BP神经网络算法结合超高效液相色谱-质谱联用技术研究红花治疗慢性酒精性肝损伤的作用机制  

Study on the Mechanism of Flos Carthami Treatment for Chronic Alcoholic Liver Injury Using BP Neural Network Algorithm Combined with UHPLC-MS Technology

作　　者：王曦烨[1] 韩晓静 姜明洋[3] 白梅荣[2] 许良[1] WANG Xi-ye;HAN Xiao-jing;JIANG Ming-yang;BAI Mei-rong;XU Liang(Inner Mongolia Key Laboratory for the Natural Products Chemistry and Functional Molecular Synthesis,College of Chemistry and Materials Science,Inner Mongolia Minzu University,Tongliao 028000,China;Key Laboratory of Mongolian Research and Development Project in Ministry of Education,Mongolian Medical College,Inner Mongolia Minzu University,Tongliao 028000,China;College of Computer Science and Technology,Inner Mongolia Minzu University,Tongliao 028000,China)

机构地区：[1]内蒙古民族大学化学与材料学院,天然产物化学与功能分子合成自治区重点实验室,内蒙古通辽028000 [2]内蒙古民族大学蒙医药学院,蒙医药研发工程教育部重点实验室,内蒙古通辽028000 [3]内蒙古民族大学计算机科学与技术学院,内蒙古通辽028000

出　　处：《质谱学报》2024年第6期897-906,I0006,共11页Journal of Chinese Mass Spectrometry Society

基　　金：国家自然科学基金地区项目(82260844,62162049);内蒙古自治区高等学校青年科技英才支持计划(NJYT23136);内蒙古自治区高等学校创新团队发展计划(NMGIRT2216);内蒙古自治区高等学校科学研究项目(NJZY21430);内蒙古自治区高等学校创新团队发展计划项目(NMGIRT2417);内蒙古自治区直属高校基本科研业务费项目(GXKY22118);内蒙古自然科学基金(2021MS08072);蒙医药研发工程教育部重点实验室开放基金(MDK2021041)。

摘　　要：临床上,红花对慢性酒精性肝损伤(chronic alcoholic liver injury,CALI)有很好的疗效,但治疗机制不甚明确。因此,阐明红花治疗CALI的分子作用机制对药物的进一步开发及应用具有重要意义。以雄性Wistar大鼠为研究对象,模型组大鼠以8 mL/kg酒精连续灌胃28天,建立CALI模型;给药组大鼠分别以高(4.290 3 g/kg)、中(1.430 1 g/kg)、低(0.476 7 g/kg)剂量灌胃红花提取物。采用大鼠血清代谢组学分析方法结合超高效液相色谱-质谱技术鉴定与CALI相关的潜在生物标志物,并研究红花对这些生物标志物的调控机制。利用MATLAB软件建立BP神经网络模型处理组学数据的分类问题。从苏木精和伊红(H&E)染色实验发现,高剂量红花提取物减轻了肝细胞的损伤程度;与模型组相比,高剂量红花组中的丙氨酸氨基转移酶(ALT)和天冬氨酸氨基转移酶(AST)的表达水平降低,表明高剂量红花提取物具有肝保护作用。BP神经网络模型的分类准确率为95.8%,分类效果良好。通过火山图分析共鉴定出20种与CALI相关的生物标志物,红花可以对这些生物标志物产生回调效果。研究表明,红花可能通过对甘油三酯、脂肪酸、磷脂、胆汁酸、氨基酸、维生素E代谢的调控作用而对CALI产生治疗效果。本研究可为红花的推广和临床应用提供了理论基础。Flos Carthami(FC)has a good therapeutic effect on chronic alcoholic liver injury(CALI)in clinical practice,but the treatment mechanism is not very clear.Therefore,elucidating the molecular mechanism of action of FC in treating CALI is of great significance for the further development and application of drugs.Male Wistar rats were as the research object.The model group rats were orally administered with 8 mL/kg alcohol for 28 days to establish a CALI model,while the treatment group rats were orally administered with high(4.2903 g/kg),medium(1.4301 g/kg),and low(0.4767 g/kg)doses of FC extract.Potential biomarkers related to CALI were identified using rat serum metabonomics analysis methods combined with ultra-high performance liquid chromatography mass spectrometry(UHPLC-MS)technology,and the regulatory mechanisms of FC on these biomarkers were investigated.The BP neural network model was established by MATLAB software to deal with the classification problem of omics data.The H&E staining experiment found that after gavage,high-dose FC extract can reduce the degree of liver cell damage.Compared with the model group,the expression levels of alanine aminotransferase(ALT)and aspartate aminotransferase(AST)in the high-dose FC group decrease,indicating that the high-dose FC extract has liver protective effect.The classification accuracy of the BP neural network model is 95.8%,and the classification effect is good.Through volcanic map analysis,a total of twenty biomarkers related to CALI are identified,and FC can have a callback effect on these biomarkers.The results indicated that FC may exert therapeutic effect on CALI by regulating the metabolism of triglycerides,fatty acids,phospholipids,bile acids,amino acids and Vitamin E.This study provides a theoretical foundation for the promotion and clinical application.

关 键 词：红花 慢性酒精性肝损伤 代谢组学 超高效液相色谱-质谱(UHPLC-MS) BP神经网络算法 

分 类 号：O657.63[理学—分析化学]