Inter-Antibody Variability in the Clinical Pharmacokinetics of Monoclonal Antibodies Characterized Using Population Physiologically Based Pharmacokinetic Modeling

The objective of this work was to develop a population physiologically based pharmacokinetic (popPBPK) model to characterize the variability in the clinical PK of monoclonal antibodies (mAbs) following intravenous (IV) and subcutaneous (SC) administration. An extensive literature search was conducted and clinical PK data for FDA-approved as well as non-approved mAbs were collected. Training and validation datasets of 44 and 9 mAbs exhibiting linear pharmacokinetics were used for model development. The variability in antibody PK was captured by accounting for different rate constants of pinocytosis (CL_up) and intracellular degradation (k_deg) for different mAbs. Typical values for CL_up and k_deg and their respective inter-antibody variabilities ((Formula presented.), (Formula presented.) were estimated to be 0.32 L/h/L and 26.1 (Formula presented.) (73% and 46%). Varied absorption profiles following SC dosing were characterized by incorporating inter-antibody variability in local degradation (k_SC) and rate of lymphatic uptake (S_Lu) of mAbs. Estimates for typical k_SC and S_Lu values, and (Formula presented.) were found to be 0.0015 (Formula presented.) and 0.54 (193%, and 49%). FDA-approved mAbs showed less local degradation (0.0014 (Formula presented.) vs. 0.0038 (Formula presented.)) compared with other clinically tested mAbs, whereas no substantial differences in physiological processes involved in disposition were observed. To evaluate the generalizability of estimated PK parameters and model validation, the final popPBPK model was used to simulate the range of expected PK for mAbs following SC administration of nine different mAbs that were not used for model-building purposes. The predicted PK of all nine mAbs was within the expected range specified a priori. Thus, the popPBPK model presented here may serve as a tool to predict the clinical PK of mAbs with linear disposition before administering them to humans. The model may also support preclinical-to-clinical translation and ‘first-in-human’ dose determination for mAbs.