Towards Model-Based Optimization for Quality by Design in Biotherapeutics Production

Abstract

Abstract Quality by Design (QbD) is the concept of designing a process by gathered knowledge to fulfill desired quality of the product. This requires monitoring and controlling product quality throughout process development. In biotherapeutics production, glycosylation is one of the major quality attributes. In general, the glycoform generated by the cell glycosylation machinery is highly dependent on the cell performance and the variations in its surrounding. Mathematical modelling of this machinery can facilitate process understanding, describe the complexity of different layers of cellular metabolism and enable model-based optimization of product formation. In this work, we investigate possible feeding scenarios utilizing a calibrated model to help understanding the possible effects of process changes in the paradigm of QbD for biotherapeutics. Our QbD objective is to maximize product yield while satisfying product quality given by criteria on the protein glycosylation pattern. For this we use a mechanistic model comprised of three major layers (scales) including bioreactor dynamics, intracellular reaction network and kinetic reactions inside Golgi apparatus. We apply design of experiments method followed by in silico experiments using the mechanistic model to identify optimal feeding profiles and evaluate the respective formed glycopatterns. In a future step, these predictions can be experimentally tested. Thereby, in a series of steps, a model can be developed to be used for model-based optimization.