Kinetic reaction modeling for antibody-drug conjugate process development

By combining the specificity of monoclonal antibodies (mAbs) and the efficacy of cytotoxic drugs in one molecule, antibody-drug conjugates (ADCs) form a promising class of anti-cancer therapeutics. This is emphasized by around 65 molecules in clinical trials and four marketed products. The conjugation reaction of mAbs with small-molecule drugs is a central step during production of ADCs. A detailed kinetic model for the conjugation reaction grants enhanced process understanding and can be profitably applied to process optimization. One example is the identification of the optimal amount of drug excess, which should be minimized due to its high toxicity and high cost. In this work, we set up six different kinetic model structures for the conjugation of a cysteine-engineered mAb with a maleimide-functionalized surrogate drug. All models consisted of a set of differential equations. The models were fit to an experimental data set, and the best model was selected based on cross-validation. The selected model was successfully validated with an external validation dataset (R² of prediction: 0.978). Based on the modeling results, process understanding was improved. The model shows that the binding of the second drug to the mAb is influenced by the attachment of the first drug molecule. Additionally, an increase in reaction rate was observed for the addition of different salts to the reaction. In a next step, the model was applied to an in silico screening and optimization, which illustrates its potential for making ADC process development more efficient. Finally, the combination of the kinetic model with a PAT tool for reaction monitoring was demonstrated. In summary, the proposed modeling approach provides a powerful tool for the investigation of ADC conjugation reactions and establishes a valuable in silico decision support for process development.