目的：通过TCGA和GEO数据库筛选与食管腺癌相关的关键基因，并分析其生物学功能、相关信号通路和临床意义。方法：综合TCGA数据库食管腺癌数据和GEO数据库GSE92396芯片数据，使用R软件的DEseq2包和Limma包进行差异表达基因分析，获得共同差异表达基因。利用R软件的clusterProfiler包对共同差异表达基因进行GO功能富集分析及KEGG通路富集分析。运用string网站和Cytoscape3.7.2软件进行蛋白互作网络分析，筛选出调节食管腺癌蛋白表达量的关键节点基因，再结合TCGA数据库分析关键节点基因与患者生存的关系。结果：通过数据库中90例食管腺癌组织和18例正常食管组织标本的基因芯片数据的分析，获得共同差异表达基因521个，其中高表达基因356个，低表达基因165个，它们主要与表皮发育和表皮细胞分化的代谢过程等相关功能和细胞因子及其受体相互作用等信号通路密切相关。蛋白互作网络分析得出15个关键节点基因，其中CXCL8和CCL20低表达的食管腺癌患者生存期显著长于高表达者（中位生存期32.4 vs 19.7个月，P<0.05；32.4 vs 13.9个月，P <0.05）。结论：数据库挖掘显示CXCL8与CCL20基因可能在食管腺癌的发生发展及预后中起着重要作用，可以作为判断患者预后的潜在指标。

Objective: To screen the key genes associated with esophageal adenocarcinoma by using TCGA and GEO databases, and to analyze their biological functions, relevant signaling pathways and clinical significance. Methods: The esophageal adenocarcinoma data downloaded from TCGA database and GSE92396 microarray data from GEO database were integrated. The analysis of differentially expressed genes (DEGs) were performed by using DEseq2 and Limma packages of R software to obtain the co-differentially expressed genes, which were then chosen for the GO function enrichment analysis and KEGG pathway analysis with clusterProfiler package of R software. The key node genes that regulate the protein expressions in esophageal adenocarcinoma were screened out by protein-protein interaction (PPI) network analysis using the string website and Cytoscape 3.7.2 software. The correlation between key node genes and the survival of patients was further analyzed by combining with TCGA database. Results: By analyzing the chip data of 90 cases of adenocarcinoma tissues and 18 cases of normal esophageal tissues from databases, a total of 521 co-differentially expressed genes were obtained, including 356 upregulated genes and 165 downregulated genes. These genes were closely related to the metabolicassociated functions mainly involving epidermis development, epidermal cell differentiation and signaling pathways involving cytokinecytokine receptor interaction, etc. The PPI network analysis revealed 15 key node genes. The survival time for patients with low CXCL8 and CCL20 expression was significantly longer compared with the patients with high expression level (median survival: 32.4 vs 19.7 months, P<0.05; 32.4 vs 13.9 months, P<0.05). Conclusion: These results show that CXCL8 and CCL20 may play an important role in the occurrence, development and prognosis of esophageal adenocarcinoma, and may be used as potential indicators to judge the prognosis of patients.

河北省自然科 学 基 金 资 助 项 目（No. H2019206664）