人工智能结合生物网络对“养血清脑制剂”和“逍遥丸”组方的优化及验证  

Optimization and Validation of Formulas of Yangxue Qingnao(养血清脑)Preparation and Xiaoyao(逍遥)Pills Based on Artificial Intelligence and Biological Networks

作　　者：赵倩 魏宇 郭凯敏 王文佳 周水平 孙鹤 方坚松[3] 胡蕴慧 ZHAO Qian;WEI Yu;GUO Kaimin;WANG Wenjia;ZHOU Shuiping;SUN He;FANG Jiansong;HU Yunhui(Tasly Pharmaceutical Group Co.,Ltd.,Tianjin 300410;National Key Laboratory of Chinese Medicine Modernization,Tianjin 300193;Science and Technology Innovation Center,Guangzhou University of Chinese Medicine,Guangzhou 510405)

机构地区：[1]天士力医药集团股份有限公司,天津300410 [2]现代中药创制全国重点实验室,天津300193 [3]广州中医药大学科技创新中心,广州510405

出　　处：《中药药理与临床》2024年第4期27-33,共7页Pharmacology and Clinics of Chinese Materia Medica

基　　金：国家自然科学基金面上项目(编号:82074278);天津市自然科学基金(编号:22JCYBJC00180)。

摘　　要：目的:利用人工智能和大数据探寻符合中医药传统的复方中药组方优化思路,为复方中药二次开发提供参考。方法:围绕“君臣佐使”理论,基于中药材安全性预测模型,预测养血清脑制剂、逍遥丸中佐药和使药的安全性:(1)剔除安全性低的药材后获得改良方;(2)保留安全性高的药材,结合中药材成分吸收、分布、代谢和排泄性质(ADME)及靶点网络分析获得改良方。随机游走算法计算养血清脑制剂改良方、逍遥丸改良方与其原方的相关性。取ICR小鼠112只,随机分为模型对照组、阿司匹林肠溶片0.2 g/kg组、养血清脑制剂原方7、14、28 g/kg组和改良方7、14、28 g/kg组,每组14只。末次给药后1 h各小鼠腹腔注射0.6%醋酸溶液,观察药物对小鼠扭体反应的影响。随机将78只SD大鼠分为正常对照组、盐酸氟西汀0.01 g/kg组、慢性不可预知温和刺激(CUMS)结合孤养抑郁大鼠模型对照组、逍遥丸原方4.34 g/kg组、改良方4.44 g/kg和4.40 g/kg组,每组13只。在试验第0、7、14、21、28、35、42、49、56、63 d分别称体质量,进行糖水消耗试验和敞箱试验;采用酶联免疫吸附(ELISA)法检测大鼠海马中多巴胺(DA)、去甲肾上腺素(NE)、5-羟色胺(5-HT)、脑源性神经营养因子(BDNF)的含量。结果:养血清脑制剂改良方和原发性头痛显著相关(Z score=11.95);与模型对照组相比,养血清脑制剂原方7、14、28 g/kg组和改良方7、14、28 g/kg组均能明显抑制醋酸所致小鼠扭体次数,呈现明显镇痛作用(P<0.05或P<0.01)。逍遥丸改良方与抑郁症显著相关(Z score=23.61);与正常对照组相比,第63 d模型对照组大鼠出现显著抑郁症状(P<0.01);与模型对照组相比,逍遥丸改良方(减薄荷和茯苓)4.40 g/kg组显著改善模型大鼠的抑郁症状(P<0.01),大鼠海马中DA、NE、5-HT、BDNF的含量显著增加(P<0.01),优于逍遥丸原方4.34 g/kg组和改良方(减薄荷)4.44 g/kg组。结论:养血清脑�Objective:This study used artificial intelligence and big data to explore ideas for optimizing compound traditional Chinese medicine(TCM)formulas,so as to provide a reference for the secondary development of compound Chinese medicine.Methods:The study focused on the traditional Chinese medicine theory of Jun(君)-Chen(臣)-Zuo(佐)-Shi(使)and predicted the safety of Zhuo and Shi drugs in Yangxue Qingnao(养血清脑)Preparation and Xiaoyao(逍遥)Pills based on the TCM safety prediction model.1)The improved formula was obtained after removing the herbals with low safety.2)The herbals with high safety were retained.Based on the absorption,distribution,metabolism,and excretion(ADME)properties of herbals and target network correlation analysis,improved formulas were obtained.The random walk algorithm was used to calculate the correlation between the improved formulas of Yangxue Qingnao Preparation and Xiaoyao Pills and their original formulas,respectively.A total of 112 ICR mice were randomly divided into a model group,enteric-coated aspirin tablet group of 0.2 g/kg,original Yangxue Qingnao Preparation groups of 28,14,and 7 g/kg,and improved formula groups of 28,14,and 7 g/kg,with 14 mice in each group.The mice were intraperitoneally injected with 0.6%acetic acid solution 1 h after the last administration,and the effect of the drug on the writhing reaction of mice was observed.78 SD mice were randomly divided into a control group,fluoxetine hydrochloride group of 0.01 g/kg,chronic unpredictable mild stimulation(CUMS)combined with solitary depression rat model group,original Xiaoyao Pill formula group of 4.34 g/kg,and improved formula groups of 4.44 g/kg and 4.40 g/kg,with 13 mice in each group.Weight measurement,sugar water consumption,and open field tests were conducted on the 0th,7th,14th,21st,28th,35th,42nd,49th,56th,and 63rd days of the experiment.Enzyme-linked immunosorbent assay(ELISA)was used to detect the contents of dopamine(DA),norepinephrine(NE),5-hydroxytryptamine(5-HT),and brain-derived neurotrophic fa

关 键 词：养血清脑制剂 逍遥丸 人工智能 君臣佐使 组方优化 验证 

分 类 号：R285.5[医药卫生—中药学]