Biological activity of IgE specific for cross-reactive carbohydrate determinants

Biological activity of IgE specific for cross-reactive carbohydrate determinants

Biological activity of IgE specific for cross-reactive carbohydrate determinants Kay Foetisch, PhD,a Sandra Westphal, PhD,a Iris Lauer, PhD,a Mechthil...

514KB Sizes 0 Downloads 24 Views

Biological activity of IgE specific for cross-reactive carbohydrate determinants Kay Foetisch, PhD,a Sandra Westphal, PhD,a Iris Lauer, PhD,a Mechthild Retzek,a Fritz Altmann, PhD,b Daniel Kolarich, PhD,b Stephan Scheurer, PhD,a and Stefan Vieths, PhDa Langen, Germany, and Vienna, Austria

From athe Department of Allergology, Paul-Ehrlich-Institut, Langen; and bthe Institute of Chemistry, University of Agriculture, Vienna. Supported in part by a grant from the Deutsche Forschungsgemeinschaft, DFG SCHE-637/1-1, and by the Austrian Science Fund (FSP S8803). Received for publication November 5, 2002; revised December 5, 2002; accepted for publication December 11, 2002. Reprint requests: Stefan Vieths, PhD, Paul-Ehrlich-Institut, Department of Allergology, Paul-Ehrlich Str. 51-59, D-63225, Langen, Germany. © 2003 Mosby, Inc. All rights reserved. 0091-6749/2003 $30.00 + 0 doi:10.1067/mai.2003.173

Key words: Cross-reactive carbohydrate determinant, IgE reactivity, histamine release, clinical relevance

The clinical relevance of IgE to cross-reactive carbohydrate determinants (CCDs) is still controversial. Because of the widespread occurrence of structurally similar CCDs in plants and invertebrates, IgE to carbohydrates often leads to false-positive diagnostic results when only the IgE binding to extracts or allergens (CAP, RAST, EAST, and immunoblot) in relation to the initiated symptoms is considered.1 Many glycoproteins, such as bromelain, which only carry one IgE-binding glycan, are unable to cross-link IgE bound to the receptors of mast cells and basophils, and clinical symptoms might not appear in patients whose IgE response is restricted to CCDs. On the basis of this view, several authors regard anti-CCD IgE as specific antibodies without clinical relevance.1-3 However, there are also many glycoproteins with more than one N-linked glycan4 (eg, horseradish peroxidase [HRP]) or 3 potential glycoprotein allergens (eg, β-fructofuranosidase, polygalacturonase 2A and pectinesterase) that we have recently identified in tomato fruit extract.5,6 Until now, studies to prove whether or not natural multivalent glycoproteins might be able to induce histamine release and thus might also contribute to clinical symptoms of the allergic patients are very rare,4 most likely due to difficulties in the experimental design: first, the lack of suitable purified multivalent allergenic glycoproteins, and second, the need of fresh blood from donors sensitized to these glycoproteins for testing in histamine release assay. In most cases only the sera but no cells of these patients are available. Therefore the present study was aimed at investigating the allergenic activity of anti-CCD IgE by using defined glycoproteins with known composition of their N-linked glycans in histamine release tests. For this purpose, food allergy to tomato was used as a model. In a previous study, 28 (35.9%) of 78 sera from patients with tomato allergy showed IgE to high-molecular-weight bands in immunoblots of tomato extract.5 This percentage of CCD-positive sera from patients with tomato allergy was also confirmed by a glycan ELISA used in an earlier study, with 11 (35.5%) of 31 patients having positive bromelain-type-glycopeptide Manα1-6(Xylβ1-2)Man β1-4GlcNAcβ1-4(Fucα1-3)GlcNAc (MUXF-glycopeptide) results.6 Ten of these sera were used to resensitize stripped basophils from normal donors according to the 889

Food and drug reactions and anaphylaxis

Background: The clinical relevance of IgE specific for crossreactive carbohydrate determinants (CCDs) has been a matter of controversy. Until now, no convincing experiments have been performed to test the biologic significance of individual multivalent allergens that carry multiple carbohydrate epitopes. Objective: We sought to contribute to the understanding of the role of CCD-specific IgE antibodies and to study whether CCD-specific IgE antibodies are able to activate basophils using different multivalent glycoproteins as antigens. Methods: Purified natural tomato β-fructofuranosidase (nLyc e 2) and rLyc e 2.02 expressed in Escherichia coli were compared by means of histamine release tests. In addition, native and deglycosylated horseradish peroxidase and a neoglycoprotein consisting of a defined glycopeptide (Manα1-6[Xylβ12]Manβ1-4GlcNAcβ1-4[Fucα1-3]GlcNAc) coupled to BSA were used in histamine release assays using stripped basophils from normal donors resensitized with IgE from CCD-reactive patients with food allergy to tomato. Results: Ten CCD-positive and 2 CCD-negative sera from patients with tomato allergy underwent histamine release testing by using the glycoproteins and nonglycosylated controls as antigens, respectively. All sera induced histamine release with tomato extract (up to 100%), confirming the allergic status of the donors. Four of the CCD-positive sera induced releases ranging from 12% to 82% with all of the glycoproteins but not with the nonglycosylated or monovalent controls. All other sera showed no response or only very weak response to the glycoproteins. Conclusion: Approximately one third of the CCD-positive sera from patients with tomato allergy have biologically relevant CCD-specific IgE antibodies. Therefore the general claim that CCD-specific IgE is clinically irrelevant has to be reconsidered critically. Hence IgE specific for CCDs should be taken into consideration in the diagnosis and therapy of certain allergies. In the subgroup of patients sensitized to CCDs, the use of natural allergens should be preferred over the use of recombinant allergens expressed in prokaryotic organisms. (J Allergy Clin Immunol 2003;111:889-96.)

890 Foetisch et al

Abbreviations used CCD: Cross-reactive carbohydrate determinant HR25: Concentration required to stimulate 25% histamine release HRP: Horseradish peroxidase LTP: Lipid transfer protein MMXF: (Manα1-3)Manα1-6(Xylβ1-2)Manβ1-4GlcNAcβ14(Fucα1-3)GlcNAc MUXF: Manα1-6(Xylβ1-2)Manβ1-4GlcNAcβ1-4 (Fucα1-3)GlcNAc

method of Kleine-Budde et al.7 We were also able to demonstrate that the IgE-binding capacity of β-fructofuranosidase (Lyc e 2) from tomato almost exclusively depends on the presence of the N-glycans.5 Therefore we selected this molecule as a glycosylated model antigen. The native form of β-fructofuranosidase from tomato was purified and used in histamine release tests in comparison with the recombinant form expressed in Escherichia coli. In addition, both native HRP and the deglycosylated form of HRP, as well as a neoglycoprotein consisting of a defined glycopeptide (MUXF-BSA) coupled to BSA, were used as model glycoallergens in these histamine release tests to confirm the results obtained with the natural glycoproteins.

J ALLERGY CLIN IMMUNOL APRIL 2003

(peroxidase-CM) was prepared by means of S-alkylation with iodoacetic acid using standard procedures. The deglycosylation of peroxidase was performed with the GlycoFree deglycosylation kit K-500 (Oxford GlycoSystems, Abingdon, United Kingdom). The natural β-fructofuranosidase (nLyc e 2) was purified from tomato extract by a 2-step purification method with hydrophobic interaction chromatography, followed by SDS-PAGE and electroelution. The identity of the protein was confirmed by N-terminal sequencing, and the IgE reactivity was verified by immunoblot analysis and ELISA. Recombinant β-fructofuranosidase (rLyc e 2.02) was expressed as His-tagged fusion protein in E coli and purified as previously described.5,12 The purity of all glycoproteins and their nonglycosylated controls was proved by SDS-PAGE (Fig 1). All extracts and allergens were stored freeze-dried at –20°C until used.

N-glycan analysis From the cDNA sequence, it was known that fructofuranosidase contains 4 potential N-glycosylation sites. Glycan analysis with mass spectrometry showed13 that nLyc e 2 carries approximately 92% structures containing α-linked fucose and β-linked xylose and approximately 8% structures with only xylose. The main structure was found to be the glycan (Manα1-3)Manα1-6(Xylβ1-2)Manβ14GlcNAcβ1-4(Fucα1-3)GlcNAc (MMXF; 84%), whereas MUXF occurred with only approximately 5%. In HRP no fewer than 6 of 8 glycosylation sites were identified, all of which contained MMXF glycans.13 The neoglycoprotein consisting of a defined glycopeptide (MUXF) coupled to BSA was found to contain 9 mol of glycopeptide per mole of protein (Table II).

Determination of specific IgE METHODS Human sera and antibodies Sera were obtained from allergic donors of the Paul-EhrlichInstitut, Langen, Germany; the Hospital Borkum Riff, Department of Dermatology and Allergology, Borkum, Germany (Dr H. Aulepp); and the Hospital Universitario Doce de Octubre, Madrid, Spain (Dr J. F. Crespo). Ten sera of patients allergic to tomato were chosen according to their positive reactivity to defined glycopeptides in a glycan ELISA (see below and Table I). Furthermore, 2 sera with multiple serologic sensitizations (including tomato) and high IgE titers that also showed strong IgE reactivity in the glycan ELISA were included in the study (sera received from MAST DIAGNOSTICA, Reinfeld, Germany). Further clinical data of the donors were not available. Two sera of patients allergic to tomato but without reactivity to CCD and one serum from a nonallergic patient served as controls, respectively (Table I). In addition, a rabbit serum against the recombinant pear profilin Pyr c 4 was used.

Preparation and purification of the N-glycans Food and drug reactions and anaphylaxis

An N-linked glycopeptide with the glycan structure MUXF was prepared from pineapple stem bromelain, and a glycopeptide containing the glycan Manα1-6(Manα1-3)Manβ1-4GlcNAcβ1-4GlcNAc was prepared from bovine fibrin, as previously described.8,9 Both glycopeptides were coupled to BSA using 1,5-difluoro-2,4dinitrobenzene.9 The identity and purity of the glycopeptides and their BSA conjugates were verified, as previously described (Fig 1).10 The glycopeptides and their BSA conjugates were stored freeze-dried at –20°C until used.

Allergens and extracts Tomato extract was prepared by a low-temperature extraction protocol, as described for apple.11 Inactive carboxymethylated HRP

Measurement of allergen-specific IgE was performed with the CAP FEIA system (Pharmacia Diagnostics, Uppsala, Sweden), according to the manufacturer’s instructions. IgE specific for carbohydrate structures was assayed with a glycan ELISA, as previously described.10 The same ELISA format was used with 250 ng per well of nLyc e 2 instead of 1 µg/mL glycopeptide as the solid-phase antigen, and bound IgE was detected with HRP-labeled instead of alkaline phosphatase–labeled anti-IgE antibody. The absorption values of the sera were considered positive if the corrected values (absorption – absorption of the negative control) were 2-fold higher than those of the negative control.

Electrophoresis and immunoblotting SDS-PAGE of tomato extract (20 µg/cm) and nLyc e 2 (0.4 µg/cm) was performed as previously described.14 After transferring the separated glycoproteins to nitrocellulose membranes by means of semidry blotting, the membranes were probed with sera from the allergic patients or from a nonallergic control subject (all diluted 1:10). Immunostaining of bound IgE antibodies was performed with alkaline phosphatase coupled to mouse anti-human IgE (1:1000, 3 hours; Pharmingen, Hamburg, Germany) and 4-nitroblue tetrazolium chloride and 5-bromo-4-chloro-3-indolylphosphate as the substrate. For inhibition of IgE binding, a 1:10 diluted serum pool (n = 3) was preincubated with 20 µg of the purified glycopeptides or 100 µg of allergen extract before incubation of the blot strips.

Basophil histamine release The histamine release was performed as described elsewhere7 with several modifications. The conditions for stripping of the nonspecific IgE and for the passive sensitization procedure were chosen according to the recommendations of Pruzansky et al.15 Cells sensitized with serum of a nonallergic subject served as negative controls. The stimulation of the cells was performed according to the

Foetisch et al 891

J ALLERGY CLIN IMMUNOL VOLUME 111, NUMBER 4

FIG 1. Purity control of the used glycoproteins by SDS-PAGE and Coomassie staining: lanes 1 and 10 (marker; molecular weights of the marker proteins are indicated on the left); lane 2, tomato extract; lane 3, nLyc e 2; lane 4, rLyc e 2.02 (6× His-tagged fusion protein); lane 5, HRP; lane 6, deglycosylated HRP; lane 7, MUXFglycopeptide; lane 8, MUXF-BSA; and lane 9, BSA. Sample load: tomato extract, 2.5 µg; purified proteins, 1 µg; MUXF-glycopeptide, 4 µg.

TABLE I. Patient data5,6,12 Serum no.

1 2 3 4 5 6 7 8 9 10 MAST1 MAST2 11 12 13

Symptoms of tomato allergy

OAS, atopic dermatitis Urticaria OAS OAS OAS Abdominal pain OAS, anaphylaxis Conjunctivitis Urticaria OAS Not known Not known OAS OAS Nonallergic

SPT tomato

CAP tomato

ELISA nLyc e 2*

CCD ELISA*

Blot CCD

Blot nLyc e 2

Blot profilin

Blot LTP

ND (+) +++ + ND ND + + + + ND ND ND ND ND

4 3 4 3 3 3 4 3 4 5 3 5 3 2 0

1.24 1.77 2.22 1.68 1.52 1.34 0.57 0.41 0.4 ND 2.19 2.17 – ND –

1.2 1.3 2.6 0.6 0.5 2.0 3.2 1.7 2.3 0.43 1.8 3.6 – – –

+ + + + + + + + + + + + – – –

+ + + + + + + (+) + + + + – – –

– – – + + – (+) – – + (+) + + – –

– – – – – – + + + + – – + + –

SPT, Skin prick test; ND, not done; OAS, oral allergy syndrome; MAST, serum from MAST Diagnostica. *Considered positive when 0.4 or greater (2-fold of the negative control).

RESULTS

TABLE II. Characteristics of the used glycoproteins: Composition and number of the N-linked glycans Glycan type*

MMXF MUXF GnMXF/MGnXF GnGnXF MMX No. of glycans

HRP

nLyc e 2

100%

83.6% 5.3% 2.3% 0.6% 8.2% >1

6

*Structures depicted in Kolarich

MUXF-BSA

100%

9

and Altmann.13

Patient data and sensitization profiles All sera of allergic subjects showed specific IgE to tomato in the CAP system (CAP class 2-5) and IgE reactivity by means of immunoblot analysis. The IgE antibodies of the 12 CCD-positive sera recognized high-mo-

lecular-weight bands in tomato extract and also the band of the isolated natural (glycosylated) form of the tomato fructofuranosidase nLyc e 2.5,6,12 Blot inhibition studies with a serum pool (n = 3) showed specific and almost

Food and drug reactions and anaphylaxis

instructions of the manufacturer of the histamine kit (Immunotech, Marseille, France), with 10-fold dilutions of the allergens starting at 10 µg/mL. Histamine release was measured by a competitive enzyme immunoassay (Immunotech). After subtraction of the spontaneous release, the allergen-induced histamine release was calculated as a percentage of the total amount of histamine determined after lysis of the basophils by twice freezing and thawing of the cells. A histamine release of greater than 10% was considered positive. All experiments were done at least twice on separate days.

892 Foetisch et al

J ALLERGY CLIN IMMUNOL APRIL 2003

FIG 2. IgE binding of a serum pool (n = 3) to tomato extract proteins (20 µg/cm; A) or to purified natural tomato fructofuranosidase (0.4 µg/cm; B) separated by SDS-PAGE and transferred to nitrocellulose membranes and blot inhibition studies with different inhibitors. The numbers on the left correspond to the molecular weight of the marker proteins in kilodaltons. Lane 1, serum pool (n = 3); lane 2, serum pool plus tomato extract (100 µg); lane 3, serum pool plus MUXF-glycopeptide (20 µg); lane 4, serum pool plus Manα1-6(Manα1-3)Manβ1-4GlcNAcβ1-4GlcNAc glycopeptide (20 µg); lane 5, serum pool plus extract of lowfat milk (100 µg); and lane 6, nonallergic control serum.

Food and drug reactions and anaphylaxis

complete inhibitions of the IgE reactivity to tomato extract proteins (Fig 2, A) and nLyc e 2 (Fig 2, B) after preincubation with tomato extract (100 µg; Fig 2, lanes 2) and with the MUXF-glycopeptide (20 µg; Fig 2, lanes 3) but no inhibitions with the Manα1-6(Manα13)Manβ1-4GlcNAcβ1-4GlcNAc glycopeptide (20 µg [without fucose and xylose]; Fig 2, lanes 4) and with 100 µg of milk extract (Fig 2, lanes 5). Furthermore, all CCD-positive sera also reacted in a CovaLink IgE ELISA with covalently coupled MUXFglycopeptide isolated from bromelain and with the covalently linked nLyc e 2. In contrast, the IgE of the 2 CCDnegative sera bound to distinct bands on the tomato blots and showed no reactivity to the glycopeptide or nLyc e 2 in immunoblot analysis and ELISA, respectively. In addition, nearly 50% of the 14 sera of patients with tomato allergy also showed IgE binding to other tomato allergens, either to the putative tomato profilin (7/14), the putative lipid transfer protein (LTP; 6/14, data not shown), or both. The identity of these proteins was verified by immunoblot analysis using human serum 11, which was already known to be IgE reactive to the cherry LTP Pru av 3 and a rabbit serum against the recombinant pear profilin Pyr c 4. Furthermore, the IgE antibody reactivity of some of these sera to purified cherry LTP and tomato profilin confirmed the identity of these bands (data not shown). The serum of a nonallergic control subject did not show any unspecific reactivity (Table I).

Histamine release indicates biologic activity of anti-CCD IgE in a subgroup of patients with tomato allergy The purified multivalent glycoprotein from tomato and other glycoproteins, such as HRP and MUXF-BSA, were used as tools to investigate to which extent IgE specific for CCD is biologically active. Because tomato contains a high amount of endogenous histamine, the allergen extract was dialyzed extensively, resulting in a reduction of the concentration of histamine to one sixth of the crude preparation. However, all values had to be corrected for the residual amount of tomato histamine per concentration that was measured simultaneously. Table III shows the maximum histamine release results of all tested sera achieved with tomato extract and with the glycoproteins. The IgE of all 14 sera from patients with tomato allergy was able to induce histamine release up to 100% with tomato extract used as the antigen. In 5 of the 12 sera with IgE specific for CCD (sera from the 4 patients with tomato allergy labeled 1 to 4 and serum 1 from MAST Diagnostica), all 3 glycoproteins released histamine from the passively sensitized basophils in a dose-dependent manner. In contrast, the deglycosylated HRP and the recombinant Lyc e 2.02, BSA, and the monovalent MUXF-glycopeptide induced no release with these sera. Fig 3 shows histamine release plots for 3 CCD-positive sera and one CCD-negative serum from a patient with tomato allergy.

Foetisch et al 893

Food and drug reactions and anaphylaxis

J ALLERGY CLIN IMMUNOL VOLUME 111, NUMBER 4

FIG 3. Histamine release results of 3 CCD-positive sera (nos. 1-3) and one CCD-negative serum (no. 12) from patients with tomato allergy. The left plot contains the response curves with tomato extract and tomato allergens (serum 1), and the right plot contains the curves obtained with HRP, MUXF-BSA, and the control proteins, respectively (serum 2). The error bars represent the SEM of single measurements repeated on different days (n = 2 for all allergens, except for nLyc e 2 and HRP for which n = 3).

894 Foetisch et al

J ALLERGY CLIN IMMUNOL APRIL 2003

TABLE III. Histamine release results IgE reactivity to: Serum No.

1 2 3 4 5 6 7 8 9 10 MAST1 MAST2 11 12 13

Tomato

CCD

+ + + + + + + + + + + + + + –

+ + + + + + + + + + + + – – –

Maximum histamine release results (%) with: Tomato

38.3 39.2 89.9 86.1 44.7 43.9 59.3 88.1 31.1 72.6 58.4 100 74.3 38.1 4.0

nLyc e 2

rLyc e 2.02

39.2 44.1 57.9 21.2 6.8 3.2 4.4 4.8 4.2 6.9 15.9 0 5.2 1.1 2.0

4.2 5.6 7.5 8.3 ND 5.0 2.7 4.8 2.2 5.5 0.6 1.3 0 0.6 ND

HRP

38.2 11.7 81.6 13.9 0.9 23.8 4.8 7.7 3.3 10.5 39 2.5 0 0 2.8

dHRP

8.1 6.9 1.2 8.5 ND 20.6 0 6.6 0.7 7.2 0 0 2.8 2.9 ND

MUXF-BSA

MUXF-GP

BSA

42.5 24.7 43.6 14.8 14.8 14.9 15.8 9.5 5.8 7.8 46 11.9 0 0 2.5

0.5 7.2 2.1 0.5 0.4 2.1 0 1.7 1.1 1.8 2.9 ND 3.6 1.5 2.1

0 5.2 2.6 7.6 0 0 5.8 3.5 2.4 0.1 1.7 0 0 1.3 2.6

GP, Glycoprotein; ND, not done; MAST, serum from MAST Diagnostica.

TABLE IV. HR25 concentrations of tomato extract and the glycoproteins Serum no.

1 2 3 4 MAST1 11 12

Tomato

nLyc e 2

4 3 0.1 0.02 1 0.03 3

1 3 0.1 5 –§ – –

HRP

0.2 –* 0.006 –† 0.1 – –

MUXF-BSA

0.5 0.05 0.01 –‡ 0.04 – –

these remaining 7 sera with positive CCD reactivity responded weakly to the MUXF-BSA conjugate, but neither responded to tomato fructofuranosidase, HRP, or unconjugated MUXF. The IgE of the 2 sera from patients with tomato allergy and CCD-negative control sera reacted only with tomato extract but not with any of the other antigens (Fig 3, serum 12). There was no antigen-stimulated significant histamine release with the nonallergic control serum (Table III).

DISCUSSION Values are given as micrograms per milliliter. MAST, serum from MAST Diagnostica. *11.7% at 0.05 µg/mL. †13.9% at 0.1 µg/mL. ‡14.8% at 0.1 µg/mL. §15.9% at 0.2 µg/mL. CCD-negative sera, no HR25 value.

Food and drug reactions and anaphylaxis

For a better comparison of the cell sensitivity of the basophils, the concentration required to stimulate 25% histamine release (HR25) was estimated from the release plots. Table IV summarizes the HR25 concentrations of the allergens for the 5 abovementioned sera showing histamine release after CCD stimulation and for the 2 CCDnegative sera. With the exception of 3 sera (nos. 3, 4, and 11 with HR25 ≤0.1 µg/mL), relatively high HR25 concentrations of tomato extract (1-4 µg/mL) were required. With these sera, similar HR25 concentrations of nLyc e 2 were obtained, with HR25 values reaching from 0.1 µg/mL (serum 3) to 5 µg/mL (serum 4). In contrast, drastically lower HR25 amounts (factor 10 to factor 60) of the multivalent glycoproteins (HRP and MUXF-BSA) were calculated for the sera from patients with positive histamine release values (Fig 3 and Table IV). Despite their positive reactivity to CCDs in blots and ELISAs, all other sera showed no significant response to the glycoproteins. The histamine release of serum 6 with HRP is probably caused by the recognition of protein epitopes because a comparable release was obtained with the deglycosylated HRP (Table III). The IgE of some of

Recently, the first systematic attempt to show that carbohydrate chains of natural glycoproteins other than HRP are able to induce clinical symptoms by cross-linking of IgE was made by comparing the histamine release achieved with the natural and recombinant forms of Cup a 1, a newly discovered glycoprotein allergen of cypress pollen.16 Only the native glycosylated molecule induced histamine release from basophils. Similar to the design of this study, we sought to investigate whether multivalent glycoproteins are able to induce histamine release. Recently, we detected a potential glycoprotein allergen, β-fructofuranosidase, in tomato fruit extract.6,12 Subsequently, this glycoprotein was purified, its glycan structure was analyzed,5 and natural β-fructofuranosidase (nLyc e 2) was selected as a model allergen for this study. In the cell wall β-fructofuranosidase from carrot, 3 of the 6 potential N-glycosylation sites are occupied by 2 complex-type glycans and one high-mannose-type glycan.17 The β-fructofuranosidase from tomato also carries 4 potential N-glycosylation sites. One glycan (MMXF) was found at amino acid position 52 of the mature protein. Until now, the remaining 3 potential N-glycosylation sites were not identified in the mass spectra,5 but due to the results of mediator release experiments presented in this study, the existence of at least one additional glycan can be deduced.

When tomato extract was used as cross-linking antigen, all sera from 12 patients with tomato allergy and 2 patients with multiple sensitizations, but not the serum from the nonallergic control subject, released histamine from passively sensitized basophils in a dose-dependent manner. This confirms, on the one hand, the allergic condition of these patients and, on the other hand, the specificity of the used test method. Furthermore, 5 (41.7%) of 12 CCD-positive sera showed specific releases of up to 82% with all 3 glycoproteins but not with the corresponding controls (Fig 3), indicating the biological activity of the CCD-specific IgE of these 5 sera. Four of the 5 donors of these sera had a food allergy to tomato, with symptoms ranging from oral allergy syndrome to urticaria. One reason for the clinical effects of tomato on these patients could be the IgE reactivity with the native tomato fructofuranosidase. Because none of the 4 patient sera reacted to the recombinant protein (rLyc e 2.02) in either the ELISA (results not shown) or in histamine release (Fig 3), the involvement of protein epitopes was excluded. Thus the reactivity to nLyc e 2 (and to HRP and MUXF-BSA) is exclusively caused by the carbohydrate moiety of the glycoproteins. Therefore it is likely that the CCD-specific IgE is also responsible for the reported symptoms and, consequently, clinically relevant in these patients. However, as depicted in Table IV, differences were observed among these 5 sera. In serum 4 the IgE response to CCDs was not sufficient to explain the strong histamine release with tomato extract. Because this serum showed IgE binding to tomato profilin on tomato extract immunoblots, this would explain the difference in the release results between tomato extract and the glycoproteins. In the other 4 glycoprotein-reactive sera, the histamine release stimulated by nLyc e 2 was similar to that stimulated by tomato extract and even stronger (10- to 60-fold on the basis of HR25) than that stimulated by MUXF-BSA and HRP (Fig 3 and Table IV). This finding might indicate the effect of the number of glycans attached to multivalent glycoproteins for a more efficient cross-linking of the receptorbound IgE. Hence, N-glycan–mediated IgE binding to nLyc e 2 nearly accounted for the whole biological activity of tomato extract in 3 of 10 CCD-positive sera from patients with tomato allergy. By contrast, glycoproteins and control proteins induced no or only weak histamine release with the remaining 7 CCD-positive sera (Table III), indicating low biologic activity of anti-CCD IgE in these subjects. Various reasons might be discussed to explain the observed differences in the biological activity of IgE antibodies directed against CCDs. First, in addition to CCD-bearing glycoproteins, tomato extract contains several other nonglycosylated potential allergens. As indicated by immunoblotting, approximately 50% of the 14 sera from patients with tomato allergy also had IgE to tomato profilin (7/14) and to LTP (6/14). Furthermore, serum 6 probably recognizes protein epitopes of peroxidase (Table III), a glyco-

Foetisch et al 895

protein that is also expressed in tomatoes.18 Only for 3 sera (sera 1-3), all of which showed positive histamine release on stimulation with all glycoproteins, was no other tomato-specific IgE than the CCD-specific IgE detectable. Thus the ratio of the level of CCD-specific IgE to that of total allergen-specific IgE or total IgE in the serum used for passive sensitization might be important.19 During the passive sensitization procedure, the CCD-specific IgE competes with the residual IgE for the limited number of free IgE receptors on the surface of the stripped basophils from the nonallergic donor. Even if some CCD-specific IgE is bound to the receptors, the IgE density might not be high enough for cross-linking of the specific antigen. Thus a high level of CCD-specific IgE (as in serum 3) would possibly favor the release with the glycoproteins. Second, the affinity and specificity of the IgE might play a role. Recently, a strong correlation was found between the affinity of the IgE for its antigen and the sensitivity of the histamine release. Therefore the IgE antibody affinities to Der p 2 varied approximately 30-fold among 21 patients with mite allergy.20 Similarly, the CCD-specific IgE of the 5 patients with positive release to the glycoproteins might have higher affinities to CCDs than the IgE of the other patients and would therefore trigger a stronger histamine release. Furthermore, the specificity of the IgE antibodies might also play a role. The additional mannose residue in the glycans of nLyc e 2 and HRP has been suggested to inhibit the binding of some IgE to the thus masked xylose residue in the MMXF structure, which is not true for the MUXF structure in the MUXF-BSA conjugate.21 This would explain the lack of responses to HRP and nLyc e 2 and the better reactivity of several of the CCD-positive sera to MUXF-BSA. Because this conjugate is an artificial product with unknown distribution of the binding sites, it is tempting to speculate that a native glycoprotein with multivalent MUXF glycans would possibly lead to more efficient histamine release, even if the correct distance between the glycans might be of importance for cross-linking of the IgE. Nevertheless, our data demonstrate that CCD-specific IgE of approximately one third of the patients with (tomato) allergy investigated in this study is biologically active and mediates histamine release of up to 82% if stimulated with different glycoproteins. In conclusion, the general claim that CCD-specific IgE has low biologic activity3 has to be reconsidered critically. Our results indicate that detailed investigations are required on an individual case basis. IgE specific for CCD might also be taken into consideration in the diagnosis and therapy of certain allergies. In patients with a clinically relevant CCD sensitization, the use of native or recombinant allergens produced in plant expression systems that are able to generate the specific N-glycans should be preferred over the use of recombinant allergens produced in E coli. We thank Dr Gerald Reese for kindly reviewing the manuscript. We thank Ingrid Weissmann for help with the neoglycoprotein synthesis.

Food and drug reactions and anaphylaxis

J ALLERGY CLIN IMMUNOL VOLUME 111, NUMBER 4

896 Foetisch et al

J ALLERGY CLIN IMMUNOL APRIL 2003

REFERENCES 1. Mari A, Iacovacci P, Afferni C, Barletta B, Tinghino R, Di Felice G, et al. Specific IgE to cross-reactive carbohydrate determinants strongly affect the in vitro diagnosis of allergenic diseases. J Allergy Clin Immunol 1999;103:1005-11. 2. Van der Veen MJ, van Ree R, Aalberse RC, Akkerdaas J, Koppelman SJ, Jansen HM, et al. Poor biologic activity of cross-reactive IgE directed to carbohydrate determinants of glycoproteins. J Allergy Clin Immunol 1997;100:327-34. 3. Van Ree R. Specific IgE without clinical allergy. J Allergy Clin Immunol 1999;103:1000-1. 4. Foetisch K, Vieths S. N- and O-linked oligosaccharides of allergenic glycoproteins. Glycoconj J 2001;18:373-90. 5. Westphal S, Kolarich D, Foetisch K, Lauer I, Altmann F, Conti A, et al. Molecular characterization and allergenic activity of Lyc e 2 (β-fructofuranosidase), a glycosylated allergen of tomato. Eur J Biochem 2003;270:1327-37. 6. Foetisch K, Son DY, Altmann F, Aulepp H, Conti A, Haustein D, et al. Tomato (Lycopersicon esculentum) allergens in pollen-allergic patients. Eur Food Res Technol 2001;213:259-66. 7. Kleine-Budde I, de Heer PG, van der Zee JS, Aalberse RC. The stripped basophil histamine release bioassay as a tool for the detection of allergenspecific IgE in serum. Int Arch Allergy Immunol 2001;126:277-85. 8. Tretter V, Altmann F, März L. Peptide-N4-(N-acetyl-β-glucosaminyl) asparaginase amidase F cannot release glycans with fucose attached α1→3 to the asparagine-linked N-acetyl-glucosamine residue. Eur J Biochem 1991;199:647-52. 9. Wilson IBH, Harthill JE, Mullin NP, Ashford DA, Altmann F. Core α1,3fucose is key part of the epitope recognized by antibodies reacting against plant N-linked oligosaccharides and is present in a wide variety of plant extracts. Glycobiology 1998;8:651-61. 10. Foetisch K, Altmann F, Haustein D, Vieths S. Involvement of carbohydrate epitopes in the IgE response of celery-allergic patients. Int Arch Allergy Clin Immunol 1999;120:30-42. 11. Vieths S, Schöning B, Petersen A. Characterization of the 18-kDa apple

12.

13.

14. 15.

16.

17.

18.

19.

20. 21.

allergen by two-dimensional immunoblotting and microsequencing. Int Arch Allergy Immunol 1994;104:399-404. Westphal S, Foetisch K, Lauer I, Conti A, Kolarich D, Altmann F, et al. Cloning and characterization of fructofuranosidase, a newly identified glycoprotein as a minor allergen of tomato (Lycopersicon esculentum). J Allergy Clin Immunol 2002;109:308. Kolarich D, Altmann F. N-glycan analysis by matrix-assisted laser desorption/ionization mass spectrometry of electrophoretically separated nonmammalian proteins: application to peanut allergen Ara h 1 and olive pollen allergen Ole e 1. Anal Biochem 2000;285:64-75. Laemmli UK. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 1970;227:680-5. Pruzansky JJ, Grammer LC, Patterson R, Roberts M. Dissociation of IgE from receptors on human basophils, I. Enhanced passive sensitization for histamine release. J Immunol 1983;131:1949-53. Pini C, Iacovacci P, Butteroni C, Afferni C, Puggioni EMR, Tinghino R, et al. Role of carbohydrates in the histamine release from basophils: comparison between the glycosylated native and recombinant Cup a 1 allergen. Allergy 2002;57(suppl 73):12. Sturm A. Heterogeneity of the complex N-linked oligosaccharides at specific glycosylation sites of two secreted carrot glycoproteins. Eur J Biochem 1991;199:169-79. Botella MA, Quesada MA, Hasegawa PM, Valpuesta V. Nucleotide sequences of two peroxidase genes from tomato (Lycopersicon esculentum). Plant Physiol 1993;103:665-6. Mita H, Yasueda H, Ishii T, Akiyama K. IgE-mediated basophil releasability is influenced by intrinsic factors and by IgE on the cell surface. Allergy 1995;50:952-8. Mita H, Yasueda H, Akiyama K. Affinity of IgE antibody to antigen influences allergen-induced histamine release. Clin Exp Allergy 2000;30:1582-9. van Ree R, Cabanes-Macheteau M, Akkerdaas J, Milazzo J-P, LoutelierBourhis C, Rayon C, et al. β(1,2)-Xylose and α(1,3)-fucose residues have a strong contribution in IgE binding to plant glycoallergens. J Biol Chem 2000;275:11451-8.

Food and drug reactions and anaphylaxis