Development of a Physiological-Based Pharmacokinetic Model for Ritonavir: Unraveling its Triphasic Effects as CYP3A Inhibitor, CYP3A Inducer, and Pgp Inhibitor.
Lien Thi Ngo (1), Tham Thi Bui (1), Hwi-Yeol Yun (1,2), Jung-Woo Chae (1,2)
(1) College of Pharmacy, Chungnam National University, South Korea, (2) Bio-AI Convergence Research Center, Chungnam National University, South Korea
Introductions: Ritonavir (RTV) was originally approved for the treatment of human immunodeficiency virus (HIV) infection. Currently, the drug is no longer used for its activity against HIV. Instead, it is used as a pharmacokinetic (PK) enhancer since RTV is well-known as a potent CYP3A inhibitor. RTV increased the area under the concentration curve (AUC) of midazolam (a strong CYP3A4 substrate) by 28.4 folds when concomitantly used1. In addition, RTV has the CYP3A4 induction effect at the steady-state condition. The drug clinically significantly decreases the AUC of ethinyl estradiol by 40%2. Simultaneously, RTV shows a potent inhibition effect on the Pgp transporter. Short-term and long-term administration of RTV increased the AUC of fexofenadine (a Pgp substrate) 2.8-fold and 1.4-fold, respectively3. However, there have not any available models accounting for all the induction/inhibition effects of RTV on these enzymes and the Pgp transporter. Considering the therapeutic indications, RTV could be administered in co-medications in many cases. Therefore, developing a model to predict drug-drug interactions (DDI) between RTV and other CYP3A/Pgp perpetrators is essential.
Objectives: Development of a physiological-based pharmacokinetic (PBPK) model for RTV that implements both the competitive- and mechanism-based inhibition effects and induction effect on CYP3A. The inhibition effect on Pgp of RTV was also described.
Methods: The development and evaluation of the whole-body PBPK model for RTV were performed using PK-Sim®. A clinical dataset covering a wide single and multiple RTV doses (from 100 to 600 mg) from 16 clinical PK profiles of RTV alone and 3 clinical DDI studies following the administration of RIV and alprazolam, midazolam, and clarithromycin in humans was used. Drug-dependent parameters (i.e., physicochemical and ADME properties) were extracted from the literature. Physiological-dependent parameters from PK-Sim® at default values, except for CYP3A4 and Pgp expression profiles, were used. The PK of RIV was modeled with the hepatic metabolism by CYP3A4/5 and CYP2D6, and the transportation by the Pgp transporter. The competitive inhibition effect of RTV on CYP3A4/5 and Pgp was modeled by Ki (µM) parameter. The mechanism-based inhibition effect of RTV on CYP3A4/5 was modeled by Kinact (1/min) and Kinact-half (µM) parameters. In addition, the induction effect of RTV on CYP3A4/5 was modeled by EC50 (µM) and Emax (no-unit) parameters. Input parameters not available from the literature were optimized by fitting the model to the observed dataset. Finally, some model parameters were refined to get the best fit. For model evaluation, plasma concentration-time profiles, AUC, and maximum concentrations (Cmax) from predicted and observed data were compared.
Results: This study successfully developed a PBPK model for RTV. The drug-dependent parameters, including molecular weight, pKa, and fraction unbound (0.6%) were fixed at the literature values. A high solubility of 100 mg/mL was fixed to ensure the free solubility of RTV in the commercial product. The lipophilicity (logP 3.44), specific intestinal permeability (1.01 E-7 cm/s), and organ permeability (1.68E-5 cm/s) were estimated by model identification steps. For enzymatic reactions, both Km and Vmax parameters for CYP3A5 and CYP2D6 reactions, and Km for the CYP3A4 reaction, were fixed at the literature values. Only the Vmax parameter (2.23 µM) for the CYP3A4 reaction was estimated. For the inhibition/induction effects, the parameters for the inhibition effects on Pgp and competitive inhibition effects on CYP3A4/5 were fixed at the literature values. Only parameters for the mechanism-based inhibition (Kinact 0.32 1/min; Kinact-half 0.10 µM) and induction effects on CYP3A4/5 (EC50 0.048 µM; Emax 153.9) were estimated to get the best fit. The developed PBPK model predicts the PK of RTV after the administration of RTV alone and RTV in DDI studies with all predicted AUC and Cmax ratios within 2-fold of the observed values. In addition, 80.5% of the predicted concentrations lie between the 2-fold border of the observed values.
Conclusions:
A PBPK model for RTV was successfully developed and evaluated in this present study. The model describes the PK profile of RTV and the DDI profiles of RTV with CYP3A/Pgp substrates well. This model could be applied to predict the DDI profiles of RTV with other CYP3A/Pgp transporters when co-administered. Accordingly, dose adjustment for individuals based on individual information could be performed.
References:
[1] Greenblatt DJ, Peters DE, Oleson LE, et al. Inhibition of oral midazolam clearance by boosting doses of ritonavir, and by 4,4-dimethyl-benziso-(2H)-selenazine (ALT-2074), an experimental catalytic mimic of glutathione oxidase. Br J Clin Pharmacol. 2009;68(6):920-927. doi:10.1111/j.1365-2125.2009.03545.x
[2] Ouellet* D, Hsu A, Qian J, et al. Effect of ritonavir on the pharmacokinetics of ethinyl oestradiol in healthy female volunteers. Br J Clin Pharmacol. 1998;46(2):111-116. doi:10.1046/j.1365-2125.1998.00749.x
[3] Kharasch ED, Bedynek PS, Walker A, Whittington D, Hoffer C. Mechanism of ritonavir changes in methadone pharmacokinetics and pharmacodynamics: II. Ritonavir effects on CYP3A and P-glycoprotein activities. Clin Pharmacol Ther. 2008;84(4):506-512. doi:10.1038/CLPT.2008.102
[4] Ghim J, Chin MC, Jung J, et al. Pharmacokinetics and Pharmacodynamics of Tegoprazan Coadministered With Amoxicillin and Clarithromycin in Healthy Subjects. J Clin Pharmacol. 2021;61(7):913-922. doi:10.1002/jcph.1805