Abstract
Regulatory T cells (Tregs) have been found to be related to immune therapeutic resistance in kidney cancer. However, the potential Tregs-related genes still need to be explored. Our study found that patients with high Tregs activity show poor prognosis. Through co-expression and differential expression analysis, we screened several Tregs-related genes (KTRGs) in kidney renal clear cell carcinoma. We further conducted the univariate Cox regression analysis and determined the prognosis-related KTRGs. Through the machine learning algorithm-Boruta, the potentially important KTRGs were screened further and submitted to construct a risk model. The risk model could predict the prognosis of RCC patients well, high risk patients show a poorer outcomes than low risk patients. Multivariate Cox regression analysis reveals that risk score is an independent prognostic factor. Then, the nomogram model based on KTRG risk score and other clinical variables was further established, which shows a high predicted accuracy and clinical benefit based on model validation methods. In addition, we found EMT, JAK/STAT3, and immune-related pathways highly enriched in high risk groups, while metabolism-related pathways show a low enrichment. Through analyzing two other external immune therapeutic datasets, we found that the risk score could predict the patient's immune therapeutic response. High-risk groups represent a worse therapeutic response than low-risk groups. In summary, we identified several Tregs-related genes and constructed a risk model to predict prognosis and immune therapeutic response. We hope these organized data can provide a theoretical basis for exploring potential Tregs' targets to synergize the immune therapy for RCC patients.