INTEGRATIVE GENOMIC PROFILING TO IDENTIFY GENES AND PATHWAYS ASSOCIATED WITH VASCULAR RISK IN AD

Abstract

Amyloid-beta peptide (Ab) can be found deposited in the brain cerebrovasculature; known as cerebral amyloid angiopathy (CAA). Presence of CAA in AD patients has been associated with greater ante-mortem cognitive impairment, independent of the underlying AD neuropathological burden. The APOEe4 allele and male sex have previously been reported to be associated with increased CAA. Identifying genetic risk factors and transcriptomic changes associated with CAA, in the context of APOE and Sex, may nominate novel genes and pathways underlying this vascular component of AD etiology; providing key insights necessary for the development of targeted therapies and/or biomarkers. Genome-wide genotypes (Infinium Omni 2.5 Exome 8 v1.3 genotyping array) and brain gene expression measures (RNAseq) were collected for a cohort of 477 AD cases with CAA scores. Following QC, and imputation to the haplotype reference consortium (HRC) panel, 5,814,739 variants were tested for association with CAA, using linear regression, adjusting for Age, Sex and population substructure (Eigenstrat). RNA sequencing (RNAseq) measures (Illumina TruSeq, 100bp PE, HiSeq4000) were collected from temporal cortex (TCX) for all 477 AD samples, and from Cerebellum (CER) tissue for a subset of the cohort (N=200). We confirmed APOEe4 dose (b = 0.21, p-value = 4.85E-14) and male sex (b=0.11, p-value = 3.68E-03) as significant variables associated with greater CAA pathology amongst AD cases. Sex stratified analysis, likewise confirmed more significant association for APOEe4 dose in females (b = 0.23, p-value = 7.16E-09) than males (b = 0.19, p-value = 1.16E-06). Genome-wide association analysis (GWAS) identified variants at 184 independent loci with a smaller p-value than those at the APOE locus in males. Additional variants with small p-values, but do not meet genome-wide significance, were also found; including variants at the PLCG2 locus, where rare coding variants were recently implicated in AD. Further insights will be gained by integration of genetic association results with expression profiling studies, currently underway. Pathway enrichment analysis will be used to identify biological pathways implicated by transcriptomic and/or genetic association with CAA. We are also seeking replication of GWAS results in an additional independent cohort of AD patients scored for CAA.