Identification of a distinct tumor endothelial cell-related gene expression signature associated with patient prognosis and immunotherapy response in multiple cancers.

BACKGROUND: Tumor endothelial cells (TECs) play a significant role in regulating the tumor microenvironment, drug response, and immune cell activities in various cancers. However, the association between TEC gene expression signature and patient prognosis or therapeutic response remains poorly understood. METHODS: We analyzed transcriptomics data of normal and tumor endothelial cells obtained from the GEO database to identify differentially expressed genes (DEGs) associated with TECs. We then compared these DEGs with those commonly found across five different tumor types from the TCGA database to determine their prognostic relevance. Using these genes, we constructed a prognostic risk model integrated with clinical features to develop a nomogram model, which we validated through biological experiments. RESULTS: We identified 12 TEC-related prognostic genes across multiple tumor types, of which five genes were sufficient to construct a prognostic risk model with an AUC of 0.682. The risk scores effectively predicted patient prognosis and immunotherapeutic response. Our newly developed nomogram model provided more accurate prognostic estimates of cancer patients than the TNM staging method (AUC = 0.735) and was validated using external patient cohorts. Finally, RT-PCR and immunohistochemical analyses indicated that the expression of these 5 TEC-related prognostic genes was up-regulated in both patient-derived tumors and cancer cell lines, while depletion of the hub genes reduced cancer cell growth, migration and invasion, and enhanced their sensitivity to gemcitabine or cytarabine. CONCLUSIONS: Our study discovered the first TEC-related gene expression signature that can be used to construct a prognostic risk model for guiding treatment options in multiple cancers.