부산대학교 도서관

Type 2 diabetes (T2D) is a metabolic disease characterised by insulin resistance and beta-cell failure and a leading cause of mortality and morbidity. GWAS have identified pancreatic beta-cell dysfunction as a major player in T2D pathogenesis, but since the majority of signals reside in non- coding sequence the causal variant(s) and effector transcripts often remain undefined. Combining human genetic discoveries with genomic interrogation of relevant tissues was shown to provide a powerful strategy to understand the molecular disease mechanisms. Thus, the main aim of my DPhil project is to generate and integrate genetic data, such as rare coding CNVs, and epigenomic data, including human pancreatic islet DNA methylation and chromatin state data, to characterise existing T2D GWAS loci and identify novel genetic T2D susceptibility signals. Using WGBS (n=10) and ATAC-seq (n=29), I extended the epigenomic characterisation of human pancreatic islets. Specifically, I characterised for the first time the whole-genome human islet DNA methylome and provided information about the variability in open chromatin in human islets. Enrichment analysis in T2D-related GWAS data showed that hypomethylated enhancer-like regions (n=37.1k, FE=3.2) and open chromatin (n=247.2k, FE=3.1) regulatory regions are strongly enriched in T2D GWAS data. These annotations were integrated with existing islet ChIP-seq annotations to extend islet chromatin states. Through this approach, I identified enhancer subtypes based on DNA methylation and accessibility, which differed markedly in their enrichment levels in T2D-related GWAS signals. The refined enhancer states were then used to prioritise T2D-associated variants for testing of allelic imbalance in open chromatin. After correcting for technical artefacts, significant allelic (FDR < 0.05) imbalance in open chromatin was discovered at 3 T2D GWAS loci (ADCY5, CDC123, KLHDC5); thus highlighting potential molecular mechanisms at these loci governing T2D risk. I also used ATAC-seq to characterise the epigenomic landscape of human iPSC subtypes derived from beta- and fibroblast-cells that differed in their ability to differentiate into endodermal tissue. The identification of significant (FDR < 0.05) differentially open chromatin sites (n=8.3k) in these iPSCs, which were found to be associated with genes linked to both pancreas development and mature function, could be used to improve current in vivo 'iPSC to endodermal' differentiation protocols. Ultimately, this may lead to enhanced generation of human islet cells relevant for both T2D disease modelling and potential clinical use. Finally, I determined the role of rare coding CNVs, identified from large-scale T2D-case-control exome-sequencing (n=14.5k) and exome chip array (n=11.6k) datasets, in T2D development. This analysis found at the known C2CD4A/VPS13C T2D GWAS locus, rare VPS13C gene deletions and duplications consistently associated with increased and decreased T2D-risk, respectively; hence, prioritising VPS13C for potential functional follow-up at this GWAS locus. These data have extended the catalogue of genetic variants (rare coding CNVs) and the epigenomic landscape of human islets and iPSCs to improve understanding of islet development and regulation as well as T2D genetic susceptibility.