Identification of anoikis-related molecular patterns and the novel risk model to predict prognosis, tumor microenvironment infiltration and immunotherapy response in bladder cancer.

Background: Anoikis, a unique form of cell death, serves as a vital part of the organism's defense by preventing shedding cells from re-attaching to the incorrect positions, and plays pivotal role in cancer metastasis. Nonetheless, the specific mechanisms among anoikis, the clinical prognosis and tumor microenvironment (TME) of bladder cancer (BLCA) are insufficiently understood. Method: BLCA patients were classified into different anoikis subtypes based on the expression of candidate anoikis-related genes (ARGs), and differences in the clinicopathological features, TME, immune cell infiltration, and immune checkpoints between two anoikis subtypes were analyzed. Next, patients in the TCGA cohort were randomized into the train and test groups in a 1:1 ratio. Subsequently, the anoikis-related model was constructed to predict the prognosis via utilizing the univariate Cox, LASSO and multivariate Cox analyses, and validated internally and externally. Moreover, the relationships between the risk score and clinicopathologic features, immune cell infiltration, immunotherapy response, and antitumor drug sensitivity were also analyzed. In addition, representative genes were evaluated using immunohistochemistry in clinical specimens, and in BLCA cell lines, functional experiments were performed to determine the biological behavior of hub gene PLOD1. Result: Two definite anoikis subgroups were identified. Compared to ARGcluster A, patients assigned to ARGcluster B were characterized by an immunosuppressive microenvironment and worse prognosis. Then, the anoikis-related model, including PLOD1, EHBP1, and CSPG4, was constructed, and BLCA patients in the low-risk group were characterized by a better prognosis. Next, the accurate nomogram was built to improve the clinical applicability by combining the age, tumor stage and risk Score. Moreover, immune infiltration and clinical features differed significantly between high- and low-risk groups. We also found that the low-risk group exhibited a lower tumor immune dysfunction and exclusion score, a higher immunophenoscore (IPS), had more sensitivity to immunotherapy. Eventually, the expression levels of three genes were verified by our experiment, and knockdown of PLOD1 could inhibit invasion and migration abilities in BLCA cell lines. Conclusion: These results demonstrated a new direction in precision therapy for BLCA, and indicated that the ARGs might be helpful to in predicting prognosis and as therapeutic targets in BLCA.