Associations between migrasome-related genes and long non-coding rnas in glioma and their prognostic relevance to the...

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.

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.

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.

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.