Associations between migrasome-related genes and long non-coding rnas in glioma and their prognostic relevance to the tumor microenvironment.
The study found that an eight-lncRNA risk score is associated with overall survival in glioma, but full validation of the model is needed before any clinical implications can be drawn.
Where it sits
this study against the rest of the ara 290 corpusSummary and findings
This study evaluated associations between migrasome-related genes and long non-coding RNAs (lncRNAs) in glioma, focusing on their prognostic relevance. A total of 674 glioma cases were analyzed, leading to the construction of an eight-lncRNA risk model. The study found significant associations with overall survival, but full external validation of the model was not feasible.
Abstract
<h4>Objective</h4>This exploratory study evaluated associations between migrasome-related genes and long non-coding RNAs (lncRNAs) in glioma and examined whether migrasome-related lncRNA patterns were associated with prognosis and the tumor microenvironment.<h4>Methods</h4>Transcriptomic, clinical, survival, and mutation data were obtained from the TCGA GDC portal. Migrasome-related genes were identified from published literature and GeneCards, and co-expressed lncRNAs were screened using |cor| > 0.4 and P < 0.05. A total of 674 glioma cases with complete survival information were randomly divided into training and internal testing cohorts (1:1; n = 337 each). Univariate Cox regression, LASSO Cox regression with tenfold cross-validation, and multivariate Cox regression were used to construct an eight-lncRNA risk model. External cohorts (CGGA325, CGGA693, and GSE16011) were screened for model-gene coverage. Because these platforms covered only part of the eight-lncRNA signature, we performed available-gene external survival analyses and available-gene partial-score analyses after within-cohort z-score normalization. Immune infiltration, TME scores, TMB, TIDE scores, and predicted drug sensitivity were analyzed as exploratory correlates.<h4>Results</h4>The eight-lncRNA risk score was associated with overall survival in the training, internal testing, and overall TCGA cohorts (P < 0.001). In the overall TCGA cohort, the risk score showed an AUC of 0.880. Full external validation of the eight-lncRNA signature was not feasible because CGGA325 covered three model lncRNAs (CRNDE, AC007879.2, and LINC00092), CGGA693 covered two (CRNDE and LINC00092), and GSE16011 covered only CRNDE. Available-gene partial scores remained associated with overall survival in CGGA325 (n = 313, HR = 1.54, 95% CI: 1.37-1.74, P = 4.78 ×10^-13), CGGA693 (n = 657, HR = 1.34, 95% CI: 1.27-1.42, P = 8.98 ×10^-25), and GSE16011 (CRNDE only; n = 240, HR = 1.43, 95% CI: 1.25-1.64, P = 3.85 ×10^-7). High-risk tumors showed higher stromal, immune, and ESTIMATE scores and increased macrophage infiltration, including M2 macrophages. Higher TMB was associated with worse survival, and predicted drug-response analyses suggested candidate compounds for further validation.<h4>Conclusion</h4>Migrasome-related lncRNA expression patterns were associated with glioma prognosis and immune microenvironment features in TCGA data. External cohorts provided partial support for the prognostic relevance of available model lncRNAs, particularly CRNDE, but did not permit full eight-lncRNA signature validation because of incomplete gene coverage. These findings remain hypothesis-generating and require complete external cohort validation and experimental studies before mechanistic or clinical application can be inferred.
Background
This exploratory study addresses the relationship between migrasome-related genes and long non-coding RNAs (lncRNAs) in glioma, a type of brain tumor. Previous research has indicated that lncRNAs may play a role in tumor biology and prognosis. Understanding these associations could provide insights into the tumor microenvironment and potential prognostic markers.
Methods
The study utilized transcriptomic, clinical, survival, and mutation data from the TCGA GDC portal, analyzing 674 glioma cases divided into training and internal testing cohorts (n=337 each). The researchers employed univariate and multivariate Cox regression, as well as LASSO Cox regression with tenfold cross-validation, to develop an eight-lncRNA risk model.
Results
The primary endpoint revealed that the eight-lncRNA risk score was significantly associated with overall survival across the training, internal testing, and overall TCGA cohorts (P < 0.001). The risk score demonstrated an AUC of 0.880 in the overall TCGA cohort. External validation was limited, with only partial coverage of the model lncRNAs in the external cohorts.
Interpretation
The findings suggest a potential link between migrasome-related lncRNA patterns and glioma prognosis, aligning with previous literature that highlights the role of lncRNAs in cancer. However, the effect sizes, while statistically significant, may not be clinically meaningful without full validation. Limitations include the reliance on incomplete external validation and the potential for confounding factors.
Key findings
- In the overall TCGA cohort, the risk score showed an AUC of 0.880.
- In CGGA325 (n=313), HR = 1.54, 95% CI: 1.37-1.74, P = 4.78 × 10^-13.
- In CGGA693 (n=657), HR = 1.34, 95% CI: 1.27-1.42, P = 8.98 × 10^-25.
- In GSE16011 (n=240), HR = 1.43, 95% CI: 1.25-1.64, P = 3.85 × 10^-7.
Limitations
- Incomplete external validation of the eight-lncRNA signature.
- Reliance on retrospective data from TCGA.
- Potential confounding factors not fully addressed.
- Short follow-up duration for some cohorts.