Directed evolution of CHO host and production cell lines

Background.

Cellular performance during the industrial production of biopharmaceuticals is of key importance for several reasons. The final cell concentration and the cell viability determine the integral of viable cells of the culture, which controls process yield. A high concentration of dead cells in the bioreactor negatively impacts protein quality, due to the release of proteases or glycosidases that degrade the (glyco)protein. The cellular state and physiology has a direct influence on the quality of protein glycosylation, thus severely impacting the biological activity and immunological properties of the produced proteins (Valley et al., 1999; Sung et al., 2004).

The cells predominantly used for production of post-translationally modified therapeutic proteins are Chinese Hamster Ovary (CHO) cells. Due to their long history in culture and the fact that different labs worldwide cultivate them in different media and under different conditions, there is high variation in cellular properties amongst the different strains. In addition every recombination event and the subsequent amplification steps are highly mutagenic and cause genetic variation per se. Thus clones coming out of the same transfection may be highly heterogeneous. This heterogeneity is one of the causes for the high work load during cell line development, as selection and screening is currently an empirical endeavour, where thousands of clones are tested. Screening has been automated by robotics in the industry, however, the criteria for selection of clones are still typically based on a single parameter (= productivity) and do not in any way reflect or predict later behaviour of cells in the bioprocess (Birch and Racher, 2006).

The major bottleneck is our lack of understanding of which cellular processes are required for proper functioning of the cellular factory or how they work in detail, how they are regulated and interconnected. One of the few cellular processes that have been well studied is the unfolded protein response (UPR) caused by overproduction of recombinant proteins (Schröder and Kaufman, 2005). Otherwise, very little is known about the control of metabolism, growth and additional factors that contribute to cellular performance.

The major features of interest in the context of mammalian cell factories are: high growth rate in protein-free medium, high final cell concentration, length of viable culture phase, high specific productivity, and of course, high product quality.

Aims and methods.

We have currently established two model cell lines derived from of the original CHO-K1 cell line. These lines were adapted to growth in protein-free medium within three weeks and then were further optimised using two cell-sorting based approaches: (i) cells were transiently transfected with human monoclonal antibodies and the highest producers repeatedly sorted out of the population. In the final round of sorting cells were subcloned. The obtained subclones were able to reproducibly produce threefold higher amounts of antibody upon transient expression. (ii) As glutamine is the preferred energy substrate of immortalised cells in culture, it is usually added to media to improve cell concentration and viability. The resulting by-product ammonia however, has a negative impact on the glycosylation status of the produced proteins. CHO-K1 cells were therefore adapted to growth in glutamine-free medium by seeding cells at high concentrations in media with progressively reduced glutamine content and isolating those cells that survived after 24 hours (typically less than 5%). The resulting cells are able to grow in glutamine-free medium, reach a comparable to slightly higher final cell concentration in batch culture, and their viability remains above 90% for 3-5 days longer than that of the initial control cell line under bioreactor conditions.

In the present project we will analyse the molecular differences between the parental and the optimised cell lines and try to understand how these cells achieve their improved performance on a molecular level. This information can then be used both to define cellular markers that allow multiparameter screening at early clone selection stages and to engineer cells to become better performers. For this, miRNA and mRNA transcriptomics will identify global regulators that push the cell towards better growth or productivity, and intracellular concentration of relevant proteins will be measured, such as transcription and translation factors, chaperones, and proteins associated with cell cycle progression and apoptosis. Initial focus will be put on proteins identified as regulated by transcriptomics, these will be analysed by flow cytometry. Finally, the phosphorylation state of regulatory proteins (e.g. MAP kinases) will be analysed, using antibodies that differentiate between the phosphorylated and non-phosphorylated state of these proteins.

Valley U., Nimtz M., Conradt H. S., Wagner R. (1999) Incorporation of ammonium into intracellular UDP-activated N-acetylhexosamines and into carbohydrate structures in glycoproteins. Biotechnol. Bioeng. 64, 401-417
Sung, K. Y., Jong, K. H., Gyun, M. L. (2004) Effect of simultaneous application of stressful culture conditions on specific productivity and heterogeneity of erythropoietin in Chinese hamster ovary cells. Biotechnol. Progr. 20, 1293-1296
Birch, J. R., Racher A. J. (2006) Antibody production. Adv. Drug Deliv. Rev. 58, 671-685
Schröder M., Kaufman R. J. (2005) The mammalian unfolded protein response. Annu. Rev. Biochem. 74, 739-789