Tumor-infiltrating immune cells are closely related to the prognosis of bladder cancer. Analysis of tumor infiltrating immune cells is usually based on immunohistochemical analysis. Since many immune cell marker proteins are not specific for different immune cells, which may induce misleading or incomplete. CIBERSORT is an algorithm to estimate specific cell types in a mixed cell population using gene expression data. In this study, the CIBERSORT algorithm was used to identify the immune cell infiltration signatures. The gene expression profiles, mutation data, and clinical data were collected from The Cancer Genome Atlas (TCGA) database. Unsupervised consensus clustering was used to acquire the immune cell infiltration subtypes of bladder cancer based on the fractions of 22 immune cell types. Four immune cell clusters with different immune infiltrate and mutation characteristics were identified. In addition, this stratification has a prognostic relevance, with cluster 2 having the best outcome, cluster 1 the worst. These clusters showed distinct mRNA expression patterns. The characteristic genes in subtype cluster 1 were mainly involved in cell division, those in subtype cluster 2 were mainly related in antigen processing and presentation, those in subtype cluster 3 were mainly involved in epidermal cell differentiation, and those in subtype cluster 4 were mainly related in the humoral immune response. These differences may affect the development of the bladder cancer, the sensitivity to treatment as well as the prognosis. Through further validation, this study may contribute to the development of personalized therapy and precision medical treatments.
Keywords: CIBERSORT algorithm; The Cancer Genome Atlas; bladder cancer; gene expression; immune infiltration subtypes; personalized therapy.
© The author(s).