OR WAIT 15 SECS
The authors review methods for generating monoclonal antibodies for research and development.
Antibodies are key players in the body’s defense against foreign agents, such as microorganisms and viruses, and have been used for research and diagnostics since the early 1900s. But for many decades their utility in the laboratory was limited by their polyclonal nature.
When confronted with a foreign agent (antigen), the human immune system generates a large number of antibodies, each of which recognizes a different part (epitope) of the antigen. This results in a vast, heterogeneous pool of polyclonal antibodies ideal for fighting off invaders but poorly suited for research and diagnostic applications that require a steady supply of antibodies that recognize the same exact epitope on an antigen.
In 1975, Koehler and Milstein overcame this major hurdle by developing a method to generate near- unlimited supplies of identical monoclonal antibodies (mAbs) that all bind to the same epitope of a given antigen (1). This breakthrough turned antibodies into indispensable tools; mAbs are now routinely used in the laboratory, serve as diagnostics in applications that range from blood typing (2) to HIV diagnosis (3, 4), and are driving a global therapeutic mAb market expected to be valued at $125 billion by 2020 (5). Here, the authors discuss commonly used mAb-generation technologies including hybridoma methods and recombinant antibody generation platforms.
Koehler and Milstein’s groundbreaking hybridoma approach allows generation of pools of cells that secrete an unlimited supply of identical antibodies by fusing B cells from an animal host that produce a desired antibody with an immortal myeloma tumor cell line. This method can be used to generate research-grade antibodies for applications that range from Western blotting to flow cytometry as well as antibodies that act as critical reagents in diagnostic kits.
For generation of research-grade mAbs, an adaption of the traditional hybridoma approach, the mouse iliac lymph node method (6), when paired with immunization using full length antigens, and downstream assays that allow mAb specificity analysis, enables short development times and fast and efficient production of mAbs for many different applications.
To generate antibodies for inclusion in diagnostic kits, which must meet more rigorous quality standards than research-grade antibodies, a funnel-shaped screening process can be employed. Such approaches increase stringency as the thousands of clones that are often required to identify a single promising mAb proceed from one stage to the next. First, mAbs should be screened for positive reactivity with the antigen of interest and counter-screened against any relevant antigens to eliminate all cross-reacting clones; epitope mapping can be part of the screening when repertoire diversity and/or specificity are needed. Next, screening should focus on the mAb’s affinity for the target antigen using assays that range from basic supernatant dilution assay to more sophisticated surface plasmon resonance techniques. Selected mAbs are then sub-cloned several times to achieve monoclonality and to eliminate clones with unstable chromosomal rearrangements. The final step is lab-scale production and purification of the mAb. This step eliminates poor mAb producers, as these clones won’t be suitable for manufacturing. At the end of this process, one has in hand highly specific and well-characterized mAbs ready for assay optimization and validation with clinical samples.
A key concern for immunoassay diagnostic kits is the obligation to maintain their performance throughout the lifetime of the commercial product. This means that mAbs included in diagnostic kits have to be secured for often more than 20 years. To ensure batch-to-batch consistency, a master cell bank must be established and the antibody stock safely stored.
One of the main limitations of traditional hybridoma-based methods is that they rely on non-human hosts for antibody generation. This limitation can prove challenging for therapeutic applications as mouse antibodies are sometimes regarded as foreign by the human immune system, triggering the production of human anti-mouse antibodies that can damage the patient’s kidneys (7).
Several different approaches have been taken to generate antibodies that do not trigger this damaging immune response, including generating chimeric or humanized antibodies by combining mouse-generated mAb antigen binding sites with human antibody sequences (8, 9) and engineering humanized mice that produce antibodies with fully human sequences. This latter approach has been used to bring several therapeutics to the market; the first therapeutic mAb generated using humanized mice, panitumumab, was approved by FDA in 2006 and several more such drugs have since been approved including nivolumab, a treatment for melanoma and squamous cell carcinoma.
An alternate approach, recombinant antibody generation, provides even greater control of antibody sequence and thus specificity and affinity; many platforms are completely animal-free. Recombinant antibody methods take advantage of vast libraries of synthetic antibody genes that can be easily manipulated to produce antibodies with desired specificities and affinities. These libraries are generated either using B cells from non-immunized donors or by de novo gene synthesis. Their large size (> 1 billion genes) enables selection of antibodies against virtually any antigen but also requires a new approach to identifying highly specific antibodies. Instead of screening for antibodies that bind an antigen of interest, antibodies with high affinity are identified through selection methods, such as ribosome (10), yeast, bacterial (11), or phage display (12).
By far the most popular screening method are phage display platforms, which fuse antibody heavy- and light-chain variable domains to a phage coat protein gene (13). The phages display this fusion protein on their surface, making it accessible for in vitro selection. During the selection process, the antigen is immobilized on a solid surface and exposed to the phage library. Phages expressing a high-affinity binding antibody interact with the immobilized antigen and are considered a specific target binder; low affinity and non-specific binders are removed during wash steps (12). Specific binders can then be amplified in expression hosts such as Escherichia coli. If desired, resulting mAbs can be further manipulated in vitro to optimize binding strength, or tested and selected for optimal performance in specific assay formats.
One popular approach is the generation of single-domain or VHH antibodies that consist of a single immunoglobulin domain. This technology is based on the discovery that camelid and cartilaginous fish antibodies consist only of a heavy chain (VH) domain and allows generation of much smaller antibodies that are easily manipulated in vitro and expressed in both prokaryotic and eukaryotic expression hosts, are highly stable even under extreme conditions, and have excellent penetrability for therapeutic applications. One disadvantage of these systems is that because these antibodies are of non-human origin resulting antibodies need to be humanized for therapeutic applications.
The first fully synthetic Human Combinatorial Antibody Library (HuCAL) was developed by Knappik et al. in 1999 (14). Detailed analysis of the antigen binding Fv moiety of the human antibody repertoire revealed that its structural diversity can be captured by seven heavy-chain and seven light-chain variable region genes. A master library of just 49 antibody genes can thus capture the structural diversity of more than 95% of the human antibody repertoire.
To generate diversity, HuCAL technology modifies the complementarity determining region (CDR) of this master library, the source of greatest variability in the antigen binding domain (Figure 1).
By engineering restrictions sites to flank the CDRs in the master library, synthetically engineered CDRs can be inserted to generate libraries of billions of functional human antibody specificities, which can then be selected against using in-vitro methods (15).
Because the HuCAL library is a human antibody framework, it can eliminate the need for antibody humanization. Because it relies on swapping and randomizing CDRs in the antigen binding domain, binding affinity can be manipulated and optimized. Using guided selection methods, highly specific antibodies can be generated to recognize epitopes with specific modifications, such as phosphorylation, mutation, or oxidation. HuCAL antibodies can be engineered in various formats, including monovalent and bivalent Fab fragments that are suited for different applications. Lastly, HuCAL antibody selection and screening relies on phage-display, and high-throughput, automated selection methods have been developed to enable the selection, screening and purification of Fab antibodies from the library in eight weeks compared to at least four months for traditional methods that rely on the immunization of animals.
Unlike traditional methods, which generate mAbs in full immunoglobulin (Ig) format, recombinant technology allows generation of antibodies in non-traditional formats such as monovalent and bivalent Fab fragments with purification and detection tags, as well as functionalized antibody fragments. With the elimination of the constant FC region from the Fab antibody, nonspecific binding to Fc receptors is prevented and antibody diffusion is improved, due to their smaller size. Traditional anti-mouse and anti-rat secondary antibodies do not bind these human antibodies, so anti-human and anti-tag secondary reagents are used instead. Fab antibodies can be converted to fully human immunoglobulins of any isotype offering advantages for the design and optimization of immunodiagnostic assays. For example, human recombinant monoclonal antibodies can provide a long-term stable source of calibrator or control to replace human disease state serum, which can be variable in quality and often difficult to source consistently. Recombinant antibodies can be easily engineered, so the format can be chosen to suit the intended application. Bivalent Fab fragments are generally preferred for applications that detect surface-bound antigens (Western blotting, flow cytometry, immunohistochemistry, etc.) because they possess two antigen binding sites, thereby increasing overall sensitivity due to avidity effects. Monovalent antibodies are ideal for crystallography and for cellular assays because they avoid cross-linking of antigens.
Regardless of the method of antibody generation, it is of utmost importance to validate it for the intended application. It has been demonstrated that a high percentage (>50%) of commercially available antibodies are not specific for their intended target or are not sufficiently sensitive to detect endogenous native protein targets (16). Various initiatives are currently underway to propose best practices for validation that could be agreed upon by antibody users in the research community and commercial antibody producers (17). To ensure consistent and reproducible results, antibody validation should be performed for each unique assay; demonstrate that the antibody can recognize the antigen of interest in the context of those assays; and include negative and positive controls whenever possible to demonstrate specificity of the antibody for the antigen.
The authors thank Amanda Turner, product manager, custom antibody products, Bio-Rad Laboratories, UK for reviewing this article and providing expert opinion.
1. G. Köhler and C. Milstein, Nature 256 (5517), 495-497 (1975).
2. L. Marks L, mAbs 6 (6), 1362-1367 (2014).
3. E. Piwowar-Manning, et al. J Clin Virol 62, 75-9 (2015).
4. E.O. Mitchell et al., J Clin Virol 58, Suppl 1:e79-84 (2013).
5. D.M. Ecker, S.D. Jones, and H.L. Levine, mAbs 7 (1), 9-14 (2015).
6. Y Sado, S. Inoue, Y. Tomono, and H. Omori, Acta Histochem Cytochem 39 (3), 89-94 (2007).
7. J.E. Frödin, A.K. Lefvert, and H. Mellstedt, Cell Biophys 21 (1-3), 153-165 (1992).
8. S.L. Morrison, M.J. Johnson, L.A. Herzenberg, and V.T. Oi, Proc Natl Acad Sci USA 81 (21), 6851-6855 (1984).
9. M.S. Neuberger, G.T. Williams, and R.O. Fox, Nature 312 (5995), 604-608 (1984).
10. D. Lipovsek and S.Plückthun, J Immnol Methods 290 (1-2), 51-67 (2004).
11. L.C. Mattheakis, R.R. Bhatt, and W.J. Dower, Proc Natl Acad Sci USA 91(19), 9022-9026 (1994).
12. H.R. Hoogenboom and G. Winter, J Mol Biol 227 (2), 381-388 (1992).
13. J.D. Marks, et al., J Mol Biol 222 (3), 581-597 (1991).
14. A. Knappik, et al., J Mol Biol 296 (1), 57-86 (2000).
15. B. Alberts and A. Johnson, Molecular Biology of the Cell, (Garland Science, New York, NY, 6th ed., 2014).
16. G. Roncador G, et al., mAbs 8 (1), 27-36 (2016).
17. M. Uhlen et al., Nat Methods 13 (10), 823-7 (2016).