BOSTON—Genome-wide association studies (GWAS) of non-Caucasian and non-European populations could help researchers detect new genes associated with the risk of Alzheimer’s disease, Lindsay A. Farrer, PhD, reported at the 2013 Alzheimer’s Association International Conference. In these populations, GWAS can identify Alzheimer’s disease risk loci that have little or no influence on the risk of disease in Caucasian populations of European descent.

GWAS may be more effective in non-Caucasian and non-European populations when there are multiple functional alleles, said Dr. Farrer, Chief of Biomedical Genetics at the Boston University School of Medicine. The ability to detect association with a particular gene may vary across populations as a result of differences in frequency of the disease-causing alleles and the influence of genetic and environmental backgrounds, he added.

Recent Studies Examined African American and Japanese Persons
Earlier this year, the Alzheimer’s Disease Genetic Consortium conducted a GWAS of a sample of approximately 5,000 African Americans. The sample size may have been smaller than that of recent GWAS of Caucasians because fewer African Americans are available to participate in genetic studies of Alzheimer’s disease, said Dr. Farrer. The investigators observed genome-wide significant associations between Alzheimer’s disease and the APOE and ABCA7 genes. Although ABCA7 also had been associated with the disease in Caucasians, the association among African Americans involved a different single nucleotide polymorphism (SNP). The investigators also found that the association strength was similar for ABCA7 and APOE in African Americans. The ABCA7 SNP associated with Alzheimer’s disease among African Americans was monomorphic among Caucasians, and vice versa, added Dr. Farrer.

Another GWAS conducted this year included 1,000 Japanese patients with Alzheimer’s disease and 1,000 Japanese controls. Aside from APOE, the only gene with a genome-wide significant association with Alzheimer’s disease was SORL1, which Farrer and his colleagues had identified as genetically and functionally associated with Alzheimer’s disease several years earlier in a multiethnic candidate gene study. The effect size of the relevant SNP was approximately the same among Asians and Caucasians, yet the strength of the evidence supporting this association was much greater in Asians than in Caucasians. The fact that the relevant SNP was about 10 times more frequent among Asians than among Caucasians may partially explain this difference, said Dr. Farrer.

GWAS Provides High-Resolution Information About Genetic Variability
GWAS permits researchers to examine genetic variability at an exceptionally high degree of resolution. It also allows researchers to investigate the genome without having a prior hypothesis in mind. “Most of the robust association studies using the GWAS approach in other disorders have identified genes that were not previously suspected to be related to the disease,” said Dr. Farrer. “Many of the localizations have been to regions that are not known to harbor any genes.”

It can be challenging to establish significant results of GWAS because of the large sample sizes required. The first GWAS of Alzheimer’s disease were conducted in 2007. Early studies involved sample sizes smaller than 4,000 individuals, including fewer than 2,080 cases of Alzheimer’s disease. Studies performed in 2009 and 2010 had sample sizes greater than 7,000 individuals and included more than 2,000 cases. These investigations led to discoveries of novel loci with genome-wide significance, including CR1. One recent GWAS examined more than 50,000 individuals in its discovery phase and nearly 20,000 in its replication phase. This study identified a large number of novel loci, including SORL1, that have genome-wide significance.

Because of their large sample sizes, GWAS allows researchers to examine gene variants that are prevalent in the population but have a small effect on the risk for Alzheimer’s disease. Many loci associated with Alzheimer’s disease have been identified this way. On the other hand, it is not feasible to use epidemiologic approaches to find rare gene variants that have a small effect on risk for Alzheimer’s disease, said Dr. Farrer. Instead, the variants could be identified in silico using bioinformatic approaches and studied in vitro.

Most GWAS to date have focused on Caucasians of European origin, but studying non-Caucasian populations could advance scientific understanding of Alzheimer’s disease. Given the established heterogeneity within Caucasian populations, researchers would likely find further heterogeneity across groups that have more distant genetic relationships, said Dr. Farrer. Also, if a gene is associated with Alzheimer’s disease in specific groups, it may have several functional variants that could be revealed through the study of non-Caucasian populations. Finally, the ability to detect an association is based not only on sample size, but also on the risk allele’s frequency and interactions with other genes and environmental factors, which may vary across populations.

—Erik Greb
Senior Associate Editor

Genomic Studies of Non-Caucasians Could Deepen Understanding of Alzheimer's Disease