Six Glycolysis-Related Genes as Prognostic Risk Markers Can Predict the Prognosis of Patients with Head and Neck Squamous Cell Carcinoma

Abstract

Objective: Head and neck squamous cell carcinoma (HNSCC) is one of the worst-prognosis malignant tumors. This study used bioinformatic analysis of the transcriptome sequencing data of HNSCC and the patients' survival and clinical data to construct a prediction signature of glycolysis-related genes as the prognostic risk markers.

Methods: Gene expression profile data about HNSCC tissues (n = 498) and normal tissues in the head and neck (n = 44) were got from The Cancer Genome Atlas (TCGA), as well as patients' survival and clinical data. Then, we obtained core genes; their expression in head and neck squamous cell carcinoma tissues is significantly different from that in normal head and neck tissues. The predicted glycolysis-related genes are screened through univariate Cox regression analysis, and then, the prognostic risk markers were constructed through further correction of multivariate Cox regression analysis. The Kaplan-Meier curve and receiver operating characteristic curve are used to analyze the potential value of these risk markers in diagnosis and prognosis. We also evaluated that the glycolysis-related prognostic risk markers composed of 6 oncogenes are correlated with clinical features, such as age, gender, grade, and clinical stage of the tumor, by univariate and multivariate Cox regression analyses.

Results: Differentially expressed glycolytic genes in HNSCC tissues and normal head and neck tissues were screened from TCGA databases using the bioinformatic method. We confirmed a set of six glycolytic genes that were significantly associated with OS in the test series. According to our analysis, the prognostic risk markers composed of HPRT1, STC2, PLCB3, GPR87, PYGL, and SLC5A12 may be an independent risk factor for the prognosis of HNSCC.

Conclusions: Through this analysis, we constructed new prognostic risk markers related to glycolysis as a prognostic risk marker for patients with HNSCC and provided new ideas and molecular targets for the research and individualized treatment of HNSCC.

Figure 1

GSEA of glycolysis-related gene sets. Enrichment plots of four glycolysis-related gene sets between HNSCC and paired normal tissues identified by GSEA.

Figure 2

The functional enrichment analysis of the core genes. Gene Ontology (GO) terms (a) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways (b) were significantly enriched by the core genes.

Figure 3

Risk score based on the 6 glycolysis-related gene prognostic risk markers in patients with HNSCC. (a) Kaplan-Meier curve of OS in the high- and low-risk groups. (b) Time-dependent ROC curves of the 6 glycolysis-related gene signatures for prediction of 3-, 5-, and 10-year OS. (c) The distribution of the 6 glycolysis-related gene risk scores and survival status for each patient.

Figure 4

Univariable and multivariable analyses for the risk score and each clinical feature.

Figure 5

Stratification analysis of various clinicopathological factors by Kaplan-Meier curves for the patients with HNSCC in the TCGA dataset. Kaplan-Meier curves of OS in different subgroups stratified by (a) age, (b) gender, (c) grade, (d) T stage, (e) N stage, and (f) AJCC stage.

Figure 6

Analysis of the expression of 6 genes in the sample (a) to screen the mutations of genes in sample of M cancer patients (b); the differential expression analysis of the 6 screened genes (∗means < 0.05, ∗∗means < 0.01, and ∗∗∗means < 0.001).

Figure 7

The expression based on each glycolysis-related gene predicts OS in patients with HNSCC and Kaplan-Meier curve of OS in the high- and low-expression groups. (b) Time-dependent ROC curves of each glycolysis-related gene for prediction of 3-, 5-, and 10-year OS.