There are several mammalian systems capable of yielding high-level expression of recombinant proteins that are suitable for use in the manufacture of protein pharmaceuticals. Cell lines derived from rodents such as Chinese hamster ovary (CHO) cells, baby hamster kidney (BHK) cells, and lymphoma cell lines such as NSO, SP2, and YB2/0 are the most commonly used. However, there are no restrictions on the type of host cell that can be used for commercial manufacturing provided the cells: (1) support high-level product expression over many months in culture, (2) can be scaled to large volumes (100-10,000 L), (3) can reach and maintain high viable cell density in an acceptable manufacturing process (>5 x 106 cells/mL in batch culture, >108 cells/mL in perfusion), (4) provide appropriate posttranslational processing capabilities, and (5) can be appropriately characterized to assure freedom of the cell substrate from adventitious agents, especially viruses.
Reports in the scientific literature suggest that most of the cell lines commonly used for biopharmaceutical production are capable of high-level gene expression on the order of 10-100 pg/cell/day (4,5), although in practice yields >10pg/cell/day have been more commonly achieved with monoclonal antibody products than with other types of recombinant proteins. Events at the transcriptional, translational, or posttranslational level have the potential to limit the cell specific productivity that can be achieved with any given protein. For many monoclonal antibodies, protein secretion appears to mirror mRNA expression levels up to specific productivities of 50 pg/cell/day or more (4). However, for other classes of proteins such as multimeric, cross-linked, or highly glycosylated growth factors and fusion proteins, posttranslational processing and protein folding may be the primary rate-limiting step.
The ability to achieve high-level mRNA expression appears to depend primarily on selecting and optimizing the right combination of host cell, expression vector, and selection strategy rather than on the attributes of any one particular host cell. For example, a panel of CHO cells deficient in dihydrofolate reductase (DHFR) are available that can be combined with a spectrum of vectors that combine promoters highly active in CHO cells with strategies to accomplish efficient transfectant selection using DHFR gene replacement (3,6). Similarly, for murine myeloma NSO cells deficient in expression of glutamine synthase (GS), use of vectors combining strong promoters/enhancers with efficient GS selection systems has allowed optimization of this system (7).
There are accumulating examples of engineering or selecting host cells for attributes that further increase the efficiency or robustness of protein production. These include coexpression of genes that enhance expression from a given promoter (trans- and cis-activators), promote posttranslational protein processing and folding, enhance ability to grow in nutritionally defined media (SUPER-CHO (8), VEGGIE-CHO (9)), improve cell cycle characteristics or provide improvements in product quality (10). Moreover, there is ongoing research directed towards marking transcriptionally active sites in the genome with a reporter gene flanked by recognition sequences that allow for the repetitive replacement of product genes into this same site. Some of these technology refinements will be touched upon briefly later in the chapter. These improvements seem equally suited to most host cells and can be applied to further enhance and optimize any well-developed expression system.
Glycosylation of recombinant glycoproteins is an attribute that has received much attention regarding the ability of various host cell lines to provide appropriate product quality (11-13). Data collected on glycoforms produced by CHO, BHK, and NSO cells indicate that the complex N-linked carbohydrate structures they synthesize contain the same basic oligosaccharide structure naturally occurring on human proteins (Fig. 2). However, each cell type has its own unique qualities. Murine cell lines, unlike hamster lines, may introduce significant quantities
Figure 2 Examples of typical N-linked glycans found on recombinant proteinsproduced in mammalian cells. Structures generally contain a variable number of lactosamine branches containing N-acetylglucosamine (GlcNAc, squares) followed by a galactose (triangles) residue. These branches may be capped with a terminal N-acetyl neuraminic acid residue (sialic acid, NANA, stars, see bottom of figure). Biantennary (A), triantennary (B), and tetraantennary structures are common (C), although these structures may be undersialylated (D) or underga-lactosylated (E). Cell line specific attributes, such as substitution of N-glycolyl neuraminic acid (NGNA) for NANA or addition of a bisecting GlcNAc structure (F), or variable fucosylation
Figure 2 Examples of typical N-linked glycans found on recombinant proteinsproduced in mammalian cells. Structures generally contain a variable number of lactosamine branches containing N-acetylglucosamine (GlcNAc, squares) followed by a galactose (triangles) residue. These branches may be capped with a terminal N-acetyl neuraminic acid residue (sialic acid, NANA, stars, see bottom of figure). Biantennary (A), triantennary (B), and tetraantennary structures are common (C), although these structures may be undersialylated (D) or underga-lactosylated (E). Cell line specific attributes, such as substitution of N-glycolyl neuraminic acid (NGNA) for NANA or addition of a bisecting GlcNAc structure (F), or variable fucosylation may occur.
of terminal N-glycolylneuramic acid (NGNA) in place of terminal NANA, or display an a-1,3-galactosyltransferase activity, that may lead to synthesis of glycan structures potentially immunogenic in humans. However, CHO and BHK cells lack the ability to provide certain structures normally found on some human glycoproteins, including terminal sialic acid residues attached in an a-2,6 linkage to galactose or a bisecting GlcNAc residue. Since the nature of the protein itself as well as the cell culture conditions also greatly influence the quality of the glycans attached to any given recombinant protein (14), the choice of cell line that provides acceptable product quality needs to be made in practice on a protein-by-protein basis.
Was this article helpful?