Physiologically-based pharmacokinetic modeling to support submission of brigimadlin (BI 907828) in dedifferentiated liposarcoma
Tobias Kanacher1, Dr. David Busse4, Dr. Reinhard Sailer3, Dr. Erik Sjögren1, Dr. Enrica Mezzalana1, Dr. Moriah Pellowe1, Dr. Marylore Chenel1, Dr. Jose David Gómez-Mantilla4, Dr. Ibrahim Ince4
1Pharmetheus AB, 2Boehringer Ingelheim Pharma Inc., 3Boehringer Ingelheim Pharma, 4Boehringer Ingelheim Pharma
Introduction: Brigimadlin is a highly potent, oral MDM2-p53 antagonist [1] for the treatment of cancer patients. After single dose administration, brigimadlin shows a long terminal half-life of ~40 h and high variability in time to maximum plasma concentration (tmax ) 4 h (2-24h). In vitro studies indicate minor metabolism via CYP3A4 and UGT1A3, while rat studies suggest biliary excretion as a major clearance route. Brigimadlin is a substrate and inhibitor of OATP1B1/1B3 (IC50: 0.70/0.66 µM) and potential inducer of CYP1A2/CYP3A4. A drug-drug interaction (DDI) study with rifampicin was conducted to assess OATP1B-mediated effects, although potential CYP3A4 induction complicated read-out interpretations. A physiologically based pharmacokinetic model (PBPK) model for brigimadlin was developed to investigate the complex transport-enzyme interplay and support further DDI investigations in patients with advanced solid tumors receiving brigimadlin. Objectives: This analysis aimed to inform brigimadlin’s regulatory submission by: 1.)Developing a PBPK model using available preclinical and clinical data. 2.)Evaluating brigimadlin’s PK and its potential DDIs as both a victim and perpetrator of OATP1B1, OATP1B3, and CYP3A4 interactions. Methods: A PBPK model for Brigimadlin was developed using the Open Systems Pharmacology Software Suite (PK-Sim® and MoBi®, v11.2). The modeling process followed a stepwise approach: 1.)Structural base PBPK model Development: Physicochemical, in vitro, and preclinical data informed brigimadlin’s disposition (distribution, metabolism, and elimination) following intravenous (IV) and oral (PO) administration. Brigimadlin elimination was modeled via CYP3A4 and UGT1A3 metabolism, OATP1B1/-B3 hepatic uptake, and passive glomerular filtration. 2.)Model Optimization & Qualification: The PBPK model performance was evaluated and refined using clinical data, including a rifampicin DDI study and human ADME study. External qualification was performed using independent clinical data (including a midazolam DDI subgroup). Brigimadlin’s PBPK model was linked with published PBPK models for rifampicin [2] and midazolam [3] to simulate DDI scenarios chosen for model qualification. Predicted vs. observed plasma profiles, area under the plasma concentration-time curve (AUClast), and maximum plasma concentration (Cmax) values were compared for qualification. 3.)Model Application The qualified model was then used to simulate prospective DDI scenarios with brigimadlin as victim or perpetrator by linking it with published PBPK models for itraconazole [4], clarithromycin [5] and ethinylestradiol [6]. Simulations were performed mimicking the trial design and trial population of each reference study. Results: A middle-out approach was used, incorporating clinical data to refine key parameters such as tissue distribution (logP), intestinal permeability, OATP1B1/-B3 kinetics, glomerular filtration rate (GFR) fraction, and CYP3A4/UGT1A3 metabolism. The model accurately captured brigimadlin’s concentration-time profiles and interactions with rifampicin. External qualification (5-80 mg) showed strong predictive performance, with an absolute average fold error (AAFE) of 1.22 for AUC (AFD 1.02-1.62) and 1.22 for Cmax (AFD 1.05-2.11) being well below a 2 fold prediction range. DDI trial predictions (10 mg Brigimadlin ± rifampicin) were accurate, with AAFE values of 1.15 for AUC and 1.12 for Cmax. DDI Predictions: Brigimadlin showed a low DDI risk as a CYP3A4 victim (=15% metabolism via CYP3A4), with minor exposure changes when co-administered with CYP3A4 inhibitor itraconazole (+11%), CYP3A4 inducer rifampicin (-6%), or CYP3A4-OATP1B1/-B3 inhibitor clarithromycin. As a CYP3A4 perpetrator, brigimadlin had minimal effects on midazolam (-24%) and ethinylestradiol (-2%) exposure. Conclusions: A PBPK model for brigimadlin was successfully developed and applied for DDI investigation. The model accurately predicted brigimadlin disposition across various dosing regimens and potential DDIs, including CYP3A4/OATP1B1/-B3 interactions. Overall, the analysis suggests brigimadlin is unlikely to significantly affect CYP3A4 substrates or be strongly impacted by CYP3A4 induction/inhibition or OATP1B1/-B3 inhibition. This study highlights the applicability of PBPK modeling for analysis and predictions of complex interplay of induction/inhibition on enzymatic and transporter-mediated processes that cannot be achieved with less mechanistic modeling approaches.
[1] LoRusso P et al. Cancer Discov 2023 ; 13:1802–13. [2] Lippert J et al. CPT pharmacometrics Syst. Pharmacol. (2019) 8, 878–882. [3] https://github.com/Open-Systems-Pharmacology/OSP-PBPK-Model-Library/tree/master/Rifampicin [4] https://github.com/Open-Systems-Pharmacology/OSP-PBPK-Model-Library/tree/master/Midazolam [5] https://github.com/Open-Systems-Pharmacology/OSP-PBPK-Model-Library/tree/master/Itraconazole [6] https://github.com/Open-Systems-Pharmacology/OSP-PBPK-Model-Library/tree/master/Clarithromycin [7] https://github.com/Open-Systems-Pharmacology/OSP-PBPK-Model-Library/tree/master/Ethinylestradiol