Alzheimer risk loci and associated neuropathology in a population-based study (Vantaa 85+)

OBJECTIVE: To test the association of distinct neuropathologic features of Alzheimer disease (AD) with risk loci identified in genome-wide association studies. METHODS: Vantaa 85+ is a population-based study that includes 601 participants aged ≥85 years, of which 256 were neuropathologically examined. We analyzed 29 AD risk loci in addition to APOE ε4, which was studied separately and used as a covariate. Genotyping was performed using a single nucleotide polymorphism (SNP) array (341 variants) and imputation (6,038 variants). Participants with Consortium to Establish a Registry for Alzheimer Disease (CERAD) (neuritic Aβ plaques) scores 0 (n = 65) vs score M + F (n = 171) and Braak (neurofibrillary tangle pathology) stages 0–II (n = 74) vs stages IV–VI (n = 119), and with capillary Aβ (CapAβ, n = 77) vs without (n = 179) were compared. Cerebral amyloid angiopathy (CAA) percentage was analyzed as a continuous variable. RESULTS: Altogether, 24 of the 29 loci were associated (at p < 0.05) with one or more AD-related neuropathologic features in either SNP array or imputation data. Fifteen loci associated with CERAD score, smallest p = 0.0002122, odds ratio (OR) 2.67 (1.58–4.49) at MEF2C locus. Fifteen loci associated with Braak stage, smallest p = 0.004372, OR 0.31 (0.14–0.69) at GAB2 locus. Twenty loci associated with CAA, smallest p = 7.17E-07, β 14.4 (8.88–20) at CR1 locus. Fifteen loci associated with CapAβ smallest p = 0.002594, OR 0.54 (0.37–0.81) at HLA-DRB1 locus. Certain loci associated with specific neuropathologic features. CASS4, CLU, and ZCWPW1 associated only with CAA, while TREM2 and HLA-DRB5 associated only with CapAβ. CONCLUSIONS: AD risk loci differ in their association with neuropathologic features, and we show for the first time distinct risk loci for CAA and CapAβ.