SCHRES-05838; No of Pages 7 Schizophrenia Research xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Schizophrenia Research journal homepage: www.elsevier.com/locate/schres

Genetic liability, prenatal health, stress and family environment: Risk factors in the Harvard Adolescent Family High Risk for Schizophrenia Study Deborah J. Walder a,b,c,⁎, Stephen V. Faraone d,e, Stephen J. Glatt d,e, Ming T. Tsuang f,g, Larry J. Seidman c,h,⁎⁎ a

Brooklyn College, Department of Psychology, United States The Graduate Center of The City University of New York (CUNY), United States c Harvard Medical School, Department of Psychiatry at Beth Israel Deaconess Medical Center, United States d SUNY Upstate Medical University, Department of Psychiatry and Behavioral Sciences, United States e SUNY Upstate Medical University, Center for Neuropsychiatric Genetics, Biomedical Sciences Program, Neuroscience and Physiology, United States f Center for Behavioral Genomics and Institute of Genomic Medicine, Department of Psychiatry at University of California — San Diego, United States g Harvard Institute of Psychiatric Epidemiology and Genetics, Harvard School of Public Health, United States h Harvard Medical School, Department of Psychiatry at Massachusetts General Hospital, United States b

a r t i c l e

i n f o

Article history: Received 8 January 2014 Received in revised form 6 April 2014 Accepted 11 April 2014 Available online xxxx Keywords: Stress Psychosis Relatives Obstetric complications Neurodevelopment Family environment

a b s t r a c t Objectives: The familial (“genetic”) high-risk (FHR) paradigm enables assessment of individuals at risk for schizophrenia based on a positive family history of schizophrenia in first-degree, biological relatives. This strategy presumes genetic transmission of abnormal traits given high heritability of the illness. It is plausible, however, that adverse environmental factors are also transmitted in these families. Few studies have evaluated both biological and environmental factors within a FHR study of adolescents. Methods: We conceptualize four precursors to psychosis pathogenesis: two biological (genetic predisposition, prenatal health issues (PHIs)) and two environmental (family environment, stressful life events (SLEs)). Participants assessed between 1998 and 2007 (ages 13–25) included 40 (20F/20M) adolescents at FHR for schizophrenia (FHRs) and 55 (31F/24M) community controls. ‘Genetic load’ indexed number of affected family members relative to pedigree size. Results: PHI was significantly greater among FHRs, and family cohesion and expressiveness were less (and family conflict was higher) among FHRs; however, groups did not significantly differ in SLE indices. Among FHRs, genetic liability was significantly associated with PHI and family expressiveness. Conclusions: Prenatal and family environmental disruptions are elevated in families with a first-degree relative with schizophrenia. Findings support our proposed ‘polygenic neurodevelopmental diathesis–stress model’ whereby psychosis susceptibility (and resilience) involves the independent and synergistic confluence of (temporally-sensitive) biological and environmental factors across development. Recognition of biological and social environmental influences across critical developmental periods points to key issues relevant for enhanced identification of psychosis susceptibility, facilitation of more precise models of illness risk, and development of novel prevention strategies. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Over the last few decades it has become firmly established that schizophrenia has early neurodevelopmental origins (Lewis and Murray, 1987; Weinberger, 1987) that later manifest in illness expression through ⁎ Correspondence to: D.J. Walder, Department of Psychology, Rm 5315 James Hall, Brooklyn College, 2900 Bedford Avenue, Brooklyn, NY 11210, United States. Tel.: +1 718 951 5000; fax: +1 718 951 4814. ⁎⁎ Correspondence to: L.J. Seidman, Massachusetts Mental Health Center, 75 Fenwood Road, Boston, MA 02115, United States. Tel.: +1 617 754 1238; fax: +1 617 754 1250. E-mail addresses: [email protected] (D.J. Walder), [email protected] (L.J. Seidman).

disruptions of normal neuromaturational processes (Walker and Bollini, 2002). Biological susceptibility is reflected in 1) behavioral (family, twin, adoption) genetic studies yielding heritability estimates of approximately .65–.70 (Gottesman and Shields, 1967), confirmed by national population-based and registry studies in Denmark (Wray and Gottesman, 2012) and Sweden (Lichtenstein et al., 2009) and 2) elevated rates of perinatal complications in schizophrenia (Cannon, Jones et al., 2002; Cannon, van Erp et al., 2002). Increasingly, molecular genetic origins are being tested with large-scale consortia (Cross-Disorder Group of the Psychiatric Genomics Consortium, 2013), pointing to complex polygenic influences involving many common single nucleotide variants and rare events such as copy number variants. Perinatal complications

http://dx.doi.org/10.1016/j.schres.2014.04.015 0920-9964/© 2014 Elsevier B.V. All rights reserved.

Please cite this article as: Walder, D.J., et al., Genetic liability, prenatal health, stress and family environment: Risk factors in the Harvard Adolescent Family High Risk for Schizophrenia Study, Schizophr. Res. (2014), http://dx.doi.org/10.1016/j.schres.2014.04.015

2

D.J. Walder et al. / Schizophrenia Research xxx (2014) xxx–xxx

and genetics represent two important risk domains given that they exert effects early, impacting brain development. Robust evidence of structural and functional brain abnormalities in nonpsychotic, biological relatives between 8 and 30 years of age (Thermenos et al., 2013) supports the notion that disrupted neurodevelopment precedes onset of frank psychosis. For example, gray matter volume abnormalities exist in youth at familial high-risk (FHR) compared to controls (Rosso et al., 2010), with greater volume reduction over time associated with increasing symptoms and cognitive deficits in those who develop schizophrenia (McIntosh et al., 2011). Prefrontal cortex alterations and smaller hippocampal volume are the most consistently reported neuroimaging findings in FHR youth, observed in pre-teen, teenage and adult relatives (Boos et al., 2007; Thermenos et al., 2013). In contrast to neurobiological studies of schizophrenia patients and their relatives in family studies, relatively less attention has been paid to environmental influences, particularly the social environment. Environmental factors are emphasized in contemporary conceptualizations of schizophrenia, most prominently in the ‘diathesis–stress’ model (Zubin and Spring, 1977). Accordingly, biological vulnerability presumably interacts with environmental risk toward precipitating psychosis (Tsuang, 2000). Despite high heritability, concordance for schizophrenia in monozygotic twins is only around 0.50 (Cardno and Gottesman, 2000). This phenotypic discordance implicates environmental factors, which are important because they are likely more malleable than genetic risk factors, particularly in the context of new approaches to early intervention and prevention strategies for psychosis. Two high-risk paradigms have evolved to identify precursors of psychosis. The clinical (or ultra) high-risk paradigm involves ascertainment of youth with subclinical psychotic symptoms. The FHR approach selects nonpsychotic biological relatives to assess liabilities expressed across a range of phenotypes presumably reflecting vulnerability. Hallmark phenotypes (e.g., odd thinking, smaller hippocampi, stress sensitivity) can be studied at different ages in FHR studies to evaluate developmental effects, and in different subpopulations (higher vs. lower genetic loading) to study subgroup expression. The latter approach captures an important proportion of individuals at heightened risk while avoiding confounds associated with illness and assumes a cumulative, non-specific, polygenic liability of genetic and environmental risk factors. Previously, we demonstrated that compared to controls, Harvard Adolescent FHR youth have neurocognitive difficulties (Seidman et al., 2006; Phillips et al., 2011; Seidman et al., 2012; Scala et al., 2013), more physical anhedonia (but not magical ideation or perceptual

aberration) and more social difficulties and reward dependence (Glatt et al., 2006; Rosso et al., 2010), the latter of which were associated with higher genetic loading (Glatt et al., 2006). We did not report on key environmental variables that may influence these outcomes, such as perinatal health issues and later life stressors. In the present paper, we propose a ‘polygenic neurodevelopmental diathesis–stress model’ that targets four early developmental perturbations demonstrated to play a role in psychosis vulnerability in a temporally-sensitive manner, not previously examined together in a FHR context. We examine two classes of biological precursors (genetic predisposition/loading; prenatal health issues (PHIs)) and two classes of social–environmental factors (family environment; stressful life events (SLEs)) (see Fig. 1). Regarding biological precursors, first, prevailing genetic hypotheses utilize polygenic models wherein many susceptibility genes of small effect (and a few rare genes with larger effects), rather than single major genes, predispose to schizophrenia (Gottesman and Shields, 1967). We utilize a proxy measure of genetic loading (Glatt et al., 2006) to approximate polygenic liability. Second, obstetric complications are one of the strongest predictors of psychosis risk. Evidence indicates higher rates of adverse prenatal events across the psychosis spectrum, such as prenatal maternal viral exposure, malnutrition, stress, and complications of pregnancy and delivery (see Cannon, Jones et al., 2002; Cannon, van Erp et al., 2002; Walder et al., 2012). Surprisingly, we are aware of only one FHR study that evaluated PHI (Gilbert et al., 2003); accordingly, high-risk offspring (compared to controls) had a higher frequency of birth complications. Stressful life events occurring during development are strongly implicated in psychosis risk. Literature demonstrates 1) relationships among major life events, daily stressors and symptomatology in schizophrenia (Norman and Malla, 1993) and 2) social environmental context modulates impact of stressful life events (Ventura et al., 1989). Undesirable life events are linked with prodromal symptoms, and daily stressors predict increased positive prodromal symptoms (Tessner et al., 2011). Strikingly few studies have examined the influence of stressful life events among youth at FHR for psychosis (Binbay et al., 2012). The one study we are aware of found that social disadvantage increases risk more for FHR offspring than non-risk offspring (Wicks et al., 2010). Finally, family environment plays a pivotal role in psychosis. Negative family environment contributes to poor prognosis (Myin-Germeys et al., 2001) and increases risk independent of family history of psychosis (González-Pinto et al., 2011). Patient exposure to hostile, critical and emotionally over-involved attitudes by relatives (Lukoff et al., 1984)

Fig. 1. Polygenic neurodevelopmental model.

Please cite this article as: Walder, D.J., et al., Genetic liability, prenatal health, stress and family environment: Risk factors in the Harvard Adolescent Family High Risk for Schizophrenia Study, Schizophr. Res. (2014), http://dx.doi.org/10.1016/j.schres.2014.04.015

D.J. Walder et al. / Schizophrenia Research xxx (2014) xxx–xxx

and high expressed emotion families (Butzlaff and Hooley, 1998) are associated with relapse; whereas, low expressed emotion mitigates effects of stressful events (Nuechterlein et al., 1994). Family intervention minimizing expressed emotion prevents relapse (Leff et al., 1982) and family interaction style combined with patient symptoms predicts symptom relapse (Levene et al., 2009). One unique study found that patients managed stressful life events better when perceiving their families as higher in cohesion, expressiveness, independence and organization, and lower in conflict (Gretchen-Doorly et al., 2011). Lastly, positive family environment is protective among individuals with a family history of psychosis (González-Pinto et al., 2011) and predicts symptom reduction and increased social functioning among high risk adolescents (O'Brien et al., 2006), though research on FHR adolescents is scarce. In the present study we hypothesized that, compared to community controls, FHRs would demonstrate 1) greater PHI and SLEs, 2) greater family conflict and less family cohesion, 3) differences in expressiveness and 4) a significant relationship of genetic liability with early (PHI) and later (SLE; family environment) developmental factors.

2. Methods 2.1. Sample The current sample was ascertained as part of the Harvard Adolescent FHR study between 1998 and 2007, described previously in detail (Glatt et al., 2006; Seidman et al., 2012). The analyses herein are novel, as are group comparisons on PHI, FE and SLE. Participants 13–25 years of age consisted of two groups; biological offspring and siblings of schizophrenia probands (FHRs), and a community control (CC) group who were the biological offspring and siblings of control probands. FHRs included 40% offspring and 60% siblings of 31 families with adult probands (at least 18 years of age) who met the DSM-IV criteria (American Psychiatric Association, 1994) for schizophrenia (n = 25) and schizoaffective, depressed type (n = 6). There were 11 families with parent–offspring data. Ten of these 11 families had 1 proband parent; 6 fathers and 4 mothers. One of these 11 families had 2 proband parents; 1 father and 1 mother. The CCs from 35 families consisted of children of parents diagnosed according to the DSM-IV criteria with no mental illness (n = 25), major depressive disorder (n = 8), mood disorder due to a general medical condition (n = 1), or cannabis abuse (n = 1), using the Diagnostic Interview for Genetic Studies (DIGS; Nurnberger, 1994) and Family Interview for Genetic Studies (FIGS; Maxwell, 1992). FHR and CC groups were comparable on sex distribution, educational level and ethnic

3

composition. FHRs were significantly older and of lower SES than CCs (see Table 1). Participants were excluded if they had any lifetime history of psychotic illness, substance dependence, neurological disease, head injury or medical illness with documented cognitive sequelae, sensory impairments, current psychotropic medication use, or a full-scale IQ estimate less than 70 based on eight sub-tests of the third editions of the Wechsler Intelligence Scale for Children (Wechsler, 1991) or Wechsler Adult Intelligence Scale (Wechsler, 1997). CCs were additionally excluded if they had any first- or second-degree biological relative with lifetime history of a psychotic disorder. 2.2. Measures and procedures After probands gave consent, their children and siblings were contacted to determine eligibility and willingness to participate. Participants age 18 years and older gave informed consent. Subjects younger than 18 years gave assent in conjunction with parental informed consent. Subjects received payment for participation. The study was approved by the human research committees of Massachusetts Mental Health Center, Massachusetts General Hospital and Harvard Medical School. 2.2.1. Diagnostic assessment Probands were administered several measures to screen for psychosis, substance use, mood disturbance and other inclusion and exclusion criteria. Measures included the Psychosis, Substance Abuse and Mood Disorders modules of the Washington University Kiddie Schedule for Affective Disorders and Schizophrenia (Geller et al., 1994), the DIGS, and the Neurodevelopmental Questionnaire (Faraone et al., 1995). For all measures, higher scores reflect greater risk, more severe symptomatology or poorer functioning, except Youth Self Report — Competence for Activities scale upon which a higher score reflects better functioning. 2.2.2. Measures of early biological and environmental susceptibility The FIGS was administered to parents to assess family history of psychiatric illness and to derive a genetic liability index. The ‘genetic load’ index (or incremental degree of presumed genetic risk) was derived based on the ‘allele-sharing’ method to compute the relative proportion of alleles individuals are expected to share with their affected biological relatives versus unaffected biological relatives, while accounting for the overall pedigree size. Similar to relative risk and genetic liability methods, this method assumes a tight correspondence between traits under study and risk genes for schizophrenia. Values of genetic loading using this model ranged from 0 to 1, with higher values reflecting greater genetic loading (Glatt et al., 2006).

Table 1 Demographic characteristics of the control and familial high risk groups. Group

Sex (N) Male Female Mean age (SD) Education SESa Ethnicity (N) Caucasian African American Hispanic, Caucasian Hispanic, Black Asian Portuguese, Cape Verdean

Group difference

Community control (n = 55)

Familial high risk (n = 40)

24 31 17.2 (3.7) 11.0 (3.3) 48.1 (15.5)

20 20 19.4 (3.9) 11.4 (2.7) 38.0 (16.4)

34 6 9 5 1 0

23 7 8 0 1 1

X2(1) = .5, p = .68 t(93) = −2.9, p = .005** t(93) = −.6, p = .55 t(90) = 3.0, p = .003** X2(5) = 6.0, p = .30

SES = socioeconomic status based on Hollingshead, 1975. **p b .01, *p b .05; two-tailed tests. a n = 92.

Please cite this article as: Walder, D.J., et al., Genetic liability, prenatal health, stress and family environment: Risk factors in the Harvard Adolescent Family High Risk for Schizophrenia Study, Schizophr. Res. (2014), http://dx.doi.org/10.1016/j.schres.2014.04.015

4

D.J. Walder et al. / Schizophrenia Research xxx (2014) xxx–xxx

The Pregnancy History Instrument — Revised (Buka et al., 2004) is a brief structured interview administered to mothers designed to obtain pregnancy and neonatal history. Items cover events related to maternal history, pregnancy, delivery and the neonatal period. An index of prenatal health issues (PHIs) was calculated by summing all endorsed items corresponding to the prenatal period. PHI included, for example, infections, gestational, placental, maternal weight, cardiovascular, immunological/endocrinological, neurological and psychiatric problems, RH incompatibility, thyroid disorder, exposure to x-rays and maternal smoking. There were a total of 25 prenatal items, yielding a maximum total raw score of 25. 2.2.3. Measures of stress and family environment The Adolescent Life Change Event Scale (Yeaworth et al., 1980), a self-report measure, was administered to assess adverse, stressful life events (SLEs). Two indices were derived. First, Total Number of Life Events (SLE-Total) included number of life events experienced by the adolescent in the past 6 months (27 items) (e.g., failing one or more subjects in school), plus the number of four significant events experienced by the adolescent ever (i.e., death of a parent, brother or sister, or close friend), yielding a maximum total raw score of 31. Second, an Impact (or total life change unit; SLE-Impact) score was computed, which weighted the 31 events based on severity, according to a predetermined rank-order scale (least to most upsetting events). The Impact score was the sum of the scores for each event experienced. The three scales underlying the relationship dimension of the Family Environment Scale — Form R-Current (FES) (Moos and Moos, 1994), a self-report questionnaire, were administered to measure current (within the last 6 months) social and environmental characteristics of the family, based on individuals' perceptions of their actual family environments. The three relationship dimensions included Cohesion, Expressiveness and Conflict. This measure was administered primarily to a caregiver/parent of the adolescent in 56 families and, when not available in five families, self-report from the adolescent was used. Data were missing for five families. 2.3. Statistical analyses Independent sample t-tests were employed to examine group differences (FHR vs. CC) in continuously distributed demographic variables, including age, education and parental SES. Chi-square tests were employed to assess group differences in sex and ethnicity. Independent sample t-tests were employed to examine group differences in predictors. Pearson bivariate correlations were employed to examine relationship of genetic load with PHI, SLE, and FES indices. Given group differences in age and SES, analyses were repeated controlling for age and SES, using ANCOVA (for the PHI index), MANCOVA (for the SLE and FES indices, respectively) and partial correlations. All tests were two-tailed.

3. Results 3.1. Normality testing All measures were normally distributed with exception of SLEImpact, which became normally distributed after applying the Box– Cox transformation. 3.2. Group differences in predictor variables PHI, but not SLE-Total or SLE-Impact, was significantly greater among FHRs than CCs. Patterns of family interaction differed such that Cohesion and Expressiveness were significantly less among families of FHRs compared to CCs, whereas Conflict was non-significantly higher in FHRs (p = .07). Findings remained significant for PHI, FES-Cohesion and FES-Expressiveness (though non-significant for the SLE indices) after controlling for both age and SES. Findings became significant for FES-Conflict after controlling for both age and SES (p b .001) (see Table 2). Among FHRs alone and CCs alone, 34 (85%) and 50 (90.9%) experienced at least one SLE, and 20 (50%) and 36 (65%) experienced at least one PHI, respectively. 3.3. Associations of genetic liability with early biological–environmental precursors and later environmental stress factor within the FHR group Genetic liability was significantly positively associated with PHI and FES-Expressiveness, though not with SLE-Total, SLE-Impact, FES-Cohesion or FES-Conflict (Table 3). After adjusting for SES and age, correlations of genetic liability with PHI and FES-Expressiveness remained significant, and SLE indices, FES-Cohesion and FES-Conflict remained non-significant. 4. Discussion As hypothesized, adolescents at FHR for psychosis differed significantly from comparisons regarding a number of biological and environmental risk factors from conception through young adulthood. Most markedly, prenatal health issues and family conflict were significantly greater, whereas family cohesion and expressiveness were significantly less among FHRs. Among FHR adolescents, greater genetic liability was associated with more prenatal health issues and family expressiveness. The number of stressful life events was somewhat (albeit nonsignificantly) greater among FHR families. Overall, these data indicate that biological and social environmental risk factors are important within FHR families. Our finding that greater genetic liability is associated with greater family environment disruption fits a cumulative exposure (or “behavioral sensitization”) (see van Winkel et al., 2008) model of psychosis risk.

Table 2 Means, standard deviations, effect sizes (Cohen's D) and group differences in precursors of psychosis susceptibility as a function of familial risk. Group Community control Prenatal health issuesb (n = 45;23) Stressful life eventsc Total score (n = 55;39) Impact factord (n = 55;40) Family Environment Scale Cohesion (n = 55;34) Expressiveness (n = 55;35) Conflict (n = 55;35)

Cohen's Da

Group difference

Group difference covaried for age & SES

Familial high risk

1.7 (1.4)

3.3 (2.6)

−0.86

t(28.4) = −2.7, p = .01*

F(3) = 6.8, p b .001***

2.9 (1.8) 138.9 (111.8)

3.3 (2.5) 202.8 (170.8)

−0.19 −0.46

t(92) = −.98, p = .33 t(93) = −1.9, p = .06

F(3) = .30, p N .10 F(3) = 1.9, p N .10

58.9 (12.2) 56.4 (12.1) 44.3 (11.0)

46.0 (16.6) 48.3 (12.1) 48.7 (10.8)

0.93 0.75 −0.41

t(54.9) = 3.9, p b .001*** t(88) = 3.1, p b .005** t(88) = −1.9, p = .07

F(3) = 8.5, p b .001*** F(3) = 4.1, p b .01** F(3) = 6.6, p b .001***

*p b .05; **p b .01; ***p b .001; all tests are two-tailed. a Cohen's D = [(CCMean − FHRMean) / sqrt S2pooled] based on unadjusted data. b Pregnancy History Instrument — Revised. c Adolescent Life Change Event Scale. d Box–Cox transformed data.

Please cite this article as: Walder, D.J., et al., Genetic liability, prenatal health, stress and family environment: Risk factors in the Harvard Adolescent Family High Risk for Schizophrenia Study, Schizophr. Res. (2014), http://dx.doi.org/10.1016/j.schres.2014.04.015

D.J. Walder et al. / Schizophrenia Research xxx (2014) xxx–xxx Table 3 Pearson bivariate correlations among biological and environmental predictors among adolescents at familial high-risk. Genetic liability index Prenatal health issuesa (n = 66) Stressful life eventsb Total score (n = 89) c

Impact factor (n = 89) Family Environment Scale Cohesion (n = 85) Expressiveness (n = 86) Conflict (n = 86)

.49 (n = 21) (.03)* .17 (n = 34) (.35) .03 (n = 34) (.85) .06 (n = 30) (.77) .41 (n = 31) (.02)* .07 (n = 31) (.72)

*p b .05; **p b .01; ***p b .001; all tests are two-tailed. a Pregnancy History Instrument — Revised. b Adolescent Life Change Event Scale. c Box–Cox transformed data.

Finally, our finding that high risk families are marked by a more conflict-laden and less cohesive style of family interaction, coupled with genetic liability being associated with a more highly expressive family pattern, is consistent with the notion that having a first-degree family member increases likelihood of a “risky family” environment. In turn, chronically stressful family environments – marked by conflict, deficient nurturing, harshness, neglect or aggression – may serve as a risk precursor for adverse health outcomes via allostatic load (Repetti et al., 2011). Moreover, social networks of recently diagnosed schizophrenia patients tend to be smaller and marked by relatively more family members than other families (Horan et al., 2006). Such social network discrepancies underscore the family environment as important for patients in managing illness (Gretchen-Doorly et al., 2011). Specifically, the family social environment (beyond familial genetic vulnerability) may modulate stress effects (Ventura et al., 1989). Inferences about causal directionality of effects are made cautiously. Genetic liability contributes to individual differences (e.g., inherent traits/personality) that impact likelihood of exposure to environmental adversities (Rutter et al., 2001). Moreover, having a family member with psychiatric illness may increase the likelihood of stressful life event exposure. In addition, genetic effects on behavior may emerge via individual likelihood of experiencing environmental adversities involving psychosocial stress and family environment (see van Winkel et al., 2008). Genetic liability and environmental stress also may modulate outcome (Brown, 2011), as evidenced in the Finnish Adoptive Family Study (Tienari et al., 2004). Genes also influence hypothalamic–pituitary–adrenal axis sensitivity (Walder et al., 2010), the primary neural system implicated in the biological stress response, magnifying stress impact. This follows Walker & Diforio's (1997) neural diathesis–stress model, positing hypothalamic–pituitary–adrenal axis as a plausible candidate neurobiological substrate regulating susceptibility toward psychosis via sensitization of striatal pathways/dopaminergic neurotransmission. This study offers a unique window to understanding individual effects of genetic predisposition, early prenatal factors and later environmental influences on psychosis vulnerability from a developmental perspective. Findings accentuate the need for research aimed at disentangling the biological ‘synergism’ versus ‘parallelism’ debate (see Darroch, 1997). That is, elucidating the complex, variable pathways to psychosis risk paved by the interplay of genes and environment via (multiplicative and/or additive) interaction (GxE; see van Winkel et al., 2008) and/or covariation (rGE) remains crucial. Study limitations include data loss on some variables (such as prenatal complications and FES) and substantial variability in stress measures. Moreover, we could not determine whether there were any differences

5

between individuals within families who participated and those who did not. Also, the family environment stress measure was administered primarily to a caregiver/parent of the adolescent (~80% of the sample) or, when not available, to the adolescent. Adolescents may have different perceptions of stressful life events in the family environment not accounted for in the current study, rendering this an important consideration for future investigation. In addition, unaffected relatives of control probands (some of whom carried a history of depression, for example) were included to avoid a ‘supernormal’ control group; this may have reduced the magnitude of group differences (see Seidman et al., 2006) and limited generalizability. Significant group differences despite this selection strategy, however, render results all the more striking. Similarly, findings after corrections for differences in family SES remained significant. The PHI measure may have been subject to maternal recall bias. However, prior studies assessing the current PHI measure demonstrated no evidence of positive maternal recall bias (Buka et al., 2000), or inaccuracy was limited to certain types of PHI (Buka et al., 2004). Our study included prenatal factors that extended beyond those identified by Buka et al. (2004) as particularly susceptible to recall bias. Thus, positive recall bias may not likely play a significant role in the current study, bolstering confidence in the current interpretation of findings. Nonetheless, research data drawn from current perinatal observation is necessary to confirm this finding. Although at least 50% of the FHRs and CCs each experienced at least one SLE and/or one PHI, the mean occurrences of SLEs (CC = 2.9(1.8); FHR = 3.3(2.5)) and PHIs (CC = 1.7(1.4); FHR = 3.3(2.6)) were relatively low. Thus, despite restricted statistical power, detection of significant group differences for PHIs was all the more striking; limited power may partly account for the absence of significant findings among SLE data. Finally, although we assessed 3 categories of outcomes (6 variables) raising the risk of Type 1 error, all would be significant with Bonferroni correction by category of measure. The current study supports the early components of our proposed ‘polygenic neurodevelopmental diathesis–stress model’ (Fig. 1) whereby psychosis susceptibility (and resilience) involves the independent and synergistic incremental confluence of biological and environmental factors, in a temporally sensitive (and potentially dependent) manner across important developmental periods. Arguably, the more (crucial) sensitive windows include prenatal through adolescence, given prominent neuromaturational processes (e.g., apoptosis, synaptic pruning, synaptogenesis, neural/cellular migration, neurohormonal surges) during these periods, which may affect target features and neuropathological signs of risk and illness. The currently employed multivariate approach draws attention to a vicissitude of mechanisms for future investigation, by which biological and environmental influences may converge toward enhancing identification of putative psychosis susceptibility. First, our findings support the possibility that genetic liability is linked with increased risk of obstetric complications among offspring. It remains unclear whether greater PHI is due to greater risk of adverse pregnancy among mothers with schizophrenia (psychological distress) (Sacker et al., 1996), or genetic liability brings its own direct risk to fetal development. Evidence points to familial vulnerability for perinatal stress and schizophrenia (see Walder et al., 2012) including a 5-fold increased schizophrenia risk among individuals with prenatal infection exposure and positive family history of psychosis (Clarke et al., 2009). Genetic liability may enhance sensitivity to (or modulate) prenatal complications toward heightened illness susceptibility and expression (Jablensky et al., 2005). The case–control design does not distinguish mechanisms via which having a first-degree relative with psychosis impacts pre- and post-natal developmental factors, with full consideration of rGE and GxE explanations. Although precise mechanisms remain unclear, there is mounting evidence of ‘early life programming’ (Bale et al., 2010), whereby PHIrelated perturbations may alter neurodevelopment during a critical period of heightened sensitivity with prominent downstream effects. This is one possible explanation of the origin of neuroanatomic

Please cite this article as: Walder, D.J., et al., Genetic liability, prenatal health, stress and family environment: Risk factors in the Harvard Adolescent Family High Risk for Schizophrenia Study, Schizophr. Res. (2014), http://dx.doi.org/10.1016/j.schres.2014.04.015

6

D.J. Walder et al. / Schizophrenia Research xxx (2014) xxx–xxx

abnormalities in biological relatives (e.g., Thermenos et al., 2013). Alternatively, evidence from genetic association studies implicates immunerelated pathways in schizophrenia (see Haavik et al., 2011), which may represent direct or secondary effects (due to the intrauterine environment via the maternal immune system). Given that psychosis risk appears to increase with increased number of adverse life events (Shevlin et al., 2008), future FHR studies examining isolated and cumulative effects of stress exposure and genetic liability across development – with consideration of psychiatric, functional and neurobiological outcome – arguably hold optimal promise. Our findings bolster the likelihood that factors beyond genetics are important to consider among individuals with a family member with psychosis. FHR studies aimed at differentiating these complex, temporally (and directionally) sensitive factors may facilitate more precise and balanced models of illness risk, and early detection of individuals in greatest need of preventive intervention. Growing evidence of not only adverse effects of dysfunctional family rearing environment, but also protective effects of positive family environment (O'Brien et al., 2006; González-Pinto et al., 2011), including among prodromal adolescents, is encouraging. These studies highlight malleability of risk and resilience and, in turn, that multi-pronged treatment approaches are key to yielding robust impact on the lives of individuals directly and indirectly affected by psychosis. Targeted family interventions (family support programs) may provide a potentially fruitful direction for preventive treatment (via modulating stress effects). Future studies comparing unaffected individuals with second- versus first- degree relatives with psychosis may further disentangle genetic from environmental influences among individuals at FHR for psychosis.Moreover, studies examining neuroanatomic correlates of risk factors posited in the current polygenic neurodevelopmental model may help further elucidate putative psychosis susceptibility with an eye toward relevant preventive interventions. Role of funding source This work was supported by the following: Fellowship Leave, The City University of New York (DJW); Stanley Medical Research Institute (LJS); National Association for Research on Schizophrenia and Depression (NARSAD; LJS, MTT); Mental Illness and Neuroscience Discovery (MIND) Institute (LJS); MH 43518 and MH 65562 (MTT, LJS); MH 63951 (LJS); MH 46318 (MTT); The Commonwealth Research Center of the Massachusetts Department of Mental Health, SCDMH82101008006 (LJS). Contributors Authors LJS, SVF and MTT designed the overarching study and protocol and received funding. LJS supervised all data collection. SVF supervised statistical analyses. SJG contributed statistical approaches. DJW generated the current study hypotheses and model, conducted the literature search/review and statistical analyses, and wrote the first draft of the manuscript. All authors contributed to the contents of the manuscript and approved the final manuscript. Conflict of interest The authors have declared that there are no conflicts of interest in relation to the subject of this study. Acknowledgments We thank the patients with schizophrenia and their family members, control families, and project staff for their generous contributions to the study. Staff included Maryan Augusta Mimi Braude, Joanne Donatelli, Lisa Gabel, Jennifer Koch, Marc Korczykowski, Erica Lee, Virna Merino, Elon Mesholam, Raquelle Mesholam-Gately, Caroline Patterson, Nicole Peace, Laura Phillips, Lynda Tucker, and Sharon White. Thanks also to Stephen Buka and Jill Goldstein for use of the PHI.

References American Psychiatric Association, 1994. Diagnostic and Statistical Manual of Mental Disorders: 4th Edition (DSM-IV). American Psychiatric Association, Washington, DC. Bale, T.L., Baram, T.Z., Brown, A.S., Goldstein, J.M., Insel, T.R., McCarthy, M.M., Nemeroff, C. B., Reyes, T.M., Simerly, R.B., Susser, E.S., Nestler, E.J., 2010. Early life programming and neurodevelopmental disorders. Biol. Psychiatry 68, 314–319. Binbay, T., Drukker, M., Alptekin, K., Elbi, H., Aksu Tanık, F., Özkınay, F., Onay, H., Zağlı, N., van Os, J., 2012. Evidence that the wider social environment moderates the association between familial liability and psychosis spectrum outcome. Psychol. Med. 42, 2499–2510.

Boos, H.B.M., Aleman, A., Cahn, W., Hulshoff Pol, H., Kahn, R.S., 2007. Brain volumes in relatives of patients with schizophrenia: a meta-analysis. Arch. Gen. Psychiatry 64, 297–304. Brown, A.S., 2011. The environment and susceptibility to schizophrenia. Prog. Neurobiol. 93, 23–58. Buka, S.L., Goldstein, J.M., Seidman, L.J., Tsuang, M.T., 2000. Maternal recall of pregnancy history: accuracy and bias in schizophrenia research. Schizophr. Bull. 26, 335–350. Buka, S.L., Goldstein, J.M., Spartos, E., Tsuang, M.T., 2004. The retrospective measurement of prenatal and perinatal events: accuracy of maternal recall. Schizophr. Res. 71, 417–426. Butzlaff, R.L., Hooley, J.M., 1998. Expressed emotion and psychiatric relapse: a metaanalysis. Arch. Gen. Psychiatry 55, 547–552. Cannon, M., Jones, P.B., Murray, R.M., 2002a. Obstetric complications and schizophrenia: historical and meta-analytic review. Am. J. Psychiatry 159, 1080–1092. Cannon, T.D., van Erp, T.G.M., Rosso, I.M., Huttunen, M., Lönnqvist, J., Pirkola, T., Salonen, O., Valanne, L., Poutanen, V.-P., Standertskjöld-Nordenstam, C.-G., 2002b. Fetal hypoxia and structural brain abnormalities in schizophrenic patients, their siblings, and controls. Arch. Gen. Psychiatry 59, 35–41. Cardno, A.G., Gottesman, I.I., 2000. Twin studies of schizophrenia: from bow-and-arrow concordances to star wars Mx and functional genomics. Am. J. Med. Genet. 97, 12–17. Clarke, M.C., Tanskanen, A., Huttunen, M., Whittaker, J.C., Cannon, M., 2009. Evidence for an interaction between familial liability and prenatal exposure to infection in the causation of schizophrenia. Am. J. Psychiatry 166, 1025–1030. Cross-Disorder Group of the Psychiatric Genomics Consortium; Genetic Risk Outcome of Psychosis (GROUP) Consortium, 2013. Identification of risk loci with shared effects on five major psychiatric disorders: a genome-wide analysis. Lancet 381, 1371–1379. Darroch, J., 1997. Biologic synergism and parallelism. Am. J. Epidemiol. 145, 661–668. Faraone, S.V., Biederman, J., Milberger, S., 1995. How reliable are maternal reports of their children's psychopathology? One-year recall of psychiatric diagnoses of ADHD children. J. Am. Acad. Child Adolesc. Psychiatry 34, 1001–1008. Geller, B., Zimmerman, B., Williams, M., Frazier, J., 1994. WASH-U-KSADS: Washington University at St. Louis Kiddie and Young Adult Schedule for Affective Disorders and Schizophrenia—Lifetime and Present Episode Version for DSM-IV. Washington University School of Medicine, St. Louis, MO. Gilbert, A.R., Montrose, D.M., Sahni, S.D., Diwadkar, V.A., Keshavan, M.S., 2003. Obstetric complications correlate with neurobehavioral and brain structural alterations in young relatives at risk for schizophrenia. Ann. N. Y. Acad. Sci. 1008, 269–275. Glatt, S.J., Stone, W.S., Faraone, S.V., Seidman, L.J., Tsuang, M.T., 2006. Psychopathology, personality traits and social development of young first-degree relatives of patients with schizophrenia. Br. J. Psychiatry 189, 337–345. González-Pinto, A., Ruiz de Azúa, S., Ibáñez, B., Otero-Cuesta, S., Castro-Fornieles, J., GraellBerna, M., Ugarte, A., Parellada, M., Moreno, D., Soutullo, C., Baeza, I., Arango, C., 2011. Can positive family factors be protective against the development of psychosis? Psychiatry Res. 186, 28–33. Gottesman, I.I., Shields, J., 1967. A polygenic theory of schizophrenia. Proc. Natl. Acad. Sci. U. S. A. 58, 199–205. Gretchen-Doorly, D., Detore, N.R., Ventura, J., Hellemann, G., Subotnik, K.L., Nuechterlein, K.H., 2011. Relationships between perceptions of the family environment and of negative life events in recent-onset schizophrenia patients. Schizophr. Res. 127, 266–267. Haavik, J., Halmøy, A., Hegvik, T., Johansson, S., 2011. Maternal genotypes as predictors of offspring mental health: the next frontier of genomic medicine? Futur. Neurol. 6, 731–743. Hollingshead, A.B., 1975. Four factor index of social status. Yale University Department of Sociology, New Haven, CT. Horan, W.P., Subotnik, K.L., Snyder, K.S., Nuechterlein, K.H., 2006. Do recent-onset schizophrenia patients experience a “social network crisis”? Psychiatry 69, 115–129. Jablensky, A.V., Morgan, V., Zubrick, S.R., Bower, C., Yellachich, L.-A., 2005. Pregnancy, delivery, and neonatal complications in a population cohort of women with schizophrenia and major affective disorders. Am. J. Psychiatry 162, 79–91. Leff, J., Kuipers, L., Berkowitz, R., Eberlein-Vries, R., Sturgeon, D., 1982. A controlled trial of social intervention in the families of schizophrenic patients. Br. J. Psychiatry 141, 121–134. Levene, J.E., Lancee, W., Seeman, M.V., Skinner, H., Freeman, S.J.J., 2009. Family and patient predictors of symptomatic status in schizophrenia. Can. J. Psychiatry 54, 446–451. Lewis, S.W., Murray, R.M., 1987. Obstetric complications, neurodevelopmental deviance, and risk of schizophrenia. J. Psychiatr. Res. 21, 413–421. Lichtenstein, P., Yip, B.H., Björk, C., Pawitan, Y., Cannon, T.D., Sullivan, P.F., Hultman, C.M., 2009. Common genetic determinants of schizophrenia and bipolar disorder in Swedish families: a population-based study. Lancet 373, 234–239. Lukoff, D., Snyder, K., Ventura, J., Nuechterlein, K.H., 1984. Life events, familial stress, and coping in the developmental course of schizophrenia. Schizophr. Bull. 10, 258–292. Maxwell, E., 1992. Manual for the Family Interview for Genetic Studies (FIGS). Clinical Neurogenetics Branch, Intramural Research Program NIMH. McIntosh, A.M., Owens, D.C., Moorhead, W.J., Whalley, H.C., Stanfield, A.C., Hall, J., Johnstone, E.C., Lawrie, S.M., 2011. Longitudinal volume reductions in people at high genetic risk of schizophrenia as they develop psychosis. Biol. Psychiatry 69, 953–958. Moos, R.H., Moos, B.S., 1994. Family Environment Scale Manual: A Social Climate Scale, 3rd ed. Consulting Psychologists Press. Myin-Germeys, I., van Os, J., Schwartz, J.E., Stone, A.A., Delespaul, P.A., 2001. Emotional reactivity to daily life stress in psychosis. Arch. Gen. Psychiatry 58, 1137–1144. Norman, R.M., Malla, A.K., 1993. Stressful life events and schizophrenia. I: A review of the research. Br. J. Psychiatry 162, 161–166. Nuechterlein, K.H., Dawson, M.E., Ventura, J., Gitlin, M., Subotnik, K.L., Snyder, K.S., Mintz, J., Bartzokis, G., 1994. The vulnerability/stress model of schizophrenic relapse: a longitudinal study. Acta Psychiatr. Scand. Suppl. 382, 58–64.

Please cite this article as: Walder, D.J., et al., Genetic liability, prenatal health, stress and family environment: Risk factors in the Harvard Adolescent Family High Risk for Schizophrenia Study, Schizophr. Res. (2014), http://dx.doi.org/10.1016/j.schres.2014.04.015

D.J. Walder et al. / Schizophrenia Research xxx (2014) xxx–xxx Nurnberger, J., 1994. The Diagnostic Interview for Genetic Studies (DIGS). Arch. Gen. Psychiatry 51, 849–859. O'Brien, M.P., Gordon, J.L., Bearden, C.E., Lopez, S.R., Kopelowicz, A., Cannon, T.D., 2006. Positive family environment predicts improvement in symptoms and social functioning among adolescents at imminent risk for onset of psychosis. Schizophr. Res. 81, 269–275. Phillips, L.K., Giuliano, A.J., Lee, E.H., Faraone, S.V., Tsuang, M.T., Seidman, L.J., 2011. Emotion–cognition interaction in people at familial high risk for schizophrenia: the impact of sex differences. J. Abnorm. Psychol. 120, 993–998. Repetti, R.L., Robles, T.F., Reynolds, B., 2011. Allostatic processes in the family. Dev. Psychopathol. 23, 921–938. Rosso, I.M., Makris, N., Thermenos, H.W., Hodge, S.M., Brown, A., Kennedy, D., Caviness, V. S., Faraone, S.V., Tsuang, M.T., Seidman, L.J., 2010. Regional prefrontal cortex gray matter volumes in youth at familial risk for schizophrenia from the Harvard Adolescent High Risk Study. Schizophr. Res. 123, 15–21. Rutter, M., Pickles, A., Murray, R., Eaves, L., 2001. Testing hypotheses on specific environmental causal effects on behavior. Psychol. Bull. 127, 291–324. Sacker, A., Done, D.J., Crow, T.J., 1996. Obstetric complications in children born to parents with schizophrenia: a meta-analysis of case–control studies. Psychol. Med. 26, 279–287. Scala, S., Pousada, A., Stone, W.S., Thermenos, H.W., Manschreck, T.C., Tsuang, M.T., Faraone, S.V., Seidman, L.J., 2013. Verbal and visual–spatial memory impairment in youth at familial risk for schizophrenia or affective psychosis: a pilot study. Schizophr. Res. 144, 122–128. Seidman, L.J., Giuliano, A.J., Smith, C.W., Stone, W.S., Glatt, S.J., Meyer, E., Faraone, S.V., Tsuang, M.T., Cornblatt, B., 2006. Neuropsychological functioning in adolescents and young adults at genetic risk for schizophrenia and affective psychoses: results from the Harvard and Hillside Adolescent High Risk Studies. Schizophr. Bull. 32, 507–524. Seidman, L.J., Meyer, E.C., Giuliano, A.J., Breiter, H.C., Goldstein, J.M., Kremen, W.S., Thermenos, H.W., Toomey, R., Stone, W.S., Tsuang, M.T., Faraone, S.V., 2012. Auditory working memory impairments in individuals at familial high risk for schizophrenia. Neuropsychology 26, 288–303. Shevlin, M., Houston, J.E., Dorahy, M.J., Adamson, G., 2008. Cumulative traumas and psychosis: an analysis of the national comorbidity survey and the British Psychiatric Morbidity Survey. Schizophr. Bull. 34, 193–199. Tessner, K.D., Mittal, V., Walker, E.F., 2011. Longitudinal study of stressful life events and daily stressors among adolescents at high risk for psychotic disorders. Schizophr. Bull. 37, 432–441. Thermenos, H.W., Keshavan, M.S., Juelich, R.J., Molokotos, E., Whitfield-Gabrieli, S., Brent, B.K., Seidman, L.J., 2013. A review of neuroimaging studies of young relatives of individuals with schizophrenia: a developmental perspective from schizotaxia to

7

schizophrenia. American Journal of Medical Genetics B Neuropsychiatric Genetics 162 (7), 604–635. Tienari, P., Wynne, L.C., Sorri, A., Lahti, I., Läksy, K., Moring, J., Naarala, M., Nieminen, P., Wahlberg, K.-E., 2004. Genotype–environment interaction in schizophrenia-spectrum disorder. Long-term follow-up study of Finnish adoptees. Br. J. Psychiatry 184, 216–222. Tsuang, M., 2000. Schizophrenia: genes and environment. Biol. Psychiatry 47, 210–220. Van Winkel, R., Henquet, C., Rosa, A., Papiol, S., Fananás, L., De Hert, M., Peuskens, J., van Os, J., Myin-Germeys, I., 2008. Evidence that the COMT(Val158Met) polymorphism moderates sensitivity to stress in psychosis: an experience-sampling study. Am. J. Med. Genet. B Neuropsychiatr. Genet. 147B, 10–17. Ventura, J., Nuechterlein, K.H., Lukoff, D., Hardesty, J.P., 1989. A prospective study of stressful life events and schizophrenic relapse. J. Abnorm. Psychol. 98, 407–411. Walder, D.J., Trotman, H.D., Cubells, J.F., Brasfield, J., Tang, Y.-L., Walker, E.F., 2010. Catechol-O-methyltransferase modulation of cortisol secretion in psychiatrically at-risk and healthy adolescents. Psychiatr. Genet. 20, 166–170. Walder, D.J., Ospina, L., Daly, M.P., Statucka, M., Raparia, E., 2012. Early neurodevelopment and psychosis risk: role of neurohormones and biological sex in modulating genetic, prenatal and sensory processing factors in brain development. In: AnastassiouHadjicharalambous, X. (Ed.), Psychosis: Causes, Diagnosis and Treatment. Nova Science. Walker, E., Bollini, A.M., 2002. Pubertal neurodevelopment and the emergence of psychotic symptoms. Schizophr. Res. 54, 17–23. Walker, E.F., Diforio, D., 1997. Schizophrenia: a neural diathesis-stress model. Psychological Review 104 (4), 667–685. Wechsler, D., 1991. Wechsler Intelligence Scales for Children, Third Edition (WISC-III). The Psychological Corporation, Inc., San Antonio, TX. Wechsler, D., 1997. Wechsler Adult Intelligence Scale — Third Edition (WAIS-III). Psychological Corporation, San Antonio, TX. Weinberger, D.R., 1987. Implications of normal brain development for the pathogenesis of schizophrenia. Arch Gen Psychiatry 44 (7), 660–669. Wicks, S., Hjern, A., Dalman, C., 2010. Social risk or genetic liability for psychosis? A study of children born in Sweden and reared by adoptive parents. Am. J. Psychiatry 167, 1240–1246. Wray, N.R., Gottesman, I.I., 2012. Using summary data from the Danish national registers to estimate heritabilities for schizophrenia, bipolar disorder, and major depressive disorder. Front. Genet. 3, 118. Yeaworth, R.C., York, J., Hussey, M.A., Ingle, M.E., Goodwin, T., 1980. The development of an Adolescent Life Change Event Scale. Adolescence 15, 91–97. Zubin, J., Spring, B., 1977. Vulnerability–a new view of schizophrenia. J. Abnorm. Psychol. 86, 103–126.

Please cite this article as: Walder, D.J., et al., Genetic liability, prenatal health, stress and family environment: Risk factors in the Harvard Adolescent Family High Risk for Schizophrenia Study, Schizophr. Res. (2014), http://dx.doi.org/10.1016/j.schres.2014.04.015