A Novel Signature Constructed by RNA-Binding Protein Coding Genes to Improve Overall Survival Prediction of Glioma Patients

Zewei Tu; Lei Shu; Jingying Li; Lei Wu; Chuming Tao; Minhua Ye; Xingen Zhu; Kai Huang

doi:10.3389/fcell.2020.588368

A Novel Signature Constructed by RNA-Binding Protein Coding Genes to Improve Overall Survival Prediction of Glioma Patients

Front Cell Dev Biol. 2021 Jan 28:8:588368. doi: 10.3389/fcell.2020.588368. eCollection 2020.

Authors

Zewei Tu¹, Lei Shu^{1

2

3}, Jingying Li⁴, Lei Wu¹, Chuming Tao^{1

2

3}, Minhua Ye¹, Xingen Zhu^{1

3}, Kai Huang^{1

3}

Affiliations

¹ Department of Neurosurgery, The Second Affiliated Hospital of Nanchang University, Nanchang, China.
² East China Institute of Digital Medical Engineering, Shangrao, China.
³ Institute of Neuroscience, Nanchang University, Nanchang, China.
⁴ Department of Comprehensive Intensive Care Unit, The Second Affiliated Hospital of Nanchang University, Nanchang, China.

Abstract

RNA binding proteins (RBPs) have been reported to be involved in cancer malignancy but related functions in glioma have been less studied. Herein, we screened 14 prognostic RBP genes and constructed a risk signature to predict the prognosis of glioma patients. Univariate Cox regression was used to identify overall survival (OS)-related RBP genes. Prognostic RBP genes were screened and used to establish the RBP-signature using the least absolute shrinkage and selection operator (Lasso) method in The Cancer Genome Atlas (TCGA) cohort. The 14 RBP genes signature showed robust and stable prognostic value in the TCGA training (n = 562) cohort and in three independent validation cohorts (Chinese Glioma Genome Atlas [CGGA]seq1, CGGAseq2, and GSE16011 datasets comprising 303, 619, and 250 glioma patients, respectively). Risk scores were calculated for each patient and high-risk gliomas were defined by the median risk score in each cohort. Survival analysis in subgroups of glioma patients showed that the RBP-signature retained its prognostic value in low-grade gliomas (LGGs) and glioblastomas (GBM)s. Univariate and multivariate Cox regression analysis in each dataset and the meta cohort revealed that the RBP-signature stratification could efficiently recognize high-risk gliomas [Hazard Ratio (HR):3.662, 95% confidence interval (CI): 3.187-4.208, p < 0.001] and was an independent prognostic factor for OS (HR:1.594, 95% CI: 1.244-2.043, p < 0.001). Biological process and KEGG pathway analysis revealed the RBP gene signature was associated with immune cell activation, the p53 signaling pathway, and the PI3K-Akt signaling pathway and so on. Moreover, a nomogram model was constructed for clinical application of the RBP-signature, which showed stable predictive ability. In summary, the RBP-signature could be a robust indicator for prognostic evaluation and identifying high-risk glioma patients.

Keywords: RNA binding protein; biomarker; glioma; overall survival; prognostic signature.