GWAS-by-Subtraction unveil the genetic architecture of non-IOP-dependent mechanism in POAG

Abstract

Purpose : Primary open-angle glaucoma (POAG) is a consequence of both intraocular pressure (IOP) and non-IOP-dependent mechanisms. Although genome-wide association studies (GWAS) on IOP are prevalent, little is known about the genetic architecture of traits affecting POAG other than IOP. In this study, we subtract the genetic effect of IOP from POAG to investigate the genetic architecture of the remaining non-IOP-dependent mechanisms in POAG.

Methods : We used genomic structural equation modelling and prior GWAS of POAG (n=216257) and IOP (n=132116) to estimate SNP associations with POAG variation that is independent of IOP. Post-GWAS analyses were then performed to provide functional annotations to the results. We constructed two genetic risk scores (GRSs) that represent the IOP and non-IOP dependent mechanisms respectively and explored their associations with the pan-retinal vascular morphological features (RVF).

Results : GWAS of a non-IOP-dependent mechanism in POAG contained 15 significant independent SNPs from 6 leading regions. The overall heritability was 2.1%. SNPs located on chromosome 9 and 14 replicated previous GWAS findings on normal tension glaucoma (NTG), while the rest 4 region- rs80233695 (encoding ANAPC1), rs12208086 (encoding STK38-SRSF3), rs7803764 (encoding TMEM196) and rs2667477 (encoding RP11-384P14.1)- were not reported as a significant region for NTG. The non-IOP-dependent mechanism demonstrated a stronger genetic correlation (rg) with the optical morphology than the IOP-dependent mechanism (rg: 0.43 [se=0.07, P=3.53E-08] vs 0.20 [se=0.05, P=3.21E-05] for 'optic cup area'; 0.24 [se=0.07, P=1.28E-03] vs 0.04 [se=0.05, P=0.4] for 'optic disc area' and 0.39 [se=0.08, P=4.88E-07] vs. 0.19 [se=0.05, P=5.32E-05] for 'Vertical cup-disc ratio'), but share less genetic architecture with blood pressure (rg: -0.05 [se=0.03, P=0.08] vs 0.10 [se=0.02, P=2.55E-05]). IOP-dependent mechanism was associated with the pathway of 'regulation_of_vascular_permeability' (P=5.32E-05) and vascular endothelial function, while non-IOP-dependent mechanism tends to be related with retinal vascular morphology as 16 RVFs were associated with non-IOP-dependent GRS but none were with IOP-dependent GRS.

Conclusions : By GWAS-by-Subtraction analysis, we identified different genetic architectures of IOP and non-IOP mechanisms of POAG, which might improve the understanding of the etiology of POAG

This abstract was presented at the 2023 ARVO Annual Meeting, held in New Orleans, LA, April 23-27, 2023.

 

Graphical abstract