Purchase this article with an account.
Monique Courtenay, William Cade, Stephen Schwartz, Jaclyn Kovach, Anita Agarwal, Gaofeng Wang, Jonathan Haines, Margaret Pericak-Vance, William Scott; Interaction analysis of exogenous estrogen in age-related macular degeneration (AMD) finds novel associations within the VEGF, Complement, and TGFB pathways. Invest. Ophthalmol. Vis. Sci. 2013;54(15):6171.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
AMD’s multifactorial etiology includes genetic and environmental risk factors. Recently the AMD Gene Consortium identified new genes that implicated the Complement, VEGF, and TGFB pathways for further analysis. Also, it has been shown that women who take estrogen in the form of hormone replacement therapy (HRT) or birth control pills (BCPs) have reduced risk of AMD. The purpose of this study was to test the association of these pathways with AMD while accounting for main effects and estrogen use.
White female AMD cases (n=540) and controls (n=239) were genotyped using Affymetrix 6.0 chipsets. History of estrogen use was collected with a self-administered questionnaire. SNPs within 20kb of each gene in the 3 pathways were chosen. Gene and pathway-based test statistics were calculated as the mean of all independent SNP (r2<0.8) chi-square statistics derived from Kraft’s 2 degree of freedom joint test of genetic and environmental factors. The joint test utilized a likelihood ratio test comparing a logistic regression model with SNP and estrogen main effects, a pairwise interaction term, age, and smoking status, and a model excluding the SNP and interaction terms. The analysis was repeated 10,000 times, permuting phenotypes with the Biased Urn sampling procedure to generate an empirical p-value.
Four known (CFH, CFB, C2, and VEGFA) and 30 novel genes were associated with AMD (p<0.05) when accounting for estrogen. Fifteen were not significant in tests of main effects suggesting their effects are moderated by estrogen. The VEGF pathway was significantly associated with neovascular AMD but only when accounting for BCP use (P= 0.013). Complement was also associated with AMD, but its results were driven solely by the main effect of CFH. The TGFB pathway was not associated with AMD when accounting for estrogen use or main effects only.
Our pathway results suggest that estrogen modifies the genetic effect of SNPs within the VEGF pathway. This pathway and all novel genes warrant closer examination for AMD association in the context of estrogen usage. These results illustrate the utility of exploring pathways of previously associated genes to find novel associations. This study also demonstrates the importance of incorporating environmental exposures in tests of genetic association, at the SNP, gene, or pathway level.
This PDF is available to Subscribers Only