目的 探究结直肠腺癌经典序列转化中的差异基因及其与免疫细胞浸润和患者生存时间的相关性，并预测具有靶向作用的中药。方法 从GEO数据库寻找符合要求的配对样本数据，使用R语言筛选序列转化过程中始终存在差异表达的关键基因，并通过基因本体（gene ontology，GO）进行功能注释。利用基因表达谱互动分析数据库（Gene Expression Profiling Interactive Analysis，GEPIA）和人类蛋白质图谱数据库验证关键基因在转录组和蛋白组水平的表达。通过Kaplan-Meier Plotter和TIMER数据库分别进行基因表达水平与结直肠癌患者生存时间、免疫细胞浸润的相关性分析。最后，利用Coremine Medical、BATMAN-TCM和草本植物数据库进行中药预测，使用Cytoscape构建“基因-中药”靶向网络，综合Network Analyzer和Swiss ADME数据库评估活性成分类药性，PyMOL软件进行分子对接。结果 获得结直肠腺癌经典序列转化的关键差异基因48个，筛选得到9个核心基因：TAGLN、FN1、ACTA2、DCN、REG4、GUCA2A、SPINK4、DEFA6、ZG16。与正常组织相比，结直肠癌组织中的核心基因在转录组和蛋白质水平上均具有统计学差异。功能性分析提示其主要参与肠道免疫、细胞外基质构成等过程。相关性分析表明核心基因的表达水平能够影响结直肠癌患者的总生存期和无复发生存期，CD4⁺T细胞、巨噬细胞、中性粒细胞和树突细胞是被核心基因调节的主要免疫细胞群。获得靶向中药129味，包括36味药食同源中药和27味毒性中药，以补益、清热、化痰为主，性味以寒、平、温、甘、苦、辛居多，多归肝、肺、脾经，其中葡萄、金荞麦、半夏、地龙、肉苁蓉的中心度值较高，筛选得到6种关键活性成分缬氨酸、鼠李糖、脯氨酸、橙花醛、蛋氨酸、甘氨酸，分子对接显示中药小分子活性成分与靶向蛋白质对接稳定。结论 筛选出“正常黏膜-肠腺瘤-肠腺癌”序列转化中始终差异表达的关键疾病基因，这些基因可能通过参与免疫调节、细胞外基质等过程影响结直肠腺癌患者的远期预后，金荞麦、葡萄、肉苁蓉等药食同源中药有望成为新药开发来源，为预防肠道早癌和癌症进展提供新的思路。

Objective To identify differentially expressed genes (DEGs) in the process of colorectal mucosa-adenoma-adenocarcinoma sequence and their relationships with immune cell infiltration and survival time of patients with colorectal cancer (CRC), and to predict potential targeted traditional Chinese medicines (TCMs). Methods The paired data was downloaded from the GEO database to obtain DEGs using R. Biological pathways were explored through gene ontology (GO). Expression levels of DEGs were validated using GEPIA and the Human Protein Atlas. Kaplan-Meier Plotter and TIMER databases were used to assess the correlation between DEGs and survival time, as well as immune cell infiltration. Finally, Coremine Medical, BATMAN, and Herb databases were used to predict TCMs. A “Gene-TCM” network was constructed with Cytoscape. Combined with Network Analyzer, the drug-likeness of the active compounds was evaluated using Swiss ADME databases. Molecular docking was performed with PyMOL software. Results Forty-eight key DGEs were identified in transformation, with nine hub genes: TAGLN, FN1, ACTA2, DCN, REG4, GUCA2A, SPINK4, DEFA6, and ZG16. At both at the transcriptome and proteome levels, the hub genes exhibited statistically significant differences in expression between tumor and normal tissues. Functional analysis indicated they were involved in intestinal immunity and extracellular matrix composition. Correlation analysis suggested that the expression levels of hub genes can affect the overall survival (OS) and relapse-free survival (RFS) of patients with CRC. CD4⁺ T cells, macrophages, neutrophils, and dendritic cells were primary immune cells regulated by hub genes. There were 130 potential TCMs, including 36 Medicine-food homologous TCMs and 27 toxic TCMs. The efficacy was mainly to tonify, clear heat, and resolve phlegm. Moreover, their four natures and five flavors were primarily cold, neutral, warm, sweet, bitter, and pungent. Most of them belonged to the liver, lung, and spleen meridians. Putao (Vitis vinifera), Jinqiaomai (Fagopyri Dibotryis Rhizoma), Banxia (Pinelliae Rhizoma), Dilong (Pheretima), Roucongrong (Cistanches Herba) had the strongest connections to hub genes. Six key active compounds were valine, L-rhamnose, proline, neral, D-methionine, and glycine. Molecular docking showed that they were stable in docking with markers. Conclusion This study identified genes that were consistently differentially expressed during the sequence. These DEGs may affect the prognosis of patients with CRC by regulating immune cells and extracellular matrix. Medicine-food homologous TCMs such as Fagopyri Dibotryis Rhizoma, Vitis vinifera, and Cistanches Herba could be promising sources for drug development. This study sheds new light on the prevention of colorectal adenoma and cancer.

Q811.4;R285

北京市自然科学基金项目（7242237）