目的 基于铜死亡相关长链非编码RNA(cuproptosis-related long noncoding RNA,CRL)构建子宫颈癌预后模型并分析不同风险组间药物敏感性差异,为子宫颈癌患者预后预测及个体化治疗提供理论依据.方法 从癌症基因组图谱(The Cancer Genome Atlas,TCGA)数据库中下载304例子宫颈癌患者的基因表达谱、突变数据和临床数据,使用随机抽样的方法将患者分为训练集(n=152例)和测试集(n=152).采用Pearson相关性分析鉴定CRL.应用单因素Cox、LASSO和多因素Cox回归分析在训练集中构建CRL风险评分模型,在测试集和整个队列中进行验证,并根据风险评分中位数将训练集和测试集患者分为高风险组(训练集76例和测试集83例)和低风险组(训练集76例和测试集69例).使用Kaplan-Meier(K-M)生存分析、受试者工作特征(receiver operating characteristic,ROC)曲线、单因素与多因素Cox回归和主成分分析(principal component analysis,PCA)评估CRL风险评分模型,并构建结合临床病理特征和CRL风险评分模型的列线图和校准曲线.通过基因集富集分析(gene set enrichment analysis,GSEA)探索该模型的潜在分子机制.使用Spearman相关分析探讨免疫细胞浸润与风险评分之间的相关性.绘制子宫颈癌患者基因突变图谱,分析CRL风险评分模型与体细胞变异之间的相关性.分析免疫治疗药物的敏感性和20种化疗药物在不同风险群体中的半抑制浓度(half maximal inhibitory concentration,IC50)值差异.结果 共获得704个CRL,经单因素Cox、LASSO和多因素Cox回归分析最终构建包含6个CRL(AC103591.4、AC021851.1、MNX1-AS1、FAM27E3、AL603832.1 和 AC097505.1)的风险评分预测模型.K-M 生存曲线、ROC 曲线下面积(area under the curve,AUC)和PCA分析均验证该模型具有良好的预测能力.多因素Cox回归显示,CRL风险评分可作为独立预后因子(P＜0.05).列线图对子宫颈癌患者的1、3和5年总生存(overall survival,OS)具有较好的预测能力.GSEA结果显示,高风险组与癌症通路相关.免疫细胞浸润结果表明,多数免疫细胞与CRL风险评分呈正相关(均r＞0,均P＜0.05).免疫检查点分析结果显示,低风险组患者免疫检查点表达较高.基因突变图谱结果表明,高低风险组间肿瘤突变负荷(tumor mutation burden,TMB)比较,差异无统计学意义(P＞0.05).药物敏感性分析结果显示,免疫治疗药物细胞毒T淋巴细胞相关抗原4对低风险组患者疗效较好,阿卡地新、二甲基草酰甘氨酸、多柔比星、索拉非尼和阿糖胞苷5种药物的IC50值在高低风险组间比较,差异均具有统计学意义(均P＜0.05).结论 6个CRL的风险评分特征可独立预测子宫颈癌患者的预后,有助于阐明子宫颈癌中CRL的机制,并为患者临床个体化治疗提供理论指导.
Construction of prognostic model and drug sensitivity analysis of cervical cancer based on cuproptosis-related long noncoding RNAs
Objective To construct a prognostic model of cervical cancer based on cuproptosis-related long noncoding RNAs(CRLs)and analyze the differences in drug sensitivity among different risk groups,in order to provide theoretical basis for the prediction of prog-nosis and individualized treatment of cervical cancer.Methods The gene-expression profiles,mutation data,and clinical data of 304 cervical cancer patients were downloaded from The Cancer Genome Atlas(TCGA)database.The patients were divided into a training set(n=152)and a test set(n=152)by random sampling.Pearson correlation analysis was used to identify CRLs.Using univariate Cox,LASSO,and multivariate Cox regression analysis,a CRL risk score model was constructed in the training set,and validated in the test set and the entire queue.Based on the median risk score,the patients in the training and test sets were divided into high-risk groups(the training set:n=76,the test set:n=83)and low-risk groups(the training set:n=76,the test set:n=69).The CRL risk score model was evaluated using Kaplan-Meier(K-M)survival analysis,receiver operating characteristic(ROC)curves,univariate and multivariate Cox regression,and principal component analysis(PCA).A nomogram predicting the prognosis of cervical cancer patients using CRL risk score model com-bined with clinicopathological characteristics and its calibration curve were constructed.The potential molecular mechanisms of this model was explored through gene set enrichment analysis(GSEA).The correlation between immune cell infiltration and risk score was analyzed by Spearman correlation analysis.The gene mutation map of cervical cancer patients was used to analyze the correlation between the CRL risk score model and somatic mutations.The sensitivity of immunotherapy drugs and the difference in half maximal inhibitory concentration(IC50)values of 20 chemotherapy drugs among different risk groups were evaluated.Results A total of 704 CRLs were obtained,and a risk score prediction model consisting of 6 CRLs(AC103591.4,AC021851.1,MNX1-AS1,FAM27E3,AL603832.1,and AC097505.1)was constructed through univariate Cox,LASSO,and multivariate Cox regression analysis.The K-M survival curve,area under the ROC curve(ROC AUC),and PCA analysis all validated the model's good predictive ability.Multivariate Cox regression showed that CRL risk score served as an independent prognostic factor(P＜0.05).The nomogram predicting the 1-,3-,and 5-year OS for cervical cancer patients had good predictive ability.GSEA results showed that the high-risk group was associated with the cancer pathway.The analysis of immune cell infiltration showed that most immune cells were positively correlated with CRL risk score(all r＞0,all P＜0.05).Immune checkpoint analysis show that the expression of immune checkpoints was higher in the low-risk group.The gene mutation map showed that there was no statis-tically significant difference in tumor mutation burden(TMB)between the high-and low-risk groups(P＞0.05).The drug sensitivity analy-sis showed that the immunotherapy drug cytotoxic T lymphocyte-associated antigen-4 had a better therapeutic effect in the low-risk group.The IC50 values of five drugs,acadesine,dimethyloxallyl glycine,doxorubicin,sorafenib,and cytarabine,in the high-and low-risk groups were all significantly different(all P＜0.05).Conclusions The risk score based on the six CRLs can independently predict the prognosis of cervical cancer patients,help clarify the mechanisms of CRLs in cervical cancer,and provide theoretical guidance for clinical individual-ized treatment of patients.
万方数据
维普
