Novel tumor necrosis factor-related long non-coding RNAs signature for risk stratification and prognosis in glioblastoma

Shengrong Long, Bingbing Wu, Yang Liu, Lesheng Wang, Bo Wang, Yu Yan, Jiazhi Jiang, Bin Yang, Qiangqiang Zhou, Min Shi, Liang Wu, Wei Wei, Xiang LiFrontiers in Neurology 2023 · DOI

Abstract: This analysis provides a comprehensive understanding of the role of TNF-related characters, which may improve the clinical outcome of GBM patients.

Introduction

Glioblastoma (GBM) remains one of the most aggressive and difficult-to-treat brain tumors, with a five-year survival rate of less than 10%. While immunotherapy has revolutionized treatment for other cancers, it has largely failed in GBM clinical trials, often due to the "cold" or immunosuppressive nature of the tumor microenvironment (TME).

Tumor Necrosis Factor (TNF) is a double-edged sword in oncology; it can promote cell death but also drive tumor proliferation and invasion. While the role of protein-coding genes in the TNF pathway is well-studied, the role of long non-coding RNAs (lncRNAs)—transcripts that regulate gene expression without being translated into proteins—remains a "black box" in GBM. This study identifies a novel six-gene signature of lncRNAs associated with the TNF pathway to serve as prognostic biomarkers and therapeutic targets.

Methods

The researchers employed a robust bioinformatics pipeline across multiple patient cohorts, including The Cancer Genome Atlas (TCGA-GBM) as a training set and the Chinese Glioma Genome Atlas (CGGA) for validation.

Overview of the bioinformatics pipeline used to identify and validate the TNF-related lncRNA signature.

Key methodological steps included:

  1. Pathway Scoring: Using single-sample Gene Set Enrichment Analysis (ssGSEA), they calculated "TNF scores" for each patient based on hallmark gene sets.
  2. Signature Construction: They identified 138 lncRNAs significantly correlated with TNF activity. Through multivariate Cox regression, they narrowed this down to a six-lncRNA risk signature: C1RL-AS1, LINC00968, MIR155HG, CPB2-AS1, LINC00906, and WDR11-AS1.
  3. Subtyping: Using ConsensusClusterPlus, they divided patients into three molecular subtypes (C1, C2, and C3).
  4. Immune Profiling: Algorithms like ESTIMATE, CIBERSORT, and TIDE were used to quantify immune cell infiltration and predict response to immune checkpoint blockade.
  5. Validation: The expression of these six lncRNAs was validated in human glioma cell lines (U251, T98G) compared to normal astroglia using quantitative real-time PCR (qRT-PCR).

Key Results and Insights

Molecular Subtyping and the "Immune Paradox"

Using the identified signature, patients were stratified into three distinct clusters.

Consensus clustering matrix (k=3) used to define three distinct molecular subtypes (C1, C2, C3) based on the lncRNA signature.

A striking finding was that Subtype C2 exhibited the highest immune scores (highest infiltration of immune cells) but the worst prognosis. This suggests that these tumors are not "cold" but rather "exhausted." The high-risk group showed significantly higher expression of immune checkpoints like PDCD1 (PD-1), CTLA4, and LAG3, explaining why high immune infiltration correlated with poor survival—the immune cells present are functionally inactivated by the tumor.

Prognostic Power

The 6-lncRNA risk score proved to be an independent predictor of survival. In the TCGA cohort, the Area Under the Curve (AUC) for predicting 1-year survival was 0.70, and 0.60 in the CGGA validation cohort. Patients in the high-risk group had significantly shorter overall survival (p=0.013 in TCGA; p=0.0041 in CGGA).

Biological Drivers

High TNF pathway activity was tightly linked to pathways involved in tumor progression, such as hypoxia, epithelial-mesenchymal transition (EMT), and angiogenesis. Lab validation via qRT-PCR confirmed that C1RL-AS1 and LINC00968 were significantly upregulated in glioma cells, while the other four lncRNAs were downregulated, matching the computational predictions.

Conclusion and Future Directions

This study provides a novel risk stratification model that predicts survival outcomes and immune microenvironment characteristics in GBM patients. While the retrospective nature of the study and the need for deeper mechanistic discovery of lncRNA-TNF interactions are limitations, the findings offer a potential guide for personalized immunotherapy. Future work should investigate whether targeting high-risk lncRNAs like C1RL-AS1 can sensitize GBM tumors to existing treatments.

Summary updated

Abstract

This analysis provides a comprehensive understanding of the role of TNF-related characters, which may improve the clinical outcome of GBM patients.
BibTeX
@inproceedings{long2023novel,
  author = {Shengrong Long and Bingbing Wu and Yang Liu and Lesheng Wang and Bo Wang and Yu Yan and Jiazhi Jiang and Bin Yang and Qiangqiang Zhou and Min Shi and Liang Wu and Wei Wei and Xiang Li},
  title = {{Novel tumor necrosis factor-related long non-coding RNAs signature for risk stratification and prognosis in glioblastoma}},
  booktitle = {Frontiers in Neurology},
  year = {2023},
  doi = {10.3389/fneur.2023.1054686},
}
Figure