Purpose:
Age-related macular degeneration (AMD) is a leading cause of irreversible visual loss in the elderly of European ancestry. Genome-wide association studies (GWAS) have confirmed the genetic susceptibility to AMD, and identified multiple AMD-associated loci on chromosomes 1 and 10, such as the CFH, C2, C3, CFB, ARMS2 and HTRA1 genes. Recent GWAS with thousands of cases and controls discovered several new AMD-associated loci, including rs9621532 near the TIMP-3 and SYN3 region of chromosome 22q12.3. We aim to replicate SNPs of this region, assess and characterize genotype-phenotype correlations.
Methods:
Three independent sample sets included two case-control studies from the National Eye Institute Clinical Center (NEI, n=363) and the Age-Related Eye Disease Study (AREDS, n=483), and a third sample, a nested case-control (n=852) from the Blue Mountains Eye Study (BMES), a population-based study. The cases from NEI and AREDS included geographic atrophy, neovascular and intermediate AMD, while BMES cases included all types in addition to early stage AMD. Controls were matched for age, sex and smoking status in BMES. Confounding factors in NEI and AREDS samples were adjusted by logistic regression. Extracted DNA was genotyped for "TIMP-3" rs9621532, 3 nearby tag SNPs and 4 well-known AMD SNPs. All cases and controls are unrelated, self-identified as white. The associations and model fit estimates were computerized using Golden Helix software and SAS.
Results:
We found no association between TIMP-3/SYN1 SNPs and AMD in all three samples. However, we uncovered a moderate protective association of TIMP-3/SYN1 rs9621532 with neovascular AMD (odds ratio=0.32, 95% confidence intervals 0.12-0.89; P =0.02) in stratified analyses of NEI samples (153 neovascular AMD cases and 142 controls), following a dominant model. The BMES sample showed the same direction of the same association with neovascular AMD as observed in NEI samples but did not reach statistical significance due to small number (32) of neovascular AMD cases. Four previously reported AMD loci in ARMS2/HTRA1 (rs10490924 and rs11200638) and CFH (rs380390 and rs1061170) were all replicated in the three samples. Of the neovascular AMD cases in the NEI sample, rs9621532 variant carriers had a 2 year older median age than wildtype controls. We found no significant interaction between rs9621532 and any of the 4 known loci tested (ARMS2/HTRA1 and CFH ).
Conclusions:
Using the previously verified sample sets, we found that rs9621532 carriers might have a lower risk of developing exudative AMD and exhibit probably later onset of the disease. Further study on other factors that might influence this association is warranted
Keywords: age-related macular degeneration • candidate gene analysis • clinical (human) or epidemiologic studies: risk factor assessment