LSCC, which is of epithelial origin, is a most common tumor of the upper respiratory tract[23]. Poor living habits, such as imbalanced diet, smoking, and alcohol consumption, are the main risk factors contributing to laryngeal cancer[24]. As an important feature of tumors, metabolic abnormality could adjust the microenvironment to meet the requirement of the constantly growing tumor cells[25]. Recently, more and more studies have focused on tumor metabolism to elucidate the pathogenesis of LSCC. Hu reported that PCK2 down-regulation inhibited the invasion, migration, and proliferation of laryngeal cancer under hypoxia, and therefore could be used as a new strategy for laryngeal cancer therapy[26]. Using liquid chromatograph-mass spectrometry (LC-MS) technology and molecular biology experiments, Zhao et al. found that fatty acid desaturase 1 (FADS1) promoted the progression of LSCC through activating AKT/mTOR signaling pathway[27].
In this study, we utilized several public databases and screened out prognostic metabolic markers of LSCC by bioinformatics analysis. It was worth noting that the results of GO and KEGG pathway enrichment analyses were all related to metabolism, such as purine metabolism, glycometabolism, lipid metabolism, and pyrimidine metabolism. Prognostic MRGs were screened out and a novel prognostic signature was then constructed based on the two MRGs (GPT and SMS).
GPT (glutamic pyruvic transaminase), also known as alanine aminotransferase, is strongly involved in cellular metabolism and cancer cell turnover[28]. GPT catalyzes the transfer of amino groups to generate products in gluconeogenesis and amino acid metabolism[29]. Accumulating evidence has shown that the expression level of GPT may be related to the prognosis of various cancers, including bladder cancer[30], pancreatic cancer[31], urothelial carcinoma[32], and small cell lung cancer[33]. The abnormal expression of GPT has a variety of causes, and long-term drinking can increase GPT expression[34]. Alcohol consumption is also an accepted risk factor for laryngeal cancer. Therefore, we speculated that chronic alcohol abuse might lead to the abnormal expression of GPT, thereby promoting the occurrence and development of laryngeal cancer. GPT is a common index to evaluate liver function, and the detection method is simple and cheap. Here, we found for the first time that GPT could be used as an important predictor of LSCC. If the role of GPT can be further proved by cohort and molecular biology experiments, it will bring great convenience for the diagnosis and prognosis of laryngeal cancer.
SMS (spermine synthase) is often associated with Snyder-Robinson syndrome (SRS), because SRS is caused by alterations in the human SMS gene[35, 36]. The level of polyamines, the aliphatic cations present in all cells, is intricately controlled by biosynthetic and catabolic enzymes. SMS, a highly specific aminopropyltransferase[37], is a biosynthetic enzyme[38]. Previous studies have shown that the increased activity of polyamine-producing enzymes leads to the up-regulation of polyamine level, which maintains the growth and proliferation of tumor cells[39]. Guo[40]found that SMS was overexpressed in colorectal cancer and maintained the colorectal cancer cell survival. Also, the high expression of SMS mRNA was related to a low overall survival rate and metastasis of triple-negative breast cancer[41]. As a spermine metabolism-related gene, the role of SMS in LSCC has not been reported, and its specific molecular mechanism is not clear. In our study, SMS expression was up-regulated in LSCC tissues and related to the prognosis of LSCC patients. We also found SMS could be a prognostic signature to predict the survival of LSCC patients. These results may shed light on the relationship between spermine metabolism and LSCC.
Concordant with previous reports, SMS was up-regulated while GPT was down-regulated in LSCC tissues compared to adjacent normal tissues as verified by clinical samples and multiple datasets in our study. Based on the expressions of SMS and GPT, LSCC patients were divided into the high and low risk groups. Five different cohorts, including internal and external validation cohorts, were analyzed and tested to verify the accuracy and reliability of the prognosis signature. The risk score of the signature was the best independent prognostic factor in both TCGA and FDEENT cohorts. Therefore, a nomogram based on the risk score was developed, which might contribute to treatment decision making.
In the previous studies, Chen[42] and Li[43] developed metabolism-related signatures to predict the prognosis of head and neck squamous cell carcinoma, while Hu[44] identified a combined lipid metabolism-related signature for oral squamous cell carcinoma (OSCC). To our knowledge, our study is the first to develop and validate an MRG signature for laryngeal squamous cell carcinoma (LSCC). This signature showed a higher AUC value than the other studies based on TCGA cohort (Figure 8I), indicating a higher sensitivity and specificity. In addition, compared with other prognostic signatures composed of multiple genes, this signature derived from two MRGs is easier to apply in clinical practice.
Also, there are some deficiencies in our research. First, only five cohorts were incorporated to verify this prognostic signature. In future study, more cohorts and LSCC patients should be included to reduce the deviation of racial and geographic distribution. Second, our findings were based on retrospective studies, prospective clinical trials should be conducted. More importantly, in vivo and in vitro experiments on GPT and SMS are needed.