Sergei Vavilov1, Elin Boger1, Joel Zirkle1, Mounier Almett1, Nina Lawrence1, Lassina Badolo1
1Drug Metabolism and Pharmacokinetics, Respiratory & Inflammation, BioPharmaceuticals R&D, AstraZeneca
Introduction: In early drug discovery, clinical data on monoclonal antibodies targeting a specific protein ligand may be unavailable before the phase 1 trials. Instead, models of ligand turnover can be calibrated based on available clinical data for antibodies that target either receptors or other proteins from the same signalling pathway. This approach allows one to predict ligand concentration, receptor occupancy and signalling suppression early in the drug development pipeline, guiding antibody engineering and preclinical studies. The model can be augmented to account for soluble receptors and downstream ligand-receptor interactions. Methods: A synthetic PKPD dataset for receptor-targeting monoclonal antibody was assembled. Assuming typical PK parameters of an antibody from [1], typical SPR measurements for binding kinetics from [2] and pre-set values for ligand and receptor synthesis rates, we generated synthetic time profiles (n = 8 time points) for receptor-targeting antibody concentration, receptor occupancy, total soluble receptor concentration and total ligand concentration. In the second step, the synthetic dataset was used to estimate the synthesis rates for ligand and receptor, which could then be validated against the original values. The mathematical model was implemented and simulated in MATLAB R2023b using ode15s as a solver and lsqnonlin as a parameter estimation tool. Results: An ODE-based model that accounts for synthesis and degradation of a ligand, membrane-bound receptor and soluble-bound receptor was successfully developed. The model was calibrated based on a synthetic dataset comprised of receptor-targeting antibody concentration, receptor occupancy, total ligand concentration and total soluble receptor concentration. Parameter estimation based on these data allows one to correctly recapture the ligand synthesis rate (with CV = 0.03%), the soluble receptor synthesis rate (with CV = 0.002%), and the membrane-bound receptor synthesis rate (with CV = 0.26%). Conclusion: The model development and calibration were based on a dataset for receptor-targeting antibody, and therefore, model-based simulations of ligand turnover, suppression, or accumulation can serve as an early-stage guideline for target validation, antibody engineering, or preclinical study design in the process of developing a ligand-targeting antibody. Preclinical in vivo studies of protein ligands only determine the total plasma concentration of a ligand routinely; in this case, a model-based approach can be useful to extract baseline free ligand concentration and baseline free receptor concentration values that would otherwise be obscured. For ligands that have soluble receptors acting as a potential drug sink, this model-based approach allows one to identify baseline free soluble receptor concentration separately. In this data-driven approach, the uncertainty of parameter estimates depends on the granularity of time profiles. Better parameter estimates can come from more frequent sampling, from data on additional downstream protein ligands and receptors that form the same signalling pathway, and from data on changes in protein ligand concentrations caused by disease progression. In the last case, modelling ligand concentration changes under antibody treatment can be used as a guide for disease-modifying properties of a drug and for dose setting in preclinical and clinical studies.
[1] Betts A, Keunecke A, van Steeg TJ, van der Graaf PH, Avery LB, Jones H, Berkhout J. Linear pharmacokinetic parameters for monoclonal antibodies are similar within a species and across different pharmacological targets: A comparison between human, cynomolgus monkey and hFcRn Tg32 transgenic mouse using a population-modeling approach. MAbs. 2018 Jul;10(5):751-764. doi: 10.1080/19420862.2018.1462429. Epub 2018 May 14. PMID: 29634430; PMCID: PMC6150614. [2] Hearty S, Leonard P, O’Kennedy R. Measuring antibody-antigen binding kinetics using surface plasmon resonance. Methods Mol Biol. 2012;907:411-42. doi: 10.1007/978-1-61779-974-7_24. PMID: 22907366.
Reference: PAGE 33 (2025) Abstr 11729 [www.page-meeting.org/?abstract=11729]
Poster: Methodology - New Modelling Approaches