Influence of genetic variants on gene expression in human pancreatic islets – implications for type 2 diabetes

Abstract

Most signals detected by genome-wide association studies map to non-coding sequence and their tissue-specific effects influence transcriptional regulation. However, many key tissues and cell-types required for appropriate functional inference are absent from large-scale resources such as ENCODE and GTEx. We explored the relationship between genetic variants influencing predisposition to type 2 diabetes (T2D) and related glycemic traits, and human pancreatic islet transcription using RNA-Seq and genotyping data from 420 islet donors. We find: (a) eQTLs have a variable replication rate across the 44 GTEx tissues (<73%), indicating that our study captured islet-specific cis-eQTL signals; (b) islet eQTL signals show marked overlap with islet epigenome annotation, though eQTL effect size is reduced in the stretch enhancers most strongly implicated in GWAS signal location; (c) selective enrichment of islet eQTL overlap with the subset of T2D variants implicated in islet dysfunction; and (d) colocalization between islet eQTLs and variants influencing T2D or related glycemic traits, delivering candidate effector transcripts at 23 loci, including DGKB and TCF7L2. Our findings illustrate the advantages of performing functional and regulatory studies in tissues of greatest disease-relevance while expanding our mechanistic insights into complex traits association loci activity with an expanded list of putative transcripts implicated in T2D development.

Molecular and functional profiling of human islets: from heterogeneity to human phenotypes

“…A web tool from our own human islet isolation programme (ADI IsletCore) was recently released (www.isletcore.ca, accessed 1 April 2020) and provides publicly available insulin secretion and quality control data on human islet preparations (now standing at 360 donors) in real time, with gene expression panels available for a subset of these. Other initiatives focus primarily on the molecular phenotyping of human islets, such as the Integrated Network for Systematic analysis of Pancreatic Islet RNA Expression (InsPIRE) group [35]. These reveal important insights into islet gene regulation (as previously reviewed [7]) and make important contributions to understanding the impact of islet-localised genetic signals on in vivo glycaemic traits [9], but no programme has yet released data on the combined molecular and functional profiling of a large number of human islet preparations.…”

TCF7L2 plays a complex role in human adipose progenitor biology which may contribute to genetic susceptibility to type 2 diabetes

“…It has been demonstrated through analysis of GWAS metadata of T2D-related traits (38,39) and human physiological studies (40) that the risk allele at this SNV increases T2D susceptibility via impaired insulin secretion. Additionally, this variant was shown to increase TCF7L2 expression in pancreatic islets (31,41). These data have conclusively established that rs7903146 increases T2D risk primarily through islet dysfunction probably driven by increased TCF7L2 expression in pancreatic beta-cells.…”

Pancreatic Islet Transcriptional Enhancers and Diabetes

“…Surprisingly though, in 75% of the loci where at least one risk SNP falls in an islet enhancer, the authors linked it to one or more distal genes [13••]. This observation has been partially supported by human islet eQTL studies, where a few T2D-associated SNPs have been linked to genes other than their closest, including variants at CDC123 (linked to CAMK1D ), ARAP1 ( STARD10 ) and ZBED3 ( PDE8B ) [13••, 57, 58, 59•], and by CRISPR-mediated perturbations of diabetes-associated enhancers [13••].…”

“…A few studies have aimed to link noncoding T2D risk variants to target genes using human islet expression quantitative trait loci (eQTL) analysis [13••, 56–58, 59•], the latest of which included pancreatic islet RNA-seq from 420 donors [59•] and identified candidate effector transcripts for 23 loci. While the results are encouraging, they still fall short from delivering a comprehensive assignment of regulatory variants to effector transcripts.…”