Prof. (Dr) Fatima Ferreira Department of Molecular Biology University of Salzburg Hellbrunnerstr. 34 A-5020 Salzburg Austria
Pollinosis patients often display adverse reactions upon the ingestion of plant-derived foods as a result of immunoglobulin E (IgE) cross-reactive structures shared by pollen and food allergen sources. The symptoms of such pollen-food syndromes (PFS) or class 2 food allergies range from local oral allergy syndrome to severe systemic anaphylaxis. Two clinical syndromes, the celery-mugwort-spice syndrome and the mugwort-mustard-allergy syndrome have been described in association with weed pollinosis. However, other associations between weed pollinosis and hypersensitivity to certain kinds of food have also been observed, like the mugwort–peach, the ragweed–melon–banana, the plantain–melon, the pellitory–pistachio, the goosefoot–fruit, the Russian thistle–saffron, and the hop–celery association. The number of allergen sources involved, the allergens, and influencing factors including geography, diet, and food preparation contribute to the high clinical complexity of PFS. So far, known causative cross-reactive allergens include profilins, lipid transfer proteins, and high-molecular weight allergens and/or glycoallergens. The current usage of nonstandardized allergen extracts poses additional problems for both diagnosis and therapy of PFS patients. Further identification and characterization of involved allergens is inescapable for better understanding of PFS and vaccine development. Panels of recombinant allergens and/or hypo-allergens are promising tools to improve both PFS diagnostics and therapy.
Patients suffering from pollinosis often display adverse reactions after the ingestion of a wide variety of plant-derived foods. Being first described as far back as 1948 (1), this kind of food allergy found special attention during the last decades, obviously because of the steadily increasing prevalence of inhalant allergies over this period (2). As specific clinical manifestations are usually restricted to the oral cavity (e.g. oral pruritus, hoarseness, swelling of lips, tongue and throat, pharyngitis, laryngeal edema), the term oral allergy syndrome (OAS) has been frequently used to describe associations between allergy to pollen and the concomitant hypersensitive reactions to certain kinds of fruits, vegetables, and spices (3–8). Because some patients experience reactions to foods without pollinosis and the symptoms are not obligatory limited to the oral cavity but may range from oral and gastrointestinal to severe systemic reactions, such as severe laryngeal edema, urticaria, bronchial asthma, or even food-induced anaphylaxis, the term OAS seems inappropriate. Recently, Mari et al. (9) defined OAS as a complex of symptoms induced by exposure of the oral and pharyngeal mucosa to food allergens including symptoms of increasing severity. According to this definition, OAS is not restricted to pollen-associated food allergies. Nevertheless, the word ‘oral’ could still be misleading. Thus, the proposed term pollen-food syndrome (PFS) seems to be less ambiguous (10). So far, several clinical syndromes have been described (11), such as the birch-fruit (12), the celery-mugwort-spice (13), and the latex-fruit syndrome (14), which by its molecular background, is comparable with PFS (10).
The term class 2 food allergy was also coined to describe the relationship between sensitivity to certain food and aeroallergens (2). In contrast to class 1 food allergy that mainly affects young children, class 2 food incompatibility is observed in adults as a consequence of sensitization to aeroallergens. Allergens eliciting class 1 food allergy (also termed complete food allergens) share special features, like resistance to gastric digestion, leading to the postulation that the sensitization process takes place in the gastrointestinal tract. In general, class 2 food allergens seem to be more sensitive to heat and digestive enzymes and therefore cannot cause per-orally sensitizations, but instead provoke allergic reactions in already sensitized patients. Thus, they are often called incomplete food allergens or nonsensitizing elicitors. According to their stability during the digestive process, they can cause symptoms ranging from mild oral reactions (typical for the birch-fruit syndrome) to anaphylactic shock, which is not rare within the celery-mugwort-spice syndrome. Therefore, the term PFS encompasses class 2 food allergy.
The association between pollinosis and food sensitization has been explained by different hypothesis. It has been suggested that lectins present in pollen and food are able to induce histamine release through unspecific interactions with immunoglobulin E (IgE) carbohydrate moieties (15), thus leading to allergy-like symptoms. However, IgE directed against common cross-reactive structures shared by pollen and plant-derived food is the most widely accepted and experimentally supported explanation (16, 17). Several lines of evidence suggest pollen as the primary source of sensitization leading to the induction of IgE antibodies that are capable to cross-react with homologous food allergens (6, 18). The advent of the idea that allergic reactions caused by the cross-reactivity between a sensitizer and an elicitor raises the question of which antigens should be called allergens: the sensitizers, the symptom elicitors, or both.
The IgE cross-reactivity might be clinically manifest or irrelevant. Clinical manifestations seem to be influenced by a number of factors including the host's immune response, allergen exposure, and the allergen itself. Structural characteristics of proteins are major determinants of cross-reactivity, thus PFS develop as a consequence of shared features at the level of primary and tertiary protein structure. It has been proposed that proteins sharing >70% sequence identity are often cross-reactive, while those displaying <50% rarely cross-react (19, 20). However, this notion shall be reconsidered. The cross-reactive birch and celery allergens Bet v 1 and Api g 1, for example, display only 40% sequence identity (21). In accordance, recent bioinformatics supported guidelines for assessment of genetically modified crops suggest a sequence identity of 35% as cut-off for potential cross-reactivity (22). Exceptions occur when postsynthetic modifications are involved in cross-reactivity between unrelated proteins, e.g. carbohydrate chains of allergenic glycoproteins. Although the clinical relevance of cross-reactive carbohydrate determinants (CCDs) is still doubted (20, 23–26), IgE antibodies directed toward glycans seem to show the widest pattern of cross-reactivity among allergenic extracts and in fact are often responsible for observed in vitro cross-reactions within PFS (19).
Additional factors influencing the clinical correlations of PFS comprise allergen concentration, differential expression of allergens during ripening, stability to cooking, and geographical and dietary factors (2, 10, 11). The geographical distribution of pollen allergens and regional dietary habits influence the frequency and development of distinct PFS. For example, allergy to fruits of the Rosaceae family is attributed to grass pollen sensitization in southern Europe and to birch pollen sensitization in Northern Europe (27–30). In this context, allergy to melon and banana has been associated with ragweed (31), apple with birch (12), celery with birch and mugwort (13, 17, 32), etc. Additionally, hypersensitivity to peach, the most frequent fruit allergy in Spain (29), is clinically not associated with any kind of particular pollinosis in southern Europe, but sensitization to taxonomically diverse pollen instead is linked to peach allergy (29, 33). Interestingly, there seems to be a higher prevalence of bronchial asthma in pollinosis patients with peach allergy (34). But it is not clear how peach allergy contributes as a risk factor for bronchial asthma.
A study on allergy to plant-derived fresh foods in a birch- and ragweed-free area showed that one in five pollen-sensitized patients was allergic to some of the tested foods. This is a much lower rate than reported in areas in which birch and/or ragweed pollens are predominant (35). Thus, the different pollen and dietary habits that predominate in a given geographic area influence the prevalence of different pollen sensitizations and associated food allergies. Geographic and dietary influences complicate epidemiologic studies on pollinosis-associated food allergy. So far, no exact data on the frequency of PFS is available. For different European countries it has been reported that one to two-thirds of birch pollinosis patients suffer from concomitant food allergy. Food intolerances are observed less frequently among mugwort or grass pollen patients. Klein-Tebbe and Herold (36) estimated that pollen-associated food allergy affects 5% of the population in central Europe.
Pollen-food syndromes associated with hypersensitivity to weed pollen
As reviewed by Gadermaier et al. (37) allergenic weed sources can be found in the botanical families of Asteraceae, Amaranthaceae, Urticaceae, Euphorbiaceae, Plantaginaceae, and Cannabaceae (38), with mainly plants of the Asteraceae family giving rise to PFS (Fig. 1). Table 1 provides a brief overview on weed pollinosis and food associations described so far. In Table 2, an overview of allergens presumed to be involved in weed pollinosis-associated food allergies is given.
Table 1. Overview on weed pollen–food associations
*Lacking evidence: nuts, legumes, Rosaceae fruits, and corn allergy are assumed to be involved within the mugwort-mustard-allergy syndrome.
†The Russian thistle has been demonstrated to cross-react with saffron flower; there is no evidence of cross-reaction between Salsola and saffron as a food allergen source.
‡Castor bean, mercury, and latex belong to the same botanical family of Euphorbiaceae and share common allergens; although no associations between castor bean and mercury with certain kinds of food have been observed, cross-reactions with foods involved in the latex-fruit syndrome cannot be ruled out.
Despite being one of the largest families of flowering plants, the Asteraceae or Compositae family contains only a few genera that constitute allergenic sources. These are Artemisia (mugwort), Ambrosia (ragweed), Helianthus (sunflower), and Parthenium (feverfew). Several PFS associated with pollinosis to the botanical family of Asteraceae weeds have been described, with mugwort being the genus most often involved (Fig. 1, Table 1).
The celery-mugwort-spice syndrome. Intolerance to celery and other vegetables as well as spices of the Apiaceae or Umbelliferous family (e.g. carrot, caraway seeds, parsley, fennel seeds, coriander seeds, aniseed) is commonly observed in birch and less frequently in mugwort (Ar. vulgaris) pollen-allergic patients. The term celery-mugwort-spice syndrome has been established to describe cross-reactivity between Artemisia and Apiaceae food as well as group sensitization within Umbelliferous plants (13, 39). Members of botanical families other than Apiaceae are also associated with the celery-mugwort-spice syndrome, e.g. Solanaceae (paprika), Piperaceae (pepper), Anacardiaceae (mango), and Liliaceae (garlic, onion; 13, 40, 41; Fig. 1). According to Moneret-Vautrin et al. (41), 10 different food allergies are related to the celery-mugwort-spice syndrome including different botanical families, like Apiaceae (e.g. coriander, caraway, and fennel) but also Liliaceae (e.g. garlic, onion, leek). Although considered rare, reports on allergic reactions after the ingestion of foods belonging to the Liliaceae family (42–47) and associations between Liliaceae and pollen hypersensitivity have been described (48, 49). Patients at the risk of spice allergy are young adults sensitized to mugwort and birch pollen, sharing cross-sensitization with various vegetables. In general, food allergy to spices is infrequent, thus it only amounts to approximately 2% of the totality of food allergy (41).
The celery-mugwort-spice syndrome cannot be discussed without mentioning the association between birch pollinosis and celery hypersensitivity as both clinical syndromes share common antigenic determinants (32). It has even been suggested that there are no celery allergens exclusively involved in the mugwort–celery association (50). In contrast, the Bet v 1 homologs Api g 1 (celery) and Dau c 1 (carrot) are exclusively responsible for the association between birch pollinosis and Apiaceae hypersensitivity (21, 51), as no Bet v 1-homologous protein has been found in mugwort pollen. Thus, based on the serologic data, it has been proposed to extend the celery-mugwort-spice syndrome to the celery-birch-mugwort-spice syndrome (17). However, Stager et al. (52) demonstrated that the celery–birch association only involves species from the Apiaceae family, whereas the celery–mugwort or the celery–birch–mugwort association comprises additional botanical families. An interesting observation pointing to differences between the celery–mugwort and the celery–birch association was published by Wuthrich et al. (53). Patients with positive skin tests to celery–birch displayed negative or low radioallergosorbent tests (RAST) to heated celery. By contrast, the celery–mugwort patients were clearly RAST-positive with heated celery extracts. These findings suggest the involvement of different allergens in the celery–mugwort and celery–birch associations.
Which are the cross-reactive allergens or structures possibly involved in the celery-mugwort-spice syndrome? Vallier et al. (49) described a 15 kDa celery component that was (later on) identified as celery profilin (Api g 4). Api g 4 could play a role within the celery–birch–mugwort association through its cross-reactivity with profilins from mugwort (Art v 4) and birch (Bet v 2; 54, 55). Bauer et al. (17) attributed a role for profilin and Bet v 1 homologs in the celery–birch association. In addition, other IgE cross-reactive proteins shared by celery, birch, and mugwort pollen were identified in the 46–60 kDa molecular weight (MW) range. These three groups of allergenic molecules were also identified in Apiaceae spices (56).
The IgE antibodies against carbohydrates are frequently observed in celery-allergic patients. Inhibition experiments using purified carbohydrate moieties (57) demonstrated the presence of N-glycans containing α1,3-fucose and β1,2-xylose in celery extracts. These sugar residues were shown to strongly contribute to IgE binding to plant glycoallergens (58). The role of CCDs in celery allergy was investigated by Fotisch et al. (59) IgE inhibition immunoblot analysis and deglycosylation experiments revealed α1,3-fucose as the key IgE-binding structure of N-glycans that are present on multiple celery glycoproteins with MWs higher than 40 kDa (59). Celery tuber proteins of 55 and 58 kDa, which were shown to be size variants of a new allergen tentatively designated Api g 5 (60), seemed to be especially important for patients with primary sensitization to mugwort pollen. Bublin et al. (61) concluded that the high MW glycoallergen Api g 5 is capable of binding human IgE and of activating basophils derived from a mugwort and celery-allergic patient exclusively via its N-glycans. Whether Api g 5 and the mugwort glycoallergen Art v 60 kDa (62), are homologous or similar proteins remains unclear. However, both are likely to be involved in CCD reactivity. Although several studies showed that glycoproteins containing CCDs are biologically active allergens (24, 57, 59, 61, 63), others demonstrated that CCDs have poor biologic activity (20, 23–26).
As already mentioned, allergen sources other than Apiaceae are often involved in the celery-mugwort-birch syndrome but not in the celery–birch association. Although Ebner et al. (64, 65) detected profilin and Bet v 1 homologous proteins in almost every source of spice analyzed, allergy to spices rarely represents an independent sensitization, but is rather a consequence of pollen sensitization and immunologic cross-reactivity (65). Leitner et al. (40) investigated allergens originating from pepper (Piperaceae) and paprika (Solanaceae) with relevance for the celery-mugwort-spice syndrome. The IgE Immunoblots and N-terminal sequencing revealed a 28-kDa pepper allergen homologous to a wheat germin protein and a 23-kDa paprika allergen homologous to an osmotin-like or pathogenesis-related (PR) protein in tomato. Furthermore, a 60-kDa allergen in pepper and other high MW allergens in paprika-bound patient's IgE antibodies. Thus, IgE cross-reactivity in the celery-birch-mugwort-spice syndrome to spices other than Apiaceae is not exclusively caused by Bet v 1 homologs and/or profilins.
An association between the celery-birch-mugwort-spice syndrome and mango fruit hypersensitivity was reported earlier (13). Enzyme allergosorbent test (EAST) and immunoblot inhibition experiments (66) demonstrated cross-reactions between mango fruit, mugwort pollen, birch pollen, celery, and carrot, all based on Bet v 1 and the Art v 60 kDa group of allergens.
Allergy to Liliaceae vegetables and spices occurs relatively rare but seems to affect young subjects with concomitant pollen hypersensitivity. Electrophoresis of garlic extract revealed two major protein bands at approximately 12 and 54 kDa, both being recognized by patient's IgE (48). A 12-kDa allergenic component was shown to be shared by several Liliaceae members (onion, leek; 67). Furthermore, similar IgE-binding proteins could be detected in mugwort pollen and hazelnut after sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and IgE immunoblot. These results indicate that garlic, onion, and certain pollens share similar allergenic epitopes. However, further studies on the role of Liliaceae vegetables and spices within the celery-mugwort-spice syndrome are still necessary.
The evaluation of recombinant Api g 1 for diagnostic purposes (51) revealed dramatic geographical differences concerning allergens involved in the celery-birch-mugwort-spice syndrome. Comparison of a central European and a Mediterranean group of celery-allergic patients showed that all central European patients displayed skin reactions with recombinant Api g 1, whereas only two of 12 patients from the Mediterranean group were Api g 1-positive. Immunoblot analysis of the Mediterranean sera showed no IgE-binding to recombinant Bet v 1 or Api g 1. In contrast, a number of different proteins ranging from 28 to 69 kDa were recognized in extracts of celery, birch, and mugwort pollen. These differences raise questions on the source of primary sensitization. In areas where birch and other Fagales trees grow, reactivity to Bet v 1 and homologous proteins is frequent. The reactivity to Api g 1 reported for the central European group perfectly fits into this scheme. Patients living in the Mediterranean area, where birch trees are rare, do not display IgE against Api g 1 but rather against high MW proteins that are also present in mugwort pollen. In fact, all but one of these patients had positive RAST scores with mugwort pollen extracts. Therefore, it is likely that cross-reactivity mediated by high MW allergens present in celery and mugwort is responsible for celery allergy in the Mediterranean area. Thus, in areas where birch trees are common celery allergy seems to be due to birch pollen sensitization whereas in areas where birches are rare mugwort pollen seems to be the primary source of sensitization.
The spectrum of allergens involved in the celery-birch-mugwort-spice syndrome (Table 2, Fig. 2) can be divided into at least four groups: first, there is the group of Bet v 1 homologous proteins (21). It has been demonstrated that IgE reactivity to Api g 1 is based on primary sensitization to Bet v 1 in a central European population (51). Secondly, the pan allergen profilin present in celery and designated Api g 4 may display IgE cross-reactivity with birch Bet v 2 and mugwort Art v 4 profilins (54, 55). The third group is composed by high MW allergens and/or glycoallergens containing CCDs that have been shown to be thermostable (17, 50, 57, 68). Thus, bearing in mind the variable frequency of anaphylaxis in the celery-birch-mugwort-spice syndrome (39, 69), high MW allergens or glycoallergens could be considered an especially important cross-reactive structure in the celery-birch-mugwort-spice syndrome. Cross-reactive carbohydrate determinants are frequently recognized by IgEs from celery-allergic patients. In this context, Art v 60 kDa (62) and the recently characterized celery allergen Api g 5 (61) are of interest. Little is known about the fourth group of mugwort and celery allergens involved in the cross-reactions with plants other than Apiaceae, like those belonging to the Solanaceae or Piperaceae family (40). A number of proteins in the higher molecular mass range have been identified as IgE-binding proteins in mugwort, birch, and celery (17), as well as being common allergenic structures in ragweed and mugwort pollen (70). Interestingly, a case report described celery-induced anaphylactic shock because of cross-reactivity with ragweed, a member of the Compositae family that is usually associated with allergy to melon (71).
As described above, the number of allergen sources involved, the nature of the allergens, and influencing factors, like geography, dietary habits, and food preparation, render the celery-birch-mugwort-spice syndrome a clinical feature of high complexity. Therefore, an allergen-based classification (72) would facilitate the analysis of allergen sources bearing clinically relevant cross-reactive allergenic structures.
The mugwort-mustard allergy syndrome. Mustard belongs to the botanical family of Cruciferae or Brassicaceae comprising two major types of seeds, the white or yellow (Sinapis alba), as main ingredient of American-style mustards, and the brown or oriental mustard (S. juncea) that is used for European and Chinese products. The major allergens of white (Sin a 1; 73) and oriental mustard (Bra j 1; 74) are seed storage proteins belonging to the 2-S albumin protein family and share common epitopes (75). Mustard allergy is a rare but not an uncommon disorder that might induce systemic reactions. Several case reports of severe anaphylaxis upon ingestion of mustard have been published. Regarded as a masked allergen in processed foods, mustard allergy can be misdiagnosed as idiopathic anaphylaxis due to the lack of full ingredient labeling. Thus, mustard sensitivity should be routinely tested in patients with idiopathic anaphylaxis (76–81).
Caballero et al. (82) reported that most of their mustard hypersensitive patients displayed systemic reactions and suffered from associated pollinosis or allergy with other vegetable foods. Among patients developing adverse reactions (mainly OAS, 10% systemic anaphylaxis) upon ingestion of mustard, more than 97% were sensitized to Ar. vulgaris pollen. Furthermore, 100% of the tested patients showed associated sensitizations with Brassicaceae vegetables excluding mustard (e.g. broccoli, cabbage, and cauliflower). Based on these observations, Figueroa et al. (83) proposed a novel mustard-mugwort-allergy syndrome to describe associations with mugwort pollinosis and several botanically unrelated plant-derived foods (e.g. nuts, legumes, Rosaceae fruits, corn). Although the causative allergens have not yet been investigated, Art v 60 kDa, Art v 4 (profilin), and Art v 3 [mugwort nonspecific lipid transfer protein (LTP)] may come into question.
The mugwort–peach association. Sensitization to peach and related Rosaceae fruits without clinical expression is commonly observed as a result of extensive cross-reactivity of IgE directed toward LTPs, Bet v 1 homologs, profilins, and CCDs. Peach (Prunus persica) allergy is the most frequent fruit allergy in Spain (29), but unlike other geographical areas where Bet v 1 homologs are responsible for the birch-fruit syndrome, the LTP Pru p 3 it is considered the major peach allergen for the Spanish population and IgE response to this allergen is related to the clinical expression of peach allergy (84). Sensitization to profilin (Pru p 4) was observed in patients with associated pollinosis but did not appear to be related to the clinical reactivity to peach (85). According to the seriousness of the symptoms, peach allergy in southern Europe can be divided into two groups. The first group, comprising nonpollen-allergic patients, is more predisposed to the occurrence of systemic symptoms whereas the second group with concomitant pollen allergy features a higher prevalence of bronchial asthma (34).
Cuesta-Herranz et al. (29) investigated cross-reactive allergy patterns considering sensitization to peach and several taxonomically unrelated pollen sources in Spain. Skin test results revealed that peach-allergic patients frequently react to most pollen of trees, grasses, and weeds, even when the species of origin were not encountered in the geographic area. Thus, pollinosis might be considered a risk factor in the development of peach allergy. Taken together these results suggest that peach allergy in Spain does not seem to be clinically associated with sensitization to any particular pollen but instead sensitization to taxonomically diverse pollens is rather linked to peach allergy. Within this study the pan allergen profilin (Pru p 4) was considered to be a relevant IgE cross-reactive antigen (85). Although profilins have been shown to be strong sensitizers, it seems that frequently they do not elicit clinical symptoms. Nonetheless, a possible role of profilin in the higher prevalence of asthma in peach-allergic patients with concomitant pollinosis has been suggested (34). Well conducted clinical investigations are still necessary to determine the clinical relevance of profilins.
Interestingly, Ar. vulgaris pollen extracts produced the maximal inhibition of specific IgE to peach in IgE inhibition experiments. In this context, Pastorello et al. (33) demonstrated that hypersensitivity to mugwort in patients with peach allergy is due to a common LTP and often without clinical expression. In IgE immunoblotting some of the patients with OAS for peach and specific IgE to mugwort reacted only to mugwort and peach LTP (Art v 3 and Pru p 3, respectively). These patients did not present hay fever symptoms and the IgE reactivity to Art v 3 seemed to be a consequence of peach sensitization. These results contrast with those reported by Lombardero et al. (86) showing that Art v 3 behaves as the primary sensitizing allergen in some patients with IgE to both Art v 3 and Pru p 3 (87). Diaz-Peralez et al. (84) also reported that LTP from Artemisia pollen cross-reacts with LTPs originating from Rosaceae fruits. However, among Rosaceae-allergic patients there are significant differences in their IgE-binding capacity to members of the plant LTP family. Thus, further studies are needed to evaluate the clinical significance of the observed cross-reactivity and clinical association.
The ragweed–melon–banana association. Anderson et al. (31) reported a series of cases where patients with ragweed (Am. artemisiifolia) allergy also experienced oral symptoms after eating various members of the Cucurbitaceae or gourd family (e.g. watermelon, cantaloupe, honeydew melon, zucchini, and cucumber) and banana. Measurements of specific IgE (88) revealed that up to 50% of ragweed patients displayed IgE directed against any single gourd family member. In addition, cross-reactivity between watermelon (Citrullus lanatus) and ragweed (Am. artemisiifolia) pollen has been shown in enzyme-linked immunosorbent assay (ELISA)-based IgE inhibition experiments. More recently Cuesta-Herranz et al. (35) showed that melon allergy occurred mainly in patients with pollinosis, even in ragweed-free areas. The most common clinical manifestations associated with melon (Cucumis melo) allergy seem to be restricted to the oropharynx and rarely affect other target organs (89). However, a report on patients experiencing life-threatening systemic reactions, including respiratory symptoms and hypotension was also published (90). In the present study (90), melon-allergic patients were shown to frequently display positive skin tests and reported to experience symptoms to other fruits such as peach, fig and kiwi. In addition, up to 23% of the patients displayed concomitant latex sensitization. As all patients suffered from pollinosis, pollen allergy was especially linked to melon hypersensitivity. Compared with a pollen-allergic control group, a significant higher frequency of sensitization to several tree and weed pollens, including Ulmus and Ambrosia as well as a higher prevalence of bronchial asthma was reported within the melon hypersensitive group (90).
Despite the lack of molecular data on melon allergens, possibly cross-reactive allergens were identified by IgE immunoblotting experiments using sera from patients displaying OAS after the ingestion of melon (91). A 13-kDa component was identified as melon profilin and considered a major allergen. Profilins were also detected as important allergenic compounds in other Cucurbitaceae fruits and vegetables (e.g. zucchini, Cucurbita pepo; 92). Digestibility analysis of melon profilin revealed its stability in human saliva, probably because of the lack digestive proteases in human saliva, such as pepsin. In contrast, simulated gastric fluid readily digested melon profilin within a few seconds. This may lead to the speculation that melon profilin may be responsible for the local oral symptoms. However, it should be considered that about 10% of patients with melon allergy display severe anaphylactic reactions (90). The clinical findings suggest that some patients allergic to melon could be concomitantly sensitized to allergens highly resistant to pepsin digestion, such as LTPs. However, no melon LTP has been identified yet. In addition, a number of components with MW ranging from 15 to 60 kDa were identified as allergens in melon, zucchini, cucumber, and watermelon. Most of these allergens seem to harbor complex asparagine-linked glycans comprising xyloxyl and fucosyl residues, which may act as CCDs (58).
Much less is known about banana allergens. Grob et al. (93) identified banana profilin as a putative cross-reactive allergen in a patient suffering from latex-fruit syndrome. Additionally, a 70-kDa protein was recognized by IgE from a banana-allergic patient without concomitant latex sensitization.
Taken together, the pan allergen profilin, high MW allergens or glycoallergens, as well as LTPs seem to be possible candidates involved in the clinical manifestation of the ragweed–melon–banana association. However, further studies should be performed to identify and characterize the cross-reactive allergens and/or structures.
In Australia and Mediterranean countries, 20–40% of pollen-allergic patients are found to be allergic to plantain (Plantago lanceolata).
The plantain–melon association. The clustering of allergy to melon (Cu. melo), plantain (P. lanceolata), and grass (Dactylis glomerata) pollen was reported by Garcia Ortiz et al. (94, 95). The IgE immunoblot analysis revealed that the three species share several distinct proteins of 14 and 31 kDa, and a spectrum of proteins migrating between 40 and 70 kDa. Inhibition experiments demonstrated that melon allergens inhibit IgE binding to grass and plantain pollen proteins. Although these results suggest that plantain, grass, and melon share common allergens, clinical and molecular data are still lacking.
Parietaria sp. (pellitory) is the allergenic genus of the Urticaceae family.
The pellitory–pistachio association. Pistachio nuts, belonging to the botanical family of Anacardiaceae, are widely used in the catering industry to produce ice creams, cakes, and mortadella or are simply eaten roasted. Very few cases of pistachio nut sensitization have been reported, with the patients concerned often displaying reactions to other dried fruits and pollen. Furthermore, different members of the Anacardiaceae family (pistachio, mango, and cashew) have been demonstrated to share common allergens (96). Four different IgE-binding components of pistachio nut with apparent MWs of 34, 41, 52 and 60 kDa have been identified (97).
Liccardi et al. (98) described two uncommon cases of severe OAS after the ingestion of pistachio nuts in Italian subjects with exclusive positive skin prick test to Parietaria and pistachio. The IgE immunoblot experiments suggested some degree of cross-reactivity between Parietaria and pistachio allergens. A subsequent study (99) was designed to investigate whether sensitization to pistachio nut is more frequent in patients with pellitory pollinosis. Within this study all patients displaying positive skin prick tests to pistachio were sensitized to Parietaria and suffered from respiratory symptoms. Another finding was that sensitization to pistachio nut allergens rarely induced severe clinical symptoms. Nonetheless, considering the high frequency of Parietaria pollinosis in the Mediterranean area and the increasing consumption of pistachio nut, it is suggested that Parietaria-sensitized patients should routinely undergo skin tests for pistachio.
Goosefoot (Chenopodium album) and the Russian thistle (Salsola kali) are the two important allergenic weeds of the Amaranthaceae family.
The goosefoot–fruit association. Most studies on chenopod allergy have shown a high degree of cross-reactivity with Amaranthaceae and taxonomically less related pollens, as well as foods from the Liliaceae family, such as garlic, onion, and Asparagus. Immunoblot and IgE inhibition analysis revealed that preincubating sera from garlic-allergic patients with onion, Phleum, and Chenopodium extracts partially abolished IgE binding to garlic allergens (48).
Several cases of patients with pollen allergy and positive skin prick test to C. album who displayed OAS after eating fresh fruits (e.g. banana, melon, and peach; 100) were reported. Immunoblot assays and IgE inhibition experiments using goosefoot patient's sera revealed several cross-reactive allergens ranging from 10 to 20 kDa and a 14 kDa major allergenic compound. Banana extract was able to totally inhibit patient's IgE recognition of Chenopodium extract. However, more extensive studies are necessary to elucidate and characterize the molecular structures involved. The recently isolated pan allergens profilin (Che a 2) and polcalcin (Che a 3) might play a role in goosefoot IgE cross-reactivity. Chenopodium profilin Che a 2 might be involved in cross-reactions to foods as it shows higher identity to latex and to food profilins than to profilins originating from other pollen (101). Furthermore, recombinant goosefoot profilin displayed a high degree of cross-reactivity with different allergenic sources (102).
The Russian thistle–saffron association. No cross-reactivity involving Russian thistle (Salsola) pollen and foods has been described so far. However, a significant degree of cross-reactivity between Salsola and saffron has been reported in a study investigating saffron flower sensitization and its clinical significance in Spanish saffron workers (103).
Japanese hop (Humulus japonicus) belongs to the botanical family of Cannabaceae and is one of the major allergenic weed pollen in Korea. An association between Japanese hops pollinosis and allergy to common hop (H. lupulus) and celery, as determined by skin prick tests have been suggested (38). Interestingly, no significant associations with ragweed or mugwort pollen were observed.
The most important allergenic weeds of the Euphorbiaceae family are castor bean (Ricinus communis) and mercury (Mercurialis annua). The latex tree (Hevea brasiliensis) also belongs to this botanical family. In vitro studies showed that latex, castor bean, and mercury contain cross-reactive allergens (104). However, the clinical significance remains to be elucidated. Although, latex has been demonstrated to be a potent source of allergens cross-reacting to foods within the latex-fruit syndrome (105), no associations between the botanically related castor bean and mercury and certain kinds of food have been observed so far.
Allergens involved in pollen-food syndromes
As a matter of evolution, cross-reactive and/or homologous molecules are expected to occur in taxonomically related allergen sources. For example, mugwort (Ar. vulgaris)-allergic patients were reported to display adverse reactions after oral provocation with chamomile (Matricaria chamomilla) infusion, both plants belonging to the same botanical family (106). A patient suffering from rhinoconjunctivitis caused by Asteraceae pollinosis elicited dyspnea after the ingestion of sunflower (H. annuus) kernels, a species also belonging to the Asteraceae family (107). In honey-allergic patients, primary sensitization may be due either to honey itself, to airborne Compositae pollen, or even to cross-reacting bee venom components (108). Pollen might act as hidden allergen source leading to adverse reactions upon honey ingestion (109).
Interestingly, the more frequent food sources involved in PFS are botanically unrelated to the pollen source, but nonetheless contain conserved homologous proteins (1). A few cross-reactive structures that are responsible for the development of PFS have been described in detail. So far, there is evidence that Bet v 1 homologs, the pan allergens profilin and LTPs, as well as high MW allergens and/or glycoallergens are involved in pollen allergy with concomitant food hypersensitivity. Although their clinical relevance is still questioned, CCDs seem to play a role within pollen food allergy (16, 26, 68, 110–112). The clinical relevance of cross-reactions based on the recognition of CCDs and profilin is limited for the population of pollen-allergic patients as a whole. For selected food-allergic patients, however, N-glycans and in particular profilin are potentially of clinical relevance. Because of their stability to proteolysis and processing, LTPs have been established as allergens in several food sources causing clinically more severe reactions (110). Finally, lectins present in many plant-derived foods that are able to induce histamine release, thus allergy-like symptoms in an immunologic unspecific manner, should be mentioned (2, 15, 16).
Below we give a short overview on known cross-reactive allergens possibly involved in pollen-associated food allergies (Fig. 2), with especial emphasis on weed PFS. The group of pan-allergens includes profilins and LTPs. In addition, high MW allergens and CCDs are discussed. Because of their involvement in the celery-birch-mugwort-spice syndrome, the Bet v 1 family is also briefly mentioned.
As reviewed by Breiteneder and Ebner (2) plant-derived proteins responsible for food allergy include various families, like PR proteins, proteinase and α-amylase inhibitors, peroxidases, profilins, seed storage proteins, thiol proteases, and lectins. Several of these proteins are widely distributed throughout the plant kingdom, thus possibly being involved in extensive IgE cross-reactivity between antigens from taxonomically unrelated plant species, a phenomenon described by the pan-allergen theory (10, 110, 113). Below, we provide a short description of cross-reactive pan-allergens involved in PFS.
Profilin. Profilin, a 12–15 kDa actin-binding and cytoskeleton regulating protein, is an ubiquitous pan-allergen involved in many cross-reactions between inhalant and nutritive allergen sources (2, 19, 110, 112, 114–117). Mugwort profilin Art v 4 (118), for example, has been demonstrated to play a role in the celery-mugwort-spice syndrome, whereas birch profilin Bet v 2 seems to be involved within the birch–celery association because of cross-reaction with celery profilin Api g 4 (17, 49, 54). The importance of latex profilin Hev b 8 within the latex-fruit syndrome has been demonstrated (119). In addition, profilin is considered an important mediator in IgE cross-reactivity between pollen and exotic fruit. A case report (107) described a Compositae-allergic patient displaying anaphylactic reactions to lychee fruit, probably due to sensitization to profilin. Banana and pineapple profilin have been demonstrated to cross-react with Hev b 8 and Bet v 2, the latex and birch profilins, respectively (120). Asero et al. (121) even concluded that allergy to melon, watermelon, citrus fruits, tomato, and banana can be used as a marker of profilin hypersensitivity if natural rubber latex and LTP sensitization can be ruled out. In patients with OAS to melon, profilin is considered a major allergen highly susceptible to pepsin digestion but not to human saliva (91). However, the role of profilins in triggering allergic symptoms has not been addressed in carefully designed clinical studies.
Lipid transfer proteins. Plant LTPs, named after their ability to transfer phospholipids from liposomes to mitochondria, form a family of 9 kDa polypeptides widely distributed throughout the plant kingdom. Thus, LTPs represent a potential pan-allergen family. They belong to the PR-14 type proteins with a role in plant defense because of their antifungal and antibacterial activities (2) and are likely to act as potent food allergens because of their thermostability and extreme resistance to pepsin digestion (7). Plant LTPs are considered major allergens especially in the Mediterranean area where they are considered the most important allergens of Rosaceae fruits, such as apple (Mal d 3), peach (Pru p 3), apricot (Pru ar 3), and cherry (Pru av 3; 19, 84, 85).
Apple allergy is commonly associated with birch pollinosis, mainly due to the cross-reactivity between Bet v 1 and Mal d 1. However, apple-allergic patients with no concomitant birch hypersensitivity frequently display sensitization to apple LTP, an apple-specific allergen in southern Europe (122). Furthermore, mugwort pollen LTP (Art v 3) has been demonstrated to cross-react with peach LTP (Pru p 3), thus being involved in the mugwort–peach association that especially occurs in the Mediterranean area (86).
High-molecular weight allergens and cross-reactive carbohydrate determinants
A group of cross-reactive high MW allergens (45–60 kDa) were identified in various pollen and food sources and presumed to play an important role in PFS (68). For example, the high prevalence of sensitization to a 60-kDa glycoprotein from mugwort pollen has been demonstrated (86, 123). Heiss et al. (62) identified cross-reactive proteins in ragweed, timothy grass, and birch pollen, as well as in fruits and vegetables, such as apples, peanuts, kiwi, and celery.
The CCD containing glycoallergen Api g 5, a protein of 55–58 kDa (60), was shown to bind IgE exclusively via its carbohydrate moiety (61). Whether Art v 60 kDa and Api g 5 are homologous proteins remains unclear. However, they might be involved in carbohydrate cross-reactivity as many high MW celery allergens contain CCDs. In this context, IgE antibodies directed against carbohydrates occur in 25% of celery-allergic patients and it has been demonstrated that N-glycans containing α1,3-fucose and β1,2-xylose (58) form the key IgE-binding epitopes of celery proteins with apparent MWs above 40 kDa (59). Cross-reactive high MW allergens have also been detected in other Apiaceae vegetables and spices as well as in paprika, pepper, and mango (17, 40, 56, 66). Thus, these allergens are likely candidates to be involved in the celery-mugwort-spice syndrome as well as in the birch–celery association (17), and their further structural characterization is warranted.
Bet v 1 homologs
As discussed above, many families of plant-derived allergens are homologous to PR proteins that are induced upon pathogen infection, wounding, or certain environmental stress (124). Bet v 1, the major allergen of birch, for example, belongs to the PR-10 type protein family with homologs in Rosaceae fruits, like apple (Mal d 1), cherry (Pru av 1), apricot (Pru ar 1), and pear (Pyr c 1), which all may contribute to the clinical manifestations of the birch-fruit syndrome (2). Bet v 1 homologs are also present in Apiaceae vegetables, e.g. celery Api g 1 and carrot Dau c 1 (21, 64, 125). These homologous proteins and other allergen families involved in the celery-mugwort-spice syndrome (e.g. profilin and high MW glycoallergens) seem to be responsible for the birch–celery association (19, 21, 49, 50, 56).
Diagnosis and therapy aspects
The diagnosis of allergic reactions to food is sometimes disregarded for varying reasons. As the clinical manifestations of the OAS are often not severe, patients visiting an Allergy Clinic for other reactions, e.g. concomitant pollinosis, do not mention their food allergy symptoms unless specifically questioned. Furthermore, as most class 2 food allergens are easily degraded during storage or extraction procedure, the biologic activity of allergenic extracts currently used for diagnosis is often very low or even absent. The use of a panel of recombinant food allergens might help to overcome this problem (126). Nonetheless, the diagnosis of food allergy is important because it may worsen the prognosis of pollinosis or elicit systemic reactions. Another problem that is consistent with earlier studies, was demonstrated by Cuesta-Herranz et al. (35). Plant-derived fresh food often elicits positive responses to skin testing, although in many cases, these are not associated with any allergy symptoms. This fact conditions a high number of false-positive results (positive skin tests in nonallergic patients) and a low positive predictive value (35). Serologic tests and skin prick tests solely indicate the presence of allergen-specific IgE but do not diagnose the clinical appearance of food allergy. Oral provocation tests have to be performed in certain cases to confirm or to refute the patient's food allergy. At present, the double-blind placebo-controlled food challenge (DBPCFC) is considered still the only conclusive evidence of a food allergy (126, 127).
As PFS develop as a consequence of cross-reactive allergens, it would be predictable that pollen allergen immunotherapy could lead to a resolution not only of allergic rhinitis symptoms but also of the associated food allergy. Indeed, a few successful cases have been reported (1, 128–130). Thus, for some patients suffering from PFS, standard pollen immunotherapy will alleviate both their pollen- and food-induced symptoms. On the contrary, other studies indicated that although specific immunotherapy (SIT) with pollen preparations effectively improves hay fever symptoms in PFS patients, a parallel reduction of clinical sensitivity to foods is not always observed or is rather limited (131, 132).
Specific immunotherapy is routinely performed with highly complex pollen allergen extracts and bears the risk of side effects, such as systemic anaphylaxis as well as sensitization to new and/or cross-reactive allergens. Thus, panels of recombinant allergens or hypo-allergens are promising tools to overcome these problems. Future studies should be directed toward the investigation and characterization of clinically relevant allergens involved in both pollinosis and concomitant food hypersensitivity. Investigation of PFS at the molecular level seems inescapable considering the high complexity of the allergen sources and the number of pan-allergens and/or other potential cross-reactive structures that are involved. In addition, detailed investigation of PFS at the allergen level will contribute to our understanding of allergic cross-reactivity and its clinical manifestation as PFS. As suggested by Mothes et al. (72), future terminology might rather describe, for example, a profilin-sensitized patient and not a celery-birch-mugwort-spice patient.
The PFS have been estimated to affect 5% of the central European population (36). Nevertheless, little is known about PFS associated with weed pollinosis and the molecules involved. Several weed pollen-associated food allergies have been individually described at the level of allergen sources. However, at the molecular level almost exclusively pan-allergens, which occur in almost all kind of pollen, have been implicated so far. Therefore, it cannot be excluded that the described syndromes are in fact weed, grass, and tree pollen associated. For example, it is well established that allergy to Apiaceae is associated with birch but also with mugwort pollinosis, both associations sharing common antigenic determinants leading to a proposed extension of the term celery-mugwort-spice syndrome to celery-birch-mugwort-spice syndrome (17). However, Bet v 1 and Api g 1 are exclusively involved in the celery–birch (21) whereas Art v 60 kDa and Api g 5 might only play a role in the celery–mugwort association (61). Furthermore, the celery–mugwort association has been demonstrated to include additional botanical families other than Apiaceae (52). Thus, there seem to exist similarities but also differences between weed- and tree pollen-associated food allergies.
The celery-mugwort-spice syndrome is the best characterized weed PFS, but still little is known about the cross-reactive structures and even less data are available for other weed pollen-associated food allergies. Thus, the current status of research does not offer satisfactory explanations for these complex clinical features. Further investigation of PFS at the allergen level seems inescapable for better understanding the molecular background and for improvement of both diagnosis and therapeutic approaches.
The work of the authors was supported by the Austrian ‘Fonds zur Förderung der Wissenschaftlichen Forschung, FWF’ (Projects S8802-B01 and P16456-B05) and the ‘Österreichische Nationalbank, ÖNB’ (Project 10150).