Product quality considerations for mammalian cell culture process development and manufacturing.

The manufacturing of a biologic drug from mammalian cells results in not a single substance, but an array of product isoforms, also known as variants. These isoforms arise due to intracellular or extracellular events as a result of biological or chemical modification. The most common examples related to biomanufacturing include amino acid modifications (glycosylation, isomerization, oxidation, adduct formation, pyroglutamate formation, phosphorylation, sulfation, amidation), amino acid sequence variants (genetic mutations, amino acid misincorporation, N- and C-terminal heterogeneity, clipping), and higher-order structure modifications (misfolding, aggregation, disulfide pairing). Process-related impurities (HCP, DNA, media components, viral particles) are also important quality attributes related to product safety. The observed ranges associated with each quality attribute define the product quality profile. A biologic drug must have a correct and consistent quality profile throughout clinical development and scale-up to commercial production to ensure product safety and efficacy. In general, the upstream process (cell culture) defines the quality of product-related substances, whereas the downstream process (purification) defines the residual level of process- and product-related impurities. The purpose of this chapter is to review the impact of the cell culture process on product quality. Emphasis is placed on studies with industrial significance and where the direct mechanism of product quality impact was determined. Where possible, recommendations for maintaining consistent or improved quality are provided.

Production of stable bispecific IgG1 by controlled Fab-arm exchange: scalability from bench to large-scale manufacturing by application of standard approaches.

The manufacturing of bispecific antibodies can be challenging for a variety of reasons. For example, protein expression problems, stability issues, or the use of non-standard approaches for manufacturing can result in poor yield or poor facility fit. In this paper, we demonstrate the use of standard antibody platforms for large-scale manufacturing of bispecific IgG1 by controlled Fab-arm exchange. Two parental antibodies that each contain a single matched point mutation in the CH3 region were separately expressed in Chinese hamster ovary cells and manufactured at 1000 L scale using a platform fed-batch and purification process that was designed for standard antibody production. The bispecific antibody was generated by mixing the two parental molecules under controlled reducing conditions, resulting in efficient Fab-arm exchange of>95% at kg scale. The reductant was removed via diafiltration, resulting in spontaneous reoxidation of interchain disulfide bonds. Aside from the bispecific nature of the molecule, extensive characterization demonstrated that the IgG1 structural integrity was maintained, including function and stability. These results demonstrate the suitability of this bispecific IgG1 format for commercial-scale manufacturing using standard antibody manufacturing techniques.

Efficient generation of stable bispecific IgG1 by controlled Fab-arm exchange.

The promise of bispecific antibodies (bsAbs) to yield more effective therapeutics is well recognized; however, the generation of bsAbs in a practical and cost-effective manner has been a formidable challenge. Here we present a technology for the efficient generation of bsAbs with normal IgG structures that is amenable to both antibody drug discovery and development. The process involves separate expression of two parental antibodies, each containing single matched point mutations in the CH3 domains. The parental antibodies are mixed and subjected to controlled reducing conditions in vitro that separate the antibodies into HL half-molecules and allow reassembly and reoxidation to form highly pure bsAbs. The technology is compatible with standard large-scale antibody manufacturing and ensures bsAbs with Fc-mediated effector functions and in vivo stability typical of IgG1 antibodies. Proof-of-concept studies with HER2×CD3 (T-cell recruitment) and HER2×HER2 (dual epitope targeting) bsAbs demonstrate superior in vivo activity compared with parental antibody pairs.

Modulation of antibody galactosylation through feeding of uridine, manganese chloride, and galactose.

Through process transfer and optimization for increased antibody production to 3 g/L for a GS-CHO cell line, an undesirable drop in antibody Fc galactosylation was observed. Uridine (U), manganese chloride (M), and galactose (G), constituents involved in the intracellular galactosylation process, were evaluated in 2-L bioreactors for their potential to specifically increase antibody galactosylation. These components were placed in the feed medium at proportionally increasing concentrations from 0 to 20 × UMG, where a 1× concentration of U was 1 mM, a 1× concentration of M was 0.002 mM, and a 1× concentration of G was 5 mM. Antibody galactosylation increased rapidly from 3% at 0× UMG up to 21% at 8× UMG and then more slowly to 23% at 20× UMG. The increase was primarily due to a shift from G0F to G1F, with minimal impact on other glycoforms or product quality attributes. Cell culture performance was largely not impacted by addition of up to 20× UMG except for suppression of glucose consumption and lactate production at 16 and 20× UMG and a slight drop in antibody concentration at 20× UMG. Higher accumulation of free galactose in the medium was observed at 8× UMG and above, coincident with achieving the plateau of maximal galactosylation. A concentration of 4× UMG resulted in achieving the target of 18% galactosylation at 2-L scale, a result that was reproduced in a 1,000-L run. Follow-up studies to evaluate the addition of each component individually up to 12× concentration revealed that the effect was synergistic; the combination of all three components gave a higher level of galactosylation than addition of the each effect independently. The approach was found generally useful since a second cell line responded similarly, with an increase in galactosylation from 5% to 29% from 0 to 8× UMG and no further increase or impact on culture performance up to 12× UMG. These results demonstrate a useful approach to provide exact and specific control of antibody galactosylation through manipulation of the concentrations of uridine, manganese chloride, and galactose in the cell culture medium.

A semi-empirical mathematical model useful for describing the relationship between carbon dioxide, pH, lactate and base in a bicarbonate-buffered cell-culture process.

The purpose of the present study was to develop a quantitative relationship between the primary factors of state affecting pH control in a bicarbonate-buffered medium. Starting with the Henderson-Hasselbach equation, several assumptions led to the following equation: L = B(T)-s x dCO2 x 10(pH-pK) where L is the lactate concentration (mM), B(T) is the total amount of base added (mM), s is the solubility of CO(2) (mM/%), dCO(2) is the dissolved CO(2) concentration (%) and pK is the acid ionization constant for bicarbonate. This equation appropriately described the relationship of these factors when using bicarbonate, carbonate and HCl (as a lactic acid surrogate) in water. However, the equation required modification to describe the relationship in cell culture medium, due presumably to the presence of other buffers and components; the final form of the equation from an empirical fit in the absence of cells was: L = B(T)-0.88 x dCO2(0.79) x 10(pH-6038) This equation was tested against actual cell culture data, from inoculum preparation in a T-flask through a 10000-litre fed-batch bioreactor, by comparing the lactate concentration calculated from base, pH and dCO(2) data with that actually measured in the bioreactor using a YSI 8500 SELECT Biochemistry Analyzer (YSI Inc., Yellow Springs, OH, U.S.A.). In every case, the calculated and actual lactate concentrations were in good agreement. The equation was useful for isolating the mechanisms leading to varied base addition across 2-, 600- and 10 000-litre-scale bioreactors. This procedure enables a new approach for quantitatively evaluating and understanding factors associated with bioreactor pH control.

Optimal NS0 cell growth in a hollow fiber bioreactor requires increased serum concentration or a cholesterol supplement on the cell side of the fiber.

An NSO/GS cell line secreting a humanized antibody was routinely propagated in a T-flask using 2% serum. For scale-up of antibody production, this cell line was inoculated into a hollow fiber system using the same serum concentration. The metabolic activity increased for a few days in the hollow fiber system, but invariably the activity dropped dramatically as the cells died by day 7. A hollow fiber micro-bioreactor was used as a screening tool to examine possible reasons for cell death in the large-scale system. As seen in the hollow fiber system, cells died when 2% serum was used either on the cell side only or on both sides of the fiber in the micro-bioreactor. In contrast, the use of 20% serum on the cell side of the fiber and basal medium on the non-cell side resulted in good cell expansion at high viability. Regardless of the cell side serum concentration, no further growth enhancements were seen when up to 20% serum was placed on the non-cell side of the fiber. These results suggest that a serum component that does not readily cross the fiber is limiting cell growth in the hollow fiber bioreactors. The addition of a cholesterol-rich lipid supplement resulted in better cell growth in the micro-bioreactor, while the addition of other non-cholesterol lipid supplements resulted in no growth enhancement. The growth-enhancing properties of the cholesterol supplement were more pronounced at lower serum concentrations, suggesting that poor growth at low serum concentration was due to suboptimal cholesterol levels. When the cell side serum concentration was increased to 20% in the hollow fiber system, cells grew and filled the bioreactor, allowing a 39-day production run. These results demonstrate that this NSO cell line requires an increased cell side serum concentration for optimal growth and that this requirement is likely due to the inherent cholesterol dependency of this cell line.

Use of hollow fiber systems for rapid and direct scale up of antibody production from hybridoma cell lines cultured in CL-1000 flasks using BD Cell MAb medium.

The combination of BD Cell MAb medium with the CL-1000 flask is increasingly being used to generate a few hundred milligram of antibody for early stage research projects. Cells are inoculated at 2 million per ml, and the antibody is harvested after 15 days or when the antibody concentration reaches above 10 mg ml(-1), whichever comes first. Currently, there is no means to scale up beyond this production level using this technology. In this study, we evaluated hollow fiber technology as the scale up alternative. The hollow fiber system was run in batch mode to mimic the method used for the CL-1000 with BD MAb medium. The FL-NS murine hybridoma cell line was simultaneously inoculated at 2 million cells per ml in a CL-1000 and the Maximizer hollow fiber bioreactor system, a 21-fold theoretical scale up over the CL-1000. The Maximizer produced 23-fold more antibody, very close to the expected theoretical amount. However, production was complete after 9 days in the Maximizer, while the CL-1000 required the full 15 days for production. In summary, these results demonstrate successful scale up of antibody production from the CL-1000 to a hollow fiber system.

Antibody production by a hybridoma cell line at high cell density is limited by two independent mechanisms.

Our previous attempt to model the stationary phase of production-scale hollow-fiber bioreactors using a scaled-down micro hollow-fiber bioreactor resulted in a predicted antibody production rate that was three- to fourfold lower than the actual value (Gramer and Poeschl, 2000). Medium limitations were suspected as the reason for the discrepancy. In this study, various increases in medium feed rate were implemented in the micro bioreactor by increasing the diameter of the silicone tubing that houses the hollow fibers. Because larger diameter tubing may induce oxygen limitations, we also explored the effect of medium recirculation to enhance oxygenation. Antibody production in the micro bioreactor increased both as a result of increased medium supply and due to medium recirculation. However, these parameters increased antibody production through two independent mechanisms. The increased medium supply resulted in a higher cell-specific antibody production rate, but not a higher viable cell density. Medium circulation resulted in the support of a higher viable cell density, but had little effect on the cell-specific secretion rate. The two mechanisms of enhanced antibody production were additive, demonstrating that simultaneous parameters can limit antibody production by this cell line in a hollow-fiber system. When the medium feed and circulation rates were increased to a volumetrically proportional scale, scale-up predictions from the micro bioreactor matched the actual data from the production-scale system to within 15%. These data demonstrate the usefulness of the micro bioreactor for characterizing cell growth and limiting mechanisms at high cell densities.