PI-096 - LEVERAGING PBPK MODELING TO ENHANCE THE TRANSLATABILITY OF HALF-LIFE EXTENDED MONOCLONAL ANTIBODIES

A. Haque¹, S. Robey², G. Fayad³, B. Maas⁴; ¹Quantitative Pharmacology & Pharmacometrics, Translational Medicine, Merck & Co., Inc., Rahway, NJ, USA, Rahway, NJ, USA, ²Quantitative Pharmacology & Pharmacometrics, Translational Medicine, Merck & Co., Inc., Rahway, NJ, USA, Rahway, NJ, United States, ³ADME, TMED, Merck & Co., Inc., Rahway, NJ, USA, Boston, MA, USA, ⁴Quantitative Pharmacology & Pharmacometrics, Translational Medicine, Merck & Co., Inc., Rahway, NJ, USA, West Point, PA, USA.

Background: Fc engineering extends monoclonal antibody (mAb) half-life by enhancing binding with the neonatal Fc receptor (FcRn) in the endosome and rescuing the antibody from lysosomal degradation. However, the impact of half-life extension (HLE) on cross-species translation remains understudied, limiting the utility of traditional allometric-based scaling approaches. To overcome this challenge, we employed a minimal PBPK (mPBPK) model that incorporates Fc-engineering concepts and molecular properties to improve clinical PK predictions for HLE mAbs.

Methods: We extracted clinical and pre-clinical PK data on HLE and wild-type (WT) mAbs with linear profiles from PubMed and developed a predictive model based on a previously published model for WT mAbs. We standardized mAb-FcRn binding affinities for each Fc mutation using representative mAbs with published data for both WT and HLE versions, scaling the original binding affinity values based on the fold-change in binding affinity introduced by the HLE mutation at acidic pH. Importantly, we also accounted for reduced mAb recycling due to increased binding at physiological pH in HLE mAbs. Model performance was verified through comparison with clinical PK estimates using NCA.

Results: The literature review identified 6 WT and 10 HLE mAbs (including 6 with YTE, 3 with LS and 1 with LA mutations). After incorporating FcRn binding at physiological pH, our mPBPK model generated improved predictions for HLE mAb PK parameters, with 96%, 87%, and 84% of the predicted parameters falling within a 2-fold range of observed values for half-life (t1/2) (Figure 1), clearance (CL), and steady-state volume (Vss), respectively. The model predicted a mean t1/2 of 96 days (CV 2%), 70 days (CV 1.7%), 109 days and 23.6 days (CV 0.2%) for YTE, LS, LA & WT mAbs respectively, while the mean clinical t1/2 was 90 days (CV 20%), 48 days (CV 11.2%) and 84 days and 25 days (CV 24%), where CV is the Coefficient of Variation.

Conclusion: Our study utilized mPBPK modeling and FcRn binding affinity data to improve predictions for HLE mAb PK. These findings highlight the potential of mPBPK modeling in improving first-in-human predictions and identifying promising HLE mAbs in discovery.