"CHO cells" or Chinese Hamster Ovary cells have a long history as tools for scientific discovery. These cells were first introduced to biomedical research in the '50s and currently represent the most frequently used non-human mammalian cell line for the generation of biological therapeutics (i.e., monoclonal antibodies, enzymes, cytokines, and hormones). For example, the majority (~84%) of monoclonal antibody therapeutics approved between 2014 and 2018 were produced using CHO cells (Walsh, 2018). CHO cell-derived monoclonal antibodies have been approved to treat a broad range of indications, including cancer, multiple sclerosis, asthma, HIV, and neuroblastoma.
Other mammalian cell lines typically used in bioprocessing include baby hamster kidney (BHK21) cells and NS0 or Sp2/0 murine myeloma cells (Chin et al. 2019). Nonetheless, several favorable properties have driven CHO cells' increased use in bioprocessing, including their resilience to growth conditions, resistance to viral infections, and high protein synthesis capacity. Significantly, protein-processing by CHO cells more closely conserves the post-translational modifications (e.g., glycosylation) and folding found in human proteins.
Increased use of human cell lines is a more recent bioprocessing trend, including human embryonic kidney 293 (HEK293) cells and HT-1080 human sarcoma cells, among many others (Chin et al. 2019). This approach is increasingly favored due to non-human mammalian cell lines' inability to fully replicate human-like glycosylation patterns. For example, CHO cells do not express several glycosylating enzymes, including GnT-III, Gal alpha2,6 ST, and alpha 1,3/4 fucosyltransferase, present in humans cells (Goh & Ng, 2018). Moreover, mammalian cell lines like CHO cells produce post-translational modifications not present in human proteins, such as the addition of alpha-gal and NGNA glycans. These non-human glycosylations ultimately increase the risk for adverse immunological reactions towards biotherapeutics (Dumont et al. 2016). Additionally, because post-translational modifications influence the yield, activity, and pharmacokinetic properties of recombinant proteins, deficiencies in human-like glycosylation patterns may negatively affect production efforts and therapeutic efficacy.
Engineering yeast to improve recombinant protein production
As a host system for recombinant protein expression, one main advantage of the yeast is being amenable to genetic manipulation. Therefore, to better enable producing proteins with desirable properties, such as the right amount and type of glycosylation, variants of each S. cerevisiae and P. pastoris have been developed. For instance, P. pastoris generally modifies recombinant proteins with shorter high-mannose oligosaccharides, having ~20 mannose residues; nevertheless, hypermannosylation occurs with some proteins (e.g., HIVgp120, and neuraminidase of the A/Victoria/3/75 influenza virus). Therefore, deletion of genes involved in glycosylation may prove an effective strategy, as demonstrated for the PNO1 (Phosphomannosylation of N-linked Oligosaccharides) gene. Deleting PNO1 in P. pastoris helped reduce the extentof glycosylation of human antithrombin, resulting in a product with reduced immunogenicity (Miura et al. 2004, Gomes et al. 2018). Lastly, genetic engineering of secretion factors, such as the alpha-mating factor, a signal peptide within the N-terminus of proteins produced by P. pastoris, has helped increase the secretion of recombinant proteins (e.g., horseradish peroxidase and lipase) (Gomes et al. 2018).
Overall, while yeast strains and variants provide several advantages over bacteria for bioprocessing, expression system selection must be guided by the specific recombinant protein’s requirements. For example, compared to Chinese Hamster Ovarian or CHO cells, yeast cannot introduce some of the more complex glycosylation patterns typically seen in mammalian proteins. While this represents a challenge for therapeutic protein production in this eukaryotic system, different approaches, such as yeast strain selection, gene codon optimization, promoter strength selection, and other genetic engineering strategies, may help achieve functional proteins at high yields (Dalton and Barton, 2014).