Ozoralizumab (ATN-103), a novel tumor necrosis factor inhibitor (TNFi), is a humanized, trivalent, bi-specific nanobody containing two human TNF‑binding domains linked to a human serum albumin-binding domain. Pharmacokinetic-pharmacodynamic (PK-PD) modeling of data from a seamless phase 1/2 study in subject with active rheumatoid arthritis (RA) was performed to understand the relationship of ATN-103 exposure to clinical efficacy measures ACR20 and ACR50, and to predict clinical responses at other doses and dose regimens.
In this seamless phase 1/2 study, subjects (253) received total 4 subcutaneous injections of placebo or ATN-103 doses of 10 mg, 30 mg, or 80 mg every 4 weeks (Q4W) or 10 mg or 80 mg every 8 weeks (Q8W) for 16 weeks. A total of 1988 serum ATN-103 samples, intensive first and last dosing cycle with troughs at other event, were analyzed using NONMEM. Longitudinal models of ACR20 and ACR50, data up to week 16, as ordered categorical were tested using published latent variable approach with direct, indirect, and effect-compartment PK-PD models. Simulations were performed to predict possible ACR responses at other doses, regimens, and for longer treatment durations.
A 2-compartment open population PK model with first-order absorption and elimination adequately described the observed concentration-time data. For a typical patient weighing 70 kg, ATN-103 PK parameter estimates were: clearance = 0.325 L/d, central volume = 1.14 L, inter-compartmental clearance = 0.642 L/day, peripheral volume = 2.65 L and the absorption rate constant = 0.152 1/d. Between-subject variability was 38% and 33% for clearance and central volume, respectively. Body weight was identified as a significant covariate on clearance and central volume of distribution and included allometrically. Higher trough ATN-103 concentrations were associated with higher ACR responses, suggesting that exposure may be predictive of clinical response. Observed ACR responses over time were best described by the effect-compartment PK-PD model at all dose/regimens (Figure 1a). Week 16 ACR20 and ACR50 responses at 80 mg Q4W were estimated to be similar to that of other anti-TNFi. For a Q8W regimen, clinical response similar to that of TNFi was predicted at a 160 mg dose (Figure 1 b).
A longitudinal effect-compartment PK-PD model utilizing a latent variable approach was used to describe the ACR20 and ACR50 responses across a range of doses investigated and subsequently used for simulations of other desired treatment regimens. Analysis results suggest that a 160 mg Q8W regimen is likely to be as efficacious as the 80 mg Q4W regimen in terms of ACR20 and ACR50 responses.
Disclosure: C. Udata, Pfizer Inc, 3 ; A. Mukherjee, Pfizer Inc, 3 ; M. Hutmacher, Pfizer Inc, 5 ; M. Peterson, Pfizer Inc, 3 ; K. H. Liao, Pfizer Inc, 3 ; T. Checchio, Pfizer Inc, 3 ; K. Shields, Pfizer Inc ; X. Meng, Pfizer Inc, 3 ; G. Comer, Pfizer Inc .