Purpose: To identify and characterize gefitinib resistance-related (GefR-related) lncRNAs and construct a prediction model for lung adenocarcinoma (LUAD).
Methods: Differential expression analysis between PC9 and gefitinib-resistant PC9 (PC9GR) cell samples was performed to screen GefR-related lncRNAs and mRNAs based on the GSE34228 dataset. These lncRNAs, mRNAs, and their corresponding microRNAs (miRNAs) were used to construct the GefR-related network and PPI networks. Functional enrichment analyses were conducted using the STRING database. A prognostic signature was developed using the TCGA dataset. The reliability of the signature was tested using the Kaplan–Meier method and ROC curve. Lastly, the FZD4-associated ceRNA subnetwork was selected to confirm the in vitro expressions of the GefR-related lncRNAs using RT-qPCR assay.
Results: A GefR-related ceRNA network that consists of 35 miRNAs, 26 lncRNAs, and 179 mRNAs was constructed. Then, 20 hub genes were screened from the targeted mRNAs of the constructed PPI network, and enrichment analysis identified relevant enriched pathways. We also constructed a prognostic signature for LUAD based on nine mRNAs in the GefR-related ceRNA network. The 9-mRNA signature was an independent predictor of LUAD, the AUC produced by ROC analysis showed a good predictive power of the model, and Kaplan–Meier analysis showed poorer outcomes in the high-risk group, relative to the low-risk group. Lastly, MIR137HG and ZNF295-AS1 levels were found to be associated with gefitinib resistance and exerted their functions through the ceRNA mechanism.
Conclusion: We established a prognostic signature and identified two lncRNAs (MIR137HG and ZNF295-AS1) with potential significant roles in gefitinib resistance.
Keywords: ceRNA, lncRNA, prognostic signature, EGFR TKIs, NSCLC, lung adenocarcinoma