A Novel PBPK Approach for mAbs and its Implications in the Interpretation of Classical Compartment Models
Ludivine Fronton (1,3) and Wilhelm Huisinga (1,2)
(1) Institute of Biochemistry and Biology, Universitaet Potsdam, Germany; (2) Institute of Mathematics, Universitaet Potsdam, Germany; (3) Graduate Research Training Program PharMetrX: Pharmacometrics & Computational Disease Modeling, Freie Universitaet Berlin and Universitaet Potsdam, Germany
Objectives: Classical compartment models are successfully used to characterize the pharmacokinetics (PK) of therapeutic monoclonal antibodies (mAbs). These models, however, still lack a theoretically sound interpretation of their PK parameters. Physiologically-based pharmacokinetic (PBPK) models are an alternative approach because they allow one to integrate in-vitro data (e.g. the neonatal Fc Receptor (FcRn) affinity, KD), physiological data (e.g. endogenous IgG (IgGendo) concentration, plasma and lymph flows, plasma, residual blood and tissue volumes) and in-vivo data (e.g. plasma and tissue concentrations). Existing PBPK models are very complex and parameter identifiability is a major issue as documented by the high parameter sensitivity reported by several authors [1, 2, 3]. The objectives were (i) to develop a novel PBPK model for mAbs which complexity is adapted to the available experimental data, in mice, in absence of target and (ii) to establish its link to classical compartment models.
Methods: We used the physiological parameters reported in [1] and [4]. The experimental venous plasma and tissue data of the mAb (7E3), administered intravenously at 8 mg/kg, were extracted from [2] for wild-type mice using the software DigitizeIt, version 1.5.8a. The residual blood volumes were reported in [5]. MATLAB R2010a was used for modeling (lsqcurvefit) and simulation (ode15s solver). The unknown parameters, i.e. tissue-to-plasma partition coefficients and tissue extraction ratios, were estimated for wild-type mice data.
Results: Because of their large molecular mass ($\sim$ 150 KDa), mAbs exhibit a poor extravasation into tissues. Based on the detailed PBPK model published in [2] and previous insights into mAbs- and IgGendo-binding to FcRn [6], we derived a PBPK model which implicitly considers IgGendo and FcRn-mediated salvage. In short, the characteristics of the resulting PBPK model are similar to those of a PBPK model for small molecule drugs showing a low volume of distribution, permeability-limited tissue distribution and a linear clearance in various tissues. The PBPK model allowed estimating reliably a minimum number of parameters: tissue-to-plasma partition coefficients and tissue extraction ratios. We extended the lumping methodology described in [7] to mAbs. The resulting minimal lumped model can be related to a simple 2-compartment model with a linear clearance from the peripheral compartment that comprises the eliminating tissues.
Conclusions: Our PBPK model suggests a new interpretation of classical compartment models. A very recent article by Shah and Betts [8] supports the generalization of our approach to non-target expressing tissues for rat, monkey and human.
References:
[1] Baxter etal. Cancer Res, 54: 1917, 1994.
[2] Garg and Balthasar. J Pharmacokinet Pharmacodyn, 34:687, 2007.
[3] Shah and Betts. J Pharmacokinet Pharmacodyn, 39:67-86, 2012.
[4] Brown etal Toxicol Ind Health, 13:407, 1997.
[5] Garg (2007), Chap. 3. PhD Thesis, Department of Pharmaceutical Sciences. 71–111.
[6] Fronton and Huisinga. PAGE 21 (2012) Abstr 2571.
[7] S. Pilari and W. Huisinga. J Pharmacokinet Pharmacodyn, 37:1-41, 2010.
[8] Shah and Betts. mAbs 5:2, 297–305, 2013.
