Mohammed Cherkaoui, Stuart Paine, Cyril Rauch, Peter Littlewood
University of Nottingham
Objectives: To generate a mechanistic dynamic model for the prediction of Drug-Drug Interactions (DDIs), which results from time-processes within hepatocytes, taking into account the spatial distribution of the drugs in a lobule, the uptake at the sinusoidal membrane, the enzyme inhibition/induction [1] and the up-regulation of the enzyme gene.
Methods: Over 70 clinical DDIs [2,3], including inhibitors, inducers and mixed interactions, were compared with the prediction using a static model and the new dynamic model. This was implemented in MATLAB® and inserted into a PBPK model with 4 compartments (Blood/Gut/Liver/Rest).
The Blood and Rest compartments are simple compartments with a physiological volume and a partition coefficient. The Blood compartment has a partition coefficient of one, whereas the Rest compartment depends on the drug.
The Gut compartment comprises two sub-compartments: the first represents the gut wall with a first order absorption for the oral dose and takes into account DDIs within the enterocytes assuming a well-stirred compartment; and the second represents the portal vein.
The estimations of the drug parameters (inhibition/inactivation/induction/uptake) were obtained with in vitro experiments and adjusted for the human liver size. The PK parameters (clearance/absorption rate) were obtained from the literature [4,5].
For each clinical case, the AUC ratio of the victim drug was estimated with the dynamic model and compared to the static model along with the clinical outcome.
Results: The preliminary results show that the model accurately predicts the DDI of the compounds which are purely inhibitors (reversible or time-dependent) or inducers. For compounds which are both, the prediction is less accurate. Overall, more than 50% of the DDIs have been predicted within 2-fold and more than 89% within 5-fold. The Geometric mean fold error (GMFE) has been estimated as 2.19, which is in the same range as the current static model ([2,3]: GMFE=1.7-2.5).
Conclusions: The model is consistent with those in the literature. It also provides a dynamic description of the DDIs, such as the enzyme level and spatial distribution within a lobule. Furthermore, the perpetrator dose regimen can be changed to observe its influences on the AUC ratio.
Finally, as in the static model [2,3], the DDIs prediction of compounds demonstrating inhibition and induction in-vitro is poor. These could be the result of a more complex mechanism occurring in the liver and/or intestine as an MDR1 induction or the perpetrator metabolite playing a role in the DDI.
References:
[1] Z. Zhang et al., “Enzyme Kinetics for Clinically Relevant CYP inhibition.” Current Drug Metabolism, 6(3), pp. 241-257, 2005.
[2] O. A. Fahim et al., “A Combined Model for Predicting CYP3A4 Clinical Net Drug-Drug Interaction Based on CYP3A4 Inhibition, Inactivation and Induction Determined in Vitro.”, Drug Metabolism and Disposition, 36(8), pp. 1698-1708, 2008.
[3] O. A. Fahim et al., “Comparison of different algorithms for predicting clinical drug-drug interactions, based on the use of CYP3A4 in vitro data: Prediction of compounds as precipitants of interaction.”, Drug Metabolism and Disposition, 37(8), pp. 1658-1666, 2009.
[4] S. K. Quinney et al., “Physiologically Based Pharmacokinetics Model of Mechanism-Based Inhibition of CYP3A by Clarithromycin.”, Drug Metabolism and Disposition, 38(2), pp. 241-248, 2010.
[5] K. Ito et al., “Prediction of the in vivo interaction between midazolam and macrolides based on in vitro studies using human liver microsomes.”, Drug Metabolism and Disposition, 31(7), pp. 945-954, 2003.
Reference: PAGE 24 () Abstr 3481 [www.page-meeting.org/?abstract=3481]
Poster: Drug/Disease modeling - Absorption & PBPK