Molecular structure-based prediction of the toxicokinetics of inhaled vapors in humans

The objectives of the present study were: (1) to evaluate the adequacy of setting hepatic extraction ratio (E) equal to 0 or 1 in physiologically based toxicokinetic (PBTK) models to generate the theoretically plausible envelope of venous blood concentration (C-v) profiles, and (2) to couple this approach with molecular structure-based estimation of blood:air and tissue:blood partition coefficients (PCs) to predict the C-v profiles of volatile organic chemicals (VOCs) in humans. Setting E = 0 or 1 in PBTK models provided simulations of C-v envelopes that contained the C-v values determined in humans exposed to low concentrations of dichloromethane (DCM), ethylbenzene (EBZ), toluene (TOL), m-xylene (XYL), trichloroethylene (TCE), and 1,1,1-trichloroethane (TRI), Following the validation of using E = 0 or 1 in conventional PBTK models to predict the theoretically plausible envelope of C-v, a quantitative structure-toxicokinetic relationship (QST(k)R) model was constructed. The QST(k)R model used molecular structure information as the sole input to predict the PCs and considered E = 0 or 1 to generate simulations of the envelope of C-v, The experimental data on C-v were in most cases within the envelopes simulated using QSTkR model for DCM, EBZ, TOL, and XYL, but were outside the envelopes for TCE and TRI. The discrepancy observed between the C-v envelopes obtained using PBTK and QSTkR models can be explained by the fact that blood:air PCs of some VOCs were under-predicted while using molecular structure information. The modeling framework presented in this article represents the first animal-replacement tool that can provide a priori predictions of the toxicokinetic profiles of VOCs prior to laboratory experiments.