Understanding sex‐specific molecular mechanisms in preclinical and Alzheimer’s disease through brain proteomics and transcriptomics

Abstract

AbstractBackgroundSex‐specific molecular mechanisms active during the transition from the preclinical stage to clinically evident Alzheimer’s disease (AD) pathobiology are not yet fully understood. In this study, we used large‐scale multi‐omics data integration of brain proteomics, brain transcriptomics, and genomics to study sex‐specific co‐expression changes related to the AD continuum.MethodsThe study included postmortem brain proteomics samples from the dorsolateral prefrontal cortex (DLPFC) of cases with AD dementia (N = 200, 55% female), cognitively unimpaired individuals with amyloid‐beta pathology (CU A+, N = 206, 70% female), and without (CU A‐, N = 110, 50% female) from the ROSMAP and Banner studies [Figure 1]. Additionally, RNAseq data was obtained from cases with AD dementia (N = 203, 70% female), CU A+ (N = 204, 63% female), and Cu A‐ (N = 125, 56% female) from ROSMAP. We performed co‐expression network analysis to identify functional sex‐specific molecular modules that differentially behave in AD different stages along the AD continuum (AD dementia vs. CU A‐, CU A+ vs. CU A‐). We assessed the preservation of these modules in the brain transcriptome. Additionally, brain protein quantitative trait loci (pQTLs) enrichment analysis was performed to characterize significant modules enriched for AD risk variants. Finally, we assessed the association of sex‐specific modules with AD traits.ResultsProteomic network analysis showed modules significantly associated with pathological traits differentially by sex [Figure 2]. For instance, a module related to cellular localization showed upregulation in AD and CU A+ compared with CU A‐ [Figure 3A]. Association with pathological traits showed a positive correlation with global pathology and a negative correlation with global cognition only in female [Figure 3B]. In males, two modules related to ubiquitin proteolysis were changed in AD dementia and CU A+. Brain transcriptome showed preservation of the cellular localization module in female [Figure 3C]. This module was enriched for AD pQTLs in the brain, including variants from SYN3, FLOT2 and PACSIN2, which are involved in processing of amyloid and vesicle trafficking [Figure 3D] ConclusionsThis multi‐omics system network analysis reveals changes of a cellular localization‐related module, specifically in the females AD brain proteome and transcriptome, which could be used as potential therapeutic targets.