The identification of a two-gene prognostic model based on cisplatin resistance-related ceRNA network in small cell lung cancer

Abstract

Abstract Background Small cell lung cancer (SCLC) is a very malignant tumor with rapid growth and early metastasis. Platinum-based chemo-resistance is the major issue for SCLC treatment failure. Identifying a new prognostic model will help to make an accurate treatment decision for SCLC patients. Methods Using the genomics of drug sensitivity in cancer (GDSC) database, we identified cisplatin resistance-related lncRNAs in SCLC cells. Based on the competing endogenous RNA (ceRNA) network, we identified the mRNAs correlated with the lncRNAs. Using Cox and LASSO regression analysis, a prognostic model was established. The survival prediction accuracy was evaluated by receiver operating characteristic (ROC) curve and Kaplan–Meier analysis. GSEA, GO, KEGG and CIBERSORT tools were used for functional enrichment and immune cells infiltration analysis. Results We first screened out 10 differentially expressed lncRNAs between cisplatin resistant and sensitive SCLC cells from GDSC database. Based on ceRNA network, 31 mRNAs were identified with a correlation with the 10 lncRNAs. Furthermore, two genes (LIMK2 and PI4K2B) were identified by Cox and LASSO regression analysis to construct a prognostic model. Kaplan–Meier analysis indicated that the high-risk group had a poor overall survival compared with the low-risk group. The predicted area under the ROC curve (AUC) was 0.853 in the training set, and the AUC was 0.671 in the validation set. In the meanwhile, the low expression of LIMK2 or the high expression of PI4K2B in SCLC tumors was also significantly associated with poor overall survival in both training and validation sets. Functional enrichment analysis showed that the low-risk group was enriched in the apoptosis pathway and high immune infiltration of T cells. Finally, an apoptosis-related gene Cathepsin D (CTSD) was identified to be up-regulated in the low-risk group, and its higher expression correlated with better overall survival in SCLC. Conclusion We established a prognostic model and potential biomarkers (LIMK2, PI4K2B and CTSD), which could help to improve the risk stratification of SCLC patients.