211-OR: Massively Parallel Characterization of Thousands of Type 2 Diabetes–Associated Variants in Multiple Cell Types Nominates Candidate Causal Regulatory Alleles and Mechanisms

• Published in: Diabetes (New York, N.Y.) Vol. 73; p. 1
• Main Authors: Kothari, Madhav, Bhuiyan, Redwan M, Ulirsch, Jacob C, Dewey, Hannah, Nerurkar, Niketa, Kursawe, Romy, Kales, Susan, Berenzy, Daniel, Tewhey, Ryan, Stitzel, Michael L
• Format: Journal Article
• Language: English
• Published: New York American Diabetes Association 01.06.2024
• Subjects: Adipocytes; Alleles; Beta cells; Chromatin; Diabetes; Diabetes mellitus (non-insulin dependent); Genome-wide association studies; Hepatocytes; Nominations
• ISSN: 0012-1797; 1939-327X
• DOI: 10.2337/db24-211-OR

Genome-wide association studies have identified >500 loci associated with type 2 diabetes (T2D). Most of these variants are in non-coding regions and may play regulatory roles, but they have not been comprehensively tested for this role. Here, we assessed transcriptional regulatory activity of 6,109 fine-mapped variants in 204 T2D-associated loci from DIAGRAM, UKBB, and BBJ cohorts using massively parallel reporter assays (MPRA) in HepG2 hepatocytes, SGBS (pre)adipocytes, and EndoC-βH3 beta-cells, three cell types implicated in T2D pathophysiology, as well as K562, SK-N-SH, and A549 cells. MPRA identified 946 expression-modulating variants (emVars) in a T2D-relevant cell type. 277 T2D-associated emVars in 81 loci overlapped an open chromatin region in T2D-relevant cell types; precision recall analysis revealed precision of 0.83 and recall of 0.3 for identifying high posterior inclusion probability (PIP > 0.9) variants using this integrated information. Importantly, MPRA nominated as high priority candidate causal variants a single functional variant at 45 loci, the majority of which were incompletely resolved using fine-mapping (PIP>0.9 (n=6), PIP 0.5-0.9 (n=3) and PIP< 0.5 (n=36)). These include well-characterized variants such as rs11257655 (CDC123/CAMK1D locus), whose risk allele increases MPRA activity via enhanced FOXA2 binding, and new variants like rs528350911 (PIP 0.99), whose T2D risk allele decreases MPRA activity through predicted disruption of a GATA binding site, and rs917195 (PIP 0.96), whose T2D risk allele in the CRHR2 locus is predicted to disrupt binding of RFXs. Together, this study provides new insights into the regulatory effects of non-coding T2D-associated variants across multiple cell types and nominates a set of high-confidence candidate causal variants for 39.7% of T2D-associated loci tested.