Sex differences in the genetic architecture underlying resilience in AD.

Abstract

Approximately 30% of older adults are cognitively normal at death despite presence of Alzheimer's disease (AD) neuropathology at autopsy. Studying these "resilient" individuals may lead to the discovery of novel therapeutic targets. In addition, growing evidence suggests sex differences in downstream neurodegenerative consequences of AD neuropathology, with recent studies highlighting notable sex-specific genetic drivers of AD pathogenesis. We sought to extend this work by elucidating sex-specific genetic factors underlying resilience to AD. We used our published genetic resilience pipeline to assess sex-specific genetic predictors (Dumitrescu et al., 2020). Briefly, we used modern psychometric approaches to harmonize cognitive measures across four cohorts of cognitive aging (N=5054), in-vivo amyloid PET across two studies, and leveraged autopsy measures of amyloidosis (CERAD staging) across two studies. A continuous measure of resilience was quantified using a latent variable framework whereby higher scores reflected better-than-predicted cognitive performance given amyloidosis level and lower scores reflected worse-than-predicted performance. We then performed sex-stratified GWASs and sex-interaction GWASs, covarying for age and the first three principal components and meta-analyzed across cohorts. Finally, we performed sex-stratified genetic correlation analyses (GNOVA) between our meta-analysis results and summary statistics from 63 complex traits. Among individuals with normal cognition, we identified a female-specific locus on chromosome 10 (rs827389, [p(females)=7.4E-09, p(males)=0.64, p(interaction)=8.3E-05, MAF=0.46]). This variant is a modest eQTL for KIN (p=0.003), a gene encoding a DNA/RNA binding protein (http://www.braineac.org). In our genetic correlation analyses, we observed male-specific correlations between resilience and two heart rate-related traits, whereby higher resilience was associated with lower genetic risk for poor heart health. We also observed opposing genetic correlations between resilience and multiple sclerosis such that females with higher resilience scores had lower susceptibility for multiple sclerosis (p.FDR=0.009), whereas males with higher resilience had higher susceptibility (p.FDR=0.001). Our results highlight sex-specific genes and pathways that may drive resilience in a biological sex-dependent manner, although independent replication is needed. The best target to enhance resilience to AD neuropathology may depend on sex and genetic context of an individual. Future work should continue to evaluate sex differences in the genetic architecture of the AD neuropathological cascade.