A Comprehensive Analysis Of The Prognosis And Immune Microenvironment Of Patients With Glioma By A Risk Model Constructed From 30 Histone-related Genes

Objective Gliomas,the most prevalent tumours of the central nervous system,are characterised by a poor prognosis and limited therapeutic choices.This calls for the search for new molecular targets for glioma to improve the prognosis of glioma patients.Oncohistones and histone modifications are closely related to the development of glioma,and the regulation of histone biological functions cannot be achieved without histone-related genes.Therefore,we attempted to construct a risk model based on histone-related genes to predict the prognosis of glioma patients using some public database data and relevant bioinformatics analysis methods and to analyse the immune microenvironment profile of the high and low-risk groups.Methods We collected histone-related genes from the GO database.In the TCGA dataset,we scored the effect of these genes on the degree of infiltration of different types of immune cells using seven commonly used immune scoring tools.Eligible histone-related genes were obtained by restricting gene screening conditions and LASSO regression analysis.Based on the screened genes,a risk model was constructed.We used the TCGA data as the training set,scored all samples for risk,and divided patients into the high and low-risk groups using the middle value of the risk score as the cut-off point.We validated the differentiation between high and low-risk groups using CGGA693,CGGA325 and GSE16011 as external validation sets.Survival analysis and ROC curves in the training and validation sets were used to verify the prediction capacity of the risk model for the prognosis of glioma patients.Univariate and multifactorial COX analysis methods were used to determine whether they were independent prognostic factors for glioma patients.The predictive power of the risk score for the prognosis of glioma patients was compared to clinical factors such as age,gender and tumour grade.By establishing a weighted gene co-expression analysis network,GO,and KEGG analyses were performed on the modules with the highest correlation to the risk score to compare the differences in molecular processes between the high and low-risk groups.Next,Seven regularly employed analytical approaches were utilised to examine changes in the immune microenvironment and the link between risk ratings and different types of immune cells in the high and low-risk groups,and the relationship between risk scores and tumour stemness scores was explored.We then analysed the sensitivity of patients in the high and low-risk groups to commonly used chemotherapeutic agents.Using consistent clustering of the TCGA dataset,we concluded that k=2 was the ideal number of clusters and categorised the glioma samples in the TCGA dataset as C1 and C2.Immunological differences between the two subtypes and survival analysis were assessed,and the relationship between the two subtypes and the high and low-risk groups was explored.Result We collected 953 histone-related genes from the GO database and further obtained 30 eligible histone-related genes by restriction gene screening conditions and LASSO regression analysis.We successfully constructed a risk model using the 30histone-related genes.In the training set,all samples were assigned risk scores,and patients were separated into high and low-risk groups based on the median risk score value.The analysis of the training set and the external validation set revealed excellent discriminating between the expression levels of these 30 genes in the high and low-risk groups.Survival analysis and ROC curves revealed the risk model’s capacity to predict prognosis,and the validation set’s findings confirmed the risk model’s precision.We also evaluated the risk model about clinical factors and demonstrated that risk score was an independent prognostic factor for glioma patients by univariate and multifactorial COX analyses.Using GO and KEGG analysis,we developed a weighted gene co-expression network to identify potential variations in biological processes between high-and low-risk groups.We hypothesised that histone-related genes may influence the prognosis of glioma patients by influencing the quantity and polarisation status of TAMs in the glioma microenvironment.We also analysed the sensitivity of different patients’ commonly used chemotherapeutic agents in the high and low-risk groups,and the results suggested significant discrimination.Consistent clustering identified two subtypes of glioma patients with different prognostic and clinical characteristics.Principal component analysis(PCA)and t-SNE results showed a high degree of overlap between the C1 and C2 subtypes and the high and low-risk groups in the same dimension,which also suggested a significant association between tumour subtype groupings and the high and low-risk groups.Conclusion We report a risk model based on 30 histone-related genes to predict the prognosis of glioma patients,validate the significant differences in prognosis between high-risk and low-risk glioma patients based on the risk score,and identify 2 subgroups of glioma patients with different prognostic and clinical characteristics.Combined with the analysis of the differences in the immune microenvironment between the high and low-risk groups,we reasonably conclude that histone-related genes may affect the prognosis of glioma patients by influencing the number and polarisation status of TAMs in the glioma microenvironment,which provides an idea and a partial theoretical basis for subsequent studies on histone and GAMs.This study may be of value to clinicians in analysing the prognosis of glioma patients,personalising treatment and finding effective gene targets.