Differentiation of human induced pluripotent stem cells as a tool to study the effects of type 2 diabetes loci.

National Institute Of Diabetes And Digestive And Kidney Diseases

Investigators

Linked publications & trials

Abstract

A 35-day protocol to reprogram human induced pluripotent stem cells (hiPSCs) into glucose-responsive pancreatic beta-like cells has been developed. Differentiating cells are being used to study the effects of the type 2 diabetes (T2D)-associated variants on beta cell development and insulin secretory function by using CRISPR/Cas9 technology to correct the mutation and observe whether a reversal in the phenotype is seen, in the subjects own genetic background. Following CRISPR/Cas9 editing, single cell cloning, expansion and verification of edit by Sanger sequencing are conducted and clonal cell lines are expanded and characterized for pluripotency and ability to generate pancreatic progenitors using flow cytometry and a hPSC scorecard assay to verify the expression of appropriate markers. Cell lines are also analyzed for potential karyotypic abnormalities and off-target CRISPR effects. AT 7 time points reflecting different developmental stages during the differentiation procedure, cells are collected for RNA analysis. The strongest association for type 2 diabetes (T2D) in Southwestern American Indians maps to intron 15 of KCNQ1 where the lead SNP (rs2299620) is in a region of strong linkage disequilibrium (LD). This region is highly imprinted (genes expressed solely from maternal or paternal allele) and several genes show tissue and developmental stage-specific imprinting. A previous study found that this T2D signal has a parent-of-origin effect, with increased T2D risk and, among normal glucose tolerant individuals a lower insulin secretion, when the risk allele is inherited maternally.To identify the effector gene and causal SNPs at this locus, we utilized our induced pluripotent stem cells (iPSC) to analyze imprinting for KCNQ1, CDKN1C, TRPM5 and TH during different stages of differentiation. KCNQ1 had monoallelic expression during the iPSC stage (day 0), pancreatic progenitor stage (PP, day 10) and endocrine progenitor stage (day 13-16) while there was a gradual loss of imprinting thereafter. CDKN1C had monoallelic expression during all stages of differentiation. In contrast, TRPM5 started losing imprinting during the PP stage and we observed bi-allelic expression during later stages while TH had biallelic expression during all stages. RNA sequencing data from different stages of pancreatic islet development identified 20 genes in a 1.2MB region around the diabetes signal that are expressed during various stages of islet development. We currently believe we have mapped the functional region at KCNQ1 to an 400 bp fragment containing rs2299620 and 3 additional SNPs. TO demonstrate the functionality of this fragment, we used CRISPR-CAS9 to generate iPSCs with targeted edits at these 4 SNPs, and 3 additional SNPs in strong LD, and used these isogenic iPSC-derived pancreatic islets as a model system to study the effect of the variants. Flow cytometry of these islets identified a lower percentage of beta-like cells (cells co-expressing INS and NKX6-1) in islets generated using iPSCs with the risk haplotype, suggesting an effect of these T2D associated SNPs on beta cell mass. Stage specific gene expression analyses identified a significant difference in the fold increase in INS gene expression during the endocrine progenitor (EPs) stage. As endocrine cell type commitment occurs during EP stage and increase in INS expression correlates with beta cell commitment, the lower increase in insulin expression during this stage with the risk haplotype could in part explain the effect of the variants on mature islet composition. In EPs with the risk haplotype, we also identified increased expression of H19 which maps upstream of the SNPs and is known to affect beta-cell mass. Methylation at enhancers regulates gene expression. Therefore, we assessed stage specific methylation at all 13 CpG sites (CpG-func) that encompass the functional region (between rs2299620 and rs74606911) and 3 CpG sites (CpG-ctrl) immediately upstream of rs2299620. We identified dynamic methylation changes at CpG-func whereas no changes at CpG-ctrl (remained hypermethylated, average methylation >90%) during different stages of pancreatic islet differentiation. CpG-func were hypermethylated during the iPSC stage but became hypomethylated during the EP stage. Interestingly, a strong difference in methylation was seen in CpG-func only during the EP stage when compared between the risk and non-risk haplotypes. These data support that the T2D SNPs at KCNQ1 affect islet composition (beta-cell mass) and this effect is manifested early during development, likely via an effect on CpG methylation and gene regulation.