基于TARGET和GEO数据库构建神经母细胞瘤DNA损伤修复相关多基因预后模型  

The Construction of a Polygenic Prognostic Risk Model Related to DNA DamageRepair in Neuroblastoma Based on TARGET and GEO Databases

作　　者：杨文发 杨业然 蒋持怡 初平 杨慧 任慧敏 于永波 郭永丽[1] YANG Wenfa;YANG Yeran;JIANG Chiyi;CHU Ping;YANG Hui;REN Huimin;YU Yongbo;GUO Yongli(Beijing Key Laboratory for Pediatric Diseases of Otolaryngology,Head and Neck Surgery,Beijing Pediatric Research Institute,Beijing Children’s Hospital,Capital Medical University,National Center for Children’s Health(NCCH),Beijing 100045,China;Biobank for Clinical Data and Samples in Pediatrics,Beijing Children’s Hospital,Capital Medical University,National Center for Children’s Health(NCCH),Beijing 100045,China)

机构地区：[1]国家儿童医学中心首都医科大学首都医科大学附属北京儿童医院北京市儿科研究所儿童耳鼻咽喉头颈外科疾病北京市重点实验室,北京100045 [2]国家儿童医学中心首都医科大学首都医科大学附属北京儿童医院儿科临床数据和样本生物库,北京100045

出　　处：《标记免疫分析与临床》2023年第10期1734-1743,共10页Labeled Immunoassays and Clinical Medicine

基　　金：北京市研究型病房建设示范单位项目(编号:BCRW202101)。

摘　　要：目的 利用生物信息学方法挖掘并筛选与神经母细胞瘤(neuroblastoma, NB)生存预后相关的DNA损伤修复相关基因,构建NB生存的多基因预后模型。方法 从GSEA(gene set enrichment analysis)数据库中下载DNA损伤修复相关基因。TARGET(therapeutically applicable research to generate effective treatments)数据集作为训练集,结合基因表达谱数据和临床信息,运用最小绝对值收敛和选择算子(least absolute shrinkage and selection operator, LASSO)回归分析和多因素Cox回归分析构建NB的DNA损伤修复相关的多基因预后模型。GEO(gene expression omnibus)数据集作为验证集,评估预后模型的准确性。使用受试者工作特征曲线(receiver operating cure, ROC)计算预后模型曲线下面积(area under the curve, AUC)。使用Kaplan-Meier生存曲线进行生存分析。通过单因素Cox回归和多因素Cox回归分析评估预后模型的预测性能。利用公共数据库数据进行Kaplan-Meier生存曲线分析,验证COA8对NB预后的影响。CCK8实验和集落形成实验验证COA8表达量对NB细胞增殖的影响。结果 构建了包含10个DNA损伤修复相关基因的NB预后模型(IL4、HBE1、WRNIP1、RSF1、MDM4、COA8、PDCD10、HIST3H2A、PRODH和ATR)。模型风险评分可将患儿分为低风险组和高风险组,且低风险组比高风险组患儿预后好(P<0.0001)。生存分析证明COA8高表达导致NB预后不良。同时CCK8实验和集落形成实验证明降低COA8表达量,可抑制NB细胞增殖。结论 构建了具有良好预测功能的DNA损伤修复相关的NB多基因预后模型。验证了COA8是NB不良预后的潜在生物标志物。Objective To construct and evaluate the prognostic model of neuroblastoma(NB).In the study,bioinformatics tools were used to mine and screen DNA damage repair genes related to survival and prognosis of NB.Methods We downloaded DNA damage repair genes from Gene Set Enrichment Analysis(GSEA).TARGET database was regarded as the training dataset.Combined with expression profile data and clinical data,LASSO(Least absolute shrinkage and selection operator)and multivariate Cox regression methods were used to construct a prognosis model for DNA damage repair related genes in NB.GEO database was regarded as the validation dataset to evaluate the accuracy of the prognostic model.The area under the curve(AUC)of the prognostic model was calculated using the receiver operating cure(ROC).The predictive performance of the prognostic model was evaluated by univariate Cox regression and multivariate Cox regression analysis.Kaplan-Meier survival curve was performed using data from public database to verify effects of COA8 and PDCD10 on the prognosis of NB.The effect of COA8 and PDCD10 expression on the proliferation of NB cells was verified by CCK8 assay and colony formation assay.Results A NB prognostic model was constructed consisting of ten DNA damage repair related genes:IL4,HBE1,WRNIP1,RSF1,MDM4,COA8,PDCD10,HIST3H2A,PRODH and ATR.The model risk score classified patients into the low-risk and high-risk groups and the patients in the high-risk group had a lower survival rate than those in the low-risk group(P<0.0001).Kaplan-Meier survival curve demonstrated that high expression of COA8 led to a poor prognosis of NB,while CCK8 and colony formation experiments showed that reducing expression levels of COA8 could inhibit the proliferation of NB cells.Conclusion We have developed a prognostic model consisting of ten DNA damage repair genes which demonstrated a reliable predictive efficacy.COA8 is identified as a potential biomarker of the poor prognosis of NB.

关 键 词：神经母细胞瘤 DNA损伤修复 预后模型 生物信息学 

分 类 号：R739.4[医药卫生—肿瘤]