Identification of cross-reactive allergens in cashew- and pistachio-allergic children during oral immunotherapy.

To the editor,

It has been estimated that around 8% of the children in the U.S. suffer from food allergy and of those, 40% are allergic to multiple foods.¹ Among tree nuts, allergies to pistachios are common in those with cashew nut allergy and multiple homologous allergenic components are shared between the two nuts.² Three major allergens from cashew (Ana o 1 at 50 kDa³, Ana o 2 major band at 33 kDa and minor band at 53 kDa⁴, and Ana o 3 at 10 kDa⁵) and five major allergens from pistachio (Pis v 1 at 7 kDa⁶, Pis v 2 at 32 kDa, Pis v 3 at 50 kDa⁷, Pis v 4 at 23 kDa, and Pis v 5 at 36 kDa⁸) have been identified. Of those, Ana o 1 and Pis v 3, Ana o 2 and Pis v 2, Ana o 2 and Pis v 5, Ana o 3 and Pis v 1 have been recognized as homologues based on their sequence similarity and cross reactivity to IgE from the patients.^7,8

We previously reported, in two independent multi-food oral immunotherapy (OIT) trials, that per respective trial, 83% (20 of 24) and 100% (8 of 8) of participants with cashew and pistachio allergies achieved desensitization to both foods by cashew-OIT alone (i.e. cross-desensitization).^9,10 However, how OIT affects the allergenicity of specific components in both foods during successful cross-desensitization is largely unknown. We hypothesized that desensitization to pistachio occurs through desensitization to homologous cashew allergens during cashew-OIT. To test our hypothesis, we investigated the specific cashew and pistachio components recognized by plasma IgE and IgG₄ from cross-desensitized participants.

Plasma samples at baseline and post-OIT from 25 individuals who participated in either one of the two multi-food OIT clinical trials ^9,10 (Table 1) and received cashew but not pistachio as one of the allergens in the respective OIT mix were used in this study (Supplemental table 1). All participants gave written informed consent. The research protocol was approved by the relevant institutional review boards. All participants were allergic to both cashew and pistachio as determined by a cumulative tolerated dose of less than 500mg of each food protein at baseline in double-blind, placebo-control food challenges (DBPCFC). In the first cohort, all 8 participants received omalizumab from week 0 to week 16 and multi-food OIT from week 8 to week 30.⁹ In the second cohort, of 17 participants, 15 received omalizumab and 2 received placebo from week 0 to week 16 and multi-food OIT was administrated to all 17 from week 8 to week 36¹⁰. All participants achieved cross-desensitization as determined by tolerating at least 2g of both cashew and pistachio proteins, respectively, in DBPCFCs at the post-OIT timepoints. The plasma samples collected at post-OIT timepoints were at least 14 weeks after omalizumab treatment which is 3.8 half-lives of omalizumab.¹¹ The cashew and pistachio components that reacted to plasma IgE and IgG₄ from the participants at baseline and post cashew-OIT were determined by western blots (Supplemental methods). McNemar’s exact tests were applied for the association between OIT and the presence or absence of each IgE and IgG₄ reactive protein component in cashew and pistachio. Benjamini–Hochberg procedures were applied for multiple hypotheses tests correction within each immunoglobin and food, respectively.

Table 1.

Demographics, DBPCFC results, and food-specific immunoglobin levels of the participants in the study.

Sex Female (n, %)	11 (44%)
Age (median, range) in years	8 (4-15)
Cashew cumulative tolerated dose (CTD) (median, range) in DBPCFC at baseline in mg	0 (0-175)
Pistachio CTD (median, range) in DBPCFC at baseline in mg	0 (0-275)
Cashew CTD in DBPCFC post-OIT in mg†	≥2000
Pistachio CTD in DBPCFC post-OIT in mg†	≥2000
Cashew-specific IgE (median, range) at baseline in kU/L‡	3.47 (0.41-85.2)
Cashew-specific IgE (median, range) post-OIT in kU/L‡	5.54 (0.32-60.8)
Pistachio-specific IgE (median, range) at baseline in kU/L ‡	11.6 (0.4-165)
Pistachio-specific IgE (median, range) post-OIT in kU/L‡	10.8 (0.68- 67.8)
Cashew-specific IgG4 (median, range) at baseline in mg/L‡	0.31 (0.01- 10.1)
Cashew-specific IgG4 (median, range) post-OIT in mg/L‡	7.95 (1.39- 203)
Pistachio-specific IgG4 (median, range) at baseline in mg/L‡	0.30 (0.02-9.05)
Pistachio-specific IgG4 (median, range) post-OIT in mg/L‡	4.3 (0.78-69.3)

^†2g was the endpoint of the trials, and challenges were often discontinued after a CTD of 2g was reached but before a reaction occurred.

^‡In 17 out of 25 participants food-specific immunoglobin levels were available.

We identified protein components from cashew and pistachio at molecular weights (MWs) corresponding to known allergens that reacted to the plasma IgE and IgG₄ from the participants (Figure 1). The numbers of participants whose IgG₄ reacted to protein components at MWs corresponding to the known allergens in cashew (i.e. Ana o 1, Ana o 2, and Ana o 3) were all significantly positively associated with OIT at post-OIT compared to baseline. (adjusted p<0.05, Figure 2). In pistachio, the numbers of participants whose IgG₄ reacted to Pis v 1 was found significantly positively associated with OIT (adjusted p<0.001). No significant association was found for the other known allergens in pistachio. In contrast, the number of participants whose IgE reacted to cashew and pistachio allergens remained similar between baseline and post-OIT.

Figure 1.

Distribution of cashew and pistachio protein components that recognize plasma IgE and IgG₄ at baseline and post-OIT.

Protein components from cashew (A) and pistachio (B) recognized by IgE at baseline (pink) and post-OIT (red), as well as IgG₄ at baseline (light blue) and post-OIT (dark blue) were identified by western blot. The molecular weights of the western blot bands were estimated by comparing with the location of markers and the bands at MWs corresponding to known allergens were labeled. Number of participants (out of 25) whose IgE and IgG₄ recognized each protein band are shown.

Figure 2.

Individual IgE and IgG₄ reactive protein components from cashew and pistachio for all 25 participants.

The heatmap shows protein components from cashew (A) and pistachio (B) that reacted to IgE at baseline (pink) and post-OIT (red), as well as IgG₄ at baseline (light blue) and post-OIT (dark blue) or those that not reacted (grey) for all 25 participants. Participants were ordered by subject ID in supplemental table 1. Known allergens as well as protein components to which IgE or IgG₄ from at least 40% of participants at baseline or post-OIT reacted are shown. Each row indicated one participant. Adjusted p values show the McNemar’s exact test results for the association between OIT with the presence or absence of each IgE and IgG₄ reactive protein component in cashew and pistachio from baseline to post-OIT and adjusted for multiple hypotheses testing. The adjusted p<0.05 were considered significant and highlighted in red.

In addition to the known allergens in cashew and pistachio, we identified IgG₄ reactive protein components in cashew and pistachio to which an increasing number of participants reacted post-OIT at MWs previously not recognized as major allergens. IgG₄ reacting to cashew components at approximately 37 kDa (adjusted p=0.012) and 60 kDa (adjusted p=0.046) and to a pistachio component at approximately 15 kDa (adjusted p<0.001) were significantly positively associated with OIT (Figure 2). Additional unknown protein components for which IgE or IgG4 from more than 40% of the participants at baseline or post-OIT reacted to, but which did not show a significant association with OIT, were identified in cashew at approximately 18 kDa and in pistachio at 42.5 kDa and 53 kDa. (Figure 2) The unknown protein components identified in our study could be similar to the proteins previously reported in cashew allergic patients at 15 and 60 kDa, and in pistachio allergic patients at 15, 41, 51 kDa.⁶

We previously reported that plasma cashew- or pistachio-specific IgE levels did not change significantly post-OIT compared to baseline; however, the IgG₄ levels and the IgG₄/IgE ratios significantly increased.¹⁰ In line with this observation, we noted a decreasing trend, but not a significant difference, in the numbers of IgE reactive protein components in cashew or pistachio between baseline and post-OIT, however the numbers of IgG₄ reactive protein components per participant increased significantly in both cashew (p<0.001) and pistachio (p=0.02) (Supplemental Figure 1).

Overall, we identified new cashew and pistachio components, as well as confirming components consistent with previous reports ^3–7, that reacted to IgE and IgG₄ in cashew and pistachio allergic participants. The cashew-OIT, which also cross-desensitized these participants to pistachio, was significantly positively associated with the numbers of participants whose IgG₄ reacted to the known cashew allergens (i.e. Ana o 1, Ana o 2, and Ana o 3). Interestingly, we found the number of participants whose IgG₄ reacted to Pis v 1 also significantly increased post-OIT. Ana o 3 and Pis v 1 are homologs in cashew and pistachio with the highest reported reactivity prevalence rate in cashew and pistachio allergic patients, resepectively.^5,6 Thus our results suggest that the cross-desensitization to pistachio was potentially facilitated through desensitization to Pis v 1 in these participants. Furthermore, we identified protein components at previously unidentified MWs in both cashew and pistachio following a similar pattern of changes post-OIT as described for the known allergens in cashew and pistachio, suggesting potentially previously unrecognized homologous allergens in cashew and pistachio. One limitation of this study is that the identities of the protein components were determined by estimating their MWs from western blots. Additional analysis of the protein sequences will be of great interest to further characterize the identities of these protein components, as well as for comparing the sequences of the components in cashew and pistachio. Furthermore, it is possible that other tree nuts (including hazelnut and walnut) in the multi-food OIT regimen received by some of participants may have had a limited effect on pistachio cross-desensitization (supplemental table 1). Future studies will be needed to elucidate the effects of multi-food OIT on the participant’s IgE and IgG4 repertoires.

In summary, the increasing number of IgG₄ reactive protein components in pistachio following cashew-OIT globally per participant (Supplement Fig. 1) as well as in individual components (Figure 1–2) suggest that cross-desensitization is achieved through desensitization of cashew-homologous components in pistachio. Our results provide important information for future component-specific diagnosis of patients allergic to both cashew and pistachio, as well as potential prognostic biomarkers for cross-desensitization of cashew and pistachio allergy during treatment.