A Prognostic Model Of Cuproptosis Related Genes In Glioma Was Constructed Based On Chinese Glioma Genome Atlas

Background and Objective :Glioma,originated from neuroepithelial tissue,is the most common primary malignant tumor in the brain.It has the characteristics of rapid growth,easy invasion and easy recurrence.Existing treatments(surgical resection,radiotherapy,chemotherapy)have not significantly improved the prognosis of patients.Finding new diagnostic,predictive,and therapeutic targets for glioma has become an urgent task.With the development of science and technology,bioinformatics has been greatly developed.Studying new targets for diagnosis,prediction and treatment of glioma by bioinformatics has become one of the important means for researchers.Regulatory cell death(RCD)plays an important role in the development of glioma.Clarifying the effect of RCD in glioma is of great significance for the discovery of new targets.Cuproptosis is a newly discovered RCD,which is different from the known cell death modes(such as apoptosis,necrosis,autophagy,pyroptosis,ferroptosis).Cuproptosis is caused by the direct binding of copper to the acylated components of the tricarboxylic acid(TCA)cycle,which leads to the aggregation of lipidated proteins and the subsequent loss of iron-sulfur cluster proteins,resulting in protein water toxicity stress and ultimately cell death.The discovery of cuproptosis provides a new idea for the study of new targets of glioma.At present,there are few studies on cuproptosis in glioma.The purpose of this study was to screen cuproptosis-related genes related to the prognosis of glioma based on Chinese Glioma Genome Atlas by bioinformatics methods,and to construct a multi-gene risk scoring model.Subsequently,the risk score and clinical factors were analyzed to screen independent prognostic factors,and then a nomogram model was constructed to predict the prognosis of patients.Materials and Methods :Tsvetkov et al.’ s paper(DOI : 10.1126 / science.abf0529)was searched on the Pubmed official website.FDR<0.05,|log FC|>1 was used as a condition to screen the experimental data provided by the article to obtain copper ionophore-related cuproptosis-related genes.Gene sequencing information and clinical information in the GTEX dataset were downloaded from the UCSC Xena website(https://xena.ucsc.edu/)).The GTEX data ID was converted by the Perl programming language,and then the normal cerebral cortex sequencing data was extracted.The sequencing data and clinical information data of m RNAseq＿325 dataset and m RNAseq＿693 dataset were obtained from the CGGA database.The m RNAseq＿325 dataset information that meet the inclusion and exclusion criteria was used as the training group,and the m RNAseq＿693 dataset information was used as the verification group.The sequencing data of normal cerebral cortex,training group and validation group were converted to log2(TPM + 1).FDR<0.05,|log FC|>1 was used as the threshold to analyze the difference between the training group sequencing data and the normal cerebral cortex sequencing data,and the differential genes were obtained.The cuproptosis-related genes were intersected with the differential genes to obtain the cuproptosis-related genes expressed in the differential genes.Subsequently,univariate COX regression,LASSO COX regression,and multivariate COX regression analysis were performed to screen cuproptosis-related genes associated with prognosis.Based on multivariate COX regression analysis,a risk score model of cuproptosis-related genes related to prognosis was constructed,and the risk value of each patient was calculated.The median risk of the training group was used as the cutoff value,and the training group and the verification group were divided into high and low risk groups.KM analysis,AUC value,risk value curve,survival state scatter plot and prognosis-related cuproptosis gene expression heat map were performed according to high and low risk groups.The risk score,clinical factors and survival data were combined,and p < 0.05 was used as the screening condition.Univariate/ multivariate COX regression analysis was performed to extract independent prognostic factors.In the training group,a COX regression model was constructed for independent prognostic factors,and it was visualized as a nomogram.The C-index value was calculated,and the calibration curves of 1,3 and 5 years were drawn to evaluate the effectiveness of the model.Verify in the verification group.Results :There were 102 cuproptosis genes related to copper ionophores.Differential analysis obtained 3982 genes.A total of 14 cuproptosis-related genes(COPB2,MBTPS2,WDR12,FAM111 A,GADD45GIP1,SLC16A1,RPLP0,CDKN2 A,CDK1,C6orf136,NAPA,PITRM1,RAD9 A,COX4I1)were expressed in the differential genes.Five cuproptosis-related genes(MBTPS2,C6orf136,CDK1,SLC16A1,RAD9A)were screened by univariate COX regression,Lasso-COX regression and multivariate COX regression.The risk scoring model was constructed according to the prognostic related genes,and the high and low risk groups were grouped with the median risk score of the training group-1.453 as the cutoff value.In the training group and the verification group,the KM analysis of the high and low risk groups showed that the survival time of glioma patients in the high risk group was significantly shorter than that in the low risk group,and the p value was less than 0.001.The AUC values of 3,5 and 10 years in the training group were 0.847,0.868 and 0.894,and 0.767,0.769 and 0.731 in the validation group.The survival state diagram shows a trend of increasing the number of deaths with the increase of risk value.Five independent prognostic factors were screened : risk score,primary / recurrent / secondary type,WHO grade,pathological diagnosis of anaplastic astrocytoma(AA),and chemotherapy.The HR value of risk score,primary / recurrent / secondary type,WHO grade,pathological diagnosis of AA was greater than 1,and the HR value of chemotherapy was less than 1.Based on the constructed nomogram model,the C-index values in the training group and the verification group were 0.782 and 0.768,respectively.And the 1-year,3-year and 5-year calibration curve fitting is higher.Conclusion :1.In this study,five cuproptosis-related genes(MBTPS2,C6orf136,CDK1,SLC16A1,RAD9A)were screened to be related to the prognosis of glioma.2.This study shows that risk score,primary / recurrent / secondary type,WHO grade,pathological diagnosis of AA,and chemotherapy can be used as independent prognostic factors for patients with glioma.3.The risk scoring model and nomogram model constructed in this study have good predictive efficacy and are expected to become one of the tools for clinical prediction of glioma prognosis.