Identification of a novel stemness-related signature with appealing implications in discriminating the prognosis and therapy responses for prostate cancer

Abstract

PurposeCancer stemness represents the tumor-initiation and self-renewal potentials of cancer stem cells. It is involved in prostate cancer progression and resistance to therapy. Herein, we aimed to unveil the stemness features, establish a novel prognostic model, and identify potential therapeutic targets. Methods26 stemness-related signatures were obtained from StemChecker. The expression profiles and clinical traits of TCGA-PRAD were obtained from TCGA and cBioPortal, respectively. GSE5446 and GSE70769 cohorts were acquired from GEO. PRAD_MSKCC cohort was also retrieved via the cBioPortal. The consensus clustering method was used for stemness subclusters classification. WGCNA was used to identify hub genes related to the stemness subcluster. The most important feature was explored in vitro. ResultsProstate cancer patients of TCGA-PRAD were divided into two subclusters (C1 and C2) based on the enrichment scores of the 26 stemness-related signatures. C1 was characterized by decreased survival, rich infiltrations of M0 macrophages and regulatory T cells, minimum sensitivity to chemotherapy, and a low response to immunotherapy. Hub genes of the red module with the highest correlation with C1 were subsequently identified by WGCNA and subjected to stemness-related risk model construction based on the machine-learning framework. Prostate cancer patients with high stemness scores had unfavorable prognosis, immunosuppressive tumor microenvironment, minimum sensitivity to chemotherapy, and a low response to immunotherapy. MXD3, the most important factor of the model, can regulate the stemness traits of prostate cancer cells. ConclusionsOur study depicted the stemness landscapes of prostate cancer and characterized two subclusters with diverse prognoses and tumor immune microenvironments. A stemness-risk signature was developed and demonstrated prospective implications in predicting prognosis and precision medicine.