T36 - Genome-Wide Association Study Of Borderline Personality Disorder Reveals Genetic Overlap With Bipolar Disorder And Schizophrenia

Abstract

Background Borderline personality disorder (BPD) is characterized by mood lability and impulsivity – symptoms which overlap with symptoms of Bipolar Disorder (BD) – a disorder showing a substantial comorbidity. Research suggests that its etiology involves both environmental and genetic factors. Formal genetic studies indicate a heritability of up to 65%. However, to date genetic research into BPD has been limited to candidate gene studies, and no case-control genome-wide association study (GWAS) has yet been performed. Systematic genome-wide screening for BPD personality features has been performed in individuals from the general population. Here, we present the first case-control GWAS of BPD, which was performed in one of the largest BPD patients samples worldwide. Given the heritability estimates for BPD, no significant single marker results were expected. Rather, we were interested in findings pointing to genes, gene-sets, and potential overlap with other psychiatric disorders. Methods GWAS was performed in 1,034 BPD patients and 1,545 controls recruited at four German academic institutions (Mannheim, Berlin, Mainz, Munich). After quality control and imputation, association was tested under an additive logistic regression model using PLINK and the derived principal components as covariates. Gene-based tests were performed using VEGAS and MAGMA. Gene set analyses were performed with GSEA for GWAS using GO. Further gene set analyses are currently being performed. The LD regression score method was used to calculate genetic overlap between BPD, BD, and Schizophrenia (SCZ). Results The top hit of the SNP-based analysis was Developmental Pluripotency Associated 3 (DPPA3, p=1.65x10-7). The top hit of the gene-based analysis was Plakophilin4 (PKP4) which reached genome-wide significance using MAGMA (p=5.26x10-7). The gene set analysis also yielded a significant finding after correction for multiple testing, i.e., exocytosis (GO:0006887; p=0.001). The genetic correlation between BPD and BD was rg=0.34 (p=4.37x10-5), and that between BPD and SCZ was rg=0.28 (p=2.99x10-3). Discussion The present study is the first case-control GWAS of BPD. As expected, no significant association was found with any single marker or gene. The top SNP was in the gene DPPA3, which is implicated in epigenetic mechanisms. The top gene of the gene-based test – PKP4 – was also one of the top hits of the single marker analysis. PKP4 is involved in the regulation of cell adhesion and cytoskeletal organization, processes which have been linked to BD and SCZ. The most promising gene of the previously reported GWAS on BPD personality features - SERINC5 - showed nominally significant association. Previous research has implicated the top hit of the gene set analysis in the molecular mechanisms of BD and SCZ, and may now represent a promising starting point for further research into BPD. The most interesting finding of the present study was the genetic overlap between BPD and BD as well SCZ. Our study is the first to demonstrate on the genetic level that BPD is not a discrete entity but overlaps with major psychoses not only on the clinical level. Future studies are warranted to determine commonalities and specificities.