The role of common and rare variants in risk for Alzheimer's disease across populations

Abstract

Until recently, APOE was the only well-established risk gene for late-onset Alzheimer disease (LOAD). Efforts to identify additional LOAD loci have concentrated on testing the common disease-common variant hypothesis by conducting genome-wide association studies (GWAS) using large case/control datasets. Thus far, these studies, conducted primarily in non-Hispanic white populations, have identified robust associations of AD with non-coding common variants (CVs) in 21 genes. The effect sizes for these variants are small and account for only a portion of the heritability of LOAD. An alternative hypothesis suggests that multiple rare variants (RVs) in these and other genes contribute to AD risk, each with stronger individual effects than CVs. Moreover, although nearly all of the known functional RVs linked to AD are deleterious, the observation that some individuals with high genetic liability (e.g., APOE ε4/ε4 genotype) remain cognitively intact after age 85 indicates that some AD-related CVs and RVs may be protective. This presentation will summarize the evidence emerging from GWAS, candidate gene and next generation sequencing (NGS) studies in diverse ethnic populations supporting roles for CVs and RVs in LOAD. The effect of some loci are age and sex dependent, whereas others are modulated by epistatic influences of other loci or environmental factors Trans-ethnic studies have yielded examples of variants influencing LOAD risk in multiple populations (albeit with variable effect sizes), allelic heterogeneity and locus heterogeneity. Interpretations are complicated by variable risk allele frequencies and linkage disequilibrium patterns across populations. Recent studies using NGS approaches have identified highly penetrant risk and protective RVs, some of which appear to be population-specific. Ambitious studies such as the United States AD Sequencing Project will likely uncover many additional genes and RVs for AD. Very large samples will be needed to verify association with RVs and small-effect CVs, as well as demonstrate gene-gene and gene-environment interaction. Trans-ethnic studies may be very useful to localize complex association signals and add power because of increased effect sizes (CVs) and frequencies (RVs) in some groups. These studies will likely suggest new pathways and novel targets that may lead to new therapeutic and preventive strategies.