Colorectal cancer is a highly malignant tumor, which is particularly prone to liver and lung metastasis, seriously affecting the survival prognosis of patients[24]. Therefore, finding a prognostic marker with high specificity and sensitivity is the advisable choice applied in the clinical practice. Extensive evidence displayed RBPs play a pivotal role in the development and progression of various human tumors[25–27]. However, only a small part of RBPs have been studied in-depth and partially confirmed that they contributed to occurrence and development of cancers[28].
In the current study, we download mRNA expression profiles and corresponding patients’ clinical information of CRC from TCGA database. 224 RBPs identified as significantly different by using the Wilcox Test for the differential analysis, consisting of 146 up-regulated RBPs and 78 down-regulated RBPs. Then, we systematically explored the potential functional pathways and constructed a PPI network of these differently expressed RBPs. Moreover, univariate Cox regression analysis, survival analysis, multivariate stepwise Cox regression analysis and ROC analysis were performed on Hub RBPs to explore its biological function and clinical significance further. These findings may contribute to developing novel biomarkers for prognosis of CRC patients.
The function enrichment analysis demonstrated that the differently expressed RBPs were significantly related to regulation of translation, RNA splicing, cytoplasmic ribonucleoprotein granule, ribonucleoprotein granule, catalytic activity, acting on RNA, mRNA 3'−UTR binding. The KEGG pathways analysis showed that the aberrantly expressed RBPs regulate the tumorigenesis and progression of CRC by affecting mRNA surveillance pathway, RNA transport, TGF − beta signaling pathway. The above information suggests that most of the enrichment pathways relate to RNA modification, which also indicates that the modification and processing of RNA is related to tumorigenesis. RNA metabolism and RNA processing have been increasingly recognized in various diseases[29, 30], Post-transcriptional regulation of RNA stability is an important procedure in gene expression processing. Previous studies have proved that RBPs as a key participant in post-transcriptional regulation, and are involved in most processes such as cell proliferation, differentiation, migration, apoptosis and angiogenesis by regulating the expression of related proteins. HuR, a member of ELA V family of RBPs ,High HuR protein expression is found in both the nucleus and cytoplasm of human colorectal cancer [14]. HuR protein can recognize and combined messenger RNA of target genes, thus stable target genes to inhibit the degradation of messenger RNA, participate in the post-transcriptional regulation of target genes, and regulate a variety of physiological and biochemical processes such as tumor cell growth, proliferation, differentiation, signal transduction, transcription and translation, angiogenesis, protein transport, cell apoptosis and its metabolism[31]. The insulin-like growth factor-2 mRNA binding proteins (IGF2BPs or IMPs) belong to a conserved subfamily of RBPs. The IMPs have been studied for their roles in regulation of post-transcriptional processes such as mRNA localization, turnover, and translational control [32]. IMP1 plays a functional role in the RNA stability by binding and shielding several mRNAs that play critical roles in cell growth and proliferation from proteolytic degradation [33]. IMP1 is overexpressed in more than 80% of human CRC[10], which regulates cell cycle progression and migration [34], and correlates with invasion, lymph node metastasis, and worse prognosis[11, 35, 36]. Moreover, ribonucleoprotein granule is an important area that implements protein synthesis. The mutation of ribonucleoprotein regulates the translation process and associated with tumor development [37]. By applying whole-exome sequencing, RNA seq, or whole-genome sequencing, ribonucleoprotein regulates gene mutations have been detected in the genome of cancer cells, including from endometrial cancer, T-cell acute lymphoblastic leukemia (T-ALL), chronic lymphocytic leukemia (CLL), colorectal carcinomas, and high-grade gliomas [38–41]. In brief, RBPs can bind to mRNAs and regulate their expression, which play an important role in the development of many diseases.
Furthermore, the hub RBPs were selected based on univariate Cox regression analysis, and multiple Cox regression analysis. 8 feature genes associated with CRC patient’s survival were identified, including BRCA1, TERT, TDRD7, PPARGC1A, LUZP4, CELF4, ZC3H12C, and PNLDC1. Among these genes, several RBPs have been reported to be closely associated with colorectal cancer. BRCA1 associated hereditary breast cancer, and recently, meta-analysis results provide clinicians and healthcare regulatory agencies with evidence of the increased risk of colorectal cancer in BRCA1 mutation carriers[42]. The human telomerase reverse transcriptase (TERT) also plays an important role in colorectal cancer growth, initially affecting the tumor's stromal microenvironment[43]. CUG binding protein 4 (CUBP4) or CELF4 is a multifunctional RBP studied primarily for its role in RNA metabolism related processes like decay, translation and splicing. Previous studies have shown that CELF2 expression is consistently reduced during neoplastic transformation, suggesting that it might play a crucial role in tumor initiation and progression[44]. It is consistent with our findings. PPARGC1A was indicated as a tumor suppressor in colorectal cancer[45] and ovarian cancer[46], as well as a negative prognostic biomarker for CRC. Although the connection between most of the differently expressed RBPs and CRC remains unclear, some RBPs have been reported to be associated with other tumors. LUZP is a regulatory protein that contains 20 amino acids, Excessive expression or mutation of LUZP may lead to the occurrence of malignant tumors, and abnormal expression of LUZP can be detected in solid tumors such as cervical cancer, breast cancer and pancreatic cancer[47–49]. ZC3H12A, also known as MCPIP1 or Regnase-1, is a novel RNA-binding protein (RBP) that plays a key role in post-transcriptional regulation and immune homeostasis[50, 51], a study shows that ZC3H12A is involved in circHECTD1/HECTD1-mediated macrophage activation in lung cancer[52]. PNLDC1 is a PARN-like 3’-to-5’ exonuclease localized on the mitochondrial surface in mouse, Associated with the emergence of the mature piRNA[53]. TUDOR domain-containing proteins (TDRDs) are chiefly responsible for recognizing methyl-lysine/arginine residue. TDRD dysregulation contributes to breast tumorigenesis[54].
Then, an 8-gene signature-based risk score formula was developed for survival analysis in the training and test set. High expression of BRCA1, TDRD7, PPARGC1A, and ZC3H12C was associated with a good prognosis in patients with CRC, whereas that of TERT, LUZP4, CELF4, and PNLDC1 were related to poor prognosis. Next, the eight RBPs were used to construct a risk model by multiple stepwise Cox regression analysis to predict prognosis in CRC patients. ROC analysis was performed and showed the higher accuracy of the risk score model both for the training set (1-year AUC = 0.670; 3-year AUC = 0.675;5-year AUC = 0.678) and test set (1-year AUC = 0.701; 3-year AUC = 0.0.709;5-year AUC = 0.730). According to our analysis, the signature-based risk score was an independent prognostic factor in CRC patients. A nomogram was constructed to enable practitioners to predict1-year, 3-year, and 5-year OS of CRC patients.
Nevertheless, our study had several limitations. Firstly, our prognostic model was only based on the data from TCGA database, this study was a pure bioinformatics study, and the scientific hypothesis was not proved by biological experiments. Secondly, our study was designed based on a retrospective analysis, and a prospective study should be performed to verify the results. Thirdly, the lack of some clinical characteristics in the datasets from TCGA may have decreased the statistical effectiveness and reliability of the multivariate stepwise Cox regression analysis.
In summary, our study not only constructed a new predictive model of RBPs signature prognosis but also by grouping to verify and evaluate the predictive ability of the model. The most important thing is can be used as an independent prognostic factor in CRC. As far as we know, this is the first report of developing a RBPs-associated prognostic model for CRC. Our results would greatly contribute to show the pathogenesis of CRC and enhance our comprehension to treatment and progression of colorectal cancer.