Lekshmi Dharmarajan, Andreas Weiss, Christian Schnell, Rowan Stringer, Ruben De Kanter, Victoria Head, Johannes Voshol, Simona Cotesta, Saskia M Brachmann, Jeffrey D Kearns
Novartis Institute of Biomedical Sciences; Basel, Switzerland and Cambridge MA USA
Objectives:
JDQ443 is an irreversible, covalent KRASG12C inhibitor that is currently being evaluated in clinical trials in patients with non-small cell lung and colorectal cancers [1]. Covalent inhibitors often show prolonged target inhibition despite their short half-life in the systemic circulation. This is observed as a decoupling of pharmacokinetics (PK) and target occupancy (TO) and is mechanistically driven by the separation of timescales between drug PK and the resynthesis rate of the target. Herein, we describe a modeling approach to assess preclinical JDQ443 PK and KRASG12C TO and the relationship to anti-cancer activity in murine xenograft models.
Methods:
The preclinical PK was modelled using a pop-PK approach in Monolix parameterized with data obtained following oral administration in tumor-bearing and non-tumor-bearing mice. The PK model was validated using data from an independent experiment where tumor-bearing mice were dosed using a micro-infusion pump. Next, the decoupled PK/TO for JDQ443 was characterized in multiple xenograft models in mice. The PK model parameters were fixed to the estimated population mean parameter estimates and an indirect response model was fitted to the in vivo TO data. To reconcile differences in TO observed in the different xenograft models, the model parameterization assumed a xenograft-specific target resynthesis rate, while the target binding parameters were assumed to be identical across all xenografts. Simulations of the model were then applied to characterize the relationship of PK and PK/TO with tumor growth inhibition data.
Results:
Results: Decoupled PK-TO was observed following oral administration of JDQ443 and was quantified using a mathematical model that accounts for varied KRASG12C resynthesis rates across multiple murine xenograft models. The modeling informed the design and interpretation of additional in vivo studies that varied the PK behavior of the drug, like daily exposure (or area-under-the-curve in a day (AUC24h)), time-over threshold, and maximal concentration using a micro-infusion pump and different dosing regimens. By integrating the PK-TO simulations with in vivo tumor growth inhibition data, we observed that the in vivo activity best correlated with TO and drug AUC24h.
Conclusions: The joint experimental and modeling analysis confirmed that TO is a relevant biomarker for the JDQ443 covalent inhibitor and this enabled the team to better understand how PK-TO relates to anti-cancer activity in murine xenograft studies. Additionally, a correlation between anti-cancer activity and the daily exposure (AUC24h ) was discovered and thereby highlights the potential utility of drug exposure as a more readily measurable marker in the clinic. Finally, we propose that our PK-TO-efficacy modeling approach is also generalizable to enable model-informed preclinical to clinical translation for other covalent compounds through the understanding of how PK relates to TO, which in turn drives anti-cancer activity in preclinical models.
Acknowledgments:
The authors would like to thank the broad team of Novartis associates involved with drug discovery and early clinical development of JDQ443, and specifically highlight contributors to [1].
References:
[1] Weiss A., et. al., “Discovery, Preclinical Characterization, and Early Clinical Activity of JDQ443, a Structurally Novel, Potent and Selective, Covalent Oral Inhibitor of KRASG12C.” Cancer Discovery (accepted for publication)
Reference: PAGE 30 (2022) Abstr 10227 [www.page-meeting.org/?abstract=10227]
Poster: Drug/Disease Modelling - Oncology