Identification of immune-based prostate cancer subtypes using mRNA expression

Background Immune infiltration of Prostate cancer (PCa) was highly related to clinical outcomes. However, previous works failed to elucidate the diversity of different immune cell types that make up the function of the immune response system. The aim of the study was to uncover the composition of TIICs in PCa utilizing the CIBERSORT algorithm and further reveal the molecular characteristics of PCa subtypes. Method In the present work, we employed the CIBERSORT method to evaluate the relative proportions of immune cell profiling in PCa and adjacent samples, normal samples. We analyzed the correlation between immune cell infiltration and clinical information. The tumor-infiltrating immune cells of the TCGA PCa cohort were analyzed for the first time. The fractions of 22 immune cell types were imputed to determine the correlation between each immune cell subpopulation and clinical feature. Three types of molecular classification were identified via R-package of “CancerSubtypes”. The functional enrichment was analyzed in each subtype. The submap and TIDE algorithm were used to predict the clinical response to immune checkpoint blockade, and GDSC was employed to screen chemotherapeutic targets for the potential treatment of PCa. Results In current work, we utilized the CIBERSORT algorithm to assess the relative proportions of immune cell profiling in PCa and adjacent samples, normal samples. We investigated the correlation between immune cell infiltration and clinical data. The tumor-infiltrating immune cells in the TCGA PCa cohort were analyzed. The 22 immune cells were also calculated to determine the correlation between each immune cell subpopulation and survival and response to chemotherapy. Three types of molecular classification were identified. Each subtype has specific molecular and clinical characteristics. Meanwhile, Cluster I is defined as advanced PCa, and is more likely to respond to immunotherapy. Conclusions Our results demonstrated that differences in immune response may be important drivers of PCa progression and response to treatment. The deconvolution algorithm of gene expression microarray data by CIBERSOFT provides useful information about the immune cell composition of PCa patients. In addition, we have found a subtype of immunopositive PCa subtype and will help to explore the reasons for the poor effect of PCa on immunotherapy, and it is expected that immunotherapy will be used to guide the individualized management and treatment of PCa patients.