Multimodal analysis identifies pericyte-centered signaling programs altered by sex and brain region in Alzheimer's Disease

Pericytes are critical components of the neurovascular unit (NVU), regulating endothelial cell (EC) stability, blood-brain barrier (BBB) integrity, and neuroimmune signaling. However, their role in Alzheimers Disease (AD), particularly in the context of sex differences and brain region specificity, remains poorly defined. Here, we use single-nucleus RNA sequencing (snRNA-seq) to characterize transcriptional and intercellular signaling changes in pericytes across the middle temporal gyrus (MTG) and dorsolateral prefrontal cortex (DLPFC) of the same AD and non-AD donors, stratified by sex. Using LIANA and Tensor-cell2cell, we identify latent communication programs altered in female AD donors, including a pericyte-EC signaling pattern that activates TGF{beta} via extracellular matrix ligands and is upregulated in the MTG but not the DLPFC. A second communication pattern, downregulated in female AD donors, reveals impaired estrogen pathway signaling through ligand-receptor interactions between pericytes and astrocytes. Supporting this, we observe reduced expression of pericyte-derived neuroligins and increased pericyte-astrocyte separation in a spatial transcriptomic subset. Additionally, we identify a microglia-to-pericyte signaling program conserved across brain regions, enriched for inflammatory pathways including hypoxia and p53, and elevated in both male and female AD donors with regional specificity. This result contrasts with the more sex-and region-specific pericyte signaling programs and suggests parallel mechanisms of NVU disruption between brain regions in AD. Our findings reveal brain region-specific and sex-specific pericyte signaling changes in AD and implicate vascular-, immune-, and synapse-associated pathways in NVU dysfunction. Altogether, the data suggest pericyte-driven communication as a mechanistic contributor to female-biased vulnerability in AD and support the need for sex-aware and region-specific approaches in neurodegeneration research.