Patients referred to the Allergy Unit of Malaga Hospital (south of Spain area) with a history of apple allergy that could be confirmed by positive prick-prick test and/or double-blind-placebo-controlled food challenge (DBPCFC), were included. Patients were classified according to symptoms: OAS when they were restricted to the skin or mucosal sites of direct contact with the allergen (itching of the oral mucosa and lips with or without angioedema immediately after eating apple), and GS when reactions involved organs far from the site of initial contact with the food (urticaria, with or without angioedema, and anaphylaxis) accompanied or not to OAS. A control group, from the same geographical area, comprised 25 non-allergic subjects with tolerance to apple. All participants completed a written informed consent and the ethical committee of our institution approved the study (CEI Provincial of Malaga).

The allergological evaluation included an examination of the patient's clinical history, a detailed questionnaire, skin test, specific IgE determination and DBPCFC.

The skin test was performed according to European guidelines [16] using Golden Delicious apples for the prick-prick technique with peel and pulp tested separately, and by SPT using commercialized extract from Phleum, Olea, Betula, Platanus, Cupressus, Parietaria and Artemisia pollen and apple, peach, cherry and hazelnut from ALK-Abelló (Madrid, Spain). The SPT response was considered positive if the diameter of the wheal area was 3 mm greater than that induced by the negative control.

Specific IgE antibodies to apple were measured by ImmunoCAP following the manufactureŕs recommendations (Phadia, Uppsala, Sweden). A positive result was defined as a value >0.35 kUA/l.

DBPCFC was performed as described [17], except in cases with a recent history (<1 year) of anaphylaxis after apple ingestion and SPT positive to apple. Briefly, meals containing 5, 40, and 120 g of fresh ground apple with a mixture of yogurt, orange juice, coffee dried, and oatmeal flakes were freshly prepared 5 minutes before administration to prevent the loss of allergenicity. Placebo meals consisted of the same ingredients, without fresh apple. If cutaneous and/or respiratory symptoms or alterations in vital signs appeared, the procedure was stopped and the symptoms were evaluated and treated.

The LTPs was purified following the method previous published by Diaz-Perales [18]. Briefly, LTP was purified from defatted peel peach or apple fruit by RP-HPLC on a Vydac-C4 column (22×250 mm; particle size 10 mm; The Separations Group, Hesperia, CA, USA), followed by RP-HPLC Nucleosil 300-C4 column (8×250 mm; particle size 5 mm; Sugelabor, Madrid, Spain). Finally, proteins were analyzed by mass spectrometry, SDS-PAGE and immunoblotting.

Mal d 2 was purified from apple fruit by cation-exchange chromatography on a Bio-ScaleTM Mini Macro-Preps High S column (BioRad, Hercules,CA, USA). Following this, its purity and reactivity was analyzed by mass spectrometry, SDS-PAGE and immunoblotting.

rMal d 1 and Pho d 2 was obtained from Biomay (Vienna, Austria) and ALK-Abello (Madrid, Spain), respectively.

rMal d 4 was prepared using the protocol described by Scheurer S [19]. Briefly, the protein was cloned by polymerase chain reaction and produced in Escherichia coli BL21. The profilin was purified as non-fusion proteins by affinity chromatography on poly-(L-proline)-Sepharose and its purity and IgE reactivity was analyzed by SDS-PAGE and immunoblotting.

rBet v 1 was produced in Escherichia coli BL21 (DE3). This protein was purified under denaturing conditions from inclusion bodies and subsequently subjected to Immobilized Metal Affinity Chromatography (IMAC) followed by ion-exchange chromatography (IEC). Finally, His-Tagged rBet v 1 was analyzed by SDS PAGE and Immunoblotting using an anti-His-Tag antibody, anti-Bet v 1 monoclonal antibody and sera from patients with IgE-reactivity to Bet v 1.

ELISA assays were performed as described [18]. Costar plates (Corning, NY, USA) were coated with 5 µg/mL of rMal d 1 (Biomay, Vienna, Austria), Mal d 2, rMal d 3 and rPru p 3 (Dra Díaz-Perales, Polytechnical University of Madrid), rMal d 4 and rBet v 1 (Dr Vieths, Paul-Ehrlich Institut) and Pho d 2 (ALK-Abelló). After, the plates were blocked with blocking solution (Sigma, St. Louis,USA) for 1 h. Individual serum from patients and controls at 1∶5 dilution was added to each well. We used two negative controls, a pool of sera from non-allergic subjects and blocking buffer, and as positive controls we used patients sera with known specific IgE to proteins. The assay was completed with rabbit anti-human IgE antibody (1∶3000, DAKO, Denmark) conjugated to horseradish peroxidase (HRP). IgE binding was detected using o-phenylenediamine (OPD, DAKO). Results are expressed as absorbance units, measured at 490 nm. They were considered positive when >0.18 (mean OD + 3xSD to blocking).

For ELISA inhibition, sera were pre-incubated with different inhibitors (rMal d 3, rMal d 4, rPru p 3, Pho d 2) at final concentrations of 20, 2, and 0.2 µg/mL or with phosphate buffer for the non-inhibited serum for 3 h at room temperature. Subsequently, the inhibitor mixtures were added to plates coated with rMal d 3, rMal d 4, rPru p 3 or Pho d 2, followed by the ELISA protocol described above.

To analyse the sensitization of patients positive to rMal d 1 and rBet v 1, sera from specific IgE positive patients were pre-incubated with rMal d 1, Mal d 2, rMal d 3, rMal d 4 and rBet v 1 and added in solid phases with rMal d 1 and rBet v 1.

Percentage inhibition of IgE binding was calculated as follows: % Inhibition  =  (serum not inhibited-serum inhibited/serum not inhibited) x100.

Quantitative variables are shown as medians and interquartile ranges (IR), while qualitative variables are shown as frequencies. Medians between groups were compared using Mann-Whitney or Kruskal-Wallis tests, while χ² was used to compare proportions. Differences with a p<0.05 were considered significant.

The study included 81 patients with a diagnosis of hypersensitivity to apple. Their median age was 31 (IR:25–38) years and 62 (76.44%) were female. Thirty-five patients (43.20%) had OAS and 46 (56.79%) had GS (22 anaphylaxis (47.82%) and 24 (52.12%) urticaria). Sixty-eight cases (83.95%) showed positive prick by prick skin tests with apple peel and pulp and 55 (67.90%) positive skin test with the commercial extract for 77.14% of cases tested using prick by prick and for 57.14% with skin test. In GS patients we found that 76.08% showed skin test positivity with commercial extract and 89.13% with prick by prick. Moreover, we did not find any patient who was positive with commercial extracts and negative with prick by prick. In addition, 65 (80.24%) had serum specific IgE to apple and 62 a positive DBPCFC to apple. DBPCFC was not performed in 19 cases who had recent apple-related anaphylaxis (<1 year) and a positive skin test to apple.

Seventy-one (87.65%) were sensitized to pollen and from these 54 (76.05%) had clinical symptoms consisting of rhinitis and/or asthma; 77 (95.06%) presented symptoms with peach and 31 (38.27%) with hazelnut. The median age at symptom onset was 15 (IR:10–25) years to peach, 17 (IR:11–27) to apple, 17.5 (IR:7.75–25) to pollen and 18 (IR:13–25) to hazelnut.

Comparisons between patients with OAS and GS (Table 1) showed that although the percentage of cases with a positive SPT and specific IgE to apple was higher in those with GS, there were no significant differences between the groups in any of the variables evaluated.

Comparisons of clinical symptoms between patients with (N = 54, 66.66%) and without pollen allergy (N = 27, 33.33%) showed that in those with pollen allergy 24 (44.44%) developed GS and 30 (55.55%) OAS, whereas in those without pollen allergy, 18 (66.66%) developed GS and 9 (33.33%) OAS (Data not shown).

Analysis of the SPT showed that 49 (60.49%) cases were positive to Phleum, 54 (66.67%) to Olea, 30 (65.22%) to Betula, 44 (54.32%) to Platanus, 41 (50.62%) to Cupressus, 38 (46.91%) to Parietaria, 35 (43.21%) to Artemisa, 46 (56.79%) to peach, 34 (41.97%) to cherry and 32 (39.51%) to hazelnut. Comparison between patients with OAS and those with GS showed a significantly higher percentage of positive results only with peach, cherry and hazelnut in those with GS (Table 2).

Comparisons of specific IgE determinations (Figure 1A) in OAS, GS, and controls showed significant differences between the three groups for rMal d 3 (p<0.001), rMal d 4 (p<0.001), rPru p 3 (p = 0.015) and rBet v 1 (p = 0.02). In more detail, there was an increase in levels of IgE to four allergens in OAS when compared to controls (rMal d 3: p<0.001, rMal d 4: p<0.001, rPru p 3: p = 0.007, and rBet v 1: 0.010) and in GS compared to controls (rMal d 3: p<0.001, rMal d 4: p = 0.001, rPru p 3: p = 0.013, and rBet v 1: p = 0.041), with no differences between OAS and GS. Comparisons in terms of positive percentage (Figure 1B) showed significant differences between OAS and controls for rMal d 3 (p = 0.036), rMal d 4 (p = 0.042), rPru p 3 (p = 0.009), rBet v 1 (p = 0.038) and Pho d 2 (p = 0.021) and between GS and controls for rMal d 3 (p = 0.005), rMal d 4 (p = 0.009), rPru p 3 (p = 0.007) and Pho d 2 (p = 0.045). Moreover, rBet v 1 was higher in OAS compared to GS (p = 0.011).

We detected that 27% of the patients recognized Mal d 1, 5% to Mal d 2, 37% to Mal d 3 and 30% to Mal d 4. When we analyzed the sensitization percentages taking into account symptoms we observed no significant differences between OAS and GS. In patients with OAS, 31% of them had specific IgE to Mal d 1, 2% to Mal d 2, 38% to Mal d 3 and 28% to Mal d 4. In patients with GS, 23% had specific IgE to Mal d 1, 9% to Mal d 2, 37.50% to Mal d 3 and 35% to Mal d 4.

Based on the IgE response to allergens from LTP and profillin families, the patients were classified in three groups (Table 3): Group 1 (LTP pattern), positive to rMal d 3 and/or rPru p 3; Group 2 (Profilin pattern), positive to rMal d 4 and/or Pho d 2; and Group 3 (LTP-Profilin pattern), positive to both. We found a significant increase in the percentage of patients positive to LTP-profilin in the group with GS. Moreover, if we consider all those with an LTP response, regardless of the profilin results, there were 11 (31.42%) positive cases in the OAS group and 27 (58.69%) in the GS (p = 0.024). On the other hand, considering all those who were positive to profilin, regardless of the LTP results, 9 (25.71%) were positive in the OAS group and 18 (39.13%) in the GS group (p = 0.240).

ELISA inhibition was performed for all cases showing a sIgE level greater than 0.5 of O.D. Figure 2 shows the ELISA inhibition pattern for representative cases from each group of patients. Results from the Group 1, patients with GS, showed that using either rMal d 3 or rPru p 3 in the solid phase the strongest inhibitor was rPru p 3, followed by rMal d 3, with no inhibition by the other allergens (Figure 2A).

In Group 2, patients with either OAS or GS, using rMal d 4 in the solid phase, the highest inhibition was found with itself followed by Pho d 2; similar results were seen using Pho d 2 in the solid phase (Figure 2B and C).

Figures 2D and 2E show sera from Group 3 for patients with OAS or GS, respectively. In both cases, when using rMal d 3 in the solid phase the highest inhibition was found with rMal d 3 followed by rPru p 3. Using Pho d 2 in the solid phase we detected the highest inhibition with itself and rMal d 4.

In order to further analyse sensitization to Bet v 1 observed in the patients allergic to apple, ELISA inhibition studies were performed by analyzing the IgE recognition to the PR-10 proteins, rMal d 1 and rBet v 1 (Figure 3). Data showed that rMal d 1 was the most potent inhibitor regardless of the solid phase used (rMal d 1 or rBet v 1). However, rBet v 1 only showed high inhibition when the same allergen was used in the solid phase. Moreover, no inhibition was detected with Mal d 2, rMal d 3 or rMal d 4.