基于毒性通路扰动的药物线粒体毒性预测的研究

Mitochondrion is a major toxic target of drug-induced serious adverse reactions, and drug-induced mitochondrial toxicity has caused great concern from drug R&D enterprises and the regulatory authorities. However, there is still no applicable predictive methods for mitochondrial toxicity. Toxicity pathway perturbation-based toxicity testing strategy provides a new way for drug safety prediction. In the present project, we focus on the prediction difficulty of mitochondrial toxicity of candidate compounds in the early stage of R&D, and introduces the strategy of toxicity pathway perturbation-based toxicity testing. PGC-1α and Nrf2 signaling pathways, which are closely correlated to the regulation of mitochondrial functions, were selected as toxicity pathway candidates. To demonstrate the feasibility that the two pathways are applicable for the prediction of drug-induced mitochondrial toxicity, the effects of various drugs with mitochondrial toxicity on PGC-1α and Nrf2 signaling pathways were analyzed in several human-derived cell lines, mainly by high content analysis. These analyses covered the key steps of drug safety prediction, including the identification and verification of the toxicity pathway modeling framework, establishment of high throughput assays, toxicity pathway dose-response modeling for the prediction of the threshold dose of pathway perturbation, and PBPK model for the data extrapolation between in vitro and in vivo. The execution of the project will provide a new strategy and way for the prediction of drug-induced mitochondrial toxicity in the early stages of R&D, and helps to reveal the potential mechanisms of drug-induced mitochondrial toxicity, and provides guidance on the structure optimization of candidate compounds.

线粒体是许多药物导致严重不良反应的毒性作用靶标，但目前国内外仍缺乏理想的药物线粒体毒性预测方法。基于“毒性通路”扰动的体外毒性测试策略为药物安全性预测提供了全新的评价途径。本课题针对在药物研发早期实现候选化合物潜在线粒体毒性预测，引入基于“毒性通路”扰动的药物毒性测试策略，选择与线粒体功能调控密切相关的信号通路—PGC-1α通路和Nrf2通路作为候选毒性通路，重点应用细胞高内涵技术，在多种人源靶细胞中分析不同线粒体毒性药物对信号通路的影响，研究其作为线粒体毒性预测的毒性通路的可行性。项目研究内容涵盖了安全性预测的关键环节，包括毒性通路框架的构建、高通量分析方法的建立、毒性通路剂量-反应模型预测药物致毒性通路扰动的阈剂量以及PBPK模型实现体外体内数据的外推。本课题的实施将为药物早期线粒体毒性预测提供新的评价策略与方法，并有助于揭示药物致线粒体毒性的可能机制，为候选化合物的结构优化提供依据。

线粒体毒性是导致药物研发失败和药品撤市的重要原因。本项目旨在建立基于毒性通路扰动的药物线粒体毒性测试策略与技术方法。按项目计划完成了申请书的全部研究工作，主要包括1） 建立基于毒性通路-关键毒性反应-有害结局关联分析的药物线粒体毒性评估方法。主要应用体外替代方法阐明药物致靶器官线粒体损伤的特征，识别发现多柔比星、氟他胺和异烟肼等药物致肝脏和心脏线粒体毒性的毒性通路与关键毒性反应，揭示毒性通路与关键毒性反应以及线粒体损伤结局之间的相互关联，为阐明药物致靶器官线粒体毒性机制提供有力证据。2） 建立了快速识别选择性损伤线粒体药物的筛查技术。针对药物致靶器官细胞毒性的机制与因果关系分析，本课题组建立了基于细胞毒表型谱的定量和动态分析技术，通过同时检测细胞死亡数量和毒性机制事件的浓度和时间关系，准确识别线粒体毒性在特定药物毒性中的作用，为快速识别药物线粒体毒性机制、建立具有科学因果关系的药物不良反应路径奠定了基础。3） 构建基于体外毒性测试和毒性通路的计算毒理学模型的药物线粒体毒性预测技术。本项目基于PGC-1α通路的生物学网络，构建PGC-1α通路计算毒理学模型。综合利用实验室获得的体外实验数据，分析阐明药物对关键毒性通路的扰动作用特征，同时利用PGC-1α通路计算模型，实现对药物的心肌线粒体毒性进行预测，并采用基因沉默等体外实验进行验证，表明基于PGC-1α通路的计算毒理学模型在预测药物心肌线粒体毒性中具有较好应用价值。4） 建立基于PBPK的药物线粒体毒性体外-体内数据外推模型。针对药物致心脏和肝脏线粒体损伤的特点及临床相关资料，整合药物的理化特征及其在心脏或肝脏中的吸收、分布和代谢等特征，以及心脏和肝脏的生理学特征，建立PBPK模型，并应用这些模型对分别预测外推多柔比星、对乙酰氨基酚和曲格列酮等药物诱导的靶器官线粒体毒性。