Background Atacicept is a recombinant TACI/IgG1-Fc fusion protein that binds and antagonizes the biological activity of the B lymphocyte stimulating factor (BLyS) and a proliferation inducing ligand (APRIL). BLyS and APRIL appear to play a role in the pathogenesis of systemic lupus erythematosus (SLE).
Objectives The aim of the analysis was to develop a semi-mechanistic binding population pharmacokinetic (PK) model of atacicept in healthy volunteers (HV) and SLE patients, to characterize the influence of covariates on the PK of atacicept.
Methods 37 subjects from a Phase I study in Caucasian and Japanese HV where single doses of 25, 75 or 150 mg were administered and 301 subjects from a Phase II study in SLE where 75 or 150 mg doses of atacicept were administered weekly for 52 weeks (bi-weekly during the first four weeks of treatment) contributed 2261 measurements of atacicept in serum. The population PK analysis was conducted with NONMEM software. A Quasi-Steady-State (QSS) approximation  of the target-mediated drug disposition (TMDD)  model was used to describe drug concentrations. The analyzed covariates were weight, age, gender, race, creatinine clearance, serum BLyS and APRIL at baseline, dose, and SLE vs HV population. Model based exposure metrics (e.g. area under the concentration curve, AUC) were derived.
Results A two-compartment QSS TMDD binding model with first-order absorption described atacicept concentrations of the two trials adequately capturing the central tendency and variability in the data. The model provided precise (relative standard errors <20%) estimates of all parameters, including binding (Kss=14.5 ng/mL), target turnover (Rmax=617 ng/mL; Kdeg=0.00431 h-1), and drug-target complex elimination (Kint=0.00066 h-1) parameters. The typical estimate of linear apparent clearances and volumes of distribution of the drug were: CL=0.334 L/h, Vc=32.2 L, Q=0.125 L/h and Vp=37.6 L. Residual variability was moderate, slightly higher in the phase II/III study (CV=25%) than in the phase I study (CV=20%). Drug CL and central volume Vc significantly increased with body weight following allometric relationships (exponents of 0.75 and 1.00 respectively) while absorption rate slightly decreased with age (as Ka ∼ (Age/29)-0.689). No significant differences in PK between HV and SLE patients or ethnicities were detected.
Conclusions The developed population PK model allowed the description of the complete atacicept concentration time profile in SLE patients, a first step in the identification of relevance to the indication exposure-response relationships for PD/clinical/safety endpoints that can inform on future study design.
Gibiansky L, Gibiansky E, Kakkar T, Ma P. Approximations of the target-mediated drug disposition model and identifiability of model parameters. J Pharmacokinet Pharmacodyn, 2008; 35(5):573-591.
Mager DE, Jusko WJ. General pharmacokinetic model for drugs exhibiting target-mediated drug disposition. J. Pharmacokinet Pharmacodyn, 2001; 28:507-532.
Disclosure of Interest O. Papasouliotis Employee of: Merck Serono, Ö. Yalkinoglu Employee of: Merck KGaA, M. Golob Employee of: Merck KGaA, D. Willen Employee of: Merck KGaA, P. Girard Employee of: Merck Serono