IDENTIFICATION OF AD RISK VARIANTS BY GENOME WIDE ASSOCIATION OF GENE EXPRESSIO

Late-onset Alzheimer's disease (LOAD) has a substantial genetic component estimated to be as high as 80%, much of which remains unexplained. Genome-wide association studies (GWAS) are emerging as a powerful approach in deciphering the genetic risk factors for common-complex diseases. We recently genotyped 318,237 single-nucleotide polymorphisms (SNPs) in 2,465 subjects from LOAD case-control series. Of these subjects, there are 200 pathologically confirmed ADs and 197 non-ADs with cerebellar RNA samples. We measured cerebellar mRNA levels of 12 AD candidate genes in these 200 autopsy-confirmed AD subjects. We extracted the c/s-SNP genotypes for these 12 candidate genes from the GWAS and performed associations utilizing their expression levels as endophenotypes. We identified 3 SNPs that associate significantly with IDE (insulin degrading enzyme) expression levels after correcting for multiple testing. One SNP had IDE expression level association at p=2.73 x 108, which would be significant even at the genome-wide level. Minor allele carriers of all 3 SNPs had IDE expression levels 2.43-2.66 fold higher than the major homozygotes. Minor allele carriers of all 3 /DESNPs had protective odds ratio estimates, as expected biologically from their effects on IDE expression levels. These SNPs were in linkage disequilibrium with IDE SNPs residing in regions conserved between human and mouse. These results suggest the existence of functional IDE variants that modify risk of AD via effects on gene expression. Importantly, they provide strong proof of principle that use of expression levels as endophenotypes may be a powerful approach in the identification of disease susceptibility alleles in GWAS. Our 200 AD cases and 197 non-AD subjects with whole genome SNP genotypes and brain RNA provide a highly valuable resource to pursue whole genome expression analysis. Assessment of whole-genome SNP associations with expression levels will generate a valuable resource for mapping complex diseases. Our data will enable simultaneous assessment of whole-genome variation for their effects on gene expression and the AD phenotype. SNPs that associate with both AD and expression levels will be candidate susceptibility variants for AD with plausible regulatory effects. In this proposal our specific aims are: 1. To obtain whole transcriptome expression levels from subjects with whole genome SNP genotypes. 2. To perform GWAS of whole transcriptome expression levels. 3. To identify and validate variants that associate with both AD risk and gene expression levels. 4. To validate the /DESNP-expression and AD associations. RELEVANCE (See instructions): Alzheimer's disease (AD) is an epidemic that accounts for 60% of all dementias and affects an estimate of 13.5 million individuals worldwide. Understanding the underlying genetics of this common disease will help understand its formation, may provide advancement for its prevention as well as potential drug targets for its cure. Our proposed work is aimed at the discovery of AD susceptibility variants that work through regulation of gene expression using a genome-wide association study design.