Involvement Of Genetic Variations And Allele-specific Expression In The Pathogenesis Of Major Psychosis

PartⅠ Multi-SNP-based integrative analysis identified functional genetic risk loci underlying susceptibility to schizophreniaDespite exhibiting genome-wide level of statistical significance,most risk SNPs identified from the PGC2 schizophrenia GWAS showed moderate effect size.We believe there are authentic risk signals among the nominal SNPs showing large effect sizes(e.g.,P value<1.0e-5 and OR>1.2 or<1/1.2).To test this hypothesis,we performed multi-SNP-based integrative analysis of PGC2 schizophrenia GWAS data using Multi-marker Analysis of GenoMic Annotation(MAGMA)method and identified 435 independent schizophrenia risk loci with P value<7.22e-6,each of which containing at least one SNP with large effect size(OR>1.2 or<1/1.2).The results that 127 out of our 435 identified loci were located within 500 kb around the reported genome-wide significant SNPs,together with pathway analysis of our 435 identified loci involved in neuronal and synaptic related pathways,demonstrated that our multi-SNP based approach provided a novel strategy to identify disease-associated variants with large effect size.To identify functional target genes of the 435 identified risk loci,we then employed a multi-SNP strategy with Summary-data-based Mendelian Randomization(SMR)method by integrating PFC eQTL data with PGC2 schizophrenia GWAS data,followed by topological analysis using PFC Hi-C data.After validated the target genes,which are located within the same topological domain with the risk loci based on the Hi-C data and show eQTL association with the risk loci,we identified that the disease association patterns of 23 risk loci each showed matched signature with the brain eQTL patterns from one of 22 target genes.Of the 23 identified target genes,USMG5 displayed the strongest multi-SNPbased SMR-multi P value and showed matched colocalized signatures between the USMG5 eQTL and GWAS associations of more than two loci,including the uncharacterized CNNM2 and the previously reported AS3MT in the 10q24.32 locus,which was also in top significant GWAS loci(PMAGMA=8.34e-13)in our 23 identified loci and a previous report on schizophrenia and other psychiatric disorders.Furthermore,we observed significant downregulation of USMG5 expression in postmortem PFC brain tissues from schizophrenia cases compared to those of nonpsychiatry controls in the Gene Expression Omnibus database.Importantly,knockdown of endogenous USMG5 expression by shRN A in human induced pluripotent stem cell-derived cortical neurons significantly dysregulated neural development and ATP synthesis.Finally,we observed USMG5 downregulated by the risk C allele of rs1926032 in the CNNM2 intronic region from both brain eQTL data and the luciferase reporter-transfected SK-N-SH cell lines,as well as and decreased ATP synthesis in the rs1926032-contained region deleted SK-N-SH cells.4C experiment also verified the chromatin interaction between rs1926032 and USMG5 promoter,which supported the transcriptional regulation of rs1926032 on USMG5 expression.Our findings provided a multi-SNP-analysis strategy in identification of GWAS risk loci and functional target genes,and delineated the molecular mechanism for the CNNM2 GWAS locus on 10q24.32 associations with schizophrenia by mediating longdistance USMG5 gene expression underlying the developmental biology of schizophrenia.PartⅡ Allele-specific expression is involved in the pathogenesis of major psychosis monozygotic twinsMajor psychosis including schizophrenia and bipolar disorder are complex heritable mental illnesses.The incidence of monozygotic twins of patients with major psychosis is only about 50%,we suspected that there may be interactions between genetic and environmental factors that play an important role in the development of major psychosis.Although a large number of genetic risk factors for these disorders is successfully identified,it remains unclear how genetic variants interact with environmental and epigenetic risk factors.Based on the phenomena that monozygotic twin pairs display identical genetics background but discordant or concordant for diagnostic phenotype,we performed whole-genome wide analysis of genetic variants,DNA methylation and RNA expression with blood of monozygotic twins.Through intergrative analysis of genomic,methylation,and transcriptome with public databasets,we observed that regulatory SNPs are highly linked with loss-of-function(LOF)SNPs in discordant twins and display allelic expression pattern,thus reducing allelic expression of LOF SNPs and risk of major psychosis.However,altered DNA methylation will induce the changes in the cis-mediated allele-specific expression pattern and increase the allelic expression level of LOF SNPs,resulting increased risk of major psychosis.This finding pointed the role of LOF SNPs and their interaction with epigenetic modification underlying the development of major psychosis.