Gliomas are the most common intracranial tumors with the highest degree of malignancy. Disturbed cholesterol metabolism is one of the key features of many malignant tumors, including gliomas. This study aimed to investigate the significance of cholesterol metabolism-related genes in prognostic prediction and in guiding individualized treatment of patients with gliomas. Transcriptional data and clinicopathological data were obtained from The Cancer Genome Atlas (TCGA) and Chinese Glioma Genome Atlas (CGGA) databases. Intraoperative glioma samples retained in our unit and the corresponding clinicopathological information were also collected with the patients' knowledge. Firstly, cholesterol metabolism-related gene signatures (CMRGS) were identified and constructed based on difference analysis, least absolute shrinkage and selection operator (LASSO) regression analysis, and univariate/multivariate COX analysis. Then, the role of CMRGS in predicting the prognosis of gliomas and distinguishing immune landscapes was evaluated by using nomograms, survival analysis, enrichment analysis, and immune-infiltration analysis. Finally, the drug sensitivity of gliomas in different risk groups was evaluated using the oncoPredict algorithm, and potentially sensitive chemotherapeutic and molecular-targeted drugs were identified. The prognostic CMRGS contained seven genes: APOE, SCD, CXCL16, FABP5, S100A11, TNFRSF12A, and ELOVL2. Patients were divided into high- and low-risk groups based on the median cholesterol metabolic index (CMI). There were significant differences in clinicopathological characteristics and overall survival between groups. COX analysis suggested that CMRGS was an independent risk factor for glioma prognosis and had a better predictive performance than several classical indicators. In addition, GSEA, immune infiltration analysis showed that CMRGS could differentiate the immune landscapes of patients in groups. The reliability of CMRGS was validated in the CGGA cohort and our Gusu cohort. Finally, 14 drugs sensitive to high-risk patients and 16 drugs sensitive to low-risk patients were identified. The CMRGS reliably predicts glioma prognosis in multiple cohorts and may be useful in guiding individualized treatment.

Glioblastoma is the most common type of glioma with a high incidence and poor prognosis, and effective medical treatment remains challenging. Pseudouridine (Ψ) is the first post-transcriptional modification discovered and one of the most abundant modifications to RNA. However, the prognostic value of Ψ-related lncRNAs (ΨrLs) for glioma patients has never been systematically evaluated. This study aims to construct a risk model based on ΨrLs signature and to validate the predictive efficiency of the model. Transcriptomic data, genomic data, and relevant clinical data of glioma patients were extracted from the Cancer Genome Atlas (TCGA) and the Chinese Glioma Genome Atlas (CGGA). ΨrLs with significant correlation with Ψ-related genes were identified, and univariate Cox regression, least absolute shrinkage and selection operator (LASSO) regression, and multivariate Cox regression were used to further select biomarkers and construct a ΨrLs signature risk model. Then, the expression of lncRNAs of ΨrLs signature in multiple glioma cell lines was detected by qPCR. Further, ROC analysis, stratification analysis, correlation analysis, survival analysis, nomogram, enrichment analysis, immune infiltration analysis, chemoradiotherapy sensitivity analysis, somatic mutation, and recurrent copy number variation (CNV) analysis were used to validate the predictive efficiency of ΨrLs signature in TCGA and CGGA datasets. A four-lncRNA ΨrLs signature (DNAJC27-AS1, GDNF-AS1, ZBTB20-AS4, and DNMBP-AS1) risk model was constructed. By ROC analysis, stratified analysis, correlation analysis, survival analysis, and nomogram, the signature showed satisfactory predictive efficiency. Functional enrichment analysis revealed the differences in immune-related biological processes between high- and low-risk groups. Immune infiltration analysis showed that the high-risk group had lower tumor purity and higher stromal, immune and ESTIMATE scores. Mitoxantrone was identified as effective drug for low-risk group of glioma patients. Key genes in glioma development, including IDH1, EGFR, PTEN, etc., were differentially mutated between risk groups. The main recurrent CNVs in low-risk groups were 19q13.42 deletion and 7q34 amplification; 10q23.31 deletion and 12q14.1 in the high-risk group. Our study identified a four-lncRNA ΨrLs signature that effectively predicts the prognosis of glioma patients and may serve as a diagnostic tool. Risk scores of glioma patients generated by the signature is associated with immune-related biological processes and chemoradiotherapy sensitivity. These findings may inform the development of more targeted and effective therapies for glioma patients.

Glioblastoma (GBM) is the most aggressive and malignant type of primary brain tumor, with a median survival time of less than two years and a uniformly poor prognosis, despite multimodal therapeutic approaches, which highlights an urgent need for novel therapeutic targets. In this study, by integrative multi-omics analysis from CPTAC database, DepMap database and seven independent GBM cohorts, four hub genes (CD44, SURF4, IGSF3 and RALGAPA1) were identified as essential genes regulated by cancer driver genes with robust prognostic value. GBM multi-omics data from public and in-house cohorts validated that CD44 and SURF4 might be synthetic lethal partners of loss-of-function tumor suppressor genes. Analysis for immune-related pathway activity revealed complex regulation relationships of the four hub genes in tumor microenvironment (TME). Further investigation on SURF4 in pathway activity, immune therapy response and drug sensitivity proposed that SURF4 emerged as a promising therapeutic target for GBM, even for pan-cancer. Pan-cancer multi-omics exploration suggested that RALGAPA1 may be a tumor suppressor gene. By screening the first-generation and second-generation DepMap database, four genes (CCDC106, GAL3ST1, GDI2 and HSF1) might be considered as synthetic targets after mutation of RALGAPA1 as a tumor suppressor gene.