Identification of a Novel Five-Gene Prognostic Model for Laryngeal Cancer Associated with Mitophagy Using Integrated Bioinformatics Analysis and Experimental Verification

Abstract

Laryngeal cancer (LC) is a prevailing tumor with a high mortality rate. The pivotal role of mitophagy in LC is acknowledged; however, a comprehensive analysis of the corresponding genes has not been conducted. In the present study, we proposed a prognostic model consisting of mitophagy-related genes in LC. Clinical information and transcriptome profiling of patients with LC and mitophagy-related genes were retrieved from open-source databases. Gene set variation analysis (GSVA) and Weighted Gene Co-expression Network Analysis (WGCNA) were used to identify core mitophagy-related genes and construct gene co-expression networks. Functional enrichment analysis was employed to analyze the enriched regulatory pathways of the mitophagy-related genes. Kaplan-Meier curves (KM), Cox, and LASSO regression were applied to explore their prognostic effects. Finally, quantitative real-time PCR (RT-qPCR) further verified the bioinformatics prediction. A total of 45 genes related to mitochondrial pathways was collected. GSVA analysis demonstrated that these genes in tumor samples mainly referred to the mitochondrial pathway. Among these genes, five mitophagy-related-gene signatures (CERCAM, CHPF, EPHX3, EXT2, and MED15) were further identified to construct the prognostic model. KM and Cox regression analyses indicated that this model had an accurate prognostic prediction for LC. RT-qPCR showed that CERCAM, CHPF, EXT2, and MED15 expression were upregulated, and EPHX3 level was decreased in LC cells. The present study established a five-mitophagy-related-gene model that can predict the prognosis of LC patients, thus laying the foundation for a better understanding and potential advancements in clinical treatments for LC.