Mathematical modeling coupled with lab experiments was used to optimize process development of bispecific mAbs. Application of QbD elements to the formation of bispecific antibodies through controlled Fab-arm exchange is demonstrated by considering first principles and kinetic data coupled with lab bench and manufacturing scale data to predict bispecific formation yield and product quality from an input of 25 different parameters. Multiple process parameters (ex. temperature, pH, ionic strength, reducing agent, IgG concentration, and dissolved oxygen) have direct effects on the final product quality and yield. Although the individual effects of each parameter can be known, because of the complex nature of the reaction (reduction and dissociation of Fab homodimers, association of Fab heterodimers, and reformation of disulfide bonds), and the combined effects of multiple process parameters simultaneously, it can be extremely difficult to optimize and predict the outcome of the reaction under a set of specific conditions.<br/>To answer this limitation and to explore the design space <i>in silico</i>, a kinetic-based mathematical model was developed that considers the effects of these process parameters to predict the bispecific formation yield and quality attributes.  Using the model, we were able to optimize the yield and reducing agent and IgG concentration, while simultaneously decrease the reaction time to fit into one manufacturing shift. In addition, the model was valuable to define a robust design space for the operating range of multiple parameters. Comparison of model results with a Definitive Screening Design (DSD) data for two IgG4 case studies and published kinetic data demonstrate a &lt; 5% difference to yield and quality attributes.  This model can be used as a tool for investigating deviations in manufacturing to help predict product quality when one or more variables exceed manufacturing recommended operating (MOR) ranges and has applications for continuous manufacturing.

Development of a predictive model for Fab arm exchange of IgG1 and IgG4 bispecific mAbs: Implementation of QbD design space elements for bispecific abs

2019 Spring National Meeting

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.