DPYSL2 and mTOR Signaling in Schizophrenia
MetadataShow full item record
Background Located in a schizophrenia susceptibility locus on chr8p21, DPYSL2 has been implicated in schizophrenia by numerous linkage and association studies; however research on its biology in relation to the risk for SZ has been sparse. DPYSL2 encodes CRMP2, a protein that functions in axon growth and maintenance. A first targeted sequencing study of this gene revealed several non-coding variants that altered expression in transient transfection assays. Most interestingly, a polymorphic CT di-nucleotide repeat (DNR) variant located in the 5’-untranslated region of DPYSL2 is associated with SZ risk and disrupts regulation of the gene by mTOR signaling pathway. Given the biological relevance of DPYSL2 to schizophrenia, the aim of my thesis, composed of three separate investigations, was to further explore the molecular mechanisms on which variations in DPYSL2 contribute risks for SZ. Methods I extended the sequencing screen of this gene on a larger scale using a high-throughput pooling strategy, with the goal to identify rare functional variants in DPYSL2. The targeted regions for sequencing included the 14 exons and 27 conserved noncoding regions in and around this gene. To expand our knowledge on the polymorphic DNR variant, we performed protein-RNA interaction studies that identified differential binding of mTOR effectors to the DNR alleles. And to study the DNR variant in its native genomic context, we used CRISPR/cas9 genome editing to create two isogenic cell lines that differ only at the DNR locus. To study loss of function of Dpysl2 in vivo, we created a conditional knockout mouse model of Dpysl2 to test the hypothesis that loss of function of Dpysl2 alters axonogenesis, perturbing morphological and behavioral phenotypes. Results From the extended sequencing screen of this DPYSL2, we identified a total of 286 variants, of which 66 were novel, and 8 were synonymous. The resulting comprehensive variant catalogue of this gene showed a modest enrichment of rare variants in SZ patients with no functional biases. Because the enrichment was small (1.28 fold-change) and not variant-specific, we decided to pursue the aforementioned DNR variant instead. In protein-RNA interaction studies, we observed differential bindings of proteins to the DNR alleles. Specifically, the longer risk allele lost binding to ELAVL4/HuD, a neuronal ribosomal binding protein regulated by the mTOR pathway. Additionally, the risk allele also had decreased binding to two mTOR effectors, 4E-BP and eIF4E. In the CRISPR-edited cells that carry the homozygous DNR alleles, we showed that the longer CT repeat leads to significant reduction of CRMP2 levels, accompanied by striking shortening of cellular projections, and significant transcriptome changes that link this transcript back to SZ. We also successfully created a conditional Dpysl2 knockout mouse, which shows interesting phenotypic changes in preliminary behavioral tests. The results will be confirmed and extended with additional testing, and this mouse model has the potential to be a powerful tool in the study of SZ. Conclusion Together my work contributes a significant body of evidence that link DPYSL2 and mTOR to risk for schizophrenia. My results support the functional importance of the DNR, suggest a role for mTOR signaling, and further implicate DPYSL2 as a schizophrenia susceptibility gene.