Machine Learning Characterization of Immunometabolism in the Tumor Microenvironment and Immunotherapy Responses in Bladder Cancer.

Abstract

Immune dysregulation and metabolism reprogramming are implicated in bladder cancer (BLCA), the relationships between immunometabolism (IMB) and BLCA remain poorly understood. We identified the expression patterns of IMB-related genes and their relationship with prognosis, ultimately developing a machine learning prognostic model. We performed a comprehensive investigation into UCN2 function in BLCA by qPCR, immunohistochemistry, Western blot, Transwell migration assay, and flow cytometry analysis. Two BLCA subclasses were identified, each exhibiting distinctive molecular patterns. Then, an IMB.score was conducted, the IMB.score not only reflected the characteristics of the clinical but also provided insights into immunotherapy efficacy. Specifically, high IMB.score category exhibited a more active TME and unfavorable prognosis; those in the high IMB.score category were more responsive to immunotherapy, suggesting an "immunity tidal model" phenotype. Besides, UCN2 is overexpressed in BLCA tissues, and was found to be positively associated with malignant phenotypes and a poorer prognosis for BLCA. Furthermore, by silencing the expression of UCN2, we observed a significant reduction in the proliferation, migration, and invasion of BLCA cells in vitro. UCN2 is considered a crucial gene in IMB that plays a significant role in the onset and development of BLCA.