Objective: To deeply investigate the gene expression profiles of esophageal squamous cell carcinoma (ESCC) and the relationship of gene expression levels with prognosis from The Cancer Genome Atlas (TCGA) database. Methods: RNA-seq V2 data of 11 normal samples and 81 esophageal squamous cell carcinoma patients, and their corresponding clinical data were downloaded from The Cancer Genome Atlas database. Differentially expressed genes between normal and tumor samples were identified by using edgeR package. Gene function enrichment analyses of differentially expressed genes were conducted. A protein-protein interaction network based on differentially expressed genes was constructed by using STRING database and the hub genes were identified based on the created gene co-expression network. In addition, survival analysis was performed. Results: Totally, 2 788 genes were identified as differential expression. Among these, 1 168 genes were up-regulated and 1 620 genes were down-regulated in tumor cases compared with normal samples. Up-regulated genes were enriched in cell cycle, DNA replication and mismatch repair pathways, while down-regulated genes were enriched in metabolic pathways. 707 genes and their 3 428 interactions were identified by protein-protein interaction analysis. Genes with copy number amplifications were considered to interact with other crucial genes. 10 co-expression modules were identified based on the gene co-expression network analysis and the ribosomal protein genes were illustrated to be correlated with tumor locations of ESCC patients (P=0.003). The 3-years survival rates of high and low expression of TNFRSF10B groups were 82.5% and 15.1%, respectively. Similarly, the 3-years survival rates of high and low expression of DDX18 groups were 82.4% and 15.2%, respectively. The survival differences stratified by these two genes were statistically significant (both P<0.1). Conclusions: The analysis results of TCGA database showed that ribosomal protein genes are correlated with tumor locations of ESCC patients. Low expressions of TNFRSF10B and DDX18 are associated with poor prognose of ESCC patients. Consequently, TNFRSF10B and DDX18 may serve as predictive markers for ESCC patients.
目的： 利用肿瘤基因图谱计划(TCGA)数据库中食管鳞癌数据，探讨正常食管上皮细胞和食管鳞癌细胞中差异表达的基因及其与患者预后的关系。 方法： 在TCGA数据库中检索81例食管鳞癌的转录组数据和患者的临床资料。采用edgeR软件鉴定正常组织和肿瘤组织间差异表达的基因，对差异表达基因行功能富集分析，应用String数据库构建蛋白间的互相作用网络。基于表达值构建基因共表达网络和基因表达水平的高低对患者进行分组，分析各组患者的预后。 结果： 在TCGA数据库中共鉴定到2 788个差异表达基因，其中1 168个基因在肿瘤组织中表达水平上调，1 620个基因在肿瘤组织中表达水平下调。上调的基因富集到细胞周期、DNA复制和错配修复等通路，下调的基因富集到代谢相关通路。蛋白互相作用网络分析获得包含707个基因及其3 428个互作关系，食管鳞癌组织中发生拷贝数扩增的基因与其他一些重要基因存在相互作用。基因共表达分析检测到10个共表达模块，其中棕色模块的核糖体蛋白基因与肿瘤所在食管部位有关(P=0.003)。TNFRSF10B高表达组和TNFRSF10B低表达组患者的3年生存率分别为82.5%和15.1%，DDX18高表达组和DDX18低表达组患者的3年生存率分别为82.4%和15.2%，差异均有统计学意义(均P<0.1)。 结论： TCGA食管鳞癌数据库中，与核糖体蛋白相关的基因与食管鳞癌发生的部位有关。TNFRSF10B和DDX18的低表达与患者的预后差有关，其可能成为食管鳞癌患者潜在的预后分子标志物。.
Keywords: Differentially expressed genes; Esophageal neoplasms; Neoplasms, squamous cell; Prognosis.