Dataset Information

An Immune-Related Gene Pairs Signature Predicts Prognosis and Immune Heterogeneity in Glioblastoma.

ABSTRACT:

Purpose

Glioblastoma is one of the most aggressive nervous system neoplasms. Immunotherapy represents a hot spot and has not been included in standard treatments of glioblastoma. So in this study, we aim to filtrate an immune-related gene pairs (IRGPs) signature for predicting survival and immune heterogeneity.

Methods

We used gene expression profiles and clinical information of glioblastoma patients in the TCGA and CGGA datasets, dividing into discovery and validation cohorts. IRGPs significantly correlative with prognosis were selected to conduct an IRGPs signature. Low and high risk groups were separated by this IRGPs signature. Univariate and multivariate cox analysis were adopted to check whether risk can be a independent prognostic factor. Immune heterogeneity between different risk groups was analyzed via immune infiltration and gene set enrichment analysis (GSEA). Some different expressed genes between groups were selected to determine their relationship with immune cells and immune checkpoints.

Results

We found an IRGPs signature consisting of 5 IRGPs. Different risk based on IRGPs signature is a independent prognostic factor both in the discovery and validation cohorts. High risk group has some immune positive cells and more immune repressive cells than low risk group by means of immune infiltration. We discovered some pathways are more active in the high risk group, leading to immune suppression, drug resistance and tumor evasion. In two specific signaling, some genes are over expressed in high risk group and positive related to immune repressive cells and immune checkpoints, which indicate aggression and immunotherapy resistance.

Conclusion

We identified a robust IRGPs signature to predict prognosis and immune heterogeneity in glioblastoma patients. Some potential targets and pathways need to be further researched to make different patients benefit from personalized immunotherapy.

SUBMITTER: Zhang N

PROVIDER: S-EPMC8076680 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:BackgroundLung adenocarcinoma (LUAD) is the most common form of lung cancer, accounting for 30% of all cases and 40% of all non-small-cell lung cancer cases. Immune-related genes play a significant role in predicting the overall survival and monitoring the status of the cancer immune microenvironment. The present study was aimed at finding an immune-related gene signature for predicting LUAD patient outcomes.MethodsFirst, we chose the TCGA-LUAD project in the TCGA database as the training cohort for model training. For model validating, we found the datasets of GSE72094 and GSE68465 in the GEO database and took them as the candidate cohorts. We obtained 1793 immune-related genes from the ImmPort database and put them into a univariate Cox proportional hazard model to initially look for the genes with potential prognostic ability using the data of the training cohort. These identified genes then entered into a random survival forests-variable hunting algorithm for the best combination of genes for prognosis. In addition, the LASSO Cox regression model tested whether the gene combination can be further shrinkage, thereby constructing a gene signature. The Kaplan-Meier, Cox model, and ROC curve were deployed to examine the gene signature's prognosis in both cohorts. We conducted GSEA analysis to study further the mechanisms and pathways that involved the gene signature. Finally, we performed integrating analyses about the 22 TICs, fully interpreted the relationship between our signature and each TIC, and highlighted some TICs playing vital roles in the signature's prognostic ability.ResultsA nine-gene signature was produced from the data of the training cohort. The Kaplan-Meier estimator, Cox proportional hazard model, and ROC curve confirmed the independence and predictive ability of the signature, using the data from the validation cohort. The GSEA analysis results illustrated the gene signature's mechanism and emphasized the importance of immune-related pathways for the gene signature. 22 TICs immune infiltration analysis revealed resting mast cells' key roles in contributing to gene signature's prognostic ability.ConclusionsThis study discovered a novel immune-related nine-gene signature (BTK, CCR6, S100A10, SEMA3C, GPI, SCG2, TNFRSF11A, CCL20, and DKK1) that predicts LUAD prognosis precisely and associates with resting mast cells strongly.