Fatemeh Aghai (1), Bettina Gerner (3), Nora Isberner (1), Sabrina Krause (1), Götz Ulrich Grigoleit (1, 2), Sebastian Zimmermann (3), Max Kurlbaum (4), Hartwig Klinker (1), Oliver Scherf-Clavel (3)
(1) University of Wu¨rzburg Medical Center, Department of Internal Medicine II, Oberdu¨rrbacher Str. 6, 97080 Wu¨rzburg, Germany, (2) Division of Hematology and Oncology, Department of Medicine II, Helios Hospital Duisburg, Germany, (3)Institute for Pharmacy and Food Chemistry, University of Würzburg, Germany, (4) Division of Endocrinology and Diabetology, Department of Internal Medicine I and Core Unit Clinical Mass Spectrometry, University Hospital Würzburg, Germany
Introduction: Ruxolitinib is an orally administered small molecule multi-kinase inhibitor with potent and selective inhibition activity against janus kinase 1/2 [1]. It is metabolized by hepatic enzymes of the cytochrome P450 (CYP) family, predominantly by CYP3A4 and to a lesser extent by CYP2C9 [2]. Drug-drug-interaction (DDI) with the potent CYP3A4 inhibitors ketoconazole and fluconazole have been described in existing literature [3,4]. However, posaconazole as another potent inhibitor of CYP3A4 is frequently used for antifungal prophylaxis in patients receiving ruxolitinib in the treatment of graft-versus-host-disease (GvHD) [5].
Objectives: A descriptive and predictive physiologically-based pharmacokinetic (PBPK) model for posaconazole and ruxolitinib was developed to evaluate the potential DDI of these concomitantly administered drugs in clinically relevant doses.
Methods: PBPK modeling was performed using PK-Sim® Version 9 as part of the Open Systems Pharmacology Suite 8.0 [6]. PBPK models for ruxolitinib and posaconazole were independently set up using physicochemical properties of each drug (e.g. pka, logP) and study demographics (age, height, weight and body mass index) of clinical studies in humans (12 for posaconazole and three for ruxolitinib). Plasma concentration-time profiles after intravenous infusion (posaconazole), single (posaconazole, ruxolitinib) and multiple oral dosing (posaconazole, ruxolitinib) were used to adjust model parameters. Further drug specific values influencing the pharmacokinetics (e.g. fraction unbound, specific intestinal permeability, enzyme kinetics) were taken from literature. Model evaluation was performed by comparing predicted and observed plasma concentration-time profiles, AUC values and maximum plasma concentrations (Cmax). To quantitate model performance, the mean relative deviation (MRD) between predicted and observed values was calculated. A MRD value ≤ 2 was considered to signify an adequate prediction. Prior to linking the PBPK models for DDI simulations, the inhibitory effect of posaconazole on CYP3A4 (Ki) was verified using the published midazolam PBPK model available on the Open Systems Pharmacology model library on GitHub [7]. Furthermore, ruxolitinib concentrations obtained from routine clinical setting were compared to predicted concentrations for the concomitant use of orally administered ruxolitinib 10 mg twice daily and 300 mg posaconazole once daily.
Results: A PBPK DDI model for ruxolitinib and posaconazole was developed and evaluated, meeting the defined criteria of MRD ≤ 2 for published data. The model can also predict observed ruxolitinib concentrations from daily clinical routine within the 90% prediction intervall. Poulin and Theil and PK Sim Standard were used as distribution and cellular permeability models for posaconazole, respectively. Formulation dependent parameters and solubility for posaconazole were modified to account for the formulation dependent systemic exposure. Metabolism of posaconazole via UDP-Glucoronosyltrasferase-1A4 was modified by fitting the catalytic rate constant kcat to16.52 min-1. Excretion was modified by adjusting biliary clearance. Ruxolitinib was modeled using Rodgers and Rowland distribution model and PK Sim Standard to describe cellular permeability. Using final model parameter values the DDI model was able to describe and predict the expected increase in ruxolitinib (DDI Cmax ratio 1.23 and DDI AUC ratio 1.56) due to inhibition of CYP3A4 by posaconazole.
Conclusion: The model allows predictions of plasma concentration-time profiles of clinically relevant ruxolitinib doses with co-administration of posaconazole. The presented model can serve as a valuable tool to describe concentration-time profiles of ruxolitinib exposure and can be used to implement existing PBPK models to evaluate the impact of other drugs affecting CYP metabolism.
Funding: Hector Foundation II, Weinheim, Germany. Fond: STIF-99 (“Individualized cancer therapy with kinase inhibitors using drug monitoring – optimization by minimally invasive at-home sampling”).
References:
[1] Quintás-Cardama A et al Blood. 2010 Apr 15;115(15):3109-17.
[2] US Food and Drug Administration. Jakafi. Prescribing Information.
[3] Aslanis V et al Cancer Chemother Pharmacol. 2019 Oct;84(4):749-757.
[4] Shi JG et al J Clin Pharmacol. 2012 Jun;52(6):809-18.
[5] Ullmann AJ et al. N Engl J Med. 2007 Jan 25;356(4):335-47.
[6] Open Systems Pharmacology. http://www.open-systems-pharmacology.org/
[7] Open Systems Pharmacology. https://github.com/Open-Systems-Pharmacology/OSP-PBPK-Model-Library.
Reference: PAGE 29 (2021) Abstr 9851 [www.page-meeting.org/?abstract=9851]
Poster: Drug/Disease Modelling - Absorption & PBPK