Surplus canine and feline sera submitted to Avacta Animal Health (Wetherby, Leeds, UK) by veterinarians for food allergen-specific IgE serological testing were used for this study. For dogs, the panel consisted of beef, pork, lamb, duck, chicken, turkey, wheat, soybean, barley, rice, potato, corn, oat, cow’s milk, whole hen’s egg, white fish, venison, salmon and rabbit. For cats, the panel included beef, pork, lamb, duck, chicken, turkey, rabbit, salmon, tuna, white fish, wheat, soybean, rice, corn, cow’s milk and whole hen’s egg. All dog and cat sera had been stored at −80 °C for a maximum of two and six years, respectively. Canine sera were selected according to their original levels of serum chicken-specific IgE (IgE reactivities expressed as optical density, OD, values at 405 nm, OD405 uncorrected for background) in the Sensitest commercially available ELISA (Avacta).
The D1 group (30 dogs) contained the following sera:
D1-LCR: ten dogs with low chicken IgE reactivity and a previous OD405 between 0.3 and 0.4 in the Sensitest assay.
D1-MCR: ten dogs with moderate IgE chicken reactivity (OD405 between 0.5 and 0.9)
D1-HCR: ten dogs with high chicken IgE reactivity (OD=405 > 1.0)
In their previous ELISAs done at the time of the original submission, all of the dogs in this group had detectable IgE to at least one food extract; one dog was only positive to chicken and none had any detectable IgE reactivity to beef.
D2-NCR: This group included sera from ten dogs with undetectable levels of IgE to chicken (i.e. “non-chicken reactive dogs” NCR; OD405 < 0.3). By previous ELISAs, these dogs had no visible IgE reactivity to any food tested on the panel.
Similarly, we selected feline sera based upon their previously determined levels of serum chicken-specific IgE (OD405, uncorrected for background).
The C1 group was composed of 31 cats divided into the following subgroups:
C1-LCR: eleven cats with low chicken IgE reactivity (OD405: ~0.2–0.4)
C1-MCR: ten cats with moderate chicken IgE reactivity (OD405: ~0.5–1.0)
C1-HCR: ten cats with high chicken IgE reactivity (OD=405: >1.0)
By previous ELISAs, all of the cats in this group had at detectable IgE to at least three foods. All cats had measurable IgE against beef but none had IgE reactivity to wheat.
Finally, the C2-NCR group comprised sera from nine cats with undetectable levels of IgE to chicken in the Sensitest ELISA (OD405: <0.2). By previous ELISAs, there was no IgE reactivity to any food in these cats.
Details from all selected canine and feline sera are found in the Additional file 1.
Raw chicken meat (CMT), turkey meat (TMT), duck meat (DMT), beef meat (BMT) and wheat (WHT) extracts were purchased from Greer Laboratories (Lenoir, NC, USA). Non-hydrolysed chicken meal (NHCM, the ground, rendered clean parts of the carcasses of slaughtered poultry), mildly-hydrolysed poultry feathers (MHPF) and extensively-hydrolysed poultry feathers (EHPF) were obtained from Royal Canin. The main difference between the MHPF and EHPF was the presence of residual proteins of molecular weight superior to 10 kDa in the former, but not in the latter; the EHPF were those used in the RCA. The commercial-grade material was defatted in acetone, filtered and then dried. Allergens were obtained after two series of successive extractions in phosphate-buffered saline (PBS), followed by centrifugation and, ultimately, by dialyzing the supernatant against distilled water. The protein concentration of each extract was determined at 280 nm by Trinean DropSense96 (Trinean, Gent, Belgium). The final protein concentrations of the NHCM, MHPF and EHPF extracts were 6.8, 3.1 and 2.4 mg/ml, respectively. Extracts were stored at 4 °C before use.
Canine and feline food allergen-specific IgE serum levels were determined by ELISA. Microtiter plates (Costar, Corning, NY, USA) were coated overnight at 2–8 °C with 50 μL/well of each extract, diluted to a concentration of 5 μg/mL of protein in 0.05 M carbonate/bicarbonate buffer, pH 9.6.
Plates were washed three times with 150 μL/well Tris-buffered saline (TBS) with 0.05% Tween-20 (TBST) and blocked with 150 μL/well TBS containing 0.5% sucrose and 0.5% PVP10 (block) for 2 h at room temperature. After removal of the block, plates were dried at 37 °C for 2 h. Canine and feline sera were diluted 1:10 in TBST, and incubated with coated extracts overnight at 2–8 °C (50 μL/well. For standardization of the canine and feline ELISAs, serial three-fold dilutions of pools of dog and cat reference sera with high levels of anti-beef IgE (for dog tests) or chicken (for cat tests) were included on each plate. Undiluted, these two standard pools were assigned a value of 500 arbitrary units (AU). Positive and negative serum controls, with moderate and negligible levels of beef- (dog tests) or wheat- (cat tests) specific IgE, respectively, were also diluted at 1:10 in TBST and included on each plate. Plates were washed with TBST, as before, and incubated with 0.5 μg/mL alkaline phosphatase-labelled anti-dog IgE (clone 5.91; Bruce Hammerberg, NC State University, Raleigh, NC, USA) for 2 h at RT. This mouse monoclonal antibody has been described previously  and shown to recognize IgE from several mammalian species including cats . The specificity of this antibody for dog IgE was confirmed in previous ELISAs, as a strong signal (OD405: 1.8) was found on a 0.5 μg/mL dog IgE coat without any concurrent detection (OD405: 0.11) of dog IgG coated at up to 20 μg/mL. After the final three washes with TBST, plates were developed with 50 μL/well of an alkaline phosphatase substrate (pNPP; BioFx Laboratories, Owings Mills, MD, USA) for 30 min at room temperature. The reactions were stopped by the addition of 50 μL/well 1 M NaOH, and OD405 were determined using a microplate reader (Tecan, Männedorf, Switzerland). For both ELISAs, standard curves were generated by fitting the mean standard uncorrected OD405 to a sigmoidal 4PL curve (Prism, Graphpad Software, La Jolla, CA, USA) with log10 concentrations of the standard dog or cat serum pools.
Inhibition ELISAs were performed to determine if NHCM, MHPF and EHPF extracts could block serum IgE binding to poultry extracts, thereby demonstrating the presence of shared IgE binding epitopes. Ten dog sera that showed the highest IgE reactivity to CMT, DMT and TMT in the ELISAs described above (i.e. the HCR sera) were selected for the inhibition assays. Due to insufficient volumes available, inhibitions were not performed with the feline sera. Inhibitor solutions of NHCM, MHPF and EHPF were prepared in TBST, each to a concentration of 500, 1000 and 2000 μg/mL (i.e. 100×, 200×, and 400× coat level, respectively). To serve as the positive inhibitor, the CMT, DMT and TMT were combined, as a three-meat extract in a 1:1:1 ratio of equal protein levels, to the same concentrations as that of the extracts above. As a negative inhibitor, we diluted the BMT extract to the same concentrations. Dog sera were pre-incubated overnight at 2–8 °C, at a dilution of 1:10, with 50 μL each of the three increasing concentrations of the test and control inhibitor solutions. Additionally, each dog serum was diluted at 1:10 in TBST in the absence of inhibitor. All serum-inhibitor mixes and no-inhibitor controls were then tested at 50 μL/well by ELISA on the CMT, DMT and TMT extracts, as described above. For any given concentration of inhibitor, the percentage inhibition was calculated as follows: 100 – ((OD405 of serum with inhibitor/OD405 of serum without inhibitor) ×100) where OD405 represents the background-corrected absorbance at 405 nm.
The extracts above, thawed after two years of freezing, were heated at 70 °C for 10 min in the presence of 5% Criterion XT reducing agent (Bio-Rad Laboratories, Hercules, CA, USA) before sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). Proteins (5 μg per lane) were separated on 4–12% Bis-Tris Criterion XT precast gels (Bio-Rad) at 180 V using MES (4-morpholineethane-sulfonic acid) running buffer on a Criterion Cell system (Bio-Rad). Molecular weight markers (10–190 kDa; PageRuler Plus, Thermo Scientific) were run in parallel to the extracts. Following electrophoresis, the gels were either stained using InstantBlue Coomassie stain (Sigma Aldrich, Dorset, UK) to verify protein loading and proteins were transferred to a polyvinylidene difluoride (PVDF) membrane (BioRad). This was performed at 100 V for 1 h in Tris/glycine transfer buffer (Bio-Rad) containing 20% methanol using Criterion Blotter apparatus (Bio-Rad). The membrane was blocked with TBS containing 0.5% PVP10 at room temperature for 2 h, and then incubated with a serum pool from three dogs with high chicken-specific IgE levels (>OD405: 1.0, by previous ELISAs), diluted at 1/10 in TBST. The membrane was incubated with serum at RT overnight. After washing three times with TBST, the membrane was incubated with alkaline phosphatase-conjugated anti-dog IgE antibody (clone 5.91, 0.5 μg/ml in TBST) at room temperature for 2 h with shaking. The membrane was washed three times with TBST and rinsed twice prior to addition of nitroblue tetrazolium/5-bromo-4-chloro-3-indolyl-phosphate substrate (Thermo Scientific). The reaction was stopped with deionised water and blots were then dried at room temperature. Images of gels and blots were captured using a G:BOX Chemi-XR5 gel imaging system (Syngene, Cambridge, UK).
Data were analyzed with SAS v9.3 (SAS, Cary, NC, USA). The frequencies of canine or feline positive ELISA tests with the various extracts were compared with permutation tests-based on Fisher’s exact test with a resampling size of 1000 times. Then, the alpha inflation risk subsequent to multiple extract comparisons was controlled with false discovery rate adjustments. The level of statistical significance was set atP < 0.05, for two-sided analyses.