Across Diagnostic Boundaries: Genetic Variants for Neuropsychiatric Disorders and their Association with Human Brain Structure

Abstract

With the advances in genome-wide association studies (GWAS), hundreds of genetic variants have been identified for neuropsychiatric disorders. Strikingly, many of these genetic variants showed complex associations across diagnostic groups. For example, the second cross-disorder GWAS meta-analysis by the Psychiatric Genomics Consortium (Lee et al., 2019) identified 11 antagonistic single-nucleotide polymorphisms (SNPs) that were associated with an increased risk for one neuropsychiatric disorder, while being protective against another disorder. Furthermore, the cross-disorder GWAS meta-analysis uncovered 23 highly pleiotropic SNPs that were associated with at least four neuropsychiatric disorders and 22 SNPs that were predominantly associated with schizophrenia (SCZ) but not with the other disorders. The underlying molecular mechanisms by which these genetic variants alter the risk of distinct neuropsychiatric disorders are largely unclear. The present thesis conducted two imaging genetic studies to uncover the associations between antagonistic, highly pleiotropic, and predominantly SCZ-associated SNPs with brain structure and brain-related traits. <br/> Study 1 performed a systematic characterization of the 11 antagonistic SNPs. Here, eight of the 11 antagonistic SNPs were significantly associated with at least one brain structural phenotype using the GWAS summary statistics from the ENIGMA and CHARGE consortia. Several of the implicated phenotypes were found to be altered in patients with bipolar disorder, major depression, or SCZ compared to controls. Six of the eight antagonistic SNPs were significantly associated with gene expression in brain tissue, and all eight antagonistic SNPs were significantly associated with cognitive-behavioral traits. Furthermore, rs301805 and rs1933802 were significantly associated with voxel-wise gray matter volume in data from the FOR2107 study. <br/> Study 2 used data from the UK Biobank (<em>n</em>=28,952) to examine the association of a genetic risk score of highly pleiotropic SNPs (PleioPsych-GRS) and a genetic risk score of predominantly SCZ-associated SNPs (SCZ-GRS) with brain structure and outcomes related to mental health. To prioritize individual SNPs, the association of each SNP with brain structure was investigated. Study 2 found that the PleioPsych-GRS was not significantly associated with brain structural phenotypes after multiple testing corrections, whereas the SCZ-GRS was significantly associated with left and right putamen volume and left and right lateral orbitofrontal surface area, among others. The PleioPsych-GRS and the SCZ-GRS were significantly associated with eight and four outcomes related to mental health, respectively. Furthermore, two highly pleiotropic and ten predominantly SCZ-associated SNPs were significantly associated with at least one brain structural phenotype. <br/> In conclusion, this thesis showed that antagonistic, predominantly SCZ-associated and, to a lesser extent, highly pleiotropic SNPs for neuropsychiatric disorders were associated with brain structure. In addition, the SNPs were associated with traits related to mental health, cognition, and behavior. These findings provided a notion of how these SNPs might influence disease development and led to the prioritization of selected SNPs and brain regions relevant for further investigations. Future work should extend these findings by exploring the association of these SNPs with additional brain modalities, including white matter microstructure and structural and functional connectivity of the human brain.