Analysis of neurodegenerative Mendelian genes in clinically diagnosed Alzheimer Disease

Alzheimer disease (AD), Frontotemporal lobar degeneration (FTD), Amyotrophic lateral sclerosis (ALS) and Parkinson disease (PD) have a certain degree of clinical, pathological and molecular overlap. Previous studies indicate that causative mutations in AD and FTD/ALS genes can be found in clinical familial AD. We examined the presence of causative and low frequency coding variants in the AD, FTD, ALS and PD Mendelian genes, in over 450 families with clinical history of AD and over 11,710 sporadic cases and cognitive normal participants from North America. Known pathogenic mutations were found in 1.05% of the sporadic cases, in 0.69% of the cognitively normal participants and in 4.22% of the families. A trend towards enrichment, albeit non-significant, was observed for most AD, FTD and PD genes. Only PSEN1 and PINK1 showed consistent association with AD cases when we used ExAC as the control population. These results suggest that current study designs may contain heterogeneity and contamination of the control population, and that current statistical methods for the discovery of novel genes with real pathogenic variants in complex late onset diseases may be inadequate or underpowered to identify genes carrying pathogenic mutations.
Author summary In this study we provide further genetic evidence of the clinical, pathological and molecular overlap between neurodegenerative diseases. We screened the known Mendelian genes in Alzheimer disease (AD), Frontotemporal Dementia (FTD), Parkinson disease (PD) and Amyotrophic Lateral Sclerosis (ALS) disease in over 13,292 individuals. We report the presence of known pathogenic mutations in AD, FTD and PD genes in both sporadic and familial late onset AD cases and even cognitively normal individuals, pointing to contamination of control cohorts with asymptomatic cases. We also observed an enrichment, although not statistically significant, in most of AD, FTD and PD genes, reinforcing the idea of pathologic crossover among these diseases. However, the fact that genes carrying known pathogenic mutation do not show significant association calls for a reevaluation of current statistical methods.