基于生物信息学分析非酒精性脂肪性肝病的线粒体自噬相关基因及防治中药筛选  

Analysis of mitophagy-related genes in non-alcoholic fatty liver disease based on bioinformatics and screening of traditional Chinese medicine for prevention and treatment

作　　者：高松林 韦柳婷 管晓 温文建 覃雁 刘沛恬 廖芸旭 黄贵华[2] GAO Songlin;WEI Liuting;GUAN Xiao;WEN Wenjian;QIN Yan;LIU Peitian;LIAO Yunxun;HUANG Guihua(Guangxi University of Chinese Medicine,Nanning 530000,China;The First Affiliated Hospital of Guangxi University of Chinese Medicine,Nanning 530000,China)

机构地区：[1]广西中医药大学,广西南宁530000 [2]广西中医药大学第一附属医院,广西南宁530000

出　　处：《中草药》2024年第19期6655-6665,共11页Chinese Traditional and Herbal Drugs

基　　金：广西研究生教育创新计划项目(YCBZ2023148);广西科技基地和人才专项(桂科AD19245168);国医大师黄瑾明学术思想与临床诊疗传承发展研究中心建设项目(桂中医大党[2022] 24号);国家中医药管理局第七批全国老中医药专家学术经验继承工作项目(国中医药人教函[2022] 76号);黄贵华桂派中医大师培养项目(桂中医药科教发[2022] 6号);黄贵华广西名中医传承工作室项目(桂卫中医发[2017] 2号)

摘　　要：目的 基于生物信息学探讨非酒精性脂肪性肝病(non-alcoholic fatty liver disease,NAFLD)与线粒体自噬相关的特征基因,并筛选对这些基因起调控作用的中药。方法 从GSE89632数据集中提取差异表达的线粒体自噬相关基因(differentially expressed mitophagy-related genes,DEMRGs),并对其进行相关性分析、蛋白互作网络构建和富集分析。根据DEMRGs的表达量对NAFLD患者进行共识聚类,分析亚型间DEMRGs的表达情况和基因集变异分析。构建并预测了4种机器学习模型的效能,通过列线图、校准曲线、决策曲线和外部数据集评估最佳机器学习模型的有效性,筛选最佳机器学习模型中NAFLD的线粒体自噬特征基因。预测能够调控线粒体自噬特征基因治疗NAFLD的潜在中药。结果 在NAFLD样本和正常对照样本中获得14个DEMRGs,DEMRGs具有相互调控的作用。根据DEMRGs的表达量可将NAFLD分为2个亚型。随机森林(random forest,RF)模型是最佳模型,可以准确地预测NAFLD。UNC-51样激酶1(UNC-51-like kinase 1,ULK1)、磷酸酶和张力蛋白同源物(phosphatase and tensin homologue deleted on chromosome ten,PTEN)诱导的激酶1(PTEN induced putative kinase 1,PINK1)、FUN14结构域蛋白1(FUN14 domain containing 1,FUNDC1)、微管相关蛋白1轻链3β(microtubule associated protein 1 light chain 3 beta,MAP1LC3B)和视神经蛋白(optineurin,OPTN)是NAFLD的线粒体自噬特征基因。具有NAFLD潜在治疗中药有三七、川芎、藁本等,功效主要以补虚、清热、化痰止咳平喘和活血化瘀为主。结论 DEMRGs在NAFLD的发生和发展中起重要作用,ULK1、PINK1、FUNDC1、MAP1LC3B和OPTN是NAFLD与线粒体自噬相关的特征基因,三七、川芎、藁本可能通过调节线粒体自噬治疗NAFLD,为NAFLD的诊断和治疗提供了新思路。Objective To explore the characteristic genes related to mitophagy in non-alcoholic fatty liver disease(NAFLD)based on bioinformatics,and to screen the traditional Chinese medicine which can regulate these genes.Methods Differentially expressed mitophagy-related genes(DEMRGs)were extracted from the GSE89632 dataset and analyzed for correlation,protein interaction network and enrichment analyses among DEMRGs.Consensus clustering of NAFLD patients was performed according to the expression of DEMRGs.Expression of DEMRGs between subtypes was analyzed as well as gene set variant analysis.Four machine learning models were constructed and their performance were predicted.The effectiveness of the optimal machine learning model was evaluated by nomograms,calibration curves,decision curves,and external datasets.The characteristic genes related to mitophagy for NAFLD were screened in terms of optimal machine learning models.Finally,the potential traditional Chinese medicines(TCMs)that can regulate the characteristics of mitophagy for the treatment of NAFLD was infered.Results A total of 14 DEMRGs were obtained in NAFLD samples and normal control samples.The DEMRGs have a mutual regulatory role.NAFLD can be categorized into two subtypes based on the expression of DEMRGs.Random forest(RF)model was the best model to predict NAFLD accurately.UNC-51-like kinase 1(ULK1),phosphatase and tensin homologue deleted on chromosome ten(PTEN)induced putative kinase 1(PINK1),FUN14 domain containing 1(FUNDC1),microtubule associated protein 1 light chain 3 beta(MAP1LC3B)and optineurin(OPTN)are characterized genes associated with mitophagy in NAFLD.Potential therapeutic TCMs include Sanqi(Notoginseng Radix et Rhizoma),Chuanxiong(Chuanxiong Rhizoma),and Gaoben(Ligustici Rhizoma et Radix),etc.The efficacy of these TCMs mainly focuses on supplementing deficiency,clearing heat,resolving phlegm,stopping coughs and calming asthma,and promoting blood circulation and removing blood stasis.Conclusion DEMRGs play an important role in the occurrence and

关 键 词：非酒精性脂肪性肝病 线粒体自噬相关基因 生物信息学 机器学习 三七 川芎 藁本 

分 类 号：R285[医药卫生—中药学]