In recent years, a multiple of studies have suggested the crucial roles of RBPs in the development and progression of various human cancers [23–25]. However, a comprehensive analysis of the potential function of RBPs on NSCLC tumorigenesis and prognosis has not been performed. In present study, A total of 459 differentially expressed RBPs were identified between cancer and control tissue. Functional enrichment analysis was performed by 459 differentially expressed RBPs. Thirteen RBPs were independently identified as prognostic RBPs by multivariate Cox regression analysis and were used to construct the prognostic RBPs signature. The high-risk group in this prognostic RBPs signature was associated with a shorter survival time. Furthermore, the risk score was an independent prognostic factor of NSCLC. In addition, the signature was validated by the testing cohort and differently clinical subgroup. Therefore, A RBPs-associated prognostic signature was successfully constructed with highly diagnostic value, which could apply to predict prognosis of patients with NSCLC and guide clinical decision-making.
The biological process in NSCLC was primarily associated with RNA splicing, ncRNA metabolic process, ribonucleoprotein complex biogenesis, regulation of translation, RNA catabolic process. Abnormality in RNA splicing have been associated with tumor initiation and progression [26, 27]. The ribonucleoprotein complex involved in initiation, progression and prognosis of cancer [28, 29]. In addition, in cancer cells, lncRNA were aberrantly expressed as classical oncogenes or tumor suppressors and correlate with the altered metabolism [30, 31]. Given the direct control of translation by tRNA modifications, tRNA modifying enzymes may function as regulators of cancer progression [32]. KEGG pathways were mainly enriched in mRNA surveillance pathway, RNA degradation, RNA transport, Ribosome biogenesis in eukaryotes. The previous research has reported that mRNA surveillance pathway has involved with initiation and progression of cancer [33–35]. RBPs were associated with tumorigenicity and metastasis involved in mRNA stability [11, 18]. In addition, the RNA degradation pathway was involved in drug anti-oral tumorigenesis [36]. Alterations of crucial pathways underlying mRNA nuclear export could result in genome instability and cancer initiation [37]. Given the important function of ribosome and ribosomal proteins (RPs) in cells, ribosomal dysfunction would result in a variety of diseases [38–41] and RPL15, a large ribosomal subunit protein, was remarkably upregulated in human cancer tissues and cultured cell lines and was closely correlated with clinicopathological characteristics [42]. However, in NSCLC, those molecular mechanisms and pathways need further explore in the future.
Several RBPs in our prognostic signature had been studied in human disease, especially in cancer. IGF2BP1 silencing could inhibit the occurrence of tumor event in NSCLC [43]. In addition, the CASC3 overexpressed in cancer tissue and promoted cancer cell proliferation, migration, invasion, and glycolysis [44, 45]. ZC3H12C could inhibit vascular inflammation and proinflammatory gene expression in endothelial cells [46] and regulate proinflammatory activation of macrophages [47, 48]. ZC3H12D contributed to inflammation-associated signaling pathways involved in the pathogenesis of Leukoaraiosis [49]. BOP1 could prompt cell migration and invasion in colorectal cancer [50] and loss of BOP1 conferred resistance to BRAF kinase inhibitors in melanoma [51]. DDX24 negatively regulated cytosolic RNA-mediated innate immune signaling [52]. FASTKD3 was an essential component of mitochondrial respiration that may modulate energy balance in cells exposed to adverse conditions by functionally coupling mitochondrial protein synthesis to respiration [53]. TARBP1 was overexpressed in NSCLC and hepatocellular cancer, and was associated with pathological grade, clinical stage and pathological type [54, 55]. Increased levels of INTS7 in prostate cancer was associated with an aggressive phenotype [56]. NOL12 was a multifunctional RNA binding protein at the nexus of RNA and DNA metabolism [57]. NOL12 repression induced nucleolar stress-driven cellular senescence and was associated with normative aging [58].PABPC1L might promote colon tumorigenesis by regulating mRNA splicing [59]. In addition, PABPC1L depletion inhibited proliferation and migration via blockage of AKT pathway in colorectal cancer [60]. In current bioinformatic analysis, the level of ZC3H12D, BOP1, IGF2BP1, KHDC1, FASTKD3, TARBP1, INTS7, NOL12, SNRPB, PABPC1L mRNA and protein in lung cancer tissue were higher than that in control tissue and were associated with survival time. However, the level of ZC3H12C, CASC3, DDX24 mRNA and protein in lung cancer tissue were lower compared with control tissue and were also associated with survival time. Therefore, the thirteen RBPs involved in tumorigenicity and prognosis of NSCLC, which deserved further explore the molecular mechanism in vitro and in vivo experiments in the future.
The previous research has demonstrated that prognostic RBPs signature could predict the prognosis of several of cancer [61, 62]. However, in lung cancer, the previous prognostic RBPs signature was constructed based on lung squamous cell carcinoma or lung adenocarcinoma [20–22], which only predict a part of patients. A prognostic RBPs signature was constructed by multivariable Cox regression analysis concluding 13 differentially expressed RBPs (ZC3H12C, ZC3H12D, BOP1, CASC3, DDX24, IGF2BP1, KHDC1, FASTKD3, TARBP1, INTS7, NOL12, SNRPB, PABPC1L) in the current study, which could predict most lung cancer patients. The high-risk group had a shorter survival time compared to the low-risk group in training and testing cohort (both P < 0.001). Area under ROC curve was 0.703 in training cohort and 0.636 in testing cohort with moderate accuracy and specificity. In addition, risk score was an independent prognostic factor of NSCLC by the univariate Cox analysis and the multivariate Cox analysis. Furthermore, the prognostic RBPs signature was applied to differently clinical subgroup ( > = 65, < 65, female, male, stage I-II, III- IV, T1-2, T3-4, N0, N1-3, M0 and M1), and high-risk group in all subgroup has a shorter survival time compared with the low-risk group, which further validated the prognostic RBPs signature. Therefore, the RBPs-associated prognostic signature was successfully constructed. In addition, A nomogram was established to predict 1-year, 3-year and 5-year survival of patients with NSCLC.
There were some highlights in current study. Our study provided innovative insights into the roles of RBPs during NSCLC development and progression and established a promising prognostic signature to evaluate patient prognosis. In addition, the prognostic RBPs signature was validated by the testing cohort and differently clinical subgroup, and had a good predict value. Nevertheless, this study had several limitations which could not be ignored. First, our results were based on the bioinformatics analysis of the TCGA data, which need to be validated in clinical patient cohort and multicenter and prospective study. Second, the prognostic RBPs was not validated by qPCR. Finally, further studies were required to clarify the molecular mechanism for clinical practice in the future.