Clinical data from 205 patients with CC who underwent radiotherapy were collected and analyzed. Of these, 180 patients received concurrent chemotherapy, while 25 patients did not due to health conditions preventing chemotherapy tolerance, advanced age, or personal refusal (Fig. 1A). The results revealed that SELP was predominantly expressed in TEC (Fig. 1B). High SELP expression was significantly correlated with improved local recurrence-free survival (LRFS), overall survival (OS), and progression-free survival (PFS) (P < 0.05 for all; Fig. 1C, Supplementary Fig. S2A). Multivariate analysis using the Cox proportional hazards model confirmed that elevated SELP expression was an independent predictor of enhanced LRFS, OS, and PFS in CC patients treated with radiotherapy (LRFS: hazard ratio [HR] = 0.28, P = 0.029; OS: HR = 0.38, P = 0.002; PFS: HR = 0.28, P < 0.001; Fig. 1D, Supplementary Fig. S2B). Furthermore, pathological type, tumor cell differentiation, FIGO stage, and treatment strategy were identified as significant predictors of survival in CC patients (Fig. 1D, Supplementary Fig. S2B). These findings collectively indicate that SELP is predominantly expressed in TECs and that its high expression is positively associated with improved outcomes in patients with CC undergoing radiotherapy.

Association of SELP+ TEC with enhanced radiotherapy efficacy and immune activation in patients with CC. A Clinical characteristics of 205 patients with CC. B Representative immunohistochemical images of SELP expression in TECs. C Kaplan–Meier survival curves showing local recurrence-free survival in CC patients stratified by SELP expression levels. D Multivariate Cox proportional hazards analysis assessing the relationship between clinical characteristics and local recurrence-free survival in CC patients. E Kaplan–Meier survival curve demonstrating PFS of CC patients treated with radiotherapy in the SELPhigh and SELPlow groups (with optimal cutoff value derived from the TCGA dataset). P-values from the two-sided log-rank test are shown. F Volcano plot illustrating differentially expressed genes between the SELPhigh and SELPlow groups, with the most significant genes highlighted. G GO terms enriched in the SELPhigh group. H Box plots comparing gene set scores between the SELPhigh and SELPlow groups. I Violin plots displaying the expression levels of atypical chemokine receptor genes in CC patients

To validate the critical role of SELP in immunity and radiotherapy response in CC, bulk RNA-sequencing data from 187 CC patients undergoing radiotherapy were extracted and analyzed from the TCGA database. The results demonstrated that high SELP expression was associated with a favorable prognosis (Fig. 1E). The SELPhigh group exhibited upregulation of KCNE1 and SCGB3A2 (Fig. 1F). GO enrichment analysis revealed that the SELPhigh group showed significant enrichment in processes related to cell–cell adhesion regulation, leukocyte migration, immune response activation, positive regulation of lymphocyte activation, and T cell activation. In contrast, processes related to the regulation of epithelial cell proliferation were less enriched (Fig. 1G). To further explore the immune landscape, the relationship between SELP expression and immune cell infiltration was examined. The SELPhigh tumors exhibited significantly higher infiltration of CD8+ T cells (P = 0.012) and CD68+ macrophages (P = 0.03) compared to SELPlow tumors, while no significant difference in CD56+ NK cell infiltration was observed between the two groups (P = 0.42, Supplementary Fig. S3A). These findings were corroborated by deconvolution analysis of TCGA data, which confirmed a higher proportion of CD8+ T cells and macrophages in SELPhigh tumors, with no significant difference in NK cells (Supplementary Fig. S3B). Previous research has highlighted a strong correlation between radiotherapy response and angiogenesis or hypoxia [10]. The lower level of hypoxia observed in the SELPhigh group was associated with improved radiotherapy efficacy (Fig. 1G, H). Additionally, the SELPhigh group exhibited elevated expression of chemokine receptors critical for the recruitment of CD8+ T cells, including members of the atypical, CC, CXC3, and CXC chemokine receptor families (Fig. 1I). These results suggest that SELP plays a pivotal role in enhancing anti-tumor immunity and improving radiotherapy outcomes in CC, potentially by influencing chemokine-mediated T cell infiltration and activation.

To investigate the potential role of SELP in TECs, scRNA-seq data derived from 42,159 cells obtained from eight CC patients were analyzed. Seven distinct cell populations were identified in the dataset (Supplementary Fig. S4). Based on the expression of lineage-specific markers, these populations were categorized into seven major clusters: B cells (MS4A1), TECs (VWF), epithelial cells (EPCAM, KRT19), cancer-associated fibroblasts (PDGFRB, LUM, MYH11), myeloid cells (CD68, CD14), NK/T cells (CD3D, NKG7), and plasma cells (MZB1).

TECs were subsequently classified as SELP+ or SELP– based on their SELP expression levels (Fig. 2A). Multiplex immunofluorescence staining confirmed the distinct clusters identified (Fig. 2B). The SELP+ TECs exhibited differential expression of genes such as ACKR1, SELE, C7, and IL1RI, while the SELP– TECs expressed genes including ESM1, BTNL9, HEY1, and FCN3 (Fig. 2C). Consistent with bulk RNA-sequencing analysis, SELP+ TECs showed enhanced enrichment in processes related to leukocyte migration, leukocyte-mediated immunity, positive regulation of cell–cell adhesion, and cytokine-mediated signaling pathways (Fig. 2D, Supplementary Fig. S5A). These processes were linked to a favorable prognosis in CC patients undergoing radiotherapy (Supplementary Fig. S6). In contrast, SELP+ TECs exhibited reduced enrichment in processes associated with epithelial cell proliferation, endothelial cell migration, endothelium development, and sprouting angiogenesis (Fig. 2D). Additionally, SELP+ TECs displayed higher expression levels of atypical chemokine receptors (Supplementary Fig. S5B). To further validate these findings, single-cell endothelial data from additional tumor types, including clear cell renal cell carcinoma, lung adenocarcinoma, and head and neck squamous cell carcinoma, were examined. In these datasets, SELP+ TECs consistently demonstrated immune activation characteristics (Supplementary Fig. S7). Moreover, SELP+ endothelial cells in normal cervical tissue exhibited similar features (Supplementary Fig. S8A-8C). Notably, SELPhigh cervical cancer samples displayed an even higher degree of immune activation compared to SELPhigh cells in normal cervical tissue (Supplementary Fig. S8D). These results suggest that SELP+ TECs exhibit enhanced immune-related features and increased chemokine receptor expression, which are correlated with a favorable prognosis and improved outcomes in CC patients undergoing radiotherapy. Taken together, these findings underscore the critical role of SELP in enhancing anti-tumor immunity and radiotherapy efficacy in CC.

Correlation of SELP+ TEC:CD8+ T cell crosstalk with improved radiotherapy outcomes and enhanced immune responses in patients with CC. A UMAP plots showing the expression of SELP in 2,569 TECs, color-coded by SELP expression levels, stratified by the median SELP expression value. B Representative images of immunofluorescence staining for CD62P in TECs within tumor tissue. C Volcano plot depicting differentially expressed genes between SELP+ TECs and SELP– TECs. D GO terms enriched in SELP+ TECs and SELP– TECs. E Heatmap illustrating intercellular communication between TECs and NK/T cell subclusters via the CCL signaling pathway. Color intensity reflects the probability of communication. F Bubble plots showing specific ligand–receptor interactions from the CCL signaling pathway involved in TEC–NK/T cell crosstalk. Bubble size indicates P-values, and color intensity reflects the interaction probability. G Kaplan–Meier survival curve demonstrating PFS in CC patients treated with radiotherapy, stratified by CCL signaling pathway score, CCL5 expression, and ACKR1 expression in the TCGA dataset. H Immunofluorescence staining revealing the interaction between CD62P+ endothelial cells (green for CD62P) and CD8+ T cells (red) via the ACKR1-CCL5 axis. CCL5 is shown in purple, ACKR1 (CD234) in orange, and nuclei in blue (DAPI staining). The merged image illustrates the colocalization of these markers within the tissue (scale bar = 40 μm). TECs, tumor endothelial cells; OS, overall survival; PFS, progression-free survival; CC, cervical cancer; UMAP, uniform manifold approximation and projection

Building on the aforementioned results, it was hypothesized that SELP+ TECs may enhance anti-tumor immunity by facilitating lymphocyte infiltration. To test this hypothesis, interactions between NK/T cells and TECs were examined. Based on transcriptomic profiling, NK/T cells were classified into seven distinct subclusters: LAG3_CD8_T, CD4_Naive, GZMK_CD8_T, NK, Treg, GINS2_CD8_T, and proliferating_T (Supplementary Fig. S9A-C). CD4_Naive cells exhibited characteristics of a naïve state, while LAG3_CD8_T cells displayed features of immune exhaustion, consistent with an exhausted phenotype. The GZMK_CD8_T cluster showed strong cytotoxic activity, TCR signaling, and effector function activation, highlighting its active role in anti-tumor immunity. NK cells demonstrated pronounced involvement in cytokine and chemokine receptor signaling, suggesting a role in immune cell recruitment and communication. The GINS2_CD8_T and Proliferating_T clusters were marked by heightened metabolic activity, reflecting intense cell cycle progression and elevated energy demands.

To investigate how SELP+ TECs contribute to CD8+ T cell infiltration, we analyzed interactions between TECs and NK/T cells using CellChat. Notably, signaling networks involving CCLs emerged as key mediators of TEC-NK/T cell communication, with SELP+ TECs exhibiting stronger interaction intensities (Fig. 2E). Further analysis of ligand-receptor pairs revealed that the ACKR1-CCL5 interaction was notably more active in SELP+ TECs than in SELP– TECs, particularly in the context of CD8+ T cell interactions (Fig. 2F). The CCL signaling network, along with CCL5 and ACKR1, was found to be associated with favorable PFS in patients with CC treated with radiotherapy (Fig. 2G). Validation via multiplex immunofluorescence staining confirmed the colocalization of SELP+CD234+ TECs with CCL5+CD8+ T cells in the tumor vasculature (Fig. 2H). ACKR1, as a receptor of CCL5, may facilitate the TECs-mediated recruitment and infiltration of CCL5-secreting CD8+ T cells into the tumor microenvironment [11]. Meanwhile, SELP+ TECs exhibited an up-regulation of pathways related to leukocyte cell adhesion and regulation of cellular extravasation (Supplementary Fig. S10A), suggesting enhanced capabilities for immune cell recruitment. Additionally, lower cytoskeleton organization scores of SELP+ TECs (Supplementary Fig. S10B) indicated reduced cytoskeletal stability, which is often associated with increased endothelial permeability. Further spatial transcriptomics analysis validated that SELP+ TECs enriched regions were negatively correlated with cytoskeleton organization and positively correlated with cellular extravasation scores (Supplementary Fig. S10C). In summary, SELP+ TECs bind CCL5-secreting CD8+ T cells through ACKR1, and their enhanced leukocyte adhesion properties, along with increased endothelial permeability, may contribute to the recruitment and extravasation of CD8+ T cells.