Miné de Kock (1), Chang Gong (2), Steven Eck (3), Xiaoru Chen (3), Karla Maloveste (3), Enow Enowtambong (3), Rachel Schiffenbauer (3), Kristen Pollizzi (4), Eleanor Clancy-Thompson (4), James Moynihan (4), Robin Lesley (5), Anna Hansen (6), Teresa Collins (1)
(1) AstraZeneca, Clinical & Quantitative Pharmacology, Cambridge, United Kingdom, (2) AstraZeneca, Systems medicine, Waltham, MA, United States of America, (3) AstraZeneca, Bioanalytics, Gaithersburg, MD, United States of America, (4) AstraZeneca, Immuno Oncology, Gaithersburg, MD, United States of America, (5) AstraZeneca, Translational Medicine, South San Francisco, CA, United States of America, (6) AstraZeneca, Translational Medicine, Gaithersburg, MD, United States of America
Objectives: Bispecific antibodies, targeting two distinct antigens simultaneously, are an emerging biologic format. A good understanding of the pharmacokinetic (PK) and pharmacodynamic (PD) relationships for both targets is required to optimise the dose for a given indication1. Sabestomig, a monovalent, bispecific, humanised IgG1 antibody targeting PD-1 and TIM-3 was evaluated for safety and efficacy in Phase 1 dose escalation studies in solid and lymphatic tumour indications. The goal was to implement one model for both solid and lymphatic tumour indications to describe the non-linear plasma PK of sabestomig, describe the total concentration kinetics of a peripheral soluble receptor, and predict peripheral and tumoral receptor occupancy to enable the selection of the best dose for the intended indications.
Methods: A semi-mechanistic PK model like the one implemented by Cao et al2 was applied to gain insight into sabestomig concentration kinetics in the central, lymph and solid tumour compartments, and to fit to plasma PK sample data for 25 lymphatic and 53 solid tumour patients. Target-mediated drug disposition was included at the site of the tumour for each indication (lymph and solid tumour compartments) with consideration for the receptor-expressing phenotypes. Total concentration data for the lower-affinity soluble receptor (sTIM-3) in the periphery was described by a sub-model that included target-mediated drug disposition for both the membrane-bound and soluble receptors in the central compartment. The model was used to simulate intra-tumoural target engagement predictions for both targets (PD-1 and TIM-3) on single and dual-expressing cells, as well as steady-state exposure metrics for each individual in exposure response and exposure safety analysis.
Results: The model described the plasma PK data and the peripheral soluble target exposure in the clinical dose range for both the solid and lymphatic tumour indications. Dose selection was supported by the exposure range that resulted in 1. saturated target-mediated drug disposition and linear peripheral clearance, 2. saturated peripheral soluble target concentrations, and 3. greater than 90% simulated target engagement at the tumour site for both targets over the dosing interval. Steady state Cmin below 12.2 mg/L resulted in target engagement below 90% based on the simulated intra-tumoural target engagement predictions for both targets on single and dual-expressing cells. The simulated exposure metrics, AUC0-21days, Cmin, and Cmax, were incorporated into exposure response and exposure safety analysis to inform dose selection. The lymphatic tumour indication had 17, 33, 17, and 33% serious adverse events in the ascending quartiles of simulated AUC0-21days with a range of 2-3951 mg.day/L. Efficacy of 28.5% was observed in the simulated steady state Cmin range of 0 to 196.9 mg/L.
Conclusions: The development of a semi-mechanistic PK model for both a solid and lymphatic tumour indications, incorporating target-mediated drug disposition at the respective tumour sites and peripheral soluble target kinetics, and the resulting intra-tumoural target engagement and PK exposure predictions informed a dose selection decision following a dose-escalation study.
References:
[1] Gibbs, J.P., Yuraszeck, T., Biesdorf, C., Xu, Y. and Kasichayanula, S. Informing Development of Bispecific Antibodies Using Physiologically Based Pharmacokinetic-Pharmacodynamic Models: Current Capabilities and Future Opportunities. The Journal of Clinical Pharmacology. 2020. 60: S132-S146. https://doi.org/10.1002/jcph.1706.
[2] Cao Y, Balthasar JP, Jusko WJ. Second-generation minimal physiologically-based pharmacokinetic model for monoclonal antibodies. J Pharmacokinet Pharmacodyn. 2013. 40(5):597-607.
Reference: PAGE 32 (2024) Abstr 11080 [www.page-meeting.org/?abstract=11080]
Poster: Drug/Disease Modelling - Oncology