In silico identification of the prognostic biomarkers and therapeutic targets associated with cancer stem cell characteristics of glioma

Wen Lvu, Fei Xu, Cheng Chen, Bo ZhangBioscience 2020 · DOI

TL;DR

This study uses bioinformatics to identify key genes associated with cancer stem cell (CSC) characteristics in glioma, a challenging brain cancer. They developed a predictive model using these genes that accurately estimates patient survival and identified three novel biomarkers (F2RL2, CLCNKA, and LOXL4) that could lead to better diagnostic and therapeutic strategies.

What Problem Does This Paper Address?

Glioma is a highly aggressive and deadly form of brain cancer with a high mortality rate. Cancer stem cells (CSCs) are thought to play a crucial role in glioma's initiation, progression, and resistance to treatment. However, the specific biomarkers and therapeutic targets related to CSCs in glioma are not well understood. This lack of understanding hinders the development of effective diagnostic tools and treatments. This paper aims to bridge this gap by identifying genes linked to CSC characteristics that can predict patient outcomes and potentially serve as therapeutic targets.

Methods

The researchers employed a multi-step bioinformatics approach:

  1. Data Acquisition: They obtained RNA sequencing (RNA-seq) data and clinical information for glioma patients from The Cancer Genome Atlas (TCGA) database.
  2. Stemness Index Calculation: They calculated the mRNA expression-based stemness index (mRNAsi) for each patient using an algorithm called OCLR. This index reflects the degree of dedifferentiation and CSC activity.
  3. Differential Gene Expression Analysis: They identified differentially expressed genes (DEGs) between gliomas with high and low mRNAsi across different histological grades (G2, G3, and G4).
  4. Identification of Stemness-Related Genes (SRGs): They identified overlapping DEGs across the different grades to pinpoint SRGs.
  5. Prognostic SRG Screening: They used univariate and multivariate Cox regression analyses to identify SRGs that were significantly associated with patient prognosis.
  6. Model Development: They built a predictive model (nomogram) based on the identified prognostic SRGs to estimate 3- and 5-year overall survival (OS) probabilities.
  7. Validation: The predictive model was validated using an external dataset from the Chinese Glioma Genome Atlas (CGGA).

Datasets

  • The Cancer Genome Atlas (TCGA): This database provided the primary RNA-seq data and clinical information for 705 primary brain glioma patients (532 LGGs and 173 GBMs).
  • Chinese Glioma Genome Atlas (CGGA): This database was used for external validation of the developed prognostic model.

Key Insights and Ideas

  • mRNAsi as a Reliable Indicator: The study confirms that the mRNAsi is a robust indicator of glioma's histological grade, IDH1 mutation status, and patient survival. This suggests that measuring stemness at the RNA level can provide valuable prognostic information.
  • CSC-Driven Prognosis: The findings strongly support the idea that CSC characteristics are intrinsically linked to glioma's aggressiveness and patient outcomes. Targeting these CSC-related pathways could be a promising therapeutic strategy.
  • Novel Biomarkers Identified: The identification of F2RL2, CLCNKA, and LOXL4 as novel prognostic biomarkers for glioma is a significant contribution. These genes, previously not linked to glioma prognosis, offer new avenues for diagnostic and therapeutic development.
  • Predictive Nomogram's Efficacy: The developed nomogram, based on 11 prognostic SRGs, demonstrated a high accuracy (AUC of 0.832) in predicting patient survival, suggesting its potential clinical utility.

Key Results

  • A total of 340 stemness-related genes (SRGs) were identified as differentially expressed across different glioma grades.
  • 11 prognostic SRGs were selected for the final multivariate Cox regression model.
  • The developed prognostic nomogram achieved an Area Under the Curve (AUC) of 0.832, indicating good predictive performance for overall survival.
  • The risk score derived from the nomogram was an independent prognostic factor for glioma patients.
  • Three novel prognostic biomarkers (F2RL2, CLCNKA, and LOXL4) were identified for glioma.

Limitations and Future Work

  • In Silico Nature: The study is primarily based on in silico analysis. While powerful, these findings would benefit from experimental validation in cell lines and animal models to confirm the functional roles of the identified biomarkers and therapeutic targets.
  • External Validation Scope: While the CGGA dataset provided validation, further validation across diverse patient cohorts and geographical locations would strengthen the generalizability of the findings.
  • Mechanism Elucidation: While the study identifies prognostic SRGs, further research is needed to fully elucidate the specific molecular mechanisms by which these genes contribute to glioma tumorigenesis, progression, and drug resistance, particularly in relation to CSCs.
  • Therapeutic Target Development: The identified novel biomarkers (F2RL2, CLCNKA, LOXL4) warrant further investigation as potential therapeutic targets. Pre-clinical and clinical studies could explore the efficacy of targeting these genes or their associated pathways.

Summary updated

Abstract

Glioma is the common histological subtype of malignancy in central nervous system, with a high morbidity and mortality. Cancer stem cells (CSCs) play an important role in regulating the tumorigenesis and progression of glioma; however, the prognostic biomarkers and therapeutic targets associated with CSC characteristics have not been fully acknowledged in glioma. In order to identify the prognostic stemness-related genes (SRGs) of glioma in silico, the RNA sequencing data of patients with glioma were retrieved from The Cancer Genome Atlas (TCGA) databases. The mRNA expression-based stemness index (mRNAsi) was significantly associated with the glioma histologic grade, isocitrate dehydrogenase 1 (IDH1) mutation and overall survival of glioma patients by the nonparametric test and Kaplan-Meier survival analysis. A total of 340 SRGs were identified as the overlapped stemness-related differential expressed genes (DEGs) of different histologic grade screened by the univariate Cox analysis. Based on 11 prognostic SRGs, the predict nomogram was constructed with the AUC of 0.832. Moreover, the risk score of the nomogram was an independent prognostic factor, indicating its significant applicability. Besides other eight reported biomarkers of glioma, we found that F2RL2, CLCNKA and LOXL4 were first identified as prognostic biomarkers for glioma. In conclusion, this bioinformatics study demonstrates the mRNAsi as a reliable index for the IDH1 mutation, histologic grade and OS of glioma patients and provides a well-applied model for predicting the OS for patients with glioma based on prognostic SRGs. Additionally, this in silico study also identifies three novel prognostic biomarkers (F2RL2, CLCNKA and LOXL4) for glioma patients.
BibTeX
@inproceedings{lvu2020i,
  author = {Wen Lvu and Fei Xu and Cheng Chen and Bo Zhang},
  title = {{<i>In silico</i> identification of the prognostic biomarkers and therapeutic targets associated with cancer stem cell characteristics of glioma}},
  booktitle = {Bioscience},
  year = {2020},
  doi = {10.1042/bsr20201037},
}
Figure