Smart feed design for biopharmaceutical production

Purpose and goal

In SmartFD project, the aim is to use mathematical modelling to improve the development of protein drug (biopharmaceuticals) manufacturing cultivation process for optimal solutions at reduced costs and labour. The focus is the creation of intelligent model-based tools to develop the culture media and feed strategies for manufacturing processes of fed-batch and perfusion. The optimization of the important quality attribute is addressed: the glycosylation, a post-translational modification key for the product stability, activity and immunological response of the body.

Expected results and effects

The influence of sugars on glycosylation was studied and described by new mathematical models. Novel methods for smart model identification for kinetics model were created by incremental modelling and by Gaussian modelling. These models were used to design feeds for perfusion mode and for fed-batch mode, and were then applied in perfusion runs and in fed-batch runs. Importantly, we have shown that this type of model-based strategy can improve the development outcome very efficiently. The work was presented in 2 peer-reviewed articles, 3 manuscripts and 3 conference presentations.

Planned approach and implementation

The project was a collaboration of several groups from KTH, i.e. Cell Technology Group in Industrial Biotechnology Dept. for bioprocessing and process modelling, Glycoscience Div. for glycan structure and Decision and Control Systems Div. for modelling, system identification and model-based feed, GE Healthcare and Cobra Biologics. The KTH groups performed the lab work, modelling, system identification and design, while the industrial partners provided inputs relevant for the industry, raw material and support for analyses.