Genetic analysis of type II diabetes in Finnish population

National Human Genome Research Institute

Investigators

Linked publications, trials & patents

Abstract

We have the ability to resample study participants, utilize novel genomic and epigenomic tools, and integrate multiple data types. With these capabilities, we have shown that stretch enhancers (regulatory enhancers >3kb in length) correlate with gene expression in a tissue-specific manner and are enriched with disease-associated GWAS variants. To further define the islet epigenome, we integrated whole genome transcriptomic data (RNA-seq) from 185 cadaveric islets and chromatin accessibility profiles (ATAC-seq) from two islet samples and show that T2D associated genetic variants are enriched in islet-specific regulatory regions. We also contributed our pancreatic islet reference chromatin analyses and gene expression data to the international InsPIRE consortium. This integrated meta-analysis of 420 pancreatic islet samples led to the identification of candidate effector transcripts at 23 T2D loci. The human pancreatic islet is composed of several cell types, with the insulin secreting beta cells representing only 40% of the population. To assess cell type specific gene expression under basal or environmental stimuli (i.e., high/low glucose), we have performed single-cell RNA-seq (scRNA-seq) of islets from individual cadaveric donors. Furthermore, we are utilizing spatial transcriptomics to determine the location of cell types within the structure of the pancreatic islet and the impact of cellular architecture on gene expression profiles. We have collected skin, muscle, and adipose biopsies from >300 well phenotyped and genotyped individuals with normal glucose tolerance, impaired glucose tolerance, or T2D. Analyses of the muscle RNA-seq data set have identified many expression quantitative trait loci (eQTLs), including some that link T2D-GWAS variants to their target genes. Current analyses of muscle and adipose RNA-seq data from the same individuals have identified eQTLs from both tissues colocalizing with T2D-GWAS signals. We have collected metabolomics data on 318 muscle and 309 adipose biopsy samples as well as complex lipid analysis of >1390 plasma samples taken during an oral glucose tolerance. Together with RNA-seq data, and we are integrating metabolomics data with genetic architecture, T2D-related traits and eQTLs to investigate potential dynamic metabolic interactions to identify and clarify disease mechanisms. To date, GWAS with the plasma metabolite data has identified >2000 genetic associations with T2D and/or T2D-related traits. In a recent collaborative effort, we identified genetic association for 1400 plasma metabolites sampled from a cohort of >6000 Finnish men (METSIM)1. Over 300 of these are novel associations potentially identifying new genes and mechanisms contributing to risk for diabetes or related traits. We have performed DNA methylation analysis in a subset of these human tissue samples. We have integrated genomic sequence, gene expression, and methylation data from 265 skeletal muscle biopsies with their corresponding phenotypes for eight physiological traits (height, waist, weight, waist hip ratio, body mass index, fasting serum insulin, fasting plasma glucose, and T2D). We identified gene and DNA methylation site relationships that may underlie 534 disease/quantitative traits. We have collected 20 liver samples, another diabetes-relevant tissue. As with pancreatic islets, we integrated RNA-seq and ATAC-seq data to identify critical regulators of genes relevant to diabetes risk. We identified >3000 QTLs within regions of open chromatin (caQTL) with an enrichment of QTL variants in liver promoter and enhancer states. Using a combination of various genomic data sets, we predicted target genes for 861 of the caQTL signals. A major obstacle to studying T2D pathogenesis is defining the functional consequences of >240 T2D GWAS loci and >1000 QTLs. We used a machine learning approach to model pancreatic islets enhancers to increase the accuracy of the predicted impact of islet regulatory variants. We have rigorously tested and validated the accuracy of the machine learning predictions using existing datasets and laboratory assays. Another approach involves the analysis of microRNA (miRNA) classes and abundance. miRNA expression is important for pancreatic development and is altered by disease states. Thus, we have characterized miRNA expression in islets from 74 donors, and identified several miRNAs associated with T2D or T2D-related traits. In a collaborative effort with the New York Stem Cell Foundation, we have used 52 skin tissue samples to generate induced pluripotent stem cell lines (iPSC). Our goal is to differentiate the 52 iPSC lines to mature beta cells and access functionality of these cells. For a subset of lines showing the highest and lowest differentiation efficiencies, we will repeat the differentiation process and perform transcriptomic analysis (bulk and scRNA-seq) and open chromatin structure (bulk and scATAC-seq) analyses at critical stages of the differentiation process. This will allow comparative analysis of the effects of genetic background on -cell development and function. This strategy will be also applied to 43 HipSci IPSC lines, each of which harbor a gene mutation known to cause mature onset diabetes of the young. 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 groups external to the FUSION study (i.e., GENESIS and HipSci). IPSC lines from many donors are pooled together and differentiated. The resulted fraction of mature -cells from each donor is quantified. The genetic makeup of the donor lines producing the highest fraction of mature -cells can thus be associated with the highest differentiation potential. To further understand the specific genes required for beta cells differentiation, and complement the GWAS studies, we are preparing to perform a genome-wide CRISPR/Cas9 interference (CRISPRi) screen. Single genes (targeted by a guide RNA) will be inhibited during the differentiation process. Cells not expressing critical genes required for -cell differentiation and maturation will be depleted in -cells compared to all other cell types A more targeted approach to identify variants critical to -cell differentiation or function is to use isogenic lines that differ genetically at only one T2D risk locus. In collaboration with Dr Shuibing Chen (Weill Cornell Medicine) we are using CRISPR prime editing to generate iPSC lines harboring T2D risk or non-risk variants. The edited iPSC lines will be differentiated into beta cells for functional in vitro analyses including glucose stimulating insulin secretion assays and potential high throughput drug screens under various exposures/treatments. To date, we have successfully editing the sequence at 4 risk loci/gene. We are also collaborating with Drs. Shuibing Chen and Stephen Parker to investigate the intrinsic and environmental network signature dynamics of pancreatic beta cell function common in type 1 and type 2 diabetes. Both intrinsic (beta cell in T2D) and environmental (immune cell in T1D) signals play critical roles in pancreatic dysfunction and cell death. We aim to perform comparative analysis on 100 cadaveric pancreatic islet samples under basal conditions for T2D modeling as well as inflammatory cytokine and viral perturbation for T1D modeling. We will generate single-cell resolution multi-omic (scRNA-seq, scATAC-seq) reference maps of cell/context-specific molecular genetic (e/caQTL) network and hub signatures of intrinsic and environmental signals in pancreatic islets. Our goal is to identity genetic and epigenetic changes in multiple tissues relevant to T2D and to determine their correlation with diabetes and diabetes-related phenotypes.