目的 探索胃“炎-癌”转化中的潜在作用靶点及其发病机制，并采用生物信息学技术预测潜在干预“炎-癌”进展的中药并验证。方法 从GEO数据库下载胃癌及慢性胃炎基因芯片，分为训练组及验证组，采用sva包及ComBat函数移除批次效应并整合。采用R包Limma、ClusterProfiler、DOSE、ggplot2等包筛选显著差异表达基因（differentially expressed genes，DEGs）并进行基因本体（gene ontology，GO）富集分析、京都基因与基因百科全书（Kyoto encyclopedia of genes and genomes，KEGG）通路富集分析。采用机器学习筛选特征基因并验证，分析特征基因表达与胃癌风险。采用R包E1071、PreProcessCore等进行免疫浸润及相关性。特征基因基于数据库映射潜在天然活性成分及对应中药，预测结果采用分子对接及细胞实验验证。结果 筛选出62个DEGs，DEGs主要参与粒细胞迁移、髓性白细胞迁移、顶端质膜构成等生物功能，参与代谢途径、细胞黏附分子、糖酵解/糖元生成等信号通路。经机器学习筛选并验证得到特征基因ATP结合盒转运体C5（ATP binding cassette transporter C5，ABCC5）、磷酸烯醇丙酮酸羧激酶1（phosphoenolpyruvate carboxykinase 1，PCK1）、碳酸酐酶IX（carbonic anhydrase IX，CA9）、甲酰肽受体1（formyl peptide receptor 1，FPR1）、磷酸肌醇3-激酶2类γ亚基（phosphoinositide-3-kinase class 2 gamma polypeptide，PIK3C2G）、前列腺干细胞抗原（prostate stem cell antigen，PSCA）、果糖-1,6-二磷酸酶2（fructose-bisphosphatase 2，FBP2），列线图分析显示特征基因异常表达群体存在高度转化风险；免疫浸润分析显示，胃癌患者浆细胞、CD8⁺T细胞、调节性T细胞、未活化NK细胞、M0巨噬细胞显著下降，滤泡辅助性T细胞、活化NK细胞、单核细胞、M1巨噬细胞、未活化树突状细胞、活化树突状细胞、嗜酸性粒细胞、中性粒细胞显著上调。靶点-成分筛选得活性成分姜黄素、杨梅素、槲皮素、查耳酮等，预测中药组方为姜黄、侧柏叶、草豆蔻、艾叶、枳实，对接显示成分、靶点间对接活性良好，细胞实验表明特征基因CA9、FBP2、ABCC5、PCK1、FPR1在胃“炎-癌”转化过程中表达显著，中药组方可显著调节其表达。结论 胃“炎-癌”转化机制复杂，预测所得中药组方可通过多途径发挥作用，研究可为胃癌发生机制及早期干预药物开发提供参考。

Objective To explore potential targets and pathogenesis in the “inflammation-cancer” transformation of the stomach and predict traditional Chinese medicine (TCM) interventions for modulating this progression using bioinformatics, with experimental validation. Methods Gene microarray data for gastric cancer (GC) and chronic gastritis were downloaded from the GEO database, which were divided into training and validation sets. The sva package and ComBat function were used to remove batch effects and integrate data. Differentially expressed genes (DEGs) were screened using R packages (Limma, ClusterProfiler, DOSE, ggplot2), followed by gene ontology (GO) and Kyoto encyclopedia of genes and genomes (KEGG) pathway enrichment analyses. Machine learning was used to screen and verify characteristic genes, and the expression of characteristic genes and the risk of GC were analyzed. Immune infiltration and correlation were performed using R packages (E1071, PreProcessCore). The characteristic genes are mapped based on the database to potential natural active ingredients and corresponding TCMs. The prediction results are verified by molecular docking and cell experiments.Results A total of 62 DEGs were identified, enriched in biological processes such as granulocyte migration, myeloid leukocyte migration, and apical plasma membrane composition, as well as signal pathways including metabolic pathways, cell adhesion molecules, and glycolysis/gluconeogenesis. The characteristic genes ATP binding cassette transporter C5 (ABCC5) and phosphoenolpyruvate carboxykinase 1 (PCK1), carbonic anhydrase IX (CA9), formyl peptide receptor 1 (FPR1), phosphoinositide-3-kinase class 2 gamma polypeptide (PIK3C2G), prostate stem cell antigen (PSCA), fructose-1, 6-bisphosphatase 2 (FBP2) were obtained through machine learning screening and verification. The nomogram indicated high transformation risk in groups with abnormal feature gene expression. Immune infiltration analysis showed that plasma cells, CD8⁺T cells, regulatory T cells, unactivated NK cells, and M0 macrophages in patients with gastric cancer decreased significantly, while follicular helper T cells, activated NK cells, monocytes, M1 macrophages, unactivated dendritic cells, activated dendritic cells, eosinophils, and neutrophils were significantly upreregulated. Target-component screening yielded active ingredients such as curcumin, myricetin, quercetin, and chalcone. It was predicted that the TCM formula would include Jianghuang (Curcumae Longae Rhizoma), Cebaiye (Platycladi Cacumen), Caodoukou (Alpiniae Katsumadai Semen), Aiye (Artemisiae Argyi Folium), and Zhishi (Aurantii Fructus Immaturus). The docking indicated that the docking activity between the components and the targets was good. Cell experiments indicated that the characteristic genes CA9, FBP2, ABCC5, PCK1, and FPR1 were significantly expressed during the “inflammation-cancer” transformation process of the stomach, and the TCM formula could significantly regulate their expressions. Conclusion The “inflammation-cancer” transformation involves complex mechanisms. The predicted TCM formula may act through multiple pathways. This research can provide a reference for the GC pathogenesis and the development of early intervention drugs.

Q811.4;TP18;R285

上海医药中药传承和创新平台能力建设（2020006）