Relevance of Endogenous IgG Binding to FcRn in PBPK Models of Therapeutic mAbs
Ludivine Fronton (1,3) and Wilhelm Huisinga (1,2)
(1) Institute of Biochemistry, Universitaet Potsdam, Germany; (2) Institute of Mathematics, Universitaet Potsdam, Germany; (3) Graduate Research Training Program PharMetrX: Pharmacometrics & Computational Disease Modeling, Martin-Luther-Universitaet Halle-Wittenberg, Freie Universitaet Berlin and Universitaet Potsdam, Germany
Objectives: There is a high interest in developing mechanistic pharmacokinetic models to describe the disposition of monoclonal antibodies (mAbs). Physiologically-based pharmacokinetic (PBPK) models represent a primary choice because they allow one to integrate in-vitro data (e.g. the neonatal Fc Receptor (FcRn) affinity, KD), physiological data (e.g. plasma and lymph flows, plasma and tissue volumes) and in-vivo data (e.g. plasma and tissue concentrations). So far, existing PBPK models in mice are not satisfactory because they either do not take into account tissue data or the endogenous immunoglobulin type G (IgGendo) for model building. We developed a PBPK model for mAb and IgGendo, in mice, in the absence of target which accounts for available plasma data, tissue data including the correction for residual blood and tissue-dependent FcRn expression.
Methods: We used the physiological parameters reported in [1]. The experimental venous plasma and tissue data of the mAb (7E3), administered intravenously at 8 mg/kg, were extracted from [1] for FcRn-knockout and wild-type mice using the software DigitizeIt, version 1.5.8a. The steady-state plasma concentration of total IgGendo was reported in [2]. The tissue concentrations of total FcRn were considered to be tissue-dependent and scaled from the whole-body FcRn concentration according to the different tissue expression levels of FcRn reported in [3]. MATLAB R2010a was used for modelling (lsqcurvefit) and simulations (ode15s solver with default options). The 5 unknown parameters were fitted simultaneously to all data.
Results: To assess whether FcRn-binding can be considered in quasi-steady state, we analytically assessed the concentration of mAb in the endosome considering explicitly the association and the dissociation to FcRn (detailed model) and under the quasi-steady-state assumption (QSSA), based on the association (kon) and dissociation (koff) constants values proposed in [4]. Given the data, the model based on QSSA resulted in excellent approximation of the detailed model due to the relative slow degradation rate constant (kdeg) of mAb compared to the fast association to FcRn. We previously showed [5] that the endosomal IgGendo level solely sets the saturation level of FcRn. Hence, it is essential to consider it in the PBPK model. However, the parameter estimates are highly sensitive to the IgGendo-binding to FcRn. We used these insights to translate the explicit IgGendo- and FcRn-model dependence into an implicit dependence which makes it possible to reliably perform the parameter estimation process.
Conclusions: To explain mechanistically the elimination process of IgGs (i.e. mAb and IgGendo), it is critical to integrate the saturable binding to FcRn in the PBPK model. However, the induced sensitivity of parameter estimates necessitates considering the FcRn-binding implicitly.
References:
[1] A. Garg and J.P. Balthasar. J Pharmacokinet Pharmacodyn, 34:687-709, 2007.
[2] R.J. Hansen and J.P. Balthasar. J Pharm Sci, 92:1206-1215, 2003.
[3] J. Borvak etal. Int Immunology, 10:1289–1298, 1998.
[4] D.K. Shah and A.M. Betts. J Pharmacokinet Pharmacodyn, 39:67-86, 2012.
[5] L. Fronton etal. PAGE 20 (2011) Abstr 2033.
