Presenting our findings on capillary endothelial cells in Alzheimer’s Disease at the Society for Neuroscience Meeting 2025 this November in San Diego.
Systems-level, large-scale single-cell RNA-seq analyses reveal identity shifts and lipid-related alterations in human brain capillary endothelial cells in Alzheimer’s disease
AUTHOR BLOCK: *Md Rezaul Islam,1,2 † Rebecca L. Pinals,1,2 † Aaron Choi,1,2 Eulim Kang,1,2 Emre Agbas,1,2 Nhat Truong,1,2 Masayuki Nakano,1,2 Claudia F. C. Lozano,1,2 Colin Staab,1,2 Li-Huei Tsai1,2
*presenting author; †equal contributions
1Picower Institute for Learning and Memory, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, United States; 2 Department of Brain and Cognitive Sciences, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, United States
Submitted Abstract: Blood-brain barrier (BBB) breakdown is implicated early in numerous neurodegenerative diseases. Endothelial cells, due to their localization at capillaries, arteries, and venules, are the first barrier cell type to encounter peripheral changes and challenges from the bloodstream. Here, we investigate the heterogeneity of brain capillary endothelial cells and the transcriptomic and cellular changes they undergo in the context of Alzheimer’s disease (AD). Through in silico sorting of endothelial cells from single-nucleus RNA sequencing (snRNA-seq) datasets acquired from a cohort of 427 individuals, we delineate vulnerable capillary endothelial subtypes in AD. Moreover, we identify a gene module (endoM2) that is significantly associated with Alzheimer’s pathology and APOE4 risk allele, inversely correlated with measures of cognition, and conserved across additional large-scale human snRNA-seq datasets of AD. Biological processes associated with this upregulated endoM2 module include cell migration, immune response, and lipid transport. To establish a gene-module-to-phenotype connection, we leverage human induced pluripotent stem cells (iPSCs) to differentiate brain microvascular endothelial cells (BMECs) and test various stimulation conditions informed by the identified GO terms in endoM2. We find a cytokine treatment leads to cellular elongation, increased lipid uptake and lipid droplet formation, and tight junction losses. We posit that BMECs undergo inflammation-induced changes in Alzheimer’s disease, exacerbated by APOE4 risk gene status, that suggest a loss of cellular identity of endothelial cells and result in loss of BBB integrity. Bulk sequencing of the cytokine-treated iPSC-derived BMECs from two isogenic pairs reveals that genes associated with cell migration are perturbed. From a targeted drug screen to inhibit top altered genes, we discover an exciting small molecule that limits cellular elongation, decreases lipid droplet accumulation, and reduces tight junction losses, validated in both 2-D cell monoculture and 3-D in vitro BBB co-culture. Our work motivates targeted interventions to prevent cerebrovascular dysfunction in the face of peripheral inflammation.
