A generic, cross-chemical predictive PBTK model with multiple entry routes running as application in MS Excel; design of the model and comparison of predictions with experimental results

Abstract

Aim: Physiologically based toxicokinetic (PBTK) models are computational tools, which simulate the absorption, distribution, metabolism, and excretion of chemicals. The purpose of this study was to develop a physiologically based pharmacokinetic (PBPK) model with a high level of transparency. The model should be able to predict blood and urine concentrations of environmental chemicals and metabolites, given a certain environmental or occupational exposure scenario.

Model: The model refers to a reference human of 70 kg. The partition coefficients of the parent compound and its metabolites (blood:air and tissue:blood partition coefficients of 11 organs) are estimated by means of quantitative structure-property relationship, in which five easily available physicochemical properties of the compound are the independent parameters. The model gives a prediction of the fate of the compound, based on easily available chemical properties; therefore, it can be applied as a generic model applicable to multiple compounds. Three routes of uptake are considered (inhalation, dermal, and/or oral) as well as two built-in exercise levels (at rest and at light work). Dermal uptake is estimated by the use of a dermal diffusion-based module that considers dermal deposition rate and duration of deposition. Moreover, evaporation during skin contact is fully accounted for and related to the volatility of the substance. Saturable metabolism according to Michaelis-Menten kinetics can be modelled in any of 11 organs/tissues or in liver only. Renal tubular resorption is based on a built-in algorithm, dependent on the (log) octanol:water partition coefficient. Enterohepatic circulation is optional at a user-defined rate. The generic PBTK model is available as a spreadsheet application in MS Excel. The differential equations of the model are programmed in Visual Basic. Output is presented as numerical listing over time in tabular form and in graphs. The MS Excel application of the PBTK model is available as freeware.

Experimental: The accuracy of the model prediction is illustrated by simulating experimental observations. Published experimental inhalation and dermal exposure studies on a series of different chemicals (pyrene, N-methyl-pyrrolidone, methyl-tert-butylether, heptane, 2-butoxyethanol, and ethanol) were selected to compare the observed data with the model-simulated data. The examples show that the model-predicted concentrations in blood and/or urine after inhalation and/or transdermal uptake have an accuracy of within an order of magnitude.

Conclusions: It is advocated that this PBTK model, called IndusChemFate, is suitable for 'first tier assessments' and for early explorations of the fate of chemicals and/or metabolites in the human body. The availability of a simple model with a minimum burden of input information on the parent compound and its metabolites might be a stimulation to apply PBTK modelling more often in the field of biomonitoring and exposure science.