Genetic and psychosocial stressors have independent effects on the level of subclinical psychosis: findings from the multinational EU-GEI study

Abstract

Gene x environment (GxE) interactions have not been reliably established regarding etiology of psychotic disorders, while genes-environment (G-E) associations have been displayed. We studied the role of GxE interaction between psychosocial stressors (childhood trauma, stressful life-events, self-reported discrimination experiences and low social capital) and the polygenic risk scores for schizophrenia (PRS-SZ) on subclinical psychosis in a population-based sample. Data were drawn from the EU-GEI study, in which subjects without psychotic disorders were included in six countries. The sample was restricted to European descendant subjects (N = 706). Subclinical dimensions of psychosis (positive, negative, and depressive) were measured by the Community Assessment of Psychic Experiences (CAPE) scale. For each dimension, the interactions between genes and environment were assessed comparing explained variances of “Genetic” models (solely fitted with PRS-SZ), “Environmental” models (solely fitted with each environmental stressor), “Independent” models (with PRS-SZ and each environmental factor), and “Interaction” models (with an interaction term between the PRS-SZ and each environmental factor). There were no direct G-E associations. PRS-SZ was associated with positive dimensions (ß = 0.092, R2 = 7.50%), and most psychosocial stressors were associated with all three subclinical psychotic dimensions (except for low social capital and positive dimension). Concerning the positive dimension, Independent models fitted better than Environmental and Genetic models. No significant GxE interaction was observed for any dimension. This study in healthy individuals suggests that i) the etiological continuum hypothesis could concern particularly the positive dimension of subclinical psychosis, ii) genetic and environmental factors have independent effects on the level of this positive dimension, iii) and that interactions between genetic and individual environmental factors could not be identified in this sample.