Chloé Burlot1, François Riglet1, Antonin Schmitt2, Sylvain Fouliard1
1Quantitative Pharmacology, Translational Medecine, Servier, 2Centre Georges-François Leclerc – INSERM U1231, Université Bourgogne Europe
Objectives S65487 is an intravenously administered BCL2 inhibitor which underwent clinical development in hematological malignancies. Before clinical trials at therapeutic doses were initiated, a microdose study was performed on healthy volunteers [1]. Microdose studies assess subtherapeutic exposures in first-in-human trials, providing insights to guide early development [2]. Although S65487 exhibited linear PK at therapeutic doses, PK parameters estimated after microdose administration were found to substantially differ. This work aims to reconciliate the S65847 PK of both kind of exposure in a single PK model in order to further support extrapolation from microdose studies in the future. Material and Methods A total of 1738 and 84 PK samples of S65487 were collected from cancer patients in phaseI/II clinical trials and from healthy volunteers from microdose study, respectively. Therapeutic doses ranged from 25 to 1200 mg, while microdose was 0.1 mg. These data were used to develop a non-linear mixed effect popPK model, using Monolix with SAEM algorithm. First, a compartmental model was built, and ETA distribution were compared to have an initial idea of the magnitude of the PK differences. Several mechanistic or drug-specific hypotheses were evaluated to explain these PK differences between microdose and therapeutic doses, such as, subject health conditions, potential differences in organ function, CYP3A4 auto-inhibition or drug target interaction [3] . Then, hypotheses were incorporated in the popPK model, which was evaluated further by calculating population predicted area under the curve (pAUC). These pAUC were compared with observed AUC (oAUC calculated using non-compartmental approaches), to achieve a pAUC/oAUC ratio lower than with initial popPK model. Models were also evaluated regarding goodness of fit, parameters estimations and BICs. Results A 3-compartments popPK model with linear elimination and distribution allowed good description of all individual data. Yet significant differences in PK parameters distribution between microdose and therapeutic doses, particularly elimination and distribution were observed. Also, pAUC/oAUC ratio was 4.96 and 0.71 for microdose and therapeutic doses respectively. Among the various hypotheses tested, the one that takes into account nonlinearity in drug disposition improved the predicted microdose exposure. This would be related to binding interactions with drug target, and was identified as the most plausible explanation for differences between studies. This was modeled by incorporating two Michaelis-Menten equations to the popPK model. A saturable distribution replaced the linear distribution between the central and a peripheral compartment and a saturable elimination pathway was added alongside linear elimination. The final model incorporated two different combined error models in order to describe the two groups of doses, leading to an improved model. These modifications of the initial popPK model, resulted in a significant improvement of exposure predictions with a pAUC/oAUC ratio of 1.2 for microdose data, and 0.81 at therapeutic doses. Conclusions A simultaneous modelling approach allowed the reconciliation of microdose and therapeutic data, with plausible mechanistic explanation and also ruling out some hypotheses. Bridging-micro- and therapeutic doses by a unified popPK model can be useful for future drug candidates, in order to take the most from microdose studies. Although difficult to predict, possible non-linearities in drug disposition should be accounted for this framework.
[1] Y. Sugiyama et S. Yamashita, « Impact of microdosing clinical study — Why necessary and how useful? », Adv. Drug Deliv. Rev., vol. 63, no 7, p. 494-502, juin 2011, doi: 10.1016/j.addr.2010.09.010. [2] T. Burt et al., « Microdosing and Other Phase 0 Clinical Trials: Facilitating Translation in Drug Development », Clin. Transl. Sci., vol. 9, no 2, p. 74-88, avr. 2016, doi: 10.1111/cts.12390. [3] T. Burt et al., « Phase 0/microdosing approaches: time for mainstream application in drug development? », Nat. Rev. Drug Discov., vol. 19, no 11, p. 801-818, nov. 2020, doi: 10.1038/s41573-020-0080-x.
Reference: PAGE 33 (2025) Abstr 11358 [www.page-meeting.org/?abstract=11358]
Poster: Drug/Disease Modelling - Oncology