Satoshi Shoji (1), Akiyuki Suzuki (1), Parya Nouri (2), Chun-Hua Cai (2), Puneet Gaitonde (2), & Scott Marshall (3)
(1) Pharmacometrics, Pfizer R&D Japan, (2) Clinical Pharmacology, Pfizer Inc. USA, (3) Pharmacometrics, Pfizer R&D Ltd. UK.
Introduction/Objectives: Tanezumab, a novel monoclonal antibody against nerve growth factor (NGF), is being developed to treat pain from moderate-to-severe osteoarthritis (OA). Target-mediated drug disposition (TMDD) approximation models were investigated to predict unobserved free systemic NGF concentration following tanezumab administration [1, 2]. Objective of this study was to predict free systemic NGF suppression following subcutaneous (SC) administration every 8 weeks (Q8W) of tanezumab in patients with OA based on a TMDD model updated from the previous investigation [2] using available data from pivotal three Phase 3 studies.
Methods: A TMDD approximation model, indirect response model interpretation of the Michaelis-Menten approximation to full TMDD model [3, 4], was fit to plasma tanezumab and serum NGF concentration-time data following SC administration of tanezumab 2.5 mg or 5 mg Q8W in patients with OA across three Phase 3 studies. Population pharmacokinetic (PK) parameters used for this analysis were fixed to the estimates from an intravenous/SC joint PK analysis, which was separately conducted using data from Phase 2/Phase 3 studies. In the studies, sparse PK data for plasma tanezumab were collected whereas serum NGF data was collected in a subset of the population. These analyses used the non-linear mixed-effects modeling approach. The software packages NONMEM (non-linear mixed effects modeling, version 7.3.0 with the parallel processing system) and Perl-speaks-NONMEM (a Perl module for NONMEM related programming, version 4.7.9) and R software (a language and environment for statistical computing, versions 3.2.2, 3.4.1, and 3.6.1) were also used.
Results: The number of samples for plasma tanezumab concentration and serum total NGF concentration was 7856 (2542 patients) and 3490 (790 patients), respectively. Of these patients, 706 patients had samples for both tanezumab and total NGF. The pharmacodynamic parameter estimates (RSE) for NGF production rate constant (ksyn), baseline NGF concentration (BASE), internalization rate constant (elimination rate of tanezumab-NGF complex, kint), and steady-state constant (Kss) were 174 pg/mL/day (5.96%), 27.3 pg/mL (1.51%), 0.0309 1/day (8.21%), and 0.728 ng/mL (4.30%), respectively. Inter-individual variability for ksyn, BASE, and Kss (CV) were estimated to be 22.0%, 20.2%, and 48.3%, respectively. Predictions from the PK-NGF model indicate that 2.5 mg every 8 weeks by SC administration would be predicted to suppress free NGF concentration in the systemic circulation by approximately 75% near the peak of tanezumab concentration, and by <5% at the trough (median). Another simulation considering inter-individual variability indicated that free NGF concentration was predicted to return to the baseline level at approximately 8 weeks (95% prediction interval: 5 to 16 weeks) after the last SC dose of tanezumab 2.5 mg.
Conclusions: A prior TMDD approximation model for tanezumab has been applied to data following SC administration to provide a prediction of unobservable free systemic NGF suppression, in support of the understanding of the pharmacology of tanezumab.
References:
[1] Arends RH, Kaila N, Marshall SF, Gibiansky L. Translational modeling of tanezumab pharmacokinetics (PK) and tanezumab-NGF relationship to predict free NGF concentrations in nonhuman primates (NHP) and humans. Poster presentation at the 2016 AAPS National Biotechnology Conference; May 16-18, 2016; Boston. Poster T2064.
[2] Shoji S, Marshall S, Xie R, et al. Slow drug-target complex kinetics and first dose overestimation of free target suppression in target-mediated drug disposition (TMDD) approximation models: An evaluation for tanezumab a NGF antibody for treatment of pain. PAGE 2017; 26:Abstr 7121
[3] Gibiansky L, Gibiansky E, Kakkar T, et al. Approximations of the target-mediated drug disposition model and identifiability of model parameters. J. Pharmacokinet. Pharmacodyn. 2008;35:573-91.
[4] Gibiansky L, Gibiansky E. Target-mediated drug disposition model: relationships with indirect response models and application to population PK-PD analysis. J. Pharmacokinet. Pharmacodyn. 2009;36,341-51.
Reference: PAGE 29 (2021) Abstr 9596 [www.page-meeting.org/?abstract=9596]
Poster: Drug/Disease Modelling - CNS