Liping Zhang, Partha Nandy
Model Based Drug Development, Janssen Research and Development, LLC
Objectives: To develop a mechanistic PK/PD model simultaneously describing the nonlinear interaction between antibody and its target, including drug disposition, antibody-target binding, target removal, and the hypothesized impact of antibody on the supply/synthesis of the target
Methods: Experimental biological compound, Drug X is a fully human monoclonal antibody designed specifically for binding to a growth factor. Steady state serum concentration data of Drug X and its target in cynomologus monkeys were obtained after weekly subcutaneous dosing of 1x, 3x, 10x, and 50x mg/kg. The concentration of Drug X was approximately linear to dose in the studied regimen. The concentration of drug-target conjugate increased with dose and followed the PK profile of antibody at doses lower than 10x, but for most of the dosing duration at 50x, the conjugate concentration was lower than that of 10x, until substantial amount of antibody was removed from the system after the dosing is stopped. Using NONMEM 7, a targeted-mediated drug disposition (TMDD) model was developed and integrated with a “threshold hypothesis”, in which antibody at concentration higher than a certain threshold reduces the amount of target available for binding; such an effect is gone and fully reversible after the antibody concentration falls below the threshold.
Results: A TMDD model integrated with the threshold hypothesis successfully described the observed concentration profiles of antibody and conjugate at all tested dose regimen. The PK model components included target binding, subsequent degradation, and linear first-order elimination from serum. Target was constantly generated in the system, bound to the antibody, and get cleared from serum either as free target or as antibody-target conjugate. The input rate of target available for antibody binding was not impacted by antibody at doses lower than 10x. At doses above 10x and for the duration of antibody concentration higher than a threshold, the antibody suppressed the supply of target available for binding. Parameters in the model, including the hypothesized threshold value were estimated with good precision. Model-based simulation was performed to predict the full time-course of the antibody, free target, and conjugate at various dose levels and regimen.
Conclusion: By adding one more structure parameter to a typical TMDD model, the “threshold hypothesis” successfully described the unusual antibody-target conjugate profile observed at higher doses. The model may be utilized to predict the time-course of non-quantifiable target profiles and to ascertain the complex interaction between antibody-target binding, abundance of antibody, and the supply/synthesis of target.
Reference: PAGE 23 (2014) Abstr 3144 [www.page-meeting.org/?abstract=3144]
Poster: Methodology - New Modelling Approaches