中国中医药信息杂志2026，Vol.33Issue(8)：23-31,9.DOI:10.19879/j.cnki.1005-5304.202603161
基于网络药理学和多种机器学习算法探讨健脾化痰方治疗肝细胞癌的机制
Exploration on the Mechanism of Jianpi Huatan Prescription in Treating Hepatocellular Carcinoma Based on Network Pharmacology and Multiple Machine Learning Algorithms
摘要
Abstract
Objective To systematically explore the molecular mechanism of Jianpi Huatan Prescription in treating hepatocellular carcinoma(HCC)by integrating network pharmacology and 12 machine learning algorithms;To provide a basis for the clinical translational application of this prescription.Methods The active components and corresponding targets of Jianpi Huatan Prescription were screened from the TCMSP,supplemented by literature retrieval.HCC transcriptome data were obtained from TCGA database to screen differentially expressed genes(DEGs).The potential targets of Jianpi Huatan Prescription against HCC were acquired by intersecting the two gene sets mentioned above.The importance of the intersecting genes was evaluated by integrating 12 machine learning algorithms to screen core diagnostic genes,and their diagnostic efficacy was verified using an independent dataset from the GEO database.GO and KEGG pathway enrichment analyses were performed to reveal the biological functions of the core genes.A subcutaneous xenograft tumor model was established by inoculating Hep1-6 cells into mice.The mice were randomly divided into a model control group,a lenvatinib positive control group,Jianpi Huatan Prescription low-and high-dosage groups,with 6 mice in each group.After 14 consecutive days of administration,tumor volume and weight were measured.Western blot,RT-qPCR and immunohistochemical staining were used to detect the expressions of key G2/M phase molecules(CDK1 and CCNB1)as well as apoptosis-related proteins.Results A total of 159 active components and 287 potential targets of Jianpi Huatan Prescription were screened,and 55 common targets were obtained by intersecting with HCC DEGs.20 core diagnostic genes were identified through comprehensive scoring based on 12 machine learning algorithms,with an average AUC of 0.901 in GEO validation.Enrichment analysis showed that these core genes were mainly enriched in cell cycle regulatory pathways such as G2/M phase transition.In vivo animal experiments demonstrated that Jianpi Huatan Prescription could significantly inhibit tumor growth,down-regulate the protein expressions of CDK1 and CCNB1,meanwhile reduce the protein expressions of pro-Caspase9 and pro-Caspase3,and up-regulate the protein expressions of cleaved-Caspase9 and cleaved-Caspase3(P<0.05,P<0.01,P<0.000 1).Conclusion Core diagnostic genes of Jianpi Huatan Prescription for HCC treatment are systematically screened via network pharmacology combined with 12 machine learning algorithms.This prescription may induce G2/M phase arrest by down-regulating CDK1 and CCNB1 expressions,thereby activating the mitochondrial apoptosis pathway and inhibiting HCC growth.关键词
肝细胞癌/网络药理学/机器学习/G2/M期检查点/细胞凋亡Key words
hepatocellular carcinoma/network pharmacology/machine learning/G2/M checkpoint/apoptosis分类
医药卫生引用本文复制引用
叶晨浩,刘影,邵高眩,徐汉辰,王磊..基于网络药理学和多种机器学习算法探讨健脾化痰方治疗肝细胞癌的机制[J].中国中医药信息杂志,2026,33(8):23-31,9.基金项目
上海市炎癌转化病证生物学前沿研究基地(2021KJ03-12) （2021KJ03-12）
