(T1330-02-07) Translational PK/PD Modeling to Predict Human PK and Treg Depletion of a Monoclonal Antibody Based on Preclinical Data

Purpose: Tregs promote tumor growth by suppressing the function of CD8+ and CD4+ T cells in the tumor microenvironment (TME) and can hamper the efficacy of immune checkpoint inhibitors. A Treg target X is predominantly expressed on activated Tregs, the most suppressive and proliferative Treg population in the TME. X-expressing Tregs are associated with high tumor grade and poor overall survival across many cancer types. In contrast, the expression of target X on peripheral Tregs or other critical immune cell populations is sparse. Therefore, the project hypothesis is targeting X allows for specific depletion of intratumoral Tregs without impacting peripheral Tregs or other immune cells. A fully human glycoengineered (afucosylated) IgG1 antibody was developed aiming to deplete the human Treg target X for the cancer treatment.

Methods: In this work, we present the translational PKPD modeling used (1) to predict the human PK using two-compartment PK model and allometric scaling from Tg32 mouse PK of human mAb; and (2) to predict the human Treg depletion using Treg depletion observed for the mouse target X surrogate antibody in CT-26 syngeneic mouse model. Risk of non-linear PK in humans due to target-mediated drug disposition (TMDD) was assessed for human mAb using PBPK approach, assuming target X expression is mainly constrained to the tumor compartment.

Results: Model simulations with a sensitivity analysis of a range of binding affinity and tumor volume suggested the low risk of TMDD. The linear PK human clearance (CL), inter-compartmental clearance (Q), volume of the central (V₁) and peripheral (V₂) compartments were projected to be 0.305 ml/h/kg, 1.34 ml/h/kg, 34.8 ml/kg, and 49.4 ml/kg, respectively. The predicted human half-life was approximately 9 days, lower than the typical IgG1 antibodies, which might be explained by the impact of afucosylation. A mouse PKPD model was developed using Treg depletion observed following a single intravenous dose of mouse target X surrogate antibody at 5, 20 and 200 µg in CT-26 syngeneic mouse model. Simulations of the mouse PD model after repeating dosing at the same dose levels used in a CT-26 mouse tumor efficacy study suggested at least 75% Treg depletion was required for efficacy and would be considered minimal efficacious target engagement levels based on available data. Considering the translational uncertainties from mouse to human, 95% target depletion has been targeted for human dose prediction to mitigate this uncertainty.

Conclusion: This work illustrates how modeling integrates preclinical data, fills gaps and supports human dose prediction. The approach can be applied to other programs in oncology and more broadly.

References: 1. Alison Betts, et al. MAbs. 2018 Jul;10(5):751-764.
2. Armin Sepp, et al. J Pharmacokinet Pharmacodyn. 2019 Aug;46(4):339


Two-compartment PK model described the Tg32 mouse PK data well


Mouse PD model and estimation of PD parameters



Simulated mouse Treg after repeat dosing at doses used in a CT-26 efficacy study suggested at least 75% target depletion would be required for efficacy