(T1430-02-10) A Minimal PBPK-PD Model to Predict PK, PD, and Efficacy of Anti-CCR8 Antibody in Mice

Purpose: C-C motif chemokine receptor 8 (CCR8) is selectively expressed on regulatory T (Treg) cells, a major immunosuppressive cell type in tumors.1 Targeting and depleting intratumoral Treg cells represents a therapeutic opportunity for treating solid tumors, as it can potentially reverse the suppression of CD8+ effector T-cells and enhance anti-tumor immunity in combination with other immunotherapies. Anti-CCR8 antibody is a mouse surrogate molecule that binds to murine CCR8 and is designed to preferentially eliminate CCR8+ Treg cells in the tumor microenvironment through ADCC-mediated depletion. To investigate the mechanism of action (MoA) for anti-CCR8, we built a minimal pharmacologically-based pharmacokinetic-pharmacodynamic (minPBPK-PD) model2 to assess the pharmacokinetics (PK), pharmacodynamics (PD), and efficacy in a mouse model.

Methods: We built a minimal PBPK-PD model to simulate the PK/receptor occupancy (RO) relationship as well as PD and efficacy of the mouse surrogate anti-CCR8 antibody. The model incorporates five key elements to capture pharmacologic activity of the anti-CCR8 antibody: (1) anti-CCR8 PK in the blood, tumor, and non-tumor tissue, (2) anti-CCR8-CCR8 binding, (3) CCR8+ Treg cell depletion in the tumor, (4) CD8+ T cell recruitment to tumor, and (5) tumor cell killing. In vitro binding data was used to quantify the binding affinity of anti-CCR8. PK, CCR8+ Treg cell count, and tumor size data were collected from E0771 tumor-bearing mice after a single dose of anti-CCR8 and were used to develop and calibrate the model. The model incorporates linear as well as nonlinear clearance due to explicitly represented binding to CCR8 in order to fully characterize the mouse PK across all dose levels. CCR8+ Treg cell depletion is then modeled as dependent on the amount of anti-CCR8-CCR8 complex. The model connects PD to efficacy by incorporating CD8+ T cell recruitment as a result of CCR8+ Treg cell depletion followed by tumor cell killing in a CD8+ T cell-dependent manner. All model simulations were run using gQSPSim, a MATLAB toolbox for running Simbiology models (MATLAB 2021a).

Results: We calibrated the minPBPK-PD model to single-dose mouse PK data across four dose levels for anti-CCR8 antibody. Nonlinear PK was observed that could not be explained by target-mediated drug disposition (TMDD) due to expected low expression in mice, therefore we included linear and nonlinear nonspecific clearance terms in the model (Fig. 1A). The model predicts that a dose of 10 mg/kg is required to fully saturate CCR8+ Tregs in the tumor for a period of 21 days (Fig. 1B). The rate of Treg cell depletion was fit to mouse CCR8+ Treg cell counts in the tumor at days 3 and 7 post dose. The model captures the dose-dependent Treg depletion dynamics and accurately describes the depletion and rescue of CCR8+ Treg cells to mouse tumors after a low dose of anti-CCR8 was administered (Fig 1C). In the model, dose-dependent CD8+ T cell counts were calibrated to data from mouse tumors at days 3 and 7 post-dose, and tumor cell dynamics were calibrated to tumor size data for 24 days post-dose (Fig. 1D). The model accurately captures the time delay between CCR8+ Treg cell depletion, which occurs within the first three days post-dose, and tumor cell killing, which occurs over a longer time scale of 8-24 days. The model demonstrates that doses causing only 70% RO at a maximum in the tumor (0.1 mg/kg) are sufficient to fully eliminate the tumor by day 21.

Conclusion: Using in vitro and in vivo mouse data to model anti-CCR8 PK, downstream Treg depletion, CD8+ T cell response, and tumor killing, we provide a quantitative mechanistic framework to capture PD and efficacy of anti-CCR8. The model proposes that the long time delay between anti-CCR8-CCR8 binding and tumor cell killing is due to the slow recruitment of CD8+ T cells to the tumor. The totality of mouse data can be explained by the model, and suggests that the proposed mechanisms yield the observed durable response.

References: [1] Barsheshet Y, et al. CCR8+ FOXp3+ Treg cells as master drivers of immune regulation. PNAS, 114 (23) 6086-6091. 2017.
[2] Cao Y, Balthasar JP, Jusko WJ. Second-generation minimal physiologically-based pharmacokinetic model for monoclonal antibodies. JPKPD, (40) 597-607. 2013.

Acknowledgments: This work was funded by Genentech.

Figure 1. Minimal PBPK-PD modeling captures anti-CCR8 PK, PD, and efficacy in mice. The model captures (A) serum PK, (B) receptor occupancy in mouse tumors, (C) CCR8+ Treg depletion in mouse tumors, and (D) tumor growth inhibition.