2008 - Marseille - France

PAGE 2008: Applications- Biologicals/vaccines
Daren Austin

Use of mechanistic models to estimate target antigen load for monoclonal antibodies

Daren Austin, Isabelle Pouliquen, Mark Baker

Clinical Pharmacology, Modelling & Simulation. GlaxoSmithKline, Greenford

Objectives: Monoclonal antibodies to cell surface targets are characterised by target-mediated non-linear clearance that is saturable at high molar ratios with subsequent linear pharmacokinetics and constant clearance with increasing dose. For targets such as the endothelial growth factor receptor (EGFR), target mediated clearance is the predominant clearance pathway at clinical doses. These nonlinear kinetic characteristics are clearly demonstrated for cetuximab (ErbituxTM) at clinical doses of 50 mg/m2, with a saturating value of 20 mL/h/m2[1] We propose, a model-based understanding of the binding kinetics, receptor turnover and pharmacokinetics of a potential candidate antibody, and show how the observed pharmacokinetics can be used to differentiate potential candidate antibodies and estimate the magnitude of the target antigen pool.

Methods: we constructed a reaction-kinetic, Pharmacokinetic/Pharmacodynamic (PK/PD) model of antibody binding to a pool of EGFR with published receptor turnover rates. The model was validated using cetuximab pharmacokinetic data and extrapolated to the potential candidate molecules of interest.

Results: The model correctly predicted the published cetuximab kinetics (without model fitting) with a target EGFR load of 2-4 mcg/mL and highly non-linear pharmacokinetic profiles. Linear pharmacokinetics at doses 10-fold lower than clinical cetuximab doses would require a much higher target specificity with an antigen load of at least least 24-5 times smaller then cetuximab.

Conclusions: The model can be used where target mediated clearance of proteins is observed to infer Proof of Pharmacology (binding), target turnover, and target load. Extention to a population-based model is possible, where the expected variability in target expression is likely to be the key driver and often known.

References:
[1] Fracasso et al. Clin Cancer Res 2007; 13(3) 986-93




Reference: PAGE 17 (2008) Abstr 1399 [www.page-meeting.org/?abstract=1399]
Poster: Applications- Biologicals/vaccines
Top