(W-109) Comparative Analysis of Minimal Physiologically Based Pharmacokinetic (mPBPK) and Traditional Pharmacokinetic (PK) Models in the Context of Quantitative Systems Pharmacology (QSP) for T-cell Engagers

Cesar Pichardo, PhD – Senior Director, Systems Medicine, AstraZeneca; Meghna Verma, PhD – Associate Director, Systems Medicine, AstraZeneca; Massimo Lai, PhD – Associate Director, Systems Medicine, AstraZeneca; Holly Kimko, PhD – Executive Director, Systems Medicine, AstraZeneca

Objective: The development of T-cell engager (TCE) therapeutics necessitates advanced pharmacokinetic (PK) modeling techniques to accurately predict drug behavior and optimize dose strategies. Quantitative Systems Pharmacology (QSP) models for T-cell engagers (TCEs) represent a sophisticated approach to understanding the complex interactions between drug compounds, T-cells, and tumor cells, as well as the broader immune system. Traditional PK models and minimal Physiologically Based Pharmacokinetic (mPBPK) models represent two distinct approaches, each with its advantages and limitations. The study aims to compare these systems models in predicting antibody concentration of TCEs.

Methods: We implemented QSP models incorporating both traditional PK model [1] and mPBPK models [2,3,4] to simulate the dynamics of TCEs. The QSP models are based on ordinary differential equations (ODE) and built using MATLAB/SimBiology [5]. The traditional PK model was based on compartmental analysis, while the mPBPK model included simplified physiological parameters. The work presented here is focused on comparing PK profiles with different systems models using CD3xBCMA T-cell engagers [2]. The models were evaluated against clinical PK data [2] for validation.

Results: The mPBPK model provided more detailed predictions of tissue-specific drug concentrations and demonstrated superior accuracy in capturing the complex pharmacokinetics of TCEs, particularly in scenarios involving significant physiological variability among patients. However, the traditional PK model offered greater simplicity leading to a faster computational implementation, making it suitable for early-stage drug development and preliminary analyses. This comparative analysis helps us understand what type of model may be useful to predict drug concentration in mice to human and their drug effects on immune response dynamics and inform the choice of appropriate dose regimens in patients.

Conclusions: While the mPBPK model within a QSP framework offers a more detailed and physiologically relevant prediction of TCE pharmacokinetics, traditional PK models remain valuable for their simplicity and computational efficiency. The choice between models should be guided by the specific requirements of the study, including the required level of physiological detail, the type and availability of experimental data, and/or the need to estimate the compound concentration in specific tissues to evaluate target engagement or safety. Future research should focus on refining these models to enhance their predictive power and utility in the development of TCE therapeutics.

Citations: Citations:
[1] Weddell J. Mechanistically modeling peripheral cytokine dynamics following bispecific dosing in solid tumors. CPT Pharmacometrics Syst Pharmacol. 2023 Nov;12(11):1726-1737
[2] Yoneyama, Tomoki, et al. "Leveraging a physiologically-based quantitative translational modeling platform for designing B cell maturation antigen-targeting bispecific T cell engagers for treatment of multiple myeloma." PLoS Computational Biology 18.7 (2022): e1009715.
[3] Cao, Yanguang, Joseph P. Balthasar, and William J. Jusko. "Second-generation minimal physiologically-based pharmacokinetic model for monoclonal antibodies." Journal of pharmacokinetics and pharmacodynamics 40 (2013): 597-607.
[4] Jiang, Xiling, et al. "Development of a minimal physiologically-based pharmacokinetic/pharmacodynamic model to characterize target cell depletion and cytokine release for T cell-redirecting bispecific agents in humans." European Journal of Pharmaceutical Sciences 146 (2020): 105260.
[5] https://www.mathworks.com/products/matlab.html