Production Strategies for Antibody Fragment Therapeutics

ABSTRACT

Full-length antibodies have captured a significant fraction of the sales volume and value of the biopharmaceuticals market. With increasing knowledge, however, it has been recognized that full-length antibodies may not always be necessary or even desirable. Antibody fragments provide the opportunity for new therapeutic possibilities. Moreover, antibody fragments have the potential for simpler, high-yielding production processes, which can translate into a lower manufacturing cost-of-goods and extended therapeutic benefit. This article discusses the structure and production strategies available for various types of antibody fragment therapeutics.

With estimated sales of $30 billion in 2007, therapeutic antibodies as a class have validated the effectiveness of using highly specific binding properties as therapies.¹ Most therapeutic antibodies today are full-length IgG molecules because they are structurally stable, have long in vivo half-life and Fc-mediated biological properties. They have a number of drawbacks as therapeutics, however, which include:

• They have high production costs arising from their expression in mammalian cells.

• They are unsuitable for intracellular targets because of their size.

• Binding can be variable depending on the protein targeted (best suited to relatively flat target surfaces).

• Their size can lead to poor tissue penetration arising from their size, which can limit their effectiveness in cases such as tumor penetration.

• The intellectual property landscape around antibody-based therapeutics discovery and development is complex. Prohibitive "royalty stacks" often affect the profitability of a resultant drug.

Avecia Biologics Ltd.

Antibody fragments, on the other hand, can provide highly specific binding as well as a number of therapeutic advantages including:

• flexibility in structure: a diverse range of binding structures can be assembled from antibody-binding domains giving bispecific or greater binding possibilities per molecule

• enhanced penetration of tissues (particularly of solid tumors) because of their smaller size

• elimination of immunogenicity to the Fc region (fragment structures can be designed without an Fc region)

• simpler production systems. Fragment antibody production is possible with microbial systems such as E. coli and yeasts, translating into lower cost-of-goods (CoGs) and wider application of the benefits of such therapeutic molecules.

The therapeutic effectiveness of antibody fragments has been demonstrated by the licensure of the first antibody fragments. Genentech's Lucentis, a fragment of Avastin, was approved by the FDA in 2006 and is used for the treatment of wet age-related macular degeneration.² In 2008, UCB Pharma's certolizumab pegol (Cimzia), a PEGylated antibody fragment, was approved for the treatment of Crohn's disease in the US.³ Table 1 shows examples of antibody fragment products that have been launched or are in development. This article discusses production strategies available for antibody fragment therapeutics.

Table 1. Examples of antibody fragment products on the market or in development

Background

The term antibody fragments covers a large number of molecular structures derived from full-length antibodies. Some clarification of structural types is useful before discussing production aspects.

IgG Structure

IgGs are the most abundant immunoglobulins in the blood and are commonly used as antibody therapeutics. They have a molecular weight of approximately 160 kDa comprising 12 domains of roughly equal size. They are made up of four separate peptide chains, two identical heavy (H) chain polypeptides and two identical light (L) chain polypeptides (Figure 1). The L and H chains are composed of two and four domains, respectively, with each domain having a similar β-barrel structure and containing one disulfide bridge. The H and L chains are linked together by a combination of disulfide and non-covalent bonds.⁴ Each chain contains both variable and constant domains. The variable domains of the H and L chain (VH and VL) contain the variable complementarity-determining regions (CDRs). These regions of extremely variable amino acid sequences are located at the N-terminal part of the antibody molecule. Together, VH and VL form the unique antigen-recognition site. The amino acid sequences of the remaining C-terminal domains are much less variable.

Figure 1. An explanatory schematic of whole antibody domain structure

Fc Region

The Fc region, also referred to as the constant domain, is the nonantigen-binding part of the antibody molecule. Fc mediates several immunological functions, such as binding to receptors on target cells and triggering effector functions that eliminate the antigen. There are applications where the Fc-mediated effects are not required or are even desirable. The Fc fragment also is the site of glycosylation in IgGs.

Antigen-Binding Region

The unique antigen-binding site of an antibody consists of the H and L chain variable domains (VH and VL). Each domain contains four conserved framework regions and three CDR regions, which determine the specificity of the antibody. The VL and VH domains together form a binding site, which binds a specific antigen.

Antigen-Binding Fragments

A wide variety of functional antigen-binding fragments has been constructed from whole antibody domains, some of which are illustrated in Figure 2 (see also reference 5). Selected fragment types are discussed in detail below.

Figure 2. Antibody fragment diversity

Fab Fragments

Fab fragments (fragment antigen binding) are the antigen-binding domains of an antibody molecule, containing two amino acid chains composed of two domains, VH + CH1 and CL + VL. An interchain disulfide bond present at the C terminal of each constant domain links the CL and CH1 domains together, which gives a molecular weight of the heterodimer around 50 kDa.⁶ Lucentis and Cimzia are both examples of Fab antibody fragments.

Single-Chain Fv Fragments

A single-chain Fv fragment (scFv) is the smallest fragment (~30 kDa) that still contains the complete antigen-binding site (VH + VL) of a whole IgG antibody.⁷ Linker sequences joining the VH and VL domains in a single amino acid chain have been used to overcome the fact that native scFv fragments are unstable and tend to dissociate from one another.

Antibody Fab and scFv fragments, comprising both VH and VL domains, retain the specific monovalent antigen-binding affinity of the parent IgG while showing improved pharmokinetics for tissue penetration. However, these are not the smallest form of antibody fragments.

Single Domains

The smallest antibody-derived binding structures are the separate variable domains. Earlier, isolated domains were not of practical interest because of poor solubility and low affinity compared to the parent antibody.⁸

However, discoveries were made later that certain types of organisms, the camelids and cartilaginous fish, possessed high affinity single V-like domains mounted on an Fc equivalent domain structure as part of their immune system.^9,10 The V-like domains (called VhH in camelids and V-NAR in sharks) typically display long surface loops, which allow penetration of cavities of target antigens. They also stabilize isolated VH domains by masking hydrophobic surface patches.¹¹

Human V domain variants have been designed using selection from phage libraries and other approaches that have resulted in stable, high binding VL- and VH-derived domains.

Multivalent Designs

Engineering the monovalent structures (Fab, scFv, V-domain) in multivalent structures can increase functional affinity (avidity). Such multispecific molecules allow direct association of two or more different targets engineered into dimeric, trimeric, or tetrameric conjugates, either chemically of genetically.^12,13

Other Structures

Other antibody-derived structures include Genentech's one-armed antibodies and Genmab's Unibody, which omit parts of the IgG structure to give specific therapeutic benefits.^14.15

Production Technologies

Expression Options

Mammalian cell expression has been used extensively to produce full-length IgGs that have been appropriately glycosylated. However, for antibody fragments, which lack the Fc region with its N-linked glycans, microbial systems are the most effective production system for the following reasons:¹⁶

• easy and rapid strain generation

• short fermentation time

• simple fermentation media

• robustness and scalability of microbial fermentation

• no virus testing requirements or viral reduction steps.

These factors can reduce therapeutic costs by streamlining development timelines and reducing manufacturing CoGs.

Fragment production with microbial systems has been demonstrated with E. coli, yeasts, and fungi. Filamentous fungi has been reported for antibody fragment expression but is not a recommended system because of high risks of product proteolysis and high culture viscosities.¹⁷ From a regulatory perspective, a fungal system lacks the history of use in therapeutic proteins, which exists for E. coli and yeasts.

With yeast systems, correctly folded product is secreted into the culture medium.^18–20 The disadvantages of yeast expression compared with E. coli expression include longer fermentation time, the potential for N glycosylation that will be nonauthentic and high in mannose, and the potential for proteolytic clipping by host proteases.

Fragment production in E. coli is preferably achieved through secretion into the oxidizing periplasmic space, which results in an authentically formed antibody fragment. The technology has been demonstrated not just for single-chain containing fragments but also for more complex structures such as Fabs. In the latter case, both H and L chains express separately into the periplasm where self assembly and refolding of the Fab takes place.

Thus, E. coli is a highly suitable and versatile expression system with many desirable features for antibody fragment production, including significant regulatory experience for therapeutic protein production.

Antibody Fragment Production by E. coli Expression

To facilitate developers of antibody fragment-based therapeutics, Avecia Biologics has developed the pAVEway system, an expression system in E. coli.

The pAVEway system offers biopharmaceutical developers key development needs including the delivery of clinical material for clinical and nonclinical development; processes providing optimized CoGs for commercial manufacture; and the application of generic methods.

Production strain selection is a critical early target in process development. Although it is possible to change the production strain at any stage in clinical development, this can lead to repeated clinical trials, with subsequent additional costs in both time and money. These costs increase substantially for later changes. It is also important to select an appropriate strain to avoid problems later in development, including process development, fermentation scale-up, purification, and manufacturing. Many expression systems and strains used at laboratory scale are not suitable for large-scale fermentation.

The pAVEway system consists of a combination of expression vectors, selected host strains, and generic fermentation processes, which can be used to go from a product gene to a high yield fermentation process for proteins, including antibody fragments, within one month. A typical pAVEway expression development study is outlined in Table 2.

Table 2. A typical pAVEway expression development study-achievable timeline to from gene to process.

Expression Vector

The pAVEway system is tightly controlled, which means that protein expression in the absence of induction is very low, allowing tight control of fermentation conditions. This tight control is achieved by optimizing the spacing between the operator sequences (as perfect palindromes) to allow DNA looping to take place (Figure 3). In addition, the rate of protein expression can be directly controlled by the concentration of the inducer (Figure 5). This is particularly important for the expression of soluble and secreted proteins, in which the best yields are obtained when the rate of expression can be tuned to the folding and secretion capacity of the cell. This effect on expression rate can optimize the folding and secretion of a number of different antibody fragment structures and other proteins.

Fig 3. DNA loop formation to achieve tight expression shut off

The basis for this tight expression control is the use of perfect palindromic operator sequences. The use of these sequences from T7 RNA polymerase-based promoters (as used in the pET system) has been extended to a range of E. coli RNA polymerase dependent promoters. These sequences have a much wider host range and avoid the use of the λDE3 lysogen that can give rise to lytic phage, which would present problems for multiproduct manufacturing facilities. Using these control elements in combination with strong promoters gives protein titers equivalent to those seen for T7-based systems in shake flasks. Additionally, the pAVEway system also produces low levels of product in the absence of induction (Figure 4). This reduced leakiness leads to superior stability of the strains and enables rationally designed fermentation protocols.

Fig 4. Properties of pAVEway expression vectors. No leaky expression is observed, even after overnight incubation.

To confirm the usefulness of the pAVEway systems, the cloning and expression of a variety of antibody fragment types (single chain and Fab) have been demonstrated and then moved to fermentation (Table 3). Intracellular soluble and insoluble protein expression have been demonstrated and are fully scaleable. Note that the active secretion protocol is different from the intracellular protocol.

Fig 5. Properties of pAVEway expression vectors - the modulatable response of expression rate to the inducer concentration.

Post-fermentation Processing

Avecia also has assembled an antibody fragment toolkit to address the needs of post-fermentation processing. Components of the toolkit include those described below:

Table 3. Antibody fragment structures expressed to date using pAVEway

Refolding Screens

Although soluble expression is targeted as the preferred option, the diversity of antibody fragment properties means that cannot always be achieved and insoluble expression may be the most effective way forward. There is a need, therefore, to have refolding approaches that can identify high-yielding refolding approaches and are suitable for large-scale applications.

Pegylation

As discussed earlier, the small size of some antibody fragments allows more effective access to disease targets than full-length antibodies. However, this also means that serum permanence is low (~2 h).²² In some therapeutic applications, this rapid clearance is desirable, while for others a longer in vivo half-life is necessary. Protein conjugation to polyethylene glycol (PEG) is a recognized approach to extend half-life and has a 20-year history of use in human therapies.²³ Cimzia is an example of the application of PEGylation to an antibody fragment for desired pharmokinetics.

The application of novel approaches to PEGylation is beneficial for production. The PEGylation of disulphide bridges developed by Polytherics (Figure 6) is a technology which is likely to have application in the field of antibody fragments.^24,25

Fig 6. Polytherics' novel disulphide bond pegylation chemistry

Conclusion

Antibody fragments are a rapidly developing therapeutic area underpinned by the success of whole antibodies and scientific insights, allowing the generation of a diverse range of antibody fragment domain-based therapeutics. Fragment antibody production is possible with microbial systems such as E. coli and yeasts, translating into lower CoGs and wider application of the benefits of such therapeutic molecules.

J.M.LIDDELL, PhD, is head of process science at Avecia Biologics Ltd, Billingham, UK, +44 (0)1642 364016, john.liddell@avecia.com

References

1. Levine HL. Presentation at Cambridge Healthcare Institute 8th Annual PepTalk Conference; 2009 Jan; San Diego, CA.

2. Kenneth TE, Kertes PJ Ranibizumab in neovascular age-related macular degeneration. Clin Interv Aging. 2006;1:451–66.

3. Rutgeerts P, Schreiber S, Feagan B, Keininger DL, O'Neil L, Fedorak RN Certolizumab pegol. A monthly subcutaneously administered Fc-free ant-TNF alpha, improves health related quality of life in patients with moderate to severe Crohn's disease. Int J Colorectal Dis. 2008;23:289–96.

4. Padlan EA. Anatomy of the antibody molecule. Mol Immunol. 1994;31169–217.

5. Holliger P, Hudson PJ. Engineered antibody fragments and the rise of single domains. Nat Biotechnol. 2005;23:1126–36.

6. Better M, Chang CP, Robinson RR, Horwitz AH. Escherichia coli secretion of an active chimeric antibody fragment. Science. 1988; 240:1041–43.

7. Skerra A, Pluckthun A. Assembly of a functional immunoglobulin Fv fragment in Escherichia coli. Science. 1988;240:1038–41.

8. Ward ES, Gussow D, Griffiths AD, Jones PT, Winter G. Binding activities of a repertoire of single immunoglobulin variable domains secreted from Escherichia coli. Nature. 1989;341:544–46.

9. Woolven BP, Frenken L, van der Logt P, Nicholls PJ. The structure of the llama heavy chain constant genes reveals a mechanism for heavy-chain antibody formation. Immunogenetics. 1999;50:98–101.

10. Streltsov VA, Varghese JN, Carmichael JA, Irving RA, Hudson PJ, Nuttall SD. Structural evidence for evolution of shark Ig new antigen receptor variable domain antibodies from a cell-surface receptor. Proc Natl Acad Sci USA. 2004;101:12444–49.

11. Jespers L, Schon, O, James LC, Vesprintsev, D, Winter G. Crystal structure of HEL4, a soluble, refoldable human VH single domain with a germ-line scaffold. J Mol Biol. 2004;337:893–903.

12. Teeling JL, French RR, Cragg MS, van den Brakel J, Pluyter M, Huang H, et al. Characterization of new human CD20 monoclonal antibodies with potent cytolytic activity against non-Hodgkin lymphomas. Blood. 2004;104:1793–1800.

13. Casey JL, Napier MP, King DJ, Pedley RB, Chaplin LC, Weir N, et al. Tumour targeting of humanised cross-linked divalent-Fab' antibody fragments: a clinical phase I/II study. Br J Cancer. 2002;86:1401–10.

14. Martens T, Schmidt NO, Eckerich C, Fillbrandt R, Merchant M, Schwall R, et al. A novel one-armed anti-c-Met antibody inhibits glioblastoma growth in vivo. Clin Cancer Res. 2006;12:6144–152.

15. Labrijn AF, Aalberse RC, Schuurman J. When binding is enough: nonactivating antibody formats. Curr Opin Immunol. 2008;20:479–85.

16. Arbabi-Ghahroudi M, Tanha J, MacKenzie R Prokaryotic expression of antibodies. Cancer Metastasis Rev. 2005;24:501–19.

17. Nyyssönen E, Penttila M, Harkki A, Saloheimo A, Knowles JK, Keranen S. Efficient production of antibody fragments by the filamentous fungus Trichoderma reesei. Biotechnol (NY). 1993;11:591–95.

18. Freyre FM, Vazquez JE, Ayala M, Canaan-Haden L, Bell H, Rodriguez I, et al. Very high expression of an anti-carcinoembryonic antigen single chain Fv antibody fragment in the yeast Pichia pastoris. J Biotechnol. 2000;6:157–63.

19. Rahbarizadeh F, Rasaee MJ, Forouzandeh M, Allameh AA. Over expression of anti-MUC1 single-domain antibody fragments in the yeast Pichia pastoris. Mol Immunol. 2006;43:426–35.

20. Harmsen MM, Van Solt CB, Fijten HPD, Van Setten MC. Prolonged in vivo residence times of llama single-domain antibody fragments in pigs by binding to porcine immunoglobulins. Vaccine. 2005; 23:4926–934.

21. Frenken LGJ, van der Linden RH, Hermans PW, Bos JW, Ruuls RC, de Geus B, Verrips CT. Isolation of antigen specific llama VHH antibody fragments and their high level secretion by Saccharomyces cerevisiae. J Biotechnol. 2000;78:11–21.

22. Humphreys DP, Heywood SP, Henry A, Ait-Lhadj L, Antoniw P, Palframan R, et al. Alternative antibody Fab' fragment PEGylation strategies: combination of strong reducing agents, disruption of the interchain disulphide bond and disulphide engineering. Protein Eng Des Sel. 2007;20:227–34.

23. Chapman AP. PEGylated antibodies and antibody fragments for improved therapy: a review. Adv Drug Deliv Rev. 2002;54:531–45.

24. Shaunak S, Godwin A, Choi JW, Balan S, Pedone E, Vijayarangam D, et al. Site-specific PEGylation of native disulfide bonds in therapeutic proteins. Nat Chem Biol. 2006;2:312–13.

25. Brocchini S, Godwin A, Balan S, Choi JW, Zloh M, Shaunak S. Disulfide bridge based PEGylation of proteins. Adv Drug Deliv Rev. 2008;60:3–12.