Online Issues

ILAR Journal Vol 46(3)

<< All Back-issues

<< This Issue's Table of Contents

Industrial Implementation of in Vitro Production of Monoclonal Antibodies

Vincent Dewar, Pierre Voet, Françcoise Denamur, and Jean Smal

Vincent Dewar, Pierre Voet, M.S., Françcoise Denamur, B.S., and Jean Smal, Ph.D., are members of the Scientific Staff of GlaxoSmithKline Biologicals, Rixensart, Belgium.

Abstract

Monoclonal antibodies are widely used at GlaxoSmithKline Biologicals (GSK Bio) for the quantification and characterization of antigens and for the release of vaccine lots. In 1998, GSK Bio decided to change the production of monoclonal antibodies (MAbs) designed for immunological tools from in vivo to in vitro technology. In 2004, all MAbs used at GSK Bio were produced in vitro. These MAbs cover more than 100 different targets with a variety of 1500 hybridomas, and approximately 60 to 90 MAbs are produced every year. This article describes the development process, including a description of the different systems tested based on double membrane or hollow fiber technology. The productivity, assets, and drawbacks of the different technologies are presented, and evaluation strategies for the choice of in vitro systems are discussed. Binding kinetics displayed by MAbs produced in vitro and in vivo were found to be similar, and MAbs produced in vitro are suitable tools for various immunological applications.

Key Words: ascites; hybridoma; hollow fiber systems; in vitro production; monoclonal antibodies; suspension systems, productivity

Introduction

Monoclonal antibodies (MAbs¹) are widely used in biomedical research, diagnosis, and therapy. In vaccine research and development, they are important tools for antigen discovery, quantification and characterization, and for release of vaccine lots. Since 1975, when monoclonal antibodies were introduced with hybridoma technology (Köhler and Milstein 1975), they have been produced in animals, usually mice. Hybridoma cells are hybrid constructs obtained by fusion between a myeloma cell and an antibody-producing lymphoid cell. To multiply antigen-specific hybridoma cells, they are injected into the peritoneal cavity of a mouse primed with pristane (Bruce et al. 2002), where they produce fluid containing a high concentration of monoclonal antibody. The mouse ascites method is easy to set up but generates pain and distress in animals. Considerable research on alternative in vitro methods to the mouse ascites has been carried out in the 1990s.

The expansion of hybridomas in animals has been the topic of much discussion in the scientific community (NRC 1999), and it is becoming less acceptable worldwide. To date, the European countries that have imposed full or partial bans on the mouse ascites technique of producing MAbs are Switzerland, The Netherlands, Germany, the United Kingdom, Sweden (Falkenberg 1998), and, most recently, Belgium. GlaxoSmithKline Biologicals (GSK Bio¹, Rixensart, Belgium), as part of its long-term efforts to reduce the numbers of animals used, has taken the initiative to convert from in vivo to in vitro methods for the production of MAbs. This decision was made in 1998, several years before local legislation limiting the in vivo ascites method was incorporated.

Various in vitro MAb production systems were evaluated, including some that were based on the hollow fiber principle and others, on suspension or double membrane technology. Implementation of the in vitro technology required a basic-to-high degree of expertise in cell culture systems. The implementation and validation of alternative in vitro technologies along with their productivities are described below.

Materials and Methods

Hybridomas and Cell Culture Conditions

The hybridoma cell lines used in this study were fusions of B lymphocytes from rat or mouse with Sp2/0 myeloma cells. Within 8 yr, we have developed at least 1500 different hybridomas that cover more than 100 target antigens. Despite the great variety of the hybridomas, all cell lines were cultured according to a same basic protocol using Dulbecco's modified Eagle culture medium at high glucose concentration, supplemented with glutamine, sodium pyruvate, essential and nonessential amino acids, a cocktail of antibiotics, and 5% fetal bovine serum. In some cases, to avoid the presence of bovine-derived antibodies, serum-free medium (SFM¹; Invitrogen, Carlsbad, CA) enriched with commercially available cell culture supplement (Roche, Basel, Switzerland) was used.

Evaluation of Different in Vitro Systems of Producing Monoclonal Antibodies

Various in vitro MAb production systems that have been evaluated between 2000 and 2003 at GSK Bio, based on suspension and hollow fiber technologies (Jackson et al. 1996), are described below. In suspension systems, the gas exchange occurs by simple diffusion, which may be enhanced by rotation of the culture flask. Hollow fiber systems may be considered as an artificial capillary system in which gas exchange to the cells as well as transport of nutrients and waste metabolites are optimized. Examples of these systems that were evaluated and include the following: Stationary suspension culture systems, rotation suspension culture systems, and hollow fiber bioreactor systems.

Stationary Suspension Culture Systems

CELLine^TM 1000. The CELLine^TM 1000 (Integra Bioscience, Chur, Switzerland) device is a membrane-based disposable cell culture system that is easy to use. It is composed of two compartments, a cultivation chamber (20 mL) and a nutrient supply compartment (1000 mL) separated by a semipermeable dialysis membrane (10 kD molecular weight cut-off), which allows small nutrients and growth factors to diffuse to the production chamber. Oxygen supply of the cells and CO₂ diffusion occur through a gas-permeable silicone membrane. Antibodies concentrate in the production medium. This culture system requires a CO₂ incubator.

For optimal production levels, the device was inoculated with 50x10⁶ cells, and 80% of the production medium and the entire nutrition medium were changed twice a week. Cells were counted, and the viability of cells was measured by trypan blue exclusion to ensure an optimal balance between cell density and high viability (Figure 1). At least 12 CELLine^TM 1000 devices can be placed in a 180-L CO₂ incubator, requiring no additional investment. A technician with basic cell culture skills can operate 16 CELLine^TM 1000 bioreactors simultaneously. This cell culture system is appropriate for use in a wide spectrum of laboratories. It meets the needs of both small research groups and professionals who must cover a large spectrum of MAbs.

Figure 1 Viability of hybridoma cells in the CELLine^TMPt 1000 bioreactor during culture. Time course of the number of living hybridomas in the cell culture system in relation to their viability (%) and the cell concentration (number of cells/mL) in the production compartment.

Rotation Suspension Culture Systems

miniPERM.

The miniPERM (Vivascience, Hannover, Germany) is a modified roller bottle two-compartment bioreactor in which the production module (35 mL) is separated from the nutrient module (450 mL) by a semipermeable dialysis membrane. Nutrients and metabolites diffuse through the membrane, and secreted antibodies concentrate in the production module (Falkenberg et al. 1995). Oxygenation and CO₂ supply occur through a gas-permeable silicone membrane at the outer side of the production module and through a second silicone membrane extended into the nutrition module. The miniPERM must be placed on a roller base inside a CO₂ incubator.

The device was inoculated with 100x10⁶ cells. Twice a week, 30% of the production medium was harvested, and the nutrient medium was replaced entirely to maintain an optimal balance between cell density and viability. It is possible to place two roller bases together in a 180-L CO₂ incubator, each holding a maximum of four bioreactors (i.e., the same amount of space is occupied for 1-4 incubations). A technician with mid-level cell culture skills can run 16 miniPERMs in parallel. The miniPERM technology is dedicated to a laboratory involved in a small- to medium-scale routine MAb production. Although this device enables cost-efficient production of antibodies, it requires some additional basic investment for the roller base and the command unit.

Hollow Fiber Bioreactor Systems

Hollow fibers are small tube-like filters with a predefined molecular weight cutoff. Large bundles of these fibers can be packed into cylindrical modules, which provide an absolute barrier to cells and antibodies while ensuring perfusion of the liquid (Cadwell 2004). Hollow fiber modules can provide a large surface area in a small volume. The walls of the hollow fibers serve as semipermeable ultrafiltration membranes. Cells are grown in the extracapillary space (ECS¹) that surrounds the fibers, and medium is perfused continuously inside the fibers. Metabolites and small nutrients freely perfuse between extra- and intracapillary space according to concentration gradients. Culture monitoring can be performed by lactate measurement.

CP100. The Cell-Pharm® system 100 (CP100, BioVest, Minneapolis, MN) is a fully integrated 0.14 m² hollow fiber cell culture system. It does not require a CO₂ incubator because air and CO₂ bottles are connected directly to the system while temperature and pH (air/CO₂ flow) are set on the control unit. The cell culture unit consists of two cartridges: one that serves as a cell compartment and the other, as an oxygenation unit.

The system should be inoculated with 400x10⁶ cells. Twice a week, the entire ECS medium (12 mL) containing MAb was harvested and the perfusion medium (1500 mL) was changed. Personnel who are highly skilled in cell culture work are required to operate the CP100 bioreactor, but an experienced technician can handle four CP100 bioreactors concurrently. Increasing this recommended maximum number of parallel runs complicates the handling of the equipment components. Some additional costs for the CP100 control unit and recurrent costs for the disposable cell culture unit represent an important expense. However, the CP100 is suitable for professional MAb production laboratories.

CP2500.The Cell-Pharm® system 2500 (CP2500, BioVest) is a hollow fiber cell culture production system that can produce high-scale quantities of MAbs. Unlike CP100, it consists of two fiber cartridges for the cells and hence offers a large cell growth surface (3.25 m²). A third cartridge serves for oxygenation of the medium.

The systems should be inoculated with 2x10⁹ cells. Feeding (8 L/day) and harvesting (35 mL/day) are carried out automatically by a computerized unit. No incubator is needed, but air and CO₂ gas bottles must be connected directly to the system. The cell culture cartridge volume is 2x60 mL and a 20-L reservoir feeds the system. Because the starting investment is quite costly, the CP2500 bioreactor is typically suitable for professional MAb production units. Although it is essential to employ personnel who are highly skilled in cell culture, the maintenance of the cell culture requires minimal manpower during the operation of the CP2500. To ensure efficient production of MAbs with the CP2500 system, it is advisable to maintain the cell culture for 3 mo.

FiberCell.The FiberCell^TM (Fibercell Systems Inc., Frederick, MD) hollow-fiber cell culture system is composed of a culture medium reservoir (250 mL) and a 60-mL fiber cartridge (1.2 m²), both connected to a single microprocessor-controlled pump (Fibercell^TMsolo pump). It is possible to prolong the media supply cycles by replacing the original medium reservoir with a 5-L flask. In contrast to the Cell-Pharm® systems, the FiberCell^TM bioreactor is used inside a CO₂ incubator. Oxygenation occurs by a gas-permeable tubing.

Optimal inoculation of the system requires 400x10⁶ cells. One to two FiberCell^TM bioreactors together with the pump and the media reservoir can fit into a standard 180 L CO₂ incubator. Operation requires mid-level cell culture skills and a moderate investment. Production capacities, handling, and investment make the Fibercell^TM system suitable for routine MAb production units.

Tecnomouse. The Tecnomouse (Integra Biosciences, Chur, Switzerland) cell production unit provides separation of cultivation (12 mL) and nutrient (10 L) chambers via hollow fibers in combination with two thin gas-permeable silicone membranes to enable oxygenation. Tecnomouse is the only compartmentalized system in which five different hybridoma cell lines can be cultured in parallel in separate cell culture cassettes.

A single cassette should be inoculated with 400×10⁶ cells. The system can be used as basic equipment running in a CO₂ incubator. With the thermohood accessory, no extra incubator is required for the cell unit. One Tecnomouse bioreactor fits into a standard 180-L CO₂ incubator. The system, which requires personnel who are highly skilled in cell culture and a large starting investment, is well suited to professional MAb production units.

Infrastructure

It is possible to produce many different laboratory-scale amounts of MAbs in vitro with a reasonable investment in man power and equipment. For example, the infrastructure used at GSK Bio for MAb production includes a laboratory of 40 m² equipped with two laminar hoods and four incubators, of which three have a capacity of 180 L, and one, of 820 L. The role of the team is to generate hybridoma cell lines, produce and purify monoclonal antibodies, and conjugate monoclonal and polyclonal antibodies to biomarkers. At least one technician works full time to produce in vitro MAbs of high quality. Except for SFM, culture media and additives are prepared in house by a dedicated service.

Results

Development of in Vitro MAb Production for Immunological Applications

Prior to 1998 at GSK Bio, monoclonal antibodies designed for use as immunological tools were produced by the mouse ascites method. In 1998, alternative in vitro technologies were sufficiently advanced to allow for their evaluation and eventually implementation. The evaluations began with two suspension systems that had a production capacity suitable for use in small- to medium-scale research applications. In 1999, 90% of the MAbs were produced in vitro, and the ascites tumor technique was used only for some medium-scale productions (150 mg of MAb) of hybridomas with low-secretion capacity. A hollow fiber bioreactor system was used routinely (e.g., for scaling-up MAb productions with weak secretors). Since 2000, all productions of monoclonal antibodies have been carried out in vitro (Figure 2). In 2003, an advanced hollow fiber cell culture system was introduced that allowed for production quantities of 250 to >500 mg of MAb per run. The trend toward higher production scales continued in 2004 (Figure 3) because the good quality of the MAbs that were obtained allowed their expanded use in tests for vaccine release.

Figure 2 Conversion of the production process at GlaxoSmithKline Biologicals from in vitro to in vivo. The number of monoclonal antibody productions performed by in vivo and in vitro techniques per year are shown for the development phase (1997-1999) and the period in which the in vitro method was implemented for routine use (2000-2004).

Figure 3 Change of in vitro monoclonal antibody production scales within 5 yr at GlaxoSmithKline Biologicals.

Of the MAbs used in routine applications, 50% are currently produced at scales ranging from 150 to 500 mg per run. In general, high-production quantities can be obtained with hollow fiber systems or by multiplication of the less complex suspension systems. Therefore, for medium-scale productions, the hollow fiber technology is applicable to hybridomas with low to high capacity for MAb production, whereas suspension systems are advisable for medium to high producers only.

Productivity of Hybridomas in Different in Vitro MAb Production Systems

The productivity of a cell culture system is the amount of MAbs produced per month, as determined by spectrophotometric measurement of absorbance at 280 nm after affinity purification on protein A sepharose. The productivity of the different in vitro MAb production systems tested is given in Table 1. Numerous different hybridomas of variable production capacities were used for this analysis, and no major difference was noticed when comparing CELLine^TM 1000, miniPERM, and CP100. As indicated, the highest level of productivity was obtained with the CP2500. With all bioreactor systems evaluated, the concentrations of antibodies obtained were 50 to 200 times greater than with T-flask (not shown). Supernatant concentrations were between 100 and 2100 μg/mL reflecting the high variability of hybridomas in terms of secretion capacity. The MAb concentrations obtained in the pooled cell culture supernatants were suitable for use in common immunological applications (e.g., Western blot, enzyme-linked immunosorbent assay [ELISA¹], fluorescent-activated cell sorter [FACS¹], immunohistochemistry, and neutralization assays). The consumption of medium was higher for the hollow fiber systems than for the suspension systems. It can be further deduced from the monthly spent medium listed in Table 1, that miniPERM consumed less medium (about 60%) than CELLine^TM 1000 per milligram of antibody. Yet on a milligram produced basis, CELLine^TM 1000 was more cost effective in medium than CP100 (270 %), CP2500 (330 %), and Fibercell (240%).

Comparing the productivity of different systems with the same hybridoma under the chosen conditions, CELLine^TM 1000, miniPERM and CP100 provided similar results (Table 2). The highest productivity level was obtained with the complex CP2500 system, which was 10 times higher than that obtained with an easy-to-use system such as the CELLine^TM 1000. The productivity of the Tecnomouse system, also reflecting a high level, can even be increased by using its full capacity (5 cassettes) in a single run. With the hybridoma used, the productivity observed with the FiberCell^TM system was three times higher than the productivity obtained with the CELLine^TM 1000 bioreactor.

Productivity of Hybridomas Under SFM Conditions

For some applications, the use of SFM was required to avoid unwanted antigen/antibody reactions. Three different hybridomas were evaluated under serum-free conditions with two different cell culture systems (Table 3). Although the number of tests is too small to allow for a statistical analysis, it appears that all three hybridomas displayed MAb secretion capabilities that were comparable to those observed for other hybridomas cultured in the presence of serum (see Table 1). The results indicate that in vitro production systems might be suitable for MAb productions even under serum-free conditions. Other hybridomas should also be tested because the risk of losing secretion capability under serum-free conditions is not negligible.

Productivity of Rat-Mouse Heterohybridomas

Some investigators have reported that rat-mouse heterohybridomas are less stable than rat-rat hybridomas, and for that reason, they have selected the mouse ascites method to obtain high concentrations of MAbs (Ohlin and Borrebaeck 1994). However, we have not encountered the problem of instability with heterohybridomas under standard experimental conditions. Indeed, the productivity of two rat-mouse hybridomas cultured in the FiberCell^TM system can be seen in Table 4. The FiberCell^TM was chosen to compensate the low secretion capacity of the hybridomas tested. We were able to generate laboratory-scale amounts of MAbs with both hybridomas. The capacity for secretion of MAbs was maintained during the whole culture period, lasting more than 4 wk.

Comparison of the Binding Properties of MAbs Produced in Vitro and in Vivo

GSK Bio considered optimization of the cell culture process to be very important during implementation of in vitro technology in 1998. A total of 69 monoclonal antibody productions were performed by ascites technology, and 55 productions by in vitro methods. The reactivity of 20 antibodies produced by either technique was evaluated by ELISA or Western blot, and their binding properties were shown to be comparable to those of antibodies produced in vivo. The example in Figure 4 indicates how MAbs produced by either technique were used in ELISA as coating (A) or as detection antibody (B). In both cases, the binding curve exhibited the same shape for both antibodies. To date, all monoclonal antibodies produced in vitro have been found to be functional in the various immunological applications tested.

Figure 4 Comparison of binding kinetics measured by enzyme-linked immunosorbent assay (ELISA) with purified monoclonal antibodies (MAbs) produced by in vivo and in vitro techniques. MAbs produced by in vivo or in vitro techniques and purified by affinity chromatography (protein A sepharose) were used as either coating (A) or detection antibodies (B). A: The concentration of coating MAb was varied whereas the concentration of the antigen and the detection antibody were kept constant. B: The concentration of the antigen was varied whereas the concentration of the detection MAb and the coating antibody were kept constant. The ELISAs were performed in 96-well plates according to standard protocol. In both experiments, the detection antibody was coupled to horseradish peroxidase, and the signal of the oxidized chromogen was measured spectrophotometrically at 490 nm.

Discussion

During the introduction of in vitro methods for production of monoclonal antibodies at GSK Bio, the suitability for a given application was tested with each batch of MAb produced by in vitro technology and compared with that of MAbs produced by ascites tumors. New hybridomas generated after 1999 did not undergo such comparative analysis because the ascites tumor technology had been almost fully replaced by in vitro technologies. We report herein on our experiences with sp2/0-derived hybridoma cells cultured in various bioreactor systems. MAbs produced in vitro displayed similar binding kinetics to MAbs produced in vivo, and they were found be suitable tools for 12 different immunological applications (data not shown), including ELISA, Western blot, immunohistochemistry, affinity chromatography, FACS, as neutralizing antibodies, and in bactericidal or opsonophagocytosis applications.

A key factor for the choice of the appropriate in vitro bioreactor is the production capacity of the system, which must meet quantitative requirements at an affordable price. Concentration of the antibodies in the production medium is important when a small liquid volume is required to reduce the time needed for further purification. In addition, the concentration of MAbs obtained with a system should be applicable to immunological applications. It is also important to consider requirements for laboratory space, supplementary material, and technical expertise. Systems that operate outside an incubator save valuable space that can be allocated to alternative cell culture activities.

The productivity of an in vitro system depends on several variables, the most important of which are the culture conditions and the hybridoma cell line inoculated. For example, to achieve production quantities of 250 mg with low secretors (≤ 30 mg/mo; 250×10⁶ cells), it is not advisable to use a suspension system due to the significant manpower and time requirements of that system (for adequate multiplication of the systems). In such cases, a more complex bioreactor that is based on hollow fiber technology is preferable for an economic production-to-investment ratio, despite the high starting costs and media consumption. Suspension systems are more appropriate for small- to medium-scale productions with hybridomas that are characterized by medium to high MAb production capacity (>30 mg/mo, 250×10⁶ cells). In our laboratory, many laboratory-scale productions ranging from 10 to 150 mg have been successful using miniPERM or CELLine^TM 1000, a system that requires little space and is easy to handle. Tecnomouse (1 cassette) and FiberCell^TM (60-mL ECS cartridge) have been very convenient for productions ≥ 150 mg. CP100 was also suitable for medium-scale productions but at a higher expenditure in terms of preculture of hybridomas and set-up of the bioreactor. Although the Tecnomouse at its full capacity (5 cassettes) has not been tested, the most appropriate system for antibody productions beyond 500 mg has been the CP2500. It is advisable for a laboratory that is specialized in MAb productions to adopt several different in vitro methods to meet different needs.

Conclusion

In vitro bioreactor systems are a viable alternative to murine ascites for laboratory-scale MAb production. Various in vitro culture systems exist that are qualitatively equal to the ascites production method. It is generally accepted that in vitro production of MAbs is preferable to producing antibodies in ascites tumor cells (NRC 1999). Members of an official ethics committee might justify the use of the ascites method as an exception when, for example, hybridoma cells must be recovered because they have failed to grow in vitro, they have become infected, or the antibodies are needed for established therapeutic purposes. The existence of validated in vitro replacements for ascites in rodents has prompted governments in Europe to impose full or partial bans on animal-based MAb production. Although the in vitro MAb production technologies are more complex than in vivo ascites methods because they require additional cell culture, they have the clear advantage that animal pain and distress are avoided.

Acknowledgments

We are grateful to Virginie Hermans and Omar Bel-Haj-Touzani, for excellent technical assistance, and to Ulrike Krause, for support in the preparation of the manuscript.

¹Abbreviations used in this article: ECS, extracapillary space; ELISA, enzyme-linked immunosorbent assay; FACS, fluorescent-activated cell sorter; GSK Bio, GlaxoSmithKline Biologicals; MAb, monoclonal antibody; SFM, serum-free medium.

References

Bruce MP, Boyd V, Duch C, White JR. 2002. Dialysis-based bioreactor systems for the production of monoclonal antibodies--Alternatives to ascites production in mice. J Immunol Methods 264:59-68.

Cadwell JJS. 2004. New development in hollow-fiber cell culture. Am Biotechnol Lab July:1-5.

Falkenberg FW. 1998. Monoclonal antibody production: Problems and solutions. 74th Forum in Immunology. Res Immunol 149:542-547.

Falkenberg FW, Weichert H, Krane M, Bartels I, Palme M, Nagels H-O, Fiebig H. 1995. In vitro production of monoclonal antibodies in high concentration in a new and easy to handle modular minifermenter. J Immunol Methods 179:13-29.

Jackson LR, Trudel LJ, Fox JG, Lipman NS. 1996. Evaluation of hollow fiber bioreactors as an altenative to murine ascites production for small scale monoclonal antibody production. J Immunol Methods 189:217-231.

Köhler G, Milstein C. 1975. Continuous cultures of fused cells secreting antibody of predefined specificity. Nature 256:495-497.

NRC [National Research Council]. 1999. Monoclonal Antibody Production. Washington DC: National Academy Press.

Ohlin M, Borrebaeck CA. 1994. Flow cytometry analysis of the stability of antibody production by human x human x mouse heterohybridoma subclones. J Immunol Methods 170:75-82.