Childhood adversity lays a fragile foundation for health across the life span. Adverse childhood experiences including child maltreatment, trauma, and exposure to other contextual stressors associated with poverty are major risk factors for the development of psychiatric disorders as well as other medical conditions in children and adults (Benjet, Borges, & Medina-Mora, Reference Benjet, Borges and Medina-Mora2010; Cohen, Janicki-Deverts, & Miller, Reference Cohen, Janicki-Deverts and Miller2007; Felitti et al., Reference Felitti, Anda, Nordenberg, Williamson, Spitz and Edwards1998; Green et al., Reference Green, McLaughlin, Berglund, Gruber, Sampson and Zaslavsky2010; Slopen, Koenen, & Kubzansky, Reference Slopen, Koenen and Kubzansky2014). One study found that adults with numerous adverse childhood experiences died nearly 20 years earlier than others (Brown et al., Reference Brown, Anda, Tiemeier, Felitti, Edwards and Croft2009). The mechanisms underlying the links between childhood adversity and poor health outcomes are not fully understood. It is clear that childhood adversity can alter the physiologic stress response, and it has been posited that changes to the stress response system may underlie the connection between early adversity and psychiatric and other health consequences (McEwen, Reference McEwen2013; Ridout, Carpenter, & Tyrka, Reference Ridout, Carpenter and Tyrka2016; Ridout et al., Reference Ridout, Ridout, Kao, Carpenter, Philip and Tyrka2015; Tyrka, Burgers, Philip, Price, & Carpenter, Reference Tyrka, Burgers, Philip, Price and Carpenter2013). There is growing appreciation that epigenetic modifications to genes that regulate the stress response are a likely mechanism by which the early environment has a long-lasting impact on stress biology. In this selective review, we describe the role of glucocorticoid signaling and epigenetic modifications in the biological response to environmental exposures and review emerging findings from our laboratory and others that suggest this mechanism underlying risk for psychiatric and other health problems.
Alterations of Hypothalamus–Pituitary–Adrenal (HPA) Axis Function and Glucocorticoid Signaling With Childhood Adversity
Research examining the biologic underpinnings of the associations between childhood maltreatment and psychopathology highlights the importance of the physiologic stress response system, and in particular, the HPA axis. In response to stressful stimuli, glucocorticoids are released and exert cellular responses by binding at the intracellular glucocorticoid receptor (GR). GRs are distributed throughout the body and brain where they regulate basal physiologic function and effect changes in various organ systems and tissues that promote adaptive responding to acute stressors (de Kloet, Joels, & Holsboer, Reference de Kloet, Joels and Holsboer2005; Kadmiel & Cidlowski, Reference Kadmiel and Cidlowski2013). Activation of the GR through cortisol binding at the hypothalamus and pituitary engages a negative feedback loop that inhibits further release of cortisol and prevents damaging effects of extreme or chronic activation (Herman, McKlveen, Solomon, Carvalho-Netto, & Myers, Reference Herman, McKlveen, Solomon, Carvalho-Netto and Myers2012; Laryea, Muglia, Arnett, & Muglia, Reference Laryea, Muglia, Arnett and Muglia2015).
Excessive stimulation by severe or prolonged stress may result in adaptive changes that alter function of the HPA axis (Doom, Cicchetti, & Rogosch, Reference Doom, Cicchetti and Rogosch2014; Fries, Hesse, Hellhammer, & Hellhammer, Reference Fries, Hesse, Hellhammer and Hellhammer2005; Heim, Ehlert, & Hellhammer, Reference Heim, Ehlert and Hellhammer2000; McEwen, Reference McEwen2007; Pryce et al., Reference Pryce, Ruedi-Bettschen, Dettling, Weston, Russig and Ferger2005; Tyrka et al., Reference Tyrka, Wier, Price, Ross, Anderson and Wilkinson2008). Frequent or excessive activation of the HPA axis in response to stress exposure can progress to a counterregulatory state of chronic adrenal stress hyporeactivity (Fries et al., Reference Fries, Hesse, Hellhammer and Hellhammer2005; Heim et al., Reference Heim, Ehlert and Hellhammer2000; McEwen, Reference McEwen2007; Pryce et al., Reference Pryce, Ruedi-Bettschen, Dettling, Weston, Russig and Ferger2005). The timing of adversity exposure during development may impact HPA axis programming (Bosch et al., Reference Bosch, Riese, Reijneveld, Bakker, Verhulst and Ormel2012), and chronic stress exposure may have the most profound effects, as repeated attempts at maintaining homeostasis alter set points and response characteristics of stress-responsive physiologic systems (Lee & Sawa, Reference Lee and Sawa2014; McEwen, Nasca, & Gray, Reference McEwen, Nasca and Gray2016).
Several studies from our group and others have revealed that early adversity is linked to abnormalities in HPA axis function in both children and adults (Gonzalez, Reference Gonzalez2013; Gunnar & Vazquez, Reference Gunnar and Vazquez2001; McCrory, De Brito, & Viding, Reference McCrory, De Brito and Viding2010). Early work in our laboratory with a sample of 50 healthy adults demonstrated that those with a history of moderate to severe childhood maltreatment (n = 23) exhibited blunted cortisol reactivity to a standardized psychosocial stress paradigm, the Trier Social Stress Test, compared to adults with no maltreatment history (Carpenter et al., Reference Carpenter, Carvalho, Tyrka, Wier, Mello and Mello2007). That these adults did not have any active psychiatric conditions, including major depressive disorder (MDD) or posttraumatic stress disorder (PTSD), suggests that early adversity poses a significant independent risk factor for altered HPA functioning even among individuals who are otherwise healthy. Several studies have replicated these findings (Carpenter, Shattuck, Tyrka, Geracioti, & Price, Reference Carpenter, Shattuck, Tyrka, Geracioti and Price2011; Carpenter et al., Reference Carpenter, Tyrka, Ross, Khoury, Anderson and Price2009; Elzinga et al., Reference Elzinga, Roelofs, Tollenaar, Bakvis, van Pelt and Spinhoven2008; Klaassens, Giltay, van Veen, Veen, & Zitman, Reference Klaassens, Giltay, van Veen, Veen and Zitman2010; Klaassens et al., Reference Klaassens, van Noorden, Giltay, van Pelt, van Veen and Zitman2009; Tyrka et al., Reference Tyrka, Wier, Price, Ross, Anderson and Wilkinson2008). In another study of 230 adults, our group demonstrated that a history of childhood emotional abuse was also associated with a blunted cortisol response to a pharmacological challenge (Carpenter et al., Reference Carpenter, Tyrka, Ross, Khoury, Anderson and Price2009). Similar to earlier work in our laboratory, none of the adult participants currently met criteria for mood or anxiety disorders, including MDD and PTSD, and these links were observed even when controlling for subthreshold symptoms of psychopathology and past psychiatric conditions. Furthermore, the association of emotional maltreatment and cortisol reactivity was stronger among older adults in the sample, suggesting that the effects of early adversity contribute to “wear and tear” on this system across the life span. Exaggerated cortisol responses have also sometimes been seen in association with early stress. In a study of 88 healthy adults, we demonstrated that the nature of childhood experiences moderated the cortisol response to pharmacologic challenge (Tyrka et al., Reference Tyrka, Wier, Price, Ross, Anderson and Wilkinson2008). Childhood parental death or desertion (N = 44) was linked with an exaggerated cortisol response to the dexamethasone corticotropin-releasing hormone test, a pharmacological challenge designed to assess HPA function. This response was moderated by the type of loss and the level of parental care. Those who experienced parental desertion coupled with low parental care demonstrated a blunted cortisol response, supporting the hypothesis that severe or chronic forms of stress may be more likely to lead to cortisol hyporeactivity. Taken together, this work from our laboratory underscores the persistent effects of childhood maltreatment on neuroendocrine function into adulthood, and highlights the importance of adversity characteristics as determinants of the pattern of HPA axis dysfunction.
The long-lasting effects of early adversity on HPA axis function have also been observed in numerous studies of children (Doom & Gunnar, Reference Doom and Gunnar2013; Gonzalez, Reference Gonzalez2013; van Andel, Jansen, Grietens, Knorth, & van der Gaag, Reference van Andel, Jansen, Grietens, Knorth and van der Gaag2014). Similar to work with adults, several studies have shown blunted diurnal cortisol concentrations with adversity, but others have shown no difference or elevated cortisol levels. A number of studies have examined the determinants of diurnal cortisol patterns in children who experience significant adversity such as maltreatment. A study of 187 maltreated and 154 nonmaltreated 5- to 13-year-old children found that overall cortisol levels across 20 weeks did not differ between maltreated and nonmaltreated children, but the maltreated group showed elevated within-person variation in cortisol values, and those with higher initial cortisol levels had cortisol suppression over time (Doom et al., Reference Doom, Cicchetti and Rogosch2014). These findings are consistent with the hypothesis that glucocorticoid downregulation may occur over time in response to elevations in cortisol levels. In addition, maltreatment severity, timing, and the number of maltreatment subtypes predicted cortisol variability. Both maltreatment and greater cortisol variability were associated with more behavior problems (Doom et al., Reference Doom, Cicchetti and Rogosch2014). Attenuation of HPA function has been reported in children referred for child protective services. Blunted cortisol was linked to more externalizing behaviors and mediated the relationship between Child Protective Service involvement and externalizing symptoms (Bernard, Zwerling, & Dozier, Reference Bernard, Zwerling and Dozier2015), and an attachment-based intervention normalized wake-up and diurnal cortisol levels in this study (Bernard, Hostinar, & Dozier, Reference Bernard, Hostinar and Dozier2015). The role of HPA axis dysfunction in the pathogenesis of psychopathology and the response to interventions is discussed further below.
Role of the HPA Axis and Glucocorticoid Signaling in Risk for Psychopathology
Several lines of evidence support the hypothesis that excessive exposure to stress-induced glucocorticoid activity may be involved in the pathogenesis of stress-related psychiatric disorders, including depressive and anxiety disorders. In rodent models, prolonged stress or glucocorticoid exposure produces anxiety- and depressive-like behaviors (Maccari, Krugers, Morley-Fletcher, Szyf, & Brunton, Reference Maccari, Krugers, Morley-Fletcher, Szyf and Brunton2014; Skupio et al., Reference Skupio, Tertil, Sikora, Golda, Wawrzczak-Bargiela and Przewlocki2015; van Donkelaar et al., Reference van Donkelaar, Vaessen, Pawluski, Sierksma, Blokland and Canete2014). In children and adolescents, altered cortisol responses are associated with internalizing behaviors, externalizing behaviors, suicidal ideation, PTSD, and depression (Braquehais, Picouto, Casas, & Sher, Reference Braquehais, Picouto, Casas and Sher2012; Doom & Gunnar, Reference Doom and Gunnar2013; Faravelli et al., Reference Faravelli, Lo Sauro, Lelli, Pietrini, Lazzeretti and Godini2012; Guerry & Hastings, Reference Guerry and Hastings2011; Ruttle et al., Reference Ruttle, Shirtcliff, Serbin, Fisher, Stack and Schwartzman2011). In a racially mixed community sample of 102 boys aged 8–11, we found that afternoon basal cortisol concentrations were positively associated with internalizing behavior problems, social problems, and emotionality. In addition, greater declines across a home visit challenge task were significantly associated with internalizing behavior, as well as social, attention, and thought problems (Tyrka et al., Reference Tyrka, Kelly, Graber, DeRose, Lee and Warren2010). In a 2-year follow-up assessment of 78 of the boys, greater cortisol declines across the home visit task were predictive of internalizing and externalizing behaviors, as well as attention and social problems. Moreover, morning and afternoon cortisol concentrations at the initial assessment significantly predicted the later development of child depressive symptoms (Tyrka, Lee, et al., Reference Tyrka, Lee, Graber, Clement, Kelly and DeRose2012).
As discussed above, adversity and trauma can be associated with both exaggerated and attenuated profiles of neuroendocrine function (Doom & Gunnar, Reference Doom and Gunnar2013; Miller, Chen, & Zhou, Reference Miller, Chen and Zhou2007; Morris, Compas, & Garber, Reference Morris, Compas and Garber2012). A meta-analysis of 6,000 trauma-exposed participants with and without PTSD (Morris et al., Reference Morris, Compas and Garber2012) showed that relative to nonexposed participants, those with PTSD as well as those with both PTSD and MDD had lower morning cortisol, lower daily output of cortisol, and lower cortisol response to dexamethasone. Trauma-exposed participants (including those with childhood neglect and abuse) without PTSD or MDD also showed blunted post-dexamethasone cortisol and afternoon/evening cortisol (Morris et al., Reference Morris, Compas and Garber2012). The only group to show elevated cortisol had comorbid PTSD/MDD, and only evening cortisol was elevated. It is interesting that this was also the only group to also show an effect of developmental timing: after controlling for the time elapsed since the focal trauma, exposure during adulthood was associated with large negative effects on morning cortisol and post-dexamethasone cortisol, whereas exposure in childhood had negligible effects in this group. However, it is important to note that overall, in both PTSD groups, time elapsed since exposure to the focal trauma was significantly associated with lower daily and post-dexamethasone cortisol levels, and with lower afternoon/evening cortisol at trend level. Other work has identified additional characteristics that influence patterns of adrenocortical dysregulation, including the nature of the trauma and other contextual factors, timing of the exposure, comorbid psychiatric and other conditions, and genetic background (De Bellis & Zisk, Reference De Bellis and Zisk2014; Doom & Gunnar, Reference Doom and Gunnar2013; Struber, Struber, & Roth, Reference Struber, Struber and Roth2014).
Given the critical negative feedback role of the GR, alterations of the number and sensitivity of the GR may explain abnormalities of HPA function associated with adversity and psychopathology. Depressed patients who have nonsuppression to the dexamethasone-suppression test have impaired leukocyte GR responses to GR agonists (Gormley et al., Reference Gormley, Lowy, Reder, Hospelhorn, Antel and Meltzer1985; Lowy, Reder, Gormley, & Meltzer, Reference Lowy, Reder, Gormley and Meltzer1988). Patients with PTSD, who often have blunted cortisol responses to stress, show “supersuppression” to dexamethasone and associated negative feedback sensitivity of lymphocyte GRs (Rohleder, Joksimovic, Wolf, & Kirschbaum, Reference Rohleder, Joksimovic, Wolf and Kirschbaum2004; Yehuda, Reference Yehuda2001; Yehuda, Golier, et al., Reference Yehuda, Flory, Pratchett, Buxbaum, Ising and Holsboer2010; Yehuda, Halligan, Grossman, Golier, & Wong, Reference Yehuda, Halligan, Grossman, Golier and Wong2002; Yehuda et al., Reference Yehuda, Yang, Golier, Tischler, Liong and Decker2004; Yehuda, Yang, Guo, Makotkine, & Singh, Reference Yehuda, Yang, Guo, Makotkine and Singh2003). High premorbid GR expression is a risk factor for PTSD (van Zuiden et al., Reference van Zuiden, Geuze, Willemen, Vermetten, Maas and Heijnen2011). Taken together, these findings indicate that abnormalities in leukocyte GR number and function may be risk factors for the development of stress-related psychiatric disorders, and may reflect effects due to early life stress exposure (Anacker, Zunszain, Carvalho, & Pariante, Reference Anacker, Zunszain, Carvalho and Pariante2011).
The GR is distributed throughout limbic brain regions and numerous other organ systems, so that excessive glucocorticoid activation at these receptors or changes to the sensitivity of this system may underlie psychiatric and other stress-related conditions. Findings from animal models and human neuroimaging studies indicate that the hippocampus and amygdala are highly vulnerable to the effects of early life stress and trauma (Lupien, McEwen, Gunnar, & Heim, Reference Lupien, McEwen, Gunnar and Heim2009; Rifkin-Graboi et al., Reference Rifkin-Graboi, Kong, Sim, Sanmugam, Broekman and Chen2015; Tottenham & Sheridan, Reference Tottenham and Sheridan2009). The hippocampus, amygdala, and prefrontal cortex have a high density of GRs, and animal models show that exposure to stress or glucocorticoids, along with effects on other stress mediators, alters neural structure in these regions (McEwen et al., Reference McEwen, Nasca and Gray2016). Chronic stress and glucocorticoids impair neuronal growth and survival in these brain regions; this may explain neuroimaging findings of reduced brain region volumes in association with early stress (McEwen et al., Reference McEwen, Bowles, Gray, Hill, Hunter and Karatsoreos2015). In the hippocampus, chronic stress and glucocorticoid treatment inhibit neurogenesis, cell proliferation, and dendritic branching, and induce cell loss and atrophy (Duman, Reference Duman2009; McEwen et al., Reference McEwen, Nasca and Gray2016; van der Kooij, Grosse, Zanoletti, Papilloud, & Sandi, Reference van der Kooij, Grosse, Zanoletti, Papilloud and Sandi2015). In contrast, in the amygdala, which mediates fear responses, chronic stress may induce a proliferative effect on neuronal dendritic branching and spine density (Vyas, Jadhav, & Chattarji, Reference Vyas, Jadhav and Chattarji2006). These changes in dendritic length are accompanied by elevations in glucocorticoids (Lakshminarasimhan & Chattarji, Reference Lakshminarasimhan and Chattarji2012), and administration of glucocorticoids can elicit similar dendritic lengthening while increasing anxiety behaviors (Mitra & Sapolsky, Reference Mitra and Sapolsky2008). It is interesting that administration of low to moderate doses of glucocorticoids at the time of acute or chronic stress exposure prevents dendritic changes in the amygdala and the development of anxiety (Rao, Anilkumar, McEwen, & Chattarji, Reference Rao, Anilkumar, McEwen and Chattarji2012; Zohar et al., Reference Zohar, Yahalom, Kozlovsky, Cwikel-Hamzany, Matar and Kaplan2011), suggesting that glucocorticoids may impart protection under conditions of moderate stress but risk with more extreme exposures. Recent evidence shows that glucocorticoid signaling in the prefrontal cortex modifies fear conditioning and responses in animal models (Reis, Almada, Fogaca, & Brandao, Reference Reis, Almada, Fogaca and Brandao2015; Wislowska-Stanek et al., Reference Wislowska-Stanek, Lehner, Skorzewska, Maciejak, Szyndler and Turzynska2013). Similar to the hippocampus, chronic stress also causes structural remodeling of the prefrontal cortex that is reversible after the termination of stress; this is not as readily reversible in aged animals (Bloss, Janssen, McEwen, & Morrison, Reference Bloss, Janssen, McEwen and Morrison2010). Thus, animal models show that stress and glucocorticoid exposure have the potential to cause structural changes to the brain circuitry critical to affective and behavioral adaptation, and thus likely underlie associations between early adversity and risk for psychopathology.
Epigenetic Mechanisms of Risk
Epigenetic changes to DNA are a mechanism by which the environment can impact gene activity and expression (Hernando-Herraez, Garcia-Perez, Sharp, & Marques-Bonet, Reference Hernando-Herraez, Garcia-Perez, Sharp and Marques-Bonet2015). Methylation, thought to be the most stable form of epigenetic change, is a likely mechanism by which stress exposure has long-lived effects. In mammals, methylation mainly occurs at cytosine nucleotide–phosphate–guanine nucleotide (CpG) sites, which are regions in the DNA where a cytosine nucleotide occurs next to a guanine nucleotide. Regions of densely clustered CpGs, known as CpG islands (CGIs), occur where gene transcription is initiated (Deaton & Bird, Reference Deaton and Bird2011). Methylation at CGIs can lead to alterations in chromatin architecture and inhibit transcription factor binding to gene promoter regions, resulting in reduced gene expression. Consistent with this, genes such as the GR gene (nuclear receptor subfamily 3, group C, member 1 [NR3C1]) that are highly expressed typically have low levels of promoter methylation (Brenet et al., Reference Brenet, Moh, Funk, Feierstein, Viale and Socci2011; Moore, Le, & Fan, Reference Moore, Le and Fan2013). Low levels of methylation typically characterize regions of the genome that are open to transcriptional regulation by methylation, such as CGIs (Deaton & Bird, Reference Deaton and Bird2011; Liyanage et al., Reference Liyanage, Jarmasz, Murugeshan, Del Bigio, Rastegar and Davie2014). In contrast, high methylation is often seen at CpG sites outside of CGIs and may have roles in genome regulation outside of genetic transcription (Hernando-Herraez et al., Reference Hernando-Herraez, Garcia-Perez, Sharp and Marques-Bonet2015).
Epigenetic Processes Related to HPA Functioning
In response to stressful stimuli, the HPA axis is triggered and cortisol is released from the adrenal cortex. Cortisol exerts cellular responses by binding at the intracellular GR (Kadmiel & Cidlowski, Reference Kadmiel and Cidlowski2013). GRs are distributed throughout the body and brain, where they regulate basal physiologic function and effect changes that promote adaptive responses to acute stressors (de Kloet et al., Reference de Kloet, Joels and Holsboer2005; Kadmiel & Cidlowski, Reference Kadmiel and Cidlowski2013). Cortisol binding to the GR in the cytosol induces translocation of the GR into the nucleus, where it can activate and repress expression of a wide range of genes (Galon et al., Reference Galon, Franchimont, Hiroi, Frey, Boettner and Ehrhart-Bornstein2002), thereby regulating systems necessary to cope with stressors. It is interesting that glucocorticoids induce the expression of different genes depending on the stress history of the organism (McEwen et al., Reference McEwen, Bowles, Gray, Hill, Hunter and Karatsoreos2015). In addition to impacting gene expression, cortisol binding at GRs affects cellular signaling pathways and mitochondrial function (McEwen, Reference McEwen2015).
(Color online) FK506 binding protein 51 (FKBP5) and glucocorticoid signaling. Cortisol circulating through the blood stream enters the intracellular space and binds to the glucocorticoid receptor (GR). The cortisol-bound receptor then translocates to the nucleus to exert its effects on gene expression. Methylation of the GR gene (nuclear receptor subfamily 3, group C, member 1 [NR3C1]) reduces GR gene expression, so there are fewer GRs available to bind to cortisol. FKBP5 decreases sensitivity of the GR to cortisol. GR activation by cortisol binding results in rapid induction of FKBP5 gene expression, and FKBP5 then binds to the GR and decreases its ability to bind cortisol and to translocate to the nucleus. Thus, FKBP5 exerts a negative feedback loop on cortisol activity. Methylation interferes with gene expression, and demethylation at intron 7 of FKBP5 is associated with increased FKBP5 gene expression and decreased GR sensitivity. Childhood maltreatment and chronic glucocorticoid administration have been linked with demethylation of FKBP5 intron 7.
As discussed above, activation of the GR through cortisol binding at the hypothalamus and pituitary also triggers a negative feedback mechanism that inhibits further cortisol release, preventing the damaging effects of chronic HPA axis activation (Herman et al., Reference Herman, McKlveen, Solomon, Carvalho-Netto and Myers2012; Laryea et al., Reference Laryea, Muglia, Arnett and Muglia2015). Changes in GR number and function in the brain and in peripheral cells such as leukocytes have been shown with PTSD, MDD, and early stress exposure (Barden, Reference Barden2004; Klengel et al., Reference Klengel, Mehta, Anacker, Rex-Haffner, Pruessner and Pariante2013; Provencal et al., Reference Provencal, Suderman, Guillemin, Massart, Ruggiero and Wang2012; van Zuiden et al., Reference van Zuiden, Geuze, Willemen, Vermetten, Maas and Heijnen2011; Yehuda, Flory, et al., Reference Yehuda, Golier, Bierer, Mikhno, Pratchett and Burton2010, Reference Yehuda, Daskalakis, Bierer, Bader, Klengel and Holsboer2015; Yehuda & Seckl, Reference Yehuda and Seckl2011). GR-mediated negative feedback is critical to regulate the activity of the HPA axis. Alterations in GR number or function in the cell can influence the activity of this system and biological adaptation to stressful and traumatic experiences. Below, we describe how epigenetic modifications to two HPA regulatory genes, NR3C1 and FK506 binding protein 51 (FKBP5), affect GR transcription and activity, and may alter responses to stress and impart risk for psychopathology.
Childhood adversity and methylation of NR3C1
Human GR is encoded by the NR3C1 gene (which is located on chromosome 5q31–32) and contains eight translated exons (numbered 2 through 9) and nine untranslated alternative first exons (Daskalakis & Yehuda, Reference Daskalakis and Yehuda2014). Methylation of CGIs in the GR alternative first exons controls tissue-specific expression of the GR (Turner, Pelascini, Macedo, & Muller, Reference Turner, Pelascini, Macedo and Muller2008) and have a role in the translational control of the GR, influencing total GR levels, as well as trafficking to the cell surface (Turner, Vernocchi, Schmitz, & Muller, Reference Turner, Vernocchi, Schmitz and Muller2014). There is now substantial evidence that methylation of NR3C1 is responsive to environmental exposures in both the prenatal period and during early childhood. Animal models have been highly informative for understanding mechanisms underlying links between environmental exposures, NR3C1 methylation, and subsequent gene expression. In rodents, low levels of maternal care (low frequency of licking and grooming behaviors and arched-back nursing) have been linked to greater methylation of the rodent GR gene nr3c1 in the hippocampus and cerebellum (Kosten & Nielsen, Reference Kosten and Nielsen2014; Weaver et al., Reference Weaver, Cervoni, Champagne, D'Alessio, Sharma and Seckl2004), specifically of the region homologous to the human alternative exon 1F. Methylation of nr3c1 in turn contributes to reduced nr3c1 gene expression. When methylation occurs at the binding site for the transcription factor nerve growth factor induced protein A (NGFI-A), it interferes with gene transcription; methylation at other CpG sites may interfere with transcription through other mechanisms (Armstrong, Lesseur, Conradt, Lester, & Marsit, Reference Armstrong, Lesseur, Conradt, Lester and Marsit2014; Turner et al., Reference Turner, Alt, Cao, Vernocchi, Trifonova and Battello2010). Methylation of this region of nr3c1 and the associated reductions in GR number and GR-mediated negative feedback, have been linked to increased glucocorticoid secretion and behavioral distress (for a review, see(Zhang, Labonte, Wen, Turecki, & Meaney, Reference Zhang, Labonte, Wen, Turecki and Meaney2013). This groundbreaking mechanistic work, coupled with evidence that both pre- and postnatal stressors contribute to methylation of nr3c1 in rodents (Kundakovic, Lim, Gudsnuk, & Champagne, Reference Kundakovic, Lim, Gudsnuk and Champagne2013; Lillycrop et al., Reference Lillycrop, Slater-Jefferies, Hanson, Godfrey, Jackson and Burdge2007; Szyf, Reference Szyf2013; Witzmann, Turner, Meriaux, Meijer, & Muller, Reference Witzmann, Turner, Meriaux, Meijer and Muller2012), complements research in humans reviewed below.
In humans, there is now compelling evidence that NR3C1 methylation is responsive to stress in both the pre- and postnatal periods. The majority of this work has focused on promoter methylation at exon 1F. Prenatal exposure to adverse conditions including maternal depression and anxiety, intimate partner violence, and war-related stressors (such as rape and refugee status) has been linked to increased methylation of NR3C1 at exon 1F in several studies (Braithwaite, Kundakovic, Ramchandani, Murphy, & Chamoagne, Reference Braithwaite, Kundakovic, Ramchandani, Murphy and Champagne2015; Conradt, Lester, Appleton, Armstrong, & Marsit, Reference Conradt, Lester, Appleton, Armstrong and Marsit2013; Hompes et al., Reference Hompes, Izzi, Gellens, Morreels, Fieuws and Pexsters2013; Kertes et al., Reference Kertes, Kamin, Hughes, Rodney, Bhatt and Mulligan2016; Mulligan, D'Errico, Stees, & Hughes, Reference Mulligan, D'Errico, Stees and Hughes2012; Oberlander et al., Reference Oberlander, Weinberg, Papsdorf, Grunau, Misri and Devlin2008; Radtke et al., Reference Radtke, Ruf, Gunter, Dohrmann, Schauer and Meyer2011). Links with prenatal exposures have been demonstrated in DNA from a variety of cell types and tissues including umbilical cord blood (Hompes et al., Reference Hompes, Izzi, Gellens, Morreels, Fieuws and Pexsters2013; Kertes et al., Reference Kertes, Kamin, Hughes, Rodney, Bhatt and Mulligan2016; Mulligan et al., Reference Mulligan, D'Errico, Stees and Hughes2012; Oberlander et al., Reference Oberlander, Weinberg, Papsdorf, Grunau, Misri and Devlin2008), placenta (Conradt et al., Reference Conradt, Lester, Appleton, Armstrong and Marsit2013; Kertes et al., Reference Kertes, Kamin, Hughes, Rodney, Bhatt and Mulligan2016), buccal cells in infancy (Braithwaite et al., Reference Braithwaite, Kundakovic, Ramchandani, Murphy and Champagne2015), and whole blood in adolescents (Radtke et al., Reference Radtke, Ruf, Gunter, Dohrmann, Schauer and Meyer2011).
Work from our laboratory and others has also demonstrated that stress exposure in childhood is linked with methylation of NR3C1. In a sample of 99 healthy adults with no history of psychiatric disorders, our group found that early adversity, including childhood maltreatment, parental loss, and low levels of parental care, was associated with increased methylation of NR3C1 at exon 1F (Tyrka, Price, Marsit, Walters, & Carpenter, Reference Tyrka, Price, Marsit, Walters and Carpenter2012). That these links were observed in adults with no history of psychopathology is consistent with work in our laboratory linking early adversity to cortisol production in healthy adults as described above (Carpenter et al., Reference Carpenter, Carvalho, Tyrka, Wier, Mello and Mello2007, Reference Carpenter, Tyrka, Ross, Khoury, Anderson and Price2009), and it suggests that early adversity exerts a lasting effect on the biological stress response system that is independent of effects of stress-related disorders or medications.
This work has been more recently applied to children. In a study of 184 impoverished maltreated and nonmaltreated preschool-aged children, we found that exposure to early adversity was linked with increased methylation of saliva DNA NR3C1 at exon 1F among preschoolers (Tyrka et al., Reference Tyrka, Parade, Eslinger, Marsit, Lesseur and Armstrong2015). Both past month and lifetime contextual stress assessed during an interview in the home, as well as a composite measure of adversity exposure, were positively associated with mean methylation across the region. These links were also observed at several individual CpG sites in this region, including CpG sites that are known to exert a functional effect on the HPA axis given their role as a transcription factor binding site. These findings are consistent with a recent study of older children between 11 and 14 years of age demonstrating that physical maltreatment was associated with increased methylation of NR3C1 at exon 1F in whole blood (Romens, McDonald, Svaren, & Pollak, Reference Romens, McDonald, Svaren and Pollak2015). Associations of childhood adversity and increased methylation of NR3C1 in blood at exon 1F have also been recently observed in a population-based sample of adolescents (van der Knaap et al., Reference van der Knaap, Riese, Hudziak, Verbiest, Verhulst and Oldehinkel2014). Collectively, this emerging work demonstrates compelling evidence that methylation of NR3C1 at exon 1F, is sensitive to childhood stress exposure.
A few studies have also examined methylation of alternate first exons of NR3C1 in relation to stress exposure. In our study of preschoolers with early adversity described above, we demonstrated that maltreated children had greater methylation at exon 1D in saliva DNA than preschoolers with no maltreatment history (Tyrka et al., Reference Tyrka, Parade, Eslinger, Marsit, Lesseur and Armstrong2015). This effect was consistent with the links observed between early adversity and methylation at exon 1F described above. Methylation at exon 1D assessed in cord blood is also sensitive to maternal pregnancy-related anxiety symptoms, and is associated with maternal diurnal cortisol levels in the first trimester (Hompes et al., Reference Hompes, Izzi, Gellens, Morreels, Fieuws and Pexsters2013). Exons 1B, 1C, and 1H have also been demonstrated to be sensitive to child maltreatment in postmortem hippocampal tissue of adult suicide victims (Labonte et al., Reference Labonte, Yerko, Gross, Mechawar, Meaney and Szyf2012). Taken together, these studies suggest that although most prior work has focused on methylation of NR3C1 at exon 1F, alternate first exons in the promoter of NR3C1 also respond to stress exposure. Important questions remain regarding the functional role of each of these alternate first exons in the stress response in various tissues, and precisely how this relates to risk for psychiatric disorders.
Emerging work supports the hypothesis that methylation of NR3C1 may be a mechanism of risk for psychopathology or maladaptive behavioral outcomes among children and adults (for a review, see Palma-Gudiel, Cordova-Palomera, Leza, & Fananas, Reference Palma-Gudiel, Cordova-Palomera, Leza and Fananas2015). For example, elevations of NR3C1 methylation in DNA from whole blood have been observed in adults with borderline personality disorder (Dammann et al., Reference Dammann, Teschler, Haag, Altmuller, Tuczek and Dammann2011), and adults with both bulimia and borderline personality disorder demonstrate greater NR3C1 methylation at exon 1C and lower methylation at exon 1H (Steiger, Labonte, Groleau, Turecki, & Israel, Reference Steiger, Labonte, Groleau, Turecki and Israel2013). Among adults with borderline personality disorder, methylation of NR3C1 in blood is positively associated with clinical severity (Martin-Blanco et al., Reference Martin-Blanco, Ferrer, Soler, Salazar, Vega and Andion2014). In contrast, lower methylation of NR3C1 in blood has also been observed in adults with MDD, PTSD, and externalizing disorders (Heinrich et al., Reference Heinrich, Buchmann, Zohsel, Dukal, Frank and Treutlein2015; Labonte, Azoulay, Yerko, Turecki, & Brunet, Reference Labonte, Azoulay, Yerko, Turecki and Brunet2014; Na et al., Reference Na, Chang, Won, Han, Choi and Tae2014; Yehuda, Golier, et al., Reference Yehuda, Flory, Bierer, Henn-Haase, Lehrner and Desarnaud2015).
These links to behavioral outcomes have also been seen in childhood and adolescence. In newborns, NR3C1 methylation is associated with decreased quality of movement and self-regulation, increased arousal and excitability, and increased nonoptimal reflexes and stress abstinence scores (Paquette et al., Reference Paquette, Lester, Lesseur, Armstrong, Guerin and Appleton2015). In our own work with preschoolers, we recently demonstrated that methylation of NR3C1 in saliva DNA is associated with the development of internalizing behavior problems, and methylation of NR3C1 mediates links between early adversity and internalizing behavior problems (Parade et al., Reference Parade, Ridout, Seifer, Armstrong, Marsit and McWilliams2016). Likewise, NR3C1 methylation is positively associated with internalizing problems in adolescence (van der Knaap, van Oort, Verhulst, Oldehunkel, & Riese, Reference van der Knaap, van Oort, Verhulst, Oldehinkel and Riese2015) and both internalizing problems and morning cortisol levels in childhood and adolescence (Dadds, Moul, Hawes, Mendoza Diaz, & Marsit, Reference Dadds, Moul, Hawes, Mendoza Diaz and Brennan2015). This work is complemented by animal models demonstrating that methylation of NR3C1 is linked with increased anxiety-like behaviors in rodents (Kosten, Huang, & Nielsen, Reference Kosten, Huang and Nielsen2014; Lutz & Turecki, Reference Lutz and Turecki2014; Pan, Fleming, Lawson, Jenkind, & McGowan, Reference Pan, Fleming, Lawson, Jenkins and McGowan2014). Taken together, this emerging body of literature supports the hypothesis that methylation of NR3C1 is linked with the development of behavioral difficulties in childhood. Furthermore, this work suggests that NR3C1 methylation is potential mechanism underlying the development of psychopathology among children and adults exposed to early adversity.
Childhood adversity and methylation of FKBP5
In addition to NR3C1, an important regulator of the GR is FKBP5, which mediates an additional negative feedback loop on glucocorticoids. GR activation results in rapid induction of FKBP5, which binds to the GR and decreases its ability to bind cortisol and to translocate to the nucleus (Figure 1). Thus, FKBP5 decreases systemic sensitivity to cortisol and may also impair GR-mediated negative feedback modulation of the HPA axis (Binder, Reference Binder2009; Cioffi, Hubler, & Scammell, Reference Cioffi, Hubler and Scammell2011; Schmidt et al., Reference Schmidt, Buell, Ionescu, Gassen, Holsboer and Cox2015; Tatro, Everall, Kaul, & Achim, Reference Tatro, Everall, Kaul and Achim2009), and methylation of the FKBP5 gene, with associated reductions in transcription, might limit these effects. Genetic variation in FKBP5 confers altered GR function and a poorly regulated neuroendocrine response to stress (Zannas & Binder, Reference Zannas and Binder2014). A single nucleotide polymorphism (SNP) in FKBP5 (C to T SNP in intron 2, rs1360780) enhances the ability of the GR to bind to the glucocorticoid response elements and induce FKBP5 expression (Zannas & Binder, Reference Zannas and Binder2014). This “risk” T allele is associated with GR resistance (Hohne et al., Reference Hohne, Poidinger, Merz, Pfister, Bruckl and Zimmermann2015; Ising et al., Reference Ising, Depping, Siebertz, Lucae, Unschuld and Kloiber2008; Menke et al., Reference Menke, Klengel, Rubel, Bruckl, Pfister and Lucae2013) and has been linked with PTSD, depressive and anxiety symptoms and disorders, and suicide (Leszczynska-Rodziewicz et al., Reference Leszczynska-Rodziewicz, Szczepankiewicz, Narozna, Skibinska, Pawlak and Dmitrzak-Weglarz2014; Suzuki et al., Reference Suzuki, Matsumoto, Sadahiro, Enokido, Goto and Otani2014; Szczepankiewicz et al., Reference Szczepankiewicz, Leszczynska-Rodziewicz, Pawlak, Narozna, Rajewska-Rager and Wilkosc2014; Van Zomeren-Dohm, Pitula, Koss, Thomas, & Gunnar, Reference Van Zomeren-Dohm, Pitula, Koss, Thomas and Gunnar2015; Zannas & Binder, Reference Zannas and Binder2014).
Recent work by Klengel et al. (Reference Klengel, Mehta, Anacker, Rex-Haffner, Pruessner and Pariante2013) examined the rs1360780 SNP and methylation of FKBP5 in relation to childhood maltreatment. Compared to adults with no childhood abuse, those with a history of childhood maltreatment had lower levels of methylation in the regulatory regions of intron 7 of FKBP5 in those with the rs1360780 risk allele, and this was associated with decreased GR sensitivity. Moreover, treatment of human hippocampal progenitor cells with glucocorticoids induced long-lasting demethylation of this regulatory region and increased FKBP5 gene expression, suggesting that prolonged cortisol exposure may be a mechanism by which this region is demethylated (Klengel et al., Reference Klengel, Mehta, Anacker, Rex-Haffner, Pruessner and Pariante2013). In our study of preschoolers with early adversity, we also found that child maltreatment was associated with lower levels of FKBP5 intron 7 methylation in saliva DNA; however, rs1360780 was not a significant moderator of this association (Tyrka et al., Reference Tyrka, Parade, Eslinger, Marsit, Lesseur and Armstrong2015). These findings of FKBP5 demethyation in association with childhood maltreatment are in contrast to recent work suggesting that low childhood socioeconomic status is associated with increased FKBP5 methylation (Needham et al., Reference Needham, Smith, Zhao, Wang, Mukherjee and Kardia2015). It is possible that low socioeconomic status in the absence of other adversities activates the HPA system to initially increase methylation but does not contribute to demethylation over time. A recent study found methylation of FKBP5 was correlated in Holocaust survivors and their offspring; survivors also showed greater methylation of FKBP5 in comparison with participants who were not Holocaust survivors, whereas the children of these survivors exhibited low levels of methylation (Yehuda, Daskalakis, et al., Reference Yehuda, Flory, Bierer, Henn-Haase, Lehrner and Desarnaud2015). This study suggested the possibility of intergenerational effects of trauma exposure. Weder et al. (Reference Weder, Zhang, Jensen, Yang, Simen and Jackowski2014) found links between child maltreatment and methylation of a different region of FKBP5 in saliva DNA among children 5–14 years of age using a 450K methylation array. Although more work is needed to understand the conditions under which adversity is linked with hyper- or hypomethylation of FKBP5 and which regions and variants of this gene are critically involved, this emerging literature suggests that stress-induced changes in methylation of FKBP5 may play a key role in long-term alterations of glucocorticoid activity.
Methylation of FKBP5 at intron 7 has also been linked to behavioral outcomes. Among neonates, placental methylation of FKBP5 at intron 7 was associated with higher levels of arousal during a physical examination (Paquette et al., Reference Paquette, Lester, Koestler, Lesseur, Armstrong and Marsit2014). Patients with bipolar disorder, particularly those with a late stage of illness, had increased post-dexamethasone cortisol levels compared to those without bipolar disorder, as well as higher levels of the FKBP5 protein and higher levels of FKBP5 methylation (Fries et al., Reference Fries, Vasconcelos-Moreno, Gubert, dos Santos, Sartori and Eisele2015). These studies suggest the possibility that methylation of FKBP5 could be a mechanism underlying links between adversity and behavioral health outcomes. However, more work is needed to replicate these findings and clarify patterns of methylation and demethylation in relation to stress exposures, glucocorticoid activation, and risk for various forms of psychopathology.
Summary of childhood adversity and methylation of glucocorticoid signaling genes
Taken together, this emerging literature indicates that childhood adversity is associated with altered methylation patterns in regulatory regions of both NR3C1 and FKBP5. Activity of these genes plays a vital role in the regulation and function of the HPA axis, and could account for some of the findings of adversity-related abnormalities in basal and provoked cortisol levels and risk for psychopathology. Regulation of these systems is complex and adaptive, and more work is needed to understand the conditions under which methylation is promoted or inhibited. It is also critical to recognize that epigenetic modulation of other genes involved in glucocorticoid signaling and other pathways are likely involved in these processes (e.g., Weder et al., Reference Weder, Zhang, Jensen, Yang, Simen and Jackowski2014).
Relevance of Epigenetics for Inventions to Address Childhood Adversity
Recent calls have been made to integrate biological measures of program efficacy into evaluations of preventative interventions for children at risk (Bruce et al., Reference Bruce, Fisher, Graham, Moore, Peake and Mannering2013; Cicchetti & Gunnar, Reference Cicchetti and Gunnar2008; Moffitt & Klaus-Grawe 2012 Think Tank, Reference Moffitt2013). Several studies have shown normalizing effects of psychosocial interventions on diurnal or provoked cortisol levels in children exposed to significant adversity (e.g., Cicchetti, Rogosch, Toth, & Sturge-Apple, Reference Cicchetti, Rogosch, Toth and Sturge-Apple2011; Dozier, Peloso, Lewis, Laurenceau, & Levine, Reference Dozier, Peloso, Lewis, Laurenceau and Levine2008; Fisher, Stoolmiller, Gunnar, & Burraston, Reference Fisher, Stoolmiller, Gunnar and Burraston2007; Fisher, Van Ryzin, & Gunnar, Reference Fisher, Van Ryzin and Gunnar2011; Laurent, Gilliam, Bruce, & Fisher, Reference Laurent, Gilliam, Bruce and Fisher2014; Nelson & Spieker, Reference Nelson and Spieker2013; Slopen, McLaughlin, & Shonkoff, Reference Slopen, McLaughlin and Shonkoff2014; van Andel et al., Reference van Andel, Jansen, Grietens, Knorth and van der Gaag2014). Altered methylation of glucocorticoid signaling genes including NR3C1 and FKBP5 may underlie adversity-induced HPA axis dysfunction, and interventions might stably alter these methylation patterns. Given the complexity in the genetic regulation of HPA axis function, different methylation patterns may explain inconsistencies in the literature on cortisol, early stress, and psychiatric illness. Consideration of alterations in gene methylation would complement examination of behavioral and endocrine outcomes in intervention trials. Furthermore, understanding whether effective behavioral interventions reverse effects of adversity on gene methylation will inform the basic science of biological pathways contributing to risk and resilience.
Although methylation is thought to be the most stable form of epigenetic modification, there is evidence that gene methylation may be plastic during childhood and into adulthood. Therapeutic interventions may alter methylation patterns and could reduce the biologic risk engendered by early life adversity (Szyf, Reference Szyf2015). There is preliminary evidence that methylation of genes involved in the stress response system might serve as biomarkers of treatment response for psychiatric disorders. In children, FKBP5 DNA methylation was significantly associated with treatment response to a cognitive behavioral therapy-based intervention for anxiety disorders (Roberts et al., Reference Roberts, Keers, Lester, Coleman, Breen and Arendt2015). Specifically, children with a smaller reduction in symptoms showed an increase in DNA methylation, while children with a larger symptom reduction showed a decrease in DNA methylation. When corrected for multiple testing, the association between DNA methylation and treatment response remained significant in participants with the “risk” T FKPB5 genotype. Likewise, in a pilot study, pretreatment methylation of the NR3C1 exon 1F promoter predicted treatment response to prolonged exposure therapy for PTSD in combat veterans; in addition, FKBP5 promoter methylation decreased and FKBP5 expression increased in association with recovery (Yehuda et al., Reference Yehuda, Daskalakis, Desarnaud, Makotkine, Lehrner and Koch2013). Future studies are warranted to understand how methylation changes in response to psychosocial interventions, and if methylation is a potential marker for the efficacy of intervention.
Several drugs can alter methylation patterns across the life span. Trichostatin A is a compound that inhibits enzymes involved in epigenetic processes that decrease gene expression. In rats, treatment with trichostatin A reversed the decreased gene expression that was imparted by neglect in early life; this reversal was associated with decreased behavioral phenotypes of depression and anxiety (Weaver, Meaney, & Szyf, Reference Weaver, Meaney and Szyf2006). Conversely, treatment with methionine, a compound that is known to increase gene methylation, suppressed gene expression in rats with high levels of early life maternal care and was associated with increased depressive and anxious behaviors (Weaver et al., Reference Weaver, Meaney and Szyf2006). A number of medications commonly used in psychiatry, including clozapine, sulpiride, and valproic acid, have been shown to actively promote demethylation in cell culture (Detich, Bovenzi, & Szyf, Reference Detich, Bovenzi and Szyf2003; Milutinovic, D'Alessio, Detich, & Szyf, Reference Milutinovic, D'Alessio, Detich and Szyf2007) and in the mouse brain (Dong, Chen, Gayin, Grayson, & Guidotti, Reference Dong, Chen, Gavin, Grayson and Guidotti2010; Dong, Nelson, Grayson, Costa, & Guidotti, Reference Dong, Nelson, Grayson, Costa and Guidotti2008). Antidepressants have been associated with changes in DNA methylation (Menke & Binder, Reference Menke and Binder2014). The DNA methylation inhibitors zebularine and 5-aza-2-deoxycytidine reverse DNA methylation and block synaptic long-term potentiation in mouse hippocampal slices (Levenson et al., Reference Levenson, Roth, Lubin, Miller, Huang and Desai2006) as well as fear memory formation (Miller & Sweatt, Reference Miller and Sweatt2007). Another potential therapeutic approach is to modify the intracellular environment such that it promotes demethylation. The methyl donor, S-adenosylmethionine inhibits demethylase activity (Detich, Hamm, Just, Knox, & Szyf, Reference Detich, Hamm, Just, Knox and Szyf2003). Levels of S-adenosylmethionine are controlled in part by dietary intake of folic acid and vitamin B12 (Bottiglieri, Reference Bottiglieri2013). Both the presence of Ca2+ and the redox state of the cell can influence demethylase activity (Szyf, Reference Szyf2015). Finally, exercise modulates changes in DNA methylation associated with stress exposure (Kashimoto et al., Reference Kashimoto, Toffoli, Manfredo, Volpini, Martins-Pinge and Pelosi2016; Rodrigues et al., Reference Rodrigues, Toffoli, Manfredo, Francis-Oliveira, Silva and Raquel2015), suggesting that physical activity interventions may have a role in modifying the effects of early adversity on gene methylation. Together, these findings suggest that interventions or treatments that interfere with or reverse DNA methylation could potentially modify the risk for psychopathology imparted by early life adversity.
Directions for Future Research
Despite accumulating knowledge of the role of epigenetics and glucocorticoid signaling in the biological response to environmental risk and protection, many questions remain regarding the biological processes underlying these links. A basic understanding of how methylation of glucocorticoid signaling genes longitudinally changes over time has yet to be achieved, with little understanding of the role of risk and protective factors that impact these developmental trajectories. Understanding how methylation patterns change across developmental stages, and whether there are sensitive periods of development for changes in methylation as a consequence of environmental exposures, is a critical next step. Knowledge of how quickly methylation occurs in response to adversity and which genes and gene networks have the largest effects is lacking. Elucidation of these influences and determinants would serve to identify the most critical developmental periods for intervention, as well as the length of time and intensity of services needed to reverse effects of adversity on methylation. Thus, many questions remain to be answered before knowledge of the basic science of epigenetic processes can be translated into the most effective clinical interventions for children exposed to adversity.
In addition to the longitudinal examination of change in methylation over time, examining links between methylation and environmental exposures from a transactional perspective is an important next step. It is well established that development is best characterized by transactional relationships between the child and his or her environment, such that the environment exerts influence on the child and the child in turn exerts influence on the environment over time (Cicchetti & Lynch, Reference Cicchetti and Lynch1993; Combs-Ronto, Olson, Lukenheimer, & Sameroff, Reference Combs-Ronto, Olson, Lunkenheimer and Sameroff2009). Although historically focused on child behavior such as temperament, emerging work considering transactive effects suggests that child biological stress responding has the potential to exert influence on the environment, in this case parental behavior, through transactional relationships as well (e.g., Perry, Mackler, Calkins, & Keane, Reference Perry, Mackler, Calkins and Keane2014). Examining the longitudinal course of methylation through a transactional lens would allow us to understand how the environment influences methylation and methylation in turn potentially exerts influence on the environment. As an example, compromised parental behavior may contribute to methylation of glucocorticoid signaling genes in the child, resulting in child behavior problems that evoke parental stress and further undermine sensitive parenting. Parenting stress related to child behavior problems might contribute to change in gene methylation patterns among parents as well.
Future work should also aim to understand the role of epigenetics in resilience processes. It is possible that alterations in methylation associated with adversity have the potential to be adaptive. For example, in male Rwandan genocide survivors, increased NR3C1 promoter DNA methylation was associated with reduced risk for PTSD, and less intrusive memories of trauma (Vukojevic et al., Reference Vukojevic, Kolassa, Fastenrath, Gschwind, Spalek and Milnik2014). This work in conjunction with emerging literature suggesting that MDD, PTSD, and externalizing disorders can be linked with reduced NR3C1 methylation (Heinrich et al., Reference Heinrich, Buchmann, Zohsel, Dukal, Frank and Treutlein2015; Labonte et al., Reference Labonte, Azoulay, Yerko, Turecki and Brunet2014; Na et al., Reference Na, Chang, Won, Han, Choi and Tae2014; Yehuda, Golier, et al., Reference Yehuda, Daskalakis, Bierer, Bader, Klengel and Holsboer2015) suggests that increases in NR3C1 methylation are not always markers of risk. Future work should also begin to examine potential moderators of the links between early adversity and methylation, and methylation and behavioral outcomes. For example, parental sensitivity and responsiveness may buffer children from effects of early adversity on methylation, and may promote adaptive behavioral outcomes even when adversity alters methylation (Conradt et al., Reference Conradt, Hawes, Guerin, Armstrong, Marsit and Tronick2016).
Finally, the consideration of epigenetic processes in response to environmental exposures has typically focused on either environment in the prenatal period or environment following birth, with few studies concurrently examining both prenatal and postnatal factors to contribute to methylation of glucocorticoid signaling genes. Consideration of environmental risk from the preconception period through the postnatal period and childhood is critical for understanding potential sensitive periods during which adversity is biologically encoded, as well as to uncover the most opportune times to intervene with children and families at risk.
Conclusion
Accumulating knowledge supports the perspective that childhood adversity undermines children's health across the life span. As reviewed above, methylation of glucocorticoid signaling genes including NR3C1 and FKBP5 are increasingly recognized as potential mechanisms by which childhood adversity is biologically encoded. Although exponential progress has been made over the past decade to understand these epigenetic processes, critical gaps in knowledge remain in our understanding of methylation across typical development and in the face of early adversity. Advancing understanding of these processes will inform prevention and intervention efforts aimed at enhancing the lives of the most vulnerable children and families.
