IgE reactivity to components of fish allergens in Pacific cod ( Gadus macrocephalus ) in atopic dogs

Background IgE reactivity to fish allergens in atopic dogs, which are used as models for food allergy, has not been elucidated to date. We investigated IgE reactivity to crude and purified Pacific cod (Gadus macrocephalus) allergens in atopic dogs to identify the allergenic components of cod. Methods Specific IgE to crude cod allergens in the sera of 179 atopic dogs, including 27 dogs with cod allergy, were measured using enzyme-linked immunosorbent assay (ELISA). The allergenic components of crude cod antigen were analyzed by ELISA, immunoblotting, and liquid chromatography-tandem mass spectrometry (LC-MS/MS). IgE reactivity to parvalbumin, collagen, and tropomyosin was evaluated using the sera of atopic dogs that were positive for specific IgE to crude cod allergens. Results Specific IgE to crude cod allergens were present in 36 (20%) of the 179 atopic dogs and 12 (44%) of the 27 dogs with cod allergy. In allergen component analysis, IgE reactivity to tropomyosin and enolase was observed in the sera of dogs with cod allergy. Among the 36 dogs with IgE reactivity to crude cod extracts, 9 (25%), 14 (39%), and 18 (50%) dogs had specific IgE to parvalbumin, collagen, and tropomyosin, respectively. Conclusions The dogs exhibited IgE reactivity to the cod allergen components similar to that observed in humans, providing support for the use of atopic dogs with fish allergy as a model for fish allergy in humans.


Background
The prevalence of fish allergy, which affects approximately 0.2% of the world population [1], is over ten times higher in geographic regions where fish is an essential dietary component such as Japan [2,3]. Fish allergy is typically known to be a life-long condition in contrast to other food allergies [3]. Since clinical crossreactivity to different fish species is a widely accepted feature of fish allergy, affected individuals have to avoid all fish species for long periods, and inadvertent exposure to fish allergens, and the consequent severe or fatal reactions remain a grave risk for fish-sensitive individuals [3]. Among the fish allergens, parvalbumin is the best characterized major allergen that is found in many species [4][5][6]. In a previous study, we identified fish gelatin (type I collagen) as an allergen [7]. More recently, other proteins such as tropomyosin, enolases, and aldolases were also considered as relevant allergens in fish [8]. To resolve the issue of quality of life in fish-allergic patients, it is necessary not only to investigate the mechanisms underlying the acquisition of allergenicity but also to elucidate the approaches to reduce allergenicity.
Animal models are beneficial to test food allergies as they provide a more rapid and extensive evaluation of allergenicity of certain foods [9]. Dogs are recognized as a useful model to study IgE-mediated hypersensitivity [10,11] as they are one of the few species other than humans in which allergies develop naturally following environmental exposure to a broad spectrum of allergens, including foods [12,13] Disease states in dogs include atopic dermatitis, gastroenteric inflammation, and anaphylaxis [14]. Based on the many similarities between canine atopic dermatitis and humans [13], atopic dogs have been utilized as animal models for food allergies to cow's milk [15], corn [16], and nuts [11].
The allergenic components of atopic dogs in food are assumed to correspond with those of humans [17]. However, Kubota et al. reported that the allergenic components in atopic dogs were different from those in humans [18]. Previous studies reported that atopic dogs were appropriate models for food allergies by demonstrating the production of specific immunoglobulin (Ig) E to crude allergens and positive oral challenges similar to those observed in human subjects [11,15].
These dogs might be suitable models to elucidate the mechanisms underlying the allergenicity of the food components if the allergenic components in the dogs are analyzed. Therefore, in the present study, we investigated IgE reactivity to crude and purified Pacific cod (Gadus macrocephalus) allergens in atopic dogs and elucidated similarities in fish allergy between humans and dogs to assess the potential of atopic dogs with fish allergy as a suitable animal model.

Sera of atopic dogs
To examine IgE reactivity in atopic dogs, we obtained sera from 179 dogs that were diagnosed with atopic dermatitis based on the criteria by Willemse [19] and Prelaud et al [20] among dogs visiting Fujimura Animal Hospital (Osaka, Japan). Twenty samples from laboratory dogs were used as negative controls. The dogs were housed indoors as experimental laboratory animals and had never been exposed to fish antigens. None of the laboratory dogs exhibited signs of atopic dermatitis.
Enzyme-linked immunosorbent assay (ELISA) for parvalbumin or collagen from salmon (Atlantic salmon; Salmo salar), sardine (Japanese pilchard; Sardinops melanostictus), and mackerel (Chub mackerel; Scomber japonicus) were performed using individual sera of nine negative control dogs that provided a large amount of serum. All sera were stored at −80°C before use. Oral informed consent was obtained from the dog owners. All experimental procedures were carried out in accordance with Japanese law and approved by the animal care and user committee of Azabu University.

Food elimination and oral provocation tests
Among the 179 dogs with atopic dermatitis, 27 were confirmed to have cod reactivity by the oral provocation test after the elimination diet test; consent for food provocation test was obtained from the dog owners. These 27 dogs were fed commercial hydrolyzed protein diets as elimination diets for 6-8 weeks by the dog owners. When the veterinary physician recognized the complete resolution of the clinical signs during the food elimination test, the dog was admitted to the animal hospital and challenged with various cod ingredients, including grilled cod meat and cod-containing dog foods. The cod provocation test was discontinued immediately upon the relapse of the clinical signs including vomiting, diarrhea, erythema, pruritic urticaria, and conjunctival hyperemia [21].

Preparation of crude cod allergens
Cod is not only one of the most commonly consumed fish species in Europe and Japan [22] but also one of the most characterized fish species with allergen components [8]. Thus, Pacific cod was purchased from a fish market in Japan to be used in the study. The fresh, raw meat of four fishes (500 µg) was homogenized in 500 µl phosphate-buffered saline (PBS, 10 mM pH 7.2) and rotated overnight at 4°C.
After centrifugation at 21500 g for 5 min at 4°C, the supernatant was collected, and the protein was quantified using the BCA protein assay (Bio-Rad, Hercules, CA, USA).
Tropomyosin was purified from the ether powder by Bailey's method with slight modification [19]. Briefly, fish ether powder was stirred in a beaker with 75 ml extraction buffer containing 15 mM Tris HCl pH 7.6 (Sigma Aldrich, St Louis, MO, USA), 1 M KCl (Kanto Kagaku, Tokyo, Japan), and 2 mM dithiothreitol (Sigma Aldrich) overnight at 4°C. The extract was collected by centrifugation at 5400 g for 10 min at 4°C. The supernatant pH was adjusted to 4.5 with 1 N HCl to precipitate tropomyosin, and the precipitation was collected by centrifugation at 5400 g for 10 min at 4°C. The isoelectric precipitation was repeated once, and the precipitated material was dissolved in the extraction buffer. The supernatant after the extraction was collected by centrifugation and fractionated by ammonium sulfate at a concentration of 50%. The sample precipitated by ammonium sulfate was dissolved and dialyzed against PBS. The obtained protein extracts were confirmed by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE).

SDS-PAGE and immunoblotting
Among the 20 atopic dogs with IgE sensitivity to crude cod allergens that did not react to parvalbumin or collagen, sera were used from two dogs (no. 34 and no. 128) that provided a large amount of sera and reacted to crude cod allergens but not cod parvalbumin or collagen, as confirmed with the provocation test using codcontaining foods (Table 1). SDS-PAGE was performed according to the method of Laemmli [24]. Precision plus protein standards (Bio-Rad) were used as molecularmass markers. Protein components were electrophoretically separated using 5%-20% gradient polyacrylamide gels, and proteins were visualized either by Coomassie brilliant blue R250 (Bio-Rad) or transferring onto polyvinylidene difluoride membranes (GE Healthcare, Chicago, IL, USA). Immunoblotting was performed as described previously [7]. IgE in patient dog sera were used as primary antibodies, which were diluted 1:10 in tris buffered saline containing 0.1% Tween-20 and 5% nonfat dried milk. Mouse monoclonal anti-dog IgE antibodies (0.5 μg/ml) were used as secondary antibodies [21]. Detection was performed using an enhanced chemiluminescence immunoblotting detection reagent (GE Healthcare).
Extracts were injected into a reversed-phase column (electrospray ionization column [octa decyl silyl]; particle inner diameter, 75 mm; length, 100 mm; diameter 3 mm; LC Assist, Tokyo, Japan) that was eluted with a 5%-45% gradient of acetonitrile containing 0.1% formic acid for 60 min at 300 nl/min. Ionization was performed using an ion-spray voltage of 2000 V at a capillary temperature of 200°C. The mass spectrometry instrument was operated in the positive ion mode over the range of 450-1200 m/z. The MS/MS spectra were obtained in the enhanced production scan mode, and two higher-intensity peaks in each mass spectrometry scan were chosen for collision-induced dissociation.
The MS/MS data were used to search in entries under the Liza aurata category of the UniProt database using the Mascot peptide search engine. An MS tolerance of 1.0 Da for-precursor ion and an MS/MS tolerance of 0.8 Da were set as windows of processing parameters for matching peptide mass values.

Atopic dogs with cod allergy: clinical characteristics and IgE reactivity to crude cod allergens
The present study included 179 atopic dogs from 34 breeds, and there were 79 males and 100 females (age, 2 months-11 years; mean age, 3.9 years). We first examined specific IgE reactivity to crude cod allergens in the atopic dogs using ELISA and found that 20% (36/179) of the dogs exhibited increased levels of specific IgE to crude cod allergens (Figure 1). Based on diet history, the clinical reactivity to cod meat was confirmed by oral provocation in 27 atopic dogs (15%). Among the 27 dogs that underwent the food provocation test, 8 dogs that were administered grilled cod meat and 19 dogs that were administered cod-containing dog foods were diagnosed with cod allergy (i.e. dogs with cod allergy). IgE reactivity to crude cod allergens was positive in 44% (12/27) of the dogs with cod allergy (Figure 1). We next performed a field survey of commercial canine dry food products to estimate the difference in fish allergen exposure within and outside Japan and found that 75% (117/157) of the Japanese canine dry food products contained fish. In contrast, 9% (7/82) of the products produced in Australia and the USA included fish.

IgE reactivity to cod parvalbumin and collagen among atopic dogs exhibiting IgE reactivity to crude cod allergens
IgE reactivity to the purified cod allergens parvalbumin and collagen were tested by ELISA using the sera of the 36 dogs with IgE reactivity to crude cod allergens. IgE reactivity to parvalbumin was observed in 25% (9/36) of the dogs, whereas IgE reactivity to collagen was present in 33% (12/36) of the dogs. However, 56% (20/36) of the dogs showed no IgE reactivity to these purified cod allergens (Figure 2A), indicating that reaction occurred to other fish proteins. We also compared IgE reactivity to parvalbumin and collagen among different fish species by analyzing IgE reactivity to parvalbumin and collagen in four fish species in six dogs that were reactive to cod parvalbumin and eight dogs that were reactive to cod collagen.
These dogs reacted strongly to crude cod antigen and had sufficient sera to conduct the tests. IgE reactivity to parvalbumin from all fish species excluding mackerel occurred in 100% (6/6) of the dogs ( Figure 2B), whereas IgE reactivity to collagen from all fish species occurred in 100% (8/8) of the dogs ( Figure 2C).

Identification of other cod allergen components in atopic dogs with cod allergy
Two dogs with a documented clinical history of food allergy to cod meat and IgE reactivity to crude cod allergens but not cod parvalbumin or collagen were recruited (Table 1) Figure S1), which revealed that the IgE level to mite tropomyosin in the serum of dog no. 34 was also significantly higher than those in the sera of the 20 control dogs.

IgE reactivity to cod tropomyosin and crude cod allergens in atopic dogs
Using the sera of 36 atopic dogs with IgE reactivity to crude cod allergens, we determined IgE reactivity to cod tropomyosin using ELISA (Figure 4), which revealed that the IgE reactivity to tropomyosin was present in 50% (18/36) of the dogs. Therefore, 67% (12/18) of the dogs with atopic dermatitis that had high IgE levels to crude cod antigen and tropomyosin. Conversely, 25% (9/36) of the dogs showed no IgE reactivity for any of these allergens ( Figure 5).

Discussion
Although one study reported that the rate of food allergy or intolerance due to fish was only 1.3% (4/297) in dogs [23], the present study results imply that the prevalence of fish allergy might be higher (Fig. 1). Our field survey of commercial canine dry food products suggested that dogs in Japan might be exposed to fish more frequently compared with dogs in other countries such as USA. These results suggest that atopic dogs might at a higher risk of developing fish allergy due to an increase in the frequency of daily exposure for fish, implicating that atopic dogs mimic the human condition. Moreover, we showed that the rate of IgE reactivity to crude cod allergens among dogs with cod allergy was 44% (12/27) (Fig. 1), which was comparable to that reported in humans [26,27]. Atopic dermatitis with food allergy can be a manifestation of an IgE-or a non-IgE-mediated reaction to food [28,29]. Non-IgE-mediated reactions are more often delayed, in contrast to IgEmediated reactions [28]. Delayed symptoms associated with reactions were frequently observed during the provocation test in the present study. For further elucidation of these aspects of atopic dermatitis, atopic dogs that are not sensitive to crude cod allergens might be useful as a spontaneous animal model of non-IgEmediated allergy.
To our knowledge, this is the first study to describe the allergenic potency of parvalbumin and collagen in dogs. Parvalbumin has a higher allergenic potency than collagen in humans with cod allergy [2]. The present study revealed that the rate of collagen allergy was higher in dogs and that collagen elicited a stronger reactivity based on specific IgE levels compared with parvalbumin in these animals ( Fig. 2A).
One possible interpretation of this discrepancy might be due to a loss of parvalbumin from dog food via physical and chemical steps in food processing, because parvalbumin is a water-soluble protein, unlike collagen. Additionally, the sera of the dogs that exhibited reactivity to crude cod as strongly as collagen and parvalbumin showed reactivity to collagen and parvalbumin from the other fish species tested in the current study, including salmon, mackerel, and sardine ( Fig. 2B and 2C). Humans exhibit broad cross-reactivity to parvalbumin and collagen from distinct fish species [22,30], which we predict might be occurring in dogs with fish allergy as well.
Of note, the level of specific IgE to tropomyosin was higher than that for other allergenic components of cod, i.e., parvalbumin and collagen, in the dogs with cod allergy (Fig. 5). Tropomyosin was demonstrated to be a fish allergen in a study using the sera of human patients with tilapia allergy [31]. Additionally, our comparison of the protein sequence of cod tropomyosin with those of other fish species revealed that cod tropomyosin exhibited 94-99% sequence similarity with tropomyosin of other fish species that are commercially available on the market (Table S1). Tropomyosin is a major allergen that is the cause of many forms of crustacean allergy [32] as well as mite allergy [33] in humans. Although comparison of the tropomyosin protein sequence revealed a low sequence similarity between cod and shrimp (Table S1), fish-shrimp cross-reactivity was previously reported in humans [34,35]. Additionally, mite-crustacean cross-reactivity was widely reported in humans, and the tropomyosin sequence similarity between the two species is over 90% [36]. In dogs, mite is one of the most frequent sensitizing allergens, and mite tropomyosin is the allergenic component of mite allergy in canine atopic dermatitis [37]. Additionally, the serum of dog no. 34 exhibited reactivity to mite tropomyosin as strongly as cod tropomyosin ( Figure S1). Taken together, these findings raise the possibility that mite allergy in dogs with IgE reactivity to mite tropomyosin might be associated with increased cod allergy in dogs with IgE reactivity to cod tropomyosin.
The results of the current study also revealed enolase is a potential allergen associated with canine fish allergy ( Fig. 3 and Table 1). Enolase was recently defined as a fish allergen that exhibited cross-reactivity to chicken in humans and dogs [35,38]. Numerous fish proteins, other than those purified proteins that are recognized as critical allergenic components in humans, have been registered in the International Union of Immunological Societies allergen database [39]. In some cases, minor allergens in humans can be dominant allergens in dogs [18,40].
Moreover, in the present study, 25% (9/36) of the dogs with specific IgE to crude cod antigens did not show IgE reactivity to any of the three major purified allergens by ELISA, implying that other purified cod allergens might underlie fish allergy in dogs.
In conclusion, the current study revealed the presence of specific IgE to crude cod allergens as well as parvalbumin, collagen, and tropomyosin in atopic dogs. The observed IgE reactivity also revealed similarities in fish allergens between dogs and humans. Previous studies utilized experimental models of canine atopic dermatitis created in laboratory facilities [9,11,15,16]. However, globally, atopic dermatitis affects approximately 10-20% of the canine population, which often visit animal hospitals [41]. The utility of atopic dogs with reactivity to allergens that correspond with components of human allergies as spontaneous animal models can facilitate and contribute significantly to efforts to elucidate food allergy, and future studies should focus on identification of allergenic components of foods in atopic dogs.

Ethics approval and consent to participate
All experimental procedures were carried out in accordance with Japanese law and approved by the animal care and user committee of Azabu University.

Consent for publication
reviewing a draft of this manuscript.  Figure 1 Immunoglobulin (Ig) E reactivity to cod crude allergens in atopic dogs based on the levels of IgE reactivity to parvalbumin and collagen among atopic dogs exhibiting significant elevation Immunoblotting for crude cod meat allergen. In the left column, the molecular standard is sh Figure 4 Determination of reactivity to cod tropomyosin among 36 dogs with specific IgE to crude cod Figure 5 Venn diagram of the number of dogs harboring fish allergens specific IgE.

Supplementary Files
This is a list of supplementary files associated with the primary manuscript. Click to download.