目的 通过网络药理学和分子对接，预测柴胡疏肝散（Chai Hu Shu Gan San，CHSGS）治疗肝细胞癌（Hepatocellular carcinoma，HCC）的分子靶点，构建预后模型和列线图。方法 从TCGA和GEO数据库筛选出差异表达基因、加权共表达基因和聚类模块与正常组的差异基因；将四者取交集，得到HCC发生发展密切相关的基因；从TCMSP和SymMap数据库下载CHSGS的化合物和靶点，进一步将CHSGS的靶点与HCC相关基因进行取交集得到治疗HCC相关靶点，将其治疗HCC相关靶点进行GO和KEGG功能富集分析；使用SymMap数据库，构建CHSGS治疗HCC“病-证-方/药”网络。将PPI和构建的药物靶点网络取交集，得到核心的靶点，进一步使用cox LASSO回归分析构建相关的预后模型；结合临床性状使用单因素和多因素cox分析构建列线图；使用分子对接评估分子靶点的结合情况。使用人类蛋白质图谱（HPA）数据库进行核心靶点的免疫组织化学染色验证。结果 得到了HCC发生发展密切相关基因890个。CHSGS治疗HCC靶点38个。GO和KEGG分析显示，CHSGS治疗HCC可能与对无机物的反应、类固醇代谢过程和对细胞外刺激的反应有关；CHSGS治疗HCC可能与FOXM1信号通路有关。cox LASSO回归构建了预后模型，其模型公式为：（0.1165）*CCNB1 + （-0.0396）*PON1 + （0.0888）*CHEK1 + （0.0521）*SPP1 + （0.0093）*NQO1 + （0.0102）*SERPINE1 + （0.0012）*IGFBP3。单因素cox联合多因素cox分析显示，ESR1、T分期是独立的预后危险因素。列线图结果显示，其c-index为0.68，P<0.001。分子对接结果显示，异鼠李素与ESR1的结合良好。免疫组织化学染色结果在蛋白质水平验证了结果的准确性。结论 异鼠李素可能是CHSGS治疗HCC的主要成分之一，CHSGS治疗HCC的靶点用以构建的预后模型可以用于患者预后的评估，根据ESR1的表达水平、T分期构建的列线图可用于患者总体生存率的估算。

Objective To predict the molecular targets of Chai Hu Su Gan San (CHSGS) in the treatment of hepatocellular carcinoma (HCC) through network pharmacology and molecular docking, and to construct a prognostic model and nomogram.Methods We screened the differentially expressed genes, weighted co-expressed genes and clustering modules from the TCGA and GEO databases and the differential genes of the normal group. We took the intersection of the four to obtain genes closely related to the occurrence and development of HCC. We downloaded CHSGS from the TCMSP and SymMap databases. The target of CHSGS and HCC-related genes were further crossed to obtain therapeutic HCC-related targets, and the therapeutic HCC-related targets were subjected to GO and KEGG functional enrichment analysis; we used the SymMap database to construct a "disease-syndrome-prescription/medicine" network for the treatment of HCC by CHSGS. We took the intersection of PPI and the constructed drug target network to obtain the core target, and further used cox LASSO regression analysis to construct related prognostic models; combined clinical traits to use single-factor and multi-factor cox analysis to construct nomograms; We used molecular docking to evaluate the binding of molecular targets. We used the Human Protein Atlas (HPA) database for immunohistochemical staining verification.Results 890 genes closely related to the occurrence and development of HCC were obtained. CHSGS treated 38 HCC targets. GO and KEGG analysis showed that CHSGS treatment of HCC may be related to the response to inorganic substances, steroid metabolism and response to extracellular stimulation; CHSGS treatment of HCC may be related to the FOXM1 signaling pathway. cox LASSO regression constructed a prognostic model, and the model formula was: (0.1165)*CCNB1 +(-0.0396)*PON1 +(0.0888)*CHEK1 +(0.0521)*SPP1 +(0.0093)*NQO1 +(0.0102)*SERPINE1 +(0.0012)*IGFBP3. Single factor cox combined with multivariate cox analysis showed that ESR1 and T stage were independent prognostic risk factors. The result of the nomogram showed that its c-index was 0.68 (P<0.001). The molecular docking results showed that the combination of isorhamnetin and ESR1 was good. The results of immunohistochemical staining also verified the accuracy of the results at the protein level.Conclusion Isorhamnetin may be one of the main components of CHSGS in the treatment of HCC. The prognostic model constructed by the target of CHSGS in the treatment of HCC can be used to evaluate the prognosis of patients. A nomogram constructed based on the expression level of ESR1 and T staging can be used to estimate the overall survival rate of patients.

R285

国家自然科学基金委员会面上项目（81873303）：基于外泌体介导miR-23a-3p调控EMT探讨健脾化瘀方抑制肝癌转移机制，负责人：钟崇；湖南省卫健委2020年度科研立项课题（20200949）：基于Ras-MAPK信号通路探讨疏肝祛瘀解毒方诱导自噬治疗肝癌的分子作用机制，负责人：李菁。