Fig. 2: Characterization of endothelial subtypes revealed the immunomodulatory function of venular ECs.

a The sub-clustering of ECs in the in-house scRNA-seq dataset, with each EC subtype being distinguished by a different color. b Differential expression of canonical endothelium identity markers (PECAM1, IGFBP3, ACKR1, PLVAP) in three EC subtypes. c Heatmap showing the DEGs between venular ECs (vECs), arterial ECs (aECs), and capillary ECs (cECs). Top fifteen DEGs for each EC subtype are shown based on their fold change value. d The GO BP-terms (biological process terms) enriched in the DEGs between vECs, aECs, and cECs. Top six GO BP-terms for each EC subtype are shown based on their adjusted p-value. e Heatmap displaying smoothed pseudotime-dependent differential gene expression in vECs and cECs, with each row indicating a gene and each column indicating a cell. The genes are grouped into three clusters based on their pseudotime-dependent expression patterns. f Differential gene expression analysis of HEVs or non-HEV venular ECs, in comparison to cECs and aECs. The x-axis and y-axis, respectively, indicate the log₂(fold change) value of the DEGs in HEVs and non-HEV venules. g Spatial maps of ECs annotated by RCTD and IHC staining of CD31 illustrating the highly consistent location of ECs in the representative areas on the ST slides of NCC-BR5. h UMAP plot showing the bin50 spots labeled as ECs across 30 ST slides. Three subtypes of ECs are identified by their signature genes and annotated by different colors. i Relative proportions of vECs, aECs, and cECs in the ST slides of 23 primary breast tumors. Red, green, and yellow indicate vECs, aECs, and cECs, respectively. j Relative proportions of vECs, aECs, and cECs in the ST slides of paired primary tumors and lymph node metastases. Red, green, and yellow indicate vECs, aECs, and cECs, respectively.