Identifying genes that underlie Primary Open-Angle Glaucoma using genetically determined gene expression

Abstract

Purpose : One of the main limiting factors in developing effective therapies for Primary Open-Angle Glaucoma (POAG) is lack of clear understanding of the molecular mechanism in disease etiologies. There have been half a dozen genome wide association studies (GWAS) on POAG that have cumulatively identified 28 potentially causal variants, but the biological mechanism underlying these associations are not clear. All except one of these GWAS variants are noncoding, and might play a regulatory role. We hypothesized that the correlation of genetically determined component of gene expression with POAG can provide a powerful method to identify genes involved in the etiology of the POAG

Methods : We used two gene-based methods that provide a framework for directly correlating imputed genetically determined gene expression with disease risk, PrediXcan and MetaXcan. We applied PrediXcan using models trained on 44 GTEx tissues to estimate individual level gene expression in 430 cases and 2219 controls (Vanderbilt University Medical Center biobank - BioVU). We applied MetaXcan on GWAS summary statistics data from Cooke Bailey & Loomis et al., 2016 from POAG 3,853 cases and 33,480 controls. In order to aggregate evidence across the different tissues we determined the joint effects of gene expression variation imputed across the tissues using Multi-Tissue PrediXcan, an F-test of a multi-variate regression.

Results : We identified four novel genes that fall in regions with sub-genome wide significant SNPs in GWAS; MEAF6 (p=3.39E-07), RTL10 (p=1.36E-07), TMED8 (p=2.35E-06) and DNALI1 (p=2.01E-12), and confirmed genes that were previously suggested to be causal in POAG and in vicinity of loci identified in GWAS: TMCO1, GAS7 and SIX6. For High Tension Glaucoma we find novel associations with UFSP2 (p=5.3E-8) and confirm associations with TMCO1, while in Normal Tension Glaucoma we find novel associations with PGLS (p=2.9E-17) and confirm association with C12orf23. MEAF6 has been shown in mice to be part of chromatin regulator complex that is crucial for normal vasculature, brain development, erythropoiesis and eye development. It is also implicated in intellectual developmental disorder with dysmorphic facies and ptosis in human.

Conclusions : Our results points to novel pathophysiology that underlie POAG. If confirmed by experiment in model animals, these findings would offer novel gene targets for POAG therapies.

This is an abstract that was submitted for the 2018 ARVO Annual Meeting, held in Honolulu, Hawaii, April 29 - May 3, 2018.