Oleg Milberg 1, Hisham Abdallah 1, Rachel Kudgus Lokken 2, Lutz.O Harnisch 1, Robert Dingman 1, Karoline.A. Meagher 1, Poulabi Banerjee 1, Michael. E. Burczynski 1, Ethan Marin 1, Aaron. P. Kithcart 1, David. E. Gutstein 1
1 Regeneron Pharmaceuticals, Inc. (Tarrytown, United States of America), 2 Allucent (Cary, United States of America)
Introduction:
REGN7508Cat is a high-affinity human monoclonal antibody targeting the catalytic domain of factor XI (FXI). Given the target-mediated nature of FXI inhibition and its nonlinear pharmacokinetics (PK), mechanistic population PK/pharmacodynamics (PD) modelling can quantify exposure–target–response relationships and support translation across clinical populations, including patients with altered coagulation biology and/or comorbidities such as cancer.
Objectives:
To develop and evaluate a joint structural target-mediated drug disposition (TMDD) population PK/PD model describing REGN7508Cat PK, FXI turnover, FXI activity suppression and activated partial thromboplastin time (aPTT) response across healthy volunteers (HVs) and post-orthopaedic surgery (VTE)-prevention participants, and to assess translational consistency in patients with peripherally inserted central catheters (PICCs), including individuals with and without active cancer.
Methods:
Data from a first-in-human HV study (NCT05603195) and a Phase 2 study in post-orthopaedic surgery VTE-prophylaxis study (NCT06454630) were jointly analysed using nonlinear mixed-effects modelling (Monolix 2024R1); simulations were performed in Simulx/RsSimulx (R 4.2.2).
A modified full TMDD structural model [1] was implemented, incorporating:
– Two-compartment free antibody disposition
– Target (FXI) synthesis (ksyn) and degradation (kdeg)
– Drug–target binding (kon, koff)
– Formation and clearance of drug–target complexes
– Parallel first-order systemic clearance and saturable target-mediated elimination
Estimation was performed sequentially in a block-wise manner: PK parameters were estimated first, followed by estimation of FXI activity parameters with PK fixed, and lastly, aPTT parameters with PK and FXI activity parameters fixed. Total drug was composed of free antibody, single arm-bound antibody, and two arm-bound antibody, whereas functional drug was composed of free antibody and single arm-bound antibody only. FXI activity was described as a function of functional drug, and aPTT was linked to FXI activity through a direct response relationship relative to baseline. Between-subject variability was included on key PK, activity and aPTT-related parameters, with allometric scaling on body weight for clearance and volume terms. Covariate selection followed a COnditional Sampling use for Stepwise Approach based on Correlation (COSSAC) procedure. Model qualification included goodness-of-fit diagnostics, prediction-corrected visual predictive checks, parameter precision assessment and evaluation of structural plausibility. In an exploratory translational assessment, preliminary PD data from the ongoing Phase 2 ROXI-CATH PICC study (NCT06299111) were overlaid with model-based predictions for the post-operative population to evaluate consistency of FXI suppression and aPTT response across populations.
Results:
The joint model quantitatively linked REGN7508Cat exposure to FXI inhibition and aPTT response in HVs and post-operative participants, reproducing nonlinear PK and the observed PD time course across dose levels (HVs) and the single-dose post-operative regimen. In post-operative participants receiving 250 mg IV, simulations predicted rapid and near-maximal FXI suppression with corresponding aPTT prolongation early post-dose. The model predicted that the majority of participants (95%) maintain a 2.5-fold increase in aPTT through approximately Day 12, with gradual recovery thereafter; by Days 12–14, FXI levels and aPTT prolongation were projected to be less than maximal, while remaining meaningfully suppressed (aPTT ~2-fold higher) relative to baseline. Across populations, FXI activity suppression and aPTT fold-change trajectories were consistent with robust target engagement over the intended early post-operative window.
Population differences were best explained by shifts in FXI turnover (elevated ksyn and kdeg) and modestly lower aPTT baseline/maximum-effect parameters in post-operative participants, while linear free-antibody clearance was similar between populations – supporting disease- and surgery-related modulation of target biology rather than altered systemic antibody disposition as the primary driver of PK/PD differences. Exploratory overlay of preliminary ROXI-CATH PICC patient PD data demonstrated general consistency with the model-predicted FXI suppression and aPTT trajectories from the post-operative population, supporting structural robustness of the mechanistic framework across populations, including patients with active cancer.
Conclusions:
A mechanistic TMDD population PK/PD model quantitatively characterised REGN7508Cat nonlinear disposition, FXI turnover dynamics and downstream aPTT response across HV and orthopaedic surgical VTE-prevention participants. In post-operative participants receiving 250 mg IV, model-based simulations support rapid near-maximal PD through approximately Day 12, followed by gradual recovery over the subsequent days. Estimated population differences were driven primarily by FXI turnover dynamics rather than changes in systemic antibody clearance. Translational evaluation in PICC study participants – including those with cancer – supported cross-population applicability of the mechanistic framework and its use for model-informed dose selection, cross-indication extrapolation and quantitative evaluation of disease-related modulation of FXI biology in FXI-mediated thrombotic indications.
References:
[1] Gibiansky L and Gibiansky E. J Pharmacokinet Pharmacodyn 2017;44:463–475.
Reference: PAGE 34 (2026) Abstr 12071 [www.page-meeting.org/?abstract=12071]
Poster: Drug/Disease Modelling - Other Topics