Using neural networks to determine the contribution of danshensu to its multiple cardiovascular activities in acute myocardial infarction rats
Using neural networks to determine the contribution of danshensu to its multiple cardiovascular activities in acute myocardial infarction rats
复制标题
DOI：
10.1016/j.jep.2011.08.069
复制
发表时间：
2011-10-31
影响因子：
5.4
通讯作者：
Liu, Xiao-Quan
中科院分区：
文献类型：
--
作者：
Chen, Yuan-Cheng;Cao, Wan-Wen;Liu, Xiao-Quan
民族药理学相关性：丹参素是丹参的水溶性活性成分，具有多种心血管调节作用。然而，丹参素的相对贡献，其多种心血管activity. Objective的研究仍然在很大程度上是未知的：建立人工神经网络（NN）模型，同时表征丹参素的药代动力学和多种心血管活动在急性心肌梗死（AMI）大鼠。材料与方法：采用冠状动脉结扎法建立大鼠急性心肌梗死模型，单次腹腔注射丹参素20 mg/kg，观察丹参素对大鼠心肌缺血再灌注后的药代动力学（PK）和药效学（PD）的影响。检测丹参素、心肌肌钙蛋白T（cTnT）、总同型半胱氨酸（Hcy）和还原型谷胱甘肽（GSH）的水平。以丹参素的血药浓度-时间曲线下面积（AUC）和大鼠体重（协变量）为输入变量，建立了丹参素药代动力学和药效学的BP神经网络模型。根据加尔森的算法，利用神经元连接权来评估输入变量对输出神经元的相对贡献。结果：丹参素具有显著的降低cTnT、升高Hcy和GSH的作用，神经网络模型能很好地捕捉到这些指标。相对贡献率计算结果显示，丹参素对PD指标的影响顺序为cTnT > GSH > Hcy，而AMI对PD指标的影响顺序为cTnT > Hcy > GSH。结论：神经网络是丹参素PK和PD谱与多种心脏保护机制联系的有力工具，为识别和排序药物多种治疗作用的相对贡献提供了一种简单的方法。(C)2011爱思唯尔爱尔兰有限公司保留所有权利。
Ethnopharmacological relevance: Danshensu is an active water-soluble component from Salvia Miltiorrhiza, which has been demonstrated holding multiple mechanisms for the regulation of cardiovascular system. However, the relative contribution of danshensu to its multiple cardiovascular activities remains largely unknown.Aim of the study: To develop an artificial neural network (NN) model simultaneously characterizing danshensu pharmacokinetics and multiple cardiovascular activities in acute myocardial infarction (AMI) rats. The relationship between danshensu pharmacokinetics (PK) and pharmacodynamics ( PD) were evaluated using contribution values.Materials and methods: Danshensu was intraperitoneally injected at a single dose of 20 mg/kg to AMI rats induced by coronary artery ligation. Plasma levels of danshensu, cardiac troponin T (cTnT), total homocysteine (Hcy) and reduced glutathione (GSH) were quantified. A back-propagation NN model was developed to characterize the PK and PD profiles of danshensu, in which the input variables contained time, area under plasma concentration-time curve (AUC) of danshensu and rat weights (covariate). Relative contribution of input variable to the output neurons was evaluated using neuron connection weights according to Garson's algorithm. The kinetics of contribution values was also compared and was validated using bootstrap resampling method.Results: Danshensu exerted significant cTnT-lowering, Hcy- and GSH-elevating effect, and these marker profiles were well captured by the trained NN model. The calculation of relative contributions revealed that the effect of danshensu on the PD marker could be ranked as cTnT > GSH > Hcy, while the effect of AMI disease on the PD marker could be ranked in the following order: cTnT > Hcy > GSH. The activity of transsulfuration pathway was quite obvious under the AMI state.Conclusion: NN is a powerful tool linking PK and PD profiles of danshensu with multiple cardioprotective mechanisms, it provides a simple method for identifying and ranking relative contribution to the multiple therapeutic effects of the drug. (C) 2011 Elsevier Ireland Ltd. All rights reserved.