US Environmental Protection Agency, Office of Research and Development, Center for Computational Toxicology and Exposure, Great Lakes Toxicology and Ecology Division, Duluth, Minnesota, USA.
Environ Toxicol Chem. 2023 Jul;42(7):1501-1515. doi: 10.1002/etc.5626. Epub 2023 May 17.
Computational models that predict chemical bioaccumulation in fish generally account for biotransformation using an apparent first-order whole-body rate constant (k ; d ). The use of such models requires, therefore, that methods exist for estimating k , ideally without the need to expose live animals. One promising approach for estimating k involves the extrapolation of measured in vitro intrinsic clearance (CL ) to the whole animal (in vitro-in vivo extrapolation, [IVIVE]). To date, however, the accuracy of such predictions has been difficult to assess due to uncertainties associated with one or more extrapolation factors and/or a mismatch between fish used to generate in vitro data and those used to conduct in vivo exposures. In the present study we employed a combined in vitro and in vivo experimental approach to evaluate the IVIVE procedure using pyrene (PYR) as a model chemical. To the extent possible, measured rates of CL were extrapolated to estimates of k using extrapolation factors based on measured values. In vitro material (liver S9 fraction) was obtained from fish exposed to PYR in a controlled bioconcentration study protocol. Fish from the same study were then used to estimate in vivo k values from an analysis of chemical depuration data. Averaged across four study groups, k values estimated by IVIVE underestimated those determined from in vivo data by 2.6-fold. This difference corresponds to a 4.1-fold underestimation of true in vivo intrinsic clearance, assuming the liver is the only site of biotransformation. These findings are consistent with previous work performed using mammals and have important implications for use of measured CL values in bioaccumulation assessments with fish. Environ Toxicol Chem 2023;42:1501-1515. Published 2023. This article is a U.S. Government work and is in the public domain in the USA.
计算模型通常使用表观一级全身速率常数（k; d）来预测鱼类的化学生物累积。因此，使用此类模型需要存在估计 k 的方法，理想情况下无需暴露活体动物。一种有前途的估计 k 的方法涉及将测量的体外固有清除率（CL）外推到整个动物（体外-体内外推，[IVIVE]）。然而，迄今为止，由于与一个或多个外推因子相关的不确定性和/或用于产生体外数据的鱼类与用于进行体内暴露的鱼类之间不匹配，因此很难评估此类预测的准确性。在本研究中，我们采用了一种结合体外和体内实验的方法，使用芘（PYR）作为模型化学物质来评估 IVIVE 程序。在可能的范围内，使用基于测量值的外推因子将测量的 CL 速率外推至 k 的估计值。体外材料（肝脏 S9 级分）是从按照受控生物浓缩研究方案暴露于 PYR 的鱼类中获得的。然后，从同一研究中选择鱼类，从化学净化数据的分析中估计体内 k 值。在四个研究组中平均，通过 IVIVE 估计的 k 值比从体内数据确定的 k 值低估了 2.6 倍。如果假设肝脏是唯一的生物转化部位，则该差异对应于真实体内固有清除率的 4.1 倍低估。这些发现与以前使用哺乳动物进行的工作一致，并对使用鱼类生物累积评估中的测量 CL 值具有重要意义。环境毒理化学 2023;42:1501-1515. 2023 年出版。本文是美国政府的工作，在美国属于公有领域。
