Late-Onset Alzheimer Disease Neuropathology Genomic Screen Identifies Novel Loci For Neuritic Plaque And Other AD Neuropathology Features (I5-2.001)

OBJECTIVE: To investigate the genetics of Late-onset Alzheimer disease by reducing phenotypic heterogeneity and studying AD coincident phenotypes. BACKGROUND: To investigate the underlying genetic mechanisms of Late-onset Alzheimer Disease (LOAD), we have performed a genome-wide association study of AD neuropathology and coincident phenotypes, including a neuropathology-confirmed case-control analysis, analyses of coincident features, including neuritic plaques (NP), lewy bodies (LB), amyloid angiopathy (AA), medial temporal sclerosis (MTS), AD Braak stage, and vascular brain injury (VBI). We used this expanded neuropathology approach to limit the effects of phenotypic heterogeneity, and provide additional insights into AD subphenotypes. DESIGN/METHODS: We examined 4914 samples with neuropathology data from 11 datasets in the Alzheimer Disease Genetics Consortium genotyped with high-density chips. Statistical aassociation was performed using logistic regression for binary traits and polytomous logistic regression for ordinal traits, followed by meta-analysis. Subjects examined included 3,887 neuropathologically-confirmed LOAD cases and 1,027 neuropathologically-confirmed cognitive controls. RESULTS: Associations of APOE and BIN1 with LOAD were confirmed. Additionally, several novel LOAD associations were found, including PHF21B (P=2.0×10-8), and SMOX (P=9.0×10-7). Multiple loci were associated with the presence of neuritic plaques, including APOE (P=1.8x10-30), GALNT7 (P=6.0×10-9), ABCG1 (P=8.0×10-9), and a region near LMX1B (P=4.3×10-8). Additional loci were found to be associated with several ordinal neuropathology traits including LB, AA, MTS, VBI, and Braak. Twelve of the 21 genetic risk loci for clinically-defined AD dementia were confirmed in our clinico-pathologic sample, with nine of these showing larger ORs in the clinic-pathologic sample. At the known loci there was strong positive correlation between AD dementia effect sizes and NFT/NP effect sizes, while VBI effect sizes showed a moderate negative correlation; other coincident features showed only nominal association with known loci. CONCLUSIONS: These results confirm several known AD risk loci and implicate novel loci in the etiology of LOAD neuropathology features, particularly neuritic plaque. Additionally, they suggest a role of AD risk loci in the development of VBI, but not other coincident features.