Genome-wide association analysis of dementia and its clinical endophenotypes reveal novel loci associated with Alzheimer disease and three causality networks of AD: the GR@ACE project.

Background: Genetics plays a major role in Alzheimer Disease (AD). To date, 40 genes associated with AD have been identified, although most remain undiscovered. Clinical, neuropathological and genetic variability might impact genetic discoveries and complicate dissection of the biological pathwaysunderlying AD. Methods: GR@ACE is a genome-wide study of dementia and its clinical endophenotypesthat encompasses 4,120 cases and 3,289 controls from Spain. GR@ACE phenotypes were defined according to AD clinical certainty and the presence of vascular co-morbidity. To explore whether clinical endophenotypesreflect variation in underlying biological pathways, we first assessed the impact of known AD loci across endophenotypesto generate three loci categories. Next, we incorporated gene co-expression data and conducted pathway analysison each category. To assess the impact of heterogeneity in the GWAS findings, the GR@ACE series were meta-analyzed with: 1) genotype-level data from dbGaP (N=21,235); and 2) summary statisticsfrom IGAP Stages I and II (n=61,571 and n=81,455 respectively). Findings: We classified known AD loci in three categories, which might reflect the disease clinical heterogeneity, from vascular and mixed forms to pure AD pathology. Immune system pathways were detected in all categories. Intriguingly, vascular processes were only detected as a causal mechanism in probable AD. A meta-analysis of GR@ACE with additional GWAS datasets revealed the ANKRD31-rs4704171 signal in the HMGCR genomic region. We confirmed NDUFAF6-rs10098778 and SCIMP-rs7225151, which were previously detected by IGAP, to be suggestive signals. We also confirmed CD33-rs3865444 to be genome-wide significant. Interpretation: The regulation of vasculature is a prominent causal component of probable AD. In that context, cerebral amyloid angiopathy, the unique identifiedlink between the vascular and amyloid hypotheses, deserves further investigation. The GR@ACE meta-analysis revealed novel AD genetic signals. GWAS results are strongly driven by the presence of clinical heterogeneity in the AD series.