ENGINEERING ANTIBODY FRAGMENTS: REPLICATING THE IMMUNE SYSTEM AND BEYOND

LA INGENIERÍA DE FRAGMENTOS DE ANTICUERPOS: IMITANDO Y EXPANDIENDO EL SISTEMA INMUNE

Felipe García Quiroz^1,2, S. Michael Sinclair¹

1 Biomedical Engineering Department, Duke University, Durham, North Carolina, EE. UU.
2 Contact e-mail: felipe.garcia@duke.edu.

Received May 25, 2010. Accepted June 29, 2010

ABSTRACT

KEY WORDS: Antibody engineering, Phase display, Cell display, Ribosome display, Antibody humanization.

RESUMEN

Las tecnologías para la ingeniería de anticuerpos han evolucionado durante las últimas dos décadas, desde la demostración de la posibilidad de humanizar anticuerpos monoclonales de ratón mediante ingeniería genética, apoyadas en el creciente entendimiento de los mecanismos involucrados en la generación de anticuerpos in vivo, y en una búsqueda constante de rutas alternativas para evolucionar y explotar sus características. Es así como los ingenieros de anticuerpos han desarrollado estrategias innovadoras para la evolución y selección de anticuerpos y de novedosos diseños de anticuerpos conocidos como fragmentos de anticuerpos. Esta revisión se enfoca en tres tecnologías que comprenden el núcleo de las tecnologías actualmente disponibles para el descubrimiento y preparación de tales anticuerpos: la presentación en fagos, la presentación en células, y la presentación en ribosomas. Este artículo busca presentar el estado del arte de estas tecnologías a un grupo general de lectores interesados en este campo, por lo que inicialmente se introducen importantes conceptos de inmunología requeridos para comprender en detalle las tecnologías discutidas. Una comparación de estas metodologías para la ingeniería de anticuerpos sugiere que a pesar del dominio de las tecnologías basadas en la presentación en fagos durante los últimos 20 años, en los próximos años la presentación en células y la presentación en ribosomas probablemente ganarán importancia para la selección y descubrimiento de fragmentos de anticuerpos. Finalmente, es probable que estas tecnologías jueguen un papel importante en la producción de la siguiente generación de terapéuticos basados en anticuerpos.

PALABRAS CLAVE: Ingeniería de anticuerpos, Presentación en fagos, Presentación en células, Presentación en ribosomas, humanización de anticuerpos.

I. INTRODUCTION

Most recent reviews in the field of antibody engineering have examined in great detail the dynamics of the clinical transfer of antibody engineering technology developed for therapeutic purposes. Substantial emphasis has been placed on the characteristics of the antibodies being used, their targets and mechanisms, and the opportunities and challenges for the continuous progress of the field, particularly the remaining limitations of the state-of-the-art technology for antibody production [1-4]. Because of this emphasis, previous reviews have been directed toward a relatively specialized audience of antibody engineers in need of constant feedback on the increasing number of antibody-based therapeutic strategies under clinical trials, since the outcome of these trials significantly affects new research initiatives and thus the evolution of the field. However, the possibility to engineer human antibodies and novel related proteins against virtually any target has broad biomedical impact, providing for a means to neutralize (i.e., render inactive through antibody binding) key soluble proteins or receptors involved in the onset or progression of disease (e.g., chronic inflammation, cancer), or develop a means to target and release additional therapeutic cargos to specific cell populations (e.g., cancer cells) in the body. Hence, this short review article is aimed at a more general readership, who may have an interest in this technology but may not be acquainted with the immunology concepts required for understanding the relevant literature in this field. This review surveys the current technologies for engineering antibodies with a focus on the methodologies for developing antibody fragments and novel engineered proteins inspired by the structural components of complete antibodies. These novel technologies provide an important alternative to traditional antibody-based technologies and are often better suited for certain biomedical applications than conventional monoclonal antibodies.

This section introduces important immunology concepts essential to understanding antibody engineering strategies, their rationale, relevance, challenges and limitations. In some cases, the in vivo processes are contrasted with their engineered counterparts, although additional analogies will become evident throughout later sections of the article. These concepts may lie in any of three categories: (i) antibody structure and (ii) function, and (iii) diversity of the immune repertoires. Figure 1 summarizes basic information regarding antibody structure and function, and Fig. 2 and Table 1 detail the concepts related to antibody diversity.

Figure 1

Figure 2

Table 1

Antibodies, or immunoglobulins, are heterodimers composed of two identical light (L) chains and two identical heavy (H) chains. One light chain is covalently linked to one heavy chain by a disulfide bond, and the resulting H-L structures are joined as a dimer of dimers (i.e., H₂L₂) by additional disulfide bonds between heavy chains (Fig. 1A). The heterodimeric structure is further stabilized by non-covalent interactions, such as hydrophobic interactions, hydrogen bonds, and salt-linkages. Early investigations into the structure of antibodies using enzymatic digestion helped to elucidate the Y-shaped structure of antibodies. Digestion with papain resulted into two antigen-binding fragments (Fab) and one crystallizable fragment (Fc), while digestion with pepsin resulted in a single antigen binding fragment comprised of two antigen-binding domains (F(ab')₂) [6].

Genetic analysis of antibodies isolated from human subjects provided further understanding of the immunoglobulin structure and variability. The first 110 amino acids of the N-terminal segments of H and L chains are highly variable sequences called the V_L and V_H domains, which account for most of the differences in specificity displayed by native antibodies [6]. The unique sequences of the V_L and V_H for a given antibody determine its idiotype (i.e., antigenic determinants). The cleft between a V_L and V_H chain is the antigen binding pocket, and the specificity of antibody-antigen binding is predominantly controlled by 6 segmented, hypervariable loops called the complementarity-determining regions (CDRs) that extend from a highly ordered β-pleated structure characteristic of the immunoglobulin folding (Fig. 1B). While the CDRs are primarily responsible for antigen specificity, the whole variable domain serves as a scaffold for the correct presentation of the binding site, and mutations along its sequence also influence, to a minor extent, antibody affinity [4,6].

The remaining amino acids of the H and L chains are highly conserved regions known as constant domains (C_H or C_L). Heavy chains have 3 to 4 C_H domains, whereas L chains have a single C_L domain (Fig. 1) encoded by one of two light-chain genes, kappa (κ) or lambda (λ). The class of an antibody is determined by its heavy chain, of which there are five different chains or isotypes: α, δ, ε, γ and µ. Immunoglobulin G (IgG) is made up of two γ heavy chains and is the most abundant (~80% of total serum immunoglobulin) and most studied immunoglobulin class for antibody engineering (Fig. 1A). The structure and functions of the other immunoglobulin classes (i.e., IgA, IgD, IgE and IgM), which play important roles in adaptive immunity, will not be discussed due to their minor role in current antibody engineering applications. Subtle amino acid differences encoded in the C_H germ-line genes lead to a further division of isotypes into subisotypes or subclasses. In humans, for instance, there are four subisotypes of γ heavy chains (γ1, γ2, γ3, and γ4) with 90-95% homology between their genes. Additionally, different members of the same species may have multiple alleles for the same isotype genes, which determine the antibody allotype.

The isotype and subisotype of an antibody strongly impact the structure and effector functions of the Fc region. Because of this, the selection of the isotype is relevant for engineering antibodies, since different applications may require the mediation of different effector functions or, even more, their absence [4,5]. The existence of different Fc regions modulates the binding to specific Fc receptors found in immune effector cells; Fc gamma receptors (FcγR) in the case of IgG;, which trigger different effector functions upon binding of the antibody-antigen complexes, such as complement activation (component C1), antibody-dependent cell-mediated cytotoxicity (ADCC), opsonization (phagocytosis by macrophages and neutrophils) and transcytosis (crossing of epithelial layers). In the case of IgG, the Fc region also has the ability to bind to the neonatal Fc receptor (FcR_N), which plays a critical role in the regulation of IgG pharmacokinetics, since the binding to the FcR_N constitutes a salvage mechanism that recycles IgG and therefore allows for prolonged serum half-lives. Despite the importance of the Fc fragment in the modulation of effector functions, and although it is amenable to tailoring antibody pharmacokinetics (i.e., select antibodies with increased affinity to FcR_N) and has the ability to trigger specific effector functions (i.e., ADCC to tumor cells expressing the target antigen), the antibody engineering technologies discussed in this article focus on the antibody-antigen interaction, and are optimized and selected in formats devoid of Fc regions [2,7]. However, it should be noted that the modularity of the antibody structures also allows for the grafting of Fc regions into optimized antibody fragments (e.g., variable regions), although this usually requires the expression of the antibody fragment in eukaryotic expression systems [8].

The ability of the immune system to generate antibodies against virtually any antigen depends on its ability to generate a sufficient number of antibodies that can be selected based on their affinity for binding the antigen. The mechanisms involved in the generation of such diversity span different levels of cell physiology and are tightly associated with the maturation and differentiation of B cells, which are responsible for their production and secretion in vivo [6]. The main mechanisms involved in the generation of antibody diversity, as depicted in Fig. 2, are further explained in Table 1, which account for the tremendous diversity (>10¹⁰) of the immune repertoire. In addition, the role of these mechanisms or their analogues in generating antibody diversity in existing antibody engineering technologies is indicated.

III. ENGINEERING ANTIBODY FRAGMENTS

The modular structure of antibodies has enabled the customization and engineering of high affinity binders in a variety of ways. Before discussing the technologies developed for the design and discovery of antibody fragments, the available antibody fragment formats are presented, since those technologies, as will be noted in section IV, are only suited for particular antibody formats.

Figure 3 depicts the available battery of antibody fragments derived from the parental IgG structure. The seminal work on the engineering of these new sets of antibody formats was conducted on Fab fragments -comprised of one antigen binding site of an IgG (V_H-C_H+V_L-C_L)-, and on single-chain variable fragments (scFv), a further simplification of the Fab structure achieved by removing the constant domains and linking the V_H and V_L fragments with a peptide linker [1]. The scFv format rapidly popularized, and is probably the most widely used antibody fragment today, mainly due to the advantages of directly linking the heavy and light domain genes. Linking these domains at the genetic level not only simplified the recombinant DNA methods involved in their processing, but significantly increased the stability of the structure and eliminated the folding problems encountered with prokaryotic expression systems (e.g., E. coli) during selection and production of antibodies with disulfide bonds [2,8]. Interestingly, the incorporation of the peptide linker and the variation of its length has been found to control the dimerization properties of the scFv fragments, with shorter sequences resulting in increasing valency (diabody, triabody and tetrabody formats have been produced). The absence of linker, which prevents the self-folding of the V_H and V_L domains of one scFv promotes the formation of bispecific scFv by noncovalent interactions between the variable domains of a second scFv [3].

Figure 3

The maximum simplification of the antibody structure, known as domain antibody (dAb), consists of a single V_H or V_L domain (i.e., only 3 CDRs). The initial attempts to derive high affinity binders using dAb were not encouraging, resulting in the selection of fragments displaying significant decreases in binding affinity, but most importantly, poor stability and a tendency to aggregate [1]. Nevertheless, the finding of dAb naturally occurring in camels, which displayed high affinity and stability, inspired the design of new dAb circumventing these problems, in a process termed "camelization" [12]. Despite the success of the camelization approach, the therapeutic applications of such antibodies were limited due to the potential immunogenicity associated with using non-human scaffolds in the variable region design [2]. Only recently, Winter and coworkers, in their efforts to characterize a set of dAbs produced against hen egg lysozyme (HEL), discovered an antibody domain displaying similar properties to those found in camel and llama V_HH Abs but without recurring to camel-based scaffolds (i.e., camelising mutations). The same group also devised a methodology for the generation of equally stable and aggregation-resistant domain antibodies [13]. One important realization of this work was the increased understanding of the role of the CDRs in determining the thermodynamic stability, as well as expression and purification yields, of antibodies [11,12].

IV. ANTIBODY ENGINEERING TECHNOLOGIES

The discovery of the hybridoma technology in 1975 enabled the production of monoclonal antibodies (mAbs) and paved the way for the evolution of the Antibody Engineering field [14]. The therapeutic potential of such technology became evident in 1984, when Winter and collaborators demonstrated an ability to form chimeric antibodies with murine antigen-binding domains and complete human effector functions (i.e., Fc region). In 1986, the same group developed the groundbreaking antibody humanization technology, which involves transferring the CDR regions of a murine monoclonal antibody into a human immunoglobulin scaffold, significantly reducing the immunogenicity issues associated with murine antibodies [15]. Current antibody engineering technologies have surpassed many of the challenges imposed by the selection of antibodies using murine cell lines (i.e., hybridoma technology), eliminating the need for humanization by enabling the production of fully human antibodies in vitro or in other engineered animal models. Hence, the technologies presented in this section will focus on these alternative methods for the selection and production of human antibody fragments.

4.1 Antibody libraries

Due to the complexity of the immune repertoire, the initial approaches for the construction of antibody libraries followed a simple strategy: the amplification of assembled antibody genes after mice immunization by means of RT-PCR using a set of primers designed for the amplification of all antibody genes and based on the variable region frameworks (already known by the time and deposited in data bases: Kabat and V-base database) [19]. However, this approach still used mice for the generation of the assembled antibody genes after immunization, therefore presenting only partial advantages. An additional level of complexity was included by applying a similar strategy for the amplification of naïve libraries (i.e., gene segments before recombination) from non-immunized animals followed by in vitro combinatorial assembly of the antibody repertoire [10].

Despite these significant advances, the antibody fragment screening and production technologies relied on non-mammalian systems, which suffer from inadequate expression levels and other problems derived from differences in codon usage. As a result, the development of semisynthetic and later of fully synthetic human antibody libraries represented an important achievement for the antibody engineering field. These libraries can now be optimized for expression according to the selection technology and desired expression system, have modular designs that allow relatively easy interconversion between different antibody formats, and significantly simplify laborious DNA manipulation steps. In addition, synthetic libraries are not limited by the bias introduced in germ-line repertoires throughout evolution, such as the tolerance mechanism against selection of self-antigens, and therefore enable, at least in theory, the discovery and selection of antibodies with no representation in natural immune repertoires [4,20].

Figure 4 presents the designs of the most advanced human synthetic libraries currently available, which are known as Human Combinatorial Antibody Libraries (HuCAL) [17,21]. Initially introduced in 2000, this synthetic library implemented innovative concepts for the generation of diversity, including diversity not only in the CDRs but also in the framework regions, which are known to play a role in CDR conformation. In addition, the diversity introduced in the CDR libraries is biased towards sequences predominant in the human immune repertoire (by using trinucleotide cassette mutagenesis), which facilitates the selection of antibody fragments with minimal or no immunogenicity (human anti-human antibody, HAHA) [9,16]. It is worth noting that the diversity in the V_H and V_L gene families, as well as the families selected, were carefully analyzed by bioinformatics means to achieve sufficient diversity while preventing excessive complexity of the library. Indeed, this library only uses 7 master genes for heavy chains and 7 genes for light chains corresponding to consensus sequences for seven V_H and seven V_L germ-line families which were found to account for more than 95% of the human antibody diversity observed in vivo. The library was initially developed in scFv format, but is now also available for Fab fragments [10]. Some characteristics of the newest versions of the HuCAL library (Fig. 4B and 4C), HuCAL Fab 1 and HuCAL GOLD, that require special attention are: (i) the absence of cysteine residues in the constant domains (eliminated to avoid problems during expression), (ii) that only Fd (V_H + C_H) is covalently attached to pIII (for phage display, reviewed in next section), so that the system depends on the non-covalent interactions with the light chain, and (iii) the absence of cysteine residues in the CDR regions in the HuCAL GOLD library (to avoid problems with the CysDisplayTM technology). Although (ii) may be complicated by light chain exchange in a given phage preparation, thereby losing the linkage of genotype to phenotype, the authors claim that after extensive use of the library this non-covalent interaction proved very stable [20,21].

Figure 4

Two powerful technologies for screening antibody libraries and selecting antibodies of high affinity for a particular antigen involve the display of one (monovalent) or several (multivalent) antibody particles on the surface of either phage virion, phage display (section 4.2), or on the cell surface (i.e., cell wall, cell membrane) of a prokaryotic or eukaryotic host, cell display (section 4.3).

4.2 Phage display

Phage display is a powerful biomolecular engineering technique for selecting high affinity binders to biologically relevant targets by several rounds of affinity selection. Foreign DNA encoding recombinant peptides or proteins is fused to coat protein DNA of bacteriophage such that recombinant molecules will be expressed and displayed on the outer surface of the phage. This strategy effectively links the protein phenotype and genotype (i.e., the corresponding DNA carried by the phage) thereby enabling the simple identification of the selected proteins at the DNA level. In the seminal publication of phage display [22], fragment genes of the endonuclease EcoRI were fused to the gene III protein (g3p) of fd filamentous phage to produce "fusion phage" capable of yielding peptides with 1000-fold higher affinity for anti-EcoRI antibody. Winter's group then demonstrated the possibility to display functional antigen-binding sites on the surface of these phage particles for their evolution [23,24], and subsequently contributed to the seminal work on the construction of large phage antibody libraries [25]. By linking phenotype to genotype, vast libraries of phage (10⁹-10¹² clones) displaying different fusion proteins can be assembled, selected for with simple affinity techniques (Fig. 5A), and quickly identified by conventional DNA sequencing [26]. Marks et al. (1991) prepared a library of scFv genes from peripheral blood lymphocytes isolated from unimmunized human donors by RT-PCR, which contained randomly generated heavy and light chain variable fragments. After affinity selection, phage displaying scFv demonstrated affinity for their target on par with the affinity of native antibodies for the same target, indicating the utility of phage display for bypassing immunization and producing high affinity binders. Today, phage display constitutes the most developed in vitro system for the selection of antibody fragments [4,27].

Figure 5

Filamentous phage (Ff), bacteriophage T4, and phage λ are the three general types of phage used for phage display. Initial bacteriophage studies used filamentous M13 phage to amplify and isolate single-stranded DNA [28], which led to M13-based plasmid constructs (M13mp18/19) that have become the basis for most phage display systems [29]. While M13-based systems are the most developed, bacteriophage T4 and phage λ systems have shown equal promise. Bacteriophage T4 are able to house significantly larger quantities of DNA [30], and phage λ display has shown increased ability to incorporate larger proteins and cDNA libraries compared with Ff phage display [31].

Phage display methods are named depending upon which phage protein has been fused, how many copies of the fusion are present in one phage particle, and what type of phage vectors are employed, phage and/or phagemid vectors [26]. One important characteristic of the phage display technology is the control it offers over valency of display, since this parameter determines the avidity-functional affinity determined by the number of binding sites-of the selection strategy and significantly affects the affinity of the selected clones. For instance, high valency often results in antibodies with moderate affinity, since the higher, uncontrolled avidity increases the stability of the antigen-antibody complex independently of the antibody affinity, whereas monovalent display ensures the selection of the antibodies with the highest affinity for the antigen [36]. Type 3 phage display uses phage vectors with one copy of recombinant gene III protein (g3p) and will generate phage that display 3 to 5 copies of a recombinant p3. Type 33 phage display systems contain two copies of g3p gene, one recombinant and one wild type and will yield multivalent phage particles with recombinant and wild-type p3 proteins displayed on the phage surface. Type 3+3 phage display systems use two different vectors, helper phage vectors and phagemid vectors. Phagemid vectors are small plasmids with high transformation efficiency that contain all the necessary components for infection, house recombinant DNA encoding the antibody-g3p fusion protein, but lack assembly and export genes. Phagemid systems, therefore, require helper phage, which retain the genes for packaging proteins and wild-type protein genes. Phagemid vectors are packaged in preference to helper phage vectors, but both wild-type and recombinant proteins will be displayed. In fact, more than 90% of the recovered phagemid particles do not display antibodies. However, this mechanism ensures that phage displaying recombinant g3p generally display a single copy on the virion surface [16,36].

Phage and phagemid vectors have been engineered extensively and now include antibiotic resistance genes for selection, a multiple cloning site (MCS) for easy generation of libraries in frame, and optimal promoter and packaging signals [22,37]. Phagemid systems are generally more stable than phage vectors, which can spontaneously delete foreign DNA fragments. Phagemid systems are also more tolerant to larger peptide inserts and generation of larger libraries is easier [38]. Phagemid display, which presents a single copy of a recombinant g3p protein, is required for optimal affinity maturation studies like antibody engineering, as display of single proteins results in the selection of fewer unique binders without interference from the effects of avidity [24]. However, phagemid display requires the addition of helper phage at specific points of the bacterial growth cycle, and therefore presents operational difficulties not characteristic of pure phage display.

Recently, a variant of the traditional phage display strategy, CysDisplay^TM (http://www.morphosys.com/), was introduced that facilitates the recovery of antibody fragments with ultra-high affinity [16]. In CysDisplay^TM, the antibody fragment is linked to the coat protein through a disulfide bond instead of the direct peptide linkage when expressed as a fusion protein. Because of this, the recovery of the phage particles is easily achieved by adding reducing agents and is rendered independent of the antibody affinity to the antigen.

4.3. Cell display

Cell display, on the other hand, uses an analogous system wherein the antibody gene is fused to a protein naturally displayed on the outer membrane of a cell, as shown in Fig. 5B. Although cell display has been used with prokaryotic cells, the most current technology mainly uses eukaryotic systems such as yeasts and mammalian cells [18,39]. This strategy, in opposition to phage display, only allows for multivalent display, whereby a large number of antibody copies are displayed on the outer membrane of the cell. While this strategy may be complicated by increased avidity compared to monovalent display systems, it benefits from existing technologies for studying eukaryotic cells, such as flow-cytometry and fluorescence activated cell sorting (FACS) ,which enable simultaneous selection and characterization of antibody fragments kinetics [40]. The basic strategy incorporates additional tags in the fusion protein in order to quantify the expression levels of the antibody fragment (i.e., the valency of the display), thus allowing the normalization of the antigen-antibody fluorescence signal to discriminate avidity effects, and also for selecting antibody fragments already optimized for high expression and the early removal of truncated products during the selection process [40].

The implementation of the cell display technology in a yeast model may also offer additional advantages compared with screening in prokaryotes by phage display. For instance, expression biases and growth selection introduced by bacteria can skew library diversity. In contrast, yeast models have been shown to propagate full library diversity along the selection process [18]. Figure 5B includes additional details about the construction of the display system, and compares the design of the fusion proteins involved in the two variants of the cell display technology herein discussed with the phage display method. In addition, a comprehensive comparison of the phage display, cell display and ribosome display (reviewed in next section) technologies is presented in Table 2.

4.4 Ribosome display

The development of the ribosome display technology by Hanes and Plückthun in 1997, based on earlier work by Mattheakis et al., 1994, represented a breakthrough in the protein engineering field by demonstrating the ability to screen libraries of unprecedented size in a cell-free system [41-43]. The advantages and disadvantages of this novel technology, in contrast with phage and cell display, are presented in Table 2; cell-free systems, for instance, overcome the basic limitation of library size imposed by transformation efficiency for the previously discussed methods [44]. Ribosome display, which has evolved over the past 10 years, is now actively used for the development of new antibody fragments in various applications [17,45,46].

Table 2

Figure 6

Ribosome display has been developed using either prokaryotic gene expression machinery (e.g., bacterial extracts) [17] or eukaryotic systems [44,47]. In both cases, the technologies are almost indistinguishable, although the use of eukaryotic ribosome display allows direct cDNA synthesis from the Antibody -Ribosome-mRNA (ARM) complexes by in-situ RT-PCR, as demonstrated by He and Taussig in 1997, and avoids some technical issues of prokaryotic ribosome display related to ARM complex disruption during elution of mRNA [44,48]. Hence, this approach reduces loss of material during ribosome disruption and mRNA recovery, which is critical to maximize diversity.

In 2007, Contreras-Martínez and Delisa proposed a variation of the ribosome display technology for the evolution of intracellular antibodies [49]. Intracellular ribosome display, which makes use of the recently discovered E. coli SecM translation arrest mechanism to allow the formation of mRNA-ribosome-antibody (fused to SecM signal), enables stable stalling of the ribosome (intracellularly) at the SecM sequence (serving also as spacer for appropriate folding) and subsequent recovery of the complexes by centrifugation. Additional selection steps proceed in a similar fashion to standard ribosome display [49].

4.5 Other aspects of the antibody engineering process

The flexibility in the design of antibody fragments (Fig. 3) and the variety of technologies available for their discovery and optimization, as discussed in Table 2, allow for myriad possibilities when tailoring the properties of these unique proteins according to the design parameters dictated by the intended application. The readers are strongly encouraged to examine excellent recent review articles discussing some important aspects of the antibody engineering process that were not the subject of this review, including: antibody characterization (i.e., quantification of kinetics, stability, immunogenicity, etc.) [1,4], pharmacokinetics of monoclonal antibodies and antibody fragments (of particular relevance for imaging applications and cancer therapy) [50,51], current expression systems for the production of antibody fragments [8,52], and finally, design of antibody-based therapeutics, an active area of research that involves engineering of the antibody fragments at all levels, particularly tailoring the effector functions of the antibody fragments which may involve the engineering of the Fc region [3,53-55].

V. CONCLUSION

The ability to rapidly engineer antibodies against virtually any antigenic biomolecule, from mRNA to small haptens to big antigenic particles or even molecules with cryptic epitopes (as in the case of dAb), with unprecedented affinities, effector functions (i.e., bispecific antibodies, catalytic antibodies, intrabodies, etc.) and stability, has been the result of the powerful antibody discovery technologies implemented in the past 20 years. Although phage display has governed the production of antibody fragments since its introduction, the new advances in ribosome display technology and cell display will probably continue to gain importance for the selection of the antibody fragments. Expected improvements in cell sorting technology might significantly increase the throughput of the cell display system, its major limitation as of today, although the recent demonstration of antibody maturation by mammalian cell display may also stimulate the use of this technology in combination with phage display in the late stages of the antibody maturation process (i.e., stability, expression, etc.). Finally, these technologies are likely to play an important role in the production of the next generation of antibodybased therapeutics.

[1] Maynard J., Georgiou G. Antibody engineering. Annual review of biomedical engineering, 2:339-376, 2000.         [ Links ]

[2] Holliger P., Hudson P.J. Engineered antibody fragments and the rise of single domains. Nature biotechnology, 23:1126-1136, 2005.         [ Links ]

[3] Carter P.J. Potent antibody therapeutics by design. Nature reviews. Immunology, 6:343-357, 2006.         [ Links ]

[4] Filpula D. Antibody engineering and modification technologies. Biomolecular engineering, 24:201-215, 2007.         [ Links ]

[5] Salfeld J.G. Isotype selection in antibody engineering. Nature biotechnology, 25:1369-1372, 2007.         [ Links ]

[6] Goldsby R., Kindt T., Osborne B., Kuby J. Immunology (5th edn): WH Freeman, New York, NY, 2003.         [ Links ]

[7] Hoogenboom H.R. Selecting and screening recombinant antibody libraries. Nature biotechnology, 23:1105-1116, 2005.         [ Links ]

[8] Leong S., Chen W. Preparing recombinant single chain antibodies. Chemical Engineering Science, 63:1401-1414, 2008.         [ Links ]

[9] Lonberg N. Human antibodies from transgenic animals. Nature biotechnology, 23:1117-1125, 2005.         [ Links ]

[10] Knappik A., Ge L., Honegger A., Pack P., Fischer M., Wellnhofer G., Hoess A., Wolle J., Pluckthun A., Virnekas B. Fully synthetic human combinatorial antibody libraries (HuCAL) based on modular consensus frameworks and CDRs randomized with trinucleotides. Journal of Molelcular Biology, 296:57-86, 2000.         [ Links ]

[11] Christ D., Famm K., Winter G. Repertoires of aggregationresistant human antibody domains. Protein engineering, design and selection: PEDS, 20:413-416, 2007.         [ Links ]

[12] Jespers L., Schon O., Famm K., Winter G. Aggregation-resistant domain antibodies selected on phage by heat denaturation. Nature biotechnology, 22:1161-1165, 2004.         [ Links ]

[13] Jespers L., Schon O., James L., Veprintsev D., Winter G. Crystal structure of hel4, a soluble, refoldable human vh single domain with a germ-line scaffold. Journal of molecular biology, 337:893-903, 2004.         [ Links ]

[14] Weiner L., Surana R., Wang S. Monoclonal antibodies: Versatile platforms for cancer immunotherapy. Nature Reviews Immunology, 10:317-327, 2010.         [ Links ]

[15] Winter G.P. Antibody engineering. Philosophical transactions of the Royal Society of London. Series B, Biological sciences, 324:537-546; discussion 547, 1989.         [ Links ]

[16] Kretzschmar T., von Rüden T. Antibody discovery: Phage display. Current opinion in biotechnology, 13:598-602, 2002.         [ Links ]

[17] Hanes J., Schaffitzel C., Knappik A., Pluckthun A. Picomolar affinity antibodies from a fully synthetic naive library selected and evolved by ribosome display. Nature biotechnology, 18:1287-1292, 2000.         [ Links ]

[18] Feldhaus M., Siegel R., Opresko L., Coleman J., Feldhaus J., Yeung Y., Cochran J., Heinzelman P., Colby D., Swers J. Flowcytometric isolation of human antibodies from a nonimmune saccharomyces cerevisiae surface display library. Nature biotechnology, 21:163-170, 2003.         [ Links ]

[19] Sblattero D., Bradbury A. A definitive set of oligonucleotide primers for amplifying human v regions. Immunotechnology, 3:271-278, 1998.         [ Links ]

[20] Rauchenberger R., Borges E., Thomassen-Wolf E., Rom E, Adar R., Yaniv Y, Malka M., Chumakov I., Kotzer S., Resnitzky D., Knappik A., Reiffert S., Prassler J., Jury K., Waldherr D., Bauer S., Kretzschmar T., Yayon A., Rothe C. Human combinatorial Fab library yielding specific and functional antibodies against the human fibroblast growth factor receptor 3. Journal of Biological Chemistry, 278:38194-38205, 2003.         [ Links ]

[21] Rothe C., Urlinger S., Lohning C., Prassler J., Stark Y., Jager U., Hubner B., Bardroff M., Pradel I., Boss M., Bittlingmaier R., Bataa T., Frisch C., Brocks B., Honegger A., Urban M. The human combinatorial antibody library HuCAL gold combines diversification of all six CDRs according to the natural immune system with a novel display method for efficient selection of high-affinity antibodies. Journal of molecular biology, 376:1182-1200, 2008.         [ Links ]

[22] Smith G. Filamentous fusion phage: Novel expression vectors that display cloned antigens on the virion surface. Science, 228:1315-1317, 1985.         [ Links ]

[23] McCafferty J., Griffiths A.D., Winter G., Chiswell D.J. Phage antibodies: Filamentous phage displaying antibody variable domains. Nature, 348:552-554, 1990.         [ Links ]

[24] Marks J., Hoogenboom H., Bonnert T., McCafferty J., Griffiths A., Winter G. By-passing immunization: Human antibodies from v-gene libraries displayed on phage. Journal of molecular biology, 222:581-597, 1991.         [ Links ]

[25] Sheets M.D., Amersdorfer P., Finnern R., Sargent P., Lindquist E., Schier R., Hemingsen G., Wong C., Gerhart J.C., Marks J.D. Efficient construction of a large nonimmune phage antibody library: The production of high-affinity human single-chain antibodies to protein antigens. Proceedings of the National Academy of Sciences of the United States of America, 95:6157-6162, 1998.         [ Links ]

[26] Smith G., Petrenko V. Phage display. Chemical Reviews, 97:391-410, 1997.         [ Links ]

[27] Lee C.M., Iorno N., Sierro F., Christ D. Selection of human antibody fragments by phage display. Nature Protocols, 2:3001-3008, 2007.         [ Links ]

[28] Messing J., Gronenborn B., Müller-Hill B., Hans Hopschneider P. Filamentous coliphage M13 as a cloning vehicle: Insertion of a HindII fragment of the Lac regulatory region in M13 replicative form in vitro. Proceedings of the National Academy of Sciences of the United States of America, 74:3642, 1977.         [ Links ]

[29] Yanisch-Perron C., Vieira J., Messing J. Improved M13 phage cloning vectors and host strains: Nucleotide sequences of the m13mpl8 and puc19 vectors. Gene, 33:103-119, 1985.         [ Links ]

[30] Efimov V., Nepluev I., Mesyanzhinov V. Bacteriophage T4 as a surface display vector. Virus Genes, 10:173-177, 1995.         [ Links ]

[31] Santini C., Brennan D., Mennuni C., Hoess R.H., Nicosia A., Cortese R., Luzzago A. Efficient display of an HCV cDNA expression library as C-terminal fusion to the capsid protein d of bacteriophage lambda. Journal of molecular biology, 282:125-135, 1998.         [ Links ]

[32] Clackson T., Wells J. In vitro selection from protein and peptide libraries. Trends in biotechnology, 12:173-184, 1994.         [ Links ]

[33] Sioud M., Dybwad A., Jespersen L., Suleyman S., Natvig J., Førre O. Characterization of naturally occurring autoantibodies against tumour necrosis factor-alpha (TNF-alpha): In vitro function and precise epitope mapping by phage epitope library. Clinical and experimental immunology, 98:520, 1994.         [ Links ]

[34] Rickles R., Botfield M., Zhou X., Henry P., Brugge J., Zoller M. Phage display selection of ligand residues important for Src homology 3 domain binding specificity. Proceedings of the National Academy of Sciences of the United States of America, 92:10909, 1995.         [ Links ]

[35] Cwirla S., Peters E., Barrett R., Dower W. Peptides on phage: A vast library of peptides for identifying ligands. Proceedings of the National Academy of Sciences of the United States of America, 87:6378-6382, 1990.         [ Links ]

[36] Griffiths A., Duncan A. Strategies for selection of antibodies by phage display. Current opinion in biotechnology, 9:102-108, 1998.         [ Links ]

[37] Wilson D., Finlay B. Phage display: Applications, innovations, and issues in phage and host biology. Canadian journal of microbiology, 44:313-329, 1998.         [ Links ]

[38] Chasteen L., Ayriss J., Pavlik P., Bradbury A.R. Eliminating helper phage from phage display. Nucleic acids research, 34:e145, 2006.         [ Links ]

[39] Ho M., Nagata S., Pastan I. Isolation of anti-cd22 fv with high affinity by fv display on human cells. Proceedings of the National Academy of Sciences of the United States of America, 103:9637-9642, 2006.         [ Links ]

[40] Feldhaus M.J., Siegel R.W. Yeast display of antibody fragments: A discovery and characterization platform. Journal of immunological methods, 290:69-80, 2004.         [ Links ]

[41] Hanes J., Plückthun A. In vitro selection and evolution of functional proteins by using ribosome display. Proceedings of the National Academy of Sciences of the United States of America, 94:4937, 1997.         [ Links ]

[42] Schaffitzel C., Hanes J., Jermutus L., Pluckthun A. Ribosome display: An in vitro method for selection and evolution of antibodies from libraries. Journal of immunological methods, 231:119-135, 1999.         [ Links ]

[43] Mattheakis L., Bhatt R., Dower W. An in vitro polysome display system for identifying ligands from very large peptide libraries. Proceedings of the National Academy of Sciences of the United States of America, 91:9022, 1994.         [ Links ]

[44] He M., Taussig M.J. Ribosome display of antibodies: Expression, specificity and recovery in a eukaryotic system. Journal of immunological methods, 297:73-82, 2005.         [ Links ]

[45] Yan X., Xu Z. Ribosome-display technology: Applications for directed evolution of functional proteins. Drug discovery today, 11:911-916, 2006.         [ Links ]

[46] Zahnd C., Amstutz P., Plückthun A. Ribosome display: Selecting and evolving proteins in vitro that specifically bind to a target. Nature Methods, 4:269-279, 2007.         [ Links ]

[47] He M., Taussig M.J. Eukaryotic ribosome display with in situ DNA recovery. Nature Methods, 4:281-288, 2007.         [ Links ]

[48] He M., Taussig M. Antibody-ribosome-mRNA (ARM) complexes as efficient selection particles for in vitro display and evolution of antibody combining sites. Nucleic acids research, 25:5132, 1997.         [ Links ]

[49] Contreras-Martinez L.M., DeLisa M.P. Intracellular ribosome display via SecM translation arrest as a selection for antibodies with enhanced cytosolic stability. Journal of molecular biology, 372:513-524, 2007.         [ Links ]

[50] Batra S.K., Jain M., Wittel U.A., Chauhan S.C., Colcher D. Pharmacokinetics and biodistribution of genetically engineered antibodies. Current opinion in biotechnology, 13:603-608, 2002.         [ Links ]

[51] Wu A.M., Olafsen T. Antibodies for molecular imaging of cancer. The Cancer Journal, 14:191-197, 2008.         [ Links ]

[52] Birch J.R., Racher A.J. Antibody production. Advanced drug delivery reviews, 58:671-685, 2006.         [ Links ]

[53] Weiner L.M. Building better magic bullets-improving unconjugated monoclonal antibody therapy for cancer. Nature reviews. Cancer, 7:701-706, 2007.         [ Links ]

[54] Fischer N. New magic bullets can hit more than one target. Expert Opinion on Drug Discovery, 3:833-839, 2008.         [ Links ]

[55] Cardinale A., Biocca S. The potential of intracellular antibodies for therapeutic targeting of protein-misfolding diseases. Trends in molecular medicine, 14:373-380, 2008.         [ Links ]