The use of HepaRG and human hepatocyte data in predicting CYP induction drug-drug interactions via static equation and dynamic mechanistic modelling approaches

Abstract

The method of predicting CYP induction drug-drug interactions (DDIs) from a relative induction score (RIS) calibration has been developed to provide a novel model facilitating predictions for any CYP-inducer substrate combination by inclusion of parameters such as the fraction of hepatic clearance mediated by a specific CYP and fraction of the dose escaping intestinal extraction. In vitro HepaRG CYP3A4 induction data were used as a basis for the approach and a large number of DDIs were well predicted. Primary human hepatocyte data were also used to make predictions, using the HepaRG calibration as a foundation. Similar predictive accuracy suggests that HepaRG and primary hepatocyte data can be used inter-changeably within the same laboratory. A comparison of this 'indirect' calibration method with a direct in vitro-in vivo scaling approach was made and investigations undertaken to define the most appropriate in vivo inducer concentration to use. Additionally, a reasonably effective prediction model based on F(2) (the concentration of inducer taken to increase the CYP mRNA 2-fold above background) was established. An accurate prediction for the CYP1A2-dependent omeprazole-caffeine interaction was also made, demonstrating that the methods are useful for the evaluation of DDIs from induction involving mechanisms other than PXR activation. Finally, a dynamic mechanistic model accounting for the simultaneous influence of CYP induction and reversible and irreversible CYP inhibition in both the liver and intestine was written to provide a prediction of the overall DDI when several interactions occur concurrently. The rationale for using the various models described, alongside commercially available prediction tools, at various stages of the drug discovery process is described.