Complex genetic interactions affect susceptibility to Alzheimer's disease risk in the BIN1 and MS4A6A loci

Abstract

Genetics is the second strongest risk factor for Alzheimer's disease (AD) after age. More than 70 loci have been implicated in AD susceptibility so far, and the genetic architecture of AD entails both additive and nonadditive contributions from these loci. To better understand nonadditive impact of single-nucleotide polymorphisms (SNPs) on AD risk, we examined individual, joint, and interacting (SNPxSNP) effects of 139 and 66 SNPs mapped to the BIN1 and MS4A6A AD-associated loci, respectively. The analyses were conducted by fitting three respective dominant allelic-effect models using data from four independent studies. Joint effects were analyzed by considering pairwise combinations of genotypes of the selected SNPs, i.e., compound genotypes (CompG). The individual SNP analyses showed associations of 18 BIN1 SNPs and 4 MS4A6A SNPs with AD. We identified 589 BIN1 and 217 MS4A6A SNP pairs associated with AD in the CompG analysis, although their individual SNPs were not linked to AD independently. Notably, 34 BIN1 and 10 MS4A6A SNP pairs exhibited both significant SNPxSNP interaction effects and significant CompG effects. The vast majority of nonadditive effects were captured through the CompG analysis. These results expand the current understanding of the contributions of the BIN1 and MS4A6A loci to AD susceptibility. The identified nonadditive effects suggest a significant genetic modulation mechanism underlying the genetic heterogeneity of AD in these loci. Our findings highlight the importance of considering nonadditive genetic impacts on AD risk beyond the traditional SNPxSNP approximation, as they may uncover critical mechanisms not apparent when examining SNPs individually.

Keywords: Aging; Alzheimer’s disease; Compound genotypes; Dementia; Genetic associations; Genetic heterogeneity; Genetic interactions.

Fig. 1

Volcano plot for CompG analyses of the 9591 SNP pairs mapped to the BIN1 gene locus on chromosome 2q14.3. The x-axis displays CompG regression beta coefficients, and the y-axis shows the respective minus-logarithm-base-10 of SGoF-adjusted P-values. The dashed line indicates the significance threshold [i.e., − log₁₀(P_SGoF = 0.05) = 1.3]. The red dots above the cutoff line indicate significant CompG effects whose comprising SNPs were not associated with Alzheimer’s disease (AD) individually. The blue dots above the cutoff line denote significant CompG effects, in which one or both comprising SNPs were associated with AD individually. The light-gray or dark-gray dots below the cutoff line show non-significant CompG effects, in which none or at least one of the comprising SNPs were associated with AD individually

Fig. 2

Volcano plot for CompG analyses of the 2145 SNP pairs mapped to the MS4A6A gene locus on chromosome 11q12.2. The x-axis displays CompG regression beta coefficients, and the y-axis shows the respective minus-logarithm-base-10 of SGoF-adjusted P-values. The dashed line indicates the significance threshold [i.e., − log₁₀(P_SGoF = 0.05) = 1.3]. The red dots above the cutoff line indicate significant CompG effects whose comprising SNPs were not associated with Alzheimer’s disease (AD) individually. The blue dots above the cutoff line denote significant CompG effects, in which one or both comprising SNPs were associated with AD individually. The light-gray or dark-gray dots below the cutoff line show non-significant CompG effects, in which none or at least one of the comprising SNPs were associated with AD individually