Pieter Van Brantegem 1, Xavier Woot de Trixhe 1, Huybrecht T'jollyn 1, Jeroen Verhoeven 1, Juan Jose Perez Ruixo 1, An Vermeulen 1
1 Johnson & Johnson (Beerse, Belgium)
Introduction: Auto-antibodies produced by B cells drive autoimmune diseases such as rheumatoid arthritis (1). B cell depletion aiming to restore normal B cell functioning has demonstrated clinical efficacy. T cell engagers create an immune synapse, linking CD20 receptors on B cells and CD3 receptors on T cells thereby forming the TBE complex that induces potent B cell killing (2), and potentially enables durable, drug-free remission.
Objectives: To quantify the influence of model parameters on outcomes under the mechanism evaluated.
Methods: We developed a mechanistic PK/PD model exploring the relationship between in vitro measured TBE kinetics at the receptor level and B cell depletion observed in patients. B and T cell kinetics were described by turnover models. TBE‑mediated B cell killing followed a sigmoidal Emax relationship. T cells were activated by the TBE complex concentration and regenerated using first-order kinetics. CD20 and CD3 synthesis and degradation were modeled explicitly, with drug‑bound targets assumed to internalize. A scaling factor was incorporated, linking the receptor-level and cell-level models by representing the hypothetical microenvironment volume per B and T cell where TBE complex formation occurs (ξ_B and ξ_T, respectively).
Digitized PK and B cell data derived from the mosunetuzumab Phase 1 first‑in‑human trial, presented at the 2025 EULAR conference was used to fit the model (3). T cell data was not available. A population PK and K-PD models were fitted using NONMEM to extract the necessary parameters. The lower limit of quantification for B cells in blood was 0.4 cells/µL. A Sobol global sensitivity analysis was conducted using the sensobol package in R to quantify the influence of model parameters on the outcomes under the mechanism evaluated (4). Evaluated outcomes included the peripheral B cell depletion AUC over 365 days, time to nadir, nadir depth, duration of depletion below 0.4 cells/µL, and time to 50% recovery. Only PD parameters not representing drug properties, that are relevant to decision making, and the potential to justify additional in vitro work were considered. All parameters were assumed to be log normally distributed with 30% CV. Parameter matrices were generated using quasi random sampling for 100,000 simulations and total order Sobol’ indices were computed. The sensitivity analysis was conducted using two dosing scenarios: (1) a single 5 mg SC dose, representing the highest single dose, and (2) a combination regimen consisting of a 5 mg SC dose followed one week later by a 60 mg SC dose, representing the highest administered dosing scenario.
Results: Across outcomes, the top three most influential parameters in terms of total‑order Sobol’ index were as follows. For the peripheral B cell depletion AUC over 365 days, the influential parameters for both dosing scenarios were ξ_T, CD3 expression and baseline B cell death rate. For time to nadir, the dominant parameters for the single SC dosing scenario were baseline T cell concentration, CD3 expression and B cell killing EC₅₀, while for the 5 mg + 60 mg dosing scenario baseline T cell concentration, B cell killing EC₅₀ and TBE‑mediated killing rate were most influential. For nadir depth, the top parameters at 5 mg were ξ_T, CD3 expression and ξ_B, whereas at 5 mg + 60 mg the ranking shifted to CD20 expression, ξ_T and ξ_B. For the duration of depletion below 0.4 cells/µL, the influential parameters at 5 mg and 5mg + 60 mg were ξ_T, CD3 expression and TBE‑mediated killing rate. For the time to 50% recovery, the top three parameters at 5 mg were ξ_T, CD3 expression and baseline B cell death rate, and at 5 mg + 60 mg CD3 expression, ξ_T and baseline B cell death rate.
Conclusion: ξ_T and CD3 expression consistently drive B cell depletion, supporting the need for a mechanistic model that explicitly includes T‑cell dynamics, even though no T cell data was available. Future work includes conducting a structural sensitivity analysis by altering the T‑cell component of the model and re-evaluating the sensitivity outcomes under each structural scenario.
References:
1. Albayrak G, Wan PKT, Fisher K, Seymour LW. T cell engagers: expanding horizons in oncology and beyond. Br J Cancer. 23 juli 2025. doi:10.1038/s41416-025-03125-y PubMed PMID: 40702106.
2. Cech P, Skórka K, Dziki L, Giannopoulos K. T-Cell Engagers-The Structure and Functional Principle and Application in Hematological Malignancies. Cancers (Basel). 20 april 2024;16(8):1580. doi:10.3390/cancers16081580 PubMed PMID: 38672662; PubMed Central PMCID: PMC11048836.
3. Chindalore VL, Martins E, Pendergraft III WF, Sheng XR, Schroeder A, Gearhart L, e.a. A Phase 1b, multicentre, open-label, dose-escalation study to evaluate the safety, tolerability, pharmacokinetics and pharmacodynamics of subcutaneously administered mosunetuzumab in participants with systemic lupus erythematosus. In. Barcelona, Spain [Poster]: EULAR 2025; 2025.
4. Puy A, Piano SL, Saltelli A, Levin SA. sensobol: An R Package to Compute Variance-Based Sensitivity Indices. Journal of Statistical Software. 30 april 2022;102:1-37. doi:10.18637/jss.v102.i05
Reference: PAGE 34 (2026) Abstr 12258 [www.page-meeting.org/?abstract=12258]
Poster: Drug/Disease Modelling - Other Topics