John Libbey Eurotext

Environnement, Risques & Santé

MENU

Predicting in vivo toxicokinetics of chemicals from in vitro data and QSAR models Volume 9, issue 6, Novembre-Décembre 2010

Figures

See all figures

Authors
Ineris Unité METO (Modèles pour l'écotoxicologie et la toxicologie) Parc technologique ALATA PB2 60550 Verneuil-en-Halatte France, Université de technologie de Compiègne CNRS-UMR 6600 Centre de recherches de Royallieu Rue Personne de Roberval BP 20.529 60205 Compiègne cedex France
  • Key words: dose-response relationship, in vitro, models, chemical, models, theoretical, test, toxicokinetics
  • DOI : 10.1684/ers.2010.0396
  • Page(s) : 489-501
  • Published in: 2010

Dose-response relationships in chemical risk assessment are commonly derived through simple mathematical models that link effects directly to exposure dose. These models, usually calibrated with animal data, are specific to the chemical, the endpoint and the experimental protocol. Taking toxicokinetics into account makes it possible to extrapolate results for different chemicals and different exposure scenarios. Among the various toxicokinetic models, physiologically based pharmacokinetic (PBPK) models are based on a mechanistic description of anatomy, physiology and the processes involved in the disposition of a compound within an organism, i.e. absorption, distribution, metabolism and excretion (ADME). Although physiological parameters have been well described for a large range of species, the literature contains little information about the parameters specific to individual chemicals. In vitro tests and in silico models based on physicochemical properties (QSAR : quantitative structure activity relationships) are a promising alternative to animal testing for estimating these parameters. In this paper, we review the use of PBPK models as an integrative tool to predict toxicokinetics based on in vitro tests and QSAR models. We illustrate this review by predicting the toxicokinetics of the volatile organic compound, 1,3-butadiene, and by comparing predictions and data observed in a human population with inter-individual variability. Integration of alternative methods into PBPK models should provide more realistic models for predictive toxicology and help deal with the lack of in vivo data for numerous marketed chemicals.