Identification of a novel defined inflammation-related long noncoding RNA signature contributes to predicting prognosis and distinction between the cold and hot tumors in bladder cancer

PurposeBladder cancer (BLCA) is one of the most frequently diagnosed urological malignancies and is the 4th most common cancer in men worldwide. Molecular targets expressed in bladder cancer (BLCA) are usually used for developing targeted drug treatments. However, poor prognosis and poor immunotherapy efficacy remain major challenges for BLCA. Numerous studies have shown that long non-coding RNAs (LncRNAs) play an important role in the development of cancer. However, the role of lncRNAs related to inflammation in BLCA and their prognostic value remain unclear. Therefore, this study is aimed to explore new potential biomarkers that can predict cancer prognosis. MethodsWe downloaded BLCA-related RNA sequencing data from The Cancer Genome Atlas (TCGA) and searched for inflammation-related prognostic long non-coding RNAs (lncRNAs) by univariate Cox (uniCox) regression and co-expression analysis. We used the least absolute shrinkage and selection operator (LASSO) analysis to construct an inflammation-related lncRNA prognosis risk model. Samples were divided into high-risk score (HRS) group and low-risk score (LRS) group based on the median value of risk scores. The independent variable factors were identified by univariate Cox (uni-Cox) and multivariate Cox (multi-Cox) regression analyses, and receiver operating characteristic (ROC) curves were used to compare the role of different factors in predicting outcomes. Nomogram and Calibration Plot were generated by the R package rms to analyze whether the prediction results are correct and show good consistency. Correlation coefficients were calculated by Pearson analysis. The Kaplan-Meier method was used to assess the prognostic value. The expression of 7 lncRNAs related with inflammation was also confirmed by qRT-PCR in BLCA cell lines. Kyoto Encyclopedia of Gene and Genome (KEGG) pathways that were significantly enriched (P < 0.05) in each risk group were identified by the GSEA software. The R package pRRophetic was used to predict the IC50 of common chemotherapeutic agents. TIMER, XCELL, QUANTISEQ, MCPCOUNTER, EPIC and CIBERSORT were applied to quantify the relative proportions of infiltrating immune cells. We also used package ggpubr to evaluate TME scores and immune checkpoint activation in LRS and HRS populations. R package GSEABase was used to analyze the activity of immune cells or immune function. Different clusters of principal component analysis (PCA), t-distribution random neighborhood embedding (t-SNE), and Kaplan-Meier survival were analyzed using R package Rtsne's. The R package ConsensesClusterPlus was used to class the inflammation-related lncRNAs. ResultsIn this study, a model containing 7 inflammation-related lncRNAs was constructed. The calibration plot of the model was consistent with the prognosis prediction outcomes. The 1-, 3-, and 5-year ROC curve (AUC) were 0.699, 0.689, and 0.699, respectively. High-risk patients were enriched in lncRNAs related with tumor invasion and immunity, and had higher levels of immune cell infiltration and immune checkpoint activation. Hot tumors and cold tumors were effectively distinguished by clusters 2 and 3 and cluster 1, respectively, which indicated that hot tumors are more susceptible to immunotherapy. ConclusionOur study showed that inflammation-related LncRNAs are closely related with BLCA, and inflammation-related lncRNA can accurately predict patient prognosis and effectively differentiate between hot and cold tumors, thus improving individualized immunotherapy for BLCA patients. Therefore, this study provides an effective predictive model and a new therapeutic target for the prognosis and clinical treatment of BLCA, thus facilitating the development of individualized tumor therapy.

Automated summary

In this study, a model containing 7 inflammation-related lncRNAs was constructed to accurately predict the prognosis of BLCA patients and differentiate between hot and cold tumors. This provides a new therapeutic target for individualized tumor therapy.