A tissue-resolved endothelial surface proteome atlas informs organ-selective vascular targeting

Abstract

BACKGROUND Endothelial cells (ECs) exhibit organ-specific functional diversity that shapes vascular homeostasis, disease susceptibility, and therapeutic accessibility. Although single-cell transcriptomic studies have defined endothelial heterogeneity at the RNA level, the in vivo cell-surface protein landscape that directly mediates vascular signaling and drug targeting remains incompletely characterized. Here, we mapped organ-specific endothelial surface proteomes atlas in vivo to define tissue-enriched vascular protein candidates relevant to organ-selective therapeutic design. METHODS We generated Cdh5-CreERT; Cre-iPEEL mice, referred to here as CHRP mice, in which membrane-tethered horseradish peroxidase is induced selectively in ECs after tamoxifen treatment, and compared CHRP labeling with non-selective NHS-Biotin vascular labeling. Following in vivo biotin-phenol perfusion, endothelial surface proteins were enriched by streptavidin affinity purification and analyzed by mass spectrometry across six organs. Proteomic profiles were used to resolve tissue- and subtype-associated endothelial surface signatures, compare protein and transcript detection patterns, and nominate tissue-selective endothelial membrane candidates, which were annotated using ChEMBL compound-target information. RESULTS Compared with non-selective NHS-Biotin labeling, CHRP improved endothelial specificity and produced clearer separation of tissue-resolved endothelial surface proteomes across brain, white adipose tissue, small intestine, kidney, lung, and skeletal muscle. CHRP proteomics revealed pronounced organ-specific heterogeneity and resolved canonical arterial, venous, and capillary programs, as well as specialized endothelial signatures including blood-brain barrier and glomerular endothelial features. Comparison with single-cell endothelial references revealed systematic differences between transcriptomic and proteomic detection of endothelial membrane proteins. Further analysis identified tissue-selective endothelial membrane candidates, and ChEMBL annotation linked a subset of these candidates to existing compound-target records, supporting the candidate atlas as a resource for future tissue-selective vascular targeting studies. CONCLUSIONS CHRP-based in vivo proximity labeling enables systematic, protein-level mapping of organ-specific endothelial surface proteomes. Together with transcriptomic comparison and compound-target annotation, this study provides a tissue-resolved endothelial surfaceome resource for vascular biology and future organ-selective therapeutic target evaluation.