ARCHIVAL REPORT

Lower Methylation of Glucocorticoid Receptor Gene Promoter 1F in Peripheral Blood of Veterans with Posttraumatic Stress Disorder Rachel Yehuda, Janine D. Flory, Linda M. Bierer, Clare Henn-Haase, Amy Lehrner, Frank Desarnaud, Iouri Makotkine, Nikolaos P. Daskalakis, Charles R. Marmar, and Michael J. Meaney Background: Enhanced glucocorticoid receptor (GR) sensitivity is present in people with posttraumatic stress disorder (PTSD), but the molecular mechanisms of GR sensitivity are not understood. Epigenetic factors have emerged as one potential mechanism that account for how trauma exposure leads to sustained PTSD symptoms given that PTSD develops in only a subset of trauma survivors. Methods: Cytosine methylation of a relevant promoter of the GR gene (NR3C1-1F promoter) and three functional neuroendocrine markers of hypothalamic-pituitary-adrenal axis function were examined in a sample of 122 combat veterans. Results: Lower NR3C1-1F promoter methylation in peripheral blood mononuclear cells (PBMCs) was observed in combat veterans with PTSD compared with combat-exposed veterans who did not develop PTSD. NR3C1-1F promoter methylation was also associated with three functional measures of glucocorticoid activity that have been associated with PTSD in combat veterans: PBMCs’ lysozyme inhibition on the lysozyme suppression test, plasma cortisol decline on the low-dose (.50 mg) dexamethasone suppression test, and 24-hour urinary cortisol excretion. Finally, NR3C1-1F promoter methylation was inversely correlated with clinical markers and symptoms associated with PTSD. Conclusions: Alterations in NR3C1-1F promoter methylation may reflect enduring changes resulting from combat exposure that lead to functional neuroendocrine alterations. Because epigenetic measures are thought to reflect enduring effects of environmental exposures, they may be useful in distinguishing combat-exposed veterans who do or do not develop PTSD. Key Words: Biomarkers, combat, cortisol, glucocorticoid receptor, HPA axis, methylation, NR3C1 gene, NR3C1-1F promoter, PTSD

C

ytosine methylation is an epigenetic mechanism that can mediate the sustained effects of environmental conditions on genomic transcription, moderating genetic predispositions (1). In rodents (2) and humans (3), early environmental experience is associated with hippocampal methylation of the glucocorticoid receptor (GR) gene (NR3C1) exon 17 promoter (rat) and its human ortholog, the exon 1F promoter, and with regulation of the hypothalamic-pituitary-adrenal (HPA) axis. These findings are of interest for the study of posttraumatic stress disorder (PTSD), which has been related to enhanced GR sensitivity, low From the Mental Health Care Center (RY, JDF, LMB, AL, FD, IM, NPD), James J. Peters Veterans Affairs Medical Center, Bronx, New York; Traumatic Stress Studies Division, Department of Psychiatry (RY, JDF, LMB, AL, FD, IM, NPD), and Department of Neuroscience (RY), Icahn School of Medicine at Mount Sinai, New York, New York; Steven and Alexandra Cohen Veterans Center for the Study of Posttraumatic Stress and Traumatic Brain Injury, Department of Psychiatry, New York University School of Medicine (CH-H, CRM), New York, New York; Sackler Program for Epigenetics Psychobiology and Departments of Psychiatry and Neurology & Neurosurgery (MJM), McGill University, Montreal, Canada; and Singapore Institute for Clinical Sciences (MJM), Singapore, Singapore. Address correspondence to Rachel Yehuda, Ph.D., Traumatic Stress Studies Division, Department of Psychiatry, Icahn School of Medicine at Mount Sinai, James J. Peters Veterans Affairs Medical Center, 526 OOMH PTSD 116/A, 130 West Kingsbridge Road, Bronx, New York 10468; E-mail: [email protected]. Received Nov 26, 2013; revised Jan 17, 2014; accepted Feb 6, 2014.

0006-3223/$36.00 http://dx.doi.org/10.1016/j.biopsych.2014.02.006

glucocorticoid levels, and a history of childhood abuse and neglect. Epigenetic signals associated with early life experiences offer a potential explanation for why stress responses do not abate in certain individuals after the stressor is removed and for the fact that not all people exposed to trauma develop PTSD (4). In Dutch soldiers who served in Afghanistan, higher GR levels in peripheral blood mononuclear cells (PBMCs) before deployment predicted PTSD (5). The goal of the current study was to examine methylation of the NR3C1-1F promoter in combat veterans with or without PTSD who served in Operation Enduring Freedom or Operation Iraqi Freedom or both. We hypothesized that NR3C1-1F promoter methylation would be associated with functional markers of glucocorticoid sensitivity. Because PTSD is associated with enhanced GR sensitivity, greater feedback regulation, and lower basal HPA axis tone (e.g., lower ambient cortisol levels), we predicted lower GR promoter methylation in combat veterans with PTSD (PTSD⫹) compared to veterans with similar combat exposure without PTSD (PTSD
). Three functional and relatively stable neuroendocrine measures previously demonstrated to vary in association with PTSD—the lysozyme suppression test (an in vitro measure of GR sensitivity; see Methods), cortisol decline in response to a low-dose (.50 mg) dexamethasone (DEX) suppression test, and 24-hour urinary cortisol excretion—were examined to determine their association with PTSD and NR3C1-1F promoter methylation. We hypothesized lower lysozyme IC50-DEX and a greater cortisol decline after DEX administration, both reflecting increased GR sensitivity, would occur in combat veterans with PTSD. NR3C1-1F promoter methylation regulates transcriptional capacity in response to transcription factor binding (3). We hypothesized that NR3C1-1F promoter methylation would be inversely associated with NR3C1-1F expression and GR functional outcomes such that lower methylation would be associated with indices reflecting enhanced GR sensitivity. Additionally, because

BIOL PSYCHIATRY 2014;]:]]]–]]] Published by Elsevier Inc on behalf of Society of Biological Psychiatry

2 BIOL PSYCHIATRY 2014;]:]]]–]]] higher GR sensitivity is thought to underpin lower ambient 24hour cortisol excretion, frequently observed in individuals with combat-related PTSD, we further hypothesized an association between NR3C1-1F promoter methylation and basal urinary cortisol.

Methods and Materials Participants and Clinical Assessment This article reports on data from 122 participants (61 with PTSD, 61 without PTSD) in whom NR3C1-1F promoter methylation and gene expression was determined in PBMCs, in addition to other neuroendocrine biomarkers. Participants were recruited as part of a larger study that was designed as an initial investigation of male combat veterans, to be followed by a replication and longitudinal validation study in men and women. Combat veterans were recruited at two sites, the James J. Peters Veterans Affairs Medical Center affiliated with Icahn School of Medicine at Mount Sinai and New York University Langone Medical Center affiliated with New York University School of Medicine through advertising in the clinic (Veterans Affairs Medical Center) and community (New York University). The study was approved by the institutional review boards of the James J. Peters Veterans Affairs Medical Center, Icahn School of Medicine at Mount Sinai, New York University Langone Medical Center, and New York University School of Medicine; all participants provided written, informed consent. The presence versus absence of a PTSD diagnosis was determined by doctoral-level psychologists using the Clinician Administered PTSD Scale (CAPS) (6). Because it was of interest to evaluate biomarkers that distinguished effects of PTSD from effects of exposure, all participants had exposure to a combatrelated criterion A trauma. Participants in the group with war zone exposure and no PTSD diagnosis had current CAPS scores #20 and had never met criteria for PTSD. Participants in the PTSD group were required to have a CAPS score $40 and to meet full DSM-IV criteria for war zone–related PTSD; combat veterans with CAPS scores between 20 and 40 were not included in the study. The Structured Clinical Interview for DSM-IV (7) was used by the same clinician to determine other DSM-IV diagnoses including substance use disorders. All participants remained eligible if they met criteria for a current or lifetime mood or anxiety disorder other than PTSD; all subjects with lifetime history of any psychiatric disorder with psychotic features, bipolar disorder, or obsessive-compulsive disorder; prominent suicidal or homicidal ideation; or a suicide attempt in the past year were excluded. Participants with current alcohol dependence or a current drug abuse or dependence diagnosis were also excluded from participation. Diagnostic information was reviewed in a weekly teleconference including all diagnostic interviewers, which ensured a similar use of diagnostic measures and convergent application of inclusion and exclusion criteria between the two sites. Medical exclusions included neurologic disorder, loss of consciousness ⬎10 min, or other systemic illness affecting central nervous system function. Relevant self-reported clinical assessments were obtained to determine severity of PTSD [with the Posttraumatic Stress Checklist (8) and depressive symptoms [with the Beck Depression Inventory (9)]. The Early Trauma Inventory (10) was used to assess childhood trauma exposure. Other measures included the Pittsburgh Sleep Quality Index (11), the Peritraumatic Dissociative Experiences Questionnaire (to assess the extent of dissociation at the time of the focal deployment trauma) (12), and the Symptom Checklist-90-R (13). www.sobp.org/journal

R. Yehuda et al. Biological Methods Blood was drawn before and after .5 mg DEX ingestion between 8:00 AM and 8:30 AM. An aliquot of whole blood was reserved from the day 1 blood sample for measurement of leukocyte count and cell type distribution. Plasma samples were collected into ethylenediamine tetraacetate–containing tubes, and an aliquot was frozen for subsequent hormone analysis. Using ACCUSPIN tubes (Sigma-Aldrich, St. Louis, Missouri), PBMCs were purified by Ficoll-Paque (Amersham, Piscataway, New Jersey). After two washes in Hanks’ Balanced Salt Solution (Life Technologies, Grand Island, New York), PBMCs were counted with a hemocytometer. The purified PBMCs were divided into aliquots for DNA extraction and cell culture for determination of the IC50-DEX of lysozyme inhibition. Assessment of the distribution of white blood cell populations was performed in a clinical certified (Clinical Laboratories Improvement Act) laboratory. Because cell type composition in PBMC samples covaries with DNA methylation (14), the PBMC ratio (lymphocyte-to-monocyte ratio), a proxy of PBMC type, was used as a covariate in DNA methylation analyses. Participants were also instructed to collect 24-hour urine samples (15), and completeness of collections was determined by participants’ report and assessment of creatinine levels. Urine volumes ⬍500 mL were deemed incomplete, and results were not used in subsequent analyses. Five main variables were measured: 1) NR3C1-1F promoter methylation; 2) NR3C1-1F expression; 3) in vitro sensitivity to DEX of lysozyme, a GR-regulated enzyme, assessed in cultured PBMCs; 4) cortisol decline in response to .50 mg DEX; and 5) 24-hour urinary cortisol excretion. DNA Cytosine Methylation of NR3C1-1F Promoter. Genomic DNA was extracted from the frozen PBMC pellets following the FlexiGene DNA Kit protocol (Qiagen, Valencia, California). Methylation mapping of the 39 C-phosphate-G (CpG) sites in the NR3C1-1F promoter was performed as previously described (3,16) using 30 clones per sample. Briefly, sodium bisulfite conversion was carried out according to the EpiTect Bisulfite Kit protocol (Qiagen) using 0.8 mg of genomic DNA in each conversion reaction and 0.8 mg of Universal Methylated DNA Standard (Zymo Research, Orange, California) to check completion of the reaction. The genomic region of the human GR exon 1F promoter was subjected to polymerase chain reaction (PCR) amplification using the following primer sequences, per previously published procedures (3,16): 50 -GTGGTGGGGGATTTG-30 (forward); 50 -ACCTAATCTCTCTAAAAC-30 (reverse). The resulting PCR product was subjected to another round of PCR, using the following nested primers, per previously published procedures (3,16): 50 -TTTTTGAAGTTTTTTTAGAGGG-30 (forward); 50 -AATTTCTCCAATTTCTTTTCTC-30 (reverse). The resulting PCR products were analyzed on a 2% agarose gel and purified using QIAquick PCR Purification Kit (Qiagen). The PCR products were subcloned using a PCR product cloning kit (Qiagen, Carlsbad, California), and individual plasmids containing the ligated promoter regions were extracted and sequenced (GENEWIZ, South Plainfield, New Jersey). The sequences for 30 individual clones were aligned and analyzed in the DNA alignment software BioEdit (Ibis Biosciences, Carlsbad, California). The DNA samples were analyzed in batches of 20–30 samples. Variability in the DNA bisulfite treatment did not exceed 2% between the batches. Two measures were calculated: 1) the number of clones with at least one methylated CpG site divided by the total number of clones was calculated to yield an estimate of the percentage of methylated clones, and the number of methylated clones (out of 30) at each of the 39 CpG sites was converted to a percentage, and 2) percentages across sites were

R. Yehuda et al.

BIOL PSYCHIATRY 2014;]:]]]–]]] 3

summed to create a total percentage of methylation across the NR3C1-1F promoter sequence. NR3C1-1F Expression. NR3C1-1F expression was measured as previously described (3,16). A portion of PBMCs was dissolved in TRIzol Reagent (Life Technologies) by adding 1 mL of the reagent per 1  107 cells, quickly frozen, stored at
801C, and later used for RNA extraction using PureLink RNA Mini Kit (Life Technologies) with on-column deoxyribonucleic acid digestion (Qiagen). The quality and yield of RNA was determined using NanoDrop 1000 (Thermo Fisher Scientific, Wilmington, Delaware). Complementary DNA synthesis was completed using Maxima Reverse Transcriptase (Thermo Fisher Scientific, West Palm Beach, Florida) and GR target oligo (CAGGGGTGCAGAGTTCGATG). Quantitative real-time PCR was performed with a LightCycler 480 System (Roche Applied Science, Indianapolis, Indiana) and RT2 PCR primer set for beta-2microglobulin gene (B2M) (Catalog no. PPH01094E; SABiosciences, Qiagen) and Nr3c1-1F (forward primer 50 -AAGAAACTGGAGAAACTCGGTGGC-30 , reverse primer 50 -TGAGGGTGAAGACGCAGAAACCTT-30 ). All reactions were done in triplicate and only one complementary DNA was amplified in each PCR (monoplex). The relative expression of Nr3c1-1F expression over B2M was calculated using the 2−ΔΔCt method and used for subsequent analyses. Lysozyme Suppression Test. The determination of the IC50-DEX for lysozyme inhibition was performed as previously described (17) by incubating cultured PBMCs with varying doses of DEX. The lysozyme IC50-DEX represents the concentration of DEX at which there is 50% inhibition of lysozyme activity and constitutes a measure of glucocorticoid sensitivity in PBMCs. Hormone Determination for DEX Suppression Test and 24-Hour Urine Collection. Cortisol and DEX levels were determined by radioimmunoassay as previously described (18). The intra-assay and inter-assay coefficients of variation were 2.3% and

6.1% for cortisol and 8.0% and 9.0% for DEX. Radioimmunoassay was also used for assay of 24-hour urinary cortisol as previously described (18). Statistical Analysis Cases and controls were selected from a larger pool based on comparability in age, race, and ethnicity. Group comparisons (PTSD vs. no-PTSD groups) were conducted using one-way analysis of variance (ANOVA) or analysis of covariance for continuous variables and χ2 analyses for categorical variables. The distributions of the raw data of NR3C1-1F promoter methylation and expression were normalized using natural log transformation before analysis. Results of analyses using transformed values are presented with the raw data shown in tables and figures for ease of interpretation. A multivariate ANOVA was also performed to investigate potential group differences in CpG sites. Methylation levels of all 39 sites were entered into the multivariate ANOVA because the associations among the dependent variables were assumed to be moderately correlated. Bivariate and partial correlations were used to assess the association of NR3C1-1F promoter methylation with neuroendocrine and clinical measures. The PBMC ratio was used as a covariate for the analyses using variables derived from PBMCs (methylation and expression measures and lysozyme IC50-DEX). Statistical significance for all analyses was set at p ⬍ .05.

Results Demographic and Clinical Information Table 1 presents demographic and clinical characteristics of the PTSD⫹ and PTSD
groups. Because the sample was intentionally selected to be comparable in age, race, and ethnicity,

Table 1. Comparison of Demographic and Clinical Characteristics For PTSD and Matched, Trauma Exposed, Non-PTSD Comparison Group

Age (Years) Race/Ethnicity Hispanic Non-Hispanic White Non-Hispanic Black Non-Hispanic Other Early Trauma Exposurea PTSD Severity (Clinician-Rated)b PTSD Severity (Self-Reported)c Depression Severityd Poor Sleep Qualitye Peritraumatic Dissociationf Psychiatric Distressg BMI (kg/m2) Lymphocyte Number (1000/μL) Monocyte Number (1000/μL) PBMC Ratio DEX Levels

PTSD (n ¼ 61) Mean (SD) or %

Non-PTSD (n ¼ 61) Mean (SD) or %

34.2 (8.8)

33.0 (8.6)

47.5% 26.2% 26.2% .0% 7.1 (5.6) 67.9 (16.45) 61.1 (11.7) 24.1 (10.8) 13.5 (3.3) 1.9 (.9) 2.3 (.6) 30.06 (5.1) 1.9 (.6) .5 (.1) 4.3 (1.5) 124.4 (92.5)

32.8% 37.3% 26.2% 3.3% 5.4 (4.3) 3.3 (4.9) 25.6 (8.2) 5.7 (6.3) 6.2 (4.1) .5 (.6) 1.4 (.4) 28.49 (4.8) 1.8 (.5) .4 (.2) 4.4 (1.4) 107.1 (75.7)

F1,120 ¼ .551 χ2 (3) ¼ 4.91

F1,119 F1,120 F1,119 F1,118 F1,117 F1,119 F1,119 F1,121 F1,118 F1,118 F1,118 F1,112

¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼

3.27 863.4 376.5 131.2 116.2 92.1 100.59 3.22 1.22 1.26 .073 1.19

NS NS

p p p p p p

NS ⬍ .0005 ⬍ .0005 ⬍ .0005 ⬍ .0005 ⬍ .0005 ⬍ .0005 NS NS NS NS NS

BMI, body mass index; DEX, dexamethasone; NS, not significant; PBMC, peripheral blood mononuclear cell; PTSD, posttraumatic stress disorder. Early Trauma Inventory (10). b Clinician Administered PTSD Scale (6). c Military Version of the PTSD Checklist (8). d Beck Depression Inventory-II (9). e Pittsburgh Sleep Quality Index (11). f Peritraumatic Dissociative Experiences Questionnaire (12). g Symptom Checklist-90-R (13). a

www.sobp.org/journal

4 BIOL PSYCHIATRY 2014;]:]]]–]]] there were no group differences for these variables. Early trauma exposure scores (on the Early Trauma Inventory) ranged from 0
21, and the groups did not differ. The PTSD⫹ group reported higher scores on measures of PTSD-related symptoms (Table 1). There were no differences in the lymphocyte or monocyte numbers or their ratio between the PTSD⫹ and PTSD
groups (Table 1). Among the PTSD⫹ sample, 52.5% of subjects met criteria for comorbid major depressive disorder (MDD), and 21.3% of these subjects were taking antidepressants or other psychotropic medications. Antidepressants were being taken by 17.2% of PTSD⫹ (without MDD) subjects and 4.9% of PTSD
subjects. NR3C1-1F Promoter Methylation A significantly lower percentage of methylated clones [F1,120 ¼ 5.46, p ¼ .021] (Figure 1A) was observed in the NR3C1-1F promoter across the 39 CpG sites in the PTSD⫹ subjects compared with the PTSD
subjects. This difference persisted when controlling for PBMC ratio [F1,117 ¼ 6.28, p ¼ .014], when controlling for use of psychotropic medications [F1,119 ¼ 4.10, p ¼ .045], and when the data were expressed as the sum percent methylation across the 39 sites [F1,120 ¼ 4.70, p ¼ .032] (Figure 1B). Figure 1C illustrates percent methylation at each of the 39 CpG sites of exon 1F. The multivariate ANOVA of the 39 CpG sites revealed significant group differences for sites 23 and

R. Yehuda et al. 39 (p ¼ .021 and p ¼ .022, respectively). When the analyses were repeated accounting for PBMC ratio, significant group effects remained for sites 23 and 39. With respect to NR3C1-1F expression, the PTSD⫹ subjects showed a 39% higher relative expression level compared with the PTSD
subjects, an effect that was not statistically significant [F1,112 ¼ 2.495, p ¼ .12]; the correlation between NR3C1-1F promoter methylation and expression was not statistically significant. However, methylation at site 23 was negatively correlated with NR3C1-1F expression (r ¼
.219, p ¼ .023, n ¼ 106). Functional Neuroendocrine Markers The PTSD⫹ subjects showed greater evidence of GR sensitivity in PBMCs as reflected by the lysozyme stimulation test compared with the PTSD
subjects (lysozyme IC50-DEX [F1,116 ¼ 5.13, p ¼ .025]) (Figure 2A). These differences persisted when controlling for the ratio of PBMCs [F1,113 ¼ 5.39, p ¼ .022] and psychotropic medication [F1,112 ¼ 7.06, p ¼ .009]. Similarly, the PTSD⫹ subjects showed a greater decline in cortisol in response to .50 mg DEX [F1,112 ¼ 7.95, p ¼ .006] (Figure 2B), a finding that persisted when the analysis was repeated using plasma levels of DEX as a covariate [F1,111 ¼ 7.68, p ¼ .007] and when additionally covarying for use of any psychotropic medication [F1,110 ¼ 4.80, p ¼ .031]. Results of 24-hour urinary cortisol excretion are shown

Figure 1. Lower NR3C1-1F promoter methylation in subjects with posttraumatic stress disorder (PTSD) (PTSD⫹) compared with subjects without PTSD (PTSD
). NR3C1-1F promoter methylation is expressed as a percentage of methylated clones (A), sum percent methylation across the 39 sites (B), and percent methylation at each of the 39 C-phosphate-G (CpG) sites (C). Boxes represent the location of known or putative canonical (solid-lined box) and noncanonical (broken-lined box) NGFI-A–binding sites and the location of the beginning of the NR3C1-exon 2 (gray box) in relation to the 39 CpG sites of NR3C1-1F (3). Data are presented as mean ⫾ SEM. Significance was set at p ⬍ .05. Parentheses signify effects at a trend level (p ⬍ .10).

www.sobp.org/journal

R. Yehuda et al.

BIOL PSYCHIATRY 2014;]:]]]–]]] 5 methylation was not significantly associated with lysozyme IC50-DEX or urinary cortisol excretion. NR3C1-1F expression was positively correlated with cortisol decline (r ¼ .179, p ¼ .068, n ¼ 105), although not at a conventional level of statistical significance, and was not associated with IC50-DEX or 24-hour urinary cortisol. There were also associations among the neuroendocrine measures. Because the lysozyme stimulation test was developed as an “in vitro” test of glucocorticoid sensitivity using varying doses of DEX administered to cultured lymphocytes, a correlation between this measure and other measures of glucocorticoid sensitivity was expected. The lysozyme IC50-DEX was negatively correlated with percent cortisol suppression (r ¼
.218, p ¼ .022, n ¼ 110) and positively correlated with 24-hour urinary cortisol excretion (r ¼ .269, p ¼ .004, n ¼ 110), reflecting an association of greater GR sensitivity with lower basal cortisol levels.

Figure 2. Higher glucocorticoid receptor responsiveness in subjects with posttraumatic stress disorder (PTSD) (PTSD⫹) compared with subjects without PTSD (PTSD
). In vitro sensitivity to dexamethasone (IC50-DEX) of the lysozyme activity performed on cultured peripheral blood mononuclear cells (A), plasma cortisol decline after digestion of a low dose (0.5 mg) of dexamethasone (B), and 24-hour urinary cortisol excretion (C). Data are presented as mean ⫾ SEM. Significance was set at p ⬍ .05. PBMC, peripheral blood mononuclear cell.

in graph form in Figure 2C. The PTSD⫹ subjects showed a significantly lower urinary cortisol excretion [F1,112 ¼ 3.97, p ¼ .049], an effect that was no longer below the significance threshold when covaried for use of psychotropic medication [F1,111 ¼ 3.65, p ¼ .057]. Associations Between Molecular and Neuroendocrine Markers Both percent methylated clones (r ¼
.220, p ¼ .019, n ¼ 114) and sum percent methylation (r ¼
.198, p ¼ .034, n ¼ 114) were inversely correlated with cortisol decline in response to DEX. DNA

NR3C1-1F Promoter Methylation Associated with Clinical Measures NR3C1-1F promoter methylation levels were negatively correlated with poorer sleep quality (r ¼
.187, p ¼ .041, n ¼ 119), peritraumatic dissociation (r ¼
.192, p ¼ .035, n ¼ 121), and psychiatric distress as reflected by the total number of symptoms endorsed on the Symptom Checklist-90-R (r ¼
.226, p ¼ .015, n ¼ 116). Early life trauma exposure as assessed by the Early Trauma Inventory was not associated with NR3C1-1F promoter methylation in this sample (r ¼ .003, p ¼ not significant, n ¼ 121). Roughly half of the PTSD⫹ subjects met criteria for current MDD. Insofar as this disorder is related to indices of the HPA axis in the opposite direction as in PTSD (19) and has also been studied in association with GR methylation (20,21), least significant difference post-hoc analyses of the clinical and biological data were performed in which the PTSD⫹ subjects were further subdivided into subjects with and without MDD. As illustrated in Figure 3, a main effect of group (percent methylated clones measure [F2,119 ¼ 4.72, p ¼ .011] [Figure 3A] and sum percent methylation measure [F2,119 ¼ 4.22, p ¼ .017] [Figure 3B]) reflected that NR3C1-1F promoter methylation was the lowest for veterans with PTSD in the absence of MDD. Post-hoc testing confirmed that the PTSD and MDD group was not significantly different than the PTSD
group on GR methylation. For neuroendocrine measures, the PTSD and MDD group means were similarly in between the means of the PTSD⫹ and PTSD
groups (IC50-DEX [F2,115 ¼ 2.68, p ¼.073] [Figure 3C] and cortisol decline [F2,111 ¼ 4.24, p ¼ .017] [Figure 3D]). With respect to clinical presentation, combat veterans who met diagnostic criteria for both PTSD and MDD reported the highest symptom levels relative to the other two groups (Table 2).

Discussion This is the first report of NR3C1-1F promoter methylation in purified PBMCs in individuals with PTSD. The primary focus was on DNA methylation, which is a comparatively stable biochemical modification (as opposed to GR transcription, which can vary dynamically in association with prevailing contextual influences). The NR3C1-1F promoter was examined because it is also highly active in brain tissue (22). The NR3C1-1F promoter lies within a CpG island located immediately upstream of the exon NR3C1-1F. Cytosine methylation levels across CpG islands are typically low, in part as a function of genomic sequence (23). Our data and that of others show low levels of methylation within this region (3,24). The current findings show a significantly lower level of NR3C1-1F promoter methylation in trauma-exposed PTSD⫹ combat www.sobp.org/journal

R. Yehuda et al.

6 BIOL PSYCHIATRY 2014;]:]]]–]]]

Figure 3. NR3C1-1F promoter methylation and glucocorticoid receptor (GR) sensitivity for subjects with posttraumatic stress disorder (PTSD) (PTSD⫹) with or without major depressive disorder (MDD) and subjects with trauma exposure without PTSD. NR3C1-1F promoter methylation is expressed as a percentage of methylated clones (A) and sum percent methylation across the 39 sites (B). In vitro GR sensitivity to dexamethasone is expressed as IC50-DEX of lysozyme activity inhibition, performed on cultured peripheral blood mononuclear cells (PBMCs) (C). Pituitary GR sensitivity is expressed as plasma cortisol decline after digestion of a low dose (.5 mg) of dexamethasone (D). For 24-hour urinary cortisol excretion, group differences were not observed [F2,111 ¼ 1.97, p ¼ .144]. Data are presented as mean ⫾ SEM. Significance was set at p ⬍ .05. Parentheses signify effects at a trend level (p ⬍ .10).

veterans compared with veterans with similar combat exposure without PTSD. These effects persisted after controlling for differences in PBMC composition. Sodium bisulfite sequencing revealed overall lower methylation in the NR3C1-1F promoter in subjects with PTSD and significant differences at specific sites (Figure 1C). These affected sites lie in immediate proximity to an nerve growth factor-induced gene A (NGFI-A) response element, the activation of which induces gene

transcription through the exon 1F promoter (3) or its rat ortholog, the exon Nr3c1-17 promoter (2). Chromatin immunoprecipitation assays with constructs containing the NR3C1-1F promoter show that methylation at these sites regulates NGFI-A binding to the promoter and NGFI-A–induced transcription (3). The NR3C1-1F promoter in postmortem brain tissue or PBMCs is differentially methylated in association with developmental history (3,21,25–27). In prior work, the strongest effects were observed at CpG islands in

Table 2. Comparison of Clinical Characteristics for Subjects with PTSD with or without MDD and Subjects with Trauma Exposure without PTSD Non-PTSD (A) PTSD without MDD (B) PTSD with MDD (C) (n ¼ 61) Mean (SD) (n ¼ 29) Mean (SD) (n ¼ 32) Mean (SD) PTSD Severity (Clinician-Rated)a

3.3 (4.9)

61.24 (13.9)

73.8 (16.5)

PTSD Severity (Self-Reported)b

25.6 (8.2)

55.2 (10.3)

66.2 (10.5)

Depression Severityc

5.7 (6.3)

17.4 (8.0)

30.0 (9.4)

Poor Sleep Qualityd

6.2 (4.1)

12.8 (3.6)

14.1 (2.9)

.5 (.6)

1.5 (.9)

2.2 (.9)

1.4 (.4)

2.2 (.6)

2.4 (.6)

Peritraumatic Dissociatione Psychiatric Distressf

MDD, major depressive disorder; NS, not significant; PTSD, posttraumatic stress disorder. Clinician Administered PTSD Scale (6). b Military Version of the PTSD Checklist (8). c Beck Depression Inventory-II (9). d Pittsburgh Sleep Quality Index (11). e Peritraumatic Dissociative Experiences Questionnaire (12). f Symptom Checklist-90-R (13). a

www.sobp.org/journal

A vs. B F2,119 ¼ 505.3, p ⬍ .0005 F2,118 ¼ 229.8, p ⬍ .0005 F2,120 ¼ 107.7, p ⬍ .0005 F2,120 ¼ 59.3, p ⬍ .0005 F2,120 ¼ 57.3, p ⬍ .0005 F2,120 ¼ 863.4, p ⬍ .0005

A vs. C

B vs. C

p ⬍ .0005 p ⬍ .0005 p ⬍ .0005 p ⬍ .0005 p ⬍ .0005 p ⬍ .0005 p ⬍ .0005 p ⬍ .0005 p ⬍ .0005 p ⬍ .0005 p ⬍ .0005

NS

p ⬍ .0005 p ⬍ .0005 p ⬍ .0005 p ⬍ .0005 p ⬍ .0005 p ¼ .039

R. Yehuda et al. close proximity to the NGFI-A response element (3,21), as observed in this sample. The sites most strongly associated with PTSD diagnosis in the current sample are in close proximity with the canonical NGFI-A binding sites (3). Methylation at one of these sites was also correlated with NR3C1-1F expression in the current sample. In vitro studies with cultured neurons from rat hippocampus show that NGFI-A overexpression results in demethylation of the exon 17 promoter (28), reflecting the capacity of proximal CpG islands for epigenetic remodeling. DNA methylation in regulatory DNA regions, as opposed to gene body regions, commonly associates with transcriptional silencing (29,30). There is evidence for higher GR receptor binding in PBMCs from PTSD⫹ patients (31), consistent with the reduced NR3C1-1F promoter methylation shown in the present study. Lower FKBP5 gene expression has also been observed in patients with PTSD and is thought to contribute to findings of altered molecular and neuroendocrine measures associated with HPA axis activity in patients with PTSD (32–34). Lower FKBP5 activity has been associated with both lower cortisol levels and enhanced suppression on the DEX suppression test (35). In the present study, NR3C1-1F promoter methylation levels were significantly correlated with cortisol decline in response to DEX and with PTSD-related symptoms, including psychological distress, peritraumatic dissociation, and poorer sleep quality. These findings are in accord with a more recent report showing that higher NR3C1-1F promoter methylation levels before treatment predicted treatment response in veterans with PTSD (16). In that study, NR3C1-1F promoter methylation did not differ between responders (veterans who no longer met diagnostic criteria for PTSD after treatment) and nonresponders, suggesting that this marker may reflect a tendency to develop PTSD. The differences in GR sensitivity in patients with PTSD contrast strikingly with the differences observed in patients with MDD. There is clear evidence of lower GR sensitivity among depressed patients (36,37), accompanied by reduced cortisol suppression after DEX administration, in contrast to observations in patients with PTSD. The GR resistance associated with lower glucocorticoid feedback sensitivity in depressed patients is apparent in reduced GR function in leukocytes reflected by decreased nuclear translocation after DEX exposure (38) and diminished DEX-induced inhibition of mitogen-associated PBMC proliferation (39). Glucocorticoid resistance in peripheral immune cells seen in depression correlates with glucocorticoid resistance in the DEX suppression test (40). The obvious contrast between these findings and the evidence for enhanced GR sensitivity in PTSD is particularly interesting considering the frequent comorbidity of depression and PTSD. Likewise, among depressed patients, early life trauma is associated with NR3C1-1F promoter methylation (3,27). We found no such association for the PTSD⫹ subjects in this study. To examine the potential impact of comorbid MDD, the data were subjected to post-hoc analyses, facilitated by the fact that more than half of the veterans who met diagnostic criteria for PTSD also met criteria for current MDD. This approach revealed that although subjects with comorbid PTSD and MDD reported higher levels of symptoms than subjects with PTSD alone, biological dysregulation of the HPA axis was most extreme in the PTSD subjects without MDD. This finding is plausible given that PTSD and MDD have been associated with distinct and, in some cases opposing, neuroendocrine profiles (19), reflecting different regulatory influences, possibly resulting from different risk factors. The literature presents different findings concerning neuroendocrine alterations in persons with PTSD in the absence or presence of comorbid MDD (41). The current data indicate that

BIOL PSYCHIATRY 2014;]:]]]–]]] 7 a molecular measure such as NR3C1-1F promoter methylation may be more sensitive to PTSD than more downstream endocrine measures, which reflect integrated regulatory function in response to multiple cellular inputs and feedback mechanisms. The results of this study further suggest that biomarkers that distinguish combat survivors with PTSD from combat survivors without PTSD may be substantially less robust in the presence of comorbid MDD. Prior studies have reported higher methylation in the NR3C1-1F promoter region in postmortem hippocampal tissue in association with childhood abuse (3) but not in hippocampi of persons with MDD compared with normal controls, despite differences in NR3C1-1F promoter expression (3,20). Higher leukocyte NR3C1-1F promoter methylation has also been observed in relation to the number of early traumatic events in patients with borderline personality disorder (27) and in patients with bulimia nervosa who were suicidal or met criteria for borderline personality disorder (42). In two other studies examining methylation in peripheral whole blood or oral rinse in subjects without psychiatric diagnoses, NR3C1-1F promoter methylation was correlated with negative feedback inhibition, in the direction reported here (43,44). In one of these studies, higher NR3C1-1F promoter methylation was associated with early adversity (44). The strongest association reported to date reflects methylation measured in brain tissue (3); in the studies of PBMCs, the correlations between methylation and adversity were weaker and included only a few sites (one to eight) rather than sites across the entire NR3C1-1F promoter region (27,44). In conclusion, there was no overall association in the current study between NR3C1-1F promoter methylation and childhood trauma exposure as measured by self-report. The study sample differs from the samples described earlier in that all subjects were male and were exposed to extremely high levels of combatrelated stress, with half of the subjects developing PTSD. The groups did not differ on levels of overall childhood trauma exposure, which may also contribute to the lack of association. A review of the above-cited studies also indicates that the effect does not extend to all types of early adversity, suggesting that only specific types of early trauma exposure result in epigenetic changes. The timing of exposure to adversity according to developmental stage may also be a critical feature that deserves further study. The absence of an association with early trauma in the current sample supports the interpretation of lower NR3C1-1F promoter methylation as a biomarker for combat-related PTSD that is not due to the predisposing influence of early trauma exposure per se but may reflect other risk markers in PTSD that are associated with negative feedback inhibition of the HPA axis. Such risk markers may include combat exposure or aspects of the acute response to trauma (e.g., peritraumatic dissociation), which may represent dispositional characteristics that predate combat exposure. This work was supported by grant funding from the Department of Defense (Grant No. W911NF-09-1-0298 to RY and Grant No. W81XWH-09-2-0044 to CRM). We thank Dr. Owen Wolkowitz for scientific collaboration and for funding that permitted additional recruitment (Department of Defense Grant No. W81XWH-10-1-0021). The authors report no biomedical financial interests or potential conflicts of interest. 1. Klengel T, Mehta D, Anacker C, Rex-Haffner M, Pruessner JC, Pariante CM, et al. (2013): Allele-specific FKBP5 DNA demethylation mediates gene-childhood trauma interactions. Nat Neurosci 16:33–41.

www.sobp.org/journal

R. Yehuda et al.

8 BIOL PSYCHIATRY 2014;]:]]]–]]] 2. Weaver IC, Cervoni N, Champagne FA, D’Alessio AC, Sharma S, Seckl JR, et al. (2004): Epigenetic programming by maternal behavior. Nat Neurosci 7:847–854. 3. McGowan PO, Sasaki A, D’Alessio AC, Dymov S, Labonte B, Szyf M, et al. (2009): Epigenetic regulation of the glucocorticoid receptor in human brain associates with childhood abuse. Nat Neurosci 12:342–348. 4. Yehuda R, Seckl J (2011): Stress-related psychiatric disorders with low cortisol levels: A metabolic hypothesis. Endocrinology 152:4496–4503. 5. van Zuiden M, Geuze E, Willemen HL, Vermetten E, Maas M, Heijnen CJ, et al. (2011): Pre-existing high glucocorticoid receptor number predicting development of posttraumatic stress symptoms after military deployment. Am J Psychiatry 168:89–96. 6. Blake DD, Weathers FW, Nagy LM, Kaloupek DG, Gusman FD, Charney DS, et al. (1995): The development of a Clinician-Administered PTSD Scale. J Trauma Stress 8:75–90. 7. Spitzer RL, Williams JBW, Gibbon M (1995): Structured Clinical Interview for DSM-IV (SCID). New York: New York State Psychiatric Institute Biometrics Research. 8. Weathers F, Huska J, Keane T (1991): The PTSD Checklist Military Version (PCL-M). Boston: National Center for PTSD. 9. Brown GP, Hammen CL, Craske MG, Wickens TD (1995): Dimensions of dysfunctional attitudes as vulnerabilities to depressive symptoms. J Abnorm Psychol 104:431–435. 10. Bremner JD, Vermetten E, Mazure CM (2000): Development and preliminary psychometric properties of an instrument for the measurement of childhood trauma: The Early Trauma Inventory. Depress Anxiety 12:1–12. 11. Buysse DJ, Reynolds CF 3rd, Monk TH, Berman SR, Kupfer DJ (1989): The Pittsburgh Sleep Quality Index: A new instrument for psychiatric practice and research. Psychiatry Res 28:193–213. 12. Marmar CR, Weiss DS, Metzler TJ (1997): The Peritraumatic Dissociative Experiences Questionnaire. In: Wilson JP, Keane TM, editors. Assessing Psychological Trauma and PTSD: A Handbook for Practitioners. New York: Guilford Press. 13. Derogatis LR (1992): SCL-90-R: Administration, scoring and procedures manual for the revised version and other instruments of the psychopathology rating scale series. Baltimore: Clinical Psychometric Research, Inc. 14. Lam LL, Emberly E, Fraser HB, Neumann SM, Chen E, Miller GE, et al. (2012): Factors underlying variable DNA methylation in a human community cohort. Proc Natl Acad Sci U S A 109(suppl 2):17253–17260. 15. Yehuda R, Bierer LM, Andrew R, Schmeidler J, Seckl JR (2009): Enduring effects of severe developmental adversity, including nutritional deprivation, on cortisol metabolism in aging Holocaust survivors. J Psychiatr Res 43:877–883. 16. Yehuda R, Daskalakis NP, Desarnaud F, Makotkine I, Lehrner AL, Koch E, et al. (2013): Epigenetic biomarkers as predictors and correlates of symptom improvement following psychotherapy in combat veterans with PTSD. Front Psychiatry 4:118. 17. Yehuda R, Yang RK, Guo SL, Makotkine I, Singh B (2003): Relationship between dexamethasone-inhibited lysozyme activity in peripheral mononuclear leukocytes and the cortisol and glucocorticoid receptor response to dexamethasone. J Psychiatr Res 37:471–477. 18. Golier JA, Schmeidler J, Legge J, Yehuda R (2006): Enhanced cortisol suppression to dexamethasone associated with Gulf War deployment. Psychoneuroendocrinology 31:1181–1189. 19. Yehuda R (2002): Post-traumatic stress disorder. N Engl J Med 346: 108–114. 20. Alt SR, Turner JD, Klok MD, Meijer OC, Lakke EA, Derijk RH, et al. (2010): Differential expression of glucocorticoid receptor transcripts in major depressive disorder is not epigenetically programmed. Psychoneuroendocrinology 35:544–556. 21. Oberlander TF, Weinberg J, Papsdorf M, Grunau R, Misri S, Devlin AM (2008): Prenatal exposure to maternal depression, neonatal methylation of human glucocorticoid receptor gene (NR3C1) and infant cortisol stress responses. Epigenetics 3:97–106. 22. Turner JD, Muller CP (2005): Structure of the glucocorticoid receptor (NR3C1) gene 50 untranslated region: Identification, and tissue distribution of multiple new human exon 1. J Mol Endocrinol 35:283–292. 23. Bird A (2002): DNA methylation patterns and epigenetic memory. Genes Dev 16:6–21. 24. Labonte B, Yerko V, Gross J, Mechawar N, Meaney MJ, Szyf M, et al. (2012): Differential glucocorticoid receptor exon 1(B), 1(C), and 1(H)

www.sobp.org/journal

25.

26.

27.

28.

29. 30.

31.

32.

33. 34.

35. 36. 37. 38. 39.

40.

41. 42.

43. 44.

expression and methylation in suicide completers with a history of childhood abuse. Biol Psychiatry 72:41–48. Hompes T, Izzi B, Gellens E, Morreels M, Fieuws S, Pexsters A, et al. (2013): Investigating the influence of maternal cortisol and emotional state during pregnancy on the DNA methylation status of the glucocorticoid receptor gene (NR3C1) promoter region in cord blood. J Psychiatr Res 47:880–891. Radtke KM, Ruf M, Gunter HM, Dohrmann K, Schauer M, Meyer A, et al. (2011): Transgenerational impact of intimate partner violence on methylation in the promoter of the glucocorticoid receptor. Transl Psychiatry 1:e21. Perroud N, Paoloni-Giacobino A, Prada P, Olie E, Salzmann A, Nicastro R, et al. (2011): Increased methylation of glucocorticoid receptor gene (NR3C1) in adults with a history of childhood maltreatment: A link with the severity and type of trauma. Transl Psychiatry 1:e59. Weaver IC, D’Alessio AC, Brown SE, Hellstrom IC, Dymov S, Sharma S, et al. (2007): The transcription factor nerve growth factor-inducible protein a mediates epigenetic programming: Altering epigenetic marks by immediate-early genes. J Neurosci 27:1756–1768. Klose RJ, Bird AP (2006): Genomic DNA methylation: The mark and its mediators. Trends Biochem Sci 31:89–97. Weber M, Hellmann I, Stadler MB, Ramos L, Paabo S, Rebhan M, et al. (2007): Distribution, silencing potential and evolutionary impact of promoter DNA methylation in the human genome. Nat Genet 39: 457–466. Yehuda R, Boisoneau D, Lowy MT, Giller EL Jr (1995): Dose-response changes in plasma cortisol and lymphocyte glucocorticoid receptors following dexamethasone administration in combat veterans with and without posttraumatic stress disorder. Arch Gen Psychiatry 52:583–593. Segman RH, Shefi N, Goltser-Dubner T, Friedman N, Kaminski N, Shalev AY (2005): Peripheral blood mononuclear cell gene expression profiles identify emergent post-traumatic stress disorder among trauma survivors. Mol Psychiatry 10:500–513. Sarapas C, Cai G, Bierer LM, Golier JA, Galea S, Ising M, et al. (2011): Genetic markers for PTSD risk and resilience among survivors of the World Trade Center attacks. Dis Markers 30:101–110. Yehuda R, Cai G, Golier JA, Sarapas C, Galea S, Ising M, et al. (2009): Gene expression patterns associated with posttraumatic stress disorder following exposure to the World Trade Center attacks. Biol Psychiatry 66:708–711. Binder EB (2009): The role of FKBP5, a co-chaperone of the glucocorticoid receptor in the pathogenesis and therapy of affective and anxiety disorders. Psychoneuroendocrinology 34(suppl 1):S186–S195. Pariante CM, Miller AH (2001): Glucocorticoid receptors in major depression: Relevance to pathophysiology and treatment. Biol Psychiatry 49:391–404. Holsboer F (2000): The corticosteroid receptor hypothesis of depression. Neuropsychopharmacology 23:477–501. Gormley GJ, Hospelhorn VD, Khan MK, Jensen EV (1985): A controlled pore glass bead assay for the measurement of cytoplasmic and nuclear glucocorticoid receptors. J Steroid Biochem 22:693–698. Lowy MT, Reder AT, Antel JP, Meltzer HY (1984): Glucocorticoid resistance in depression: The dexamethasone suppression test and lymphocyte sensitivity to dexamethasone. Am J Psychiatry 141: 1365–1370. Wodarz N, Rupprecht R, Kornhuber J, Schmitz B, Wild K, Braner HU, et al. (1991): Normal lymphocyte responsiveness to lectins but impaired sensitivity to in vitro glucocorticoids in major depression. J Affect Disord 22:241–248. Morris MC, Compas BE, Garber J (2012): Relations among posttraumatic stress disorder, comorbid major depression, and HPA function: A systematic review and meta-analysis. Clin Psychol Rev 32:301–315. Steiger H, Labonte B, Groleau P, Turecki G, Israel M (2013): Methylation of the glucocorticoid receptor gene promoter in bulimic women: Associations with borderline personality disorder, suicidality, and exposure to childhood abuse. Int J Eat Disord 46:246–255. Edelman S, Shalev I, Uzefovsky F, Israel S, Knafo A, Kremer I, et al. (2012): Epigenetic and genetic factors predict women’s salivary cortisol following a threat to the social self. PLoS One 7:e48597. Tyrka AR, Price LH, Marsit C, Walters OC, Carpenter LL (2012): Childhood adversity and epigenetic modulation of the leukocyte glucocorticoid receptor: Preliminary findings in healthy adults. PLoS One 7:e30148.