Downstream Processing of Monoclonal Antibodies: from High Dilution to High Purity

Monoclonal antibodies (MAbs) represent the fastest growing pharmaceutical market segment. Even with conservative assumptions about growing attrition rates, substitution pressure, and margin squeezes, MAb sales will probably reach a stable plateau of $20 billion by 2010. While the commercialization of MAbs is gathering momentum, the sector is facing a worldwide shortfall of available biomanufacturing capacity that is becoming a critical strategic limitation, especially for companies without established market access.

Uwe Gottschalk

Improvements are due in all areas of the pharmaceutical supply chain but especially in downstream processing to manage current manufacturing challenges and are therefore vital for the long-term success of the sector.

Antibodies represent a class of flexible molecular adaptors that play a vital role in the adaptive immune systems of vertebrates.¹ They are bifunctional molecules with a basic symmetrical structure consisting of pairs of identical heavy and light chains linked through disulfide bridges (Figure 1).²

Figure 1. Form Follows Function: Domain Structure of an IgG Antibody

The individual chains are globular domains that are highly conserved between different immunoglobulins (constant region) or contain sequence variability (variable region). Antigen binding sites are formed by the interaction of hypervariable loop regions exposed near the N-terminus of the polypeptide chains — the complementary determining regions (CDRs), which are surrounded by relatively invariant framework residues. Their diversity (idiotypic variation) represents the central aspect of the humoral immune response. Functions that are essential for the cellular immune response, such as complement activation and lymphocyte binding, reside in the constant domains.³

Immunoglobulins (Igs) are glycoproteins that contain 3 to 12 percent carbohydrates.⁴ In an IgG molecule, the sugar part is N-linked to a highly conserved site at Asn 297 in the C_H2 domain of both heavy chains (Figure 2). Although N-linked glycosylation does not interfere with antigen recognition, a number of implications are linked to this functionality, such as: stability, pharmacokinetics, antigenicity, Fc-related effector functions, and serum stability of antibodies.⁵ Removing terminal sialic acid results in drastically reduced half-life and increased liver uptake through the asialoglycoprotein receptor.

Figure 2. Glycosylation Pattern of a Human IgG Antibody

The traditional method of MAb production generates molecules of murine origin with an immunogenic potential and a short serum half-life in humans.^7,8 Although a number of murine antibodies are still commercially available, the focus today is on chimeric, humanized, and (eventually) human antibodies, generated with the use of transgenic mice or large combinatorial libraries in bacteriophages or yeast.⁹

In a mere 30 years of development, a total of 23 MAbs and MAb-related proteins have been approved for medical treatments. MAbs represent the fastest growing segment within biopharmaceuticals and are outperforming recombinant proteins with a compound annual growth rate of 20 percent.¹⁰ Hundreds of second and third generation antibody-based products are in preclinical and clinical development.

The vast majority of today's antibody treatment concepts rely on active immunization directed to specific disease targets. A few have the agonistic effect, namely providing enzyme-like support of a physiological function. A more common effect is antagonistic through the neutralization of a signaling molecule or its specific receptor. Antagonistic treatment concepts typically require high doses in the >1mg/kg area. Common targets are soluble cytokines like IL-1 and TNF to prevent pro-inflammatory mechanisms in disorders like rheumatoid arthritis, inflammatory bowel disease, psoriasis, and autoimmune diseases such as multiple sclerosis and Crohn's disease. As a consequence, antagonistic antibodies will be required in the hundreds of kilograms with ton requirements coming in sight (Figure 3).¹¹

Figure 3. Annual Sales and Manufacturing Demands for Marketed MAbs.11

BASIC REQUIREMENTS FOR MAB MANUFACTURING

MAbs are biopharmaceuticals, and their manufacturing processes are subject to the same basic requirements as any other drug. The assessment of a licensing application focuses on the quality, safety, efficacy, and consistency of the product. The production process must be designed in such a way that it meets the highest requirements with regard to consistency and reproducibility. Process-related questions like characterization of raw materials, in-process controls, change policies, and process and assay validation are therefore very demanding, and result in the rigorous control of various critical parameters during production and testing.

¹²

Like other biological products, antibodies are large and complex molecules with isoforms and microheterogeneities. Analysis of the bulk product in ways similar to small molecules is therefore not sufficient to assure consistency.

The manufacturing of full-sized antibodies is currently restricted to mammalian cell systems that possess the correct biosynthesis and glycoslylation machinery. Of course, mammalian cell culture requires strict safety precautions due to the risk of pathogenic contamination. Traditional methods like generation of mouse ascites are no longer accepted for quality reasons and nowadays banned in most countries.

Process Development Plan

Process development follows an overall plan that coordinates all the various disciplines and facilitates a scale-up of 1,000-fold and more.

¹³

During the exploratory phase, a research method is used to generate small amounts of MAb to investigate target-specific effects

in vitro.

Important technical goals at this stage are feasibility and speed.

Manufacturing the clinical material for an Investigational New Drug (IND) requires a method established under current Good Manufacturing Practice (cGMP) with validation of some critical elements. The whole process should be designed for robustness and scalability from this point onward.

A Biologics License Application (BLA)-enabling process is typically used for the manufacturing of clinical Phase III material and for the full-scale plant. The process qualification documents contain the full validation package for the Chemistry Manufacturing and Controls (CMC) part of the application dossier, including a virus-safety concept. As one transfers into the process scale, there is a shift of emphasis towards quality and process economy, resulting in an integrated, robust, cGMP-compliant manufacturing process. At this stage, most process changes are considered to be major, leading to biochemical comparability and clinical bridging studies, and possibly to a new licence application.

FERMENTATION IS SETTING THE PACE

Routine production of MAbs under cGMP became feasible with the implementation of cell culture in industrial biotechnology. Antibody-secreting mammalian cells derived from a comprehensively characterized Manufacturers Working Cell Bank (MWCB) are cultivated in bioreactors under optimized conditions with various parameters monitored on a continuous basis. Expression of recombinant MAbs in Chinese hamster ovary (CHO) cells is the industry standard. Other common immortalized cell lines are the mouse myeloma derived NS0, baby hamster kidney cells (BHK) and PerC6 derived from human kidney.

¹⁵

Fed-batch fermentation with expression rates in the g-MAb/L range is the norm and much higher levels are in sight (Figure 4).¹⁶ Compared to the early days of antibody manufacturing, this represents a 1,000-fold improvement in volumetric productivity. Recent developments in cell culture are leading to both higher cell densities as well as higher specific production rates. There is more room for improvement, and we foresee developments achieving higher biomass and more efficient gene transfer using the whole molecular genetic repertoire.¹⁷

Figure 4. Development of MAb Expression Rate in Mammalian Cell Culture Showing Technological Advance. Adapted from Reference 15.

Glycosylation

Different posttranslational modifications may lead to the synthesis of different glycosylation variants within a cell clone and even within a single cell.

¹⁸

As a result, analysis of the MAb glycosylation pattern is a sensitive tool to demonstrate the effective control of critical parameters within a manufacturing process. It also can be a central element in biochemical comparability studies during scale up from laboratory to greater than 10,000-L fermentation runs.

¹⁹

Glycosylation is a complex biosynthetic event that is largely influenced by cell culture conditions such as the cell bank, the fermentation process (with its various control parameters), and the medium composition.²⁰ Intrinsic differences arise from the endogenous properties of the expression system related to the type of glycoenzymes and sialic acids the cells are using.

Even within a validated manufacturing process, one must accept microheterogeneities in the glycosylation pattern (with truncated forms and different overall monosaccharide compositions) as long as pre-determined acceptance criteria are met. It is vital to use validatable specifications throughout the clinical development and freeze them after the market launch of a biopharmaceutical. This requirement usually excludes major changes in the manufacturing process such as a different expression system or cultivation method unless new preclinical and clinical studies are performed.

Carbohydrate Funtionality

The first decision to make for the appropriate expression and manufacturing system is whether a full-length antibody, including the full carbohydrate functionality, is required for the biological function. In its aglycosylated form, an antibody can support all antagonistic functions that are required without the need for a human-like glycosylation. Mutants can easily be expressed in a number of host cells without the need for additional post-translational modifications.

If, however, the carbohydrate functionality is required, hybridoma or myeloma expression systems (for genetically engineered antibodies) are the expression systems of choice for full-length antibodies.²¹ Most of these cell lines are of rodent origin and this accounts for various host-cell related differences in the glycan pattern, with the use of different forms of sialic acid being the predominant one. CHO cells are able to perform most glycosylation steps comparable to the human profile. Even using human cell lines cannot guarantee a reproducible pattern because of the high influence of the manufacturing process parameters. Increase of cell density and yield optimization interfere with the overall metabolism, including induction of glycosyltransferases and oligosaccharide formation.

One aspect of modern cell line development focuses on the glycosylation pattern, as overexpressed glycoproteins tend to contain truncated oligosaccharide chains due to incomplete posttranslational modification.²² Cell lines with overexpressed glycosyltransferases have the potential to generate a greater homogeneity through the reduction of terminal GlcNAc and the increase of sialylation.²³

DOWNSTREAM PROCESSING: SLOW EVOLUTION OF TECHNOLOGY

Although antibodies from the same class and subclass are considerably different, basic elements of the purification strategy have a generic character and allow for a rational approach in process development (Figure 5). Typically every process step addresses a certain task in the overall strategy, and redundancy is avoided wherever possible. Development tools like design of experiments, process modeling, and simulation are employed to adapt and optimize the individual steps and to identify potential bottlenecks in the overall process. Basic principles like the Capturing – Intermediate Purification – Polishing (CIPP) strategy are key in order to optimize quality, yield, and overall productivity.

²⁴

Figure 5. A Generic Manufacturing Strategy for Monoclonal Antibodies.

Moving from high dilution to high purity is the principle of a good downstream process. Its priority is to concentrate and thus stabilize the product very early on and to remove critical impurities throughout the process. We also keep a focus on selectivity and resolution towards the final steps of bulk manufacturing (drug substance).

Antibody molecules vary in their solubility and stability against chemical and physical influences. Isoelectric points — the pH at which a molecule in a solution will no longer move in an electric field because it no longer has a net electric charge — are typically between 5 and 9, mainly due to the different sialic acid content. Purification techniques are not considerably different from bioseparation of other proteins as they focus mainly on bioaffinity, charge, hydrophobicity, and size. See Jungbauer for a good review.²⁵

Initial recovery (capturing) and polishing are the most critical phases that address the key features of biotech products. The high dilution of the target molecule after biosynthesis is one of the reasons there is higher processing cost when compared to making chemical drugs. Modern bioseparation processes focus on a fast and efficient isolation of the product in a robust and productive capturing step to facilitate rapid volume reduction as well as separation, and therefore stabilization of the antibody.

In a typical MAb purification process, the first step after the harvest is a clarification step that yields a cell-free feed stream to be applied to initial recovery. From a processing standpoint, the harvest is a highly diluted solution that causes handling issues. Modern capturing-supports in chromatography combine some selectivity for initial purification with excellent flow properties (low back pressure at high linear flowrates), and high dynamic capacity. Typical linear velocities are above in the area of 500 cm/hr and can be used for feed streams of up to 500 column volumes.

There are a lot of proposed schemes for the initial recovery of Mabs, but none of them has provided a real breakthrough yet. We instead rely upon an affinity chromatography step. This workhorse technology provides very high selectivity for the target molecule.²⁶ The older common packing for the purification of most IgG species is Protein A, a 42,000-Dalton protein derived from a strain of Staphylococcus aureus. Immobilized Protein A has been used extensively as an affinity support for the purification of a wide variety of IgG molecules from many different species of mammals.²⁷

Protein A consists of a single polypeptide chain that is structurally and functionally composed of two parts. The N-terminal region contains four homogenous sub-regions, each of which can bind human IgG molecules, with a total of two active binding sites at a time. (The approximate association constant [K_a], used to measure the strength of a binding between an antibody and antigen, is 10⁸ M^-1.)²⁸ The natural molecule is located on the outer membrane surface of S. aureus and is linked to the cell surface through its c-terminal region.²⁹ This allows the pathogen to bind between the C_H2 and C_H3 domains of the IgG, and thus prevent interaction of the Fc part of the antibody with effector cells, thereby circumventing activation of the immune system.³⁰

Today a truncated, recombinant version of Protein A with a molecular mass of approximately 32,000 Daltons is the standard. It exhibits the same affinity for IgG molecules as the native protein but has considerably lower non-specific binding properties. This protein is genetically engineered, allowing for a site directed immobilization through introduced, N-terminal bridging groups. Recombinant Protein A no longer requires affinity purification on immobilized virus-inactivated human IgG, leading to an increased virus safety of the product.

Protein A as a biological reagent is still raising concerns as to the high cost of the matrix, the problem of leakage into the product, the cleanability after repeated use, and especially the sensitivity to caustic solutions that are widely used as bacteriostatic agents.³¹ Most recent variants of Protein A support focus on the long-term stability in caustic cleaning agents like 0.1 M NaOH and increased dynamic capacity.³²

Meanwhile, a number of other characterized immunoglobulin-binding proteins like Protein G from Streptococcus pyogenes, Protein L from Peptostreptococcus magnus, and the genetically engineered Protein Z offer different specificities and can be used for other antibody classes and also fragments.^33,34 However, protein affinity sorbents are responsible for a significant part of the variable costs in manufacturing, and various strategies have been evaluated to identify chemical pseudoaffinity ligands that bind and elute MAbs selectively.^35,36 A number of biomimetics directed to the Protein A binding domain in the C_H2-domain, based on peptide libraries, combinatorial chemistry, and rational ligand design, have been synthesized and studied for MAb purification.³⁷

Polishing

Final purification via polishing is vital. The biomanufacturing environment provides an excellent basis for the growth of all kinds of organisms and their metabolic products such as viruses, DNA, host cell proteins (HCP), and endotoxins as well as the growth of process-derived contaminants and impurities. In addition, antibodies are complex macromolecules with isoforms and microheterogeneities that require consistency within predefined limits.

Reliable polishing tools are generic platforms that provide high resolution and meet the highest safety standards for selective removal of endogenous and adventitious viruses, prions, and other human pathogenic agents.³⁸ Most contaminants have an acidic PI and can be removed through either binding the antibody on a cation exchange (CEX) support or further downstream in a flow-through mode of an anion exchange (AEX) chromatography at very high linear velocities (Table 1).³⁹

Table 1. Physical Properties of MAbs Compared to Typical Processing-derived Contaminants39

Chromatographic supports are diffusion limited and flow-through columns must be scaled based on the flow rate rather than on binding capacity. Membrane adsorbers with a more open structure and virtually no diffusion limitation provide a robust alternative. They can remove a whole set of contaminants in one step and a number of concepts have been described (Figure 6).⁴⁰ The maximum separation power with membrane chromatography is achieved at high dilution of the target molecules and thus is suitable in capturing and polishing.⁴¹ In addition, these membrane-based tools can be discarded after a single use, making them very interesting from a cleanability and from a process economy standpoint.⁴²

Figure 6. Structural Features of Membrane Adsorbers versus Resin-based Media40

Clear Out the Viruses

For final product analysis, evidence must be provided with validated bioanalytical quality control methods that besides the correct identity and homogeneity, critical impurities have been reduced below the specified limits.

⁴³

Table 2 offers one method of setting up the release specifications. For a validated, product-related HCP assay with sensitivity in the 10 to 100 ppm range, mock fermentation and purification runs are performed to obtain an HCP-specific antiserum.

⁴⁴

Table 2. Major Process-derived Contaminants in MAb Production

Most immortalized cell lines have been shown to express endogenous retrovirus-like particles that may be replication-competent and infectious. Therefore MAbs derived from mammalian cell culture require a reliable virus safety strategy. A good one covers three levels:

• Characterization of host cell line and raw material sources

• Testing of product from various manufacturing stages

• Validation of the virus clearance capability of the production process.

The robust and reliable ability to eliminate viruses must be demonstrated in a risk-based approach.⁴⁶ Virus validation studies employ model viruses that are suitable for representing known risks from the sources involved.^47,48 Managing the unknown is playing an important role in the design of these studies.

Today's requirements ask for a statistically independent combination of methods for the clearance of enveloped and non-enveloped viruses based on the different physical principles of removal and inactivation and yet are complementary to each other.^49,50 Size-exclusion based virus filtration is used in most of the existing processes. Among the different virus inactivation methods, treating with high temperature for a short time (HTST) has been used for retrovirus clearance, whereas UV-irradiation is most powerful for the elimination of small, non-enveloped and otherwise very resistant viruses like Porcine Parvovirus (PPV).^51,52

Partitioning steps are considered less robust as they are influenced by the actual process conditions. In any case, scaled-down models must be designed by carefully representing the process conditions and accounting for aging of support during column lifetime.⁵³

EMERGING ISSUES IN MAB PRODUCTION

Biomanufacturing is a risky and fixed-cost driven business.

⁵⁴

A highly regulated environment, long lead times, high capital investments, and demanding processes dictate actions. The fallout from the recent failure of Antegren — a MAb for Multiple Sclerosis and Crohn's Disease treatment that received fast track review status in the US after only one year of Phase III human trials — is a good example.

There are forces driving us towards higher efficiency and productivity in MAb production. Among these are genomic research and still unmet medical needs; limited GMP manufacturing capacity; growing competition between companies and products; economic problems of the healthcare systems; and unrelenting demands for higher quality. Product development and revenue generation are major drivers for the growth of biopharmaceutical companies (colloquially expressed as wanting a full pipeline).^55,56

Efficient development and use of technology is an important factor in both upstream and downstream processing. Right now the sector is lacking the infrastructure for the manufacture of antibodies in the development stage. Despite major investments, the high demand and long lead times prohibit a short-term solution to close the current gaps in biopharmaceutical supply chains.

The key measurable aspects of desirable advances are low cost of goods, higher expression rates, optimized product yields, and robust, scalable manufacturing operations that do not compromise product quality. Proteinanalytical methods that allow for comparability studies between different production versions, scales, and manufacturing sites are needed as part of the technology package. Whole-process design with the integration of upstream and downstream processing will become the preferred scenario. Industry will adapt simple generic-platform technology modules that will provide the basis for economically viable biomanufacturing — a vital essential for the eventual success of MAbs.

Improved Process Platforms

Recent advances in fermentation development will bring mid-term relief in biomanufacturing. In addition, alternative expression systems for antibodies and fragments may provide the basis for very large-scale applications at low cost.

⁵⁷

For antibody fragments, microbial expression in

Escherichia coli, Pichia pastoris

and

Hansenula polymorpha

is state of the art with production levels at 1 to 2 g/L in high cell-density fermentation.

^58,59

It is unlikely that mammalian cell culture (today's standard technology) is capable of overcoming some of the inherent limitations such as the scale up, long culture periods, capital costs, and sterility requirements for these very large-scale applications.

Transgenic animals and plants combine the advantage of full-length expression of MAbs with very high biomass generation, and they output a cell density one hundred times higher than in any bioreactor.⁶⁰ Cell lines from animals suffer from some unresolved issues like pathogen threat and complex matrices that are difficult to process.⁶¹ As a result, there are no biopharmaceuticals from such transgenic sources on the market.

Species-specific glycosylation is another problem with the transgenic production of human MAbs when Fc-related properties are required. Plant-derived glycoproteins contain beta-linked xylose sugars that may have some immunogenic potential.⁶² With growing demand for low-cost antagonistic antibodies, the full-length expression of aglycosylated MAbs in non-edible plants such as alfalfa, tobacco, lemna, safflower, and moss will become a vital alternative.⁶³

Industrial-scale culture of plant cells is an interesting option.⁶⁴ MAb applications with annual production of more than 1,000 kg are in sight for the continuous treatment of autoimmune diseases; for topical and oral administration as vaccines; for antiinfective agents; and outside the medical field in the industrial genre, for use as catalytic and immunopurification reagents.⁶⁵ Looking ahead to robust large-scale manufacturing processes, MAbs will be produced in bulk amounts and deliver on the potential to be eventually used as therapeutic agents in commodities like toothpastes and shampoos.⁶⁶

The current situation in biomanufacturing is that fermentation development is setting the pace in terms of productivity. This will be further accentuated with alternative expression systems. Innovative downstream processing technology that has the potential to accommodate these improvements is desperately needed to tackle the current backlog in bioseparation. However, up- and downstream processing are following different paths and there is little doubt that productivity increase in fermentation is shifting the bottleneck downstream. Since downstream processing costs account for up to 80 percent of overall production costs, there is no way to ignore them.⁶⁷

Revisiting robust technology of the "low-end" type is a potential solution to this dilemma. Examples are extraction, precipitation, crystallization, and filtration. These work well with small molecules and commodity proteins. A robust downstream process must be simple, with quality, yield, productivity, and overall economics as the primary goals. An integrated design with compatible unit operations looks to be the answer.

Think Big

The unit manufacturing costs of industrial goods all follow a clear economy of scale relationship. Steiner made a calculation for a MAb with an annual demand of approximately 100 kg, resulting in an overall cost range in the vicinity of $100/g. Approximately 75 percent is related to fixed costs and the remainder from variable costs related to process consumables.

⁶⁸

Time-to-market is still the most important driver, but return-on-investment calculations are vital for the management of a full pipeline. It may be bad business for companies to lock up their financial resources into hardware investments on the basis of preliminary data with the risk that capital costs will kill the product and eventually also the company.

The result is a clear trend towards disposable manufacturing as another paradigm shift in biomanufacturing.⁶⁹ Use of disposables during early production phases has recently been reported as a strategy to avoid the economic and quality risks accompanying upfront hardware investments.⁷⁰

Uwe Gottschalk, Ph.D., is vice president of purification technologies at Sartorius AG, Biotechnology Division, Weender Landstrasse 94-108, Göttingen, Germany, 37075, tel: 49.551.3082016, fax: 49.551.3082835, Uwe.Gottschalk@sartorius.com.

REFERENCES:

1. Jerne KJ. The Immune System.

Scientific American.

1973; 229:52-60.

2. Padlan EA. Anatomy of the antibody molecule. Mol. Immunol. 1994; 31:169-217.

3. Hulett MD, Hogarth PM. Molecular basis of Fc receptor function. Adv. Immunol. 1994; 57:1-127.

4. Jefferis R, Lund J. Glycsilation of antibody molecules, Structure and functional significance. Chem. Immunol. 1997; 65:111-128.

5. Ibid.

6. Siberil S, Teillaud JL. Future Prospects in antibody engineering and therapy. In: Subramanian G. editor. Antibodies, Volume 2, Novel Technologies and Therapeutic Use. New York: Kluwer Academic, 2004: 199-215.

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

8. Shawler DL, Bartholomew RM, Smith LM. Human immune response to multiple injections of murine monoclonal IgG. J. Immunol. 1985; 135:1-12.

9. Little M, Kipriyanov SM, Le Gall F, Moldenhauer G. Of mice and men, hybridoma and recombinant antibodies. Immunol. Today 2000; 21(8):364-70.

10. Robinson K., An industry comes of age. BioPharm International 2002 Nov; 15:20-24.

11. Milroy D, Auchincloss C. Monoclonal Antibodies – on the Crest of a Wave. Horizons. Wood Mackenzie, Boston MA; 2003 Nov; 6.

12. Federici MM. The quality control of biotechnology products. Biologicals 1994; 22:151-159.

13. Rathore A, Velayudhan A. Guidelines for optimisation and scale up in preparative chromatography. BioPharm International 2003 Jan; 16(1):34-42.

14. Little M, et al. op. cit.

15. Wurm FM. Production of recombinant protein therapeutics in cultivated mammalian cells. Nature Biotechnology 2004; 22:1393-1398.

16. Wurm FM, Griffith J. Mammalian Cell Culture. In: Meyers RA editor. The Encyclopaedia of Physical Science and Technology, Third Edition. Burlington MA; Academic Press; 2002; 9:31-47.

17. Wurm FM, Jordan M. Gene transfer and gene amplification in mammalian cells. In: Makrides SC editor. Gene Transfer and Expression in Mammalian Cells. New Comprehensive Biochemistry, Volume 38; Burlington MA; Academic Press; 2003:309-335.

18. Deng XK, Raju TS, Morrow KJ. Achieving appropriate glycosilation during the scale up of antibody production. In: Subramanian G. editor. Antibodies – Novel Technologies and Therapeutic Use, Volume 2. New York, Kluwer Academic; 2004: 53-78.

19. Jenkins N, Parekh RB, James DC. Getting the glycosilation right, implications for the biotechnology industry. Nat. Biotechnol. 1996; 14:975-981.

20. Patel TP, Parekh RB, Moellering BJ, Prior CP. Different culture methods lead to differences in glycosilation of a murine IgG monoclonal antibody. Biochem. J. 1992; 285:839-845.

21 Yoo EM, Koteswara R, Chintalacharuvu ML, Morrison SL. Myeloma expression systems. J Immun. Methods 2002; 261: 1-20.

22. Stanley P. Glycosilation engineering. Glycobiology 1992; 2:99-107.

23. Weikert S, et al. Engineering Chinese hamster ovary cells to maximize sialic acid content of recombinant glycoproteins. Nat. Biotechnol. 1993; 17:1116-1121.

24. Grund E. Advances in downstream processing. Presented at Scaling up of Biopharmaceutical Proteins. IBC Life Sciences, Basel, January 23., 2003.

25. Jungbauer A. Preparative chromatography of biomolecules. J. Chrom. 1993; 639:3-16.

26. Wilchek M, Chaiken I. An overview of affinity chromatography. Methods Mol. Biol. 2000; 147:1-6.

27. Huse K, BÖ HJ, Scholz GH. Purification of antibodies by affinity chromatography. J. Biochem. Biophys. Methods 2002; 51: 217-231.

28. Suralia A. Interaction of Protein A with the domains of the Fc-Fragment. TIBS 1982; 7:318-323.

29. SjÖst J, Movitz J, Johansson I-B. Localization of protein A in Staphylococcus aureus. Eur. J. Biochem. 1972; 30:190.

30. Suralia A, op. cit.; 318-323.

31. Iyer H, et al. Considerations during development of a protein A-based antibody purification process. Biopharm International 2002 Jan; 20:14-20.

32. Bergander T, Chirica L, LjunglÖof, Malmquist G. Novel high capacity Protein A affinity chromatography media. 3rd International Symposium on Downstream Processing of Genetically Engineered Antibodies and Related Molecules. Nice, France; 2004.

33. Raeder R, Boyle MD. Analysis of immunoglobulin G-binding –protein expression by invasive isolates of Streptococcus pyogenes. Clin. Diagn. Lab. Immunol 1995; 2:484-486.

34. Bjorck L, Protein L. A novel bacterial cell wall protein with affinity for IgL chains. J. Immunol. 1988; 140:1194-1197.

35. Curling J. Affinity chromatography — from textile dyes to synthetic ligands by design, Parts I and II, BioPharm International 2004, July and August; 17(7):34-42 and 17(8):60-66.

36. Boschetti E. The use of thiophilic chromatography for antibody purification, a review. J. Biochem. Biophys. Methods 2001; 49:361-389.

37. Li R, Dowd V, Stewart DJ, Burton SJ, Lowe CR. Design, synthesis, and application of a Protein A mimetic. Nature Biotechnology 1998; 16:190-195.

38. Fish B, Concepts in development of manufacturing strategies for monoclonal antibodies. In, Subramanian G. editor. Antibodies, Volume 1, Production and Purification. New York: Kluwer Academic, 2004: 1-23.

39. Hanna LS, Pine P., Reuzinsky G, Nigam S, Omstead DR. Removing specific cell culture contaminants in a Mab purification process. BioPharm International 1991 Oct; 4:33-37.

40. Gosh R. Protein separation using membrane chromatography, opportunities and challenges. J. Chrom. 2002; 952:13-27.

41. Gottschalk U, Fischer-Fruehholz S, Reif O. Membrane adsorbers, a cutting edge process technology at the threshold. BioProcess Intl. 2004; 2:56-65.

42. Warner TN, Nochumshon S. Rethinking the economics of chromatography. BioPharm International 2003 Jan; 16:58-60.

43. FDA. Points to consider in the manufacture and testing of monoclonal antibody products for human use. Bethesda MD 1997 Feb. 27. Available at: www.fda.gov/cber/gdlns/ptc_mab.txt

44. Hoffmann K. Strategies for host cell protein analysis. BioPharm International 2000; 13:38-45.

45. Richter A, Jostameling M., MÜller K, Herrmann A, Pitschke M. Quality control of antibodies for human use. In: Subramanian G. editor. Antibodies, Volume 1, Production and Purification. New York; Kluwer Academic, 2004:169-188.

46. EMEA. Committee for Proprietary Medicinal Products (CPMP). ICH Q5A. Note for guidance on quality of biotechnology products, viral safety evaluation of biotechnology products derived from cell lines of human or animal origin. CPMP/ICH/295/95; London; 1997.

47. EMEA. CPMP. Note for guidance on virus validation studies, the design, contribution and interpretation of studies validating the inactivation and removal of viruses. CPMP/BWP/268/95; London; 1995.

48. Sofer G. Virus Inactivation in the 1990s – Part 4, culture media, biotechnology products, and vaccines. BioPharm International 2003 Jan; 16:50-57.

49. Walter JK, Nothelfer F, Werz W. Validation of viral safety for pharmaceutical proteins. In: Subramanian G. editor, Bioseparation and Bioprocessing, Volume 1. Weinhemim Germany; Wiley-VCH, 1998; 465-496.

50. Sofer G, Lister DC, Boose JA. Inactivation methods grouped by virus. BioPharm International. Virus inactivation in the 1990s – and into the 21st Century. 2003 June, Supplement S37-S42.

51. Plavsic AM, Bolin S. Resistance of porcine circovirus to gamma irradiation. BioPharm International 2001 April; 14:32-36.

52. Wang J, et al. Virus inactivation and protein recovery in a novel ultraviolet-C reactor, Vox Sanguinis 2004; 86:230-238.

53. Sofer G. Validation, Ensuring the accuracy of scaled-down chromatography models, BioPharm International 1996 Oct.:51-54.

54. Gottschalk U. New and unknown challenges facing biomanufacturing, BioPharm International 2005 March; 18(3):24-28.

55. Sinclair, A, Biomanufacturing capacity, Current and future requirements. J. Commercial Biotechnology, 2001, 8, 43-50.

56. Gottschalk U. Biotech manufacturing is coming of age. BioProcess Intl. 2004 April; 1:54-61.

57. Verma R, Boleti E, George AJ. Antibody engineering, comparison of bacterial, yeast, insect and mammalian expression systems. J. Immunol. Methods 1993; 216:165-181.

58. Humphreys DP. Production of antibodies and antibody fragments in Escherichia coli and comparison of their functions, uses and modifications. Curr. Opin. Drug Discov. Develop. 2003; 6:188-196.

59. Joosten V, Lokman C, Hondel C, Punt PJ. The production of antibody fragments and antibody fusion proteins by yeasts and filamentous fungi. Microbial. Cell Factories 2003; 2:1-15.

60. Fischer R, Eman N. Molecular farming of pharmaceutical proteins. Transgenic Res. 2000; 9:279-299.

61. Pollock DP, Kutzko JP, Brick-Wilson E, Williams JL, Echelard Y, Meade HM. Transgenic milk as a method for the production of recombinant antibodies. J. Immunol. Methods 1999; 231:147-157.

62. Hiatt S. Monoclonal antibody engineering in plants. FEBS Lett. 1992; 307:71-75.

63. Peeters K, De Wilde C, De Jaeger G, Angenon G, Depicker A. Production of antibodies and antibody fragments in plants. Vaccine 2001; 19:2756-2761.

64. Hellwig S, Drossard J, Twyman RM, Fischer R. Plant cell cultures for the production of recombinant proteins. Nature Biotechnology 2004; 22:1393-1398.

65. Abiko Y. Passive immunization against dental caries and periodontic disease, development of recombinant and human monoclonal antibodies. Crit. Rev. Oral Biol. Med. 2000; 11:140-158.

66. Joosten V, et al. op. cit.

67. Sadana A, Beelaram A. Efficiency and economics of bioseparation, some case studies. Bioseparation 1994; 4:221-235.

68. Steiner U. The business case for plant factories. Plant Conference Quebec 2003.

69. Hodge G.: Disposable components enable a new approach to biopharmaceutical manufacturing. BioPharm. International 2004; 15:38-49.

70. Warner TN, Nochumshon S, Rethinking the economics of chromatography. BioPharm International 2003 Jan; 16 58-60.