Despite advances in improving the roles of genomic and immunity in HNSCC, to date biomarkers is still lacked. Additional biomarkers in the clinical will be needed to guide the prognosis in this cancer. Recently, studies found some mRNAs could be biomarkers to predict the prognosis for HNSCC and indicated they have the clinical significance. For example, Qun Li et al. [20] confirm that levels of P4HA1 mRNA and protein in tissues were obviously higher in HNSCC than that in controls. High expression was related to tumor category, lymphatic metastasis and pathological stage. Xu He et al. [21] downloaded the GSE9716, GSE61772 and GSE20549 from the Gene Expression Omnibus (GEO) database and reported that high POPDC3 expression associated with poor a prognosis for patients with HNSCC. But these genes cannot accurately predict patient survival because a single gene is modulated by many factors. So, gene signature including multiple genes has been the important model to predict outcome in cancers. Results from these signatures reflect more effects and present the more accurate prediction compared to biomarker of a single gene.
Some signatures have been discovered in cancers, a gene glycolysis-related signature consists of HMMR, B4GALT1, SLC16A3, ANGPTL4, EXT1, GPC1, RBCK1, SOD1, and AGRN could be a biomarker which were correlated with metastasis and overall survival in lung adenocarcinoma [22]. Using multivariate analysis, Pankaj Ahluwalia et al. [23] identified a novel prognostic signature which has clinical utility in colorectal cancer. In this study, we also conducted the GSEA to analyze the mRNA expression data from TCGA database and nine mRNAs expression levels have the differences. Additionally, Furthermore, in univariate Cox regression and multivariate Cox regression analyses, we identified that a signature of the nine genes pose the value to predict the prognosis to HNSCC. Consequently, the risk signature also can confirm the clinical results. Interestingly, the high-risk score was correlated with a poor prognosis. Glucose-6-phosphate dehydrogenase (G6PD) is a vital enzyme in preventing cells from oxidative damage. Some studies have found that G6PD play an important role in cancers to influence its occurrence and development. Chin An Yang et al. [24] suggested that G6PD could be a marker to predict risk, prognosis and chemo-sensitivity of glioma. Epidermal growth factor receptor (EGFR) is a key gene to regulate biological behavior of tumor cells. EGFR-targeted drugs have widely been utilized in therapy of lung and colorectal cancers. Xiaozhen Liu et al. [25] recruited 200 molecular apocrine breast cancer patients and found that EGFR may be helpful to assess the prognosis of the cancer. Expression of ALDH2 were significantly associated with poor prognosis in hepatocellular carcinoma patients [26]. We also found the protective role in prognosis of HNSCC. Intravesical recurrence-free survival of patients with positive GPR87 tumors is shorter than those with negative GPR87 tumors [27]. Overexpression of STC2 can suppress cell apoptosis and promote cell proliferation, migration, invasion, and make cell cycle arresting at the G1/S transition in HNSCC [28]. Consistent with this, in this study, the over expression impacted the poor prognosis of HNSCC. Multivariate analysis indicated that high expression of PDK3 acted an independent risk factor in event-free survival and overall survival of acute myeloid leukemia [29]. ELF3 can regulate the cell cycle and epithelial-mesenchymal transition (EMT) in non-small cell lung cancer through PI3K/AKT and ERK signaling pathways and downstream effectors [30]. H M H Abaza et al. [31] reported expression of STC1 gene could be a useful biomarker for predicting clinical outcome and monitoring therapeutic response in acute leukemia patients. GNPDA1 had good effects on CD4 + memory T cells, which were affected by the gemcitabine therapeutic effect in gemcitabine-resistant pancreatic cancer cell [32]. These results are in line with our that these genes served as carcinogenic roles or cancerocidal roles. As far as know, the mRNAs signature can perform better predict the prognosis of HNSCC.
Increased glycolysis is one of metabolic characteristics named Warburg effect. Deregulation or alteration of energy metabolism has been considered an important hallmark of cancer. Gangcai Zhu et al. [33] revealed that hypoxia can promote migration, invasion and glycolysis in HNSCC by HIF-1α-metadherin loop. Decreased expression of GRIM-19 enhanced aerobic glycolysis in HNSCC through DNA hypermethylation [34]. Moreover, high expression of HK2 was tested in head and neck cancerous tissues compared to normal counterparts. HK2-deficient HNSCC cells had greater sensitivity to cisplatin and 5-fluorouracil [35]. It is crucial that hexokinase is a key enzyme glycolytic prognosis. Hence, metabolism of glycolysis may involve the development and progression of HNSCC. Some single gene involving glycolysis which predict prognosis of HNSCC has been reported in the study, but no report in the glycolysis-related gene signatures. We first reported and illustrated that the glycolysis-gene signature including G6PD, EGFR, ALDH2, GPR87, STC2, PDK3, ELF3, STC1 and GNPDA1 which has prognostic value for HNSCC.
Over all, this study firstly identified that a glycolysis related gene signature (G6PD, EGFR, ALDH2, GPR87, STC2, PDK3, ELF3, STC1 and GNPDA1) can predict the prognosis of patients with HNSCC. This signature could be a new target for the therapy of HNSCC.