Objective: We examined the epistatic effect between haplotypes of glycogen synthase kinase-3beta (GSK3B) gene and microtubule-associated protein Tau (MAPT) gene in Alzheimer's disease (AD).
Methods: A genetic association study of three AD cohorts was made. Linear regression analyses were used to examine effects of MAPT polymorphisms on gene expression and alternative splicing. beta-Catenin levels and signaling were determined using Western blot and luciferase reporter assays in cells transfected with a combination of GSK3B and MAPT complementary DNA.
Results: Consistent interaction between GSK3B and MAPT genes in three late-onset AD cohorts was observed, with the GSK3B haplotype (T-T) significantly increasing the risk for AD in individuals with at least one H2 haplotype (odds ratio, 1.68-2.33; p = 0.005-0.036). The GSK3B haplotype was significantly protective in the Chinese cohort (odds ratio, 0.33; p = 0.016), after adjusting for the effect of age and sex. There are significant differences in in vivo transcriptional efficiency between the two MAPT haplotypes (H1 and H2) as determined by measurement of cerebellar transcripts (p < 0.001). Overexpression of either MAPT or GSK3B resulted in decreased beta-catenin levels compared with a control vector (p < 0.001). Conversely, cotransfection of both of these molecules increased beta-catenin signaling.
Interpretation: Our genetic and biochemical analyses have identified a novel interaction between Tau and GSK-3beta in late-onset AD causative factors. Our data are consistent with an epistatic model of interaction where discordant levels of GSK3B and MAPT gene expression can lead to altered beta-catenin levels and pathogenicity.