Cornstarch is less allergenic than corn flour in dogs and cats previously sensitized to corn

Cutaneous adverse food reactions (CAFR) represent one of the most common diagnoses given to dogs and cats with pruritus or allergic skin diseases [1]. To echo the terminology used in human medicine [2], the term CAFR was first coined for animals with skin manifestations of food reactions that included both food allergies (that is, adverse health effects arising from a specific immune response occurring reproducibly on exposure to a given food) and food intolerances (e.g., reactions occurring following non-immune mechanisms). While we could not identify any well-documented case of non-immunologic food intolerance in dogs and cats in the veterinary literature, there are numerous studies suggesting an immunologic foundation to canine and feline CAFRs. Indeed, there is documentation of either food-specific IgE hypersensitivity and/or food-activated T cells in such patients (reviewed in [3]). That most dogs and cats with CAFR likely suffer from bona fide food allergies rather than food intolerances opens new avenues for a better diagnosis and treatment. For example, characterizing allergenic food sources and the main culprit allergens would likely lead to more accurate restriction-provocation trials, IgE serological and cell-mediated diagnostic assays. Furthermore, one could then also envision the possible development of oral allergen immunotherapy and the design of hypoallergenic pet foods for the long-term management of food allergic dogs and cats.

A recent critically appraised topic concluded that the most common allergenic food sources in dogs and cats are derived from mammalian, poultry, and fish meats, while grains are uncommon culprits [4]. Among the latter, corn was found to be an offending allergen in less than 5% of dogs and cats with food allergies [4].

Recent proteomic studies markedly expanded the list of possible IgE-targeted corn allergenic proteins for humans [5–7]. An important observation of the most recent studies was that corn allergens appear to vary between plant cultivars as well as between allergic human individuals [6, 7]. The major corn allergen Zea m 14, recognized by nearly 90% of patients, is a 9 kDa lipid transfer protein present in corn kernels [8].

To date, there is only limited information available on the identity of corn allergens in dogs, while none has been reported yet in cats. Cornstarch is frequently added to commercial pet foods, including extensively-hydrolyzed ones. Whilst there is no clinical evidence to suggest its inclusion in such diets would be problematic for animals sensitized to corn, in a recent study evaluating the proteins present in three commercial hydrolysate-containing pet foods, granule-bound starch synthase 1 (GBSS-I, an NDP-glucose-starch glucosyltransferase) was the only corn protein identified as an allergen by serum IgE from a small fraction of dogs with suspected nonseasonal allergy [9]; of importance is that GBSS-I is present in cornstarch.

The objectives of this study were to determine if IgE from corn-sensitized dogs and cats recognized proteins present in corn flour and starch. We will show herein that, presumably due to its very low protein content and limited proteome, cornstarch is hypoallergenic compared to its flour.

In this serological study using two different techniques, IgE detected allergens in corn flour only in dogs and cats having moderate and high IgE serum levels against a corn kernel extract. In contrast, none of the tested sera had detectable IgE against cornstarch proteins. As such, cornstarch should be considered hypoallergenic compared to its flour in dogs and cats with detectable serum corn-specific IgE.

These results are not surprising, as they echo the clinical indings of experimental challenges in corn-sensitized dogs [11]. In this study, 14 Maltese-beagle atopic dogs with historical clinical hypersensitivity to corn and soy were challenged, in three separate phases of two consecutive administration days, with 200 mg/kg of cornmeal (i.e., kernels ground to a coarser texture than flour), cornstarch and soy [11]. While ten of these 14 dogs (71%) exhibited a flare of clinical signs after eating the cornmeal, only three (21%) flared with cornstarch, these three dogs also having reacted to cornmeal. Consequently, only three of ten dogs (30%) clinically-reactive to cornmeal had a flare after eating the cornstarch [11].

Even though cornstarch contains approximately less than one-twentieth of the protein amount of corn flour (0.3 vs 6.9 g of protein/100 g; http://nutritiondata.self.com; page last accessed October 31, 2017), the lower allergenicity that was seen with the starch is unlikely due to a difference in protein concentration between these foodstuffs. Indeed, we performed ELISAs and immunobloting with the same protein concentration (5 μg/ml) for all extracts. Consequently, the difference in the allergenicity of corn byproducts most likely resides in the different protein composition existing between flour and starch.

While corn flour is made from milling whole dried kernels, cornstarch mostly contains the kernel’s starch granule-rich endosperm. In addition to amylose and amylopectin, seed endosperms also include storage proteins for the embryo as well as structural, metabolic and starch granule-associated proteins (SGAPs) [12].

Recent proteomic studies employed 2D-electrophoresis and mass spectrometry to characterize corn kernel allergens recognized by IgE from corn-allergic human sera [5–7]. Altogether, more than 20 allergens were uncovered in corn flour [5–7]. At the time of this writing, the allergen standardization international registry (www.allergen.org; site last accessed October 31, 2017) only lists three of these corn kernel allergens (Zea m 8, 14 and 25).

Matching the recently identified corn kernel allergens [5–7] to proteins characterized in a cornstarch proteome confirms the relative hypoallergenicity of cornstarch compared to its parent flour [13]. Indeed, the starch appears to contain only half of the newly-recognized corn kernel allergens [13]. Five of these are enzymes involved in carbohydrate metabolism: the GBSS-I (NDP-glucose-starch glucosyltransferase), the UTP-glucose-1-phosphate uridylyltransferase, the cytoplasmic isozyme of fructose-bisphosphate aldolase, the cytosolic triosephosphate isomerase and the malate dehydrogenase [5–7, 13]. Other cornstarch allergens are seed storage proteins: the globulin 2 (Zea m G2), the vicilin-like embryo storage protein (Zea m G1) and several proteins of the zein and glutelin families (including Zea m 27). Finally, the starch allergen thioredoxin H2 (Zea m 25) is involved in protein folding, sorting and degradation [13]. Importantly, cornstarch does not appear to contain the major corn kernel allergen, the 9 kDa lipid-transfer protein Zea m 14 [8], a finding that we confirmed with the likely detection of this allergen with its typical 9–10 kDa band in the flour but not in the two tested starches.

While the specific amounts of each of these allergenic proteins were not reported in the two cornstarch proteomic articles [13, 14], one can infer from the published data that over half of the SGAPs are enzymes involved in carbohydrate metabolism [13]. In the earlier study, the protein(s) present in largest amount in cornstarch were the granule-bound starch synthase(s) (GBSSs) [14]. The granule-bound starch synthase family was later confirmed to compose up to 85% of the total internal SGAPs [13]. Similarly, mass spectrometry analysis of the Royal Canin Anallergenic itself, which contains hydrolyzed poultry feathers and the purified cornstarch tested herein, identified only residual GBSS-I [15]. It is likely that, during the desiccation of the endosperm for starch processing, these normally cytoplasmic amylogenesis enzymes remained bound to the remnants of the amyloplast membrane of the starch granules [16].

Importantly, the GBSS-I was reported recently to be an allergen for both corn-sensitized humans [7, 17] and dogs [9]. Unfortunately, because of the high homology existing between GBSSs from starch-containing plants, that the GBSS-I is an allergen leads to a high risk of IgE cross-reactivity, not only among grains [18], but also likely between grains and tubers [9, 19]. Whether such cross-reactivity would be clinically-relevant in corn-allergic humans or dogs has not been reported yet.

That proteins like GBSS-I are recognized corn allergens does not mean that they could induce flares in corn-allergic individuals, be it an animal or a human. There are at least three factors that could reduce its allergenicity: 1) the nature of the targeted epitopes, 2) food processing and 3) the amount of allergen consumed.

While most allergic patients have IgE that recognize epitopes on the peptidic moiety of proteins, others bind to glycans on complex glycoproteins (reviewed in [20]). The most common situation is when IgE targets classic xylose-containing cross-reactive carbohydrate determinants (CCDs) that are shared among proteins of a plant, insect and parasite origin [20]. Identifying that IgE target classic CCDs is essential, as they could be the source of nonspecific IgE seropositivity as such IgE are believed to be mostly nonpathogenic in humans [20]. To address this issue, we previously evaluated if the reactivity of corn-specific IgE in dogs was specific for glycans. Sera from seven dogs with high corn-specific IgE were tested against the same SCE as above, the CCD-rich bromelain (a protein from pineapple) and a high-amylose cornstarch different from those examined in the current study; testing was done by ELISA before and after deglycosylation of the extracts using a standard periodate oxidation protocol. While the median post-deglycosylation reduction of IgE reactivity to the SCE was only about 20%, that to bromelain was close to 50%, and that to the high-amylose cornstarch was above 80% (J. Bexley: unpublished data). Results of this pilot study suggest that, even if present, most of the IgE reactivity to cornstarch proteins appears directed against CCDs, and, as such, might not be pathogenic or relevant.

Food processing is another reason why allergens would no longer be recognized by IgE from clinically-reactive patients, for example, if allergens were thermolabile. This phenomenon has been shown recently to also occur in dogs suspected of having a food allergy: raw proteins seemed more allergenic than cooked ones [21, 22]. Conversely, it is possible, at least theoretically, that the detection of the GBSS-I by IgE in the study by Roitel and colleagues [9] might be due to allergenic neo-epitopes being induced by heat and processing during pet food manufacturing, while, in our study, we used corn starches before their inclusion and processing in the diet itself.

Furthermore, the clinical reactivity of food-allergic humans is mainly allergen-dose-dependent, with protein doses eliciting reactions varying markedly between children and adults or between food sources [23]. For example, while less than 0.1 mg of milk proteins might be sufficient for some milk-allergic children to flare, adults might need 50 times that amount; for shrimp, more than 10 mg of protein would be required for 5% of adults to react [23]. How much GBSS-I would be eaten by a dog fed a cornstarch-containing pet food and whether its allergenicity would survive after food processing is unknown. Nevertheless, our immunoblotting studies showed that, while some corn-sensitized dogs had IgE directed against SCE and SCF proteins of a molecular weight compatible with that of the GBSS-I (55–60 kDa), such bands were remarkably absent with the two tested starch extracts. These results suggest that, even if detectable by mass spectrometry in the Royal Canin Anallergenic [15], the amount of residual GBSS-I present is likely minute. It is only by challenging corn-allergic dogs with cornstarch processed in pet foods that the real determination of the allergenicity of GBSS-I will be thoroughly assessed.

Finally, the immunoblotting experiments provided valuable data on the molecular weights of several major corn allergens recognized by corn-sensitized dogs and cats. It is hoped that the sequencing of these bands will help precisely determine the identity of such allergens and their respective amount in corn-derived products such as flour and starch.

In this study, we established, using both ELISA and immunoblotting, that IgE from corn-sensitized dogs is less likely to recognize cornstarch proteins than those present in kernel and flour extracts. As corn is not yet known as a common allergen source in dogs and cats, and as its starch seems to be less allergenic than the flour, pet foods containing cornstarch as a carbohydrate source might be preferable for pets suspected of corn allergies.