LncRNAs have been confirmed to be in relation to cancer immunity regulation as well as tumor microenvironment in bladder cancer 17. Several immune-related lncRNA models have been constructed in published literature 14,18,19. Nevertheless, these signature models are developed on the basis of expression quantifications of immune-related lncRNAs. Herein, this study recognized the immune-related lncRNA pairs and constructed a reliable and independent risk model through combining two lncRNAs, not adopting their expression levels 20.
Here, we firstly screened immune-related lncRNAs by Pearson correlation analyses. Different from previous research, dysregulated immune-related lncRNAs were further identified by comparing their expressions between bladder cancer and normal specimens 14,21. With cyclically single pairing methods with 0-or-1 matrices, we identified immune-related lncRNA pairs. Combining univariate analyses and LASSO model, we developed a risk model for bladder cancer. Not using the median RS as the cutoff value that differentiated bladder cancer subjects into high- and low-risk subgroups, the AIC value of one-year survival was determined as the optimal cut-off value 14,22,23. Furthermore, this risk model possessed distinct associations with survival status, age, stage and TNM of bladder cancer. As depicted in multivariate regression analyses, RS might independently predict bladder cancer patients’ OS. In comparison to other clinical features, RS displayed the highest AUC of one-year OS, indicating that this RS possessed the potential as a favorable predictor of bladder cancer. Moreover, this risk model was in relation to immune cell infiltrations and immune checkpoints. Intertumoral tumor-infiltrating immune cells may impact the responses to ICIs. Here, by comprehensively utilizing XCELL, TIMER, QUANTISEQ, MCPCOUNTER, EPIC, CIBERSORT-ABS and CIBERSORT algorithms, we characterized the correlations between risk model and immune cell infiltrations. High-risk specimens possessed elevated infiltration levels of myeloid dendritic cell, B cell native, macrophage M0 and M2, neutrophil and T cell CD8 in bladder cancer. Also, GAL9 displayed elevated expression in low- than high-risk specimens while higher TIM-3 and PD1LG2 expressions were found in high-risk specimens than low-risk specimens. These data indicated that this risk model might be utilized for predicting immunotherapy response of bladder cancer.
Bladder cancer represents a complex malignancy correlated to high morbidity and mortality risks. if not treated optimally. Neoadjuvant chemotherapy has been recommended prior to radical cystectomy for bladder cancer. Although the survival benefit is nearly 5-10%, some subjects cannot respond to chemotherapy 24. Thus, identifying predictors may distinctly reduce side effects and miss the optimal time for surgery. Here, our data suggested that high-risk patients exhibited higher sensitivity to cisplatin in comparison to low-risk individuals. Inversely, subjects with low-risk were more sensitive to methotrexate than those with high-risk. Above data were indicative that this risk model might possess the potential to predict the sensitivity to cisplatin and methotrexate for bladder cancer.
Due to high abundance, lncRNAs have distinct biological functions. Our methods identified dysregulated immune-related lncRNAs as well as established the optimal immune-related lncRNA pairs. Hence, pairs with high or low expressions only were tested not detecting expression levels of each lncRNA. Our risk signature possessed the superiority in clinical practice for distinguishing high- and low-risk patients. Due to the closely correlations to immune-related genes, the selected lncRNAs potentially participated in mediating immune microenvironment shape in bladder cancer. However, there are several limitations in our study. Firstly, more independent bladder cancer cohorts should be utilized for validating the identified prognostic immune-related lncRNA model. Furthermore, the functions of these lncRNAs and their interactions with immune-related genes will be confirmed based on in vitro and in vivo experiments.
Collectively, this prognostic signature constructed by 15 immune-related lncRNA pairs as an independent predictor displayed the favorable performance in predicting prognoses of bladder cancer. Also, it had the potential to predict immune landscape and chemotherapeutic response for bladder cancer patients.