Ioannis Loisios-Konstantinidis (1), Alexandros Kourentas (2), Felix Huth (1), Heidi Einolf (3), Matthias Hoch (1)
(1) Pharmacokinetic Sciences, Translational Medicine, Biomedical Research, Novartis, Basel, Switzerland; (2) Analytical Research and Development, Development, Novartis, Basel, Switzerland; (3) Pharmacokinetic Sciences, Translational Medicine, Biomedical Research, Novartis, New Jersey, USA
Objectives:
In a clinical study, co-administration of multiple doses of an itraconazole oral solution containing hydroxypropyl-beta-cyclodextrin (HP-beta-CD) at a total of 8000 mg per dose with a 40 mg dose of asciminib decreased asciminib exposure (expressed as AUCinf) by 40.2% in healthy subjects [1]. However, other cyclodextrins than HP-beta-CD are used as excipients in medicinal use and include alpha-CD, beta-CD, gamma-CD, sulfobutylether (SBE)-beta-CD and randomly methylated-beta-CD [2[. In a post-marketing requirement FDA and EMA requested a PBPK modeling analysis to assess the clinical relevance of all cyclodextrins in medicinal use on the exposure of asciminib. The objective of this study was to establish a cutoff for the amount of each cyclodextrin (CD) that would not reduce asciminib exposure by more than 20% (i.e., based on the bioequivalence criteria) using an in vitro-in silico-in vivo extrapolation approach.
Methods:
For this purpose, at first the affinity of asciminib to cyclodextrin was quantified in vitro by determining binding constants (KD) for HP-beta-CD, SBE-beta-CD, alpha-, beta- and gamma-CD (food-grade), when possible based on their solubility in fasted state simulated intestinal fluid (FaSSIF). In addition, the in vitro permeability of asciminib was determined in a MDCK-based transwell experiment with increasing concentrations of the respective cyclodextrins, (i.e., alpha-CD, beta-CD, gamma-CD, HP‑beta-CD, and SBE-beta-CD) being added to the donor compartment together with 10 µM of asciminib free base. A nonlinear equation was fitted to the data in order to obtain the quantitative relationship between the apparent permeability of asciminib and the concentration of each cyclodextrin.
Once the in vitro data were generated, three different PBPK modeling approaches were performed to evaluate the effect of the respective cyclodextrin on the exposure of asciminib at different dose levels. In the first approach, the in vitro binding constants were used directly within the mechanistic excipient binding model of the Simcyp version 21.1.. In the second approach, the theoretical equations by Dahan et al (2010)[3] assuming quasi-equilibrium and a 1:1 stoichiometry between drug and CD were used in combination with the in vitro binding constants to calculate the free fraction of asciminib in each segment of the gastrointestinal tract (GIT) at different doses of the respective cyclodextrin. Based on the free fraction hypothesis, the effective permeability of asciminib was corrected in a linear fashion. In the third approach, the relationship between experimentally measured apparent permeability and the concentration of the respective cyclodextrin was used to calculate the apparent permeability of asciminib in each segment of the GIT at different doses of the respective cyclodextrin and then the apparent permeability was converted to the effective permeability using a correlation curve internally developed by Novartis. The regional effective permeability values which were calculated using either the second or third approach were then used as input in Simcyp to simulate the effect of different doses of the respective cyclodextrin on the asciminib exposure.
Results:
The available clinical data of the interaction of asciminib with the itraconazole solution containing 8,000 mg HP-beta-CD were used to confirm the PBPK model performance[1]. The PBPK model was able to describe the observed PK parameters of asciminib with or without co-administration of itraconazole/cyclodextrin solution with prediction errors less than 28%. Furthermore, the PBPK model was also able to capture the Cmax and AUCinf ratios of the interaction with the itraconazole/cyclodextrin solution within 20% of the observed values. The verified PBPK model was applied to predict the effect of a wide range of dose of different cyclodextrins (i.e., alpha-, beta-, gamma-, HP-beta and SBE-beta-CD).
Conclusions:
Overall, the PBPK model predicted that alpha-, beta- or gamma-CD is not expected to have an effect on the PK of asciminib at their maximum soluble dose or dose available in oral products (including food and drinks). For HP-beta and SBE-beta cyclodextrin a maximum dose of 2,000 mg was predicted to not have a relevant effect (i.e., less than 20% decrease) on the exposure of asciminib. Both FDA and EMA, accepted the results of the PBPK modeling analysis without the need to conduct any clinical study or additional in vitro experiments.
References:
[1] M. Hoch et al., ‘Pharmacokinetics of asciminib in the presence of CYP3A or P-gp inhibitors, CYP3A inducers, and acid-reducing agents’, Clin Transl Sci, vol. 15, no. 7, pp. 1698–1712, Jul. 2022, doi: 10.1111/cts.13285.
[2] Gonzalez Pereira A, Carpena M, García Oliveira P, et al (2021) Main Applications of Cyclodextrins in the Food Industry as the Compounds of Choice to Form Host-Guest Complexes. Int J Mol Sci; 22(3):1339.
[3] Dahan A, Miller JM, Hoffman A, et al (2010) The Solubility–Permeability Interplay in Using Cyclodextrins as Pharmaceutical Solubilizers: Mechanistic Modeling and Application to Progesterone. J Pharm Sci; 99(6):2739-49.
Reference: PAGE 32 (2024) Abstr 11058 [www.page-meeting.org/?abstract=11058]
Poster: Drug/Disease Modelling - Absorption & PBPK