- Top of page
- Conflict of interest
- Supporting Information
A type I hypersensitivity response is initiated when allergen-specific IgE bound to high-affinity IgE receptors (FcεRI) on the surface of effector cells, such as mast cells and basophils, are cross-linked by allergen. This triggers a cascade of events eventually giving rise to the release of biologically active compounds that exert their effects both locally and systemically [1, 2]. It is thus evident that IgE plays a key role, as the allergen recognizer, in the mechanisms behind the symptoms associated with an allergic response, thereby influencing the quality of life for as many as one-third of the population of the industrialized world .
Over the years, a range of studies have characterized the genetic composition of human IgE repertoires, in terms of V(D)J germline gene usage, mutational status and evidence of antigenic selection [4-13]. However, much work remains before we can understand the products of these antibody-encoding genes in terms of development and function.
One intriguing outcome of several of the past IgE-repertoire studies is a significant difference in the utilization of different immunoglobulin germline gene subgroups to produce the heavy chain variable (VH) domains of IgE in certain tissue as compared to IgE and other isotypes encoded by peripheral blood lymphocytes. More specifically, an overrepresentation of IgE-encoding transcripts derived from the immunoglobulin heavy variable (IGHV) 5 germline gene subgroup has been observed in such tissue samples [4-6]. The reason behind and outcome of such biased use of certain genes are a matter of debate. Two possible explanations are that such skewed repertoires are the result of either a polyclonal expansion of B cells expressing surface IgE with origin in these genes, for example, by bacterial superantigens  or of a clonal selection process where certain allergens favour the selection of clones originating in a limited set of germline genes .
The second partner in the initial interaction triggering an allergic response, the allergen, adds further complexity to the task of understanding the molecular basis of the early events in such a response. Not only do most allergens exist as numerous different isoforms with varying potencies to provoke allergic reactions [15, 16], but the existence of homologous allergens in related species also adds complexity to the analysis of development of human IgE repertoires. IgE-binding epitopes are to a certain extent shared between such homologous members of a protein family [17, 18] explaining the often broad sensitivity profiles of allergic individuals. A detailed analysis of monoclonal human IgE has the potential to address these matters in great detail as we have carried out in the past for groups 1 and 5 grass pollen allergens .
The major birch pollen allergen, Bet v 1, is one of the most potent, and therefore also the most studied, tree allergens in Europe . It belongs to the widely represented PR-10 family [SCOPe database (http://scop.berkeley.edu/) entry d.129.3.1]  within the pathogenesis-related protein Bet v I family (Pfam database: http://pfam.sanger.ac.uk/family/PF00407). This allergen exists in at least 18 different isoallergens and isoforms, as defined by IUIS . Although several studies have dealt with characterization of responses between polyclonal IgE preparations and different Bet v 1 isoforms and homologues [15, 16], the molecular basis of such interactions or discrimination of human IgE between PR-10 proteins of different origins still remains largely unresolved.
We here approach these issues and present a molecular characterization of Bet v 1-specific human antibody fragments derived from the IgE repertoire of allergic individuals. These antibody fragments have a genetic origin in IGHV5, the germline gene subgroup sometimes implicated as overrepresented in IgE repertoires [4-6]. We show that such a repertoire of clonally different antibodies, despite their origin in a common immunoglobulin VH gene, is able to target more than one epitope on Bet v 1, thereby in principle fulfilling the prerequisites for cross-linkage and initiation of the allergic cascade. We also describe one of the epitopes recognized by such an antibody fragment with origin in the IgE repertoire in detail. From that assessment, we are able to define the molecular basis for the inability of this antibody fragment to recognize certain isoforms of Bet v 1 and to cross-react with many other members of the PR-10 subfamily.
- Top of page
- Conflict of interest
- Supporting Information
Despite the proven usefulness of human monoclonal antibodies in studies of biomedically important antibody responses, for instance in viral and bacterial infection as well as in autoimmunity, researchers within the field of allergic disease have only recently started to make use of the huge potential contained in human monoclonal IgE. We have in two recent studies deconvoluted the information content of two sets of human monoclonal IgE-derived scFv specific for the two major grass pollen allergens Phl p 1  and Phl p 5, respectively, allowing for in-depth studies of the interaction between such antibody fragments and their corresponding allergens. On the same path, others have utilized both human and murine monoclonal antibodies aiming at a deeper understanding of the molecular interactions taking place at the initiation of an allergic response (exemplified by [43-47]). Studies of these kinds have, among other things, led to identification of new IgE-binding epitopes and a better understanding of aspects of the IgE response that define the biological effects exerted by such antibodies. We in this study continue on that path by presenting isolation of four novel human antibody fragments with origin in the IgE repertoire, all specific for the major birch allergen Bet v 1, and characterization of their interaction with the allergen.
Since the structure of Bet v 1 was solved almost two decades ago , numerous studies have identified single amino acids, peptide stretches or surface patches that to some extent are involved in the interaction between human IgE and the allergen. Recently, Hecker et al.  reported the first human IgE-binding epitope on Bet v 1 recognized by a scFv based on a VH with human IgE origin and a synthetic VL. However, the two IgE epitopes we report within this study are both distinctly different from the epitope reported by Hecker et al., and unlike previously characterized IgE antibodies several of the binders included in this study target epitopes that to some extent can be defined by short linear peptides. The more narrowly defined epitope of IgE-derived clones M0418 and B14 (residues I56 to K65) encompasses the loop connecting β-strands 3 and 4 of Bet v 1, and belongs to a conformationally flexible region of the allergen . This epitope partially overlaps that of the mouse monoclonal IgG1 antibody mAb2 shown to exhibit a substantial inhibition of binding of human serum IgE to Bet v 1 . In addition, earlier studies using mouse monoclonal antibodies derived by immunization with birch pollen identified this area as important for IgE binding . The second novel epitope described in this study, covering part of α-helix 2 and the loop connecting it to β-strand 2, belongs to a more structurally conserved region of the allergen and is recognized by B10 and most likely B13. This epitope resides in an area previously not identified as important for human IgE binding using human or murine monoclonals, but in vitro mutagenesis studies did identify the phenylalanine in position 30 to have influence on IgE binding . It is, however, difficult to conclude whether this mutation affects IgE binding directly or via induction of conformational changes in the protein.
The identification of novel and more precise characterization of such biologically relevant epitopes made possible by usage of antibody fragments, derived from human IgE repertoires, provides useful information that could aid in rational design of new hypoallergenic versions of Bet v 1 to be used as safer alternatives to the allergen extracts currently used in specific immunotherapy. The epitopes identified by us and previously by Hecker and colleagues  should be suitable targets for directed mutagenesis in attempts to reduce the IgE reactivity of the protein, in a strategy previously described for the major timothy grass pollen allergen Phl p 1 . Further, detailed knowledge on the location of IgE-binding epitopes would be valuable in future attempts to increase the resolution of allergy diagnosis, from component-resolved to epitope-resolved diagnosis.
Although adjacent to one another on the surface of Bet v 1, the two identified epitopes could allow for simultaneous binding of two scFv targeting their respective epitopes (Figure S3), thereby fulfilling the basic criteria of FcεRI cross-linkage. The optimal separation distance for such FcεRI cross-linkage, as determined by degranulation assays, is suggested to be 44-51Å , which may be rationalized by the disposition of Fabs in the asymmetrically bent FcεRI bound form of IgE . In the Bet v 1 monomer, the most distant Cα atoms from the M0418/ B14 and B13 epitopes are only ~20Å apart. However, the ability of Bet v 1 to cross-link receptor on effector cells was shown to require a dimer rather than a monomer . In the model of the proposed Bet v 1 dimer (Figure S8) , the distance between Cα atoms of Phe62, a critical residue for M0418 epitope, is 40Å, approaching the favourable distance for FcεRI cross-linking . While the distance between the M0418 and B13 epitopes of different Bet v 1 molecules within the dimer varies from ~13-35Å, conformational flexibility of the M0418 epitope, together with the protruding CDRH3 of the antigen binding site of M0418 (Fig. 5), could aid in optimal positioning of receptors on the cell surface, when different epitopes are engaged in cross-linking.
The presence of skewed IgE repertoires, with overrepresentation of VH-encoding transcripts with origin in only certain germline genes or gene subgroups, has been suggested in several previous studies [4-6], but we here, to our knowledge for the first time, show that such a repertoire, here restricted to the usage of transcripts with origin in IGHV5-51, despite its theoretical loss of diversity of antigen binding sites, still should be able to initiate allergic responses. Although there is a positive correlation between the clonal size of allergen-specific IgE and the extent of the biological effect (i.e. basophil degranulation), a single pair of IgE antibodies, targeting non-overlapping epitopes , as is the case with M0418/B14 and B10/B13, is enough to achieve a strong degranulation of effector cells.
The four IgE-derived antibody fragments included in this study were, as all other allergen-specific human monoclonal IgE available to date, selected from combinatorial antibody libraries . The usage of such antibody fragments has been successfully employed in several previous studies of human IgE [9, 10, 17, 18, 40, 57] and of responses involving other human isotypes and relies on the fact that the random pairing of sequences encoding the VH domain of human IgE with random sequences encoding VL domain can recreate VH-VL pairs with antigen specificity. As a large portion of the specificity of antibodies resides within the VH , especially in the CDRH3 , it should be considered likely that naturally occurring specificities are largely retained among selected antigen-specific binders. One should, nevertheless, be aware of the potential risk of creating in particular slightly modified specificities that do not match those found in vivo through the random pairing of VH and VL. The fact that several similar scFv with quite different VL were isolated in this study argues in favour of the presence of specificity determining VH domains that have originally been selected in vivo.
Access to human monoclonal IgE allows for extensive characterization of such antibodies at a much more detailed resolution than that afforded by studies of polyclonal serum IgE preparations. As human and mouse antibody responses, and human IgE and IgG responses have been shown to differ in many respects [18, 59], availability of human monoclonal IgE offers new opportunities to more faithfully study antibody–allergen interactions critical to allergic disease. Only in a few instances [46, 47] has it been possible to study human IgE–allergen interactions using X-ray crystallography. In this study, we were able to obtain a first high-resolution structure of a human allergen-specific IgE fragment in the scFv format, a format that may be more readily used in the selection of specific binders as compared to Fab fragments used in the past to determine human IgE–allergen interactions. Unfortunately, despite efforts using both intact allergen and allergen-derived peptides, it was not possible to obtain crystals of a complex, attributed to the involvement of CDRH3 in maintaining crystal contacts. Nevertheless, the structure of the unbound IgE scFv M0418 demonstrates that a human allergen-specific antigen binding site may display a protruding CDRH3, similar to human antibodies of other isotypes [60-62].
Even in the absence of allergen-IgE scFv structures, the availability of human IgE-derived antibody fragments allows for very detailed analyses of human IgE–allergen interactions at a level not easily obtained with polyclonal antibodies. In this study, we have defined two human IgE epitopes on Bet v 1 and we have been able to define their very different cross-reactive potential in terms of recognition of Bet v 1 isoforms and other proteins of the PR-10 family. Importantly, it has also been possible to connect genetic studies of IgE-repertoire composition and its involvement in responses against an important allergen family. We foresee that human monoclonal IgE as they become more readily available will continue to prove their extensive value in defining allergy-causing immune responses and in providing new solutions (e.g. new hypoallergenic variants as described in the past ) that may be used to treat allergic disease in the clinic.
In summary, we have in this study isolated four novel Bet v 1-specific antibody fragment with origin in human IgE repertoires and defined two peptides, each containing the epitope of one of the two clonally related IgE pairs. We have identified certain allergen residues that are crucial for the interaction with one of these human IgE-derived antibody fragments, providing an explanation, at a molecular level, for the inability of this allergen-specific IgE fragment to recognize certain allergen isoforms and to cross-react with many members of the PR-10 protein family. We also show that IgE repertoires restricted in their genetic composition, in this case to the IGHV5-51 germline gene, a major member of a gene subgroup reported [4-6] to be overrepresented in some IgE responses, still retains the ability to generate IgE capable of receptor cross-linkage. We foresee that the strategy used in this and other studies making use of human monoclonal allergen-specific IgE will help us decipher unanswered questions concerning the crucial interactions between IgE and the allergen at the onset of allergic reactions.