Psychiatric risk and resilience: Plasticity genes and positive mental health

Abstract Objective The at‐risk mental state (ARMS) for psychosis has long played a key role in diathesis‐stress models of schizophrenia. More recent studies, however, have called for extending the boundaries of the ARMS construct beyond attenuated psychosis in nonhelp‐seeking samples to include not only other vulnerability indicators but also protective factors related to genotype, mental health, personality, and cognition. Method Accordingly, we assessed in a sample of 100 college students, the ARMS construct with the Brief Prodromal Questionnaire (PQ‐B) for psychosis, in conjunction with measures of positive mental health, childhood adversity, psychiatric symptoms, personality traits, social cognition, and genetic variables derived from assays of the serotonin transporter (5‐HTTLPR) and the brain‐derived neurotrophic factor (BDNF). Results Higher PQ‐B scores correlated positively with vulnerability indicators of childhood adversity and heightened levels of a wide variety of psychiatric symptoms but correlated negatively with protective factors of better overall mental health, social cognition as well as with a distinct NEO profile marked by reduced neuroticism and elevated agreeableness and conscientiousness. Multivariate analyses indicated that a composite ARMS measure comprised of PQ‐B scores plus anxiety and depression symptoms revealed significant genotype differences, with lowest risk and highest resilience for allelic carriers of 5‐HTTLPR‐short and BDNF Met polymorphisms. Conclusions Results provided support for extending the ARMS construct, pointing to important contributions of personality, social cognition, and genes that support neural plasticity in mitigating vulnerability and enhancing resilience and well‐being.


| INTRODUC TI ON
The concept of risk or vulnerability has long played a critical role in both mental health research and practice. Its historical roots may be traced to the early work of Harry Stack Sullivan (1927) who first described the prodromal state of schizophrenia as characterized by subclinical symptoms and signs of psychosis (Kirch et al., 1992). The subsequent identification of the so-called schizophrenia prodrome held the promise of providing a window for targeted treatment and prevention, an invaluable opportunity for therapeutic intervention during a presumed critical period that too often had been viewed solely as an initial step in an inevitable disease progression. No doubt inspired by the potential clinical impact of these early efforts, more contemporary studies of the past 15 years have developed rigorous measures of the schizophrenia prodrome, now cast as an at-risk mental state (ARMS) that can be reliably and validly assessed with structured interview, self-report, or combination thereof (Fusar-Poli et al., 2012) In fact, meta-analyses of longitudinal studies showed, independent of risk instrument, individuals assessed as at-risk had transition rates to psychosis of 18% after 6-month follow-up, 22% after 1 year, 29% after 2 years, and 36% after 3 years (Fusar-Poli et al., 2012).
Recent studies have called for extending the boundaries of the ARMS construct beyond attenuated psychosis to include other psychiatric symptoms such as depression and anxiety as well as other measures of stress and adversity in nonhelp-seeking samples (Lee et al., 2018;Linscott & Van Os, 2013;van Os, & Murray, 2013;van Os & Reininghaus, 2016). From this transdiagnostic perspective, ARMS may be viewed as a dynamic and malleable condition, which might remit, persist without worsening, or diverge along several different illness trajectories (McGorry et al., 2018;McGorry & Nelson, 2016). While the factors that may influence etiology and ultimate outcome or resolution remain an active area of research, risk models have traditionally emphasized a diathesis-stress framework in which vulnerability for developing a disorder can emanate from biological, psychological, or sociocultural sources triggered by adverse environmental events or experiences perceived as overwhelming one's personal resources (e.g., Hooley et al., 2017;Nolen-Hoeksema & Watkins, 2011). For example, studies have shown that individuals with a particular genetic diathesis when exposed to environmental adversity are more likely to develop major depression than those who experience the same level of major life stressors but without the genetic vulnerability (Caspi et al., 2003).
More recent formulations, however, have provided a new perspective on risk, one less psychopathological in focus and more resonant with positive psychology, examining what may be viewed as the "bright side" of these gene-environment interactions (Ellis & Boyce, 2011). Known as the differential-susceptibility hypothesis, it proposes that the very same characteristics such as a "risky" genotype or a "reactive" temperament that make individuals disproportionately vulnerable to stress also make them disproportionately more likely to benefit from positive experiences and environmental supports (Belsky & Hartman, 2014). Here, vulnerability genes are reconceptualized as plasticity alleles that confer sensitivity to both positive and negative experiences and environments. Unknown, however, is whether such variation in plasticity or malleability "for better and for worse" contributes to individual differences in mental health in mitigating or protecting against clinical risk.
Accordingly, adopting a positive psychology framework (see e.g., Layous et al., 2014), we examine the concept of ARMS through the lens of plasticity, broadly defined, in relation to genotype, mental health, personality, and neuropsychological functioning. We targeted specific allelic variants arising from a priori selected, single nucleotide polymorphisms (SNPs) in two candidate plasticity genes, serotonin (5-HTTLPR) and neurotrophic factor (BDNF), each linked to individual differences in emotionality, stress, cognition, and personality (Belsky & Hartman, 2014). In so doing, single and polygenic effects of these alleles on ARMS can be directly tested in conjunction with measures of childhood stress and current levels of psychiatric vulnerability, positive mental health, personality traits, and neuropsychological functioning. Together, these psychological and biological measures provide a multimethod design to test the unifying hypothesis that reduced ARMS may be expressed genetically, by increased plasticity alleles, and, behaviorally, by higher mental health, specific personality profiles and better neuropsychological performance.

| Participants
One hundred participants, recruited from the greater Boston area, primarily at the University of Massachusetts, Boston (UMB) were between the ages of 18 and 25 (M = 21.22 years, SD = 1.99) and identified as English speaking for at least 5 years prior to study enrollment. Seventy percent of participants identified as biologically female, 42% racially identified as White, 72% reported the United States of America as their country of origin, and 63% endorsed 1-3 years of college as their level of education. The Institutional Review Board at University of Massachusetts Boston approved all research study procedures. Consenting participants completed selfreport measures and neuropsychological tests, and then provided a DNA sample via a cheek swab for the assaying of genotypes.
Participants were compensated $25 for their time or received extra credit in psychology courses.

| Self-report measures
Prodromal Questionnaire-Brief (PQ-B). The PQ-B is a 21-item selfreport questionnaire designed to assess the presence or absence of psychosis-risk syndromes (Loewy et al., 2011). Based on the items from the Structured Interview for Prodromal Syndromes (McGlashan et al., 2001), the PQ-B assesses positive symptoms of psychosis experienced in the past month PQ-B has been shown to be an effective and efficient instrument for screening psychosis risk across a wide variety of samples and settings, including Chinese-speaking mental health referrals (Xu et al., 2016), Nigerian secondary school students (Okewole et al., 2015), Dutch-speaking young adults ages 18-35 recruited from a general help-seeking population (Ising et al., 2012), and male prisoners (Jarrett et al., 2015).
Brief Symptom Inventory (BSI). The BSI is a 53-item scale that measures psychiatric symptoms status across nine distinct domains: Somatization, Obsessive-Compulsive, Interpersonal Sensitivity, Anxiety, Hostility, Depression, Paranoid Ideation, Psychoticism, and Phobic Anxiety (Derogatis & Spencer, 1982). The BSI also includes measures of overall Global Severity Index (GSI) and a Positive Symptom Distress Index (PSDI). BSI scores reflect current psychiatric status on a Likert scale ranging from 0 (not at all) to 4 (extremely). The BSI has demonstrated good internal consistency among nonpsychiatric populations (0.71-0.85 across scales) and moderate to high test-retest reliability (0.68-0.91 across scales) and convergent and discriminant validity with the Minnesota Multiphasic Personality Inventory (MMPI) (Derogatis & Melisaratos, 1983;Derogatis & Spencer, 1982).
Adverse childhood experiences (ACE). This scale is a 10-item measure that assesses eight categories of adverse experiences in childhood, including emotional, physical, and sexual abuse, and household dysfunction (i.e., substance abuse, mental illness, mother treated violently, and incarcerated household member). Participants are asked to provide "Yes" or "No" responses to each of the 10 items.
Scores range from 0 to 10, with higher scores indicative of greater number of adverse events in childhood (Anda et al., 2006). Revised NEO personality test (NEO-PI-R). The NEO-PI-R is an objective, self-report measure of five distinct and presumably universal personality traits: neuroticism, extraversion, openness, agreeableness, and conscientiousness (Costa & McCrae, 1992). DeYoung et al. (2010) reported alpha reliabilities of internal consistency for the five NEO trait scales as 0.92 for Neuroticism; 0.87 for Extraversion; 0.89 for Openness; 0.91 for Agreeableness; and 0.91 for Conscientiousness.

| Neuropsychological measures
Reading the mind's eyes test (RMET; Baron-Cohen et al., 1997), used as a computerized task of social cognition, specifically Theory of Mind, consists of 36 trials, presented in randomized order, with each trial displaying a black-and-white photograph of the eyes, eyebrows, and bridge of the nose of a White male or female individual making a facial expression, and below the photograph are four adjectives that describe a complex emotion (e.g., reflective, aghast, irritated, and impatient). Test-retest reliability has been reported in nonclinical populations to range from 0.7 to 0.8 (Hallerbäck et al., 2009)

| DNA collection and extraction
Cytobrush swabs (Coopersurgical Inc.) were used to collect buccal cells. Participants were instructed to brush the swab 30 times against the inside of their cheek while slowly rotating the swab.
Swabs were immediately placed on ice and stored at −80 degrees C until DNA extraction. Buccal samples were extracted using a Zymo Quick DNA Universal Kit per the manufacturer's instructions (Zymo Research). DNA yield from buccal samples ranged from 0.48μg to 14.4μg of DNA. Extracted DNA was stored in molecular biologygrade water at −80°C until genotyping analysis.

| 5-HTTLPR genotyping
Genotyping for 5-HTTLPR polymorphisms was performed using polymerase chain reaction and resolution using gel electrophoresis

| BDNF genotyping
TaqMan SNP genotyping was used to determine BDNF val66met genotype (rs6265). 25 μl PCR reactions were performed using a predesigned 1X Taqman allelic discrimination assay (assay number: C__11592758_10; Applied Biosystems), containing forward and reverse primers and allele-specific probe with 5ng of sample DNA. Genotypic amplification was achieved using the StepOne Plus Real-Time (Applied Biosystems) PCR System with programming as follows: 95°C for 10m, followed by 42 cycles of 95°C for 15 s and 60°C for 1m. Genotype was determined from the resulting allelic discrimination plot.

| Statistical analysis
For the behavioral measures, Pearson correlations tested for associations of PQ-B scores with ACE, NEO, MHC-SF, RMET, and cognitive neuropsychological tests. We then performed to a mixed-model 2 × 5 ANOVA with PQ-B group as the betweensubjects factor, median split in two levels (low, high), and NEO personality traits as the within-subjects factor with five levels  Table 1). Table 2 presents the correlations of PQ-B scores with behavioral measures. As seen in Table 2 Thus, a mixed-model ANOVA with PQ-B group as the betweensubjects factor with two levels (low, high) and NEO personality traits as the within-subjects factor with five levels (neuroticism, extraversion, openness, agreeableness, and conscientiousness) revealed a highly significant PQ-B group × personality interaction, As shown in Table 1, PQ-B scores did not differ for 5-HTTLPR To examine the effects of these two candidate genes simultaneously on ARMS, we submitted PQ-B total symptoms scores to a 2 × 2 ANOVA with two between-subjects factors of serotonin (long, short) and BDNF (Val/Val, Met). The 2 × 2 ANOVA revealed a statistically significant interaction of BDNF × serotonin, F(1, 96) = 3.82, p = .05, partial eta squared = 0.038. As shown in Figure 1

| D ISCUSS I ON
The current study investigated the ARMS construct broadly, focus- have also reported similar associations between psychotic risk measures and symptoms and cognitive changes (Karcher et al., 2018;Kelleher et al., 2012;Laurens et al., 2007). Consistent with these population-based investigations, the current findings revealed strong associations between PQ-B scores and increased ratings for both psychotic and nonpsychotic symptoms. The latter relationship conforms to prior research that has emphasized the importance of including nonpsychotic affective symptoms of depression and anxiety in risk formulation as well as measures of childhood adversity in nonhelp-seeking samples (e.g., Linscott & Van Os, 2013;Yamasaki et al., 2018).
While providing strong support for extending the clinical dimensions of ARMs, the present investigation also provided clear evi- of a key aspect of social cognition, known as the theory of mind. The theory of mind reflects a specific set of abilities that are distinct from cognitive intelligence and essential for the development of effective social interaction and communication (e.g., Heyes & Firth, 2014), and frequently lowered in at-risk individuals (Fusar-Poli et al., 2012). As such, these data suggested that well-developed social cognitive abilities may reduce risk and enhance mental health.
These behavioral data also conformed well to the differentialsusceptibility hypothesis that specific serotonin and BDNF genes confer plasticity, thereby allowing carriers of these alleles greater sensitivity to environmental factors, whether negative or positive, as well as to a wider range of developmental outcomes. Of particular relevance is the idea that more plasticity alleles, the greater the likelihood of benefit from positive experiences and events. Indeed, neural plasticity is defined as a fundamental property of the brain that allows it to change in response to external input, learning, and training (Forsyth & Lewis, 2017). Here, we focused on the serotonin short and BDNF met allelic pair as plasticity genes with our results In line with the differential-susceptibility hypothesis, such genetic variation may mediate the development of emotional processes that reduce risk and contribute to adaptive mental health. However, this hypothesis also emphasizes the importance of positive environmental influences in triggering plasticity genes. In this regard, the current study did not include an index of positive childhood experiences.
And while carriers of the serotonin short and BDNF met allelic pair had lowest exposure to childhood adversity, whether they also had increased exposure to positive environmental experiences and events is unknown, and would need to be independently assessed in future studies of large representative samples.
Clinically, the current findings align well with recent treatment studies focusing on the enhancement of healthy personality either via psychopharmacology (e.g., Ilieva, 2015;Knutson et al., 1998) or psychological therapy (Armstrong & Rimes, 2016  also key contributors to risk and resilience.

| CON CLUS IONS
In summary, the results provided support for not only extending the ARMS to include other psychiatric symptoms namely anxiety and depression, but also pointed to a number of protective factors that may reduce vulnerability and enhance mental health. Specifically, a personality configuration characterized by low neuroticism and elevated extraversion, agreeableness, and conscientiousness along with higher social cognitive abilities and greater positive mental health, as well as a particular set of BDNF and 5HTTLPR plasticity alleles may increase both illness resistance and well-being.

This work was supported by the University of Massachusetts
Boston Department of Psychology. The authors declare no conflict of interest.

AUTH O R CO NTR I B UTI O N
PGN, VCH, KOD, and RH contributed to the conception and design of this study. KOD, VCH, SBB, and HEL collected and organized the data. All authors analyzed the data, contributed to the drafting and editing of the manuscript and read and approved the final manuscript.

PEER R E V I E W
The peer review history for this article is available at https://publo ns.com/publo n/10.1002/brb3.2137.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.