A multi-cohort genome-wide association study in African ancestry individuals reveals risk loci for primary open-angle glaucoma

Abstract

Primary open-angle glaucoma (POAG), the leading cause of irreversible blindness worldwide, disproportionately affects individuals of African ancestry. We conducted a genome-wide association study (GWAS) for POAG in 11,275 individuals of African ancestry (6,003 cases; 5,272 controls). We detected 46 risk loci associated with POAG at genome-wide significance. Replication and post-GWAS analyses, including functionally informed fine-mapping, multiple trait co-localization, and in silico validation, implicated two previously undescribed variants (rs1666698 mapping to DBF4P2; rs34957764 mapping to ROCK1P1) and one previously associated variant (rs11824032 mapping to ARHGEF12) as likely causal. For individuals of African ancestry, a polygenic risk score (PRS) for POAG from our mega-analysis (African ancestry individuals) outperformed a PRS from summary statistics of a much larger GWAS derived from European ancestry individuals. This study quantifies the genetic architecture similarities and differences between African and non-African ancestry populations for this blinding disease.

Keywords: African ancestry; Black; genetic risk factors; genome-wide association study; glaucoma; health disparities; neurodegeneration; ophthalmology; optic nerve; primary open-angle glaucoma.

Figure 1.. Study designs and subject characteristics

POAAGG, Primary Open-Angle African American Glaucoma Genetics; GGLAD, Genetics of Glaucoma in People of African Descent; ADAGES, African Descent and Glaucoma Evaluation Study; AOU, All of Us; GBMI, Global BioBank Meta-Analysis Initiative; EAS, East Asian Ancestry; AFR, African Ancestry; AMR, Admixed American Ancestry; EUR, European Ancestry; SAS, South Asian Ancestry.

Figure 2.. Discovery of known and previously undescribed loci from the discovery mega-analysis of African ancestry individuals

For a Figure360 author presentation of this figure, see https://doi.org/10.1016/j.cell.2023.12.006. The circular plot depicts genes previously associated with POAG (blue) and previously undescribed genes revealed by the discovery mega-analysis (orange). The inner section (black) shows the mega-analysis Manhattan plot in a circular fashion. The green section demonstrates heterogeneity I2 values from sex-stratified meta-analyses. The inner scatter plots display the relationship between the absolute beta and minor allele frequency (MAF) among the mega-analysis data, including African ancestry individuals only and GBMI results for European individuals, East Asian individuals, and cross-ancestry meta-analyses. See also Tables S1–S4 and S7 and Figures S1, S2, and S4.

Figure 3.. Trait co-localization results

(A) Mega-analysis of discovery cohort, colored by trait GWAS data for baseline CDR, with separation of congruent SNPs (increased expression associated with increase in the trait) versus incongruent SNPs (increase in expression associated with a decrease in the trait). (B) Enrichment of SNPs associated with baseline CDR among significant SNPs from the mega-analysis. (C) P-P plot, with congruent SNPs on top and incongruent SNPs at the bottom. (D) Violin boxplot, with rs11824032 genotypes on the x axis and baseline CDR on the y axis. (E) Violin boxplot, with rs10892564 genotypes on the x axis and baseline CDR the on the y axis. See also Table S5.

Figure 4.. Forest plot showing effect estimates for discovery mega-analyses and replication studies, as well as the pooled effect meta-analyses for two replicated risk variants

Pooled estimates for odds ratio and 95% confidence intervals of all replication datasets and meta-analyses (discovery + all replication sets) were calculated by random effect analyses. Results for individual datasets are denoted by rectangles with lines indicating the 95% confidence intervals and black diamonds indicating the final pooled summary values. All replication datasets include GGLAD-2 + PMBB + AOU + MVP datasets. Meta-analyses p values calculated from the ForestPM plot package are denoted on the top of each plot. (A) SNP rs34957764 in the ROCK1P1 gene was tested in all datasets. (B) SNP rs1666698 in the DBF4P2 gene was tested in all datasets. GGLAD-2, Genetics of Glaucoma in People of African Descent; PMBB, Penn Medicine BioBank; AOU, All of Us; MVP, Million Veteran Program. See also Table S6 and Figures S2 and S3.

Figure 5.. Quantitative gene expression analyses results and proposed mechanisms of genes

Solid arrows indicate the proposed mechanisms of variants identified in our dataset and dashed arrows indicate a mechanism that is known in the literature. *ROCK1P1 is a pseudogene and is a result of the partial duplication of the ROCK1 gene. A variant that maps to ROCK1P1 was identified in our discovery mega-analysis and replicated in the AOU dataset. The quantitative expression profiles of the nearest genes to the three likely causal variants from the mega-analysis (ARHGEF12, ROCK1P1, and DBF4P2) were analyzed in human eye tissues (hTMs, iPSC-RGCs, and retinal cells) using quantitative RT-PCR, with and without H₂O₂ treatment. (A) hTMs were treated with or without 100 μM H₂O₂. (B) iPSC-RGCs were treated with or without 650 μM H₂O₂. (C) Human retinal tissues from normal and glaucoma donors were used. GAPDH was used as the housekeeping gene. Data are presented as mean ± SEM. The data were normalized to control cells and analyzed using Student’s t test statistical analysis (*p < 0.05, **p < 0.01, ***p < 0.001). (D) Schematic representation of RhoA/ROCK downstream signaling pathway involving ARHGEF12 and ROCK1 and leading to glaucoma in individuals of African ancestry. We propose that ARHGEF12 functions downstream of the TGF-β and RhoA/Rho-associated kinase signaling pathways to activate ROCK1 via RhoA, which decreases TM plasticity and aqueous humor outflow and increases RGC death and optic nerve atrophy.

Figure 6.. Performance of PRS generated from summary statistics from GBMI (European individuals) and the discovery mega-analysis (African individuals) in independent datasets of African ancestry individuals

Summary statistics from GBMI (European individuals) and discovery mega-analysis (African individuals) were each used to create a PRS. Each PRS was measured in three independent datasets of African ancestry individuals (eMERGE, GGLAD-2, PMBB). (A) The Nagelkerke R² is shown on the x axis, and the test datasets are shown on the y axis, with the p value corresponding to the Nagelkerke p value. (B) The odds ratio (x axis) was calculated by comparing individuals in top 20% PRS with rest of the test datasets (y axis). The p value corresponds to regression p value. See also Table S8.