Predicting drugs' in vivo effects from in vitro testing requires modelling processes that are not reproduced by in vitro systems. The most obvious difference between the two situations is the absence of the absorption, distribution, metabolism, and excretion (ADME) processes that govern target tissue exposure in vivo. For identical input doses, the concentrations to which in vitro systems are exposed may not correspond to those found in vivo. The partners of Predict-IV work package 5 will develop prediction models for ADME processes on the basis of in vitro information. They will assemble all of these models into a global, generic, physiologically-based pharmacokinetic (PBPK) model able to simulate concentration-time profiles in human blood or tissues. In vivo drug toxicity will then be predicted by coupling the predicted profiles to the pharmacodynamic (PD) relationships observed in vitro. Human inter- or intraindividual variability will be evaluated for each component of the approach. The progress made on the following specific objectives will be discussed: - Predicting ADME processes based upon in vitro data or in silico tools; - Developing a generic PBPK model for humans and rats; - Developing dose-response models for in vitro toxicity endpoints in liver, kidneys and central nervous system; - Coupling the PBPK and dose-response models; - Predicting the impact of human variability on the drugs' toxicity; - Simulating the PK/PD of drugs on sensitive sub-populations; - Confronting predictions to published observations using a meta-models approach.
