Abstract
Purpose:
To utilize a large cohort of extended, multi-generational pedigrees to estimate heritability and test common variants previously associated with age-related macular degeneration (AMD) within the contexts of various phenotypic models.
Methods:
Families of northern European descent with a history of multiple members with AMD were recruited at the Casey Eye Institute (Portland, OR, USA) and expanded by recruitment throughout the United States. Using clinical data and fundus photographs of the macula, AMD was classified and scored based on 6 phenotype models: 3 (liability threshold determined) dichotomous models, 1 polychotomous model, and 2 novel quantitative traits that evaluate drusen morphology and macular pigment changes. Genetic variants (n=24) were chosen due to known AMD association studied in unrelated cases and controls, and genotyped by LGC Genomics (Beverly, MA, USA). Heritability estimates, bivariate phenotype correlation analyses and measured genotype association analyses were performed in SOLAR.
Results:
The cohort comprised 152 pedigrees that included 1,287 ascertained family members. Additive genetic heritability estimates for our 6 phenotype models ranged from 22-49% (p<0.0004). Bivariate phenotype correlation analyses indicated a large proportion (97%) of the additive genetic mechanisms influencing the drusen and pigment quantitative traits were shared. There was nominal evidence (p=0.03) to suggest a mutually exclusive set of genes also influence the phenotype variance observed in these 2 quantitative traits. Six SNPs showed a nominal statistical significant association (p<0.005) in 6/6 AMD models: CFH rs1061170/rs10737680, C2 rs9332739, VEGFA-C6orf223 rs943080, ARMS2 rs10490924 and HTRA1 rs11200628. Two SNPs were significant in 5/6 models: C3 rs2230199 and SKIV2L rs429608; and the PLA2G12A SNP rs2285714 was associated in 4/6 models. The CFH, ARMS2 and HTRA1 SNPs remained statistically significant after our multiple testing criterion (6 models; Bonferroni-adjusted p=0.00083).
Conclusions:
We herein provide quantifiable estimates of AMD heritability in a large cohort of extended pedigrees. The SNPs tested do not explain all of the additive genetic liability observed in our pedigrees suggesting that other, undiscovered, variants influence AMD susceptibility in this cohort. This study design may broaden our insight into the underlying genetic architecture influencing AMD.