http://orcid.org/0000-0002-8671-1238
Figures
Abstract
Background
Gene-environment interaction (GxE) research has highlighted the importance of investigating the FK506 binding protein 51 (FKBP5) gene as a sensitivity gene. However, previous GxE studies with FKBP5 have not measured the full environmental spectrum or applied statistical tests to discern whether the GxE interaction fits better with the differential-susceptibility or diathesis-stress hypotheses. This study examined whether single nucleotide polymorphisms (SNPs) on FKBP5 gene moderate the association of positive and negative recent life events (LEs) with depressive symptoms, state-anxiety, neuroticism, and social anxiety traits.
Methods
A total of 86 nonclinical young adults were administered psychological measures and were genotyped for five FKBP5 SNPs (rs3800373, rs9296158, rs1360780, rs9470080 and rs4713916).
Results
Regression analyses indicated significant GxE interactions for social anxiety and neuroticism. The interactions predicting neuroticism fit different models for different SNPs, although the overall effect indicated by the haplotype was consistent with the differential-susceptibility hypothesis: the risk-haplotype group presented higher neuroticism in the presence of more negative LEs and lower neuroticism in the presence of more positive LEs. The GxE interactions for social anxiety were consistent with the diathesis-stress model. The lack of significance in the for-better side for social anxiety might be related to the fact that it mapped onto low extraversion, which is associated with a lower permeability to positive experiences.
Citation: Pérez-Pérez B, Cristóbal-Narváez P, Sheinbaum T, Kwapil TR, Ballespí S, Peña E, et al. (2018) Interaction between FKBP5 variability and recent life events in the anxiety spectrum: Evidence for the differential susceptibility model. PLoS ONE 13(2): e0193044. https://doi.org/10.1371/journal.pone.0193044
Editor: Igor Branchi, Istituto Superiore Di Sanita, ITALY
Received: July 28, 2017; Accepted: February 2, 2018; Published: February 21, 2018
Copyright: © 2018 Pérez-Pérez 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.
Data Availability: The authors of the present study confirm that some access restrictions apply to the data underlying the findings. This is given to the fact that the consent form that participants signed before taking part in the study imposes restrictions for making the data publicly available. This study was approved by the Institutional Review Board of the Universitat Autònoma de Barcelona (Comissió d'Ètica en l'Experimentació Animal i Humana; CEEAH); protocol reference 701H-JS; http://www.recerca.uab.es/ceeah/). Participants agreed for all the data collected to be available to the members of the research group Person-Environment Interaction in Psychopathology led by Prof. Neus Barrantes-Vidal (Address: Departament de Psicologia Clínica i de la Salut, Facultat de Psicologia, Edifici B, Universitat Autònoma de Barcelona, 08193 Cerdanyola del Vallès, Spain; telephone: +34 93 5813864; email: neus.barrantes@uab.cat). Participants were informed that data were available on request (requests should be addressed to the contact details just provided).
Funding: This work was supported by the Spanish Ministerio de Economía y Competitividad (Plan Nacional de I+D PSI2011-30321-C02-01 and C02-02; PSI2014-54009-R; Red de Excelencia PSI2014-56303-REDT, PROMOSAM: Investigación en Procesos, Mecanismos y Tratamientos Psicológicos para la Promoción de la Salud Mental), Fundació La Marató de TV3 (091110) and Comissionat per a Universitats i Recerca of the Generalitat de Catalunya (2014SGR1070 and 2014SGR1636). Neus Barrantes-Vidal is supported by the Institució Catalana de Recerca i Estudis Avançats (ICREA) Academia Award.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Gene-Environment interaction (GxE) research has been primarily guided by the diathesis-stress model [1–3], which establishes that individuals carrying genetic-risk variants are more vulnerable to the effect of environmental adversity and thus more prone to develop psychopathology. Therefore, GxE has predominantly focused on the assessment of the negative side of the environment (adversity) and has almost neglected the positive side (supportive experiences). More recently, Belsky & Pluess [4] have underscored the biasing emphasis of GxE on environmental adversities and proposed an alternative hypothesis, a differential-susceptibility to environmental influences. It poses that, due to evolutionary reasons, individuals with different genetic background should differ in the degree to which they are affected by the whole environmental spectrum (from positive to negative) and not only by the degree in which they are affected by adverse environments, as predicted by the diathesis-stress hypothesis [5]. Therefore, more plastic individuals are expected to be more susceptible to both the negative effects of adverse environments and the beneficial effects of positive environments, while less plastic individuals are expected to be less affected by the environment [6].
Although a fast-growing research has recently pointed to the critical role of positive environments [7], the differential-susceptibility hypothesis has been much less employed to investigate GxE on psychopathology. The problem of GxE being exclusively focused on negative environments is that vulnerability and susceptibility can become indistinguishable, since susceptibility, in some cases, only becomes evident when measuring the full environmental spectrum [4]. Failure to distinguish between diathesis-stress and differential-susceptibility models has critical implications, as it could lead to an inadequate characterization of the etiology of mental disorders and resilience processes. Such biased accounts can result in interventions that are based on an incomplete or even erroneous understanding of human development, which in turn can produce inefficacy, iatrogenic effects, and unnecessary economical costs. Indeed, differential-susceptibility research demonstrates that measuring the full environmental spectrum, genetic variants that were robustly consistent with a diathesis-stress pattern and consequently understood as “risk-variants” (e.g., the short allele of 5HTTLPR), may be better conceptualized as genetic susceptibility factors as they match with differential-susceptibility predictions [8].
The gene encoding FK506 binding protein 51 (FKBP5), located on chromosome 6p21.31, is a highly interesting target for GxE as it is considered a shared etiological factor underlying stress-related disorders [9, 10]. The FKBP5 protein, through the inhibition of glucocorticoid receptors activity, promotes the homeostatic regulation of the hypothalamic-pituitary-adrenal (HPA) axis, the principal biological mechanism of the stress response [10, 11]. Several FKBP5 single nucleotide polymorphisms (SNPs) have been robustly associated with individual differences in the stress response of healthy adults (e.g., prolonged recovery period of the HPA axis after exposure to a stressor and increased glucocorticoid receptor resistance; [10]). FKBP5 variability has also been shown to be associated with heightened amygdala reactivity in the context of emotional neglect, increased attentional threat bias and differences in hippocampal shape [10–13], and with anxiety-proneness trait levels [14].
As diathesis-stress research highlighted, the interaction of FKBP5 variability with childhood, but not adult trauma [15], has been found to confer risk for several psychopathological phenotypes [16], including depression [17, 18], psychosis [19, 20], anxiety [21], suicidal attempts [22, 23], aggression [24] and post-traumatic stress disorder [15, 25]. However, Belsky & Pluess [26] suggested that some of these results may in fact represent differential susceptibility effects. For instance, Xie et al. [27] found that individuals homozygous for the risk T allele of the rs9470080 had the highest risk of post-traumatic stress disorder if they were exposed to early trauma, but the lowest if they were not exposed. Likewise, Zimmermann et al. [28] found that the cumulative incidence of first-episode major depression was the highest for individuals possessing risk alleles of two FKBP5 SNPs (rs3000377 and rs47139611) if they had experienced severe trauma, but the lowest if they had not. Similar results were found in the interaction between childhood trauma and risk genotypes of four SNPs (rs3800373, rs9296158, rs1360780, rs9470080) in relation to aggression [24]. Finally, in a sample of post-institutionalized adolescents, Vanzomeren-Dohm et al. [18] found that the rs1360780 did not moderate the relation between early adversity and depressive symptoms. However, consistent with a differential-susceptibility pattern, it moderated the association between current peer victimization and depressive symptoms for girls carrying the risk T allele.
These GxE studies highlight the importance of investigating the FKBP5 as a sensitivity gene. Nonetheless, to the best of our knowledge, none of them measured the interplay between FKBP5 variability and the full environmental spectrum or applied statistical tests to discern whether this interaction better fits the differential-susceptibility or diathesis-stress models. In addition, most GxE research has focused on the interaction between childhood trauma and FKBP5 variants, whereas there is scant knowledge about the interaction of FKBP5 variants with adult recent life events (LEs). Therefore, the aim of this study was to address these gaps by examining whether FKBP5 stress-related polymorphisms moderate the association between the full spectrum of adult LEs (from positive to negative) and anxious-depressive trait and state measures in nonclinical young adults. On the basis of the diathesis-stress model, it would be predicted that individuals carrying the risk alleles would show higher scores on the anxiety-depression measures when they experienced more negative LEs (as compared with participants homozygous for the non-risk alleles). In the differential-susceptibility model, the same individuals carrying the risk alleles would also show lower anxiety-depression levels when exposed to more positive LEs (as compared with participants homozygous for the non-risk alleles).
Methods
Ethics statement
The study was approved by the Ethics Committee of the Universitat Autònoma de Barcelona (Comissió d'Ètica en l'Experimentació Animal i Humana) and conformed to the Helsinki Declaration. The participants in this interview study were over eighteen years of age and had full capacity to consent to participation in research. All participants provided written informed consent and were paid for their participation.
Participants
Data were collected as part of an ongoing longitudinal investigation examining psychosis risk and resilience in youth adults from the Universitat Autònoma de Barcelona (n = 547; 16.82% male; mean age = 20.6). A smaller subsample was selected to conduct exhaustive interview and laboratory follow-up measurements. Participants with high schizotypy scores were oversampled to ensure that the selected subsample was representative of the unselected sample but still contained enough variance on schizotypy measures (see [29]). The sample analyzed in this study comprised 86 participants with valid data targeted at the third wave assessment (mean age = 24.8; SD = 2.68; range = 22–34 years; 39.53% male). We examined whether scores of the unselected participants assessed in the first screening sample who were not retained for the third follow-up (group 0; n = 461) differed from those of the present sample (group 1; n = 86). Results of multivariate regression analyses showed that, after controlling for sex and age, the effect of group was nonsignificant for the two schizotypy scores used as screening criteria (i.e., positive schizotypy, p = 0.290; negative schizotypy, p = 0.457), indicating that the present sample is unbiased in terms of schizotypy scores.
Materials and procedure
Participants completed all psychological measures used in this study through an online procedure and provided buccal mucosa on cotton swabs.
Psychological measures.
All measures showed an internal consistency between good and excellent (see Table 1). Depressive symptoms were evaluated using the 21-item Beck Depression Inventory Second Edition (BDI-II; [30]), which measures the severity of depressive symptoms (last two weeks) on a Likert-scale from 0 to 3. State-anxiety was measured with the 21-item Beck Anxiety Inventory (BAI; [31]), which assesses the intensity of anxiety symptoms over the last week with ratings ranging from 0 (not at all) to 3 (could barely stand it).
Neuroticism was assessed using the neuroticism subscale from the NEO Personality Inventory-Revised (NEO-PI-R; [32]), which captures the five-factor model of personality. It measures 6 facets (anxiety, angry-hostility, depression, self-consciousness, impulsiveness and vulnerability) via 48 items ranging from 0 (strongly disagree) to 4 (strongly agree). Social anxiety traits were measured with the Spanish version of the Social Phobia and Anxiety Inventory-Brief form (SPAI-B; [33]), a 16-item self-report ranging from 1 (never) to 5 (always).
LEs were measured with the Life Experiences Survey (LES; [34]), a self-report assessing the subjective impact of a variety of LEs over the last year. It is composed of two sections: section 1 contains 47 common LEs plus 3 empty spaces for participants to register events not collected in the questionnaire; section 2, specifically designed for university students, contains 10 items related to academic events. Participants rated the perceived impact of experienced events on a 7-point Likert-scale ranging from -3 (extremely negative) to +3 (extremely positive). The LES provides two measures of subjective impact: the Negative Change Score (the sum of the subjective impact ratings with negative values) and the Positive Change Score (the sum of the subjective impact ratings with positive values). In this study we used the LES mean score (LESms), which is the mean of the subjective impact ratings of the experienced events for each individual, including positive and negative ones. Positive scores indicate that LEs had a positive subjective impact.
Genotyping.
Genomic DNA was extracted using the Real Extraction DNA kit (Durviz S.L.U., Valencia, Spain). The rs3800373, rs929615, rs1360780, rs9470080 and rs4713916 SNPs on FKBP5 gene were genotyped using TaqMan 5'-exonuclease allelic discrimination assay (Applied Biosystems) via 5 custom assays. All SNPs were in accordance with Hardy-Weinberg Equilibrium.
Haplotypes were estimated using a Bayesian approach implemented with PHASE software [35]. Linkage disequilibrium, which is the tendency of SNPs to be inherited together, was examined by pair-wise comparisons of r2 and D’ using Haploview version 4.2 [36]. We computed individual’s haplotypes considering three tag SNPs (rs3800373, rs929615, rs1360780) reported in previous studies (see Table 2 for comparison groups).
Statistical method
In order to avoid misinterpreting possible gene-environment association as gene-environment interactions, the effect of FKBP5 variants on LESms was examined by T-Test comparisons. We also used T-Tests to analyze whether criteria (symptoms and traits) and predictor (LESms) variables were associated with sex.
Multiple regression analyses were conducted separately to examine the main-effects of the genetic (five FKBP5 SNPs and haplotype) and environmental (LESms) variables, as well as GxE interactions, on the anxiety-depression measures. Given that discarding the presence of nonlinear phenomena masked by a significant linear interaction is an important step to ensure that data are consistent with differential-susceptibility [37], we estimated additional models for all significant interactions to ensure that data are consistent with differential-susceptibility, including X2 (LESms2) and X2Z (LESms2 x FKBP5 variant). Then, simple slope analyses were conducted to test the statistical significance and magnitude of the environmental effects (LESms) over the criteria at the two levels of the FKBP5 moderators (risk and nonrisk carriers). These analyses were performed using STATA-12 [38].
In order to compute the Diathesis-Stress and Differential-Susceptibility Indices (DE/DS-Indices) recommended by Roisman et al. [37], significant interactions were entered in the Web-based program (http://www.yourpersonality.net/interaction/). These indices included the analysis of Regions of Significance on X (RoS X), the Proportion of Interaction (PoI) and the Proportion Affected (PA). RoS on X indicates the specific values of the predictor (LESms) below and above which the regression lines for the two levels of the moderator (FKBP5 variant) differ significantly in terms of the criterion. If RoS on X is statistically significant for the right and left sides of the environmental measure then differential-susceptibility is supported. The PoI index expresses the proportion of the total interaction represented on the right side of the crossover point, and indicates the area for which the effect of the predictor on the criterion is “for-better”. PoI values between 0.40 and 0.60 represent an interaction-effect highly consistent with differential-susceptibility, whereas values approaching 0 are highly consistent with diathesis-stress. Given that RoS on X and PoI are affected by the range of observations of the predictor (LESms), they should be computed by convention at ±2 SD from the mean of the predictor [37].
PA represents the proportion of individuals differentially affected by the interaction, that is, those falling above the crossover point of the interaction. Roisman et al. [37] suggested considering differential-susceptibility when PA is above 0.16 and definitely concluding diathesis-stress when PA is below 0.02. Given that PA needs to be computed on the assumption of a normal distribution of the predictor, we used the Shapiro-Wilk test to show that the LESms did not differ from the normal distribution (p = 0.452).
Results
T-test comparisons revealed that none of the FKBP5 variants were associated with the LESms. None of the 4 criteria were associated with sex, whereas LESms was higher in women than in men, indicating that women reported a greater positive impact of LEs (t (84) = 2.31, p = 0.023, d = 0.49, 95% CI [0.07, 0.91]).
Multiple regression main-effects analyses showed that genetic variables did not predict criterion measures (all p > 0.11). In contrast, positive LEs predicted lower levels of depressive symptoms (p < 0.001, 95% CI [-3.84, -1.25]), state anxiety (p = 0.027, 95% CI [-2.66, -0.17]), neuroticism (p = 0.011, 95% CI [-12.41, -1.68]) and social anxiety (p = 0.009, 95% CI [-5.85, -0.76]). Regarding GxE interactions, we found 7 significant results (see Table 3). In all cases, additional models showed that neither X2 nor X2Z (or their combination) were statistically significant (all p > 0.18), confirming the absence of nonlinear phenomena masked by the significant linear-interactions. No significant GxE interactions emerged for depressive symptoms or state-anxiety, although rs9296158, rs1360780 and the haplotype reached a trend towards significance for state-anxiety. The interaction of rs3800373, rs9296158 and the haplotype with LESms predicted both neuroticism and social anxiety, whereas the interaction of 1360780 with LESms only predicted neuroticism. Simple slopes analyses indicated that in all cases the effect of LESms on neuroticism and social anxiety was only significant for those carrying the risk alleles (non-risk carriers, all p > 0.455).
Table 3 also shows the DE/DS indices. Concerning neuroticism, the RoS on X analysis revealed vantage-sensitivity for rs3800373 in favor of risk-carriers, that is, individuals carrying risk-alleles (C/C or A/C) only differed from those carrying the non-risk alleles (A/A) in that they showed less neuroticism when experiencing more positive LEs. PoI > 0.60 and PA > 0.16 also supported the RoS on X analysis, highlighting the importance of the “for-better side” in this interaction. The interaction of LESms and rs1360780 on neuroticism was consistent with a diathesis-stress pattern evaluated with the RoS on X and the PoI indices (PoI 0.40 > 0.60), that is, the risk group showed higher neuroticism when experiencing more negative LEs. However, it was consistent with differential-susceptibility as evaluated with the PA, which indicated that 21% (criterion > 16%) of individuals were differentially affected by the interaction. Regarding rs9296158 and the haplotype, all DE/DS-indices converged to be highly consistent with a differential-susceptibility pattern. Individuals carrying either risk alleles or the risk-haplotype had higher neuroticism when experiencing more negative LEs as well as lower neuroticism when experiencing more positive LEs as compared with non-risk carriers (see Fig 1).
Description: Blue lines represent nonrisk-groups and black lines risk-groups. Graph (a) shows a vantage sensitivity effect: the group of individuals carrying the risk-allele (C) only differ from the nonrisk-group (A/A), showing significantly lower neuroticism when they experienced more positive LEs. Graphs (b) and (d) demonstrate a differential-susceptibility effect: risk-groups, with respect to nonrisk-groups, have significantly higher neuroticism if they experienced more negative LEs, as well as significantly lower neuroticism if they experienced more positive LEs. Graph (c) shows a diathesis-stress effect: the risk-group differed from the nonrisk-group, showing significantly higher neuroticism when they experienced more negative LEs. All graphs were plotted at 2 SD from LESms (range; -1.401, 2.532).
Regarding social anxiety, rs3800373, rs9296158 and the haplotype also showed significant interactions with LESms but, unlike neuroticism, all DE/DS-indices converged to be highly consistent with a diathesis-stress pattern; that is, individuals carrying the FKBP5 risk alleles only differed from those carrying the non-risk alleles in showing higher social anxiety when they experienced more negative LEs (see Fig 2).
Description: Blue lines represent nonrisk-groups and black lines risk-groups. All 3 graphs demonstrate a diathesis-stress effect: the risk-groups only differed from the nonrisk-groups by showing significantly higher social anxiety if they experienced more negative life events. All graphs were plotted at 2 SD from the mean of LESms (range; -1.401, 2.532).
Additional analyses
We performed exploratory analyses partialing out BDI-II and BAI variables to estimate the impact of depressive-anxiety symptoms on the significant interactions. Specifically, we partialed BDI-II and BAI variables out of the analysis of the LES x FKBP5 interaction. The results showed that the interaction with rs1360780 (diathesis-stress) lost statistical significance (p = 0.26), whereas all the other interactions involving a for-better side (rs3800373, 9296158 and haplotype) remained almost the same (only one dropped statistical significance to p = 0.06; see S1 Table). This may suggest that depressive-anxiety symptoms may be deeply involved in the for-worse effect found for rs1360780, but may not be involved in the for-better effects for rs3800373, rs9296158, and the haplotype. This would be consistent with research showing that the changes induced by negative LEs on neuroticism occur via an increase in anxiety-depressive symptoms, which is not the case for positive LEs [39]. Also, this would fit with evidence indicating that rs1360780 is one of the most relevant SNPs involved in the stress response [10]. However, this should be interpreted cautiously in the current study because of the post hoc nature of the analyses and because the BDI-II and BAI variables were entered in the model only as main effects. Following the suggestion of a reviewer, we reran all of the analyses in S1 Table using the recommendations by Keller [40], who suggested that in order to properly control for potential confounders, all the covariate-by-environment and the covariate-by-gene interaction terms should be entered in the same model that tests the gene-by-environment interaction term (see S1 Table). In this case, none of GxE interactions remain significant.
Discussion
The present study examined whether FKBP5 variability moderated the association of LEs with depressive symptoms, state-anxiety, neuroticism, and social anxiety. To the best of our knowledge, this is the first GxE study examining FKBP5 variability in the full spectrum of environmental exposures and directly testing differential-susceptibility versus diathesis-stress models. Significant GxE interactions emerged for neuroticism and social anxiety, whereas no interactions were found regarding depressive symptoms and state-anxiety.
Different moderation profiles were found for neuroticism: rs3800373 was consistent with a vantage-sensitivity effect, rs9296158 with differential-sensitivity, rs1360780 with diathesis-stress and, importantly, the haplotype (that accounted for all the accumulative genetic variance of those SNPs) was consistent with differential-susceptibility. The overall differential-susceptibility pattern was consistent with the cross-sectional study by Pluess et al. [41], who found the same pattern for homozygous subjects on the stress-related risk allele of the 5-HTTLPR gene and suggested that neuroticism could be a stable or unstable trait depending on the genotype. Furthermore, a recent non-genetic longitudinal study evidenced long-lasting changes on neuroticism induced by LEs, such that positive LEs decreased neuroticism scores whereas negative LEs increased them. Interestingly, the relation was mediated by depressive and anxiety symptoms for negative but not for positive LEs, which had the strongest effect [39].
Unlike the differential-susceptibility effect in neuroticism, all significant interactions for social anxiety were consistent with a diathesis-stress pattern. This discrepancy might be explained by the interplay of social anxiety with neuroticism, extraversion, and resilience. Social anxiety is a combination of high neuroticism and low extraversion; actually, the genetic factors that influence those dimensions account entirely for the genetic liability to social anxiety [42]. Interestingly, both personality dimensions, which are moderately heritable and also shaped by LEs, are linked to resilience [39, 43, 44]. High extraversion is associated with high positive affect, social support, and enhanced learning from positive reinforcement [32, 45]. The impact of positive LEs on building-up resilience, which is stronger than the impact of negative LEs, is mediated by decreases in neuroticism and increases in extraversion [44]. Therefore, given that social anxiety is defined by low extraversion and this involves a lower permeability to positive experiences, it is not surprising that, unlike neuroticism, the for-better side of the interaction indicating benefit from positive experiences was not significant. This resonate with the notion that additional environmental/genetic factors influencing extraversion would be involved in social anxiety. In this sense, the dopaminergic DRD4 gene might be an interesting candidate as it is linked to extraversion and has been proposed as a sensitivity marker [46, 47].
In the diathesis-stress model, being affected by environmental adversities is equivalent to being vulnerable, whereas being unaffected is equivalent to being resilient. By contrast, in the differential-susceptibility model, putatively vulnerable individuals are considered to be plastic in a for-better-and-for-worse pattern, whereas the putatively resilient individuals are thought to be “unmalleable” to both supportive and adverse environments [26]. It seems possible thus to distinguish two different GxE processes underlying a resilient outcome (i.e., overcoming adversity) depending on the degree of environmental sensitivity. The first would be a relatively stable resilience process derived from being unaffected by environmental influences (either positive or negative) and would characterize individuals who are considered “resilient” under both models. The second would be a highly dynamic process arising from enhanced responsiveness to environmental influences and would characterize individuals considered to be “vulnerable” (diathesis-stress) or “plastic” (differential-susceptibility). In this dynamic process, the detrimental effects of adversity would be compensated by benefiting from the positive effects of subsequent supportive experiences. The plausibility of these processes appears to be supported by research suggesting that individuals carrying “sensitivity” genes benefit the most from certain therapeutic interventions [48].
Previous studies reporting GxE interactions on adult depressive outcomes mostly evaluated the effect of distant and severe childhood adversities [10]. In those studies, the most replicated GxE interaction was found for rs1360780, which has also been shown to mediate childhood trauma interactions via epigenetic processes [49, 50]. Unlike previous work, the present study evaluated the effect of common (not necessarily traumatic) and recent LEs in early adulthood. Therefore, it is attractive to speculate that the lack of significant FKPB5 x adult LEs interactions on depressive symptoms found in this study is partially related to the fact that epigenetic changes that mediate the association of early-life adversity with adult depression are no longer (or less dramatically) induced by the impact of negative (though not necessarily severe) LEs in adulthood. Regarding state anxiety, the measure used in our study is largely characterized by somatic complaints and might not be as sensitive to FKPB5 x LEs interactions as the trait and social component measures.
The strengths of this study include the estimation of a haplotype that increases the power to detect genetic associations [51] and the use of specific statistical tests to formally investigate the pattern of GxE interactions [37]. Limitations of the study include the cross-sectional design, and therefore the speculative nature of any causal inference; the use of self-report measures, which entail memory biases; and the relatively small sample size, which limited the statistical power to detect significant results. It is important to note that we did not control for multiple testing in the present study given that our analyses were limited to testing a priori hypothesized relationships and the serious limitations related to post hoc alpha adjustment [52]. However, following a Reviewer’s suggestion about controlling for covariate interaction terms as recommended by Keller [40], we have included a table containing these results as supplementary material. Finally, we also estimated additional models for all significant interactions to ensure that data were consistent with differential-susceptibility. In this sense, although we used the three commonly used tests of differential-susceptibility proposed by Roisman et al. [37] that have been extensively used in the GxE field, recent methodological work [53] suggests that these tests present relevant limitations that can be minimized with novel approaches. More empirical research is needed to analyze the performance and limitations of Roisman’s guidelines.
From a clinical standpoint, this avenue of research with the FKBP5 gene may have interesting implications. For instance, there have been efforts to develop drugs that block FKBP5 activity in order to modify the risk for the development of stress-related disorders [10, 16] based on research conducted within a diathesis stress framework. Further examination of whether FKBP5 fits better a differential susceptibility model might be relevant to inform whether interventions aimed at blocking FKBP5 activity in individuals carrying ‘risk’ genotypes might have the undesirable effect of altering a natural resilience mechanism: the potential of benefitting from positive and supportive environments.
To conclude, our results indicated that the interactions predicting neuroticism fit different models for different polymorphisms, although the overall effect indicated by the haplotype was consistent with the differential-susceptibility hypothesis. The interactions predicting social anxiety (rs3800373, rs9296158 and haplotype) were consistent with the diathesis-stress model, possibly reflecting that the low extraversion component of social anxiety is associated with a lower permeability to positive experiences.
Supporting information
S1 Table. Comparison of the interaction terms and explained variance between the original and the post-hoc regressive models for the neuroticism and the social anxiety statistical significant interactions.
https://doi.org/10.1371/journal.pone.0193044.s001
(DOCX)
References
- 1. Gottesman II, Shields J. A polygenic theory of schizophrenia. Proc Natl Acad Sci U S A. 1967;58: 199–205. pmid:5231600
- 2. Monroe SM, Simons AD. Diathesis-stress theories in the context of life stress research: implications for the depressive disorders. Psychol Bull. 1991;110: 406–425. pmid:1758917
- 3.
Zuckerman M. Vulnerability to Psychopathology: A Biosocial Model. Washington: American Psychological Association; 1999.
- 4. Belsky J, Pluess M. Beyond diathesis stress: differential susceptibility to environmental influences. Psychol Bull. 2009;135: 885–908. pmid:19883141
- 5. Belsky J, Bakermans-Kranenburg MJ, van Ijzendoorn MH. For better and for worse: Differential susceptibility to environmental influences. Curr Dir Psychol Sci. 2007;16: 300–304.
- 6. Belsky J, Hartman S. Gene-environment interaction in evolutionary perspective: differential susceptibility to environmental influences. World Psychiatry. 2014;13: 87–89. pmid:24497257
- 7. Pluess M. Individual Differences in Environmental Sensitivity. Child Dev Perspect. 2015;9: 138–143.
- 8. van IJzendoorn MH, Belsky J, Bakermans-Kranenburg MJ. Serotonin transporter genotype 5HTTLPR as a marker of differential susceptibility? A meta-analysis of child and adolescent gene-by-environment studies. Transl Psychiatry. 2012;2: e147. pmid:22872162
- 9. Daskalakis NP, Binder EB. Schizophrenia in the spectrum of gene-stress interactions: the FKBP5 example. Schizophr Bull. 2015;41: 323–329. pmid:25592294
- 10. Zannas AS, Binder EB. Gene-environment interactions at the FKBP5 locus: Sensitive periods, mechanisms and pleiotropism. Genes Brain Behav. 2014;13: 25–37. pmid:24219237
- 11. Binder EB. The role of FKBP5, a co-chaperone of the glucocorticoid receptor in the pathogenesis and therapy of affective and anxiety disorders. Psychoneuroendocrinology. 2009;34 Suppl 1: S186–S195.
- 12. Fani N, Gutman D, Tone EB, Almli L, Mercer KB, Davis J, et al. FKBP5 and Attention Bias for Threat. JAMA Psychiatry. 2013;70: 392–400. pmid:23407841
- 13. White MG, Bogdan R, Fisher PM, Muñoz KE, Williamson DE, Hariri AR. FKBP5 and emotional neglect interact to predict individual differences in amygdala reactivity. Genes Brain Behav. 2012;11: 869–878. pmid:22979952
- 14. Shibuya N, Suzuki A, Sadahiro R, Kamata M, Matsumoto Y, Goto K, et al. Association study between a functional polymorphism of FK506-binding protein 51 (FKBP5) gene and personality traits in healthy subjects. Neurosci Lett. 2010;485: 194–197. pmid:20849924
- 15. Binder EB, Bradley RG, Liu W, Epstein MP, Deveau TC Mercer KB, et al. Association of FKBP5 polymorphisms and childhood abuse with risk of posttraumatic stress disorder symptoms in adults. JAMA. 2008;299: 1291–1305. pmid:18349090
- 16. Zannas AS, Wiechmann T, Gassen NC, Binder EB. Gene-Stress-Epigenetic Regulation of FKBP5: Clinical and Translational Implications. Neuropsychopharmacology. 2016;41: 261–274. pmid:26250598
- 17. Appel K, Schwahn C, Mahler J, Schulz A, Spitzer C, Fenske K, et al. Moderation of adult depression by a polymorphism in the FKBP5 gene and childhood physical abuse in the general population. Neuropsychopharmacology. 2011;36: 1982–1991. pmid:21654733
- 18. Vanzomeren-Dohm A, Pitula CE, Koss KJ, Thomas K, Gunnar MR. FKBP5 moderation of depressive symptoms in peer victimized, post-institutionalized children. Psychoneuroendocrinology. 2015;51: 426–430. pmid:25462914
- 19. Collip D, Myin-Germeys I, Wichers M, Jacobs N, Derom C, Thiery E, et al. FKBP5 as a possible moderator of the psychosis-inducing effects of childhood trauma. Br J Psychiatry. 2013; 202: 261–268. pmid:23429203
- 20. Cristóbal-Narváez P, Sheinbaum T, Rosa A, Ballespí S, de Castro-Catala M, Peña E, et al. The Interaction between Childhood Bullying and the FKBP5 Gene on Psychotic-Like Experiences and Stress Reactivity in Real Life. PLoS One. 2016; 11: e0158809. pmid:27389186
- 21. Scheuer S, Ising M, Uhr M, Otto Y, von Klitzing K, Klein AM. FKBP5 polymorphisms moderate the influence of adverse life events on the risk of anxiety and depressive disorders in preschool children. J Psychiatr Res. 2016;72: 30–36. pmid:26521051
- 22. Roy A, Gorodetsky E, Yuan Q, Goldman D, Enoch MA. Interaction of FKBP5, a stress-related gene, with childhood trauma increases the risk for attempting suicide. Neuropsychopharmacology. 2010;35: 1674–1683. pmid:20090668
- 23. Roy A, Hodgkinson CA, Deluca V, Goldman D, Enoch MA. Two HPA axis genes, CRHBP and FKBP5, interact with childhood trauma to increase the risk for suicidal behavior. J Psychiatr Res. 2012;46: 72–79. pmid:21978546
- 24. Bevilacqua L, Carli V, Sarchiapone M, George DK, Goldman D, Roy A, et al. Interaction Between FKBP5 and Childhood Trauma and Risk of Aggressive Behavior. Arch Gen Psychiatry. 2012;69: 62–70. pmid:22213790
- 25. Klengel T, Binder EB. Allele-specific epigenetic modification: a molecular mechanism for gene-environment interactions in stress-related psychiatric disorders? Epigenomics. 2013;5: 109–112. pmid:23566086
- 26. Belsky J, Pluess M. Beyond risk, resilience, and dysregulation: phenotypic plasticity and human development. Dev Psychopathol. 2013;25: 1243–1261. pmid:24342838
- 27. Xie P, Kranzler HR, Poling J, Stein MB, Anton RF, Farrer LA, et al. Interaction of FKBP5 with childhood adversity on risk for post-traumatic stress disorder. Neuropsychopharmacology. 2010;35: 1684–1692. pmid:20393453
- 28. Zimmermann P, Brückl T, Nocon A, Pfister H, Binder EB, Uhr M, et al. Interaction of FKBP5 Gene Variants and Adverse Life Events in Predicting Depression Onset: Results From a 10-Year Prospective Community Study. Am J Psychiatry. 2011;168: 1107–1116. pmid:21865530
- 29. Barrantes-Vidal N, Gross GM, Sheinbaum T, Mitjavila M, Ballespí S, Kwapil TR. Positive and negative schizotypy are associated with prodromal and schizophrenia-spectrum symptoms. Schizophr Res. 2013;145: 50–55. pmid:23402694
- 30.
Beck AT, Steer RA, Brown GK. Manual for the Beck Depression Inventory-II. 2nd ed. San Antonio: Psychological Corporation; 1996.
- 31.
Beck & Steer , 2011 Manual BAI. Inventario de Ansiedad de Beck (Adaptación española de Sanz, J.). Madrid: Pearson Educación; 2011.
- 32.
Costa PT, McCrae RR. Inventario de Personalidad NEO Revisado (NEO PI-R). Inventario NEO reducido de Cinco Factores (NEO-FFI) Manual. Madrid: TEA; 1999.
- 33. Piqueras JA, Espinosa-Fernández L, García-López LJ, Beidel DC. Validation of the Social Phobia and Anxiety Inventory-Brief form (SPAI-B) in Spanish young adults. Behav Psychol. 2012;20: 505–528.
- 34. Sarason IG, Johnson JH, Siegel JM. Assessing the impact of life changes: development of the Life Experiences Survey. J Consult Clin Psychol. 1978;46: 932–946. pmid:701572
- 35. Stephens M, Donnelly P. A comparison of bayesian methods for haplotype reconstruction from population genotype data. Am J Hum Genet. 2003;73: 1162–1169. pmid:14574645
- 36. Barrett JC, Fry B, Maller J, Daly MJ. Haploview: analysis and visualization of LD and haplotype maps. Bioinformatics. 2005;21: 263–265. pmid:15297300
- 37. Roisman GI, Newman DA, Fraley RC, Haltigan JD, Groh AM, Haydon KC. Distinguishing differential susceptibility from diathesis–stress: Recommendations for evaluating interaction effects. Dev Psychopathol. 2012;24: 389–409. pmid:22559121
- 38.
StataCorp. Stata Statistical Software: Release 12. StataCorp LP: College Station, TX; 2011.
- 39. Jeronimus BF, Ormel J, Aleman A, Penninx BW, Riese H. Negative and positive life events are associated with small but lasting change in neuroticism. Psychol Med. 2013;43: 2403–2415. pmid:23410535
- 40. Keller MC. Gene × environment interaction studies have not properly controlled for potential confounders: the problem and the (simple) solution. Biol Psychiatry. 2014;75: 18–24. pmid:24135711
- 41. Pluess M, Belsky J, Way BM, Taylor SE. 5-HTTLPR moderates effects of current life events on neuroticism: Differential susceptibility to environmental influences. Prog Neuropsychopharmacol Biol Psychiatry. 2010;34: 1070–1074. pmid:20573579
- 42. Bienvenu OJ, Hettema JM, Neale MC, Prescott CA, Kendler KS. Low extraversion and high neuroticism as indices of genetic and environmental risk for social phobia, agoraphobia, and animal phobia. Am J Psychiatry. 2007;164: 1714–1721. pmid:17974937
- 43. Ogle CM, Rubin DC, Siegler IC. Changes in neuroticism following trauma exposure. J Pers. 2014;82: 93–102. pmid:23550961
- 44. Sarubin N, Wolf M, Giegling I, Hilbert S, Naumann F, Gutt D, et al. Neuroticism and extraversion as mediators between positive/negative life events and resilience. Pers Indiv Differ. 2015;82: 193–198.
- 45. Pangallo A, Zibarras L, Lewis R, Flaxman P. Resilience through the lens of interactionism: A systematic review. Psychol Assess. 2014;27:1–20. pmid:25222438
- 46. Belsky J, Hartman S. Gene-environment interaction in evolutionary perspective: differential susceptibility to environmental influences. World Psychiatry. 2014;13: 87–89. pmid:24497257
- 47. Eichhammer P, Sand PG, Stoertebecker P, Langguth B, Zowe M, Hajak G. Variation at the DRD4 promoter modulates extraversion in Caucasians. Mol Psychiatry. 2005;10: 520–522. pmid:15738935
- 48. van Ijzendoorn MH, Bakermans-Kranenburg MJ. Genetic differential susceptibility on trial: meta-analytic support from randomized controlled experiments. Dev Psychopathol. 2015;27: 151–162. pmid:25640837
- 49. Klengel T, Mehta D, Anacker C, Rex-Haffner M, Pruessner JC, Pariante CM, et al. Allele-specific FKBP5 DNA demethylation mediates gene-childhood trauma interactions. Nat Neurosci. 2013;16: 33–41. pmid:23201972
- 50. Klengel T, Binder EB. FKBP5 allele-specific epigenetic modification in gene by environment interaction. Neuropsychopharmacology. 2015;40: 244–246. pmid:25482174
- 51. Crawford DC, Nickerson DA. Definition and clinical importance of haplotypes. Annu Rev Med. 2005;56: 303–320. pmid:15660514
- 52. O’Keefe D. Colloquy: Should familywise alpha be adjusted? Against familywise alpha adjustment. Hum Commun Res. 2003;29: 431–447.
- 53. Del Guidice M. Statistical tests of differential susceptibility: Performance, limitations, and improvements. Dev Psychopathol. 2017; 29: 1267–1278. pmid:28052783