428472 Clone and Media Selection in Mab Biosimilar Process Development

Monday, November 9, 2015: 1:04 PM
Ballroom D (Salt Palace Convention Center)
Michael Sokolov1, Jonathan Ritscher1, Miroslav Soos1, Nicola MacKinnon2, Arnaud Perilleux2, Matthieu Stettler2, Hervé Broly2, Massimo Morbidelli1 and Alessandro Butté1, (1)Institute for Chemical and Bioengineering, ETH Zurich, Zurich, Switzerland, (2)Biotech Process Sciences, Merck Serono S.A., Corsier-sur-Vevey, Switzerland

In the biopharmaceutical sector the topic of biosimilars remains a highly controversial issue. Until 2014, eleven different biosimilar active ingredients have been approved by the European Medicines Agency (EMA), while issuing a European marketing authorization to the first biosimilar monoclonal antibody (mAb) in 2013. Experience indicates that 7 to 8 years are required to launch a biosimilar on the market, at a cost of $100 – 250 million, while antibodies are likely to cost more. From the bioengineering perspective, this requires a target-directed process development approach.

This work presents a novel approach for the screening stage of the process development of a biosimilar. The objective of this stage is to evaluate and select both, cell line and media candidates, as well as media supplements and their levels. The procedure relies on the use of high-throughput experiments performed in deep well-plates. The results are analyzed by advanced multivariate data analysis tools, with the aim of maximizing the probability of reaching the bio-target in the following development stages. Regarding the manifold quality characteristics of the target molecule including its glycosylation and charge profiles, as well as aggregated and high and low molecular weight species, an approach combining principal component analysis (PCA) and decision trees is applied to provide selection criteria. The robustness of the approach is then tested on an external validation data set. Finally, the scalability of the derived process design space is discussed by a multivariate comparison to micro-bioreactor and lab scale experiments.


Extended Abstract: File Not Uploaded