SU23 - ASSOCIATIONS OF POLYGENIC RISK OF SCHIZOPHRENIA AND MAJOR DEPRESSION IN STUDIES OF BIPOLAR DISORDER AND MAJOR DEPRESSION

Abstract

Background Large Genome-Wide Association Studies (GWAS) have identified common genetic variants that underlie Schizophrenia (SCZ) and Major Depressive Disorder (MDD), although Bipolar Disorder (BD) remains elusive. Cross disorder analysis has shown a strong genetic correlation of SCZ with BD, with modest correlations of SCZ with MDD, and MDD with BD. We used polygenic risk scores to examine shared genetic variation among disorders and to dissect the genetic contribution to shared clinical features, such as psychotic symptoms. Methods Participants from the Kaiser Permanente Research Program on Genes, Environment, and Health (9,080 MDD cases; 54,808 controls) and a multi-ethnic GWAS of bipolar 1 disorder (BD1) (5,056 cases; 54,808 controls) were genotyped on the Affymetrix Axiom system, using ancestry-specific arrays, and the data imputed to the 1000 Genome Project (March 2012 release). We calculated standardized polygenic risk scores for SCZ (SCZ-GRS) and MDD (MDD-GRS) as the weighted sum of risk alleles using 128 SNPs and 47 SNPs that were significantly associated with SCZ and MDD, respectively, in independent GWAS. Two SNPs were common to both scores. Logistic regression analyses of the association of SCZ-GRS and MDD-GRS with BD1 and MDD, separately, included gender, age, and principal components for adjustment of ancestry. Analyses were conducted separately in race-ethnicity groups and by gender. Results The SCZ-GRS and MDD-GRS were significantly and independently associated with BD1 in non-Hispanic whites (3,725 cases, 52,430 controls) in an analysis including both (Odds Ratio (OR)-MDD-GRS=1.138; 95% CI=1.099, 1.178; p= 2.646×10–13; OR-SCZ-GRS=1.146 95% CI=1.107, 1.187; p= 1.241×10–14), with similar results in the smaller samples of minorities, and in men and women. There was no interaction between the scores. Among BD1 cases, the SCZ-GRS was significantly associated with psychotic symptoms and hospitalization for BD; the MDD-GRS was also modestly associated with psychotic symptoms and nominally with rapid cycling of moods, a hallmark of BD. In analyses of MDD in non-Hispanic whites (7,833 cases; 43,581 controls), the MDD-GRS was significantly associated with MDD (OR=1.147; 95% CI=1.119,1.176; p=3.13×10–27), but the SCZ-GRS was not (OR=1.017; 95% CI=0.992, 1.042; p=0.189). Similar results were found in race-ethnic minorities, and in men and women. There was no interaction. Neither the MDD-GRS nor the SCZ-GRS was associated with hospitalization or psychotic symptoms in MDD, although the prevalence was much lower in MDD than BD. Discussion BD has both cognitive and affective components, most often including depressive episodes like those in MDD, but also frequently including psychotic symptoms, as does SCZ. The independent association of the MDD-GRS and SCZ-GRS with BD1 suggests that different genetic variants may contribute to the cognitive and affective clinical features of BD. In this study, genetic variation associated with SCZ was not associated with MDD, or with psychotic symptoms in MDD. Identification of genetic variation that influences specific disorders and specific features of disorders increases knowledge about the genetic etiology of disorders and may contribute to understanding the underlying biology of disorders and identification of new targets for treatment.