162. Shared Genetic Risk of Schizophrenia and Gray Matter Reduction in 6p22.1

Abstract

Background: Genetics is known to influence both schizophrenia risk and brain structure. Thereby a question emerges as to whether the genetic underpinnings of these two traits overlap. Methods: We focused on 1402 common single nucleotide polymorphisms (SNPs) which were mapped from the psychiatric genomics consortium (PGC) 108 regions and showed relatively strong group differences (P < 1.00E-04 in PGC; Ripke et al., 2014). We used parallel independent component analysis (Liu et al., 2009) to study multivariate associations of these SNPs with whole-brain gray matter volume (GMV) variation (429655 voxels) in 1401 individuals. The discovery sample consisted of 355 schizophrenia (SZ) patients and 437 healthy controls not having been used in PGC. The components identified in the discovery sample were evaluated in an independent sample (272 SZ patients and 337 controls) for replication. Results: The component number was estimated to be 28 and 65 for the SNP and GMV modality respectively using minimal description length (Rissanen, 1978). One pair of SNP-GMV components were identified to show a significant association passing Bonferroni correction (r = −.16, P = 4.45E-06) which was further replicated when evaluated in the independent sample (r = −.08, P = 4.75E-02), both controlled for diagnosis. The GMV component presented a significant GMV reduction in the inferior parietal and superior temporal regions in the SZ group (P = 8.52E-11) and this reduction was also replicated (P = 2.04E-05). The SNP component included 39 dominant loci residing in chr6:28308034-28684183, echoing the most significant region of PGC 108 SZ relevant regions. Though in the discovery sample no significant group difference was noted in the SNP component loading, it indeed highly correlated (r = .97) with the cumulative polygenic risk score (PRS) for SZ (leveraging the PGC odds ratios) of the same 39 SNPs (Purcell et al., 2009; Ripke et al., 2014). Between the PRS for SZ and GMV component loading, a direct correlation of r = −.14 (P = 9.06E-05) was noted. The higher the PRS for SZ, the lower the GMV. Conclusion: Collectively, we provide evidence that a set of SNPs in 6p22.1, a region previously implicated for SZ susceptibility, likely influence both SZ and gray matter volume in temporal and parietal regions which are among the regions thought to be altered in this illness. The findings emphasize the need for dissecting SZ susceptibility loci to better understand their associations with focal brain measures.