Optimizing Molecule Design for Bispecific Antibodies in Immuno-Oncology: A Case Study with FAP-4-1BBL
Javier Sanchez (1, 2), Christina Claus(3), Christine McIntyre(4), Tamara Tanos(1), Axel Boehnke(1), Lena E. Friberg(2), Siv Jönsson(2), Nicolas Frances(1)
(1) Roche Pharma Research and Early Development (pRED), Roche Innovation Center Basel, Basel, Switzerland. (2) Department of Pharmacy, Uppsala University, Uppsala, Sweden. (3) Roche Pharma Research and Early Development (pRED), Roche Innovation Center Zurich, Schlieren, Switzerland. (4) Roche Pharma Research and Early Development, Roche Innovation Center Welwyn, Welwyn Garden City, UK.
Objectives: FAP-4-1BBL is a T-cell bispecific antibody exerting 4-1BB costimulation only while simultaneously bound to the Fibroblast Activation Protein (FAP) expressed by fibroblasts of inflammatory tissues such as the tumor microenvironment [1]. The trimeric complex (TC) formed by FAP-4-1BBL bound simultaneously to FAP and to 4-1BB is expected to be the driver of FAP-4-1BBL pharmacological effect [2]. TC formation versus bispecific antibody concentration has been observed to follow a bell-shaped curve [3]. The number of TC that can be formed depends on the number of accessible receptors on the cell surface, the bispecific antibody concentration, and the binding affinities to both targets. Using in vitro data, translational modeling and simulation, and receptor expression data, we suggest an optimized molecule design (in terms of binding affinities) with potentially improved efficacy.
Methods: In a previously published model, [2], we have explored via simulations TC formation with FAP-4-1BBL as a function of dose and dosing schedule. We now further investigate the impact of the binding affinity on TC formation and expected pharmacology to identify a refined molecule design with potential to maximize benefit to the patient population.
PK and TC/tumor cell killing relationship were assumed equal to that of the original molecule. For each virtual molecule, the minimum number of FAP receptors on the fibroblast surface to exert 90% of the maximum tumor cell killing effect was calculated. Furthermore, the proportion of virtual patients achieving such levels of TC formation was derived, as well as the expected dose at which such TC formation is achieved. All simulations included inter-individual variability in PK parameters and FAP expression.
All model simulations and model fitting were performed in R 4.2.1 [4] using the rxode2 package [5]. The model developed on in vitro data was translated to the clinic using FAP expression values from different solid tumor types [6]. Human PK for FAP-4-1BBL was available from the literature [7], and plasma:tumor ratios were fixed to 2.2 or 10. The model-simulated average number of TCs on T cells formed over a 30-week dosing period, and its associated expected tumor cell killing, was considered as a surrogate for clinical efficacy.
Results: With the original FAP-4-1BBL molecule, we demonstrated that an average of 2.6 × 10^-2 trimeric complexes per T cell are required to reach 90% of the maximum tumor cell killing [2]. This would require a minimum of 13,400 FAP receptors per fibroblast, even at the exposure maximizing TC formation. In the current work, we demonstrate that an optimized molecule on the FAP binding affinity requires less target expression to form the same amount of TC. For instance, a molecule with a 10-fold increase in affinity would require as little as 4,000 FAP receptors per fibroblast to achieve 90% of the maximum effect. Though the dose maximizing TC formation for the high affinity molecule would be lower than that of the original molecule, TC formation at any dose is expected to be higher for the high-affinity molecule. At their respective TC-maximizing doses, the expected percentage of maximum tumor cell killing was found to be 90% for the enhanced affinity molecule, versus 75% for the original molecule.
Conclusions: We show how modeling and simulation can support optimal molecule design. Our results indicate that, for the case of bispecific antibodies in immuno-oncology, increased binding affinity on the target epitope could lead to increased response (higher efficacy and higher number of patients with benefit) in the patient population. In addition, increased binding affinity widens the range of exposures that generate the number of TCs needed to achieve 90% of the maximum benefit, which could in turn ease dose selection during clinical development. However, the affinity design with bispecifics needs to also consider off-tumor target expression to allow a therapeutic index.
References:
1. Claus, C. et al. Tumor-targeted 4-1BB agonists for combination with T cell bispecific
antibodies as off-the-shelf therapy. Sci Trans Med (2019)
2. Sanchez, J. et al. A model-based approach leveraging in vitro data to support dose
selection from the outset: A framework for bispecific antibodies in immuno-oncology.
CPT:PSP (2023)
3. Betts, A. et al. Mechanistic Quantitative Pharmacology Strategies for the Early Clinical
Development of Bispecific Antibodies in Oncology. Clin Pharmacol Ther. 2020
4. R Core Team. R: A Language and Environment for Statistical Computing. R
Foundation for Statistical Computing; 2022
5. Fidler M, et al. RxODE: Facilities for Simulating from ODE-Based Models. R package
version 2.0.13 ed2023
6. Zboralski, D. et al. Preclinical evaluation of FAP-2286 for fibroblast activation protein
targeted radionuclide imaging and therapy. Eur J Nuc Med Mol (2022)
7. Micallef S, et al. Using population pharmacokinetics to capture the binding of a novel
bispecific fibroblast activation protein (FAP) to its 4-1BB ligand (4-1BBL) and support
phase 2 dose selection in oncology patients. ACoP11; Virtual2020.