Abstract
<jats:title>Abstract</jats:title><jats:p>Transfectants with a wide range of cellular phenotypes are obtained during the process of cell line generation. For the successful manufacture of a therapeutic protein, a means is required to identify a cell line with desirable growth and productivity characteristics from this phenotypically wide‐ranging transfectant population. This identification process is on the critical path for first‐in‐human studies. We have stringently examined a typical selection strategy used to isolate cell lines suitable for cGMP manufacturing. One‐hundred and seventy‐five transfectants were evaluated as they progressed through the different assessment stages of the selection strategy. High producing cell lines, suitable for cGMP manufacturing, were identified. However, our analyses showed that the frequency of isolation of the highest producing cell lines was low and that ranking positions were not consistent between each assessment stage, suggesting that there is potential to improve upon the strategy. Attempts to increase the frequency of isolation of the 10 highest producing cell lines, by in silico analysis of alternative selection strategies, were unsuccessful. We identified alternative strategies with similar predictive capabilities to the typical selection strategy. One alternate strategy required fewer cell lines to be progressed at the assessment stages but the stochastic nature of the models means that cell line numbers are likely to change between programs. In summary, our studies illuminate the potential for improvement to this and future selection strategies, based around use of assessments that are more informative or that reduce variance, paving the way to improved efficiency of generation of manufacturing cell lines. © 2010 American Institute of Chemical Engineers Biotechnol. Prog., 2010</jats:p>
Original language | English |
---|---|
Pages (from-to) | 1455-1464 |
Number of pages | 0 |
Journal | Biotechnology Progress |
Volume | 26 |
Issue number | 5 |
Early online date | 22 Apr 2010 |
DOIs | |
Publication status | Published - Sept 2010 |