Basic Science and Pathogenesis.

Background: The cerebrovasculature is an essential component of brain homeostasis. Cerebrovascular disorders are associated with an increased risk for neurodegenerative diseases, including Alzheimer's disease (AD). However, the mechanisms by which cerebrovascular dysfunction contributes to neurodegeneration are poorly understood.

Method: We optimized nuclei isolation from human brains to increase the representation of cerebrovascular cells (CVC) and performed single-nucleus transcriptomic profiles (snRNA-seq) of parietal cortex from healthy and AD donors from the Australian Brain Bank Network (ABBN; n = 72 brains), which are richly annotated for cerebrovascular phenotypes, including cerebral amyloid angiopathy (CAA). In parallel, we generated spatially resolved transcriptomic profiles for a subset of these samples. We integrated our new data with seven public snRNA-seq datasets to create a comprehensive cerebrovascular atlas. Simultaneously, we reprocessed public snATAC-seq data from cases and controls to identify regulatory elements in CVC likely mediating AD gene risk.

Result: Our cerebrovascular atlas encompassed ∼133K CVC across five major cell types (endothelial, smooth muscle, pericytes, fibroblasts, ependymal). This high resolution allowed the identification of perturbed CVC transcriptional programs between cases and controls, including pronounced disruptions in cellular homeostasis via heat shock proteins and smooth-muscle-specific downregulation of ANKRD36 in early-onset AD patients. Rare variants in ANKRD36 were recently implicated in AD in a whole-exome sequencing study. We also identified transcriptional signatures associated with CAA in vascular and glial cells. We validated these results using spatial transcriptomics and data from the Seattle AD cohort. Lastly, we prioritized eight independent AD risk loci, including APP and APOE, where at least one fine-mapped risk variant (95% credible set) overlapped a regulatory element active in vascular cells. At the APH1B locus, we identified a vascular enhancer likely regulating TPM1 and USP3. All loci putatively acting through vascular cells were shared with at least one other cell type (e.g., CASS4 with microglia), supporting reports of shared genetic risk between AD and cerebrovascular disease but suggesting the shared risk is mediated in a cell type-specific manner (Figure 1).

Conclusion: Our work provides novel insights into the role of cerebrovascular dysfunction in AD and identifies genes and regulatory elements mediating AD genetic risk through CVC.