Online Issues

<< All Back-issues

<< This Issue's Table of Contents

ILAR Journal V37(3) 1995
Adjuvants and Antibody Production

Monoclonal antibody methods give us the ability to derive individual antibodies of invariant specificity and selectivity, and to immortalize the antibody-producing cells, ensuring a virtually infinite supply (Köhler and Milstein, 1975). However, hybridoma technology requires substantial time, labor, expense, specialized cell culture facilities, the use of animals, and the expertise to prepare and screen large numbers of hybridomas to select the best ones. The number of substances that are immunogenic in mammals is limited, and the maximum diversity expected from the mammalian immune response is on the order of 6 x 10⁶ different antibodies (Harlow and Lane, 1988, p. 16). More important, there is no practical way to alter the properties of antibodies produced by hybridomas.

Interest in isolating and expressing antibody genes developed after the first descriptions of hybridoma production by Köhler and Milstein 1975). Recombinant antibody technology is based on advances in the understanding of antibody structure and function, the biology of bacteriophage replication, and new techniques for DNA manipulation and mutagenesis. After the DNA sequences of many immunoglobulin heavy and light chain variable domains (V_H and V_L) were determined, consensus oligonucleotide primers were designed to recover the genes from new antibodies (Orlandi et al., 1989; Coloma et al., 1991). The polymerase chain reaction (PCR) was used to amplify the genes from a single-stranded DNA copy of the antibody messenger RNA (mRNA) (Mullis, 1990). This was accompanied by the development of new plasmid and bacteriophage cloning vectors for the selection and expression of antibodies. Additional necessary methods were quickly adapted from other applications.

In this article we present a brief review of antibody architecture; an introduction to the major steps in deriving, selecting, and expressing recombinant antibodies; an overview of the present status of antibody engineering and semi-synthetic combinatorial antibody libraries; and our perspectives on how these emerging technologies will affect the use of animals in research.

GENERAL PRINCIPLES

Antibody Structure

This article includes references to monoclonal antibodies (MAbs), Fabs, and Fv fragments. These are diagrammed in Figure 1. Although antibodies are very specific in their recognition of a particular antigenic structure (epitope), X-ray crystallography has revealed that all antibodies share similar structural features (Alzari et al., 1988; Padlan, 1994). Antibodies (immunoglobulins, or Igs) comprise two large (heavy, H) and two small (light, L) polypeptide chains. The H-L and H-H pairs are joined by disulfide bridges (Figure 1). Intact IgG antibodies have two combining sites, which enable them to bind polyvalent or surface-bound antigens with higher avidity. Sequence analysis of antibodies reveals that both chains are composed of regions of variable sequence at the amino terminus (V_H[, V_L), followed by amino acids that are the same or constant (C_H, C_L) among antibodies of the same class. Each VH and VL consists of four relatively conserved framework (FR) segments separating three loops that are hypervariable in sequence, called the complementarity-determining regions (CDRs) (Figure 2).

The tertiary structure of the combining site is also similar among antibodies. The V_H and V_L sequences are folded to form a cup-like site for binding the antigen. The constant domains help to stabilize this structure, which is very similar among antibodies from different species. The affinity and specificity of an antibody is almost entirely due to the size and sequence of the CDR loops. The unique shape of the combining site is maintained by structural interactions between amino acids of the framework and CDRs.

The antibodies are divided into subclasses (IgG₁, IgG_2a, IgG_2b, IgG₃, IgM, IgA, IgD, IgE in the mouse) based on the sequence and function of the constant region. There also are groups within the subclasses that have minor differences in the primary sequence of the framework regions (Kabat et al., 1991).

Fab fragments consist of the variable domains and the corresponding first constant domain: V_H-C_H1 and V_L-C_L (Figure 1). Fab fragments can be prepared by cleavage of MAbs with papain or by expression of the Fab sequences in Escherichia coli bacteria. Proteolytic and recombinant Fab fragments usually have the same antigen-binding affinities as the intact antibody (Anand et al., 1991a). The C_H1 and C_L domains promote the physical association of the H and L chains, which become covalently joined by interchain disulfide bonds. Recombinant Fab fragments can be as thermostable as the intact antibody (Alfthan et al., 1993). The constant domains of Fabs provide conserved antigenic determinants, which make it easy to select Fabs with specific antibodies.

Fv fragments consist of only the variable domains of the heavy and light chains. This is the smallest structure that retains the binding properties of the parent antibody. It is difficult to generate Fv fragments from whole antibodies. Recombinant V_H and V_L polypeptides that associate to become Fv fragments can be prepared in E. coli, but they are unstable in the physiological pH range. In early studies, recombinant Fvs were stabilized in vitro by crosslinking them with glutaraldehyde or by creating a disulfide bond between the polypeptides. However, stable recombinant Fvs can also be expressed as a single polypeptide in which the V_H and V_L domains are joined by a peptide linker of 14-16 amino acids to create a single chain antibody or ScFv (Bird et al., 1988). The length and composition of the linker may affect the binding affinity and stability of the ScFv (Whitlow et al., 1993). ScFvs can have binding constants similar to that of the corresponding intact antibody (Whitlow and Filpula, 1991).

Recombinant Antibody Methods

Recombinant antibody technology involves recovering the antibody genes from the source cells, amplifying and cloning the genes into an appropriate vector, introducing the vector into a host, and achieving expression of adequate amounts of functional antibody. Additional techniques are required to select the desired antibody when a large repertoire of genes has been cloned. Recombinant antibodies can be cloned from any species of antibody-producing animal, if the appropriate oligonucleotide primers or hybridization probes are available (see below).

The ability to manipulate the antibody genes makes it possible to generate new antibodies in vitro. This can be done at the level of the whole combining site by making new combinations of H and L chains (Collet et al., 1992; Kang et al., 1991a; Marks et al., 1992). It can also be done by mutating individual CDRs (Cheetham, 1988; Garrard and Henner, 1993; Kettleborough et al., 1991). The variable regions can also be joined with sequences that have other properties, such as chimeric antibody-toxin molecules for targeted therapy of tumors (Reiter et al., 1994). Since Fab and ScFv fragments are smaller than whole antibodies, they are superior for medical imaging and other applications. Another advantage is that recombinant antibodies can be expressed in new hosts. When antibodies to small molecules are required, it is often difficult to design and synthesize haptens and hapten-carrier conjugates. Mutation of cloned antibodies is an alternate way to obtain the desired specificities. The value of these approaches to generate new immunoreagents for diagnostic and therapeutic applications is just beginning to be explored.

The host in which most recombinant antibody methods were developed is the bacterium Escherichia coli. Growth of bacteria is rapid and inexpensive, and a number of vectors are available for expression and manipulation of cloned genes. DNA can be introduced directly into E. coli (the process known as transformation) or by infectious bacteriophage (transfection). Genetic constructions of antibody fragments (Fab and ScFv) can be quickly assessed and various selection methods can be applied. It is easier to scale up the production of antibody fragments in E. coli than in mass culture of mammalian cells such as hybridomas.

Obtaining Antibody Genes

Antibody genes can be derived in several ways, one of which is schematized in Figure 3. The starting material is mRNA derived from antibody-producing cells such as antigen-stimulated B lymphocytes from the spleen or peripheral blood. Hybridomas are also a source of mRNA that predominantly or exclusively encodes a single antibody. Because mRNAs have a polyadenylate (polyA) sequence at their 3' end, they can be purified from the total RNA population by affinity chromatography on oligodeoxythymidylate-cellulose. A complementary DNA (cDNA) copy of the mRNA is then made using the enzyme reverse transcriptase. Commercial kits are available that provide all the materials for mRNA isolation and preparation of cDNA.

Selective Amplification of Antibody Genes

The mRNA population contains messages for a vast number of genes in addition to antibody genes. The Fab or Fv sequences of antibodies can be selectively amplified from the cDNA by PCR using specific complementary oligonucleotide primers. This procedure, which revolutionized in vitro gene amplification, is simple to perform, but the choice of primers is very important. Figure 2 illustrates where these primers bind. One primer is required for the 5' end and another for the 3' end of the gene. The 5' primers are complementary either to the antibody leader sequence (Coloma et al., 1991; Larrick et al., 1989a,b,) or to conserved parts of the first framework region FR1 (Huse et al.,1989; Orlandi et al., 1989). Because the FR1 domains of immunoglobulins differ in sequence, mixtures of primers or degenerate primers are needed to accommodate these differences. The 3' primers of Fabs recognize the end of the C_H1 of each Ig subclass (IgG₁, IgG_2a, IgG_2b, IgG₃, etc.) and the end of the C_L domain. Fv domains are amplified using a primer that binds to the 3' end of the V_H and V_L regions. Several laboratories have designed sets of primers that allow amplification of Fab or Fv sequences from mouse, human, rabbit, and other species.

The products of PCR amplification are populations of different H and L sequences (Figure 3A). The amplified DNA is purified, digested with restriction enzymes to create ends compatible with sites in the cloning vector, and ligated in place. Vectors that have been used include bacterial plasmids, variants of the lytic bacteriophage lambda, and (more recently) phagemid variants of the non-lytic filamentous bacteriophages M13, fd, and f1.

The versatile phagemid vectors normally replicate in E. coli as an extrachromosomal element (plasmid). However, when the host is infected with an M13 helper phage, the phagemid DNA is packaged into complete, infective bacteriophage. An antibiotic resistance marker facilitates selection of host cells that carry the phagemid. Transcription of the cloned antibody sequences is under the control of a bacterial promoter. The sequences, preceded by a bacterial "leader sequence" such as pelB, are produced as a fusion protein with a minor coat protein (gIII, the phage gene III product) located at the tip of the phage. In cloned Fabs only the H chain is fused with gene III. ScFvs are produced as a single protein linked to gIII. The leader sequence directs the proteins to the periplasmic space where the antibody fragment-gIII fusion is incorporated into viable phage. With Fabs, the free L chain in the periplasm associates with the fused H chain-gIII to form a functional Fab. Phage are secreted through the bacterial outer membrane and display one to three copies of the encoded antibody. Antibody fusions have also been devised with the major coat protein, gVIII (Kang et al., 1991b). The gIII and gVIII phage display systems were designed to allow selection and enrichment of functional antibody fragments from a mixture, based upon their affinity for the antigen.

Selection of Phage Displaying Antibodies

Phage displaying the desired antibodies are selected by binding to antigen in a format similar to solid-phase immunoassay (Barbas and Lerner, 1991). The process, called "phage panning," is depicted at the top of Figure 3B. The antigen or hapten conjugate is immobilized on microplate wells or on magnetic beads or solid material packed into a column. Bound phage can be eluted and amplified by replication in new host cells. Those that bind weakly or not at all are washed away before the elution step. After several rounds of binding and amplification, the phage population should consist almost entirely of those that express the desired antibodies.

Expression of Soluble Antibody from Selected Clones

In order to produce soluble antibody, the antibody genes cloned in the phagemid must be expressed free of the phage gene III protein. This is achieved differently in two widely used systems. With cloned Fabs, the DNA containing the gIII fragment is excised by digestion with restriction enzymes and the vector is re-ligated and introduced into new host cells (Barbas et al., 1991). Expression of the antibody leads to the synthesis and export of soluble Fab into the periplasmic space. Soluble Fab can be recovered by gentle lysis of the cells (Figure 3B). A different method has been used for expression of ScFvs. Termination codons in an mRNA will halt protein synthesis, but these codons are not recognized as translational stop signals in E. coli that have a suppressor mutation. The phagemid for the ScFv is constructed with a translational stop codon, UAG, at the junction between the antibody and gIII protein sequences. When display phage are desired, the phagemid is grown in E. coli TG1, which has a suppressor mutation that does not recognize the stop codon. This allows the ScFv-gIII fusion protein to be synthesized. To obtain the soluble antibody, the phagemid is grown in a non-suppressor strain, E. coli HB2151. This host recognizes the stop codon at the end of the ScFv gene, and soluble antibody is produced (Hoogenboom et al., 1991). Bacterial colonies producing soluble Fab or ScFv can be detected either by in situ immunostaining of colony lifts in a procedure similar to western blotting (Skerra et al., 1991) or by immunoassay.

Difficulties with Large-scale Antibody Production in E. coli

Bacterial systems that have proven to be very efficient for expression of some proteins have sometimes given disappointing results with recombinant antibodies. Efficiencies of recombinant antibody expression in E. coli vary with the antibody. In some early experiments expression of soluble Fabs and ScFvs was under 10 mg/l, which is considerably less than the amounts expected from flask cultures of hybridomas. This is sufficient for applications such as enzyme immunoassay and western blotting, but not for industrial needs. Two known problems are misfolding of the antibody chains and incompatibility with bacterial secretory pathways. In some cases the amounts of functional antibody are related to the order in which cloned domains are synthesized in ScFv constructs, that is, H-linker-L vs. L-linker-H (Anand et al., 1991b; Tsumoto et al., 1994). Different immunoglobulin subclasses have somewhat different structures that may influence expression. For example, an IgG3 Fab specific for 2-phenyloxazolone was only assembled properly if most of the CH1 domain was removed, while an IgG1 Fab with the same specificity was expressed correctly (Alfthan et al., 1993). Another group reported that the light chain subclass (k or l) of two different carbohydrate-specific Fabs had a pronounced effect on yields of functional Fab in E. coli (MacKenzie et al., 1994). Even single amino acid changes greatly alter the yield of some antibody fragments (Brummell et al., 1993; McManus and Reichmann, 1991).

Despite these problems, some laboratories have successfully produced usable amounts of antibodies. One strategy has been to optimize the export of Fabs to the periplasmic space where they are more likely to fold correctly (Carter et al., 1992). This resulted in expression of 1-2 g/l, an efficiency practical for most research and industrial applications. Another approach was to place the transcription of the antibody genes under the control of the bacteriophage T7 promoter. This causes antibody to be produced in the form of insoluble inclusion bodies in the cytoplasm of E. coli. The insoluble material is recovered, solubilized with strong denaturing agents, and allowed to refold slowly in vitro during dialysis or dilution (Whitlow and Filpula, 1991). A recent refinement of this method reportedly yielded 50 to 120 mg/l of active ScFv from inclusion bodies after renaturation and affinity chromatography (Burks and Iverson, 1995). Although this method appears to be very efficient, in vitro refolding requires additional time and expense. In summary, the production of large amounts of functional antibody fragments in E. coli is still the exception rather than the rule, and expression of antibody fragments in bacteria may have to be optimized empirically (Carter et al., 1992). Many laboratories are working on ways to improve bacterial expression of soluble antibodies.

Expression of Antibodies in Other Hosts

Functional antibody fragments have also been expressed in other prokaryotic and eukaryotic hosts, including Bacillus subtilis WB600 (Wu et al., 1993), the yeast strains Saccharomyces pombe (Davis et al., 1991) and Pichia pastoris, Xenopus (toad) oocytes (Biocca et al., 1993), insect cells (Page and Murphy, 1990), and plants (Hiatt et al., 1989). Several ScFv constructs have now been expressed in mammalian cells (Dorai et al., 1994). Each of these hosts has advantages and disadvantages. Prokaryotic hosts eliminate the risk of adventitious mammalian viruses in the antibody preparation, but may produce endotoxins or other contaminants. In eukaryotic cells whole antibody molecules retain normal post-translational modifications as well as effector functions such as complement-fixing sites. However, eukaryotic cells are more difficult and expensive to culture, and the amount of product is usually lower.

ANTIBODY ENGINEERING

Antibody engineering is the process of altering antibody structure and functional properties by recombinant DNA methods. Once the DNA sequences of the variable regions are known, the amino acid sequence can be deduced. Methods of in vitro mutagenesis can be applied to insert, delete, or change one or several amino acids, or to exchange entire variable domains. Many laboratories worldwide are now using these techniques to produce antibodies that would be difficult or impossible to obtain from animals. Some key steps in recombinant antibody engineering are described below.

DNA Sequencing

After the desired antibody genes have been cloned and selected as described in Figure 3, the DNA sequence is determined using chain termination sequencing methods that are now available in commercial kits (Maniatis et al., 1989). The sequencing requires oligonucleotide primers complementary to the 5' ends of the region to be sequenced. As portions of the sequence are determined, additional primers may be needed to extend the sequence. Overlapping regions of sequence determined in separate runs are aligned to obtain the entire sequence. Both strands of the DNA are sequenced in order to verify the results. A frequently updated database of all reported antibody sequences has been compiled (Kabat et al., 1991). New antibody sequences can be analyzed by first comparing them to the most similar counterparts in this database.

Modeling the Combining Site

The ability to model the combining site and visualize antibody-antigen interactions in three-dimensional space is a powerful tool for antibody engineering. The structures of many antibodies have been determined by X-ray crystallography at atomic resolution, and the coordinates are stored as files in the Brookhaven Protein Data Bank. These files may be retrieved through the Internet and displayed in many programs available on minicomputers and modeling workstations. Because there is considerable homology among antibody framework domains and the secondary structures of some CDRs, it is possible to construct a computational model of a new antibody in order to predict which CDRs and other residues are important to epitope binding. The model can then be used to guide subsequent engineering and mutagenesis steps (Roberts et al., 1994). The software package AbM (Oxford Molecular, Ltd.) uses established crystallographic structures to build antibody models from amino acid sequence data. The resulting model, which consists of a set of atomic coordinates in three-dimensional space, can be compared to known antibody structures. We used AbM as an aid to building a model of an antibody that binds an herbicide (Bell et al., 1995). However, considerable experience is required to interpret and refine such computational models.

Changing the Structure

Since a model is only a hypothesis, mutated variant antibodies must be created and analyzed to test which residues are important for binding or structural properties. The mutations are introduced into the antibody sequences in the cloning vector. Phage displaying altered antibodies may be selected, or soluble variant antibodies can be produced as shown in Figure 3B. The consequences of the mutation can be assessed by immunoassay or methods such as affinity electrophoresis (Brummell et al., 1993).

It may be possible to alter affinity and specificity or both by changing the relative orientations of V_H and V_L domains at their interface, lengthening or shortening particular CDRs to enlarge or shrink the binding pocket, increasing the flexibility of CDRs in the combining site, removing or re-spacing some of the side chains that form the combining site, or altering residues that do not contact antigen but help to form the combining site through CDR-CDR and CDR-framework interactions (Roberts et al., 1987). The structure can be changed to accommodate water molecules, metal atoms, and other functional groups. In principle other properties of the antibody-antigen interactions can be changed by altering close-contact residues. The antibody can also be fused with other antibody molecules, toxins, or enzymes.

HUMANIZED ANTIBODIES

One of the most valuable applications of antibody engineering has been the preparation of antibodies that are less antigenic and more stable for human clinical diagnostic and therapeutic applications. Many MAbs that are useful for medical imaging or therapy were derived in rodents and evoke an undesired immune response in patients. Human MAbs have proven to be very difficult to produce by conventional hybridoma methods. The alternative has been to "humanize" mouse MAbs using recombinant antibody techniques. One of the early approaches involved splicing the CDRs from the V regions of a useful rodent antibody into a human antibody framework. Although ability to bind the antigen was lost in this direct substitution of the CDRs, knowledge of antibody structure was used to change other residues that restored binding in the humanized antibody (Kettleborough et al., 1991; Reichmann et al., 1988).

Recently, a novel strategy was described in which rodent Fabs were used as a template to guide the selection of fully human Fab fragments. This humanization was done by selecting an antibody from the human repertoire that matched the binding property of the rodent antibody. The H chain from a mouse MAb to human tumor necrosis factor alpha (TNFa) was cloned with a library of human L chains in a phagemid vector. The resulting chimeric mouse-human Fabs displayed on phage were selected by panning on TNFa. The human L chain from the strongest-binding clone was recovered, re-cloned with a library of human H chains, and the strongest TNFa-binding phage were selected once again. The entirely human Fab identified in this way bound to the same epitope as the rodent MAb, with similar affinity (Jespers et al., 1994).

SYNTHETIC COMBINATORIAL LIBRARIES--ANTIBODIES WITHOUT IMMUNIZATION

Synthetic combinatorial antibody libraries were developed as an alternative to the use of animals for antibody production (Barbas et al., 1992; Hoogenboom and Winter, 1992; Lerner et al., 1992). Based on the knowledge that antigen binding is due primarily to interactions with the six CDR loops in the combining site, it became feasible to generate diversity in vitro by limited mutagenesis of one or more CDRs. Such a library can be orders of magnitude more diverse than the 10⁵ to 10⁶ antibodies expressed by the mammalian genome. A library can be displayed on the surface of phage and new antibodies selected by panning as described previously.

Construction of Synthetic Combinatorial Libraries

Combinatorial library design is based on the unique structural similarities in all antibodies, irrespective of their binding specificity. CDRH3 has the greatest flexibility and conformational variability, and may have the greatest influence on antigen binding. One of the first combinatorial Fab libraries was constructed by making sequence and length variations in CDRH3 of a human Fab specific for tetanus toxoid (Barbas et al., 1992). It is also known that amino acid side chains on at least five of the six CDRs generally contact the epitope (Roberts et al., 1993; Wilson and Stanfield, 1993). Consequently other combinatorial libraries have been constructed by introducing diversity in several of the CDRs (Garrard and Henner, 1993). Another strategy made use of the natural diversity of ScFv libraries prepared with CDRs from the immunologically naive germline antibody genes (Griffiths et al., 1993; Marks et al., 1991). A third approach to generating diversity seeks to improve the affinity of selected antibodies by allowing random or directed reassortment of cloned V_H and V_L regions (Marks et al., 1992). This type of diversity has been called "chain shuffling." Another recently published method allows combinatorial shuffling of individual intact CDRs (Crameri and Stemmer, 1995).

Advantages and Limitations

In theory virtually any specificity could be recovered from a combinatorial library, including antibodies to substances that might not evoke a conventional immune response. For example, semi-synthetic combinatorial libraries with randomized CDRH3 domains have yielded Fabs that bind haptens (Barbas et al., 1993a) and Fabs that form coordination complexes with metal ions (Barbas et al., 1993b). However, to date there have been relatively few published reports of high affinity antibodies derived from combinatorial libraries. ScFv libraries with CDRs from a human germline V gene repertoire were used to select antibodies that bind to a variety of haptens and self-antigens (Hoogenboom and Winter, 1992). The affinities of these antibodies were relatively low, as might be expected in a primary immune response (Griffiths et al., 1993; Marks et al., 1991). The synthetic combinatorial libraries and screening methods that have been used to date have several limitations that must be overcome if this technology is to replace conventional antibody methods. Some of the important problems are:

• The theoretical diversity in a phage library is much greater than the efficiencies with which the antibody genes can be introduced into bacteria. There have been attempts to increase this by a mechanism involving recombination (Waterhouse et al., 1993).
• Because of the complex structural requirements, most of the diversity will result in misfolded protein that is not able to bind the epitope. Critical CDR-CDR and CDR-framework interactions need to be preserved if the antibodies are to be of practical value (Padlan, 1994; Roberts et al, 1993).
• The diversity must be where it best influences binding. A combinatorial library in CDRH3 may not be a source of useful antibodies to molecules that have binding interactions with other CDRs.
• If higher affinity antibodies are sought, it would be desirable to develop in vitro methods to mimic the mutation and recombination processes that lead to affinity maturation in vivo. Some recombinant antibody techniques of this type have already been developed (Gram et al., 1992; Hawkins et al., 1992; Söderlund et al., 1992)
• Crystallographic structure studies have shown differences in the shape of the combining pockets for protein epitopes and small organic molecules (Wilson and Stanfield, 1993). This suggests that frameworks suitable for epitopes on proteins may not be suitable for recognition of small molecules. It may be preferable to develop separate combinatorial libraries optimized for small molecules in addition to libraries for epitopes on protein.

IMPACT ON ANIMAL USE

Antibody engineering methods are evolving that may eventually replace, reduce, or refine animal use. This process will be slow for several reasons. The new technologies are still too complex, expensive, and time consuming to compete with conventional production of antisera and hybridomas. There are technical problems, such as the need for oligonucleotide primers for unbiased amplification of all immunoglobulin variants, and problems experienced with the expression of antibody fragments in E. coli. These will have to be overcome before it becomes possible to clone some antibodies or obtain complete repertoires from animals. The goal of recombinant antibody expression in bacterial and eukaryotic hosts on a level comparable to ascites production in rodents has not yet been achieved. Many recombinant antibodies that have useful practical applications have been cloned from hybridomas or cells from immunized animals. Synthetic combinatorial library technology will have to be improved before it can be used to obtain antibodies with properties comparable to those from immunized animals.

Recombinant antibody techniques may reduce some types of animal use but increase others. Cloned antibody genes have been used to create transgenic mice that express murine antibodies with human constant domains (Zou et al., 1993), and more recently a complete human antibody repertoire (Jakobovits et al., 1994; Lonberg et al., 1994). Klaus Rajewsky and coworkers made a major breakthrough in this area. They discovered how to insert an altered immunoglobulin domain transgene at the correct chromosomal location so that B lymphocytes can perform normal recombination, class switching, and affinity maturation (Taki et al., 1993). Such transgenic animals will greatly facilitate research on basic mechanisms of the mammalian immune response, including autoimmune diseases. They create a system for producing "human antisera" to antigens that could not safely be administered to human subjects. These mice are also a source of cells for making hybridomas that stably secrete human antibodies.

Recombinant antibody techniques are being increasingly applied in veterinary and human medical research. Antibody libraries cloned from autoimmune mice (Calcutt et al., 1994) and humans (Falorni et al., 1994) will increase our understanding of the number of different epitopes and the idiotypic diversity involved, and may open the way to new therapies. Similarly, repertoire cloning from animals will reveal new aspects of their immune systems and may influence the use of particular species as research models. A recent study of this type demonstrated that swine have the largest number of IgG constant domains of any species and the least sequence heterogeneity between IgG subclasses (Kacskovics et al., 1994). Bacteria, yeast, insect cells, and higher plants are promising hosts for commercial antibody production and should eventually replace animals for industrial-scale expression of cloned antibodies (Hiatt et al., 1989).

In summary, the technologies introduced in this article have already begun to change immunological research and antibody production. As the methods improve, combinatorial libraries and antibody engineering should reduce the need for live animals for these purposes. We imagine that the earliest investigators of mammalian humoral immunity would be gratified to learn that the understanding and application of antibodies has evolved to this stage after nearly a century of research requiring the use of animals.

REFERENCES

Alfthan, K., K. Takkinen, D. Sizman, I. Seppala, T. Immonen, L. Vanne, S. Keranen, M. Kaartinen, J. K. C. Knowles, and T. T. Teeri. 1993. Efficient secretion of murine Fab fragments by Escherichia coli is determined by the first constant domain of the heavy chain. Gene 128:203-209.

Alzari, P. M., M-B. Lascombe, and R. J. Poljak. 1988. Three-dimensional structure of antibodies. Ann. Rev. Immunol. 6:555-580.

Anand, N. N., G. Dubuc, J. Phipps, C. R. MacKenzie, J. Sadowska, N. M. Young, D. R. Bundle, and S. A. Narang. 1991a. Synthesis and expression in Escherichia coli of cistronic DNA encoding an antibody fragment specific for a Salmonella serotype B O-antigen. Gene 100:39-44.

Anand, N. N., S. Mandal, C. R. MacKenzie, J. Sadowska, B. Sigurskjold, N. M. Young, D. R. Bundle, and S. A. Narang. 1991b. Bacterial expression and secretion of various single-chain Fv genes encoding proteins specific for a Salmonella serotype B O-antigen. J. Biol. Chem. 266 (32):21874-21879.

Barbas, C. F., and R. A. Lerner. 1991. Combinatorial immunoglobulin libraries on the surface of phage (phabs): Rapid selection of antigen-specific Fabs. METHODS: A Companion to Methods in Enzymology 2(2):119-124.

Barbas, C. F., A. S. Kang, R. A. Lerner, and S. J. Benkovic. 1991. Assembly of combinatorial libraries on phage surfaces: The gene III site. Proc. Natl. Acad. Sci. USA 88:7978-7982.

Barbas, C. F., J. D. Bain, D. M. Hoekstra, and R. A. Lerner. 1992. Semisynthetic combinatorial antibody libraries: A chemical solution to the diversity problem. Proc. Natl. Acad. Sci. USA 89(10):4457-4461.

Barbas, C. F., W. Amberg, A. Simoncsits, T. M. Jones, and R. A. Lerner. 1993a. Selection of human anti-hapten antibodies from semisynthetic libraries. Gene 137:57-62.

Barbas, C. F, J. S. Rosenblum, and R. A. Lerner. 1993b. Direct selection of antibodies that coordinate metals from semisynthetic combinatorial libraries. Proc. Natl. Acad. Sci. USA 90:6385-6389.

Bell, C. W., V. A. Roberts, K-B. G. Scholthof, G. Zhang, and A. E. Karu. 1995. Recombinant antibodies to diuron: A model for the phenylurea combining site. Pp. 50-71 in Immunoanalysis of Agrochemicals: Emerging Technologies. J. Nelson, A. E. Karu, and R. Wong, eds. Washington D.C.: American Chemical Society (Symposium No. 586).

Biocca, S., P. Pierandreiamaldi, and A. Cattaneo. 1993. Intracellular expression of anti-P2l(ras) single chain Fv fragments inhibits meiotic maturation of Xenopus oocytes. Biochem. Biophys. Res. Comm. 197(2):422-427.

Bird, R. E., K. D. Hardman, J. W. Jacobson, S. Johnson, B. M. Kaufman, S. M. Lee, T. Lee, S. H. Pope, G. S. Riordan, and M. Whitlow. 1988. Single-chain antigen-binding proteins. Science 242:423-426.

Brummell, D. A., V. P. Sharma, N. N. Anand, D. Bilous, G. Dubuc, J. Michniewicz, C. R. MacKenzie, J. Sadowska, B. W. Siguskjold, B. Sinnot, N. M. Young, D. R. Bundle, and S. A. Narang. 1993. Probing the combining site of an anti-carbohydrate antibody by saturation mutagenesis: Role of the heavy-chain CDR3 residues. Biochemistry 32: 1180-1187.

Burks, E. A., and B. L. Iverson. 1995. Rapid, high-yield recovery of recombinant digoxin binding single chain Fv from Escherichia coli. Biotechnol. Prog. 11:112-114.

Calcutt, M. J., T. P. Quinn, and S. L. Deutscher. 1994. Isolation and characterization of nucleic acid-binding antibody fragments from autoimmune mice-derived bacteriophage display libraries. J. Cell. Biochem. Abstr. Suppl. 18D(Abstract T-303):197.

Carter, P., R. F. Kelley, M. L. Rodrigues, B. Snedecor, M. Covarrubias, M. D. Velligan, W. L. T. Wong, A. M. Rowland, C. E. Kotts, M. L. Carver, M. Yang, J. H. Bourell, H. M. Shepard, and D. Henner. 1992. High level Escherichia coli expression and production of a bivalent humanized antibody fragment. Bio/Technol. 10:163-167.

Cheetham, J. 1988. Reshaping the antibody combining site by CDR replacementÑtailoring or tinkering to fit? Protein Eng. 2:170-172.

Collet, T. A., P. Roben, R. OÕKennedy, C. F. Barbas, D. R. Burton, and R. L. Lerner. 1992. A binary plasmid system for shuffling combinatorial antibody libraries. Proc. Natl. Acad. Sci. USA 89:10026-10030.

Coloma, M. J., J. W. Larrick, M. Ayala, and J. V. Gavilondo-Cowley. 1991. Primer design for the cloning of immunoglobulin heavy-chain leader-variable regions from mouse hybridoma cells using the PCR. BioTechniques 11(2):152-156.

Crameri, A., and W. P. C. Stemmer. 1995. Combinatorial multiple-casette mutagenesis creates all the permutations of mutant and wild-type sequences. BioTechniques 18(2):194-196.

Davis, G. T., W. D. Bedzyk, E. W. Voss, and T. W. Jacobs. 1991. Single chain antibody (SCA) encoding genes: One-step construction and expression in eukaryotic cells. Bio/Technol. 9:165-169.

Dorai, H., J. E. McCartney, R. M. Hudziak, M-S. Tai, A. A. Laminet, L. L. Houston, J. S. Huston, and H. Oppermann. 1994. Mammalian cell expression of single-chain Fv (sFv) antibody proteins and their c-terminal fusions with interleukin-2 and other effector domains. Bio/Technol. 12:890-897.

Falorni, A., M. A. A. Persson, A. Borch, B. Persson, and Å. Lernmark. 1994. Construction of a combinatorial antibody library from an insulin-dependent diabetic patient. J. Cell. Biochem. Abstr. Suppl. 18D(Abstract T-310):199.

Garrard, L. J., and D. J. Henner. 1993. Selection of an anti-IGF-1 Fab from a Fab phage library created by mutagenesis of multiple CDR loops. Gene 128:103-109.

Gram, H., L- A. Marconi, C. F. Barbas, T. A. Collet, R. A. Lerner, and A. S. Kang. 1992. In vitro selection and affinity maturation of antibodies from a naive combinatorial immunoglobulin library. Proc. Natl. Acad. Sci. USA 89:3576-3580.

Griffiths, A. D., M. Malmqvist, J. D. Marks, J. M. Bye, M. J. Embleton, J. McCafferty, M. Baier, K. P. Holliger, B. D. Gorick, N. C. Hughes-Jones, H. R. Hoogenboom, and G. Winter. 1993. Human anti-self antibodies with high specificity from phage display libraries. EMBO J. 12(2):725-734.

Harlow, E., and D. Lane. 1988. Antibodies: A Laboratory Manual. Cold Spring Harbor, N.Y.: Cold Spring Harbor Laboratory.

Hawkins, R. E., S. J. Russell, and G. Winter. 1992. Selection of phage antibodies by binding affinity: Mimicking affinity maturation. J. Mol. Biol. 226:889-896.

Hiatt, A., R. Cafferkey, and K. Bowdish. 1989. Production of antibodies in transgenic plants. Nature 342:76-78.

Hoogenboom, H. R., and G. Winter. 1992. By-passing immunization: Human antibodies from synthetic repertoires of germline VH gene segments rearranged in vitro. J. Mol. Biol. 227:381-388.

Hoogenboom, H. R., A. D. Griffiths, K. S. Johnson, D. J. Chiswell, P. Hudson, and G. Winter. 1991. Multi-subunit proteins on the surface of filamentous phage: Methodologies for displaying antibody (Fab) heavy and light chains. Nucl. Acids Res. 19(15):4133-4137.

Huse, W. D., L. Sastry, S. A. Iverson, A. S. Kang, M. Alting-Mees, D. R. Burton, S. J. Benkovic, and R. A. Lerner. 1989. Generation of a large combinatorial library of the immunoglobulin repertoire in phage lambda. Science 246:1275-1281.

Jakobovits, A., L. L. Green, M. C. Hardy, C. E. Maynard-Currie, H. Tsuda, D. M. Louie, D. Smith, M. J. Mendez, H. Abderrahim, M. Noguchi, and S. Klapholz. 1994. Expression of human immunoglobulin loci-derived YACs in mice: Towards mice producing a large repertoire of human antibodies. J. Cell. Biochem. Abstr. Suppl. 18D (Abstract T-017):185.

Jespers, L. S., A. Roberts, S. M. Mahler, G. Winter, and H. R. Hoogenboom. 1994. Guiding the selection of human antibodies from phage display repertoires to a single epitope of an antigen. Bio/Technol. 12:899-903.

Kabat, E. A., T. T. Wu, H. M. Perry, K. S. Gottesman, and C. Foeller. 1991. Sequences of Proteins of Immunological Interest, 5th ed. Washington, D.C.: U.S. Public Health Service, National Institutes of Health.

Kacskovics, I., J. Sun, and J. E. Butler. 1994. Four subclasses of swine IgG identified from the cDNA sequences of a single animal. J. Cell. Biochem. Abstr. Suppl. 18D(Abstract T-520):214.

Kang, A. S., T. M. Jones, and D. R. Burton. 1991a. Antibody redesign by chain shuffling from random combinatorial libraries. Proc. Natl. Acad. Sci. USA 88:11120-11123.

Kang, A. S., C. F. Barbas, K. D. Janda, S. J. Benkovic, and R. A. Lerner. 1991b. Linkage of recognition and replication functions by assembling combinatorial antibody Fab libraries along phage surfaces. Proc. Natl. Acad. Sci. USA 88:4363-4366.

Kettleborough, C. A., J. Saldanha, V. J. Heath, J. Morrisson, and M. M. Bendig. 1991. Humanization of a mouse monoclonal antibody by CDR grafting: The importance of framework residues on loop conformation. Protein Eng. 4(7):773-783.

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

Larrick, J. W., L. Danielsson, C. A. Brenner, M. Abrahamson, K. E. Fry, and C. A. Borrebaeck. 1989a. Rapid cloning of rearranged immunoglobulin genes from human hybridoma cells using mixed primers and the polymerase chain reaction. Biochem. Biophys. Res. Comm. 160:1250-1256.

Larrick, J. W., L. Danielsson, C. A. Brenner, E. F. Wallace, M. Abrahamson, K. E. Fry, and C. A. K. Borrebaeck. 1989b. Polymerase chain reaction using mixed primers: Cloning of human monoclonal antibody variable region genes from single hybridoma cells. Bio/Technol. 7:934-938.

Lerner, R. A., A. S. Kang, J. D. Bain, D. R. Burton, and C. F. Barbas. 1992. Antibodies without immunization. Science 258:1313-1314.

Lonberg, N., L. D. Taylor, C. E. Carmack, and D. Huszar. 1994. Human sequence antibodies from transgenic mice. J. Cell. Biochem. Abstr. Suppl. 18D(Abstract T-018):185.

MacKenzie, R. C., V. Sharma, D. Brummell, D. Bilous, G. Dubuc, J. Sadowska, N. M. Young, D. R. Bundle, and S. A. Narang. 1994. Effect of Cl-Ck domain switching on Fab activity and yield in Eschericia coli: Synthesis and expression of genes encoding two anti-carbohydrate Fabs. Bio/Technol. 12:390-395.

Maniatis, T., E. F. Fritsch, and J. Sambrook. 1989. Molecular Cloning: A Laboratory Manual. 2nd ed. Cold Spring Harbor, N.Y.: Cold Spring Harbor Laboratory.

Marks, J. D., H. R. Hoogenboom, T. P. Bonnert, J. McCafferty, A. D. Griffiths, and G. Winter. 1991. Bypassing immunization: Human antibodies from V-gene libraries displayed on phage. J. Mol. Biol. 222:581-597.

Marks, J. D., A. D. Griffiths, M. Malmquist, T. Clackson, J. M. Bye, and G. Winter. 1992. Bypassing immunization: Building high affinity human antibodies by chain shuffling. Bio/Technol. 10:779-783.

McManus, S., and L. Reichmann. 1991. Use of 2D NMR, protein engineering, and molecular modeling to study the hapten-binding site of an antibody Fv fragment against 2-phenyloxazolone. Biochemistry 30: 5851-5857.

Mullis, K. B. 1990. The unusual origin of the polymerase chain reaction. Sci. Am. April 1990:56.

Orlandi, R., D. H. Gssow, P. T. Jones, and G. Winter. 1989. Cloning immunoglobulin variable domains for expression by the polymerase chain reaction. Proc. Natl. Acad. Sci. USA 86(May 89):3833-3837.

Padlan, E. A. 1994. Anatomy of the antibody molecule. Mol. Immunol. 31(3):169-217.

Page, M. J., and V. F. Murphy. 1990. Expression of foreign genes in cultured insect cells using a recombinant baculovirus vector. Pp. 573-588 in Animal Cell Culture (Methods in Molecular Biology, Vol 5). J. Pollard and J. M. Walker, eds. Clifton, N.J.: Humana Press.

Reichmann, L., M. Clark, H. Waldmann, and G. Winter. 1988. Reshaping human antibodies for therapy. Nature 332:323-327.

Reiter, Y., U. Brinkmann, K. O. Webber, S- H. Jung, B. Lee, and I. Pastan. 1994. Engineering interchain disulfide bonds into conserved framework regions of Fv fragments: Improved biochemical characteristics of recombinant immunotoxins containing disulfide-stabilized Fv. Protein Eng. 7(5):667-704.

Roberts, S., J. C. Cheetham, and A. R. Rees. 1987. Generation of an antibody with enhanced affinity and specificity for its antigen by protein engineering. Nature 328:731-734.

Roberts, V. A., E. D. Getzoff, and J. A. Tainer. 1993. Structural basis of antigenic cross-reactivity. Pp. 31-53 in Structure of Antigens, Vol. 2. M. H. V. van Regenmortel, ed. Boca Raton, Fla.: CRC Press.

Roberts, V. A., J. Stewart, S. J. Benkovic, and E. D. Getzoff. 1994. Catalytic antibody model and mutagenesis implicate arginine in transition-state stabilization. J. Mol. Biol. 235:1098-1116.

Skerra, A., M. L. Dreher, and G. Winter. 1991. Filter screening of antibody Fab fragments secreted from individual bacterial colonies: Specific detection of antigen binding with a two-membrane system. Anal. Biochem. 196:151-155.

Söderlund, E., A. C. Simonsson, and C. A. K. Borrebaeck. 1992. Phage display technology in antibody engineering: Design of phagemid vectors and in vitro maturation systems. Immunol. Revs. 130:109-124.

Taki, S., M. Meiering, and K. Rajewsky. 1993. Targeted insertion of a variable region gene into the immunoglobulin heavy chain locus. Science 262:1268-1270.

Tsumoto, K., Y. Nakaoki, Y. Ueda, K. Ogasahara, K. Yutani, K. Watanabe, and I. Kumagai. 1994. Effect of the order of antibody variable regions on the expression of the single-chain HyHEL10 Fv fragment in E. coli and the thermodynamic analysis of its antigen-binding properties. Biochem. Biophys. Res. Comm. 201(2):546-551.

Waterhouse, P., A. D. Griffiths, K. S. Johnson, and G. Winter. 1993. Combinatorial infection and in vivo recombination: A strategy for making large phage antibody repertoires. Nucl. Acids Res. 21(9):2265-2266.

Whitlow, M., and D. Filpula. 1991. Single-chain Fv proteins and their fusion proteins. METHODS: A Companion to Methods in Enzymology 2(2):97-105.

Whitlow, M., B. A. Bell, S- L. Feng, D. Filpula, K. D. Hardman, S. L. Hubert, M. L. Rollence, J. F. Wood, M. E. Schott, D. E. Milenic, T. Yokota, and J. Schlom. 1993. An improved linker for single-chain Fv with reduced aggregation and enhanced proteolytic stability. Protein Eng. 6(8):989-995.

Wilson, I. A., and R. L. Stanfield. 1993. Antibody-antigen interactions. Curr. Opin. Struct. Biol. 3:113-118.

Wu, X- C., S- C. Ng, R. I. Near, and S- L. Wong. 1993. Efficient production of a functional single-chain antidigoxin antibody via an engineered Bacillus subtilis expression-secretion system. Bio/Technol. 11:71-76.

Zou, Y- R., H. Gu, and K. Rajewsky. 1993. Generation of a mouse strain that produces immunoglobulin k chains with human constant regions. Science 262:1271-1274.

FIGURE 1 The overall structure and domains in whole monoclonal antibody molecules (MAbs), Fab fragments, and single-chain Fvs (ScFv) discussed in this article.

FIGURE 2 Organization of the four framework (FR) and three complementarity-determining loops (CDRs) in variable regions, and the constant domains in heavy and light chains. Each diagram may be thought of as representing the coding regions in cDNA shown in Figure 3. The arrows indicate where oligonucleotide primers would hybridize to the strand, and direction in which a new complementary strand would be propogated in the PCR. For Fabs (top two structures) the primers would amplify the constant domain adjacent to the variable domain. For Fvs (bottom two structures) only the variable domains are amplified.

FIGURE 3 Steps in the derivation (A), selection, and expression (B) of recombinant Fabs, as described in the text. The process for selecting and expressing ScFvs is very similar, but the arrangement and transcription of genes in the vector, and the method of making soluble ScFv are somewhat different.

A. Total RNA is extracted from splenocytes, peripheral lymphocytes, or hybridoma cells. The polyadenylated mRNA, which includes the antibody coding mRNAs, is purified, and a cDNA is made. The cDNAs encoding H and L domains are amplified by PCR using specific oligonucleotide primers. The PCR products are purified, treated with restriction enzymes, and ligated into the cloning vector. The vector contains appropriate transcriptional promoters, leader, ribosome-binding, and translational stop sequences, as well as an antibiotic resistance gene to facilitate selection. E. coli cells are transformed with the vector and the diverse population is allowed to replicate.

B. Co-infection of the host with helper phage results in replication of phage that display the antibodies encoded in their DNA. The phage are concentrated and selected by binding to antigen. Phage enriched by selection are amplified by growth in E. coli. After several rounds of selection and amplification, phage displaying the desired antibodies are identified by immunoassay. These can be used to infect new host cells under conditions where soluble Fab will be expressed.

Copyright © 2008. National Academy of Sciences.
All rights reserved.
500 Fifth St. N.W., Washington, D.C. 20001.
Terms of Use and Privacy Statement