Mapping adipose and muscle tissue expression quantitative trait loci in African Americans to identify genes for type 2 diabetes and obesity

Abstract

Relative to European Americans, type 2 diabetes (T2D) is more prevalent in African Americans (AAs). Genetic variation may modulate transcript abundance in insulin-responsive tissues and contribute to risk; yet published studies identifying expression quantitative trait loci (eQTLs) in African ancestry populations are restricted to blood cells. This study aims to develop a map of genetically regulated transcripts expressed in tissues important for glucose homeostasis in AAs, critical for identifying the genetic etiology of T2D and related traits. Quantitative measures of adipose and muscle gene expression, and genotypic data were integrated in 260 non-diabetic AAs to identify expression regulatory variants. Their roles in genetic susceptibility to T2D, and related metabolic phenotypes were evaluated by mining GWAS datasets. eQTL analysis identified 1,971 and 2,078 cis-eGenes in adipose and muscle, respectively. Cis-eQTLs for 885 transcripts including top cis-eGenes CHURC1, USMG5, and ERAP2, were identified in both tissues. 62.1% of top cis-eSNPs were within ±50kb of transcription start sites and cis-eGenes were enriched for mitochondrial transcripts. Mining GWAS databases revealed association of cis-eSNPs for more than 50 genes with T2D (e.g. PIK3C2A, RBMS1, UFSP1), gluco-metabolic phenotypes, (e.g. INPP5E, SNX17, ERAP2, FN3KRP), and obesity (e.g. POMC, CPEB4). Integration of GWAS meta-analysis data from AA cohorts revealed the most significant association for cis-eSNPs of ATP5SL and MCCC1 genes, with T2D and BMI, respectively. This study developed the first comprehensive map of adipose and muscle tissue eQTLs in AAs (publically accessible at https://mdsetaa.phs.wakehealth.edu) and identified genetically-regulated transcripts for delineating genetic causes of T2D, and related metabolic phenotypes.

Genetic regulation of adipose tissue transcript expression is involved in modulating serum triglyceride and HDL-cholesterol

“…This study utilized gluco-metabolic phenotype and multiomic data from 260 unrelated non-diabetic individuals from the “African American Genetics of Metabolism and Expression” (AAGMEx) cohort, previously recruited by our laboratory 15, 16 . Studies were performed at the Wake Forest School of Medicine (WFSM) Clinical Research Unit.…”

“…Infinium HumanOmni5Exome-4 v1.1 DNA Analysis BeadChips (Illumina) were used to genotype DNA samples based on the manufacturer’s recommendations. Additional technical details of standard gene expression analyses and genotyping methods are described in our published studies 15, 16 . Expression data was submitted to Gene Expression omnibus (GEO id #GSE95674)…”

“…Data quality control methods are presented in our previously published studies 15, 16 . Briefly, measures of gluco-metabolic phenotypes were examined for outliers in a univariate fashion and as correlated pairs.…”

“…Models included age, gender, and ancestry proportion (admixture) as covariates. As described elsewhere 15, 16 , admixture estimates used in this analysis was computed using the ADMIXTURE program 20 . In a secondary analysis, BMI was used as an additional covariate.…”

“…To identify cis-expression quantitative trait loci (cis-eQTLs), linear regression model was computed with the log 2 transformed expression values as the outcome and an additive genetic model for the SNP as implemented in the R-package MatrixEQTL 23 ; age, gender, and African ancestry proportion were covariates as described elsewhere 15 . SNPs within ±500kb around the respective expressed transcript were considered for this analysis.…”