Identifying causal mechanisms for beta cell dysfunction at type 2 diabetes risk loci

Genome-wide association studies (GWAS) have identified more than 150 loci associated with type 2 diabetes (T2D) risk. Most association signals are located in noncoding regions of the genome, and uncertainty remains over the causal gene in the vast majority of cases. Motivated by the overall aim of accelerating the translation of genetic discoveries into molecular mechanisms for disease pathogenesis, the work presented in this thesis sought to identify candidate causal genes, and to study their roles in human beta cell function. The causal mechanism for non-coding variants at the CDKN2A/B locus has remained elusive, though animal models have implicated Cdkn2a in beta cell function. To determine the effect of CDKN2A haploinsufficiency on glucose homeostasis in humans, I analysed data from oral and intravenous glucose tolerance tests in individuals carrying rare CDKN2A loss-offunction mutations. Compared with controls, carriers displayed increased insulin secretion, impaired insulin sensitivity, and reduced hepatic insulin clearance. Follow-up studies in the human beta cell line, EndoC-βH1, demonstrated cell-cycle independent effects of CDKN2A on insulin secretion. At the PAM/PPIP5K2 locus, two non-synonymous variants in PAM are associated with both T2D risk and insulinogenic index. To elucidate a causal mechanism, I investigated possible local and extra-pancreatic roles of PAM in beta cell function. I demonstrated that PAM deficiency results in reduced insulin content and altered dynamics of insulin secretion in EndoC-βH1 and primary beta cells. I also identified the amidated granular packaging protein Chromogranin A as an endogenous PAM substrate, establishing a direct link to granulogenesis in human beta cells. Lastly, to enable systematic, large-scale prioritization of causal genes for T2D risk, I performed a high-throughput RNAi screen for beta cell dysfunction. Among 300 genes selected from 75 risk loci, I identified significant hits at half of these regions. The hits were highly enriched for known monogenic diabetes genes, but also revealed the poorly characterised transcription factor ZMIZ1 to be one of the strongest regulators of insulin secretion. Silencing of ZMIZ1 in primary human islets subsequently revealed a core beta cell network to be negatively affected, highlighting a plausible causal mechanism. Overall, this thesis has addressed a significant outstanding challenge in the translation of genetic signals into disease mechanisms. Using a range of different strategies, including genetic data, candidate gene biology, and functional genetic screening, the work has successfully identified several candidate causal mechanisms for beta cell dysfunction at T2D risk loci. These results provide a deeper understanding of the fundamental mechanisms underpinning disease predisposition, and thus highlight potential new therapeutic applications.