(T1430-04-22) Novel PBPK-IVIVC Model to Predict Human Pharmacokinetics of Ruxolitinib Sustained-Release Formulations

Purpose: In vitro-in vivo correlation (IVIVC) allows prediction of the in vivo pharmacokinetics (PK) of sustained-release (SR) formulations from the in vitro dissolution patterns by correlating the in vitro properties of a dosage form and the in vivo response. This study aimed to develop a novel IVIVC model combined with physiologically based pharmacokinetics (PBPK), which can predict the pharmacokinetics of SR formulations in humans via interspecies scaling from the in vitro dissolution and in vivo animal data.

Methods: Three types of SR tablets containing the model drug ruxolitinib with different drug release rates were prepared using the wet granulation method. The in vitro dissolutions of the prepared SR tablets were assessed via the USP Apparatus 2 method. In vivo pharmacokinetic data were obtained following intravenous (IV) injection (4 mg/kg, n=4) in rats and following IV injection (1 mg/kg, n=4) and oral administration of three SR tablets (25 mg, n=3) in beagle dogs. Human PK data following oral administration of a ruxolitinib SR tablet was obtained from the literature (1). Based on the PK data from dogs, a population PK model was developed to predict the in vivo dissolution of SR tablets. IVIVC was then established by correlating the in vitro and in vivo dissolution parameters. A PBPK model for ruxolitinib was first developed in rats and extended to beagle dogs and humans. Finally, the IVIVC model was incorporated into the PBPK model to describe the dissolution process following oral administration and to predict the pharmacokinetics of SR tablets in humans.

Results: The time for 80% in vitro dissolution (T₈₀) was 1.5 h, 7.0 h, and 11.5 h for the SR tablet with fast, medium, and slow dissolution rates, respectively. Following IV injection, plasma ruxolitinib concentrations rapidly declined with an average half-life (t_1/2) of 0.38 h in rats and 0.94 h in dogs. The average oral bioavailability of ruxolitinib SR tablets ranged from 27.81 to 35.54% in dogs. The Michaelis-Menten equation adequately described drug release from the SR tablet. IVIVC was established by correlating the experimentally determined in vitro dissolution parameter, V_{max, in vitro}, and a PK model-estimated in vivo dissolution parameter, V_{max, in vivo}, for three SR tablets in beagle dogs. The PBPK model involving 11 tissues could appropriately describe the plasma concentration vs. time profiles of rats and dogs. The PBPK-IVIVC model that applied the IVIVC equation to the PBPK model predicted the maximum plasma concentration (C_max) and the area under the plasma concentration-time curve (AUC) of ruxolitinib following oral administration of SR tablets with the average prediction error (PE%) of 5.65% and 8.15%, respectively, in dogs. Finally, the PBPK-IVIVC model could predict the time course of plasma concentration following oral administration of ruxolitinib SR tablet (25 mg) in humans. Compared to the observed data (1), the average PE% of the final PBPK-IVIVC model for human PK prediction was 14.15% for C_max and 1.89% for AUC.

Conclusion: A novel PBPK-IVIVC model was developed by combining the IVIVC and PBPK models, which allows prediction of the time course of plasma concentration following oral administration of SR tablets in humans from its in vitro dissolution and animal data. The present model-based approach would provide a valuable tool to predict the clinical pharmacokinetics of new dosage forms for their efficient design and development.

References: 1. Yong Ni, Bhavnish Parikh, Krishnaswamy Yeleswaram, Susan Erickson-Viitanen, William V. Williams (2019). U.S. Patent No. 10,166,191. Incyte Corporation, Wilmington, DE.

Acknowledgements: This work was supported by the National Research Foundation of Korea(NRF) grant funded by the Korean government (MSIT) (Nos. 2021R1A2B5B01001970 and NRF-2023R1A2C1007013).