Background: Genome-wide association studies (GWAS) identified a number of single nucleotide polymorphisms (SNPs) that are associated with some aspect of the pathogenesis of Alzheimer's disease (AD) or age-related cognitive impairment. As independent replications remain the only way to validate proposed GWAS signals, we measured SNPs in candidate genes with known involvement in either different aspects of cognition or AD pathogenesis (APOE, BDNF, COMT, SORL1, TOMM40 and KIBRA) and investigate their associations with mild cognitive impairment (MCI) and dementia in this sample. Methods: We analyzed 96 SNPs across six genes in 2,857 women (age>65) from the WHIMS; 552 (19%) were cases of MCI (N=165) or dementia (N=387), the remaining 81% were controls. Procedures for the diagnosis of MCI and AD are described elsewhere (Shumaker et al. 1998; 2003; 2004). A majority of the SNPs were selected from the International Haplotype Map project (www.hapmap.org) to tag the candidate genes; a few were missense SNPs with previously reported associations with related phenotypes. All assays were performed using a custom Illumina GoldenGate assay. SNP associations were evaluated for both dementia and combined dementia+MCI phenotypes using logistic regression under an additive genetic model adjusting for age, race-ethnicity, and hormone therapy.
Results: One SNP in TOMM40 (rs157582) gene was associated with MCI+dementia phenotype, even after correcting for multiple comparisons (P =0.0007). Given the a priori hypotheses involved in selection of candidate genes, an additional 16 SNPs (Figure 1) were found to be associated with combined dementia+MCI phenotype at P = =0.05 (APOE: rs405509, rs439401; TOMM40: rs8106922, rs157580, rs405697; COMT: rs4646316, rs1544325, rs4680; and KIBRA: rs4320284, rs11740112, rs10040267, rs13171394, rs6555802, rs2241368, rs244904 and rs10475878). Dementia results were similar. Conclusions: Our results support the existing view that APOE/TOMM40 genes constitutes a well-established dementia risk (Kehoe et al. 1999) and provide a validation for our general approach to GWAS signal validation, though we fully acknowledge that most of the SNPs associated in our study likely are not causative. Additional loci in COMT and KIBRA genes provide important insights into cognitive impairment associated with aging and may potentially lead to a better understanding of pathology and target pathways for therapy.