African Ancestry Proportion Influences Ileal Gene Expression in Inflammatory Bowel Disease

Abstract

The influence of ancestry on inflammatory bowel disease (IBD) susceptibility has recently been examined via several large-scale genome-wide association studies,1Huang C. et al.Gastroenterology. 2015; 149: 1575-1586Abstract Full Text Full Text PDF PubMed Scopus (50) Google Scholar,2Brant S.R. et al.Gastroenterology. 2017; 152: 206-217Abstract Full Text Full Text PDF PubMed Scopus (86) Google Scholar but the effects of ancestry-specific variation in risk and modifier gene expression on prognosis in IBD remain poorly characterized. It has been estimated fairly consistently that the prevalence of IBD is approximately 2 to 3 times greater in individuals of European versus African American (AA) descent,3Nguyen G.C. et al.J Crohns Colitis. 2014; 8: 288-295Abstract Full Text Full Text PDF PubMed Scopus (111) Google Scholar but the literature offers conflicting reports on complications and outcomes in AA versus European ancestry patients.4Walker C.H. et al.Gastroenterol Rep (Oxf). 2017; 5: 288-292Crossref PubMed Scopus (6) Google Scholar,5Kugathasan S. et al.Lancet. 2017; 389: 1710-1718Abstract Full Text Full Text PDF PubMed Scopus (335) Google Scholar AA individuals tend to be admixed, with approximately 80% African (YRI) and 20% European ancestry (CEU).6Tishkoff S.A. et al.Science. 2009; 324: 1035-1044Crossref PubMed Scopus (1057) Google Scholar Although genome-wide association studies have identified hundreds of variants associated with IBD risk, most research has been conducted on cohorts of exclusively European ancestry.7de Lange K.M. et al.Nat Genet. 2017; 49: 256-261Crossref PubMed Scopus (539) Google Scholar Earlier studies on IBD risk variants, such as NOD2, concluded that mutations in AA individuals result from European admixture, and thus confer similar increases in risk.8Adeyanju O. et al.Inflamm Bowel Dis. 2012; 18: 2357-2359Crossref PubMed Scopus (16) Google Scholar However, the largest genome-wide association studies to date of IBD in AA identified 2 novel African-specific loci associated with IBD, hinting at the existence of African-specific contributions to disease that remain to be elucidated.2Brant S.R. et al.Gastroenterology. 2017; 152: 206-217Abstract Full Text Full Text PDF PubMed Scopus (86) Google Scholar In this study, we evaluated ileal transcriptomic profiles of 154 individuals of AA and European ancestry with IBD, and characterized differential gene expression between populations. We then examined the effect of proportions of African and European ancestry in AA patients on gene expression, demonstrating that observed variation in gene expression between populations is heritable and not solely caused by environmental differences. We performed RNAseq of ileal biopsies sampled from control subjects (n = 25, no intestinal inflammation and normal histology) and 129 patients with ulcerative colitis (n = 36) and Crohn’s disease (n = 93). Differential gene expression analysis revealed 1360 upregulated and 1345 downregulated genes at an FDR cutoff of 0.05 in AA compared with European ancestry patients (Figure 1A). Hierarchical clustering based on these 2705 genes shows separation of transcriptomic profiles by ancestry into 2 clusters (Figure 1B). To explore functional pathways implicated by differentially expressed genes, we performed Gene Set Enrichment Analysis. Oxidative phosphorylation, adipogenesis, and xenobiotic metabolism were among the gene sets enriched for genes upregulated in AA, whereas tumor necrosis factor-α signaling, inflammatory response, and interferon-γ response were enriched in downregulated genes (Figure 1C, Supplementary Table 1). Each of these pathways is highly relevant to the development and progression of pathology in IBD, which highlights the importance of better understanding of genetic contributions to disease and possible personalized treatment options. Following differential expression analyses, we sought to validate that the population-based variation in gene expression we observed is heritable and not solely attributable to environmental differences. By studying an admixed population, we can contrast the influences of European and African ancestry among individuals of mixed ancestry to the difference between the 2 predominant ancestry groups. Proportions of admixture in AA were estimated from common genotypes called from the RNAseq reads, using ADMIXTURE software9Alexander D.H. et al.Genome Res. 2009; 19: 1655-1664Crossref PubMed Scopus (4073) Google Scholar with 1000 Genomes CEU and YRI individuals as reference European and African populations (Figure 2A). AA who group with European ancestry individuals in cluster 2 exhibit a highly significant (P = 6×10-5) trend of increased proportions of European admixture compared with AA grouping in cluster 1 (Figure 2B). We then applied methods introduced by Price et al10Price A.L. et al.PLoS Genet. 2008; 4e1000294Crossref PubMed Scopus (70) Google Scholar to estimate the percentage of gene expression variation between populations that can be attributed to genetic effects. They argued that the slope of the regression of gene expression against ancestry proportion should be the same as the regression of the difference between the 2 predominant ancestry groups if the effect of ancestry is purely genetic, whereas if the 2 measures are uncorrelated it is purely environmental. The slope c of the 2 regressions assessed across all genes thus provides an estimate of the average heritability. By applying this method, we obtained c = 0.43, which we validated as statistically significant (P = .05) relative to 1000 permutations (Figure 2C, Supplementary Table 2). This value implies a substantial heritable component to differences in gene expression observed between AA and European ancestry IBD patients. In summary, our study shows strong differential gene expression in key pathways based on African versus European ancestry. We further demonstrate that this variation in gene expression among populations can be partially attributed to heritable, genetic effects, rather than solely to differences in environmental factors. The pathways highlighted here are known to be critical in IBD pathogenesis, and elevation of inflammatory and tumor necrosis factor-α signaling seems to be consistent with evidence of worse prognosis in AA. Further investigation into ancestry-specific variation in disease is necessary for the development of personalized therapeutics. Download .pdf (.25 MB) Help with pdf files Supplementary Figure 1 and Supplementary Tables 1 and 2 Download .pdf (.13 MB) Help with pdf files Supplementary Data

Topics & Concepts