基于生物信息大数据研究牡丹皮对自身免疫病“异病同治”作用机制及Q-marker预测

Mechanism of "Treating Different Diseases with Same Method" of Mudanpi(Cortex Moutan) in Treatment of Autoimmune Diseases and Q-Marker Prediction Based on Bioinformatics Data

目的 基于中医病机理论与系统药理学的方法探讨牡丹皮防治系统性红斑狼疮(systemic lupus erythematosus,SLE)与类风湿性关节炎(rheumatoid arthritis,RA)的“异病同治”功效网络和作用机制并对其Q-marker进行预测。方法 通过检索TCMSP、BATMAN、TCMID、ETCM、SymMap、中科院化学数据库等多个中药资源数据库结合中国知网文献挖掘建立牡丹皮化学成分库,利用ADME参数OB、DL值筛选活性成分。搜集TCMSP、BATMAN、TCMID数据库的活性成分靶点,STITCH、SIB、SEA数据库预测的活性成分靶点加以补充。Comparative toxicogenomics database(CTD)数据库收集SLE与RA相关靶点。根据韦恩图找出牡丹皮-SLE-RA共有靶点,构建蛋白互作网络,依据拓扑参数找出核心靶点。利用DAVID平台对共同交集靶点进行GO和KEGG分析,采用Cytoscape的插件ClueGO对共同核心靶点的GO-BP和KEGG分析结果进行可视化展示。基于“中药-成分-靶点-通路-疾病”网络分析预测Q-marker。结果 在建立的牡丹皮成分库中经过筛选共确定了28个有效成分,对应826个靶点,化合物靶点与两种疾病靶点共有795个交集靶点,依据蛋白互作网络拓扑特征共筛选出20个核心靶点,GO分析及KEGG富集分析筛选得到相关生物过程188个,细胞组分33个,分子功能45个(P<0.01,FDR<0.01),相关信号通路8条(P<0.01,FDR<0.01),主要涉及HIF-1、PI3K-Akt、MAPK、Toll、TNF、雌激素等途径,其中缺氧诱导因子-1(HIF-1)富集度最高,是“异病同治”的主要通路。网络药理学成分有效性分析预测得到10个成分可作为牡丹皮的Q-marker。结论 牡丹皮可通过多个成分协同作用于多个靶点、多条途径起到共同防治SLE与RA的效果,从靶点和通路两方面发挥“异病同治”机制,筛选出的Q-marker可用于牡丹皮药材及其相关复方制剂的质量控制,以适应中药现代化和国际化的要求。

Objective To explore the efficacy network and mechanism of Mudanpi(Cortex Moutan) to prevent and treat both systematic lupus erythematosis(SLE) and rheumatoid arthritis(RA) and predict Q-markers based on pathogenesis theory and systematic pharmacology of traditional Chinese medicine(TCM).Methods By searching multiple TCM resource databases such as TCMSP,BATMAN,TCMID,ETCM,SymMap and Chinese academy of sciences chemical database,combined with literature mining,the chemical classification database of Mudanpi(Cortex Moutan) was established,and the active ingredients were screened by using ADME parameters oral bioavailability(OB) and drug likeness(DL).The active ingredient targets of TCMSP,BATMAN and TCMID databases were collected,and the active ingredient targets predicted by STITCH,SIB and SEA databases were supplemented.The CTD database collected SLE and RA related targets.The common targets of compound and disease were identified according to Venn diagram,the protein interaction network was constructed,and the core targets were identified according to the topological parameters.The DAVID platform was used for GO and KEGG analysis of intersection targets,and the GO analysis and KEGG analysis results of core targets were visualized by using ClueGO of Cytoscape.Q-marker was predicted based on the network analysis of “traditional Chinese medicine-component-target-pathway-disease”.Results A total of 28 active components were identified in the established Mudanpi(Cortex Moutan) component library,corresponding to the 826 targets.There were 795 intersection targets between compound targets and targets of two diseases.A total of 20 core targets were screened according to the topological characteristics of protein-protein interaction network.Go analysis and KEGG enrichment analysis screened 188 related biological processes,33 cell components and 45 molecular functions(P<0.01,FDR<0.01),8 relevant signal paths(P<0.01,FDR<0.01).It mainly involved HIF-1,PI3 K Akt,MAPK,toll,TNF,estrogen and other pathways,of which hypoxia inducible factor-1(HIF-1) had the highest enrichment and was the main pathway of “treating different diseases with the same method”.The validity analysis of network pharmacological components predicted that 10 components could be used as Q-marker of Mudanpi(Cortex Moutan).