In this work, we present a mathematical model that described the bio-distribution of TDBs in preclinical solid tumor xenograft mouse models. This model is an extension of the minimal Physiological based pharmacokinetic model (mPBPK) developed by Cao et al (2). The model includes representation of the antibody distribution to and from the systemic circulation to up to two tumor compartments (to represent dual-flank tumors, T cell rich tissues of spleen and lymph nodes, representative leaky tissues (e.g. liver) and representative tight tissues (e.g. muscle). The model is capable of representing the presence/effect of the target in all of these tissues (each with tissue specific levels) and the presence/effect of CD3 in T cell rich tissues. Explicit binding kinetics of the antibody to both the target and CD3 is included to explore the impact of affinity to both target and CD3 on the bio-distribution to the different tissues. The model was developed based on bio-distribution data of a collection of antibodies from Mandikian et al. (1). These included molecules with a range of affinities to CD3 (0.05 – 50 nM) and antibodies binding HER2 as the target and non-target binding molecules. Further, data from different tumor xenograft models with differences in target expression levels was included to test the model robustness to variability in target levels. The radiochemistry of dual-tracer captures both intact antibody and catabolized antibody and both these data forms were included in the model to explore mechanisms that influence bio-distribution. Model simulations suggested that CD3 affinity had minimal influence on the bio-distribution to tumor but significantly affected the bio-distribution to secondary lymphatics including spleen and lymph nodes. Model simulations for radiolabeled dosed material with and without a CD3 blocking agent showed the higher exposure in T cell rich tissues due to increased CD3 binding for high affinity molecules with KD<5 nM. Non-target binding antibody simulation captures nonspecific uptake and catabolism of the molecules in the tumor and normal tissues. Model based sensitivity analyses was performed to explore the effect of target binding affinity, target expression levels and turnover and CD3 binding affinity on the bio-distribution of the molecules in the tumor tissues. These predictive simulations demonstrate how the target characteristics influence what the optimal antibody characteristics are to maximize the bio-distribution to the desired site of action. Further, the models serve as a platform to explore the impact of uncertainty and variability in the underlying biology on exposures; these analyses include effect of variability in T cells in the experimental mouse models, effect of tumor size, etc. Overall, the TDB bio-distribution platform serves as a robust tool for candidate molecule selection and design of optimal in vivo pharmacology experiments in mouse efficacy models to support translational efforts.
1) Mandikian et al., Relative Target Affinities of T-Cell-Dependent Bispecific Antibodies Determine Biodistribution in a Solid Tumor Mouse Model, Mol Cancer Ther. 2018 Apr; 17(4):776-785. 2) Cao et al., Second-generation minimal physiologically-based pharmacokinetic model for monoclonal antibodies, J Pharmacokinet Pharmacodyn. 2013 Oct; 40(5):597-607
Describe the effects of target and molecule characteristics on the bio-distribution of TDBs in preclinical animal models
Conduct in-silico research using the model to quantify the bio-distribution of TDBs, including predictions to evaluate effect of target and CD3 affinity, target levels and dynamics, and dose levels
Design optimal preclinical in-vivo pharmacology experiments to test activity and efficacy of TBD antibodies in appropriate tumor xenograft mouse models