Mutations in human Gli-similar (GLIS) 3 protein cause neonatal diabetes. The GLIS3 gene region has also been identified as a susceptibility risk locus for both type 1 and type 2 diabetes. GLIS3 plays a role in the generation of pancreatic beta cells and in insulin gene expression, but there is no information on the role of this gene on beta cell viability and/or susceptibility to immune- and metabolic-induced stress. GLIS3 knockdown (KD) in INS-1E cells, primary FACS-purified rat beta cells, and human islet cells decreased expression of MafA, Ins2, and Glut2 and inhibited glucose oxidation and insulin secretion, confirming the role of this transcription factor for the beta cell differentiated phenotype. GLIS3 KD increased beta cell apoptosis basally and sensitized the cells to death induced by pro-inflammatory cytokines (interleukin 1β + interferon-γ) or palmitate, agents that may contribute to beta cell loss in respectively type 1 and 2 diabetes. The increased cell death was due to activation of the intrinsic (mitochondrial) pathway of apoptosis, as indicated by cytochrome c release to the cytosol, Bax translocation to the mitochondria and activation of caspases 9 and 3. Analysis of the pathways implicated in beta cell apoptosis following GLIS3 KD indicated modulation of alternative splicing of the pro-apoptotic BH3-only protein Bim, favouring expression of the pro-death variant Bim S via inhibition of the splicing factor SRp55. KD of Bim abrogated the pro-apoptotic effect of GLIS3 loss of function alone or in combination with cytokines or palmitate. The present data suggest that altered expression of the candidate gene GLIS3 may contribute to both type 1 and 2 type diabetes by favouring beta cell apoptosis. This is mediated by alternative splicing of the pro-apoptotic protein Bim and exacerbated formation of the most pro-apoptotic variant Bim S.
Pancreatic beta cell dysfunction and death is a central event in the pathogenesis of diabetes. Genome-wide association studies have identified a large number of associations between specific loci and the two main forms of diabetes, namely type 1 and type 2 diabetes, but the mechanisms by which these candidate genes predispose to diabetes remain to be clarified. The GLIS3 gene region has been identified as a susceptibility risk locus for both type 1 and type 2 diabetes—it is actually the only locus showing association with both forms of diabetes and the regulation of blood glucose. We show that decreased expression of GLIS3 may contribute to diabetes by favouring beta cell apoptosis. This is mediated by the mitochondrial pathway of apoptosis, activated via alternative splicing (a process by which exons are joined in multiple ways, leading to the generation of several proteins by a single gene) of the pro-apoptotic protein Bim, which favours formation of the most pro-apoptotic variant. The present data provides the first evidence that a susceptibility gene for diabetes may contribute to disease via regulation of alternative splicing of a pro-apoptotic gene in pancreatic beta cells.