The Rise and Fall of IgE

by Molllllly Vernersson                     

      !"

#$ %&''$ (")*")&*

++ *!(")!+! ,+'!"*! ,,+'+ -- Dissertation for the Degree of Doctor of Philosophy in Molecular Immunology presented at Uppsala University in 2002

Abstract

Vernersson, M., 2002. The rise and fall of IgE. Acta Universitatis Upsaliensis. Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology. 742. 61 pp. Uppsala. ISBN 91-554-5388-0.

Immunoglobulin E (IgE) occurs exclusively in mammals and is one of five immunoglobulin (Ig) classes found in man. Unlike other isotypes, IgE is best known for its pathological effects, whereas its physiological role remains somewhat elusive. To trace the emergence of IgE and other post-switch isotypes we have studied Ig expression in two monotreme species, the duck-billed platypus (Ornithorhynchus anatinus) and the short-beaked echidna (Tachyglossus aculeatus), leading to the cloning of IgE, two IgG isotypes in platypus and echidna IgE. The presence of IgE and the conservation of the overall structure in all extant mammalian lineages indicates an early appearance in mammalian evolution and a selective advantage of structural maintenance. Furthermore, both of the two highly divergent platypus γ chains have three constant domains. Hence, the major evolutionary changes that gave rise to the IgE and IgG isotypes of present day mammals occurred before the separation of monotremes from the marsupial and placental lineages, estimated to have occurred 150-170 million years ago. As the central mediator in atopic allergy, IgE is a prime target in the development of preventive treatments. This thesis describes an active immunization strategy that has the potential to reduce IgE to a clinically significant extent. The active vaccine component is a chimeric IgE molecule, Cε2-Cε3-Cε4. The receptor-binding target domain, Cε3, is derived from the recipient species, whereas the flanking domains, acting both as structural support and to break T-cell tolerance, are derived from an evolutionarily distant mammal. Vaccination of ovalbumin-sensitized rats resulted in a substantial reduction in total IgE in three out of four strains, accompanied by a significant reduction in skin-reactivity upon allergen challenge. No cross-linking activity was observed and the response to vaccination was reversible with time. The apparent safety and efficacy of the vaccine suggest that active immunization against IgE has the potential to become a therapeutic method for humans. Furthermore, the cloning and expression of the pig (Sus scrufa) ε chain will facilitate the development of sensitive and specific assays for pig IgE, thus increasing the possibilities of using the pig model in future studies of IgE-mediated reactions.

Key words: Immunoglobulin, IgE, evolution, atopic allergy, vaccine

Molly Vernersson, Department of Cell and Molecular Biology, Immunology Program, Uppsala University, Biomedical Center, Box 596, SE-751 24 Uppsala, Sweden.

© Molly Vernersson 2002 ISSN 1104-232X ISBN 91-554-5388-0 Printed in Sweden by Eklundshofs Grafiska, Uppsala 2002

2 Problems worthy of attack prove their worth by hitting back

Grooks by Piet Hein

3 4 To my Family

5 List of Publications

This thesis is based on the following publications and manuscripts, which are referred to in the text by their roman numerals.

I . Vernersson M, Aveskogh M, Munday B and Hellman L. Evidence for an early appearance of modern post-switch immunoglobulin isotypes in mammalian evolution (II); cloning of IgE, IgG1 and IgG2 from a monotreme, the duck-billed platypus, Ornithorhynchus anatinus. Eur J Immunol. 2002; 32:2145-2155.

II. Vernersson M, Aveskogh M and Hellman L. Cloning of IgE from the echidna (Tachyglossus aculeatus) and a comparative analysis of ε chains from all three extant mammalian lineages. Manuscript. 2002.

III. Vernersson M, Pejler G, Kristersson T, Alving K and Hellman L. Cloning, structural analysis, and expression of the pig IgE ε chain. Immunogenetics. 1997; 46:461-468.

IV. Vernersson M, Ledin A, Johansson J and Hellman L. Generation of Therapeutic Antibody Responses against IgE through Vaccination. FASEB J. 2002; 16:875-877. FASEB J. express article 10.1096/fj.01-0879fje, 2002.

Reprints were made with permission from the publishers. © Springer-Verlag GmbH & Co. KG © WILEY-VCH Verlag GmbH 2002

6 Contents Abstract ...... 2 List of Publications...... 6 Contents...... 7 Abbreviations and definitions ...... 8 Introduction ...... 9 Immunoglobulins...... 10 Immunoglobulin features ...... 10 Immunoglobulins in evolution...... 11 What about Mammals ...... 13 The rise of IgE...... 14 IgE features...... 15 Switching to IgE...... 16 The IgE Network...... 16 FcεRI...... 17 FcεRII/CD23...... 17 Regulation of and by IgE ...... 18 Why do we have IgE?...... 19 IgE and parasitic defense...... 19 Other functions? ...... 20 IgE deficiency ...... 21 Atopic allergy ...... 21 The allergic response...... 22 Why are allergies increasing? ...... 25 Current therapeutic strategies...... 26 Hyposensitization...... 27 Allergen independent treatments...... 28 Passive immunization against IgE ...... 28 Active Immunization against IgE...... 29 Present study...... 32 Aim ...... 32 Results and discussion...... 32 Cloning and phylogenetic analysis of monotreme ε- and γ chains (I and II)...... 32 Comparative analysis of mammalian ε chains (Paper II)...... 34 Cloning, structural analysis and expression of Pig IgE (Paper III) ...... 37 Active immunization against IgE (paper IV) ...... 39 Loose Ends ...... 45 Concluding remarks...... 47 Acknowledgements...... 49 References...... 51

7 Abbreviations and Definitions

APC antigen presenting cell BCR B cell receptor C constant CD23 the low affinity receptor for IgE CDR complementarity-determining region CH1, -2, -3 and -4 constant heavy chain domain 1, -2, -3 and -4, respectively. Cε1, -2, -3 and -4 constant ε chain domain 1, -2, -3 and -4, respectively. ε chain the heavy chain of IgE ELISA enzyme-linked immunosorbent assay eutherians placental mammals FcεRI the high affinity receptor for IgE FcεRII the low affinity receptor for IgE, also called CD23 FcεRIα the IgE binding alpha chain of the high affinity receptor γ chain the heavy chain of IgG GST-C2C3 a fusion protein consisting of the Cε2-Cε3 region of rat IgE coupled to glutathione-S-transferase of S. japonicum H heavy IFN interferon Ig immunoglobulin IL interleukin κ kappa (light chain isotype) λ lambda (light chain isotype) L light MAb monoclonal antibody metatherians marsupial mammals MHC major histocompatibility complex mIg membrane bound immunoglobulin monotremes a small group of egg-laying mammals OMO recombinant protein: OpossumCε2-MouseCε3-OpossumCε4 OOO recombinant protein: OpossumCε2-Cε3-Cε4 ORO recombinant protein: OpossumCε2-RatCε3-OpossumCε4 OROO recombinant protein: OpossumCε2-Rat/OpossumCε3- OpossumCε4 Ova ovalbumin PCR polymerase chain reaction prototherians monotremes TCR T cell receptor Th cell T helper cell therian mammals placental and marsupial mammals Vvariable

8 Introduction

We are constantly exposed to a great variety of potentially pathogenic microbes, such as bacteria, viruses, protozoa and multicellular parasites. Survival in this rather hostile environment requires effective and versatile means of protection. The first lines of defense are physical and chemical barriers, such as the skin and lysozyme in saliva, which passively, but effectively, prevent entry of unwanted intruders. However, if these barriers fail, the cells and molecules that constitute our immune system take a more offensive course of action to destroy the intruder.

Our immune system consists of a wide range of distinct cell types and molecules, each with important roles. Some components are present and armed before the infection, like phagocytic cells and complement proteins in blood, which recognize and respond to non-self molecular patterns commonly displayed by microbes. These cells and molecules are responsible for innate immunity, characterized by rapid responses that are essentially identical from infection to infection with a given microbe.

Other components develop after the infection and adapt to it; namely, the lymphocytes and their products, which constitute adaptive immunity. The hallmarks of adaptive immunity are specific recognition of a virtually infinite variety of distinct non-self structures, antigens, and the ability to “remember” and respond more vigorously upon repeated exposure to any given microbe. The lymphocytes fall into two major populations, B cells and T cells, which specifically recognize antigens through their membrane-bound antigen receptors, the B cell receptor (BCR) and the T cell receptor (TCR). The immunoglobulins (Igs) are the soluble forms of the BCR and they operate as adaptors in the immune system by linking the specific recognition of microbes to the cells and molecules that ultimately destroy the intruder, like the phagocytic cells and complement proteins of innate immunity. The various components of the immune system work in concert to eliminate the intruder using the cells and molecules that are most efficient against the microbe in question.

However, sometimes the immune system reacts unnecessarily vigorously against otherwise harmless environmental agents, allergens. These overreactions are termed hypersensitivity reactions and fall into four subgroups, types I – IV, depending on the mechanism behind the reaction. The type I hypersensitivity reactions are mediated by allergen-specific IgE and are responsible for the clinical manifestations of atopic allergies, such as hay fever, food allergies, eczema, asthma and anaphylaxis. More than 20% of the general population suffers from allergic diseases in one form or another and the alarming increase in their prevalence over the past decades have placed them among the most common human diseases in developing countries.

9 Although IgE is best known for its detrimental effects, causing allergic disease can hardly be the primary function of IgE. When did IgE appear as a separate isotype, and why? What is the selective advantage of having IgE? IgE may have evolved as a protective mechanism against certain parasitic infections, but its overall contribution to host defense remains a subject of controversy. Do we still need IgE? Is it possible to vaccinate against IgE? This thesis features IgE, focusing on some of the unresolved questions, opening with an evolutionary perspective and ending with the development of an anti-IgE vaccine as a potential therapeutic strategy against atopic allergy.

Immunoglobulins Immunoglobulins, or antibodies, are heterodimeric glycoproteins produced by B cells. They share several characteristics with the TCR, including antigen recognition and various mechanisms that generate diversity, as well as structural features. However, immunoglobulins recognize native antigens, such as proteins, polysaccarides and lipids, whereas the TCR recognizes peptide fragments presented together with the major histocompatibility complex (MHC) class I and II molecules on the surface of infected and antigen-presenting cells (APC). Furthermore, in contrast to the TCR, immunoglobulins occur in a bifunctional secreted form. In addition to antigen binding, secreted immunoglobulins exhibit one or more effector functions accounted for by distinct structures spatially separated from the antigen-binding sites.

Immunoglobulin features Immunoglobulins are usually composed of two identical heavy (H) chains and two identical light (L) chains held together by disulfide bonds and stable non-covalent interactions. Both H chains and L chains consist of a variable (V) region responsible for antigen binding and a constant (C) region which determines the isotype. Mammals express five major immunoglobulin classes; IgM, IgD, IgA, IgG, and IgE, which are defined by their H-chain C-region genes; µ, δ, α, γ, and ε. The H-chain C-region genes occur in a tandem array in the H-chain locus and share V-region genes. The L chain also occurs in two different isotypes, kappa (κ) and lambda (λ), but they have separate loci and, accordingly, combine with separate sets of V-region genes.

Both L chains and H chains are folded into repeated globular motifs, immunoglobulin domains, each consisting of about 110 amino acids and composed of two layers of anti- parallel β-pleated sheets that are joined together by a disulphide bond. The L-chain isotypes are composed of two such Ig domains, one V domain and one C domain. The mammalian H-chain isotypes are composed of a V domain and, depending on the isotype, three (α and γ) or four (µ and ε) C domains.

In germline configuration, the genes encoding the V regions are composed of multiple copies of variable (V), diversity (D) and joining (J) segments (H-chain) or V and J segments (L-chain). The enormous V region variability is, in part, generated by the

10 random rearrangement of these segments into a functional V region by a genetic somatic event called V(D)J recombination. The H chain V region (VH) and L chain V region (VL) are juxtaposed in the protein and together they form the antigen-binding site. The numerous potential VH and VL combinations generate an amazing diversity of the Ig repertoire, which enables recognition of virtually all foreign and potentially pathogenic structures.

Initially, the VDJ exon encoding the VH is spliced to the Cµ exon, forming a µ chain, which subsequently assembles with a successfully rearranged L chain. Hence, IgM is the initial isotype expressed during B-cell development. The antigen specificity of a given B-cell clone is constant: that is, the Ig V regions are retained. However, through a mechanism called isotype switching the constant region, Cµ, can be replaced by Cα, C γ or Cε, resulting in expression of IgA, IgG or IgE isotypes that have the same antigen specificity as the original IgM.

Together, the immunoglobulins mediate the effector functions of humoral immunity, the principal defense against extracellular microbes. However, each isotype engages distinct effector mechanisms for disposing of the antigen, including interaction with isotype specific receptors on various immune cells, activation of complement and transfer across the placenta or through epithelial layers. In short, mammalian IgM serves as the predominant BCR and it is the principal antibody of primary immune responses. IgA is the primary isotype in mucosal immunity, providing a frontline defense at epithelial surfaces. IgG is the predominant serum antibody and the role of mast cell activation, and thus of mediating anaphylactic reactions, falls mainly on IgE.

Immunoglobulins in Evolution The immunoglobulins and all other components of adaptive immunity (TCR, MHC and recombination enzymes) seem to have emerged at the appearance of the jawed vertebrates, gnatosomes [1-3]. Since these components appear to be absent in invertebrates as well as agnathans, the adaptive immune system has been proposed to have appeared in a relatively short period of evolutionary time, possibly in a single step catalyzed by horizontal transfer of recombination genes from prokaryotes into the vertebrate genome. This putative event has been described as a “Big Bang” in molecular evolution [2, 3]. Since the time of this putative “Big Bang”, the complexity of the adaptive immune system has increased gradually throughout vertebrate evolution (Fig. 1). IgM, the only isotype that occurs in all investigated gnatosome vertebrates, appears to be the first and most ancient isotype. Accordingly, IgM is the major serum isotype found in elasmobranches (sharks and rays) and bony fish [4]. Fish B cells apparently do not undergo isotype switching [4]. However, depending on the species, an IgD-like Ig, IgNAR, IgNARC, IgW or IgR may also be expressed, either through differential splicing (the IgD-like Ig) or by their presence on loci separate from that of IgM, reviewed in [4].

11 IgM Gnatosomate ancestor

IgD IgM

Bony fish

IgM IgY IgX ?

Amphibians

IgMIgY IgA ?

Birds

IgM IgD IgE IgG IgA

Mammals

Figure 1. Scheme for the evolutionary development of immunoglobulins. Putative gene duplication events that gave rise to new isotypes are indicated by dashed arrows. So far, there is no evidence of IgD in amphibians or in birds, but the number of investigated species is limited.

12 Amphibians (Xenopus) express two non-IgM post-switch isotypes, IgY and IgX [5]. The IgY isotype has a wide distribution in non-mammalian vertebrates, including amphibians; axolotl [6] and Xenopus [7], reptiles; lizard [8] and turtle [9] and birds; chicken [10] and duck [11]. In contrast, the IgX isotype appears to be restricted to Xenopus, but the number of species investigated so far is limited. The two Xenopus post-switch isotypes may have originated from independent duplications of the Cµ gene, and based on membrane exon analysis, IgY appeared before IgX [12]. The latter hypothesis is further supported by the presence of IgY, but not IgX, in the more primitive axolotl [6] and indications of an IgY-like isotype in lungfish [6, 13, 14]. Hence, IgY is believed to be the first post-switch isotype to appear in evolution and the mechanism of isotype switching may have emerged with the tetrapods, or possibly in the putative ancestors of tetrapods, the lungfish or the coelacanth (Latimeria) [6, 12].

Although Xenopus IgX appears to have a distribution and effector functions similar to those of the IgA isotype expressed by birds and mammals, it is probably not the direct homologue of mammalian IgA [15]. Hence, in addition to IgM and IgY, birds express IgA [16]. Like IgM and IgY, avian IgA contains four constant domains, whereas mammalian IgA has only three. Sequence comparisons suggest a common progenitor Cα gene for avian (chicken) and mammalian IgA and that the hinge region found in mammalian IgA has replaced the Cα2 domain in the putatively more primitive avian IgA [16].

What about mammals? Most studies of immunoglobulin expression in mammals have focused on placental (eutherian) mammals, and all investigated species express IgM, IgA, IgG and IgE. So far, IgD has been found only in rodents and primates. Some Ig classes; namely, IgG and IgA, are often expressed from more than one H-chain gene; that is, several subclasses of IgG and IgA occur in many species. For example, humans express four IgG isotypes (IgG1 through IgG4) and two IgA isotypes (IgA1 and IgA2). Only one IgG isotype occurs in rabbits [17], but at least ten out of thirteen different Cα genes are expressed in vivo [18, 19]. Conversely, only one Cα gene occurs in the pig [20], whereas eight C γ genes have been detected in genomic Southern blots and at least five of these are expressed [21].

Apart from the eutherians, there are two other extant mammalian lineages, the marsupials (metatheria) and the monotremes (prototheria). The evolutionary separation of the marsupials from the placental mammals is generally dated back to approximately 130 to 175 million years ago [22-24], whereas the major radiation of the placental mammals probably occurred 60 to 120 million years ago [24].

Immunoglobulin expression has only been studied in a few of the approximately 270 existing marsupial species. The cloning of IgG, IgE and IgA from the gray short-tailed opossum (Monodelphis domestica) revealed that all of the eutherian post-switch

13 isotypes occur also in marsupials [25]. Furthermore, as in placental mammals, multiple subclasses of IgG are reportedly present in several marsupial species, as reviewed in [26].

The monotremes, which are represented by only three species; namely, the duck-billed platypus and two echidna species, have been regarded as "reptile-like" or "primitive" mammals simply because they lay eggs. However, they do possess several of the major mammalian features. Histological studies show that the spleen, thymus and gut- associated lymphoid tissues in both the platypus and the echidna are well developed and comparable in histological structure to those in therian mammals (eutherians and marsupials) [27, 28]. However, in sites where lymph nodes would be expected in therian mammals, monotremes were found to have lymphoid nodules that resemble the jugular bodies of the amphibians, which indicates that monotremes have a somewhat more primitive immune system [27, 28]. Early studies of immunoglobulin expression in the echidna established that the major immunoglobulin classes were very similar to human IgM and IgG [29]. This has been confirmed recently by the cloning of echidna IgM and IgG1 [22, 30].

The evolutionary relationship between marsupials and monotremes is not fully resolved. The Marsupionata hypothesis, supported mainly by mitochondrial data, suggests that marsupials and monotremes are sister lineages [31-34]. In opposition, the Theria hypothesis proposes that marsupials and eutherians are sister lineages and that the monotremes constitute the most divergent lineage. The Theria hypothesis initially emerged from morphological data and has more recently gained additional support from paleontological data [35, 36] and a number of nuclear gene sequences [22, 23, 37-40]. Studies based on IgM and protamine sequences, both of which support the Theria hypothesis, estimate the time of divergence of the monotremes from the other mammalian lineages at around 150 to170 million years ago [22, 23].

The Rise of IgE So far, IgE as well as IgG have been found exclusively in mammals. However, the IgY isotype found in amphibians, reptiles and birds (see above) shares characteristics with both IgE and IgG, and based on several observations, a common phylogenic history for the IgY isotypes and mammalian IgE and IgG has been suggested (Fig. 1) [6, 12, 41, 42].

In structure, IgY is more similar to IgE. Both IgY and IgE have four constant domains, whereas the IgG isotype has only three. Sequence comparisons among the different domains of IgY and IgG suggest that the equivalent of the Cυ2 domain is absent in mammalian IgG [41, 42]. However, the Cυ2 domain shares features with the mammalian hinge, suggesting that a truncated Cυ2 domain formed the hinge in what evolved into mammalian IgG [41, 42]. Two additional cysteine residues, putatively involved in the formation of a second intra-domain disulfide-bridge, are present in the

14 Cυ1 domain of IgY. These additional cysteine residues also occur in most of the characterized ε chains, excluding those of possum and rodents, as well as in some γ chains, including those of rabbit and pig [17, 21, 42].

In analogy with IgG in mammals, IgY is the major serum antibody in birds and amphibians, and hence is the major isotype involved in the defense against systemic infections. However, IgY can also mediate anaphylactic reactions [43], which usually, but not exclusively, is ascribed to IgE in mammals. In some mammalian species anaphylactic reactions can be mediated by IgG, e.g., in the guinea pig [44] and to a lesser extent in mice [45]. Nevertheless, it may have been of selective advantage to separate the two functions of IgY into two separate isotypes, IgG and IgE, and thereby restrict potentially dangerous reactions to an isotype that normally occurs in low concentrations.

IgE Features (or what makes IgE so special?) IgE is the least abundant of the human immunoglobulin classes and was accordingly the last to be discovered in the late 1960s [46-48]. The concentration of IgE in normal human sera is between 10 and 400 ng/ml, whereas IgG is present at concentrations that are thirty thousand times higher (9.5-12.5 mg/ml) [49]. Furthermore, at least in humans, the turnover rate of IgE is much more rapid than that of IgG. The half-life of IgE in the circulation is estimated at 2-2.5 days, whereas serum IgG has a half-life of about 3 weeks [49].

As previously mentioned, mammalian species often express multiple subclasses of the other post-switch isotypes, IgG and IgA. In contrast, to my knowledge, only one functional ε gene occurs in all investigated species. However, several alternatively spliced ε chain mRNA transcripts have been identified in humans [50-57] and a few, non-homologous to the human isoforms, in mice [58]. A few of the human splice variants seem to be translated and give rise to stable protein products in vivo [53], including the so-called tailpiece isoform, IgEtp [56, 59, 60]. In IgEtp, the two carboxy- teminal amino acids of the classical secreted IgE are replaced by eight novel amino acids, including a carboxy-terminal cysteine residue. In a previous study, Diaz-Sanchez and colleagues reported a positive correlation between the amount of IgEtp mRNA and high total IgE levels, as observed in atopy or parasitic infections [59]. However, recent studies by the same group did not enable assignment of any unique characteristics to IgEtp and no correlation between serum IgE levels and IgEtp expression was observed [60]. Hence, the contribution of the splice variants to the characteristics of human IgE appears to be limited. Furthermore, most of the alternatively spliced transcripts appear to be by-products of the splicing machinery, encoding non-functional protein products that are retained within the cells and degraded [55].

15 No reports of alternative splicing of IgE are available for any mammalian species other than humans and mice and efforts to identify splice variants of rat IgE were fruitless [61]. Hence, in most mammalian species, IgE probably occur in the two classical forms only, that is, the secreted and the membrane bound form of IgE.

Finally, IgE is shaped like no other isotype. Unlike IgM and IgA, the secreted form of IgE is monomeric. Unlike IgA and IgG, the other post-switch isotypes, IgE has no hinge and is therefore less flexible. However, unlike IgM, the other isotype with four constant domains, IgE adopts a bent structure in solution [62].

Switching to IgE The production of IgE antibodies is initiated when an antigen is recognized by the mIg on the B cell surface. However, to further B-cell proliferation and differentiation into IgE-secreting plasma cells, two additional signals are required [63]. One is the CD40/CD40L interaction and the other is interleukin-4 (IL-4) and/or IL-13, which direct recombination to the Cε gene, resulting in ε-chain transcription.

A subset of T helper (Th) cells, Th2 cells, can supply both of these signals [63]. Hence, the antigen is internalized and processed into peptide that become associated with MHC class II molecules and transported to the B-cell surface. Specific T cells that carry TCRs with antigenic epitopes recognize and bind to the processed antigen. This recognition results in a transient expression of CD40L on the T cell. CD40L/CD40 interaction upregulates B7-1/B7-2 (CD80/CD86) expression by the B cell and stimulates B-cell proliferation. Binding of B7 to CD28 on the T cell activates the T cell, which starts to produce IL-4 and other cytokines and thus triggers the differentiation of B cells into IgE- secreting plasma cells. However, this is a very simplified scenario. The Th2 subset may well be the main source of IL-4, but these cells also require IL-4 to become polarized into Th2 cells, a hen and egg situation. What is the original source of IL-4? Mast cells, basophils and eosinophils, the effector cells of IgE, all produce type 2 cytokines, including IL-4, as reviewed in [64]. In addition, both mast cells and basophils express the CD40L and can induce IgE synthesis [65]. Nevertheless, IgE synthesis involves Th2-type cytokines and IgE responses against a given antigen are regulated predominantly by Th2-polarized T cells [66]. In contrast, interferon-γ (IFNγ) produced by Th1 polarized cells inhibit both Th2 responses and IgE synthesis.

The IgE Network Although the IgE isotype constitutes only a minuscule fraction of the total antibody content in human sera, its actions are potent due to interaction with, and the actions of, two structurally unrelated receptors: the high affinity receptor, FcεRI, and the low affinity receptor, FcεRII, also known as CD23. FcεRI engages IgE with an 10 -1 exceptionally high affinity, Ka~10 M , which is 1000-fold higher than that of CD23.

16 FcεRI FcεRI belongs to the immunoglobulin superfamily and is highly homologous to the Fcγ receptors. It is expressed primarily by mast cells in tissues and basophils in blood, on which it has a tetrameric structure composed of four transmembrane peptides; an α- chain, a β-chain and two disulfide linked γ-chains (αβγ2) [67]. The α-chain contains two extracellular Ig domains that are responsible for IgE binding, whereas the β- and γ- chains are involved in receptor assembly and signal transduction [67, 68]. High “constitutive” expression of FcεRI is restricted to mast cells and basophils [69]. However, in humans, low levels of FcεRI expression have been detected also on other cell types, including eosinophils, neutrophils and platelets, and in a in putative trimeric αγ2 structure also on monocytes, Langerhans cells and dendritic cells, as reviewed in [68]. The functional significance of FcεRI on some of these cell types is uncertain. For instance, the level of FcεRI expression on eosinophils is very low, only about 0.5% of that on basophils, and apparently IgE does not mediate eosinophil effector functions [70]. Considering the central role of the β-chain in receptor activation, the importance of the putative trimeric form (αγ2) is also questionable [68]. Nevertheless, its presence on professional APCs is suggestive of a role in antigen uptake and presentation [71]. In rodents, FcεRI is expressed exclusively on mast cells and basophils and has an obligatory αβγ2 structure [68].

The crystal structure of the human IgE-Fc fragment bound to FcεRIα has been solved and it reveals that one dimeric IgE-Fc molecule binds asymmetrically to two separate sites on one FcεRIα, each site involving one Cε3 domain [72]. Twelve amino acids from one of the Cε3 domains are involved in the interaction with receptor site 1, whereas ten amino acids from the other Cε3 domain interact with site 2. The residues in the respective Cε3 domains that are involved in binding to site 1 and site 2 overlap, but are not identical. Hence, explaining the 1:1 stoichiometry of the interaction [72]. Furthermore, binding to FcεRI appears to induce a conformal change in the Fc region of IgE, indicating a conformal flexibility of this region and suggesting that the high affinity of the IgE/FcεRI interaction might be achieved by an induced fit [73, 74]. In addition, although binding interactions occur exclusively between the Cε3 domains of the IgE and FcεRIα, the conformal change also involves the Cε2 domain and its presence markedly reduces the dissociation rate of the IgE/FcεRIα complex [74, 75]. The dissociation rate of the IgE/FcεRI complex is extraordinary slow, about two weeks [76]. In comparison, the off rate of the interaction between IgG and its high affinity receptor is only a few minutes [76]. In addition, the Cε4 domain appears to be important for dimerization and proper folding of the Cε3 domain [77-79].

FcεRII/CD23 In contrast to most immunoglobulin receptors, CD23 does not belong to the immunoglobulin superfamily, but is member of the C-type (calcium-dependent) lectin superfamily [67]. The CD23 molecule consists of an extracellular C-terminal lectin domain, a stalk region, a transmembrane region and a cytoplasmic N-terminus, which

17 self-associate and form trimers by coiling around each other at the stalk region [67]. The lectin domain comprises the binding site for IgE and the interaction is thought to involve two lectin “heads” [67, 80]. The binding site on human IgE for CD23 has been mapped to the Cε3 domain and appears to overlap with, but is distinct from, that of FcεRI [80, 81].

Differential splicing of CD23 results in two different isoforms; CD23a and CD23b, which differ only by a few amino acid residues in the cytoplasmic N-terminus [80]. Both isoforms are found on B cells: CD23a is constitutively expressed and appears to be restricted to B cells, whereas CD23b is induced in particular by IL-4 and is expressed on a variety of cells, including monocytes, macrophages, eosinophils, platelets, natural killer cells, T cells, follicular dendritic cells (FDC) and Langerhans cells, as reviewed in [67, 82, 83]. The CD23a isoform is associated with endocytosis of IgE-coated particles by B cells and IgE-dependent antigen presentation to T cells [84], whereas the CD23b isoform on macrophages, eosinophils and platelets mediates IgE-dependent cytotoxicity and promotes phagocytosis of soluble IgE complexes [82].

To complicate things further, proteolytic cleavage in the stalk region of CD23 generates a series of soluble fragments, some of which appear to have cytokine-like properties [80, 84]. In addition to IgE, CD23 also interacts with CD21 on B cells and FDCs, as well as other members of the integrin family, suggesting a role for CD23 in IgE- independent inflammatory mechanisms [85].

Regulation of and by IgE IL-4 is the spider in the IgE network. It promotes the development of Th2-type responses, enhances FcεRI expression, induces the expression of CD23 and most importantly, together with IL-13, it directs the switch towards IgE synthesis [67, 68]. However, once induced, IgE appears to regulate its own responses via interactions with its receptors.

To begin with, IgE can modulate the surface expression of both FcεRI and CD23. Several studies, both in vitro and in vivo, have demonstrated that IgE enhances FcεRI expression on mast cells and basophils, both in humans [86-88] and in mice [89, 90]. The FcεRI-expression level in peritoneal mast cells from IgE-deficient (IgE-/-) mice is fourfold to fivefold lower than that in wild-type mice [89]. Similarly, B cells derived from IgE-/- mice express fourfold to sixfold less CD23 than those of wild-type animals, as reviewed in [91]. Accordingly, intravenous administration of IgE enhances surface expression of both receptors in the IgE-/- mice [89, 91].

The enhanced FcεRI-surface expression is mediated by IgE/FcεRI interaction and appears to be relatively insensitive to protein synthesis inhibitors [88, 89]. This suggests that IgE acts by slowing depletion, thus prolonging the survival of preformed FcεRI on the cell surface, rather than by inducing FcεRI synthesis [88, 89].

18 Furthermore, IgE-mediated upregulation of FcεRI on mast cells appears to enhance mast-cell-effector functions by increasing both the sensitivity to antigen challenge and the intensity of the response [86, 89]. In addition, binding of monomeric IgE to FcεRI, even in the absence of antigen, promotes mast cell survival by preventing growth- factor-deprivation-induced apoptosis [92, 93].

In contrast, CD23 appears to be the most important receptor affecting IgE production. Interaction with CD23 may either inhibit or promote IgE responses, depending on the expressing cell type, the expression level and the form of CD23. Binding of IgE/antigen complexes to CD23 on IL-4 stimulated B cells, which express high levels of CD23, inhibits IgE production [94, 95]. Under other conditions, possibly involving lower degrees of receptor cross-linking, IgE/antigen complexes enhance antigen-specific IgE responses via CD23 engagement on B cells, reviewed in [95]. Furthermore, IgE binding to membrane-bound CD23 inhibits proteolytic cleavage and the release of soluble CD23 products, some of which promote IgE synthesis (larger trimeric fragments) possibly via CD21, whereas others appear to inhibit IgE production [96].

In conclusion, by engaging FcεRI IgE enhances its own effects, whereas by engaging CD23 IgE can keep the responses under control.

Why do we have IgE? The mere existence of a complex and powerful biological system such as the IgE network implies a strong selective pressure to establish and maintain it. However, IgE is best known for its pathological effects in allergic diseases and the beneficial role of IgE in host defense remains somewhat uncertain. Nevertheless, IgE and its receptors are believed to have evolved as a defense mechanism against parasitic infections, in particular against parasitic worms, i.e., helminth infections [67].

IgE and parasitic defense Immune responses against helminth infections, such as Schistosoma mansoni and S. haematobium, are characterized by elevated serum IgE concentrations and eosinophilia [97]. A protective role for IgE was demonstrated by passive transfer of parasite-specific IgE in a rat model of schistosomiasis [97]. Furthermore, several human seroepidemiological studies have shown that the presence of parasite-specific IgE correlates with protective immunity against shistosomiasis [98-100] and other helminth infections [101]. In human schistosomiasis, antibody-dependent, cell-mediated cytotoxicity involving eosinophils and IgE appears to be the main mechanism for killing parasite larvae [97]. Hence, eosinophils may have evolved as a defense against parasites (such as parasitic worms) that are too large for phagocytic cells. In this context, the principle function of the IgE-mediated mast cell reaction may be to direct eosinophils to the site of parasite infection and to enhance their anti-parasitic functions.

19 However, the importance of IgE in parasitic infections remains a controversial issue, mainly due to contrasting findings in murine models of S. mansoni. In one study, IgE- deficient mice developed two-fold greater S. mansoni worm burdens than did wild-type mice [102], whereas no difference in worm burden or egg load was observed in another, similar study [103]. In a third study, neither experimental increase of IgE production through IL-4 stimulation nor passive transfer of parasite specific IgE was associated with any significant changes in worm burden or egg load, indicating that IgE has no relevant function in primary murine S. mansoni infections [104]. Similarly, schistosome infections develop normally in FcεRI deficient mice, with similar parasite recovery and egg load as in wild-type mice, suggesting a limited role for the IgE/FcεRI signaling pathway in protection against S. mansoni infection [105].

Then again, whereas Th2-type responses and eosinophils appear to be essential components for parasite expulsion in humans, Th1-type responses and macrophages appear to play a more important role in mice, at least in schistosomiasis [97], suggesting that, in this context, IgE may be more important in humans than in mice.

Other functions? IgE appears to function as an enhancer of immune responses against antigens that are present at low concentrations. This has been demonstrated in vivo in a series of experiments in mice, as reviewed in [95]. Intravenous administration of small protein antigens together with antigen-specific IgE can induce antibody responses that are more than 1000-fold higher than those induced by antigen alone [95]. The response to IgE/antigen is strictly antigen specific, but not isotype specific, as it involves upregulation of IgM-, IgG1-, IgG2a- as well as IgE responses. The proposed mechanism behind IgE-mediated enhancement is endocytosis of IgE/antigen complexes via CD23 by B cells and presentation of antigen peptides to antigen-specific T cells [95]. A requirement for CD23 in IgE-mediated upregulation was demonstrated by a complete lack of enhancement in mice pretreated with anti-CD23 antibodies as well as in CD23 deficient mice [106, 107]. The ability of IgE to augment responses under suboptimal conditions is suggestive of a physiological function in early responses [95]. This, together with the comparatively high turnover rate of IgE in serum, raises questions regarding the positive value of IgE. Does IgE function as a “doorkeeper”, scanning the antigen repertoire of the environment and preparing the individual for potentially harmful pathogens? [108]. A rapid turnover may serve to adequately prepare the individual to deal with its present environment.

Furthermore, surgery and other forms of tissue injury, such as burns and heart attacks, characteristically evoke a transient rise in serum IgE levels [109-113]. Serum IgE levels appear to rise shortly after surgery, peak by day five and then decline again [110, 113]. In contrast, serum IgG as well IgA and IgM fall rapidly after surgical intervention, but by postoperative day 7 they have returned to initial levels [110, 113]. In addition to IgE, an early increase in serum IL-6, a cytokine involved in acute phase protein induction, is

20 also observed after surgical intervention [109]. Hence, the behavior of IgE in situations involving tissue injury led Szczeklik and colleagues to suggest a possible role for IgE in acute-phase responses [111].

Finally, individuals with high IgE levels, such as atopics, apparently have a mild haemostatic imbalance, which means that their blood takes longer than normal to clot [114, 115]. This imbalance is reflected by a slightly prolonged bleeding time, depressed platelet aggregation and a delayed generation of thrombin, the final enzyme in the coagulation cascade. These phenomena could prolong the generation of a clot in a critically obstructed coronary artery and thus protect individuals with high IgE levels from sudden cardiac arrest associated with coronary thrombosis (blocking of vessels that provide the heart muscle with blood) [114, 115].

IgE deficiency One way to increase our understanding of the physiological role of IgE is to study the clinical characteristics of IgE deficiency. One study linked IgE deficiency with an increased prevalence of additional Ig deficiencies, autoimmune disease and non-allergic reactive airway disease [116]. IgE deficiency has also been associated with recurrent or chronic sinopulmonary infection, indicating a role for IgE in mucosal immunity [117]. Lack of IgE could lead to impaired mucosal defense and thereby increase susceptibility to respiratory tract infections [117]. However, in studies such as these it is always difficult to establish cause and effect. Why do these individuals have low IgE? In contrast to the human situation, animal models, i.e., mice knock-out models, allow us to study the effect of a single deficiency. Mice with a targeted deletion of IgE, IgE-/- mice, show no apparent signs of disease, the animals do not die prematurely and lymphocyte development is normal [45, 95]. Hence, life is possible in the complete absence of IgE, at least for laboratory mice. More importantly, IgE deficiency may also occur in humans with an otherwise normal immunologic profile without evidence of illness [118].

Atopic Allergy That allergies are widespread and that they may cause serious problems is common knowledge for anybody who has arranged a birthday party for kids recently. Furthermore, all of us have friends or family members that are allergic to one thing or another. In fact, at least 20% of the Swedish population suffers from one or several of the clinical manifestations of atopic allergy, ranging from hay fever to asthma, eczema and, in the worst case, systemic anaphylactic reactions. Asthma is one of the most common chronic childhood diseases, at least in the industrialized part of the world, and thousands of people die from asthma annually. Furthermore, asthma is a major cause of hospitalization and school/work absenteeism. Although hay fever and eczema are not life-threatening conditions, the quality of life for, and the activity of, affected individuals

21 can be seriously reduced. Taken together, atopic diseases constitute one of the major problems of modern day medicine.

Atopic allergies, or hypersensitivity type 1 reactions, are caused by adverse IgE-antibody responses against otherwise harmless environmental agents, allergens. Curiously, our allergic or atopic friends are often allergic to the same things. For instance, birch pollen, cats, nuts and dust mites are common allergen sources. What makes these things more allergenic (better inducers of IgE responses) than others? In other words: What makes an allergen an allergen? Some physical characteristics appear to be more common among allergens than among other proteins. Most allergens are soluble proteins and many are exceptionally heat stable [119]. They are often glycosylated and several allergens have enzymatic activities, as reviewed in [119]. A neat example of the latter, is Der p 1, the major house-dust-mite allergen, which has proteolytic activity and has been shown to cleave CD23 and promote IgE production [119]. However, none of the above-mentioned characteristics are unique to allergens, and whether a protein is allergenic appears to have no single structural cause. Furthermore, although none of us can avoid exposure to birch pollen we are not all allergic. Hence, other factors in addition to the nature of the allergen predispose to allergy. Nevertheless, the detrimental role of IgE is well established and the mechanism behind the allergic response is fairly well understood.

The allergic response Allergic responses involve two separate phases, a sensitization phase and an effector phase (Fig. 2). During the sensitization phase allergen-specific IgE antibodies are produced and become localized to the effector cells of allergic responses, mast cells and basophils. The effector phase involves the activation of these cells, which results in the release of mediators and the recruitment of other inflammatory cells that cause the clinical symptoms of atopic allergy. Both phases begin with the presence of an allergen.

The sensitization phase An allergen may enter the body via inhalation, ingestion, injection or through the skin. APCs encounter the allergen at, or close to, epithelial surfaces (skin, airways, intestines) and different cell types function as APCs depending on the site of allergen encounter. Macrophages, activated B cells and dendritic cells act in tissues, such as the airways, whereas Langerhans' cells are the most important APCs in the skin [71]. The APCs then migrate to a draining lymph node where they present allergen-derived peptides to naïve T cells and induce T-cell activation. Depending on the local conditions, which include the cytokine environment, the type of antigen, the antigen dose and the type of APCs, Th cells polarize to a greater or lesser extent into either a Th1- or Th2-like phenotype. The phenotype is defined by the cytokine repertoire and highly polarized Th1 cells produce mainly IL-2 and IFNγ, whereas Th2 cells produce IL-3, IL-4, IL-5, IL-6, IL-9, IL-10 and IL-13. As previously mentioned, polarization to and activation of the Th2 phenotype promotes the development of IgE responses. The

22 key cytokine that directs naïve Th cells to become Th1 effector cells is IL-12 [71]. The loss of IL-12 responsiveness appears to be an important factor in the polarization to Th2 cells [64]. IL-4 inhibits the expression of an IL-12 receptor subunit, IL-12R β2, and thus directs the loss of IL-12 responsiveness and drives polarization to the Th2 phenotype [120, 121]. In contrast, IFNγ promotes IL-12 responsiveness and consequently, in the presence of IL-12, also commitment to the Th1 pathway. The main source of IL-12 is activated macrophages, but other APCs, such as dendritic cells, also produce IL-12 [71]. The capacity of dendritic cells to induce either Th1 or Th2 responses may be due to their IL-12 production at the time, which in turn may be influenced by the pathogen (allergen) [71].

The cytokine environment in the draining lymph node also influences the switching pattern of allergen-specific B cells. If IL-4 and IL-13 as well as allergen-specific Th2 cells are present, i.e., if T cell help is available, a substantial fraction of the allergen- specific B cells will switch to IgE production and mature into IgE-secreting plasma cells. The secreted allergen-specific IgE antibodies leave the lymph nodes via the blood stream and become localized to mast cells and basophils by binding to FcεRI expressed on the surface of the respective cell-types. In addition, there appears to be an ongoing local production of allergen-specific IgE by resident B cells in the nasal and lung mucosa of hay fever and asthmatic patients, respectively [122]. An individual that has allergen-specific IgE bound to his/her mast cells in the tissues (basophils in the blood) is sensitized to the allergen in question and subsequent encounters with the allergen initiate the effector phase.

The effector phase When a multivalent allergen associates with receptor-bound IgE on the surface of mast cells and basophils, it cross-links the receptor molecules on the cell membrane. This cross-linking initiates signal transduction events and causes an influx of Ca2+ ions into the target cell that triggers degranulation and the release of preformed chemical mediators, such as histamine, proteolytic enzymes and heparin. The preformed mediators give rise to an immediate response, usually within 20 minutes. In addition, the degranulating cell releases de novo synthesized lipid mediators (leukotrienes and prostaglandins) and cytokines (IL-5 and IL-8) that recruit other inflammatory cells (eosinophils, neutrophils and lymphocytes) to the reaction site. The influx of inflammatory cells causes the late phase of the allergic response, which occurs within 6- 24 hours after allergen exposure. Together, the released mediators are responsible for the physiological effects that characterize allergic reactions, such as increased vascular permeability, vasodilation, itching, edema, cell infiltration, smooth muscle contraction and hyper production of mucus.

23 (effector phase) of allergen induced, IgE-mediated responses, which lead to the clinical presentations of atopic allergy. Figure 2. MHCII APC The allergic response. TCR Th0 IL-4 Allergen Th2 Bcell CD40L BCR IL-13 IL-4 Sensitization phase CD40 Scheme of molecular and cellular events behind the development (sensitization phase) expression Bcell Epithelial barrier IL-5 Plasma cell IgE Basophil In tissues Mast cell In blood Fc ε RI Eosinophil Infiltration & Effector phase activation Allergen IL-5 Release of mediators Degranulation e.g. Histamine Clinical effect: Hay fever Eczema Asthma

24 Why are allergies increasing? Before trying to find answers to the above question one must first try to establish whether allergies really are increasing. According to several epidemiological studies, conducted in different parts of the world, the prevalence of allergic diseases has increased dramatically (twofold to fourfold) during recent decades, as reviewed in [123, 124]. However, one might argue that this only reflects that we are better at diagnosis today than we were 30 years ago. True or not, several studies indicate that allergic diseases have continued to increase even during the last decade, ruling out improved diagnosis as an important parameter. For instance, one survey of Australian schoolchildren recently reported an increase in diagnosed asthma (including non-atopic asthma) from 30,5% in 1992 to a staggering 38.6% in 1997 [125]. Similar increases in atopic diseases have been reported among schoolchildren in Aberdeen (Scotland) [126]. Between 1989 and 1994, the prevalence of diagnosed asthma increased from 10.2% to 19.6%, that of eczema from 12.0% to 17.7% and that of hay fever from 11.9 to 12.7% [126]. In conclusion, allergies really are increasing and they are reaching epidemical proportions, but why?

The tendency to develop atopic diseases appears to be due to both genetic and environmental factors. Atopic diseases often cluster within families and seem to have a strong, but heterogeneous genetic component. More than 20 genes appear to be involved and polymorphisms in several candidate genes, including the IL-4 promoter, the IL-4 receptor and the FcεRI β-chain, have been associated with asthma and atopic disease, as reviewed in [127, 128]. Even so, genetic changes alone cannot account for a fourfold increase in one generation. However, a genetic predisposition in combination with certain so-called environmental factors that have emerged or become more pronounced during recent decades might well suffice.

Several studies have linked the rise in atopy with increased living standards, a “westernized” lifestyle and a reduced frequency of childhood infections, as reviewed in [123, 124, 129]. For instance, studies conducted in 1991-92, roughly one year after the unification of Germany, revealed that airway hyper-responsiveness and atopic sensitization were less prevalent among children living in Leipzig (former East Germany) than among children living in Munich (former West Germany) [130]. In 1995-96, after five years of integration with Western society, the prevalence of atopic sensitization in schoolchildren living in Leipzig had increased from 19.3% (in 1991-92) to 26.7% [130]. In 1991-92, Leipzig was far more polluted than Munich, indicating that although air pollution is known to aggravate already existing asthma, pollution in general cannot be responsible for the increased prevalence of asthma and other atopic diseases [131]. Nevertheless, certain pollutants; namely, diesel exhaust particles (DEP), appear to induce allergic sensitization and thus may, in part, explain why allergic diseases often are more common in cities than in rural areas within a given country, e.g., in Sweden, as reviewed in [129, 132]. Moreover, several epidemiological studies

25 have shown an inverse association between the risk of atopy and family size, birth order and nursery school attendance, all factors that will increase the risk of infections in early childhood, reviewed in [123, 124, 129]. The apparent inverse relationship between childhood infections and atopic diseases is the basis for the Hygiene hypothesis, which states that infections during early life, bacterial and/or viral, redirect the maturing immune system toward Th1 responses and thereby prevent the Th2-polarized allergic responses [132]. Hence, a reduction in the overall microbial burden, as observed in modern “Western” societies, may result in unrestrained Th2 responses and an increase in allergic diseases.

Current Therapeutic Strategies The best way to avoid allergy is simply to avoid the allergen. However, this is only feasible with certain allergens, such as food allergens, and it requires that the allergen is known. A lot of research is therefore devoted to identifying allergens and developing diagnostic tests. However, many of the most common allergens are difficult to avoid, e.g., inhalant allergens, demonstrating the need for treatments.

Most of the currently available treatments for atopic allergies involve symptom relieving drugs which target the effects of the mediators rather than preventing the allergic response. For instance, the actions of histamines and leukotrienes are prevented by histamine H1 receptor antagonists, i.e., anti-histamines (Clarityn and Teldanex) and leukotriene receptor antagonists (Monteleukast), respectively. However, a specific antagonist arrests the effect of only one in the array of mediators involved in the allergic response. The antihistamines, are effective in rhinitis and against itch in atopic dermatitis (eczema), but have limited effect in the management of asthma [133]. The corticosteroids (Pulmicort, Becotide and Prednisolone) have a more general anti- inflammatory effect and, at present, they constitute the most effective treatment in both asthma and rhinitis [133]. However, high doses or long-term treatment with corticosteroids, even topical corticosteroids, may have systemic side effects [133, 134] and lead to permanent insufficiency of the adrenal cortex. Furthermore, although corticosteroids efficiently abolish the inflammation characteristic of the late phase response, they have no effect on the immediate response. Fast relief from the immediate response in asthma is achieved with β2-stimulators (Bricanyl and Ventolin) that act on β2-receptors and relax bronchial muscles. Sodium chromoglycates (Lomudal) prevent both the immediate and late phase responses by stabilizing mast cells and preventing degranulation. The side effects of sodium chromoglycates are minor, but they have no effect at all in about 1/3 of all patients with allergic asthma. Finally, potentially fatal systemic reactions, anaphylactic shock, are alleviated by prompt administration of adrenaline [134].

26 Although the current pharmacological treatment of atopic disease often is effective, the drugs have only a short-term effect (symptoms recur within weeks of withdrawal) and the adverse side effects of long-term treatment are of great concern.

Hyposensitization When treatment with drugs give unsatisfactory results, hyposensitization therapy is sometimes applied in an attempt to reduce the patient's sensitivity to the allergen. Hyposensitization therapy involves repeated subcutaneous injections of increasing doses of allergen over long periods of time (years), a desensitizing strategy that has been used to treat allergic diseases for nearly 100 years [135]. The exact mechanism behind hyposensitization is unknown, but it may involve immune deviation from Th2- to Th1- polarized responses and/or induction of anergy in allergen-specific T cells, as reviewed in [134, 136]. The highest success rates for this type of therapy are observed in seasonal allergic rhinitis as well as bee-venom and penicillin allergy [134, 136]. A successful treatment is often accompanied by an increase in the allergen-specific IgG4 to IgE ratio and it has been suggested that IgG4 antibodies prevent sensitization by competing with IgE for the allergen [134, 136]. Furthermore, in successful hyposensitization therapy against bee venom allergy, IL-10 secreting cells appear to be critical. IL-10 may have a role both in regulating the IgG4/IgE balance and/or in induction of anergy in allergen- specific T cells, as reviewed in [134, 136]. However, there is no clear-cut correlation between any of these findings and clinical improvement in patients. Hence, although hyposensitization is sometimes successful in reducing the sensitivity to the injected allergen, there is no guarantee of improvement. Furthermore, all forms of immunotherapy require identification of the allergen and a given patient may be sensitized to numerous allergens, which may be difficult to prepare for injection. Most importantly, hyposensitization therapy always involves a risk of triggering dangerous, and potentially fatal, anaphylactic reactions.

Consequently, several attempts have been made to minimize the risk of systemic reactions in response to hyposensitization therapy. This may be achieved by modifying the allergen in various ways, either chemically with formaldehyde, generating so-called allergoids, or by recombinant techniques, as reviewed in [136]. The aim of the modifications is to reduce the IgE-binding capacity, but unfortunately the immunogenicity is often reduced as well and thus also the efficacy of the treatment [134].

Due to the undesirable side effects, drawbacks and sometimes unsatisfactory results of the current therapeutic approaches, a lot of research is devoted to finding better alternatives.

27 Allergen-independent treatments Concomitant with improved understanding of the mechanisms behind atopic allergy, several potential allergen-independent strategies for intervention in the allergic response have emerged during the last decade.

As previously mentioned, IL-4 and IL-13 are key factors in inducing IgE synthesis and in addition, IL-4 is directly involved in the differentiation of Th2-type responses. Hence, by blocking the effects of these central cytokines one might prevent the development of allergic responses. A high-affinity human IL-4 antagonist, obtained by a single amino acid substitution in the activation domain, has been shown to inhibit both IL-4 and IL- 13 induced IgE secretion from human B cells in vitro, as reviewed in [137]. Furthermore, an analogous mouse IL-4 antagonist efficiently inhibited IL-4-dependent allergen sensitization in vivo [138].

Other allergen-independent approaches focus on interference with the IgE/FcεRI interaction and thereby prevent receptor cross-linking and mast-cell degranulation. For instance, dimeric human receptor chimeras, consisting of the FcεRI α-chain coupled to either the entire Fc-region [139] or the CH3 domain only [140] of human IgG1 have been developed. Both of these soluble receptor chimeras (receptor antagonists) bind free IgE with high affinity and inhibit IgE binding to cell-surface-bound FcεRI [139, 140]. Other laboratories have focused on the possibility of blocking the receptor with short peptides that mimic the receptor-binding properties of IgE, as reviewed in [141]. However, to be clinically effective such peptides have to be long-lived, non- immunogenic and bind FcεRI with very high affinity, thus making the process of designing them, at best, difficult.

Due to its central role in atopy and limited involvement in other biological reactions, IgE is considered to be a prime target in the development of anti-allergic treatments. Based on the assumption that a substantial reduction in total and antigen-specific IgE levels will correlate with improved conditions in atopic individuals, two distinct strategies aiming at IgE depletion have evolved. One approach is passive immunization with monoclonal anti-IgE antibodies and the other is active immunization, vaccination, against IgE.

Passive immunization against IgE Passive immunization involves repeated intravenous or subcutaneous injections with monoclonal antibodies against IgE. To achieve a beneficial pharmacological effect while avoiding anaphylactic side effects, an antibody that binds specifically to free IgE but not to receptor-bound IgE on the surface of mast cells and basophils is required. Furthermore, to avoid strong immune reactions against the monoclonal antibody (MAb) and thus allow repeated administration, the MAb must not be recognized as foreign. This can be circumvented by so-called humanization, which entails grafting of the complementarity determining regions (CDR) from the parent murine MAb into human

28 framework regions, generating reshaped human V regions which then are joined to human C regions [142]. These humanized antibodies retain the antigen specificity of the original mouse MAb, but are much less immunogenic.

Hence, different groups of investigators have independently isolated and humanized a number of high-affinity anti-IgE MAbs [142, 143]. The different research programs joined forces and one of the most promising MAbs was chosen for further clinical development, rhu-MAb-E25. rhu-MAb-E25, also called Omalizumab, is a κ IgG1 antibody that binds to the FcεRI-binding region of free IgE, and thus cannot interact with FcεRI-bound (nor CD23-bound) IgE and cause cross-linking [144, 145]. In addition, rhu-MAb-E25 may bind mIgE on the surface of B cells [144, 146, 147]. Therefore, the therapeutic effect of rhu-MAb-E25 may be two fold: by both neutralizing free IgE and down regulating IgE production by B cells [144].

In numerous clinical trials (phase I through phase III), treatment with rhu-MAb-E25 has resulted in substantial reductions in free-serum-IgE levels that correlate with beneficial effects in patients with allergic rhinitis [148-151] or asthma [152-156]. The free-serum- IgE levels decrease by ≥90% from pretreatment values within 24h of subcutaneous administration of rhu-MAb-E25, and continued treatment results in reduced expression of FcεRI on effector cells and thus reduced responsiveness to antigen stimulation, as reviewed in [145]. rhu-MAb-E25 treatment significantly reduces the severity of symptoms in patients with allergic rhinitis as well as their requirements for rescue medication [149, 151]. In one study, 21 % of the patients that received rhu-MAb-E25 considered their symptoms completely controlled, as compared to 2% in the placebo group [149]. Similar improvements have also been observed in patients with allergic asthma.

Treatment with rhu-MAb-E25 appears to be safe and well tolerated. To my knowledge, only one patient, who was included in a pilot study of aerosolized administration in the lungs, have developed antibody responses against rhu-MAb-E25, as reviewed in [145]. Furthermore, no signs of immune-complex-related diseases were reported in any of the above-mentioned studies, but no long-term studies have been conducted so far. In my opinion, the major disadvantage of passive immunization is the high dose requirement. Subcutaneous dosage of rhu-MAb-E25 is based on body weight and total serum IgE levels as follows: ≥0.016mg/kg/IU/ml. Hence, a patient that weighs 80kg and has a total serum IgE level of 500ng/ml (200IU) requires a monthly dose of about 250mg rhu-Mab-E-25 [155], which translates to a yearly dose of about 3 g, indicating that treatment may be very costly.

Active immunization against IgE The other approach to IgE depletion is active immunization, vaccination, against IgE. Usually vaccination entails injection of, e.g., attenuated microbes with the prospect of inducing strong anti-microbial immune responses that will protect the recipient from

29 future infection with the microbe in question. However, in this case the target molecule is not a foreign microbe but a self-protein, IgE. Some years ago, a study was initiated in the laboratory to explore the possibility of inducing strong and therapeutically useful antibody responses against self-IgE through vaccination.

Vaccination induces a polyclonal antibody response, but the same requirements that apply to the MAb, apply also to the polyclonal antibodies. Hence, antibodies generated in response to vaccination should bind to free IgE with high affinity, but not to FcεRI bound IgE. The crystal structure of the IgE/FcεRI interaction was not available at the time, but a 76-amino-acid region at the border between the Cε2 and Cε3 domains had been shown to interact with FcεRI and prevent binding of free IgE [157]. Therefore, with the aim of evaluating the vaccine approach in a rat model, a fusion protein consisting of the Cε2 and Cε3 domains of rat IgE coupled to glutathione-S-transferase (GST) from Schistosoma japonicum, GST-C2C3, was designed [137, 158-160].

The T cells of a mature immune system are normally self-tolerant, that is, they do not respond to self-proteins such as IgE. B cells, on the other hand, may be self-reactive, but they will remain in an inactive state in the absence of T-cell help. However, the T- cell tolerance may be circumvented by coupling a foreign-carrier protein, such as GST, to the target self-protein. In theory, B cells with mIg that is specific for the self-protein (the receptor binding region of rat IgE) will internalize and degrade the entire GST-C2C3 protein and thus present peptides derived also from GST. Hence, GST-specific T cells can provide rat-IgE specific B cells with the signals required for proliferation and production of anti-rat IgE antibodies.

Vaccination with GST-C2C3 together with Freund’s adjuvant resulted in strong IgG responses against self-IgE in three of four rat strains [158]. The anti-IgE response was accompanied by a significant reduction in total serum IgE, to below the detection limit of the assay (10-20ng/ml). Furthermore, GST-C2C3 vaccination resulted in an almost complete inhibition of mast cell reactivity in the skin upon challenge with either a cross-linking polyclonal anti-IgE antiserum or a specific allergen [158]. These results were encouraging indeed.

However, harsh conditions were required to obtain the bacterially expressed GST-C2C3 vaccine component in soluble form. Consequently, some of the surface epitopes of the C2C3-derived component in the vaccine might not have retained their native conformations, which, in turn, may generate antibody responses with inadequate affinity for native IgE. Hence, although the initial results were very promising, the solubility issue and a number of additional questions had to be addressed in order to proceed with the vaccine strategy in human clinical trials.

If successful, the vaccination strategy has several advantages over that of passive immunization. Firstly, vaccination induces a polyclonal response. The risk of immune-

30 complex disease is thus minimal compared to that incurred during long-term treatment with high doses of a MAb. Secondly, vaccination is estimated to entail a yearly administration dose that is about 10 000 times lower than that of passive immunization with a MAb. This will most certainly reduce the cost of treatment significantly. Furthermore, whereas passive immunization might require monthly visits to a physician, vaccination will probably not necessitate more than four visits a year, a clear benefit that certainly will affect patient compliance with treatment.

31 Present study

AIM The major aim of this study was to develop an active immunization strategy against IgE as a preventive, allergen-independent treatment for atopic allergy. The central role of IgE in allergic reactions is well established, but its normal physiological function remains somewhat elusive. Another important objective was thus to increase our understanding of IgE, focusing on structural aspects from an evolutionary perspective.

Results and Discussion

The rise …

Cloning and structural- and phylogenetic analysis of monotreme ε- and γ chains (Paper I and II) To trace the emergence of IgE and other post-switch isotypes in vertebrate evolution we have studied Ig expression in two monotreme species: the duck-billed platypus (Ornithorhynchus anatinus) and the short-beaked echidna (Tachyglossus aculeatus).

Partial platypus IgE and IgG cDNA clones were isolated using a PCR-based cloning strategy and degenerate primers, as described previously [25]. The clones obtained were used as probes to screen a platypus-spleen cDNA library. Positive clones were analyzed further and the complete nucleotide sequences were determined for one full-length clone each of platypus IgE and IgG. Upon further screening using a full-length γ-chain clone as probe, a novel γ-chain isotype was isolated and sequenced. The two platypus γ-chain isotypes were designated IgG1 and IgG2 in order of “appearance”.

In addition, a partial platypus IgE clone was used as a probe to screen an echidna spleen cDNA library. Again, positive clones were selected for further analysis and the complete nucleotide sequence of one full-length echidna ε-chain clone was determined.

Sequence comparisons reveal that both echidna and platypus IgE conform to the general structure displayed by eutherian and marsupial ε chains (paper I and II). Hence, both echidna and platypus IgE have four constant domains and all of the generally conserved cysteine residues involved in the various intra- and inter-chain disulfide bridge formations occur in the monotreme ε chains. Furthermore, the potential N-linked glycosylation site in the CH3 domain, conserved among all previously characterized ε chains, is present also in echidna and platypus IgE (Fig. 3).

32 S VH CH1 CH2 CH3 CH4 CDR1 2 3 Echidna IgE

LHH

S VH CH1 CH2 CH3 CH4 CDR1 2 3 Platypus IgE

LHH

SVH CH1 Hinge CH2 CH3 CDR1 2 3 Platypus IgG1

L H

SVH CH1 Hinge CH2 CH3 CDR1 2 3 Platypus IgG2

L H

Figure 3. Schematic maps of the echidna IgE, and platypus IgE, IgG1 and IgG2 heavy chain clones. The signal sequence (S) is represented in black. The light gray regions in the variable region (VH) denote the positions of the CDRs. The constant domains are illustrated by striped boxes and the hinge region (in IgG) is indicated in dark gray. The approximate positions of potential N-glycosylation sites are indicated by forks, and brackets indicate the positions of putative intra-chain disulfide bridges. Cysteine residues involved in putative inter-chain disulfide bridges connecting the two heavy chains and the heavy chain with the light chain are denoted by H and L, respectively.

Like the previously characterized eutherian and marsupial γ chains, the platypus γ1- and γ2 chains contain three constant domains and a hinge region between the CH1 and CH2 domains (Fig. 3). However, the sequence divergence between the two platypus IgG isotypes is remarkably high and they display several structural differences. In platypus IgG1, as in most, if not all, mammalian γ chains, the cysteine residue involved in the disulfide bridge connecting the heavy chain with the light chain occurs in the C-terminal region of the CH1 domain. In contrast, in platypus IgG2 this cysteine residue occurs in the N-terminal region of the same domain, which may be a unique feature among mammalian γ chains (Fig. 3). Furthermore, one potential inter-chain disulfide-bridge- forming cysteine residue occurs in the hinge region of IgG1, whereas two cysteine residues are present in the same region of IgG2 (Fig. 3). The two platypus γ chains also differ with respect to the number of potential N-linked glycosylation sites (Fig. 3). However, the lone site in the CH2 domain of IgG1 is conserved between the two platypus isotypes and among other γ chains as well. Moreover, this site corresponds to the generally conserved N-linked glycosylation site in the CH3 domain of mammalian IgE. The sequence divergence between the platypus γ chains is actually greater than that

33 between rodent and human IgG (Fig. 4). Hence, the presence of three constant domains in both of the two highly divergent platypus γ chains suggests that the deletion of the second domain in an ancestral IgG occurred very early in mammalian evolution.

The primary structures of the monotreme ε- and γ chains suggest that all of the major evolutionary changes that gave rise to the mammalian Ig isotypes occurred before the separation of the three extant mammalian lineages. However, the results from the phylogenetic analyses are not conclusive with respect to the evolutionary relationship between monotremes and marsupials. Based on the ε-chain sequences, monotremes and marsupials appear to be sister lineages, which is in agreement with the Marsupionata hypothesis [31-34]. In contrast, the platypus IgG isotypes, together with IgM and IgA sequences support the opposing Theria hypothesis, according to which the monotremes constitute the most divergent mammalian lineage [22, 23, 35-40] (Fig. 4). Estimates based on protamine and IgM sequences indicate that the monotremes separated from the ancestors of marsupials and placentals about 150-170 million years ago [22, 23], whereas the separation of early mammalian ancestors from early reptiles may have occurred as far back as 310-330 million years ago [24]. Provided that the Theria hypothesis is correct, this leaves us with an evolutionary window for the appearance of modern post-switch isotypes of about 150 million years (between 150-170 to 310-330 million years ago).

Comparative analysis of mammalian ε chains (Paper II) Including platypus and echidna IgE, fifteen different mammalian ε-chain sequences have been isolated so far. Furthermore, each of the three mammalian lineages (placentals, marsupials and monotremes) are now represented by at least two ε-chain sequences, providing us with decent grounds for a detailed comparative analysis of mammalian IgE.

Comparative amino acid sequence analysis reveals that the overall IgE structure is conserved among the various ε chains. All ε chains have four constant domains and, with a few exceptions, the various cysteine residues involved in putative disulfide bridge formations occur in conserved positions. The most striking deviations occur in the rodent ε chains, which lack one of the intra-chain disulphide bridges in the CH1 domain and have an extended C-terminus.

34 999 Axolotl 829 Xenopus Turtle 933 Duck 1000 1000 Chicken 869 Cow Mouse IgM 952 947 Human Possum 1000 1000 Opossum ECHIDNA

995 PLATYPUS-A1 ECHIDNA 1000 PLATYPUS-A2 1000 1000 Possum Opossum 692 Rabbit-A11 1000 Rabbit-A10 892 764 Rabbit-A3 Rabbit-A6 882 Rabbit-A2 1000 Mouse IgA 908 591 Dog 998 Pig 562 Macaque 1000 Human-A2 1000 Human-A1 991 Gorilla-A1 Duck 990 1000 Chicken

662 Axolotl Duck IgY 1000 Chicken Sheep-G2 436 1000 Pig-G1 Pig-G4 1000 Pig-G2b 601 665 Human-G4 1000 Human-G2 597 Human-G3 1000 672 Human-G1 Rabbit 1000 Rat-G2a Mouse-G1 IgG 987 845 Mouse-G3 611 Hamster 812 Rat-G2b 998 881 Mouse-G2a Possum 1000 Opossum PLATYPUS-G2 996 ECHIDNA-G1 1000 PLATYPUS-G1 787 1000 ECHIDNA 985 PLATYPUS Possum 1000 Opossum 1000 Rat 1000 Mouse 883 Pig IgE 962 Sheep 1000 Cow 1000 Dog 928 980 Cat Horse 0.05 720 Orangutan 1000 Human 883 Chimpanzee

Figure 4. Ig sequence comparisons presented as a phylogenetic tree. A comparative analysis of all presently available IgE heavy chains sequences, along with a selection of mammalian IgG, IgA and IgM sequences as well as some avian and amphibian IgM and IgY heavy chain sequences. The bootstrap tree is based on 1000 independent comparisons and the bootstrap values are added at each branch point. The sequence distance is represented by the length of the branches according to the time scale (lower left), which corresponds to a sequence divergence of 5%.

35 The number and positions of potential N-glycosylation sites differ considerably among the various ε chains (Fig. 5). In general, the monotreme and marsupial ε chains contain fewer N-glycosylation sites than the eutherian ε chains. Furthermore, no N- glycosylation site occurs in the CH1 domain of monotreme and marsupial ε chains, whereas two or three sites are present in this domain of eutherian IgE. Conversely, some eutherian ε chains (those of primates, pig and horse) lack glycosylation sites in the CH4 domain. However, one potential N-glycosylation site in the CH3 domain occurs in all ε chains, suggesting that glycosylation at this position is important for protein structure and/or solubility. This site actually occurs in the FcεRI binding region of human IgE, but the involvement of the attached carbohydrate in the interaction appears to be limited [72]. However, glycosylation at this site may be important to ensure a proper conformation of the receptor-interacting region [161, 162]. Although additional glycosylation may be of some importance, the differences observed among the various ε chains suggest that their positioning is less restricted.

L H H

CH1 CH2 CH3 CH4 Echidna Platypus Opossum Possum Mouse Rat Sheep Pig Horse Cow Dog Cat Chimpanze Orangutan ( ) Human

Figure 5. Variation in glycosylation patterns among mammalian ε chains. The constant regions of the presently available ε chains are indicated as lines below a schematic map of the overall ε-chain structure. The approximate positions of potential N-glycosylation sites in the respective ε chains are indicated by forks.

Interaction between IgE and FcεRI involves both of the CH3 domains, each binding asymmetrically to two separate sites on one FcεRIα [72]. All together, eighteen amino acid positions in the CH3 domains are involved in the interaction [72]. Six of these amino acid residues are identical among all fifteen ε chains and most positions contain analogous amino acid residues (paper II). The putative receptor-binding regions of the rodent ε chains are the most divergent, sharing only eight to ten residues with the corresponding regions of monotreme and primate IgE (paper II). Despite the rather

36 significant differences in the receptor-binding regions, rodent IgE can bind to the human FcεRI receptor [163]. However, mouse IgE cannot engage human CD23 and human IgE cannot bind to either of the rodent receptors [81, 164]. Inter-specie reactivity has also been reported between dog IgE and primate FcεRI, whereas both human and rhesus monkey IgE failed to sensitize canine skin [165]. The structural basis for these inter- species interactions and the apparent specie specificity of human IgE is unknown. However, it is tempting to link the latter to three non-analogous substitutions that occur exclusively in the receptor-binding region of primate IgE (paper II).

Furthermore, we compared the various ε chains with respect to charge distribution at pH 7.0 and isoelectric point (pI) values calculated from their amino acid sequences. The monotreme and marsupial ε chains all have a considerable negative net charge, resulting in pI values at or below 6.0, whereas the eutherian ε chains have pI values between 6.5 and 8.2. Interestingly, there appears to be some correlation between charge distribution and potential N-glycosylation sites (Fig. 5). Domains that lack, or have few, N- glycosylation sites often have a more pronounced negative net charge and vice versa (paper II), thus suggesting that the actual pI values for the various ε chains may be more similar than what can be discerned from the amino acid sequence alone due to the presence of sialic acid-containing complex oligosaccharides.

The presence of IgE and the conservation of the overall structure in all extant mammalian lineages indicates an early appearance in mammalian evolution and a selective advantage of structural maintenance. The assumed duplication of an ancestral IgY in the lineage leading to mammals separated the major effector functions characteristic of present day IgG and IgE. This separation, which limited the role of the major plasma Ig isotype (IgG) in mast cell and basophil activation, may have had a selective advantage that could explain the maintenance of both IgG and IgE in all extant mammalian lineages.

Cloning, structural analysis and expression of Pig IgE (Paper III) The pig has been used extensively in biomedical science for many years, especially in surgical and cardiovascular research. Furthermore, due to several similarities with human physiology, together with other advantages, the pig is a prime choice as a potential xenograft donor in human transplantations [166, 167]. Consequently, immunological studies in the pig have gained much attention and the pig has been developed as a large model for studies of inflammatory reactions [168-170]. However, relatively few of the reagents required for studies of allergic inflammation are available, including sensitive and specific assays for total and antigen-specific IgE. To increase our knowledge of IgE, as well as the possibility to use the pig model in studies of IgE-mediated inflammatory reactions, we have cloned and expressed the ε chain of IgE in the domestic pig (Sus scrufa).

37 Two highly conserved regions among the previously characterized ε chains, one in the 3’-end of the CH2 domain other in the 5’-end of the CH4 domain, were selected for synthesis of degenerate oligonucleotide primers (paper III). A partial pig IgE DNA fragment was isolated through PCR amplification using the degenerate primers and single stranded cDNA derived from spleen and lymph nodes of Ascaris-sensitized pigs as template. The clone obtained was used as probe to screen a cDNA library constructed from mRNA isolated from the spleen and lymph nodes of the same Ascaris-sensitized animals. Positive clones were analyzed further and the complete nucleotide sequences were determined for one full-length clone of porcine IgE.

Amino acid sequence comparison with the previously characterized ε chains revealed that the pig ε chain conforms to the general IgE structure and that it shares the highest degree of sequence similarity with the sheep ε chain (Fig. 4).

The CH2-CH3-CH4 region of pig IgE was expressed in insect cells using a baculovirus expression system. In brief, a DNA fragment encoding the AcNPV gp64 signal, which enables secretion of the target protein, followed by a histidine hexapeptide and the CH2- CH3-CH4 domains of the pig ε chain was obtained through PCR amplifications (paper III). The final PCR product was ligated into the baculovirus transfer vector pVL1392 (PharMingen) and the obtained vector was used as a rescue plasmid in the BaculoGold (PharMingen) expression system. After co-transfection of Sf 9 insect cells and two additional rounds of virus amplification, a high-titer recombinant virus stock was obtained and used to infect High-Five insect cells. Recombinant pig IgE (6xHis-CH2- CH3-CH4) was purified from the conditioned medium on Ni2+-NTA-agarose (QIAGEN) columns.

SDS-PAGE analysis revealed that under reducing conditions the recombinant pig IgE migrated as a protein with a molecular weight of ~43kD, which corresponds to the estimated size of the 6xHis-CH2-CH3-CH4 protein. However, under non-reducing conditions only a small fraction of the recombinant protein migrated as the expected dimer. Instead, most of the material migrated as large multimeric complexes. The aggregation was probably caused by the presence of an unpaired cysteine residue in the N-terminal region of the CH2 domain.

The recombinant pig IgE protein can be used to raise polyclonal and monoclonal antibodies. This will facilitate the development of sensitive and specific assays for pig IgE and thus increase the possibility to use the pig model in future studies of allergic reactions.

At present, rodent models are generally employed in studies of IgE-mediated reactions, owing in part to the availability of the reagents required to monitor the reactions. However, there are several differences between the rodent and human systems that require consideration. For instance, rodent FcεRI is expressed exclusively on mast cells and

38 basophils and has an obligatory αβγ2 structure, whereas human FcεRI appears to have a wider distribution including eosinophils and platelets as well as monocytes, Langerhans cells and dendritic cells in a putative trimeric αγ2 structure [68]. Furthermore, human mast cells appear to be regulated exclusively by IgE, and no cross- talk between IgE- and IgG mediated responses occurs [171]. In contrast, IgE-deficient mice can display anaphylactic reactions upon antigen challenge, demonstrating that non- IgE pathways for mast cell activation exist in rodents [45]. In addition, in mice, IgE can interact with inhibitory and activatory Fcγ receptors on the surface of mast cells [172]. Hence, although rodent models are useful, extrapolation from rodents to the human situation may have limitations. So far, the situation in the pig is unknown, but, physiologically, the pig appears to be compatible with humans and the need for more compatible animal models is apparent.

The fall …

Active immunization against IgE (paper IV)

The design and production of a soluble vaccine component against rat IgE Based on the assumption that a substantial reduction in total and antigen-specific IgE levels will correlate with improved clinical outcomes in atopic individuals, an active immunization strategy for IgE depletion was developed in our laboratory [158]. In the previous studies, a fusion protein consisting of the Cε2-Cε3 region of rat IgE coupled to GST, GST-C2C3, was used as the active component in the vaccine. The initial results were very promising, but GST-C2C3 could only be obtained in a soluble form if harsh denaturing conditions were applied [158]. This solubility problem as well as several other issues, including the risk of generating cross-linking antibody responses, had to be resolved to proceed with the vaccine strategy in view of human clinical trials.

Initially, we invested a lot of time and effort to obtain the GST-C2C3 vaccine component in a soluble form. However, following several fruitless attempts using different bacterial expression systems, a yeast system (Pichia pastoris) and a baculovirus expression system we concluded that solubility requires maintenance of the IgE framework, which in turn requires correct folding, dimerization, disulfide bridge formation and probably also glycosylation [173]. Considering this, the vaccine design itself, a fusion of two radically different proteins with respect to processing, was probably a major impediment. GST is an intracellular protein devoid of post- translational modifications, whereas IgE is a glycosylated membrane-bound or secreted protein. Therefore, it was essential to find an alternative, more suitable, carrier protein. Furthermore, the GST-C2C3 vaccine contained both the Cε2 and Cε3 domains of rat IgE, that is, self-IgE. To limit the risk of generating cross-linking responses we aimed at reducing the self component in the next-generation vaccine.

39 Evolutionary studies of IgE in the laboratory had recently led to the cloning of IgE from a marsupial, the opossum (Monodelphis domestica) [25]. Opossum IgE and rat IgE are distantly related, with a sequence identity of ~43 % on the amino acid level, suggesting that domains derived from opossum IgE are sufficiently foreign to function as a carrier. However, despite the sequence divergence, the overall structures of opossum and rat IgE appear to be similar [25]. Therefore, domains derived from opossum IgE might also serve as structural support for the self-IgE-derived target domain. The Cε2 domain of IgE is not required for binding to FcεRI, that is, the active conformation of the receptor- binding region occurs in the absence of the Cε2 domain [77]. However, the Cε2 domain contains both of the cysteine residues involved in the inter-chain disulfide-bridge formations that occur in the IgE dimer. The Cε4 domains, on the other hand, appear to be important for the maintenance of the active conformation of the receptor-interacting Cε3 domain [77]. Based on these considerations, a sequel vaccine component containing the third constant domain of the rat ε chain flanked by the second and fourth constant domain of the opossum ε chain, OpossumCε2-RatCε3-OpossumCε4, ORO, was designed and expressed in a mammalian expression system (Fig. 6).

6x Opossum Rat Opossum S His CH2 CH3 CH4 AAAA

Signal seq. H H cleavage *

Figure 6. DNA construct of the ORO vaccine component. The signal sequence (S) is derived from human BM40. An asterisk denotes an unpaired cysteine residue that was changed to serine by point mutagenesis.

In brief, the DNA construct of ORO was made from cDNA clones using PCR amplification and subsequent subcloning. For purification purposes a region encoding a histidine hexapeptide was included in the 5’ primer directed against the region encoding the N-terminal of the opossum Cε2 domain. DNA fragments corresponding to the ORO protein were cloned into a modified expression vector named pCEP-Pu2, which contains a signal sequence that targets the expressed protein for secretion. Human embryonic kidney cells, 293-EBNA (Invitrogen), were transfected with the recombinant plasmid using Lipofectamine (Gibco BRL) and subsequent selection for the plasmid was effected by the addition of 0.5µg/ml puromycin (P8833, Sigma). The recombinant ORO protein was purified from the conditioned media on Ni-NTA-agarose (QIAGEN) using a batch procedure which entails protein binding to the adsorbent in solution.

SDS-PAGE analysis revealed that under reducing conditions ORO migrates as a protein with a molecular weight of ~48kD. However, under non-reducing conditions most of the protein migrates at ~97kD, suggesting that recombinant ORO is produced as a disulfide- linked dimer. Furthermore, preliminary Biacore measurements revealed that both ORO

40 and its human counterpart (replacing rat Cε3 with human Cε3) bind to the human FcεRI. Hence, the hybrid proteins appear to fold properly and the receptor-binding target-region occurs in an active and thereby a nearly native conformation. Since B cells recognize epitopes of native antigens, a nearly native conformation of the target region is probably required to evoke an antibody response with sufficiently high affinity for native IgE, especially considering the high affinity of the competing IgE-FcεRI interaction (Ka~1010 M-1)

Another objective of the present study was to define the requirements of the target region in the vaccine, that is, the receptor-binding region derived from self-IgE. With this aim, two additional vaccine components, OROO and OMO, were designed and expressed (Fig. 7). The purpose of OROO, which contains a part rat, part opossum Cε3 domain, was to investigate whether we could trim down the self-component in the vaccine even further without compromising efficacy. The aim of OMO, which contains a mouse Cε3 domain, was to study the efficacy of a vaccine component with an altered target region. The mouse Cε3 domain differs from that of the rat in 21 positions, corresponding to 19% difference in sequence, and can thus be considered as an altered target region. In addition, recombinant OOO (opossum Cε2-Cε4) was produced to facilitate studies of the response against the carrier protein alone (Fig. 7).

ORO OROO OMO OOO IgE 6xHis 6xHis 6xHis 6xHis CH2

CH3

CH4

Opossum Rat Mouse

Figure 7. Schematic structure of the various recombinant proteins.

Vaccination of ovalbumin-sensitized rats To evaluate the efficacy of the ORO vaccine in a more clinically significant setting, an ovalbumin (Ova) sensitized rat model was established through repeated intraperitoneal (i.p.) injections of Ova in PBS prior to vaccination. The study involved four different rat strains that demonstrate significant, strain-dependant differences in normal serum IgE levels: Brown Norway, Louvain, Wistar and Lewis. The initial IgE levels in the Brown Norway strain were so high (500-110,000 ng/ml) that no additional effect of sensitization was observed, whereas the Lewis strain had very little IgE (1-16 ng/ml)

41 but did not seem to respond to sensitization with this isotype. However, the initial IgE levels in the Louvain and Wistar strains were 1-27ng/ml and 7-155ng/ml, respectively, and sensitization resulted in an approximately eightfold increase in total serum IgE in both of these strains. To mimic the deregulated control of IgE expression in atopic patients, Ova sensitization was continued throughout the vaccination program. The treatment program for the Wistar strain is shown in figure 8.

Groups of animals from the four different strains were vaccinated with ORO or treated with BSA as negative control. In addition, groups of Wistar rats were vaccinated with OROO, OMO or OOO according to the treatment program (Fig. 8). Blood samples were collected at several time points (Fig. 8) and comparative IgG anti-IgE titers and total IgE levels in the sera were determined by ELISA analysis.

Sensitization (10µg Ova) Sensitization (3µg Ova) Blood draw -10 -5 0 5 10 15 20 Vaccination Week number

Figure 8. Time line of the treatment program for the Wistar strain. At the initial vaccination, the animals received 250µg recombinant protein (ORO, OROO, OMO or OOO) or BSA in 100µl neutral PBS mixed 50:50 with Freund’s Complete Adjuvant. In week 3 , 100µg recombinant protein or BSA in 100µl PBS mixed 50:50 with Incomplete Freund’s adjuvant (IFA) was administered. At the following two occasions, the animals received 25µg recombinant protein or BSA in 50µl PBS mixed 50:50 with IFA. At all vaccinations were administered i.p.

Antibody responses against IgE were detected in all ORO-vaccinated animals (paper IV). Comparative IgG anti-IgE titers detected in Wistar sera from week 10 of the treatment program are shown in figure 9. The average response in the Wistar strain was ~80% of that in the Lewis strain, the strain which generated the most prominent response (paper IV). The average responses against IgE in the Brown Norway and Louvain strains were much weaker, about 20-fold and 13-fold lower, respectively, than that in the Lewis strain (paper IV). Therapeutic effects of the antibody response generated against ORO were observed in three of the vaccinated strains (paper IV). The serum IgE levels were reduced to below 10 ng/ml in all vaccinated Wistar and Lewis rats and in four out of nine ORO-vaccinated Louvain rats. In comparison with control animals, and based on geometrical mean values, the reduction in serum IgE in ORO-vaccinated Wistar and Lewis rats was 90% and 92%, respectively. The total serum IgE levels determined in the Wistar sera at different time points are shown in figure 9. In addition, a significant effect of ORO vaccination was observed in the skin of Ova-sensitized Wistar rats challenged with the allergen, Ova (paper IV). The apparent mast cell activity was significantly lower in ORO-vaccinated animals compared to BSA-treated control animals, indicating

42 that the antibody response generated against ORO also has a clinical effect on mast cells in the periphery.

BSA ORO A Controls Vaccinated 2 Anti-ΙgE OD405 1.5

1 1=1:5 2=1:25 3=1:125 0.5 4=1:625 5=1:3125 6=1:15625 0 1 23456 1 23456 1:5 Serial dilutions

B 1000 [IgE] ng/ml 100

10

1 -7 0 5 10 14 -7 0 5 10 14 Week number

Figure 9. Therapeutic anti-IgE titers and total serum IgE in vaccinated Wistar rats. A) Comparative analyses of anti-IgE antibody titers in 1:5 serial dilutions (in horse sera) of sera collected in week 10 of the treatment program from BSA-treated and ORO-vaccinated animals. B) Total serum IgE concentrations in sera collected throughout the treatment program from the BSA-treated and ORO-vaccinated Wistar rats. The same symbol is used for the same animal in the anti-IgE (A) and total IgE (B) assays.

In comparison with ORO, vaccination with the modified vaccine components, OROO and OMO, generated considerably lower responses against IgE and no significant effects on total serum IgE levels were observed (paper IV). Hence, the size and the number of available epitopes appear to be of major importance for the efficacy of the vaccine. However, it should be mentioned that recombinant OROO was produced at lower yields than the other proteins and that upon SDS-PAGE analysis under non-reducing conditions the band corresponding to OROO was more smeared and less of the material migrated as the dimer (paper IV). It is possible that the design of the dual species Cε3 domain caused problems in protein folding and that the reduced efficacy of OROO is due to a somewhat misfolded target region rather than to the size of the self-component. Nevertheless, the OMO protein appears to be properly folded (paper IV), and considering that 81% of the amino acid residues in the rat and mouse Cε3 domains are identical and that only two of the 18 residues shown to be involved in the receptor interaction differ between the two species [72] (paper II), the failure of OMO is notable. The outcome of these preliminary studies is that, to reach optimal efficacy the target region should be as

43 large as possible without running the risk of cross-linking antibody responses, and the species-specific sequence should not be altered.

Reversibility of treatment Given our limited understanding of what future challenges our immune system may face and the role IgE may play in these challenges, it was important for us to demonstrate a return to the pre-vaccination state after discontinuation of treatment. The most prominent anti-IgE response was observed in the Lewis strain and by week 45 of our study, about 11 months after the booster injection, the anti-IgE titers in this strain were reduced by 89-99%. Other modified self-proteins have shown similar reversibility patterns [174-176], suggesting that therapeutic antibody responses against a self-protein are reversible with time and that the antibody response is not potentiated by the presence of the non-modified self-protein.

Safety of ORO vaccination Our main concern was the risk of generating potentially cross-linking antibody responses against the ORO vaccine. Firstly, antibody responses against the carrier, opossum IgE, might be cross-reactive and thereby potentially cross-linking. Most of the sequence identity between opossum and rat IgE is concentrated around cysteine residues involved in putative disulphide bridge formations and may not be easily available as B- cell epitopes. However, the overall structure of opossum and rat IgE appear to be similar and therefore, structural epitopes that are difficult to discern from the sequence may generate cross-linking responses. Secondly, although the self-component is significantly reduced in the ORO vaccine, the risk of generating potentially cross- linking responses against epitopes outside of the receptor-binding regions in the self- derived Cε3 domain remains. The amino acid regions involved in the FcεRI interaction are spread across the Cε3 domain, but they represent only a fraction of the residues in this domain [72].

A skin reactivity assay in high responder Brown Norway rats was used to determine the mast-cell activating potential of the antibody responses generated in vaccinated Wistar rats. No skin reactivity was detected upon provocation with sera from Wistar rats vaccinated with either OOO or ORO, whereas the positive control, a cross-linking mouse anti-rat IgE MAb, caused extensive mast cell cross-linking and degranulation (paper IV). Crystallization studies of human IgE-Fc Cε3-Cε4 have revealed that FcεRI binding induces a conformal change in the Cε3 domain [73], which may prevent antibodies directed against the Cε3 domain of free IgE from binding to receptor-bound IgE. Hence, conformal changes, together with steric hindrance upon receptor binding, may operate to limit the cross-linking potential of antibody responses generated against the entire Cε3 domain. The sequence identity between human and opossum IgE is similar to that between rat and opossum IgE, which strongly suggests that opossum IgE will be safe to use as a carrier protein in humans as well.

44 Loose ends

Here I will focus on the allergy vaccine and an important unresolved issue, that is, the apparent shortcomings of the vaccine in individuals with initially very high IgE levels.

Vaccination with ORO generated detectable antibody responses against IgE in all rat strains, including the high-responder strain Brown Norway (paper IV). Although this is a significant improvement, since no anti-IgE antibody responses were detected in this strain following vaccination with GST-C2C3 [158], the response generated against ORO had no effect on the serum IgE level. Our results indicate that the therapeutic capacity of the elicited response is dependent on the initial serum IgE concentration. Such high serum IgE levels as those observed in the Brown Norway strain probably induce a strong B-cell tolerance in the host that will impede the effects of the vaccine. Since the vaccine is intended as a therapeutic rather than a prophylactic treatment and that, more often than not, serum IgE levels are elevated in allergic individuals, ongoing studies are aimed at overcoming this potential limitation of the vaccine.

A successful pretreatment that reduces the IgE level significantly prior to vaccination might serve to evade tolerance induction in newly recruited B cells. For human purposes, humanized monoclonal anti-IgE antibodies, such as rhu-MAb-E25, might afford effective pretreatment [142, 143, 145], but such antibodies are not available for rats. In the absence of a rat anti-rat IgE we thought that a non-cross-linking monoclonal mouse IgG1 anti-rat IgE, MAS-314 (Harlan, Loughborough, England) might suffice to put our pretreatment hypothesis to the test.

In brief, Ova-sensitized brown Norway rats were treated with MAS-314 or PBS (control animals) weekly for six weeks. At each occasion, 1mg MAS-314 in 500µl PBS was administered intramuscularly in the thighs. In addition, a separate group of MAb-treated animals were vaccinated with ORO using a similar protocol as in figure 8. The initial dose of ORO was administered two days after the initial MAb treatment. Blood was collected at several time points throughout the study and analyzed for total serum IgE by ELISA.

Four days after the initial MAb treatment (week 1), an almost complete disappearance of total IgE was observed in the MAb treated animals, but the reduction was only temporary (Fig. 10). By week 3, the IgE levels in the MAb treated animals were actually elevated in comparison with those in the control animals, and at subsequent time points the IgE levels were similar in the two groups (Fig. 10). A second ELISA analysis, to detect antibodies against MAS-314, revealed that a strong response was elicited against the mouse MAb (Fig. 10). This response presumably induced the clearance of subsequent MAb doses and thus prevented further clearance of IgE. No therapeutic anti-IgE antibody responses were elicited in the ORO-vaccinated MAb-treated animals, which was not surprising in view of the brief effect of the MAb treatment.

45 Hence, although mouse and rat IgG1 share ~80% sequence identity over the constant regions, the mouse MAb did not work for pretreatment in our rat model.

PBS Control MAb Treated (MAS 314) 1000000 100000 10000 [IgE] ng/ml 1000 100 10 1 -4 -1 1 3 6 8101722 -4 -1 1 3 6 8101722 1

αMAS 314 0.75 OD 405 0.5

0.25

0 -4 -1 1 3 6 8101722 -4 -1 1 3 6 8101722 Week number

Figure 10. Treatment with MAS 314, a mouse anti-rat IgE MAb. Total serum IgE concentrations (upper graphs) and antibody titers against MAS 314 (lower graphs) in sera collected throughout the treatment program from PBS-treated (control animals) and MAS 314-treated, unvaccinated Brown Norway rats.

A chimeric molecule consisting of the extracellular portion of the human FcεRI α- chain coupled to a truncated human IgG1 heavy chain has been described previously [139]. When expressed in mammalian cells this chimeric molecule was secreted as an antibody-like disulfide-linked homodimer [139]. The chimeric molecule was capable of binding free IgE, and since it cannot bind receptor-bound IgE it may, in this respect, function as a non-cross-linking MAb against IgE. Hence, we have designed a rat version of this chimeric molecule as an alternative to a rat anti-rat IgE monoclonal. Whether this chimeric molecule can serve to reduce the IgE level prior to vaccination and if this, in turn, can serve to break the putative B-cell tolerance and, thereby, to enhance the therapeutic effects of the vaccine, remains to be seen.

Another important issue with regard to the vaccine is the adjuvant problem. In order to apply the vaccine strategy to humans it is essential to find an alternative to Freund’s adjuvant, which is too toxic for human use. At present, aluminum compounds are the only adjuvants approved in the US, and consequently the only widely used adjuvants for clinical use in humans. However, in a pilot adjuvant study, no antibody responses

46 against IgE were detected in Wistar rats vaccinated with ORO together with Al(OH)3 (Johansson et. al, manuscript in preparation). Aluminum adjuvants favor Th2-type responses and can induce IgE-antibody responses [177]. Hence, aluminum adjuvants may promote IgE-mediated reactions, which seems incompatible with the intended effect of the ORO vaccine. Our results using a water-in-oil formulated, mineral oil-based adjuvant analogous to Freund’s incomplete adjuvant, Montanide ISA 51 (Seppic, Paris, France) were more encouraging. Vaccination with ORO together with Montanide ISA 51 generated therapeutic antibody responses that were at least comparable to those generated with Freund’s adjuvant (Johansson et. al, manuscript in preparation). Montanide ISA 51 has been tested in AIDS, malaria and cancer immunotherapeutic clinical trials and the vaccine formulations were generally well-tolerated [178-180]. However, the acceptable level of toxicity is likely to be higher for AIDS, malaria and cancer treatments than for an allergy vaccine. Nevertheless, Montanide ISA 51 is a potential adjuvant candidate for the human version of the ORO vaccine.

Concluding remarks

When did IgE appear? IgE expression appears to be restricted to mammals, but it occurs in all three extant mammalian lineages: placentals, marsupials and monotremes. The conservation of the overall IgE structure among the available ε chains suggests that IgE was present, in essentially its present form, in a common mammalian ancestor before the evolutionary separation of the extant mammalian lineages. According to the Theria hypothesis, the monotremes constitute the most divergent lineage and the estimated time of divergence from the therian lineages is about 150-170 million years ago [22, 23]. Hence, an ancestral IgE was present at least 150-170 million years ago, but may have appeared already at, or shortly after, the time of divergence of the early mammalian ancestors from the early reptiles lineages, about 310-330 million years ago [24].

Do we need IgE? Mice with a targeted deletion of the Cε gene, IgE-/- mice, do not suffer from any apparent disease, they do not die prematurely, lymphocyte development is normal as are serum levels of the other Ig isotypes (IgG, IgA and IgM) [45, 95]. IgE deficiency may also occur in humans with an otherwise normal immunologic profile without evidence of illness [118]. However, laboratory mice, as well as humans in modern Western society, live in an environment that is almost devoid of parasites and the general microbial load is low. Under these circumstances IgE may well be redundant. However, these microbe-depleted environments have emerged relatively recently, especially in relation to evolutionary time. The presence of IgE and the conservation of the overall IgE structure in all extant mammalian lineages indicate an early appearance in

47 mammalian evolution and a selective advantage of structural maintenance. IgE probably played a beneficial role in host defense against parasitic infections during the evolution of mammalian species and may still do so. Furthermore, IgE functions as an enhancer of an immune response against antigens that occur in low concentrations, as reviewed in [95]. In addition, very low amounts of antigens are required to activate mast cells and basophils if antigen-specific IgE is present on the cell surface. In an environment where microbes are plentiful, the ability to respond to low amount of potentially fatal microbes, to enhance the response and induce mast cell effector functions may be crucial. However, these potentially beneficial effects of IgE might be of limited value in populations living in modern Western society, where allergic diseases constitute a far more important problem than do parasitic infections. This assumption is the basis for the development of therapeutic strategies aimed at IgE depletion. The clinical studies of passive immunization with rhu-MAb-E25 support this assumption in that no adverse side effects of IgE-depletion has been observed in any of the patients. Hence, at least in atopic individuals IgE appears to be dispensable.

Vaccination against IgE – is it possible? The present study suggests that vaccination against IgE is indeed a possible approach to prevent IgE-mediated immune responses. We have successfully produced a soluble vaccine component, ORO, which retains the receptor-binding target-region in a native or nearly native conformation. Vaccination resulted in a 90% decrease in serum IgE levels in sensitized Wistar rats. This is encouraging, since similar reductions in free serum IgE upon passive immunization with rhu-MAb-E25 have led to clinical improvements in atopic patients [148, 149, 154, 155]. In addition, ORO vaccination resulted in a significant inhibition of mast cell reactivity in the skin upon allergen challenge. A vaccine containing the entire Cε3 domain from self-IgE appears to be safe to use as an antigen. No cross-linking activity was observed in the sera of vaccinated animals and the elicited response was reversible with time. The allergic-rat model has its limitations and there are still some unresolved questions. However, the apparent efficacy and safety of ORO vaccination, together with the clinical benefits of vaccination over passive immunization with a MAb, suggests that the vaccination strategy has potential to become a therapeutic method for humans.

48 Acknowledgements

I would like to take the opportunity to thank all of you, who in one way or another, have contributed to the completion of this thesis. No one who has not been through this can possibly imagine how many people that are affected by the ordeal. In particular, I would like to acknowledge:

Professor Lars Hellman, my supervisor and mentor in the field of immunology, for letting me join the group and the vaccine project, your never-ending ideas, your encouragement and most of all for your confidence in my work. It has been a privilege to work for you!

Professor Lars Pilström, for sharing your wisdom of the small (Igs) and big things in life, for your encouragement and for leaving me licorice when I couldn’t have the real stuff.

Maria Aveskogh, our outstanding technician, for many shared laughs and for just being you. There are bad ways, better ways and Maria’s way to do lab work. Thank you for taking the time to teach me your way!

Everybody at Resistentia Pharmaceuticals: Markus, Bengt, Christina, Birgitta, Mats, Åsa, Stefan, Anki and Alexis, for your support and for believing in the vaccine enough to make it your bread and butter.

Everybody at ICM and IMBIM, especially Sigrid, Christer, Joakim, Eva, Tony and Ylva, who made my life easier by taking care of the practical details.

The vaccine girls: Anna, for your positive attitude and for taking the time to read this and making me feel like Shakespeare with your “suvves”. Jeannette, for being a good friend and for sharing the late hours and the agonies of writing.

The rest of ”Hellmans flickor”, Parvin, Camilla, Ulrika, Sara and Maike for making this place a good one, and coming to work a good thing.

Coauthors and collaborators, especially Gunnar Pejler, for help with the pig manuscript and Therese Kristersson, for helping me with the pig situation, and Ingrid Dahlbom.

The girls that came after, but left before me: Maryam, my first wing-mate, it was a pleasure (hire me!), Carolina, for your dimples, Lotta, for your wonderful theories of life, Otti, and Frida, for many laughs and for writing the “how to become a PhD” manual, it has been a big help.

The Cod group: Siv, for looking so happy to see me (it’s not only the coffee is it?), Niklas, especially for the cake you pretended to know nothing about, Sirje, Mats, and Anna-Carin.

Anders and Erika, the additional lunch dates, for fruitful discussions…

49 Former group members and colleagues: Magnus, Claudia, Tomas and Mårten for taking care of me when I was a novice.

All my friends outside of this world, no one named, no one forgotten.

My mother-in-law, Margareta Vernersson, also known as “världens bästa farmor” for your help with the kids, I could not have done this without you.

Sonja, my sister, for your support and for being an invaluable friend.

David Eaker (also known as Dad) for linguistic revision.

Mom and Dad, for your endless love and support, for giving me the tools to complete this, for your encouragement and for always believing in me. In short, for being my parents and for doing it so well.

Per, my wonderful husband, for taking also the first part of “i nöd och lust” seriously during the last month. I owe you big time.

Emil and Anna, my beloved children. I will not lie and say that you made this easier, but you are without competition the joys of my life.

50 References

[1] Du Pasquier L. Origin and evolution of the vertebrate immune system. Apmis 1992; 100:383-392. [2] Marchalonis J J, Schluter S F. A stochastic model for the rapid emergence of specific vertebrate immunity incorporating horizontal transfer of systems enabling duplication and combinational diversification. J Theor Biol 1998; 193:429-444. [3] Schluter S F, Bernstein R M, Bernstein H, Marchalonis J J. 'Big Bang' emergence of the combinatorial immune system. Dev Comp Immunol 1999; 23:107-111. [4] Bengtén E, Wilson M, Miller N, Clem L W, Pilström L, Warr G W. Immunoglobulin Isotypes: Structure, Function, and Genetics. In: Du Pasquier L, Litman G W, eds. Origin and Evolution of the Vertebrate Immune System. Vol. 248. Berlin Heidelberg New York: Springer-Verlag, 2000. p. 189-219. [5] Hsu E, Flajnik M F, Du Pasquier L. A third immunoglobulin class in amphibians. J Immunol 1985; 135:1998-2004. [6] Fellah J S, Kerfourn F, Wiles M V, Schwager J, Charlemagne J. Phylogeny of immunoglobulin heavy chain isotypes: structure of the constant region of Ambystoma mexicanum upsilon chain deduced from cDNA sequence. Immunogenetics 1993; 38:311-317. [7] Amemiya C T, Haire R N, Litman G W. Nucleotide sequence of a cDNA encoding a third distinct Xenopus immunoglobulin heavy chain isotype. Nucleic Acids Res 1989; 17:5388. [8] Natarajan K, Muthukkaruppan V R. Immunoglobulin classes in the garden lizard, Calotes versicolor. Dev Comp Immunol 1984; 8:845-854. [9] Leslie G A, Clem L W. Phylogeny of immunoglobulin structure and function. VI. 17S, 7.5S and 5.7S anti-DNP of the turtle, Pseudamys scripta. J Immunol 1972; 108:1656- 1664. [10] Parvari R, Avivi A, Lentner F, Ziv E, Tel-Or S, Burstein Y, Schechter I. Chicken immunoglobulin gamma-heavy chains: limited VH gene repertoire, combinatorial diversification by D gene segments and evolution of the heavy chain locus. Embo J 1988; 7:739-744. [11] Magor K E, Warr G W, Middleton D, Wilson M R, Higgins D A. Structural relationship between the two IgY of the duck, Anas platyrhynchos: molecular genetic evidence. J Immunol 1992; 149:2627-2633. [12] Mussmann R, Wilson M, Marcuz A, Courtet M, Du Pasquier L. Membrane exon sequences of the three Xenopus Ig classes explain the evolutionary origin of mammalian isotypes. Eur J Immunol 1996; 26:409-414. [13] Marchalonis J J. Isolation and characterization of immunoglobulin-like proteins of the Australian lungfish (Neoceratodus forsteri). Aust J Exp Biol Med Sci 1969; 47:405-419. [14] Litman G W, Wang A C, Fudenberg H H, Good R A. N-terminal amino-acid sequence of African lungfish immunoglobulin light chains. Proc Natl Acad Sci U S A 1971; 68:2321-2324. [15] Mussmann R, Du Pasquier L, Hsu E. Is Xenopus IgX an analog of IgA? Eur J Immunol 1996; 26:2823-2830. [16] Mansikka A. Chicken IgA H chains. Implications concerning the evolution of H chain genes. J Immunol 1992; 149:855-861.

51 [17] Martens C L, Currier S J, Knight K L. Molecular genetic analysis of genes encoding the heavy chains of rabbit IgG. J Immunol 1984; 133:1022-1027. [18] Burnett R C, Hanly W C, Zhai S K, Knight K L. The IgA heavy-chain gene family in rabbit: cloning and sequence analysis of 13 C alpha genes. Embo J 1989; 8:4041- 4047. [19] Spieker-Polet H, Yam P C, Knight K L. Differential expression of 13 IgA-heavy chain genes in rabbit lymphoid tissues. J Immunol 1993; 150:5457-5465. [20] Brown W R, Butler J E. Characterization of a C alpha gene of swine. Mol Immunol 1994; 31:633-642. [21] Kacskovics I, Sun J, Butler J E. Five putative subclasses of swine IgG identified from the cDNA sequences of a single animal. J Immunol 1994; 153:3565-3573. [22] Belov K, Hellman L, Cooper D W. Characterisation of echidna IgM provides insights into the time of divergence of extant mammals. Dev Comp Immunol 2002; in press. [23] Retief J D, Winkfein R J, Dixon G H. Evolution of the monotremes. The sequences of the protamine P1 genes of platypus and echidna. Eur J Biochem 1993; 218:457-461. [24] Kumar S, Hedges S B. A molecular timescale for vertebrate evolution. Nature 1998; 392:917-920. [25] Aveskogh M, Hellman L. Evidence for an early appearance of modern post-switch isotypes in mammalian evolution; cloning of IgE, IgG and IgA from the marsupial Monodelphis domestica. Eur J Immunol 1998; 28:2738-2750. [26] Miller R D, Belov K. Immunoglobulin genetics of marsupials. Dev Comp Immunol 2000; 24:485-490. [27] Diener E, Ealey E H. Immune system in a monotreme: studies on the Australian echidna (Tachyglossus aculeatus). Nature 1965; 208:950-953. [28] Connolly J H, Canfield P J, McClure S J, Whittington R J. Histological and immunohistological investigation of lymphoid tissue in the platypus (Ornithorhynchus anatinus). J Anat 1999; 195:161-171. [29] Atwell J L, Marchalonis J J, Ealey E H. Major immunoglobulin classes of the Echidna (Tachyglossus aculeatus). Immunology 1973; 25:835-846. [30] Belov K, Hellman L, Cooper D W. Characterization of immunoglobulin gamma 1 from a monotreme, Tachyglossus aculeatus. Immunogenetics 2002; 53:1065-1071. [31] Janke A, Gemmell N J, Feldmaier-Fuchs G, von Haeseler A, Paabo S. The mitochondrial genome of a monotreme--the platypus (Ornithorhynchus anatinus). J Mol Evol 1996; 42:153-159. [32] Janke A, Xu X, Arnason U. The complete mitochondrial genome of the wallaroo (Macropus robustus) and the phylogenetic relationship among Monotremata, Marsupialia, and Eutheria. Proc Natl Acad Sci U S A 1997; 94:1276-1281. [33] Penny D, Hasegawa M. Molecular systematics. The platypus put in its place. Nature 1997; 387:549-550. [34] Kirsch J A, Mayer G C. The platypus is not a rodent: DNA hybridization, amniote phylogeny and the palimpsest theory. Philos Trans R Soc Lond B Biol Sci 1998; 353:1221-1237. [35] Hu Y, Wang Y, Luo Z, Li C. A new symmetrodont mammal from China and its implications for mammalian evolution. Nature 1997; 390:137-142. [36] Luo Z X, Cifelli R L, Kielan-Jaworowska Z. Dual origin of tribosphenic mammals. Nature 2001; 409:53-57. [37] Kullander K, Carlson B, Hallbook F. Molecular phylogeny and evolution of the neurotrophins from monotremes and marsupials. J Mol Evol 1997; 45:311-321. [38] Messer M, Weiss A S, Shaw D C, Westerman M. Evolution of the Monotremes: Phylogenetic Relationship to Marsupials and Eutherians, and Estimation of

52 Divergence Dates Based on α-Lactalbumin Amino Acid Sequences. J Mamm Evol 1998; 5:95-105. [39] Lee M H, Shroff R, Cooper S J, Hope R. Evolution and molecular characterization of a beta-globin gene from the Australian Echidna Tachyglossus aculeatus (Monotremata). Mol Phylogenet Evol 1999; 12:205-214. [40] Killian J K, Buckley T R, Stewart N, Munday B L, Jirtle R L. Marsupials and Eutherians reunited: genetic evidence for the Theria hypothesis of mammalian evolution. Mamm Genome 2001; 12:513-517. [41] Magor K E, Higgins D A, Middleton D L, Warr G W. One gene encodes the heavy chains for three different forms of IgY in the duck. J Immunol 1994; 153:5549-5555. [42] Warr G W, Magor K E, Higgins D A. IgY: clues to the origins of modern antibodies. Immunol Today 1995; 16:392-398. [43] Faith R E, Clem L W. Passive cutaneous anaphylaxis in the chicken. Biological fractionation of the mediating antibody population. Immunology 1973; 25:151-164. [44] Andersson P. Antigen-induced bronchial anaphylaxis in actively sensitized guinea- pigs. Pattern of response in relation to immunization regimen. Allergy 1980; 35:65- 71. [45] Oettgen H C, Martin T R, Wynshaw B A, Deng C, Drazen J M, Leder P. Active anaphylaxis in IgE-deficient mice. Nature 1994; 370:367-370. [46] Ishizaka K, Ishizaka T. Identification of gamma-E-antibodies as a carrier of reaginic activity. J Immunol 1967; 99:1187-1198. [47] Johansson S G, Bennich H, Wide L. A new class of immunoglobulin in human serum. Immunology 1968; 14:265-272. [48] Bennich H H, Ishizaka K, Johansson S G, Rowe D S, Stanworth D R, Terry W D. Immunoglobulin E. A new class of human immunoglobulin. Immunochemistry 1968; 5:327-328. [49] Capra J D, Frazer, J K. Immunoglobulins: Structure and function. In: Paul W E, ed. Fundamental Immunology. Fourth Edition. Philadelphia: Lippincott-Raven Publishers, 1999. p. 60. [50] Peng C, Davis F M, Sun L K, Liou R S, Kim Y W, Chang T W. A new isoform of human membrane-bound IgE. J Immunol 1992; 148:129-136. [51] Zhang K, Saxon A, Max E E. Two unusual forms of human immunoglobulin E encoded by alternative RNA splicing of epsilon heavy chain membrane exons. J Exp Med 1992; 176:233-243. [52] Hellman L. Characterization of four novel epsilon chain mRNA and a comparative analysis of genes for immunoglobulin E in rodents and man. Eur J Immunol 1993; 23:159-167. [53] Zhang K, Max E E, Cheah H K, Saxon A. Complex alternative RNA splicing of epsilon-immunoglobulin transcripts produces mRNAs encoding four potential secreted protein isoforms. J Biol Chem 1994; 269:31322. [54] Saxon A, Max E E, Diaz-Sanchez D, Zhang K. Alternative RNA of epsilon transcripts produces mRNAs encoding two membrane and four secreted IgE isoforms. Int Arch Allergy Immunol 1995; 107:45-47. [55] Batista F D, Efremov D G, Burrone O R. Characterization and expression of alternatively spliced IgE heavy chain transcripts produced by peripheral blood lymphocytes. J Immunol 1995; 154:209-218. [56] Batista F D, Efremov D G, Burrone O R. Characterization of a second secreted IgE isoform and identification of an asymmetric pathway of IgE assembly. Proc Natl Acad Sci U S A 1996; 93:3399-3404.

53 [57] Batista F D, Anand S, Presani G, Efremov D G, Burrone O R. The two membrane isoforms of human IgE assemble into functionally distinct B cell antigen receptors. J Exp Med 1996; 184:2197-2205. [58] Anand S, Batista F D, Tkach T, Efremov D G, Burrone O R. Multiple transcripts of the murine immunoglobulin epsilon membrane locus are generated by alternative splicing and differential usage of two polyadenylation sites. Mol Immunol 1997; 34:175-183. [59] Diaz-Sanchez D, Zhang K, Nutman T B, Saxon A. Differential regulation of alternative 3' splicing of epsilon messenger RNA variants. J Immunol 1995; 155:1930-1941. [60] Chan L A, Lyczak J B, Zhang K, Morrison S L, Saxon A. The novel human IgE epsilon heavy chain, epsilon tailpiece, is present in plasma as part of a covalent complex. Mol Immunol 2000; 37:241-252. [61] Aveskogh M, Hellman L. A single major transcript encodes the membrane-bound form of rat immunoglobulin E. Scand J Immunol 1995; 42:535-539. [62] Beavil A J, Young R J, Sutton B J, Perkins S J. Bent domain structure of recombinant human IgE-Fc in solution by X-ray and neutron scattering in conjunction with an automated curve fitting procedure. Biochemistry 1995; 34:14449-14461. [63] Gould H J, Beavil R L, Vercelli D. IgE isotype determination: epsilon-germline gene transcription, DNA recombination and B-cell differentiation. Br Med Bull 2000; 56:908-924. [64] Robinson D S. Th-2 cytokines in allergic disease. Br Med Bull 2000; 56:956-968. [65] Gauchat J F, Henchoz S, Mazzei G, Aubry J P, Brunner T, Blasey H, Life P, Talabot D, Floresromo L, Thompson J, Kishi K, Butterfield J, Dahinden C, Bonnefoy J Y. Induction of human IgE synthesis in B-cells by mast cells and basophils. Nature 1993; 365:340-343. [66] Wierenga E A, Snoek M, de Groot C, Chretien I, Bos J D, Jansen H M, Kapsenberg M L. Evidence for compartmentalization of functional subsets of CD2+ T lymphocytes in atopic patients. J Immunol 1990; 144:4651-4656. [67] Sutton B J, Gould H J. The human IgE network. Nature 1993; 366:421-428. [68] Kinet J P. The high-affinity IgE receptor (Fc epsilon RI): from physiology to pathology. Annu Rev Immunol 1999; 17:931-972. [69] Wedemeyer J, Galli S J. Mast cells and basophils in acquired immunity. Br Med Bull 2000; 56:936-955. [70] Kita H, Kaneko M, Bartemes K R, Weiler D A, Schimming A W, Reed C E, Gleich G J . Does IgE bind to and activate eosinophils from patients with allergy? J Immunol 1999; 162:6901-6911. [71] von Bubnoff D, Geiger E, Bieber T. Antigen-presenting cells in allergy. J Allergy Clin Immunol 2001; 108:329-339. [72] Garman S C, Wurzburg B A, Tarchevskaya S S, Kinet J P, Jardetzky T S. Structure of the Fc fragment of human IgE bound to its high-affinity receptor Fc epsilonRI alpha. Nature 2000; 406:259-266. [73] Wurzburg B A, Garman S C, Jardetzky T S. Structure of the human IgE-Fc C epsilon 3- C epsilon 4 reveals conformational flexibility in the antibody effector domains. Immunity 2000; 13:375-385. [74] Wan T, Beavil R L, Fabiane S M, Beavil A J, Sohi M K, Keown M, Young R J, Henry A J, Owens R J, Gould H J, Sutton B J. The crystal structure of IgE Fc reveals an asymmetrically bent conformation. Nat Immunol 2002; 3:681-686. [75] McDonnell J M, Calvert R, Beavil R L, Beavil A J, Henry A J, Sutton B J, Gould H J , Cowburn D. The structure of the IgE Cepsilon2 domain and its role in stabilizing the complex with its high-affinity receptor FcepsilonRIalpha. Nat Struct Biol 2001; 8:437-441.

54 [76] Novak N, Bieber T. To bend or not to bend. Nat Immunol 2002; 3:607-608. [77] Basu M, Hakimi J, Dharm E, Kondas J A, Tsien W H, Pilson R S, Lin P, Gilfillan A, Haring P, Braswell E H, et al. Purification and characterization of human recombinant IgE-Fc fragments that bind to the human high affinity IgE receptor. J Biol Chem 1993; 268:13118-13127. [78] Helm B A, Sayers I, Higginbottom A, Machado D C, Ling Y, Ahmad K, Padlan E A, Wilson A P. Identification of the high affinity receptor binding region in human immunoglobulin E. J Biol Chem 1996; 271:7494-7500. [79] Bjorklund J E, Schmidt M, Magnusson C G. Characterisation of recombinant human IgE-Fc fragments expressed in baculovirus-infected insect cells. Mol Immunol 2000; 37:169-177. [80] Sayers I, Helm B A. The structural basis of human IgE-Fc receptor interactions. Clin Exp Allergy 1999; 29:585-594. [81] Nissim A, Schwarzbaum S, Siraganian R, Eshhar Z. Fine specificity of the IgE interaction with the low and high affinity Fc receptor. J Immunol 1993; 150:1365- 1374. [82] Capron A, Dessaint J P, Capron M, Joseph M, Ameisen J C, Tonnel A B. From parasites to allergy: a second receptor for IgE. Immunol Today 1986; 7:15-18. [83] Bonnefoy J Y, Aubry J P, Gauchat J F, Graber P, Life P, Flores R L, Mazzei G. Receptors for IgE. Curr Opin Immunol 1993; 5:944-949. [84] Mudde G C, Bheekha R, Bruijnzeel K C. Consequences of IgE/CD23-mediated antigen presentation in allergy. Immunol Today 1995; 16:380-383. [85] Bonnefoy J Y, Lecoanet H S, Gauchat J F, Graber P, Aubry J P, Jeannin P, Plater Z C. Structure and functions of CD23. Int Rev Immunol 1997; 16:113-128. [86] Yamaguchi M, Sayama K, Yano K, Lantz C S, Noben T N, Ra C, Costa J J, Galli S J . IgE enhances Fc epsilon receptor I expression and IgE-dependent release of histamine and lipid mediators from human umbilical cord blood-derived mast cells: synergistic effect of IL-4 and IgE on human mast cell Fc epsilon receptor I expression and mediator release. J Immunol 1999; 162:5455-5465. [87] MacGlashan D W, Jr., Bochner B S, Adelman D C, Jardieu P M, Togias A, McKenzie- White J, Sterbinsky S A, Hamilton R G, Lichtenstein L M. Down-regulation of Fc(epsilon)RI expression on human basophils during in vivo treatment of atopic patients with anti-IgE antibody. J Immunol 1997; 158:1438-1445. [88] MacGlashan D J, Lichtenstein L M, McKenzie W J, Chichester K, Henry A J, Sutton B J, Gould H J. Upregulation of FcepsilonRI on human basophils by IgE antibody is mediated by interaction of IgE with FcepsilonRI. J Allergy Clin Immunol 1999; 104:492-498. [89] Yamaguchi M, Lantz C S, Oettgen H C, Katona I M, Fleming T, Miyajima I, Kinet J P, Galli S J. IgE enhances mouse mast cell Fc(epsilon)RI expression in vitro and in vivo: evidence for a novel amplification mechanism in IgE-dependent reactions. J Exp Med 1997; 185:663-672. [90] Lantz C S, Yamaguchi M, Oettgen H C, Katona I M, Miyajima I, Kinet J P, Galli S J . IgE regulates mouse basophil Fc epsilon RI expression in vivo. J Immunol 1997; 158:2517-2521. [91] Oettgen H C, Geha R S. IgE in asthma and atopy: cellular and molecular connections. J Clin Invest 1999; 104:829-835. [92] Asai K, Kitaura J, Kawakami Y, Yamagata N, Tsai M, Carbone D P, Liu F T, Galli S J , Kawakami T. Regulation of mast cell survival by IgE. Immunity 2001; 14:791-800. [93] Kalesnikoff J, Huber M, Lam V, Damen J E, Zhang J, Siraganian R P, Krystal G. Monomeric IgE stimulates signaling pathways in mast cells that lead to cytokine production and cell survival. Immunity 2001; 14:801-811.

55 [94] Luo H Y, Hofstetter H, Banchereau J, Delespesse G. Cross-linking of CD23 antigen by its natural ligand (IgE) or by anti- CD23 antibody prevents B lymphocyte proliferation and differentiation. J Immunol 1991; 146:2122-2129. [95] Heyman B. IgE-mediated enhancement of antibody responses: the beneficial function of IgE? Allergy 2002; 57:577-585. [96] Sutton B J, Beavil R L, Beavil A J. Inhibition of IgE-receptor interactions. Br Med Bull 2000; 56:1004-1018. [97] Capron M, Capron A. Immunoglobulin E and effector cells in schistosomiasis. Science 1994; 264:1876-1877. [98] Hagan P, Blumenthal U J, Dunn D, Simpson A J G, Wilkins H A. Human IgE, IgG4 and resistance to infection with Schistosoma haematobium. Nature 1991; 349:243-245. [99] Rihet P, Demeure C E, Bourgois A, Prata A, Dessein A J. Evidence for an association between human resistance to Schistosoma mansoni and high anti-larval IgE levels. Eur J Immunol 1991; 21:2679-2686. [100] Dunne D W, Butterworth A E, Fulford A J C, Kariuki H C, Langley J G, Ouma J H, Capron A, Pierce R J, Sturrock R F. Immunity after treatment of human schistosomiasis: association between IgE antibodies to adult worm antigens and resistance to infection. Eur. J. Immunol. 1992; 22:1483-1494. [101] Faulkner H, Turner J, Kamgno J, Pion S D, Boussinesq M, Bradley J E. Age- and infection intensity-dependent cytokine and antibody production in human trichuriasis: the importance of IgE. J Infect Dis 2002; 185:665-672. [102] King C L, Xianli J, Malhotra I, Liu S, Mahmoud A A, Oettgen H C. Mice with a targeted deletion of the IgE gene have increased worm burdens and reduced granulomatous inflammation following primary infection with Schistosoma mansoni. J Immunol 1997; 158:294-300. [103] El Ridi R, Ozaki T, Kamiya H. Schistosoma mansoni infection in IgE-producing and IgE-deficient mice. J Parasitol 1998; 84:171-174. [104] El Ridi R, Ragab S, Lewis S, Afifi A. Role of IgE in primary murine Schistosomiasis mansoni. Scand J Immunol 2001; 53:24-31. [105] Jankovic D, Kullberg M C, Dombrowicz D, Barbieri S, Caspar P, Wynn T A, Paul W E, Cheever A W, Kinet J P, Sher A. Fc epsilonRI-deficient mice infected with Schistosoma mansoni mount normal Th2-type responses while displaying enhanced liver pathology. J Immunol 1997; 159:1868-1875. [106] Gustavsson S, Hjulstrom S, Liu T, Heyman B. CD23/IgE-mediated regulation of the specific antibody response in vivo. J Immunol 1994; 152:4793-4800. [107] Gustavsson S, Wernersson S, Heyman B. Restoration of the antibody response to IgE/antigen complexes in CD23-deficient mice by CD23+ spleen or bone marrow cells. J Immunol 2000; 164:3990-3995. [108] Hellman L. Genes for Immunoglobulin E: Structure, Rearrangement and Expression. Doctoral thesis. ISSN 0345-0058, ISBN 91-554-1721-3, The Department of Microbiology, Uppsala University, Uppsala 1985. [109] Szczeklik A, Jawien J, Stadler B M, Radwan J, Piwowarska W, Dziatkowiak A. Possible relationship between interleukin-6 and response of immunoglobulin E to surgical trauma. Ann N Y Acad Sci 1995; 762:477-479. [110] Szczeklik A, Jawien J. Immunoglobulin E in acute phase response to surgical stress. Clin Exp Allergy 1996; 26:303-307. [111] Szczeklik A, Jawien J. Possible role of IgE in acute-phase response. Allergy 1997; 52:1149-1150. [112] Kampen G T, Poulsen L K, Nielsen H J, Schulze S, Petersen L J. IgE levels in surgery: effect of ranitidine and prednisolone. Allergy 1999; 54:171-176.

56 [113] Navarro-Zorraquino M, Lozano R, Deus J, Pastor C, Larrad L, Tejero E, Roman J , Palacios M J, Torcal J, Salinas J C. Determination of the immunoglobulin E postoperative variation as a measure of surgical injury. World J Surg 2001; 25:585- 591. [114] Szczeklik A. Mast cell, IgE and atherothrombosis. Int Arch Allergy Immunol 1998; 116:166. [115] Szczeklik A. Atopy and sudden cardiac death. Lancet 2000; 355:2254. [116] Smith J K, Krishnaswamy G H, Dykes R, Reynolds S, Berk S L. Clinical manifestations of IgE hypogammaglobulinemia. Ann Allergy Asthma Immunol 1997; 78:313-318. [117] Schoettler J J, Schleissner L A, Heiner D C. Familial IgE deficiency associated with sinopulmonary disease. Chest 1989; 96:516-521. [118] Levy D A, Chen J. Healthy IgE-deficient person. N. Engl. J. Med. 1970; 283:541. [119] Huby R D, Dearman R J, Kimber I. Why are some proteins allergens? Toxicol Sci 2000; 55:235-246. [120] Rogge L, Barberis-Maino L, Biffi M, Passini N, Presky D H, Gubler U, Sinigaglia F. Selective expression of an interleukin-12 receptor component by human T helper 1 cells. J Exp Med 1997; 185:825-831. [121] Szabo S J, Dighe A S, Gubler U, Murphy K M. Regulation of the interleukin (IL)-12R beta 2 subunit expression in developing T helper 1 (Th1) and Th2 cells. J Exp Med 1997; 185:817-824. [122] Smurthwaite L, Walker S N, Wilson D R, Birch D S, Merrett T G, Durham S R, Gould H J. Persistent IgE synthesis in the nasal mucosa of hay fever patients. Eur J Immunol 2001; 31:3422-3431. [123] Davies R J, Rusznak C, Devalia J L. Why is allergy increasing?--environmental factors. Clin Exp Allergy 1998; 28 Suppl 6:8-14. [124] Howarth P H. Is allergy increasing?--early life influences. Clin Exp Allergy 1998; 28 Suppl 6:2-7. [125] Downs S H, Marks G B, Sporik R, Belosouva E G, Car N G, Peat J K. Continued increase in the prevalence of asthma and atopy. Arch Dis Child 2001; 84:20-23. [126] Omran M, Russell G. Continuing increase in respiratory symptoms and atopy in Aberdeen schoolchildren. Bmj 1996; 312:34. [127] Leung D Y. Molecular basis of allergic diseases. Mol Genet Metab 1998; 63:157-167. [128] Ono S J. Molecular genetics of allergic diseases. Annu Rev Immunol 2000; 18:347- 366. [129] Diaz-Sanchez D. Pollution and the immune response: atopic diseases--are we too dirty or too clean? Immunology 2000; 101:11-18. [130] von Mutius E, Weiland S K, Fritzsch C, Duhme H, Keil U. Increasing prevalence of hay fever and atopy among children in Leipzig, East Germany. Lancet 1998; 351:862-866. [131] Cookson W O, Moffatt M F. Asthma: an epidemic in the absence of infection? [comment]. Science 1997; 275:41-42. [132] Yazdanbakhsh M, Kremsner P G, van Ree R. Allergy, parasites, and the hygiene hypothesis. Science 2002; 296:490-494. [133] Barnes P J. Therapeutic strategies for allergic diseases. Nature 1999; 402:B31-38. [134] Kay A B. Overview of 'allergy and allergic diseases: with a view to the future'. Br Med Bull 2000; 56:843-864. [135] Noon L. Prophylactic inoculation against hay fever. Lancet 1911; 1:1572-1573. [136] Larche M. Specific immunotherapy. Br Med Bull 2000; 56:1019-1036.

57 [137] Hellman L. Vaccines Against Allergies. In: Perlmann P, Wigzell H, eds. Handbook of Experimental Pharmacology, Vaccines. Vol. 133: Springer-Verlag, 1999. p. 499- 526. [138] Grunewald S M, Werthmann A, Schnarr B, Klein C E, Brocker E B, Mohrs M, Brombacher F, Sebald W, Duschl A. An antagonistic IL-4 mutant prevents type I allergy in the mouse: inhibition of the IL-4/IL-13 receptor system completely abrogates humoral immune response to allergen and development of allergic symptoms in vivo. J Immunol 1998; 160:4004-4009. [139] Haak F M, Ridgway J, Shields R, Robbins K, Gorman C, Jardieu P. Human IgE receptor alpha-chain IgG chimera blocks passive cutaneous anaphylaxis reaction in vivo. J Immunol 1993; 151:351-358. [140] Vangelista L, Cesco-Gaspere M, Lorenzi R, Burrone O. A minimal receptor-Ig chimera of human FcepsilonRI alpha-chain efficiently binds secretory and membrane IgE. Protein Eng 2002; 15:51-57. [141] Helm B A, Spivey A C, Padlan E A. Peptide blocking of IgE/receptor interaction: possibilities and pitfalls. Allergy 1997; 52:1155-1169. [142] Kolbinger F, Saldanha J, Hardman N, Bendig M M. Humanization of a mouse anti- human IgE antibody: a potential therapeutic for IgE-mediated allergies. Protein Eng 1993; 6:971-980. [143] Presta L G, Lahr S J, Shields R L, Porter J P, Gorman C M, Fendly B M, Jardieu P M. Humanization of an antibody directed against IgE. J. Immunol. 1993; 151:2623- 2632. [144] Chang T W. The pharmacological basis of anti-IgE therapy. Nat Biotechnol 2000; 18:157-162. [145] Babu K S, Arshad S H, Holgate S T. Anti-IgE treatment: an update. Allergy 2001; 56:1121-1128. [146] Davis F M, Gossett L A, Pinkston K L, Liou R S, Sun L K, Kim Y W, Chang N T, Chang T W, Wagner K, Bews J, Brinkmann V, Towbin H, Subramanian N, Heusser C. Can anti-IgE be used to treat allergy? In Springer Semin. Immunopathol. 1993; 15:51-73. [147] Heusser C, Jardieu P. Therapeutic potential of anti-IgE antibodies. Curr Opin Immunol 1997; 9:805-813. [148] Casale T B, Bernstein I L, Busse W W, LaForce C F, Tinkelman D G, Stoltz R R, Dockhorn R J, Reimann J, Su J Q, Fick R J, Adelman D C. Use of an anti-IgE humanized monoclonal antibody in ragweed-induced allergic rhinitis. J Allergy Clin Immunol 1997; 100:110-121. [149] Adelroth E, Rak S, Haahtela T, Aasand G, Rosenhall L, Zetterstrom O, Byrne A, Champain K, Thirlwell J, Cioppa G D, Sandstrom T. Recombinant humanized mAb- E25, an anti-IgE mAb, in birch pollen-induced seasonal allergic rhinitis. J Allergy Clin Immunol 2000; 106:253-259. [150] Casale T B. Anti-immunoglobulin E (omalizumab) therapy in seasonal allergic rhinitis. Am J Respir Crit Care Med 2001; 164:S18-21. [151] Kuehr J, Brauburger J, Zielen S, Schauer U, Kamin W, Von Berg A, Leupold W, Bergmann K C, Rolinck-Werninghaus C, Grave M, Hultsch T, Wahn U. Efficacy of combination treatment with anti-IgE plus specific immunotherapy in polysensitized children and adolescents with seasonal allergic rhinitis. J Allergy Clin Immunol 2002; 109:274-280. [152] Boulet L P, Chapman K R, Cote J, Kalra S, Bhagat R, Swystun V A, Laviolette M, Cleland L D, Deschesnes F, Su J Q, DeVault A, Fick R J, Cockcroft D W. Inhibitory effects of an anti-IgE antibody E25 on allergen-induced early asthmatic response [see comments]. Am J Respir Crit Care Med 1997; 155:1835-1840.

58 [153] Fahy J V, Fleming H E, Wong H H, Liu J T, Su J Q, Reimann J, Fick R J, Boushey H A. The effect of an anti-IgE monoclonal antibody on the early- and late-phase responses to allergen inhalation in asthmatic subjects [see comments]. Am J Respir Crit Care Med 1997; 155:1828-1834. [154] Milgrom H, Fick R J, Su J Q, Reimann J D, Bush R K, Watrous M L, Metzger W J. Treatment of allergic asthma with monoclonal anti-IgE antibody. rhuMAb-E25 Study Group [see comments]. N Engl J Med 1999; 341:1966-1973. [155] Soler M, Matz J, Townley R, Buhl R, O'Brien J, Fox H, Thirlwell J, Gupta N, Della Cioppa G. The anti-IgE antibody omalizumab reduces exacerbations and steroid requirement in allergic asthmatics. Eur Respir J 2001; 18:254-261. [156] Busse W W. Anti-immunoglobulin E (omalizumab) therapy in allergic asthma. Am J Respir Crit Care Med 2001; 164:S12-17. [157] Helm B, Kebo D, Vercelli D, Glovsky M, Gould H, Ishizaka K, Geha R, Ishizaka T. Blocking of passive sensitization of human mast cells and basophil granulocytes with IgE antibodies by a recombinant human ε -chain fragment of 76 amino acids. Proc. Natl. Acad. Sci. USA 1989; 86:9465-9469. [158] Hellman L. Profound reduction in allergen sensitivity following treatment with a novel allergy vaccine. Eur. J. Immunol. 1994; 24:415-420. [159] Hellman L, Carlsson M. Allergy vaccines: A review of developments. Clin. Immunother. 1996; 2:130-142. [160] Hellman L. Is vaccination against IgE possible? Adv Exp Med Biol 1996; 409:337- 342. [161] Sayers I, Cain S A, Swan J R, Pickett M A, Watt P J, Holgate S T, Padlan E A, Schuck P, Helm B A. Amino acid residues that influence Fc epsilon RI-mediated effector functions of human immunoglobulin E. Biochemistry 1998; 37:16152-16164. [162] Da Luz J. Aspects of N-Glycosylation in Human IgE. Doctoral thesis. ISBN 91-7349- 130-6. Department of Medicin, Karolinska Institutet, Stockholn 2002. [163] Conrad D H, Wingard J R, Ishizaka T. The interaction of human and rodent IgE with the human basophil IgE receptor. J Immunol 1983; 130:327-333. [164] Kulczycki A, Jr., Isersky C, Metzger H. The interaction of IgE with rat basophilic leukemia cells. I. Evidence for specific binding of IgE. J Exp Med 1974; 139:600- 616. [165] Lowenthal M, Patterson R, Harris K E. Passive transfer of IgE-mediated cutaneous reactivity in heterologous species. Ann Allergy 1993; 71:481-484. [166] Sachs D H. The pig as a potential xenograft donor. Vet Immunol Immunopathol 1994; 43:185-191. [167] Cooper D K, Gollackner B, Sachs D H. Will the pig solve the transplantation backlog? Annu Rev Med 2002; 53:133-147. [168] Fornhem C, Peterson C G, Dahlback M, Scheynius A, Alving K. Granulocyte function in the airways of allergen-challenged pigs: effects of inhaled and systemic budesonide. Clin Exp Allergy 1996; 26:1436-1448. [169] Fornhem C, Dahlback M, Kumlin M, Lundberg J M, Alving K. Effects of local and systemic budesonide on allergen-induced airway reactions in the pig. Br J Pharmacol 1996; 118:989-997. [170] Fornhem C, Peterson C G, Scheynius A, Alving K. Influence of endogenous cortisol on eosinophil function in sensitized pigs: direct measurements of eosinophil peroxidase. Clin Exp Allergy 1996; 26:469-478. [171] Turner H, Kinet J P. Signalling through the high-affinity IgE receptor Fc epsilonRI. Nature 1999; 402:B24-30.

59 [172] Ujike A, Ishikawa Y, Ono M, Yuasa T, Yoshino T, Fukumoto M, Ravetch J V, Takai T. Modulation of immunoglobulin (Ig)E-mediated systemic anaphylaxis by low-affinity Fc receptors for IgG. J Exp Med 1999; 189:1573-1579. [173] Young R J, Owens R J, Mackay G A, Chan C M, Shi J, Hide M, Francis D M, Henry A J, Sutton B J, Gould H J. Secretion of recombinant human IgE-Fc by mammalian cells and biological activity of glycosylation site mutants. Protein Eng 1995; 8:193-199. [174] Dalum I, Butler D M, Jensen M R, Hindersson P, Steinaa L, Waterston A M, Grell S N, Feldmann M, Elsner H I, Mouritsen S. Therapeutic antibodies elicited by immunization against TNF-alpha. Nat Biotechnol 1999; 17:666-669. [175] Talwar G P. Vaccines and passive immunological approaches for the control of fertility and hormone-dependent cancers. Immunol Rev 1999; 171:173-192. [176] Chackerian B, Lowy D R, Schiller J T. Conjugation of a self-antigen to papillomavirus-like particles allows for efficient induction of protective autoantibodies. J Clin Invest 2001; 108:415-423. [177] Gupta R K. Aluminum compounds as vaccine adjuvants. Adv Drug Deliv Rev 1998; 32:155-172. [178] Jerome Aucouturier L D, S Deville, S Ascarateil, V Ganne. Montanide ISA 720 and 51: a new generation of water in oil emulsions as adjuvants for human vaccines. Expert Rev. Vaccines 2002; 1:111-118. [179] van Driel W J, Ressing M E, Kenter G G, Brandt R M, Krul E J, van Rossum A B, Schuuring E, Offringa R, Bauknecht T, Tamm-Hermelink A, van Dam P A, Fleuren G J , Kast W M, Melief C J, Trimbos J B. Vaccination with HPV16 peptides of patients with advanced cervical carcinoma: clinical evaluation of a phase I-II trial. Eur J Cancer 1999; 35:946-952. [180] Slingluff C L, Jr., Yamshchikov G, Neese P, Galavotti H, Eastham S, Engelhard V H, Kittlesen D, Deacon D, Hibbitts S, Grosh W W, Petroni G, Cohen R, Wiernasz C, Patterson J W, Conway B P, Ross W G. Phase I trial of a melanoma vaccine with gp100(280-288) peptide and tetanus helper peptide in adjuvant: immunologic and clinical outcomes. Clin Cancer Res 2001; 7:3012-3024.

60 Here is a fact that should help you to fight a bit longer: Things that don’t act ually kill you outright make you stronger

Grooks by Piet Hein

61