Construction of an lncRNA model for prognostic prediction of bladder cancer

Abstract

Objective

We aimed to investigate the role and potential mechanisms of long non-coding RNAs (lncRNAs) in bladder cancer (BC), as well as determine their prognostic value.

Methods

LncRNA expression data and clinical data from BC patients were downloaded from The Cancer Genome Atlas (TCGA) database. R software was used to carry out principal component analysis (PCA), differential analysis, and prognostic analysis. Lasso regression and multivariate Cox regression analyses were performed to identify potential prognostic genes. The expression of five identified genes and their correlation with prognosis were verified using TCGA and GSE13507 datasets. In addition, quantitative real-time polymerase chain reaction (qRT-PCR) was used to confirm the expression of these five genes in cell lines (two human BC cell lines and one human bladder epithelial cell line) and tissues (84 pairs of BC tissues and the corresponding paracancerous tissues). Risk scores that had been generated from the five genes and their prognostic ability were assessed by receiver operating characteristic (ROC) and Kaplan–Meier (KM) curves. Co-expressed genes were screened by WGCNA and analyzed by GO and KEGG, while functional enrichment and immune infiltration analyses were performed using STRING ( https://cn.string-db.org/) and TIMER2.0 ( http://timer.cistrome.org/) online tools, respectively.

Results

CYP4F8, FAR2P1, LINC01518, LINC01764, and DTNA were identified as potential prognostic genes. We found that these five genes were differentially expressed in BC tissue, as well as in BC cell lines, and were significantly correlated with the prognosis of BC patients. KM analysis considering risk scores as independent parameters revealed differences in overall survival (OS) by subgroups. The ROC curve revealed that a combined model consisting of all five genes had good predictive ability at 1, 3, and 5 years. GO and KEGG analyses of 567 co-expressed genes revealed that these genes were significantly associated with muscle function.

Conclusion

LncRNAs can be good predictors of BC development and prognosis, and may act as potential tumor markers and therapeutic targets that may be beneficial in helping clinicians decide the most effective treatment strategies.