Metabolomics: Markers of Drug-Induced Liver Injury

描述(由申请人提供)： 开发临床用新药的成本是巨大的，而且在药物开发的后期阶段失败会对其产生很大影响。提早从开发流水线中移除在治疗剂量下可能对人体有毒性的候选药物，对药物开发成本产生了巨大影响。肝脏损伤是药物下架或增加安全警示的主要原因之一。这项建议的目的之一是确定一组从尿液中排泄的内源性代谢物，这些代谢物可用于筛选药物诱导的肝损伤的早期标志。这项建议的第二个目的是更深入地了解反映肝损伤特定机制的标志物。这两个目标都有可能更好地定义临床前使用的敏感标记物，并为患者群体开发标记物提供潜力。核磁共振和GC-MS代谢谱将用于给药大鼠的肝脏和尿液，这些大鼠服用安慰剂、无效水平或药物诱导的肝损伤水平的氯贝特、丙戊酸、异烟肼、苯妥英钠和对乙酰氨基酚。除代谢组学外，还将获得肝损伤的常规测量(肝脏重量、血清酶升高和组织病理学)。这些药物的尿液和肝脏代谢组学特征将被减少和分析，以提供预测每种药物的反应测量的信号模式。此外，单个药物的预测模式的联合将被用于根据所用药物、剂量水平、暴露时间以及与不良反应的相关性来区分所有组。用于识别信号子集(S)的方法将通过丢弃一出方法进行交叉验证。定义这些模式的信号将使用GC-MS、核磁共振和LC-MS/MS方法进行鉴定，然后分配给生化途径，以确定标记特征与作用模式的相关性。

DESCRIPTION (provided by applicant): The cost of developing new drug entities for clinical use is substantial, and is greatly impacted by failure during the later stages of drug development. The early removal from the development pipeline of drug candidates that are likely to be toxic at therapeutic doses in humans has a huge impact on the cost of drug development. Liver injury is one of the major reasons for removal of a drug from the market, or addition of safety alerts. One aim of this proposal is to identify a set of endogenous metabolites excreted in urine that can be used to screen, as an early marker, for drug-induced liver injury. A second aim of this proposal is to gain more insight into markers that are reflective of specific mechanisms of liver injury. Both aims have the potential of better defining sensitive markers for pre-clinical use as well as provide potential for development of markers for patient populations. NMR and GC-MS metabolomic profiles will be developed for liver and urine from rats administered vehicle, no-effect levels, or drug induced-liver injury levels of clofibrate, valproic acid, isoniazid, phenytoin, and acetaminophen. Conventional measures of liver injury (liver weights, elevation in serum enzymes, and histopathology) will be obtained in addition to metabolomic profiles. The urine and liver metabolomics profiles for these drugs will be reduced and analyzed to provide the pattern of signals that are predictive of the response measurements for each drug. In addition, the union of the predictive patterns for individual drugs will be used to differentiate all groups based on the drug administered, dose level, exposure duration, and correlation with adverse response. The method used to identify the sub-set(s) of signals will be cross-validated by the drop-one-out approach. The signals defining these patterns will be identified using GC-MS, NMR, and LC-MS/MS methods and then assigned to biochemical pathways for defining the relevancy of the marker profiles to mode of action.