Sex-heterogenous effect on Alzheimer's disease risk at the BIN1 locus.

Abstract

Among Alzheimer's Disease (AD) tier 1 genes, BIN1 shows the greatest sex-biased expression in GTEx RNASeq, notably in brain tissues. Fine-mapping studies suggest that the BIN1 locus harbors at least two independent risk variants. We considered a region ±200kb around BIN1 and performed sex-stratified analyses to identify genome-wide significant variants with a sex-heterogenous effect in imputed data from the AD Genetics Consortium. We ran conditional analyses on rs6733839 to show that variants with sex-heterogenous effects were independent from the lead variant at this locus. Additionally, we performed sex- and rs6733839-genotype-stratified analyses to understand which haplotype drives this sex-heterogenous effect on AD risk and on BIN1 expression in brain tissue from the ROSMAP study. Rs10200967 has a significant sex-heterogenous effect on AD risk and is genome-wide significant in females but not males (Table 1). In the conditional analysis the association remains significant (pfemale = 6.5×10-3 , Table 2). The linkage disequilibrium between these two variants is low (r2 = 0.12). The protective association of rs10200967 is strongest in females homozygous for the major allele of rs6733839 (p = 1.1×10-3 ). Among individuals homozygous for the major allele of rs6733839, the effect of the interaction between rs10200967 dosage and sex on AD risk is significant (p = 3.2×10-3 ). In the full sample, the three-way interaction between these two variants and sex is significant (p = 0.021, Table 3). The rs10200967 minor allele is associated with an increased expression in GTEx (p = 6.0×10-15 , Figure 1) and ROSMAP (p = 9.1×10-3 , Table 4). Among rs6733839 reference allele homozygotes, the rs10200967 interaction with sex on BIN1 expression is significant (p = 0.0495). In the full ROSMAP sample, the three-way interaction is trending significant (p = 0.062, Table 5). Interestingly, rs10200967 is located in a histone peak and a start-exon of a BIN1 transcript (Figure 1) reinforcing its putative regulatory role. Our sex- and rs6733839-genotype stratified analyses, demonstrate that rs10200967 at the BIN1 locus is genome-wide significant, with a sex-heterogenous effect on AD risk and on BIN1 expression. These results support the growing consensus that there are two separate signals at the locus and suggest that rs10200967 contributes to the signal independent of rs6733839.