基于生物信息学筛选结肠癌失巢凋亡相关的lncRNA及预后风险模型的构建

Establishment of anoikis-related long non-coding RNA and prognostic risks models of COAD prognosis based on bioinformatics

目的 通过生物信息学方法对癌证基因图谱数据库(TCGA,https://portal.gdc.cancer.gov/)的结肠癌(colon adenocarcinoma,TCGA-COAD)数据进行挖掘与结肠癌失巢凋亡相关的lncRNA并进行效能的验证,预测潜在干预中药。方法从TCGA数据库下载COAD基因和转录组数据及临床资料,获取LncRNA表达数据,然后从GSEA数据库(http://www.gsea-msigdb.org/gsea/index.jsp)获得失巢凋亡基因与LncRNA进行共表达分析确定失巢凋亡相关的LncRNA。通过对失巢凋亡相关LncRNA的单因素Cox分析、Lasso回归分析和多因素Cox分析,最终确定用于构建风险模型的失巢凋亡相关LncRNA。根据训练组风险评分中位数将病人分为高、低风险组,再利用独立预后分析、ROC分析、列线图、C-index分析、无进展生存期分析、生存分析、PCA分析来验证模型的预测效能和区分高、低风险人群的能力。对高、低风险组进行差异基因分析和通路富集分析,对在高风险组高表达的差异基因与TCGA-COAD基因集差异分析,取得在肿瘤中高表达的基因与高风险高表达基因组取交集基因,通过Coremine数据库查找交集基因对应中药(P值小于0.05)。结果 通过单因素Cox分析获得9个与COAD患者总生存时间显著相关的失巢凋亡相关LncRNA,运用Lasso-Cox方法构建并优化模型,确定了7个LncRNA(SUCLG2-AS1、LINC02163、AC012313.5、LINC02257、AC008494.3、AP001619.1、AP003555.1)用于构建风险模型,经验证,该模型可有效区分高、低风险人群,具备良好的生存预测能力;差异基因富集于血浆血红素的清除、清道夫受体的配体的结合和摄取、CD22介导B细胞受体调控、补体的激活等免疫信号通路;10个基因(CDKN2A、KLK6、TNNT1、CST6、FGF19、H19、AL137800.1、PRSS2、AC105460.1、RNU4ATAC)在肿瘤和高风险组中均高表达,通过对Coremine Medical数据库查找结肠癌关键基因潜在干预结肠癌的中药有茶树根、薏苡仁、姜黄、黄芩、黄芪、苦参等。结论 本研究建立了一个由7个LncRNA组成的结肠癌预后模型,预测了具有干预结肠癌潜力的中药,为结肠癌患者的个体化治疗及预后判断提供借鉴价值。

Objective Colon adenocarcinoma data from The Cancer Genome Atlas(TCGA) database was mined by bioinformatics analysis methods to construct and validate a risk prediction model,predicting the relationship between the model and traditional Chinese medicine.Methods COAD gene and transcriptome data and clinical data from TCGA database were downloaded to obtain LncRNA expression,and then anoikis-related genes obtained from GSEA database were co-expressed with LncRNA for analysis to determine the anoikis-related LncRNA.Through univariate Cox analysis,Lasso regression analysis and multifactor Cox analysis of anoikis-related LncRNA,7 anoikis-related LncRNAs were finally determined to construct risk models.Patients were divided into high and low risk group according to the median risk score of the training group,and the predictive efficacy of the model and the ability to distinguish between high and low risk group were verified by independent prognosis analysis,C-index analysis,ROC analysis,survival analysis,progression-free survival analysis,PCA analysis.Differential gene analysis and pathway enrichment analysis were performed on high-risk and low-risk groups to observe which gene pathways the differential genes were enriched.Prediction of Traditional Chinese medicine were carried out through the Coremine Medical database.Results 9 anoikis-related LncRNAs significantly associated with overall survival in COAD patients were obtained by univariate Cox analysis.Using the Lasso-Cox method to construct and optimize the model,7 LncRNAs(SUCLG2-AS1、LINC02163、AC012313.5、LINC02257、AC008494.3、AP001619.1、AP003555.1) were finally determined to construct the risk model.The model could effectively distinguish between high and low risk,and had good survival prediction ability;Differential genes analysis between high risk and low risk groups,and 35 differential genes were enriched in gene pathways related to scavenging of heme from plasma、binding and uptake of ligands by scavenger receptors、CD22 mediated BCR regulation、FCGR activation、creation of C4 and C2 activators、role of LAT2 NTALLAB on calcium mobilization、initial triggering of complement、role of phospholipids in phagocytosis.7893 differential genes were screened out from gene expression of colon cancer,and we screened out that 10 genes had higher expression in cancer and high-risk group.10 genes(CDKN2A、KLK6、TNNT1、CST6、FGF19、H19、AL137800.1、PRSS2、AC105460.1、RNU4ATAC) were screened out from Coremine Medical database,and the potential effective TCM against COAD were Camellia sinensis(Chashugen)、Semen coicis(Yiyiren)、Curcuma longae Rhizoma(Jianghuang)、Radix scutellaria(Huangqin)、Astragali radix(Huangqi)、Sophora flavescens(kushen),etc.Conclusion A prognostic model for colon cancer composed of 7 LncRNAs was established,and the traditional Chinese medicine with the potential to intervene in colon cancer was predicted,which provides reference value for the individualized treatment and prognosis of colon cancer patients.