IDENTIFICATION OF GENES INVOLVED IN PROINSULIN PROCESSING AND SECRETION

Type 2 diabetes (T2D) affects ~500 million individuals worldwide and is projected to rise 46% by 2045. Patient care is complicated by T2D’s heterogeneous nature and substantial costs. As the disease progresses, insulin secretion diminishes, and insulin demand rises in tissues such as the liver, muscle, and adipose. Each tissue plays a distinct role in glucose homeostasis and metabolism regulation. Beta cells, the primary insulin source, attempt to meet this demand, leading to insulin processing errors and elevated proinsulin secretion. Many genes regulating this process are known, but the process is incompletely understood. Family- and population-based studies have demonstrated that genetics contribute to an individual’s risk of developing T2D. Genome-wide association studies identify regions of the genome associated with disease but do not reveal which genes drive these signals or the tissues they act through. Evaluating T2D-relevant traits more likely to act through a specific tissue, such as plasma proinsulin levels and beta cells, increases the likelihood of detecting signals relevant to particular mechanisms, like insulin processing and secretion. A recent genome-wide association study (GWAS) meta-analysis of plasma proinsulin levels identified 36 independent signals at 30 loci. At one locus on Chromosome 12, we investigated if the lead variant, rs150781447, and the nearest gene, TBC1D30, influenced proinsulin secretion. Using CRISPR Cas9 editing, we generated four classes of Tbc1d30 mutant cell lines and found evidence that reduced gene function increased proinsulin secretion, suggesting that TBC1D30 is likely driving the association signal at this locus and that rs150781447 may be the causal variant. Next, we performed a targeted prime editing screen to evaluate ten putative causal genes and two variant substitutions across seven loci. We evaluated 129 guide combinations and achieved editing efficiencies ≥10% for 11 of 12 targets and ≥48% at nine targets (Sgsm2, rs61741902, Madd, Kank1, Larp6, Lrrc49, Stx16, Wipi1, Arsg). We identified guide characteristics correlated with efficient editing and demonstrated a role for Sgsm2 and rs61741902 in insulin secretion. Together, we have established a prime editing strategy to functionally identify genes and variants at GWAS loci and demonstrated a role for two genes in proinsulin secretion.