Genetic analysis of type II diabetes

National Human Genome Research Institute

Investigators

Linked publications, trials & patents

Abstract

Defining the functional consequences of >600 T2D GWAS loci and >1000 QTLs is challenging. We have made significant contributions to identifying the mechanism by which non-coding variants confer increased risk for T2D. One approach is to quantify allelic contribution of heterozygous SNPs in regulatory regions to open chromatin accessibility. We generated the largest set of ATAC-seq data in skeletal muscle, pancreatic islet, adipose, and liver in a total of 493 subjects. We performed ATAC-seq allelic imbalance analyses and identified tens of thousands of SNPs in regulatory regions that break down linkage disequilibrium of T2D GWAS associated loci and identified likely causal variants at over 1/3 of GWAS association signals. In collaboration with the New York Stem Cell Foundation, we aim to identify genetic factors associated with differentiation of induced pluripotent stem cells (iPSCs) to insulin secreting beta cells. The TCF7L2 locus contains one of the first identified and strongest common variant genetic associations with type 2 diabetes (T2D). Recent fine-mapping studies have refined this signal to a single candidate causal intronic variant, rs7903146. We generated isogenic iPSCs representing the full allelic series of rs7903146 (CC, CT, TT) as well as heterozygous and homozygous TCF7L2 loss of function (LoF). We differentiated these lines to stem cell-derived islet-like organoids (SC-islets) that contain beta cells. Throughout differentiation, we conducted extensive phenotyping, including (i) flow cytometry and immunofluorescence staining for critical, stage-specific marker genes at the definitive endoderm (S1) and pancreatic progenitor (S4) stages, (ii) cell painting at S1 and S4, (iii) multimodal single-cell sequencing of RNA and chromatin accessibility at S1, S4, and SC-islets, (iv) brightfield morphology imaging of SC-islets, and (v) insulin secretion assays of SC-islets. Our initial results confirm a large effect of TCF7L2 LoF mutations consistent with the essential role of TCF7L2 during beta cell differentiation. We also observe a strong effect of rs7903146 at later stages of differentiation, including SC-islet morphology and insulin secretion capacity. These results comport with previous studies that suggest rs7903146 resides within a pancreatic islet beta cell specific enhancer. To understand genetic effects that drive beta cell differentiation, we are also undertaking a GWAS study with several hundred iPSC lines, including those previously generated by other groups (i.e., GENESIS and iPSCORE). As an initial analysis, we analyzed the genetics of iPSC growth rates. Briefly, we measured the growth of iPSC lines derived from 602 unique donors using high-throughput time-lapse imaging, quantified proliferation through a growth Area-Under-the-Curve (gAUC) phenotype, and correlated gAUC with the gene expression and genotype of the cell lines. We identified 3,091 genes associated with gAUC, many of which are well established regulators of cell proliferation. We also found that rare deleterious variants in WDR54 were associated with reduced iPSC growth and that WDR54 was differentially expressed with respect to gAUC. Although no common variants showed a genome-wide association with gAUC, iPSC lines from monozygotic twins were highly correlated, and common genetic variation explained approximately 71-75% of the variance in iPSC growth rates. These results indicate a complex genetic architecture of iPSC growth rates, where rare, large-effect variants in important growth regulators, including WDR54, are layered onto a highly polygenic background. These findings have important implications for the design of pooled iPSC-based studies and disease models, which may be confounded by intrinsic growth differences. It has been shown that N6-methyladenosine (m6A) of RNAs coding for genes that regulate insulin secretion were reduced in T2D islets compared to controls. Direct RNA nanopore sequencing (DRS) potentially enables measuring all 13 known mammalian RNA modifications simultaneously, preserving modifications present in the native RNA. We examined m6A, 5-methylcytosine (m5C), inosine, and pseudouridine RNA modifications and expression levels in two human pancreatic beta-cell lines, EndoC-BH1 and EndoC-BH3, after one hour of glucose stimulation. We identified 1,697 differentially modified sites (DMSs) across all modifications. These DMSs were largely independent of changes in gene expression levels and enriched in transcripts for type 2 diabetes (T2D) genes. Our study demonstrates RNA modification changes in response to cellular stimuli at the single-nucleotide level and provides new insights into RNA-mediated mechanisms that may contribute to normal beta-cell response and potential dysfunction in T2D. In collaboration with Drs Vukašin Jovanović and Carlos Tristan, Claudia Doege, and Shuibing Chen, we are investigating the role of the hypothalamus in glycemic homeostasis. We have differentiated iPSC lines derived from T2D and non-T2D individuals to hypothalamic arcuate-like neurons and characterized these lines using single-nuclei multiomic sequencing. We have identified hundreds of differentially expressed genes and differentially accessible regions between cells generated from diabetic and healthy individuals. Additionally, iPSC lines with knockouts (KO) of T2D risk associated genes (TCF7L2, IGF2BP2) have been differentiated to arcuate neurons. We detected that the KO of TCF7L2 resulted in severe loss of key marker genes in the arcuate-like neuron (POMC), suggesting importance of TCF7L2 in development of hypothalamic neurons. The differentiated neuronal lines will be used in in vitro functional analyses to assess genomic and proteomic changes in response to perturbations that may affect cellular metabolism. To spatially contextualize genomic features of diabetes, we generated a rigorously validated spatial transcriptomic dataset from human pancreas samples spanning normal glucose tolerance and T2D. We established computational approaches to correct for technical artifacts (such as transcript bleed) and published raw and processed data as an open-access resource. Building upon this spatial transcriptomics dataset, we investigated how islet size and cellular composition—particularly the proportions of alpha, beta, and delta cells—correlate with transcriptomic profiles, highlighting the unique molecular signature of small islets primarily composed of insulin-secreting beta cells and the depletion of non-beta endocrine cells in these structures. Finally, we have also contributed analyses to improve our understanding of T2D-associated loci. In collaboration with Profs. John Danesh and Adam Butterworth (Cambridge), we performed a T2D GWAS in 75,000 people from Bangladesh and combined summary data with two previous studies, identifying 52 T2D-associated loci. After fine-mapping, we found three novel, protective signals enriched in Bangladeshis, including two at the HNF4A locus. Given the known role of HNF4A in beta cell function, we are characterizing the functional impact of these variants during beta cell differentiation using the same isogenic iPSC model described above for rs7903146 at TCF7L2. Also, in collaboration with Dr. Joshua Denny (NHGRI), we are exploring the role of structural variants (SVs) at T2D-associated loci in the NIH All of Us Program. We imputed genotypes for >100,000 SVs, tested for association of SVs and nearby SNPs with T2D, and performed fine-mapping at associated loci. These analyses are still underway, but preliminary results nominate a 53 base pair deletion that may underlie the T2D association at LAMA1, highlighting a novel mechanism at T2D-associated loci. Our goal going forward is to identity genetic and epigenetic changes in multiple tissues relevant to T2D and to determine their correlation with diabetes and diabetes-related phenotypes.