The role of endothelial cell-related gene COL1A1 in prostate cancer diagnosis and immunotherapy: insights from machine learning and single-cell analysis.

Background: Endothelial cells are integral components of the tumor microenvironment and play a multifaceted role in tumor immunotherapy. Targeting endothelial cells and related signaling pathways can improve the effectiveness of immunotherapy by normalizing tumor blood vessels and promoting immune cell infiltration. However, to date, there have been no comprehensive studies analyzing the role of endothelial cells in the diagnosis and treatment of prostate adenocarcinoma (PRAD).

Method: By integrating clinical and transcriptomic data from TCGA-PRAD, we initially identified key endothelial cell-related genes in PRAD samples through single-cell analysis. Subsequently, cluster analysis was employed to classify PRAD samples based on the expression of these endothelial cell-related genes, allowing us to explore their correlation with patient prognosis and immunotherapy outcomes. A diagnostic model was then constructed and validated using a combination of 108 machine learning algorithms. The XGBoost and Random Forest algorithms highlighted the significant role of COL1A1, and we further analyzed the expression and correlation of COL1A1, AR, and EGFR through multiplex immunofluorescence staining. In vitro experimental analysis of the impact of COL1A1 on the progression of PRAD.

Results: Single-cell analysis identified 12 differential prognostic genes associated with endothelial cells. Cluster analysis confirmed a strong correlation between endothelial cell-related genes and both prostate cancer prognosis and immunotherapy responses. Diagnostic models developed using various machine learning techniques demonstrated the significant predictive capability of these 12 genes in the diagnosis of prostate cancer. Furthermore, based on patients' prognostic information, multiple machine learning analyses highlighted the critical role of COL1A1. Immunofluorescence analysis results confirmed that COL1A1 is highly expressed in prostate cancer and is positively correlated with both AR and EGFR. In vitro experiments confirm that reducing COL1A1 expression levels can inhibit PRAD progression.

Conclusion: This study provides a comprehensive analysis of the role of endothelial cell-related genes in the diagnosis, prognosis, and immunotherapy of prostate cancer. The findings, supported by various machine learning algorithms and experimental results, highlight COL1A1 as a significant target for the diagnosis and immunotherapy of PRAD.