Region‐Specific Associations of Human Astrocytic Endfoot Genes with Alzheimer’s Disease Neuropathology and Cognitive Decline

Abstract

AbstractBackgroundWe completed a well‐powered bulk transcriptomic analysis of the astrocytic water channel aquaporin‐4 (AQP4) and related glymphatic molecules and discovered significant associations with Alzheimer’s disease (AD) neuropathological and cognitive outcomes. Genes included AQP4 and dystrophin‐associated complex components, namely DMD, DTNA, DAG1, and SNTA1, which encode proteins responsible for anchoring the AQP4 channel to the astrocytic endfoot membrane.MethodLeveraging regional brain RNA sequencing from the ROS/MAP dataset (Table 1), we determined the effects of transcript‐level variation on neuropathological outcomes at autopsy and longitudinal cognition using multiple linear regression models and mixed effects models, respectively. Discrete brain regions included: the dorsolateral prefrontal cortex (DLPFC, N = 921), caudate nucleus (CN, N = 705), and posterior cingulate cortex (PCC, N = 516). Square‐root transformed neuropathological measures of amyloid and tau burden at autopsy were quantified using immunohistochemistry. Cognitive performance was assessed longitudinally with global cognition calculated as a composite score derived from 19 neuropsychological tests. All model covariates included age at death, sex, educational attainment, and post‐mortem interval. Longitudinal models also included latency to death and the fixed and random effect of interval between current and last visit.ResultHigher AQP4 expression in the DLPFC related to higher amyloid pathology (β = 0.21, P.fdr = 0.04). Similarly, higher DTNA expression was associated with higher tau burden (βDLPFC = 0.36, P.fdrDLPFC = 0.008) and a faster rate of cognitive decline (βDLPFC = ‐0.036, P.fdrDLPFC = 0.0004; βPCC = ‐0.037, P.fdrPCC = 0.02). Higher SNTA1 in the CN related to a faster rate of cognitive decline (β = ‐0.038, P.fdr = 0.03). A summary of significant associations is presented in Table 2 and select plots are provided in Figure 2. AQP4, DTNA, and SNTA1 showed comparable associations with clinical AD in independent datasets within the AMP‐AD project (https://agora.adknowledgeportal.org/).ConclusionOur results indicate that gene expression at the AQP4‐dystrophin axis relates to both AD neuropathology and cognition, particularly in the prefrontal cortex. Future analyses will incorporate single‐nucleus measures of gene expression and cell‐type deconvolution to clarify cell‐type specific contributions in this pathway. Together, these results highlight a set of genes relevant to glymphatic function that should be prioritized for future mechanistic studies.