A Five-lncRNAs Signature-Derived Risk Score Based on TCGA and CGGA for Glioblastoma: Potential Prospects for Treatment Evaluation and Prognostic Prediction.

Xuegang Niu,Tongshuo Zhang,Tao Fang,Lingyin Meng,Jiangnan Sun,Jipeng Jiang,Huanming Li

doi:10.3389/fonc.2020.590352

Abstract

Accumulating studies have confirmed the crucial role of long non-coding RNAs (ncRNAs) as favorable biomarkers for cancer diagnosis, therapy, and prognosis prediction. In our recent study, we established a robust model which is based on multi-gene signature to predict the therapeutic efficacy and prognosis in glioblastoma (GBM), based on Chinese Glioma Genome Atlas (CGGA) and The Cancer Genome Atlas (TCGA) databases. lncRNA-seq data of GBM from TCGA and CGGA datasets were used to identify differentially expressed genes (DEGs) compared to normal brain tissues. The DEGs were then used for survival analysis by univariate and multivariate COX regression. Then we established a risk score model, depending on the gene signature of multiple survival-associated DEGs. Subsequently, Kaplan-Meier analysis was used for estimating the prognostic and predictive role of the model. Gene set enrichment analysis (GSEA) was applied to investigate the potential pathways associated to high-risk score by the R package “cluster profile” and Wiki-pathway. And five survival associated lncRNAs of GBM were identified: LNC01545, WDR11-AS1, NDUFA6-DT, FRY-AS1, TBX5-AS1. Then the risk score model was established and shows a desirable function for predicting overall survival (OS) in the GBM patients, which means the high-risk score significantly correlated with lower OS both in TCGA and CGGA cohort. GSEA showed that the high-risk score was enriched with PI3K-Akt, VEGFA-VEGFR2, TGF-beta, Notch, T-Cell pathways. Collectively, the five-lncRNAs signature-derived risk score presented satisfactory efficacies in predicting the therapeutic efficacy and prognosis in GBM and will be significant for guiding therapeutic strategies and research direction for GBM.

Highlights

As one of the most common malignant brain tumor [1, 2], the 5year survival rate for glioblastoma (GBM) in patients is less than 5%
Can we create a model on the basis of a multiple-gene signature that can advance the effectiveness of treatment evaluation and prognostic prediction for GBM? In the present study, we collected high-throughput data in The Cancer Genome Atlas (TCGA) and Chinese Glioma Genome Atlas (CGGA) databases which were generated by microarrays and nextgeneration sequencing, we identified some survival-related differentially expressed genes (DEGs) by analyzing the data of Long-noncoding RNA (lncRNA) expression in GBM, subsequently, a risk model for treatment evaluation and prognostic prediction was established on the basis of the identified gene signature
We selected the common genes of the TCGA and CGGA data sets, including 930 common lncRNAs

Summary

Introduction

As one of the most common malignant brain tumor [1, 2], the 5year survival rate for glioblastoma (GBM) in patients is less than 5%. The abnormal expression of lncRNA is tightly connected to the occurrence, prognosis, and survival of patients with cancer [6, 7]. Numerous studies have illustrated that certain lncRNAs are aberrantly expressed in GBM tissue, and many of them have been confirmed to be involved in tumor invasion, immune escape and radiation resistance. Specific lncRNAs were identified as prognostic biomarkers and therapeutic targets for GBM, while some of them were proposed to be novel indicators for survival prediction in GBM patients

Methods

Results

Conclusion