Ayatallah Saleh (1,2), Josefine Schulz (1), Franziska Kluwe (1,2), Linda B.S. Aulin (1), Wilhelm Huisinga (4), Gerd Mikus (1,3), Charlotte Kloft (1)* and Robin Michelet (1)*
(1) Department of Clinical Pharmacy and Biochemistry, Institute of Pharmacy, Freie Universität Berlin, Germany, (2) Graduate Research Training program PharMetrX, Germany, (3) Department of Clinical Pharmacology and Pharmacoepidemiology University Hospital Heidelberg, Heidelberg, Germany, (4) Institute of Mathematics, University of Potsdam, Germany
Introduction: Voriconazole (VRC), a broad-spectrum antifungal, exhibits nonlinear pharmacokinetics (PK), with complex metabolism and large inter- and intraindividual variability. It is extensively metabolised via CYP2C19 and CYP3A4, and slightly by CYP2C9 [1], with <2% of unchanged drug excreted renally [2,3]. VRC N-oxide (NO), the main circulating metabolite in plasma, does not add to the antifungal activity but is suspected of causing adverse events (AEs) [4,5]. To date, there is no consensus regarding the contribution of each metabolic pathway to VRC clearance (CL). Moreover, VRC and its metabolites are potent CYP inhibitors, leading to complex drug-drug interactions. Standard dosing might lead to under- or overexposure of VRC, causing therapeutic failure or increasing the risk for AEs [4].
Objectives: This work aimed to (i) elucidate the complex metabolism of VRC and its metabolites and (ii) assess their impact on VRC PK using a sequential middle-out approach, combining bottom-up (physiologically-based pharmacokinetic, PBPK) and top-down (nonlinear mixed-effects, NLME) approaches.
Methods: Initially, in vitro assays based on recombinant CYPs enzymes were carried out to get kinetic parameters [6]. Next, a whole-body PBPK model for VRC was developed using PK-Sim® (v. 10) [7]. Physicochemical properties were gathered [8] and the Poulin and Theil method was used to determine the organ-to-plasma partition coefficients. The PBPK model assumed that VRC is metabolised via CYP2C19, CYP3A4 and CYP2C9 and excreted via glomerular filtration. Enzymes’ tissue expression distribution was extracted from the PK-Sim® expression database [9] with a reference value of 0.76, 4.32 and 3.84 μmol/L for CYP2C19, CYP3A4 and CYP2C9, respectively [10]. A middle-out approach was applied using a clinical database comprising 10 clinical studies conducted at the Medical University of Vienna and the University Hospital of Heidelberg. A total of 153 individuals were enrolled, covering a broad spectrum of covariates and study designs, patient characteristics, CYP2C19 genotype, dosing routes and regimens, and co-medications. VRC and its metabolites NO, hydroxyvoriconazole (OH-VRC), and di-hydroxyvoriconazole were measured in plasma and retrieved in urine after a deglucuronidation step. Model evaluation was performed by comparing predicted and observed plasma concentration-time profiles, area under the concentration-time curve (AUC) values and maximum plasma concentrations.
Results: First, using a bottom-up approach, a priori predictions of drug exposure comparing subtherapeutic (50 mg infused over 2 h) vs therapeutic (400 mg infused over 2 h) single i.v. dosing were obtained. VRC showed fast elimination with linear PK at low doses, and the in vivo trajectories were well captured by the PBPK model’s a priori prediction. The observed in vivo trajectories of high-dose VRC indicated decreasing elimination over time, which was not captured by the PBPK model’s a priori prediction due to the (auto-)inhibition processes not yet accounted for. Second, CYP2C19 genotype-predicted phenotype was implemented into the model. The CYP2C19 enzyme abundance of poor metabolisers (PM), intermediate metabolisers (IM), normal metabolisers (NM) and rapid metabolisers (RM), were in the range of 0–0.24, 0.25–0.55, 0.56–1.05 and 1.06-3.55 μmol/L. Due to the significant overprediction observed for the CYP2C19 PM, a sensitivity analysis was carried out. Results showed that the predicted AUC was sensitive to fup and renal CL. Upon investigating the % of VRC and glucuronidated VRC excreted in the urine within 24 h after administration of 400 mg i.v., a significant difference in contribution of the renal and glucuronidation pathway was seen among different phenotypes (PM>IM>NM>RM). Therefore, literature in vitro data [11] were used to account for the direct glucuronidation of VRC via UDP‑glucuronosyltransferases (UGT1A4), with a reference concentration of 2.32 μmol of UGT1A4 per litre of liver tissue [10]. The turnover number of UGT1A4 (Kcat=13.74 1/min) was inferred from in vivo data of single i.v. dose in CYP2C19 PM, leading to better model performance in those individuals.
Conclusions: Combining in vitro, in silico and in vivo data, the proposed middle-out approach assisted in characterising VRC PK and will continue to in further iterations. In contrast to previous PBPK models [12,13], direct glucuronidation of VRC was identified as crucial metabolic pathway for CYP2C19 PM.
References:
[1] R. Hyland, B.C. Jones, D.A. Smith. Identification of the cytochrome P450 enzymes involved in the N-oxidation of voriconazole. Drug Metab. Dispos. 31: 540–7 (2003).
[2] N. Hohmann, R. Kreuter, A. Blank et al. Autoinhibitory properties of the parent but not of the N-oxide metabolite contribute to infusion rate-dependent voriconazole pharmacokinetics. Br. J. Clin. Pharmacol. 83: 1954–1965 (2017).
[3] S.J. Roffey, S. Cole, P. Comby et al. The disposition of voriconazole in mouse, rat, rabbit, guinea pig, dog, and human. Drug Metab. Dispos. 31: 731–41 (2003).
[4] J. Schulz, F. Kluwe, G. Mikus et al. Novel insights into the complex pharmacokinetics of voriconazole: a review of its metabolism. Drug Metab. Rev. 51: 247–265 (2019).
[5] O. Epaulard, M.-T. Leccia, S. Blanche et al. Phototoxicity and photocarcinogenesis associated with voriconazole. Med. Mal. Infect. 41: 639–45 (2011).
[6] J. Schulz, A. Thomas, A. Saleh et al. Towards the Elucidation of the Pharmacokinetics of Voriconazole: A Quantitative Characterization of Its Metabolism. Pharm. 2022, Vol. 14, Page 477 14: 477 (2022).
[7] http://www.open-systems-pharmacology.org/
[8] K. Vanstraelen, J. Wauters, H. De Loor et al. Protein-binding characteristics of voriconazole determined by high-throughput equilibrium dialysis. J. Pharm. Sci. 103: 2565–70 (2014).
[9] M. Meyer, S. Schneckener, B. Ludewig et al. Using expression data for quantification of active processes in physiologically based pharmacokinetic modeling. Drug Metab. Dispos. 40: 892–901 (2012).
[10] B. Achour, M.R. Russell, J. Barber et al. Simultaneous Quantification of the Abundance of Several Cytochrome P450 and Uridine 5′-Diphospho-Glucuronosyltransferase Enzymes in Human Liver Microsomes Using Multiplexed Targeted Proteomics. Drug Metab. Dispos. 42: 500–510 (2014).
[11] K. Bourcier, R. Hyland, S. Kempshall et al. Investigation into UDP-glucuronosyltransferase (UGT) enzyme kinetics of imidazole- and triazole-containing antifungal drugs in human liver microsomes and recombinant UGT enzymes. Drug Metab. Dispos. 38: 923–9 (2010).
[12] X. Li, S. Frechen, D. Moj et al. A Physiologically Based Pharmacokinetic Model of Voriconazole Integrating Time-Dependent Inhibition of CYP3A4, Genetic Polymorphisms of CYP2C19 and Predictions of Drug-Drug Interactions. Clin. Pharmacokinet. 59: 781–808 (2020).
[13] Y. Zhang, S. Zhao, C. Wang et al. Application of a Physiologically Based Pharmacokinetic Model to Characterize Time-dependent Metabolism of Voriconazole in Children and Support Dose Optimization. Front. Pharmacol. 12: 636097 (2021).
Reference: PAGE 30 (2022) Abstr 10184 [www.page-meeting.org/?abstract=10184]
Poster: Drug/Disease Modelling - Absorption & PBPK