Receptor Mediated Disposition PK/PD Model of Filgrastim in Healthy Adults following Intravenous and Subcutaneous Administrations.
W. Krzyzanski (1), P. Kawczak (1), P. Wiczling (2), P. Lowe (3), E. Pigeolet (3), F. Ludicke (4), S. Balser (4).
(1) Department of Pharmaceutical Sciences, University at Buffalo, State University of New York, Buffalo, (2) Department of Biopharmaceutics and Pharmacodynamics, Medical University of Gdansk, Gdansk, Poland, (3) Novartis Pharma AG, (4) Sandoz Biopharmaceutical Development.
Objectives: To develop a mechanism based pharmacokinetic (PK) and pharmacodynamic (PD) model of human granulocyte-colony stimulating factor (G-CSF), that would account for the change in the filgrastim clearance upon multiple dosing due to an increase of the G-CSF receptor mediated endocytosis.
Methods: The data consisted of filgrastim (Neupogen®) plasma concentrations and absolute neutrophil counts (ANC) which were obtained from clinical studies in healthy volunteers. Three subcutaneous (SC) doses of 2.5 µg/kg (n = 14), 5 µg/kg (n = 14) and 10 µg/kg (n = 19) µg/kg and one intravenous (IV) infusion (n = 12) of 5 µg/kg over 0.5 h were studied. The PK model included first-order elimination from plasma, receptor binding, and internalization of drug receptors complexes. The PD model consisted of a series of transit compartments that represented the aging populations of neutrophil precursors in the bone marrow pool, a neutrophil blood compartment, and a marginal pool. The serum filgrastim concentration stimulated the differentiation and maturation of the precursor cells as well as the neutrophil mobilization from the bone marrow, and the marginalization. All effects were described by the stimulatory Hill functions with a common SC50 and process specific Smax parameters. A feedback process from ANC controlling the total G-CSF receptor pool was included. Population nonlinear mixed-effect modeling was done using NONMEM VI. The first-order conditional estimation with interaction (FOCE) method was used. The confidence intervals of population parameter estimates were determine by a non-parametric bootstrap procedure.
Results: The estimate of the filgrastim volume of distribution (VD) was 3.4 L. The typical value of the linear elimination rate constant (kel) was 0.3 1/h whereas the internalization rate constant was 0.8 1/h. The equilibrium dissociation constant (KD) was estimated as 1.7 ng/mL (94.4 pM), and the filgrastim bioavailability was 78 %. The typical value of the SC50 was 13.3 ng/mL, and the estimates of the Smax values corresponding to the filgrastim mobilizing effect was 30.0. The PD model also predicted a strong effect of filgrastim on the acceleration of the differentiation and maturation of the neutrophil precursors in the bone marrow. The inter-individual variability was determined for the neutrophil baseline counts, kel, and VD.
Conclusions: The presented model expanded previously published TMDD PK model for filgrastim. The increase in filgrastim clearance upon multiple dosing was attributed to the increased in ANC paralleled by an increase in the total G-CSF receptor density. Simultaneous modeling of filgrastim plasma concentrations and ANC was necessary to adequately describe PK data.
