A dominant model of occupational stress posits the combination of high job demands and limited control over job-related decision making (low decision latitude) as pernicious attributes of the work environment, and job strain defined in this manner has been associated with both elevated blood pressure and heart disease (e.g., Kivimaki et al, 2006; Landsbergis et al, 2003; Ohlin et al, 2007). Because the sympathetic nervous system is a key determinant of cardiac and vascular function, obvious candidate genes for studies of gene-environment interactions affecting blood pressure are those of the several adren-ergic receptors. In a large, middle-aged sample of employed individuals, for instance, homozy-gosity for the deletion allele of a common insertion/deletion polymorphism in the a2B-adrenoreceptor (ADRA2B) was found associated with higher systolic and diastolic blood pressure in men, relative to those of other ADRA2B genotypes, but only when accompanied by job strain (Ohlin et al, 2007). In addition, the decision latitude component of job strain alone correlated inversely with systolic blood pressure in men. In another study of the same sample, job strain interacted only marginally with genotypes of an argenine/glycine substitution in the P 1-adrenoreceptor gene (ADRB1), although systolic blood pressure covaried inversely with extent of job demands among men carrying the glycine-encoding allele of this polymorphism (Ohlin et al, 2008). Another model of occupational stress construes the work environment as a balance of implied obligations whereby the employee's work effort is compensated by salary, recognition, security, and opportunities for advancement (Siegrist et al, 2004). Job stress is defined in this context as a relative imbalance between effort and reward. In the one study following this model, presence of hypertension was predicted by an interaction of job stress and another argenine/glycine substitution, in the p2-adrenoreceptor gene (ADRB2:
P2 -AR-16) (Yu et al, 2008). Hypertension was about 3.5 times more prevalent among individuals with any glycine-encoding allele than in those homozygous for the alternate arge-nine allele, but only in conjunction with high job stress (effort/reward imbalance) and, again, only in men. All three of these studies suggest that attributes of the work environment moderate effects of adrenoreceptor gene variation on blood pressure and do so in men alone, although variability in their operational definitions of job stress (job strain, effort-reward imbalance) and in the components of genotype-dependent job strain associated with blood pressure (latitude, job demands) leaves these gene-environment interactions preliminary.
Activation or dysregulation of the HPA system is the most frequently invoked mechanism to explain effects of psychological stress on disease, both physical and psychiatric (see Chapter 43). This is undoubtedly due to actions of the glucocorticoid hormone, cortisol, on diverse physiological functions and tissues of the body, including metabolism and activities of the cardiovascular, immune, and central nervous systems. Some HPA phenotypes are also heritable. For instance, genetic variance appears to account for half or more of interindividual variability in basal cortisol levels aggregated across multiple measurements in the morning (Bartels et al, 2003; Meikle et al, 1988) and similar, if somewhat lesser, heritabilities are seen in the cortisol response to morning awakening (Kupper et al, 2005; Wust et al, 2000). With respect to molecular variation, a series of investigations has focused on the corticotropin-releasing hormone type 1 receptor (CRHR1) gene. In the first of these, individual single nucleotide polymorphisms (SNPs) and a TAT haplotype at markers rs7209436, rs110402, and rs242924 were found associated with lower levels of depressive symptomatology, when compared to other CRHR1 genotypes, among African American women who were abused in childhood (Bradley et al, 2008). The apparently protective effects of this CRHR1 variation against early adversity/abuse were replicated in relation to lifetime history of major depressive disorder in an independent sample of predominantly Caucasian women (Bradley et al, 2008) and in relation to both past year and recurrent depression in a midlife sample of British women (Polanczyk et al, 2009). A similar interaction was not observed among participants of the Dunedin (New Zealand) Multidisciplinary Health and Development Study (Polanczyk et al, 2009). However, one other investigation has shown the "protective" alleles at rs110402 and rs242924 to mitigate the heightened cortisol responses to a dexamethasone/corticotropin-releasing hormone (dex/CRH) challenge otherwise seen among adults maltreated as children (Tyrka et al, 2009).
Another series of studies has examined polymorphic variation in FKBP5, which encodes a co-chaperone of heat stress protein-90, FK506 binding protein-5, that helps regulate binding affinity of the glucocorticoid receptor (GR). In a depressed patient sample, TT genotype at rs1360780 was found to be associated with a more rapid response to antidepressant treatment and greater recurrence of depressive episodes, higher levels of FKBP5 expression in lymphocytes, and a blunted adrenocorticotropin hormone (ACTH) response to dex/CRH challenge, relative to those carrying one or more copies of the alternate C allele (Binder et al, 2004). This polymorphism and another FKBP5 SNP, rs3800373, also predicted peritraumatic dissociation, a risk factor for PTSD, among medically injured children (Koenen et al, 2005), and these and two other SNPs analogously predicted adult PTSD symptomatology in interaction with childhood abuse (but not adult trauma exposure) among predominantly low-income, African American men and women (Binder et al, 2008). Interestingly, in the latter investigation, CRHR1 SNPs previously reported to interact with childhood abuse in the prediction of depressive symptoms in the same study sample (Bradley et al, 2008) did not similarly predict PTSD symptoms here, suggesting some specificity of association.
Other recent studies have investigated PTSD risk in relation to varying exposure to hurricanes occurring in Florida in 2004. In one, the C allele of a SNP in the "regulator of G-protein signaling 2" (RGS2) gene, rs4606, predicted post-hurricane PTSD symptomatology, but only among individuals with high hurricane exposure (viz., a combination of hurricane force winds or flooding, material losses, and extended displacement) and low levels of social support (Amstadter et al, 2009). And in the same study, lifetime PTSD symptoms were likewise predicted by the interaction of RGS2 variation with lifetime exposure to traumatic (life-threatening) events. Two prior investigations of the same study cohort genotyped common regulatory variation in the serotonin transporter gene-linked polymorphic region (5-HTTLPR), one variant of which (typically termed the short [S] allele) reduces transcriptional efficiency of the serotonin transporter gene compared to the alternate, long (L) allele. In one study, the low expression allele increased risk of post-hurricane PTSD in interaction with area-level socioeconomic indicators (crime and unemployment rates) (Koenen et al, 2009), and in the second, the same genetic variation predicted both PTSD and major depression in persons with high hurricane exposure and low social support, relative to all other combinations of genotype, exposure and support (Kilpatrick et al, 2007). To the extent that a high exposure to hurricanes, as defined by these investigators, is an impactful life event, it is perhaps surprising that the RGS2 and 5-HTTLPR polymorphisms did not interact with exposure alone, but only in three-way interaction with social support or community characteristics. This may indicate that the presence of social support or socioeconomic advantage ameliorates genotype-dependent risk of PTSD and depression associated with hurricane exposure.
5.1.1 Gene-Stress Interaction and the Challenge of Replication
At present, studies of gene-stress interactions have produced only small literatures, often just one or two investigations addressed to a particular gene, environmental moderator, and phe-notype or to phenotypes studied in multiple studies, but with respect to different genes (e.g.,
PTSD). Two exceptions are literatures generated by two early gene-environment studies (Caspi et al, 2002, 2003), in which a sufficient number of investigations testing the same hypothesis are available to evaluate the replica-bility of gene-stress interactions. In the first of these, the initial study found self-rated aggressiveness, conduct disorder, symptoms of adult antisocial personality disorder, and commission of a violent crime potentiated by childhood maltreatment among males with low-transcription variants of a functional promoter polymorphism of monoamine oxidase-A (MAOA), relative to men having an alternate, high-activity MAOA allele (Caspi et al, 2002). Two meta-analyses, the latter of which included the initial study and seven attempted replications, have now been reported (Kim-Cohen et al, 2006; Taylor and Kim-Cohen, 2007). In the second meta-analysis, the authors derived the correlation of adversity with antisocial outcomes for each MAOA genotype and expressed the MAOA x Adversity interaction as the difference in correlation between genotypes. The pooled estimate of effect size (correlation) across studies was 0.30 for the low-activity MAOA genotype and 0.13 for the high-activity genotype. The test of the interaction yielded a modest, but significant, effect size (for the difference in correlation) of 0.17 (95% CI: 0.09, 0.24; p < 0.001). This outcome was only minimally affected by the removal of either the original investigation or the two studies having effect sizes larger than the sentinel report, or by serial deletion of each study individually. This suggests the presence of a robust gene-stress interaction. More recent studies have tended also to confirm the interaction of MAOA variation and childhood maltreatment on later externalizing and antisocial behaviors (Ducci et al, 2008; Enoch et al, 2010), although this relationship may be less stable in women (Prom-Wormley et al, 2009) and less apparent at very severe levels of early adversity (Weder et al, 2009).
The second sizable literature on gene-stress interactions derives from a widely publicized study in which childhood maltreatment and recent stressful life events predicted depression in proportion to the number of S, relative to L, alleles of 5-HTTLPR that individuals possessed (Caspi et al, 2003). 0ver 30 investigations of widely varying design, methodology, and study population have been reported since, and this literature has been the subject of numerous commentaries, narrative reviews, and meta-analyses. With respect to the latter, two meta-analyses reported in 2009 covered largely (but not entirely) overlapping studies and both failed to confirm an interactive effect of 5-HTTLPR variation and life events on depression (Munafo et al, 2009; Risch et al, 2009). These results cast a long shadow over this most frequently cited instance of gene-stress interaction and may seem especially disconcerting given strong rationale for the hypothesis. For instance: (1) serotonergic dysregulation has long been implicated in depression and mood regulation (Thase, 2009);
(2) the serotonin transporter is a primary target of pharmacotherapy for depression (Gitlin, 2009);
(3) the S allele of orthologous 5-HTTLPR variation in the rhesus macaque lowers brain serotonin turnover in monkeys, but only among animals reared without maternal contact (Bennett et al, 2002); and (4) in vivo brain serotonergic responsivity (assessed by neuropharmacologic challenge) is attenuated in euthymic individuals with a history of depression (Bhagwagar et al, 2002; Flory et al, 1998) and among carriers of the 5-HTTLPR S allele (Reist et al, 2001; Whale et al, 2000). 0ne environmental parameter, socioeconomic status, has also been shown to modulate the influence of 5-HTTLPR variation on central serotonergic responsivity (Manuck et al, 2004).
Stressful life events likewise warrant consideration as an environmental "pathogen" due to abundant evidence that they predict depression onset, particularly first episodes (Kendler et al, 2000; Uher and McGuffin, 2008) and, in twin research, have been found to do so in interaction with genetic liability (Kendler et al, 1995). A 30-year literature indicates that life events most germane to depression are those that occur abruptly and within 3-6 months preceding onset, that are impactful and directly affect the respondent, and that commonly involve situations of significant threat, loss, or humiliation (Brown and Harris, 1978; Brown, et al, 1995; Kendler et al, 2003; Monroe et al, 2009). Yet almost none of the studies of 5-HTTLPR-stress interactions for depression appear to have been informed by these observations. In reviewing the methodologies of 13 5-HTTLPR-life stress studies, for instance, Monroe and Reid (2008) found that only one study could appropriately distinguish acute stressors from chronic conditions, only three could appropriately distinguish major from minor events, only five restricted event assessments to the last 6 months (others extending measurements even to 5 years or lifetime exposure), only three assessed participant-focused events, and only one study satisfied all four of these conditions. They also report that instruments for assessing life events were diverse in format and content across studies, that nearly all lacked provenance in the life stress/depression literature, that procedures for determining total exposure to life events were different in every study, and that in only four studies could it be assumed that life events even preceded the onset of depression. Regarding this last consideration, a significant main effect of life events on depression, as reported in the two cited meta-analyses, could reflect, in part, reverse causation (events consequent on depression) or, where events and outcomes were assessed at the same time, biased event recall (Monroe and Reid, 2008).
Also addressing the heterogeneity of study methods, Uher and McGuffin (2010) recently stratified, by format of life event assessment, 34 studies that purposed to test interactive effects of 5-HTTLPR variation and environmental adversity on depression. By their analysis, the hypothesized interaction was demonstrated in 11 of 15 studies employing event measurements that were either "objective" (i.e., ascertained independently of participants' reports) or derived from contextually sensitive life stress interviews, while the four remaining studies reported at least partial replications (e.g., findings delimited by gender or event kind). On the other hand, all non-replications emerged in investigations using life event checklists or other self-report instruments, yielding a distribution of six replications, four partial replications, and ten non-replications. The authors note as well that the two previously cited meta-analyses (Munafo et al, 2009; Risch et al, 2009) included only a minority of all studies and, by sampling disproportionately from those relying on participants' self-reported life stress, may have confounded study outcomes with variability and quality of investigational methods (Uher and McGuffin, 2010).
As noted previously, the sentinel study in this literature (Caspi et al, 2003) reported that both childhood maltreatment and life events predicted depression in interaction with 5-HTTLPR variation. The interval of event reporting in this study was 5 years, suggesting that events per se were not a proximal cause of depression. In another recent commentary, Brown and Harris (2008) note that childhood maltreatment is itself a potent risk factor for depression and that in the study of Caspi and colleagues (2003) the genotype-dependent relationship between childhood adversity and depression appears to have been stronger than that for life events. Presenting evidence that early maltreatment also predicts the frequency of events experienced over a 45-year period in adulthood, Brown and Harris argue that life events occurring outside the canonical range for life event/depression associations may be a marker of early childhood adversities that, in interaction with allelic variation at 5-HTTLPR, affect neurodevelopmen-tal processes. In turn, these neurodevelopmental changes may promote behaviors that increase the likelihood of experiencing stressful life events over the lifecourse and also heighten risk of depression in adulthood. In contrast, recent life events may predict depressive episodes independently or interact only minimally with 5-HTTLPR genotype. Such speculation is also consistent with a paucity of evidence showing allelic differences at 5-HTTLPR to affect serotonin transporter availability, binding, and mRNA expression in the adult brain (e.g., Mann et al, 2000; Parsey et al, 2006; Shioe et al, 2003; Sibille and Lewis, 2006; van Dyck et al, 2004).
Was this article helpful?
Everything you need to know to control your stress is included in this special report. It may be a symptom of the strains of modern life: the prevalence of stress andanxiety. It seems that stress is ever-present, attacking us at work, at school, on the road, and at home. To a large extent, it is impossible to thrive on stress.