The most frequently investigated polymorphism (n = 18) was the 5-HTTLPR (serotonin transporter linked polymorphic region) which is located in the promoter of the human serotonin transporter gene (SLC6A4). Out of these studies, 12 found the L′/L′ allele to be associated with resilience (37, 59, 102, 114, 117, 126, 130–134, 144) and four studies found the S or S′ allele to be associated with resilience (103, 104, 123, 145), and one study found that only haplotypes of the serotonin transporter were associated with resilience, the single genetic variants were not (115). There are a several factors that might contribute to these inconsistent findings.

First, the studies used different allelic variation classification. The A 44 bp deletion/insertion results in a short (S) or a long (L) allele, creating a biallelic classification of S/L alleles (149). The rs25531 is a SNP in the L allele, which leads to an A-G polymorphism. This provides a functional triallelic polymorphism as S, L_G, and L_A. The L_G allele is functionally equivalent to the S allele, having the same expression levels, both of which are lower than L_A (69). Consequently, the L′/L′ can be reclassified as equal to L_A/L_A; L′/S′ = L_A/S and L_A/L_G; and S′/S′ = S/S & L_G/S and L_G/L_G. This is the triallelic classification system. 12 of the studies excluded the SNP rs25531 from their analyses (37, 102, 103, 114, 115, 117, 126, 130, 132, 134, 144, 145), and five studies genotyped for rs25531 and recoded accordingly (59, 104, 123, 131, 133). One study found a significant association with the biallelic classification system, but not with the triallelic (37), and another study excluded the results of the triallelic system due to a low number of participants carrying the L_G allele (102).

A second factor contributing to inconsistencies in findings between studies are differences in the developmental stage at which effects were studied. Eight of the studies investigated children or adolescents, while ten investigated adults. Within the studies that did use the triallelic classification system, those that found that the L/L or L′/L′ genotype of 5-HTTLPR contributed to resilience, investigated children/adolescents (59, 131, 133), while those that found the S/S or S′/S′ genotype investigated adults (104, 123). This could be explained by the evolutionary genetic theory of antagonistic pleiotropy (150), which was first proposed as an explanation for multiple effects of a gene with opposing effects on fitness depending on the developmental stage. Notably, genes enhancing reproductive success in early life are predicted to become deleterious in later life. This opposite effect in young vs. old participants should be interpreted with caution because, although there is some evidence emerging of the age-genotype interaction [e.g., (151, 152)], the evidence is limited.

A third factor that may explain discrepancies in consensus are sex differences in the 5-HTTLPR polymorphism. Previous research from both animals and humans suggests that males and females carrying the short 5-HTTLPR allele respond to environmental stress factors in opposite directions (153, 154). A recent review of sex differences in affective disorders found the S allele to be differentially associated with an increased risk of depression, anxiety, and internalizing behavior in women, and increased risk of aggressiveness, conduct disorder, and externalizing behavior in men (155). The two studies that investigated female-only populations (114, 132) found the L allele to be associated with resilience, while one study investigating males only found the S/S homozygous genotype to be associated with resilience.

The second most frequently investigated gene was Brain-Derived Neurotrophic Factor (BDNF) (n = 5). All five of the studies reviewed investigated SNP rs6265 in exon 11 of the BDNF gene (105, 116, 122, 129, 131). Of the four studies, three found Val/Val homozygous genotype to be associated with resilience, while two studies found the Met allele to be associated with resilience. This inconsistency in results could, again, be due to a number of confounds. Allele frequencies of the BDNF Val66Met polymorphism are dependent on ethnic background: frequency of the Met allele among Caucasians is 25–32% but is more common in Asians at 40–50% (156). Of the five studies, Aizawa et al. (129) and Kang et al. (105) both studied Asian populations, Nederhof et al. (131) and Surtees et al. (122) both studied European Caucasians, and Peters et al. (116) studied Brazilian nationals. Sex differences in the populations studied may also result in inconsistent findings. A recent review found some action mechanisms or functions of BDNF vary according to sex and encourage considering sex influences when drawing conclusions from studies (157). One study reviewed only observed a significant association between resilience and the Val allele in males (105).

Two of the studies assessed polygenic resilience with a gene-gene interaction effect. One study suggests a sex-dependent result, where the BDNF Val66Met polymorphism modulates the effects of the Val158Met polymorphism of COMT on resilience, but only in males (105). The other study investigated the gene-gene interaction between BDNF Val66Met polymorphism and 5-HTTLPR, specifically investigating resilience depending on the genotype of the other gene variant (131).

Three studies investigated the rs53576 polymorphism in the oxytocin receptor gene (OXTR). Cicchetti and Rogosch (144) found the impact of child maltreatment on resilient functioning was decreased in children with the GG genotype when compared to GA or AA genotypes, assuming the dominant effect of the A allele. Saphire-Bernstein et al. (124) reported a link between the GG genotype and higher levels of optimism, mastery, and self-esteem, suggesting that the influence of this OXTR polymorphism on psychological resources may primarily mediate the effects of OXTR on depressive symptoms.

In contrast, Bradley et al. (158) reported that, regardless of family environment, adults with the AA genotype had higher levels of positive affect and resilient coping scores. The authors reported that GG and AG genotypes were more influenced positively and negatively by the family environment, in keeping with the differential susceptibility hypothesis.

One study investigated the OXTR variant rs2254298 on psychological resilience as measured by CD-RISC (106). The authors found a dose-dependent effect of the A allele in ethnically Korean adolescents and young adults, with highest resilience in carriers of the GG genotype, decreasing by 3.84 on the CD-RISC scale with a one-copy increase in the A allele. An OXTR haplotype containing the rs2254298 A allele was also associated with lower CD-RISC scores.

Four studies investigated Apolipoprotein E (APOE). The presence of the ε4 allele is a strong genetic risk factor for Alzheimer's disease (AD), whereas the ε2 allele is protective against AD. Two of the studies found both the ε3 and ε2 allele to be associated with resilience (136, 143), and the other study found ε3 allele to be associated with resilience (137). However, this study did not genotype for the APOE ε2 allele, so there is consensus amongst the studies reviewed about which allele is associated with resilience despite the studies using different instruments to measure resilience and associating the allele with different outcomes: AD, PTSD, and epilepsy. A fourth study found US veterans with mild traumatic brain injuries (TBIs) had significantly better scores on the resilience scale of the TBI- Quality of Life measure (120), but again, the study did not genotype for ε2 or ε3 isoforms of APOE so it cannot be included in the consensus of the APOE ε2 allele conferring resilience.

Three studies investigated catechol-O-methyltransferase (COMT). All studies investigated a functional polymorphism on rs3680, Val¹⁵⁸Met (105, 107, 126), yet again the results were inconsistent. Of the two studies which used the same instrument to measure resilience, the CD-RISC, van Rooij et al. (107) found childhood trauma load was associated with increased inhibition-related hippocampal activation in Val/Val carriers, with the relationship between childhood trauma and resilience mediated by hippocampal activation. This implies that Val/Val carriers may develop a mechanism to cope with early adversity by relying more on contextual information processed in the hippocampal region in order to regulate behavior. However, Kang et al. (105) found a conflicting, polygenic, and sex-dependent result. Males with homozygous Val allele of BDNF and COMT Met-present genotype were found to have higher resilience than BDNF Met carriers. Amstadter et al. (126) also found associated resilience with the Met allele. This inconsistency could again be attributed to sex differences: van Rooij et al. (107) only studied women, while Kang et al. (105) only found a significant association with resilience in males. COMT enzyme activity and the neurochemistry and behavior of COMT knock out mice are both markedly sexually dimorphic (159). In addition, the Val¹⁵⁸Met polymorphism shows sexual dimorphism in predisposition to psychiatric disorders (159).

Results could also be conflicting due to the different ethnic groups analyzed by the researchers. COMT Val/Met allele frequencies differ across ethnic groups, with higher heterozygosity in European populations than other geographical regions (160).

Further the assumed mode of inheritance in each study may contribute to inconsistencies in results. Both Kang et al. (105) and van Rooij et al. (107) assumed dominance of the Met allele and grouped the sample into subjects with the Val/Val genotype vs. Val/Met and Met/Met. By contrast, Amstadter et al. (126) assumed dominance of the Val allele (Met/Met vs. Val/Met and Val/Val). Due to this difference in modes of inheritance, it is unclear whether heterozygotes of rs4680 in COMT (Val/Met carriers) would be associated with increased or decreased resilience.

The human dopamine receptor D4 gene (DRD4) was investigated by three studies (108, 135, 144). Das et al. (108) found an interaction effect of childhood adversity, the DRD4 7-repeat allele, and resilience. A decline in resilience was associated with increased childhood adversity in individuals who lacked the 7-repeat allele, therefore suggesting that the 7-repeat allele is protective. Conflicting results come from Gervai's et al. (135) study, which analyzed a haplotype of the 7-repeat allele and the −521C/T promoter polymorphism of DRD4 which found an absence of the T.7 haplotype acts as a resilience factor. Cicchetti and Rogosch (144) studied the effects of the −521C/T promoter polymorphism of DRD4 (rs1800955) and revealed another interaction effect where maltreated children with CC or CT genotypes were protected from the significantly detrimental effect of extensive maltreatment, assuming the dominant effect of the C allele (CC and CT vs. TT). This is in some way consistent with Gervai et al. (135).

Two studies investigated tumor necrosis factor α (TNFα) (113, 146). Bruenig et al. (113) examined a SNP (rs1800629) in the promoter region of TNFα, which has been identified with potential functional consequences for gene transcription. While no significant effect was observed between CD-RISC and genotype, there was a significant correlation between resilience and decreased TNFα serum levels. Dunn et al. (146) found two SNPs to be significantly associated with a resilient latent class: rs2229094 and rs1800629. Individuals who possess the A rare allele for rs1800629 had higher odds of belonging to the resilient class, while individuals who possess the T allele in rs2229094 had higher odds of belonging to the resilient class.

Interleukin-10 (IL10), was associated with resilience in two studies (110, 146). Rana et al. (110) found the SNP rs1800896 in IL10 to be associated with optimism in an exploratory multi-locus polygenic analysis with two other SNPs: rs6323 of MAOA and rs1800792 of Fibrinogen Gamma Chain (FGG). Dunn et al. (146) found participants who possessed a G allele of SNP rs1518111 of IL10 to have higher odds of belonging to the resilient latent class. Both studies used a strategy of selecting high-probability candidate genes to analyse. While scientifically justifiable, selecting genes from a list of “usual suspects” introduces bias, and many genetic variants identified by GWAS have included genes previously not thought to be involved in the target disease etiology (161).

Two studies, although of the same population, found two significant genome-wide associations between SNPs (rs3748348 and rsrs3648346) of Ribonuclease A Family Member 13 (RNASE13), and executive functioning resilience (141, 142), suggesting a cognitive basis for resilience. Pathway-based analysis by gene set enrichment analysis implicated EF resilience with genetic pathways involving dendritic/neuron spine, presynaptic membrane, and post-synaptic density (142). RNASE13 has been linked to susceptibility for aortic aneurysm in Japanese individuals (162).

Two studies investigated monoamine oxidase A (MAOA) genetic variants (110, 163). In addition, Azadmarzabadi et al. (111) examined mRNA levels of MAOA and found that MAOA expression is downregulated in individuals with low stress resilience, as measured by the CD-RISC. Rana et al. (110) found SNP rs6323 in MAOA to be associated with both resilience and optimism in an exploratory multi-locus polygenic analysis. Tikkanen et al. (163) found alcoholic violent offenders carrying the low MAOA activity genotype (3-repeat allele) had no increase in risk of reoffending when exposed to the independent predictors of recidivism, heavy drinking and childhood physical abuse. Significant associations have previously been reported between MAOA and anger-related traits (164).

The most frequently investigated polymorphism associated with the operationalisation of resilience as levels of reactivity, was again the 5-HTTLPR (n = 18). The environmental measures included stressful life events, stroke, war zone combat, victimization, racial discrimination, child abandonment, interferon-alpha treatment, and neuroticism. Of these 16 studies, 13 found the L allele to be associated with resilience (132, 148, 165, 166, 174, 180, 182, 191, 194, 196, 212, 214, 216), whereas three studies found the S allele to be associated with resilience (177, 197, 203), and two studies found the SL heterozygous allele to be associated with resilience (176, 220).

Two studies investigated APOE, both concerning the severity of PTSD symptoms among veterans exposed to combat trauma (198, 201). Both studies found that absence of the ε4 allele was associated with significantly less severity of PTSD symptoms than carriers of the ε4 allele.

Eight studies investigated the BDNF Val66Met polymorphism (172, 173, 178, 179, 200, 203, 212, 219) in environmental measures of childhood adversity, childhood abandonment, traumatic life events, newly diagnosed diabetes, and physical exercise. Of the eight studies, 4 associated the Met allele with resilience (178, 179, 203, 212), and the other three found Val/Val or Val/Met genotype to be associated with resilience (172, 173, 200, 219). Interestingly, the four studies that found the Met allele to be associated with resilience all studied children or young adults. Min et al. (172) found a gender-specific effect with the Val/Met genotype in males appearing to be resilient against the higher levels of anxiety associated with exposure to childhood maltreatment. Mata et al. (178), also found gender-specific effects where being physically active was protective for girls with a BDNF met allele against the development of depressive symptoms, but not for girls with a Val/Val genotype—examining how physical activity contributes to resilience to psychopathology.

Five studies investigated COMT polymorphisms, and all studies reported the Val allele as associated with resilience, reaching consensus (186, 204, 215, 217, 229). Of the five studies, two studies were fMRI studies assessing the effects of COMT Val158Met genotype on emotional stimuli processing. They found the presence of a Val158 allele results in an increase in emotional resilience against negative mood states via decreased neuronal activation to unpleasant stimuli (215, 217). Two studies used psychological measures; subjective stress experience (204), and sadness score (186), and one study assessed resilience via cortisol response levels (229).

Five studies investigated the interaction between child abuse/maltreatment and polymorphisms in the corticotropin-releasing hormone receptor 1 gene (CHRH1). CRHR1 is a major regulator of neuroendocrine, autonomic, and behavioral responses to stress, and its ligand, CRH, has been associated with mood and anxiety disorders (239). The studies found significant GxE interactions with multiple individual SNPs, as well as with a common three allele haplotype, the TAT haplotype, formed by SNPs rs7209436, rs110402, and rs242924. Bradley et al. (240) found that carriers of the TAT haplotype were found to be protected from the adverse impact of severe childhood abuse, as measured by the Childhood Trauma Questionnaire (CTQ). In an extended sample of the Bradley et al. study, Heim et al. (241) confirmed the protective effect of the rs110402 A-allele against the negative emotional consequences of childhood abuse in the male subsample only. They hypothesize that this effect may not be due to sex differences but due to the differences in the abuse experienced, with physical abuse most commonly experienced by men, and sexual abuse most commonly experienced by women. Although, Polanczyk et al. (242) only replicated the protective effect of the TAT haplotype in a sample of women. The second sample in Polanczyk's et al. study consisted of men and women, and no significant result was found. The authors suggest that this difference is attributable to the different types of measurement of maltreatment between the samples, with maltreatment reported via the CTQ more likely to elicit emotional memories, and the hypothesized link between CRHR1's protective effect is related to its function in the consolidation of memories of emotionally arousing experiences.

Grabe et al. (183) identified a previously unreported SNP, rs17689882, in which the minor allele (A) had a protective effect against depressive symptoms among individuals who had experienced moderate to severe neglect. Laucht et al. (184) replicated this result, finding the minor A allele as protective.

Two studies investigated GxE interactions associated with the dopamine transporter 1 gene (DAT1, SLC6A3). Felten et al. (186) found carriers of the homozygous 9-repeat allele showed dramatically reduced sadness scores in a sample of healthy Caucasians. In contrast, Satterfield et al. (188) found that subjects homozygous for the 10-repeat allele were resilient to the build-up of cognitive performance impairment due to total sleep deprivation.

Three studies investigated GxE interactions associated with the glucocorticoid receptor co-chaperone, FKBP5, which modulates glucocorticoid receptor sensitivity. Two of the three studies examined the moderating effects of FKBP5 rs3800373 on suicide risk (209) and attachment security and depressive symptoms (58). Roy et al. (209) found that individuals that lacked either major homozygote were resilient to the effects of childhood trauma (suicide attempt prevalence 0.24). Similarly, Borelli et al. (58) found that minor C allele homozygotes had the most optimal psychological profiles in positive caregiving contexts, but the most maladaptive outcomes in adverse caregiving contexts. This is supportive of the DSH. Buchmann et al. (230) investigated the interaction between the FKBP5 rs1360780 genotype, childhood adversity, and cortisol response. The authors found a blunted cortisol response to environmental challenges, as observed in homozygous carriers of the 9-repeat allele, indicative of an enhanced biological resilience to healthy individuals with a history of childhood adversity.

Two studies investigated how the IL6 rs1800795 polymorphism interacts with adverse environmental factors to promote chronic inflammation. The first study found that older adults with a C allele of the polymorphism are protected from the increased risk of inflammation-related disease and mortality shown in individuals with the G allele of the polymorphism and high socioeconomic adversity exposure (101). The second study identified a protective effect of the G allele in adolescents, which desensitizes individual physiology to the pro-inflammatory effects of adverse socioenvironmental conditions (151). This age-dependent antagonistic pleiotropy suggests that genetic resilience is dependent on the developmental environment that interacts with social conditions.

Six studies investigated GxE interactions associated with the polymorphism rs16147 of Neuropeptide Y. Five of the six studies associated individuals homozygous for the T allele with resilience (82, 119, 168, 232). Sommer et al. (82) found that childhood adversity was significantly associated with increased depression and anxiety for homozygote carriers of the C allele, but homozygote carriers of the T allele were protected. Amstadter et al. (168) found the TT genotype interacts with impact level of hurricane exposure to be protective against anxiety. Chang et al. (232) reported an interaction between chronic high stress, the TT genotype, and high cardiac vagal tone, a physiological measure that is gaining support as a marker of resilience by capturing the integration of the varying regulatory mechanisms of the autonomic nervous system (243). Gan et al. (119) reported two separate studies within the same paper. Study 1 found that T allele carrier earthquake survivors reported consistent levels of resilience in low, moderate and high trauma exposure. Study 2 found T allele carriers who reported a high number of early stressful life events were associated with a greater positive future focus.

Donner et al. (169), in contrast to the other studies, found no GxE interaction effect with rs16147, childhood adversity, and anxiety susceptibility, but instead found risk haplotypes significantly associated with anxiety that carry the T-allele of the SNP.

Resnick et al. (118) found an association between NPY and psychological resilience following hip fracture but they did not report allelic variation.

Two studies investigated the A118G (SNP rs1799971) polymorphism of the μ-opioid receptor gene (OPRM1). Levran et al. (236) assessed whether opioid-related genetic variants contribute to reduced vulnerability to relapse to heroin in persons without agonist treatment. The authors found that carrying at least one copy of the G allele is significantly higher in subjects with long term abstinence from heroin, suggesting this genotype may blunt the endogenous stress response and contribute to resilience against relapse.

Troisi et al. (213) found that G allele carriers had similarly high scores on fearful attachment, regardless of the quality of maternal care. By comparison, AA genotype carriers were more affected by early experience with those recalling high levels of maternal care exhibiting less fearful attachment behavior, and those recalling low levels of maternal care exhibiting high levels of fearful attachment. This is again in line with the DSH and suggests G allele carriers have a decreased sensitivity to adverse rearing environments.

Three studies investigated the interactions between polymorphisms in OXTR and environments on a range of outcomes. Two of the studies investigated the rs53576 polymorphism in OXTR, and both studies associated A allele carriers with resilience. Bradley's et al. (211) findings suggest that A allele carriers are more resilient to the adverse effects of severe childhood adversity by being protected against emotional dysregulation and disorganized attachment. Hostinar et al. (167) replicated this finding, showing maltreated adolescent A-carriers exhibit the same level of psychological symptoms as non-maltreated adolescents, supporting resilience research that the experience of maltreatment does not deterministically lead to psychopathology.

Loth et al. (192) investigated the rs237915 SNP of OXTR and found that CC homozygotes were more resilient against the effect of stressful life experiences than other genotypes, partly mediated by genotype-dependent sensitivity to the reinforcement values of negative social cues.

Two studies investigated the TNFα G308A polymorphism on performance impairment during total sleep deprivation (TSD) (189, 190). Skeiky et al. corroborated Satterfield and colleagues' earlier findings that indicated that the A allele at the TNFα 308 locus (SNP rs1800629) is associated with resilience to psychomotor vigilance performance impairment during TSD, in comparison with the more common G allele.

Briand et al. (244) used a transgenic mouse model, with C57BL/6 mice possessing the functionally equivalent SNP in the mouse Oprm1 gene (A112G) as the human OPRM1 A118G SNP. To determine the role of this polymorphism in behavioral interactions and to investigate the functional mechanisms underlying this behavior, the authors used a social defeat paradigm. The authors found that G-allele carriers were resilient to the deleterious consequences of stress, as measured by positive social interactions and reduced anhedonia. This is consistent with the human studies of OPRM1 also investigated in this review, which found G-allele carriers are resilient against relapse to heroin use (236) and show decreased levels of fearful attachment in low maternal care environments (213). An increase in brain activation in stress and reward circuitry is also evident in resilient Oprm1 G/G mice in Briand's et al. (244) study, thus suggesting that modeling this SNP in mice provides construct validity for this model of resilience These models are in direct contrast with results in a recent meta-analysis which found no moderating effect of OPRM1 A118G, on response to naltrexone treatment in individuals with alcohol use disorder (245).

Cohen et al. examined the Npy-ergic system and its association with behavioral responses to stress in a rat model of PTSD (246). The authors classified subpopulations of Sprague-Dawley rats based on their performance in elevated plus maze and acoustic startle response tests using a pre-set cut-off criterion after exposure to predator-scent stress. The study found protective qualities of Npy. Notably, an extreme reduction in Npy expression in selected brain regions of the animals severely affected by the stressor. This behavioral disruption was reversed by administering exogenous Npy directly into the hippocampus of the stressed animals 1 h post-exposure. While the study did not examine genetic variation associated with NPY polymorphisms, the results are still worth noting due to homologous findings in humans.

CACNA1C has been identified as one of the strongest genetic risk factors for the development of affective disorders in both GWAS and candidate gene studies, [e.g., (18, 19, 247)]. Three of the 18 animal models examined this gene in mouse models using different approaches.

Terrillion et al. (248) used a subpopulation approach, classifying male wild-type or conditional Cacna1c knockout mice on a C57BL/6 background into resilient or susceptible based on their performance in a chronic social defeat test. They found expression of Cacna1c in the nucleus accumbens (NAc) is unchanged in resilient mice compared with control mice not exposed to social stress. Susceptible mice show a decrease in Cacna1c expression in the NAc after the social defeat paradigm, suggesting normal Cacna1c function in the NAc is crucial for resiliency to social stressors.

Dedic et al. (249) used a transgenic mouse model to assess the impact of Cacna1c deletion at different stages of development (See Supplementary Table 1 for genetic background and breeding timelines). If embryonic Cacna1c deletion in forebrain glutamatergic neurons occurs, then endophenotypes associated with psychiatric disorders manifest themselves. If the deletion of Cacna1c from forebrain glutamatergic neurons occurs in adulthood, the opposite effect occurs, inducing improved cognitive flexibility, strengthened synaptic plasticity, and increased stress resilience.

Michels et al. (250) examined a gene-stress interaction using immortalized mouse hippocampal cells, a well-established model system that reflects a common cellular response to environmental stress. In contrast with Terrillion's et al. (248) work, the authors show that reduced Cacna1c expression mediates neuroprotective effects against oxidative stress, predominantly at the level of mitochondria. This implies that reduced CACNA1C expression converges to control mitochondrial function, resulting in cellular resilience against oxidative stress. There is a growing body of evidence that major psychiatric illnesses are associated with impaired cellular resilience and synaptic dysfunction at various levels (251). These results suggest that alterations in mitochondrial function could affect critical cellular processes, dampening synaptic plasticity, and contributing to the pathophysiology of psychiatric disorders.

Decreased or increased CACNA1C expression levels have been associated with the SNP rs1006737, the same SNP investigated in studies included in this review. This, and Dedic's work, suggests that alterations in CACNA1C expression may be developmental-stage, brain-region stage, cell-type or sex-specific.

Manitt et al. used a transgenic mouse model to examine the netrin-1 guidance cue receptor gene, deleted in colorectal cancer (dcc), and resilience (252). The dcc haploinsufficient male mice were maintained on a BL/6 background and bred with wild-type BL/6 female mice. Using dcc loss-of-function transgenic mice, the study found reduced dcc expression confers resilience against developing neuroanatomical, neurochemical, and behavioral traits associated with mental disorders involving medial prefrontal cortex (mPFC) dysfunction. Significantly, these protective phenotypes are only present in adult but not pre-pubertal mice and most likely result from selective alterations in the reorganization of mPFC dopamine circuitry. One human study found in this review examined how DCC contributes to the risk of schizophrenia (202). The authors found that individuals heterozygous at one SNP, rs2270954, had decreased risk of schizophrenia. Further studies have demonstrated that microRNA regulation of DCC by miR-218, a posttranscriptional repressor of DCC, maybe a switch of susceptibility vs. resilience to stress-related disorders (253). Together, these studies suggest that DCC could be a promising candidate gene that contributes to the genetic basis of resilience.

Three studies investigated transgenic mouse models of Fkbp5, a co-chaperone protein of the Hsp90 complex that regulates the glucocorticoid receptor. FKBP5 SNPs have been associated with psychiatric disorders including depression, PTSD, bipolar disorder, and anxiety (254). FKBP51, encoded by the FKBP5 gene, and FKBP52, encoded by the FKBP4 gene, are components of the chaperone-receptor heterocomplex and differentially regulate the glucocorticoid receptors or mineralocorticoid receptors. Hartmann et al. (255) characterized heterozygous Fkbp52 knockout mice kept on a mixed 129SvJ × C57BL/6 background, under basal and chronic social defeat stress conditions (CSDS). Fkbp52 heterozygous knockout mice demonstrated stress resilience, such as reduced basal corticosterone levels and more active stress-coping behavior in the forced swim test (FST) following CSDS. This mimics the phenotype of Fkbp51 deficient mice in the FST, as found by O'Leary et al. (254). O'Leary et al. tested how Fkbp5 deletion in aged mice affects behavior, given that FKBP51 levels increase with age. They found that aged Fkbp51 deficient mice displayed enhanced active stress-coping behavior following the acute severe stress exposure of the FST, similar to levels of control mice treated with antidepressants. This antidepressant behavior in the Fkbp5^−/− did not affect cognition or other basic motor functions, and reduced corticosterone levels following stress were also observed.

Kwon et al. (256) exposed wild-type and Fkbp5 knock-out mice to chronic restraint stress. They found Fkbp5-deficient mice were resilient to the depressive-like behavior exhibited by wild-type mice following chronic restraint stress. Further RNA sequencing analyses revealed a module of co-expressed genes, M55, that was downregulated in the WT group and restored in the KO group. Gene ontology enrichment analysis of M55 revealed biological functions involved with inflammatory response, gland morphogenesis and nervous system development.

Together, these findings suggest that absence of FKBP5 and regulation of the immune response results in a decrease of corticosterone levels and HPA-axis activity, which increases resilience to depressive phenotypes.

Depression has been reported to be more common in patients with Parkinson's disease (PD) than in the general population, often appearing earlier than the onset of motor symptoms (257). The G2019S mutation in LRRK2 causes late-onset PD, but it is unclear how this mutation impacts depression-related behaviors (258). Matikainen-Ankney et al. (258) addressed this by subjecting Lrrk2-G2019S knock-in and wild-type mice to CSDS. Mice used in the study were congenic on C57BL/6NTac background. The study showed that young adult G2019S knock-in mice were highly resilient to CSDS, failing to exhibit social avoidance compared to wild-type mice. There were no behavioral differences between the genotypes found in the absence of CSDS.

Given the well-documented research on the APOE ε4 genotype association with genetic risk for Alzheimer's Disease (AD), and human imaging studies suggesting APOE ε4 is associated with brain structures linked to cognitive decline, Kulkarni et al. (259) assessed brain structure and function in APOE ε4 knock-in rats. Male and female WT and human APOE ε4 knock in Sprague Dawley rats were studied for changes in brain structure and behavioral tests. There was evidence of sex differences in brain structure and function in the APOE ε4 knock in rats with the data suggesting female carriers are more resilient to cognitive/emotional problems possibly due to altered brain volumes and enhanced connectivity.

Longevity was used as a resilience phenotype in three studies (226–228) and has also been previously associated with individuals with protective variants, buffering the effect of disease-associated variants and adverse environmental factors, decreasing morbidity at the end of extreme longevity (29–32). A recent review and meta-analysis confirmed the association of five polymorphisms that were significantly associated with exceptional longevity: ACE rs4340, APOE ε2/3/4, FOXO3A rs2802292, the KLOTHO KL-VS variant and IL6 rs1800795 (281). Of these five polymorphisms, three had genetic associations with resilience within our review, with resilience conceptualized as decreased sensitivity to trauma (143), neuronal resilience to AD (137), epilepsy (136), elevated serum klotho levels that promote a resilient brain (218), and desensitized physiology to pro-inflammatory effects (101, 151).

Sense of coherence, referring to a consistent but dynamic feeling of confidence, defined as the belief that life is comprehensive, manageable and meaningful, was operationalised as a resilience phenotype in three studies (121–123). The first GWA analysis recently conducted for SOC yielded a significant association, rs74920024 on chromosome 2 (282). The association locus is close to Glycerol-3-Phosphate Dehydrogenase 2 (GPD2), a gene encoding a protein that localizes to the inner mitochondrial membrane and is responsible for the conversion of glycerol-3-phosphate to dihydroxyacetone phosphate. While there is no overlap of this gene with candidate genes found in this review, GPD2 has been previously associated with large GWAS's on cognitive ability (283) and educational attainment from the Social Science Genetic Consortium 2018 GWAS (267), and cognitive aspects of educational attainment from a large GWAS meta-analysis combining a cognitive consortium and human intelligence GWAS (284). While these traits have not been used as a proxy for resilience within this review, they have wide support as individual level resilience factors [e.g., (285–287)], with high cognitive skills in childhood predicting resilience to adversity in adolescence (288).

Two studies within this review operationalised resilience as stress coping (129, 130). Çarkaxhiu Bulut et al. (130), evaluated the SLC64A 5-HTTLPR polymorphism and resilience components on the development of psychopathology in adolescent sexual abuse cases, while Aizawa et al. (129) examined genetic association of BDNF and variation in stress-coping style. A recent systematic review of genetic influences on coping (289) found multiple associations between SLC64A and Adrenoceptor Beta 2 (ADRB2) and coping phenotypes. One important finding of their review that echoes the sentiments within our own review, is the lack of consensus regarding instruments for coping phenotypes. Again, despite the lack of consensus of instruments within the reviews, there is overlap on the candidate genes associated with coping and resiliency phenotypes. A study included in this review found a novel association of a SNP within the promoter region of the ADRB2 gene that was associated with relative resilience to childhood adversity (199). Previous research has revealed links between ADRB2 and resting blood pressure (BP) and regulation of BP in stressful situations (290, 291). This genetic association of ADRB2 found in both coping and resiliency phenotypes is interesting given the known links between stress and risk factors for cardiovascular disease.

Gervai et al. (135) examined the link between DRD4 gene polymorphism and infant attachment and found that not carrying the T.7 haplotype of the DRD4 acts as a resilience factor in the optimal development of early attachment. Within a recent systematic review of gene-environment interactions on infant and child attachment (292), DRD4 was associated with the interaction between maternal sensitivity and parenting style in five studies, with contradictory results. While there was also a lack of consensus of which DRD4 allele was associated with resilience in the studies meeting our selection criteria in this review, later research found the DRD4 7-repeat allele to be protective against the risk of a dysfunction in emotional regulation associated with birthweight (293). This is consistent with Das's et al. (108) finding but contradictory to Gervai's et al. (135) and Cicchetti and Rogosch's (144). While these associations are suggestive of reoccurring patterns of associations, ultimately, there is insufficient evidence to meaningfully comment on the direct influence of DRD4 on attachment or resilience, a position reiterated in a later systematic review of the genetics of adult attachment (294).

While the conceptualization of resilience in the literature is highly variable, the relationship between PTG and resilience is not clear. Some researchers equate resilience with PTG (295), while others assume it is a form of resilience (296). Only one study included in this review met the inclusion criteria (127) and examined the association between PTG, PTSD and genetic variants. The study found a GxE association between PTG and Regulator of G Protein Signaling 2 (RGS2) (rs4606), driven by disaster exposure. However, a recent systematic review found no gene expression changes associated with PTG (297), again highlighting a focus on pathogenic research, which fails to adequately measure or acknowledge positive post-trauma outcomes, whether they be conceptualized at resilience or PTG.