the degree of difficulty). The manipulation tests included Picture Completion (PC, point out lack of stroke and the name in the figure), Picture Arrangement (PA, arrange randomly sorted figures in a meaningful story), Block Pattern (BP, decompose cubic blocks into the pattern the experimenters presented), Object Assembly (OA, compose four disassembled pieces into complete figures within a certain time), and Coding (Cd, numbers from 1 to 9 were given a specific mark, and the subjects were required to fill in the blank spaces under each number with the corresponding marks without skipping any).
During intelligence testing, the raw score from each subtest was first transformed into a scaled score, and then the sum of scaled scores-V (I+C+S+A+D+V), the sum of scaled scores-P (PC+PA+B-P+OA+Cd), and the total sum of scaled scores were calculated. Finally, the scaled scores were transformed into Verbal intelligence quotient (VIQ), procedure intelligence quotient (PIQ), and full intelligence quotient (FIQ).
The following IQ factors were also calculated accordingly: verbal comprehension factor (VC = I+V+C+S), perceptual organization factor (PO = PC+PA+BP+OA), and memory/caution factor (M/C = A+D+Cd). All intelligence tests were administered by trained professionals who were blinded to the study grouping.
## Fmri data acquisition
All studies were performed using a 3.0-T scanner (Intera Achieva; Philips Medical Systems, Best, Netherlands) with an 8-
# Statistical analysis
Data obtained from intelligence testing was analyzed using SPSS 17.0 (IBM, USA) software package. FIQ, VIQ, PIQ, VC, PO, and M/C were expressed as mean6SD. Student's t-tests or Mann-Whitney U tests were used to compare the two independent groups, depending on if the data were normally or non-normally distributed, respectively. Because we employed Bonferroni corrections for multiple comparisons, only values of P,0.0083 were considered statistically significant.
All fMRI data preprocessing was performed using the SPM8 software package (Welcome Department of Cognitive Neurology, London, UK) on the MATLAB (Mathworks, USA) platform. Images were realigned, and differences in slice acquisition time were corrected using temporal realignment to the middle slice. Data were discarded if fMRI head motion was .2 mm or 3u. Images were then resampled to 3 mm 6 3 mm The resultant data were further filtered using a bandpass temporal filter (0.01-0.1 Hz) to reduce the effects of low frequency drift and high frequency physiological noise using the REST software package; nuisance signals such as global mean and head motion parameters were regressed out [bib_ref] REST: a toolkit for resting-state functional magnetic resonance imaging data processing, Song [/bib_ref].
Whole-brain functional connectivity was analyzed using the method developed by Liu [bib_ref] Whole brain functional connectivity in the early blind, Liu [/bib_ref]. The registered fMRI data were segmented into 116 regions using the AAL template described by Tzourio-Mazoyer et al. [bib_ref] Automated anatomical labeling of activations in SPM using a macroscopic anatomical parcellation..., Tzourio-Mazoyer [/bib_ref] , which was used in some previous studies [bib_ref] Widespread functional disconnectivity in schizophrenia with resting-state functional magnetic resonance imaging, Liang [/bib_ref] [bib_ref] Whole brain functional connectivity in the early blind, Liu [/bib_ref] [bib_ref] A resilient, low-frequency, small-world human brain functional network with highly connected association..., Achard [/bib_ref]. This parcellation divided the cerebra into 90 regions (45 in each hemisphere) and the cerebella into 26 regions (9 in each cerebellar hemisphere and 8 in the vermis). [fig_ref] Table 1: Abbreviations and MNI coordinates of AAL [/fig_ref] lists their abbreviations and the MNI coordinates of the center of each region.
Regional mean time series were estimated by first averaging the time series of all voxels by region. Then, Pearson correlation analysis was conducted for each pair of regions in each subject. After correlation coefficients were computed, Fisher's r-to-z transformation was applied to the correlation coefficients. Individual z-scores were compared with two-tailed t-tests to determine the significance of functional connectivities between the two groups. The false discovery rate (FDR) approach was applied to identify the restriction threshold capable of reducing the proportion of type I errors to ,0.05. Pearson correlation analysis was performed between the FIQ, VIQ, PIQ, VC, PO, and M/C of all PNE children, with z-scores representing connectivities with significant differences between the PNE and control groups.
# Results
Based on observed head motion .2 mm or 3u in 8 children of the PNE group and 6 children of the control group, C-WISC and fMRI data were discarded for 14 subjects. The fMRI data of the remaining 67 children in the PNE group (M/F = 39:28; average age 10.561.2 y; 4.8760.65 y of education) and 66 children in the control group (M/F = 37:29; 10.161.1 y; 4.7660.68 y of education) were included in the study, resulting in a total of 133 subjects. There were no significant differences in age or years of education (P = 0.1275 and P = 0.3410, respectively).
## Intelligence testing
FIQ, VIQ, and PIQ values were normal in the PNE group, showing no significant variation compared with normal controls (P.0.0083). The M/C factor, however, was significantly different between the two groups (P = 0.0060), but no significant difference was observed for VC or PO factors (P.0.0083). The intelligence results are shown in detail in .
## Functional connectivity differences
Similar functional connectivity patterns were observed in both groups. The majority of strong functional connectivities (large zscores) were observed between symmetrical interhemispheric regions, within a lobe, or adjacent to lobular regions [fig_ref] Figure 1: Mean z-score matrices for the control and PNE groups [/fig_ref]. A total of 12 significantly different functional connectivities were identified in the cerebellum, frontal lobe, and thalamus between the PNE and control groups at an FDR corrected threshold of P,0.05. We observed 10 decreased and 2 increased functional connectivities in the PNE group. Aberrant functional connectivities were visualized with the BrainNet Viewer (http://www.nitrc. org/projects/bnv/) [fig_ref] Figure 2: Differences in functional connectivity [/fig_ref] ; [fig_ref] Table 3: Differences in functional connectivity [/fig_ref]. The M/C factor also showed a significant correlation with the z-scores representing connectivity between the Cerebellum_Crus1_L and Frontal_Mid_R (r 2 = 0.787, P,0.001) [fig_ref] Figure 3: Relationship between M/C factor and connectivity z-scores [/fig_ref].
# Discussion
## C-wisc testing reveals potential attention deficits in pne
Intelligence testing was performed using the C-WISC method, which assesses parameters representative of speech, operation, and overall intelligence level. The current findings demonstrate that the overall intelligence level and verbal and manipulation abilities of PNE children were within normal ranges and were not significantly different from controls, indicating that PNE children have normal intelligence levels. Despite overall normal intelligence levels, children with PNE exhibited a markedly different M/C factor. The intelligence test revealed that PNE children had intelligence structure abnormality and attention deficits.
## Atypical cerebello-thalamo-frontal functional connectivity in children with pne
The cerebello-thalamo-frontal pathway, including the left and right DLPFC, thalamus, and cerebellar hemisphere, were the predominant sites of altered functional connectivity. The cerebellum has been shown to have strong functional connectivities involved in working memory control and execution. Thus, this region has been associated with cerebellum dysfunctions related with attention deficit disorders in ADHD children using fMRI methodologies [bib_ref] Diffusion tensor imaging (DTI) and tractography of the cerebellar projections to prefrontal..., Jissendi [/bib_ref] [bib_ref] Cortical hubs revealed by intrinsic functional connectivity: mapping, assessment of stability, and..., Buckner [/bib_ref] [bib_ref] Segregated fronto-cerebellar circuits revealed by intrinsic functional connectivity, Krienen [/bib_ref] [bib_ref] Cognitive and motor loops of the human cerebro-cerebellar system, Salmi [/bib_ref]. Our data suggest that aberrant connectivity of the cerebello-thalamo-frontal circuit may also be involved in the etiology underlying the initial development of attention deficit disorders in PNE children.
The thalamus is involved in normal attention functions; however, it also plays an important role in sleep cycle regulation and relaying sensory afferent information from the bladder. In the currently accepted model for bladder control during urine storage, this information is transmitted through the periaqueductal gray (PAG), to the ACC, the insula, and the lateral prefrontal cortex (LPFC) [bib_ref] A decade of functional brain imaging applied to bladder control, Fowler [/bib_ref]. The current findings suggest that altered thalamic connectivity may result in an inability to wake during sleep in response to the need to void, and this is compounded by abnormally large urine storage volume in the bladders of children with PNE [bib_ref] The role of the thalamus in sleep, pineal melatonin production, and circadian..., Jan [/bib_ref].
The ACC is involved in a wide range of cerebral functions, including cognitive function control, mediation of bodily arousal states, and interoceptive awareness [bib_ref] Neural systems supporting interoceptive awareness, Critchley [/bib_ref]. The ACC is also involved in attention and introspection, which allows an individual to develop a sense of conscious awareness that the bladder is full. It also plays a role in executive control, a process involved in voiding the bladder in appropriate places and times [bib_ref] Neurophysiology of the lower urinary tract, Beckel [/bib_ref]. The current study suggests that connectivity between the right ACC and left cerebellum (crus I) is increased in children with PNE, which may indicate functional changes in the cerebello-thalamo-frontal pathway that also affect other ACC functionalities. For example, the development of attention deficit disorders and nocturnal enuresis may involve similar mechanisms [bib_ref] Changes in the brain microstructure of children with primary monosymptomatic nocturnal enuresis:..., Lei [/bib_ref]. While the current study shows persuasive initial evidence for this hypothesis, the connectivities between the three distinguishable functional subregions of the ACC (dorsal, ventral, and circumcallosal) must be further examined in order to define the mechanistic role of the ACC in both PNE and attention deficit pathogeneses.
Children with PNE also exhibited a pattern of increased connectivity in the right inferior frontal gyrus. It has been suggested that neural reorganization may compensate for deficient regions during manipulation and memory maintenance in ADHD patients [bib_ref] Efficiency of the prefrontal cortex during working memory in attention-deficit/hyperactivity disorder, Sheridan [/bib_ref]. Thus, it follows that attention dysfunctions in children with PNE may also be correlated with abnormal cerebello-thalamo-frontal functional connectivity. The presence of attention deficit disorders would play a causative role in the failure to optimize the sense transduction pathways of the brain associated with bladder filling. Thus, inhibition of functional defects and subsequent induced functional disorder occur, ultimately resulting in PNE symptoms.
The functional connectivity abnormalities described here provide evidence initial to support these variations in children with PNE; however, the current study is limited by the relatively small, homogeneous sample. To confirm these results and their specific mechanism of action, future tests on large, varied cohorts will be required. The automated anatomical labeling atlas applied in the current study includes 116 regions, covering the entire cerebrum and cerebellum but excluding the brainstem. Thus, the current findings cannot rule out the involvement of this region in the pathogenesis of PNE; previous studies have suggested deficits in brainstem inhibition function in PNE children. Thus, brainstem functional connectivity should be investigated [bib_ref] The impact of attention deficit hyperactivity disorders on brainstem dysfunction in nocturnal..., Baeyens [/bib_ref]. We should also note that subjects with PNE and ADHD were excluded to avoid the possible confounding effects of attention impairment caused by ADHD on experimental results. This exclusion may introduce certain elements of selection bias that must be considered when reviewing these results and designing future studies.
# Conclusions
We identified and attention deficit-based intelligence structure abnormality associated with variation in functional connectivities of the cerebello-thalamo-frontal region of the brain in children with PNE. This association between attention and PNE may provide valuable information to researchers interested in developing novel therapeutic treatments for persistent PNE in children and young adults.
[fig] Figure 1: Mean z-score matrices for the control and PNE groups. Each figure shows a 1166116 square matrix in which the x-and y-axes correspond to AAL regions. Each entry indicates the mean z-score representing functional connectivity between each pair of regions in the brain. doi:10.1371/journal.pone.0051924.g001 [/fig]
[fig] Figure 2: Differences in functional connectivity. Edges of each brain region (nodes) with indicated corresponding regions of decreased (blue) and increased (pink) functional connectivity in the PNE group. doi:10.1371/journal.pone.0051924.g002 [/fig]
[fig] Figure 3: Relationship between M/C factor and connectivity z-scores. Relationship between M/C factor and connectivity z-scores between Cerebellum_Crus1_L and Frontal_Mid_R in PNE children (r 2 = 0.787, P,0.001). doi:10.1371/journal.pone.0051924.g003 [/fig]
[table] Table 1: Abbreviations and MNI coordinates of AAL. and to stabilize the magnetic field. Thus, each fMRI scan was performed over 496 s. In order to stabilize patients during fMRI scanning, a thick ear cushion was applied to fix head placement and to eliminate the effects of constructed defects caused by head movements. [/table]
[table] Table 3: Differences in functional connectivity. [/table]
|
Disulfide mapping the voltage-sensing mechanism of a voltage-dependent potassium channel
Voltage-dependent potassium (Kv) channels allow for the selective permeability of potassium ions in a membrane potential dependent manner, playing crucial roles in neurotransmission and muscle contraction. Kv channel is a tetramer, in which each subunit possesses a voltage-sensing domain (VSD) and a pore domain (PD). Although several lines of evidence indicated that membrane depolarization is sensed as the movement of helix S4 of the VSD, the detailed voltage-sensing mechanism remained elusive, due to the difficulty of structural analyses at resting potential. In this study, we conducted a comprehensive disulfide locking analysis of the VSD using 36 double Cys mutants, in order to identify the proximal residue pairs of the VSD in the presence or absence of a membrane potential. An intramolecular SS-bond was formed between 6 Cys pairs under both polarized and depolarized environment, and one pair only under depolarized environment. The multiple conformations captured by the SS-bond can be divided by two states, up and down, where S4 lies on the extracellular and intracellular sides of the membrane, respectively, with axial rotation of 180°. The transition between these two states is caused by the S4 translocation of 12 Å, enabling allosteric regulation of the gating at the PD.
Voltage-dependent potassium (Kv) channels are membrane proteins that are selectively permeable to potassium ions (K + ) in a membrane potential dependent manner. Macroscopic current of the Kv channels can be described as a cycle of four stages 1 : (i) no K + current is observed at resting potential (the resting state), (ii) maximum peak current is observed upon depolarization (the activated state), (iii) the current exponentially decays to the constant K + current (the inactivated state), and (iv) the current stops upon repolarization (return to (i)). Through this functional cycle, Kv channels control the action potential, which plays crucial roles in neurotransmission and muscle contraction.
Kv channels are tetrameric proteins, in which each subunit possesses six transmembrane helices, S1-S6. Each subunit consists of a voltage-sensing domain (VSD) that includes S1-S4, and a pore domain (PD) comprised of S5-S6. The center of the tetramer has a K + conducting pore that possesses two gates: a crossing of four S6 helices (helix bundle crossing, HBC) at the intracellular exit of the pore, and a K + selectivity filter (SF) located on the extracellular side of the pore. While the S4 helix of the VSD resides on the intracellular side of the membrane in the resting state, it moves to the extracellular side upon depolarization 2 . This voltage dependent conformational change of the VSD allosterically opens the HBC gate in the PD, leading to channel activation 2 .
To date, structural information at an atomic resolution has been reported for the Kv channels in the absence of a membrane potential. These include the crystal structures of rat Kv1.2 [bib_ref] Structure of the full-length Shaker potassium channel Kv1.2 by normal-mode-based X-ray crystallographic..., Chen [/bib_ref] [bib_ref] Crystal structure of a mammalian voltage-dependent Shaker family K+ channel, Long [/bib_ref] , the chimera of rat Kv1.2 and Kv2.1 [bib_ref] Atomic structure of a voltage-dependent K+ channel in a lipid membrane-like environment, Long [/bib_ref] , and the Kv channels from Aeropyrum pernix (KvAP) [bib_ref] X-ray structure of a voltage-dependent K+ channel, Jiang [/bib_ref] , as well as a model structure of KvAP 7 . These structures are assumed to represent the activated state, where the S4 helix lies on the extracellular side and the HBC is open. The conformation of the VSD of KvAP is essentially the same as that of the isolated VSD in crystal [bib_ref] X-ray structure of a voltage-dependent K+ channel, Jiang [/bib_ref] and in solution [bib_ref] Solution structure and phospholipid interactions of the isolated voltage-sensor domain from KvAP, Butterwick [/bib_ref] , which is consistent with previous electron paramagnetic resonance study showing that the isolated VSD from KvAP in the lipid bilayer retains a very similar conformation to that in the full length KvAP 9 .
Scientific RepoRts | 6:37303 | DOI: [bib_ref] The principle of gating charge movement in a voltage-dependent K+ channel, Jiang [/bib_ref].1038/srep37303
However, the structure of the resting state and the voltage-dependent conformational changes have not been determined, because of the difficulty in analyzing the structure at resting potential. Thus, the gating mechanism of the HBC remains elusive.
Several biochemical analyses have revealed the voltage dependent conformational change of VSD. The distance between the membrane surface and each residue of VSD was analyzed by avidin binding to a biotin modified Kv channel, which provided the insights that S3 and S4 move vertical to the membrane, depending on the membrane potential [bib_ref] The principle of gating charge movement in a voltage-dependent K+ channel, Jiang [/bib_ref] [bib_ref] Calibrated measurement of gating-charge arginine displacement in the KvAP voltage-dependent K+ channel, Ruta [/bib_ref]. Disulfide locking analyses [bib_ref] Two atomic constraints unambiguously position the S4 segment relative to S1 and..., Campos [/bib_ref] [bib_ref] Sensing voltage across lipid membranes, Swartz [/bib_ref] [bib_ref] Structural basis for gating charge movement in the voltage sensor of a..., Yarov-Yarovoy [/bib_ref] and metal ion bridge analysis [bib_ref] Tracking a complete voltage-sensor cycle with metal-ion bridges, Henrion [/bib_ref] showed the movement distances and the rotation angles (30-180°) of S4 upon conformational change. EPR analysis showed that the VSD of KvAP changes its conformation, depending on the lipid environment [bib_ref] Structural basis of lipid-driven conformational transitions in the KvAP voltagesensing domain, Li [/bib_ref].
However, there are several problems with these analyses: (1) one or more mutations were introduced to the voltage-sensing Arg residue in S4 and/or its interacting counterpart residues, Asp or Glu, which may modify the voltage-sensing properties of the VSD 12-14 , (2) the conformational change of VSD was detected indirectly, through the change in the K + current of Kv channels [bib_ref] The principle of gating charge movement in a voltage-dependent K+ channel, Jiang [/bib_ref] [bib_ref] Calibrated measurement of gating-charge arginine displacement in the KvAP voltage-dependent K+ channel, Ruta [/bib_ref] [bib_ref] Two atomic constraints unambiguously position the S4 segment relative to S1 and..., Campos [/bib_ref] [bib_ref] Sensing voltage across lipid membranes, Swartz [/bib_ref] [bib_ref] Structural basis for gating charge movement in the voltage sensor of a..., Yarov-Yarovoy [/bib_ref] [bib_ref] Tracking a complete voltage-sensor cycle with metal-ion bridges, Henrion [/bib_ref] , (3) the conformational change of VSD was deduced from the results of studying a limited number of mutants [bib_ref] Two atomic constraints unambiguously position the S4 segment relative to S1 and..., Campos [/bib_ref] [bib_ref] Sensing voltage across lipid membranes, Swartz [/bib_ref] , and (4) the events were investigated during the change of the membrane potential, rather than at a specific potential [bib_ref] The principle of gating charge movement in a voltage-dependent K+ channel, Jiang [/bib_ref] [bib_ref] Calibrated measurement of gating-charge arginine displacement in the KvAP voltage-dependent K+ channel, Ruta [/bib_ref] [bib_ref] Two atomic constraints unambiguously position the S4 segment relative to S1 and..., Campos [/bib_ref] [bib_ref] Sensing voltage across lipid membranes, Swartz [/bib_ref] [bib_ref] Structural basis for gating charge movement in the voltage sensor of a..., Yarov-Yarovoy [/bib_ref] [bib_ref] Tracking a complete voltage-sensor cycle with metal-ion bridges, Henrion [/bib_ref].
In this study, we conducted a comprehensive disulfide locking (SS-locking) analysis of VSD, using 36 double Cys mutants that possess mutations at residues excluding Arg, Asp, and Glu. We identified the proximal Cys residues of VSD by detecting the SS-bond formation in the presence and absence of membrane potential, suggesting the existence of a conformational equilibrium at each membrane potential. Conformational changes of the VSD required for the voltage-sensing is described, based on the results of these analyses.
# Results
In order to analyze the voltage-dependent conformational change of the VSD in liposomes, we first observed the membrane potential of the liposomes reconstituted with VSD. Next, we characterized the voltage dependence of the conformations of the prepared VSD, by observing the fluorescence change of a chemically attached fluorescence moiety. We developed a method for detecting intramolecular SS-bond formation between proximal Cys residues in the double Cys mutant of VSD (SS-locking analysis). This method was applied to 36 double Cys mutants of VSD in liposomes, in the presence and absence of a membrane potential.
Measurement of the membrane potential on VSD-reconstituted liposomes. Membrane potential on the VSD-reconstituted liposomes was measured based on the fluorescence of 3-(4-(2-(6-(dibutylamino)-2-naphthyl)-trans-ethenyl) pyridinium) propanesulfonate (di-4-ANEPPS) [fig_ref] Figure 1: Measurement of membrane potentials of liposomes [/fig_ref] , which has reportedly been used for measuring the membrane potentials of empty liposomes and cells [bib_ref] A naphthyl analog of the aminostyryl pyridinium class of potentiometric membrane dyes..., Loew [/bib_ref].
Di-4-ANEPPS is a membrane potential-sensitive fluorescent dye. Its hydrophobic site is inserted into the membrane while its hydrophilic site is exposed to the aqueous solution. The change in its fluorescence intensity is proportional to the change in the membrane potential. A change in the fluorescence intensity of 9.6% was reported for a 100 mV change in the membrane potential in empty EggPC liposomes, with excitation at 530 nm and emission at 610 nm [bib_ref] A naphthyl analog of the aminostyryl pyridinium class of potentiometric membrane dyes..., Loew [/bib_ref]. Furthermore, in a variety of cells, 7-10% change in the fluorescence intensity was observed when the membrane potential was changed by 100 mV [bib_ref] A naphthyl analog of the aminostyryl pyridinium class of potentiometric membrane dyes..., Loew [/bib_ref].
In this study, the membrane potential was formed by selective K + influx from high K + buffer (referred to as membrane potential buffer) to the inside of the liposomes with 0.10 mM K + . We observed the fluorescence time course for the membrane potential buffer in the presence of di-4-ANEPPS, at various K + concentrations that provide the desired theoretical membrane potential. The 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoethanolamine (POPE)/1-palmitoyl-2-oleoyl-sn-glycero-3-phospho-(1′ -rac-glycerol) (POPG) liposome solution was added to the cuvette at 0 s, and then 1 μ L of 40 μ M valinomycin, which is a K + selective ionophore, was added at 60 s, to form a membrane potential [bib_ref] Proton conduction and bafilomycin binding by the V0 domain of the coated..., Zhang [/bib_ref] [bib_ref] Functional reconstitution of purified human Hv1 H+ channels, Lee [/bib_ref].
When the theoretical membrane potentials were 0 mV, − 60 mV, − 120 mV, and − 180 mV, the observed reductions in fluorescence intensity of empty POPE/POPG liposomes were 0.90, 4.8, 10.8, and 16.7%, respectively [fig_ref] Figure 1: Measurement of membrane potentials of liposomes [/fig_ref]. These results indicated that the intensity reductions were proportional to the theoretical values of the membrane potential, and the observed rate of reduction was 0.09% per − 1 mV [fig_ref] Figure 1: Measurement of membrane potentials of liposomes [/fig_ref]. This relationship between the reduction of di-4-ANEPPS fluorescence intensity and the membrane potential was utilized to estimate the membrane potential of numerous VSD-reconstituted POPE/POPG liposomes. [fig_ref] Figure 1: Measurement of membrane potentials of liposomes [/fig_ref] shows an example of the estimation of the membrane potential of liposomes reconstituted with a VSD mutant, V42C/ I131C. While the theoretical potential was − 187 mV, the potentials estimated from the fluorescence intensity reduction were − 140 and − 124 mV, just after the addition of valinomycin and at 90 s after the addition of valinomycin, respectively. The difference between the theoretical value and the experimental value is probably due to the incomplete K + shield of the liposome membrane reconstituted with the VSD mutant.
Voltage-dependent conformational change of VSD. In order to investigate the voltage-dependent conformational change of VSD, we used the V119C mutant, in which the Cys119 residue was chemically modified by a fluorescent dye, monobromobimane (mBBr), as previously reported [bib_ref] Structural basis for the inhibition of voltage-dependent K+ channel by gating modifier..., Ozawa [/bib_ref] [fig_ref] Figure 2: Fluorescence analysis of mBBr to examine voltage-dependent conformational changes of VSD [/fig_ref]. The conformational change can be detected by the change in the intensity of the fluorescence from mBBr, since the intensity decreases when the fluorophore moves to a more polar environment [bib_ref] The effect of ligand efficacy on the formation and stability of a..., Yao [/bib_ref]. The values of membrane potential were estimated for each sample by di-4-ANEPPS analysis.
We prepared POPE/POPG liposomes reconstituted with mBBr-modified VSD. The membrane potential dependent fluorescence changes are shown in [fig_ref] Figure 2: Fluorescence analysis of mBBr to examine voltage-dependent conformational changes of VSD [/fig_ref] -d. Upon membrane potential formation, instantaneous reduction in fluorescence intensity, followed by time dependent recovery was observed [fig_ref] Figure 2: Fluorescence analysis of mBBr to examine voltage-dependent conformational changes of VSD [/fig_ref]. The instantaneous intensity reduction indicates that the fluorophore at Cys119 moves to a more polar environment. The reduction ratios of the fluorescence intensity depended on the membrane potential, indicating that the VSD on the liposome experiences a voltage-dependent conformational change. Sigmoidal curve fitting of the fluorescence intensity reductions against the membrane potential revealed that the activation potential (V 1/2 ) of VSD was − 104 mV [fig_ref] Figure 2: Fluorescence analysis of mBBr to examine voltage-dependent conformational changes of VSD [/fig_ref]. It should be noted that time dependent recovery of fluorescence intensity [fig_ref] Figure 2: Fluorescence analysis of mBBr to examine voltage-dependent conformational changes of VSD [/fig_ref] also reflects the conformational change of the VSD, corresponding to the decay of membrane potential [fig_ref] Figure 1: Measurement of membrane potentials of liposomes [/fig_ref]. For example, in [fig_ref] Figure 2: Fluorescence analysis of mBBr to examine voltage-dependent conformational changes of VSD [/fig_ref] , the intensity reduction ratios were 13.7% and 6.6% at 100 s (upon the formation of the membrane potential) and 300 s (200 s after the formation of the membrane potential), respectively. The membrane potentials estimated by di-4-ANEPPS analysis were − 147 mV upon the formation of the membrane potential and − 116 mV in 200 s (data not shown), respectively. The relationship between the membrane potential of − 116 mV and the intensity reduction of the mBBr fluorescence by 6.6% at 200 s after the membrane potential formation [fig_ref] Figure 2: Fluorescence analysis of mBBr to examine voltage-dependent conformational changes of VSD [/fig_ref] , is consistent with the sigmoidal curve shown in [fig_ref] Figure 2: Fluorescence analysis of mBBr to examine voltage-dependent conformational changes of VSD [/fig_ref] , which was constructed from a number of experiments with different membrane potential values upon the membrane potential formation.
The activation potential, which is lower by about 40 mV than that obtained from the G-V curve of full length KvAP [bib_ref] A gating model for the archeal voltage-dependent K(+ ) channel KvAP in..., Schmidt [/bib_ref] , is consistent with the reports that the activation potential of VSD is lower than that of Kv channels, in analyses of KvAP and other Kv channels [bib_ref] The size of gating charge in wild-type and mutant Shaker potassium channels, Schoppa [/bib_ref] [bib_ref] Voltage-Sensing Residues in the S2 and S4 Segments of the Shaker K+..., Seoh [/bib_ref] [bib_ref] Lipid-dependent gating of a voltage-gated potassium channel, Zheng [/bib_ref]. Therefore, the voltage-dependent conformational change in the VSDs used in this study is likely to be essentially the same as that in the full length KvAP.
## Detection of an intramolecular ss-bond in vsd double cys mutants.
The free thiol groups in a protein can be modified by maleimide polyethylene glycol (mal-PEG). The modification can be detected as a mobility shift in the SDS-PAGE analysis [bib_ref] Mapping the membrane-aqueous border for the voltage-sensing domain of a potassium channel, Neale [/bib_ref] [bib_ref] Pegylation: A Method for Assessing Topological Accessibilities in Kv1.3, Lu [/bib_ref]. In order to estimate the intramolecular SS-bond formation in the VSD double Cys mutants by detecting free thiol groups, VSD double Cys mutants, which were pre-denatured in SDS and urea, were treated with mal-PEG with a molecular weight of ca. 2,000 [fig_ref] Figure 3: Detection of an intramolecular SS-bond [/fig_ref].
First, a double Cys mutant (V42C/I130C) in n-decyl-β -D-maltopyranoside (DM) detergent micelles was denatured after reduction or oxidation, and then treated with mal-PEG. For the reduced VSD double Cys mutant, the mal-PEG treatment caused mobility shift in SDS-PAGE, and the band without the PEG modification (PEGylation) disappeared [fig_ref] Figure 3: Detection of an intramolecular SS-bond [/fig_ref] , Reductive), indicating that the two thiol groups were PEGylated and no SS-bond was formed between the two Cys residues. However, for the oxidized VSD double Cys mutant, a band without PEGylation was observed [fig_ref] Figure 3: Detection of an intramolecular SS-bond [/fig_ref] , indicating that the two Cys residues formed an SS-bond that precluded PEGylation. The SS-bond formation between C42 and C130 caused slight mobility shift in SDS-PAGE, compared to that of the SH form, since the molecular shape of VSD in the SDS micelles is modified by the intramolecular SS-bond. However, the difference in the mobility between the SS and SH forms depends on the Cys positions, and in some cases, no significant difference in the mobility was observed. PEGylation used in this study helped the detection of SH form clearly. These results also indicated that the reaction of the mal-PEG modification proceeds faster than the intramolecular SS-bond formation between the free thiol groups in the denaturation buffer.
Next, we analyzed the intramolecular SS-bond formation of the VSD double Cys mutants reconstituted in liposomes, in the presence or absence of a membrane potential. Since the negative membrane potential was applied to the outside of the liposomes [fig_ref] Figure 1: Measurement of membrane potentials of liposomes [/fig_ref] , the N and C termini of the VSD needed to be on the outer surface of the liposomes. In order to analyze the VSD that is exposed to membrane potential in the right direction, we reconstituted the maltose binding protein (MBP)-fused VSD in liposomes and digested the MBP outside the liposomes by human rhinovirus (HRV) 3C protease. This procedure enables the discrimination of the VSD with an MBP-tag outside of the liposomes from that with the tag inside, in SDS-PAGE, since the latter possesses higher molecular weight because of the uncleaved MBP [fig_ref] Figure 2: Fluorescence analysis of mBBr to examine voltage-dependent conformational changes of VSD [/fig_ref]. The digested VSD was analyzed by SS-locking after negative membrane potential was created outside the liposomes by K + influx. We applied this method to the 36 VSD double Cys mutants [fig_ref] Figure 4: Locations of Cys mutation on VSD [/fig_ref].
## Voltage-dependent ss-locking analysis of 36 vsd double cys mutants.
In order to comprehensively search for the residue pairs that are proximate to each other at resting potential, we prepared 36 double Cys mutants, in which mutations were introduced at one of 4 residues on S1 (residues located on the S4 side; S32, L36, V39, or V42), and one of 9 residues on S4 (from L121 to I131 except for Arg; L121, L122, F124, L125, I127, L128, L129, I130, or I131) [fig_ref] Figure 4: Locations of Cys mutation on VSD [/fig_ref]. It should be noted that the conformation of S4 and not that of S1, is assumed to change drastically depending on the membrane potential.
Liposomes reconstituted with a VSD double Cys mutant were diluted with the membrane potential buffer containing valinomycin, in which the K + concentrations were the same as and 1000-fold higher than that on the inside of the liposomes, to create theoretical membrane potentials of 0 and − 187 mV, respectively. The values of membrane potential were estimated for each sample by the di-4-ANEPPS analysis. These VSD double Cys mutants were incubated in the presence of 30 μ M Cu 2+ -o-phenanthroline (CuP) for 1 min at room temperature (25 °C), to allow intramolecular SS-bond formation, between two Cys residues that are in close proximity of each other (when two Cys residues form an SS-bond, the distance between the two Cβ atoms of the Cys residues is within 4.6 Å 28 ). The incubation for 1 min was the shortest period needed for handling the samples in the experiments. The temperature was set to the same as that for the fluorescence experiments to evaluate membrane potential. The oxidative effect of CuP was quenched after 1 min by 5 mM ethylenediaminetetraacetic acid (EDTA). After the removal of the liposomal lipids, the VSD mutants were denatured with SDS and urea, treated simultaneously with mal-PEG, and then analyzed by SDS-PAGE. An unPEGylated band provides good evidence of the SS-bond formation that protects the protein from the PEGylation. The SS-bond formation was judged as "+ ", when the intensity of the unPEGylated band is higher than 25% of the total intensity of the PEGylated and unPEGylated bands. It should be noted that this analysis could not discriminate between transitions with high and low probability, since the SS-bond is formed even at a single occurrence of close proximity. Therefore, we did not analyze the extent of the SS-bond formation quantitatively, but evaluated qualitatively the possibility of two Cys residues coming into close proximity.
Three representative results of the SS-locking analysis are shown in [fig_ref] Figure 5: Voltage-dependent SS-locking analysis [/fig_ref]. In [fig_ref] Figure 5: Voltage-dependent SS-locking analysis [/fig_ref] , for the V42C/I131C mutant, an SDS-PAGE band for an unPEGylated VSD was observed at 0 and − 140 mV, even with mal-PEG treatment. In contrast, in [fig_ref] Figure 5: Voltage-dependent SS-locking analysis [/fig_ref] , for the S32C/L122C mutant, the SDS-PAGE band corresponding to unPE-Gylated protein disappeared with the mal-PEG treatment, and two bands corresponding to PEGylated protein were observed at both 0 and − 175 mV, indicating that one or two Cys residues are subject to the PEGylation. Considering the fact that the unPEGylated band disappeared with the mal-PEG treatment in [fig_ref] Figure 5: Voltage-dependent SS-locking analysis [/fig_ref] , the first PEGylation for one of the free thiol groups proceeds sufficiently quickly, while the second PEGylation seems slower than the first one. Since the molecular weight of mal-PEG is ca. 2,000, steric hindrance between a A comparison of the results in [fig_ref] Figure 5: Voltage-dependent SS-locking analysis [/fig_ref] ,b revealed that not all, but a significant amount, of the V42C/I131C mutant formed an SS-bond, which protected the protein from PEGylation, and that no SS-bond was formed in the V32C/L122C mutant between the two thiol groups of the Cys residues. Thus, in the wild type VSD, V42 and I131 would approach each other at 0 and − 140 mV, while S32 and L122 do not, at either 0 or − 175 mV.
In [fig_ref] Figure 5: Voltage-dependent SS-locking analysis [/fig_ref] , for the V42C/I130C mutant, an unPEGylated band was observed only at 0 mV (depolarized), indicating that SS-bonds were formed between V42C and I130C, under the depolarized conditions. Thus, these results suggested that V42 and I130 could approach each other only under depolarized conditions. While V42C/ I130C exhibited an unPEGylated band without the mal-PEG treatment [fig_ref] Figure 5: Voltage-dependent SS-locking analysis [/fig_ref] , the SS-bond was expected to be formed not during the incubation period at the membrane potential of − 130 mV, but under the denatured condition afterwards, because the sample with mal-PEG treatment did not show unPEGylated band.
We next investigated whether each of the other 32 double Cys mutants could form an SS-bond under the depolarized and/or polarized conditions, based on the existence of the unPEGylated band in SDS-PAGE .
The results are schematically summarized in [fig_ref] Figure 6: Schematic diagram of the results of the voltage-dependent SS-locking analysis [/fig_ref]. Under the depolarized conditions (0 mV), the intramolecular SS-bond formation was observed for 7 mutants L36C/L129C, V42C/L121C, V42C/L122C, V42C/I127C, V42C/L129C, V42C/I130C, and V42C/I131C. Under the polarized conditions (< − 100 mV), the intramolecular SS-bond formation was observed for 6 mutants L36C/L129C, V42C/L121C, V42C/L122C, V42C/I127C, V42C/ L129C, and V42C/I131C.
The unPEGylated band might be due to oxidation of a thiol group to a sulfino or sulfonic group without the formation of the intramolecular SS-bond. In order to investigate whether the PEGylation was protected by the SS-bond or not, we reduced SS-bonds after CuP oxidation, followed by mal-PEG modification [fig_ref] Figure 7: Conformational equilibrium of the VSD [/fig_ref]. Reduced samples did not show unPEGylated bands, but showed the PEGylated bands. These results clearly showed that these mutants formed intramolecular SS-bond and were not oxidized to a sulfino or sulfonic group.
Among these mutants, all but one residue pair formed an SS-bond at both 0 and < − 100 mV (bold black lines in [fig_ref] Figure 6: Schematic diagram of the results of the voltage-dependent SS-locking analysis [/fig_ref] connecting the two Cys residues). In particular, the residues L121C and I131C, which formed an SS-bond with V42C both at 0 and < − 100 mV, are 10-residues apart on the S4 helix. If a single conformation of VSD is assumed, it would be impossible for L121 and I131 to simultaneously approach V42, strongly suggesting that the VSD adopts multiple conformations at both membrane potentials.
Only one mutant, V42C/I130C, exhibited significant differences in the SS-bond formation between 0 mV and < − 100 mV: an SS-bond was formed at 0 mV (orange line in [fig_ref] Figure 6: Schematic diagram of the results of the voltage-dependent SS-locking analysis [/fig_ref] , and not at − 130 mV.
Considering the fact that the VSD adopts multiple conformations under the polarized and depolarized conditions, the voltage-dependent difference in the SS-bond formation suggests that the multiple conformations are differently populated under the polarized and depolarized conditions.
# Discussion
In this study, we analyzed the conformations of the isolated VSD in the lipid bilayer in the presence or absence of a membrane potential. Previous studies indicate that the isolated VSD from KvAP in the lipid bilayer adopts a similar conformation to that from the full length KvAP 9 . Consistent with this finding, our fluorescence analysis indicated that the sigmoidal relationship between the membrane potential and the conformational change of the VSD [fig_ref] Figure 2: Fluorescence analysis of mBBr to examine voltage-dependent conformational changes of VSD [/fig_ref] is also similar to that of the full length KvAP. Time dependent recovery of fluorescence intensity after instantaneous reduction [fig_ref] Figure 2: Fluorescence analysis of mBBr to examine voltage-dependent conformational changes of VSD [/fig_ref] seems to correspond to the decay of membrane potential [fig_ref] Figure 1: Measurement of membrane potentials of liposomes [/fig_ref].
In order to detect proximal residue pairs of the VSD, we conducted voltage-dependent SS-locking analyses for VSD possessing double Cys mutations in one of four residues in S1 (S32, L36, V39, V42) and one of nine residues in S4 (L121 -I131 except the two Arg residues) [fig_ref] Figure 4: Locations of Cys mutation on VSD [/fig_ref]. The residue pairs that formed an SS-bond at 0 and < − 100 mV are summarized in [fig_ref] Figure 6: Schematic diagram of the results of the voltage-dependent SS-locking analysis [/fig_ref]. While one residue pair exhibited voltage-dependent difference in the SS-bond formation (orange), six residue pairs formed SS-bonds at both membrane potentials (bold black lines). In particular, V42C, located on the extracellular side of the S1 helix, formed an SS-bond with the S4 residues at the N-terminal end (L121C and L122C) and those at the C-terminal end (I127C, L129C, and I131C), while no SS-bond was observed for V42C with F124C or L125C, located in the middle of S4. These results strongly suggest that the S4 helix can exist stably in the intracellular (down) and extracellular (up) positions, where the N and C-terminal ends of S4 are close to the extracellular end of S1, respectively, and that the S4 helix is unable to remain in the middle position.
Assuming the two states where the S4 helix resides in the up and down positions (referred to as the up and down states, respectively, [fig_ref] Figure 7: Conformational equilibrium of the VSD [/fig_ref] , all of the residue pairs forming an SS-bond are close to each other in either state. In the up state, V42 in S1 is close to I127, L129, I130, and I131 in S4 [fig_ref] Figure 7: Conformational equilibrium of the VSD [/fig_ref]. In the down state, V42 is close to L121 and L122 in S4, and at the same time, L36 in S1 is close to I129 in S4 [fig_ref] Figure 7: Conformational equilibrium of the VSD [/fig_ref]. In each state, however, 2-3 continuous residues in S4 formed an SS-bond with one of the residues in S1, indicating that the S4 helix possesses flexibility with an axial rotation of about 180°. These results suggest that each state is not composed of a single conformation, but rather contains an ensemble of multiple conformations. Since the residue pairs connected by bold black lines in Figs 6 and 7 formed SS-bonds at both membrane potentials, the VSD seems to exist in a conformational equilibrium between the up and down states, where each state is also in equilibrium among multiple conformations.
While six Cys residue pairs formed an SS-bond regardless of the magnitude of the membrane potential, the SS-bond between V42C-I130C was observed only at 0 mV (orange in [fig_ref] Figure 6: Schematic diagram of the results of the voltage-dependent SS-locking analysis [/fig_ref]. Assuming that the VSD adopts either of the up or down state, this result suggests that depolarization of the membrane increases the proportion of the up state.
L129C forms an SS-bond with V42C in the up state and with L36C in the down state. Since the Cβ atoms of L36 and V42 are 12 Å apart, the distance of the S4 translational shift is estimated at about 12 Å [fig_ref] Figure 7: Conformational equilibrium of the VSD [/fig_ref]. Therefore, these results suggest that changes of the membrane potential are sensed by the S4 translocation that are 12 Å at the largest, leading to the allosteric regulation of the gating at the PD [bib_ref] Mechanisms of activation of voltage-gated potassium channels, Grizel [/bib_ref] [bib_ref] Voltage-Gated Potassium Channels: A Structural Examination of Selectivity and Gating, Kim [/bib_ref].
Previous studies also detected voltage-dependent conformational changes [bib_ref] The principle of gating charge movement in a voltage-dependent K+ channel, Jiang [/bib_ref] [bib_ref] Calibrated measurement of gating-charge arginine displacement in the KvAP voltage-dependent K+ channel, Ruta [/bib_ref] [bib_ref] Two atomic constraints unambiguously position the S4 segment relative to S1 and..., Campos [/bib_ref] [bib_ref] Sensing voltage across lipid membranes, Swartz [/bib_ref] [bib_ref] Structural basis for gating charge movement in the voltage sensor of a..., Yarov-Yarovoy [/bib_ref] [bib_ref] Tracking a complete voltage-sensor cycle with metal-ion bridges, Henrion [/bib_ref] and lipid-driven conformational change [bib_ref] Structural basis of lipid-driven conformational transitions in the KvAP voltagesensing domain, Li [/bib_ref] of the VSD. The voltage-dependent conformational change suggested from this study seems to be different from the lipid-driven conformational change [bib_ref] Structural basis of lipid-driven conformational transitions in the KvAP voltagesensing domain, Li [/bib_ref]. We assume that the down state in [fig_ref] Figure 7: Conformational equilibrium of the VSD [/fig_ref] is different from the conformation in non-phospholipid bilayer. However, we cannot compare the conformations of VSD precisely, because we cannot acquire detailed structural data of VSD in non-phospholipid bilayer. On the other hand, our estimation of the S4 translocation is similar to, but not completely the same as those in the previously reported values ranging from 6 to 20 Å, which were obtained by different methodologies such as other disulfide locking analyses 12-14 and a metal-ion bridge analysis [bib_ref] Tracking a complete voltage-sensor cycle with metal-ion bridges, Henrion [/bib_ref]. It should be noted that, in the previous reports, Cys mutations were introduced at charged residues such as Arg, Asp, or Glu 12-15 , although the membrane potential is sensed by the Arg residues that interact with Asp/Glu residues. In this study, we avoided introducing mutations at these key residues to preserve the electrophysiological properties of VSD, since mutation of these residues may modify the mode of the voltage-dependent conformational change of VSD.
# Methods
Expression and purification of VSD. The DNA encoding Met -12 to Lys 147 of the KvAP channel, with an N-terminal decahistidine tag, followed by an HRV 3C protease cleavage site, was inserted into the pMAL-c2X plasmid (New England Biolabs). To prevent the artificial dimerization of VSD, Cys -2 was substituted with Ser by the QuikChange ® system (Strategene). Hereafter, this mutant, Cys-2Ser, is referred to as VSD. The site specific Cys mutations described below, were also introduced into the VSD by QuikChange ® . We prepared the V119C mutant for fluorescent labeling, and 36 double Cys mutants for SS-locking (see below), in which we mutated one residue in S1 (S32, L36, V39, or V42) and another one in S4 (L121, L122, F124, L125, I127, L128, L129, I130, or I131) to Cys.
The VSD and its mutants were expressed in E. coli XL1-Blue cells, as fusion proteins with an N-terminal MBP and a decahistidine tag (hereafter, the expressed proteins are referred to as MBP-VSD). The cells were grown at 37 °C to an A 600 of 0.4-0.6, and protein expression was then induced with 1.0 mM isopropyl-β -D-thiogalactopyranoside for 3-6 hours at 37 °C. Purification of the MBP-VSD was described elsewhere [bib_ref] Structural basis for the inhibition of voltage-dependent K+ channel by gating modifier..., Ozawa [/bib_ref]. For the reconstitution of the purified proteins, the buffer was exchanged with "liposome buffer", containing 20 mM Hepes-NaOH, pH 8.0, 0.10 mM KCl, 149.9 mM NaCl, and 10%(v/v) glycerol, in the presence of 2 mM DM. The purified MBP-VSDs were concentrated to 100-200 μ M and stored at 4 °C, after addition of 10 mM dithiothreitol (DTT).
## Preparation of liposomes.
A 3:1 (w/w) mixture of POPE/POPG (Avanti) was dried under nitrogen stream and then suspended at a concentration of 10 mg/mL in liposome buffer containing 100 mM DM. For the preparation of empty liposomes, the lipid suspension was slowly rotated at room temperature for 3 hours. For the preparation of VSD-reconstituted liposomes, the lipid suspension was rotated at room temperature for 1 hour, followed by the addition of the MBP-VSD at a weight ratio of 1:200 for VSD/lipid, and further rotation at room temperature for 2-3 hours. Then, the solution was dialyzed against a 500-1000 times volume of the liposome buffer with a 10,000 MWCO membrane (Spectrum), at 4 °C for 3-5 days. After dialysis, the liposomal solution was extruded on an Avanti ® Mini-Extruder (Avanti), with 11 passes through a 0.1 μ m Nuclepore membrane (Whatman). The MBP residing on the outside of the liposomes was removed by digesting the linker between the MBP and the VSD with HRV 3C protease (Novagen).
Measurement of the membrane potential of the liposomes. The prepared liposomes were diluted with "membrane potential buffer", containing 20 mM Hepes-NaOH, pH 8.0, x mM KCl (x = 0.1-150), 150-x mM NaCl, and 10% (v/v) glycerol, which is the same as the liposome buffer except for the concentrations of KCl and NaCl. Then, a membrane potential was formed on the liposome membrane by the addition of valinomycin, a K + selective ionophore, in DMSO solution.
We measured the membrane potential of the liposomes by observing the change in the fluorescence intensity of the membrane potential sensitive fluorescent dye, di-4-ANEPPS. Data were collected at 25 °C on an RF-5300PC spectrofluorometer (Shimadzu).
Briefly, 10 μ L of 100 μ M di-4-ANEPPS DMSO solution and 50 μ L liposome buffer were added to 1700 μL membrane potential buffer (see for the concentrations of KCl and NaCl in the membrane potential buffer), and then fluorescence time course measurement was started with excitation at 530 nm and emission at 610 nm. At 0 s, a 40 μ L aliquot of liposomes was added, and after a 60 s interval, 1 μ L of 40 μ M valinomycin was added to form a membrane potential. The fluorescence intensity following the addition of valinomycin was estimated from the linear regression of the fluorescence intensity for 45 s.
Fluorescence analysis of mBBr-modified VSD. Monobromobimane-modified VSD was prepared according to the published procedure [bib_ref] Structural basis for the inhibition of voltage-dependent K+ channel by gating modifier..., Ozawa [/bib_ref]. The fluorescence analysis was conducted with membrane potential buffer .
The fluorescence time course was observed for 1800 μ L membrane potential buffer, with excitation at 394 nm and emission at 470 nm. At 30 s, a 40 μ L aliquot of liposomes was added, and 1 μ L of 40 μ M valinomycin was added at 90 s, to form a membrane potential. The fluorescence intensity upon the formation of the membrane potential by the addition of valinomycin (at 100 s) was estimated from the linear regression of the fluorescence intensity for the following 45 s. The values of membrane potential were estimated for each sample by di-4-ANEPPS analysis, as shown in "Measurement of the membrane potential of the liposomes". In order to obtain the relationship between the membrane potential and the fluorescence intensity reduction ratio, we conducted this analysis with identical liposome sample and different membrane potential buffers whose theoretical membrane potentials are 0, 40, 80, 100, 110, 120, 130, 140, 150, 160, 180, and 188 mV.
## Detection of an intramolecular ss-bond in the vsd in dm micelle. the vsd double cys mutants
were incubated with either 10 mM DTT as a reductant or 100 μ M CuP as an oxidant at room temperature for 1 hour. The CuP was inactivated by the addition of 500 μ M EDTA.
The VSD was collected as a precipitate by chloroform/methanol extraction [bib_ref] A method for the quantitative recovery of protein in dilute solution in..., Wessel [/bib_ref]. The VSD precipitate was solubilized in "denaturation buffer" (15 mM Tris-HCl, pH 7.4, 2% (w/v) SDS, 6 M urea) containing 0 mM or 1 mM mal-PEG, with a molecular weight of about 2,000 (SUNBRIGHT ® ME-20MA (NOF)), incubated at room temperature for 3 hours, and analyzed by SDS-PAGE.
Voltage-dependent SS-locking analysis. Liposomes reconstituted with VSD double Cys mutant were prepared as described above, with 1 mM DTT. The DTT in the liposome solution was removed by gel filtration with PD miditrap TM G-25 (GE Healthcare). It should be noted that MBP-VSD is digested by HRV 3C protease when the MBP resides on the outside of the liposome, and MBP on the inside of the liposome remains fused with the VSD. The digested MBP and the HRV 3C protease, both of which were His-tagged, were removed by His-select ® Nickel affinity gel (SIGMA-ALDRICH).
For the oxidative SS-bond formation, 760 μ L membrane potential buffer (KCl concentration 0.10 mM or 150 mM), 1 μ L of 40 μ M valinomycin, 40 μ L liposome solution, and 8 μ L of 3 mM CuP (final concentration at 30 μ M) were mixed on a shaking vortex mixer at room temperature. The oxidation was stopped with 5 mM EDTA at 1 min. In order to oxidize only at the polarized membrane potential, oxidation was conducted for 1 min, during which, about 8% of the membrane potential decayed [fig_ref] Figure 1: Measurement of membrane potentials of liposomes [/fig_ref]. The values of membrane potential were estimated for each sample by di-4-ANEPPS analysis, as shown in "Measurement of the membrane potential of the liposomes".
Next, the VSD was extracted as a precipitate with chloroform/methanol 30 , solubilized with denaturation buffer containing 0 mM or 1 mM mal-PEG, and incubated at room temperature overnight, and analyzed by SDS-PAGE. An intramolecular SS-bond was detected by the band without PEGylation, which is observed only when the two thiol groups formed the intramolecular SS-bond (n = 1-2).
## Confirmation of the ss-bond formation.
Liposomes reconstituted with VSD double Cys mutant were prepared as described above, with 1 mM DTT. The MBP-tag on the outside of the liposome was digested by HRV 3C protease. The DTT in the liposome solution was removed by gel filtration with PD miditrapTM G-25. 380 μ L liposome buffer, 20 μ L liposome solution, and 4 μ L of 3 mM CuP (final concentration at 30 μ M) were mixed on a shaking vortex mixer and incubated for 1 hour at room temperature. The oxidation was stopped with 5 mM EDTA.
Next, the VSD was extracted as a precipitate with chloroform/methanol, chloroform/methanol extraction was conducted twice, solubilized with liposome buffer containing 40 mM DM and 0 mM or 10 mM tris(2-carboxyethyl)phosphine (TCEP), and incubated at room temperature for 1 hour. VSD was extracted as a precipitate with chloroform/methanol, chloroform/methanol extraction was conducted twice, solubilized with denaturation buffer containing 0 mM or 1 mM mal-PEG, and incubated at room temperature overnight, and analyzed by SDS-PAGE.
[fig] Figure 1: Measurement of membrane potentials of liposomes. (a) Voltage dependence of the fluorescence intensity of empty liposomes. (b) Relationship between the theoretical membrane potential formed on empty liposomes and the fluorescence intensity reduction ratio of di-4-ANEPPS (panel (a)). (c) Fluorescence timecourse for the liposomes reconstituted with a VSD mutant, V42C/I131C, where the K + concentration ratio was adjusted to form a − 187 mV theoretical membrane potential (black line). The results of (a) are shown by gray lines. The potentials corresponded to − 140 and − 124 mV at a few and 90 s after the addition of valinomycin, respectively. [/fig]
[fig] Figure 2: Fluorescence analysis of mBBr to examine voltage-dependent conformational changes of VSD. (a) The location of V119, which was mutated to Cys with the mBBr modification. (b-d) Fluorescence time courses of mBBr attached VSD on the liposomes with three different membrane potentials: (b) 0 mV, (c) − 90 mV, and (d) − 147 mV. (e) Relationship between the estimated membrane potential by di-4-ANEPPS analysis and the fluorescence intensity reduction ratio of mBBr at 100 s (upon the formation of the membrane potential). [/fig]
[fig] Figure 3: Detection of an intramolecular SS-bond. (a) Schematic diagram of the intramolecular SS-bond detection. The SH form of the VSD mutant is PEGylated and exhibited a significant mobility shift (left), while the SS form of the VSD mutant is not PEGylated, leaving the mobility unchanged (right). (b) SDS-PAGE mobility of the VSD double Cys mutant (V42C/I130C) under reductive or oxidative conditions in the presence or absence of mal-PEG. [/fig]
[fig] Figure 4: Locations of Cys mutation on VSD. (a)The locations of the residues mutated to Cys are shown as spheres (residues on S1 and S4 are colored red and green, respectively) on the VSD crystal structure (PDB code 1ORS). (b) The analyzed residue pairs are connected by lines. [/fig]
[fig] Figure 5: Voltage-dependent SS-locking analysis. The SS-locking analysis results of three VSD double Cys mutants are shown. SS(+ ) stands for the SS-bond formation, as evidenced by the intensity of the unPEGylated band higher than 25% of the total intensity of the PEGylated and unPEGylated bands. SS(− ) stands for no SSbond formation, as evidenced by the intensity of the unPEGylated band lower than 25% of the total intensity of the PEGylated and unPEGylated bands. (a) V42C/I131C, (b) S32C/L122C, and (c) V42C/I130C. Scientific RepoRts | 6:37303 | DOI: 10.1038/srep37303 VSD-attached mal-PEG molecule and a free mal-PEG could reduce the reaction rate of the second mal-PEG molecule, even when the VSD is denatured. Indeed, at least one mal-PEG molecule is bound to all the mutants as shown in Fig. 5 and Supplementary Figs S3-6. The relative intensity ratio of the bands for the VSD -Mal-PEG molecule and the VSD -(Mal-PEG) 2 molecule varied, depending on the positions of the Cys residues. [/fig]
[fig] Figure 6: Schematic diagram of the results of the voltage-dependent SS-locking analysis. Black bold lines show residue pairs that formed an SS-bond. The orange line shows the residue pair, which only formed an SSbond at 0 mV. Scientific RepoRts | 6:37303 | DOI: 10.1038/srep37303 [/fig]
[fig] Figure 7: Conformational equilibrium of the VSD. VSD exists in an equilibrium between the up and down states in either polarized or depolarized environment. Scientific RepoRts | 6:37303 | DOI: 10.1038/srep37303 [/fig]
|
Human Organoids as a Promising Platform for Fighting COVID-19
The coronavirus disease 2019 global pandemic evoked by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has triggered a major public health problem with significant morbidity and mortality. Understanding the pathogenesis and molecular mechanisms underlying this novel virus is crucial for both fundamental research and clinical trials in order to devise effective therapies and vaccination regimens. Basic research on SARS-CoV-2 largely depends on ex vivo models that allow viral invasion and replication. Organoid models are now emerging as a valuable tool to investigate viral biology and disease progression, serving as an efficient platform to investigate potential therapies for COVID-19. Here, we summarize various human stem cell-derived organoid types employed in SARS-CoV-2 studies. We highlight key findings from these models, including cell tropisms and molecular mechanisms in viral infection. We also describe their use in identifying potential therapeutic agents against SARS-CoV-2. As more and more advanced organoids emerge, they will facilitate the understanding of disease pathogenesis for drug development in this dreaded pandemic.
# Introduction
The coronavirus disease 2019 (COVID- [bib_ref] Genetic Manipulation of Human Intestinal Enteroids Demonstrates the Necessity of a Functional..., Haga [/bib_ref] pneumonia epidemic caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has become the worst public health threat in the current century. Treatment strategies for prevention and intervention are a matter of urgency. Although many clinical trials are currently underway, preclinical research on in vitro models is also needed to allow us to better understand virus infections and to test drugs and vaccines for safety and efficacy.
Until now, there has only been a limited number of animal models available in SARS-CoV-2 research. The most susceptible animals to this virus are cats [bib_ref] Susceptibility of ferrets, cats, dogs, and other domesticated animals to SARS-coronavirus 2, Shi [/bib_ref] , ferrets [bib_ref] Susceptibility of ferrets, cats, dogs, and other domesticated animals to SARS-coronavirus 2, Shi [/bib_ref] [bib_ref] Antiviral Efficacies of FDA-Approved Drugs against SARS-CoV-2 Infection in Ferrets, Park [/bib_ref] , rhesus macaques [bib_ref] Infection with novel coronavirus (SARS-CoV-2) causes pneumonia in Rhesus macaques, Shan [/bib_ref] , cynomolgus macaques [bib_ref] Comparative pathogenesis of COVID-19, MERS, and SARS in a nonhuman primate model, Rockx [/bib_ref] , and golden hamsters [bib_ref] Pathogenesis and transmission of SARS-CoV-2 in golden hamsters, Sia [/bib_ref]. SARS-CoV-2 utilizes angiotensin converting enzyme 2 (ACE2) for cellular entry and the cell surface transmembrane serine protease 2 (TMPRSS2) for spike protein priming [bib_ref] SARS-CoV-2 cell entry depends on ACE2 and TMPRSS2 and is blocked by..., Hoffmann [/bib_ref] [bib_ref] Enhanced isolation of SARS-CoV-2 by TMPRSS2-expressing cells, Matsuyama [/bib_ref]. Transgenic mice expressing human ACE2 can be a good option to study COVID-19 [bib_ref] The pathogenicity of SARS-CoV-2 in hACE2 transgenic mice, Bao [/bib_ref]. Nevertheless, animal-based studies are generally expensive and lengthy, and fail to fully mimic human physiology due to species specificity.
Classic cell lines are a great alternative to animal models. In fact, much of the insight on the pathogenesis and drug responses of SARS-CoV-2 has been obtained using numerous cell lines, such as Vero (kidney epithelial cells), Caco2 (intestinal cells), Calu3 (pulmonary cells), and Huh7 (hepatic cells) [bib_ref] Enhanced isolation of SARS-CoV-2 by TMPRSS2-expressing cells, Matsuyama [/bib_ref] [bib_ref] Comparative tropism, replication kinetics, and cell damage profiling of SARS-CoV-2 and SARS-CoV..., Chu [/bib_ref] [bib_ref] Identification of Antiviral Drug Candidates against SARS-CoV-2 from FDA-Approved Drugs, Jeon [/bib_ref]. However, cell lines are mostly of malignant origin and cannot simulate cell-cell and cell-matrix interactions. The in vivo infection status and pathological features may also be poorly recapitulated in these two-dimensional (2D) cultures.
A range of recent studies on COVID-19 highlights the value of more physiological in vitro Ivyspring International Publisher models, called organoids. Organoids are miniaturized, three-dimensional (3D) tissue models that derived from induced pluripotent stem cells (iPSCs), embryonic stem cells (ESCs), or multipotent adult stem cells (ASCs) [bib_ref] Modeling Development and Disease with Organoids, Clevers [/bib_ref] [bib_ref] Organogenesis in a dish: modeling development and disease using organoid technologies, Lancaster [/bib_ref]. Importantly, they consist of various cell types and faithfully recapitulate the essential structure and physiological characteristics of the parental organ in a dish [bib_ref] Disease Modeling in Stem Cell-Derived 3D Organoid Systems, Dutta [/bib_ref] [bib_ref] Progress and potential in organoid research, Rossi [/bib_ref]. Organoids are powerful tools for disease modeling in vitro, providing a more flexible, efficient, and large-scale drug screening platform than in vivo models. They are also ethically humane tools used to study physiology and disease, including viral infections.
## Organoids in virology
In virology, organoids are particularly useful as they permit studies on viruses that are difficult to cultivate. Organoid models have been shown to be an excellent model to study viral infections and host-virus interactions [bib_ref] A three-dimensional model of human lung development and disease from pluripotent stem..., Chen [/bib_ref] [bib_ref] Oral Mucosal Organoids as a Potential Platform for Personalized Cancer Therapy, Driehuis [/bib_ref] [bib_ref] Human Norovirus Cultivation in Nontransformed Stem Cell-Derived Human Intestinal Enteroid Cultures: Success..., Estes [/bib_ref] [bib_ref] Replication of human noroviruses in stem cell-derived human enteroids, Ettayebi [/bib_ref] [bib_ref] Genetic Manipulation of Human Intestinal Enteroids Demonstrates the Necessity of a Functional..., Haga [/bib_ref] [bib_ref] Using brain organoids to understand Zika virus-induced microcephaly, Qian [/bib_ref] [bib_ref] Long-term expanding human airway organoids for disease modeling, Sachs [/bib_ref] [bib_ref] Self-organized cerebral organoids with human-specific features predict effective drugs to combat Zika..., Watanabe [/bib_ref] [bib_ref] Differentiated human airway organoids to assess infectivity of emerging influenza virus, Zhou [/bib_ref]. Generating various organoid models from human tissues has become more important in studying viral infections as patients show signs of systemic symptoms in addition to respiratory infection.
Most of airborne viruses enter human body by infecting epithelial cells. Organoids derived from oral mucosa, airway, lung, and intestines have been used to model viral infection and replication. Organoids derived from human oral mucosal maintain morphologic and functional characteristics of their parental tissues, and have been shown to be productively infected by human papillomavirus (HPV) and herpes simplex virus (HSV) [bib_ref] Oral Mucosal Organoids as a Potential Platform for Personalized Cancer Therapy, Driehuis [/bib_ref]. Human airway organoids have emerged as a valuable tool to assess the susceptibility of animal influenza viruses [bib_ref] Differentiated human airway organoids to assess infectivity of emerging influenza virus, Zhou [/bib_ref] [bib_ref] Tropism, replication competence, and innate immune responses of influenza virus: an analysis..., Hui [/bib_ref]. proved that the differentiated human airway organoids are susceptible to the human-infective influenza viruses H7N9 and H1N1, while avian H7N2 and swine H1N1 influenza viruses showed lower replication levels [bib_ref] Differentiated human airway organoids to assess infectivity of emerging influenza virus, Zhou [/bib_ref]. Another study used human airway organoid cultures to measure replication competence, tissue tropism and host response to human influenza virus [bib_ref] Tropism, replication competence, and innate immune responses of influenza virus: an analysis..., Hui [/bib_ref]. Moreover, Sachs, et al., used human airway organoids to show dramatic epithelial remodeling after respiratory syncytial virus (RSV) infection [bib_ref] Long-term expanding human airway organoids for disease modeling, Sachs [/bib_ref]. Noroviruses are a major cause of gastroenteritis in the world. Human stem cell-derived intestinal organoids have become the first ex vivo infection model to support noroviruses replication [bib_ref] Replication of human noroviruses in stem cell-derived human enteroids, Ettayebi [/bib_ref]. Indeed, the human intestinal organoid cultivation system recapitulates the physiologically active intestinal epithelium, and allows studies of norovirus replication efficiency in vitro [bib_ref] Human Norovirus Cultivation in Nontransformed Stem Cell-Derived Human Intestinal Enteroid Cultures: Success..., Estes [/bib_ref]. This model was also used to confirm that the intestines are a target organ for Middle East respiratory syndrome coronavirus (MERS-CoV) [bib_ref] Human intestinal tract serves as an alternative infection route for Middle East..., Zhou [/bib_ref].
During the 2015 Zika virus (ZIKV) outbreak, human iPSC-derived cerebral organoids were adopted to provide proof that this virus selectively replicates in the developing brain, preferentially infecting neural cell precursors, leading to congenital abnormalities including microcephaly [bib_ref] Self-organized cerebral organoids with human-specific features predict effective drugs to combat Zika..., Watanabe [/bib_ref] [bib_ref] Recent Zika Virus Isolates Induce Premature Differentiation of Neural Progenitors in Human..., Gabriel [/bib_ref] [bib_ref] Brain-region-specific organoids using mini-bioreactors for modeling ZIKV exposure, Qian [/bib_ref]. These studies provide the explanation as to why the virus has a greater detrimental effect on the fetal brain compared to the postnatal brain. By using cortical organoids to infect cytomegalovirus, Sison, et al., found organoid structure alterations and disruptions in specific marker expression [bib_ref] Human Cytomegalovirus Disruption of Calcium Signaling in Neural Progenitor Cells and Organoids, Sison [/bib_ref]. Furthermore, human liver organoids also emerged as a promising personalized infection model in the treatment of hepatitis B virus (HBV). By co-culture of human iPSCderived endodermal, mesenchymal and endothelial cells with a chemically defined medium in a 3D system, Nie and colleagues generated the functional liver organoids that can inherit the genetic background of donors and reproduce host-virus interactions by simulating HBV propagation and the virus-induced hepatic dysfunction [bib_ref] Recapitulation of hepatitis B virus-host interactions in liver organoids from human induced..., Nie [/bib_ref].
## Human organoids in studying sars-cov-2 infection
Besides lung injury caused by SARS-CoV-2, symptoms have also been noted in multiple other organs, including brain, liver, intestine, kidney, eye, heart, and blood vessels [fig_ref] Figure 1: Schematic representation of the main organs affected by SARS-COV-2 [/fig_ref]. Many research groups have utilized organoid technologies to understand the tissue tropism and cellular response of SARS-CoV-2 and the damage caused. Organoids can be established from either human pluripotent stem cells (PSCs) including ESCs and iPSCs, or multipotent adult tissue stem cells (ASCs). SARS-CoV-2 infection and replication have been studied in PSC-and ASC-based organoids derived from a wide range of organs [fig_ref] Table 1: List of human PSC-derived organoids used to study SARS-CoV-2 infection [/fig_ref]. Organoid models are proving their worth in verifying the safety and efficacy of antiviral therapies as these in vitro models can exploit a priori knowledge on virus biology, and allow for the testing of well-known viral inhibitors and the discovery of new drugs [fig_ref] Figure 2: ducts isolated from human liver biopsies ACE2 and TMPRSS2 are involved in... [/fig_ref].
## Lung organoids
SARS-CoV-2 has been demonstrated to employ two key host proteins, ACE2 and TMPRSS2, to bind and infect host cells [bib_ref] SARS-CoV-2 cell entry depends on ACE2 and TMPRSS2 and is blocked by..., Hoffmann [/bib_ref] [bib_ref] Enhanced isolation of SARS-CoV-2 by TMPRSS2-expressing cells, Matsuyama [/bib_ref]. ACE2 and TMPRSS2 have been shown to be highly expressed in human airway organoids [bib_ref] Differentiated human airway organoids to assess infectivity of emerging influenza virus, Zhou [/bib_ref] [bib_ref] The role of ACE2 in pulmonary diseases-relevance for the nephrologist, Oudit [/bib_ref] , indicating that they are suitable for SARS-CoV-2 research. Air-liquid interface (ALI) cultures of human airway organoids were readily infected by the addition of SARS-CoV-2 to the apical side [bib_ref] SARS-CoV-2 productively infects human gut enterocytes, Lamers [/bib_ref]. The virus infection was observed in ciliated cells but not in goblet cells, suggesting that ciliated cells are the main targets of the virus [bib_ref] SARS-CoV-2 productively infects human gut enterocytes, Lamers [/bib_ref]. In a recent study, alveolar cells, basal cells, and rare neuroendocrine cells were grown from human fetal lung bud tip progenitor organoids [bib_ref] An organoid-derived bronchioalveolar model for SARS-CoV-2 infection of human alveolar type II-like..., Lamers [/bib_ref]. Most of the infected cells were alveolar type II cells, and a low dose of interferon lambda 1 (IFN-λ1) was shown to reduce viral replication. Therefore, ciliated cells and alveolar cells are susceptible to SARS-CoV-2 infection, and ciliated cells could be infected prior to the alveoli [bib_ref] An organoid-derived bronchioalveolar model for SARS-CoV-2 infection of human alveolar type II-like..., Lamers [/bib_ref]. Ebisudani et al., established an efficient ASC-based alveolosphere culture system. These alveolospheres express ACE2 at both RNA and protein levels, and maintain the robust infection efficiency of SARS-CoV-2 even after long-term cultivation [bib_ref] Direct derivation of human alveolospheres for SARS-CoV-2 infection modeling and drug screening, Ebisudani [/bib_ref]. Salahudeen, et al., generated human distal lung organoids with apical-out morphology to present ACE2 on the exposed external surface, allowing them to be infected by SARS-CoV-2 [bib_ref] Progenitor identification and SARS-CoV-2 infection in human distal lung organoids, Salahudeen [/bib_ref]. Contrary to Lamers's work [bib_ref] SARS-CoV-2 productively infects human gut enterocytes, Lamers [/bib_ref] , goblet cells, but not ciliated cells, were infectable in their organoid cultures [bib_ref] Progenitor identification and SARS-CoV-2 infection in human distal lung organoids, Salahudeen [/bib_ref]. Additionally, ASC-derived human lung organoids consisting of both proximal and distal airway epithelia have been established, and the proximal airway epithelium has been shown to be more permissive to SARS-CoV-2 infection [bib_ref] Adult stem cell-derived complete lung organoid models emulate lung disease in COVID-19, Tindle [/bib_ref]. Han, et al., demonstrated SARS-CoV-2 entry and infection in human ESC-derived lung organoids, which are mainly composed of alveolar type-I (AT1) cells, alveolar type-II (AT2) cells, stroma cells, neuroendocrine cells, and airway epithelial cells [bib_ref] Identification of SARS-CoV-2 inhibitors using lung and colonic organoids, Han [/bib_ref]. Gene ontology analysis of infected lung organoids revealed that most upregulated genes were associated with immune response. Several chemokines and cytokines, including tumor necrosis factor (TNF), interleukin (IL), as well as nuclear factor kappa beta (NF-κB) were observably upregulation [bib_ref] Identification of SARS-CoV-2 inhibitors using lung and colonic organoids, Han [/bib_ref] [bib_ref] Host metabolism dysregulation and cell tropism identification in human airway and alveolar..., Pei [/bib_ref]. A progressive increase of cell death in lung organoids became evident at 72 hours post infection, and this was further validated by immunostaining experiments [bib_ref] Host metabolism dysregulation and cell tropism identification in human airway and alveolar..., Pei [/bib_ref].
Lung organoids derived from human pluripotent stem cells (hPSCs) were permissive to mass production, cryopreservation, and genetic manipulation [bib_ref] Generation of Complete Multi-Cell Type Lung Organoids From Human Embryonic and Patient-Specific..., Leibel [/bib_ref]. They also allowed researchers to assess the antiviral effects of COVID-19 candidate drugs [bib_ref] Direct derivation of human alveolospheres for SARS-CoV-2 infection modeling and drug screening, Ebisudani [/bib_ref] [bib_ref] Identification of SARS-CoV-2 inhibitors using lung and colonic organoids, Han [/bib_ref] [bib_ref] Host metabolism dysregulation and cell tropism identification in human airway and alveolar..., Pei [/bib_ref] [bib_ref] Androgen Signaling Regulates SARS-CoV-2 Receptor Levels and Is Associated with Severe COVID-19..., Samuel [/bib_ref]. Remdesivir, a nucleotide analogue prodrug used to inhibit viral replication, has been shown to greatly reduce the production of infectious virus particles in both human airway organoids and alveolar organoids, while camostat (a TMPRSS2 inhibitor) has been shown to slightly decrease the production of the virus in airway organoids but not alveolar organoids [bib_ref] Host metabolism dysregulation and cell tropism identification in human airway and alveolar..., Pei [/bib_ref]. Remdesivir has also been shown to inhibit SARS-CoV-2 replication in human alveolospheres at the concentration comparable with the circulating drug level [bib_ref] Direct derivation of human alveolospheres for SARS-CoV-2 infection modeling and drug screening, Ebisudani [/bib_ref]. Suzuki, et al., found a reduction of viral copy number in the infected bronchial organoids with treatment of camostat. Neutralizing antibody CB6, one of the promising neutralizing antibodies to treat COVID-19 [bib_ref] A human neutralizing antibody targets the receptor-binding site of SARS-CoV-2, Shi [/bib_ref] , significantly decreased the production of infectious viral in lung organoids [bib_ref] Host metabolism dysregulation and cell tropism identification in human airway and alveolar..., Pei [/bib_ref]. In addition, pre-treatment of lung organoids with imatinib, chloroquine, mycophenolic acid (MPA), and quinacrine dihydrochloride (QNHC) effectively blocked the ACE2 cleavage site, suggesting a potential role in decreasing SARS-CoV-2 infection [bib_ref] Identification of SARS-CoV-2 inhibitors using lung and colonic organoids, Han [/bib_ref]. Human lung organoids have also been used to test the efficacy of candidate antiandrogenic drugs, such as finasteride, ketoconazole, and dutasteride, in resisting SARS-CoV-2 infection. These drugs have been certified to downregulate the expression of ACE2 in lung organoids, and therefore to reduce susceptibility to the virus [bib_ref] Androgen Signaling Regulates SARS-CoV-2 Receptor Levels and Is Associated with Severe COVID-19..., Samuel [/bib_ref]. Overall, human lung organoids can serve as an excellent model to investigate SARS-CoV-2 infection and to screen and discover candidate COVID-19 therapeutics. Besides lung injury caused by SARS-CoV-2, symptoms have also been noted in multiple other organs, including brain, eye, heart, liver, intestine, kidney, and blood vessels. SARS-CoV-2 entry and infection in lung organoids, which are mainly composed of alveolar type-I, alveolar type-II, stroma, neuroendocrine, and airway epithelial cells [bib_ref] Identification of SARS-CoV-2 inhibitors using lung and colonic organoids, Han [/bib_ref] ; SARS-CoV-2 were shown to infect ciliated, club, and alveolar type 2 cells in airway and alveolar organoids, and induce the downregulation of the metabolic processes and the upregulation of immune response [bib_ref] Host metabolism dysregulation and cell tropism identification in human airway and alveolar..., Pei [/bib_ref] ; Remdesivir was shown to greatly reduce the production of infectious virus particles in airway and alveolar organoids [bib_ref] Host metabolism dysregulation and cell tropism identification in human airway and alveolar..., Pei [/bib_ref] ; Androgen signaling inhibition reduces SARS-CoV-2 infection in lung organoids [bib_ref] Androgen Signaling Regulates SARS-CoV-2 Receptor Levels and Is Associated with Severe COVID-19..., Samuel [/bib_ref] Brain organoids are permissive to infection but do not support active viral replication [bib_ref] SARS-CoV-2 targets neurons of 3D human brain organoids, Ramani [/bib_ref] ; Neuronal infection was inhibited by blocking ACE2 with antibodies or by administering cerebrospinal fluid from a COVID-19 patient [bib_ref] Neuroinvasion of SARS-CoV-2 in human and mouse brain, Song [/bib_ref] ; SARS-CoV-2 can directly target cortical neurons and neural progenitor cells in brain organoids [bib_ref] SARS-CoV-2 infects human neural progenitor cells and brain organoids, Zhang [/bib_ref] ; Extensive viral protein expression and infectious viral particles were detected in brain organoids infected with SARS-CoV-2 [bib_ref] Infectability of human BrainSphere neurons suggests neurotropism of SARS-CoV-2, Bullen [/bib_ref] ; Choroid plexus organoids are permissive to productive infection, leading to transcriptional upregulation of inflammatory genes [bib_ref] Human pluripotent stem cell-derived neural cells and brain organoids reveal SARS-CoV-2 neurotropism..., Jacob [/bib_ref] ; SARS-CoV-2 infects choroid plexus, leading to damage of this brain barrier [bib_ref] SARS-CoV-2 infects the brain choroid plexus and disrupts the blood-CSF-barrier in human..., Pellegrini [/bib_ref] ; Sofosbuvir protects brain organoid from SARS-CoV-2 infection. [bib_ref] SARS-CoV-2 productively infects human gut enterocytes, Lamers [/bib_ref] [bib_ref] An organoid-derived bronchioalveolar model for SARS-CoV-2 infection of human alveolar type II-like..., Lamers [/bib_ref] ; Human alveolospheres express ACE2 and allow robust SARS-CoV-2 infection [bib_ref] Direct derivation of human alveolospheres for SARS-CoV-2 infection modeling and drug screening, Ebisudani [/bib_ref] ; Remdesivir inhibits SARS-CoV-2 replication in alveolospheres [bib_ref] Direct derivation of human alveolospheres for SARS-CoV-2 infection modeling and drug screening, Ebisudani [/bib_ref] ; Distal lung organoids with apical-out morphology to present ACE2 on the exposed external surface, allowing SARS-CoV-2 infection [bib_ref] Progenitor identification and SARS-CoV-2 infection in human distal lung organoids, Salahudeen [/bib_ref] ; Goblet cells but not ciliated cells were infectable in organoid cultures [bib_ref] Progenitor identification and SARS-CoV-2 infection in human distal lung organoids, Salahudeen [/bib_ref] ; ASC-derived lung organoids consisting of both proximal and distal airway epithelia, and the proximal airway epithelium are more permissive to SARS-CoV-2 infection [bib_ref] Adult stem cell-derived complete lung organoid models emulate lung disease in COVID-19, Tindle [/bib_ref] ; ACE2 and TMPRSS2 are highly expressed in bronchial organoids; A reduction of viral copy number in the infected bronchial organoids with treatment of camostat; Not only intracellular viral genome, but also progeny virus, cytotoxicity, pyknotic cells, and moderate increases of the type I interferon signal can be observed after SARS-CoV-2 infection.
## Remdesivir; lopinavir; nelfinavir; camostat
Intestinal organoids/ Colonic organoids Normal intestinal and colonic samples from patients Enterocytes produced infectious viral particles, and induced a generic viral response [bib_ref] SARS-CoV-2 productively infects human gut enterocytes, Lamers [/bib_ref] ; SARS-CoV-2 was able to infect human intestinal organoids, leading to progressive cytopathic effects with virus replication over time [bib_ref] Infection of bat and human intestinal organoids by SARS-CoV-2, Zhou [/bib_ref] ; Duodenum-and ileum-derived organoids are permissive to infection and support robust viral replication [bib_ref] TMPRSS2 and TMPRSS4 promote SARS-CoV-2 infection of human small intestinal enterocytes, Zang [/bib_ref] ; Pre-treatment of colon-derived organoids with both IFN-β1 and IFN-λ significantly blocked SARS-CoV-2 infection and this was associated with a decrease in viral genome copy number [bib_ref] Critical Role of Type III Interferon in Controlling SARS-CoV-2 Infection in Human..., Stanifer [/bib_ref].
## Intestinal organoids
Apart from respiratory illnesses, gastrointestinal symptoms have also been identified in a subset of patients [bib_ref] Extrapulmonary manifestations of COVID-19, Gupta [/bib_ref] , indicating that the gastrointestinal tract may be a potential entry route for SARS-CoV-2. Many kinds of bats are disreputable carriers of zoonotic viruses that occasionally spread to humans, including coronaviruses. Based on the close relation of SARS-CoV-2 to SARS-CoV identified in bat species, intestinal organoids were established from human and the horseshoe bat Rhinolophus sinicus, and their sensitivities to SARS-CoV-2 infection were examined for the first time [bib_ref] Infection of bat and human intestinal organoids by SARS-CoV-2, Zhou [/bib_ref]. Both human and bat intestinal organoids were successfully infected by SARS-CoV-2 isolated from COVID-19 patients, leading to progressive cytopathic effects after virus replication. In addition, the expression of ACE2 and TMPRSS2 were significantly increased upon induction of organoid differentiation, indicating that SARS-CoV-2 infection occurs in differentiated cells [bib_ref] Infection of bat and human intestinal organoids by SARS-CoV-2, Zhou [/bib_ref]. These findings are consistent with an independent study that revealed infection and replication of the virus in the enterocytes of human small intestinal organoids [bib_ref] SARS-CoV-2 productively infects human gut enterocytes, Lamers [/bib_ref]. As expected, ACE2 is highly expressed in differentiated enterocytes. Infected enterocytes produced viral particles, and elicited a significant upregulation of viral response genes, probably via cytoplasmic sensing of the viral RNA genome [bib_ref] SARS-CoV-2 productively infects human gut enterocytes, Lamers [/bib_ref]. Human ESC-derived colonic organoids were also readily infected by SARS-CoV-2 [bib_ref] Identification of SARS-CoV-2 inhibitors using lung and colonic organoids, Han [/bib_ref]. Transcriptional profiling of colonic organoids indicated that enterocytes were the cell types most susceptible to SARS-CoV-2-entry virus [bib_ref] Identification of SARS-CoV-2 inhibitors using lung and colonic organoids, Han [/bib_ref]. In addition, human duodenum-and ileum-derived organoids were permissive to SARS-CoV-2 infection, and the two mucosa-specific serine proteases, TMPRSS2 and TMPRSS4, activated SARS-CoV-2 spike protein and facilitated virus entry into host cells [bib_ref] TMPRSS2 and TMPRSS4 promote SARS-CoV-2 infection of human small intestinal enterocytes, Zang [/bib_ref]. These studies used human intestinal organoids to support clinical evidence that the gastrointestinal tract is a possible transmission route of SARS-CoV-2.
Several studies have employed intestinal organoids to test drug candidates that might ameliorate gastrointestinal illnesses in the clinic. Krüger and colleagues [bib_ref] Drug inhibition of SARS-CoV-2 replication in human pluripotent stem cell-derived intestinal organoids, Krüger [/bib_ref] utilized ESC-derived intestinal organoids to test the efficiency of three candidate drugs: remdesivir [bib_ref] Remdesivir for the Treatment of Covid-19-Preliminary Report. Reply, Beigel [/bib_ref] , famotidine [bib_ref] Famotidine use is associated with improved clinical outcomes in hospitalized COVID-19 patients:..., Freedberg [/bib_ref] and EK1 [bib_ref] Inhibition of SARS-CoV-2 (previously 2019-nCoV) infection by a highly potent pan-coronavirus fusion..., Xia [/bib_ref]. Results of this study indicate that remdesivir and EK1, but not famotidine, inhibited viral infection and replication [bib_ref] Drug inhibition of SARS-CoV-2 replication in human pluripotent stem cell-derived intestinal organoids, Krüger [/bib_ref]. Similar to the drug reaction in lung organoids, the three drug candidates imatinib, mycophenolic acid and quinacrine dihydrochloride also block SARS-CoV-2 infection in colonic organoids [bib_ref] Identification of SARS-CoV-2 inhibitors using lung and colonic organoids, Han [/bib_ref]. Furthermore, using human colon-derived organoids, Stanifer, et al., demonstrated that pre-treatment of colon-derived organoids with both type I IFN (IFN-β1) and type III IFN (IFN-λ) significantly blocked SARS-CoV-2 infection and this was related to a reduction in viral genome copy number [bib_ref] Critical Role of Type III Interferon in Controlling SARS-CoV-2 Infection in Human..., Stanifer [/bib_ref].
## Brain organoids
A mounting number of COVID-19 cases exhibit neurologic symptoms and neuropsychiatric disorders [bib_ref] Neurological associations of COVID-19, Ellul [/bib_ref] [bib_ref] Effects of COVID-19 on the Nervous System, Iadecola [/bib_ref] , suggesting that the central nervous system (CNS) may be vulnerable to this virus. Human iPSCs-derived brain organoids have allowed several groups to independently examine the neurotropism and neurotoxic effects of SARS-CoV-2 in this pandemic. Ramani, et al., revealed that SARS-CoV-2 could infect human brain organoids within two days of virus exposure [bib_ref] SARS-CoV-2 targets neurons of 3D human brain organoids, Ramani [/bib_ref]. Interestingly, no productive replication of the virus could be observed in neural cells, supporting the hypothesis that SARS-CoV-2 can use the CNS as a long-term reservoir [bib_ref] Is the brain a reservoir organ for SARS-CoV2?, Gomez-Pinedo [/bib_ref]. However, neurodegeneration-like effects consisting of extensive cell death and hyperphosphorylation, were observed in these neurons infected by SARS-CoV-2, associated with misallocation of the structural protein Tau [bib_ref] SARS-CoV-2 targets neurons of 3D human brain organoids, Ramani [/bib_ref]. Similarly, several studies also found that neurons including neural progenitor cells (NPCs) and mature cortical neurons in these organoids were observed to be infected by SARS-CoV-2 [bib_ref] Neuroinvasion of SARS-CoV-2 in human and mouse brain, Song [/bib_ref] [bib_ref] SARS-CoV-2 infects human neural progenitor cells and brain organoids, Zhang [/bib_ref]. On the contrary, Bullen, et al., found an increased accumulation of viral particles in neuronal cells of brain organoids infected with SARS-CoV-2, indicating an active infection and replication of the virus in neurons [bib_ref] Infectability of human BrainSphere neurons suggests neurotropism of SARS-CoV-2, Bullen [/bib_ref]. These contradictory results may be due to differences in experimental conditions between studies by different teams, such as the time of SARS-CoV-2 infection and the adoption of iPSC-derived organoids at different differentiation and developmental stage.
A recent study systematically tested SARS-CoV-2 infection in various region-specific brain organoids of the hypothalamus, midbrain, hippocampus, and cortex [bib_ref] Human pluripotent stem cell-derived neural cells and brain organoids reveal SARS-CoV-2 neurotropism..., Jacob [/bib_ref]. Intriguingly, the choroid plexus epithelium expressed in hippocampal organoids were more vulnerable to the virus than neurons, suggesting that the choroid plexus epithelium is probably the gateway for the entry of the SARS-CoV-2 into the human brain. This possibility is supported by recent findings that show abundant expression of ACE2 and TMPRSS2 in choroid plexus epithelial cells [bib_ref] Histological Evidence for the Enteric Nervous System and the Choroid Plexus as..., Deffner [/bib_ref] [bib_ref] SARS-CoV-2 infects the brain choroid plexus and disrupts the blood-CSF-barrier in human..., Pellegrini [/bib_ref]. The productive SARS-CoV-2 infection of choroid plexus organoids led to cell death, transcriptional upregulation of inflammatory genes, and functional deficits [bib_ref] Human pluripotent stem cell-derived neural cells and brain organoids reveal SARS-CoV-2 neurotropism..., Jacob [/bib_ref]. Pellegrini, et al., also found a stronger tropism of SARS-CoV-2 towards the choroid plexus than neurons within choroid plexus organoid cultures [bib_ref] SARS-CoV-2 infects the brain choroid plexus and disrupts the blood-CSF-barrier in human..., Pellegrini [/bib_ref]. The infection by SARS-CoV-2 damages choroid plexus epithelial cells, leading to leakage of the blood-cerebrospinal fluid barrier which plays a vital role in preventing the entry of pathogens, immune cells, and cytokines into the cerebrospinal fluid and brain [bib_ref] SARS-CoV-2 infects the brain choroid plexus and disrupts the blood-CSF-barrier in human..., Pellegrini [/bib_ref].
Human PSC-based brain organoids have become permissive to high-throughput drug screenings, consequently accelerating the discovery or repurposing of drugs for preventing and treating CNS related COVID-19 symptoms. Sofosbuvir, an FDA (US Food and Drug Administration)-approved nucleotide polymerase inhibitor [bib_ref] Sofosbuvir: A novel treatment option for chronic hepatitis C infection, Bhatia [/bib_ref] , was used as a treatment for SARS-CoV-2 infection. Notably, this drug has been shown to decrease viral accumulation and reduce neuronal death in brain organoids, highlighting the potential abirritation for neurological symptoms. Song, et al., found metabolic changes in SARS-CoV-2infected neurons in brain organoids, but no evidence of type I interferons (IFN-I) response was detected [bib_ref] Neuroinvasion of SARS-CoV-2 in human and mouse brain, Song [/bib_ref]. IgG antibodies against SARS-CoV-2 present in the cerebrospinal fluid from a patient hospitalized with COVID-19 were able to prevent virus infection of brain organoids [bib_ref] Neuroinvasion of SARS-CoV-2 in human and mouse brain, Song [/bib_ref]. Overall, these studies exhibit great potential of human PSC-based brain organoids for probing the infection of SARS-CoV-2 virus in the CNS.
## Liver organoids
As liver damage has been observed in patients with COVID-19 [bib_ref] Liver injury in COVID-19: management and challenges, Zhang [/bib_ref] , experimental platforms including human liver ductal organoids, hepatocyte and cholangiocyte organoids have been used as models of SARS-CoV-2 infection. Yang, et al., established adult liver hepatocyte and cholangiocyte organoids from bile duct epithelial cells and found a high expression of both ACE2 and TMPRSS2 in these organoid derivatives [bib_ref] A human pluripotent stem cell-based platform to study SARS-CoV-2 tropism and model..., Yang [/bib_ref]. Liver hepatocyte and cholangiocyte organoids are permissive to both SARS-CoV-2 pseudo-entry virus and SARS-CoV-2 virus infection. SARS-CoV-2 infection caused significant expression of chemokines and upregulation of inflammatory pathways, as also seen in autopsy samples from patients with COVID-19 [bib_ref] A human pluripotent stem cell-based platform to study SARS-CoV-2 tropism and model..., Yang [/bib_ref]. Zhao, et al., reported a liver ductal organoid model generated from liver bile duct-derived progenitor cells grown in a 3D culture system [bib_ref] Recapitulation of SARS-CoV-2 infection and cholangiocyte damage with human liver ductal organoids, Zhao [/bib_ref]. Cholangiocytes in human liver ductal organoids express ACE2 and TMPRSS2 that enable SARS-CoV-2 infection. Infected cells in liver ductal organoids overexpressed chemokines, formed syncytia and underwent extensive apoptosis with injury of the bile acid transporting functions, resulting in the accumulation or leakage of bile acid and a series of clinical symptoms in COVID-19 patients [bib_ref] Recapitulation of SARS-CoV-2 infection and cholangiocyte damage with human liver ductal organoids, Zhao [/bib_ref]. These observations suggest that the hepatic injury caused directly by SARS-CoV-2 infection should also be taken into account in treating COVID-19 patients. Together, hPSC-derived liver organoids provide a valuable platform for understanding the tropism and pathogenesis of SARS-CoV-2 and discovering prospective anti-viral therapeutics.
## Kidney organoids
Given the presence of renal manifestations such as proteinuria, hematuria, and other classic symptoms in severe COVID-19 patients [bib_ref] Renal complications in COVID-19: Aystematic review and meta-analysis, Kunutsor [/bib_ref] [bib_ref] Acute Kidney Injury in the 2019 Novel Coronavirus Disease, Qian [/bib_ref] and the detection of this virus in the urine of infected individuals [bib_ref] Persistence and clearance of viral RNA in 2019 novel coronavirus disease rehabilitation..., Ling [/bib_ref] , human kidney organoids have also been tested as a model of SARS-CoV-2 infection. Kidney organoids were recently established from human ESCs, and can be directly infected by SARS-CoV-2 [bib_ref] Inhibition of SARS-CoV-2 infections in engineered human tissues using clinical-grade soluble human..., Monteil [/bib_ref]. Consistent with widely expressed ACE2 in human kidney biopsies, ACE2 is expressed in podocytes and proximal tubular cells of human kidney organoids [bib_ref] Inhibition of SARS-CoV-2 infections in engineered human tissues using clinical-grade soluble human..., Monteil [/bib_ref]. Another study generated long term cultures of organoids from human kidney proximal tubular epithelial cells and found that ACE2 was more highly expressed in 3D culture than in 2D cultured cells [bib_ref] Long Term Culture of Human Kidney Proximal Tubule Epithelial Cells Maintains Lineage..., Xia [/bib_ref]. To confirm whether SARS-CoV-2 invades renal cells via ACE2, human recombinant soluble ACE2 (hrsACE2) was added to competitively bind to the virus rather than host cells [bib_ref] Inhibition of SARS-CoV-2 infections in engineered human tissues using clinical-grade soluble human..., Monteil [/bib_ref] [bib_ref] Human soluble ACE2 improves the effect of remdesivir in SARS-CoV-2 infection, Monteil [/bib_ref]. As a result, hrsACE2 significantly decreased SARS-CoV-2 infection of kidney organoids in a dose-dependent manner without causing cell toxicity. also proved the neutralizing effect of their modified long-acting ACE2 variants to resist SARS-CoV-2 in human kidney organoids [bib_ref] A novel soluble ACE2 variant with prolonged duration of action neutralizes SARS-CoV-2..., Wysocki [/bib_ref]. Interestingly, hrsACE2 could markedly improve the effect of remdesivir in SARS-CoV-2 infection in kidney organoid cultures [bib_ref] Human soluble ACE2 improves the effect of remdesivir in SARS-CoV-2 infection, Monteil [/bib_ref]. These observed efficacies of ACE2 in kidney organoids encourage further clinical investigation of this drug alone or in combination with other drugs.
## Cardiac and vascular organoids
Cardiovascular complications are another common condition in patients with severe COVID-19 and increase the risk of mortality [bib_ref] Cardiovascular implications of fatal outcomes of patients with coronavirus disease 2019 (COVID-19), Guo [/bib_ref] [bib_ref] Endothelial cell infection and endotheliitis in COVID-19, Varga [/bib_ref] [bib_ref] COVID-19 and the cardiovascular system, Zheng [/bib_ref]. Human PSC-based cardiac organoids [bib_ref] SARS-CoV-2 infects human pluripotent stem cell-derived cardiomyocytes, impairing electrical and mechanical function, Marchiano [/bib_ref] [bib_ref] BET inhibition blocks inflammation-induced cardiac dysfunction and SARS-CoV-2 infection, Mills [/bib_ref] and vascular organoids [bib_ref] Inhibition of SARS-CoV-2 infections in engineered human tissues using clinical-grade soluble human..., Monteil [/bib_ref] have been established as potential in vitro models to study SARS-CoV-2 infection in the cardiovascular system. The high expression of ACE2 has been observed in cardiomyocytes, endothelial cells, vascular smooth muscle cells, and pericytes lining blood vessels [bib_ref] Cell type-specific expression of the putative SARS-CoV-2 receptor ACE2 in human hearts, Nicin [/bib_ref] [bib_ref] Myocyte-Specific Upregulation of ACE2 in Cardiovascular Disease: Implications for SARS-CoV-2-Mediated Myocarditis, Tucker [/bib_ref].
Mills et al., developed a high-throughput hPSCcardiac organoid platform, using it to screen 105 small molecules with regenerative potential. These organoids could also be a feasible platform to study virus infection and screen antiviral drugs [bib_ref] Drug screening in human PSC-cardiac organoids identifies pro-proliferative compounds acting via the..., Mills [/bib_ref]. Recently, by using human cardiac organoids, showed that bromodomain and extraterminal family inhibitors (BETi) reduced ACE2 expression, decreased transcription of genes in the viral response, and blocked SARS-CoV-2 infection of cardiomyocytes and inflammation-induced cardiac dysfunction [bib_ref] BET inhibition blocks inflammation-induced cardiac dysfunction and SARS-CoV-2 infection, Mills [/bib_ref]. By using iPSC-derived blood vessel organoids, Monteil, et al., validated that SARS-CoV-2 was able to infect human blood vessels [bib_ref] Inhibition of SARS-CoV-2 infections in engineered human tissues using clinical-grade soluble human..., Monteil [/bib_ref]. Importantly, the virus replication was detected in these organoids after infection. The addition of clinical-grade hrsACE2 dramatically reduced SARS-CoV-2 infections of human blood vessel organoids [bib_ref] Inhibition of SARS-CoV-2 infections in engineered human tissues using clinical-grade soluble human..., Monteil [/bib_ref]. Consistently, organoids infected with mixtures of SARS-CoV-2 virus particles and variable concentrations of hrsACE2 have markedly decreased levels of intracellular viral RNA [bib_ref] Inhibition of SARS-CoV-2 infections in engineered human tissues using clinical-grade soluble human..., Monteil [/bib_ref]. These findings may explain how the virus spreads through the body and leads to organ damage in severely ill individuals.
## Eye organoids
Recent clinical evidence has shown that a minority of COVID-19 patients present with ocular symptoms [bib_ref] SARS-CoV-2 Isolation From Ocular Secretions of a Patient With COVID-19 in Italy..., Colavita [/bib_ref] [bib_ref] Ocular Findings and Proportion with Conjunctival SARS-COV-2 in COVID-19 Patients, Zhou [/bib_ref]. Importantly, SARS-CoV-2 entry factors ACE2 and TMPRSS2 have been found to be expressed in the limbal, corneal, and conjunctival epithelium of the eye and their organoid derivatives [bib_ref] Expression of endogenous angiotensin-converting enzyme 2 in human induced pluripotent stem cell-derived..., Lai [/bib_ref] [bib_ref] Co-expression of SARS-CoV-2 entry genes in the superficial adult human conjunctival, limbal..., Collin [/bib_ref] [bib_ref] SARS-CoV-2 infection of ocular cells from human adult donor eyes and hESC-derived..., Makovoz [/bib_ref] [bib_ref] ACE2 and TMPRSS2 are expressed on the human ocular surface, suggesting susceptibility..., Zhou [/bib_ref]. Also, SARS-CoV-2 viruses have been detected in the conjunctiva of infected rhesus monkeys [bib_ref] Coronavirus disease 2019 (SARS-CoV-2) and colonization of ocular tissues and secretions: a..., Aiello [/bib_ref]. These findings indicate that ocular surfaces may serve as an additional entry vector for SARS-CoV-2. Therefore, eyes are at risk for infection by SARS-CoV-2 and are a route warranting protection. Human iPSC-derived whole-eye organoid models, which are comprised of cells from the retina, cornea, ciliary margin, retinal pigment epithelium, lens and iris as well as cell monolayers grown from limbal, corneal and conjunctival epithelium, have been established to study SARS-CoV-2 ocular infection [bib_ref] SARS-CoV-2 infection of ocular cells from human adult donor eyes and hESC-derived..., Makovoz [/bib_ref]. In fact, the corneal limbus seems to be the most susceptible to infection, consistent with its high expression of ACE2 and TMPRSS2, essential proteins that mediate SARS-CoV-2 viral entry. This means that the cornea may be the entrance and proliferation site of virus infection. Although IFN-I and IFN-III were not detected, human eye organoids mounted a significant inflammatory response indicative of NF-κB signaling [bib_ref] SARS-CoV-2 infection of ocular cells from human adult donor eyes and hESC-derived..., Makovoz [/bib_ref]. Similar findings were obtained from ocular biopsies from human beings infected with SARS-CoV-2, supporting the utility of eye organoids to study SARS-CoV-2 infection and identify prophylactics that may protect the eyes from infection [bib_ref] SARS-CoV-2 infection of ocular cells from human adult donor eyes and hESC-derived..., Makovoz [/bib_ref]. Future studies are required to better clarify how infection in the eye may lead to transmission to other regions of the body.
# Limitations and perspectives
Exploring the life cycle of SARS-CoV-2 and unearthing the efficacy of drugs can be made more objective by using organoid models. Organoids models can simulate the pathology of COVID-19 in corresponding tissues, and provide a valuable platform for unraveling virus infectivity, tropisms, replication kinetics and potential treatments [bib_ref] Harnessing the power of novel animal-free test methods for the development of..., Busquet [/bib_ref] [bib_ref] In vitro and animal models for SARS-CoV-2 research, Takayama [/bib_ref]. Although a range of new studies demonstrated many more advantages of human organoids, these models still have some limitations. The microenvironment of organoids differs from the native organ to some extent due to the absence of additional cell types such as fibroblasts, vascular cells, neural cells and immune cells. In general, iPSC-derived organoids can contain mesenchymal lineages such as fibroblasts, while ASC-derived organoids consist exclusively of epithelial cells [bib_ref] The Organoid Platform: Promises and Challenges as Tools in the Fight against..., Geurts [/bib_ref]. Moreover, organoid models cannot fully represent the systemic symptoms associated with whole body responses to this disease. This means virology research using organoids will still need to be validated in animal models and clinical studies. Anyhow, it is worth noting that organoids exhibit greater complexity and resemble the true organ more closely than any other 2D cultured cells.
Immune cells play a vital role in the fight against SARS-CoV-2 infection [bib_ref] Immunology of COVID-19: Current State of the Science, Vabret [/bib_ref]. Organoids derived from non-small-cell lung cancer and colorectal cancer have been co-cultured with T cells from individual patients [bib_ref] Tumor organoid-T-cell coculture systems, Cattaneo [/bib_ref] [bib_ref] Generation of Tumor-Reactive T Cells by Co-culture of Peripheral Blood Lymphocytes and..., Dijkstra [/bib_ref]. The co-culture system showed successful antigen presentation of tumor organoid cells to T cells, which subsequently became activated [bib_ref] Generation of Tumor-Reactive T Cells by Co-culture of Peripheral Blood Lymphocytes and..., Dijkstra [/bib_ref]. Immune co-cultured organoids can also be applied to virus infection research to understanding COVID-19 immune responses. Furthermore, several studies have reported promising organoid-based findings, including the bioengineering of a scaffold-guided functional intestine using a bioreactor [bib_ref] Engineering transplantable jejunal mucosal grafts using patient-derived organoids from children with intestinal..., Meran [/bib_ref] [bib_ref] Homeostatic mini-intestines through scaffold-guided organoid morphogenesis, Nikolaev [/bib_ref] and organoids-on-a-chip [bib_ref] Organoids-on-a-chip, Park [/bib_ref]. Such physiologically relevant culture systems, as well as functional simulation of immune responses in bioengineered organoids, will open up new perspectives for disease modelling to help fight the current pandemic.
## Abbreviations
[fig] Figure 1: Schematic representation of the main organs affected by SARS-COV-2. [/fig]
[fig] Figure 2: ducts isolated from human liver biopsies ACE2 and TMPRSS2 are involved in viral entry[64]; SARS-CoV-2 infection impairs bile acid transportation functions[64]. / Common analyses and key applications of human PSC-and ASC-derived organoid platforms in COVID-19 research. Various organoids derived from PSCs or ASCs have been established to study SARS-CoV-2 infection. Organoid models are commonly used to investigate SARS-CoV-2 tropism and COVID-19 pathophysiology across different organs, as well as to verify the safety and efficacy of candidate drugs and screen new therapeutic strategies. Furthermore, patient-derived organoids may potentially serve as a platform to test the efficacy of antiviral drugs for individual patients. [/fig]
[table] Table 1: List of human PSC-derived organoids used to study SARS-CoV-2 infection [/table]
[table] / Table 2: Human hepatocyte and cholangiocyte organoids are highly permissive to SARS-CoV-2 infection[63]; Liver organoids show similar chemokine response as COVID-19 patients[63].Kidney organoids express ACE2 and TMPRSS2[68]; SARS-CoV-2 can directly infect kidney organoids[68]; Human recombinant soluble ACE2 reduce SARS-CoV-2 infection in kidney organoids[68,71]; Combination therapy using Remdesivir with recombinant soluble ACE2 reduces virus entry and replication[70].BET inhibitors reduce ACE2 expression, decrease transcription of genes in the viral response, and block SARS-CoV-2 infection of cardiomyocytes and inflammation-induced cardiac dysfunction[76]. List of ASC-derived organoids used to study SARS-CoV-2 infection Ciliated cells and alveolar cells are susceptible to SARS-CoV-2 infection, and ciliated cells could be infected prior to the alveoli [/table]
|
A novel weighting method to remove bias from within-subject exposure dependency in case-crossover studies
Background: Case-crossover studies have been widely used in various fields including pharmacoepidemiology. Vines and Farrington indicated in 2001 that when within-subject exposure dependency exists, conditional logistic regression can be biased. However, this bias has not been well studied.Methods:We have extended findings by Vines and Farrington to develop a weighting method for the case-crossover study which removes bias from within-subject exposure dependency. Our method calculates the exposure probability at the case period in the case-crossover study which is used to weight the likelihood formulae presented by Greenland in 1999. We simulated data for the population with a disease where most patients receive a cyclic treatment pattern with within-subject exposure dependency but no time trends while some patients stop and start treatment. Finally, the method was applied to real-world data from Japan to study the association between celecoxib and peripheral edema and to study the association between selective serotonin reuptake inhibitor (SSRI) and hip fracture in Australia.Results:When the simulated rate ratio of the outcome was 4.0 in a case-crossover study with no time-varying confounder, the proposed weighting method and the Mantel-Haenszel odds ratio reproduced the true rate ratio. When a time-varying confounder existed, the Mantel-Haenszel method was biased but the weighting method was not. When more than one control period was used, standard conditional logistic regression was biased either with or without time-varying confounding and the bias increased (up to 8.7) when the study period was extended. In real-world analysis with a binary exposure variable in Japan and Australia, the point estimate of the odds ratio (around 2.5 for the association between celecoxib and peripheral edema and around 1.6 between SSRI and hip fracture) by our weighting method was equal to the Mantel-Haenszel odds ratio and stable compared with standard conditional logistic regression.Conclusion:Case-crossover studies may be biased from within-subject exposure dependency, even without exposure time trends. This bias can be identified by comparing the odds ratio by the Mantel-Haenszel method and that by standard conditional logistic regression. We recommend using our proposed method which removes bias from within-subject exposure dependency and can account for time-varying confounders.
# Background
The case-crossover design has been widely used since it was proposed in 1991 [bib_ref] The case-crossover design: a method for studying transient effects on the risk..., Maclure [/bib_ref]. The design has been used in various fields including pharmacoepidemiology [bib_ref] Case-crossover study design in pharmacoepidemiology: systematic review and recommendations, Consiglio [/bib_ref] , occupational epidemiology [bib_ref] Selecting appropriate study designs to address specific research questions in occupational epidemiology, Checkoway [/bib_ref] , studies on traffic safety [bib_ref] The epidemiologic principles underlying traffic safety study designs, Kim [/bib_ref] and air pollution health effects [bib_ref] Case-crossover analysis of air pollution health effects: a systematic review of methodology..., Carracedo-Martínez [/bib_ref] [bib_ref] Bias in the case-crossover design: implications for studies of air pollution, Lumley [/bib_ref]. In case-crossover studies, individuals who have experienced the outcome (cases) act as their own controls by including one or more periods before the onset of the outcome. The period including the outcome is the case period, while period(s) prior to the case period act as the controls. The number of control periods can be large: for example, in the original article [bib_ref] The case-crossover design: a method for studying transient effects on the risk..., Maclure [/bib_ref] , one analysis involved 8766 control periods. The effect of the exposure should be brief; exposure in any period should affect the outcome in that period only, without any 'carryover effect' [bib_ref] The case-crossover design: a method for studying transient effects on the risk..., Maclure [/bib_ref] [bib_ref] Bias in the case-crossover design: implications for studies of air pollution, Lumley [/bib_ref] [bib_ref] Case-control studies, Rothman [/bib_ref]. In addition, in case-crossover studies, the rate of outcome occurrence is usually assumed to be unchanged during exposed or unexposed periods, respectively. However, like other case-only studies, the case-crossover study has an advantage that the effect of time-invariant confounders is automatically controlled because the case period is compared with control period(s) of the same individual [bib_ref] Case-control studies, Rothman [/bib_ref] [bib_ref] Exchangeability in the case-crossover design, Mittleman [/bib_ref] [bib_ref] When should case-only designs be used for safety monitoring of medical products?, Maclure [/bib_ref]. The case-crossover study also has unique characteristics. For example, the original unidirectional case-crossover study does not include periods after the outcome occurs. Thus, there is no bias due to the outcome influencing future exposures or future observation periods, unlike other case-only studies such as self-controlled case series [bib_ref] Case series analysis for censored, perturbed, or curtailed post-event exposures, Farrington [/bib_ref] [bib_ref] Self-controlled case series analysis with event-dependent observation periods, Farrington [/bib_ref] [bib_ref] Self controlled case series methods: an alternative to standard epidemiological study designs, Petersen [/bib_ref].
However, case-crossover studies are susceptible to at least two types of major biases. The first is bias due to time trends in the exposure which can be removed using a variant of case-crossover studies, the case-time-control design [bib_ref] When should case-only designs be used for safety monitoring of medical products?, Maclure [/bib_ref] [bib_ref] Use of self-controlled designs in pharmacoepidemiology, Hallas [/bib_ref] [bib_ref] The case-time-control design, Suissa [/bib_ref]. The second is bias due to withinsubject exposure dependency or autocorrelation in an individual's exposure history [bib_ref] Bias in the case-crossover design: implications for studies of air pollution, Lumley [/bib_ref] [bib_ref] Within-subject exposure dependency in casecrossover studies, Vines [/bib_ref] [bib_ref] On the validity of the case-time-control design for autocorrelated exposure histories, Jensen [/bib_ref]. This bias is particularly important in pharmacoepidemiology because drug use on 1 day is rarely independent from use in the preceding days. However, the potential for this bias has had little attention in the pharmacoepidemiology literature, and standard conditional logistic regression has been used without assessing whether bias due to withinsubject exposure dependency exists in many case-crossover studies [bib_ref] Dental procedures, antibiotic prophylaxis, and endocarditis among people with prosthetic heart valves:..., Tubiana [/bib_ref] [bib_ref] C-reactive protein and risk of venous thromboembolism: results from a population-based case-crossover..., Grimnes [/bib_ref] [bib_ref] A case-crossover study of urological chronic pelvic pain syndrome flare triggers in..., Sutcliffe [/bib_ref] [bib_ref] Association between ranibizumab injections and risk of acute myocardial infarction in age-related..., Ryu [/bib_ref] [bib_ref] Mortality associated with wildfire smoke exposure in Washington state, Doubleday [/bib_ref]. This may be in part because this bias has been considered rather minor when exposure is stationary [bib_ref] Within-subject exposure dependency in casecrossover studies, Vines [/bib_ref].
In this paper, we show that within-subject exposure dependency may produce unignorably large bias even if exposure is stationary. We also describe how to remove bias from within-subject exposure dependency, when time trends in the exposure do not exist.
## Motivating examples
Our motivation for investigating bias in case-crossover studies due to within-subject exposure dependency was a hypothetical drug treatment with a specific cyclic exposure pattern, where standard conditional logistic regression may produce bias. We also outline realworld pharmacoepidemiological studies in Japan and Australia where our proposed method is used.
## Drug treatment with specific exposure pattern
We illustrate our motivating example using a drug treatment with a cyclic pattern consisting of two periods of treatment followed by one period with no treatment. The exposure pattern clearly shows autocorrelation, with the probability of exposure in one treatment period dependent on previous periods. We show that estimates of the odds ratio for the exposureoutcome association may be biased, depending on the number of control periods chosen.
In a hypothetical unidirectional case-crossover study with 4 periods (one case period and 3 control periods), only 3 exposure patterns are possible. If exposure and non-exposure are indicated by 1 and 0, respectively, the 3 exposure patterns are (1101), (1011), and (0110). Therefore, subjects with an exposed case period can have only 1 unexposed and 3 exposed periods, while those with unexposed case period can have only 2 unexposed and 2 exposed periods. In such a casecrossover study, when the true odds ratio (OR) is 4.0, the OR is underestimated as 3.2 using standard conditional logistic regression for matched case-control studies.
However, if there are two control periods, the possible exposure patterns are (110), (101), and (011). In this case, irrespective of whether the case period is exposed or unexposed, all subjects have 1 unexposed and 2 exposed periods and there is no bias in the estimate of the OR. Similarly, there is no bias when the total number of periods (including the case period) is an integral number of 3 (the cyclic pattern).
## Overview of real-world examples
We describe two real world pharmacoepidemiological studies which will be analyzed by our proposed method. The first is an investigation into a previously reported association between celecoxib and peripheral edema [bib_ref] The validity of sequence symmetry analysis (SSA) for adverse drug reaction signal..., Wahab [/bib_ref] using Japanese data from 25 corporate-type health insurance plans . From 99,821 new users of celecoxib aged between 20 and 74 years from May 2013 to April 2018, who used celecoxib after at least 180 days of non-use, we selected 311 cases who experienced peripheral edema and had both exposed and unexposed days during an 84-day study period.
The second example is an Australian study of the association between hip fracture and psychoactive medicines [bib_ref] Psychoactive medicine use and the risk of hip fracture in older people:..., Leach [/bib_ref] [bib_ref] Risk of hip fracture in older people using selective serotonin reuptake inhibitors..., Leach [/bib_ref]. The data were obtained from the Australian Department of Veterans' Affairs administrative claims database. Psychoactive medicines included benzodiazepines, selective serotonin re-uptake inhibitors (SSRIs), opioids, antipsychotics and tricyclic antidepressants. The hip fracture cases were 8828 patients aged over 65 years who were hospitalized between 2009 and 2012. A previous case-crossover study of this cohort found an increased risk of hip fracture for opioids, SSRIs and antipsychotics [bib_ref] Psychoactive medicine use and the risk of hip fracture in older people:..., Leach [/bib_ref]. A related case-control study found an association between hip fracture and SSRIs when used concurrently with other psychoactive medicines [bib_ref] Risk of hip fracture in older people using selective serotonin reuptake inhibitors..., Leach [/bib_ref].
# Methods
To avoid bias due to within-subject exposure dependency, the likelihood for estimating the odds ratio for the exposure-outcome association may be modified. At least two approaches are effective to modify the likelihood and both involve two-step weighting procedures. In the first approach, the probability of each exposure permutation is estimated in Step 1 and used as weights in the likelihood in Step 2. In the special case where this probability is the same for every permutation ("global exchangeability"), the modified likelihood is equivalent to that for the standard conditional logistic regression.
In this paper, we propose another approach that can be used if "pairwise exchangeability" is assumed. Pairwise exchangeability holds where there are no time trends in the exposure i.e., the proportion exposed in the population is stationary over the study period [bib_ref] Within-subject exposure dependency in casecrossover studies, Vines [/bib_ref]. In this scenario, the probabilities that the case period is exposed and unexposed are estimated in Step 1 and used as weights in Step 2. When there is pairwise exchangeability, an unbiased OR can be obtained by the Mantel-Haenszel method as well.
## Bias due to conditional logistic regression for case-crossover studies
In 2001, Vines and Farrington proposed the likelihood for case-crossover studies as [bib_ref] Within-subject exposure dependency in casecrossover studies, Vines [/bib_ref] :
where X i0 is the exposure level at the case period (m = 0) and X im is the exposure level at the m-th control period at t (t = − m: m = 1, 2, −-, M), x i0 denotes
[formula] (1) L = N � i=1 exp � x i0 �∑ * P � X i0 = x i * (0) , − − −, X iM = x i * (M) � ∑ exp � x i (0) � P � X i0 = x i (0) , − − −, X iM = x i (M) � [/formula]
the observed exposure level at the case period and x im denotes the observed exposure level at the m-th control period, the sum in the denominator ranges over all permutations of κ of the integers {0, 1, −--, M} and the sum in the numerator ranges over the subset of the permutations for which x iκ(0) = x i0 of the individual i. In Eq. (1), X i0 may be a binary exposure variable but can be a multi-level exposure variable or a vector of an exposure and time-varying confounders. When X im denotes a binary exposure, the denominator in Eq. (1) becomes
[formula] exp ( ) ∑ 1 P � X i0 = 1, N exposed = k � + ∑ 0 P � X i0 = 0, N exposed = k � [/formula]
where N exposed is the total number of exposed (case and control) periods, given by N exposed = M m=0 X im and κ 1 P X i0 = 1, N exposed = k is the sum of probabilities for all the permutations of k exposed and (M + 1-k) unexposed periods where X i0 = 1, and κ 0 P X i0 = 0, N exposed = k is that where X i0 = 0. Data on N exposed = k is informative only when the positivity (non-zero probability) condition is satisfied or ∑P(X i0 = 1, N exposed = k) > 0 and ∑P(X i0 = 0 N exposed = k) > 0. Otherwise, they do not contribute to the estimation of exp(β).
Let OR VF be the estimate of exp(β) obtained from Eq. (1). The likelihood in Eq. (1) and OR VF are in general different from the following likelihood for the standard conditional logistic (SCL) regression for individually matched case-control studies and its estimate, OR SCL .
Vines and Farrington showed that the likelihoods in Eqs. [bib_ref] The case-crossover design: a method for studying transient effects on the risk..., Maclure [/bib_ref] and [bib_ref] Case-crossover study design in pharmacoepidemiology: systematic review and recommendations, Consiglio [/bib_ref] } for all permutations κ of {0, 1, −---, M}, that is, global exchangeability holds. For example, Eqs. (1) and (2) are equivalent when the exposure status in one period is independent from the status in any other periods and the exposure probability is the same in all of case and control periods (Appendix 1, Additional File 1). If global exchangeability does not hold, OR SCL can be biased.
[formula] are equivalent if P{X i0 = x i0 , − − −, X i M = x iM } = P{X i0 = x iκ(0) , − − −, X iM = x iK [/formula]
Vines and Farrington did not show explicitly how to estimate P{X i0 = x iκ(0) , − − −, X iM = x iκ(M) } in Eq. (1). These probabilities may be estimated by the proportion of each permutation in the population which contain cases, or in samples representing the population such as time-controls in the case-time-control design proposed by Suissa [bib_ref] The case-time-control design, Suissa [/bib_ref]. In the next section, however, we introduce a different approach to remove the bias from within-subject exposure dependency by assuming that pairwise exchangeability is satisfied but global exchangeability may not necessarily hold.
[formula] (2) L = N i=1 exp (βx i0 ) M j=0 exp βx ij [/formula]
Weighting method to remove BIAS due to within-exposure dependency Case-crossover studies with a binary exposure When pairwise exchangeability, P{X i0 = 1, X im = 0} = P(X i0 = 0, X im = 1} is satisfied for a binary exposure in all control periods m (m = 1, 2, −-, M), the estimate of exp(β) using the Mantel-Haenszel method (OR MH ) is unbiased whether within-subject exposure dependency exists or not [bib_ref] Within-subject exposure dependency in casecrossover studies, Vines [/bib_ref]. In line with this finding, we will show that when pairwise exchangeability holds, the probabilities that the individual is unexposed (π 0 ) and exposed (π 1 ) at the case period, can be estimated from the cases in a case-crossover study (without requiring data from the population or time-controls). Once π 0 , π 1 , and the relative exposure probability π 10 = π 1 / π 0 are estimated, the following likelihood for casecrossover studies proposed by Greenland [bib_ref] A unified approach to the analysis of case-distribution (case-only) studies, Greenland [/bib_ref] can be used to obtain an unbiased estimate of exp(β), defined as OR G :
In the left-hand side of Eq. (3), π ik is the probability that individual i has the k-th exposure level at the case period (k = 0, 1 for binary exposure), π ic is π ik observed, and x ic is the exposure level observed when individual i has the outcome. In the middle of Eq. (3), subscript i is omitted in the exposure probabilities as π ik is replaced by the expected value in the population in the current study. In the right-hand side of Eq. (3) π k0 = π k /π 0 . Eq. (3) stands for the model where x ik is a binary variable, multi-level exposure variable, or a vector of the observed exposure and time-varying confounders.
For a binary exposure, the right-hand side of Eq. (3) can be rewritten as where x ic = 1 and π c0 = π 10 = π 1 /π 0 when the individual i was exposed at the case period and x ic = 0 and π c0 = π 00 = π 0 /π 0 = 1 when unexposed. As Vines and Farrington showed [bib_ref] Within-subject exposure dependency in casecrossover studies, Vines [/bib_ref] , Greenland did not estimate π k for case-crossover studies in Eq. (3) when within-subject exposure dependency exists.
We outline a novel weighting method using a modified version of Greenland's likelihood. We propose that π k can be estimated, with or without within-subject exposure dependency, by assuming pairwise exchangeability. Let P kl denote the joint probability that the subject has the k-th exposure level at the case period and has the l-th exposure level at the m-th control period:
[formula] (3) L = N � i=1 exp � x ic � ic ∑ k exp � x k � ik = N � i=1 exp � x ic � c ∑ k exp � x k � k = N � i=1 exp � x ic � c0 ∑ k exp � x k � k0 (4) L = N i=1 exp βx ic π c0 1 + exp (β)π 10 (5) P kl[m] = P(X 0 = x k , X m = x l ) m = 1, 2, − − −, M [/formula]
In Eq. (5), X 0 is the exposure status at the case period and X m is the exposure status at the m-th control period. When the exposure variable is binary, both X 0 and X m have two levels (k = 0, 1 and l = 0,1) and pairwise exchangeability is equivalent to stationary exposure (no time trends): when the exposure process is stationary, P 10[m] + P = P 01[m] + P and this relationship leads to the pairwise exchangeability condition P [bib_ref] Case series analysis for censored, perturbed, or curtailed post-event exposures, Farrington [/bib_ref]
[formula] [m] = P 01[m] (m = 1, 2, −-, M). [/formula]
Using conditional probabilities, pairwise exchangeability, P{X i0 = 1, X im = 0} = P(X i0 = 0, X im = 1} can be rewritten as:
where π 0 = P(X 0 = 0) and π 1 = P(X 0 = 1). When both sides of Eq. (6) are summed up over M control periods (m = 1, 2, −-, M), we obtain:
The quantity M m=1 P( X m = 0|X 0 = 1) can be estimated by the average number of unexposed control periods (where X m = 0) in those exposed at the case period (X 0 = 1). Similarly, the quantity M m=1 P( X m = 1|X 0 = 0) can be estimated as the average number of exposed control periods (X m = 1) in those unexposed at the case period (X 0 = 0). This average, defined as PT 10 and PT 01 , respectively, can be written as:
where PT 10i is the number of unexposed control periods (person-time) of case i who is exposed at the case period, PT 01i is the number of exposed control periods (person-time) of case i who is unexposed at the case period, and a 1 is the number of discordant exposed cases with at least one unexposed control period and a 0 is the number of discordant unexposed cases with at least one exposed control period. When M m=1 P( X m = 0|X 0 = 1) and M m=1 P( X m = 1|X 0 = 0) in Eq. (7) are substituted by PT 10 and PT 01 , respectively, we obtain:
When standard statistical software is used for conditional logistic regression analysis of case-crossover studies, we introduce weighting to ensure the denominator in Eq. (2) equals that in Eq. (4). Every exposed and unexposed period in case i should be weighted by w i1 and by w i0 , respectively, defined as follows: where m 1 i is the number of exposed periods and m 0 i is the number of the unexposed periods (including both case and control periods) in case i and π 10 is estimated from Eq. [bib_ref] When should case-only designs be used for safety monitoring of medical products?, Maclure [/bib_ref]. In most statistical software, w ik (k = 0,1) may be specified as an offset variable in conditional logistic regression which uses the following likelihood:
[formula] (6) π 1 P( X m = 0|X 0 = 1) = π 0 P( X m = 1|X 0 = 0) (7) π 1 ∑ M m=1 P X m = 0|X 0 = 1 = π 0 ∑ M m=1 P X m = 1|X 0 = 0 (8) PT 10 = i PT 10i /a 1 and PT 01 i PT 01i /a 0 (9) 10 = 1 ∕ 0 = ∑ M m=1 P X m = 1|X 0 = 0 ∕ ∑ M m=1 P X m = 0|X 0 = 1 = PT 01 ∕PT 10( [/formula]
where w ij = w i1 and w ij = w i0 when j-th period is exposed and unexposed, respectively, in case i.
Using a 1 and a 0 , Eq. (4) can be rewritten as Equation (12) gives (see Additional File 1, Appendix 2) the following maximum likelihood estimate for OR G :
When π 10 estimated by Eq. (9) is considered as a constant, the variance is given by:
In order to allow for the variance of the weights estimated by Eq. (9), we recommend the use of bootstrapping to estimate the 95% confidence interval using the 2.5 to 97.5 percentiles of OR G .
From Eqs. (8), [bib_ref] When should case-only designs be used for safety monitoring of medical products?, Maclure [/bib_ref] and (13) we obtain:
Equation [bib_ref] Within-subject exposure dependency in casecrossover studies, Vines [/bib_ref] shows that when the model involves only one binary exposure variable, the point estimate of OR G is the same as OR MH .
## Case-crossover studies with a binary exposure and a binary time-varying confounder
The Mantel-Haenszel estimator is unbiased for binary exposures when there is pairwise exchangeability. When there is a binary time varying confounder (z), Maclure in his original proposal of the case-crossover design recommended further stratification when using the Mantel-Haenszel method [bib_ref] When should case-only designs be used for safety monitoring of medical products?, Maclure [/bib_ref]. In a simulation, we followed this recommendation and stratified each subject by z. As a result, control periods where z = 0 were excluded from subjects with z = 1 at the case period. Similarly, control periods where z = 1 were excluded from those with z = 0 at the case period. We have shown that excluding these [bib_ref] Self-controlled case series analysis with event-dependent observation periods, Farrington [/bib_ref]
[formula] L = N i=1 exp (βx i0 ) M j=0 w ij exp βx ij (12) L = exp (β)π 10 1 + exp(β)π 10 a 1 1 1 + exp(β)π 10 a 0 (13) OR G = exp (β) = a 1 a 0 1 π 10 (14) v(β) = 1 a 0 + 1 a 1 (15) OR G = exp (β) = a 1 a 0 1 π 10 = a 1 PT 10 a 0 PT 01 = i PT 10i i PT 01i = OR MH [/formula]
control periods can lead to bias (see [fig_ref] Table 4: Estimates of OR SCL , OR VF and OR MH [/fig_ref] in Results section).
On the other hand, our method can be extended to studies with time-varying confounder to analyze data of all case and control periods and all periods can be included in the analysis. For example, when there is a binary exposure (x) and a binary time-varying confounder variable (z), βx ij in the likelihood in Eq. (11) is
[formula] specified as (β, γ)(x ij , z ij ) T , which is equal to 0 when (x ij , z ij ) =(0,0), β when (x ij , z ij ) =(1,0), γ when (x ij , z ij ) =(0,1), and β + γ when (x ij , z ij ) = (1,1) where exp(γ) [/formula]
is an estimate of the odds ratio of z. Similarly to the finding that OR SCL in Eq. (2) is unbiased when within-subject exposure dependency does not exist (Appendix 1, Additional File 1), we may estimate an unbiased OR G in the model involving the exposure and time-varying confounder by calculating the weight from the exposure variable x (Eq. 10), if within-subject dependency does not exist for z during exposed periods (where the probability that the confounder is positive is f 1 ) and during unexposed periods (f 0 ) and the confounder is adjusted for as in the standard conditional logistic regression. Similarly as for a case-crossover study with a binary exposure, we recommend bootstrapped 2.5 to 97.5 percentiles of OR G to estimating the 95% CI to allow for the variance of the weights (see Appendix 3, Additional File 1 for the detail; as to the relevant SAS codes, see 4-1 h and 4-2f in Appendix 4, Additional File 1).
## Simulation studies
We simulated data relevant to drug therapy with no time trends and with autocorrelated exposure patterns (within-subject dependency) (Appendix 4, Additional File 1). The simulated data is created based on the following observations (i) drug treatment sometimes has a cyclic pattern which often produces within-subject exposure dependency, (ii) some outcomes (e.g., acute adverse events) tend to occur soon after the treatment is initiated but they may also occur later during the drug therapy, and (iii) some patients stop treatment for various reasons while some patients start treatment. When the rate of stopping treatment is the same as that of starting treatment, the stationarity of the exposure may be maintained in the population as follows. We simulated scenarios with and without time varying confounding.
Assume that drug treatment involves a cyclic pattern where one cycle consists of 7 days and a patient has a drug on days 1 and 4 but no drug on days 2, 3, 5, 6, and 7. [fig_ref] Figure 1: Exposure pattern in a hypothetical dynamic population of 80,000 patients [/fig_ref] depicts three cycles of drug treatment with 8 subgroups consisting of 1 case period and 21 control periods, where 1 period is 1 day. Subgroup A represents stoppers who stop treatment at the case period while Subgroup H represents new users who start treatment at the case period. Subgroups B to G represent patients being treated with a different timing relative to the start of the treatment cycle. In [fig_ref] Figure 1: Exposure pattern in a hypothetical dynamic population of 80,000 patients [/fig_ref] , the proportion of those exposed to a drug in the population is always ¼, indicating stationarity (no time trends). [fig_ref] Figure 2: Expected frequency of cases by subgroup in [/fig_ref] shows 140 cases who had the outcome at the case period when the size of each of Subgroups A to H, N = 10,000, the event rate in an unexposed period (r 0 ) is 0.001 per period and the rate ratio is 4. We assume that the expected number of cases is determined by exposure at the case period only, or Nr 0 when unexposed and N RR r 0 when exposed at the case period. [fig_ref] Figure 2: Expected frequency of cases by subgroup in [/fig_ref] shows 184 cases who had the outcome at the case period when N = 10,000, the event rate in the unexposed period without time-varying confounding is 0.001 per period and the rate ratio is 4 and 2 for the exposure and time-varying confounder, respectively. The status of the time-varying confounder in the unexposed or exposed period is related to the exposure status of that period only, and f 0 = 0.2 and f 1 = 0.4 where f 0 and f 1 are the expected values of the proportion of exposed periods and unexposed periods in the population, respectively, when the confounder is positive (See Appendix 3, Additional File 1). We assume that the number of cases is as expected, or
[formula] Nr 0 (1 − f 0 ), N RR z r 0 f 0 , N RR r 0 (1 − f 1 ), [/formula]
and N RR RR z r 0 f 1 , when the combination of the exposure and time-varying confounder variables (x, z) at the case period is (0, 0), (0, 1), (1, 0), and (1, 1), respectively. The status of the time-varying confounder in the control periods was randomly generated and is therefore independent of the exposure or time-varying confounder at different periods.
To determine the effect of the length of each time period on the estimated odds ratio, we also analyzed the data by dividing 22 days in [fig_ref] Figure 1: Exposure pattern in a hypothetical dynamic population of 80,000 patients [/fig_ref] into 11, 7, 5, 3, and 2 periods (M = 10, 6, 4, 2, and 1) where 1 period included 2, 3, 4, 7 and 11 days, respectively. The status of the exposure and time varying confounder was defined by the last day of each period (Definition I in . We analyzed the data by standard conditional regression, the Vines and Farrington method, Mantel-Haenszel methods, and our modified Greenland's method for the fixed study period of 22 days. With time-varying confounder (z), case and control periods of each subject were further stratified by z in the Mantel-Haenszel method. We used bootstrapping to estimate 2.5 to 97.5 percentiles for OR G . Data simulation and analyses were performed using SAS 9.4 and SAS codes including those for bootstrap method are shown in Appendix 4 in Additional File 1.
## Case-crossover studies of real-world data
The method was also applied to data from Japanese and Australian databases. The Japanese study on the association between celecoxib and peripheral edema from a previous study [bib_ref] The validity of sequence symmetry analysis (SSA) for adverse drug reaction signal..., Wahab [/bib_ref] was approved by the ethics committee of Tokyo University of Science (approval number: 18023) where obtaining the informed consent from study subjects was waived for the current study. The Japanese data came from 25 corporate-type health insurance plans extracted from Cross-Fact database . Claims data covering 60 months between May 2013 to April 2018 included 1,163,968 males (age (SD) = 42.5 (13.2) years old) and 1,349,901 females (42.2 (13.1) years old) who were 20 years old or older (but younger than 75 years old). As detailed in Additional File 1 (Appendix 5), we examined 99,821 new users of celecoxib, who used celecoxib after at least 180 days of non-use. The occurrence of peripheral edema was defined by new use of furosemide after at least 180 days of non-use and the index date was the day when the outcome occurred. Daily exposure during an 84-day study period was determined using a 7-day grace period. We selected 311 cases who had both exposed and unexposed days during the study period. The 84-day study period was divided into (M + 1) periods where M = 1, 2, 5, 11, 27 and 83, resulting in periods of 42, 28, 14, 7, 3 and 1 days, respectively. Four different definitions were used to determine exposure in the case and control periods as shown in . Cases who started celecoxib on the index date were excluded from the analysis since furosemide could have been prescribed for prevention rather than treatment of edema.
The data was analyzed by standard conditional logistic regression (Eq. (2)) and the Mantel-Haenszel method with their 95% CIs, and the weighting method for a binary exposure (Eq. (11)) with 2.5 to 97.5 percentiles estimated by the bootstrap method. We also extended the study period to 168 and 336 days. All the analyses were performed using SAS 9.4. Data and SAS codes to analyze the data are available upon request.
The data for the Australian study on the association between hip fracture and psychoactive medicines [bib_ref] Psychoactive medicine use and the risk of hip fracture in older people:..., Leach [/bib_ref] [bib_ref] Risk of hip fracture in older people using selective serotonin reuptake inhibitors..., Leach [/bib_ref] were obtained from the Australian Department of Veterans' Affairs administrative claims database. The study was approved by Department of Defense and Veterans' Affairs Human Research Ethics (E016-007) and University of South Australia Human Research Ethics (P203/04) where obtaining the informed consent from study subjects was waived for the current study. The cases were 8828 patients aged over 65 years who were hospitalized for hip fracture between 2009 and 2012. The index date for each case was the date of hospitalization. We have reanalyzed the data from the previous case-control study [bib_ref] Risk of hip fracture in older people using selective serotonin reuptake inhibitors..., Leach [/bib_ref] as a case-crossover study with SSRIs as the exposure using the same methods as the Japanese study.
Both of the studies in Japan and Australia were carried out in accordance with the Declaration of Helsinki, and all methods were carried out in accordance with relevant guidelines and regulations in Japan and Australia. and show OR SCL , OR VF , and OR MH with their 95% CIs, and OR G with its 2.5 to 97.5 percentiles estimated for the scenario in [fig_ref] Figure 2: Expected frequency of cases by subgroup in [/fig_ref] for M control periods (1 period = 1 day). The difference between OR SCL and [fig_ref] Figure 1: Exposure pattern in a hypothetical dynamic population of 80,000 patients [/fig_ref]. (a) 140 cases who had the outcome at the case period involving one binary exposure variable only, where the incidence rate per period at unexposed period (r0) is assumed to be 0.001 and the rate ratio of the exposure (RR) is assumed to be 4. (b) 184 cases who had the outcome at the case period involving one binary exposure variable and one binary time-varying confounder where the incidence rate per period at unexposed period without confounder (r0) is assumed to be 0.001, the rate ratio of the exposure (RR) is assumed to be 4, and that of the time-varying confounder (RRz) is assumed to be 2. The proportion of the time-varying confounder at unexposed periods (f0) is 0.2 and that at exposed periods (f1) in the population 0.4. The status of the confounder at control periods is randomly assigned . n: frequency of cases belonging to each ID_Subgroup in [fig_ref] Figure 1: Exposure pattern in a hypothetical dynamic population of 80,000 patients [/fig_ref] ; c0: exposure status at the case period; z0: status of time-varying confounder at the case period Exposure Definitions Last day = exposure status of the period was the exposure status on the last day; Half or more days = exposure status of the period is defined as 'exposed' if at least half of days during the period was exposed and 'unexposed' otherwise; Any 1 day = exposure status of the period is defined as 'exposed' if at least 1 day during the period is exposed and 'unexposed' otherwise . On the other hand, odds ratios from the remaining 3 methods (OR VF , OR MH , and OR G ) were unbiased irrespective of the value of M. The estimate of OR VF cannot be estimated when M = 7, 10, 14, 17 and 21 because the positivity condition was not satisfied for any data. For example, when M = 7, ∑P(X i0 = 0, N exposed = 1) = 0 because X i0 = 1 in Subgroup H which is only one subgroup where N exposed = 1 and similarly for N exposed = 2 or 3, X i0 was the same for all subgroups. and Appendix show OR SCL , OR VF , and OR MH with their 95% CIs, and OR G with its 2.5 to 97.5 percentiles estimated for the time varying confounding scenario in [fig_ref] Figure 2: Expected frequency of cases by subgroup in [/fig_ref] for M control periods (1 period = 1 day). The difference between OR SCL for exp(β) and the true RR (4.0) was more than 10% of the true value when M > 5 and increased when M increased. The OR SCL estimate for exp(β) was larger than the corresponding value in . When control periods were extended to 10 and 20 cycles, the point estimate of OR SCL for exp(β) was 9.18 (6.23-13.52) and 10.84 (7.09-16.56), respectively. On the other hand, OR VF and OR G for both exp(β) and exp(γ) were in general unbiased, particularly when M > 7 where the estimates of OR VF and OR G for exp(β) were within 3% of the true value. The estimate of OR MH for exp(β) varied between 3.82 (M = 2) and 4.65 (M = 13). Though not shown in and , OR MH for exp(β) estimated by ignoring the time-varying confounding was stable but overestimated as 4.67 (17% above the true value). When M = 1, OR VF was 4.27 for exp(β) (about 7% overestimated) and 1.86 for exp(γ) (7% underestimated). Similarly, when M = 1, OR G was 4.26 for exp(β) (7% overestimated) and 1.68 for exp(γ) (16% underestimated). When N was increased to 1,000,000, exp(β) and exp(γ) were 3.96 and 2.09 for OR VF and 3.97 and 2.04 for OR G when M = 1 (not shown in . [fig_ref] Table 4: Estimates of OR SCL , OR VF and OR MH [/fig_ref] and show OR SCL , OR VF , and OR MH with their 95% CIs, and OR G with its 2.5 to 97.5 percentiles for the scenarios in [fig_ref] Figure 2: Expected frequency of cases by subgroup in [/fig_ref] where the length of the study period was fixed as 22 days but the length of each time period varied between 1 and 11 days. For the scenario in [fig_ref] Figure 2: Expected frequency of cases by subgroup in [/fig_ref] with one binary exposure variable only, OR SCL for exp(β) varied between 4.00 and 6.61 when the length of the time period varied, but OR VF , OR MH , and OR G were stable and unbiased (see [fig_ref] Table 4: Estimates of OR SCL , OR VF and OR MH [/fig_ref] and . For the scenario in [fig_ref] Figure 2: Expected frequency of cases by subgroup in [/fig_ref] with one binary exposure variable and one time varying confounder randomly generated in the control periods, the OR SCL estimate for exp(β) varied between 4.34 (1 period = 11 days) and 6.79 (1 period = 7 days) when M and the length of the time period varied, while OR VF and OR G for exp(β) were stable and close to the true value, though the odds ratio was a little overestimated as 4.34 when the time period was 11 days (see [fig_ref] Table 4: Estimates of OR SCL , OR VF and OR MH [/fig_ref] and . On the other hand, OR MH for exp(β) varied 3.54 (1 period = 4 days) and 5.56 (1 period = 11 days). Though not shown in [fig_ref] Table 4: Estimates of OR SCL , OR VF and OR MH [/fig_ref] and , OR MH for exp(β) estimated by ignoring the time-varying confounding was stable but overestimated as 4.67. When M = 1 (1 period = 11 days), OR VF was 4.31 (8% overestimated) for exp(β) and 1.94 (3% underestimated) for exp(γ). Similarly, when M = 1, OR G was 4.34 (9% overestimated) for exp(β) and 1.35 (32% underestimated) for exp(γ). When N was increased to 1,000,000, exp(β) and exp(γ) were 3.97 and 1.95 for both OR VF and OR G when M = 1 (not shown in [fig_ref] Table 4: Estimates of OR SCL , OR VF and OR MH [/fig_ref]. [fig_ref] Table 5: Estimates of OR SCL and OR MH [/fig_ref] show the estimates using Exposure Definition I in in the Japanese study. In general, OR SCL The estimates of OR SCL , OR VF , OR MH , and OR G from simulated data with a binary exposure only (study period = M + 1 days). Results are shown graphically in M is the number of control periods and study period = (M + 1) days. In the Australian study, after excluding the concordant cases, there were 1316 discordant cases with daily exposed and unexposed periods to SSRIs in the 180 days before the index date. Exposure on the index date was excluded since hip fracture may have occurred the day before admission to hospital.
# Results
## Simulation studies: comparison of methods with or without time-varing confounding
## Analyses of case-crossover studies of real-world data
We used periods of 1, 5, 20, 30, 60 and 90 days with M = 179, 35, 8, 5, 2 and 1, respectively. Exposure within each period was defined using Definition I . Estimates of the OR SCL were biased upwards [fig_ref] Table 7: Estimates of OR SCL and OR MH [/fig_ref]. When M = 1 (exposure period = 90 days), estimates for OR SCL was identical to OR MH and OR G as expected. As in the Japanese study, OR SCL from Eq. (2) increased with M for M > 1.
# Discussion
## Methods to remove bias from within-subject exposure dependency with or without time-varying confonders
Using simulated data, we showed that OR SCL can be biased when there is within-subject exposure dependency but no exposure time trend, except when only one control period is used. When only one control period is used (M = 1), pairwise exchangeability is equivalent to global exchangeability and bias due to within-subject exposure dependency in standard conditional logistic regression does not occur (although bias due to time trends may occur). In , OR SCL increased with the increase of study period. This observation is similar to Estimates of OR_SCL, OR_VF, OR_MH and OR_G for simulated data in [fig_ref] Figure 1: Exposure pattern in a hypothetical dynamic population of 80,000 patients [/fig_ref]. OR_SCL, OR_VF and OR_MH with their 95% CI, and OR_G with its 2.5-97.5pct for the exposure (exp(β)) for M = 1, 3, 6, 9, 13, 20, 69, and 139, where M is the number of control periods and study period = M + 1 days (1 period = 1 day). (a) shows results for the population in [fig_ref] Figure 2: Expected frequency of cases by subgroup in [/fig_ref] with one binary exposure variable only. (b) shows results for the population in [fig_ref] Figure 2: Expected frequency of cases by subgroup in [/fig_ref] with one binary exposure variable and one binary time-varying confounder. OR_SCL: odds ratio by the standard conditional logistic regression (OR SCL ); OR_VF: odds ratio by the Vines and Farrington's method (OR VF ); OR_MH: odds ratio by the Mantel-Haenszel method (OR MH ); OR_G: odds ratio by the Greenland's method (OR G ); 95%CI: 95% confidence interval; 2.5-97.5pct: 2.5 to 97.5 percentiles Estimates of OR SCL , OR VF , OR MH , and OR G from simulated data with a binary exposure and a binary time-varying confounder (study period = M + 1 days). Results are shown graphically in M is the number of control periods and study period = (M + 1) days. that in the previous case-crossover studies, where the odds ratio increased when a longer study period was employed [bib_ref] Persistent user Bias in case-crossover studies in Pharmacoepidemiology, Hallas [/bib_ref] [bib_ref] Use of the case-crossover design to study prolonged drug exposures and insidious..., Wang [/bib_ref] [bib_ref] Prescription of antidepressants and the risk of road traffic crash in the..., Orriols [/bib_ref]. In , OR VF , OR MH , and OR G were unbiased. Of those 3 unbiased estimates, OR VF may be difficult to calculate when analyzing real-world data because it requires population data (or samples from the population) to estimate the exposure probabilities, unlike OR MH and OR G . In addition, probabilities for many exposure permutations must be reliably estimated which may be intractable. For example, in [fig_ref] Figure 1: Exposure pattern in a hypothetical dynamic population of 80,000 patients [/fig_ref] control periods, only 8 subgroups A to H were assumed to exist, but in real-world data, many more exposure patterns would occur. It is possible that the positivity condition is not satisfied for certain permutations, and these do not contribute to the estimation of OR VF . These limitations make OR VF difficult to use in practical applications. As Vines and Farrington showed, when no exposure time trend exists and there is one binary exposure variable which is pairwise exchangeable, OR MH and OR G are unbiased even when OR SCL is biased. However, when the model involves a time-varying confounder in addition to the exposure variable, OR MH may be biased when periods of each subject are further stratified by the time-varying confounder as in [fig_ref] Table 4: Estimates of OR SCL , OR VF and OR MH [/fig_ref]. In contrast, the timevarying confounder can be handled by OR G . Results shown in [fig_ref] Table 4: Estimates of OR SCL , OR VF and OR MH [/fig_ref] indicate that our proposed method is unbiased when both within-subject exposure dependency and time-varying confounding occur at the same time, provided that the time-varying confounder is independent between periods.
In [fig_ref] Table 4: Estimates of OR SCL , OR VF and OR MH [/fig_ref] , when M = 1 (i.e., only 1 control period is used), OR G and OR VF for exp(β) were overestimated and those for exp(γ) were underestimated, but the estimates approached to the true values when N was increased, suggesting that the deviation from the true values observed when M = 1 was due to random error. Conversely, it is likely that employing a larger number of control periods can produce more precise point estimates of exp(β) and exp(γ).
# Real-world data analysis
In the Japanese and Australian studies, we found a discrepancy between OR SCL and OR G when more than one control period was used. In the Japanese study, OR SCL was more than 2 times larger than OR G when the study period increased. We believe that this discrepancy occurred mainly due to within-subject exposure dependency, which increases as the study period increases. The estimates of OR G were the same as OR MH as expected.
In the Australian study, the discrepancy between OR SCL and OR G was modest compared with the Japanese study. This was compatible with the finding by Vines and Farrington that the bias due to within-subject exposure dependency is minimal when exp(β) ≈ 1 [bib_ref] Within-subject exposure dependency in casecrossover studies, Vines [/bib_ref].
## Limitations of the current study and future direction
In the current study, our focus was on bias from withinsubject exposure dependency. However, biases can occur from other sources as well. First, we have not applied the weighting method when a time trend in the exposure exists [bib_ref] The case-time-control design, Suissa [/bib_ref]. Though it was shown in the methods section that if there is no time trend, probabilities that the individual is exposed (π 1 ) and unexposed (π 0 ) at the case period can be estimated without data from population or time-controls, to check whether assumptions of pairwise exchangeability (no exposure time trend) are satisfied, we need data of the population or time-controls. When the proportion of periods exposed in the population or time-controls is roughly constant over study periods, this will support pairwise exchangeability assumptions.
When time trend exists, case-time-control design [bib_ref] The case-time-control design, Suissa [/bib_ref] may be used, but more study is needed when both exposure time trend and within-subject exposure dependency exist. Second, a 'washout period' has been used in some case-crossover studies [bib_ref] On the validity of the case-time-control design for autocorrelated exposure histories, Jensen [/bib_ref] [bib_ref] Dental procedures, antibiotic prophylaxis, and endocarditis among people with prosthetic heart valves:..., Tubiana [/bib_ref] [bib_ref] A case-crossover study of urological chronic pelvic pain syndrome flare triggers in..., Sutcliffe [/bib_ref] to allow for the uncertainty in the optimal length of one period and to reduce within subject exposure dependency. In the current study, when the length of the time period varied, OR G (and OR MH ) was stable in both the simulation study and real-world data. Since exposure was defined by the last day of the period, any period with 2 or more days was equivalent to using a 'washout period' at the beginning of the period. However, much more analyses are needed to examine the need for a 'washout period' and to determine the optimal length of one period particularly when within-subject exposure dependency exists. Another bias may occur when the event rate is not constant. For example, the event rate may be particularly high soon after the exposure is started, compared to later after treatment was started. One solution for this problem is to divide the exposed periods into high-risk and low-risk periods and considering this as different levels of exposure. When within-subject exposure dependency exists, this will require weighting for at least 3 exposure levels and the weighting method for binary exposures described in this study should be expanded for multi-level exposures.
Limitations of our study include that we assumed a time-varying confounder with no within-subject dependency. If within-subject dependency exists for a time-varying confounder, the weighting method may need to allow for both the exposure and time-varying confounders. Furthermore, when unmeasured timevarying confounders exist, the results still can be biased even when using our weighting method.
Finally, as mentioned earlier, the variance of OR G in Eq. [bib_ref] The case-time-control design, Suissa [/bib_ref] is estimated assuming the weight in Eq. (10) is a constant. To allow for uncertainty in estimating the weights in Eq. (10), bootstrapping was used to estimate to 95% CI from 2.5 to 97.5 percentiles of OR G . However, an analytical method to estimate the variance of OR G which accounts for the variance of the weights may be useful and be developed in a future study.
# Conclusion
Despite autocorrelated exposures being common in pharmacoepidemiology, bias due to within-subject exposure dependency in the case-crossover study has had little attention in the pharmacoepidemiology literature and standard conditional logistic regression has been widely used without assessing whether this bias exists. Although using only one control period can avoid bias due to within-subject exposure dependency, this will reduce the accuracy of the estimate due to random error, as seen in our simulation study (the odds ratio when M = 1 in [fig_ref] Table 4: Estimates of OR SCL , OR VF and OR MH [/fig_ref]. To assess for the possibility of bias, we recommend comparing the Mantel-Haenszel odds ratio with the standard conditional logistic regression odds ratio before starting analysis of case-crossover data. If timevarying confounders exist in data set, they may be ignored when comparing two odds ratios to assess the possibility of bias due to within-subject exposure dependency. If a substantial discrepancy is found (e.g., more than a prespecified threshold such as 5 to 10% of the Mantel-Haenszel odds ratio), standard conditional regression should not be used. Either the Mantel-Haenszel method or our weighting method should be used instead. The weighting method has less bias than the Mantel-Haenszel method when a time-varying confounder exists and in such a case we recommend using our proposed method to estimate OR G in the analysis of case-crossover data because it removes bias from within-subject exposure dependency and can account for time-varying confounders.
Future research will extend our weighting method to allow for time trends in the exposure, 'washout periods' , multi-exposure levels which are potentially important when the event rate changes during exposed periods, and analytical method to have the variance of OR G by making allowance for the uncertainty in estimating the weight.
## Supplementary information
The online version contains supplementary material available at https:// doi. org/ 10. 1186/ s12874-021-01408-5.
Additional file 1.
[fig] Figure 1: Exposure pattern in a hypothetical dynamic population of 80,000 patients. Patients receiving drug treatment with a cycle of 7 days are divided into 8 Subgroups A to H where 1 period is defined as 1 day. Subgroup A represents stoppers that stop treatment, Subgroup H represents starters that start treatment, and Subgroups B to G represent different exposure patterns of those being treated at the case period. The bolded outline indicates that 21 control periods can be divided into 3 cycles with the same exposure pattern. N: size of each subgroup; c0: exposure status at the case period; cm (m = 1, 2, −--, 21): exposure status at the m-th control period; Tx: treatment [/fig]
[fig] Figure 2: Expected frequency of cases by subgroup in [/fig]
[table] 1: period = 1 day for all values of M OR SCL : odds ratio by the standard conditional logistic regression; OR VF : odds ratio by the Vines and Farrington's method; OR MH : odds ratio by the Mantel-Haenszel method; OR G : odds ratio by the Greenland's method; 95%CI: 95% confidence interval; 2.5-97.5pct: 2.5 to 97.5 percentiles M OR SCL (95%CI) OR VF (95%CI) OR MH (95%CI) OR G (2.and Table 6 for Exposure Definitions II, III and IV (Table 1) are presented. Those results indicated, as in Tables 5 and 6, that OR SCL varied when M varied as well as when the study period varied, while OR MH and OR G were relatively stable. Detailed description for Definition II, III and IV is given in Appendix 5 in Additional File 1. [/table]
[table] Table 4: Estimates of OR SCL , OR VF and OR MH [/table]
[table] Table 5: Estimates of OR SCL and OR MH [/table]
[table] Table 6: Estimates of OR SCL and OR MH (95% CI), and OR G (2.5-97.5pct): Japanese data on celecoxib-peripheral edema with study period = 168 and 336 days and Exposure Definition I M is the number of control periods OR SCL : odds ratio by the standard conditional logistic regression; OR MH : odds ratio by the Mantel-Haenszel method; OR G : odds ratio by the Greenland's method. 95%CI: 95% confidence interval; 2.5-97.5pct: 2.5 to 97.5 percentiles [/table]
[table] Table 7: Estimates of OR SCL and OR MH (95% CI), and OR G (2.5-97.5pct): Australian data on SSRI-hip fracture (Study period = 180 days, Exposure Definition I) M is the number of control periods OR SCL : odds ratio by the standard conditional logistic regression; OR MH : odds ratio by the Mantel-Haenszel method; OR G : odds ratio by the Greenland's method; 95%CI: 95% confidence interval; 2.5-97.5pct: 2.5 to 97.5 percentiles [/table]
|
Colposcopy accuracy in diagnosing cervical precancerous lesions in western Kazakhstan
A B S T R A C TThis retrospective cohort study focused on colposcopic accuracy for the diagnosis of cervical premalignant lesions using cytology and histology, as well as HPV data not included in current cervical screening practices in Kazakhstan. Colposcopy performance was assessed using the modified Reid index in women aged 18-63 years. In total, 1,129 colposcopic-HPV-cytology triple samples and 94 histology findings were collected. The sensitivity of colposcopy was 81.6% with specificity 72.6% for LSIL but fell to 56.6% with specificity 88.3% for CIN2+ vs. 89.6% and 74.5% for cytology at CIN2+, respectively. The ORs for high-grade lesion occurrence within each colposcopy group at viral load rising vs. ORs for HPV-negative women were 3.4; 5.3; and 39.7, respectively (p < 0.0001). Total attributive agreement between the colposcopy and histology findings reached 55.3%, κ 0.47 ± 0.06 vs. 0.62 ± 0.08 for cytology, and 0.34 ± 0.13 and 0.58 ± 0.1, for specialists, respectively. Outcomes obtained for colposcopy alone failed to show satisfactory reliability. Globally adopted primary HPV screening would be the best option despite the related costs.
# Introduction
Cervical cancer (CC) is caused by the group of human papillomaviruses (HPV) for which "there is no perfect way to categorize a continuum of their carcinogenic potential" [bib_ref] Classification of weakly carcinogenic human papillomavirus types: addressing the limits of epidemiology..., Schiffman [/bib_ref]. In Kazakhstan, a sizeable post-Soviet state in Central Asia, the nationwide CC screening program was implemented in 2008; nevertheless, issues of cancer prevention remain problematic. The screening program includes a Pap test every four years in women aged 30-70 years old using liquid-based cytology (LBC) techniques with "Cell Scan" technology (South Korea manufacturing) and conventional azur-eosin staining as an opportunistic method. In research, the conventional azur-eosin method had a sensitivity of 90.4% and a specificity of 90.0% for CIN2+. Researchers failed to show the LBC "Cell Scan" technique to be superior to simple azur-eosin staining [bib_ref] Cervical screening in Western Kazakhstan: Liquid-based cytology 'Cell Scan' versus azur-eosin staining, Balmagambetova [/bib_ref]. Primary HPV testing accompanied by cytology triage in HPV-positive women aged 30+ has not been adopted in the country, despite the HPV prevalence of approximately 25-28% [bib_ref] Epidemiology of HPV infection and HPV-related cancers in Kazakhstan: a review. Asian..., Aimagambetova [/bib_ref]. Reportedly, the CC incidence rate was 18.2 per 100,000 women in Kazakhstan.
According to commonly accepted guidelines, colposcopy examination follows screening procedures upon presentation of abnormal cytology . Regarding the diagnostic accuracy of colposcopy in various metaanalyses, its sensitivity fluctuated from 29% to 100%, and its specificity ranged from 12% to 88% [bib_ref] Systematic reviews and meta-analyses of the accuracy of HPV tests, visual inspection..., Mustafa [/bib_ref]. At least a rough analysis of colposcopy capabilities for screening purposes would be reasonable given the HPV testing unavailability in the country.
In this study, we assessed colposcopy accuracy for the diagnosis of cervical precancer lesions compared to cytology and histology findings by applying HPV data not included in current cervical screening practices in Kazakhstan.
# Methods
This retrospective cohort study explored the limits of colposcopy within a broad multipurpose research effort on HPV infection across western provinces of the country. The study protocol was approved by the University's IRB and published [bib_ref] The impact of human papillomavirus genotype on colposcopic appearance: a cross-sectional analysis, Bekmukhambetov [/bib_ref]. Work was carried out under the Helsinki Declaration principles, and all participants signed the informed consent form. We enrolled women within 18-63 years old and then stratified the sample by age. A total of 1,129 of 1,166 women were asked for their clinical history and were selected for colposcopy. No vaccination history was the only inclusion criterion. HPV vaccination, as well as HIV presence, pregnancy, or any previous procedure on the cervix were exclusion criteria. Cases with verified invasive cancer were not included in the study.
Qualitative detection and quantification of human papillomavirus was performed through real-time PCR based on Russian test systems and equipment. We used the "Quantum-21 ′′ kit for typing and quantifying the DNA of low-risk HPV (6, 11, 44) and high/probable carcinogenic risk in a total of 21 types. To isolate the viral DNA, sets PROBA-NK-PLUS, the same production, were applied.
We performed colposcopy after PCR-based assays for HPV from the cervix according to standard procedures. To assess cervical condition, we applied the modified Reid colposcopy index (RCI) despite the presence of Swede scores (developed in 2005), because RCI is in use in Kazakhstan. Standard parameters were assessed as presented in the EVAH study , as the objectives and methods applied were similar. The EVAH research focused on studying colposcopic performance in diagnosing high-grade cervical lesions using colposcopic characteristics and high-risk HPV genotyping. Colposcopists identified and graded the lesions and scored the lesions' colposcopic impression, collecting multiple (up to four) biopsies at suspected CIN, including the normal tissue biopsy. Accordingly, we also scored the following parameters: the lesion color, the surface configuration and margins, presence/absence and degree of punctation and mosaicism, vessels, acetowhitening rapidity, and size of the lesions (0, <25, 25-50, >50% of the cervix). Cervical tissue biopsy selectively followed colposcopy examinations at suspected CIN. To systematize findings, we allocated three groups according to a colposcopy opinion: group 1, with scores by RCI up to 2 (supposedly, from NILM to CIN1); group 2, with scores up to 4 (LSIL and overlapping lesion, presumably CIN2); group 3, where scores five and over were referred (HSIL, likely to be CIN3), respectively. Colposcopic terminology adopted in July 2011 was used. The two team members experienced and certified in the cervical pathology were responsible for colposcopy opinion and biopsy sampling. Their experience in colposcopy performance was six and more than ten years, respectively. Cytology findings obtained through the conventional technique (azur-eosin staining) were available for all women and designated according to the TBS (Terminology Bethesda System) 2001, as the research started before the issuing of the new edition of 2016. All histology findings were also combined into three groups and designated group 1 -NILM, up to CIN1; group 2 -LSIL, up to CIN2; and group 3 -HSIL, CIN3.
## Statistical processing:
All calculations were carried out using Statistica 10 (Dell Technologies, Texas, USA). For all tests, a two-sided type I error of p < 0.05 at 95% CI was assumed to be statistically significant. Evaluation of colposcopy as a diagnostic tool for the detection of CIN was performed using Cohen's kappa calculation and ROC analysis using SPSS v.25 (IBM, Armonk, NY, USA) and www.medcalc.be. The logistic regression model with OR calculations was designed to evaluate the probability of HSIL development at viral load increments within each RCI group.
# Results
All colposcopy findings (n 1,129) were stratified by HPV status. As such, we performed 846 HPV-negative and 283 HPV-positive assays. Among the HPV-positive women, 60 (21.2%) were low-positive, 96 (33.9%) were moderate, and 127 (44.9%) were of high categories of viral load. Baseline data are provided in .
A statistically significant age difference, both for HPV-negative and HPV-positive women, has been established when analyzing across the three groups of incrementally increasing changes depending on the colposcopic impressions. The Kruskal-Wallis rank-sum test was H = 44.47 (2, p < 0.05, n 846) and H = 41.2 (2, p < 0.05, n 283). In the meantime, no difference has been found in the mean age of women's sexual debut: H = 2.93 (2, p 0.23) for HPV-negative and H = 1.85 (2, p 0.39) for HPV-positive sample, respectively.
A noteworthy trend emerged when we assessed cytology findings. In eighty (9.5%) of HPV-negative women having unfavorable colposcopic impressions (RCI scores 5+), high-grade cytology lesions were found. Conversely, almost in a quarter (23.2%) of HPV-positive women with severe colposcopic opinions, the NILM cytology conclusions were obtained. Overall, the number of high-grade lesions in cytology smears (HSIL) was 25 of 1,129 (2.2%): 7 in RCI group-1, 8 in group-2, and 10 in group-3. As to the viral load trends, the proportion of high load (5 + GE*10 3 per sample) was twice as lower in the group-1 (RCI 0-2) compared to the group RCI 5+ (34.7% vs. 68.3%), and the average viral load was expectedly lower (4.7 vs. 7.0). also demonstrates a consecutive increase in the share of the most carcinogenic HPV 16/18. The severity of colposcopy changes grows as the proportion of these types rises (Kendall's τ 0.26, p 0.035).
## Assessing the diagnostic value of colposcopy
We performed ROC analysis to estimate colposcopy diagnostic value and assessed the agreement between colposcopists' impressions and histology conclusions. ROC analysis showed that the area under the curve (AUC) was 0.78 ± 0.05 (CI 95% 0.66;0.84) for colposcopy vs. 0.83 ± 0.04 (CI 95% 0.72;0.89) for cytology (p 0.0001). [fig_ref] Table 2: Coordinates of the obtained ROC curve with 95% CI [/fig_ref] outlines the coordinates of the obtained ROC curve. The sensitivity of colposcopy with the threshold for LSIL and overlapping lesions (RCI scores up to 4) fell to 56.6% when the cut-off was raised to the high-grade lesion, CIN2+ (RCI 5+). Conversely, the specificity increased for HSIL. The positive likelihood ratio (+LR) increased from 2.7 to 4.64, and the -LR increased from 0.29 to 0.52. Although for cytology the trend was the same, baseline magnitudes were higher, while the range was significantly lower. The sensitivity of cytology at the HSIL cut-off fell to 89.6% vs. 56.6% for colposcopy, +LR increased from 2.21 to 3.23, and -LR increased from 0.08 to 0.16.
Calculation of interrater agreements between the grouped histology and colposcopy findings resulted in 55.3% for n 94 and 52.2% and 60.4% in specialists, respectively. Overall concordance between the colposcopic performance and histology reached 56.8% for benign and dubious lesions up to LSIL; 30.8% for overlapping lesions, up to CIN2; and 61.5% for HSIL. Accordingly, linear weighted Cohen's κ was found to be 0.47 ± 0.06 (95% CI 0.38;0.61); for specialist-1 0.34 ± 0.13 (95% CI 0.08;0.59), and 0.58 ± 0.1 (95% CI 0.39;0.81) for specialist-2. Meanwhile, in the mentioned study on cytology aspects of current cervical screening in the country [bib_ref] Cervical screening in Western Kazakhstan: Liquid-based cytology 'Cell Scan' versus azur-eosin staining, Balmagambetova [/bib_ref] , κ 0.62 ± 0.08 for azur-eosin staining was established.
## Probability for hsil development at viral load raising
To evaluate the probability of HSIL development at viral load increments within each RCI group, we designed a logistic regression model (Nagelkerke's R 2 0.44). The odds ratios (ORs) are presented in [fig_ref] Table 3: Effect of viral load within the colposcopy groups on HSIL development [/fig_ref]. [fig_ref] Table 3: Effect of viral load within the colposcopy groups on HSIL development [/fig_ref] , the chance for HSIL appearance rises depending on HPV load magnitudes within each group, thus promoting further deterioration of the cervix condition and related RCI. The ORs for highgrade lesion occurrence at viral load rising, for the group with scores by RCI up to 2 (supposedly, from NILM to CIN1) compared to women having severe colposcopic opinions, resulted in 3.4 vs. 39.7 (p < 0.0001).
## As shown in
# Discussion
Commonly, the risk of developing CIN conjugately rises as the viral load of highly carcinogenic HPV types increases. Moreover, reportedly, there are type-specific differences in correlations with the severity of lesions. [bib_ref] Type-specific high-risk human papillomavirus viral load as a viable triage indicator for..., Dong [/bib_ref] found that the viral loads of HPV-16, -31, -33, -52, and -58 positively correlated with the severity of cervical lesions, whereas those of HPV-18, -45, -56, -59, and other types did not. In our study, we did not separate the viral load by type.
To reveal the relationship between the colposcopy findings depending on viral loads, we determined the chance for HSIL development per RCI group. Probability for HSIL occurrence increased consecutively at viral load raising (ORs raised from 3.4 in group-1 up to 39.7, p < 0.0001 in the group with RCI 5+). Our findings turned out to be significantly inferior to those of other authors made based on a larger number of observations. [bib_ref] Implications of semi-quantative HPV viral load estimation by Hybrid capture 2 in..., Basu [/bib_ref] established ORs for CIN2+ diagnosis in women with a high level of viral load as 46, 217.4, and 3915.1 for the "probable high grade" group, respectively (p < 0.001, n 39,728).
Overall, we established cytology superiority over colposcopy in CIN detection. We found acceptable sensitivity of colposcopy for detecting LSIL (81.6%) but observed a further decrease to 56.6% at raising the threshold to HSIL. [bib_ref] Correlation of two colposcopic indices for predicting premalignant lesions of Cervix, Kushwah [/bib_ref] , comparing the performance of Swede score and RCI, established the sensitivity of RCI as 89% for any lesion, similarly falling to 56% for HSIL, while the specificity increased to 92.9%. [bib_ref] Correlation of Swede score colposcopy scoring system and histopathological results in patients..., Alan [/bib_ref] , who also studied the Swede scoring system performance, found that a cut-off value ≥ 6 had a high sensitivity for high-grade lesions, and this scoring system was a useful tool for evaluating atypical cervical cytology in women with high-risk HPV infection.
We cannot explain the similarity of our data concerning the low sensitivity of RCI for the HSIL vs. the Swede system. Nevertheless, further research comparing the performance of the two scoring systems appears not to be justified, because the data obtained in this study suggests the need for globally adopted primary HPV testing implementation.
# Conclusion
Our results are in line with a common trend stating that colposcopy alone is insufficient to provide proper detection and prediction of highgrade cervical lesions. In general, we established that the utility of colposcopy diagnosis through RCI pertains mostly to low-grade lesions. In addition, the probability of HSIL occurrence increases with increasing Descriptive statistics of the study. HPV viral load.
Overall, colposcopy efficiency in the diagnosis of cervical precancerous lesions failed to show satisfactory reliability. Globally adopted primary HPV screening would be the best option despite the related costs.
## Credit authorship contribution statement
## 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.
[table] Table 2: Coordinates of the obtained ROC curve with 95% CI. [/table]
[table] Table 3: Effect of viral load within the colposcopy groups on HSIL development. [/table]
|
Persistent bronchography in a newborn with esophageal atresia
Barium sulfateIodate medium Respiratory distress Bronchography a b s t r a c t Esophageal atresia (EA) with tracheoesophageal fistula occurs in about 1:2,500 births. We report a case of persistent bronchography in a newborn with EA and distal tracheoesophageal fistula. A large amount of barium sulfate was injected for mistake by a tube directly in the right bronchus to evaluate the patency of the esophagus. The infant showed signs of respiratory distress; he was intubated and transported at children's Hospital Santobono Pausilipon. Here, it was performed a chest X-ray that confirmed the opacification of the right bronchial tree, and it was suspected an EA type 3b. The literature recommends that: evaluation of the patency of the esophagus, with an iodinate contrast medium, should be done in a pediatric specialized center for high risk of lung aspiration.A male infant weighting 3,380 g was born through elective Cesarean section at 37þ 4 of 7 weeks of gestation in a community hospital. Pregnancy was complicated by polyhydramnios. Apgar scores were 7 I -9 V . Based on the prenatal history and the presence of copious drooling after birth, an esophageal atresia (EA) was suspected. A catheter was inserted for a length of 10 cm from the lip to assess the interruption of the esophagus. Afterward, barium sulfate was injected to confirm the diagnosis of EA. The chest X-ray showed opacification of only the right bronchial tree so it was suspected an EA with proximal tracheoesophageal fistula (TEF;Fig. 1). R a d i o l o g y C a s e R e p o r t s 1 1 ( 2 0 1 6 ) 1 1 3 e1 1 5 http://dx.
Immediately, the newborn showed signs of respiratory distress due to barium sulfate aspiration. For this reason, the infant was intubated and transported at children's Hospital Santobono Pausilipon by neonatal emergency transport system. During the transport, the baby was subjected to synchronized intermittent positive pressure ventilation with peak inspiratory pressure ¼ 22 cm H 2 O, positive end expiratory pressure ¼ 25 cm H 2 O, respiratory rate ¼ 40 b/min, FiO 2 ¼ 0.3, and aspirated continually. At the hospital, the baby continued mechanical ventilation with the same parameters and the chest X-ray with iopamidol did not reveal the proximal TEF . Because of radiologic evidence along with the presence of air into stomach, we suspected the presence of EA with distal TEF. Also, the revaluation of the first chest X-ray obtained in community hospital by a trainee pediatrician, it was observed that the tip of the catheter was inserted directly into right bronchus. At 3 days of life, the patient underwent surgical intervention for EA. An axillary incision was executed on the right side of the chest and a thoracotomy at fourth intercostal space. It was resolved the distal TEF and executed the esophageal anastomosis. The surgery had a good outcome. The infant began to breathe spontaneously at 9 days of life. There have been no complications reported during hospitalization. The baby was discharged at 24 day of life in good clinical condition. In follow-up visit at 4 months, the infant did not show any respiratory complications but persisted the bronchoalveolar opacification of the right lung [fig_ref] Figure 3 e: After surgical intervention and during the 4th month of life, thoracoabdominal X-ray... [/fig_ref].
# Discussion
EA with or without TEF is the most common congenital anomaly of the esophagus and occurs approximately to 1:2,500 births. There are 5 types of EA, they are described more frequently according to Vogt's classification. Type 1 is TEF without EA, type 2 is EA without TEF, type 3a is EA with proximal TEF, type 3b is EA with distal TEF, and finally type 3c is EA with proximal and distal TEF [bib_ref] Congenital esophageal atresia, Vogt [/bib_ref]. The diagnosis is made in the early postnatal period in the vast majority of infants [bib_ref] Oesophageal atresia and tracheo-oesophageal fistula, Smith [/bib_ref]. During prenatal age, polyhydramnios along with a small stomach has around a 55% predictive value for EA [bib_ref] Oesophageal atresia and tracheo-oesophageal fistula, Smith [/bib_ref] [bib_ref] Current knowledge on esophageal atresia, Pinheiro [/bib_ref]. If the diagnosis is suspected, a catheter should be inserted from the mouth into stomach. If there is ongoing doubt regarding the diagnosis, a contrast study may be helpful [bib_ref] Oesophageal atresia and tracheo-oesophageal fistula, Smith [/bib_ref] [bib_ref] From Vogt to Haight and Holt to now: the history of esophageal..., Muensterer [/bib_ref]. In these cases, a small amount of water soluble contrast medium should be administered [bib_ref] From Vogt to Haight and Holt to now: the history of esophageal..., Muensterer [/bib_ref] [bib_ref] Esophageal atresia with proximal tracheoesophageal fistula: a missed diagnosis, Parolini [/bib_ref] , preferably in a pediatric specialized center [6] with extreme care for the high risk of aspiration of contrast medium during the study [bib_ref] Oesophageal atresia and tracheo-oesophageal fistula, Smith [/bib_ref] [bib_ref] From Vogt to Haight and Holt to now: the history of esophageal..., Muensterer [/bib_ref]. In our case report, the infant was born with a good Apgar score but due to the evaluation of EA, it was inserted with a catheter from the lip to assess the interruption of the esophagus. Unknowingly, the catheter was inserted into right bronchus, and it was injected barium sulfate to evaluate the presence of TEF. The infant showed respiratory distress for barium sulfate aspiration, and it was suspected an EA with proximal TEF. This contrast medium was chosen because the baby was born in a community hospital where only barium sulfate is used for contrastographic examinations. Barium sulfate should be avoided if there is the suspect of obstructions, fistulas, and intestinal perforations. If it comes out on the viscera, barium sulfate becomes toxic; indeed, it can cause peritonitis or mediastinitis chemical because of its high osmolarity. It is recommended instead the use of iodinate contrast medium to evaluate the patency of esophagus because of its low osmolarity. At children's Hospital Santobono Pausilipon, we performed a chest X-ray with iopamidol, and there was not proximal TEF found, instead we suspected a distal TEF. During the surgery, our diagnosis was confirmed. In a similar case report to evaluate the patency of esophagus in an infant with suspected EA, it was administered gastrografin (diatrizoate meglumine and diatrizoate sodium contrast) by orogastric tube. The infant aspirated the contrast medium that caused a tracheobronchogram, which developed chemical pneumonitis from gastrografin aspiration after surgical operation [bib_ref] Diagnosis of esophageal atresia with tracheoesophageal fistula: is there a need for..., Mcduffie [/bib_ref].
In the past, gastrografin and barium sulfate aspiration induced chemical pneumonitis and caused death both in adults and in children [bib_ref] Barium sspiration, Albeldawi [/bib_ref] [bib_ref] Fatal aspirations in infancy during gastrointestinal series, Mcalister [/bib_ref]. If barium sulfate is aspired, lungs will metabolize it slowly. The real danger of barium sulfate aspiration is immediate because it causes pneumonitis chemical or pulmonary edema within few seconds or few hours from aspiration. Complications also depend on the amount ingested. Long-term effects of contrast media aspirated have been studied in an animal study on rats and dogs until 9 months. The study revealed that nonionic iodinecontaining agents were better tolerated and evoked less pulmonary response than barium sulfate [bib_ref] The effect of some contrast agents in the lung: an experimental study..., Mcalister [/bib_ref].
r e f e r e n c e s
[fig] Figure 1 e: At birth, thoracoabdominal X-ray showed opacification of right bronchial tree and the tip of the catheter placed in the main right bronchusFig. 2e After 2 hours of life, the chest X-ray detected an atretic esophageal pouch (one arrow), the opacification of right bronchial tree, and air in the stomach. [/fig]
[fig] Figure 3 e: After surgical intervention and during the 4th month of life, thoracoabdominal X-ray showed persistent unilateral bronchography with lower intensity. [/fig]
|
Selling one’s future: over-indebtedness and the risk of poor mental health and the role of precarious employment – results from the Scania Public Health Cohort, Sweden
# Introduction
Credit plays a paramount role in modern economies, since it increases the possibility to make investments that can generate future financial and other gains and to allocate lifetime income in a more optimal relation to need. From a health perspective, such resources are very likely related to opportunities for improved health over the life-course. However, when the credit market targets those who cannot make ends meet, this can cause increasing indebtedness, which can lead to poorer health, both due to loss of material and other resources and to exposure to growing stress and worries about one's loss of control over the household economy. Thus, the proportion of individuals who are prone to over-indebtedness will most likely expand considerably in the aftermath of the COVID-19 pandemic. Moreover, indebtedness could be expected to have a differential effect on health, much depending on factors such as socioeconomic position, country of birth (ie, migration background) and age.
## Strengths and limitations of this study
⇒ This was a prospective cohort study with 2795 vocationally active men and women randomly selected from the general adult population. ⇒ The exposure and outcome variables have been used and validated in previous studies. ⇒ Detailed information concerning employment history and details regarding current employment for each individual was used for determining precarious employment status. ⇒ Non-response and missing data among respondents may have caused some selection bias. ⇒ We lacked access to an objective measure regarding over-indebtedness, which may have caused misclassification.
## Open access
The credit market has expanded over the last decades leading to a build-up of consumer debt. However, it was not until the financial crisis in 2007-2008 that an increased proportion of households that had difficulties making ends meet was noted in the European Union.As of 2016, the household debt in relation to disposable income was more than twice as high in the OECD compared to 1996.Measured as share of gross domestic product, the indebtedness of Swedish households has doubled since the 1990s.According to public reports, the number of over-indebted individuals in Sweden amounts to between 6% and 18% of the population, depending on definition. The subsequent yearly societal cost is estimated to 30 billion SEK.The credit market has also changed in terms of type of credit providers and types of 'products', for example, so called 'pay-day loans' as a temporary financial solution when financial means are insufficient to cover daily cost of living because of inadequate income and lack of savings. [bib_ref] Protection of Consumer Rights in SMS Loan Agreements, Simovart [/bib_ref] These types of loans have at the same time become very accessible via the Internet and mobile phones. Since the interest rates and fees generally are very high for these loans, this could quickly lead to a spiral of indebtedness where the loantaker in practice ends up owing the lender a substantial proportion of his/her future income. [bib_ref] Bilaga 1: Överskuldsättningens kostnader, Ahlström [/bib_ref] Over-indebtedness can be defined as a situation, persisting for a sustained period of time, in which daily cost of living exceeds income, and in which the imbalance cannot be compensated by savings. This will inevitably lead to use of credit in one way or another, the cost of which will add to the discrepancy between expenditure on the one side and income and savings on the other. This situation could represent an emerging significant stressor and determinant of ill health, and thereby constitute a source of health inequality that is likely to be overlooked in research on income or wealth and health, since information concerning over-indebtedness can not easily be retrieved from registers on income, wealth or social transfers.
A growing number of scientific studies on the indirect and direct association between over-indebtedness and health have been published in the scientific literature. Systematic reviews have provided support for the assumption that the financial recession has been associated with an increase in suicide rates in high-income countries such as Europe and North America. [bib_ref] Systematic review of suicide in economic recession, Oyesanya [/bib_ref] Further, socioeconomic mediators such as unemployment, income decline and unmanageable debts are significantly associated with poor mental well-being, increased rates of common mental disorders, substance-related disorders and suicidal behaviours. [bib_ref] Mental health outcomes in times of economic recession: a systematic literature review, Frasquilho [/bib_ref] [bib_ref] On the relation between Over-Indebtedness and well-being: an analysis of the mechanisms..., Ferreira [/bib_ref] [bib_ref] Association between overindebtedness and antidepressant use: a cross-sectional analysis, Warth [/bib_ref] [bib_ref] The impact of savings and credit on health and health behaviours: an..., Białowolski [/bib_ref] Individuals who cannot make their loan payments have been found to suffer from depression and have suicidal ideation more often compared with those without such problems. [bib_ref] Health effects of indebtedness: a systematic review, Turunen [/bib_ref] Another study investigating predictors of mental health in people on the verge of bankruptcy found that perceived financial strain was an important predictor of poor mental health. [bib_ref] Beyond debt. A moderator analysis of the relationship between perceived financial strain..., Selenko [/bib_ref] However, most of the studies have been cross-sectional, and more longitudinal studies are needed to corroborate the mentioned findings.
High debt relative to available assets has also been associated with poorer self-assessed general health, higher diastolic blood pressure, obesity, poorer health-related behaviour and myocardial infarction. 14 The effect of overindebtedness on health has been found to be stronger for women than for men. [bib_ref] Over-indebtedness and chronic disease: a linked register-based study of Finnish men and..., Blomgren [/bib_ref] In a life-course perspective, poor self-rated health has been found to be even more prevalent among those who have experienced economic stress both early in life and later on. [bib_ref] Economic stress and perceived health among adolescents in Sweden, Hagquist [/bib_ref] According to a systematic review and meta-analysis examining the relationship between personal debt and mental health, 17 ten epidemiological studies with nationally representative samples of the general population were found, whereof seven came from the UK and nine out of ten were cross-sectional. Unsurprisingly, very high ORs for different mental health outcomes were found. In many cases the causality pattern could be described as a vicious circle, which could start either with a decrease in income or in health. Poor mental health provoked by over-indebtedness tends to increase the risk of further financial problems and so forth.Moreover, overindebtedness or coping with excessive credit is associated with subjective distress such as shame, anxiety and apathy. [bib_ref] Minuskontot: Ekonomiska villkor för personer med psykisk funktionsnedsättning, Levinsson [/bib_ref] The mentioned process could involve mediating mechanisms, as well as situations where certain exposures modify the effect of each other, such that already vulnerable groups, for example, individuals with low education or in a precarious employment situation, are hit more severely by over-indebtedness in terms of resulting poor mental health.
To the best of our knowledge, evidence regarding the role of precarious employment as a mediator or modifier of the impact of over-indebtedness on mental health is lacking.
In summary, previous research relies heavily on crosssectional data from the UK, why studies from other national contexts are needed. Moreover, because of the plausible bidirectional causal pathways, prospective studies of long duration with several assessments of both types of variables, over-indebtedness as well as mental health, are preferable. Also, as mentioned above, little is known about the role of precarious employment and the differential health effects across sociodemographic groups.
## Objectives
The aim of the study was to investigate the role of overindebtedness for the development of poor mental health in a longitudinal perspective, and whether this impact is modified by age, gender, educational level, housing tenure or being in a precarious employment situation.
## Open access
# Methods
## Study population
The present cohort was established in 1999/2000 and followed up in 2005 and 2010. [bib_ref] Representativity of a postal public health questionnaire survey in Sweden, with special..., Carlsson [/bib_ref] At baseline, a postal questionnaire was sent out to 25 000 men and women, 18-80 years old. These individuals were randomly selected from the population register, such that equal representation was achieved from all municipalities in the county of Scania (population 1.3 million), Sweden. In total, 13 589 out of 23 437 eligible individuals returned a completed questionnaire (response rate 58%). All of those who responded at baseline and still residing in the county were invited to follow-up after 5 and 10 years. Out of 12 002 respondents alive and still living in the region after 10 years, there were 8206, that is, 68%, who also participated in the 2005 and 2010 inquiries.
A comparison between these 8206 respondents and the corresponding Scania general population at baseline 20 showed that younger individuals, men and individuals with a low educational level were slightly under-represented among respondents. However, a recently published study from this cohort, showed very little proof of selection bias regarding common health outcomes, including mental health. [bib_ref] Assessment of selection bias due to dropouts in the follow-up of the..., Canivet [/bib_ref] Since one aim of the study was to look into the relationship between precarious employment and overindebtedness and mental health, we first excluded individuals who were not vocationally active, that is, at baseline were retired, on disability pension, or on longterm sick leave, and also those who answered that they did not wish to work a year from now (N=1744). Out of the remaining 6462 persons, those with lacking data on employment precariousness (N=2479) were excluded; out of these, 1106 were 55-80 years old at baseline, that is, were qualified for old-age retirement during the studied window of time. Further, we excluded individuals who lacked baseline data regarding the variables included in the multivariable analyses (rather than assigning them as 'internal' missing); that is, housing tenure (N=159) or educational level (N=158), also those lacking baseline and 2005 information on over-indebtedness (N=249), furthermore those who lacked information regarding mental health at all follow-ups (N=161). Finally, in order to reduce the influence of reverse causation, we also excluded individuals with poor mental health at baseline (N=1563), instead of controlling for this in the multivariable analysis. Thus, the final study population consisted of 2795 individuals, 1256 men and 1539 women.
Outcome variable: poor mental health Mental health was measured at all three time points of assessment, using the 12-item version of the General Health Questionnaire (GHQ-12). We used the 0-0-1-1 scoring method (range 0-12) recommended by the creators of the instrument, [bib_ref] The validity of two versions of the GHQ in the who study..., Goldberg [/bib_ref] with poor mental health or 'GHQ-caseness' determined by the population mean and defined as a score of 2 or higher. [bib_ref] Why GHQ threshold varies from one place to another, Goldberg [/bib_ref] Over-indebtedness Two items have been used previously in several Swedish studies to evaluate economic hardship, namely inability to meet expenses and lacking cash reserves. [bib_ref] Economic hardships in adulthood and mental health in Sweden. The Swedish national..., Ahnquist [/bib_ref] For the former we used the question 'How often during the last 12 months have you had problems paying your bills?' and for the latter, 'In case of an unforeseen emergency, do you have the capacity to raise 12 000 SEK (approximately 1100 EUR) in a week's time?'. The same questions were asked in 1999/2000, 2005 and 2010, but the specified amount was raised to 14 000 SEK, approximately 1300 EUR, in the 2010 survey. Over-indebtedness was defined as affirming 'about half of the months' (or more often) on the first question and answering 'no' on the second.
## Other variables
Age was classified into four groups, except in the interaction analysis, where it was dichotomised, and in the multivariate analysis, where it was used as a continuous variable. Country of origin was recorded as 'born in Sweden' (yes/no). Educational level at baseline was determined by the self-reported total years of formal education and classified into four groups. Housing tenure was dichotomised as either owned or rented residence. For employment precariousness, we used a dichotomous variable, based on a combination of data on present unemployment, episode of involuntary unemployment during the past 3 years (no/yes), currently temporarily versus permanently employed, and perceived job insecurity.
# Statistical methods
The relationships between background factors and poor mental health in 2010 are presented as percentages and age-adjusted IRRs, which are a good estimate of relative risks, using a modified Poisson regression model with robust standard errors. [bib_ref] A modified poisson regression approach to prospective studies with binary data, Zou [/bib_ref] In the multivariate analysis, over-indebtedness was tested against the outcome with the stepwise addition of educational level, country of origin and precarious employment.
The tests for effect modification were performed with simple dummy variables. The synergy indices were calculated as proposed by Rothman.according to which a synergy index >1 may indicate a synergistic effect, and a synergy index <1 an antagonistic effect between two exposure variables. In the interaction analysis, educational level was dichotomised as '12 years or less' versus '13 years or more'. There was a negative relationship between educational level and over-indebtedness (see the Results section), while the relationship with the outcome variable poor mental health at follow-up was positive (see the Results section), and therefore a high educational level was designated as exposure.
Two standard statistical analysis programmes were used, IBM SPSS Statistics for Windows, V.22.0, Armonk, NY: IBM Corp. and Stata Statistical Software: Release V.12, College Station, TX: StataCorp LP.
## Open access
Patient and public involvement statement Patient or public involvement in the design, reporting or dissemination of results was not possible. [fig_ref] Table 1: Background characteristics at baseline in 1999/2000 of 1256 men and 1539 women... [/fig_ref] shows the prevalence of over-indebtedness in the total study population at baseline and also separately by gender. Overall, the prevalence of over-indebted individuals was 4.2%, but with a clear difference between men and women (2.8 vs 5.3%, respectively). The only clear age-related difference in over-indebtedness consisted of a low prevalence in participants aged 55 and above. Overindebtedness was more common in those with an educational level of 10-12 years, versus those with less or with more years. When educational level was dichotomised as '12 years or less' versus '13 years or more', 5.5% of those with a low educational level were over-indebted at baseline, versus 2.7% of those with a high educational level (data not shown in tables). Over-indebtedness was three times as common among those born in another country than Sweden. As perhaps expected, over-indebtedness was four times as prevalent among those who reported a precarious employment situation versus those with a non-precarious situation and almost three times as prevalent among those who rented their dwelling (compared with those who owned it). The gender pattern was largely similar for all factors. shows the results of the longitudinal analysis, where poor mental health at follow-up in 2010 was used as the outcome. The age-adjusted risk for poor mental health at follow-up in over-indebted individuals, measured as IRR, was 2.2 (95% CI 1.7 to 2.8) in the entire group, 2.6 (1.8 to 3.9) for men and 1.9 (1.4 to 2.6) for women. The risk for poor mental health was four times higher in those who were 18-24 years at baseline, with an IRR of 4.1 (2.3 to 7.4), and decreased successively with increasing age, compared with those who were 55 or older at baseline, which was used as the reference group. That pattern was similar in men and women. Those with an educational level of 13-14 years at baseline had a slightly higher risk of poor mental health at follow-up (IRR=1.4; 95% CI 1.01 to 2.0) than those with 9 years or less, which was used as the reference. When educational level was dichotomised, 15.1% of those with a high educational level had poor mental health at follow-up versus 13.9% of those with a low educational level (data not shown in tables). Foreign birth was related to poor mental health in men only, with an IRR of 1.6 (1.01 to 2.6). Having a precarious employment situation during the period 1999/2000 to 2005 was statistically significantly related to poor mental health in 2010 (IRR=1.4; 95% CI 1.2 to 1.7), although also here with a gender difference; the IRR for men was 1.8 (1.3 to 2.4) and for women 1.2 (0.95 to 1.5). Housing tenure (rented dwelling) at baseline was not related to poor mental health in 2010. shows the effect on the risk estimate for poor mental health at follow-up when exposed to overindebtedness during the period 1999/2000 to 2005, with a stepwise introduction of the main covariates. In the final step, when precarious employment situation during Open access 1999/2000 to 2005 was introduced into the analysis, the risk estimate for poor mental health regarding exposure to over-indebtedness 1999/2000 to 2005 decreased about 10% for the entire group. The decrease in IRR in the fully adjusted model was slightly larger for men than for women. As a general remark, we find it possible that the decrease in IRR is a case of over-adjustment, since all the introduced covariates could be included in the same causal chains, as well as in separate ones, and we have no way to disentangle this.
# Results
As it is of considerable interest to investigate whether there are certain segments of the population that are more vulnerable concerning the risk of developing poor mental health when exposed to over-indebtedness, we examined this concerning age and level of education. As seen in table 4A, there was no indication of a synergistic effect of young age and over-indebtedness on the risk of poor mental health at follow-up. However, as seen in table 4B, while there was likewise no interaction between a high educational level and over-indebtedness in women, a higher risk for poor mental health was seen in men with this particular combination of risk factors. The IRR for men was 3.8 (95% CI 1.8 to 8.0), and the resulting synergy index 1.8.
# Discussion
This study confirmed that vocationally active individuals who are exposed to over-indebtedness have an increased risk of developing poor mental health. Age-adjusted estimates showed that the risk was almost tripled for men and doubled for women in this general adult population sample of vocationally active individuals . Considering that about 4% of all individuals in the cohort were exposed to over-indebtedness at baseline and probably substantially more during the follow-up period, overindebtedness should be regarded as an important general population risk factor for poor mental health. This is pertinent especially in the aftermath of the COVID-19 pandemic, during which the number of vulnerable individuals in the adult population is likely to increase considerably.
At the same time, especially from a health equity perspective, it must be acknowledged that the prevalence and to some extent the impact of exposure to over-indebtedness in the population depend on sociodemographic circumstances. Previous research provides few examples of how this could be integrated into a coherent theoretical framework regarding health equity. We speculate that the reason that younger individuals are more exposed to over-indebtedness than middle-aged or older ones may partly have a 'life-course' explanation (mortgage loans are taken early in life and paid off over time), but there is also a generational effect due to the expansion of credit markets over time. For the other group disparities in over-indebtedness, reasons must be sought elsewhere. Women were more exposed than men, foreign-born more than those born in Sweden, and those with a low educational level more than those with a high level. In short, it seems that the risk for over-indebtedness follows all the main societal power structures, that is, gender, social class and ethnicity. Therefore, it is an important factor to consider when formulating and implementing policies for promoting health equity.
Regarding socioeconomic factors that logically should be strongly related to over-indebtedness, that is, precarious employment and housing ownership (ie, nonownership), we found that both these factors were strongly related to over-indebtedness, but that only precarious employment was also related to the risk of developing poor mental health during our observation period. However, when we performed an analysis to control for confounding from precarious employment, the impact of over-indebtedness on poor mental health was only moderately reduced. Considering a bidirectional relation between over-indebtedness and poor mental health, the choice to exclude individuals with poor mental health at baseline could have yielded an underestimation of the true association between the variables.
Health inequality could be the result of both differential exposure and differential impact when once exposed. [bib_ref] Differential vulnerability and susceptibility: how to make use of recent development in..., Diderichsen [/bib_ref] From the discussion above, it seems obvious that there is a clear case of differential exposure regarding overindebtedness and sociodemographic factors. When we tested by means of two-way interactions whether we in addition could find any evidence for a differential impact, we concluded that this was not the case regarding age [fig_ref] Table 4A: Interaction analyses, by gender, with synergy indices regarding age group in 1999/2000... [/fig_ref]. We noted, however, that the combination of a high educational level and over-indebtedness [fig_ref] Table 4B: Interaction analyses, by gender, with synergy indices regarding educational level in 1999/2000... [/fig_ref] was associated with a synergy index of 1.0 in women, whereas the corresponding index in men was 1.8, thus indicating that these factors may act synergistically in a detrimental way for men. One could speculate that over-indebtedness may be especially hard to endure for a well-educated man, since over-indebtedness implies that one is financially not 'in control', whereas gender-based stereotypes of masculinity imply the opposite. Added to that might be the frustration caused by having invested many years in education and not being 'rewarded' by economic stability. Nevertheless, this finding relies on an analysis in which the number of exposed men is small, why the CIs are large. Therefore, this result needs to be replicated in other research.
As shown in many studies as well as in this one, the basic power relations in society, among which would be those determined by gender, social class and ethnicity, show a strong relation to the risk of becoming over-indebted. This is not surprising, because all the above mentioned power relations determine access to factors that could avoid a short-term or long-term situation where credit is needed for securing stable living conditions. Not being able to do so will induce both a loss of control and a loss of material resources, situations that have been shown to increase individuals' susceptibility to poor health development, especially poor mental health. However, the 'Faustian deal' of using credit in this situation could make Open access things worse, since a spiral of failing credit and declining mental health could place an individual in a situation where he or she in fact has sold a considerable part of future earnings to a moneylender institute. The possibility of having such debts cancelled in a regulated way, which is open to business enterprises through the legal bankruptcy framework, is very restricted for individuals in the Swedish context. This is particularly problematic in a health equity perspective because of the strong association between societal power structures and the risk of becoming over-indebted, all of which could be magnified in the wake of the COVID-19 pandemic.
The results of this study are therefore very relevant for policy, not only regarding public health and health equity, but also for policy formulation and implementation in the judicial and social realms. A specific policy recommendation is to expand the possibilities to write off debt for the individual, in a similar way that now is possible for business enterprises, in order to prevent a vicious circle of poor mental health and increasing debt, which seems maleficent for all involved stakeholders. Because of the growing problem with over-indebtedness, it seems urgent to develop the information systems concerning individual's and household's financial situation, in a manner that high quality information becomes available, not only concerning income but equally concerning financial assets and debt. This, especially since increasing inequality appears to be better mirrored by the latter, rather than the former.
# Strengths and weaknesses
This study uses a prospective design using a large randomly selected general population sample of individuals in the occupationally active age-bracket in southern Sweden. A first selection took place in the very establishment of the Scania Public Health Cohort, since the response rate was 58%.This opens up for the introduction of selection bias. However, two studies that were specifically designed to elucidate selection bias in this cohort concluded that apart from an under-representation of foreign-born individuals, the responders were not statistically different from non-responders regarding other sociodemographic characteristics, 20 nor were any signs of selection bias evident since the risk estimates for common illnesses (including mental health) were very similar between participants and drop outs. In the former study, it was also noted that healthcare usage did not differ between responders and non-responders during the baseline survey year. Moreover, such bias could in fact give rise to an underestimation, rather than an overestimation, of the observed differences, since it seems reasonable that heavily indebted individuals for several reasons (lack of postal address, etc) could be under-represented in the sample, especially if their mental health was poor.
The main exposure, over-indebtedness, was based on two questions that in combination have been deemed valid by other authors [bib_ref] Economic hardships in adulthood and mental health in Sweden. The Swedish national..., Ahnquist [/bib_ref] to identify individuals with a financial situation indicating that their income for a longer period of time was insufficient to meet their regular household bills without having other financial assets (the definition of over-indebtedness). We believe that this measure might be more objective, since it builds on two factual circumstances, than measures solely based on a subjective assessment of financial stress. The prevalence of estimated over-indebtedness in our sample seems to be reasonably on par with the prevalence of over-indebtedness in the general Swedish population at the time of assessment.A direct measure of over-indebtedness would be preferable, but this seems difficult to implement either by register studies or by surveys, since there is a whole range of credit forms to be covered, from regular bank loans to personal or even criminal types of credit that are very difficult to assess in a reliable and valid manner.
Our measure of precarious employment is based on relevant indicators and has been used in a recent study [bib_ref] Precarious employment is a risk factor for poor mental health in young..., Canivet [/bib_ref] ; however, it has not been validated in relation to other existing instruments. Given that confounding from a precarious employment situation was rather weak, and considering that this factor very well could be involved in the same causal chain as the main exposure, we find it unlikely that the discussed possible misclassification had biased our finding to any degree of importance. Although we have assessed potential confounding from main socioeconomic background factors known to be linked to over-indebtedness, such as gender, educational level, age and migration status, we cannot fully rule out residual confounding from other factors. This study was based on self-reported data. However, the GHQ instrument has been extensively validated, and recent studies carried out in Sweden have shown that the GHQ-12 performs acceptably in detecting depression in the general population [bib_ref] Validity of the 12-Item version of the general health questionnaire in detecting..., Lundin [/bib_ref] and also performs acceptably in discriminating out-patients with depression and anxiety from healthy controls. [bib_ref] Discriminant validity of the 12-Item version of the general health questionnaire in..., Lundin [/bib_ref]
# Conclusions
This prospective study showed that exposure to overindebtedness had a negative effect on vocationally active individual's mental health. It also showed that overindebtedness is not uncommon in the general working population, and it therefore seems to be a significant contributor to poor mental health at the population level, especially since loss of employment and secure income is likely to increase considerably in the aftermath of COVID-19. Moreover, over-indebtedness was found to be clearly related to an unfavourable position with regard to societal power relations. The findings are therefore of importance not only for policy makers in the area of public health and health equity, but also for policy areas that determine the risk for becoming over-indebted, such as the social and judicial realms.
## Open access
Contributors The study was conceived and designed by P-OÖ, CC and AV. The analysis was done by MM. TB drafted the first version of the manuscript. All authors took part in interpreting the results and writing subsequent drafts. P-OÖ is the guarantor, i.e. the author responsible for the overall content of this work.
Funding This work was funded by FORTE (Swedish Research Council for Health, Working Life and Welfare; Grant Number 2013-1269) and the Pufendorf Institute at Lund University (Credit Society Project). The funding institutions had no role in the design of the study, data collection, analysis and interpretation of data, or in writing of the manuscript.
Competing interests None declared.
Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
Patient consent for publication Not applicable.
Ethics approval This study involves human participants and was approved by Ethics name ID:The Regional Ethical Review Board at Lund University, Sweden, ID numbers: 1999-99; 2005-471 and 2010-392. Participants gave informed consent to participate in the study before taking part.
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement Data are available upon reasonable request.
Open access This is an open access article distributed in accordance with the Creative Commons Attribution 4.0 Unported (CC BY 4.0) license, which permits others to copy, redistribute, remix, transform and build upon this work for any purpose, provided the original work is properly cited, a link to the licence is given, and indication of whether changes were made. See: https://creativecommons.org/ licenses/by/4.0/.
## Orcid id
Per-Olof Östergren http://orcid.org/0000-0002-7903-6668
[table] Table 1: Background characteristics at baseline in 1999/2000 of 1256 men and 1539 women from the Scania Public Health Cohort, in relation to over-indebtedness Over-indebtedness at least once in 1999/2000 or 2005 in relation to the outcome poor mental health at follow-up in 2010, with forward stepwise addition of potential confounders; educational level in 1999/2000, country of origin, and precarious employment at least once in 1999/2000 or 2005 [/table]
[table] Table 4A: Interaction analyses, by gender, with synergy indices regarding age group in 1999/2000 and over-indebtedness at least once in 1999/2000 or 2005, in relation [/table]
[table] Table 4B: Interaction analyses, by gender, with synergy indices regarding educational level in 1999/2000 and over-indebtedness at least once in 1999/2000 or 2005, in relation to poor mental health in 2010 [/table]
|
Complex genetic alterations contribute to rapid disease progression in an ALK rearrangement lung adenocarcinoma patient: a case report
Anaplastic lymphoma kinase tyrosine kinase inhibitors (ALK-TKIs) have been found to significantly improve the quality of life and survival in ALK-positive non-small cell lung cancer (NSCLC) patients. However, the duration of responses is limited by drug resistance. Genetic heterogeneity of ALKpositive tumors could potentially explain the differences in individual patient outcomes. We performed nextgeneration sequencing (NGS) on plasma samples, pleural effusion samples, and tissue re-biopsy obtained at various treatment milestones from an ALK rearrangement lung adenocarcinoma patient undergoing targeted therapy. The liver metastases of the EML4-ALK NSCLC patient presented rapid progression after 3.5 months of alectinib, while the other lesions showed good partial response. Targeted NGS identified the newly emerged MET amplification except for EML4-ALK in plasma ctDNA and liver lesions. Subsequently, a clinical benefit was achieved one month after the commencement of crizotinib, a dual ALK and MET inhibitor; however, the patient experienced disease progression another month later. Several rounds of ALK-TKI combination therapy were tried but failed. Concurrent genetic alterations, including loss-of-function mutations in FBXW7 and MLL3, may mainly contribute to poor prognosis in the patient. It highlighted the molecular profiling by using NGS can be useful in identifying the heterogeneity across lesions and the resistance mechanism of targeted treatments.
# Introduction
Anaplastic lymphoma kinase (ALK) gene rearrangements have been reported in approximately 5% of non-small cell lung cancers (NSCLCs) and function as oncogenic driver event [bib_ref] Genomic alterations in lung adenocarcinoma, Devarakonda [/bib_ref]. The second-generation ALK inhibitor alectinib demonstrated superior efficacy and lower toxicity compared to the first-generation ALK inhibitor crizotinib in advanced ALK-rearranged NSCLCs (3), establishing alectinib as the new standard first-line therapy. The sequential therapy of ALK-tyrosine kinase inhibitors (TKIs) allows long survivals up to more than 7 years [bib_ref] 06 The Sequential Therapy of Crizotinib Followed by Alectinib: Real World Data..., Watanabe [/bib_ref]. Despite responding to ALK-TKIs initially, rapid progression may have occurred, thereby limiting the prolonged effectiveness of ALK-TKIs. Genetic heterogeneity of ALK-positive tumors could potentially explain the differences in individual patient outcomes [bib_ref] Impact of TP53 mutation status on systemic treatment outcome in ALKrearranged non-small-cell..., Kron [/bib_ref]. Here, we present a case of an ALK-rearranged [fig_ref] Figure 1: Clinical response to ALK-TKIs therapy of primary and liver metastasis lesions [/fig_ref]. However, the CT scan showed a dramatic progression of liver metastasis after 3.5 months of treatment, while the other lesions showed good partial response. Then the patient was treated with radiofrequency ablation of liver tumors, but the liver lesions continued to grow rapidly. Meanwhile, the next-generation sequencing detected both EML4-ALK rearrangement and MET amplification in plasma ctDNA and liver lesions . Crizotinib (250 mg twice a day), a dual inhibitor of ALK and MET, was then administered in September 2019. Symptoms such as chest and back pain significantly improved within the first month, and the diameter of the liver metastases decreased from 66 to 50 mm. But unfortunately, the patient developed multiple low-density nodules in the liver in November 2019. To explore new therapeutic strategies, we used ctDNA analysis to track the evolution of resistance during treatment. The result revealed retained EML4-ALK fusion (AF =14.5%) without the amplification of MET .
Since the patient experienced further disease progression with new bone lesions, pemetrexed, cisplatin, and bevacizumab were given as third-line treatment. After two cycles of chemotherapy, liver metastases, especially those in the left lobe, progressed again, so the patient switched to a combination treatment of alectinib and cabozantinib (60 mg twice a day), a multikinase inhibitor with activity against MET, in March 2020. He achieved stable disease (SD) after two months, and cabozantinib discontinued due to grade 3 hand-foot syndrome (HFS). At that time, a second biopsy specimen showed low PD-L1 expression with a tumor proportion score (TPS) of 1-2% by immunohistochemistry (Dako 22C3). The patient had shown increased ctDNA gene mutation frequencies while the amplification of MET was still not detected . Based on the results from the phase III ALTER-0303 trial (Clinical Trial Registry ID: NCT 02388919) of anlotinib in China, anlotinib (12 mg) and alectinib were administered, but the patient's condition continued to deteriorate. Repeated analyses indicated the presence of an inactivating mutation in FBXW7 (p.M268Dfs*18) during the disease which may be sensitive to mTOR inhibitor [bib_ref] Temsirolimus therapy in a patient with lung adenocarcinoma harboring an FBXW7 mutation, Villaruz [/bib_ref]. After a discussion with the patient and approval from his insurer, he was treated with lorlatinib (100 mg) and everolimus (10 mg) in May 2020. After 20 days of treatment, the diameter of the liver metastases decreased from 225 to 167 mm. However, the disease had substantially progressed and the patient died on June 15, 2020. [fig_ref] Figure 1: Clinical response to ALK-TKIs therapy of primary and liver metastasis lesions [/fig_ref] illustrated the flow of treatments and image evaluation.
All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee(s) and with the Helsinki Declaration (as revised in 2013). Written informed consent was obtained from the patient.
# Discussion
Though the number of ALK-rearranged lung cancer patients is relatively small, multiple ALK-TKIs have a relatively longer progression-free survival (PFS) period and have been approved for clinical use. Alectinib, a second-generation ALK-TKI, improves prognosis of treatment-naive ALKpositive NSCLCs, with an objective response rate of 82.9% and median PFS of 34.8 months [bib_ref] Alectinib versus Crizotinib in Untreated ALK-Positive Non-Small-Cell Lung Cancer, Peters [/bib_ref]. However, resistance to TKIs is inevitable and the mechanism of acquired resistance to alectinib in patients with ALK rearrangement has not yet been completely identified. In our case, the patient showed a mixed response to liver metastasis and primary lung tumors after initial alectinib treatment Transl Cancer Res 2021;10 (6) [fig_ref] Figure 1: Clinical response to ALK-TKIs therapy of primary and liver metastasis lesions [/fig_ref]. This phenomenon may attribute to intertumoral genetic heterogeneity. To clarify the resistance mechanism, a panel sequencing was performed in the liver metastasis.
In addition to the EML4-ALK fusion previously discovered in the primary lesion, MET amplification was detected.
Previous studies have found evidence that cMET activation through MET gene amplification can potentially confer resistance to alectinib but not to crizotinib [bib_ref] Crizotinib can overcome acquired resistance to CH5424802: is amplification of the MET..., Gouji [/bib_ref] [bib_ref] Non-Small Cell Lung Cancer Cells Acquire Resistance to the ALK Inhibitor Alectinib..., Isozaki [/bib_ref]. Our case showed a clinical benefit of crizotinib despite of drug resistance that occurred rapidly after. A negative result for the amplification of MET was found in the cfDNA at that time, which may be due to non-shedding of the amplified MET, real elimination of the amplified clone, or too little ctDNA in the plasma sample. Considering the absence of the amplification of MET was confirmed in rebiopsies from hepatic biopsies after the progression of alectinib and cabozantinib, we cannot exclude that crizotinib caused or contributed to the disappearance of the MET amplification tumor clone, as we did observe a response of some hepatic lesions to this drug. Except for MET amplication, some key variants were list in table1, especially those present from baseline throughout the time course. Their pathogenicity and association with ALK-TKIs resistance were assessed using the public databases and published literature, such as ClinVar, Catalogue of Somatic Mutations in Cancer (COSMIC), and PubMed. Previous studies suggested a potential role of TP53 mutations in poor therapeutic response and outcome in ALK/TP53 co-mutated patients [bib_ref] Concomitant TP53 mutations with response to crizotinib treatment in patients with ALK-rearranged..., Song [/bib_ref] [bib_ref] Correlation of baseline molecular and clinical variables with ALK inhibitor efficacy in..., Camidge [/bib_ref]. The deletion of FBXW7 in NSCLC was also found associated with poor overall survival [bib_ref] FBXW7 deletion contributes to lung tumor development and confers resistance to gefitinib..., Xiao [/bib_ref]. FBXW7 is a member of the F-box protein family, which controls proteasome-mediated degradation of oncoproteins such as rapamycin (mTOR), c-Myc, cyclin E, Mcl-1, Jun, and Notch 1 [bib_ref] FBXW7: a critical tumor suppressor of human cancers, Yeh [/bib_ref]. In vitro studies showed that the loss of FBXW7 leads to resistance to gefitinib and crizotinib [bib_ref] FBXW7 deletion contributes to lung tumor development and confers resistance to gefitinib..., Xiao [/bib_ref] [bib_ref] Targeting FBW7 as a Strategy to Overcome Resistance to Targeted Therapy in..., Ye [/bib_ref]. Villaruz et al. reported a case that harbored an FBXW7 mutation without EGFRmutant or ALK rearrangement responded to the mTOR inhibitor temsirolimus [bib_ref] Temsirolimus therapy in a patient with lung adenocarcinoma harboring an FBXW7 mutation, Villaruz [/bib_ref]. In our case, a combination of ALK-TKI and mTOR inhibitor did not seem to overcome ALK-TKI resistance, potentially due to posterior line of therapy or different choice of mTOR inhibitor. Moreover, Ye et al. reported a PI3K/Akt-and MEK/Erk-independent resistance mechanism by which loss of FBXW7 leads to targeted therapy resistance via stabilization of anti-apoptotic protein Mcl-1 [bib_ref] Targeting FBW7 as a Strategy to Overcome Resistance to Targeted Therapy in..., Ye [/bib_ref] , suggesting that FBXW7-mediated activation of multiple signaling pathways might contribute to ALK-TKIs resistance. Notably, FBXW7 inactivation is known to partly induce TKI-resistance by promoting epithelial-mesenchymal transition (EMT) [bib_ref] FBXW7 deletion contributes to lung tumor development and confers resistance to gefitinib..., Xiao [/bib_ref]. EMT has recently been implicated in resistance to lorlatinib in patient-derived cell lines [bib_ref] Diverse Resistance Mechanisms to the Third-Generation ALK Inhibitor Lorlatinib in ALK-Rearranged Lung..., Recondo [/bib_ref] and to alectinib and lorlatinib in a patient [bib_ref] Changing ALK-TKI-Resistance Mechanisms in Rebiopsies of ALK-Rearranged NSCLC: ALK-and BRAF-Mutations Followed by..., Urbanska [/bib_ref]. Interestingly, we identified a concomitant nonsense mutation in MLL3 at relapse on Crizotinib. Some studies demonstrated the function of mutant MLL3 in facilitating tumor EMT [bib_ref] KMT2C Mutations in Diffuse-Type Gastric Adenocarcinoma Promote Epithelial-to-Mesenchymal Transition, Cho [/bib_ref] and involvement in lung cancer development and survival [bib_ref] COMPASS Ascending: Emerging clues regarding the roles of MLL3/KMT2C and MLL2/ KMT2D..., Fagan [/bib_ref]. Together, these observations indicate that EMT induced by the deletion of FBXW7 and MLL3 may represent the main mechanism of resistance to ALK-TKIs. There are not enough samples to confirm the presence of EMT by using markers such a positive immunostaining for vimentin and loss of E-cadherin expression. But we found that the third tumor rebiopsy taken from hepatic metastases had almost completely lost the expression of the adenocarcinoma-marker CK7 and TTF1. It is worth-noting that mutations crossing take place between MET and MLL3, revealing the resistance heterogeneity and selection of tumor evolution.
In addition to the above speculation, non-reciprocal ALK translocation with the retaining of the 5' region of the ALK gene was also observed in the patient samples , which was reported as a poor predictive marker in first-line crizotinib-treated ALK-rearranged NSCLCs [bib_ref] Detection of Nonreciprocal/Reciprocal ALK Translocation as Poor Predictive Marker in Patients With..., Zhang [/bib_ref]. In that study, three patients who harbored non-reciprocal/ reciprocal ALK translocation also did not benefit from alectinib therapy after the failure of crizotinib. However, it is still poorly understood how 5'-ALK DNA could contribute to poor prognosis of patients.
In conclusion, we report the case of an EML4-ALK fusion-positive NSCLC patient, who progressed rapidly during the different lines of ALK-TKIs therapy in a year. Genetics variations concurrent with EML4-ALK from tissues and ctDNA during disease may be accounted for the treatment response and prognosis of the patient. Molecular profiling by using NGS can be useful for monitoring tumor heterogeneity and clonal evolution during ALK-TKIs treatment in NSCLCs. Treatment strategises for these patients need further research. Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee(s) and with the Helsinki Declaration (as revised in 2013). Written informed consent was obtained from the patient.
## Foundation of guangdong grants (2016a030310242).
## Footnote
Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the noncommercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.
[fig] Figure 1: Clinical response to ALK-TKIs therapy of primary and liver metastasis lesions (relevant changes indicated by arrows). (A) Computed tomography (CT) scans of the chest from May 2019 to June 2020; (B) Liver CT scans from May 2019 to June 2020. [/fig]
|
WHO Grade I Meningioma Metastasis to the Lung 26 Years after Initial Surgery: A Case Report and Literature Review
Metastases from meningioma grade I are especially rare. We describe a case of a 65-year-old male with meningioma WHO grade I with a history of local recurrence and distant metastasis to the lung 26 years after the initial surgery. The original tumor was localized at the occipital low convex and invaded into the venous sinus and posterior cranial fossa; it was resected. About 15 years later, the tumor recurred in the posterior cranial fossa and g -knife radiosurgery was performed. About 4 years later, the recurred tumor was resected at our hospital. Another 7 years later, the tumor recurred in the same area and right middle cranial fossa. All tumors except that inside the venous sinus were excised. All specimens obtained were classified as meningioma WHO grade I. Preoperative examination of the third operation revealed a nodule in the lower lobe of the right lung. The nodule grew gradually. Four months after the third surgery, partial resection of the right lung was performed. Histology indicated meningioma WHO grade I. The two lesions in the cranium and lung lesions were subjected to fluorescence in situ hybridization of the NF2 gene, and the three specimens had similar findings, genetically confirming them to be metastases of the intracranial meningioma. A literature review of past cases of meningioma progression revealed that the mean duration to metastasis is 12.5, 6.8, 3.7 years for grades I, II, and III, respectively. The current case therefore has an extended time frame.of malignancy. 2-6) Distant metastasis of a meningioma is extremely rare, at 0.15%, 7) but this may be an underestimation. 8) Because of the rare nature of extracranial metastases, no standard management protocol has been established and the prognosis for these patients is unknown. 9,10) Most meningiomas are WHO grade I and about 7-15% and 2-4% of meningiomas are WHO grade II and III, respectively. 11,12) In metastatic cases, there are more cases of grades II and III. We describe a case of WHO grade I meningioma that had metastasized to the lung and reviewed the relevant literature.CaseThe patient was a 65-year-old male who, 26 years previously, had undergone surgery (at a previous hospital) for occipital low convexity meningioma that had invaded the transverse sinus, sigmoid sinus, jugular vein, and posterior cranial fossa. Part of the tumor in the sinus was left untouched, the remainder of the tumor was excised and a subtotal resection was performed. The tumor was diagnosed as fibrous meningioma. Sixteen years after the operation, the tumor recurred in the posterior cranial fossa and γg-knife radiosurgery was performed at another hospital. A further 3 years later, the tumor recurred and the patient came to our hospital for a second surgery. The second operation was performed and the tumor was completely excised (Figs. 1A and 1B). Histological sections showed a proliferation of spindle cells with oval and elongated nuclei arranged in fascicular or whorl-like arrangements. No atypical features were detected and WHO grade I meningioma was considered(Fig. 1C). Seven years after the second operation, the patient suffered headaches and nausea. Magnetic resonance imaging (MRI) revealed a recurring tumor (size: 6.5 × 5.7 × 5.9 cm) in the location of the previous operation; there was also another tumor in the middle cranial fossa (5.0 × 4.7 × 3.8 cm) where the first operation had been executed (Figs. 2A and 2B). A third operation was scheduled and the preoperative examination was performed. A chest X-ray revealed a nodule shadow in the right middle lung field. In a chest computed tomography (CT), a 2 cm nodule was seen in the lower lobe of the right lung(Fig. 3A). At this time point, treatment for the intracranial tumor was performed which included embolization of the feeding artery and tumor resection (Simpson grade II; Figs. 2C and 2D). There was no macroscopically visible tumor infiltration into the bone, and no histological examination of tumor infiltration was performed. Histologically, the sections showed proliferation of meningothelial tumor cells with oval or elongated nuclei arranged in
Metastases from meningioma grade I are especially rare. We describe a case of a 65-year-old male with meningioma WHO grade I with a history of local recurrence and distant metastasis to the lung 26 years after the initial surgery. The original tumor was localized at the occipital low convex and invaded into the venous sinus and posterior cranial fossa; it was resected. About 15 years later, the tumor recurred in the posterior cranial fossa and g -knife radiosurgery was performed. About 4 years later, the recurred tumor was resected at our hospital. Another 7 years later, the tumor recurred in the same area and right middle cranial fossa. All tumors except that inside the venous sinus were excised. All specimens obtained were classified as meningioma WHO grade I. Preoperative examination of the third operation revealed a nodule in the lower lobe of the right lung. The nodule grew gradually. Four months after the third surgery, partial resection of the right lung was performed. Histology indicated meningioma WHO grade I. The two lesions in the cranium and lung lesions were subjected to fluorescence in situ hybridization of the NF2 gene, and the three specimens had similar findings, genetically confirming them to be metastases of the intracranial meningioma. A literature review of past cases of meningioma progression revealed that the mean duration to metastasis is 12.5, 6.8, 3.7 years for grades I, II, and III, respectively. The current case therefore has an extended time frame.
Keywords: meningioma, extracranial metastasis, WHO grade I, lung, sinus invasion of malignancy. [bib_ref] Two cases of atypical meningioma with pulmonary metastases: a comparative cytogenetic analysis..., Frydrychowicz [/bib_ref] [bib_ref] Two-year survival after multiple bilateral lung metastasectomies for cranial meningioma, D'aiuto [/bib_ref] [bib_ref] Surgical resection of pulmonary metastases from meningioma: report of a case, Kanzaki [/bib_ref] [bib_ref] Metastatic meningioma to the lung with multiple pleural metastases, Kaminski [/bib_ref] [bib_ref] Bone and lung metastases from intracranial meningioma, Fabi [/bib_ref] Distant metastasis of a meningioma is extremely rare, at 0.15%, [bib_ref] Meningioma metastatic to the lung, Adlakha [/bib_ref] but this may be an underestimation. [bib_ref] Distant metastases in meningioma: an underestimated problem, Surov [/bib_ref] Because of the rare nature of extracranial metastases, no standard management protocol has been established and the prognosis for these patients is unknown. [bib_ref] Pulmonary and pleural metastasis of intracranial anaplastic meningioma in a 3-year-old boy:..., Honda [/bib_ref] [bib_ref] Visceral and bone metastases of a WHO grade 2 meningioma: a case..., Paix [/bib_ref] Most meningiomas are WHO grade I and about 7-15% and 2-4% of meningiomas are WHO grade II and III, respectively. [bib_ref] Meningioma grading: an analysis of histologic parameters, Perry [/bib_ref] [bib_ref] Classic, atypical, and anaplastic meningioma: three histopathological subtypes of clinical relevance, Maier [/bib_ref] In metastatic cases, there are more cases of grades II and III. We describe a case of WHO grade I meningioma that had metastasized to the lung and reviewed the relevant literature.
## Case
The patient was a 65-year-old male who, 26 years previously, had undergone surgery (at a previous hospital) for occipital low convexity meningioma that had invaded the transverse sinus, sigmoid sinus, jugular vein, and posterior cranial fossa. Part of the tumor in the sinus was left untouched, the remainder of the tumor was excised and a subtotal resection was performed. The tumor was diagnosed as fibrous meningioma. Sixteen years after the operation, the tumor recurred in the posterior cranial fossa and γg-knife radiosurgery was performed at another hospital. A further 3 years later, the tumor recurred and the patient came to our hospital for a second surgery. The second operation was performed and the tumor was completely excised (Figs. 1A and 1B). Histological sections showed a proliferation of spindle cells with oval and elongated nuclei arranged in fascicular or whorl-like arrangements. No atypical features were detected and WHO grade I meningioma was considered . Seven years after the second operation, the patient suffered headaches and nausea. Magnetic resonance imaging (MRI) revealed a recurring tumor (size: 6.5 × 5.7 × 5.9 cm) in the location of the previous operation; there was also another tumor in the middle cranial fossa (5.0 × 4.7 × 3.8 cm) where the first operation had been executed (Figs. 2A and 2B). A third operation was scheduled and the preoperative examination was performed. A chest X-ray revealed a nodule shadow in the right middle lung field. In a chest computed tomography (CT), a 2 cm nodule was seen in the lower lobe of the right lung . At this time point, treatment for the intracranial tumor was performed which included embolization of the feeding artery and tumor resection (Simpson grade II; Figs. 2C and 2D). There was no macroscopically visible tumor infiltration into the bone, and no histological examination of tumor infiltration was performed. Histologically, the sections showed proliferation of meningothelial tumor cells with oval or elongated nuclei arranged in
# Introduction
Intracranial meningiomas are the most common tumors of the central nervous system, accounting for 13-26% of all primary intracranial tumors; 1) they are typically solitary and benign. Despite being pathologically benign, there are rare cases of metastases to extracranial sites, which take a course . Although a small focus of necrosis was seen in the posterior fossa specimen, it was considered to be an effect of vascular embolization. These findings were consistent with WHO grade I meningioma. Ki-67(MIB-1) labeling index of the posterior fossa and temporal lesions were 4.8% and 5.1%, respectively, making them slightly high. Since there is a tendency for tumorous lesions in the lung to expand, a thoracoscopic partial resection of the right lower lobe was performed 2 months after the third cranial surgery. Histologically, the section showed proliferation of spindle shaped tumor cells with round to oval nuclei forming intersecting short fascicles . Although, necrosis was seen in the center of the tumor, no other atypical features were evident . We performed fluorescence in situ hybridization (FISH) to reveal the presence or lack of heterozygosity of the NF2 gene (22q 12.2) in both the resected lung tumor and the intracranial tumor. Probes for NF2 assessment included a Fluorescein isothiocyanate (FITC, green) -labeled chromosome 22 centromeric (CEP22q) probe and Texas Red-labeled, locus-specific NF2 probe (Abnova, Walnut, CA, USA). Results revealed that both these tumors had loss of heterozygosity of the NF2 gene . These findings were consistent with metastasis of meningioma WHO grade I. No adjuvant therapy was performed. The patient is alive 12 months after surgery of the pulmonary lesion without recurrence of the lung and head lesions.
# Discussion
Meningioma was found in the patient's lung, 26 years after the first surgery. The lesion was histologically similar to the intracranial lesion, and pulmonary metastasis of the meningioma was suspected. Distant metastases of grade I meningioma are rare. Despite this rarity, primary meningioma may occur in the lung. In order to confirm the genetic identity of these tumors, FISH was performed, and it was found that the NF2 gene (22q 12.2) was deleted to the same degree in both intracranial and lung lesions [fig_ref] Table 1: Comparison of NF2 deletion pattern with FISH in lung and brain meningioma [/fig_ref]. Thus, we confirmed genetically and morphologically that the intracranial lesions and the pulmonary lesion were identical.
The higher the tumor grade, the higher the incidence of distant metastasis. The incidence of distant metastases in grades II and III are 5% and 30%, respectively. [bib_ref] Metastatic meningiomas: an unusual clinical and pathological diagnosis with highly variable outcome, Forest [/bib_ref] The route for distant metastases includes hematogenous metastasis via the internal jugular vein system and paraspinal venous plexus, lymphogenous metastasis, and cerebrospinal fluid dissemination. [bib_ref] Extraneural metastases from cranial meningioma: a case report, Abboud [/bib_ref] As far as we are aware, there has been no report summarizing the time period until the occurrence of distant metastases or the course after distant metastasis for each grade. To examine these, we reviewed the English-language literature published since 2007 using the PubMed search engine with the terms ("meningioma" AND "metastasis"). This search was performed on January 31st, 2018. We excluded cases in which metastases were found first or where simultaneous primary tumor and metastasis were identified. Cases of suspected drop metastasis or cerebrospinal fluid dissemination were also excluded. We reviewed 35 articles and 48 (present case included) cases of meningioma metastasized to extracranial sites. [bib_ref] Two cases of atypical meningioma with pulmonary metastases: a comparative cytogenetic analysis..., Frydrychowicz [/bib_ref] [bib_ref] Surgical resection of pulmonary metastases from meningioma: report of a case, Kanzaki [/bib_ref] [bib_ref] Distant metastases in meningioma: an underestimated problem, Surov [/bib_ref] [bib_ref] Pulmonary and pleural metastasis of intracranial anaplastic meningioma in a 3-year-old boy:..., Honda [/bib_ref] [bib_ref] Visceral and bone metastases of a WHO grade 2 meningioma: a case..., Paix [/bib_ref] The median age of the extracranial metastases was 61.5 years (range 3-82 years) with a female predominance (30 females, 17 males, and not described in one case). Regarding the WHO grade of the primary tumor, 19 (39.6%) cases were grade I, 14 (29.2%) cases were grade II, 12 (25%) cases were grade III, and 3 (6.3%) cases were not described. Regarding metastatic tumor grading, 12 (25%) cases were grade I, 15 (31.3%) cases were grade II, 19 (39.6%) cases were grade III, and two (4.2%) cases were not described. In 13 cases, pathological upgrade was confirmed at the time of distant metastasis from initial surgery. With regard to the time of upgrading, 10 cases were at the time of local recurrence, three cases were metastasis, and one case showed local recurrence and metastasis simultaneously.
The period of primary tumor operation to metastasis varied depending on the tumor grade. The mean period until metastasis in grades I, II, and III was 11.0, 5.4, and 2.0 years, respectively. The period until metastasis in our case was 26 years; this was fairly long compared with the mean period. Prognosis after metastasis was examined, and it was different depending on each grade. For grade I, there was one case of death with an unknown survival time, but there were no other deaths within the observation period in other cases. In grades II and III, the mean survival times were 3.3 and 0.98 years, respectively. Median survival period was 4 years and 1 year, respectively. Data indicate that when the tumor of a distant metastasis reached a high grade, the survival period was significantly shortened.
Known risk factors for metastasis and local recurrence are histopathological signs of malignant behavior. But this does not explain why grade I meningiomas metastasize. The pathophysiology leading to distant metastases may be a hematogenous spread originating from tumor invasion into the venous sinuses. In fact, 75% of patients with extracranial metastases of meningioma show an invasion of the venous sinus. [bib_ref] Extracranial metastatic meningioma, Figueroa [/bib_ref] In our review, five cases did not reveal any relationship between the sinus and tumor from the description and two cases were spinal meningioma. These seven cases were excluded when considering the association with the venous sinus. Invasion or contact with the venous sinus was observed in 60.1% (25/41) of cases. In this group, grade I accounted for 11/25 cases (44%). Grades II and III were 7/25 (28%) and 5/25 (20%), respectively. Not described was 2/25 (8%). In each grade, where an association with the sinus venosus was observed, the association was as follows: grade I, 61%; grade II, 54%; and grade III, 62.5%. These results indicate that contact with the venous sinus has the potential to develop extracranial metastases. In addition, in recent years, there are reports that CD90 becomes highly expressed in meningioma metastasis, and that chromosomal instabilities such as deletion of chromosomes 22 and 1 are associated with distant metastasis. [bib_ref] Two cases of atypical meningioma with pulmonary metastases: a comparative cytogenetic analysis..., Frydrychowicz [/bib_ref] [bib_ref] CD90 expression in atypical meningiomas and meningioma metastasis, Scognamiglio [/bib_ref] Our case was histologically WHO grade I but the MIB-1 labelling index was slightly high. At the time of first surgery, the tumor had already made extensive infiltration into the venous sinus and jugular vein. These factors may have contributed to meningioma WHO grade I metastasized to the lung.
# Conclusion
We experienced a case of meningioma WHO grade I that had metastasized to the lung 26 years after initial surgery. Confirmation was made morphologically and also genetically. However, despite the histological features indicating WHO grade I, sinus invasion may be associated with metastasis. Our review shows that the histological grading of the primary tumor was related to the period to metastasis. Furthermore, grading of the distant metastasized tumor was related to survival time.
# Limitation
In this case study, we considered the lung disease to be a metastasis, but we could not fully eliminate either metastasis or both primary. Because the somatic NF2 loss of heterozygosity was not examined, the possibility of meningiomatosis cannot be discounted.
## Conflicts of interest disclosure
All authors declare no conflicts of interest.
[fig] Figure 1, Figure 2, Figure 3: A and B) Pre-third operation. The tumor recurred in the same place (6.5 × 5.7 × 5.9 cm) as for the previous operation and additionally, it expanded from the ridge of the resected site to the middle cranial fossa (5.0 × 4.7 × 3.8 cm). (C and D) Tumors were excised for Simpson grade II. (E and F) The sections show proliferation of meningothelial tumor cells with oval or elongated nuclei arranged in intersecting short fascicles or small whorl-like structures [hematoxylin and eosin, (E) 100×, (A) In the chest CT, a 2 cm nodule was seen in the right lower lobe. (B and C) The section showed proliferation of spindle shaped tumor cells with round to oval nuclei forming intersecting short fascicles [hematoxylin and eosin, (C) 200×]. Necrosis was seen in the center of the tumor, but no other atypical features were evident. (D) NF2 gene fluorescence in situ hybridization (FISH) image. NF2 FISH predominantly demonstrated a single pair in the lung tumor. Arrows show one pair of red and green signals per cell. Red signal, NF2 (22q, 12.2); Green (FITC) signal, fascicles or small whorl-like structures (Figs. 2E and 2F). No high mitotic figures were found (1 < 10 high power fields) [/fig]
[table] Table 1: Comparison of NF2 deletion pattern with FISH in lung and brain meningioma [/table]
|
Human Monocyte-Derived Osteoclasts Are Targeted by Staphylococcal Pore-Forming Toxins and Superantigens
Staphylococcus aureus is the leading cause of bone and joint infections (BJIs). Staphylococcal pathogenesis involves numerous virulence factors including secreted toxins such as pore-forming toxins (PFTs) and superantigens. The role of these toxins on BJI outcome is largely unknown. In particular, few studies have examined how osteoclasts, the boneresorbing cells, respond to exposure to staphylococcal PFTs and superantigens. We investigated the direct impact of recombinant staphylococcal toxins on human primary mature monocyte-derived osteoclasts, in terms of cytotoxicity and cell activation with cell death and bone resorption assays, using macrophages of the corresponding donors as a reference. Monocyte-derived osteoclasts displayed similar toxin susceptibility profiles compared to macrophages. Specifically, we demonstrated that the Panton-Valentine leukocidin, known as one of the most powerful PFT which lyses myeloid cells after binding to the C5a receptor, was able to induce the death of osteoclasts. The archetypal superantigen TSST-1 was not cytotoxic but enhanced the bone resorption activity of osteoclasts, suggesting a novel mechanism by which superantigen-producing S. aureus can accelerate the destruction of bone tissue during BJI. Altogether, our data indicate that the diverse clinical presentations of BJIs could be related, at least partly, to the toxin profiles of S. aureus isolates involved in these severe infections.
# Introduction
Bone is a mineralized tissue in a constant renewal process called bone remodelling, provided by the coordinated action of two main cell types, osteoblasts, the bone-forming cells, and osteoclasts, the bone-resorbing cells [bib_ref] Interaction between osteoblast and osteoclast: impact in bone disease, Phan [/bib_ref]. « Osteoblasts derive from mesenchymal stems, while osteoclasts have a myeloid origin and share common features with macrophages including a phagocytic activity [bib_ref] Dual impact of live Staphylococcus aureus on the osteoclast lineage, leading to..., Trouillet-Assant [/bib_ref]. Osteoclast maturation involves the fusion of several mononucleated osteoclast precursors into giant multinucleated-cells endowed with the bone-matrix resorption ability. In a physiological context, bone integrity is maintained by a balance between osteoblastic and osteoclastic activities throughout life. During bone and joint infections (BJIs), including osteomyelitis and orthopaedic device infections, this process is impaired by the interaction of bacteria with bone tissue, leading to bone destruction [bib_ref] Bone and joint infections in hospitalized patients in France, 2008: clinical and..., Grammatico-Guillon [/bib_ref]. Indeed, Staphylococcus aureus, the leading cause of BJIs, is responsible for bone infections marked by progressive bone loss.
Numerous studies have investigated the direct impact of S. aureus on osteoblasts [bib_ref] Interaction of staphylococci with bone, Wright [/bib_ref] , [bib_ref] PSMs of hypervirulent Staphylococcus aureus act as intracellular toxins that kill infected..., Rasigade [/bib_ref]. It is well-known that this pathogen is able to adhere to osteoblasts, become internalized and survive intracellularly and/or induce cell death [bib_ref] Staphylococcus aureus small-colony variants are adapted phenotypes for intracellular persistence, Tuchscherr [/bib_ref]. Moreover, several studies have demonstrated the ability of live S. aureus to inhibit osteoclastogenesis and to increase bone resorption mediated by osteoclasts [bib_ref] Dual impact of live Staphylococcus aureus on the osteoclast lineage, leading to..., Trouillet-Assant [/bib_ref]. These observations suggest that S. aureus directly interacts with bone cells, modifying their functions of bone mineralization or bone resorption. Nevertheless, the pathophysiologic mechanisms, responsible for the bone destruction observed in BJI caused by S. aureus, remain incompletely understood.
S. aureus is able to act on remote target cells through secreted virulence factors, including toxins. S. aureus expresses a large panel of pore-forming toxins (PFTs) that target the host cell membrane including α-(Hla), β-(Hlb), γ-(HlgAB and HlgBC) haemolysins, and leukocidins (LukED, LukGH and the Panton Valentine Leukocidin [PVL]). PFT-induced permeabilization of the cytoplasmic membrane results in the efflux of intracellular metabolites and ultimately cell death. S. aureus also expresses superantigenic toxins such as the toxic shock syndrome toxin (TSST-1) or staphylococcal enterotoxins (SEA, SEB, etc.) responsible for a polyclonal activation and a massive proliferation of T cells independent of antigen specificity. Most of these staphylococcal toxins target immune cells derived from the myeloid lineage (monocytes, macrophages and dendritic cells), a characteristic which is thought to help S. aureus escape the immune system [bib_ref] The bicomponent pore-forming leucocidins of Staphylococcus aureus, Alonzo [/bib_ref].
Noteworthy, several of the aforementioned toxins exhibit some degree of specificity with respect to immune cells through the specific binding of cell surface receptors [bib_ref] The staphylococcal toxins γ-haemolysin AB and CB differentially target phagocytes by employing..., Spaan [/bib_ref]. Because osteoclasts derive from progenitors of the myeloid lineage, we hypothesized that their susceptibility to staphylococcal toxins share some similarities with the susceptibility of other myeloid cells such as macrophages, which could be of interest for our understanding of the pathophysiology of S. aureus BJIs. We thus tested the direct effect of a panel of recombinant staphylococcal toxins on primary human mature osteoclasts by assessing cell cytotoxicity. We also investigated the impact of the TSST-1 superantigen on the bone resorption activity of osteoclasts.
# Materials and methods
## Preparation of osteoclasts
Monocytes were purified from the blood of healthy donors (n = 3) purchased from Etablissement Français du Sang (Lyon Gerland, France), as previously described [bib_ref] Measles Virus Suppresses Cell-mediated Immunity by Interfering with the Survival and Functions..., Fugier-Vivier [/bib_ref]. Donors gave written consent to EFS for the use of blood sample for research purposes at the time of sampling (number of the agreement linking EFS and the research laboratory: 14-1820-69). Briefly, after collection, blood was loaded on a Lymphocyte Separation Medium density gradient (Eurobio, Courtaboeuf, France) to purify mononuclear cells. Cells were then centrifuged through a 50% Percoll density gradient to concentrate monocytes and purified from the light-density fraction by immunomagnetic depletion using magnetic beads (Dynabeads goat anti-mouse IgG, Invitrogen, Carlsbad, CA) and a cocktail of monoclonal antibodies (mAbs): anti-CD19, anti-CD3, anti-CD56 and anti-glycophorin A (Beckman-Coulter, Miami, FL), ensuring purification rates 95%. Monocytes were then plated in 96-well plates at a density of 10 5 cells/well and cultured in α-MEM medium (Gibco Life Technologies, Inc., Grand Island, NY) supplemented with 2 mM L-glutamine (Gibco Life Technologies), 1% penicillin/streptomycin (Gibco Life Technologies), and 10% foetal bovine serum (PAN Biotech, Aldenbach, Germany).
Monocytes were differentiated into mature osteoclasts or macrophages as described elsewhere [bib_ref] Measles Virus Suppresses Cell-mediated Immunity by Interfering with the Survival and Functions..., Fugier-Vivier [/bib_ref]. Briefly, osteoclasts were obtained by incubating monocytes with 50 ng/mL M-CSF (Monocyte Colony-stimulating Factor) (PeproTech, Rocky Hill, NJ) and 30 ng/mL of RANKL (Receptor Activator of Nuclear factor Kappa-B Ligand) (PeproTech, Rocky Hill, NJ) from day 1 to day 3, and in presence of 25 ng/mL of M-CSF and 100 ng/mL of RANKL from day 3 to day 6. Macrophages were obtained by incubation of monocytes with 50 ng/mL of M-CSF only from day 0 to day 6.
## Toxins production
Recombinant staphylococcal toxins (Hla, Hlb, HlgAB, HlgBC, LukED, LukGH, PVL, TSST-1 and SEA) were purified as described elsewhere [bib_ref] Cross-talk between Staphylococcus aureus leukocidins-intoxicated macrophages and lung epithelial cells triggers chemokine..., Perret [/bib_ref]. The endotoxin content of the recombinant protein solutions was controlled and confirmed to be less than 0.004 endotoxin units per μg of protein.
## Cytotoxicity assay
Toxin cytotoxicity to osteoclasts and macrophages was quantified by monitoring propidium iodide (PI) incorporation as previously described [bib_ref] AIM2/ASC triggers caspase-8-dependent apoptosis in Francisella-infected caspase-1-deficient macrophages, Pierini [/bib_ref]. Cells were incubated with PI (5 μg/ml) and variable concentrations of recombinant staphylococcal toxins. Propidium iodide fluorescence was measured over a 3-hour period on a microplate fluorimeter (Tecan, Lyon, France), using untreated cells as control.
## Bone resorption assay
Bone resorption was quantified as a means to assess the impact of the TSST-1 superantigen on osteoclast activity. On day 6, mature osteoclasts were detached from plastic wells by flushing after incubation with Accutase (Invitrogen Life Technologies, Gaithersburg, MD) (37°C-30 min), and then seeded at 2.10 4 cells/well on mineralized matrix Osteo Assay Surface 96-well plates (OsteoCorning1, Corning, MA, USA). TSST-1 was added to wells at 0.1, 1, 10, 100 and 1000 ng/ml. Twenty four hours later, osteoclasts were lysed by osmotic shock, and OsteoCorn-ing1 Assay plates were stained with PBS/5% silver nitrate (Sigma-Aldrich) to quantify resorption using a Leica 22 DMI6000 microscope (Nanterre, France) and Fiji software (US National Institutes of Health, Bethesda, Maryland, USA) as described elsewhere [bib_ref] Control of Bone Resorption by Semaphorin 4D Is Dependent on Ovarian Function, Dacquin [/bib_ref]. Results were expressed as the proportion of resorbed area in each condition relative to the resorbed area in untreated cells.
## Statistical analyses
Differences in means were analysed using Student's test (t-test) with a threshold of 0.05. Analyses were performed using R software, version 2.14.2 (The R foundation for statistical computing, Vienna, Austria). Results were expressed as means and 95% confidence intervals derived from three independent experiments realized in triplicate. Each experiment was realized with a different blood donor.
# Results
We first evaluated the cytotoxic effect of a wide range of staphylococcal toxins on human macrophages by measuring incorporation of PI 3 hours post treatment. Macrophages served as a reference profile of the different toxins' activities [fig_ref] Fig 1: Cytotoxicity of staphylococcal toxins on human macrophages and mature osteoclasts [/fig_ref]. As expected, we observed a high cytotoxicity of the membrane-damaging toxins (Hla, Hlb, HlgAB, HlgBC, LukGH and PVL) to human macrophages while superantigenic toxins did not cause significant cytotoxicity compared to untreated cells. These results, consistent with those obtained in literature [bib_ref] Cross-talk between Staphylococcus aureus leukocidins-intoxicated macrophages and lung epithelial cells triggers chemokine..., Perret [/bib_ref] , were used to validate our experimental protocol.
Because osteoclasts share a common myeloid origin with macrophages, we hypothesized that staphylococcal toxin cytotoxic activity profiles on osteoclasts could be similar with those observed with macrophages. To confirm this hypothesis, the experimental protocol used with macrophages was tested on mature human osteoclasts [fig_ref] Fig 1: Cytotoxicity of staphylococcal toxins on human macrophages and mature osteoclasts [/fig_ref]. Results indicated that poreforming toxins induced a significant cytotoxicity on mature human osteoclasts unlike superantigens. Moreover, cytotoxicity profiles appeared to be superimposable between macrophages and osteoclasts. Toxin-induced cytotoxicity to macrophages and osteoclasts could only be compared qualitatively (cytotoxicity profiles caused by the various toxins), and not quantitatively [fig_ref] Fig 1: Cytotoxicity of staphylococcal toxins on human macrophages and mature osteoclasts [/fig_ref]. Indeed, mature osteoclasts have dozens of nuclei so this increased DNA content per cell leads to higher values of PI incorporation-induced fluorescence as compared to macrophages, which prevented us to compare fluorescence values of macrophages and osteoclasts.
Using increasing concentrations of toxins, we showed that membrane-damaging toxins had a dose-dependent cytotoxic effect on mature human osteoclasts. For example PI incorporation in cells treated with PVL at 0.1, 1, 10 and 100 ng/ml were respectively 14%, 63%, 315% and 358% higher, compared to untreated [fig_ref] Fig 1: Cytotoxicity of staphylococcal toxins on human macrophages and mature osteoclasts [/fig_ref]. In contrast, superantigenic toxins caused significant cytotoxicity on osteoclasts only above 10 000 ng/mL.
Because it has been established that some staphylococcal toxins also have cellular activation effect [bib_ref] Cross-talk between Staphylococcus aureus leukocidins-intoxicated macrophages and lung epithelial cells triggers chemokine..., Perret [/bib_ref] , we tested the impact of the non-cytotoxic toxins on osteoclast activation. We investigated the capacity of TSST-1 to activate mature human osteoclasts by assessing bone resorption assays [fig_ref] Fig 2: Fig 2 [/fig_ref]. Results demonstrated that TSST-1 significantly enhanced osteolytic activity in a dose-dependent manner, resorbed area by cells treated with TSST-1 at 1, 10, 100, 1000 ng/ ml were respectively 13%, 17%, 24% and 26% higher, compared to untreated cells (p<0.001 for all). TSST-1 concentrations below 1 ng/mL induced no measurable effect.
# Discussion
Using an in vitro model, we evaluated the direct, specific, and independent effect of recombinant S. aureus toxins on mature human monocyte-derived osteoclasts. Using osteoclasts and macrophages differentiated from the same blood donors, we have shown that human macrophages and mature monocyte-derived osteoclasts exhibit similar susceptibility profiles with respect to staphylococcal toxins. PFTs caused significant dose-dependent cellular cytotoxicity on these two cell types whereas superantigenic toxins were not or poorly cytotoxic. Noteworthy, these results are in agreement with the fact that mature osteoclasts express the complement receptor C5a [bib_ref] Complement C3a and C5a modulate osteoclast formation and inflammatory response of osteoblasts..., Ignatius [/bib_ref] , which has recently been identified as the PVL and HlgBC receptor [bib_ref] The staphylococcal toxin Panton-Valentine Leukocidin targets human C5a receptors, Spaan [/bib_ref]. Moreover, several studies have demonstrated that human osteoclasts express CCR2, CCR5 and CXCR1 and 2 [bib_ref] Autocrine signaling is a key regulatory element during osteoclastogenesis, Kopesky [/bib_ref] , [bib_ref] Differential Expression of Chemokines, Chemokine Receptors and Proteinases by Foreign Body Giant..., Khan [/bib_ref] , [bib_ref] MIP-1alpha utilizes both CCR1 and CCR5 to induce osteoclast formation and increase..., Oba [/bib_ref] which have recently been identified as receptors respectively of LukED, HlgBC and both of these PFTs [bib_ref] The staphylococcal toxins γ-haemolysin AB and CB differentially target phagocytes by employing..., Spaan [/bib_ref]. This suggests that osteoclasts are targeted by PVL, HlgBC and LukED during BJIs. The direct cytotoxic activity of PVL highlighted in this study could play a role, which level remains to be determined, in the severity and outcome of acute BJIs due to PVL-producing S. aureus. Indeed, it is known that BJIs caused by S. aureus PVL positive strains are more severe and extensively destructive than those caused by S. aureus PVL negative ones [bib_ref] Panton-Valentine Leukocidin Enhances the Severity of Community-Associated Methicillin-Resistant Staphylococcus aureus Rabbit Osteomyelitis, Crémieux [/bib_ref]. The direct effect on osteoclasts could be added to the cytolytic indirect effect related to the release, at the infection site, of the cytoplasmic content of macrophages and neutrophils, which were the only target cells previously identified for PVL. Previous studies have demonstrated that PVL, in addition to targeting neutrophils can target macrophages and to trigger IL-1β secretion [bib_ref] Cross-talk between Staphylococcus aureus leukocidins-intoxicated macrophages and lung epithelial cells triggers chemokine..., Perret [/bib_ref] and amplify inflammation. This content has likely a direct cytotoxic effect on osteoblasts, osteoclasts and the bone tissue itself, leading to a local inflammation, tissue necrosis and thus to bone destruction.
As superantigens were not cytotoxic to osteoclasts, the second part of this study aimed to assess the ability of TSST-1, to stimulate mature monocyte-derived osteoclasts. It has been shown that staphylococcal superantigens ensure, in absence of antigen presentation, bridging between the TCR Vβ chain of T cells and MHC class II of antigen presenting cells and in particular by osteoclasts [bib_ref] MHC class II transactivator is an in vivo regulator of osteoclast differentiation..., Benasciutti [/bib_ref]. Although the presence of these two cell types appears to be required for the synergistic action of superantigens, several studies have reported the ability of superantigens to stimulate macrophage pro-inflammatory cytokine secretion in the absence of T cells [bib_ref] Superantigens augment antigen-specific Th1 responses by inducing IL-12 production in macrophages, Bright [/bib_ref]. Our data show that TSST-1 promotes bone resorption by human mature monocyte-derived osteoclasts at concentrations above 1 ng/mL, which are probably relevant in the TSST-1 stimulates bone resorption capacity of mature human osteoclasts. For bone resorption assay, mature osteoclasts were detached from plastic on day 6 and seeded at 2.10 4 cells/well on mineralized matrix Osteo Assay Surface 96-well plates. TSST-1 was added to the cell culture medium. After 24 hours of culture, the osteocorning matrices were stained with 5% silver nitrate to measure the resorbed area (white area). The percentages of matrix that were resorbed by untreated (A) or TSST-1-treated osteoclasts (B, C) were measured using the Fiji software. Bars represent 100 μm. Results of resorption area quantification (D) represent the mean of resorbed area with 95% confidence interval, of 3 independent experiments realised in triplicate on 3 different donors (*p<0.05, ** p<0.01, ***p<0.001). TSST-1: Toxic shock syndrome toxin. context of bone infection in vivo. Indeed, plasmatic TSST-1 concentrations of more than 5 ng/ mL have been observed in infected patients [bib_ref] Rapid assay for detection of toxic shock syndrome toxin 1 from human..., Miwa [/bib_ref] , and it is likely that toxin concentrations at the site of infection are greater than those in circulating blood. Importantly, S. aureus strains can harbor other superantigens than TSST-1. Whether these superantigens could trigger osteoclastogenesis similar to TSST-1 remains an open question. Indeed, the osteoclastic stimulation observed in our model might be specific to TSST-1, because this toxin has been shown to interact with several cell surface targets including ADAM17 and EGFR [bib_ref] A disintegrin and metalloproteinase 17 (ADAM17) and epidermal growth factor receptor (EGFR)..., Breshears [/bib_ref] or CD40 [bib_ref] Immunity to Staphylococcus aureus secreted proteins protects rabbits from serious illnesses, Spaulding [/bib_ref] , which are expressed by osteoclasts [bib_ref] ADAM gene expression and regulation during human osteoclast formation, Verrier [/bib_ref] [28] [bib_ref] Inhibition of epidermal growth factor receptor tyrosine kinase ameliorates collagen-induced arthritis, Swanson [/bib_ref]. Further studies are warranted to determine which osteoclastic receptors are involved in TSST-1-induced stimulation, and to determine whether this stimulation involves a canonical superantigen-MCH class II interaction which might be triggered by other superantigens.
Collectively, our results suggest that bone loss during staphylococcal BJIs might not only be driven by non-specific inflammation and local acidity, created by dead cells, but also by the specific targeting and activation of bone resorbing cells by bacterial toxins. The balance between osteoclasts killing by PFTs and the superantigen-mediated increase in osteoclasts' bone resorption activity may control the different clinical expression of BJIs associated with the toxinic profile of the different S. aureus strains.
[fig] Fig 1: Cytotoxicity of staphylococcal toxins on human macrophages and mature osteoclasts. Human monocytes were differentiated for 6 days into macrophages with macrophage colony-stimulating factor (M-CSF) (A) or into mature osteoclasts in the presence of M-CSF and receptor activator of NFκ-B ligand (RANK-L) (B and C). Staphylococcal toxins were then added to the cell culture medium containing propidium iodide (PI) and cell death was quantified by monitoring PI incorporation over a 3 hours period (A, B and C). Fluorescence in each well was normalised to the fluorescence obtained with untreated cells. Results represent the mean cytotoxicity with 95% confidence interval, of 3 independent experiments performed in triplicate on 3 different donors (*p<0.05, ** p<0.01, ***p<0.001). Hla: α haemolysin, Hlb: β haemolysin, Hlg AB and Hlg BC: γ haemolysins AB and BC, Luk ED and Luk GH: leukocidins ED and GH, PVL: Panton Valentine Leukocidin, TSST-1: toxic shock syndrome toxin, SEA: Staphylococcal enterotoxin A.doi:10.1371/journal.pone.0150693.g001 [/fig]
[fig] Fig 2: Fig 2. TSST-1 stimulates bone resorption capacity of mature human osteoclasts. For bone resorption assay, mature osteoclasts were detached from plastic on day 6 and seeded at 2.10 4 cells/well on mineralized matrix Osteo Assay Surface 96-well plates. TSST-1 was added to the cell culture medium. After 24 hours of culture, the osteocorning matrices were stained with 5% silver nitrate to measure the resorbed area (white area). The percentages of matrix that were resorbed by untreated (A) or TSST-1-treated osteoclasts (B, C) were measured using the Fiji software. Bars represent 100 μm. Results of resorption area quantification (D) represent the mean of resorbed area with 95% confidence interval, of 3 independent experiments realised in triplicate on 3 different donors (*p<0.05, ** p<0.01, ***p<0.001). TSST-1: Toxic shock syndrome toxin. [/fig]
|
Elevated CO2 Modifies N Acquisition of Medicago truncatula by Enhancing N Fixation and Reducing Nitrate Uptake from Soil
The effects of elevated CO 2 (750 ppm vs. 390 ppm) were evaluated on nitrogen (N) acquisition and assimilation by three Medicago truncatula genotypes, including two N-fixing-deficient mutants (dnf1-1 and dnf1-2) and their wild-type (Jemalong). The proportion of N acquisition from atmosphere and soil were quantified by 15 N stable isotope, and N transportation and assimilation-related genes and enzymes were determined by qPCR and biochemical analysis. Elevated CO 2 decreased nitrate uptake from soil in all three plant genotypes by down-regulating nitrate reductase (NR), nitrate transporter NRT1.1 and NR activity. Jemalong plant, however, produced more nodules, up-regulated N-fixation-related genes and enhanced percentage of N derived from fixation (%Ndf) to increase foliar N concentration and N content in whole plant (Ntotal Yield) to satisfy the requirement of larger biomass under elevated CO 2 . In contrast, both dnf1 mutants deficient in N fixation consequently decreased activity of glutamine synthetase/glutamate synthase (GS/GOGAT) and N concentration under elevated CO 2 . Our results suggest that elevated CO 2 is likely to modify N acquisition of M. truncatula by simultaneously increasing N fixation and reducing nitrate uptake from soil. We propose that elevated CO 2 causes legumes to rely more on N fixation than on N uptake from soil to satisfy N requirements. Citation: Guo H, Sun Y, Li Y, Liu X, Ren Q, et al. (2013) Elevated CO 2 Modifies N Acquisition of Medicago truncatula by Enhancing N Fixation and Reducing Nitrate Uptake from Soil. PLoS ONE 8(12): e81373.
# Introduction
Global atmospheric CO 2 concentrations have been increasing at an accelerating rate. The concentration, which was 280 ppm before industrialization and was 394 ppm in December 2012 (Mauna Loa Observatory: NOAA-ESRL), is expected to reach at least 550 ppm by the year 2050. The effects of elevated CO 2 on C3 plants are generally characterized by increased photosynthesis, growth and yield in plant tissues [bib_ref] What have we learned from 15 years of free-air CO 2 enrichment..., Ainsworth [/bib_ref]. Under elevated CO 2 , the ''extra C'' is assimilated and transported from leaves and shoots to roots, and the C:N ratio is consequently increased [bib_ref] The response of photosynthesis and stomatal conductance to rising [CO 2 ]:..., Ainsworth [/bib_ref]. Thus, plant responses to elevated CO 2 are likely to be limited by the availability of N.
Besides increases in biomass and productivity, a common characteristic of non-leguminous C3 plants in an elevated CO 2 environment is a 10-15% decrease in N concentration (g of N per g of plant tissue ) [bib_ref] Effects of elevated CO 2 on the protein concentration of food crops:..., Taub [/bib_ref]. Three major hypotheses have been proposed to explain this phenomenon [bib_ref] Why are nitrogen concentrations in plant tissues lower under elevated CO 2..., Taub [/bib_ref]. According to the reduced uptake hypothesis, N content is reduced because decreased stomatal conductance and transpiration under elevated CO 2 reduces N uptake by roots [bib_ref] Can decreased transpiration limit plant nitrogen acquisition in elevated CO 2 ?, Mcdonald [/bib_ref]. The N loss hypothesis presumes that N losses increase under elevated CO 2 because of increasing NH 3 volatilization or increasing root exudation of organic N [bib_ref] A new explanation of the N concentration decrease in tissues of rice..., Pang [/bib_ref]. The dilution hypothesis, which has received the most attention, considers that N content is diluted under elevated CO 2 by accumulation of more total non-structural carbohydrates (TNC), which results in a greater biomass for a given quantity of N [bib_ref] A meta-analysis of elevated [CO 2 ] effects on soybean (Glycine max)..., Ainsworth [/bib_ref]. Depending on the species or genotype, these hypotheses may partially or largely explain the substantial reduction in the N content in non-leguminous plants under elevated CO 2 [bib_ref] Elevated CO 2 changes the interactions between nematode and tomato genotypes differing..., Sun [/bib_ref]. Furthermore, elevated CO 2 has little effect on the N content in legumes, which might be attributed to their unique ability to utilize atmospheric N 2 [bib_ref] Effects of elevated CO 2 on the protein concentration of food crops:..., Taub [/bib_ref] , but it still lacks the experimental evidence to address the physiological mechanism underlying N metabolism of legume plants under elevated CO Leguminous plants acquire N by three major pathways. First, legumes uptake ammonia (NH 4 + ) from soil and incorporate it into organic compounds. Second, legumes uptake nitrate from soil and reduce it to NH 4 + . Third, legumes in symbioses with N-fixing bacteria can obtain N from the atmosphere by N fixation, i.e., by converting N 2 to NH 4 + [bib_ref] A chemical index of soil nitrogen availability, Keeney [/bib_ref]. Among these three pathways, N fixation is most costly in terms of energy and resources. [bib_ref] How much nitrogen do legumes fix?, Larue [/bib_ref] , for example, found that four legumes including Glycine max, Vigna unguiculata, Phaseolus vulgaris, and Pisum sativum, consume an average of 6.7 g of carbohydrate to obtain 1 g of N by symbiosis [bib_ref] How much nitrogen do legumes fix?, Larue [/bib_ref]. Acquiring N via uptake of nitrate or ammonia from soil required less carbohydrate C than acquiring N by symbiosis [bib_ref] Energy requirement for symbiotic nitrogen fixation, Silsbury [/bib_ref] [bib_ref] Symbiotic N 2 fixation activity in relation to C economy of Pisum..., Voisin [/bib_ref]. Nitrogenase activity, the most important enzyme involved in N fixation, and nodule formation are often suppressed when nitrate or ammonia availability is sufficient to meet the requirements of plant growth [bib_ref] Differential effects of combined N sources on early steps of the Nod..., Barbulova [/bib_ref]. Thus, it seems that legume plants preferentially obtain N via uptake from the soil rather than fixation from the atmosphere [bib_ref] Symbiotic N 2 fixation activity in relation to C economy of Pisum..., Voisin [/bib_ref].
To sustain and maximize growth and biomass under elevated CO 2 , legumes require additional N [bib_ref] A meta-analysis of elevated [CO 2 ] effects on soybean (Glycine max)..., Ainsworth [/bib_ref]. Owing to the high C consumption required for N fixation, elevated CO 2 helps legumes fix N from atmosphere [bib_ref] Are the carbon costs of seed production related to the quantitative and..., Munier-Jolain [/bib_ref]. After reviewing 127 studies, [bib_ref] Nitrogen dynamics in grain crop and legume pasture systems under elevated atmospheric..., Lam [/bib_ref] concluded that the amount of N fixed from the atmosphere by legumes increased 38% under elevated CO 2 , which was accompanied by increases in whole plant nodule number (+33%), nodule mass (+39%), and nitrogenase activity (+37%) [bib_ref] Nitrogen dynamics in grain crop and legume pasture systems under elevated atmospheric..., Lam [/bib_ref]. Furthermore, enhancement of N fixation in legumes is essential for overcoming the N limitation under elevated CO 2 [bib_ref] Increased C availability at elevated carbon dioxide concentration improves N assimilation in..., Rogers [/bib_ref]. However, the relative contributions of N fixation and uptake from soil to the N content of legumes under elevated CO 2 are largely unknown. It is likely that legumes adjust their means of utilizing N resources to adapt to environmental changes [bib_ref] Legume Nitrogen Fixation and Soil Abiotic Stress: From Physiology to Genomics and..., Valentine [/bib_ref] , and a CO 2 -enriched environment may affect the crosstalk between the different N acquisition pathways in legumes.
The current study examined N acquisition via N fixation and N uptake in N-fixing-deficient mutants (dnf1) and wild-type (Jemalong) of M. truncatula. We tested the hypothesis that M. truncatula plants regulate the relative contribution of N fixation and N uptake from soil to maximum the N assimilation rate to satisfy the higher N requirement under elevated CO 2 . The specific objectives were to determine: (1) how elevated CO 2 affects N fixation from the atmosphere and N uptake from soil; and (2) whether elevated CO 2 affects N assimilation of the M. truncatula genotypes. To help meet these objectives, we measured the expression of key genes and the activity of key enzymes involved in N acquisition and assimilation (glutamine synthase/glutamate synthase, GS/GOGAT cycle) [bib_ref] Continuous CO 2 enrichment leads to increased nodule biomass, carbon availability to..., Cabrerizo [/bib_ref]. Meanwhile, 15 N stable isotope technique was used to determine N acquisition and partitioning, and estimate the proportion of N fixed from atmosphere/N uptake from soil [bib_ref] Foliar and soil d 15 N values reveal increased nitrogen partitioning among..., Gubsch [/bib_ref]. [fig_ref] Figure 1: Above-and below-ground biomass of M [/fig_ref]. The atmospheric CO 2 concentration treatments were: (1) current atmospheric CO 2 levels (390 ml/L), and (2) elevated CO 2 levels (750 ml/L, the predicted level in about 100 years). Four blocks were used, and each block contained one OTC with ambient CO 2 and one with elevated CO 2 . From seedling emergence to the harvesting of M. truncatula plants (27 August to 15 October 2011, a total of 50 days), CO 2 concentrations were monitored and adjusted with an infrared CO 2 analyzer (Ventostat 8102, Telaire Company, Goleta, CA, USA) once every minute to maintain relatively stable CO 2 concentrations. The measured CO 2 concentrations throughout the experiment (mean 6 SD per day) were 391623 ppm in the ambient CO 2 chambers and 743632 ppm in the elevated CO 2 chambers. The auto-control system for maintaining the CO 2 concentrations, as well as specifications for the OTCs, is detailed in [bib_ref] An improved top-open chamber for research on the effects of elevated CO..., Chen [/bib_ref] [bib_ref] An improved top-open chamber for research on the effects of elevated CO..., Chen [/bib_ref]. The tops of the OTCs were covered with nylon net to exclude insects. Air temperatures were measured three times per day throughout the experiment and did not differ significantly between the two treatments (24.963.4uC in OTCs with ambient CO 2 vs. 26.263.9uC in OTCs with elevated CO 2 ).
# Materials and methods
## Atmospheric co 2 concentration treatments
## M. truncatula mutants and rhizobium inoculation
Three M. truncatula genotypes were studied: the N-fixationdeficient mutants dnf1-1 and dnf1-2 as well as their wild-type Jemalong. These three genotypes were obtained from the laboratory of Sharon Long, Department of Biology, Stanford University. The nodules of these dnf1 mutants are small and white and are blocked at an intermediate stage of development [bib_ref] A nodule specific protein secretory pathway required for nitrogen-fixing symbiosis, Wang [/bib_ref]. The dnf1-1 mutant allele has a large deletion of at least 20 kb around TC121074 locus, and the dnf1-2 mutant allele has an independent disruption of the TC121074 locus [bib_ref] Foliar and soil d 15 N values reveal increased nitrogen partitioning among..., Gubsch [/bib_ref]. Although both mutants can be infected in the inner cortex, both lack acetylene reduction activity and Nodulin31 expression and have only a small level of nifH expression in the symbiotic nodule [bib_ref] Nitrogen fixation mutants of Medicago truncatula fail to support plant and bacterial..., Starker [/bib_ref]. After seeds were chemically scarified and surface sterilized by immersion in concentrated H 2 SO 4 for 5 min, they were rinsed with sterilized water several times. The seeds were placed in Petri dishes filled with 0.75% agar, kept in the dark at 4uC for 2 days, and then moved to 25uC for 2 days to germinate. The germinated seeds were sown on sterilized soil and inoculated 2 days later with the bacterium Sinorhizobium meliloti Rm1021 [bib_ref] Nitrogen fixation mutants of Medicago truncatula fail to support plant and bacterial..., Starker [/bib_ref] , which was kindly provided by Professor Xinhua Sui (Department of Microbiology, College of Biological Sciences, Chinese Agricultural University). S. meliloti was cultured on YM (H 2 O 1000 ml, yeast 3 g, mannitol 10 g, KH 2 PO 4 0.25 g, K 2 HPO 4 0.25 g, MgSO 4 ?7H 2 O 0.1 g, NaCl 0.1 g, pH 7.0-7.2) for 3 days at 28uC to obtain an approximate cell density of 10 8 ml 21 . At sowing, each seedling was inoculated with 0.5 ml of this suspension. After they had grown in sterilized soil for 2 weeks, the M. truncatula seedlings were individually transplanted into plastic pots (35 cm diameter and 28 cm height) containing sterilized loamy field soil (organic carbon 75 g/kg; N 500 mg/kg; P 200 mg/kg; K 300 mg/kg) and placed in OTCs on 27 August 2011. Each OTC contained 30 plants (10 each per genotype) with 240 plants in total.
Plants were maintained in the OTCs for 50 days. Pot placement was re-randomized within each OTC once every week to avoid any effects from the position of pots in each OTC. No chemical fertilizers and insecticides were used. Water was added to each pot once every 2 days.
## Plant sampling and preparation
All the plants of M. truncatula were randomly harvested on 13-15 October 2011. Root of each plant were carefully removed from soil and washed. A stereomicroscope was used to count the nodules on the entire root system of 6 plants from each M. truncatula genotype per OTC ( = 24 plants from each genotype at each CO 2 level and 144 in total). After nodules were counted, the shoots and roots of each plant were collected, oven-dried (65uC) for 72 h, and weighed. The leaves and root tissues were then ground to a fine powder (approx. 0.85 mm size) and analyzed for total non-structural carbohydrates (TNCs), N concentration and 15 N isotopic analysis. Another three plants from each M. truncatula genotype per OTC (9 plants per OTC and 72 plants in total) were randomly selected for enzyme analysis and real-time PCR. 50 mg of mature leaves and 100 mg of lateral roots from each plant were stored in freezing tubes at 275uC until used for real-time PCR. 0.5 g of mature leaves and 1.0 g of lateral roots from the same plants were frozen for enzyme analysis as described in the following paragraph.
## Tncs, n concentration and d 15 n analysis
TNCs, mainly starch and sugars, in leaves and roots were quantified by acid hydrolysis following the method of [bib_ref] Effects of seasonal water availability on phenology and the annual shoot carbohydrate..., Tissue [/bib_ref]. N concentrations in leaves and roots were measured by Kjeltec N analysis (Foss automated Kjeltec TM instruments, Model 2100) [bib_ref] Effects of seasonal water availability on phenology and the annual shoot carbohydrate..., Tissue [/bib_ref]. d 15 N were determined from approximately 3 mg plant sample with an isotope-ratio mass spectrometer (IRMS; Delta plus XP and Delta C prototype Finnigan MAT, respectively, Finnigan MAT, Bremen, Germany; 0.1% precision). The d 15 N values represent nitrogen isotopic composition of the sample relative to that of atmospheric dinitrogen in %:
[formula] d 15 N sample~Rsample {R standard À Á =R standard |100 [/formula]
Where R standard is the 15 N/ 14 N ratio of atmospheric N 2 and R sample is the 15 N/ 14 N ratio of the sample plant. The repeated measurement precision was 0.2%.
Percentage of N derived from atmosphere and N uptake from soil Percentage of N fixed from atmosphere is a yield-independent parameter and was calculated according to [bib_ref] Use of 13 C and 15 N isotopes to investigate O 3..., Pausch [/bib_ref] [25]: . Total non-structural carbohydrate (TNC) content in leaves and roots of M. truncatula plants as affected by CO 2 level and plant genotype: dnf1-1 and dnf1-2 are deficient in N fixation, and Jemalong is their wild type. Each value represents the average (6SE) of four replicates. Different lowercase letters indicate significant differences between ambient CO 2 and elevated CO 2 within the same genotype. Different uppercase letters indicate significant differences among genotypes with the same CO 2 treatment as determined by Tukey's multiple range test at P,0.05. doi:10.1371/journal.pone.0081373.g003
[formula] %Ndf~1{d 15 N M:truncatula =d 15 N reference À Á |100 [/formula]
Where % Ndf is the percentage of N derived from atmosphere. d 15 N M. truncatula is the d 15 N of wild-type Jemalong, d 15 N reference is the average value of d 15 N of dnf1-1 and dnf1-2 in the same OTC. dnf1-1 and dnf1-2 have similar N uptake and rooting patterns as Jemalong but are deficient in nitrogen fixation, and therefore served as the reference plant for analyzing the N fixation of wildtype M. truncatula.
In order to evaluate changes in N source (i.e. as derived from Nfixation or soil) for M. truncatula plants, Nf Yield (N derived from N-fixation per plant) and Ns Yield (N derived from soil per plant) of Jemalong estimates were calculated as follows: Where Ntotal Yield is the N content in per whole plant, Biomass above-ground is the biomass of above-ground tissue in M. truncatula plants, Biomass under-ground is the biomass of under-ground tissue, N leaf is the N concentration of leaves and N root is the N concentration of root.
[formula] NtatalYield(mg=plant)~Biomass above-ground |N leaf zBiomass under-ground |N root NfYield(mg=plant)~NtotalYield|%Ndf NsYield(mg=plant)~NtotalYield{NfYield [/formula]
## Activities of enzymes involved in n uptake and assimilation
The activities of nitrate reductase (NR), glutamine synthetase (GS), and glutamate synthase (GOGAT) in leaves and roots were determined using frozen tissue (approximately 0.5 g leaf tissue and approximately 1.0 g root tissue per plant). Once the tissue was ground to a fine powder, leaves or roots from three plants of the same genotype within each OTC were combined to form one sample from each OTC. The unit of replication for statistical analyses was the OTC (n = 4). An extract was obtained by grinding each leaf sample or root sample in 50 mM Tris HCl buffer (pH 7.8, 3 ml/g of leaf tissue) containing 1 mM MgCl 2 , 1 mM EDTA, 1 mM b-mercaptoethanol, and 1% (w/v) polyvinylpolypyrro-lidone. This extract was immediately frozen for later use. For assays, the thawed extract was centrifuged at 13,000 g for 10 min, and the enzyme activities were measured in the supernatant as described by [bib_ref] Enhanced carbon dioxide leads to a modified diurnal rhythm of nitrate reductase..., Geiger [/bib_ref] for NR [bib_ref] Enhanced carbon dioxide leads to a modified diurnal rhythm of nitrate reductase..., Geiger [/bib_ref] , by [bib_ref] Respective roles of the glutamine synthetase/glutamate synthase cycle and glutamate dehydrogenase in..., Glévarec [/bib_ref] for GS [bib_ref] Respective roles of the glutamine synthetase/glutamate synthase cycle and glutamate dehydrogenase in..., Glévarec [/bib_ref] , and by [bib_ref] Regulation by light and metabolites of ferredoxin-dependent glutamate synthase in maize, Suzuki [/bib_ref] for GOGAT [bib_ref] Regulation by light and metabolites of ferredoxin-dependent glutamate synthase in maize, Suzuki [/bib_ref]. Protein concentrations of leaves and roots were measured using bovine serum albumin as a standard. One unit (U) of GS/GOGAT activities are defined as the amount of the GS or . Activities of the enzymes involved in N reduction (NR) and in N assimilation (GS and GOGAT) in the leaves and roots of M. truncatula plants as affected by CO 2 level and plant genotype: dnf1-1 and dnf1-2 are deficient in N fixation, and Jemalong is their wild type. Each value represents the average (6SE) of four replicates. Different lowercase letters indicate significant differences between ambient CO 2 and elevated CO 2 within the same genotype. Different uppercase letters indicate significant differences among genotypes within the same CO 2 treatment as determined by Tukey's multiple range test at P,0.05. doi:10.1371/journal.pone.0081373.g005
GOGAT that catalyzes 1 nmol of glutamine or glutamate per minute in the homogenate.
Expression of Genes Associated with N Fixation, Uptake, and Assimilation as Determined by Quantitative RT-PCR Each treatment combination was replicated four times for biological repeats, and each biological repeat contained three technical repeats. The RNAeasy Mini Kit (Qiagen) was used to isolate total RNAs from M. truncatula leaves and roots, and 1 mg of RNA was used to generate the cDNAs. The mRNAs of the following nine target genes were quantified by real-time quantitative PCR: early nodule-specific protein 40 (ENOD) (maintenance of nodule symbiosis) [bib_ref] EFD is an ERF transcription factor involved in the control of nodule..., Vernié [/bib_ref] , nodulation gene (nodF) (nodF genes are required for nodulation) [bib_ref] EFD is an ERF transcription factor involved in the control of nodule..., Vernié [/bib_ref] , nitrogen-fixing gene (nifH) (nifH genes control synthesis of nitrogenase) [bib_ref] Genes required for rapid expression of nitrogenase activity in Azotobacter vinelandii, Curatti [/bib_ref] , nitrate transporter NRT1.1 (NT) [bib_ref] Functional assessment of the Medicago truncatula NIP/LATD protein demonstrates that it is..., Bagchi [/bib_ref] , nitrate reductase (NR), ammonium transporter protein (AMT) [bib_ref] Stimulation of nodulation in Medicago truncatula by low concentrations of ammonium: quantitative..., Fei [/bib_ref] , glutamine synthetase 2 (GS) [bib_ref] Concerted modulation of alanine and glutamate metabolism in young Medicago truncatula seedlings..., Limami [/bib_ref] , and glutamate synthetase (GOGAT) [bib_ref] Concerted modulation of alanine and glutamate metabolism in young Medicago truncatula seedlings..., Limami [/bib_ref] [fig_ref] Figure 1: Above-and below-ground biomass of M [/fig_ref] in File S3). Specific primers for each gene were designed from the M. truncatula EST sequences using PRIMER5 software [fig_ref] Table 1: CO 2 effects on N characteristics [/fig_ref] in File S1). The PCR reactions were performed in 20 mL reaction volumes that included 10 mL of 26SYBRs Premix EX TaqTM (Qiagen) master mix, 5 mM of each gene-specific primer, and 1 mL of cDNA template. Reactions were carried out on the Mx 3500P detection system (Stratagene) as follows: 2 min at 94uC; followed by 40 cycles of 20 s at 95uC, 30 s at 56uC, and 20 s at 68uC; and finally one cycle of 30 s at 95uC, 30 s at 56uC, and 30 s at 95uC. This PCR protocol produced the melting curves, which can be used to judge the specificity of PCR products. A standard curve was derived from the serial dilutions to quantify the copy numbers of target mRNAs. b-actin and pnp were used as internal qPCR standards for the analysis of plant and bacterial gene expression, respectively [bib_ref] EFD is an ERF transcription factor involved in the control of nodule..., Vernié [/bib_ref]. The relative level of each target gene was standardized by comparing the copy numbers of target mRNA with copy numbers of b-actin or pnp (the house-keeping gene), which remain constant under different treatment conditions. The levels of b-actin or pnp mRNAs in the control were examined in every PCR plate to eliminate systematic error. The fold-changes of target genes were calculated using the 2 2DDCt normalization method.
# Statistical analysis
Statistical analyses were performed with SPSS 13.0 software (SPSS Inc., Chicago, IL). Two-way analyses of variance (ANOVA) were used to analyze the effect of CO 2 and plant genotype on M. truncatula growth traits, TNC, N concentration, Ntotal Yield and enzyme activities. If an ANOVA was significant, Tukey's multiple range test was used for mean separation (P,0.05). Significance of the effect of CO 2 on %Ndf, Nf Yield, Ns Yield of Jemlaong and genes regulating N metabolism were determined by independent ttests.
# Results
## Plant biomass and nodule number
CO 2 level, genotype and their interaction significantly affected the above-ground biomass, below-ground biomass and total biomass in File S2). Total biomass did not significantly differ among the genotypes under ambient CO 2 but was greater for the wild-type Jemalong than for the mutants under elevated CO 2 [fig_ref] Figure 1: Above-and below-ground biomass of M [/fig_ref]. In response to elevated CO 2 , above-ground biomass increased 37.1% and total biomass increased 41.9% for Jemalong plants but the biomass of dnf1 mutant plants was not significantly affected by the CO 2 treatments [fig_ref] Figure 1: Above-and below-ground biomass of M [/fig_ref]. CO 2 level and genotype significantly affected the nodule numbers . Regardless of CO 2 level, nodule number was greater for Jemalong than for the dnf1 mutants [fig_ref] Figure 2: Nodule number per root of M [/fig_ref]. Elevated CO 2 increased nodule numbers of Jemalong but not of the mutants.
## Tnc and n characteristic in plant
CO 2 level significantly affected the foliar TNC, and all factors significantly affected the root TNC . Elevated CO 2 increased the TNC content in leaves and roots of Jemalong but only in leaves of dnf1-1 and dnf1-2 . Foliar TNC content did not differ among the three M. truncatula genotypes . Regardless of CO 2 level, Jemalong had the highest root TNC content .
Genotype was significant for the foliar N concentration, and all factors significantly affected the root N concentration and Ntotal Yield . Elevated CO 2 increased the foliar N concentration and Ntotal Yield in Jemalong but reduced in both dnf1 mutants [fig_ref] Figure 4: N concentrations in leaves and roots as well as Ntotal Yield [/fig_ref]. Elevated CO 2 reduced N concentration in the roots of both dnf1 mutants but not in Jemalong [fig_ref] Figure 4: N concentrations in leaves and roots as well as Ntotal Yield [/fig_ref]. Under ambient CO 2 , foliar N and root N concentrations were not significantly different among three genotypes. Under elevated CO 2 , however, N concentration in leaves and roots were higher in Jemalong than in the mutants [fig_ref] Figure 4: N concentrations in leaves and roots as well as Ntotal Yield [/fig_ref]. Regardless of CO 2 level, Jemalong had higher Ntotal Yield than dnf1-1 and dnf1-2 mutants [fig_ref] Figure 4: N concentrations in leaves and roots as well as Ntotal Yield [/fig_ref]. Furthermore, elevated CO 2 increased %Ndf and Nf Yield but decreased Ns Yield of Jemalong [fig_ref] Table 1: CO 2 effects on N characteristics [/fig_ref].
Activities of the Enzymes NR, GS and GOGAT CO 2 level was significant for the activities of foliar NR and root NR. . Elevated CO 2 reduced NR activity in the leaves and roots of all three genotypes . Under ambient CO 2 , NR activity in both leaves and roots were higher in both dnf1 mutants than in Jemalong . Under elevated CO 2 , however, NR activity did not differ among the three genotypes .
Genotype and the interaction between CO 2 and genotype significantly affected the foliar GS and root GS. All factors significantly affected the foliar GOGAT and root GOGAT . Elevated CO 2 decreased GS and GOGAT activities in the two dnf1 mutants but not in Jemalong . GS and GOGAT activities in Jemalong leaves and GOGAT in roots were higher than dnf1-1 and dnf1-2 mutant in both CO 2 levels. GS activity in roots did not differ among the three genotypes under ambient CO 2 but were higher in Jemalong than in the mutants under elevated CO 2 .
Expression of Genes Associated with N Fixation, Uptake, and Assimilation as Determined by Quantitative RT-PCR Elevated CO 2 up-regulated the expression of N fixation related genes including ENOD, nodF, and nifH, but down-regulated the expression of nitrate uptake and transport related genes including NR and NT in Jemalong plants [fig_ref] Figure 6: Expression of genes involved in N fixation [/fig_ref]. For dnf1-1 and dnf1-2, elevated CO 2 down-regulated the gene expression of NR and NT, and ammonia transport related genes AMT, and N assimilation related gene including GS and GOGAT [fig_ref] Figure 6: Expression of genes involved in N fixation [/fig_ref].
# Discussion
The notion that elevated CO 2 can increase plant biomass and TNC content in plant tissues is widely accepted [bib_ref] A meta-analysis of elevated [CO 2 ] effects on soybean (Glycine max)..., Ainsworth [/bib_ref]. Although this concept was further supported by the current report, our results also indicate that the key element in the increase in biomass of M. truncatula under elevated CO 2 is the availability of N [fig_ref] Figure 2: Nodule number per root of M [/fig_ref]. Using dnf1-1 and dnf 2 mutants, we demonstrated that M. truncatula is able to adjust different N partitioning pathways to ensure a sufficient N supply under ambient CO 2 . Elevated CO 2 , however, reduced N uptake from soil by suppressing N uptake related gene and increased the reliance on fixation of atmospheric N 2 .
N availability is one of the key factors limiting plant growth and production. Elevated CO 2 stimulating plant growth would increase the N demand of plants [bib_ref] Yield response of Lolium perenne swards to free air CO 2 enrichment..., Daepp [/bib_ref]. The extent of the CO 2 response at the plant level could consequently be limited by N availability [bib_ref] CO 2 enhancement of forest productivity constrained by limited nitrogen availability, Norby [/bib_ref] [bib_ref] The interaction between elevated carbon dioxide and nitrogen nutrition: the physiological and..., Stitt [/bib_ref]. In current study, since elevated CO 2 increased foliar N concentration and Ntotal yield [fig_ref] Table 1: CO 2 effects on N characteristics [/fig_ref] ; [fig_ref] Figure 4: N concentrations in leaves and roots as well as Ntotal Yield [/fig_ref] , Jemalong plants were able to produce more biomass under elevated CO 2 [fig_ref] Figure 1: Above-and below-ground biomass of M [/fig_ref]. Moreover, although TNC content in leaves and roots of dnf1-1 and dnf1-2 mutants were increased , N concentration in leaves and roots as well as the Ntotal yield of dnf1 mutant plants were decreased by elevated CO 2 [fig_ref] Table 1: CO 2 effects on N characteristics [/fig_ref] ; [fig_ref] Figure 4: N concentrations in leaves and roots as well as Ntotal Yield [/fig_ref]. It may suggest that dnf1 mutants were unable to provide sufficient N to support the enhancement of biomass under elevated CO 2 [fig_ref] Figure 1: Above-and below-ground biomass of M [/fig_ref]. Thus, our results demonstrated that the symbiotic N 2 fixation provided legumes an incomparable advantage in producing larger amounts of biomass under elevated CO 2 [bib_ref] Acclimation of photosynthesis to elevated CO 2 under conditions of low N..., Rogers [/bib_ref].
The results of the current study show that legumes are very flexible in their utilization of N from soil and atmosphere under ambient CO 2 . Although the dnf1 mutants are unable to fix atmospheric N 2 , GS/GOGAT activities involved in N assimilation and N concentration in leaves and roots did not differ from those of the wild-type Jemalong under ambient CO 2 . As indicated by increased gene expression (NR, NT) and enzyme activities (NR) of essential components of the alternate N acquisition pathways [fig_ref] Table 1: CO 2 effects on N characteristics [/fig_ref] ; [fig_ref] Figure 6: Expression of genes involved in N fixation [/fig_ref] , the dnf1 mutants compensated for the loss of N fixation by enhancing their uptake of N from soil under ambient CO 2 . However, the GS/GOGAT activities and N concentration in dnf1 plants were lower than in Jemalong plants under elevated CO 2 [fig_ref] Figure 4: N concentrations in leaves and roots as well as Ntotal Yield [/fig_ref] , which indicated that elevated CO 2 limited the N availability for both dnf1 mutants. Furthermore, [bib_ref] Direct evidence that symbiotic N 2 fixation in fertile grassland is an..., Lüscher [/bib_ref] found even in the high soil N treatment, ineffectively nodulating lucerne were unable to increase the N concentration and biomass under elevated CO 2 [bib_ref] Direct evidence that symbiotic N 2 fixation in fertile grassland is an..., Lüscher [/bib_ref]. Our results confirmed that soil N is insufficient to meet the increasing N demand of M. truncatula which can fully transform increased C assimilation into biomass [bib_ref] Does nitrogen nutrition restrict the CO 2 response of fertile grassland lacking..., Zanetti [/bib_ref].
The soil N availability appears to be suppressed by elevated CO 2 for all three M. truncatula genotypes, as reflected in decreased Ns Yield in Jemalong and Ntotal Yield in both dnf1 mutants under elevated CO 2 . Furthermore, the enzyme activity of NR and the expression of NT and NR genes of all three genotypes were also down-regulated by elevated CO 2 . This indicates that elevated CO 2 suppresses N uptake of M. truncatula from soil. This is consistent with the the finding that N uptake from soil by Trifolium repens were decreased under elevated CO 2 grown in a grassland ecosystem [bib_ref] Stimulation of Symbiotic N 2 Fixation in Trifolium repens L. under Elevated..., Zanetti [/bib_ref]. In addition, our results showed that elevated CO 2 down-regulated NR and NT but was not significant for ammonia transporter AMT [fig_ref] Figure 6: Expression of genes involved in N fixation [/fig_ref]. It seems that the decreases of N uptake from soil were mainly associated with the decreases of nitrate uptake rather than ammonia uptake.
The decreased nitrate uptake under elevated CO 2 could be explained by two factors: lower soil N availability and plant NO 3 reduction. Elevated CO 2 reduced the soil N availability by increasing N immobilization and denitrification in soil. For example, elevated CO 2 increased microbial community composition in rhizosphere soil of white clover, and subsequently increased N immobilization into the expanded microbial biomass [bib_ref] Changes in microbial activity and composition in a pasture ecosystem exposed to..., Montealegre [/bib_ref]. Additionally, elevated CO 2 increased the emission of N 2 O from soil [bib_ref] Nitrous oxide emissions from grass swards during eight years of elevated atmospheric..., Baggs [/bib_ref] , and this increase of N loss caused decreases of nitrate availability in soil. On the other hand, lower plant photorespiration induced by elevated CO 2 could decrease the nicotinamide adenine dinucleotide (NADH) [bib_ref] Nitrate assimilation in plant shoots depends on photorespiration, Rachmilevitch [/bib_ref] , which provides the energy required to convert NO to NO 2 2 in the cytoplasm of leaf mesophyll cells [bib_ref] Carbon dioxide enrichment inhibits nitrate assimilation in wheat and Arabidopsis, Bloom [/bib_ref]. Moreover, elevated CO 2 increased HCO [bib_ref] The response of photosynthesis and stomatal conductance to rising [CO 2 ]:..., Ainsworth [/bib_ref] 2 , and in turn inhibited NO 2 2 transportation from cytosol into the chloroplast [bib_ref] Nitrogen assimilation and growth of wheat under elevated carbon dioxide, Bloom [/bib_ref] , which led to a decrease in plant nitrate reduction.
Insufficient soil N uptake was considered to be one of the reasons for the increased contribution of N 2 fixation under elevated CO 2 [bib_ref] Does nitrogen nutrition restrict the CO 2 response of fertile grassland lacking..., Zanetti [/bib_ref]. In agreement with higher foliar N concentration and Ntotal Yield in the Jemalong plants, there was a strong increase in the %Ndf, nodule numbers and up-regulation of N fixation-related genes (ENOD, nodF and nifH) under elevated CO 2 . Furthermore, elevated CO 2 decreased N concentration and Ntotal Yield in both dnf1 mutants, suggesting that the increased N concentration and Ntotal Yield in Jemalong was solely the result of elevated CO 2 -induced increases of N 2 fixation. In addition, the fixation of N 2 required substantial amount of C resource, and the respiration measurements showed the costs of C for N assimilation from nitrate seem to be lower than those for N 2 fixation [bib_ref] Differential effects of combined N sources on early steps of the Nod..., Barbulova [/bib_ref]. Elevated CO 2 , however, provided the sufficient C to satisfy the energy demand for N 2 fixation, and decreased soil N availability under elevated CO 2 accelerated N 2 fixation in Jemalong plants [bib_ref] Elevated atmospheric CO 2 does not affect per se the preference for..., Zanetti [/bib_ref].
Although elevated CO 2 tends to increase the N concentration and modify N acquisition patterns of legumes, there is little evidence that elevated CO 2 can affect the key enzymes involved in N assimilation [bib_ref] Continuous CO 2 enrichment leads to increased nodule biomass, carbon availability to..., Cabrerizo [/bib_ref]. GS and GOGAT are critical enzymes involved in the assimilation of ammonia, which is not only derived from nitrate reduction and N 2 fixation but also from some secondary metabolism processes, i.e. photorespiration or amino acid catabolism [bib_ref] Glutamine Synthetase in Legumes: Recent Advances in Enzyme Structure and Functional Genomics, Betti [/bib_ref]. Photorespiration is one of the most important physiological process in which high amounts of ammonium are released [bib_ref] The photorespiratory nitrogen cycle, Keys [/bib_ref] , which was likely to be suppressed by elevated CO 2 [bib_ref] Nitrogen assimilation and growth of wheat under elevated carbon dioxide, Bloom [/bib_ref]. This is probably the reason why GS and GOGAT activities were unaffected even though M. truncatula could acquire more N from fixation under elevated CO 2 . Furthermore, elevated CO 2 decreased the enzyme activity and transcripts of GS and GOGAT in both dnf1 mutants, and these decreases were accompanied by decreases in the N concentration of roots and leaves. Thus, it appears that M. truncatula and presumably other legumes require N fixation to maintain N assimilation under elevated CO 2 .
In conclusion, regardless of wild-type and N fixation mutant, elevated CO 2 decreased N uptake from soil by down-regulating the expression of NR and NT of M. truncatula. Wild-type plants, however, are able to up-regulate N fixation related genes and increase nodule numbers under elevated CO 2 to maintain sufficient N concentration for plant growth. This suggests that as atmospheric CO 2 continues to rise, legumes may rely more on N fixation due to less on N uptake from soil. This could benefit agriculture because higher N fixation may compensate N depletion from soil, which would facilitate the growth of nonleguminous plants. Although our study has important implications for agriculture and for regional and global N budgets under predicted CO 2 conditions, the enhancement of leguminous N fixation by elevated CO 2 is environment-dependent [bib_ref] Legume species identity and soil nitrogen supply determine symbiotic nitrogen-fixation responses to..., West [/bib_ref]. N fixation can be limited by the availability of other soil nutrients (i.e., molybdenum, phosphorus, potassium) or by abiotic stresses (i.e., salinity, alkalinity, acidity, drought, fertilizer, metal toxicity) [bib_ref] Element interactions limit soil carbon storage, Van Groenigen [/bib_ref]. Moreover, since N uptake from soil is constrained by elevated CO 2 , legumes are very likely to find it more difficult to maintain their growth under elevated CO 2 when they are subjected to stresses that reduce N fixation. Considering few studies have examined the interactive effects of elevated CO 2 and other abiotic stress on the N dynamics of legume, environmental variables in addition to atmospheric CO 2 concentrations should be considered when predicting future N dynamics of legumes. Besides, Understanding the N dynamics of legume plants and ensuring food security in the future also require a deeper understanding of interaction between legume plants and other organisms such as herbivorous insects.
## Supporting information
File S1 [fig_ref] Table 1: CO 2 effects on N characteristics [/fig_ref] : Primer sequences used for real-time quantitative PCR.
## (doc)
File S2 : P values from two-way ANOVAs for the effects of CO 2 level, M. truncatula genotype, and their interaction on the growth traits and foliar chemical components of alfalfa plants.
## (doc)
File S3 [fig_ref] Figure 1: Above-and below-ground biomass of M [/fig_ref] : The legume genes shown in this figure were tracked in the current study and are involved in N fixation, N uptake from soil, and N assimilation as indicated. The genes include: early nodulespecific protein 40 (ENOD), nodulation genes (nodF), nitrogenfixing genes (nifH), nitrate transporter NRT1.1 (NT), nitrate
[fig] Figure 1: Above-and below-ground biomass of M. truncatula plants as affected by CO 2 level and plant genotype: dnf1-1 and dnf1-2 are deficient in N fixation, and Jemalong is their wild type. Each value represents the average (6SE) of four replicates. Different lowercase letters indicate significant differences between ambient CO 2 and elevated CO 2 within the same genotype. Different uppercase letters indicate significant differences among genotypes within the same CO 2 treatment as determined by Tukey's multiple range test at P,0.05. doi:10.1371/journal.pone.0081373.g001 [/fig]
[fig] Figure 2: Nodule number per root of M. truncatula plant as affected by CO 2 level and plant genotype: dnf1-1 and dnf1-2 are deficient in N fixation, and Jemalong is their wild type. Each value represents the average (6SE) of four replications. Different lowercase letters indicate significant differences between ambient CO 2 and elevated CO 2 within the same genotype. Different uppercase letters indicate significant differences among M. genotypes within the same CO 2 treatment as determined by Tukey's multiple range test at P,0.05. doi:10.1371/journal.pone.0081373.g002 [/fig]
[fig] Figure 4: N concentrations in leaves and roots as well as Ntotal Yield (N content in per plant) of M. truncatula plants as affected by CO 2 level and plant genotype: dnf1-1 and dnf1-2 are deficient in N fixation, and Jemalong is their wild type. Each value represents the average (6SE) of four replicates. Different lowercase letters indicate significant differences between ambient CO 2 and elevated CO 2 within the same genotype. Different uppercase letters indicate significant differences among genotypes within the same CO 2 treatment as determined by Tukey's multiple range test at P,0.05. doi:10.1371/journal.pone.0081373.g004 [/fig]
[fig] Figure 6: Expression of genes involved in N fixation (ENOD, nodF, and nifH), nitrate transportation and reduction (NT and NR), ammonium uptake (AMT), and N assimilation (GS and GOGAT) in leaves of M. truncatula plants as affected by CO 2 level and plant genotype: dnf1-1 and dnf1-2 are deficient in N fixation, and Jemalong is their wild type. Values indicate foldchange in expression based on qPCR determination, and each value represents the average (6SE) of four replicates. An asterisk above a column indicates a significant difference in gene expression under ambient vs. elevated CO 2 (P,0.05). doi:10.1371/journal.pone.0081373.g006 [/fig]
[table] Table 1: CO 2 effects on N characteristics (%Ndf, Nf Yield, Ns Yield) of wild-type Jemalong plant. [/table]
|
The antibody response to the glycan α‐Gal correlates with COVID‐19 disease symptoms
The coronavirus disease 2019 (COVID-19) pandemic caused by severe acute re-
# | introduction
The coronavirus disease , a pandemic caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), has rapidly evolved from an epidemic outbreak to a disease affecting the global population. SARS-CoV-2 infects human host cells by binding to the angiotensin-converting enzyme 2 (ACE2) receptor. [bib_ref] A pneumonia outbreak associated with a new coronavirus of probable bat origin, Zhou [/bib_ref] It has been established that COVID-19 mainly affects the respiratory tract, but as a systemic disease, it affects multiple processes including the gastrointestinal, cardiovascular, neurological, hematopoietic, and immune systems. [bib_ref] Lymphopenia predicts disease severity of COVID-19: a descriptive and predictive study, Tan [/bib_ref] Several days after the onset of symptoms, the SARS-CoV-2 infection becomes more systemic and affects various organs with inflammatory responses and lymphocytopenia. [bib_ref] Lymphopenia predicts disease severity of COVID-19: a descriptive and predictive study, Tan [/bib_ref] Lymphocytopenia is likely caused by the direct lethal effect of SARS-CoV-2 on lymphocytes with the ACE2 receptor 3 and the release of pro-inflammatory cytokines such as tumor necrosis factor-α (TNF-α), interleukin 1 (IL-1) and IL-6 that induce apoptosis in lymphocytes. [bib_ref] Increased TNF-alpha-induced apoptosis in lymphocytes from aged humans: changes in TNF-alpha receptor..., Aggarwal [/bib_ref] The "cytokine storm syndrome (CSS)" has been associated with COVID-19 through the activation of the nuclear factor-kB (NF-kB) innate immune pathway resulting in the upregulation of proinflammatory cytokines. [bib_ref] Efficacy of glutathione therapy in relieving dyspnea associated with COVID-19 pneumonia: a..., Horowitz [/bib_ref] Lymphocytopenia in patients with along with the rise in neutrophils has been associated with worse disease prognosis. Consequently, patients with respiratory distress syndrome in intensive care unit (ICU) show lower lymphocyte counts and higher mortality when compared to other COVID-19 patients. [bib_ref] Clinical characteristics of 138 hospitalized patients with 2019 novel coronavirus-infected pneumonia in..., Wang [/bib_ref] [bib_ref] Novel coronavirus infection (COVID-19) in humans: a scoping review and meta-analysis, Borges Do Nascimento [/bib_ref] Additionally, COVID-19 patients suffer dysbacteriosis in the gut and lung microbiota due to enrichment of opportunistic pathogens and depletion of beneficial commensals, which recommends the development of interventions such as probiotics to reduce the severity of COVID-19 through modification of the microbiota composition. [bib_ref] A pneumonia outbreak associated with a new coronavirus of probable bat origin, Zhou [/bib_ref] [bib_ref] A possible probiotic (S. salivarius K12) approach to improve oral and lung..., Pierro [/bib_ref] [bib_ref] Lung microbiota in the acute respiratory disease: from coronavirus to metabolomics, Fanos [/bib_ref] Humans evolved by losing the capacity to synthesize the glycan Galα1-3Galβ1-(3)4GlcNAc-R (α-Gal), which resulted in the development of a protective response of anti-α-Gal IgM/IgG antibodies against pathogenic viruses (e.g., HIV), bacteria (e.g., Mycobacterium) and parasites (e.g., Plasmodium) containing this modification on membrane proteins. [bib_ref] Gut microbiota elicits a protective immune response against malaria transmission, Yilmaz [/bib_ref] [bib_ref] Catastrophic-selection" interplay between enveloped virus epidemics, mutated genes of enzymes synthesizing carbohydrate..., Galili [/bib_ref] [bib_ref] Environmental and molecular drivers of the α-Gal syndrome, Cabezas-Cruz [/bib_ref] [bib_ref] The alpha-Gal syndrome: new insights into the tick-host conflict and cooperation, De La Fuente [/bib_ref] [bib_ref] Vaccination with alpha-Gal protects against mycobacterial infection in the zebrafish model of..., Pacheco [/bib_ref] The natural IgM/IgG antibodies against α-Gal are produced in response to bacteria with this modification in the microbiota. [bib_ref] Gut microbiota elicits a protective immune response against malaria transmission, Yilmaz [/bib_ref] In addition to anti-α-Gal antibody-mediated pathogen opsonization, this glycan induces various immune mechanisms such as B-cell maturation, macrophage response, activation of the complement system, upregulation of pro-inflammatory cytokines through the Toll-like receptor 2 (TLR2)/NF-kB innate immune pathway, and TLRmediated induction of the anti-inflammatory nuclear factor-erythroid 2-related factor 2 signalling pathway. [bib_ref] Vaccination with alpha-Gal protects against mycobacterial infection in the zebrafish model of..., Pacheco [/bib_ref] [bib_ref] Regulation of the immune response to α-Gal and vector-borne diseases, Cabezas-Cruz [/bib_ref] [bib_ref] Toll-like receptor signaling induces Nrf2 pathway activation through p62-triggered Keap1 degradation, Yin [/bib_ref] In conjunction, the immune response to α-Gal in animal models has shown protection against infectious diseases without inflammatory responses. [bib_ref] Gut microbiota elicits a protective immune response against malaria transmission, Yilmaz [/bib_ref] [bib_ref] Environmental and molecular drivers of the α-Gal syndrome, Cabezas-Cruz [/bib_ref] [bib_ref] The alpha-Gal syndrome: new insights into the tick-host conflict and cooperation, De La Fuente [/bib_ref] [bib_ref] Vaccination with alpha-Gal protects against mycobacterial infection in the zebrafish model of..., Pacheco [/bib_ref] Based on these results, we have hypothesized that the immune response to α-Gal may play a role in the person-to-person variability in COVID-19 disease symptoms with a putative protective capacity. [bib_ref] The exquisite corpse for the advance of science, De La Fuente [/bib_ref] First, if the virus contains α-Gal, it would be possible to limit the zoonotic transmission of SARS-CoV-2 by antibody-mediated opsonization. [bib_ref] The exquisite corpse for the advance of science, De La Fuente [/bib_ref] Secondly, boosting α-Gal-mediated protective immune and antiinflammatory responses may contribute to the control of COVID-19 while increasing protection to pathogens with α-Gal on their surface that negatively affect the individual response to SARS-CoV-2. [bib_ref] Vaccination with alpha-Gal protects against mycobacterial infection in the zebrafish model of..., Pacheco [/bib_ref] [bib_ref] The exquisite corpse for the advance of science, De La Fuente [/bib_ref] To address this hypothesis, herein we characterized the antibody response to α-Gal in patients at different stages of COVID-19 and in comparison with healthy control individuals. The results showed that while the inflammatory response and the anti-SARS-CoV-2 (Spike) IgG antibody titers increased, reduction in anti-α-Gal antibody titers and alteration of anti-α-Gal antibody isotype composition correlated with COVID-19 severity. These results suggested that the inhibition of the α-Gal-induced immune response translates into more aggressive viremia and severe disease symptoms. Reference values for serum immunoglobulin levels 21 were considered in the analysis of the profile of anti-α-Gal antibody isotypes.
## | materials and methods
## | covid-19 patients and healthy control individuals
## | serum and saliva samples
Although this methodology has been previously validated, [bib_ref] Tick-host conflict: immunoglobulin E antibodies to tick proteins in patients with anaphylaxis..., Mateos-Hernández [/bib_ref] In the blood cell analysis, the ICU patients showed a higher lymphocytopenia, percentage and neutrophil counts when compared to hospital discharge and hospitalized individuals (p < .001; and . The cellular and biochemical indicators of systemic inflammation, neutrophil-lymphocyte count ratio (NLR), C-reactive protein (CRP), and D-dimer levels were higher in ICU patients when compared to other patients (p < .002; and . 3.2 | Immune response to SARS-CoV-2 increased with severity in COVID-19 patients All COVID-19 symptomatic patients showed both IgA and IgG antibody titers against SARS-CoV-2 . In asymptomatic cases, only IgG antibody titers were determined, and all tested positive . However, only the IgG titers against the SARS-CoV-2
# | statistical analysis
Spike protein significantly increased in accordance with disease symptoms (p = .02; with a positive correlation (r s > 0; p = 0; . These results showed that COVID-19 patients were immunocompetent despite the inflammatory response.
## | immune response to α-gal varied in covid-19 patients
The serum IgA, IgE, IgM and IgG antibody response to α-Gal was characterized in healthy individuals and COVID-19 patients at different disease stages [fig_ref] 2. 3 |: Determination of antibody titers against SARS-CoV-2 Antibody titers specific for the recognition... [/fig_ref]. The calculated anti-α-Gal IgE levels were below (8.3E − 5 to 3.4E − 02 kU/l) the cut-off value of 0.35 kU/l used for the diagnosis of the α-Gal syndrome. A negative correlation was observed for IgE, IgM, and IgG between anti-α-Gal antibody titers and disease severity (r s < 0; p = 0; [fig_ref] 2. 3 |: Determination of antibody titers against SARS-CoV-2 Antibody titers specific for the recognition... [/fig_ref]. The anti-α-Gal IgA antibody titers did not vary between the different groups (p = .21136; [fig_ref] 2. 3 |: Determination of antibody titers against SARS-CoV-2 Antibody titers specific for the recognition... [/fig_ref] nor correlate with disease severity (r s = 0.02; p = .91; [fig_ref] 2. 3 |: Determination of antibody titers against SARS-CoV-2 Antibody titers specific for the recognition... [/fig_ref]. For anti-α-Gal IgM and IgG antibodies, the titers decreased from healthy to ICU individuals (p < .00001; [fig_ref] 2. 3 |: Determination of antibody titers against SARS-CoV-2 Antibody titers specific for the recognition... [/fig_ref]. However, in asymptomatic cases, the anti-α-Gal IgE titers were higher than in healthy individuals and symptomatic COVID-19 patients (p < .000001; [fig_ref] 2. 3 |: Determination of antibody titers against SARS-CoV-2 Antibody titers specific for the recognition... [/fig_ref]. In COVID-19 patients, the IgE but not IgM and IgG antibody titers were higher in hospitalized patients than in hospital discharge and ICU cases (p < .05; .
The profile of anti-α-Gal antibody isotypes was qualitatively compared between groups including reference values for serum immunoglobulin levels [fig_ref] 2. 3 |: Determination of antibody titers against SARS-CoV-2 Antibody titers specific for the recognition... [/fig_ref]. The results evidenced that anti-α-Gal [fig_ref] 2. 4 |: Determination of antibody titers against α-Gal High absorption capacity polystyrene microtiter plates... [/fig_ref]. . B, Serum anti-SARS-CoV-2 IgA, IgG (spike) and IgG (nucleocapsid) antibody levels were determined by ELISA. The patients were grouped as asymptomatic (n = 10), hospital discharge (n = 27), hospitalized (n = 29) and ICU (n = 25). The results were compared between different groups by one-way ANOVA test (p < .05). A Spearman rho (r s ) correlation analysis (p < .05) was conducted between anti-Spike IgG antibody titers and disease severity (2 = asymptomatic, 3 = hospital discharge, 4 = hospitalized, 5 = ICU). ANOVA, analysis of variance; COVID-19, coronavirus disease 2019; CRP, C-reactive protein; ELISA, enzyme-linked immunosorbent assay; ICU, intensive care unit; IgA, immunoglobulin A; NLR, neutrophil-lymphocyte count ratio; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2 URRA ET AL.
## F i g u r e 1 laboratory tests in covid-19 patients. a, cellular and biochemical indicators of systemic inflammation included neutrophils (cell counts and percent), lymphocytes (cell counts and percent), nlr, d-dimer and crp levels
## | 2069
The analysis of the cytokine response was focused on anti-
# | discussion
Systemic inflammation is associated with changes in the quantity and composition of circulating blood cells and has been identified as the primary basic mechanism resulting in disability and increased mortality in COVID-19. [bib_ref] COVID-19: consider cytokine storm syndromes and immunosuppression, Mehta [/bib_ref] As previously reported, [bib_ref] Prognostic value of neutrophilto-lymphocyte ratio in sepsis: a meta-analysis, Huang [/bib_ref] [bib_ref] Selective CD8 cell reduction by SARS-CoV-2 is associated with a worse prognosis..., Urra [/bib_ref] infection.
In addition to the observed negative correlation between anti-α-Gal IgE, IgM, and IgG antibody titers and COVID-19 disease severity, our results showed differences in the profile of anti-α-Gal antibody isotypes in COVID-19 cases that may be associated with different disease stages . These results suggested that higher anti-α-Gal IgE levels in asymptomatic cases may reflect an allergic response mediated by this glycan, which reflects the trade-off associated with the immune response to α-Gal that benefits humans by providing immunity to pathogen infection while increasing the risk of developing allergic reactions to this molecule. [bib_ref] Environmental and molecular drivers of the α-Gal syndrome, Cabezas-Cruz [/bib_ref] [bib_ref] The alpha-Gal syndrome: new insights into the tick-host conflict and cooperation, De La Fuente [/bib_ref] In healthy individuals as in hospital discharge cases, the higher representation of anti-α-Gal
IgM and/or IgG antibodies may be associated with a protective response to COVID-19. However, in hospitalized patients, the representation of anti-α-Gal antibody isotypes did not vary, which could reflect the absence of protection. Finally, the higher representation of anti-α-Gal IgA antibodies in ICU patients may be associated with the inflammatory response observed in these cases.
In accordance with these results, it was recently shown in endogenous α-Gal-negative turkeys that treatment with probiotic F I G U R E 3 Serum anti-α-Gal antibody response in COVID-19 asymptomatic and symptomatic cases and healthy controls. A, The IgA, IgE, IgM and IgG anti-α-Gal antibody titers were determined by ELISA. Individuals were grouped as healthy controls (n = 37), asymptomatic (n = 10), hospital discharge (n = 27), hospitalized (n = 29) and ICU (n = 25). The results were compared between different groups by one-way ANOVA test (p < .05). A Spearman rho (r s ) correlation analysis (p < .05) was conducted between anti-α-Gal IgA, IgE, IgM and IgG antibody titers and disease severity (1 = healthy, 2 = asymptomatic, 3 = hospital discharge, 4 = hospitalized, 5 = ICU). B, Profile of anti-α-Gal antibody isotype (shown as percentage of antibody titers) for each group. Reference values for serum immunoglobulin levels were included. Antibody isotypes with highest representation on each group are highlighted in red. ANOVA, analysis of variance; COVID-19, coronavirus disease 2019; ELISA, enzyme-linked immunosorbent assay; ICU, intensive care unit; IgA, immunoglobulin A; α-Gal, Galα1-3Galβ1-(3)4GlcNAc-R bacteria with high α-Gal content results in protection against aspergillosis through reduction by still unknown mechanisms in the pro-inflammatory anti-α-Gal IgA response in the lungs.In Spain, differences have been observed in the number of reported cases per 100 000 people by age and sex, with more females at age 20-59 and males at age 60-89 with a higher mortality in males.In this study, differences in age, but not sex, were observed in symptomatic COVID-19 cases . However, the youngest cases corresponded to ICU patients and healthy control individuals, thus reducing the possible effect of age and sex on the observed F I G U R E 4 Salivary anti-α-Gal antibody response in COVID-19 asymptomatic cases and healthy controls. A, The anti-α-Gal IgA and IgG antibody titers were determined by ELISA and compared in asymptomatic cases between serum and saliva samples by Student′s t-test (p < .05; n = 10). B, The anti-α-Gal IgA antibody titers in saliva were determined by ELISA and compared between asymptomatic COVID-19 cases and healthy individuals by Student′s t-test (p < .05; n = 10). COVID-19, coronavirus disease 2019; ELISA, enzyme-linked immunosorbent assay; ICU, intensive care unit; IgA, immunoglobulin A; α-Gal, Galα1-3Galβ1-(3) 4GlcNAc-R F I G U R E 5 A negative correlation between anti-α-Gal antibody titers and COVID-19 disease severity and differences in the profile of anti-α-Gal antibody isotypes may be associated with different disease stages. Our hypothesis is that the dysbacteriosis observed in COVID-19 patients translates into a reduction in total anti-α-Gal antibody titers and alteration of anti-α-Gal antibody isotype composition due to the reduction in the microbiota of α-Gal-containing commensal bacteria. COVID-19, coronavirus disease 2019; IgA, immunoglobulin A; α-Gal, Galα1-3Galβ1-(3)4GlcNAc-R; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2 disease symptoms. Furthermore, currently a clear correlation has not been found between age, sex, and the antibody response to α-Gal. [bib_ref] Prevalence of type I sensitization to alpha-gal in forest service employees and..., Fischer [/bib_ref] [bib_ref] IgE-mediated sensitization to galactose-α-1,3-galactose (α-Gal) in urticaria and anaphylaxis in Spain: geographical..., Borrega [/bib_ref] [bib_ref] IgE antibodies to alpha-gal in the general adult population: relationship with tick..., Gonzalez-Quintela [/bib_ref] The protective response of anti-α-Gal IgM/IgG antibodies against pathogenic organisms containing this modification on membrane proteins has been well documented. [bib_ref] Gut microbiota elicits a protective immune response against malaria transmission, Yilmaz [/bib_ref] [bib_ref] Catastrophic-selection" interplay between enveloped virus epidemics, mutated genes of enzymes synthesizing carbohydrate..., Galili [/bib_ref] [bib_ref] Environmental and molecular drivers of the α-Gal syndrome, Cabezas-Cruz [/bib_ref] [bib_ref] The alpha-Gal syndrome: new insights into the tick-host conflict and cooperation, De La Fuente [/bib_ref] [bib_ref] Vaccination with alpha-Gal protects against mycobacterial infection in the zebrafish model of..., Pacheco [/bib_ref] [bib_ref] Effect of blood type on anti-α-Gal immunity and the incidence of infectious..., Cabezas-Cruz [/bib_ref] In contrast, IgE antibody response against α-Gal has been associated with the allergy to mammalian meat or α-Gal syndrome and other diseases such as atopy, coronary artery disease and atherosclerosis. [bib_ref] IgE antibodies to alpha-gal in the general adult population: relationship with tick..., Gonzalez-Quintela [/bib_ref] [bib_ref] The association between Ixodes holocyclus tick bite reactions and red meat allergy, Van Nunen [/bib_ref] [bib_ref] Galactose-α-1,3-galactose: atypical food allergen or model IgE hypersensitivity?, Wilson [/bib_ref] [bib_ref] Platts-Mills TAE. IgE, α-Gal and atherosclerosis, Wilson [/bib_ref] In this study and based on anti-α-Gal IgE levels, all individuals were negative for α-Gal syndrome. [bib_ref] Alpha-gal syndrome: challenges to understanding sensitization and clinical reactions to alpha-gal, De La Fuente [/bib_ref] The anti-α-Gal IgM/IgG antibodies can protect from infection by opsonizing pathogens with α-Gal on their surface. [bib_ref] Gut microbiota elicits a protective immune response against malaria transmission, Yilmaz [/bib_ref] Other immune-mediated mechanisms may be also activated in response to α-Gal, 14-16 which can be activated by SARS-CoV-2 expressing blood type B antigen on their envelope. Although not addressed in this study, these findings prompted us to consider that blood type O individuals could produce antibodies against A and B
antigens that in addition to IgM/IgG antibodies against α-Gal, which cross-react with the structurally similar blood B antigen, [bib_ref] IgE production to α-gal is accompanied by elevated levels of specific IgG1..., Rispens [/bib_ref] could be involved in a polyvalent recognition of the SARS-CoV-2 Spike that may be implicated in the human protection to COVID-19.
In our study, we did not collect information on ABO blood type in COVID-19 patients. However, in a related study with a similar group of patients (n = 73; 9, 40 and 24 hospital discharge, hospitalized and ICU patients, respectively), the results did not show significant differences in ABO blood group distribution. Nevertheless, the results suggest that the ABO blood factor should be considered when evaluating the antibody response to α-Gal. Blood type O individuals, who produce anti-A and anti-B antibodies, can be protected against SARS-CoV-2 particles carrying blood antigens A or B.
However, blood type A and B individuals, who produce either anti-A or anti-B antibodies, would be protected only against SARS-CoV-2 particles carrying antigen A or B, respectively. Therefore, both blood type A and B individuals will be highly susceptible to SARS-CoV-2 particles coming from blood type O individuals because these viral particles do not carry either blood antigen A or B on the envelope. [bib_ref] Harnessing the natural anti-glycan immune response to limit the transmission of enveloped..., Breiman [/bib_ref] Assuming equal replicative fitness for viruses replicating in cells expressing any of the ABO blood group enzymes, an epidemiological dynamics would be created in which, after a large proportion of the population being exposed and infected by SARS-CoV-2, the frequency of individuals with blood types B and A would be equally represented among COVID-19 patients. In contrast, blood type O individuals would be underrepresented, relative to the frequency of these blood groups in the general population. These predictions were corroborated in a large study (n = 750 000) even after adjusting for age, sex, body mass index, race, ethnicity, and co-morbidities.Therefore, the blood type's effects are not explained by other risk factors including age, sex, race, ethnicity, hypertension, diabetes mellitus, obesity, and cardiovascular and respiratory diseases,which support the immunological considerations discussed above.
Based on the fact that natural antibodies against α-Gal are produced in response to bacteria with this modification in the microbiota, [bib_ref] Gut microbiota elicits a protective immune response against malaria transmission, Yilmaz [/bib_ref] our hypothesis is that the dysbacteriosis observed in COVID-19 patients [bib_ref] Alterations in gut microbiota of patients with COVID-19 during time of hospitalization, Zuo [/bib_ref] translates into a reduction in total anti-α-Gal antibody titers and alteration of anti-α-Gal antibody isotype composition due to the reduction in the microbiota of α-Gal-containing commensal bacteria and other still uncharacterized mechanisms . Alternatively, and hypothetically, individuals with higher α-Gal content in the microbiota may be less susceptible to COVID-19. Additionally, the pulmonary microbiota can be affected with the presence of gut bacteria in the lungs. [bib_ref] Lung microbiota in the acute respiratory disease: from coronavirus to metabolomics, Fanos [/bib_ref] In conclusion, according to these results and previous findings in retrovirus, [bib_ref] A novel mechanism of retrovirus inactivation in human serum mediated by antialpha-galactosyl..., Rother [/bib_ref]
[fig] 2. 3 |: Determination of antibody titers against SARS-CoV-2 Antibody titers specific for the recognition of virus infection based on IgG against SARS-CoV-2 Spike (EI 2606-9601G) and Nucleocapsid (EI 2606-9601-2G) proteins and IgA (EI 2606-9601A) were determined by ELISA (Euroimmun) following the manufacturer′s indications. 19,20 Briefly, 100 µl of the calibrator, positive and negative controls and serum samples at 1:100 dilution was added to the 96-microwell plate coated with SARS-CoV-2 proteins and incubated for 1 h at 37°C. After washing three times with 300 µl/well of wash buffer, 100 µl/well of enzyme conjugate (peroxidase-labelled anti-human IgG or IgA) were added and incubated for 30 min at RT. Then, after 3 washes with 300 µl/well of wash buffer, 100 µl/well of chromogen substrate solution were added and incubated for 15 min (EI 2606-9601-2 G) or 30 min (EI 2606-9601 G; EI 2606-9601 A) at RT. Finally, the colorimetric reaction was stopped with 100 µl/well of stop solution and the absorbance was measured in a spectrophotometer (Thermo Fisher Scientific) at O.D. of 450 nm. Results were evaluated semiquantitatively by calculating the ratio between O.D. of the sample and the O.D of the calibrator, those under 0.8 considered as negative and those over 1.1 as positive. [/fig]
[fig] 2. 4 |: Determination of antibody titers against α-Gal High absorption capacity polystyrene microtiter plates were coated with 50 ng of BSA coated with α-Gal (BSA-α-Gal, thereafter named α-Gal; Dextra) per well in carbonate-bicarbonate buffer (Sigma-Aldrich) and used for ELISA. After an overnight incubation at 4°C, coated plates were washed once with 100 µl/well PBS with 0.05% Tween 20 (PBST; Sigma-Aldrich), blocked with 100 µl/well of 1% human serum albumin (HAS) in PBST (Sigma-Aldrich) for 1 h at RT and then washed four times with 100 µl/well of PBST. Human serum and saliva samples were diluted 1:100 and 1:2, respectively, in PBST with 1% HAS and 100 µl/well were added into the wells of the antigen-coated plates and incubated for 1 h at 37°C. Plates were washed four times with PBST and 100 µl/well of goat anti-human immunoglobulins-peroxidase IgG (FC specific; Sigma-Aldrich), IgM (µchain specific; Sigma-Aldrich), IgE (ɛ-chain specific; Sigma-Aldrich), and IgA (heavy chain specific; Bio-Rad) secondary antibodies diluted 1:1000, v/v in blocking solution were added and incubated for 1 h at RT. Plates were washed four times with 100 µl/well of PBST and 100 µl/well of 3,3,´5,5-tetramethylbenzidine TMB (Promega) were added and incubated for 20 min at RT. Finally, the reaction was stopped with 50 µl/well of 2 N H 2 SO 4 and the O.D. was measured in a spectrophotometer at 450 nm. The average of two technical replicates per sample was used for analysis after background (coated wells incubated with PBS and secondary antibodies) subtraction. [/fig]
|
A retrospective analysis of the care cascades for non‐communicable disease and mental health among people living with HIV at a tertiary‐care centre in Malaysia: opportunities to identify gaps and optimize care
Introduction:The rapidly growing epidemic of non-communicable diseases (NCDs) including mental health among aging people living with HIV (PLWH) has put a significant strain on the provision of health services in many HIV clinics globally. We constructed care cascades for specific NCDs and mental health among PLWH attending our centre to identify potential areas for programmatic improvement.Methods: This was a follow-up study of participants recruited in the Malaysian HIV & Aging study (MHIVA) from 2014 to 2016 at the University Malaya Medical Centre (n = 336). PLWH on suppressive antiretroviral therapy (ART) for a minimum of 12 months were invited to participate. At study entry, all participants underwent screening for diabetes (DM), hypertension (HTN) and dyslipidaemia; and completed assessments using the depression, anxiety and stress scale (DASS-21). Screening results were recorded in medical charts and clinical management provided as per standard of care. A subsequent review of medical records was performed at 24 months following study completion among participants who remained on active followup. Treatment pathways for NCD treatment and psychiatric referrals were assessed based on local practice guidelines to construct the care cascade. Results: A total of 329 participants (median age = 43 years, 83% male, 100% on ART) completed follow-up at 24 months. The prevalence of diabetes was 13%, dyslipidaemia 88% and hypertension 44%, whereas 23% presented with severe/extremely severe symptoms of depression, anxiety and/or stress. More than 50% of participants with dyslipidaemia and hypertension were not diagnosed until study screening, whereas over 80% with prevalent psychiatric symptoms were not previously recognized clinically. Suboptimal control of fasting lipids, sugar and blood pressure were found in the majority of participants despite optimal HIV treatment outcomes maintained over this same period. Only 32% of participants with severe/extremely severe mental health symptoms received psychiatric referrals and 83% of these attended their psychiatry clinic appointments. Conclusions: Systematic screening must be introduced to identify NCDs and mental health issues among PLWH followed by proper linkage and referrals for management of screen-positive cases. Assessment of factors associated with attrition at each step of the care cascade is critically needed to improve health outcomes in our aging patients.
# | introduction
The introduction of antiretroviral therapy has progressively improved the life expectancy of people living with HIV (PLWH). Consequently, the global demographics of PLWH has changed with approximately 7.5 million individuals over the age of 50 years. In the Asia-Pacific region, one in three PLWH are expected to be over the age of 50 years by 2025 [bib_ref] HIV and aging: demographic change in the Asia-Pacific region, Puhr [/bib_ref]. This age advancement is associated with an increased prevalence of chronic comorbidities [bib_ref] Cross-sectional comparison of the prevalence of age-associated comorbidities and their risk factors..., Schouten [/bib_ref] and geriatric syndromes [bib_ref] Major health impact of accelerated aging in young HIV-infected individuals on antiretroviral..., Rajasuriar [/bib_ref] [bib_ref] Geriatric syndromes in older HIV-infected adults, Greene [/bib_ref] among PLWH, leading to calls for HIV treatment programmes to adapt a multi-disciplinary team approach to better address the healthcare needs of the population [bib_ref] Older people with HIV are an essential part of the continuum of..., Siegler [/bib_ref].
Among the Asia-Pacific LMIC countries including Malaysia, the care of PLWH is still largely centred in tertiary care settings and in many instances, HIV specialists are the sole healthcare provider for PLWH. HIV doctors thus inadvertently carry the additional responsibility of primary care in screening for non-communicable diseases, mental health and cancer [bib_ref] Screening and management of mental health and substance use disorders in HIV..., Parcesepe [/bib_ref]. Furthermore, high clinic case loads, limited support personnel and poor integration between specialty care services make it challenging for HIV treatment programmes in the region to adopt the call to establish multidisciplinary care models to manage our patients as they age. Though various models of integrated care for PLWH have been introduced, these have largely been in resource-rich settings and evolved from well established and better resourced health systems compared to those typically found in the LMIC setting [bib_ref] Older people with HIV are an essential part of the continuum of..., Siegler [/bib_ref].
At our treatment centre, we previously identified a high burden of geriatric conditions among PLWH on stable ART on routine follow-up which was associated with poorer quality of life, higher mortality risks and increased healthcare utilization [bib_ref] Major health impact of accelerated aging in young HIV-infected individuals on antiretroviral..., Rajasuriar [/bib_ref]. This study highlighted an urgent need to expand our HIV care to beyond that of virological and immunological outcomes. To this end, we constructed care cascades for the management of specific non-communicable diseases including mental health among PLWH on routine follow-up in our clinic to identify key areas in care provision which needed improvement. This pre-implementation assessment has highlighted important constructs at the patient, provider, organizational and process levels which need to be addressed when considering strategies to integrate NCD and mental healthcare into HIV services in our setting.
# | methods
This was a follow-up study of participants from the Malaysian HIV and Aging (MHIVA) cohort where participants under routine care at the Infectious Diseases Unit, University Malaya Medical Centre (UMMC) were recruited between October 2014 and March 2016. The inclusion criteria for the study was age ≥25 years, HIV RNA < 50 copies/mL on antiretroviral therapy (ART) for at least 12 months and no acute illness on recruitment.
Details of the study and cohort characteristics (N = 336) have been previously described [bib_ref] Major health impact of accelerated aging in young HIV-infected individuals on antiretroviral..., Rajasuriar [/bib_ref]. Briefly, all participants underwent assessments for chronic comorbidities including screening for symptoms of depression, anxiety and stress as well as dyslipidaemia, hypertension and diabetes. Abnormal results from screening were reviewed by study doctors and noted in participants case notes to be addressed at their next clinic visit (within three to six months) or sooner if deemed urgent. Clinical management of abnormal screens were done as per standard of care and described in .
In this study, a review of medical records was performed of participants who remained on active follow-up in our clinic (n = 329) to construct care cascades for NCD and mental health. Participants provided written informed consent for the study team to access medical records up to five years following recruitment. The study protocol was approved by the institutional review board (MEC 20151-937).
## | construction of care cascade for non-communicable diseases
At recruitment into MHIVA, all participants had fasting bloods taken to assess lipids, glucose and HbA1c while serial blood pressures were done to assess for hypertension. Prior diagnosis of NCDs were ascertained from multiple sources; review of medical records, pharmacy records and self-reports by participants. The prevalence of each NCD was calculated from both the number of new diagnosis identified as a result of study screening as well as from previous diagnosis. Records of participants with a new or previous NCD diagnosis were reviewed at 24 months following recruitment and assessed for receipt of treatment; defined as having received either advice on lifestyle modification or medication at any point during follow-up. Control of abnormal lab values to recommended target thresholds were also assessed at 24 months and defined based on local clinical practice guidelinesand summarized in 2.2 | Construction of care cascade for mental health All participants were also screened for depression, anxiety and stress at recruitment with the DASS-21 questionnaire while prior mental health diagnosis were assessed from medical and pharmacy records. Assessment of participants with prevalent psychiatric symptoms requiring referrals are defined in . At 24 months of follow-up, medical records of those with prevalent symptoms were assessed for formal referrals by the HIV doctor to the psychiatric unit and the participants follow-up attendance at the psychiatric clinic.
We additionally explored the care cascade for HIV management in the same period. Indicators assessed were the proportion of participants on follow-up who were still receiving antiretroviral therapy, maintained HIV RNA levels at <50 copies/mL and had >80% compliance of HIV clinic attendance over the follow-up period. A missed appointment or a change in appointment dates beyond three months of the original date were assessed as non-compliant.
## | results and discussion
A total of 329 from the 336 participants initially recruited in the MHIVA cohort remained on active follow-up at our unit after 24 months. The majority of participants were males (82.7%) and the median age was 44 (interquartile range, IQR 38-51) years. Participants had received ART for a median duration of 7 (4-11) years at the point of study inclusion. Baseline CD4 T cell count was 110 (35-246) cells/µL, whereas current CD4 T-cell count was 571 (403-730) cells/µL. The majority received efavirenz (NNRTI)-based regimen (85%) followed by a PI-based regimen (12%).
The prevalence of NCDs were high and similar to other cohorts of PLWH in the region [bib_ref] Diabetes mellitus burden among people living with HIV from the Asia-Pacific region, Han [/bib_ref] [bib_ref] Greater burden of chronic comorbidities and co-medications among people living with HIV..., Ruzicka [/bib_ref] [bib_ref] Prevalence of hypertension among patients aged 50 and older living with human..., Dakum [/bib_ref] , with diabetes found in 13%, dyslipidaemia 88% and hypertension 44%. For all three NCDs assessed, a significant proportion of participants were previously undiagnosed and identified only as a result of study screening, implying a poor practice of screening for NCDs as part of routine HIV care [fig_ref] Figure 1: Treatment cascades for non-communicable diseases and HIV disease among PLWH [/fig_ref]. This observation has also been documented in other HIV clinic cohorts in the region [bib_ref] Diabetes, mortality and glucose monitoring rates in the TREAT Asia HIV Observational..., Bijker [/bib_ref] [bib_ref] High prevalence of non-communicable diseases and associated risk factors amongst adults living..., Chhoun [/bib_ref]. By 24 months of follow-up, 89% of those diagnosed with diabetes (n = 44) had received treatment (16% lifestyle only, 84% medication). However, fewer participants with a diagnosis of dyslipidaemia and hypertension were on treatment, 66% and 64%, respectively, and this could be associated with the asymptomatic nature of these conditions. Thirty-three percent of participants treated for dyslipidaemia and 23% treated for hypertension were managed with lifestyle modifications alone (results not shown). The proportion of participants achieving target laboratory outcomes for NCDs were also suboptimal in our cohort with control of lipid levels being the poorest (25% of all diagnosed).
We documented the care cascade for HIV treatmentto explore if poor NCD management over the 24-month period could have been contributed by poor participant attendance and thus the lack of opportunity to optimize NCD care. However, we did not find this to be a potential reason for the observed results as prevalence for all HIV care indicators were at least 90%. Over the 24-month observation period, participants appeared to have well-controlled HIV disease, but a large proportion experienced undiagnosed and/or sub-optimal control of hypertension, dyslipidaemia and diabetes. Poor control of NCDs has recently been described to be a significant contributor of fatal and non-fatal cardiovascular events among PLWH on ART in the Asia-Pacific region [bib_ref] Estimated risk of cardiovascular disease among the HIV-positive patients aged 40 years..., Wu [/bib_ref] and addressing this gap should be made a programmatic priority in our clinic.
Participants in this study, by nature of their inclusion criteria, were PLWH with no outstanding HIV-related medical issues and on long-term ART. The latter, is in itself associated with metabolic and CVD complications [bib_ref] Association of predicted 10 years cardiovascular mortality risk with duration of HIV..., Todowede [/bib_ref] but often overlooked. In our setting, patients of this profile tend to be seen by trainee/junior doctors who may not be familiar with the need or importance of screening for NCDs. We speculate that this oversight may have contributed to the poor management of NCDs among PLWH in our setting despite their HIV care being optimal over the same period. Additionally, as clinic loads are high, patients in queue tend to be seen by any available doctor and this system is not conducive in ensuring care plans are continued especially in the context of Malaysia where shared-decision making is not widely practiced [bib_ref] An overview of patient involvement in healthcare decision-making: a situational analysis of..., Ng [/bib_ref] [bib_ref] The state of shared decision making in Malaysia, Lee [/bib_ref]. This could potentially have contributed to participants remaining on lifestyle modification as their primary treatment for an extended period despite not achieving optimal NCD control. We also could not rule out the reluctance of participants to initiate medications for NCD due to concerns of pill burden and side effects despite being advised by their doctors (personal communication). Additionally, ART medications are government subsidized, whereas NCD medications are not. Our findings collectively imply a poor appreciation by both HIV providers and patients of the importance of optimizing care for NCDs. Whether this is due to inadequate training among providers, poor awareness among both providers and patients, gaps in health financing or other structural/process failures needs to be explored further. . Abbreviated treatment cascade for PLWH presenting with extreme to extremely severe symptoms of depression, anxiety and/or stress requiring psychiatric referrals (n = 329). The care cascade for mental health issues were assessed from screening conducted as part of the Malaysian HIV and Aging (MHIVA) study to the first presentation to a psychiatrist/psychologist within 24 months after the end of study recruitment. New diagnosis with prevalent mental health symptoms requiring referrals were defined as individuals with scores ranging from severe to extremely severe on the Depression, Anxiety and Stress questionnaire (DASS-21) Individuals with symptoms and a prior mental health diagnosis are indicated as 'recognized clinically'. The proportion of those with prevalent symptoms who subsequently received referrals to a psychiatrist or psychologist by 24 months following end of recruitment were noted as were those who then followed through with their first appointment.
previous mental health diagnosis. This is consistent with the lack of routine mental health screening practice in our HIV clinic. To our surprise, only 32% of participants requiring further psychiatric assessment following screening with DASS-21 received referrals to see a psychiatrist/psychologist. Of note and contrary to prior reports of significant drop outs in clinic attendance for mental health services following referrals [bib_ref] The depression treatment cascade: disparities by alcohol use, drug use, and panic..., Diprete [/bib_ref] [bib_ref] Falling through the cracks: the gaps between depression prevalence, diagnosis, treatment, and..., Pence [/bib_ref] , we found that the majority of participants who received referrals to the psychiatric clinic met their appointments. The management of mental health among participants showed a clear lapse in screening practices and recognition of the need for referrals by HIV care providers. Screening for mental health is not routinely performed in many HIV treatment sites in the Asia-Pacific region [bib_ref] Screening and management of mental health and substance use disorders in HIV..., Parcesepe [/bib_ref] [bib_ref] Barriers to mental healthcare and treatment for people living with HIV in..., Sohn [/bib_ref] despite high reported prevalence for a range of mental health disorders among PLWH [bib_ref] Depression, social support, and stigma as predictors of quality of life over..., Garfin [/bib_ref] [bib_ref] Anxiety and depressive symptoms among patients infected with human immunodeficiency virus in..., Kee [/bib_ref]. The sentiment and clinical experience of our HIV providers are that participants are reluctant to attend psychiatry clinics likely due to stigma (personal communication) and thus in the absence of opportunities for intervention, there is limited value to perform screening. Prior efforts to organize joint HIV and mental health/substance abuse clinics in our unit have not materialized due to staff shortages and poor coordination. Our assessment of the treatment cascade, however, challenges this sentiment as we found that the majority of participants who received referrals did attend appointments at the psychiatry clinic. However, our study did not include an assessment of the adequacy of treatment obtained via this care pathway which could take considerably longer than 24 months. Future studies will need to thoroughly explore reasons for the attrition along the mental healthcare cascade including effectiveness of treatment and the need for additional training in mental health assessment for HIV providers, their readiness and barriers to providing referrals and patients stigma experience in accessing mental healthcare services.
# | conclusions
In conclusion, we identified a high rate of NCDs and mental health issues requiring follow-up care among PLWH attending our HIV clinic. Assessment of the care cascade for these conditions have identified key areas which need to be urgently addressed; specifically improving screening practices and linkage to care, as well as optimizing treatment outcomes. Future studies will need to systematically assess factors which facilitate or impede these processes within our HIV treatment programme in order to improve health outcomes in our patients as they age. RR, AK, SFSO and AK conceived the study; CML, DR, WPL, AK and FJY contributed to recruitment and data collection, DR, FJY, CML, WPL and RR analysed the data; RR and CML wrote and formatted the manuscript; RR, SFSO and AK contributed to funding. All authors reviewed and approved the final manuscript.
## A c k n o w l e d g e m e n t s
The authors thank all the participants and investigators of the Malaysian HIV and Aging Study and Malaysian Elderly Longitudinal Research (MELOR); Rishanantini Sakthivel, Sivaraj Naidu, Nor Syuhada Ahmad Bashah, Siti Azdiah Andul Aziz and Sheron Goh for helping with participant recruitment and performing health assessments.
## D i s c l a i m e r
The abstract of this study has been submitted for consideration at the 23 International AIDS Conference 2020.
[fig] Figure 1: Treatment cascades for non-communicable diseases and HIV disease among PLWH (n = 329). Non-communicable disease treatment cascades (grey bars) were constructed for (A) hypertension, (B) dyslipidaemia and (C) diabetes mellitus from laboratory and clinical data collected following screening of the Malaysian HIV and Aging study participants until 24 months after the end of study recruitment. Participants with a diagnosis (new or previous) who received either lifestyle modification advice or medication by the end of the assessment period were considered as treated. Participants who achieved target thresholds for HbA1C or fasting blood sugar (diabetes), LDL-cholesterol (dyslipidaemia) and systolic blood pressure (hypertension), were considered as controlled. Care cascades for (D) HIV management (black bars) were also constructed for HIVrelated indicators assessed over the same period. [/fig]
[fig] Figure 2: Abbreviated treatment cascade for PLWH presenting with extreme to extremely severe symptoms of depression, anxiety and/or stress requiring psychiatric referrals (n = 329). The care cascade for mental health issues were assessed from screening conducted as part of the Malaysian HIV and Aging (MHIVA) study to the first presentation to a psychiatrist/psychologist within 24 months after the end of study recruitment. New diagnosis with prevalent mental health symptoms requiring referrals were defined as individuals with scores ranging from severe to extremely severe on the Depression, Anxiety and Stress questionnaire (DASS-21) Individuals with symptoms and a prior mental health diagnosis are indicated as 'recognized clinically'. The proportion of those with prevalent symptoms who subsequently received referrals to a psychiatrist or psychologist by 24 months following end of recruitment were noted as were those who then followed through with their first appointment. [/fig]
|
A Review on the Synthesis of Fluorescent Five- and Six-Membered Ring Azaheterocycles
# Introduction
The synthesis of fluorescent azaheterocycles continues to arouse strong interest due to their great potential for application as sensors and biosensors, luminophores on in the construction of Organic Light-Emitting Diode (OLED) devices, laser and other semiconductor devices [bib_ref] The Optoelectronic Nose: Colorimetric and Fluorometric Sensor Arrays, Li [/bib_ref] [bib_ref] Cancer Biomarker Detection: Recent Achievements and Challenges, Wu [/bib_ref] [bib_ref] Detecting Cancer by Breath Volatile Organic Compound Analysis: A Review of Array-Based..., Queralto [/bib_ref] [bib_ref] Efficient green electroluminescence from 1,3-diphenyl-1H-pyrazolo[3,4-b]quinoxaline dyes in dye-doped polymer based electroluminescent devices, Gąsiorski [/bib_ref] , as well as to their potential biological properties as antimicrobial [bib_ref] Synthesis and in vitro antibacterial activity of novel heterocyclic derivatives of 18-nor-equilenin, Karnik [/bib_ref] , antifungal [bib_ref] Azasteroids as antifungals, Burbiel [/bib_ref] , anticancer [bib_ref] Efficient synthesis and characterization of ethyl 7-acetyl-2-substituted 3-(substitutedbenzoyl) indolizine-1-carboxylates for in vitro..., Sandeep [/bib_ref] [bib_ref] Design and Synthesis of 1,2-Bis(hydroxymethyl)pyrrolo[2,1-a]phthalazine Hybrids as Potent Anticancer Agents that Inhibit..., Chang [/bib_ref] , antituberculosis [bib_ref] Synthesis of indolizine derivatives with selective antibacterial activity against Mycobacterium tuberculosis, Gundersen [/bib_ref] [bib_ref] Generation and exploration of new classes of antitubercular agents: The optimization of..., Moraski [/bib_ref] antioxidant [bib_ref] Synthesis and antioxidant activity of 3,3 -diselanediylbis (N,N-disubstituted indolizine-1-carboxamide) and derivatives, Narajji [/bib_ref] and anti-HIV [bib_ref] Indolizine derivatives as HIV-1 VIF-ElonginC interaction inhibitors, Huang [/bib_ref] agents.
The advantages of the azaheterocyclic fluorophores, such as small size, enriched photostability, a wide and tunable spectral range, and, frequently, high brightness, are the reason why these fluorophores are preferred and used in various medical application. Probe structure can be modified to adjust excitation and emission wavelengths, target-binding affinity, chemical reactivity, and subcellular localization [bib_ref] Small-molecule fluorescent probes and their design, Fu [/bib_ref] [bib_ref] Spectral-Luminescent Properties of 12-Oximino Derivatives of 8-AZA-D-Homogona-12,17a-Diones and their Concentration Dependence, Bagnich [/bib_ref] [bib_ref] Delayed Fluorescence and Phosphorescence of 8-Aza-d-Homogonane in the Gas and Condensed Phases, Borisevich [/bib_ref] [bib_ref] Specific fluorescence properties and picosecond transient absorption of 8-azasteroids, Akhrem [/bib_ref].
In this review, we try to present an overview of the newest research in the synthesis of fluorescent azaheterocyclic derivatives, from the literature reported during the last ten years. The methodology adopted for literature search and selection was to access the Reaxys database for "fluorescence", "azaheterocycles" and "fluorescent azaheterocycles" keywords. The obtained results, as well as the references cited, were selected to be included in the present review if they present the synthesis of fluorescent azaheterocycles, and the fluorescent properties of these compounds were evaluated.
The present review was organized taking into account the size of the azaheterocycle, the nature of the cycle (monocycle and fused polycycles), as well as our contribution in this research field.
## Synthesis and fluorescent properties of five-and six-membered ring azaheterocycles
## Five membered ring azaheterocycles
Five membered ring azaheterocycles, such as pyrroles, diazoles, triazoles, tetrazoles and their fused derivatives, are privileged scaffolds posing a wide range of biological activities and diverse applications in organic electronics [bib_ref] Synthesis, structure, antimycobacterial and anticancer evaluation of new pyrrolo-(phenanthroline) derivatives, Matarneh [/bib_ref] [bib_ref] An insight into the medicinal perspective of synthetic analogs of indole: A..., Thanikachalam [/bib_ref] [bib_ref] New indolizines with phenanthroline skeleton: Synthesis, structure, antimycobacterial and anticancer evaluation, Danac [/bib_ref] [bib_ref] Current status of pyrazole and its biological activities, Naim [/bib_ref] [bib_ref] Mangalagiu, I.I. Hybrid imidazole (benzimidazole)/pyridine (quinoline) derivatives and evaluation of theiranticancer and..., Mantu [/bib_ref] [bib_ref] The therapeutic journey of benzimidazoles: A review, Bansal [/bib_ref] [bib_ref] A. 1,2,3-Triazole-containing hybrids as leads in medicinal chemistry: A recent overview, Bozorov [/bib_ref] [bib_ref] Tetrazoles: Synthesis and biological activity, Kaushik [/bib_ref]. Due to these properties, many researchers were focused on the obtaining of new compounds with such skeleton.
Martínez-Lara et al. [bib_ref] Straight access to highly fluorescent angular indolocarbazoles via merging Au-and Mo-catalysis, Martínez-Lara [/bib_ref] developed a new obtaining methodology of two different angular indolocarbazole moieties by using two sequential gold-and molybdenum-catalysis to obtain new indolo carbazoles as potential luminophores with application in OLED devices.
A selection of 1,1-bis(indol-3-yl)-3-alkyn-2-ols 3, was readily prepared through the alkylation of the indole derivatives 1 with methyl or phenyl glyoxal, followed by a standard Sonogashira-type reaction with 2-nitroaryl iodides. Next, these compounds underwent a cyclization reaction, promoted by NaAuCl 4 , affording the corresponding carbazoles 4 (see Scheme 1).
## Five membered ring azaheterocycles
Five membered ring azaheterocycles, such as pyrroles, diazoles, triazoles, tetrazoles and their fused derivatives, are privileged scaffolds posing a wide range of biological activities and diverse applications in organic electronics. [bib_ref] Synthesis, structure, antimycobacterial and anticancer evaluation of new pyrrolo-(phenanthroline) derivatives, Matarneh [/bib_ref] [bib_ref] An insight into the medicinal perspective of synthetic analogs of indole: A..., Thanikachalam [/bib_ref] [bib_ref] New indolizines with phenanthroline skeleton: Synthesis, structure, antimycobacterial and anticancer evaluation, Danac [/bib_ref] [bib_ref] Current status of pyrazole and its biological activities, Naim [/bib_ref] [bib_ref] Mangalagiu, I.I. Hybrid imidazole (benzimidazole)/pyridine (quinoline) derivatives and evaluation of theiranticancer and..., Mantu [/bib_ref] [bib_ref] The therapeutic journey of benzimidazoles: A review, Bansal [/bib_ref] [bib_ref] A. 1,2,3-Triazole-containing hybrids as leads in medicinal chemistry: A recent overview, Bozorov [/bib_ref] [bib_ref] Tetrazoles: Synthesis and biological activity, Kaushik [/bib_ref]. Due to these properties, many researchers were focused on the obtaining of new compounds with such skeleton.
Martínez-Lara et al. [bib_ref] Straight access to highly fluorescent angular indolocarbazoles via merging Au-and Mo-catalysis, Martínez-Lara [/bib_ref] developed a new obtaining methodology of two different angular indolocarbazole moieties by using two sequential gold-and molybdenum-catalysis to obtain new indolo carbazoles as potential luminophores with application in OLED devices.
A selection of 1,1-bis(indol-3-yl)-3-alkyn-2-ols 3, was readily prepared through the alkylation of the indole derivatives 1 with methyl or phenyl glyoxal, followed by a standard Sonogashira-type reaction with 2-nitroaryl iodides. Next, these compounds underwent a cyclization reaction, promoted by NaAuCl4, affording the corresponding carbazoles 4 (see Scheme 1). Scheme 1. Synthetic procedure used for the obtaining of indolylcarbazole derivatives 4.
Without further purification, the crude carbazoles were directly subjected to Mo-catalyzed Cadogan reaction delivering the desired indolocarbazoles 5. All indolocarbazoles were obtained in high yields. The authors [bib_ref] Straight access to highly fluorescent angular indolocarbazoles via merging Au-and Mo-catalysis, Martínez-Lara [/bib_ref] used microwave irradiation for the synthesis of the indolocarbazole derivatives 4 and 5, and obtained final products in much shorter time and with better yields (see Scheme 2). Without further purification, the crude carbazoles were directly subjected to Mocatalyzed Cadogan reaction delivering the desired indolocarbazoles 5. All indolocarbazoles were obtained in high yields. The authors [bib_ref] Straight access to highly fluorescent angular indolocarbazoles via merging Au-and Mo-catalysis, Martínez-Lara [/bib_ref] used microwave irradiation for the synthesis of the indolocarbazole derivatives 4 and 5, and obtained final products in much shorter time and with better yields (see .
To obtain a perspective on the effect that the C-7 substituent of the indolocarbazole skeleton (methyl in 5aa, and phenyl in 5ba) the photoluminescent properties of two selected compounds, 5aa and 5ba in DMSO were investigated .
The phenyl-substituted 5ba indolo carbazole shows a wider absorption wavelength range than methylsubstituted 5aa indolo carbazole and both show fluorescence quantum yields around 70% in DMSO, double as the Φ fl of some similar compounds from the literature. These compounds are potentially useful for optoelectronic applications due to their high extinction coefficient and fluorescence quantum yields.
By combining 2H-imiazole 1-oxides 7a-l with pentafluorophenyl lithium 8 made in situ from pentafluorobenzene 6 and n-BuLi, Moseev et al. [bib_ref] Transition-Metal-Free C-H/C-Li Coupling of Nonaromatic 2H-Imidazole 1-Oxides with Pentafluorophenyl Lithium in the..., Moseev [/bib_ref] were able to create novel pentafluoroaryl-modified derivatives of 2H-imidazole type 10a-l and 11a-l, as is presented in Scheme 3. Pentafluorophenyllithium 8 first attacks the imidazole-N-oxide 7a-l, resulting in the unstable σ H -adduct 9 that can be converted into the product either by "addition-elimination" (Path A), leading to the formation of compounds 10a-l, or by . Fluorescent properties of the selected 5aa and 5ba in DMSO. The phenyl-substituted 5ba indolo carbazole shows a wider absorption wavelength range than methylsubstituted 5aa indolo carbazole and both show fluorescence quantum yields around 70 % in DMSO, double as the Φfl of some similar compounds from the literature. These compounds are potentially useful for optoelectronic applications due to their high extinction coefficient and fluorescence quantum yields.
By combining 2H-imiazole 1-oxides 7a−l with pentafluorophenyl lithium 8 made in situ from pentafluorobenzene 6 and n-BuLi, Moseev et al. [bib_ref] Transition-Metal-Free C-H/C-Li Coupling of Nonaromatic 2H-Imidazole 1-Oxides with Pentafluorophenyl Lithium in the..., Moseev [/bib_ref] were able to create novel pentafluoroaryl-modified derivatives of 2H-imidazole type 10a-l and 11a-l, as is presented in Scheme 3. Pentafluorophenyllithium 8 first attacks the imidazole-N-oxide 7a−l, resulting in the unstable σ H -adduct 9 that can be converted into the product either by "addition-elimination" (Path A), leading to the formation of compounds 10a-l, or by "addition-oxidation" (Path B), leading to the creation of compounds 11a-l containing N-oxide group.
## Scheme 2. synthesis of indolo[2,3-c]carbazole derivatives.
Molecules 2022, [bib_ref] Transition-Metal-Free C-H/C-Li Coupling of Nonaromatic 2H-Imidazole 1-Oxides with Pentafluorophenyl Lithium in the..., Moseev [/bib_ref] O a R 1 =Ph; R 2 =Me; R 3 =Me b R 1 =Ph; R 2 =Me; R 3 =Et c R 1 =Ph; R 2 ,R 3 =-(CH 2 ) 5d R 1 =C 6 H 4 Br(p); R 2 =Me; R 3 =Me e R 1 =C 6 H 4 Br(p); R 2 =Me; R 3 =Et f R 1 =2-thienyl; R 2 =Me; R 3 =Me g R 1 =2-thienyl; R 2 =Me; R 3 =Et h R 1 =C 6 H 4 NO 2 (p); R 2 =Me; R 3 =Me i R 1 =C 6 H 4 OCH 3 (p); R 2 =Me; R 3 =Me j R 1 =C 6 H 4 OCH 3 (p); R 2 ,R 3 =-(CH 2 ) 5k R 1 =C 6 H 4 CH 3 (p); R 2 =Me; R 3 =Me l R 1 =C 6 H 4 OCF 3 (p); R 2 =Me; R 3 =Me Scheme 3. Synthesis of pentafluoroaryl-modified derivatives.
As possible applications, the authors claim that the push-pull fluorophore systems containing both electron-donating (EDG) and electron-withdrawing (EWG) groups in the vicinal position could be used as push−pull fluorophore systems. The obtained molecule types 10 and 11 have undergone a photophysical study to determine the potential practical uses in the construction of photoactive materials. The authors results are shown in [fig_ref] Table 2: The photophysical properties for the obtained molecules 10 and 11 [/fig_ref]. Maximum absorption wavelength (λ abs ), molar absorptivity at the absorption maximum wavelength (ε max ), maximum emission wavelength (λ em ), fluorescence quantum yield (Φ fl ), average lifetime (τ), and fluorescence (k f ) amd non-radiative (k nr ) rate constants.
Molecules 2022, 27, 6321 4 of 37
As possible applications, the authors claim that the push-pull fluorophore systems containing both electron-donating (EDG) and electron-withdrawing (EWG) groups in the vicinal position could be used as push-pull fluorophore systems. The obtained molecule types 10 and 11 have undergone a photophysical study to determine the potential practical uses in the construction of photoactive materials. The authors results are shown in [fig_ref] Table 2: The photophysical properties for the obtained molecules 10 and 11 [/fig_ref]. The authors concluded that the presence of EWG in the para-position of the phenyl group improved the fluorescence quantum yield of the produced fluorophores by measuring the quantum yield of polyfluorinated-2Himidazole derivatives 10 and 11.
The phenomenon of intramolecular charge transfer (ICT) in polar protonic solvents (MeOH and EtOH) has been shown for several compounds, and connections between "structure-photophysical characteristics" have been found. The characteristics of the investigated photophysical properties allow them to adjust the structures of the photoactive molecules, which opens the possibility of using the synthesized push-pull systems in the development of difficult fluorometric molecular sensors.
Pu et al.synthesized antipyrine-containing diarylethenes derivatives 14 in their attempt to obtain novel multi-controllable photochromic and fluorescent derivatives (see . It was determined that 14o had an absolute fluorescence quantum yield of 1.4%. UV light (297 nm) irradiation caused the photocyclization reaction, which resulted in the synthesis of 14c with a quantum yield of 0.8%. As a result, the emission strength of 14o ab- It was determined that 14o had an absolute fluorescence quantum yield of 1.4%. UV light (297 nm) irradiation caused the photocyclization reaction, which resulted in the synthesis of 14c with a quantum yield of 0.8%. As a result, the emission strength of 14o abruptly reduced, and the fluorescence changed from bright orange to darkness. Conversely, by exposing 14o with appropriate visible light, the fluorescence may be recovered.
At 580 nm, the isomer 14o showed high brightly orange fluorescence. The emission strength of 14o reduced to 18% with the addition of 5.0 equiv of tetrabutylammonium hydroxide, and the color changed noticeably from bright orange to blackness as a result of the creation of deprotonated 14o'. The fluorescence could return to its initial form upon neutralization with 2.0 equiv of HCl. Similar to this, the stimulation of TBAH/HCl in acetonitrile could reversibly alter the fluorescence of 14c. When irradiated to UV/vis light, the fluorescence alterations between 14o' and 14c' was irreversible. The fluorescence of 14o' stayed unaltered when irradiated to UV light, however when exposed to visible light (λ > 500 nm), the fluorescence of 14c' might return to that of 14o', as may be observed in Scheme 5. The results obtained by the authors shown that the base could suppress the photochromism of diarylethene with an antipyrine unit's and acid could restore it. In order to examine the fluorescence variations of 14o, 5.0 equiv of several metal ions, including Al 3+ , Zn 2+ , Ni 2+ , K + , Cd 2+ , Ca 2+ , Cu 2+ , Ba 2+ , Pb 2+ , Cr 3+ , Co 2+ , Sr 2+ , Mg 2+ , Mn 2+ , Hg 2+ and Fe 3+ , were added. The authors pointed out that the addition of Al 3+ was the only metal ion that significantly affected the fluorescence of 14o; other metal ions had a negligible effect. As a result of this study, the diarylethene with an antipyrine unit type 14o demonstrated special base-gated photochromic properties and could be used as a highly selective naked-eye chemosensor for Al 3+ detection.
## Six membered ring azaheterocycles
Electron-deficient six membered ring azaheterocycles, such as azines and diazines, are strongly desired and studied due to their potential biological activities and also to their potential applications in organic electronics. [bib_ref] Fluorescent labeling of biomolecules with organic probes, Gonçalves [/bib_ref] [bib_ref] Chemical reporters for biological discovery, Grammel [/bib_ref] [bib_ref] Cellular incorporation of unnatural amino acids and bioorthogonal labeling of proteins, Lang [/bib_ref] [bib_ref] Design and application of receptor-targeted fluorescent probes based on small molecular fluorescent..., Zhang [/bib_ref] [bib_ref] Photoluminescent nanoparticles for chemical and biological analysis and imaging, Algar [/bib_ref].
Motivated by the small size and high quantum yield (Φ = 60%) of the unsubstituted pyridin-2-amine, being a potential scaffold for a fluorescent probe, Li et al. [bib_ref] Synthesis and Fluorescent Properties of Aminopyridines and the Application in "Click and..., Li [/bib_ref] , in their research for new multisubstituted aminopyridines, used the Rh-catalyzed coupling of vinyl azide with isonitrile to form a vinyl carbodiimide intermediate, which followed a tandem cyclization with an alkyne to the desired amine.
The reactions were carried "one-pot", using vinyl azide 15 and isonitrile 16, Rh-cata- In order to examine the fluorescence variations of 14o, 5.0 equiv of several metal ions, including Al 3+ , Zn 2+ , Ni 2+ , K + , Cd 2+ , Ca 2+ , Cu 2+ , Ba 2+ , Pb 2+ , Cr 3+ , Co 2+ , Sr 2+ , Mg 2+ , Mn 2+ , Hg 2+ and Fe 3+ , were added. The authors pointed out that the addition of Al 3+ was the only metal ion that significantly affected the fluorescence of 14o; other metal ions had a negligible effect. As a result of this study, the diarylethene with an antipyrine unit type 14o demonstrated special base-gated photochromic properties and could be used as a highly selective naked-eye chemosensor for Al 3+ detection.
## Six membered ring azaheterocycles
Electron-deficient six membered ring azaheterocycles, such as azines and diazines, are strongly desired and studied due to their potential biological activities and also to their potential applications in organic electronics [bib_ref] Fluorescent labeling of biomolecules with organic probes, Gonçalves [/bib_ref] [bib_ref] Chemical reporters for biological discovery, Grammel [/bib_ref] [bib_ref] Cellular incorporation of unnatural amino acids and bioorthogonal labeling of proteins, Lang [/bib_ref] [bib_ref] Design and application of receptor-targeted fluorescent probes based on small molecular fluorescent..., Zhang [/bib_ref] [bib_ref] Photoluminescent nanoparticles for chemical and biological analysis and imaging, Algar [/bib_ref].
Motivated by the small size and high quantum yield (Φ = 60%) of the unsubstituted pyridin-2-amine, being a potential scaffold for a fluorescent probe, Li et al. [bib_ref] Synthesis and Fluorescent Properties of Aminopyridines and the Application in "Click and..., Li [/bib_ref] , in their The reactions were carried "one-pot", using vinyl azide 15 and isonitrile 16, Rh-catalyst, proper ligand and 1,4-dioxane as solvent, under N 2 atmosphere and at room temperature in the first stage, then after the vinyl azide spot disappeared on TLC, NH 4 Cl, NaHCO 3 , and alkyne were added, and the mixture was heated to 120 - C for 8 h (see . For quantitative fluorescent detection, the solution of aminopyridine was then diluted to 10 μM. The measured parameters are presented in [fig_ref] Table 3: Fluorescent properties of aminopyridines [/fig_ref]. [fig_ref] Table 2: The photophysical properties for the obtained molecules 10 and 11 [/fig_ref] λabs-absorption maxima in ethanol (longest wavelength transition), λex-excitation wavelenght, λemmaxima of the corrected emission spectra in ethanol, Φfl-the fluorescence quantum yield, external standard: 9,10-diphenylanthracene (Φ = 95%, in cyclohexane).
Compound 17 was derivatized in order to observe the variation of the photoluminescent properties, as is presented in Scheme 7. For quantitative fluorescent detection, the solution of aminopyridine was then diluted to 10 µM. The measured parameters are presented in [fig_ref] Table 3: Fluorescent properties of aminopyridines [/fig_ref]. [fig_ref] Table 2: The photophysical properties for the obtained molecules 10 and 11 [/fig_ref] λ abs -absorption maxima in ethanol (longest wavelength transition), λ ex -excitation wavelenght, λ em -maxima of the corrected emission spectra in ethanol, Φ fl -the fluorescence quantum yield, external standard: 9,10diphenylanthracene (Φ = 95%, in cyclohexane).
Compound 17 was derivatized in order to observe the variation of the photoluminescent properties, as is presented in Scheme 7.
Thus, when the tertiary butyl group was cleaved using CF 3 COOH the obtained compound 27 showed no fluorescence. When the ester groups were hydrolyzed to carboxylic acid, the obtained compound 28 still had a good fluorescence (Φ = 31%). When compound 35 was reduced with LiAlH 4 a high fluorescent (Φ = 81%) alcohol 29 was obtained in excellent yield, but the λ em was reduced to 400 nm.
The same authors [bib_ref] Synthesis and Fluorescent Properties of Aminopyridines and the Application in "Click and..., Li [/bib_ref] , in order to apply the newly synthesized compound as fluorescent probe, used the click reaction to bind the fluorescent aminopyridine scaffold to biomolecules conjugated with alkynyl group (see . Firstly, they synthesized a Boc-NH protected compound 30 accordingly to the above-described method from corresponding starting materials, then the azido substituted derivative 31 was obtained via a Sandmeyer reaction. Finally, the azide 31 react easily with phenylacetylene or propynol (by using catalytic amounts of Cu(II)) to obtain the desired triazole products 32 or 33. Thus, when the tertiary butyl group was cleaved using CF3COOH the obtained compound 27 showed no fluorescence. When the ester groups were hydrolyzed to carboxylic acid, the obtained compound 28 still had a good fluorescence (Φ = 31%). When compound 35 was reduced with LiAlH4 a high fluorescent (Φ = 81%) alcohol 29 was obtained in excellent yield, but the λem was reduced to 400 nm.
The same authors [bib_ref] Synthesis and Fluorescent Properties of Aminopyridines and the Application in "Click and..., Li [/bib_ref] , in order to apply the newly synthesized compound as fluorescent probe, used the click reaction to bind the fluorescent aminopyridine scaffold to biomolecules conjugated with alkynyl group (see . Firstly, they synthesized a Boc-NH protected compound 30 accordingly to the above-described method from corresponding starting materials, then the azido substituted derivative 31 was obtained via a Sandmeyer reaction. Finally, the azide 31 react easily with phenylacetylene or propynol (by using catalytic amounts of Cu(II)) to obtain the desired triazole products 32 or 33. Scheme 8. Binding the fluorescent aminopyridine scaffold to biomolecules conjugated with alkynyl group. Thus, when the tertiary butyl group was cleaved using CF3COOH the obtained compound 27 showed no fluorescence. When the ester groups were hydrolyzed to carboxylic acid, the obtained compound 28 still had a good fluorescence (Φ = 31%). When compound 35 was reduced with LiAlH4 a high fluorescent (Φ = 81%) alcohol 29 was obtained in excellent yield, but the λem was reduced to 400 nm.
The same authors [bib_ref] Synthesis and Fluorescent Properties of Aminopyridines and the Application in "Click and..., Li [/bib_ref] , in order to apply the newly synthesized compound as fluorescent probe, used the click reaction to bind the fluorescent aminopyridine scaffold to biomolecules conjugated with alkynyl group (see . Firstly, they synthesized a Boc-NH protected compound 30 accordingly to the above-described method from corresponding starting materials, then the azido substituted derivative 31 was obtained via a Sandmeyer reaction. Finally, the azide 31 react easily with phenylacetylene or propynol (by using catalytic amounts of Cu(II)) to obtain the desired triazole products 32 or 33. The authors successfully applied this "click-and-probing" experiment to bovine serum albumin (BSA) conjugated with a terminal alkyne, demonstrating the biochemical application of this probe and of the fluorescent multisubstituted aminopyridines for detection and analysis.
Piloto et al. [bib_ref] Acridinyl methyl esters as photoactive precursors in the release of neurotransmitteramino acids, Piloto [/bib_ref] studied the use of acridine azaheterocycle as a photochemically removable protecting group for the carboxylic group of amino acids involved in neurotransmision. In this respect, a series of amino acids 34 (glycine, alanine, glutamic acid, β-alanine and γ-aminobutyric acid) protected on amino group with tert-butoxycarbonyl, was treated with 9-bromomethylacridine 35, the corresponding heterocyclic ester 36 being obtained (see . The coupling reaction requires potassium fluoride as base and N,N-dimethylformamide as solvent, and take place at room temperature.
Piloto et al. [bib_ref] Acridinyl methyl esters as photoactive precursors in the release of neurotransmitteramino acids, Piloto [/bib_ref] studied the use of acridine azaheterocycle as a photochemically removable protecting group for the carboxylic group of amino acids involved in neurotransmision. In this respect, a series of amino acids 34 (glycine, alanine, glutamic acid, β-alanine and γ-aminobutyric acid) protected on amino group with tert-butoxycarbonyl, was treated with 9-bromomethylacridine 35, the corresponding heterocyclic ester 36 being obtained (see . The coupling reaction requires potassium fluoride as base and N,Ndimethylformamide as solvent, and take place at room temperature. Scheme 9. The reaction of tert-butoxycarbonyl amino protected amino acids with 9-bromomethylacridine 35.
UV/Vis absorption and emission spectra for the ester conjugates 36a-e were measured on degassed 10 −5 M solutions in two solvent systems: absolute ethanol, and a methanol-HEPES buffer (80:20), the measured parameters are presented in [fig_ref] Table 4: Yields, UV/VIS and fluorescence data for esters 36a-e in absolute ethanol and... [/fig_ref]. λabs-the maximum wavelength of absorption spectra, log ε-logarithm of the molar absorption coefficient λem-the maximum wavelength of photoluminescence spectra, Φfl-photoluminescence quantum yield, calculated using 9,10-diphenylanthracene as a standard (Φfl = 0.95 in ethanol), Δλ-Stokes' shifts (Δλ = λem -λabs).
Due to the heterocyclic chromophore, the absorption maximum of compounds showed no influence of amino acid residue. Ester derivatives 36a-e displayed emission maxima between 411-435 nm in both solvents, with moderate Stokes' shifts (50-75 nm). These esters exhibited higher fluorescence quantum yields in a methanol-HEPES buffer solvent system.
The deprotection of the carboxylic group from the corresponding fluorescent derivatives in methanol-HEPES buffer (80:20) solutions was achieved by irradiation. The best results on photorelease of these amino acids from were obtained on irradiation at 350 nm. This ability of acridinyl methyl ester make them a suitable option for the photochemical Scheme 9. The reaction of tert-butoxycarbonyl amino protected amino acids with 9-bromomethylacridine 35.
UV/Vis absorption and emission spectra for the ester conjugates 36a-e were measured on degassed 10 −5 M solutions in two solvent systems: absolute ethanol, and a methanol-HEPES buffer (80:20), the measured parameters are presented in [fig_ref] Table 4: Yields, UV/VIS and fluorescence data for esters 36a-e in absolute ethanol and... [/fig_ref]. λ abs -the maximum wavelength of absorption spectra, log ε-logarithm of the molar absorption coefficient λ em -the maximum wavelength of photoluminescence spectra, Φ fl -photoluminescence quantum yield, calculated using 9,10-diphenylanthracene as a standard (Φ fl = 0.95 in ethanol), ∆λ-Stokes' shifts (∆λ = λ em − λ abs ).
Due to the heterocyclic chromophore, the absorption maximum of compounds showed no influence of amino acid residue. Ester derivatives 36a-e displayed emission maxima between 411-435 nm in both solvents, with moderate Stokes' shifts (50-75 nm). These esters exhibited higher fluorescence quantum yields in a methanol-HEPES buffer solvent system.
The deprotection of the carboxylic group from the corresponding fluorescent derivatives in methanol-HEPES buffer (80:20) solutions was achieved by irradiation. The best results on photorelease of these amino acids from were obtained on irradiation at 350 nm. This ability of acridinyl methyl ester make them a suitable option for the photochemical release of functional molecules bearing a carboxylic acid group in organic synthesis and in cellular applications.
In order to obtain new luminogens for sensing strong acids Tang et al. [bib_ref] Aromatic azaheterocycle-cored luminogens with tunable physical properties via nitrogen atoms for sensing..., Tang [/bib_ref] used Suzsuki cross-coupling reaction. The luminogens consist of one pyridine, 1,3-diazine, 1,4-diazine, 1,2-diazine and phthalazine moieties as the central cores and two AIE-active tetraphenylethene units in the lateral sides. Aggregation induced emission (AIE) effect exhibiting high emission in concentrated solution or even in the solid state, hold great promise for a luminogen in practical applications.
PY-TPE 39b, PYM-TPE 39c, PYA-TPE 39d, PYD-TPE 39e and PTZ-TPE 39f were easily synthesized by 4-(1,2,2-triphenylvinyl) phenyl boronic acid 37 reaction with 2,5-dibromopyridine 38b, 2,5-dibromopyrimidine 38c, 2,5-dibromopyrazine 38d, 3,6-dibromopyridazine 38e, 1,4dibromophthalazine 38f under Suzuki cross-coupling in high yields (see . For comparison, an analogue molecule BZ-TPE 38a fused with phenyl unit was also prepared.
for a luminogen in practical applications.
PY-TPE 39b, PYM-TPE 39c, PYA-TPE 39d, PYD-TPE 39e and PTZ-TPE 39f were easily synthesized by 4-(1,2,2-triphenylvinyl) phenyl boronic acid 37 reaction with 2,5-dibromopyridine 38b, 2,5-dibromopyrimidine 38c, 2,5-dibromopyrazine 38d, 3,6-dibromopyridazine 38e, 1,4-dibromophthalazine 38f under Suzuki cross-coupling in high yields (see . For comparison, an analogue molecule BZ-TPE 38a fused with phenyl unit was also prepared.
[formula] + B OH OH X 1 X 2 X 3 X 4 Br Br N N Br Br Pd(PPh 3 ) 4 2 N N 39a BZ-TPE (89%) 39b PY-TPE (87%) 39c PYM-TPE (80%) 39d PYA-TPE (79%) 39e PYD-TPE (75%) X 1 X 2 X 3 X 4 39f PTZ-TPE (68%) 37 [/formula]
38a _ e 38f a X 1 = C, X 2 = C, X 3 = C, X 4 =C; b X 1 = N, X 2 = C, X 3 = C, X 4 =C; c X 1 = N, X 2 = C, X 3 = C, X 4 =N; d X 1 = N, X 2 = C, X 3 = N, X 4 =C; e X 1 = N, X 2 = N, X 3 = C, X 4 =C; Scheme 10. Suzuki cross-coupling reaction of 4-(1,2,2-triphenylvinyl) phenyl boronic acid 37 with 2,5-dibromopyridine 38b, 2,5-dibromopyrimidine 38c, 2,5-dibromopyrazine 38d, 3,6-dibromopyridazine 38e, 1,4-dibromophthalazine 38f.
The photophysical properties of the luminogens were studied by UV-vis and fluorescence spectroscopies in CH3CN, and are presented in [fig_ref] Table 5: Photophysical data for the as-prepared luminogens [/fig_ref]. These luminogens show a colorless character in common organic solvents, making them good candidates for colorimetric or fluorometric detection with less self-absorption disturbance. All compounds show very weak fluorescence emission in CH3CN, caused by non-radiative fashion arising from the rotation of tetraphenylethene units. The photophysical properties of the luminogens were studied by UV-vis and fluorescence spectroscopies in CH 3 CN, and are presented in [fig_ref] Table 5: Photophysical data for the as-prepared luminogens [/fig_ref]. These luminogens show a colorless character in common organic solvents, making them good candidates for colorimetric or fluorometric detection with less self-absorption disturbance. All compounds show very weak fluorescence emission in CH 3 CN, caused by non-radiative fashion arising from the rotation of tetraphenylethene units. λ abs -the maximum wavelength of absorption spectra recorded in CH 3 CN (10 −5 M), log ε-logarithm of the molar absorption coefficient λ em -the maximum wavelength of photoluminescence spectra measured in CH 3 CN-water (1:9), Φ fl -photoluminescence quantum yield, τ-fluorescence lifetime, k f -radiation rate constant (
[formula] k f = Φ fl /τ), k nf -non-radiative rate constant (k nf = kf (1 − Φ fl )/Φ fl ). [/formula]
The emission spectra (in CH 3 CN-water) were recorded in order to determine the AIE effect in aggregate states of the as-prepared luminogens. Benzyl-, pyridine-, pyrimidineand pyrazine-tetraphenylethenes (39a-d) exhibited high fluorescence quantum yields in aggregate states due to a typical AIE effect. 1,2-Diazine-tetraphenylethenes (pyridazine-39e and phthalazine-39f) have strong yellow fluorescence emission under protonation enabling AIE performance. The synergetic effect of AIE and the protonation on the 1, 2-diazine segments allows them to detect strong acids with very low pKa values.
The similarities of the 4-hydroxy-1,3-thiazole unit, a chromophore and fluorophore, with naturally occurring luciferin, determine Menzel et al. [bib_ref] Arylamine-Modified Thiazoles as Donor-Acceptor Dyes: Quantum Chemical Evaluation of the Charge-Transfer Process..., Menzel [/bib_ref] to obtain new of arylaminemodified thiazoles as donor-acceptor dyes. The donor part of the dyes was based on the 4-methoxy-1,3-thiazole core substituted with an arylamine (phenyl-, p-anisole-, p-tolyl-, or phenothiazine-) in the 5-position as donor, and a pyridine, pyrimidine, or pyrazine moiety in the 2-position as acceptor.
The synthesis of the 4-hydroxy-1,3-thiazoles 42a-c involved a Hantzsch thiazole cyclizations between the corresponding azaheterocyclic thioamides 40a-c and ethyl 2-bromo-2-(4-nitrophenyl)acetate 41a. Similarly, starting from ethyl 2-bromo-2-(4-bromophenyl)acetate 41b and pyridine-2-carbothioamide 40a, compound 42d was prepared, then methylated via Williamson ether synthesis, with methyl iodide in DMSO. The reduction of the nitro group in 43a-c with freshly prepared Raney nickel and hydrazine in EtOH fallowed (see Scheme 11).
39e and phthalazine-39f) have strong yellow fluorescence emission under protonation enabling AIE performance. The synergetic effect of AIE and the protonation on the 1, 2diazine segments allows them to detect strong acids with very low pKa values.
The similarities of the 4-hydroxy-1,3-thiazole unit, a chromophore and fluorophore, with naturally occurring luciferin, determine Menzel et al. [bib_ref] Arylamine-Modified Thiazoles as Donor-Acceptor Dyes: Quantum Chemical Evaluation of the Charge-Transfer Process..., Menzel [/bib_ref] to obtain new of arylamine-modified thiazoles as donor-acceptor dyes. The donor part of the dyes was based on the 4-methoxy-1,3-thiazole core substituted with an arylamine (phenyl-, p-anisole-, ptolyl-, or phenothiazine-) in the 5-position as donor, and a pyridine, pyrimidine, or pyrazine moiety in the 2-position as acceptor.
The synthesis of the 4-hydroxy-1,3-thiazoles 42a-c involved a Hantzsch thiazole cyclizations between the corresponding azaheterocyclic thioamides 40a-c and ethyl 2bromo-2-(4-nitrophenyl)acetate 41a. Similarly, starting from ethyl 2-bromo-2-(4-bromophenyl)acetate 41b and pyridine-2-carbothioamide 40a, compound 42d was prepared, then methylated via Williamson ether synthesis, with methyl iodide in DMSO. The reduction of the nitro group in 43a-c with freshly prepared Raney nickel and hydrazine in EtOH fallowed (see Scheme 11). Scheme 11. Synthesis of new arylamine-modified thiazoles donor-acceptor dyes.
The reaction condition for the double N-arylation Buchwald-Hartwig cross-coupling reaction were: bis(dibenzylideneacetone)palladium(0) [Pd(dba)2] as the precatalyst, KOtBu as the base, and toluene as solvent. By using the P(tBu)3 reagent, the desired products were obtained in moderate to good yields (76-90 %). The obtaining of both disubstituted A1 and B1, and the monosubstituted products A1m (69%) and B1m (88%), demonstrates the two-step nature of the reaction (see Scheme 12).
## Scheme 11. synthesis of new arylamine-modified thiazoles donor-acceptor dyes.
The reaction condition for the double N-arylation Buchwald-Hartwig cross-coupling reaction were: bis(dibenzylideneacetone)palladium(0) [Pd(dba) 2 ] as the precatalyst, KOtBu as the base, and toluene as solvent. By using the P(tBu) 3 reagent, the desired products were obtained in moderate to good yields (76-90%). The obtaining of both disubstituted A1 and B1, and the monosubstituted products A1m (69%) and B1m (88%), demonstrates the two-step nature of the reaction (see . The Buchwald-Hartwig reaction starting with 43d, a thiazole substituted aryl halide, was not successful under the conditions of the double N-arylation of the amines 44a-c, the biarylphosphane 2-(dicyclohexylphosphanyl)-2′,6′-dimethoxy-1,1′-biphenyl (SPHOS) was required as ligand (see Scheme 13). The Buchwald-Hartwig reaction starting with 43d, a thiazole substituted aryl halide, was not successful under the conditions of the double N-arylation of the amines 44a-c, the biarylphosphane 2-(dicyclohexylphosphanyl)-2′,6′-dimethoxy-1,1′-biphenyl (SPHOS) was required as ligand (see Scheme 13). The arylamine donors has a strong influence on the emission quantum yields recorded in CH3CN solution at room temp, the phenyl-based triarylamines having good quantum yields (40-47%), while the p-anisole-based triarylamines show values below 1% for quantum yields, and consequently no fluorescence for the p-N,N'-dimethylaniline derivative. The obtained photoluminescence quantum yields (see [fig_ref] Table 6: Spectroscopic properties, in CH3CN at room temperature, of the synthesized dyes [/fig_ref] are correlated with the measured emission lifetimes (τ). λabs-the maximum wavelength of absorption, log ε-logarithm of the molar absorption coefficient λem -the maximum wavelength of photoluminescence, Φfl-photoluminescence quantum yield, τ-fluorescence lifetime, n.d-not detected, *-below demand interval, **-measured in THF, sh.-shoulder. The arylamine donors has a strong influence on the emission quantum yields recorded in CH 3 CN solution at room temp, the phenyl-based triarylamines having good quantum yields (40-47%), while the p-anisole-based triarylamines show values below 1% for quantum yields, and consequently no fluorescence for the p-N,N -dimethylaniline derivative. The obtained photoluminescence quantum yields (see [fig_ref] Table 6: Spectroscopic properties, in CH3CN at room temperature, of the synthesized dyes [/fig_ref] are correlated with the measured emission lifetimes (τ). λ abs -the maximum wavelength of absorption, log ε-logarithm of the molar absorption coefficient λ em -the maximum wavelength of photoluminescence, Φ fl -photoluminescence quantum yield, τ-fluorescence lifetime, n.d-not detected, *-below demand interval, **-measured in THF, sh.-shoulder.
Only the emission due to the carbazole moiety, as a main band, can be observed for compound D1 in the polar CH 3 CN, while for D2 and D3 the emissions are modest.
The authors [bib_ref] Arylamine-Modified Thiazoles as Donor-Acceptor Dyes: Quantum Chemical Evaluation of the Charge-Transfer Process..., Menzel [/bib_ref] carried out measurements of the absorptions (and emissions) in different solvents for compounds A2 and D1 in order to investigate them in more detail (see [fig_ref] Table 7: Spectroscopic behavior of dyes A2 and D1 in different solvents [/fig_ref]. The increasing of the solvent polarity induces a weak hypsochromic shift on the absorption maxima, but a bathochromic shift on the emission maxima. This behavior is due to a conformational transformation from a planar locally excited (LE) state to an intramolecular charge-transfer (ICT) or, consequently, to a twisted intramolecular chargetransfer (TICT) state.
Menzel et al. [bib_ref] Arylamine-Modified Thiazoles as Donor-Acceptor Dyes: Quantum Chemical Evaluation of the Charge-Transfer Process..., Menzel [/bib_ref] tested also the ability of these dyes to act as ligands and synthesized seven new ruthenium complexes and measured their emission spectra. The synthesis of the heteroleptic Ru II complexes involve the activation of the cis-(dmbpy) 2 RuCl 2 precursor with AgPF 6 prior to the complexation in acetone with the appropriate ligand (1 equiv.) for 24 h under reflux conditions (see Scheme 14). λ abs -the maximum wavelength of absorption, log ε-logarithm of the molar absorption coefficient λ em -the maximum wavelength of photoluminescence, Φ fl -photoluminescence quantum yield, τ-fluorescence lifetime, *-extinction coefficient could not be measured, due to the poor solubility. λabs-the maximum wavelength of absorption, log ε-logarithm of the molar absorption coefficient λem-the maximum wavelength of photoluminescence, Φfl-photoluminescence quantum yield, τfluorescence lifetime, *-extinction coefficient could not be measured, due to the poor solubility.
Menzel et al. [bib_ref] Arylamine-Modified Thiazoles as Donor-Acceptor Dyes: Quantum Chemical Evaluation of the Charge-Transfer Process..., Menzel [/bib_ref] tested also the ability of these dyes to act as ligands and synthesized seven new ruthenium complexes and measured their emission spectra. The synthesis of the heteroleptic Ru II complexes involve the activation of the cis-(dmbpy)2RuCl2 precursor with AgPF6 prior to the complexation in acetone with the appropriate ligand (1 equiv.) for 24 h under reflux conditions (see . Each complex shows an enhanced absorption in the visible part of the UV/Vis spectrum, due to additional ligand-centered (LC) π-π* transitions and metal-to-ligand chargetransfer (MLCT) transitions originating from 4-methoxy-1,3-thiazole ligands. Each complex shows an enhanced absorption in the visible part of the UV/Vis spectrum, due to additional ligand-centered (LC) π-π* transitions and metal-to-ligand chargetransfer (MLCT) transitions originating from 4-methoxy-1,3-thiazole ligands.
## Fused azaheterocycles
## Fused azaheterocycles
Fused azaheterocycles combine an electron-excessive pyrrole and an electron deficient azine ring which involve an uneven π-electron distribution. This uneven π-electron distribution induces interesting optical properties on the compound with such fused azaheterocycles and make them very attractive materials in optoelectronics [bib_ref] Unusual light-driven amplification through unexpected regioselective photogeneration of five-membered azaheterocyclic AIEgen, Li [/bib_ref]. Moreover, the variety of potential applications in the fields of medicinal of these compounds explains the increased interest of researchers in their study [bib_ref] Synthesis and in vitro analysis of novel dihydroxyacetophenone derivatives with antimicrobial and..., Zbancioc [/bib_ref] [bib_ref] Benzoquinoline Derivatives: A Straightforward and Efficient Route to Antibacterial and Antifungal Agents, Antoci [/bib_ref] [bib_ref] Synthesis, molecular modelling and anticancer evaluation of new pyrrolo[1,2-b]pyridazine and pyrrolo[2,1-a]phthalazinederivatives, Popovici [/bib_ref].
Rajbongshi et al.obtained new chromeno[2,3-b]indoles 47a-c as part of the series of chromenoindole derivatives by reacting 1-(α-amino-α-arylalkyl)-2-naphthol 45 with indole 46 at 100 - C in the presence of catalytic p-toluenesulfonic acid, I 2 and tert-butyl hydroperoxide (TBHP), as is presented in Scheme 15.
In various solvents, the fluorescence spectra of the produced chromeno[2,3-b]indoles 47a-c alter and show a significant Stokes shift, ranging from 250 nm in acetonitrile to 186 nm in ethyl acetate. The chromeno indoles' fluorescence quantum yield was evaluated at 280 nm using tryptophan in water as a reference and at 313 nm using naphthalene. [fig_ref] Table 8: Fluorescence quantum yield [/fig_ref] displays the fluorescence quantum yields of obtained compounds. azaheterocycles and make them very attractive materials in optoelectronics. [bib_ref] Unusual light-driven amplification through unexpected regioselective photogeneration of five-membered azaheterocyclic AIEgen, Li [/bib_ref] Moreover, the variety of potential applications in the fields of medicinal of these compounds explains the increased interest of researchers in their study [bib_ref] Synthesis and in vitro analysis of novel dihydroxyacetophenone derivatives with antimicrobial and..., Zbancioc [/bib_ref] [bib_ref] Benzoquinoline Derivatives: A Straightforward and Efficient Route to Antibacterial and Antifungal Agents, Antoci [/bib_ref] [bib_ref] Synthesis, molecular modelling and anticancer evaluation of new pyrrolo[1,2-b]pyridazine and pyrrolo[2,1-a]phthalazinederivatives, Popovici [/bib_ref].
Rajbongshi et al.obtained new chromeno indoles 47a-c as part of the series of chromenoindole derivatives by reacting 1-(α-amino-α-arylalkyl)-2-naphthol 45 with indole 46 at 100 °C in the presence of catalytic p-toluenesulfonic acid, I2 and tert-butyl hydroperoxide (TBHP), as is presented in Scheme 15.
[formula] Scheme 15. Synthesis of chromeno[2,3-b]indoles 47a-c. [/formula]
In various solvents, the fluorescence spectra of the produced chromeno[2,3-b]indoles 47a-c alter and show a significant Stokes shift, ranging from 250 nm in acetonitrile to 186 nm in ethyl acetate. The chromeno indoles' fluorescence quantum yield was evaluated at 280 nm using tryptophan in water as a reference and at 313 nm using naphthalene. [fig_ref] Table 8: Fluorescence quantum yield [/fig_ref] displays the fluorescence quantum yields of obtained compounds. The investigated chromeno[2,3-b]indoles 47a-c produce fluorescence at a significantly lower energy compared to their absorption, and their fluorescence spectra exhibit a solvatochromic shift, which suggests intramolecular charge transfer in the excited state. The three chromeno indoles' with strong Stokes-shifted fluorescence suggests that they may be used as luminescent solar concentrators or scintillators.
Li et al. [bib_ref] Unusual light-driven amplification through unexpected regioselective photogeneration of five-membered azaheterocyclic AIEgen, Li [/bib_ref] obtained a fused five-membered azaheterocycle with an aggregationinduced emission (AIE) characteristic using an unanticipated regioselective photoreaction.
The starting compound o-TPBQ 48 were easily prepared through a facile one-step synthetic route via a modified Sonogashira coupling reaction according to the previously reported literature. [bib_ref] Time-Dependent Photodynamic Therapy for Multiple Targets: A Highly Efficient AIE-Active Photosensitizer for..., Li [/bib_ref] [bib_ref] Synthesis of Benzoquinolizinium Salts by Rh(III)-Catalyzed Cascade Double N-Annulation Reactions of Allylamines,..., Han [/bib_ref] The authors proposed all of the potential products, including the common six-membered ones, 50 C6-TPBQ' and 51 C6-TPBQ'', as well as the five-membered ring product 49 (C5-TPBQ), although the likelihood of this is very low. They also considered the specific The investigated chromeno[2,3-b]indoles 47a-c produce fluorescence at a significantly lower energy compared to their absorption, and their fluorescence spectra exhibit a solvatochromic shift, which suggests intramolecular charge transfer in the excited state. The three chromeno indoles' with strong Stokes-shifted fluorescence suggests that they may be used as luminescent solar concentrators or scintillators.
Li et al. [bib_ref] Unusual light-driven amplification through unexpected regioselective photogeneration of five-membered azaheterocyclic AIEgen, Li [/bib_ref] obtained a fused five-membered azaheterocycle with an aggregationinduced emission (AIE) characteristic using an unanticipated regioselective photoreaction.
The starting compound o-TPBQ 48 were easily prepared through a facile one-step synthetic route via a modified Sonogashira coupling reaction according to the previously reported literature [bib_ref] Time-Dependent Photodynamic Therapy for Multiple Targets: A Highly Efficient AIE-Active Photosensitizer for..., Li [/bib_ref] [bib_ref] Synthesis of Benzoquinolizinium Salts by Rh(III)-Catalyzed Cascade Double N-Annulation Reactions of Allylamines,..., Han [/bib_ref].
The authors proposed all of the potential products, including the common six-membered ones, 50 C 6 -TPBQ and 51 C 6 -TPBQ", as well as the five-membered ring product 49 (C 5 -TPBQ), although the likelihood of this is very low. They also considered the specific location of the N atom in o-TPBQ and the reported literature on the photoreaction. Surprisingly, the typical six-membered cyclized product (C 6 -TPBQ' or C 6 -TPBQ") was not obtained; instead, only the five-membered cyclized product 49, C 5 -TPBQ, was obtained (see . NMR and high-resolution mass spectroscopies were used to confirm the structure of C 5 -TPBQ. In DMSO/water solutions with various water fractions, the photoluminescence spectra of C5-TPBQ were recorded. This compound has a low fluorescence quantum yield in DMSO solution (1.1%), and when water is gradually added to the solution, an increased photoluminescence signal is obtained. The AIE characteristic was demonstrated by the In DMSO/water solutions with various water fractions, the photoluminescence spectra of C 5 -TPBQ were recorded. This compound has a low fluorescence quantum yield in DMSO solution (1.1%), and when water is gradually added to the solution, an increased photoluminescence signal is obtained. The AIE characteristic was demonstrated by the 110-fold increase in emission intensity in DMSO/water mixes with 99% water compared with DMSO solution.
This research provides a method for quickly and easily creating fused five-membered azaheterocyclic compounds with unique fluorescence properties. These compounds have a wide range of uses in the biological and optoelectronic domains.
Having in view the synthesis of new 1H-pyrazolo quinoxaline (PQX) derivatives with specific photophysical properties as potential materials for optoelectronics, Wojtasik et al.tried to improve the Cadogan synthesis of these compounds. The authors used triphenylphosphine as reducing agent, in dimethylacetamide (DMAc) at 240 - C, under microwave irradiation and obtained PQX derivatives in much shorter time and with better yields of the final product (see . In DMSO/water solutions with various water fractions, the photoluminescence spectra of C5-TPBQ were recorded. This compound has a low fluorescence quantum yield in DMSO solution (1.1%), and when water is gradually added to the solution, an increased photoluminescence signal is obtained. The AIE characteristic was demonstrated by the 110-fold increase in emission intensity in DMSO/water mixes with 99% water compared with DMSO solution.
This research provides a method for quickly and easily creating fused five-membered azaheterocyclic compounds with unique fluorescence properties. These compounds have a wide range of uses in the biological and optoelectronic domains.
Having in view the synthesis of new 1H-pyrazolo quinoxaline (PQX) derivatives with specific photophysical properties as potential materials for optoelectronics, Wojtasik et al.tried to improve the Cadogan synthesis of these compounds. The authors used triphenylphosphine as reducing agent, in dimethylacetamide (DMAc) at 240 °C, under microwave irradiation and obtained PQX derivatives in much shorter time and with better yields of the final product (see .
[formula] Scheme 17. Cadogan synthesis of 1H-pyrazolo[3,4-b]quinoxaline derivatives 53a-g. [/formula]
The substrates 52a-g were obtained by the coupling of the appropriate 2-iodonitrobenzene derivatives 54 with 1,3-disubstituted 5-aminopyrrazole 55 in the presence of palladium catalyst and BINAP (2,2′-bis(diphenylphosphino)-1,1′-binaphtyl), as is presented in Scheme 18. By this reaction, authors obtained the best yields, despite the long reaction time.
[formula] Scheme 17. Cadogan synthesis of 1H-pyrazolo[3,4-b]quinoxaline derivatives 53a-g. [/formula]
The substrates 52a-g were obtained by the coupling of the appropriate 2-iodonitrobenzene derivatives 54 with 1,3-disubstituted 5-aminopyrrazole 55 in the presence of palladium catalyst and BINAP (2,2 -bis(diphenylphosphino)-1,1 -binaphtyl), as is presented in Scheme 18. By this reaction, authors obtained the best yields, despite the long reaction time. The recorded emission spectra of compounds 53e-g reveal an emission maximum between 436 nm and 483 nm, according to the nature of the substituents from the 6th position and the used solvent. These results are presented in the [fig_ref] Table 9: The photophysical constants of the compounds 23e-g in different solvents [/fig_ref]. λabs-spectral position of the first absorption maxima, ε(λabs)-magnitude of the molar absorptivity of λabs, λem-the spectral position of the fluorescence maxima, Φfl-the fluorescence quantum yield, The recorded emission spectra of compounds 53e-g reveal an emission maximum between 436 nm and 483 nm, according to the nature of the substituents from the 6th position and the used solvent. These results are presented in the [fig_ref] Table 9: The photophysical constants of the compounds 23e-g in different solvents [/fig_ref].
The introduction of the electron donor substituent (methyl group-53e) in the 6th position of 1,3-dimethyl-1H-pyrazolo[3,4-b]quinoxaline 53d did not changed the emission properties comparing with the unsubstituted 53d system, while the electron-accepting substituents (chlorine-53f and trifluoromethyl group-53g) shifted the fluorescence band twoards longer wavelenghts by 8 nm to 21 nm, respectively, and induced an increase of the quantum yield of the fluorescence.
In order to develop a new family of thermally activated delayed fluorescence (TADF) emitters, utilizing the donor-acceptor scaffold, Goya et al. [bib_ref] A New Entry to Purely Organic Thermally Activated Delayed Fluorescence Emitters Based..., Goya [/bib_ref] focused on the use of the electron-deficient azaaromatic scaffold in place of the usual dibenzo[a,j]phenazine (DBPHZ) unit. Organic compounds that can display TDAF are very intense studied as emitters for efficient OLEDs, since they can achieve theoretically 100% internal quantum efficiency (IQE) by harvesting electrically generated triplet excitation and convert into the emissive singlet excitations through reverse intersystem crossing (rISC). Pyrido[2,3-b]pyrazine (PYPZ) moiety was selected as the electron-acceptor (A) unit, due to the both pyridine and pyrazine π-deficient heterocycles. When this electron deficient core is connected with an appropriate electron-donor (D), intramolecular charge-transfer (ICT) in the excited state will occur and CT states should become the first singlet excited state (S1). As donor part was selected dihydrophenazasiline (DHPHAzSi) since DHPHAzSi compounds are moderate electron donor. The starting compounds 59 and 61 were prepared according to the literature procedures [bib_ref] A general method for the synthesis of structurally diverse quinoxalines and pyridopyrazine..., Kaur [/bib_ref] , through the condensation between 2,3-diamino-5-bromopyridine with benzyl and phenanthrene-9,10-dione, respectively.
The photophysical properties of compounds 56-58, were investigated in a non-polar polymer matrix Zeonex ® , and in small molecule hosts, 4,4′-bis(N-carbazolyl)-1,1′-biphenyl (CBP) and tris(4-(9H-carbazoyl-9-yl)phenyl)amine (TCTA) and are presented in [fig_ref] Table 10: The photophysical properties of the compounds 56-58 [/fig_ref].. The starting compounds 59 and 61 were prepared according to the literature procedures [bib_ref] A general method for the synthesis of structurally diverse quinoxalines and pyridopyrazine..., Kaur [/bib_ref] , through the condensation between 2,3-diamino-5-bromopyridine with benzyl and phenanthrene-9,10-dione, respectively.
The photophysical properties of compounds 56-58, were investigated in a non-polar polymer matrix Zeonex ® , and in small molecule hosts, 4,4 -bis(N-carbazolyl)-1,1 -biphenyl (CBP) and tris(4-(9H-carbazoyl-9-yl)phenyl)amine (TCTA) and are presented in [fig_ref] Table 10: The photophysical properties of the compounds 56-58 [/fig_ref]. λ em -the maximum wavelength of photoluminescence spectra, Φ fl -photoluminescence quantum yield in degassed, τ PF -prompt fluorescence lifetime, τ DF -delayed fluorescence lifetime, DF/PF-the ratio of delayed fluorescence to prompt fluorescence, EA-activation energy of the triplet to singlet transfer, error ± 0.01 eV, S 1 -singlet energy, error ± 0.03 eV, T 1 -triplet energy, error ± 0.03 eV, ∆E ST -energy splitting, error ± 0.03 eV, Zeonex ® -Cyclo Olefin Polymer is an engineered plastic that provides "glass-like" transparency, low protein absorption, high purity, low water absorption, and excellent moisture barrier. All parameters were estimated at 300 K.
Compounds 56-58 manifest emissions in two different time domains. The first component is the prompt fluorescence (PF) from the singlet excited state (S1), with a lifetime in the nanosecond time delay, and the second components are the delayed emissions from the high triplet excited state (T2) (as theoretical calculations revealed), decaying in micro to millisecond delay time.
Georgescu et al. [bib_ref] Synthesis and fluorescence of 1, Tatu [/bib_ref] [bib_ref] Synthesis and fluorescence of new 3-biphenylpyrrolo[1,2-c]pyrimidines. Arab, Tatu [/bib_ref] studied the synthesis and fluorescence of 3-aryl-7-benzoylpyrrolo pyrimidines in order to determinate the influence of the chemical structure and of solvent polarity on their optical properties.
The authors used, for the synthesis of 3-biphenyl-pyrrolo[1,2-c]pyrimidines, a one-pot, three-component procedure [bib_ref] Synthesis and fluorescence of 1, Tatu [/bib_ref] [bib_ref] Synthesis and fluorescence of new 3-biphenylpyrrolo[1,2-c]pyrimidines. Arab, Tatu [/bib_ref] , which involve a 1,3-dipolar cycloaddition reaction of a 4-biphenyl pyrimidinium-N-ylide with an activated alkyne 65 in 1,2-epoxybutane at reflux. The ylide is generated "in situ" from the corresponding pyrimidinium salts that was formed by the N-alkylation of the pyrimidine 63 with halogenoketone 64. The advantage of performing the reaction in a one-pot three-component approach is the direct formation of the final aromatic compounds 66a-j, avoiding the formation of dipyrimidino-pyrazinic inactivated products (see .
The absorption and emission spectra of compounds 66a-j were recorded in acetonitrile:chloroform (1:1) (3.5 * 10 −6 mol/L) solutions. The fluorescence quantum yield has been calculated [fig_ref] Table 11: The main spectral features of the compounds 4a-j in acetonitrile [/fig_ref] for all compounds 66a-j using Equation (1) using quinine sulphate as standard.
[formula] Φ fl = Φ ref × I A A ref I Ref × n n ref(1) [/formula]
In order to calculate the quantum yield (Φ fl ), according to the Equation (1) Looking at the values from [fig_ref] Table 11: The main spectral features of the compounds 4a-j in acetonitrile [/fig_ref] , it can be seen that two of the compounds (66c p-flouoro substituted and 66f 3,4-dimethoxy substituted) have higher values of quantum yield. These higher values of quantum yield can be explained by a more extended π electron conjugated system in the case of this compound. Ethyl 3-(4-biphenylyl)-7-(3,4dimethoxybenzoyl)pyrrolo[1,2-c]pyrimidine-5-carboxylate was found to have the highest quantum yield value (55%). These higher quantum yield values suggest that the studied compounds are promising candidates for fluorescent chemical sensors. pot, three-component procedure [bib_ref] Synthesis and fluorescence of 1, Tatu [/bib_ref] [bib_ref] Synthesis and fluorescence of new 3-biphenylpyrrolo[1,2-c]pyrimidines. Arab, Tatu [/bib_ref] , which involve a 1,3-dipolar cycloaddition reaction of a 4-biphenyl pyrimidinium-N-ylide with an activated alkyne 65 in 1,2-epoxybutane at reflux. The ylide is generated "in situ" from the corresponding pyrimidinium salts that was formed by the N-alkylation of the pyrimidine 63 with halogenoketone 64. The advantage of performing the reaction in a one-pot three-component approach is the direct formation of the final aromatic compounds 66a-j, avoiding the formation of dipyrimidino-pyrazinic inactivated products (see . The absorption and emission spectra of compounds 66a-j were recorded in acetonitrile:chloroform (1:1) (3.5 * 10 −6 mol/ L) solutions. The fluorescence quantum yield has been calculated [fig_ref] Table 11: The main spectral features of the compounds 4a-j in acetonitrile [/fig_ref] for all compounds 66a-j using Equation (1) using quinine sulphate as standard. λ abs -the maximum wavelength of absorption, ε-the molar absorption coefficient, λ ex -excitation wavelength, λ em -the maximum wavelength of photoluminescence, Φ fl -photoluminescence quantum yield, ∆ν-Stokes shifts.
In view of future exploitation of pyrrolo[1,2-c]pyrimidine compounds in the field of bio-imaging investigations Georgescu et al.synthesized a series of new pyrrolo[1,2c]pyrimidine derivatives and studied their photophysical properties. The synthetic approach was similar with the one described above, but this time did not use the one pot strategy, but only by the 1,3-dipolar cycloaddition of the pyrimidinium ylides generated "in situ" from the corresponding pyrimidinium bromides 67 with the alkyne dipolarophiles 68 in 1,2-epoxybutane as reaction medium and acid scavenger (see . λabs-the maximum wavelength of absorption, ε-the molar absorption coefficient, λex-excitation wavelength, λem-the maximum wavelength of photoluminescence, Φfl-photoluminescence quantum yield, Δν-Stokes shifts.
[formula] Φ = Φ × ×(1) [/formula]
In order to calculate the quantum yield (Φfl), according to the equation (1) we need to determine the maximum value of the absorbance at the emission wavelength λem2, (A), area of the emission peak (I), and refractive index (n) for the solution of investigated compound, and quantum yield (Φref), maximum value of the absorbance at the emission wavelength λem2 (Aref), area of the emission peak (Iref), and refractive index (nref) for the standard solution (quinine sulphate), respectively.
Looking at the values from [fig_ref] Table 11: The main spectral features of the compounds 4a-j in acetonitrile [/fig_ref] , it can be seen that two of the compounds (66c p-flouoro substituted and 66f 3,4-dimethoxy substituted) have higher values of quantum yield. These higher values of quantum yield can be explained by a more extended π electron conjugated system in the case of this compound. Ethyl 3-(4-biphenylyl)-7-(3,4-dimethoxybenzoyl)pyrrolo[1,2-c]pyrimidine-5-carboxylate was found to have the highest quantum yield value (55 %). These higher quantum yield values suggest that the studied compounds are promising candidates for fluorescent chemical sensors.
In view of future exploitation of pyrrolo[1,2-c]pyrimidine compounds in the field of bio-imaging investigations Georgescu et al.synthesized a series of new pyrrolo[1,2c]pyrimidine derivatives and studied their photophysical properties. The synthetic approach was similar with the one described above, but this time did not use the one pot strategy, but only by the 1,3-dipolar cycloaddition of the pyrimidinium ylides generated "in situ" from the corresponding pyrimidinium bromides 67 with the alkyne dipolarophiles 68 in 1,2-epoxybutane as reaction medium and acid scavenger (see Scheme 21). The photophysical properties of fused derivatives 69a-j were investigated by using different sorts of solvents (see [fig_ref] Table 2: The photophysical properties for the obtained molecules 10 and 11 [/fig_ref]. In chloroform, compounds derived from the ethyl propiolate, 69a-d, present better fluorescence yield than the compounds derived from the symmetric alkynes 69e-j. The excitation, during the recording of the emission spectra, was The photophysical properties of fused derivatives 69a-j were investigated by using different sorts of solvents (see [fig_ref] Table 2: The photophysical properties for the obtained molecules 10 and 11 [/fig_ref]. In chloroform, compounds derived from the ethyl propiolate, 69a-d, present better fluorescence yield than the compounds derived from the symmetric alkynes 69e-j. The excitation, during the recording of the emission spectra, was performed at λ abs2 at which a higher intensity of the fluorescence was obtained. A main emission band is located in the blue region of the visible spectrum in 437-463 nm region. λ abs -the maximum wavelength of absorption, ε-the molar absorption coefficient, λ ex -excitation wavelenght, λ em -the maximum wavelength of photoluminescence, Φ fl -photoluminescence quantum yield calculated using quinine sulphate standard, ∆ν-Stokes shifts.
The authorsinvestigate the influence of all R 1 , R 2 , R 3 and R 4 substituents on the fluorescence properties of the synthesized compounds and concluded that a substituent that determine the growth of conjugation inside the pyrrolo[1,2-c]pyrimidine fragment lead to an increase of the fluorescence, while a substituent that determine the decrease of conjugation inside the pyrrolo[1,2-c]pyrimidine fragment lead to decrease of the fluorescence.
Tomashenko et al. [bib_ref] A new heterocyclic skeleton with highly tunable absorption/emission wavelength via H-bonding, Tomashenko [/bib_ref] obtained new pyrido[2,1-a]pyrrolo[3,2-c]isoquinoline 76a-c heterocyclic system by using Pd to catalyze the intramolecular cyclization of 1-[1-benzyl-2-(2-bromophenyl)-1H-pyrrol-3-yl]pyridin-1-ium bromides 74a-c (see .
The obtained compounds 76a-c have fluorescent properties in solutions, with quantum yields for 76c in methanol reaching 81%. Photophysical data were studied in toluene, acetonitrile, dichloromethane (DCM) and methanol, respectively [fig_ref] Table 3: Fluorescent properties of aminopyridines [/fig_ref]. The obtained compounds display moderate to strong fluorescence in solution, giving minor variations in emission maxima position with changes in solvent polarity. The biggest effect was shown in methanol solutions and upon addition of proton donors to aprotic solvents. [fig_ref] Table 3: Fluorescent properties of aminopyridines [/fig_ref]. Photophysical properties of compounds 76a-c in toluene, acetonitrile, dichloromethane (DCM) and methanol solutions at room temperature, λ ex = 420 nm. Lifetimes (τ) were determined at λ max of the emission bands. The obtained compounds 76a-c have fluorescent properties in solutions, with quantum yields for 76c in methanol reaching 81%. Photophysical data were studied in toluene, acetonitrile, dichloromethane (DCM) and methanol, respectively [fig_ref] Table 3: Fluorescent properties of aminopyridines [/fig_ref]. The obtained compounds display moderate to strong fluorescence in solution, giving minor variations in emission maxima position with changes in solvent polarity. The biggest effect was shown in methanol solutions and upon addition of proton donors to aprotic solvents.
## Compound
## Our recent contribution to the field
Nitrogen ylide chemistry is a traditional research area in our group, professors Zugravescu and Petrovanu being the pioneer of this field. Several compounds with complex azaheterocyclic skeleton were synthesized through the N-ylides. Many of these compounds present practical importance either to their biological activities, having antimicrobial [bib_ref] Synthesis and in vitro analysis of novel dihydroxyacetophenone derivatives with antimicrobial and..., Zbancioc [/bib_ref] [bib_ref] Benzoquinoline Derivatives: A Straightforward and Efficient Route to Antibacterial and Antifungal Agents, Antoci [/bib_ref] , anticancer [bib_ref] Synthesis, molecular modelling and anticancer evaluation of new pyrrolo[1,2-b]pyridazine and pyrrolo[2,1-a]phthalazinederivatives, Popovici [/bib_ref] [bib_ref] Mangalagiu, I.I. Design, synthesis and in vitro anticancer activity of a new..., Zbancioc [/bib_ref] [bib_ref] Hybrid Imidazole-Pyridine Derivatives: An Approach to Novel Anticancer DNA Intercalators, Lungu [/bib_ref] [bib_ref] Microwave assisted synthesis of six member ring azaheterocycles and their antimycobacterial and..., Matarneh [/bib_ref] , antitubercular [bib_ref] Microwave assisted synthesis of six member ring azaheterocycles and their antimycobacterial and..., Matarneh [/bib_ref] [bib_ref] Mangalagiu, I.I. Design, synthesis and antituberculosis activity of some new pyridazine derivatives:..., Mantu [/bib_ref] [bib_ref] Mangalagiu, I.I. Bis-(imidazole/benzimidazole)-pyridine derivatives: Synthesis, structure and antimycobacterial activity. Part XII, Antoci [/bib_ref] [bib_ref] Synthesis and antituberculosis activity of some new pyridazine derivatives. Part II, Mantu [/bib_ref] or anti-leishmaniasis [bib_ref] New Molecules with Azaheterocycles Skeleton of Potential Interest in Leishmaniasis, Mangalagiu [/bib_ref] activity, or to their optical properties (some of this compound presenting intense blue fluorescence) [bib_ref] Microwave-Assisted Synthesis of Highly Fluorescent Pyrrolopyridazine Derivatives, Zbancioc [/bib_ref] [bib_ref] Pyrrolodiazine derivatives as blue organic luminophores: Synthesis and properties, Zbancioc [/bib_ref] [bib_ref] Conformational effects on the lowest excited states of benzoylpyrrolopyridazine: Insights from PCM..., Maftei [/bib_ref] [bib_ref] Pyrrolopyridazine derivatives as blue organic luminophores: Synthesis and properties. Part 2, Zbancioc [/bib_ref] [bib_ref] Microwave Assisted Reactions of Fluorescent Pyrrolodiazine Building Blocks, Moldoveanu [/bib_ref] [bib_ref] Fluorescent Azasteroids through Ultrasound Assisted Cycloaddition Reactions, Moldoveanu [/bib_ref] [bib_ref] Ultrasound-Assisted Synthesis of Fluorescent Azatetracyclic Derivatives: An Energy-Efficient Approach, Zbancioc [/bib_ref]. Here we will present our recent (last 12 years) achievements in the synthesis of fluorescent azaheterocyclic compound.
In one extensive work [bib_ref] Microwave-Assisted Synthesis of Highly Fluorescent Pyrrolopyridazine Derivatives, Zbancioc [/bib_ref] [bib_ref] Pyrrolodiazine derivatives as blue organic luminophores: Synthesis and properties, Zbancioc [/bib_ref] , we carried out a comprehensive investigation into the synthesis of fluorescent pyrrolodiazine (PD). All PD derivatives 78-82 were produced using a methodology which involve two steps. In the first stage, diazinium salts 78a-h were obtained by quaternization of diazine [pyridazine (PY) or phthalazine (PH)] with halogenated derivatives with increased reactivity. In the 2nd stage, a typical Huisgen [3+2] dipolar cycloaddition of diazinium ylides 78 to the corresponding dipolarophiles were performed (see .
In one extensive work, [bib_ref] Microwave-Assisted Synthesis of Highly Fluorescent Pyrrolopyridazine Derivatives, Zbancioc [/bib_ref] [bib_ref] Pyrrolodiazine derivatives as blue organic luminophores: Synthesis and properties, Zbancioc [/bib_ref] we carried out a comprehensive investigation into the synthesis of fluorescent pyrrolodiazine (PD). All PD derivatives 78-82 were produced using a methodology which involve two steps. In the first stage, diazinium salts 78a-h were obtained by quaternization of diazine [pyridazine (PY) or phthalazine (PH)] with halogenated derivatives with increased reactivity. In the 2nd stage, a typical Huisgen [3+2] dipolar cycloaddition of diazinium ylides 78′ to the corresponding dipolarophiles were performed (see The synthesis of the desired pyrrolodiazine derivatives 79-82 was performed under conventional thermal heating (TH) and using MW irradiation, by a typical Huisgen [3+2] dipolar cycloaddition. The reactions under MW irradiation have shown some important advantages such as significantly higher yields, decreasing the reaction time from hours to 5 min and five times less amount of solvent is needed.
To obtain a perspective on the effect that the C-7 substituent of the pyrrolodiazine skeleton the photoluminescent properties of the obtained compounds were measured in different solvents and are presented in [fig_ref] Table 4: Yields, UV/VIS and fluorescence data for esters 36a-e in absolute ethanol and... [/fig_ref].
While the partially saturated dihydro-pyrrolo-PY (82a-d) are red-shifted and have a poor or moderate quantum yield (around 5-40%), the tetrahydro-pyrrolodiazine (81b,c) have a negligible quantum yield (less than 5%). The fully aromatized pyrrolo-PY (79a-c and 80a-c) are highly powerful blue emitters with extremely high quantum yields (up to 90%). The highly blue emitters of completely aromatized pyrrolo-PH (79h and 80g-i) exhibit a low quantum yield of less than 10% and an unusual, blue-shifted absorption with λ max of absorption around 314-322 nm. Regarding the effect of the substituent from the 7th position of the pyrrolodiazine skeleton, the pyrrolo-PY compounds exhibit intense blue fluorescence and have a very high quantum yield when the substituent is an ester or amide group, whereas the quantum yield is negligibly low when the substituent is a ketone (see . We confirmed the experimental results through theoretical calculations into another study from our group, [bib_ref] Conformational effects on the lowest excited states of benzoylpyrrolopyridazine: Insights from PCM..., Maftei [/bib_ref]. According to computational study, the first nπ* state may be significantly stabilized when the carbonyl group from the 7th position of the pyrrolo-PY moiety is oriented so that it faces the diazine nitrogen. Given the lower (fluorescent) ππ* state expected for all conformations, such an effect does not seem to affect the fluorescence of the ester derivative. On the other hand, we found that the same nπ* state is predicted to have a lower energy than the first ππ* state for the most stable conformation of the benzoyl-substituted derivative, which is consistent with the weak fluorescence that was experimentally seen.
We demonstrate how the highly fluorescent ester-substituted pyrrolopyridazine's solvent-dependent photophysical characteristics can be accurately predicted by TD-DFT in SS-PCM solvation when hybrid functionals such as B3LYP or PBE0 are used, together with a reasonable basis set.
In order to increase the fluorescent properties of the pyrrolo-PY derivatives, we studied the influence of the substituent from phenyl ring on the 2 nd position of pyrrolo-PY skeleton, [bib_ref] Pyrrolopyridazine derivatives as blue organic luminophores: Synthesis and properties. Part 2, Zbancioc [/bib_ref] and used the strategies adopted for construction of fluorescent pyrrolo-PY derivatives are similar to those presented in the previous study [bib_ref] Pyrrolodiazine derivatives as blue organic luminophores: Synthesis and properties, Zbancioc [/bib_ref] (see . We confirmed the experimental results through theoretical calculations into another study from our group [bib_ref] Conformational effects on the lowest excited states of benzoylpyrrolopyridazine: Insights from PCM..., Maftei [/bib_ref]. According to computational study, the first nπ* state may be significantly stabilized when the carbonyl group from the 7th position of the pyrrolo-PY moiety is oriented so that it faces the diazine nitrogen. Given the lower (fluorescent) ππ* state expected for all conformations, such an effect does not seem to affect the fluorescence of the ester derivative. On the other hand, we found that the same nπ* state is predicted to have a lower energy than the first ππ* state for the most stable conformation of the benzoyl-substituted derivative, which is consistent with the weak fluorescence that was experimentally seen.
We demonstrate how the highly fluorescent ester-substituted pyrrolopyridazine's solvent-dependent photophysical characteristics can be accurately predicted by TD-DFT in SS-PCM solvation when hybrid functionals such as B3LYP or PBE0 are used, together with a reasonable basis set.
In order to increase the fluorescent properties of the pyrrolo-PY derivatives, we studied the influence of the substituent from phenyl ring on the 2nd position of pyrrolo-PY skeleton [bib_ref] Pyrrolopyridazine derivatives as blue organic luminophores: Synthesis and properties. Part 2, Zbancioc [/bib_ref] , and used the strategies adopted for construction of fluorescent pyrrolo-PY derivatives are similar to those presented in the previous study [bib_ref] Pyrrolodiazine derivatives as blue organic luminophores: Synthesis and properties, Zbancioc [/bib_ref] (see . solvent-dependent photophysical characteristics can be accurately predicted by TD-DFT in SS-PCM solvation when hybrid functionals such as B3LYP or PBE0 are used, together with a reasonable basis set.
In order to increase the fluorescent properties of the pyrrolo-PY derivatives, we studied the influence of the substituent from phenyl ring on the 2 nd position of pyrrolo-PY skeleton, [bib_ref] Pyrrolopyridazine derivatives as blue organic luminophores: Synthesis and properties. Part 2, Zbancioc [/bib_ref] and used the strategies adopted for construction of fluorescent pyrrolo-PY derivatives are similar to those presented in the previous study [bib_ref] Pyrrolodiazine derivatives as blue organic luminophores: Synthesis and properties, Zbancioc [/bib_ref] (see . For the obtained compounds 86 and 87, we investigated the photoluminescent properties (see [fig_ref] Table 5: Photophysical data for the as-prepared luminogens [/fig_ref]. The pyrrolo-PY substituents had some effect on the absorbance and fluorescence properties. Thus, the cycloadducts with carboethoxy groups, in the seventh position have a more fluorescence than those with a carbomethoxy groups, in the same position and also the cycloadducts with a para-chlorophenyl substituent in the second position of pyrrolo-PY moiety have a more fluorescence than those para-bromophenyl substituted.
Continuing our studies [bib_ref] Microwave Assisted Reactions of Fluorescent Pyrrolodiazine Building Blocks, Moldoveanu [/bib_ref] , we synthesized a new class of fused pyrrolodiazines 91-95 in order to obtain new fluorescent compounds by the same strategy-obtaining of the N-ylides fallowed by their cycloaddition with activated alkynes (see . The reactions were conducted both under conventional TH and MW irradiation. The reactions under MW irradiation have shown a series of advantages such as decreasing the reaction time from 6 hours to 10 minutes and slightly higher yields. After the synthesis of the desired pyrrolodiazines, we investigated their photoluminescent properties in non-polar solvents cyclohexane and dichloromethane. The measured parameters are presented in [fig_ref] Table 6: Spectroscopic properties, in CH3CN at room temperature, of the synthesized dyes [/fig_ref]. A relationship between structure of cycloadducts and fluorescence quantum yields was founded. Thus, the compounds with pyrrolo-pyridazine skeleton showed good quantum yield (around 25%), while compounds with pyrrolophthalazine skeleton showed a lower quantum yield.
In the next step we obtained some new bromoderivatives with increased reactivity by bromination of fluorescent pyrrolodiazines in heterogeneous catalysis in the presence After the synthesis of the desired pyrrolodiazines, we investigated their photoluminescent properties in non-polar solvents cyclohexane and dichloromethane. The measured parameters are presented in [fig_ref] Table 6: Spectroscopic properties, in CH3CN at room temperature, of the synthesized dyes [/fig_ref]. A relationship between structure of cycloadducts and fluorescence quantum yields was founded. Thus, the compounds with pyrrolo-pyridazine skeleton showed good quantum yield (around 25%), while compounds with pyrrolophthalazine skeleton showed a lower quantum yield.
In the next step we obtained some new bromoderivatives with increased reactivity by bromination of fluorescent pyrrolodiazines in heterogeneous catalysis in the presence of copper (II) bromide (see . These bromo-derivatives can fluorescently label biological macromolecules due to their increased reactivity, being easily incorporated in those bio-macromolecules. Having in view the results of the previous studies, we prepared a new family of pyrrolobenzo[f]quinoline derivatives [bib_ref] Fluorescent Azasteroids through Ultrasound Assisted Cycloaddition Reactions, Moldoveanu [/bib_ref] in order to obtain blue fluorescent azaheterocyclic derivatives. The pyrrolobenzo[f]quinolines 102a-c, 103a-c and 104c were obtained by the same setup. In the first stage, benzo[f]quinolinium salts 100a-c were obtained by quaternization of benzo[f]quinoline with bromoketones. In the 2nd stage, a typical Huisgen [3+2] dipolar cycloaddition of benzo[f]quinolinium ylides 101a-c to the alkyne as dipolarophiles was performed (see . A comparative study of these reactions was carried out under conventional thermal heating and under ultrasound (US) irradiation. The reactions under US irradiation have shown a series of advantages such as decreasing the reaction time from 2 days to 2 hours and slightly higher yields. Having in view the results of the previous studies, we prepared a new family of pyrrolobenzo[f ]quinoline derivatives [bib_ref] Fluorescent Azasteroids through Ultrasound Assisted Cycloaddition Reactions, Moldoveanu [/bib_ref] in order to obtain blue fluorescent azaheterocyclic derivatives. The pyrrolobenzo[f ]quinolines 102a-c, 103a-c and 104c were obtained by the same setup. In the first stage, benzo[f ]quinolinium salts 100a-c were obtained by quaternization of benzo[f ]quinoline with bromoketones. In the 2nd stage, a typical Huisgen [3+2] dipolar cycloaddition of benzo[f ]quinolinium ylides 101a-c to the alkyne as dipolarophiles was performed (see . A comparative study of these reactions was carried out under conventional thermal heating and under ultrasound (US) irradiation. The reactions under US irradiation have shown a series of advantages such as decreasing the reaction time from 2 days to 2 hours and slightly higher yields.
The photophysical properties of azatetracyclic derivatives 102a-c, 103a-c and 104c were studied in cyclohexane and trichloromethane, respectively, and are presented in [fig_ref] Table 7: Spectroscopic behavior of dyes A2 and D1 in different solvents [/fig_ref]. [fig_ref] Table 7: Spectroscopic behavior of dyes A2 and D1 in different solvents [/fig_ref]. λ max (nm) of absorption spectra and λ max (nm) of emission spectra of compounds 102a-c, 103a-c and 104c.
## Compound
Fluorescence (λ max , nm) Absorption (λ max , nm) The obtained cycloadducts are blue emitters having λ max of fluorescence around 430-450 nm. The cycloadducts with similar structure 102a,b and 103a,b present small differences in their electronic spectra, while the cycloadducts with bulky pivaloyl group 102c and 103c or without keto group in the 3rd position 104c shows more intense fluorescence emission.
derivatives. [fig_ref] Scheme 1: Synthetic procedure used for the obtaining of indolylcarbazole derivatives 4 [/fig_ref] were obtained by the same setup. In the first stage, benzo[f]quinolinium salts 100a-c were obtained by quaternization of benzo[f]quinoline with bromoketones. In the 2nd stage, a typical Huisgen [3+2] dipolar cycloaddition of benzo[f]quinolinium ylides 101a-c to the alkyne as dipolarophiles was performed (see . A comparative study of these reactions was carried out under conventional thermal heating and under ultrasound (US) irradiation. The reactions under US irradiation have shown a series of advantages such as decreasing the reaction time from 2 days to 2 hours and slightly higher yields.
## Scheme 28. synthesis of blue fluorescent azaheterocyclic derivatives 102-104 by huisgen [3+2] dipolar cycloaddition.
As a continuation of this work, in another study [bib_ref] Ultrasound-Assisted Synthesis of Fluorescent Azatetracyclic Derivatives: An Energy-Efficient Approach, Zbancioc [/bib_ref] we derivatized the previously obtained cycloadducts 102a,b and 103a,b in order to use this blue fluorescent azaheterocyclic derivatives as fluorescent biomarkers. In the first step, we applied this functionalization to the tetracyclic cycloadduct products 102a,b and 103a,b by the bromination in heterogeneous catalysis obtaining a mixture of mono or dibrominated product type 105a-d and 106a-d (see . This bromo-derivatives can fluorescently label biological macromolecules (peptides, proteins, DNA) due to their increased reactivity, being easily incorporated in those bio-macromolecules. For the new fluorophores [fig_ref] Scheme 1: Synthetic procedure used for the obtaining of indolylcarbazole derivatives 4 [/fig_ref] we investigated optical properties on chloroform diluted solutions (see [fig_ref] Table 8: Fluorescence quantum yield [/fig_ref]. [fig_ref] Table 8: Fluorescence quantum yield [/fig_ref]. λmax (nm) of absorption spectra, λmax (nm) of emission spectra and fluorescence quantum yield (%) of compounds 105a-d and 107a-d.
## Scheme 29. derivatization of blue fluorescent azaheterocyclic derivatives 102-103.
In the next step, by a nucleophilic substitution, the monobrominated cycloadducts 105a-d were used as starting materials for the synthesis of corresponding azides 107a-d.
The reaction was carried out in tetrabutylammonium bromide (TBAB) as phase transfer catalyst and in a chloroform/water mixture (see .
For the new fluorophores [fig_ref] Scheme 1: Synthetic procedure used for the obtaining of indolylcarbazole derivatives 4 [/fig_ref] we investigated optical properties on chloroform diluted solutions (see [fig_ref] Table 8: Fluorescence quantum yield [/fig_ref]. [fig_ref] Table 8: Fluorescence quantum yield [/fig_ref]. λ max (nm) of absorption spectra, λ max (nm) of emission spectra and fluorescence quantum yield (%) of compounds 105a-d and 107a-d.
## Compound
Fluorescence (λ max , nm) When comparing the fluorescence intensity of the bromides 105a-d to the fluorescence of the azides 107a-d, we observed a decrease in fluorescence, due the alteration of the planar structure of the molecule. The same effect is highlighted when comparing the fluorescence intensity of the underivatized cycloadducts 102a,b and 103a,b with the fluorescence intensity of the derivatized azatetracyclic bromides 105a-d or azides 107a-d.
Another research field of interest to us, is of the compounds which contain a pyrrolo[2,1-a]isoquinoline or imidazo[2,1-a]isoquinoline framework, due to their potential biological activity, but also to their extended π-conjugated systems that make them excellent candidates for photophysical applications [bib_ref] Synthesis and photophysical insights of new fused N-heterocyclic derivatives with isoquinoline skeleton, Gherasim [/bib_ref]. Thus, we synthesized several new pyrrolo [2,1-a]isoquinolines and imidazo[2,1-a]isoquinolines in order to realize a complete study regarding their photophysical properties and potential applications in the field.
The chosen method for the assembly of fused target polyheterocycles relied on 1,3-dipolar cycloaddition of different isoquinolinium ylides 111a-c (in situ generated in basic medium from corresponding salts 110a-c, which were obtained by isoquinoline 108 alkylation with halides 109a-c) to ethyl propiolate or ethyl cyanoformate. The intermediate dihydropyrrolo isoquinolines 112 a-c underwent oxidative dehydrogenation under atmospheric conditions, yielding the final compounds 112a-c in good yields (63-80%). Using ethyl cyanoformate as dipolarophile in similar conditions, we obtained imidazo[2,1-a]isoquinolines 113a-c presumably via dihydroderivatives 113 a-c (see .
The electronic absorption and emission spectra of the obtained azaheterocycles were recorded in dichlorometane (DCM) and dimethylsulphoxide (DMSO), and are presented in [fig_ref] Table 9: The photophysical constants of the compounds 23e-g in different solvents [/fig_ref]. The substituents on the 1st and 3rd positions of the pyrrole ring (COOMe, COOEt) have a low influence on the spectral pattern of pyrroloisoquinolines 112b and 112c. In addition, the solvent polarities have little to no influence on the electronic absorption spectra of pyrrolo-and imidazo-isoquinolines, suggesting that the ground state of these derivatives is not influenced by the solvent polarity. The extended π-π* conjugation in the imidazoisoquinolines system due to the substituents from the 1st and 3rd positions of the pyrrole ring induced a bathochromic shift of the absorption maxima. In DMSO, in the case of pyrroloisoquinolines, a hypsochromic shift of the absorption band occurs. dihydropyrrolo[2,1-a]isoquinolines 112′a-c underwent oxidative dehydrogenation under atmospheric conditions, yielding the final compounds 112a-c in good yields (63-80%). Using ethyl cyanoformate as dipolarophile in similar conditions, we obtained imidazo[2,1-a]isoquinolines 113a-c presumably via dihydroderivatives 113′a-c (see . The electronic absorption and emission spectra of the obtained azaheterocycles were recorded in dichlorometane (DCM) and dimethylsulphoxide (DMSO), and are presented in [fig_ref] Table 9: The photophysical constants of the compounds 23e-g in different solvents [/fig_ref]. The substituents on the 1 st and 3 rd positions of the pyrrole ring (COOMe, COOEt) have a low influence on the spectral pattern of pyrroloisoquinolines 112b and 112c. In addition, the solvent polarities have little to no influence on the electronic absorption spectra of pyrrolo-and imidazo-isoquinolines, suggesting that the ground state of these derivatives is not influenced by the solvent polarity. The extended π-π* conjugation in the imidazoisoquinolines system due to the substituents from the 1st and 3rd positions of the pyrrole ring induced a bathochromic shift of the absorption maxima. In DMSO, in the case of pyrroloisoquinolines, a hypsochromic shift of the absorption band occurs. λabs-the maximum wavelength of absorption spectra, λem-the maximum wavelength of photoluminescence spectra, Φfl-photoluminescence quantum yield, τ-fluorescence lifetime, sh-shoulder.
Scheme 30. Synthesis of fused target polyheterocycles 112-113 by 1,3-dipolar cycloaddition. λ abs -the maximum wavelength of absorption spectra, λ em -the maximum wavelength of photoluminescence spectra, Φ fl -photoluminescence quantum yield, τ-fluorescence lifetime, sh-shoulder.
In the case of the compound 112a containing CN group on the pyrrole ring, a similar behaviour of the emission spectra to absorption spectra was observed, namely a hypsochromic shift in both DCM and DMSO solvents, and no influence of the solvent polarity on the position of the emission maxima. The substituents on the pyrrole ring have no influence on the position of emission bands for derivatives 113a-c, excepting on the compound 113a which displays a hypsochromic shift in dichloromethane. The emission bands of imidazoisoquinolines 113a-c are hypsochromic shifted in DCM and DMSO comparing with the emission bands of pyrroloisoquinoline derivatives 112a-c. Time-correlated single photon counting (TCSPC) technique was used to determine the luminescence lifetimes. In DMSO, over the entire emission range, a double-exponential function describes better the emission decays for all investigated compounds. In DMSO, pyrroloisoquinolines 112a-c have longer fluorescence lifetime τ1 than those of imidazoisoquinolines 113a-c. The lifetimes of the excited-state for all compounds are in the nanosecond timescale.
Indolizine derivatives with phenanthroline skeleton [bib_ref] Steady state and time resolved fluorescence studies of new indolizine with phenanthroline..., Airinei [/bib_ref] [bib_ref] Synthesis and properties of new fused pyrrolo-1,10-phenanthroline type derivatives, Matarneh [/bib_ref] were another research field in our group due to both their potential biological activities and their extended π-electron system which make them valuable materials in the construction of new optoelectronic devices. Compounds 114a-g and 115a-e were synthesized in our group using the same 3+2 dipolar cycloaddition strategy of cycloimmonium ylides to activated alkynes (see Scheme 31) [bib_ref] Steady state and time resolved fluorescence studies of new indolizine with phenanthroline..., Airinei [/bib_ref]. In this case 4,7-phenanthrolin-4-ium and 1,7-phenanthrolin-7-ium ylides, generated from the corresponding monoquaternary 116 and 117 salts, were added to ethyl propiolate or dimethyl acetylene dicarboxylate DMAD. field in our group due to both their potential biological activities and their extended πelectron system which make them valuable materials in the construction of new optoelectronic devices. Compounds 114a-g and 115a-e were synthesized in our group using the same 3+2 dipolar cycloaddition strategy of cycloimmonium ylides to activated alkynes (see . [bib_ref] Steady state and time resolved fluorescence studies of new indolizine with phenanthroline..., Airinei [/bib_ref] In this case 4,7-phenanthrolin-4-ium and 1,7-phenanthrolin-7-ium ylides, generated from the corresponding monoquaternary 116 and 117 salts, were added to ethyl propiolate or dimethyl acetylene dicarboxylate DMAD. The absorption spectra of all the compounds present the fine structure specific to phenanthrene spectrum. The extended π-system of the 1,7-phenanthroline compounds explained the bathochromic shift to longer wavelength absorption band as compared to the UV-Vis absorption spectra of 4,7-phenanthroline derivatives. The position of the absorption and emission maxima of phenanthroline derivatives are highly influenced by the substituents on the pyrrole ring, as can be observed from [fig_ref] Table 2: The photophysical properties for the obtained molecules 10 and 11 [/fig_ref] The absorption spectra of all the compounds present the fine structure specific to phenanthrene spectrum. The extended π-system of the 1,7-phenanthroline compounds explained the bathochromic shift to longer wavelength absorption band as compared to the UV-Vis absorption spectra of 4,7-phenanthroline derivatives. The position of the absorption and emission maxima of phenanthroline derivatives are highly influenced by the substituents on the pyrrole ring, as can be observed from [fig_ref] Table 2: The photophysical properties for the obtained molecules 10 and 11 [/fig_ref]. The introduction of CN, COOMe and COOEt groups in the 9th and 7th positions of the pyrrole ring determine a hypsochromic shift of the absorption and emission maxima of phenanthroline derivatives [fig_ref] Scheme 1: Synthetic procedure used for the obtaining of indolylcarbazole derivatives 4 [/fig_ref]. The same effect on the absorption and fluorescence spectra is obtained when a third substituent (COOMe) is introduced in the 8th position of the pyrrole ring, only in the case of the compound 114e. Φfl-photoluminescence quantum yield, τ-fluorescence lifetime, f-the weighted contribution.
The phenanthroline derivatives displayed broad and structureless emission band in 425 to 480 nm depending on the substituent nature at the pyrrole moiety. For 1,7-phenanthroline derivatives the emission band vas found in the 440-450 nm region, while for 4,7phenanthrolines the emission band was found in 425-480 nm region. The position and the nature of the substituent at the pyrrole ring have a great influence on the fluorescence quantum yield of pyrrolo-phenanthroline derivatives. Scheme 32 illustrates the effect of substituents from the pyrrole ring on the fluorescence yield of pyrrolophenanthroline derivatives. The fluorescence emission of disubstituted phenanthroline derivatives (114d, 115c, 115d) is practically quenched by the presence of halogens (Cl, Br) in the para position of phenacyl group (COC 6 H 4 ) from 9th position of the pyrrole ring. The emission quantum yield decreases significantly (115e−0.044) when COC 6 H 4 OMe-(p) group is introduced in the 9th position of the pyrrole ring due to the withdrawing effect of the methoxy group which determines a decrease in conjugation in the 1,7-phenanthroline system. Time-correlated single photon counting method in dichloromethane was used for the fluorescence lifetimes (τ) estimation.
The obtaining of small organic molecules capable to be incorporated in specific biomolecules or used in biomedical optical imaging was another research field in our group. A promising scaffold able to fulfil this purpose could be the bipyridyl. Bipyridyl, having two heterocyclic cores, can involve one or both of them in the formation of ylides and/or cycloaddition products. Thus, for these compounds we can obtain mono-salts [bib_ref] Pyridyl-indolizine derivatives as DNA binders and pH-sensitive fluorescent dyes, Marangoci [/bib_ref] , monoylides [bib_ref] Pyridyl-indolizine derivatives as DNA binders and pH-sensitive fluorescent dyes, Marangoci [/bib_ref] , mono-cycloadducts (indolizines) [bib_ref] Pyridyl-indolizine derivatives as DNA binders and pH-sensitive fluorescent dyes, Marangoci [/bib_ref] , mono-indolizine mono-salts [bib_ref] Fluorescent Conjugates: pH Stability, Dye-DNA Interaction and Biological Activity, Gradinaru [/bib_ref] [bib_ref] Novel cyclodextrin-based pH-sensitive supramolecular host-guest assembly for staining acidic cellular organelles, Pricope [/bib_ref] [bib_ref] Cyclodextrin Encapsulated pH Sensitive Dyes as Fluorescent Cellular Probes: Self-Aggregation and In..., Sardaru [/bib_ref] , mono-indolizine mono-ylides [bib_ref] Fluorescent Conjugates: pH Stability, Dye-DNA Interaction and Biological Activity, Gradinaru [/bib_ref] [bib_ref] Novel cyclodextrin-based pH-sensitive supramolecular host-guest assembly for staining acidic cellular organelles, Pricope [/bib_ref] [bib_ref] Cyclodextrin Encapsulated pH Sensitive Dyes as Fluorescent Cellular Probes: Self-Aggregation and In..., Sardaru [/bib_ref] , and theoretically bis-indolizines.
In one of our studies [bib_ref] Pyridyl-indolizine derivatives as DNA binders and pH-sensitive fluorescent dyes, Marangoci [/bib_ref] we report on the synthesis, fluorescence properties and a preliminary evaluation of pyridyl-indolizine containing anthracene moiety as a DNA binding agent. The conversion of the ethyl ester group from the 1st position of the indolizine into the corresponding propargyl-based indolizine derivative facilitate the addition of anthracenyl group through a "click" reaction. Fluorescent pyridyl-indolizine derivative 124 was synthesized in moderate yield in an adapted four step synthesis. In the first step a cycloaddition of 4,4 -dipyridinium ylide, generated from the corresponding monoquaternary 4,4 -dipyridinium phenacyl salt 118, with ethyl propiolate as activated alkyne gives an indolizine intermediate 119. Isolated indolizine 119 ester was hydrolysed into the corresponding carboxylic acid derivative 120 in the second step, then in the third step was reacted with propargyl amine to yield alkyne-substituted indolizine derivative 121, suitable for further "click" reactions. Separately, chloromethylanthracene 122 was straightforwardly transformed into corresponding azide 123. In the final step, alkyne indolizine 121 was reacted together with azide 123 in a "click" type reaction to yield the final substituted pyridyl-indolizine 124 (see .
gives an indolizine intermediate 119. Isolated indolizine 119 ester was hydrolysed into the corresponding carboxylic acid derivative 120 in the second step, then in the third step was reacted with propargyl amine to yield alkyne-substituted indolizine derivative 121, suitable for further "click" reactions. Separately, chloromethylanthracene 122 was straightforwardly transformed into corresponding azide 123. In the final step, alkyne indolizine 121 was reacted together with azide 123 in a "click" type reaction to yield the final substituted pyridyl-indolizine 124 (see . [fig_ref] Table 2: The photophysical properties for the obtained molecules 10 and 11 [/fig_ref]. The excitation wavelength was 395 nm and the emission spectra were recorded in 425-700 nm domain. The emission of precursor alkyne 121 at acidic pH value of 2.0 is 6-fold stronger than at pH values of 5.0-12.0, while the fluorescence intensity of compound 124 at pH = 12.0 is 7-fold greater than at pH = 5.0. In the case of the precursor alkyne 121, the emission spectrum presents a band with a shoulder at 495 nm at pH = 12.0 and a strong bathochromic shift of the emission maximum at pH values from 8.0 to 5.0, while the emission spectrum of compound 124 at pH = 5.0-12.0 presents bands with a shoulder at 475 nm. At pH = 12.0, the solution emits strongly in the green spectral window at 475 nm.
Agarose gel electrophoresis analysis and spectroscopic investigations were performed in order to investigate the interaction of nucleic acids with pyridine-indolizines 121 and 124. Thus, deoxyribonucleic acid, low molecular weight from salmon sperm (sDNA) was used as a natural double-stranded DNA to test the binding properties of these compounds by agarose gel electrophoresis. The investigation shows that both investigated compounds 121 and 124 interact with sDNA.
The interactions of compounds 121 and 124 with sDNA was examined also by UV-visible and fluorescence investigation. Absorption and emission spectra were recorded before the addition of sDNA, immediately after the addition of sDNA solution and after the incubation of the indolizine-sDNA mixture at room temperature for 24 h. The absorption spectrum of compound 121 changes only after 24 h by the decrease of the absorption band intensity. Contrarily the absorption spectrum of compound 124 showed instantaneous changes in the shape of the band, upon the addition of sDNA, the structured band yielded by anthracene suffering modifications and after 24 h transforming into a broad band with a slightly weaker intensity. Thus, the presence of the anthracene moiety in the structure of indolizine 124 facilitates easier interaction with nucleic acids due to its higher affinity for DNA when compared to the propargyl moiety of the indolizine 121. The same conclusion is obtained after the study of the fluorescence spectra of the mixtures: in case of compound 121, alteration of the initial shape or intensity of the emission band was only observed after addition of sDNA and 24 h of incubation, while for the compound 124 immediate increase in fluorescence intensity of the mixture was observed, followed by additional increase after 24 h.
In order to investigate the ability of mono-indolizine mono-salts to interact with DNA, we used spectral methods UV-vis and fluorescence spectrometry [bib_ref] Fluorescent Conjugates: pH Stability, Dye-DNA Interaction and Biological Activity, Gradinaru [/bib_ref]. Thus, a series of mono-indolizine mono-salts 125a-e were synthesized starting from the corresponding mono-indolizines 126 by alkylating with halogeno-ketones 127. The monoindolizines 126 were obtained by the cycloaddition of the corresponding ylides, generated in situ from the mono salts 128, with ethyl propiolate (see . conclusion is obtained after the study of the fluorescence spectra of the mixtures: in case of compound 121, alteration of the initial shape or intensity of the emission band was only observed after addition of sDNA and 24 h of incubation, while for the compound 124 immediate increase in fluorescence intensity of the mixture was observed, followed by additional increase after 24 h. In order to investigate the ability of mono-indolizine mono-salts to interact with DNA, we used spectral methods UV-vis and fluorescence spectrometry. [bib_ref] Fluorescent Conjugates: pH Stability, Dye-DNA Interaction and Biological Activity, Gradinaru [/bib_ref] Thus, a series of mono-indolizine mono-salts 125a-e were synthesized starting from the corresponding mono-indolizines 126 by alkylating with halogeno-ketones 127. The monoindolizines 126 were obtained by the cycloaddition of the corresponding ylides, generated in situ from the mono salts 128, with ethyl propiolate (see Scheme 34). Scheme 34. Synthesis of mono-indolizine mono-salts 125a-e by alkylating with halogeno-ketones.
The absorption spectra of the mono-indolizine mono-salts were recorded in aqueous solution. The electron-donating methyl or methoxy group from the last inserted aromatic ring leads to a small bathochromic shift. DNA solution was progressively added to 125a and the spectra were recorded, in order to study the interaction of compound 125a with DNA. The hypsochromic shift of dye absorption maxima with the progressive addition of the DNA solution reflect the DNA binding to 125a. Moreover, in the UV region (350-360 nm) a hyperchromic effect was observed. This typical hyperchromic effect is caused by the damaging of the DNA double-helix structure due to the intercalation of 125a to DNA.
The interaction of 125a with DNA was also investigated by emission spectra in acidic Scheme 34. Synthesis of mono-indolizine mono-salts 125a-e by alkylating with halogeno-ketones.
The absorption spectra of the mono-indolizine mono-salts were recorded in aqueous solution. The electron-donating methyl or methoxy group from the last inserted aromatic ring leads to a small bathochromic shift. DNA solution was progressively added to 125a and the spectra were recorded, in order to study the interaction of compound 125a with DNA.
The hypsochromic shift of dye absorption maxima with the progressive addition of the DNA solution reflect the DNA binding to 125a. Moreover, in the UV region (350-360 nm) a hyperchromic effect was observed. This typical hyperchromic effect is caused by the damaging of the DNA double-helix structure due to the intercalation of 125a to DNA.
The interaction of 125a with DNA was also investigated by emission spectra in acidic conditions since our investigated dyes display a lower solubility and instability in weakly basic condition. The fluorescence reached a maximum in pH range 1.8-3.6. The fluorescence intensities vary as follow: 125b > 125a > 125e or 125d > 125c, being relatively higher in pH range 2.5-4.5. At higher pH values the solubility of dyes was very low due to the ylide formation.
Some preliminary investigation used herring sperm DNA (hsDNA) or control plasmid pUC19 in order to establish the interaction of the dyes with DNA. The fluorescence quenching of 125a upon addition of hsDNA shows that the dye 125a interact with DNA. The fluorescence of 125a-DNA system was investigated at different dye (125a) concentration and quenching extent has a maximum value at 6 µM concentration of dye. Supplementary investigation on the interaction mechanism for binding of 125a to hsDNA showed a similarly interaction way of 125a and ethidium bromide with hsDNA, compound 125a being binded in the minor groove of DNA.
Since the solubility the compound 125a is poor, to increase its solubility and to extend its applications as cell staining agent or cell pH sensitive dye, we proposed the incorporation of compound 125a in β-cyclodextrin (β-CD) [bib_ref] Novel cyclodextrin-based pH-sensitive supramolecular host-guest assembly for staining acidic cellular organelles, Pricope [/bib_ref]. The reversible transformation of the pyridinium moiety in compound 125a to the corresponding nitrogen ylide 129a under proper pH condition (see , influence its fluorescence emission spectra, making it a pH sensible fluorescent dye. The inclusion complex of indolizinyl-pyridinium salt in β-cyclodextrin (β-CD) was prepared by heating the equimolar amounts of components in water till 110 °C for 60 minutes, cooling down the solution under stirring for 6 h to reach the equilibrium followed by filtration through Phenex syringe filters (pore size: 0.45 μm).
The ESI-MS experiments and molecular docking studies confirmed the formation of an inclusion complex between indolizine derivative and β-cyclodextrin in 1:1 and 1:2 ratios. The cytotoxicity of the inclusion complex was considerably reduced comparing with the cytotoxicity of free indolizine on both HeLa (human cervix adenocarcinoma) and NHDF (normal human dermal fibroblasts) cells.
For the first time we demonstrated that the toxicity of a fluorescent dye was strongly reduced by the formation of cyclodextrin inclusion complex, allowing the successful application in cell staining. The study regarding cell membrane permeability showed that the nontoxic inclusion complexes mixture specifically accumulates in cell acidic organelles since could not pass through the cell plasma membrane.
Recent, [bib_ref] Cyclodextrin Encapsulated pH Sensitive Dyes as Fluorescent Cellular Probes: Self-Aggregation and In..., Sardaru [/bib_ref] we prepared three new cyclodextrin encapsulated pH sensitive dyes and investigated their ability for self-aggregation and in vitro assessments as fluorescent cellular probes. The mono-indolizine mono-salts 130a-c (their structures are presented in The inclusion complex of indolizinyl-pyridinium salt in β-cyclodextrin (β-CD) was prepared by heating the equimolar amounts of components in water till 110 - C for 60 min, cooling down the solution under stirring for 6 h to reach the equilibrium followed by filtration through Phenex syringe filters (pore size: 0.45 µm).
The ESI-MS experiments and molecular docking studies confirmed the formation of an inclusion complex between indolizine derivative and β-cyclodextrin in 1:1 and 1:2 ratios. The cytotoxicity of the inclusion complex was considerably reduced comparing with the cytotoxicity of free indolizine on both HeLa (human cervix adenocarcinoma) and NHDF (normal human dermal fibroblasts) cells.
For the first time we demonstrated that the toxicity of a fluorescent dye was strongly reduced by the formation of cyclodextrin inclusion complex, allowing the successful application in cell staining. The study regarding cell membrane permeability showed that the nontoxic inclusion complexes mixture specifically accumulates in cell acidic organelles since could not pass through the cell plasma membrane.
Recent, ref. [bib_ref] Cyclodextrin Encapsulated pH Sensitive Dyes as Fluorescent Cellular Probes: Self-Aggregation and In..., Sardaru [/bib_ref] we prepared three new cyclodextrin encapsulated pH sensitive dyes and investigated their ability for self-aggregation and in vitro assessments as fluorescent cellular probes. The mono-indolizine mono-salts 130a-c (their structures are presented in Scheme 36) were synthesized in moderate yields in an adopted two step strategy based on our methodological background.
an inclusion complex between indolizine derivative and β-cyclodextrin in 1:1 and 1:2 ratios. The cytotoxicity of the inclusion complex was considerably reduced comparing with the cytotoxicity of free indolizine on both HeLa (human cervix adenocarcinoma) and NHDF (normal human dermal fibroblasts) cells.
For the first time we demonstrated that the toxicity of a fluorescent dye was strongly reduced by the formation of cyclodextrin inclusion complex, allowing the successful application in cell staining. The study regarding cell membrane permeability showed that the nontoxic inclusion complexes mixture specifically accumulates in cell acidic organelles since could not pass through the cell plasma membrane.
Recent, [bib_ref] Cyclodextrin Encapsulated pH Sensitive Dyes as Fluorescent Cellular Probes: Self-Aggregation and In..., Sardaru [/bib_ref] we prepared three new cyclodextrin encapsulated pH sensitive dyes and investigated their ability for self-aggregation and in vitro assessments as fluorescent cellular probes. The mono-indolizine mono-salts 130a-c (their structures are presented in Scheme 36) were synthesized in moderate yields in an adopted two step strategy based on our methodological background. Scheme 36. The structure of the mono-indolizine mono-salts 130a-c.
Due to poor solubility in water of the indolizinyl-pyridinium salts 130a-c, these compounds were tested for the formation of inclusion complexes with β-CD. We started with Scheme 36. The structure of the mono-indolizine mono-salts 130a-c.
Due to poor solubility in water of the indolizinyl-pyridinium salts 130a-c, these compounds were tested for the formation of inclusion complexes with β-CD. We started with an equivalent of each compound suspended in water, followed by the addition of an excess of β-CD (5 eq.) and heating to 90 - C until the reaction solution became transparent (after 25 min of heating). In case of 130c_CD the solutions remaining completely transparent after cooling down, while in case of 130a_CD and 130b_CD the formation of slightly cloudy solutions was observed. The excess up to 10 equivalents of the CD amount in the reaction mixture still did not change the appearance for the 130a_CD and 130b_CD at room temperature. These solutions were used in the subsequent analyses only after a microfiltration procedure.
The ESI-MS experiments and molecular docking studies confirmed a 1:1 and 1:2 ratios between indolizine derivative and β-cyclodextrin in the inclusion complex.
Absorption and emission spectra of indolizine derivatives 130a-c and their inclusion complexes 130a-c_CD were recorded at 1.0, 7.4, and 13.0 pH values. (1.0, 7.4, and 13.0) and the results compared. The indolizines 130a and 130b have similar absorption spectra, but slightly different than the absorption spectra of the indolizine 130c. At pH = 13.0, all the investigated indolizines have shown similar spectra. In the case of inclusion complexes 130a-c_CD, their absorption spectra at acidic and neutral pH values (1.0 and 7.4, respectively) are similar for all the investigated complexes, but different compared to the absorption spectra of the starting indolizines 130a-c. At basic pH values (pH = 13.0), the shape of the spectra of all complexes 130a-c_CD is drastically changed in comparison to each other and to the starting indolizines 130a-c.
The excitation wavelength was 420 nm when fluorescence spectra of 130a-c and 130a-c_CD were measured at different pH values. Starting indolizines 130a-c exhibited similar emission spectra at acidic and neutral pH values, with an emission band around 550 nm. This band is approximately two times higher in intensity at acidic than neutral pH value. At basic pH values (pH = 13.0), the indolizines shown low to no fluorescence. The reversible transformation of pyridinium salts into analogous non-fluorescent ylides may explain this pH-dependent behaviour. The inclusion complexes 130a-c_CD demonstrated slightly different pH dependent behaviours, showing a similar low intensity at basic pH values, but different intensities at acidic pH values. The fluorescence intensities of the inclusion complexes 130a_CD and 130b_CD at acidic pH are only slightly higher than intensities at neutral pH values, while in the case of the inclusion complex 130c_CD, the fluorescence intensities at both the acidic and neutral pH were comparable. The formation of the inclusion complexes altered the planar structure of the indolizinyl-pyridinium salt molecules due to the specific rotation limitations induced by the steric hindrances with the CD, which determine a decrease of fluorescence intensity of the inclusion complexes at acidic pH values.
Supplementary studies shown that the inclusion complexes have no cytotoxicity, they specifically accumulate within acidic organelles or mitochondria due to cellular permeability, and their intracellular fluorescence increase over a 24 hours period with outstanding signal stability.
## Concluding remarks
In conclusion, the present review article deals with design, synthesis, and photophysical properties of azaheterocycle-based materials. Several synthetic methodologies were used for the obtaining of fluorescent 5, 6 membered and fused azaheterocycles. The structure-fluorescence relationship of the obtained compound was investigated and allowed the authors to obtain compounds with better fluorescent properties in terms of emission wavelength, emission intensity and quantum yield.
The presented approaches regarding the synthesis of fluorescent azaheterocycles, using varied methodologies, enable researchers to create a library of multifunctional derivatives, possessing high efficiency of fluorescence.
The tunable emission properties of the azaheterocyclic compound make them valuable materials as potential luminophores in OLED devices, as luminescent solar concentrators or scintillators, as fluorometric (naked-eye chemosensor) molecular sensors for detection of metal ions, strong acids and bases with high selectivity, as thermally activated delayed fluorescence emitters and as fluorescent probe.
[fig] Scheme 1: Synthetic procedure used for the obtaining of indolylcarbazole derivatives 4. [/fig]
[fig] Scheme 4: Synthesis of antipyrine-containing diarylethenes derivatives 14. [/fig]
[fig] Scheme 5: Structure changes of 14 induced by TBAH/HCl and UV/vis. [/fig]
[fig] Scheme 6: One-pot synthesis of multi-substituted aminopyridines 17-26. [/fig]
[fig] Scheme 7: Derivatization of multisubstituted aminopyridine 17. Scheme 7. Derivatization of multisubstituted aminopyridine 17. [/fig]
[fig] Scheme 8: Binding the fluorescent aminopyridine scaffold to biomolecules conjugated with alkynyl group. Scheme 8. Binding the fluorescent aminopyridine scaffold to biomolecules conjugated with alkynyl group. [/fig]
[fig] Molecules 2022 ,: 27, x 12 of 41 Scheme 12. N-arylation Buchwald-Hartwig cross-coupling reaction of arylamines 44a-c. [/fig]
[fig] λ: abs -spectral position of the first absorption maxima, ε( abs )-magnitude of the molar absorptivity of abs , em -the spectral position of the fluorescence maxima, Φ fl -the fluorescence quantum yield, MCHX-methylcyclohexane, THF-tetrahydrofuran, CAN-acetonitrile. [/fig]
[fig] The 7 -: bromo-2,3-diphenylpyrido[2,3-b]pyrazine 59 and 10-bromo-acenaphtho[1,2b]pyrido[2,3-e]pyrazine 61 donors were bonded to the corresponding DHPHAzSi acceptors through the Pd-catalyzed Buchwald-Hartwig amination in good yields (see Scheme 19). Molecules 2022, 27, x 17 of 41 Scheme 19. Buchwald-Hartwig amination of 7-bromo-2,3-diphenylpyrido[2,3-b]pyrazine 59 and 10bromo-acenaphtho[1,2-b]pyrido[2,3-e]pyrazine 61 donors with DHPHAzSi acceptors. [/fig]
[fig] Scheme 20: One-pot, three-component synthesis of 3-biphenyl-pyrrolo[1,2-c]pyrimidines 66a-j. [/fig]
[fig] Scheme 21: Synthesis of pyrrolo[1,2-c]pyrimidine 69a-j by 1,3-dipolar cycloaddition. [/fig]
[fig] Scheme 24: Influence of the substituents on the fluorescence of the pyrrolo-PY compounds. [/fig]
[fig] Scheme 25: Synthesis of aryl substituted pyrrolopyridazines 86-87 by 1,3-dipolar cycloaddition. Scheme 25. Synthesis of aryl substituted pyrrolopyridazines 86-87 by 1,3-dipolar cycloaddition.The reactions under MW irradiation have shown some important advantages such as slightly higher yields and decreasing of the reaction time from 2 h to 5 min in liquid phase and 15 minutes in solid phase. [/fig]
[fig] Scheme 26: Synthesis of fluorescent fused pyrrolodiazines 91-95 by 1,3-dipolar cycloaddition. [/fig]
[fig] Scheme 27: Derivatization of fluorescent pyrrolodiazines 91-94 by bromination. [/fig]
[fig] Scheme 29: Derivatization of blue fluorescent azaheterocyclic derivatives 102-103. [/fig]
[fig] Φ: fl -photoluminescence quantum yield, τ-fluorescence lifetime, f-the weighted contribution. The phenanthroline derivatives displayed broad and structureless emission band in 425 to 480 nm depending on the substituent nature at the pyrrole moiety. For 1,7-phenanthroline derivatives the emission band vas found in the 440-450 nm region, while for 4,7-phenanthrolines the emission band was found in 425-480 nm region. The position and the nature of the substituent at the pyrrole ring have a great influence on the fluorescence quantum yield of pyrrolo-phenanthroline derivatives. Scheme 32 illustrates the effect of substituents from the pyrrole ring on the fluorescence yield of pyrrolophenanthroline derivatives. [/fig]
[fig] Scheme 32: The influence of substituents from the pyrrole ring on the fluorescence yield of pyrrolophenanthroline derivatives. The fluorescence emission of disubstituted phenanthroline derivatives (114d, 115c, 115d) is practically quenched by the presence of halogens (Cl, Br) in the para position of phenacyl group (COC6H4) from 9 th position of the pyrrole ring. The emission quantum yield decreases significantly (115e−0.044) when COC6H4OMe-(p) group is introduced in Scheme 32. The influence of substituents from the pyrrole ring on the fluorescence yield of pyrrolophenanthroline derivatives. [/fig]
[fig] Scheme 33: Synthesis of pyridyl-indolizine containing anthracene moiety as a DNA binding agent. Scheme 33. Synthesis of pyridyl-indolizine containing anthracene moiety as a DNA binding agent. The emission properties of compounds 121 and 124 were investigated in DMF at different pH values using 1xTAE buffer solutions (40 mM Tris, 20 mM acetic acid and 1 mM EDTA), and are presented in [/fig]
[fig] Scheme 35: The reversible transformation of the pyridinium moiety in compound 125a to the corresponding nitrogen ylide 129a under proper pH condition. [/fig]
[fig] Author: Contributions: Design and conception was performed by G.Z., I.I.M. and C.M. contribute by writing, reviewing and approving the final version. All authors have read and agreed to the published version of the manuscript. [/fig]
[fig] Funding: This research received no external funding. Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. Data Availability Statement: Not applicable. [/fig]
[table] Table 2: The photophysical properties for the obtained molecules 10 and 11.Table 1. Fluorescent properties of the selected 5aa and 5ba in DMSO. [/table]
[table] Table 3: Fluorescent properties of aminopyridines [/table]
[table] Table 4: Yields, UV/VIS and fluorescence data for esters 36a-e in absolute ethanol and a methanol-HEPES buffer (80:20) solution. [/table]
[table] Table 5: Photophysical data for the as-prepared luminogens [/table]
[table] Table 6: Spectroscopic properties, in CH3CN at room temperature, of the synthesized dyes [/table]
[table] Table 7: Spectroscopic behavior of dyes A2 and D1 in different solvents. [/table]
[table] Table 8: Fluorescence quantum yield (Φfl) of the chromeno [2,3-b]indole derivatives 47a-c in acetonitrile and ethyl acetate at 280 nm (tryptophan in water as reference) and at 313 nm (naphthalene in ethanol as reference). [/table]
[table] Table 9: The photophysical constants of the compounds 23e-g in different solvents. [/table]
[table] Table 10: The photophysical properties of the compounds 56-58. [/table]
[table] Table 11: The main spectral features of the compounds 4a-j in acetonitrile:chloroform (1:1). [/table]
[table] Table 12: The spectroscopic features for absorption and fluorescence of 69a-j in chloroform solution (10 −6 M). [/table]
[table] Table 13: Photophysical properties of compounds 76a-c in toluene, acetonitrile, dichloromethane (DCM) and methanol solutions at room temperature, λex = 420 nm. Lifetimes (τ) were determined at λmax of the emission bands. [/table]
[table] Table 14: λ max (nm) of absorption spectra, fluorescence spectra, and relative quantum yields (%) of PD compounds 79-82. [/table]
[table] Table 15: The spectroscopic features for absorption and fluorescence intensity of 86a-d and 87a-d in in acetonitrile solution (5 × 10 −5 M). [/table]
[table] Table 16: λmax (nm) of absorption spectra, λmax (nm) of emission spectra, and relative quantum yields (%) of compounds 91-95. [/table]
[table] Table 19: Photophysical properties of isoquinoline derivatives. [/table]
|
Subsets and Splits