Multi-omics investigation of the GNG12-AS1/miR-138-5p/PDPK1 candidate regulatory relationship and transcriptome-derived prognostic modeling in glioma
Objective Glioma is characterized by molecular heterogeneity and variable clinical behavior. We evaluated a computationally nominated GNG12-AS1/miR-138-5p/PDPK1 relationship, its cellular and pathway context, and an analytically separate transcriptome-derived prognostic model. Methods Public glioma transcriptomic and clinical resources were integrated using Spearman correlation with Benjamini-Hochberg False Discovery Rate (FDR) control, single-cell RNA sequencing, pathway enrichment, ligand-receptor, immune and GDSC2-based ridge analyses, and a Cox-LASSO-Cox modeling workflow. RT-qPCR and flow cytometry in U251 cells assessed perturbation-associated RNA expression, PDPK1, phospho-AKT Thr308, total AKT and apoptosis. The fixed prognostic score was evaluated in CGGA_mRNAseq_325; incremental value beyond clinical variables was assessed in the pooled CGGA_mRNAseq_693 and CGGA_mRNAseq_325 complete-case dataset. Results Pairwise Spearman analyses showed a weak positive association between GNG12-AS1 and miR-138-5p (rho = 0.084, FDR = 0.042), an inverse association between miR-138-5p and PDPK1 (rho = −0.217, FDR < 0.001) and a weak positive association between GNG12-AS1 and PDPK1 that did not meet the FDR threshold (rho = 0.064, FDR = 0.086). RT-qPCR and flow cytometry confirmed perturbation efficiency, reduced PDPK1 and phospho-AKT Thr308, comparatively stable total AKT and increased Annexin-positive fractions. The 12-feature model had C-indices of 0.772 and 0.744 in the training and validation cohorts. In the pooled secondary clinical analysis (715 complete cases; 480 events), the risk score was associated with overall survival (HR per 1-SD increase = 2.085, 95% CI 1.802–2.413) and increased the C-index by 0.030 (bootstrap 95% CI 0.016–0.045). Forty-two drug labels had concordant lower predicted GDSC2 log-response values across both CGGA cohorts; AZD3759 was the principal computational candidate. Conclusion The computational and functional results provide directionally concordant support for a candidate GNG12-AS1/miR-138-5p/PDPK1 regulatory relationship but do not establish direct RNA binding or a serial ceRNA mechanism. The separate prognostic model retained discrimination in the validation cohort and showed modest incremental value in the pooled secondary analysis. Drug-response findings remain computational priorities rather than measured patient IC50 values or treatment recommendations.
Authors
- Lingtong You (ORCID: https://orcid.org/0000-0002-1399-6981)
- Jinning Zhang (ORCID: https://orcid.org/0000-0003-2861-041X)
- Huangqiang Yang
- Jiaxin Huang
- Junhuai Wei
- Zhifeng Yu
Institutions
- Fujian Medical University (CN)
Publication Details
- Journal
- Frontiers in Neurology
- Published
- 2026-11-18
- DOI
- https://doi.org/10.3389/fneur.2026.1939176
- Primary Topic
- Cancer-related molecular mechanisms research
- Type
- article
- Field-Weighted Citation Impact
- 0.00