You are here:
Assessment of in vitro high throughput pharmacokinetic data to predict in vivo pharmacokinetic data of environmental chemicals
Citation:
Hughes, M., C. Ring, T. Fennell, D. MacMillan, J. Simmons, R. Thomas, AND J. Wambaugh. Assessment of in vitro high throughput pharmacokinetic data to predict in vivo pharmacokinetic data of environmental chemicals. Society of Toxicology, New Orleans, LA, March 13 - 17, 2016.
Impact/Purpose:
The purpose of this study was to assess pharmacokinetic model predictions of in vivo pharmacokinetics of several environmental chemicals. The models used published high throughput in vitro pharmacokinetic data to make these predictions. The predictions were assessed by generating in vivo pharmacokinetic data from animals dosed with the test chemicals.
Description:
Assessing the health risks of the thousands of chemicals in use requires both toxicology and pharmacokinetic (PK) data that can be generated more quickly. For PK, in vitro clearance assays with hepatocytes and serum protein binding assays provide a means to generate high throughput (HT) PK data. A PK model is used to extrapolate the in vitro data to in vivo exposures. In this study, we assessed the ability of current in vitro HTPK models to accurately predict in vivo PK parameters, such as peak plasma concentrations (Cmax), volume of distribution (Vd), and half-life. Twenty-six ToxCast chemicals were evaluated, including a variety of pesticides (e.g., carbofuran) and industrial chemicals (e.g., bisphenol A). Adult male Sprague-Dawley rats with an indwelling jugular vein cannula were dosed by oral gavage (po) or intravenously (iv) via the tail vein to doses corresponding to those that HTPK predicted would produce serum concentrations in the range of bioactivity observed in ToxCast bioassays. Serial blood samples were collected out to 96 h. Plasma was isolated by centrifugation and stored at -80o C until analyzed for parent compound by HPLC-MS/MS. For po exposure, HTPK predictions for Cmax were within 3-fold of measured values for 36% of observations. HTPK over-predicted Cmax (a risk-conservative error) for 71% of observations. The prediction of Vd was within 3-fold for 43% of observations, but half-life was >3-fold over-predicted for 87% of observations. The data suggest that models based on in vitro HTPK data may be sufficient for ~40% of the chemicals analyzed in this study. However, for most of the discordant chemicals, the bias present for the HTPK data was conservative in a risk context. The data suggest that the discordance for many chemicals may be due to overestimation of half-life. Additional data including bioavailability and effect of transporters may aid in rendering these assays and models more predictive in the future. (This abstract does not represent US EPA policy.)