基于双硫死亡相关的lncRNAs的胶质母细胞瘤预后模型的构建和验证

Construction and validation of a glioblastoma prognostic model based on disulfidptosis-related lncRNAs

目的:胶质母细胞瘤(glioblastoma,GBM)是最常见的原发性中枢神经系统恶性肿瘤,预后差。双硫死亡是一种新发现的调节性细胞死亡形式,其失调与癌症进展及治疗相关。目前,关于双硫死亡相关的长链非编码RNAs(disulfidptosis-related long non-coding RNAs,DRLs)在GBM中的研究还很有限。本研究旨在探讨DRLs与GBM患者预后的关系,构建并验证预后模型。方法:从癌症基因组图谱网站下载153例GBM患者样本的RNA测序数据和相应的临床特征,并从既往研究中获得10个双硫死亡相关的基因。将153例GBM患者样本按照1?1的比例随机分为训练集和验证集,前者用于筛选与GBM患者预后相关的DRLs和构建预后模型,后者和全部数据集的数据用于检验预后模型的准确性。根据单因素Cox回归分析、LASSO回归分析和多因素Cox回归分析,筛选出与GBM预后相关的DRLs并构建最佳预后模型。采用Kaplan-Meier生存分析、受试者操作特征(receiver operator characteristic,ROC)曲线分析和C指数分析评估该模型的预测性能。根据风险评分中位数分别将训练集、验证集和全部数据集的GBM患者分为高、低风险组,进一步对每个数据集中不同风险组之间的基因进行基因集变异分析、基因集富集分析、免疫微环境特征和免疫检查点基因分析、抗肿瘤药物敏感性预测分析。结果:构建并验证了一个包含9个DRLs的预后模型。训练集、验证集和全部数据集中风险评分与患者预后生存状态的关系散点图显示风险评分越高,病死患者越多,符合模型的预测;Kaplan-Meier生存分析结果表明3个数据集中高风险组GBM患者的总生存期均低于低风险组;ROC曲线分析结果表明训练集1、2和3年的曲线下面积(area under the curve,AUC)均保持在0.75以上,而验证集和全部数据集1、2和3年的AUC均分别保持在0.60和0.70以上,表明该模型预测性能良好;与年龄和性别相比,该风险模型更能预测GBM患者生存时间。基因分析结果表明预后较差的高风险组富集了免疫相关的通路,具有较高的免疫得分、免疫相关功能激活、免疫细胞浸润和较高的免疫检查点基因表达,且表现出对多种抗肿瘤药物敏感。结论:基于9个DRLs构建的预后模型对GBM患者的预后评估具有潜在的临床应用价值,有助于评估不同风险组GBM患者的基因差异、免疫浸润情况和抗肿瘤药物敏感性,为实现GBM的个性化治疗提供依据。

Objective:Glioblastoma(GBM)is the most common malignant tumor of the central nervous system characterized by poor prognosis.Disulfidptosis,a newly identified form of regulatory cell death,whose dysregulation is also causally linked to cancer progression and treatment response.However,current studies screening disulfidptosis-related long noncoding RNAs(DRLs)in GBM are limited.This study aims to investigate the relationship between DRLs and the prognosis of GBM patients and construct and verify a prognosis model.Methods:RNA sequencing data and corresponding clinical features of 153 GBM patient samples were downloaded from The Cancer Genome Atlas website,and 10 genes associated with disulfidptosis were obtained from previous studies.The 153 GBM patient samples were randomly divided into a training set and a validation set in a 1꞉1 ratio.The former was used to screen DRLs related to the prognosis of GBM patients and construct a prognostic model,while the latter and the overall dataset were used to test the accuracy of the prognostic model.Multivariate Cox regression analysis,LASSO regression analysis,and univariate Cox regression analysis were used to screen DRLs related to the prognosis of GBM and construct the optimal prognosis model.Kaplan-Meier survival analysis,receiver operator characteristic(ROC)curve analysis,and C-index analysis were used to evaluate the prognostic ability of the model.The GBM patients in the training set,validation set,and overall dataset were divided into a high-risk group and a low-risk group according to the median risk score.Further analysis included gene set variation analysis,gene set enrichment analysis,immune microenvironment characteristics,immune checkpoint gene analysis,and prediction analysis of sensitivity to anti-tumor drugs for different risk groups within each dataset.Results:A prognostic model containing 9 DRLs was constructed and verified.Scatterplots showing the relationship between the risk scores and the survival status of the patients in the training set,validation set and overall dataset indicated that higher risk scores were associated with more deceased patients,confirming the predictive ability of the model.Kaplan-Meier survival analysis results showed that the overall survival of GBM patients in the high-risk group was lower than that of the low-risk group.ROC curve analysis results demonstrated that the area under the curve(AUC)for 1,2,and 3 years in the training set were above 0.75,while those in the validation set and the overall dataset were above 0.60 and 0.70,respectively,indicating good prediction performance of the model.Compared with age and gender,the risk model was more effective in predicting the survival time of GBM patients.Gene analysis results showed that the high-risk group with poorer prognosis was enriched in immune-related pathways,exhibiting higher immune scores,activation of immune-related functions,immune cell infiltration,and higher expression of immune checkpoint genes.Additionally,the high-risk group showed sensitivity to multiple antitumor drugs.Conclusion:The prognostic model based on 9 DRLs has potential clinical application value for the prognosis assessment of GBM patients.It helps evaluate genetic differences,immune infiltration,and sensitivity to anti-tumor drugs in different risk groups of GBM patients,providing a basis for personalized treatment for GBM.