Purpose:

In this study, we elucidated a novel concept of reverse translational science by utilizing model-based approach to transfer the knowledge in reverse direction from development to discovery (Fig.1). FcRn affinity profiles and endothelial endocytosis have been shown to play an important role in the disposition of therapeutic monoclonal IgG antibody (TMAb). The minimal physiologically-based pharmacokinetic (mPBPK) model was developed in this study to construct detail quantitative relationships between the endothelial endocytosis rate/FcRn binding affinity and the PK parameters of the TMAb. Such relationships are useful to identify the discovery “sweet spot” in designing future generation of TMAb with desirable PK preclinical and clinical PK properties (Reverse Translational Approach). Furthermore, the mPBPK model developed in the preclinical model was useful to predict the PK of the TMAb in adult human subjects (Forward Translational Approach).

Methods: The following PK were used in this study: exogenous/endogenous native IgG in cynomolgus monkey (CM) and human subject (HS), anti-VEGF IgG₁ wild type and its five variants in CM with a wide range of FcRn-binding at both acidic and physiological pH, and bevacizumab PK in HS. The PK of exogenous native IgG in CM was constructed using Complex Dedrick Plot method. The basic structure of the developed mPBPK models consisted of plasma, tissue and lymph compartments (Fig.2).  The tissue compartment was further subdivided into vascular, endothelial and interstitial spaces. The pH transition between endosome with acidic and physiological pH was included in the model. The following parameters were fitted to the PK of antibodies in CM: IgG synthesis rate K_syn, IgG endothelial clearance CL_e, dissociation constant K_D of the native IgG at physiological pH KD_ph7.4, modulation factor of the vascular reflection coefficient F1 and modulation factor of the endocytosis rate F2. The model was then used to simulate the concentration-time profile after a single IV dose of 5 mg/kg for anti-VEGF IgG₁ antibodies to construct the quantitative relationships between FcRn binding affinity and the clearance/terminal half-life of anti-VEGF IgG₁ antibody in CM. The predictive performance of the final model was examined by using the model to predict the PK of bevacizumab in HS. The CM mPBPK model was re-calibrated to the PK of endogenous/exogenous native IgG in HS by fitting the following parameters: IgG endothelial clearance CL_e, and dissociation constant K_D at physiological pH KD_ph7.4. By fixing these parameters, the calibrated HS mPBPK model was then used to simulate the PK of bevacizumab in adult human at dose 5, 7.5 and 10 mg/kg and was compared to the PK data of bevacizumab obtained from the literature.


Results: The estimated parameters in CM were: K_syn = 0.053 l/h (CV = 0.4%), CL_e = 0.034 l/h (CV = 2%), and KD_ph7.4 = 23 µM (CV = 3%) , F1 = 1.55 (CV = 0.2%), F2 = 0.09-0.27 (CV = 0.7-1.5%). The PK of anti-VEGF antibodies with a wide range of FcRn binding affinity at acidic and physiological pH were simulated by using the median (0.11), minimum (0.09) and maximum (0.28) value of the estimated F2. As a result, the endocytosis rate had noticeable effects on the PK parameters of the antibodies. For example, the maximum/minimum achievable half-lives of the anti-VEGF IgG1 antibodies were 150/25 d, 140/20 d, and 120/10 d, for the endocytosis rate based on F2 of 0.09, 0.11 and 0.28, respectively. The maximum/minimum achievable clearances of the anti-VEGF IgG1 antibodies were 1.9/0.6 ml/d/kg, 2.2/0.7 ml/d/kg, and 5.5/0.8 ml/d/kg, for the endocytosis rate based on F2 of 0.09, 0.11 and 0.28, respectively.  At each endocytosis rate, the effect of FcRn binding affinity at physiological pH on the plasma terminal half-life became increasingly more important as the FcRn binding at acidic pH increased. In constrast, the effect of FcRn binding affinity at acidic pH on the plasma half-life decreased as the FcRn binding affinity at physiological pH increased. The model parameters were successfully calibrated to the PK of endogenous and exogenous IgG in adult human. The fitted parameters in HS were: CL_e = 0.11 l/h (CV = 0.2%) and KD_ph7.4 = 300 µM (CV = 0.6%). Based on the simulated result, the mPBPK model developed in this study was able to adequately predict the PK of bevacizumab in adult human subjects.


Conclusion:

To our knowledge, this is the first mPBPK model developed to quantify the complex relationships between FcRn binding affinity, endothelial endocytosis and disposition of IgG. Such quantiative relationships are useful for the discovery of the “sweet spot” in designing the next generation of anti-VEGF IgG₁ antibodies with optimal PK properties. Furthermore, the mPBPK model developed in the preclincial CM model was successfully used to predict the PK of the bevacizumab in adult human subjects.

Presenting Author(s)

• DH

  Deni Hardiansyah

  – Post Doc Scholar, University of Kentucky, Lexington, Kentucky

Main Author(s)

• CN

  Chee Ng

  – University of Kentucky, Kentucky

Primary/Principal Investigator(s)

• CN

  Chee Ng

  – University of Kentucky, Kentucky

Co-Author(s)

• CN

  Chee Ng

  – University of Kentucky, Kentucky
• DH

  Deni Hardiansyah

  – Post Doc Scholar, University of Kentucky, Lexington, Kentucky