Evaluation of the theoretical risk of cross‐reactivity among recently identified food allergens for dogs

Abstract

Background

There is increasing evidence of cross‐reactivity between allergens of close or distant species. The A‐RISC (Allergens'–Relative Identity, Similarity and Cross‐reactivity) index helps evaluate the risk of theoretical cross‐reactivity between proteins of the same family among different species.

Objectives

To report the A‐RISC indices for several food allergens of dogs between multiple food sources.

Materials and Methods

We selected several recently characterised food allergens for dogs from fish and chicken (ACTA1, ALDOA, CKM, ENO3, GAPDH, PKM and TPI1), fish (TPM1/2), beef/lamb (PGM1) and corn/potato (WAXY). When quality sequence data were available, A‐RISC indices were calculated between multiple animal and plant species that can be used as food sources. For the TPM subunits, A‐RISC indices also were calculated with the environmental allergens Bla g 4 and Der f 10, and the Toxocara canis nematode.

Results

The A‐RISC indices suggest a substantial theoretical risk of cross‐reactivity between species for all allergens considered. For TPM, this risk also extends to the environmental and nematode allergens.

Conclusions and clinical relevance

There is a high theoretical risk of cross‐reactivity between allergens of different species used as food sources. The clinical relevance of these elevated A‐RISC indices should be studied further.

Background: There is increasing evidence of cross‐reactivity between allergens of close or distant species. The A‐RISC (Allergens'–Relative Identity, Similarity and Cross‐reactivity) index helps evaluate the risk of theoretical cross‐reactivity between proteins of the same family among different species. Objectives: To report the A‐RISC indices for several food allergens of dogs between multiple food sources. Conclusions and clinical relevance:

There is a high theoretical risk of cross‐reactivity between allergens of different species used as food sources. The clinical relevance of these elevated A‐RISC indices should be studied further.

Abbreviations

A‐RISC

Allergens' Relative Identity, Similarity and Cross‐reactivity

FA

food allergies

INTRODUCTION

At this time, the diagnosis of food allergies (FA) in pets relies on the performance of an open, two‐phase, restriction–provocation trial. ¹ During the restriction phase, the pet normally is given a home‐made or commercial diet formulated with ‘novel’, ‘limited’, ‘restricted’ or ‘single protein’ ingredients, or protein digests of variable degree of hydrolysis (i.e. hydrolysates). Because of the higher costs of the latter, clinicians often opt for the lower‐priced ‘novel protein’ diets. Still, such a selection presumes that the novel diet does not share any cross‐reactive allergens with the originally eaten food.

Unfortunately, each publication characterising new food allergens in dogs raises increasing concerns that an extensive cross‐reactivity might exist among food allergens. Indeed, the main food sources causing allergies in dogs are meats of mammalian, poultry or piscine origin, ² and, consequently, newly characterised immunoglobulin (Ig)E‐targeted allergens are muscle or blood proteins. While the first discovery of meat allergens in dogs suggested the immunological (IgE) cross‐reactivity between two allergens of evolutionary‐close species (e.g. beef, lamb and cow's milk), ³ a recent study established the presence of multiple IgE‐cross‐reactive allergens between chicken and two evolutionarily distant fish species. ⁴ In 2022, we reported the identification of eight chicken proteins targeted by serum IgE from dogs with clinical signs of chicken allergy. ⁵ For one of these allergens (chicken serum albumin, ALB, Gal d 5), we used a novel index the A‐RISC (Allergens'–Relative Identity, Similarity and Cross‐reactivity) ⁶ to help determine the theoretical risk of cross‐reactivity of this chicken allergen with homologous proteins from other poultry, mammalian, fish and reptile (alligator) species. As the A‐RISC indices often were higher than 50% (0.50), the chance of cross‐reactivity among albumins thus was predicted to vary from medium to very high. ⁵ While a clinical cross‐reactivity was not proven—or even investigated—in the dogs from that study, these elevated A‐RISC indices raised the question that a cross‐reactivity might exist for some food‐allergic dogs. Indeed, if a dog were IgE‐sensitised to chicken's Gal d 5, there is a risk that its clinical signs might not decrease following the feeding of, for example, alligator‐ (A‐RISC = 0.68) or rabbit‐ (A‐RISC = 0.52) based diets.

This paper expands the A‐RISC index calculations to other previously published meat and plant allergens. The values obtained suggest a very high theoretical risk of cross‐reactivity among food allergens, even from evolutionarily distant species.

MATERIALS AND METHODS

From previous reports of IgE‐targeted allergens in sensitised dogs, ³ , ⁴ , ⁵ we selected the following meat allergens: alpha‐actin (ACTA1), aldolase A (ALDOA), creatine kinase M (CKM), (beta) enolase 3 (ENO3), glyceraldehyde‐3‐phosphate dehydrogenase (GAPDH), l‐lactate dehydrogenase (LDHA), phosphoglucomutase 1 (PGM1), protein‐kinase M (PKM) and triosephosphate isomerase 1 (TPI1). We excluded Bos d 7 (bovine IgG), the first beef/milk allergen identified in dogs, ³ due to its multiple‐chain structure and the presence of IgG subclasses, making comparisons between species difficult. We also omitted Bos d 6 (bovine serum albumin), which we discussed in detail in our previous paper. ⁵ Whenever good quality amino acid sequences were available, A‐RISC indices were calculated, as done previously, ⁶ between the following species: chicken (Gallus domesticus), duck (Anas spp.), turkey (Meleagris gallopavo), ostrich (Struthio camelus), cow (Bos domesticus), sheep (Ovis aries), horse (Equus caballus), pig (Sus domesticus), rabbit (Oryctolagus cuniculus), alligator (Alligator Mississippiensis), salmon (Salmo salar) and cod (Gadus morhua).

As an illustration that cross‐reactivity phenomena are not limited to meat allergens, we added the IgE‐targeted granule‐bound starch synthase 1 (WAXY), which is found in starches from several hydrolysate‐containing commercial diets. ⁷ For this allergen, we determined the A‐RISC indices between the following plant foods: wheat (Triticum aestivum), barley (Fagopyrum esculentum), rice (Oryza sativa), corn/maize (Zea mays), potato (Solanum tuberosum) and pea (Pisum sativum).

Finally, to demonstrate that cross‐reactivity also exists between food allergens and proteins from the same family in evolutionarily distant insect, arachnid and nematode species, we calculated the A‐RISC indices for the two subunits of the tropomyosin allergen (TPM1 and TPM2) between homologous proteins of the animal species described above, to which we added those of the Dermatophagoides farinae house dust mite (Der f 10), the German cockroach (Blatella germanica, Bla g 7) and the ubiquitous Toxocara canis roundworm.

RESULTS

In Figure 1a, we show the A‐RISC indices for LDHA, which we selected as a representative meat allergen. The theoretical risk of cross‐reactivity between the examined species was high to very high, because indices ranged from 0.77 to 0.98. As can be expected, the indices were highest between evolutionarily close species (i.e. between selected mammals, poultry or fish) and lower between these different groups. This segregation between groups became more obvious when plotting the different sequences of LDHA against that of chicken (Figure 1b).

FIGURE 1.

A‐RISC (Allergens'–Relative Identity, Similarity and Cross‐reactivity) indices for l‐lactate dehydrogenase (LDHA) and granule‐bound starch synthase 1 (WAXY). (a) A‐RISC indices for LDHA between 12 animal species can be used as different food sources. (b) When compared to chicken, the closest LDHA sequences are those of birds and alligators, then mammals and then fish species. (c) A‐RISC indices for WAXY between six plants used as food sources; (d) when compared to wheat, the WAXY sequence of barley is closest; then, we find those of rice and corn and then potato and pea.

Nearly identical results were obtained when calculating A‐RISC indices for the other meat allergens ACTA1, ALDOA, CKM, ENO3, GAPDH, PGM1, PKM and TPI1 (Figure S1).

For our representative plant food allergen, the GBSS1 (WAXY), the theoretical risk of allergen cross‐reactivity also was substantial between examined species (Figure 1c); the A‐RISC indices were highest among grains and also elevated between grains, potato and pea. When compared to the wheat's WAXY sequence, that of barley was the closest, followed by corn and rice, and then potato and pea (Figure 1d).

Finally, we calculated A‐RISC indices for the two chains of TPM, a well‐known panallergen for humans and a recently found fish allergen in dogs. ⁸ As can be seen in Figure S1, the high‐value indices suggested a substantial risk of cross‐reactivity between mammal, poultry and fish TPMs. This risk also was important between the TPMs of these species and two inhalant allergens, Bla g 7 and Der f 10, the latter being a minor mite allergen for dogs. ⁸ Finally, A‐RISC indices indicated a possible cross‐reactivity, albeit lower, between all of these TPMs and that of Toxocara canis.

DISCUSSION

We report herein the theoretical risk of cross‐reactivity between several allergens from multiple animal and plant species using the newly developed A‐RISC index. ⁶

As meats represent the main sources that trigger food allergies in dogs and cats, ² we first focussed on the muscle proteins recently characterised as allergens. ³ , ⁴ , ⁵ , ⁸ The A‐RISC indices among species were high to very high, thus suggesting a substantial likelihood of cross‐reactivity. The additional A‐RISC calculations for the WAXY and TPM allergens further highlighted that such cross‐reactivity is not limited to meat allergens and also might occur between meat, nematode and environmental allergens.

It is unknown if these theoretical predictions will translate into a clinical cross‐reactivity when dogs eat foods from different origins, yet there is cause for concern.

In some cases, the pertinence of the A‐RISC already is supported by immunological and clinical data. For example, the A‐RISC index between the ENO3 of chicken (Gal d 9) and those of cod (Gad m 2) and salmon (Sal s 2) is 0.84, which suggests a high theoretical risk of cross‐reactivity between these two types of meats. The recent demonstration of cross‐reactivity between chicken and fish at the immunological (humans and dogs) ⁴ , ⁹ and clinical (humans) ⁹ levels validates the relevance of the notional risk determined by the A‐RISC indices.

Although theoretical, the potential cross‐reactivities between the allergens of all these foods should prompt the reexamination of the current practice of performing restrictive diets for diagnosing food allergies with ‘novel, limited, restricted’ food sources. Indeed, on the one hand, if clinical signs improve with these diets and recur after provocation with the original food, clinicians can confidently diagnose food allergies. However, on the other hand, if clinical signs do not diminish with the newly introduced diet, the clinician will be left to wonder if the diagnosis of food allergy can be excluded, or alternatively, if the lack of improvement is due to the feeding of cross‐reactive food, environmental or nematode allergens to which it is sensitised.

Consequently, the performance of food trials with ‘novel’ diets is inherently fraught with a sensitivity for the diagnosis of FA that probably is lower than that currently expected; it is thus evident that the real prevalence of FA in animals likely is higher than that reported. ¹⁰

In conclusion, we report herein the elevated theoretical risk of allergenic cross‐reactivity between allergens of multiple animal and species used as food sources for pets. Further studies should help determine whether these risks correlate with clinical allergy.

AUTHOR CONTRIBUTIONS

Thierry Olivry: Conceptualization; Writing – original draft; Supervision. Andrea O’Malley: Investigation; Writing – review & editing. Maksimilian Chrusz: Investigation; Methodology; Validation; Visualization; Writing – review & editing.

ACKNOWLEDGEMENT

None.

FUNDING INFORMATION

This study was self‐funded.

CONFLICT OF INTEREST

No conflicts of interest have been declared.

Supporting information

Figure S1

Olivry T, O’Malley A, Chruszcz M Evaluation of the theoretical risk of cross‐reactivity among recently identified food allergens for dogs. Vet Dermatol. 2022;33:523–526. 10.1111/vde.13110