Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Open Access

Peer-reviewed

Research Article

Epigenetic and Genetic Factors Predict Women's Salivary Cortisol following a Threat to the Social Self

  • Shany Edelman ,

    Contributed equally to this work with: Shany Edelman, Idan Shalev

    Affiliation Department of Neurobiology, Hadassah-Hebrew University Medical Center, Jerusalem, Israel

    ⨯
  • Idan Shalev ,

    Contributed equally to this work with: Shany Edelman, Idan Shalev

    Affiliation Department of Neurobiology, Hadassah-Hebrew University Medical Center, Jerusalem, Israel

    ⨯
  • Florina Uzefovsky,

    Affiliation Department of Psychology, Hebrew University, Jerusalem, Israel

    ⨯
  • Salomon Israel,

    Affiliation Department of Psychology, Hebrew University, Jerusalem, Israel

    ⨯
  • Ariel Knafo,

    Affiliation Department of Psychology, Hebrew University, Jerusalem, Israel

    ⨯
  • Ilana Kremer,

    Affiliation Technion Medical School, Haifa and Mazra Hospital, Akko, Israel

    ⨯
  • David Mankuta,

    Affiliation Department of Labor and Delivery, Hadassah Medical Organization, Jerusalem, Israel

    ⨯
  • Marsha Kaitz,

    Affiliation Department of Psychology, Hebrew University, Jerusalem, Israel

    ⨯
  • Richard P. Ebstein

    rpebstein@gmail.com

    Affiliations Department of Neurobiology, Hadassah-Hebrew University Medical Center, Jerusalem, Israel, Department of Psychology, Hebrew University, Jerusalem, Israel, Department of Psychology, National University of Singapore, Singapore, Singapore

    ⨯

Abstract

Evidence suggests that the reactivity of the Hypothalamus-Pituitary-Adrenal axis (HPAA) is modulated by both genetic and environmental variables. Of special interest are the underlying molecular mechanisms driving gender differences to psychosocial stressors. Epigenetic mechanisms that sculpt the genome are ideal candidates for mediating the effects of signals on the HPAA. In the current study, we analyzed by pyrosequencing, bisulfite-treated buccal DNA from male and female university students who participated in the Trier Social Stress Test (TSST). A linear regression model was used to ascertain the effects of sex, CpG methylation and genes on stress response. Total cortisol output (area under the curve, AUC) was significantly predicted by glucocorticoid receptor (NR3C1) exon 1F methylation (averaged across 39 CpG sites) solely in female subjects. A single CpG site located in the exon 1F noncanonical nerve growth factor-inducible protein A (NGFI-A) transcription factor was a highly significant predictor of AUC in female subjects. Additionally, variations in the estrogen receptor alpha (ESR1) and the serotonin transporter promoter (5-HTTLPR) genes were independent additive predictors of AUC. The full model accounted for half of the variance (50.06%) in total cortisol output. Notably, this is the first demonstration that epigenetic changes at the GR exon 1F correlate with HPAA reactivity. These findings have important implications for understanding the molecular mechanisms underlying gender differences in stress-related disorders and underscore the unique value of modeling both epigenetic and genetic information in conferring vulnerability to stress.

Citation: 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(11): e48597. https://doi.org/10.1371/journal.pone.0048597

Editor: Qing Song, Morehouse School of Medicine, United States of America

Received: May 9, 2012; Accepted: September 27, 2012; Published: November 14, 2012

Copyright: © 2012 Edelman et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Funding: This research was partially supported by the following agencies Israel Science Foundation (Grant No. 389/05); National University of Singapore (Grant no HSS-1001-P02) and by a grant from the Science of Generosity Initiative. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Competing interests: The authors have declared that no competing interests exist.

Introduction

Psychosocial stress along with the coping styles that people employ when challenged by stress, are considered important determinants of overall wellbeing [1]. Particularly important is the body's reaction to social stressors; reflecting the daily changes we face at home, with friends, during school and at work. Dickerson & Kemeny [2] review evidence that human cortisol responses to acute stressors are most pronounced in situations that pose a social threat to the individual. Notably, not all individuals respond similarly to social stress and as noted by McEwen [3] there are very large individual differences in stress reactivity, reflecting significant life events. While some individuals appear to be resilient to difficult conditions, others react adversely to such challenges, incurring a range of physical and mental disorders. This imbalance is expressed via dysregulation of the Hypothalamus-Pituitary-Adrenal axis (HPAA), the main pathway of the stress response (with cortisol acting as its end product). The HPAA has been meticulously investigated and found connected to a large number of stress related disorders [4].

One of the most consistent findings employing psychological stress tasks is the significantly larger salivary cortisol response in healthy adult men compared to women following short-term laboratory stress [5], [6], [7], [8], [9]. Kudielka and Kirschbaum [10] extensively reviewed gender differences in response to stress and suggest that such differences reflect causal links and are not merely epiphenomena of sex-specific prevalence rates of several disorders [11], [12]. Male stress responses may predominantly involve the traditional “fight and flight” reaction while women's stress response may be better characterized by “tend and befriend”, involving nurturant activities and the creation of social networks. Differences between genders in stress response can be attributed to circulating gonadal sex hormones, sexual dimporphism of brain functioning and corticosteroid binding [13]. However, much of the underlying neurochemical and neurogenetic mechanisms for gender differences in stress reactivity generally remain obscure.

The final target of the HPAA cortisol release is the glucocorticoid receptor (GR, NR3C1 MIM +138040) [14]. The GR is a member of the steroid receptor superfamily and is the key mediator of the majority of cortisol's tissue effects by way of direct binding to hormone-responsive elements in the DNA or via interactions with other transcription factors and regulation of gene transcription [14]. GR levels are transcriptionally controlled by multiple untranslated alternative first exons, each with its own promoter providing a mechanism for tissue-specific fine-tuning of GR levels [15]. The mechanisms regulating the differential usage of these first exons in different tissues and individuals, and the role of the 5′-UTR in the splicing of the coding exons are still poorly understood (see review by Turner et al [16]). Both individual and gender differences are observed in basal and dynamic cortisol changes [8] and common polymorphisms in the GR and other genes partially contribute to disparities in HPAA reactivity [17], [18], [19], [20].

There is increasing evidence for involvement of the epigenome in altering short and long-term status of GR and cortisol responsiveness [21]. The nerve growth factor-inducible protein A (NGFI-A), is a transcription factor that has been shown in the rat [22] and human [23] to regulate the expression of the NR3C1 promoter; its methylation down regulates gene expression [22]. In a seminal article, Weaver, Meaney and colleagues [21] showed that differential maternal care in rat pups modified the methylation pattern of the hippocampal GR exon 17 which led to significant differences in subsequent adult behavior. Importantly, the cytosine residue within the 5′ CpG dinucleotide of the noncononocal NGFI-A (CpG31, CpG32) consensus sequence was highly methylated (associated with low GR expression) in the offspring of low caring mothers, and rarely methylated (high GR expression) in the offspring of high caring dams explaining the observed differences in HPAA reactivity in the adult offspring. The impact of maternal care on the epigenome is mediated by serotonergic (5-HT) neurotransmission that drives downstream expression of NGFI-A targeting its cognate binding site on the GR exon 17 promoter. In humans, McGowan et al [23] showed that the hippocampal GR promoter 1F exon (orthologue to the rat exon 17) was characterized by increased methylation of the NGFI-A transcription factor in suicide victims who were exposed to childhood abuse. Oberlander et al [24] showed that the cord blood neonatal GR methylation pattern including exon 1F was influenced by mother's mood and SSRI treatment during pregnancy. Increased GR methylation at this site was also associated with increased salivary cortisol stress responses at 3 months. Moser et al [25] found that in post-mortem hippocampal specimens from 32 subjects with various psychiatric disorders, a noncanonical site (equivalent to McGowan CpG32 as in the current study) was poorly methylated and the adjacent sites were unmethylated. They interpret their results as indicating that the mechanism presented by Weaver and associates is probably different between humans and rat. Similarly, Alt et al [26] showed low methylation levels in both major depressive disorder and control brains. Exon 1F was uniformly unmethylated. The variability in methylation effects across studies may reflect the fact that changes in methylation, which may be reversible [27], are also likely subject to multiple influences by both stressors and genetics. As noted by Meaney et al [28], the enzymatic machinery required to alter cytosine methylation is operative in post-mitotic neurons.

Prompted by these intriguing studies, and the importance of understanding the basis for individual and gender differences in HPAA reactivity with its profound implications for overall psychological and physical wellbeing, we used pyrosequencing [29] of bisulfite treated buccal DNA to assess the methylation status of the NR3C1 exon 1F promoter sequence in nonclinical subjects.

In addition, towards examining the role of genetic factors in understanding individual differences in HPAA response to social stress, we also genotyped these subjects for two salient genes, the serotonin transporter (SLC6A4) promoter region (5-HTTLPR) indel [30] and the estrogen receptor alpha (ESR1) promoter region TA repeat [31], respectively coding for serotonergic neurotransmission and a gonadal sex hormone that based on both animal and human studies, are known to modulate HPAA [8], [32], [33].

Our strategy in this investigation is to simultaneously model both epigenetic factors as well as sequence variations of two candidate genes towards understanding how nurture and nature modulate in a gender-specific manner HPAA response to social stress. Determination of both GR methylation in the exon 1F and genotyping polymorphic promoter-regions in two salient genes that impact the HPAA will allow integration and evaluation of the interactions between distinct trans polymorphisms, methylation and stress hormone response to social stressors. The strategy employed in the current investigation illustrates an important step towards deciphering the molecular determinants of complex phenotypes.

Estradiol especially seems to exert modulating effects on HPAA functioning, including HPAA responsiveness to glucocorticoid negative feedback [34]. The ESR1 is of particular interest since maternal behavior can be modulated by ESR1 expression and intriguingly, such effects are transmitted across generations. Specifically, Champagne et al [35], [36] demonstrated that high maternal care females exhibited significantly elevated levels of ESR1 mRNA compared with low maternal care females. Levels of estrogen receptor will determine the responsiveness to this hormone and ultimately regulate the efficiency of estrogen-mediated signaling and the biological and behavioral outcomes associated with estrogen. In rats, estradiol treatment significantly increased the evening elevation in cortisol and the stress-induced rise in cortisol [37]. The ability of estradiol to inhibit glucocorticoid negative feedback occurs specifically via the estrogen receptor α (ESR1), acting at the level of the paraventricular nucleus. Altogether estrogen plays an important role in HPAA regulation [38], [39]. Hence, we hypothesized that the polymorphic ESR1 gene is a worthwhile candidate, along with epigenetic modifications of the GR receptor, to impact social stress evoked increases in salivary cortisol especially in women subjects. ESR1 was of particular interest since this polymorphism seems to be a marker for important clinical outcomes in women [40].

There is considerable evidence for the idea that serotonin (5-HT) mediates the effect of postnatal modulation by differential maternal care on hippocampal GR expression since these effects are blocked by concurrent treatment with a serotonin receptor antagonist [41]. Importantly, 5-HT is not only a key regulatory neurotransmitter of HPAA responsiveness but also enables the methylation of the NGFI-A transcription factor, important in rat maternal care [42]. The serotonin transporter (SLC6A4) is an important element in regulation of synaptic serotonin levels [30] leading us to hypothesize in our model that the serotonin transporter, along with ESR1 (‘nature’), contribute jointly with epigenetic modification of GR (‘nurture), to explain individual differences in gender-sensitive stress responses, especially regarding gender differences.

The nonclinical subjects in the current study were tested for HPAA responsiveness to acute social stress [20] obtained by using a laboratory-based paradigm, the Trier Social Stress Test (TSST) [43], that leverages a ‘threat to the social self’ [2] via public speaking and mental arithmetic, to generate an unambiguous physiological endpoint, indexed by salivary cortisol. Importantly, both the TSST response and basal cortisol levels have been shown to be substantially heritable [44], [45], providing the necessary background for the current investigation.

Results

Methylation of GR exon 1F in men and women

As previously reported by us [20] and others (9), in both men and women there is a significant increase (greater in males compared to females) in salivary cortisol levels following the TSST (Figure 1). The stress induced rise in salivary cortisol is presented in Figure 1 for each time point in a GLM repeated measures plot (SPSS) as well as AUC (see insert) for both men and women. There is a significant rise in cortisol (GLM repeated measures) for both men (tests of within subjects (F = 22.32, p<0.001; Huyhn-Feldt corrected) and women (F = 9.04, p<0.001; Huyhn-Feldt corrected). Since men and women are characterized by significantly different cortisol levels, each sex was analyzed separately and when sex is included in the analysis it is entered as an independent variable We next examined the methylation level and averaged the results across 39 assayed CpG sites in exon 1F for each subject (Figure 2). Overall, women showed significantly greater methylation levels than did men (Figure 3A) across the entire promoter region (t = 2.538, p = 0.013). Notably, marked individual differences for both men and women were observed at many individual CpG sites (Figure 2A, 2B). Overall levels of GR exon 1F methylation were similar to those previously observed by Oberlander et al [24] in peripheral tissue.

thumbnail
Download:
Figure 1. Salivary cortisol levels following the TSST.

Area under the curve following the TSST for males and females. The rise in salivary cortisol is significant (SPSS GLM repeated measures) for men (F = 22.32 p<0.001) and for females (F = 9.04 p<0.001).

https://doi.org/10.1371/journal.pone.0048597.g001

thumbnail
Download:
Figure 2. Methylation levels across 39 CpG sites in exon 1F for each subject.

Methylation level of GR exon 1F at individual CpG sites for females (n = 46) (2A) and males (n = 46) (2B). Each point in the figure represents a single individual's percentage methylation value for each of the 39 CpG sites that were analyzed by bisulfite pyrosequencing. The points were color coded by the graph option in Excel and each color and shape represents an individual subject. The graph illustrates the marked individual differences (indicated by the shape and color of markers) in percentage methylation at each CpG site. Figure (2C) presents the promoter sequence of exon 1F showing the CpG sites and their position. Highlighted regions represent known or putative canonical (italics) and noncanonical (not italics) NGFI-A–binding sites, with the shaded grey area indicating the beginning of the exon (following McGowan et al (23)).

https://doi.org/10.1371/journal.pone.0048597.g002

thumbnail
Download:
Figure 3. Methylation levels of the GR exon 1F and NFGI-A CpG #12 by gender.

Percentage methylation levels of exon 1F (A) and NFGI-A CpG #12 (B) for males and females.

https://doi.org/10.1371/journal.pone.0048597.g003

GR exon 1F methylation and total cortisol output (AUC)

We next examined the relationship between sex, average methylation level across exon 1F, the interaction (sex x methylation) and AUC (summarized in Table 1). Sex (R2Δ = 0.116 F1,89 = 11.809, p = 0.001) and 1F methylation (R2Δ = 0.065 F1,88 = 7.082 p = 0.009) were significant predictors of AUC. The interaction between sex and exon 1F methylation did not attain significance (R2Δ = 0.197, F1,88 = 1.793, p = 0.184). Nevertheless, because of the potential role of sex differences as evidenced by sex differences in AUC, all subsequent analyses were carried out separately for men and women. For men, methylation was not a significant predictor (p = 0.722). In contrast, for women (Figure 4), the average methylation level of the GR 1F exon was inversely related to the amount of salivary cortisol secreted (AUC) during the TSST (R2Δ = 0.213 F1,44 = 11.877, p = 0.001), accounting for 21.3% of the variance.

thumbnail
Download:
Figure 4. GR exon 1F methylation and total cortisol output.

Correlation between AUC and average percent methylation at GR exon 1F in female subjects.

https://doi.org/10.1371/journal.pone.0048597.g004

thumbnail
Download:
Table 1. Linear regression analysis (average % methylation).

https://doi.org/10.1371/journal.pone.0048597.t001

Methylation of the noncanonical transcription factor NGFI-A CpG #12

Since this specific NGFI-A site displayed differential methylation, and was prognostic of behavior and clinical state both in the original rat study of maternal care [21] as well as the more recent human study of suicide with child abuse [23], we examined individual CpG sites within the NGFI-A recognition sequences as predictors of AUC (Table S1 for summary). A significant difference (Figure 3B) at CpG #12, a noncanonical NFGI-A binding site, was also observed in methylation between males and females (t = 2.618, p = 0.012). In the regression model both sex (R2Δ = 0.113, F1,62 = 7.923, p = 0.007) and CpG#12 (R2Δ = 0.173, F1,61 = 14.78, p<0.001) were significant predictors of AUC. There is no significant interaction between CpG#12 and sex (p = 0.652). Moreover, it is the only CpG site among the 39 tested that remains a significant predictor of AUC following Bonferroni correction (pcorrected = 0.03).

ESR1 and 5-HTTLPR are independent predictors of AUC

All subjects were genotyped for two relevant genes, the ESR1 and 5-HTTLPR. We continued to analyze the proposed model (Table 1 for summary) in which average 1F methylation, 5-HTTLPR, ESR1 and the interaction between each gene and methylation level predict AUC (AUC = methylation + 5-HTTLPR + ESR1). In female subjects, there is a significant main effect of 5-HTTLPR (R2Δ = 0.172, F1,42 = 12.032, p = 0.001) and ESR1 (R2Δ = 0.132, F2,40 = 5.634, p = 0.007) on AUC. The interaction between 5-HTTLPR and 1F methylation did not attain significance (R2Δ<0.001, F1,42 = 0.019, p = 0.891). Nor was there any significant interaction between ESR1 x exon 1F methylation (p = 0.452). Altogether, in the full model (1F methylation, 5-HTTLPR and ESR1 polymorphisms) a total of 50.06% (adjusted R2) of the variance in total salivary cortisol output is explained. There was no significant effect of genotype on methylation.

Discussion

A fuller understanding of the molecular mechanisms underlying differences between male and female response to stress has potentially profound implications for explaining gender differences in vulnerability to both psychopathology [46] and physical disease [47], [48], [49]. We have used a well-characterized laboratory based social stress test to examine the impact of epigenetic and genetic variation on cortisol response in a group of nonclinical subjects. In women, and not in men, the averaged methylation of 39 examined CpG sites located across the GR promoter exon 1F is a highly significant predictor of total cortisol response (AUC) in the TSST.

We specifically focused on a specific CpG (CpG#12) since this single CpG site is located in the exon 1F noncanonical NGFI-A transcription factor. Importantly, NGFI-A has been shown both in the rat and human to regulate the expression of the NR3C1 promoter; its methylation down regulates gene expression [21], [23] This specific CpG explains a notable 28% of the variance in total cortisol output in women. Importantly, women show significantly greater methylation in exon 1F, and at the NGFI-A transcription factor site, compared to men. That being said, our provisional results on a relatively small sample size should be considered exploratory and we eagerly await future replications in an independently recruited sample towards validating these first findings.

Gender effects in HPAA responsiveness are likely due to the cellular environment in women and men which differs substantially regarding the hormonal milieu and the impact of sex on the penetrance and expressivity of a wide variety of quantitative traits have been underscored [50]. Several studies have indicated that in the presence of estradiol there is an enhanced responsiveness of the HPAA to a stressor that may be due in part to impairment of the glucocorticoid negative feedback loop [37]. Additionally, in the rat model of high and low maternal care, there were differences in brain ESR1 mRNA levels between offspring of high and low mothers [35]. Hence, based on the information provided by the rat model of maternal care, we thought it worthwhile to examine the contribution of the ESR1 promoter-region repeat polymorphism as a predictor of total cortisol output. We observed a significant main effect between this gene and methylation of the GR 1F exon.

The serotonin transporter promoter-region indel was also examined as a predictor of cortisol output. Serotonin plays an important role in HPAA responsiveness [51], [52] and, by a pathway independent of the GR, this transmitter was shown to mediate the methylation of the NGFI-A transcription factor in the rat model of high and low maternal care [42]. We confirm both a significant main effect of 5-HTTLPR short allele on basal cortisol [53], as well as a significantly increased overall cortisol rise (AUC) in carriers of the short allele following the TSST [54]. Similarly, in girls aged 9–14 exposure to a laboratory stress test, evoked greater cortisol response in subjects genotyped for the 5-HTTLPR short alleles [55]. Future investigations could profitably examine methylation patterns on other genes in the human stress response. Of especial interest is the serotonin transporter, since studies have observed salient epigenetic signatures in the serotonin transporter [56], [57], [58].

We observed that GR methylation and ESR1 and 5-HTTLPR genotypes independently predict AUC. The absence of a significant interaction between 5-HTTLPR and the ESR1 and 1F methylation in the regression analysis suggests that these two genes, which are known to partially mediate stress, are affecting the stress response independently of methylation of the GR. Our results might suggest that other pathways, and not solely GR exon 1F methylation, are mediating the effect of these hormones on HPAA responsiveness. For example, chronic estrogen administration has been shown to increase corticotrophin releasing hormone (CRH) mRNA levels, CRH synthesis and release in CRH-containing neurons of the paraventricular nucleus [59]. Indeed the human CRH gene contains functional half palindromic estradiol responsive element [60] and estrogen receptors are considered important modulators of CRH secretion [61]. Another mechanism for ESR1 impacting HPAA responsiveness without directly impacting the GR receptor methylation profile, is based on the report of Westberg et al [31] that ‘neuroticism’, ‘psychoticism’, and ‘non-conformity’ all appeared to be associated with the ER alpha gene. Hence, the effect of ER alpha on cortisol output might be mediated indirectly by personality traits.

Yet an additional explanation is suggested by the work of Turner's group [16]. They showed in white blood cells of healthy donors, highly variable methylation patterns among different GR promoters, suggesting that epigenetic programming may not be restricted to the 1F promoter, but operates throughout the extended CpG island [16]. It cannot, therefore, be excluded that the impact of estrogen and serotonin is mediated by other GR first exons and not solely by 1F. In such a scenario no statistic interaction would be expected between GR 1F methylation and either of the two polymorphic genes we examined. Clearly, future studies should examine the methylation patterns of other first exon alternative splicing forms with the aim to clarify their relationship to HPAA responsiveness.

An increasing volume of experimental research also indicates that 5HT can modulate HPAA responsiveness and cortisol release beyond direct actions at the level of the GR [51]. For example, Heisler et al have shown that 5HT2C receptors in the paraventricular nucleus of the hypothalamus mediate serotonin-induced release of CRH. Additionally, serotonin directly induces secretion of cortisol directly from the human adrenal gland [62]. Finally, Moser et al [63] have recently shown by PET imaging that release of cortisol by serotonin is likely mediated by 5HT1A receptor located in the hypothalamus and are known to mediate the release CRH [64]. These findings may help to understand the current finding that the serotonin transporter polymorphism has a main effect on HPAA responsiveness but shows no significant statistical interaction with the GR exon 1F methylation [61].

Failure to observe an interaction could also be an issue of power. Although the sample size in the current study is relatively large in comparison to the few studies of GR methylation and behavior so far carried out in humans, further stratifying the female subjects by ESR1 repeat length ipso facto reduces power. So we cannot rule out the possibility that failure to observe a significant interaction in the regression analysis is simply a power issue. The current results cannot distinguish between these various alternative explanations leaving such explanations to be resolved in future investigations.

Interestingly, in the rat model it has been shown that NGFI-A is required for serotonin-induced DNA demethylation and increased exon 17 expression [22]. Intriguingly, the 5-HTTLPR polymorphism is controversially linked to depression mediated by stressful life events [65], [66], [67], [68] and it has recently been suggested that the method used to assess environmental adversity, as well as clinical assessment of the subjects, could explain most discrepancies in results [65]. Studies using objective evidence or detailed interviews to assess environmental adversity in context consistently find an interaction in the expected direction, but studies relying on brief self-report measures of adversity often showed negative findings. Our findings suggest the notion that a possibly objective measure of stressful life events that could be incorporated in G×E studies, such as pioneered by Caspi and his colleagues [68], [69], would be to examine methylation patterns of the GR receptor as we and others [23], [24], [25], [26] have done. After all, stressful life events need to impact the genome at the molecular level towards effecting behavior mediated presumably by modifying gene expression.

Ideally, research on methylation with regards to stress reaction would have been performed on brain tissue especially the hippocampus [21], [23]. Such studies in people, however, are clearly restricted to the availability of post-mortem material and constrained by particular clinical diagnoses [70]. In the current study we instead examined a peripheral tissue, buccal epithelial cells, to monitor GR methylation status allowing us access to a much wider range of subjects. Some information regarding the relationship between buccal cells, white blood and brain cells is provided by Kaminsky et al [71]. Caution is needed in interpreting peripheral cells as models of brain function and it is not completely clear to what extent results in buccal cells can be extrapolated to brain despite their shared ectodermal origin. Peripheral tissues have been suggested as models or probes for studying human behavior [72] and there is evidence that gene expression in lymphocytes, for example, may partially reflect similar brain activity [73]. To our knowledge such analyses have not been carried out between buccal cells and brain.

Nevertheless, regardless of the relationship between buccal, brain and white blood cells, the peripheral GR has also been suggested to be important in its own right in the maintenance of HPAA homeostasis. In fact, there are well-established peripheral pathways that bypass or complement the HPAA [74]. For example, an increasing number of preclinical and clinical studies report dissociation of ACTH and cortisol levels in critical illness, inflammation and mental disorders [75]. Mechanisms involve altered adrenal responsiveness, aberrant receptor expression or modulation of adrenal function by cytokines, vasoactive factors or neuropeptides [75]. The current results highlight the importance of peripheral GR methylation as an important component in regulation of cortisol secretion.

It is tempting to hypothesize that in the presence of low GR concentration (high exon1F methylation [23]) only low levels of cortisol are required to activate peripheral tissues. HPAA responsiveness in such individuals is fine-tuned to the levels not only of hippocampal [23] and pituitary [76] glucocorticoid receptor expression but also, as suggested by the current findings, to the peripheral GR. Expression of the GR in peripheral tissues may in part be determined by the methylation status of the 1F exon. This notion is supported by our finding that when basal cortisol is entered as the dependent variable in the regression analysis similar results (GR methylation is a significant predictor of basal cortisol levels in women) are obtained as observed for AUC following the TSST. This suggests that increased cortisol levels are associated with decreased GR expression driven perhaps by increased GR promoter methylation. Such an explanation is consistent with GR downregulating its own receptor [77]. Of course, a limitation of the current study is that we did not measure levels of GR mRNA in the buccal washes and therefore we can only speculate regarding the role of methylation status on expression of this receptor.

A cardinal question raised by the current findings is the developmental window when the methylation pattern of GR buccal DNA is inscribed. Is the methylation pattern fixed in the first few weeks or months as observed by Oberlander et al [24] or perhaps represents years of environmental input? The current investigation is a cross-sectional study and cannot answer this important question and we can therefore only speculate regarding this issue. Clearly, longitudinal studies from early developmental times are required to resolve this questioin. In either scenario, the endpoint of these dynamic environmental challenges coupled with DNA sequence variations represented by common polymorphisms appear to explain a good percentage of individual and gender differences in HPAA responsiveness to social stress. By modeling in the current investigation epigenetic and genetic information we incrementally add to understanding how nature and nature jointly contribute to gender-specific responses to social stress.

Methods

For detailed methods see Methods S1

As described in a previous paper [20], participants were primarily college students from the Hebrew University. Selection criteria stipulated that subjects were <35 years old, had no history of psychiatric or endocrine illness, were currently non-smokers, were not pregnant, had not given birth in the past year, and were not using medication on a regular basis besides single-phase oral contraceptives. Altogether 92 subjects (46 males and 46 females, average age 25.29, S.D. = 3.6) were included in the study. The study was approved by the IRB of Herzog Hospital, Jerusalem and all subjects provided written informed consent. The subjects were tested for HPAA responsiveness to acute social stress [20] the Trier Social Stress Test (TSST) [43], that leverages a ‘threat to the social self’ [2] via public speaking and mental arithmetic, to generate an unambiguous physiological endpoint, indexed by salivary cortisol. Salivary cortisol was samples eight times. DNA was extracted from 20 ml of mouthwash samples using the Master Pure kit (Epicentre,Madison,WI). Pyrosequencing [29] of bisulfite treated buccal DNA was carried out by EpigenDX. In addition, we genotyped these subjects for two salient genes, the serotonin transporter (SLC6A4) promoter region (5-HTTLPR) [30] and the ESR1 promoter region TA repeat [31]. Altogether, we examined the methylation level across 39 CpG sites in GR exon 1F promoter sequence for each subject. All statistical tests were carried out using SPSS version 15. A linear regression model was used to ascertain the effects of sex, GR methylation level and genes on stress response.

Supporting Information

Table S1.

Linear regression analysis (noncanonical NGFI-A CpG#12).

https://doi.org/10.1371/journal.pone.0048597.s001

(DOC)

Methods S1.

https://doi.org/10.1371/journal.pone.0048597.s002

(DOC)

Author Contributions

Conceived and designed the experiments: SE IS RPE. Contributed reagents/materials/analysis tools: SE IS RPE FU SI IK DM MK AK. Wrote the paper: SE IS RPE AK SI. Data collection: FU SI IS SE. Data analysis: SE IS RPE AK IK DM MK.

References

  1. 1. Miller G, Chen E, Cole SW (2009) Health psychology: developing biologically plausible models linking the social world and physical health. Annu Rev Psychol 60: 501–524.
  2. 2. Dickerson SS, Kemeny ME (2004) Acute stressors and cortisol responses: a theoretical integration and synthesis of laboratory research. Psychol Bull 130: 355–391.
  3. 3. McEwen BS (2008) Central effects of stress hormones in health and disease: Understanding the protective and damaging effects of stress and stress mediators. Eur J Pharmacol
  4. 4. Steptoe A, Dockray S, Wardle J (2009) Positive affect and psychobiological processes relevant to health. J Pers 77: 1747–1776.
  5. 5. Steptoe A, Wardle J, Pollard TM, Canaan L, Davies GJ (1996) Stress, social support and health-related behavior: a study of smoking, alcohol consumption and physical exercise. J Psychosom Res 41: 171–180.
  6. 6. Lovallo WR, Farag NH, Vincent AS, Thomas TL, Wilson MF (2006) Cortisol responses to mental stress, exercise, and meals following caffeine intake in men and women. Pharmacol Biochem Behav 83: 441–447.
  7. 7. Nicolson N, Storms C, Ponds R, Sulon J (1997) Salivary cortisol levels and stress reactivity in human aging. J Gerontol A Biol Sci Med Sci 52: M68–75.
  8. 8. Kudielka BM, Hellhammer DH, Wust S (2009) Why do we respond so differently? Reviewing determinants of human salivary cortisol responses to challenge. Psychoneuroendocrinology 34: 2–18.
  9. 9. Kirschbaum C, Kudielka BM, Gaab J, Schommer NC, Hellhammer DH (1999) Impact of gender, menstrual cycle phase, and oral contraceptives on the activity of the hypothalamus-pituitary-adrenal axis. Psychosom Med 61: 154–162.
  10. 10. Kudielka BM, Kirschbaum C (2005) Sex differences in HPA axis responses to stress: a review. Biol Psychol 69: 113–132.
  11. 11. Klein LC, Corwin EJ (2002) Seeing the unexpected: how sex differences in stress responses may provide a new perspective on the manifestation of psychiatric disorders. Curr Psychiatry Rep 4: 441–448.
  12. 12. Chida Y, Hamer M (2008) Chronic psychosocial factors and acute physiological responses to laboratory-induced stress in healthy populations: a quantitative review of 30 years of investigations. Psychol Bull 134: 829–885.
  13. 13. Patchev VK, Almeida OF (1998) Gender specificity in the neural regulation of the response to stress: new leads from classical paradigms. Mol Neurobiol 16: 63–77.
  14. 14. Chrousos GP, Kino T (2009) Glucocorticoid signaling in the cell. Expanding clinical implications to complex human behavioral and somatic disorders. Ann N Y Acad Sci 1179: 153–166.
  15. 15. Turner JD, Muller CP (2005) Structure of the glucocorticoid receptor (NR3C1) gene 5′ untranslated region: identification, and tissue distribution of multiple new human exon 1. J Mol Endocrinol 35: 283–292.
  16. 16. Turner JD, Alt SR, Cao L, Vernocchi S, Trifonova S, et al. (2010) Transcriptional control of the glucocorticoid receptor: CpG islands, epigenetics and more. Biochem Pharmacol
  17. 17. Kumsta R, Moser D, Streit F, Koper JW, Meyer J, et al. (2009) Characterization of a glucocorticoid receptor gene (GR, NR3C1) promoter polymorphism reveals functionality and extends a haplotype with putative clinical relevance. Am J Med Genet B Neuropsychiatr Genet 150B: 476–482.
  18. 18. Kumsta R, Entringer S, Koper JW, van Rossum EF, Hellhammer DH, et al. (2008) Glucocorticoid receptor gene polymorphisms and glucocorticoid sensitivity of subdermal blood vessels and leukocytes. Biol Psychol
  19. 19. Wust S, Van Rossum EF, Federenko IS, Koper JW, Kumsta R, et al. (2004) Common polymorphisms in the glucocorticoid receptor gene are associated with adrenocortical responses to psychosocial stress. J Clin Endocrinol Metab 89: 565–573.
  20. 20. Shalev I, Lerer E, Israel S, Uzefovsky F, Gritsenko I, et al. (2009) BDNF Val66Met polymorphism is associated with HPA axis reactivity to psychological stress characterized by genotype and gender interactions. Psychoneuroendocrinology 34: 382–388.
  21. 21. Weaver IC, Cervoni N, Champagne FA, D'Alessio AC, Sharma S, et al. (2004) Epigenetic programming by maternal behavior. Nat Neurosci 7: 847–854.
  22. 22. Weaver IC, D'Alessio AC, Brown SE, Hellstrom IC, Dymov 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.
  23. 23. McGowan PO, Sasaki A, D'Alessio AC, Dymov S, Labonte B, et al. (2009) Epigenetic regulation of the glucocorticoid receptor in human brain associates with childhood abuse. Nat Neurosci 12: 342–348.
  24. 24. Oberlander TF, Weinberg J, Papsdorf M, Grunau R, Misri S, et al. (2008) Prenatal exposure to maternal depression, neonatal methylation of human glucocorticoid receptor gene (NR3C1) and infant cortisol stress responses. Epigenetics 3: 97–106.
  25. 25. Moser D, Molitor A, Kumsta R, Tatschner T, Riederer P, et al. (2007) The glucocorticoid receptor gene exon 1-F promoter is not methylated at the NGFI-A binding site in human hippocampus. World J Biol Psychiatry 8: 262–268.
  26. 26. Alt SR, Turner JD, Klok MD, Meijer OC, Lakke EA, et al. (2010) Differential expression of glucocorticoid receptor transcripts in major depressive disorder is not epigenetically programmed. Psychoneuroendocrinology 35: 544–556.
  27. 27. Meaney MJ, Szyf M, Seckl JR (2007) Epigenetic mechanisms of perinatal programming of hypothalamic-pituitary-adrenal function and health. Trends Mol Med 13: 269–277.
  28. 28. Meaney MJ, Szyf M (2005) Environmental programming of stress responses through DNA methylation: life at the interface between a dynamic environment and a fixed genome. Dialogues Clin Neurosci 7: 103–123.
  29. 29. Tost J, Gut IG (2007) DNA methylation analysis by pyrosequencing. Nat Protoc 2: 2265–2275.
  30. 30. Canli T, Lesch KP (2007) Long story short: the serotonin transporter in emotion regulation and social cognition. Nat Neurosci 10: 1103–1109.
  31. 31. Westberg L, Melke J, Landen M, Nilsson S, Baghaei F, et al. (2003) Association between a dinucleotide repeat polymorphism of the estrogen receptor alpha gene and personality traits in women. Mol Psychiatry 8: 118–122.
  32. 32. Shansky RM (2009) Estrogen, stress and the brain: progress toward unraveling gender discrepancies in major depressive disorder. Expert Rev Neurother 9: 967–973.
  33. 33. Belmaker RH, Agam G (2008) Major depressive disorder. N Engl J Med 358: 55–68.
  34. 34. Young EA (1995) The role of gonadal steroids in hypothalamic-pituitary-adrenal axis regulation. Crit Rev Neurobiol 9: 371–381.
  35. 35. Champagne FA, Weaver IC, Diorio J, Sharma S, Meaney MJ (2003) Natural variations in maternal care are associated with estrogen receptor alpha expression and estrogen sensitivity in the medial preoptic area. Endocrinology 144: 4720–4724.
  36. 36. Champagne FA, Weaver IC, Diorio J, Dymov S, Szyf M, et al. (2006) Maternal care associated with methylation of the estrogen receptor-alpha1b promoter and estrogen receptor-alpha expression in the medial preoptic area of female offspring. Endocrinology 147: 2909–2915.
  37. 37. Weiser MJ, Handa RJ (2009) Estrogen impairs glucocorticoid dependent negative feedback on the hypothalamic-pituitary-adrenal axis via estrogen receptor alpha within the hypothalamus. Neuroscience 159: 883–895.
  38. 38. Seale JV, Wood SA, Atkinson HC, Harbuz MS, Lightman SL (2004) Gonadal steroid replacement reverses gonadectomy-induced changes in the corticosterone pulse profile and stress-induced hypothalamic-pituitary-adrenal axis activity of male and female rats. J Neuroendocrinol 16: 989–998.
  39. 39. Seale JV, Wood SA, Atkinson HC, Harbuz MS, Lightman SL (2005) Postnatal masculinization alters the HPA axis phenotype in the adult female rat. J Physiol 563: 265–274.
  40. 40. Sundermann EE, Maki PM, Bishop JR (2010) A review of estrogen receptor alpha gene (ESR1) polymorphisms, mood, and cognition. Menopause 17: 874–886.
  41. 41. Mitchell JB, Iny LJ, Meaney MJ (1990) The role of serotonin in the development and environmental regulation of type II corticosteroid receptor binding in rat hippocampus. Brain Res Dev Brain Res 55: 231–235.
  42. 42. Laplante P, Diorio J, Meaney MJ (2002) Serotonin regulates hippocampal glucocorticoid receptor expression via a 5-HT7 receptor. Brain Res Dev Brain Res 139: 199–203.
  43. 43. Kirschbaum C, Pirke KM, Hellhammer DH (1993) The ‘Trier Social Stress Test’–a tool for investigating psychobiological stress responses in a laboratory setting. Neuropsychobiology 28: 76–81.
  44. 44. Federenko IS, Nagamine M, Hellhammer DH, Wadhwa PD, Wust S (2004) The heritability of hypothalamus pituitary adrenal axis responses to psychosocial stress is context dependent. J Clin Endocrinol Metab 89: 6244–6250.
  45. 45. Bartels M, Van den Berg M, Sluyter F, Boomsma DI, de Geus EJ (2003) Heritability of cortisol levels: review and simultaneous analysis of twin studies. Psychoneuroendocrinology 28: 121–137.
  46. 46. Solomon MB, Herman JP (2009) Sex differences in psychopathology: Of gonads, adrenals and mental illness. Physiol Behav
  47. 47. Miller GE, Chen E, Fok AK, Walker H, Lim A, et al. (2009) Low early-life social class leaves a biological residue manifested by decreased glucocorticoid and increased proinflammatory signaling. Proc Natl Acad Sci U S A 106: 14716–14721.
  48. 48. Weidner G, Cain VS (2003) The gender gap in heart disease: lessons from Eastern Europe. Am J Public Health 93: 768–770.
  49. 49. Sher L (2005) Type D personality: the heart, stress, and cortisol. Qjm 98: 323–329.
  50. 50. Weiss LA, Pan L, Abney M, Ober C (2006) The sex-specific genetic architecture of quantitative traits in humans. Nat Genet 38: 218–222.
  51. 51. Dinan TG (1996) Serotonin and the regulation of hypothalamic-pituitary-adrenal axis function. Life Sci 58: 1683–1694.
  52. 52. Way BM, Taylor SE (2010) Social influences on health: is serotonin a critical mediator? Psychosom Med 72: 107–112.
  53. 53. Wust S, Kumsta R, Treutlein J, Frank J, Entringer S, et al. (2009) Sex-specific association between the 5-HTT gene-linked polymorphic region and basal cortisol secretion. Psychoneuroendocrinology 34: 972–982.
  54. 54. Way BM, Taylor SE (2010) The serotonin transporter promoter polymorphism is associated with cortisol response to psychosocial stress. Biol Psychiatry 67: 487–492.
  55. 55. Gotlib IH, Joormann J, Minor KL, Hallmayer J (2007) HPA Axis Reactivity: A Mechanism Underlying the Associations Among 5-HTTLPR, Stress, and Depression. Biol Psychiatry
  56. 56. Philibert RA, Sandhu H, Hollenbeck N, Gunter T, Adams W, et al. (2008) The relationship of 5HTT (SLC6A4) methylation and genotype on mRNA expression and liability to major depression and alcohol dependence in subjects from the Iowa Adoption Studies. Am J Med Genet B Neuropsychiatr Genet 147B: 543–549.
  57. 57. Philibert R, Madan A, Andersen A, Cadoret R, Packer H, et al. (2007) Serotonin transporter mRNA levels are associated with the methylation of an upstream CpG island. Am J Med Genet B Neuropsychiatr Genet 144B: 101–105.
  58. 58. van IMH, Caspers K, Bakermans-Kranenburg MJ, Beach SR, Philibert R Methylation matters: interaction between methylation density and serotonin transporter genotype predicts unresolved loss or trauma. Biol Psychiatry 68: 405–407.
  59. 59. Patchev VK, Hayashi S, Orikasa C, Almeida OF (1995) Implications of estrogen-dependent brain organization for gender differences in hypothalamo-pituitary-adrenal regulation. FASEB J 9: 419–423.
  60. 60. Vamvakopoulos NC, Chrousos GP (1993) Evidence of direct estrogenic regulation of human corticotropin-releasing hormone gene expression. Potential implications for the sexual dimophism of the stress response and immune/inflammatory reaction. J Clin Invest 92: 1896–1902.
  61. 61. Miller WJ, Suzuki S, Miller LK, Handa R, Uht RM (2004) Estrogen receptor (ER)beta isoforms rather than ERalpha regulate corticotropin-releasing hormone promoter activity through an alternate pathway. J Neurosci 24: 10628–10635.
  62. 62. Lefebvre H, Contesse V, Delarue C, Feuilloley M, Hery F, et al. (1992) Serotonin-induced stimulation of cortisol secretion from human adrenocortical tissue is mediated through activation of a serotonin4 receptor subtype. Neuroscience 47: 999–1007.
  63. 63. Moser U, Wadsak W, Spindelegger C, Mitterhauser M, Mien LK, et al. (2010) Hypothalamic serotonin-1A receptor binding measured by PET predicts the plasma level of dehydroepiandrosterone sulfate in healthy women. Neurosci Lett 476: 161–165.
  64. 64. Pan L, Gilbert F (1992) Activation of 5-HT1A receptor subtype in the paraventricular nuclei of the hypothalamus induces CRH and ACTH release in the rat. Neuroendocrinology 56: 797–802.
  65. 65. Uher R, McGuffin P (2010) The moderation by the serotonin transporter gene of environmental adversity in the etiology of depression: 2009 update. Mol Psychiatry 15: 18–22.
  66. 66. Risch N, Herrell R, Lehner T, Liang KY, Eaves L, et al. (2009) Interaction between the serotonin transporter gene (5-HTTLPR), stressful life events, and risk of depression: a meta-analysis. Jama 301: 2462–2471.
  67. 67. Munafo MR, Durrant C, Lewis G, Flint J (2008) Gene x Environment Interactions at the Serotonin Transporter Locus. Biol Psychiatry
  68. 68. Caspi A, Sugden K, Moffitt TE, Taylor A, Craig IW, et al. (2003) Influence of life stress on depression: moderation by a polymorphism in the 5-HTT gene. Science 301: 386–389.
  69. 69. Caspi A, McClay J, Moffitt TE, Mill J, Martin J, et al. (2002) Role of genotype in the cycle of violence in maltreated children. Science 297: 851–854.
  70. 70. Lewis DA (2002) The human brain revisited: opportunities and challenges in postmortem studies of psychiatric disorders. Neuropsychopharmacology 26: 143–154.
  71. 71. Kaminsky ZA, Tang T, Wang SC, Ptak C, Oh GH, et al. (2009) DNA methylation profiles in monozygotic and dizygotic twins. Nat Genet
  72. 72. Gladkevich A, Kauffman HF, Korf J (2004) Lymphocytes as a neural probe: potential for studying psychiatric disorders. Prog Neuropsychopharmacol Biol Psychiatry 28: 559–576.
  73. 73. Rollins B, Martin MV, Morgan L, Vawter MP (2010) Analysis of whole genome biomarker expression in blood and brain. Am J Med Genet B Neuropsychiatr Genet 153B: 919–936.
  74. 74. Herman JP, Figueiredo H, Mueller NK, Ulrich-Lai Y, Ostrander MM, et al. (2003) Central mechanisms of stress integration: hierarchical circuitry controlling hypothalamo-pituitary-adrenocortical responsiveness. Front Neuroendocrinol 24: 151–180.
  75. 75. Bornstein SR, Engeland WC, Ehrhart-Bornstein M, Herman JP (2008) Dissociation of ACTH and glucocorticoids. Trends Endocrinol Metab 19: 175–180.
  76. 76. Walker JJ, Terry JR, Lightman SL (2010) Origin of ultradian pulsatility in the hypothalamic-pituitary-adrenal axis. Proc Biol Sci 277: 1627–1633.
  77. 77. Oakley RH, Cidlowski JA (1993) Homologous down regulation of the glucocorticoid receptor: the molecular machinery. Crit Rev Eukaryot Gene Expr 3: 63–88.