Curriculum in Toxicology, Division of Medicinal Chemistry and Natural Products, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599, USA.
Chem Res Toxicol. 2010 Apr 19;23(4):724-32. doi: 10.1021/tx900451r.
Adverse effects of drugs (AEDs) continue to be a major cause of drug withdrawals in both development and postmarketing. While liver-related AEDs are a major concern for drug safety, there are few in silico models for predicting human liver toxicity for drug candidates. We have applied the quantitative structure-activity relationship (QSAR) approach to model liver AEDs. In this study, we aimed to construct a QSAR model capable of binary classification (active vs inactive) of drugs for liver AEDs based on chemical structure. To build QSAR models, we have employed an FDA spontaneous reporting database of human liver AEDs (elevations in activity of serum liver enzymes), which contains data on approximately 500 approved drugs. Approximately 200 compounds with wide clinical data coverage, structural similarity, and balanced (40/60) active/inactive ratios were selected for modeling and divided into multiple training/test and external validation sets. QSAR models were developed using the k nearest neighbor method and validated using external data sets. Models with high sensitivity (>73%) and specificity (>94%) for the prediction of liver AEDs in external validation sets were developed. To test applicability of the models, three chemical databases (World Drug Index, Prestwick Chemical Library, and Biowisdom Liver Intelligence Module) were screened in silico, and the validity of predictions was determined, where possible, by comparing model-based classification with assertions in publicly available literature. Validated QSAR models of liver AEDs based on the data from the FDA spontaneous reporting system can be employed as sensitive and specific predictors of AEDs in preclinical screening of drug candidates for potential hepatotoxicity in humans.
药物的不良反应（AEDs）仍然是药物开发和上市后停药的主要原因。虽然与肝脏相关的 AEDs 是药物安全性的主要关注点，但用于预测候选药物人类肝脏毒性的计算模型却很少。我们已经应用定量构效关系（QSAR）方法来对肝脏 AEDs 进行建模。在这项研究中，我们旨在构建一个能够基于化学结构对用于肝脏 AEDs 的药物进行二分类（活性与非活性）的 QSAR 模型。为了构建 QSAR 模型，我们采用了 FDA 人类肝脏 AEDs（血清肝酶活性升高）自发报告数据库，其中包含大约 500 种已批准药物的数据。选择了大约 200 种具有广泛临床数据覆盖、结构相似性和平衡（40/60）活性/非活性比例的化合物进行建模，并将其分为多个训练/测试和外部验证集。使用 k 最近邻方法开发 QSAR 模型，并使用外部数据集进行验证。在外部验证集中，为预测肝脏 AEDs 而开发的模型具有较高的灵敏度（>73%）和特异性（>94%）。为了测试模型的适用性，对三个化学数据库（世界药物索引、Prestwick 化学文库和 Biowisdom 肝脏智能模块）进行了计算机筛选，并通过将基于模型的分类与公开文献中的断言进行比较，确定了预测的有效性（在可能的情况下）。基于 FDA 自发报告系统数据的肝脏 AEDs 的验证 QSAR 模型可作为候选药物在人类潜在肝毒性的临床前筛选中用于预测 AEDs 的敏感和特异性预测因子。
