Soybean protein (SP) extract was obtained from Glycine max L. Merr. seeds as described previously.24 Briefly, seeds were crushed and extracted with 0.01 N NaHCO₃ at 90℃. The extract was centrifuged at 2,500×g for 20 minutes at room temperature and lipids were extracted with chloroform overnight at 4℃. The extract was dialyzed against distilled water and stored at -20℃ until use.

CMP extract was obtained from commercial skimmed milk. Proteins were extracted with phosphate-buffered saline pH 7.4 (10 mg/mL) and filtered. The extract was stored at -20℃ until use. The presence of soy components in the CMP extract was previously discarded by indirect ELISA with SP-specific rabbit antiserum.

Sera from 10 pediatric patients diagnosed with milk allergies according to history, skin prick test, and serum specific IgE were used. Double-blind placebo-controlled food challenge is not performed in Argentina for diagnosis; instead, elimination of milk from the diet and open challenge is performed. Soy allergy was ruled out based on history and serum IgE against SP. Sera from healthy individuals with no allergy history and normal level of serum IgE, or from patients allergic to aeroallergens with no CMP-specific IgE antibodies or history of food allergy, were included as negative controls.

Three monoclonal antibodies (mAb) with differential specificities for α-casein (1D5), β-casein (4C3) and κ-casein (3B5), which were characterized previously,30 were used.

The cDNA coding sequence for P34 (GenBank: DQ324851) was obtained by PCR amplification of the cDNA library.31 Amplified PCR products were cloned directionally into pENTR/D TOPO (Life Technology, S.A. Argentina) and were then transferred to the pDEST-maltose-binding protein (MBP) destination vector for expression.32 E. coli BL21 Codon Plus containing the constructs pDEST His-MBP-P34 and pDEST His-MBP were induced and recombinant proteins were purified as described previously.24 Depletion of lipopolysaccharide was performed with a Sepharose-polymyxin B resin (Pierce, Rockford, IL, USA). Endotoxin determination was performed using the Limulus amoebocyte chromogenic assay (LONZA, Buenos Aires, Argentina). Protein concentration was determined using the bicinchoninic acid assay with bovine serum albumin as a standard (Pierce, Rockford, IL, USA).

Polystyrene microtitre plates were coated with 0.25 µg/well α-casein (Sigma-Aldrich, MO, USA) and blocked with 5% horse serum. Purified 1D5 mAb was separately incubated with various quantities of the soluble inhibitor α-casein, P34 or ovalbumin (OVA) as a non-related protein (2 hours at 37℃). This pre-mixed dilution was then added to the coated wells (30 minutes at 37℃). Finally, the secondary antibody was added (1:3,000, 1 hour at 37℃) and color was developed with o-phenylenediamine. Optical density (OD) was measured at 492 nm.

Basophils were obtained from whole heparinized blood collected from non-allergic human donors using ammonium-chloride-potassium (ACK) buffer (0.15 M NH₄Cl, 10 mM KHCO₃, 0.1 mM Na₂EDTA, pH 7.2) as erythrocyte lysis buffer. Total cells were re-suspended in ice-cold lactic acid buffer pH 3.9 (13.4 mM lactic acid, 140 mM NaCl and 5 mM KCl) for 5 minutes for stripping. Cells were washed and re-suspended again in 2 mL of saline buffer containing 20% serum from milk allergic patients (5 sera), 4 mM Na₂EDTA and 10 mg/mL heparin (90 minutes at 37℃). Cells were washed with 20 nM HEPES buffer pH 7.4 and incubated with HEPES buffer containing 1 mM CaCl₂ (30 minutes at 37℃).33 Cells were then challenged with whole milk proteins, SP or the recombinant P34 soy allergen at various concentrations. Saline buffer was used to stop cell activation. Finally, washed cells were incubated with 7-aminoactinomycin (7-AAD) (BD Pharmingen, USA) and biotinylated anti-human IgE monoclonal antibody (Vector Laboratories Inc., CA, USA), streptavidin-allophycocyanin (APC) (eBioscience, USA), phycoerythrin (PE)-conjugated antibody specific to human CD63 (BD Pharmingen, USA), or monoclonal antibodies specific to human CD203c (Macs, Miltenyi Biotec, Germany), followed by goat anti-mouse IgG fluorescein isothiocyanate (FITC) (Santa Cruz Biotechnology, USA). Fluorescence data were acquired with a Becton Dickinson FACSCalibur flow cytometer (Franklin Lakes, NJ, USA) and analyzed with the BD CellQuest Pro software and the FlowJo software (Tree Star Inc., Ashland, OR, USA). Controls with a non-related protein (OVA) or with sera from non-allergic subjects were included.

The research was conducted in strict agreement with the international ethical standards for animal experimentation (Helsinki Declaration and its amendments, Amsterdam Protocol of welfare and animal protection and National Institutes of Health, USA NIH, guidelines: Guide for the Care and Use of Laboratory Animals) and was approved by the local Institutional Animal Care and Use Committee at the School of Animal Science (University of La Plata).

Statistical analyses were performed using the GraphPad Prism 5 software. The significance of differences was determined using an independent-sample t-test or ANOVA test. A P value <0.05 was considered to indicate statistical significance.

Human sera containing specific IgE antibodies were used to assess the recognition of recombinant P34. Fig. 1A shows that P34 was recognized as a coated antigen by 9 of 10 sera from milk allergic patients. Additional bands due to partial degradation of P34 were detected with some sera. No binding of MBP with sera from allergic patients (#3) or of P34 with sera from non-allergic subjects and non-milk-allergic patients (#11 and 12) was detected. To rule out co-sensitization in patients, immunoblotting was performed using three casein-specific monoclonal antibodies; all mAbs revealed a band corresponding to P34 (Fig. 1B). MBP was not recognized.

To rule out the possibility of artifactual reactivity due to new epitopes created during antigen coating, P34 was used as a soluble inhibitor in a competitive ELISA using the α-casein-specific 1D5 mAb (Fig. 1C). The sigmoid-shape of the dose-response inhibition curve obtained with P34 as an inhibitor supported the specificity of the antigen-antibody reaction. Antibody binding to the immobilized α-casein (0.25 µg/well) was inhibited by 100% with 0.002 mg/mL of soluble α-casein, while 50% inhibition (IC50) was achieved with 0.0002 mg/mL of α-casein and 0.05 mg/mL of P34. No inhibition was observed over a wide range of OVA concentrations.

To confirm the cross-reactivity with human IgE antibodies, we performed BAT using sera from milk allergic patients. Stripped basophils from healthy donors were passively sensitized with individual sera containing IgE and then activated with various concentrations of milk or soy proteins. A minimum of 100,000 events per sample were analyzed by flow cytometry and cells were gated based on physical properties (Fig. 2AI), 7 AAD-live cells (Fig. 2AII), and IgE⁺ cells (Fig. 2AIII), and were further analyzed for membrane CD63 and CD203c. Isotype controls were used to establish fluorescence thresholds. Percentages of double-positive cells were compared between basophils incubated with the allergen or PBS, CMP, and SP induced a higher percentage of IgE⁺ CD63⁺ CD203c⁺ cells compared with PBS alone (18.45%±1.35% and 18.80%±0.5% vs 9.16%±1.26%, respectively) (Fig. 2B). When P34 was used for IgE-dependent basophil activation, we observed 18.5%±0.45% double-positive cells. The non-related protein OVA at various concentrations resulted in 8.37%±0.31% double-positive cells. In addition, there were no statistically significant differences in cell activation using sera from non-allergic patients (data not shown).

CMP-sensitized Balb/c mice were used to investigate the clinical relevance of the immunochemical co-recognition (Fig. 3A). Intragastically sensitized mice developed hypersensitivity reactions (Fig. 3B) that were scored immediately after oral challenge with CMP or SP (Table 1). CMP-sensitized animals showed high scores upon challenge with SP, while no symptoms were observed in sham mice. To correlate these clinical findings with a systemic parameter, histamine was measured (Fig. 3C). We observed increased plasma levels in sensitized mice, demonstrating that the administration of SP or CMP to CMP-sensitized animals developed adverse reactions. Next, we examined the functionality of IgE antibodies. Subcutaneous injection of CMP or SP resulted in a local and distant inflammation within minutes (Fig. 3D). Injection of saline induced negative results in sham and sensitized animals.

We next explored whether soy proteins were recognized by the secreted T-helper 2 (Th2)-associated antibodies. We found that SP and P34 were recognized by mouse CMP-specific IgE and IgG1 antibodies. No IgG2a specific antibodies were detected in sensitized or sham animals (Fig. 4A). To correlate the humoral immune response with the cellular immunity, the antigen-specific T-cell response and cytokine production were examined. A significant increase in the secretion of Th2 cytokines (IL-5 and IL-13) was detected in sensitized mice challenged with CMP or SP, while secretion of IFN-γ remained unchanged (Fig. 4B). Upon stimulation of spleen cells with detoxified P34 protein, IL-5 and IL-13 were secreted; in contrast, MBP did not induce cell stimulation.

Consistent with the in vitro findings, we observed a positive cutaneous test with P34 (Fig. 4C), demonstrating that this molecule has at least two B epitopes capable of cross-linking membrane-bound IgE. Furthermore, the negative results obtained with all proteins in sham mice demonstrated that oral challenge with recombinant protein did not sensitize the mice.

Since sensitization does not always correlate with the clinical situation, a sublingual challenge was performed with P34 and control proteins in sham and sensitized animals. P34 and β-Lg strongly induced immediate symptoms in CMP-sensitized mice (Fig. 4D). In contrast, no clinical signs were induced by non-related OVA.