Sandra Grañana-Castillo (1), Susanne Prothon (2), Ann Aurell Holmberg (3), Constanze Hilgendorf (4), Michael L Williams (2), David Boulton (5), Pradeep Sharma (6)
(1) Clinical Pharmacology and Quantitative Pharmacology, Clinical Pharmacology and Safety Sciences, R&D, AstraZeneca, Barcelona, Spain (2) Clinical Pharmacology and Quantitative Pharmacology, Clinical Pharmacology and Safety Sciences, R&D, AstraZeneca, Gothenburg, Sweden (3) DMPK, Early Respiratory & Immunology, R&D, AstraZeneca, Gothenburg, Sweden (4) DMPK, Early Cardiovascular, Renal & Metabolism, R&D, AstraZeneca, Gothenburg, Sweden (5) Clinical Pharmacology and Quantitative Pharmacology, Clinical Pharmacology and Safety Sciences, R&D, AstraZeneca, Gaithersburg, US (6) Clinical Pharmacology and Quantitative Pharmacology, Clinical Pharmacology and Safety Sciences, R&D, AstraZeneca, Cambridge, UK
Background and objectives: Nintedanib is a narrow therapy index drug and an essential co-medication to treat idiopathic pulmonary fibrosis (IPF). It is metabolised primarily by CES1 and to a minor extent by CYP3A4, and it is a substrate of the efflux transporter P-gp [1]. When given with strong inhibitors (e.g. ketoconazole), it can elicit a drug-drug interaction (DDI) with an AUC ratio (with vs without perpetrator) of 1.59 [2]. Therefore, the aim was to develop a nintedanib PBPK model to predict DDIs, which can be later used to assess the DDI risk with new drugs under development for IPF.
Methods: In this present work, a PBPK model for nintedanib was constructed in Simcyp (Version 23, Certara UK Ltd) by a middle-out approach using physicochemical properties, in vitro DMPK properties, and clinical PK data (hADME and DDI with itraconazole). Jmax and Km were measured in house and optimised using the SIVA tool (Version 5, Certara UK Ltd). The model was verified by simulating plasma concentration profiles and PK parameters (Cmax and AUC) at 50mg, 100mg, 150mg and 200mg nintedanib single dose. Model credibility was assessed by determining the geometric mean fold error (GMFE) for accuracy, and absolute average fold error (AAFE) and percentage prediction error (PPE) for precision of Cmax and AUC. The nintedanib PBPK model was applied to predict the DDI with rifampicin (CYP3A4 and P-gp inducer) and AZD5055 (P-gp inhibitor), mirroring the clinical study designs, and the simulations were compared with clinical data of these interactions ([2], NCT05644600).
Results: The model validation of nintedanib alone, and nintedanib with ketoconazole yielded an GMFE, AAFE, and PE of 0.82-fold, 1.24-fold, and 10% for Cmax, and 1.38, 1.39-fold, and 24% for AUC, respectively. In the rifampicin simulation, Cmax ratio and AUC ratio were predicted as 0.56 and 0.48, respectively (versus clinical study of Cmax ratio of 0.58 and AUC ratio of 0.49), capturing the CYP3A4 and P-gp based DDI magnitude. In the AZD5055 simulation, both Cmax ratio and AUC ratio for the 10mg AZD5055 dose were 1.00; and 1.01 for the 35mg dose, indicating a low risk of AZD5055 to elicit a significant DDI due to P-gp inhibition. This lack of interaction was similar to the clinical trial where the observed Cmax ratio and AUC ratio were 1.01 and 0.94, respectively for the DDI with AZD5055 10mg dose, and 1.11 and 1.03 for the DDI with the AZD5055 35mg dose. Nonetheless, multiple challenges were present in the development of this model. Nintedanib undergoes lysosomal trapping which delays its tmax and alters the distribution, and it is metabolised by CES1 and presents non-specific binding, which complicated the recovery of this drug in in vitro assays.
Conclusions: A PBPK model of nintedanib was developed which captured the potential risk of DDI due to CYP3A4 metabolism and P-gp and transport using a middle out approach. The model was applied to simulate the DDI with rifampicin and AZD5055 and compared to the corresponding clinical trials, which increased the confidence on the model predictions. As nintedanib is one of the most used drugs to treat IPF, the validated nintedanib PBPK model for could be impactful in the assessment of DDI risk with IPF compounds in clinical development.
References:
[1] Wind S, Schmid U, Freiwald M, Marzin K, Lotz R, Ebner T, Stopfer P, Dallinger C. Clinical Pharmacokinetics and Pharmacodynamics of Nintedanib. Clin Pharmacokinet. 2019 Sep;58(9):1131-1147. doi: 10.1007/s40262-019-00766-0. PMID: 31016670; PMCID: PMC6719436.
[2] Luedtke D, Marzin K, Jungnik A, von Wangenheim U, Dallinger C. Effects of Ketoconazole and Rifampicin on the Pharmacokinetics of Nintedanib in Healthy Subjects. Eur J Drug Metab Pharmacokinet. 2018 Oct;43(5):533-541. doi: 10.1007/s13318-018-0467-9. PMID: 29500603; PMCID: PMC6133080.
Reference: PAGE 32 (2024) Abstr 11050 [www.page-meeting.org/?abstract=11050]
Poster: Drug/Disease Modelling - Absorption & PBPK