Glioma, the most frequent malignant primary intracranial tumors in adults, seriously threatens human health because of its malignant progression and poor prognosis[18]. Even with the most comprehensive treatments for glioma patients (such as surgery, radiation, chemotherapy and immunotherapy), there still occupies an unsatisfying median survival of 15 months[6, 19]. Although the treatment approaches of gliomas have improved constantly over the past few years, it is still difficult to cure completely, and patients are prone to recurrence after the initial treatment, leading to poor clinical prognosis among glioma patients[20–22]. Therefore, some novel biomarkers need to be discovered as potential therapeutic targets to improve prognosis.
Reprogramming of energy metabolism is considered as one of the cancer hallmarks[23, 24]. Tumor cells primarily utilize the glycolysis pathway for energy supply and glycolysis rate is 100 times higher compared to the normal cells[25, 26]. Thus, this trait of tumor has attracted prodigious attention from researchers in recent years. Mounting studies have revealed that glycolysis pathway exhibits a high prognostic value in glioma patients[27, 28]. Multiple well-known biomarkers of glioma are associated with the glycolysis pathway. For example, compared to the IDH1 wild-type, the intermediate products of glycolysis pathway are obviously reduced in the glioma tissues with IDH1 mutation[29]. Ras inhibition can cause glycolysis pathway shutdown and might therefore block tumor cell invasiveness, survival in glioma[30]. Down-regulation of EGFR, a significant oncogenic signature in glioma, can trigger regression of the glioblastoma through reversing the Warburg effect[31]. At the same time, LncRNAs have been proven to be implicated in considerable human diseases including cancers[32]. Numerous studies have evidenced that some signaling transduction and function related LncRNAs, such as immune-related[33] and autophagy-related LncRNAs[34], can affect the development and progression of the glioma.
In 2007, Chen et al. constructed a five gene prognostic signature for NSCLC[35]. Since then, gene signature, also known as classifier, has been regularly utilized to predict prognosis of various cancers including glioma. For example, an immune-related six-LncRNAs was developed to predict the prognosis of glioma[36]. Considering the significance of the glycolysis pathway in cancer, we construct a glycolysis-related LncRNAs signature through various methods such as LASSO regression, Cox regression and this signature is independently associated with the glioma prognosis. In addition, two groups based on risk score are analyzed by Kaplan–Meier survival curves analysis, ROC curve analysis, multivariate Cox regression analysis, and GSEA. All the above results prove that our signature is significantly related to OS and can be further considered to be a novel potential molecular therapeutic target.
In our signature, we screen and identify seven LncRNAs (AC093627.7, AC093627.10, RP11-359G22.2, LINC01272, AC092484.1, AC026904.1, and RP11-294N21.3). LINC01272 is considered as a new predicter biomarkers of gastric cancer and is closely related to the cancer cell proliferation and cell migration[37]. Li et al. has reported that LncRNAs AC026904.1 could regulate SLUG expression at both transcriptional and post-transcriptional levels, exerting critical roles in epithelial-mesenchymal transition in breast cancer[38]. AC093627.7, AC093627.10, RP11-359G22.2, AC092484.1, RP11-294N21.3 have been reported for the first time in our present study. These LncRNAs may play an unknown biological role in glioma. According to the results of Kaplan–Meier survival curves analysis in our study, all the identified seven LncRNAs are associated with the survival status of glioma patients. The patients with high expression of RP11-294N21.3 can live longer time. On the contrary, the patients with low expression of AC093627.7, AC093627.10, RP11-359G22.2, LINC01272, AC092484.1, AC026904.1 live longer time.
The results of GSEA enrichment analysis demonstrate that several pathways and biology functions are positively associated with risk scores such as p53 signaling pathway, cell cycle pathway, DNA replication pathway, DNA post-replication repair pathway. The p53 is a famous tumor suppressed factor which is significantly associated with some important roles such as DNA damage and oncogenic signaling, and its dysfunction can promote cancer development and deterioration[39, 40]. Deviant cell cycle is the most common variation during tumor onset and development[41]. Abnormal DNA replication and DNA post-replication repair process may induce the accumulation of genetic aberrations that promote diseases such as cancer[42]. Thus, it is speculated that the identified seven LncRNAs might impact cell cycle, cell proliferation, DNA replication, and p53 signaling pathway through glycolytic biologic processes.
Although the seven glycolysis-related LncRNAs prognostic signature identified in our study is robust, there are still several limitations. Firstly, we need to validate functional features of this seven LncRNAs signature through basic biological experiments (in vivo or in vitro) and clinical researches. Secondly, the whole seven glycolysis-related LncRNAs, identified in our study using TCGA database are not completely comprised in other databases such as Chinese Glioma Genome Atlas (CGGA) or Gene Expression Omnibus (GEO), so we have to divide the total TCGA glioma data into a training cohort and a validation cohort by using random number manner, which limited the generalizability of our conclusions to some extent.