ZRANB3 is an African-specific type 2 diabetes locus associated with beta-cell mass and insulin response

Abstract

Genome analysis of diverse human populations has contributed to the identification of novel genomic loci for diseases of major clinical and public health impact. Here, we report a genome-wide analysis of type 2 diabetes (T2D) in sub-Saharan Africans, an understudied ancestral group. We analyze ~18 million autosomal SNPs in 5,231 individuals from Nigeria, Ghana and Kenya. We identify a previously-unreported genome-wide significant locus: ZRANB3 (Zinc Finger RANBP2-Type Containing 3, lead SNP p = 2.831 × 10^-9). Knockdown or genomic knockout of the zebrafish ortholog results in reduction in pancreatic β-cell number which we demonstrate to be due to increased apoptosis in islets. siRNA transfection of murine Zranb3 in MIN6 β-cells results in impaired insulin secretion in response to high glucose, implicating Zranb3 in β-cell functional response to high glucose conditions. We also show transferability in our study of 32 established T2D loci. Our findings advance understanding of the genetics of T2D in non-European ancestry populations.

Fig. 1

Manhattan plot of discovery GWAS: the AADM study

Fig. 2

Regional association plots for TCF7L2 and ZRANB3 in the AADM GWAS for T2D

Fig. 3

Generation of zranb3 genomic zebrafish mutants. a, b Identified mutations found in homozygous F2 animals by CRISPR/Cas9 in Tg(ins:mCherry) or wild-type animals. In both lines genomic DNA deletions in exon 4 resulted in frame shift and premature stop codons (red box), confirmed by sequencing. c Confocal microscopy images of ins:mCherry transgene (red) in principal islets of wild type or mutant 5 dpf larvae. d Confocal images of whole-mount immunostained islets of wild type or mutant 5 dpf larvae, using antibodies against insulin (red) or glucagon (green). Scale bars = 25 µM

Fig. 4

Knockdown of zranb3 reduces β-cell number in zebrafish larvae. Injection of antisense oligonucleotide morpholinos (MO) targeting zranb3 reduced β-cells in 5 dpf larval zebrafish, detected by Tg(ins:mCherry) expression imaged via whole-mount confocal microscopy (a, b). c Depth coding of confocal microscopy images of whole-mount islets reveals individual cell resolution of β-cells in 5 dpf larval islets, color coded by depth (distal to proximal) along the z-axis. d, e Whole-mount immunostaining of MO-injected Tg(ins:mCherry) larval islets using antibody against glucagon (green). f Quantification of β-cell number in 5 dpf larval islets of Tg(ins:mCherry) injected with indicated concentrations of MO against zranb. (n = 17–19). Bars represent average across groups. Error bars represent SEM. p Values are from the t test. Scale bars = 25 µM

Fig. 5

Increased islet apoptosis with loss of zranb3. a, b Co-localization in larval islets of activated Caspase-3 with insulin by whole-mount double immunostaining demonstrating significantly higher proportion of embryos exhibiting co-localization in larval (5 dpf) islets compared to controls, as quantified in (c). Data indicate average of three replicate experiments. Each experiment is n = 8–10 embryos

Fig. 6

Suppression of Zranb3 in MIN6 β-cells impairs insulin secretion in response to glucose. a Reduced Zranb3 expression upon treatment of MIN6 cells with siRNA targeting Zranb3; n = 4 independent experimental replicates for each, each dot representing independent experiment and bars represent average values across experiments. b Quantification of total insulin secreted by MIN6 cells into culture media upon low- (2.5 mM) and high-(16.7 mM) glucose conditions. n = 16–24 replicates for each experiment and averages across experiments indicated. Error bars represent SEM

Fig. 7

Significant genes on integrative GWAS and transcriptomic analysis. a Displays genes with gene-set p < 10⁻³ colored in deep red. Transcriptomic data from whole blood (11 studies), “Fat Grouped” (Grundberg et al. and GTeX adipose tissue), adipose tissue (GTeX adipose tissue only), skeletal muscle (GTeX skeletal muscle only). b Plot of gene-based versus best single variant association p values for whole blood