Epigenome-wide meta-analysis of PTSD across 10 military and civilian cohorts identifies methylation changes in AHRR - Nature Communications

Post-traumatic stress disorder (PTSD) is characterized by re-experiencing, avoidance, and hyperarousal symptoms that can negatively impact mood and physiologic health¹. Although not everyone who experiences trauma goes on to develop PTSD, those who do often experience severe and disabling symptomatology^2,3. Research suggests that both genetic and environmental factors contribute to risk for developing PTSD^4,5.

Given the necessary, but not sufficient, role of environmental exposure (i.e., trauma) in the development of PTSD, it is critical to characterize the pathways underlying differential risk and resilience. Studies that contexualize the role of environmental influences provide additional insight into modifiable factors that may promote post-trauma resilience. For example, lack of social support at the time of trauma is associated with increased risk of developing PTSD in both military and civilian contexts⁶. Similarly, studies that investigate how presumed environmental influences might affect biological pathways could provide insights into the genes whose regulation patterns differ in those with PTSD. A growing line of research aimed at elucidation of mechanisms by which environmental factors contribute to PTSD, has focused on epigenetic modifications, which regulate gene function in response to environmental triggers. Epigenetic modifications, such as DNA methylation at cytosine-guanine dinucleotides (CpG sites), can induce changes in gene expression that are maintained through each round of cell division.

Multiple reviews have linked traumatic stress to differences in the proportion of methylated DNA intensity to non-methylated DNA intensity (ß) at specific CpG sites^7,8,9,10. Indeed, a number of epigenome-wide association studies (EWAS) of individual cohorts have identified PTSD-associated CpG sites in genes and pathways related to neurotransmission and immune function^{11,12,13,14,15}. Similarly, studies using methylation of specific CpG sites that predict chronological age with reasonably high accuracy to capture age acceleration (deviations from a line of prefect prediction), demonstrate associations with PTSD and link differences in peripheral DNA methylation to memory formation and neural integrity^16,17. Although promising, the extant literature of EWAS studies on PTSD is limited by use of individual cohorts with small sample sizes and varying analytic pipelines that can make it challenging to synthesize findings. Consortia efforts can overcome these limitations by providing a shared analytic pipeline and increasing sample size and thus statistical power. The goal of this study is to capitalize on consortium strengths by conducting a meta-analysis of DNA methylation across 10 military and civilian cohorts participating in the Psychiatric Genomics Consortium (PGC) PTSD Epigenetics Workgroup. Our results suggest that lower aryl-hydrocarbon receptor repressor (AHRR) methylation in those with PTSD correlates with lower kynerunine levels, which may contribute to immune dysregulation in PTSD.

Participating cohorts

Sample characteristics for the 10 cohorts that have contributed data are listed in Table 1 (N = 1896). All participants were exposed to trauma, and 42% had a current diagnosis of PTSD. There were no significant age differences in PTSD cases and trauma-exposed controls. However, the demographic characteristics for each cohort varied substantially (Supplementary Data 1), with cases more likely to be male (X² = 22.4; p < 0.05) and current smokers (X² = 54.2; p < 0.05) across the majority of cohorts and in the overall cohort.

PTSD-associated CpG sites from Meta-Analysis

In our primary meta-analysis, we found 4 CpG sites associated with current PTSD after correction for multiple comparisons (−6.60 < z < −5.57; p < 3.6E-08; Fig. 1; Table 2; Supplementary Data 2). AHRR contains the top 4 PTSD-associated CpGs, with lower methylation in PTSD cases relative to controls (Supplementary Fig. 1). This was consistent when stratified by both sex (Supplementary Fig. 2) and ancestry (Supplementary Fig. 3). Data from iMETHYL, which reports eQTMs associated at false-discovery rate (FDR) < 0.05, suggests that methylation of both cg05575921 and cg25648203 are inversely associated with AHRR expression in peripheral blood mononucular cells (PBMCs).

a Manhattan plot for the meta-analysis of 10 cohorts (N_cases = 878, N_controls = 1,018). Association analyses of each cohort are based on empirical Bayes method. Meta-analysis is done by sample size-weighted sum of z-scores. p-values are two-sided and unadjusted for multiple testing. All of the association models are adjusted for sex (if applicable), age, CD8, CD4, NK, B cell, and monocyte cell proportions, and ancestry using principal components (PCs). The x-axis is the location of each site across the genome. The y-axis is the –log10 of the p-value for the association with PTSD. The dashed line indicates statistical significance at p < 3.6E-8.

Full size image

Sensitivity analysis with smoking status

As lower methylation of AHRR CpG sites has been associated with smoking^18,19,20, we controlled for smoking status in our sensitivity analyses of the 4 significant CpGs (Supplementary Fig. 4). The association between AHRR methylation and PTSD was attenuated for all four CpGs. Since there is a higher rate of smoking among PTSD cases, we evaluated potential differences of effect between PTSD and smoking status by testing the associated CpGs separately in smokers and non-smokers (Fig. 2) and by plotting the associated CpGs separtely by smoking status (Supplementary Fig. 5). The associations between AHRR CpGs and PTSD were most prominent in non-smokers. Of the multiple stratified analyses performed (Supplementary Figs. 2, 3, and 5), the effect sizes were consistent across different strata, with the smokers being the only one did not show an association with PTSD.

One-thousand eight-hundred and forty-four samples are examined, including 1255 non-smokers (circles) and 589 smokers (triangles). On the x-axis, error bars represent the 95% confidence interval, and the center of the error bars represent effect sizes for each of the CpGs on the y-axis.

Full size image

We further delineated the association between PTSD and smoking-associated CpG sites from a large meta-analysis on smoking conducted by the CHARGE consortium²⁰. We hypothesized that unreported smoking among those that identify as non-smokers would result in the appearance of an association between smoking-associated CpG sites and PTSD. This evaluation included 21 CpGs representing the most significantly associated CHARGE smoking sites and CpGs in AHRR. If the association between AHRR and PTSD was due to unreported smoking in PTSD controls, we would expect to see a similar effect across all smoking-CpGs. However, after adjusting for the number of tests performed, AHRR CpG sites remained associated with PTSD in the non-smokers only, and no smoking CpG showed a significant association with PTSD in the controls or cases. These findings are consistent with the hypothesis that the association between PTSD and DNA methylation of AHRR is independent of smoking status (Supplementary Fig. 6).

Tryptophan catabolism in PTSD

We next evaluated other biological factors that could contribute to aryl-hydrocarbon receptor (AHR) expression using linear regression models. AHR is activated in many immune cell types, including T cells, B cells, and NK cells, by indolamine-mediated tryptophan catabolism²¹. Inflammation can shift tryptophan metabolism away from serotonin synthesis towards kynurenine synthesis²². To evaluate the association of tryptophan catabolism in PTSD and its relationship to AHRR methylation, we leveraged data from the 116 subjects from MRS that had both DNA methylation and tryptophan metabolite data. In this group, we first note that kynurenine levels were lower in subjects with PTSD relative to controls (t = −2.00; p = 0.048; Supplementary Fig. 7). Consistent with this observation, lower methylation of AHRR CpGs associated with lower kynurenine (cg21161138; t = 2.09, p = 0.039) and lower kynurenic acid (cg05575921; t = 2.95, p = 0.004).

Further, metabolome data in the MRS confirmed that self-reported smoking status is consistent with empirical cotinine levels; 87% of self-reported non-smokers and light smokers had cotinine levels consistent with no-smoking, second-hand smoking or light exposure, while 91% of regular or heavy smokers had cotinine levels consistent with their endorsement. In addition, these data also provide insight into why controlling for smoking may attenuate the association between DNA methylation and PTSD. We noted an inverse relationship between cotinine and kynurenine (r = −0.267; p = 0.0004) that was consistent among both smokers and non-smokers (Supplementary Fig. 8).

An individual’s risk of developing PTSD depends on both the nature of the trauma and the physiological response to that trauma²³. Not all individuals exposed to trauma develop PTSD, and a better understanding of the modifiable biological factors underlying risk and resilience will inform the development of new prevention and treatment strategies. In this study we identified CpG sites associated with PTSD. We observed that, on average, PTSD cases had lower methylation at several CpG sites in the AHRR gene when compared to trauma-exposed controls. Methylation of AHRR CpGs has been strongly linked to smoking^18,19,20. As substantially more of the PTSD cases reported smoking compared to controls, this suggested that we should control for smoking in the EWAS. A comparable approach was taken by Marzi and colleagues²⁴. In their study of DNA methylation in relation to victimization stress in adolescents, they noted experiment-wide significant associations in 3 of the 4 AHRR CpGs associated with PTSD. Similar to our findings, the association of AHRR and PTSD was no longer significant after statistically controlling for smoking. Marzi and colleagues concluded that there are no robust changes in DNA methylation in victimized young people. Our study went beyond this approach to conduct a stratified analysis that revealed the most prominent PTSD-associated difference in AHRR methylation was evident in the non-smokers²⁴. To evaluate the possibility of confounding due to unreported smoking among those that identify as non-smokers, we tested top smoking-related CpGs, including AHRR, for association with PTSD, and identified no evidence of association between smoking-related CpG sites and PTSD, supporting our conclusion that the association between AHRR CpG sites and PTSD was independent of smoking.

Though it is possible that stratifying on smoking could introduce a collider bias that distorts the stratified analysis because of a SNP or other unmeasured confounder, this scenario seems unlikely given that the associations between AHRR CpGs and PTSD are stronger in non-smokers. As such, this cross-sectional analysis is unable to determine whether methylation of AHRR CpGs is likely a cause or consequence of smoking. However, two recently published studies have provided insight into this question. The first sought to identify DNA methylation differences associated with PTSD in blood of 378 PTSD cases and 135 controls from the Translational Research Center for Traumatic Brain Injury and Stress Disorders (TRACTS) cohort²⁵. This study used a methylation-based score for smoking as a covariate, and still reported a significant association between PTSD and cg05575921 (AHRR). The second examined DNA methylation differences in blood associated with polycyclic aromatic hydrocarbon (PAH) levels unrelated to smoking²⁶. In 708 non-smokers from Taiwan, the authors report associations between PM_2.5 levels, an indicator of regional air quality, and lower methylation at cg05575921. Taken together, these studies suggest that methylation of this CpG can occur through processes other than smoking.

Over the last decade, multiple studies described the role of the aryl-hydrocarbon receptor in the immune system, with specific roles in T cells, B cells, monocytes and dendritic cells (DC)^21,27,28,29. The aryl-hydrocarbon receptor (AhR) pathway works in a regulatory capacity. Briefly, when a ligand binds to the AhR, it translocates to the nucleaus where it drives expression of its target genes, including the aryl-hydrocarbon receptor repressor (AHRR), which begins a feedback loop in which it can competitively bind the AhR. The AhR pathway can either limit or stimulate an inflammatory response. One mechanism by which this occurs is by promoting differentiation of T cells into T regulatory cells (Tregs) or T helper 17 (Th17), though this appears to be done in a ligand-specific manner. For example, endogenous ligands, such as kynurenine or dietary indoles³⁰, tend to promote differentiation into Tregs, which reduce the immune response in a self-limiting manner, resulting in expression of anti-inflammatory genes, reduced inflammation, and less Treg generation. In contrast, exogenous ligands, such as dioxin or polyaromatic hydrocarbons in cigarette smoke, tend to promote differentiation into Th17, expression of pro-inflammatory cytokines, heightened inflammation, and activation the drug metabolizing enzymes^27,31,32.

In a subset of subjects with tryptophan metabolite data, we reported that kynurenine levels were significantly lower in subjects with PTSD relative to trauma-exposed controls and that lower methylation of AHRR CpGs associated with lower kynurenine and its metabolites. This pattern is similar to what is observed following chronic exposure to nicotine and possibly other AHR-stimulating ligands³³. This chronic exposure scenario is reflected in our cohort among both controls and PTSD cases by the finding that higher levels of cotinine were associated with decreased levels of kynurenine. Independent of its source, lower levels of kynurenine would likely result in a counter-regulatory increase in pro-inflammatory activity³⁴, and may help to account for the frequent observation of heightened inflammatory activity observed in subjects with PTSD³⁵. Our secondary analysis of DNA methylation and kynurenine was limited to the only cohort with cotinine and metabolomics data available on the subjects included in the methylation study, which limited sample size, and power for this analysis, and potentially generalizability to other cohorts. However, we believe that the ability to verify self-reported smoking status with a biological measure and to evaluate our kynurenine hypothesis was a considerable strength of the study. We hope these results will prompt other investigators to replicate and extend these findings.

While smokers are more common among our PTSD group, approximately half of the cases in our cohorts are non-smokers, who exhibited the most prominent associations between AHRR CpGs and PTSD. Methylation of AHRR CpGs decreases only so much even among heavy smokers, suggesting that there may be limited variability in these subjects and a limited degree to which methylation at this site may be decreased. Though the vast majority of the AHRR methylation literature is focused on characterizing its variation in the context of smoking, the types of polycyclic aromatic hydrocarbons that stimulate AHR are present from multiple exogenous sources, including burning wood or charcoal, auto emissions, industrial exhaust, and urban dust³⁶. Similarly, participants in the WTC and military cohorts are likely to have substantial levels of occupational exposure that could result in AHR activation^37,38,39. Though we were not able to directly assess exposure to endogenous or exogenous AhR ligands in this study, it is reasonable to hypothesize that smoking is just one of many potential environmental exposures that could promote inflammation following AHR stimulation.

Both epidemiologic and immunologic studies report immune dysregulation in those with PTSD compared to controls. For example, autoimmune and inflammatory disorders, such as rheumatoid arthritis, have been linked to PTSD^40,41,42. Studies of the immune system generally support higher levels of pro-inflammatory cytokines in PTSD cases relative to controls³⁵. A study investigating Tregs reported a lower number of Tregs in the blood of PTSD patients⁴³, while another reported a difference in the composition of Treg populations that suggested higher susceptibility for autoimmunity⁴⁴. Another study reported lower proportions of Tregs and higher proportions of Th17 cells in PTSD cases before linking Th17 counts to higher clinical symptom scores⁴⁵. Finally, a randomized control trial of subjects undergoing narrative exposure psychotherapy noted higher levels of Tregs and lower PTSD symptoms following treatment⁴⁶. Unfortunately, we were unable to directly measure or impute the proportion of different types of T cells in this study. Though the degree to which the immune system is involved remains speculative without functional studies, we hope these data will support more detailed immunophenotyping of subjects with PTSD.

This study has a number of additional limitations that should be considered. First, we used existing data generated on the HumanMethylation450 array, which captures only a fraction of the CpG sites in the genome. Though sequencing methods would capture a larger proportion of the epigenome that may include regions important for PTSD, focusing on this array allowed us to evaluate a larger and more diverse cohort of subjects and enabled processing and analysis of these diverse samples with a common analytic pipeline. Similarly, this meta-analysis uses only cross-sectional data. In order to establish whether these PTSD-associated differences are a cause or consequence of the disorder or both, longitudinal and functional studies will be required. Second, there is phenotypic heterogenetity among the cohorts contributing to this meta-analysis. Some cohorts assess individuals that were recently exposed to trauma while others include individuals with chronic PTSD. Also, few of the cohorts included in this meta-analysis have detailed physical or psychiatric information on subjects prior to trauma exposure, making it difficult to evaluate the role of lifestyle factors, such as obesity, or comorbidities, such as substance use. Thus, it is possible that some of the epigenetic differences observed in this study may have been in place prior to PTSD development, including the possibility of previous episode of PTSD in subjects that no longer meet current diagnostic criteria, or reflect other factors such as differences in genetic background between cases and controls. Third, our study examined whole blood-derived DNA. Though this approach likely captures part of the PTSD sequelae and may be informative for future biomarker studies, it is unlikely to reflect DNA methylation in brain regions most relevant for PTSD. As studies of tissues from PTSD Brain Banks are conducted, it will be important to look for parallels between these different tissues.

Taken together, the results of this study implicate the immune system in PTSD and suggest that epigenetic mechanisms may play a role in that process. A substantial fraction of those diagnosed with PTSD do not respond to pharmacologic or psychological interventions, and clinical and preclinical studies have begun to evaluate strategies to limit inflammation as a first line treatment^47,48. Future studies should evaluate the role of ligand-specific AHR activation in the development and progression of PTSD.

Post-traumatic stress disorder cohorts and assessments

The participating cohorts, presented in Table 1, consisted of three civilian cohorts: the Detroit Neighborhood Health Study (DNHS), the Grady Trauma Project (GTP), and the World Trade Center 9/11 First Responders study (WTC); and seven military cohorts: the Army Study to Assess Risk and Resilience in Servicemembers (Army STARRS), the Marine Resiliency Study (MRS), the Injury and Traumatic Stress study (INTRuST), the Prospective Research in Stress-Related Military Operations study (PRISMO), a European and African-American cohort from the Veterans Affairs’ Mental Illness Research, Education and Clinical Centers (VA-M), and a cohort from the National Center for PTSD (VA-NCPTSD). All subjects participating in these studies provided informed consent, and all studies were approved by respective institutional review boards. Current PTSD diagnosis was assessed by each individual cohort in accordance with the harmonization principles adopted by the PGC-PTSD Workgroup⁴. Briefly, diagnosis of current PTSD was based on the diagnostic criteria defined by each cohort’s principal investigator (see cohort descriptions in Supplementary Methods for complete details). All control subjects were trauma-exposed, and, if assessed in the respective cohort, control subjects that had a prior history of PTSD were excluded. Covariates were age, sex, genetic ancestry, and smoking status. A total of 1,896 subjects (42% cases) with DNA methylation from whole blood measured using the Illumina HumanMethylation450 BeadChip were selected for inclusion in this meta-analysis.

Quality control (QC) procedures

Data contributed to this study were part of 10 different studies, which were each designed to address a study-specific question. In developing the pipeline to analyze this existing data, the PGC-PTSD group issued recommendations for studies to balance potential confounders. All cohorts were instructed on importance of balancing plate layouts, and the potential for residual sources of technical variation was further assessed by examining the association of methylation principal components with chip, position, plate, and demographic characteristics within each study. Each study performed analyses at their site. To ensure consistent QC procedures across all participating cohorts, a set of common scripts were developed and implemented uniformly⁴⁹. ß-values, representative of the proportion of methylation at each probe, were caculated for all CpG sites in each sample. Samples with probe detection call rates <90% and those with an average intensity value of either <50% of the experiment-wide sample mean or <2000 arbitrary units (AU) were excluded. Probes with detection p-values > 0.001 or those based on less than three beads were set to missing as were probes that cross-hybridized between autosomes and sex chromosomes⁵⁰. CpG sites with missing data for >10% of samples within cohorts were excluded from analysis. Normalization of probe distribution was conducted using Beta Mixture Quantile Normalization (BMIQ)⁵⁰ after background correction. Density plots of Type I and Type II probes before and after normalization were examined to confirm probe distributions between types were similar. ComBat was used to account for sources of technical variation including batch and positional effects, while preserving variation attributable to study-specific outcomes and covariates that would be used in downstream analyses (e.g., case status or sex). Proportions of CD8, CD4, NK, B cells, monocytes, and granulocytes were estimated using each individual’s DNA methylation data based on the approach described by Houseman and colleagues and publicly available reference data (GSE36069)^51,52.

Statistical analysis

Within each cohort, logit transformed ß-values (M-values) were modeled by linear regression as a function of PTSD, adjusting for sex (if applicable), age, CD8, CD4, NK, B cell, and monocyte cell proportions, and ancestry using principal components (PCs). For cohorts with available GWAS data (Army STARRS, GTP, INTRuST, MRS, VA-M, VA-NCPTSD), PCs 1-3 were included as covariates. For cohorts without GWAS data (DNHS, PRISMO, WTC), the method described by Barfield and colleagues was used to generate ancestry PCs directly from the methylation data using CpG sites from the HumanMethylation450 array that are located within 0–100 bp of 1000 Genomes Project variants with minor allele frequency > 0.01, and PCs 2–4, which were the components that correlate most with ancestry, were used to adjust for ancestry as recommended by the authors⁵³. Using the empirical Bayes method in the R package limma⁵⁴, moderated t-statistics were calculated for each CpG site, converted first into one-sided p-values then converted into z-scores to account for direction of effects. Meta-analysis across cohorts was performed by weighting each cohorts’ z-scores by its sample size relative to the total meta-analysis sample, summing weighted z-scores across cohorts from which two-sided p-values were calculated. We used the genome-wide significance threshold suggested for HumanMethylation450 BeadChip to adjust for multiple testing (p < 3.6E-8)⁵⁵. The iMETHYL database (http://imethyl.iwate-megabank.org/) was used to evaluate whether methylation of AHRR CpGs associated with AHRR expression in PBMCs, the available tissue that most closely resembles whole blood⁵⁶.

Metabolite analysis in the Marine Resiliency Study (MRS)

Targeted, broad-spectrum metabolomics was performed as previously described^57,58 with minor modifications. Lithium heparin plasma samples were collected and stored at −80 °C until used for analyses in the MRS metabolomics study. Samples were analyzed from 116 participants exposed to military combat; 53 diagnosed with PTSD, and 63 controls without. Ninety microliters of plasma and 10 µl of added stable isotope internal standards were combined, extracted, and analyzed by hydrophilic interaction chromatography, electrospray ionization, and tandem mass spectrometry on a SCIEX 5500 QTRAP HILIC-ESI-MS/MS platform. From the 486 metabolites measured, we extracted and analyzed data for kynurenine and kynurenic acid to test our hypothesis that AHRR methylation associates with trypophan breakdown products. Cotinine, a metabolite of nicotine with a plasma half-life of about 18 h, was measured to provide independent information about tobacco exposure. A cotinine arbitrary unit (AU) of 1 × 10⁶ was used as the upper limit for a non-smoker, light smoker, or second-hand tobacco exposure, with ≥1 × 10⁶ indicative of regular or heavy smoking. Analyte AU data was log-transformed, and analyte z-scores were used for further statistical analysis. To determine if trytophan metabolite levels differed between PTSD cases and controls, we performed linear regressions of metabolites on PTSD status, including GWAS PCs 1–3 and cell type proportions as covariates. To determine if tryptophan metabolite levels were associated with AHRR methylation, we performed linear regressions of metabolites on each AHRR CpG site, including PCs and cell type proportions as covariates. Correlation between cotinine and kynurenine was measured using Pearson’s correlation coefficient.

Reporting summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.

1. Association, A. P. American Psychiatric Association. 5th edn, (American Psychiatric Association, 2013).

2. Brady, K. T., Killeen, T. K., Brewerton, T. & Lucerini, S. Comorbidity of psychiatric disorders and posttraumatic stress disorder. J. Clin. Psychiatry 61, 22–32 (2000).

3. Najt, P., Fusar-Poli, P. & Brambilla, P. Co-occurring mental and substance abuse disorders: a review on the potential predictors and clinical outcomes. Psychiatry Res. 186, 159–164 (2011).

   Article  PubMed  Google Scholar 

4. Logue, M. W. et al. The Psychiatric Genomics Consortium Posttraumatic Stress Disorder Workgroup: Posttraumatic Stress Disorder Enters the Age of Large-Scale Genomic Collaboration. Neuropsychopharmacology 40, 2287–2297 (2015).

   Article  PubMed  PubMed Central  Google Scholar 

5. Nievergelt, C. M. et al. Genomic approaches to posttraumatic stress disorder: The Psychiatric Genomic Consortium Initiative. Biol. Psychiatry 83, 831–839 (2018).

   Article  PubMed  PubMed Central  Google Scholar 

6. Brewin, C. R., Andrews, B. & Valentine, J. D. Meta-analysis of risk factors for posttraumatic stress disorder in trauma-exposed adults. J. Consult Clin. Psychol. 68, 748–766 (2000).

   Article  CAS  PubMed  Google Scholar 

7. Vinkers, C. H. et al. Traumatic stress and human DNA methylation: a critical review. Epigenomics 7, 593–608 (2015).

   Article  CAS  PubMed  Google Scholar 

8. Daskalakis, N. P., Rijal, C. M., King, C., Huckins, L. M. & Ressler, K. J. Recent genetics and epigenetics approaches to PTSD. Curr. Psychiatry Rep. 20, 30 (2018).

   Article  PubMed  PubMed Central  Google Scholar 

9. Sheerin, C. M., Lind, M. J., Bountress, K. E., Nugent, N. R. & Amstadter, A. B. The genetics and epigenetics of PTSD: overview, recent advances, and future directions. Curr. Opin. Psychol. 14, 5–11 (2017).

   Article  PubMed  PubMed Central  Google Scholar 

10. Kim, G. S., Smith, A. K., Nievergelt, C. M. & Uddin, M. Neuroepigenetics of post-traumatic stress disorder. Prog. Mol. Biol. Transl. Sci. 158, 227–253 (2018).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

11. Uddin, M. et al. Epigenetic and immune function profiles associated with posttraumatic stress disorder. Proc. Natl. Acad. Sci. USA 107, 9470–9475 (2010).

    Article  ADS  CAS  PubMed  PubMed Central  Google Scholar 

12. Mehta, D. et al. Childhood maltreatment is associated with distinct genomic and epigenetic profiles in posttraumatic stress disorder. Proc. Natl. Acad. Sci. 110, 8302–8307 (2013).

    Article  ADS  CAS  PubMed  PubMed Central  Google Scholar 

13. Smith, A. K. et al. Differential immune system DNA methylation and cytokine regulation in post‐traumatic stress disorder. Am. J. Med. Genet. Part B: Neuropsychiatr. Genet. 156, 700–708 (2011).

    Article  CAS  Google Scholar 

14. Kuan, P. F. et al. An epigenome-wide DNA methylation study of PTSD and depression in World Trade Center responders. Transl. Psychiatry 7, e1158 (2017).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

15. Rutten, B. P. F. et al. Longitudinal analyses of the DNA methylome in deployed military servicemen identify susceptibility loci for post-traumatic stress disorder. Mol. Psychiatry 23, 1145–1156 (2018).

    Article  CAS  PubMed  Google Scholar 

16. Wolf, E. J. et al. Accelerated DNA methylation age: associations with PTSD and neural integrity. Psychoneuroendocrinology 63, 155–162 (2016).

    Article  CAS  PubMed  Google Scholar 

17. Katrinli, S. et al. Evaluating the impact of trauma and PTSD on epigenetic prediction of lifespan and neural integrity. Neuropsychopharmacology, https://doi.org/10.1038/s41386-020-0700-5 (2020).

18. Wilson, R. et al. The dynamics of smoking-related disturbed methylation: a two time-point study of methylation change in smokers, non-smokers and former smokers. BMC Genomics 18, 805 (2017).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

19. Philibert, R. et al. Dose response and prediction characteristics of a methylation sensitive digital PCR assay for cigarette consumption in adults. Front. Genet. 9, 137 (2018).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

20. Joehanes, R. et al. Epigenetic signatures of cigarette smoking. Circ. Cardiovasc. Genet. 9, 436–447 (2016).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

21. Wheeler, M. A., Rothhammer, V. & Quintana, F. J. Control of immune-mediated pathology via the aryl hydrocarbon receptor. J. Biol. Chem. 292, 12383–12389 (2017).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

22. Myint, A. M. Kynurenines: from the perspective of major psychiatric disorders. FEBS J. 279, 1375–1385 (2012).

    Article  CAS  PubMed  Google Scholar 

23. Kessler, R. C. Posttraumatic stress disorder: the burden to the individual and to society. J. Clin. Psychiatry 61, 4–12 (2000).

    PubMed  Google Scholar 

24. Marzi, S. J. et al. Analysis of DNA methylation in young people: limited evidence for an association between victimization stress and epigenetic variation in blood. Am. J. Psychiatry 175, 517–529 (2018).

    Article  PubMed  PubMed Central  Google Scholar 

25. Logue, M. W. et al. An epigenome-wide association study of posttraumatic stress disorder in US veterans implicates several new DNA methylation loci. Clin. Epigenetics 12, 46 (2020).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

26. Tantoh, D. M. et al. Methylation at cg05575921 of a smoking-related gene (AHRR) in non-smoking Taiwanese adults residing in areas with different PM2.5 concentrations. Clin. Epigenetics 11, 69 (2019).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

27. Julliard, W., Fechner, J. H. & Mezrich, J. D. The aryl hydrocarbon receptor meets immunology: friend or foe? A little of both. Front Immunol. 5, 458 (2014).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

28. De Abrew, K. N., Kaminski, N. E. & Thomas, R. S. An integrated genomic analysis of aryl hydrocarbon receptor-mediated inhibition of B-cell differentiation. Toxicological Sci.: Off. J. Soc. Toxicol. 118, 454–469 (2010).

    Article  CAS  Google Scholar 

29. Goudot, C. et al. Aryl hydrocarbon receptor controls monocyte differentiation into dendritic cells versus macrophages. Immunity 47, 582–596.e586 (2017).

    Article  CAS  PubMed  Google Scholar 

30. Gutierrez-Vazquez, C. & Quintana, F. J. Regulation of the immune response by the aryl hydrocarbon receptor. Immunity 48, 19–33 (2018).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

31. Nuti, R. et al. Ligand binding and functional selectivity of L-tryptophan metabolites at the mouse aryl hydrocarbon receptor (mAhR). J. Chem. Inf. Model 54, 3373–3383 (2014).

    Article  CAS  PubMed  Google Scholar 

32. Esser, C. & Rannug, A. The aryl hydrocarbon receptor in barrier organ physiology, immunology, and toxicology. Pharmacol. Rev. 67, 259–279 (2015).

    Article  CAS  PubMed  Google Scholar 

33. Zielinska, E., Kuc, D., Zgrajka, W., Turski, W. A. & Dekundy, A. Long-term exposure to nicotine markedly reduces kynurenic acid in rat brain–in vitro and ex vivo evidence. Toxicol. Appl. Pharm. 240, 174–179 (2009).

    Article  CAS  Google Scholar 

34. Nguyen, N. T. et al. Aryl hydrocarbon receptor and kynurenine: recent advances in autoimmune disease research. Front. Immunol. 5, 551 (2014).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

35. Passos, I. C. et al. Inflammatory markers in post-traumatic stress disorder: a systematic review, meta-analysis, and meta-regression. Lancet Psychiatry 2, 1002–1012 (2015).

    Article  PubMed  Google Scholar 

36. Schmidt, J. V. & Bradfield, C. A. Ah receptor signaling pathways. Annu. Rev. Cell Dev. Biol. 12, 55–89 (1996).

    Article  CAS  PubMed  Google Scholar 

37. Magnusson, R., Hagglund, L. & Wingfors, H. Broad exposure screening of air pollutants in the occupational environment of Swedish soldiers deployed in Afghanistan. Mil. Med. 177, 318–325 (2012).

    Article  PubMed  Google Scholar 

38. Lippmann, M., Cohen, M. D. & Chen, L. C. Health effects of World Trade Center (WTC) Dust: an unprecedented disaster’s inadequate risk management. Crit. Rev. Toxicol. 45, 492–530 (2015).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

39. National Academies of Sciences, Engineering, and Medicine. Assessment of the Department of Veterans Affairs Airborne Hazards and Open Burn Pit Registry. (Washington, DC: The National Academies Press, 2017). https://doi.org/10.17226/23677.

40. Qureshi, S. U., Pyne, J. M., Magruder, K. M., Schulz, P. E. & Kunik, M. E. The link between post-traumatic stress disorder and physical comorbidities: a systematic review. Psychiatr. Q 80, 87–97 (2009).

    Article  PubMed  Google Scholar 

41. Boscarino, J. A. Posttraumatic stress disorder and physical illness: results from clinical and epidemiologic studies. Ann. N. Y Acad. Sci. 1032, 141–153 (2004).

    Article  ADS  PubMed  Google Scholar 

42. Stein, M. B. et al. Genome-wide association studies of posttraumatic stress disorder in 2 cohorts of US army soldiers. JAMA Psychiatry 73, 695–704 (2016).

    Article  PubMed  PubMed Central  Google Scholar 

43. Sommershof, A. et al. Substantial reduction of naive and regulatory T cells following traumatic stress. Brain Behav. Immun. 23, 1117–1124 (2009).

    Article  CAS  PubMed  Google Scholar 

44. Jergovic, M. et al. Patients with posttraumatic stress disorder exhibit an altered phenotype of regulatory T cells. Allergy Asthma Clin. Immunol. 10, 43 (2014).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

45. Zhou, J. et al. Dysregulation in microRNA expression is associated with alterations in immune functions in combat veterans with post-traumatic stress disorder. PLoS ONE 9, e94075 (2014).

    Article  ADS  PubMed  PubMed Central  CAS  Google Scholar 

46. Morath, J. et al. The effect of trauma-focused therapy on the altered T cell distribution in individuals with PTSD: evidence from a randomized controlled trial. J. Psychiatr. Res. 54, 1–10 (2014).

    Article  PubMed  Google Scholar 

47. Lee, B. et al. Effects of systemic administration of ibuprofen on stress response in a rat model of post-traumatic stress disorder. Korean J. Physiol. Pharm. 20, 357–366 (2016).

    Article  CAS  Google Scholar 

48. Gamble-George, J. C. et al. Cyclooxygenase-2 inhibition reduces stress-induced affective pathology. Elife 5, https://doi.org/10.7554/eLife.14137 (2016).

49. Ratanatharathorn, A. et al. Epigenome-wide association of PTSD from heterogeneous cohorts with a common multi-site analysis pipeline. Am. J. Med. Genet. B Neuropsychiatr. Genet. 174, 619–630 (2017).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

50. Teschendorff, A. E. et al. A beta-mixture quantile normalization method for correcting probe design bias in Illumina Infinium 450 k DNA methylation data. Bioinformatics 29, 189–196 (2013).

    Article  CAS  PubMed  Google Scholar 

51. Houseman, E. A. et al. DNA methylation arrays as surrogate measures of cell mixture distribution. BMC Bioinforma. 13, 1 (2012).

    Article  Google Scholar 

52. Reinius, L. E. et al. Differential DNA methylation in purified human blood cells: implications for cell lineage and studies on disease susceptibility. PLoS ONE 7, e41361 (2012).

    Article  ADS  CAS  PubMed  PubMed Central  Google Scholar 

53. Barfield, R. T. et al. Accounting for population stratification in DNA methylation studies. Genet Epidemiol. 38, 231–241 (2014).

    Article  PubMed  PubMed Central  Google Scholar 

54. Smyth G. K. limma: Linear Models for Microarray Data. In Bioinformatics and Computational Biology Solutions Using R and Bioconductor. Statistics for Biology and Health (eds. Gentleman, R., Carey, V. J., Huber, W., Irizarry, R. A. & Dudoit, S.) (Springer, New York, NY, 2005). https://doi.org/10.1007/0-387-29362-0_23.

55. Saffari, A. et al. Estimation of a significance threshold for epigenome-wide association studies. Genet Epidemiol. 42, 20–33 (2018).

    Article  PubMed  Google Scholar 

56. Hachiya, T. et al. Genome-wide identification of inter-individually variable DNA methylation sites improves the efficacy of epigenetic association studies. NPJ Genom. Med. 2, 11 (2017).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

57. Li, K., Naviaux, J. C., Bright, A. T., Wang, L. & Naviaux, R. K. A robust, single-injection method for targeted, broad-spectrum plasma metabolomics. Metabolomics: Off. J. Metabolomic Soc. 13, 122 (2017).

    Article  CAS  Google Scholar 

58. Naviaux, R. K. et al. Metabolic features of chronic fatigue syndrome. Proc. Natl. Acad. Sci. USA 113, E5472–E5480 (2016).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

59. Baker, D. G. et al. Predictors of Risk and Resilience for Posttraumatic Stress Disorder Among Ground Combat Marines: Methods of the Marine Resiliency Study. Preventing Chronic Disease. Centers for Disease Control and Prevention (Atlanta, GA, 2012).

Download references

Authors and Affiliations

1. Emory University, Department of Gynecology and Obstetrics, Atlanta, GA, USA

   Alicia K. Smith, Seyma Katrinli & Varun Kilaru

2. Emory University, Department of Psychiatry & Behavioral Sciences, Atlanta, GA, USA

   Alicia K. Smith, Adriana Lori, Kerry J. Ressler & Kerry J. Ressler

3. Columbia University, Department of Epidemiology, New York, NY, USA

   Andrew Ratanatharathorn

4. University of California San Diego, Department of Psychiatry, La Jolla, CA, USA

   Adam X. Maihofer, Dewleen G. Baker, Victoria B. Risbrough, Murray B. Stein, Murray Stein, Ariel Lang & Caroline M. Nievergelt

5. University of California, The Mitochondrial and Metabolic Disease Center, Departments of Medicine, Pediatrics, and Pathology, San Diego, CA, USA

   Robert K. Naviaux

6. University of North Carolina, Gillings School of Global Public Health, Department of Epidemiology, Chapel Hill, NC, USA

   Allison E. Aiello

7. Virginia Commonwealth University, Department of Psychiatry, Richmond, VA, USA

   Ananda B. Amstadter & Scott D. McDonald

8. Duke Molecular Physiology Institute, Duke University Medical Center, Durham, NC, USA

   Allison E. Ashley-Koch, Melanie E. Garrett, Michael A. Hauser, Xuejun Qin, Xue-Jun Qin, XueJun Qin & Kerry J. Ressler

9. Veterans Affairs San Diego Healthcare System, San Diego, CA, USA

   Dewleen G. Baker, Victoria B. Risbrough, Murray B. Stein, Murray Stein & Caroline M. Nievergelt

10. Veterans Affairs Center of Excellence for Stress and Mental Health, San Diego, CA, USA

    Dewleen G. Baker, Victoria B. Risbrough & Caroline M. Nievergelt

11. VA Mid-Atlantic, Mental Illness Research, Education, and Clinical Center, Durham, NC, USA

    Jean C. Beckham, Michelle Dennis, Nathan A. Kimbrel & Christine Marx

12. Veterans Affairs Durham Healthcare System, Durham, NC, USA

    Jean C. Beckham, Nathan A. Kimbrel, Christine Marx, Mira Brancu, Patrick S. Calhoun, Eric Dedert, Eric B. Elbogen, John A. Fairbank, Jason D. Kilts, Angela Kirby, Christine E. Marx, Scott D. Moore, Rajendra A. Morey, Jennifer C. Naylor, Cindy Swinkels, Steven T. Szabo, Larry A. Tupler & Elizabeth E. Van Voorhees

13. Duke University Medical Center, Department of Psychiatry and Behavioral Sciences, Durham, NC, USA

    Jean C. Beckham, Michelle Dennis, Nathan A. Kimbrel, Christine Marx & Anthony Zannas

14. University Medical Center Utrecht, Brain Center Rudolf Magnus, Utrecht, The Netherlands

    Marco P. Boks, Elbert Geuze, Eric Vermetten, Danny Nispeling & Dick Schijven

15. State University of New York at Stony Brook, Epidemiology Research Group, Stony Brook, NY, USA

    Evelyn Bromet

16. Boston University, School of Public Health, Boston, MA, USA

    Sandro Galea

17. Netherlands Ministry of Defence, Brain Research and Innovation Centre, Utrecht, The Netherlands

    Elbert Geuze

18. McLean Hospital, Division of Depression and Anxiety, Belmont, MA, USA

    Guia Guffanti & Kerry J. Ressler

19. Harvard Medical School, Boston, MA, USA

    Guia Guffanti, Ronald C. Kessler, Kerry J. Ressler, Nikolaos P. Daskalakis & Kerry J. Ressler

20. Duke University, Department of Medicine, Durham, NC, USA

    Michael A. Hauser, Mira Brancu, Patrick S. Calhoun, Eric Dedert, Eric B. Elbogen, John A. Fairbank, Jason D. Kilts, Christine E. Marx, Scott D. Moore, Rajendra A. Morey, Jennifer C. Naylor, Cindy Swinkels, Steven T. Szabo, Larry A. Tupler & Elizabeth E. Van Voorhees

21. Harvard T.H. Chan School of Public Health, Department of Epidemiology, Boston, MA, USA

    Karestan C. Koenen & Archana Basu

22. Massachusetts General Hospital, Psychiatric and Neurodevelopmental Genetics Unit, Center for Human Genetic Research, and Department of Psychiatry, Boston, MA, USA

    Karestan C. Koenen

23. Broad Institute of MIT and Harvard, Stanley Center for Psychiatric Research, Cambridge, MA, USA

    Karestan C. Koenen

24. State University of New York at Stony Brook, Department of Applied Mathematics and Statistics, Stony Brook, NY, USA

    Pei-Fen Kuan & Pei Fen Kuan

25. University of California, The Mitochondrial and Metabolic Disease Center, Department of Medicine, San Diego, CA, USA

    Kefeng Li & Lin Wang

26. National Center for PTSD, Behavioral Science Division at VA Boston Healthcare System, Boston, MA, USA

    Mark W. Logue, Mark W. Miller & Erika J. Wolf

27. Boston University School of Medicine, Department of Psychiatry, Boston, MA, USA

    Mark W. Logue, Mark W. Miller & Erika J. Wolf

28. Boston University School of Medicine, Department of Medicine (Biomedical Genetics), Boston, MA, USA

    Mark W. Logue

29. Boston University School of Public Health, Department of Biostatistics, Boston, MA, USA

    Mark W. Logue

30. State University of New York at Stony Brook, Department of Medicine, Stony Brook, NY, USA

    Benjamin J. Luft

31. University of California, The Mitochondrial and Metabolic Disease Center, Department of Neuroscience, San Diego, CA, USA

    Jane C. Naviaux

32. Brown University, Psychiatry and Human Behavior, Department of Pediatric Research, Providence, RI, USA

    Nicole R. Nugent, Leslie A. Brick & Audrey Tyrka

33. Maastricht University Medical Centre, School for Mental Health and Neuroscience, Department of Psychiatry and Neuropsychology, Maastricht, The Netherlands

    Bart P. F. Rutten

34. University of California San Diego, Department of Family Medicine and Public Health, La Jolla, CA, USA

    Murray B. Stein, Murray Stein & Sonia Jain

35. Uniformed Services University School of Medicine, Center for the Study of Traumatic Stress, Bethesda, MD, USA

    Robert J. Ursano & Eric Vermetten

36. Leiden University Medical Center, Department of Psychiatry, Leiden, The Netherlands

    Eric Vermetten

37. Netherlands Defense Department, Research Center, Utrecht, UT, The Netherlands

    Eric Vermetten

38. Arq Psychotrauma Expert Group, Diemen, The Netherlands

    Eric Vermetten

39. Amsterdam UMC (location VUmc), Department of Psychiatry, Amsterdam, The Netherlands

    Christiaan H. Vinkers & Mirjam Van Zuiden

40. Amsterdam UMC (location VUmc), Department of Anatomy and Neurosciences, Amsterdam, The Netherlands

    Christiaan H. Vinkers

41. Medical College of Georgia at Augusta University, Department of Psychiatry and Human Behavior and Office of Academic Affairs, Augusta, GA, USA

    Nagy A. Youssef

42. University of South Florida, College of Public Health, Tampa, FL, USA

    Agaz Wani & Monica Uddin

43. Department of Neurosurgery, Stanford University Medical Center, Stanford, CA, USA

    Gerry Grant

44. Department of Psychiatry, Indiana University, Indianapolis, IN, USA

    Thomas W. McAllister

45. Spaulding Rehabilitation Hospital and Harvard Medical School, Charleston, MA, USA

    Ross Zafonte

46. Riverside University Health System Medical Center, Moreno Valley, CA, USA

    Raul Coimbra

47. University of Louisville School of Medicine, Louisville, KY, USA

    Norberto Andaluz

48. University of Pittsburgh School of Medicine, Pittsburgh, PA, USA

    Lori Shutter

49. Medical University of South Carolina, Ralph H. Johnson VA Medical Center, Charleston, SC, USA

    Mark S. George

50. Salisbury Veterans Affairs Health Care System, Salisbury, NC, USA

    Robin A. Hurley, Jared A. Rowland, Katherine H. Taber & Ruth E. Yoash-Gantz

51. Wake Forest University, Winston-Salem, NC, USA

    Robin A. Hurley, Jared A. Rowland & Ruth E. Yoash-Gantz

52. Baylor College of Medicine, Houston, TX, USA

    Robin A. Hurley & Katherine H. Taber

53. Richmond Veterans Affairs Health Care System, Richmond, VA, USA

    Scott D. McDonald

54. Virginia College of Osteopathic Medicine, Blacksburg, VA, USA

    Katherine H. Taber

55. University of Cape Town, SA MRC Unit on Risk & Resilience in Mental Disorders, Department of Psychiatry, Cape Town, ZA, South Africa

    Shareefa Dalvie & Nastassja Koen

56. Cohen Veterans Bioscience, Cambridge, USA

    Nikolaos P. Daskalakis

57. Icahn School of Medicine at Mount Sinai, Department of Psychiatry, New York, NY, USA

    Nikolaos P. Daskalakis

58. University of Amsterdam, Department of Child and Adolescent Psychiatry, Amsterdam Public Health Research Institute, Amsterdam UMC, Location Academic Medical Center, Amsterdam, The Netherlands

    Judith B. M. Ensink

59. Academic Center for Child and Adolescent Psychiatry, De Bascule, Amsterdam, The Netherlands

    Judith B. M. Ensink

60. University of Amsterdam, Department of Clinical Epidemiology, Biostatistics and Bioinformatics, Amsterdam Public Health Research Institute, Amsterdam UMC, Location Academic Medical Center, Amsterdam, The Netherlands

    Judith B. M. Ensink

61. Stellenbosch University Faculty of Medicine and Health Sciences, Department of Psychiatry, Cape Town, ZA, South Africa

    Sian M. J. Hemmings & Soraya Seedat

62. University of Wisconsin School of Medicine and Public Health, Department of Psychiatry, Madison, WI, USA

    Ryan Herringa

63. University of Aberdeen, Department of Applied Medicine, Aberdeen, UK

    Sylvanus Ikiyo

64. Yale University School of Medicine, Division of Human Genetics, Department of Psychiatry, New Haven, CT, USA

    Janitza Montalvo-Ortiz

65. VA CT Healthcare Center, West Haven, CT, USA

    Janitza Montalvo-Ortiz

66. University of Illinois at Urbana-Champaign, Department of Psychology, Urbana, IL, USA

    John Pfeiffer

67. University of Iowa Carver College of Medicine, University of Iowa Hospitals and Clinics, Department of Psychiatry, Iowa City, IA, USA

    Gen Shinozaki

68. Columbia University Medical Center, Department of Medicine, New York, NY, USA

    Jennifer A. Sumner

69. University of Cape Town, Department of Human Biology, Cape Town, ZA, South Africa

    Patricia Swart

70. Butler Hospital, Mood Disorders Research Program and Laboratory for Clinical and Translational Neuroscience, Providence, RI, USA

    Audrey Tyrka

71. University of North Carolina, Department of Psychiatry, Chapel Hill, NC, USA

    Anthony Zannas

72. University of North Carolina, Department of Genetics, Chapel Hill, NC, USA

    Anthony Zannas

73. University of North Carolina School of Medicine, Institute for Trauma Recovery, Chapel Hill, NC, USA

    Anthony Zannas

74. University of North Carolina, Neuroscience Curriculum, Chapel Hill, NC, USA

    Anthony Zannas

Consortia

Contributions

Interpreted results, writing, and editing the manuscript: A.K.S., A.R., A.X.M., R.K.N., M.W.L., M.U., C.M.N. Conceptualization and supervision of project: A.K.S., M.W.L., M.U., C.M.N. Performed meta-analyses: A.R., A.X.M., S.K. Performed metabolomics analyses: A.X.M., R.K.N., K.L., J.C.N. Contributed to study-specific EWAS by providing phenotype, genotype and performing data analyses: A.K.S., A.R., A.X.M., A.E.A., A.B.A., A.E.A.K., D.G.B., J.C.B., M.P.B., E.B., M.D., S.G., M.E.G., E.G., G.G., M.A.H., S.K., V.K., R.C.K., N.A.K., K.C.K., P.F.K., M.W.L., A.L., B.J.L., M.W.M., N.R.N., X.Q., K.J.R., V.B.R., B.P.F.R., M.B.S., R.J.U., E.V., C.H.V., L.W., N.A.Y., M.U., C.M.N. Unless otherwise stated, INTRuST Clinical Consortium, V.A. Mid-Atlantic MIRECC Workgroup and PGC-PTSD Epigenetics Workgroup authors contributed to data collection but did not participate in analysis or writing of this report. All other authors reviewed and approved the final version of the paper.

Corresponding author

Correspondence to Alicia K. Smith.

Competing interests

Dr. Youssef’s disclosures include Speaker CME honoraria from the Georgia Department of Behavioral Health and Developmental Disabilities (DHBDD). Dr. Youssef received research support from the Department of Veteran Affairs and The Augusta Biomedical Research Corporation in the last 3 years. His current research funding (but not direct payment) include Merck pharmaceuticals (8S0073-9), and MECTA Corporation (72200S-2) and American Foundation for Suicide Prevention. Dr. Stein has in the past 3 years received payments for editorial work from UpToDate, Biological Psychiatry, and Depression and Anxiety. He has also in the past 3 years been paid as a consultant for Actelion Pharmaceuticals, Aptinyx, Bionomics, Janssen, and Pfizer. No other author declares any conflict of interest.

Peer review information Nature Communications thanks Frank Dudbridge, Anthony Zannas and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions