Copy Number Variants associated with Glaucoma in the NEIGHBOR Study

Abstract

Purpose: To characterize the role that DNA copy number variants (CNVs) play in the etiology of primary open angle glaucoma (POAG).

Methods: We identified CNVs from the large and well-characterized NEIGHBOR dataset. More than 4500 unrelated Caucasian POAG cases and controls were genotyped on the Illumina Human660W_Quad_v1 BeadChip. Quality control (QC) steps removed samples with standard deviation (SD) of the mean log R ratio greater than 0.4, mean B allele frequency SD>0.1, and number of CNVs >2 SD of the mean across the entire data set. Data from 1156 cases and 1079 controls were analyzed with PennCNV, and 1101 cases and 1036 controls with QuantiSNP. All CNVs >5kb and including >5 consecutive probes were included in a gene-based analysis. Analysis was performed with the entire dataset as well as phenotypic subsets of high tension (HTG) and normal tension (NTG) glaucoma.

Results: After correction for genome wide significance (p<0.15, as implemented in PLINK), 17 genes were significantly associated with the total POAG dataset, 7 in the HTG subset, and 22 in the NTG subset. These included several genes not previously associated with POAG. Deletions of PACS2 are associated in the full dataset (p=0.008 after genomewide correction). PACS2 is involved in caspase-mediated apoptosis, the mechanism of retinal ganglion cell death in glaucoma. Deletions of BRSK2 are associated in the NTG subgroup (p=0.02 after genome wide correction). BRSK2 is involved in cell cycle checkpoint control, as is CDKN2BAS, the locus displaying the strongest genetic association with POAG. We conducted pathway analysis of all implicated genes using Ingenuity IPA software and identified 14 genes in the TGF-β and TNF-α pathways. Both of these pathways have been extensively implicated in the pathogenesis of POAG. Finally, we examined CNVs in known glaucoma-related genes. We found deletions in SIX6 and GAS7 (1 case each) and duplications in TMCO1, GAS7, and CDKN2BAS (1 case each).

Conclusions: CNV analysis of the NEIGHBOR dataset revealed rare CNVs in known POAG susceptibility genes, as well as more common CNVs in novel candidate genes. Pathway analysis of CNVs identifies biological processes involved in POAG pathogenesis. Thus, common and rare CNVs contribute to POAG pathogenesis and merit further consideration and functional evaluation.

Keywords: 539 genetics