Purchase this article with an account.
William K Scott, Omar García Rodríguez, Patrice Whitehead-Gay, Larry D. Adams, Jacob K. Welch, Renee Laux, Milam A Brantley, Jaclyn L. Kovach, Stephen G Schwartz, Anita Agarwal, Jonathan L Haines, Margaret A Pericak-Vance; Three loci associated with risk of advanced age-related macular degeneration (AMD) also influence anti-VEGF treatment response. Invest. Ophthalmol. Vis. Sci. 2019;60(9):2968. doi: https://doi.org/.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Response to anti-vascular endothelial growth factor (anti-VEGF) treatment for choroidal neovascularization (CNV) is variable. Fritsche et al. (2016) demonstrated that 52 single nucleotide polymorphisms (SNPs) in 34 loci account for 50% of the variation in the risk of developing AMD. Most genetic studies of treatment response have examined the association of these known genetic risk factors on treatment response, with few replicated results. We tested these AMD SNPs to clarify the inconsistent results.
We examined the association between 15,161 SNPs in these 34 loci and one-year change in visual acuity (VA) in 90 non-Hispanic white individuals with CNV (107 eyes) treated monthly for the first three months and as needed afterward. Samples were genotyped using the Illumina HumanCoreExome-24 and imputed against the 1000 Genomes Project Phase I reference panel. Association of one-year change in VA (measured as the change in logMAR) with SNP dosage was analyzed using linear mixed models, adjusting for age, sex, treatment, correlation of paired eyes, and population substructure (the top 3 principal components from EIGENSTRAT analysis).
There were six variants in three AMD risk loci significantly associated with treatment response when correcting for the 34 loci tested (α=1.5 x 10-3). Each C allele at rs1951409 (p=9.62 x 10-4) and each A allele at rs199555630 (p=6.73 x 10-4) in RAD51B (pairwise linkage disequilibrium r2=0.28) were associated with worse VA one year after anti-VEGF treatment. Each T allele at rs117086110 (p=5.21 x 10-5) and each A allele at rs76904544 (p=6.00 x 10-5) in RAD51B (r2=0.94) were associated with improved VA. There was no linkage disequilibrium between the positively and negatively associated SNPs in RAD51B. Two other variants: rs2233739 (p=3.51 x 10-4) in LIPC and rs12519450 in PRLR/SPEF2 (p=1.28 x 10-3) were associated with improved VA.
Previous studies have not associated these six variants with anti-VEGF treatment response in CNV. RAD51B is involved in DNA repair and is associated with several cancers which may also be treated with anti-VEGF therapy. These findings could allow identifying individuals at risk of poorer treatment responses and particular genetic profile, which it could be used to prioritize for new treatments or altered therapeutic regimens.
This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.
This PDF is available to Subscribers Only