基于m^(6)A相关lncRNA的女性肺腺癌预后风险模型

A prognostic risk model for female lung adenocarcinomas based on m^(6)A-related lncRNAs

目的建立m^(6)A相关长链非编码RNA(lncRNA)女性肺腺癌患者的预后风险模型,并分析有代表性的m^(6)A相关lncRNA在女性肺腺癌患者治疗和预后中的作用。方法从TCGA数据库下载女性肺腺癌患者的RNA表达数据(291例肿瘤组织和34例正常组织)和临床数据(280例)。采用limma包分析23个m^(6)A甲基化修饰基因的表达情况。根据TCGA上关于基因集合ID的说明,检索并收集女性肺腺癌lncRNA表达的数据。从GENCODE下载m^(6)A调节基因的转录组学数据。采用共表达相关性分析筛选m^(6)A相关lncRNA。采用limma包对筛选的m^(6)A相关lncRNA进行差异分析。采用survival、Cox回归分析包进行Kaplan-Meier、单因素和多因素Cox回归分析、受试者操作特征曲线[曲线下面积(AUC)>0.65]分析,进一步筛选预后lncRNA。选取AUC最大的lncRNA(AP001981.2)作为代表性的m^(6)A相关lncRNA,用于确认预测特征。采用rms包,通过将AP001981.2表达水平与年龄和临床分期相结合建立列线图。采用校准曲线检验预测与现实之间的一致性。按照AP001981.2的logFC进行分组,logFC>0为高表达组(131例),logFC<0为低表达组(131例)。采用oncoPredict包进行2组患者的药物敏感性分析。通过enrichplot包行基因集富集分析,探讨AP001981.2的参与通路。结果23个m^(6)A调节器分子中有15个基因差异表达,其中有6个基因表达下调,9个基因表达上调。共识别了16876个lncRNA,并鉴定出了2910个m^(6)A相关lncRNA,其中有1156个差异表达m^(6)A相关lncRNA。经Kaplan-Meier、单因素Cox回归分析筛选,共有17个与女性肺腺癌患者预后相关的lncRNA。多因素Cox分析显示有10个lncRNA可作为独立预后因子,其中9个lncRNA可作为独立危险因子(AL157931.1、AL359853.1、AC007128.1、AC005618.1、LINC01138、AC103591.4、AP001981.2、AC243772.2和AC078983.1),1个lncRNA可作为独立保护因子(AC018529.1)。选择AUC最大(0.736)的AP001981.2作为代表性的m^(6)A相关lncRNA。构建了预测女性肺腺癌患者1、3、5年生存率的列线图,校准图显示列线图预测与实际观察之间的良好一致性。低表达组和高表达组患者对11种药物(UMI-77、BMS-754807、AZD8055、Sabutoclax、Entinostat、PRT062607、ABT737、WEHI-539、Doramapimod、Elephantin、GSK269962A)的敏感性比较,差异均有统计学意义(均P<0.05)。AP001981.2参与α-亚麻酸代谢、致心律失常性右心室心肌病、胰岛素信号通路、哺乳动物雷帕霉素靶蛋白信号通路和血管平滑肌收缩相关代谢通路。结论构建了一个m^(6)A相关lncRNA女性肺腺癌预后风险模型。AP001981.2作为代表性的m^(6)A相关lncRNA参与肿瘤相关通路,是潜在的治疗靶点,也是女性肺腺癌预后的独立危险因子。

Objective To establish a prognostic risk model for female lung adenocarcinomas based on m^(6)A-related long non-coding RNAs(lncRNAs),and to analyze the therapeutic and prognostic potentials of representative m^(6)A-related lncRNAs in female lung adenocarcinoma patients.Methods A dataset containing RNA expressions of 291 female lung adenocarcinoma specimens and 34 normal tissues as well as clinical data of 280 female lung adenocarcinoma patients was downloaded from The Cancer Genome Atlas(TCGA)database database.Using limma package in R,expressions of 23 genes with m^(6)A methylation modification were analyzed.LncRNA expressions in female lung adenocarcinoma specimens were collected based on the description of gene set IDs on TCGA.Transcriptomic data of m^(6)A regulatory genes were downloaded from GENCODE.Co-expression correlation analysis was performed to screen m^(6)A-related lncRNAs in female lung adenocarcinomas,and their differential expressions were analyzed using the limma package.Kaplan-Meier,univariate and multivariate Cox regression analysis,and receiver operating characteristic(ROC)curve(area under the curve[AUC]>0.65)analysis were performed using survival and Cox regression packages in R to further screen prognostic m^(6)A-related lncRNA in female lung adenocarcinomas.LncRNA AP001981.2 with the largest AUC was selected as the representative m^(6)A-related lncRNA in female adenocarcinomas and subjected to the validation of the prognostic potential.Using the RMS package,a nomogram was established by combining the representative m^(6)A-related lncRNA with age and tumor staging of female lung adenocarcinomas.Calibration curves were plotted to verify the consistency between predictions and actual conditions.According to the log fold change(FC)of lncRNA AP001981.2,female lung adenocarcinoma patients were divided into high-level group(logFC>0,n=131)and low-level group(logFC<0,n=131).Using oncoPredict package,drug sensitivity analysis was performed.Using the enrichplot package,the gene set enrichment analysis(GSEA)was performed to screen signaling pathways enriched in lncRNA AP001981.2.Results Among the 23 m^(6)A regulator molecules,15 genes were differentially expressed,including 6 downregulated and 9 upregulated genes.A total of 16,876 lncRNAs and 2,910 m6A-related lncRNAs in female lung adenocarcinomas were identified,including 1,156 differentially expressed m^(6)A-related lncRNAs.Kaplan-Meier and univariate Cox regression analysis identified 17 m6A-related lncRNAs with the prognostic potential in female lung adenocarcinoma patients.Multivariate Cox analysis revealed 10 m^(6)A-related lncRNAs serving as independent prognostic factors,of which 9 lncRNAs were independent risk factors(AL157931.1,AL359853.1,AC007128.1,AC005618.1,LINC01138,AC103591.4,AP001981.2,AC243772.2,and AC078983.1),and 1 was an independent protective factor(AC018529.1).LncRNA AP001981.2 had the largest AUC of 0.736,and adopted as the representative m^(6)A-related lncRNA in female lung adenocarcinomas.Nomograms were constructed to predict the 1-year,3-year,and 5-year survival of female lung adenocarcinoma patients,and the calibration curve showed a good consistency between the predicted and actual observations.There were significant differences in the sensitivity to 11 drugs(UMI-77,BMS-754807,AZD8055,Sabutoclax,Entinostat,PRT062607,ABT737,WEHI-539,Doramapimod,Elephantin and GSK269962A)between high-level group and low-level group(all P<0.05).LncRNA AP001981.2 was involved in theα-Linolenic acid metabolism,arrhythmogenic right ventricular cardiomyopathy,insulin signaling pathway,mammalian rapamycin target protein signaling pathway,and vascular smooth muscle contraction related metabolic pathway.Conclusions A prognostic risk model for female lung adenocarcinomas with m^(6)A-related lncRNAs is constructed.LncRNA AP001981.2 is a representative m^(6)A-related lncRNA involved in lung cancer related pathways,serving as a potential therapeutic target and an independent risk factor for the prognosis of female lung adenocarcinomas.