Identifying Genes that Underlie Exfoliation Syndrome using Genetically Determined Gene Expression

Abstract

Purpose : Exfoliation syndrome (XFS) is an age-related systemic disorder characterized by excessive production and progressive accumulation of an abnormal extracellular material, with conspicuous ocular manifestation. It is the most common cause of secondary glaucoma resulting in widespread global blindness. The largest GWAS meta-analysis data of exfoliation syndrome of 123,457 individuals by Aung, et al., 2017, identified 7 loci with the strongest signal in chromosome 15 at LOXL1. Biological importance of coding mutations on LOXL1 were elucidated, but mechanisms underlying associations with other variants in the locus are not clear. We hypothesize that correlation of the genetically determined component of gene expression in XFS can provide a powerful method to identify genes involved in XFS.

Methods : We used gene-based method, PrediXcan, for correlating genetically determined gene expression with disease risk. We applied PrediXcan using models trained on 48 GTEx tissues to estimate individual gene expression from GWAS summary statistics of global XFS dataset from Aung, et al. Additionally, a subset of 8056 European-ancestry individual genetic data from Aung, et al., including 2000 new genetic data from Vanderbilt BioVU were analyzed.

Results : On chromosome 15, 23 genes showed statistically significant associations in both single-tissue and multi-tissue PrediXcan analyses. An additional 5 genes were associated in only multi-tissue analysis within the same chr15q22-25 region spanning 3 mb. Additionally, 7 genes on chromosomes 1 (2 genes), 6 (1), 8 (1), 10 (2) and 19 (1) were identified in multi-tissue analysis. PrediXcan in the available tissues supported 6p21 locus as PRRT1 but not the AGPAT1. Thirteen (12 on chromosome 15) of the 35 genes identified in the global dataset were confirmed in a subset of European-ancestry individuals. Conditional PrediXcan analysis of chr15q22-25 region in the European-ancestry dataset confirmed association signals in six genes: LOXL1, STOML1, UBL7, NEO1, SCAMP2 & INSYN1. All except INSYN1 have been identified in ocular tissues.

Conclusions : Gene-based strategies offer powerful methods to prioritize genes for therapeutic targets. Conditional PrediXcan analysis of genetic signals will be confirmed in individuals of Asian ancestry from Aung, et al. If confirmed, our findings would offer novel gene targets for understanding the pathogenesis of XFS and developing therapies.

This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.