Evaluating In Vitro-In Vivo Extrapolation of Toxicokinetics

Abstract

Prioritizing the risk posed by thousands of chemicals potentially present in the environment requires exposure, toxicity, and toxicokinetic (TK) data, which are often unavailable. Relatively high throughput, in vitro TK (HTTK) assays and in vitro-to-in vivo extrapolation (IVIVE) methods have been developed to predict TK, but most of the in vivo TK data available to benchmark these methods are from pharmaceuticals. Here we report on new, in vivo rat TK experiments for 26 non-pharmaceutical chemicals with environmental relevance. Both intravenous and oral dosing were used to calculate bioavailability. These chemicals, and an additional 19 chemicals (including some pharmaceuticals) from previously published in vivo rat studies, were systematically analyzed to estimate in vivo TK parameters (e.g., volume of distribution [Vd], elimination rate). For each of the chemicals, rat-specific HTTK data were available and key TK predictions were examined: oral bioavailability, clearance, Vd, and uncertainty. For the non-pharmaceutical chemicals, predictions for bioavailability were not effective. While no pharmaceutical was absorbed at less than 10%, the fraction bioavailable for non-pharmaceutical chemicals was as low as 0.3%. Total clearance was generally more under-estimated for nonpharmaceuticals and Vd methods calibrated to pharmaceuticals may not be appropriate for other chemicals. However, the steady-state, peak, and time-integrated plasma concentrations of nonpharmaceuticals were predicted with reasonable accuracy. The plasma concentration predictions improved when experimental measurements of bioavailability were incorporated. In summary, HTTK and IVIVE methods are adequately robust to be applied to high throughput in vitro toxicity screening data of environmentally relevant chemicals for prioritizing based on human health risks.

Figure 1.

Overview of experiments and analysis. We collected new TK data and jointly analyzed that data with literature data in order to evaluate 4 key issues in IVIVE. Predictions of oral absorption, volume of distribution, clearance, and uncertainty were all evaluated. Data were modeled with both 1- and 2-compartment models (Figure 2) and if the 2-compartment model was selected, the volumes of the 2 compartments were added to make a volume of distribution and the rate for the second (elimination) phase was used as k_elim.

Figure 2.

Both the empirical 1- and 2-compartment models were investigated for each chemical time course. A single set of model parameters were optimized to describe all available plasma concentration data (i.e., simultaneously for both oral and intravenous doses and for all data sources). F_bio is the fraction (between 0 and 1) of the oral dose that is absorbed at a rate of k_gutabs into the primary compartment. Intravenous doses are added to the intravenous compartment at time 0. V_d is the volume of distribution in the 1-compartment model, and $V_{\mathrm{d}}\hspace{0.17em}=\hspace{0.17em}{V}_1\hspace{0.17em}+\hspace{0.17em}{V}_2$ for the 2-compartment model. Elimination (e.g., metabolic, renal) is characterized by the rate k_elim.

Figure 3.

A “heatmap” of physico-chemical properties, in vitro TK parameters (Wetmore et al., 2013), and TK parameters estimated from in vivo plasma concentration. Rows (chemicals) are clustered by Euclidean distance so that adjacent rows are more similar to each other. Each column (chemical properties) was scaled by the standard deviation of the column and the mean value was subtracted, such that a value of 0 corresponds to the mean and values of −1 or 1 correspond to values one standard deviation above or below the chemicals, respectively. In some cases, TK parameters could not be estimated (e.g., no oral data available for estimating F_bio and k_gutabs). Fraction of the compound that is neutral, positively, or negatively ionized at pH 7.4 is indicated by “neutral ph74,” “positive ph74,” and “negative ph74.” The bar at the left-hand side indicates pharmaceuticals in gray and other chemicals in black.

Figure 4.

Comparison of measured volumes of distribution (V_d) with predictions based upon in vitro data and in silico methods (Pearce et al., 2017a; Schmitt, 2008). The solid line in each panel indicates the identity line (1:1, perfect predictions). Chemicals can be identified by their chemical abbreviation given in Table 1.

Figure 5.

Comparison of the TK clearance estimated from the in vivo data with the clearance predictions made using HTTK data. The more likely empirical TK model (either 1 or 2 compartments, as in Figure 2) is selected for each chemical using the AIC (Akaike, 1974). Estimated standard deviation is indicated by a vertical line, and is often smaller than the plotted text. The solid line indicates the identity line (1:1, perfect predictions), while the dotted and dashed lines indicate the linear regression (log-scale) trend lines for pharmaceuticals and other chemicals, respectively. Chemicals can be identified by their chemical abbreviation given in Table 1.

Figure 6.

Distribution of gut absorption rate for environmental and pharmaceutical chemicals.

Figure 7.

In panel A, the observed fraction of oral dose absorbed from the gut is compared with predictions made using gastroplus (Simulations Plus, 2017). The solid line indicates the identity line (1:1, perfect predictions). These fractions are distrusted from nearly 0 to effectively 100% (panel B). Chemicals can be identified by their chemical abbreviation given in Table 1.

Figure 8.

Rat in vivo data were collected for diverse environmental chemicals to evaluate the predictive ability of HTTK, especially with respect to predicting steady-state serum concentration (C_ss). Chemicals can be identified by their chemical abbreviation given in Table 1. The abbreviations are centered at the measured and predicted values. The solid line indicates the identity line (1:1, perfect predictions). In panel A, the 1-compartment model is parameterized with a predicted volume of distribution and clearance, based upon in vitro measured parameters. F_bio is assumed to be 100%. In panel B, the in vivo measured F_bio is used to reduce the amount of the oral dose absorbed in the 1-compartment model, to illustrate the improvement possible if F_bio could be predicted accurately.

Figure 9.

Evaluation of Wambaugh et al. (2015) classification scheme for predicting errors made when using HTTK data to predict in vivo steady-state plasma concentration (C_ss). Chemicals were placed into categories, depending upon the size of the error. “On the order” represented the best case, where errors were within 3.2 times over or under the true value. Some chemicals were determined to be problematic due to limitations in the HTTK methods (e.g., plasma protein binding estimation) or failure to come to steady state. Wherever the chemical names overlap the vertical, gray bands, the observed errors are consistent with predicted error. Chemicals can be identified by their chemical abbreviation given in Table 1.

Figure 10.

Evaluation of the ability of in vitro HTTK data, coupled with a 1-compartment model, to predict important TK statistics like the maximum plasma concentration (C_max) for characterizing tissue exposure. A single point is plotted for each combination of chemical, dose amount, and dose route, either intravenous (iv) or per oral (po). In panel A, the 1-compartment model is parameterized with a predicted volume of distribution and clearance, based upon in vitro measured parameters. F_bio is assumed to be 100%. In panel B, the in vivo measured F_bio is used to reduce the amount of the oral dose absorbed in the 1-compartment model, to illustrate the improvement possible if F_bio could be predicted accurately.

Figure 11.

Evaluation of predictions for the time-integrated plasma concentration (i.e., AUC). A single point is plotted for each combination of chemical, dose amount, and dose route, either intravenous (iv) or per oral (po). In panel A, the 1-compartment model is parameterized with a predicted volume of distribution and clearance, based upon in vitro measured parameters. F_bio is assumed to be 100%. In panel B, the in vivo measured F_bio is used to reduce the amount of the oral dose absorbed in the 1-compartment model, to illustrate the improvement possible if F_bio could be predicted accurately. The solid line in each panel indicates the identity line (1:1, perfect predictions).