Depletion of eosinophils during sensitization but not challenge phase in mice blocks the development of food allergy early in life

Abstract

Food allergy can be life threatening and often develops early in life, especially in infants and children with atopic dermatitis. Food allergy is induced in neonatal mice with skin barrier mutations (Flaky Tail, FT^+/− mice with filaggrin and mattrin gene mutations) by epicutaneous sensitization with co-exposures to the food allergen peanut extract (PNE), the environmental allergen Alternaria alternata (Alt), and detergent (4% SDS); oral PNE-challenge induces anaphylaxis. Sensitization in these neonates also induces eosinophil infiltration into the skin and elevates skin expression of eotaxins (CCL11 and CCL24). However, roles for eosinophils in food allergy are not known. In this study, the iPhil^+/− FT^+/− pups, which have an inducible eosinophil-deficiency upon injection of diphtheria toxin (DTX), were sensitized and then received PNE by gavage to assess anaphylaxis. DTX depletion of eosinophils, during sensitization and oral PNE-challenge, blocked the recruitment and activation of mast cells, blocked the Alt+PNE-induced increase in plasma IL-33 and OSM, attenuated serum PNE-specific IgE/IgG1/IgG2b, and blocked oral-PNE-induced anaphylaxis. Anti-IL-5 depletion of eosinophils during sensitization/challenge also blocked anaphylaxis. When eosinophils were depleted during allergen-skin-sensitization and restored before oral PNE-challenge, anaphylaxis was blocked. In contrast, when eosinophils were present during allergen-skin-sensitization but then depleted during oral PNE-challenge, anaphylaxis was not blocked. Together, these data indicate that although eosinophils are not necessary during oral food allergen-induced anaphylaxis, eosinophils have a critical role during the development of food allergy early in life by regulating the sensitization-induced increase in mast cell numbers and food allergen-specific IgE.

Introduction

Food allergy is increasing rapidly at the global level.^1–13 As a public health concern, food allergy currently affects 11% of adults and 8% of children in the United States.¹⁴ Symptoms of food allergy range from mildly itching skin and abdominal discomfort to severe anaphylaxis, which can be deadly.^13,15 Food allergy has caused burdens not only to the allergic individuals but also to their health care providers.^16,17 Despite of its increasing prevalence and recent development of potential therapeutics, there is no cure for food allergy.

Food allergy develops early in life.^18–22 It was shown in previous studies that up to 35% of children with atopic dermatitis develop food allergy.^23–26 The development of atopic dermatitis involves skin barrier dysfunction, particularly loss-of-function mutations in skin barrier genes, including SPINK5, FLG (filaggrin), and Tmem79 (mattrin).^27,28 These loss-of-function mutations have been reported in children with peanut allergy.^29,30 Moreover, the ubiquitous environmental allergens, house dust mite (HDM) and Alternaria Alternata (Alt), present in nearly all homes across the United States^31–35 also play important roles in the development of food allergy and atopic dermatitis.^36–40 A neonatal mouse model exists with skin barrier dysfunction that mimics infants and young children with eczema who are susceptible to develop food allergy early in life.^28,41–45 In this food allergy mouse model, pups have heterozygous skin barrier mutations (Flakey Tail, FT^+/−).⁴⁵ We have demonstrated that FT^+/− neonatal mice develop food allergy by co-exposure to food allergen (ovalbumin or peanut), environmental allergen (HDM or Alt), and detergent.⁴⁵ These co-exposures are relevant as they that can occur in homes of children with skin barrier mutations. In contrast, wild type C57BL/6 mice do not generate food allergy with these co-exposures.⁴⁵ Although our studies focus on early life development of food allergy because food allergy in humans most often develops in children, we have reported that in Alt+PNE skin sensitized 6 week old adult FT^+/−mice, oral PNE induces anaphylaxis.⁴⁶ Our qPCR and bulk RNA sequencing analysis of the skin from Alt+PNE sensitized FT^+/− pups revealed increased expression of CCL11 (eotaxin-1) and CCL24 (eotaxin-2) ,⁴⁷ which is consistent with the increase in numbers of their skin eosinophils.

Eosinophils are granulocytes that are important in many cellular processes in maintaining homeostasis of the body’s immune responses.^48,49 Studies of allergic respiratory inflammation, gastrointestinal responses to chicken egg ovalbumin, and atopic diseases revealed that eosinophils are not simply downstream mediators of other inflammatory cells but rather contributing to the adverse effect upon allergen challenge.^50–55 However, it is unknown whether eosinophils are also associated with the development of peanut food allergy in FT^+/− mice.

In this study, we investigated the importance of eosinophils in the sensitization as well as the food allergen-challenge phase of food allergy by depleting eosinophils with anti-IL-5 or by depleting eosinophils with injection of diphtheria toxin (DTX) in iPhil^+/− FT^+/− pups. We demonstrate that food allergy is blocked when eosinophils are depleted during allergen-sensitization and allergy oral-challenge, or when eosinophils are depleted during sensitization and restored before the oral allergen challenge. However, when eosinophils are depleted after skin sensitization but before oral allergen-challenge, anaphylaxis was not blocked. Together, our data indicate that eosinophil depletion affects the sensitization but not the food allergen-challenge phase of food allergy.

Methods

Mice

All mice were on a C57BL/6J background. Flaky Tail (FT^−/−) mice with homozygous mutations in Flg^ft/ft/TMEM79^ma/ma) were from Jackson Laboratories, Bar Harbor, ME. Mice with inducible eosinophil-deficient mice (iPhil^+/+) were from Mayo Clinic Arizona.^56,57 These mice were bred to generate iPhil^+/+ FT^−/−− male mice, which were then bred with wild-type C57BL/6J females to generate iPhil^+/− FT^+/− pups. The heterozygous FT^+/− genotype is used because these mice develop food allergy and most humans are not homozygous for these mutations.^28,41–44 The iPhil heterozygous genotype is used because homozygous iPhil mice loose the expression of eosinophil peroxidase (EPX) and therefore EPX-driven expression of the receptor for DTX All mice were genotyped. All animal protocols were performed in accordance with Indiana University School of Medicine Institutional Review Committee for animals.

Allergens

Alternaria alternata (Alt) extract (catalog#XPM1D3A2.5) was from Stallergenes Greer Labs (Lenoir, North Carolina, USA), because this extract has been used in patient immunotherapy. To generate peanut extract (PNE), peanuts (Trader Joe’s roasted and unsalted peanuts) were ground, and then 25 g were homogenized in 250 ml of 20 mM Tris buffered (pH 7.2) saline (TBS).⁵⁸ This was stirred for 2 h at room temperature and centrifuged at 3,000× g for 30 min. The aqueous middle layer was collected and centrifuged at 1,600× g for 45 min to remove residual particles and fat. The aqueous layer was collected, and protein concentrations were determined by Pierce BCA Protein Assay Kit (Thermo Fisher Sci). Aliquots were stored at −20°C. We have previously reported that the Alt and PNE extracts did not exert overt protease activity as co-incubation of Alt and PNE did not alter the size of the major protein bands on SDS-PAGE chromatography.⁴⁵

Allergen sensitization and challenge protocol

iPhil^+/− FT^+/− or FT^+/− pups were skin-sensitized starting on postnatal days (PND) 6 followed by 4 more skin sensitizations, for a total of 5 skin sensitizations within 2 weeks, as indicated in the figure timelines, using our previously published sensitization protocol.⁴⁵ Briefly, pups were shaved on their back and taped using 3M Micropore surgical cloth tape (3M, Saint Paul, Minnesota, USA) to remove the residual fur without overt damage to their skin.⁴⁵ Then, 4% SDS (one wipe of skin with sterile gauze containing 4% SDS) (Sigma Aldrich, Saint Louis, Missouri, USA), Alternaria Alternata extract (10 µg protein in 5 µl sterile saline) and PNE (100 µg protein in PNE in 10 µl sterile saline) were applied on the shaved area. Treated pups were placed in a cage without their mothers for 40 min to avoid passing allergens through grooming. Pups were then washed gently with sterile water and wiped dry to remove allergens before placing them back with their mothers. On PND19, sensitized pups received oral gavage with PNE (100 µl of 10 mg/ml PNE in sterile water) using a 24-gauge gavage needle (Pet Surgical, Agoura Hills, California, USA). Their body temperatures were monitored before and every 15 min up to 60 min after the gavage using a BAT-12 Microprobe Thermometer with a RET 4 thermocouple sensor type T rectal probe for neonatal mice (Physitemp Instruments Inc, Clifton, New Jersey, USA). After gavage, pups were placed back with their mothers for 4 h, and then pups were euthanized for collection of tissues. In some experiments, sensitization and challenge timelines are as indicated in the figures.

Administration of anti-IL-5 antibodies or diphtheria toxin (DTX)

In the anti-IL-5 treatment protocol indicated in the figure timeline, anti-IL-5 or isotype control antibodies (catalog no. BE0198 and BE0088, respectively; BioXCell, Lebanon, New Hampshire, USA) were intraperitoneally injected (4 µg/g of body weight) ⁵⁹ on postnatal days (PND) 5, 8, 12, 15, and 19. In other experiments, pups were treated with or without diphtheria toxin (DTX; Item No. 19657, Cayman Chemical) at a dose of 15 ng/g of body weight on PND 3, 4, 7, 10, 12, 14, and 17, to deplete eosinophils and maintained throughout the sensitization phase as indicated in the figure timelines. In the delayed eosinophil depletion model, pups were sensitized followed by DTX administration on PND 21, 22, and 25 as indicated in the figure timelines.

PNE-specific antibodies, MCPT-1, histamine, OSM, IL-33, and Areg ELISAs

Pup sera and plasma were collected at 4 h after oral gavage with PNE. Serum PNE-specific IgE, IgG1, and IgG2b levels were determined by ELISAs with modifications from our previous study.⁴⁷ Briefly, high-binding 96-wells plates (Fisher Scientific, catalog no, 50-823-480) were coated with 60 µl of 10 µg PNE in carbonate buffer overnight at 4°C. On the next day, the plates were washed three times with PBS-0.05% Tween and then blocked with 150 µl of 3% BSA-PBS for 2 h at room temperature.

For the PNE-specific IgE ELISAs, PNE-specific IgGs were removed from samples by plating 25 µl of sera and 25 µl 1% BSA-PBS per well of an anti-mouse IgG-coated plates (Thermo Scientific, catalog no.15134) for 1 h, and then the samples were transferred from the anti-IgG coated plates into PNE-coated/3% BSA-blocked plates for an incubation overnight at 4°C. On the following day, the plates were washed, and then 100 µl goat anti-mouse IgE-HRP (Novus Biologicals, catalog no. NB7533) at a 1:500 dilution in 1% BSA-PBS was added to plates and then incubated for 1 h at room temperature and then washed. To amplify the signal for HRP for detection of neonate PNE-specific IgE, the plates were washed followed by addition of 100 µl donkey anti-goat IgG-HRP (Invitrogen, catalog no. A15999) at a 1:500 dilution with 1% BSA-PBS for 1 h at room temperature. Then the plates were washed and incubated with 100 µl TMB substrate (eBioscience, catalog no. 004201-56). The plates were read at 650 nm with a microplate reader (BioTek, Winooski, Vermont, USA).

For PNE-specific IgG1 and IgG2b, sera were plated at 1/20,000 and 1/1,000 in 1% BSA-PBS, respectively, into the PNE-coated/3% BSA-blocked plates and then incubated overnight at 4°C. On the following day, plates were washed and then 100 µl biotinylated anti-mouse IgG1 (BD Pharmingen, catalog no.553441) or anti-mouse IgG2b (BD Pharmingen, catalog no. 553393) at a 1:250 dilution in 1% BSA-PBS was added for 1 h at room temperature. Then, the plates were washed and 100 µl/well streptavidin-HRP (GE Healthcare, catalog no. RPN1231) at 1:1000 dilution in 1% PBS was added for 1 h at room temperature. Then all plates were washed and incubated with 100 µl TMB substrate (eBioscience, catalog no. 004201-56). The plates were read at 650 nm with a microplate reader (BioTek, Winooski, Vermont, USA). Plasma histamine (LDN, catalog# BA E-5800R) and MCPT-1 (Invitrogen, catalog no. 88-7503-22) was measured according to the manufacturer’s instructions.

Skin IL-33, OSM and areg RNA expression

Tissues were placed in 200 µl TriZOL, 5 µl of RNAsin Ribonuclease Inhibitor (catalog no. PRN2115; Promega), and 2 µl of SUPERase-In RNase Inhibitor (catalog no. AM2696; Invitrogen by Thermo Fisher Scientific) and homogenized using a Next Advance Bullet Blender Storm 24 Place Bead Homogenizer. Then, another 400 µl of TriZOL was added and the tubes were centrifuged at 4°C for 8 minutes at 1200 rpm. Then 800 µl of the supernatant was placed into nuclease-free tubes on ice. An equal volume of 100% ethanol was added. The rest of the RNA extraction procedure followed the protocol for the RNeasy Mini Kit for skin tissue (Qiagen, Hilden, Germany) on ice. DNA digest was done by DNase 1 (5 µl) and DNA Digestion buffer (10 µl) from the kit. RNA concentration, A260/280 ratio, and A260/230 ratio of each RNA sample was measured using a Take3 plate and Synergy H1 Microplate Reader. The cDNA was generated using the qScript cDNA Synthesis Kit (catalog no. 95047; Quanta Bio). The qRT-PCR analysis was performed with a 7500 Fast Real Time PCR System (Applied Biosystems). Reactions were set up with the TaqMan Multiplex Master Mix (catalog no. 44445544). RT-qPCR reactions were set up with TaqMan PCR probes (catalog #4331182; ThermoFisher); GAPDH (assay ID: Mm99999915_g1), IL-33 (assay ID: Mm00505403_m1), Areg (assay ID: Mm01354339_m1), and OSM (assay ID: Mm01193966_m1). Data were normalized to GAPDH, and mRNA expression fold-change relative to GAPDH control was calculated using the 2−ΔΔCt method.

Histopathology

Blood collected at 4 h after gavage or 4 h after the last sensitization were used to determine the number of blood eosinophils. In brief, blood and heparin were mixed at a 1:1 ratio. Then, the samples was then diluted 1:5 with discomb stain (1 part acetone, 1 part 2% aqueous eosin, and 8 parts distilled water) ^60,61 and counted with a hemocytometer. For assessment of skin eosinophil and mast cells numbers, 8 mm skin biopsy punches were fixed in 10% formalin followed by 70% ethanol, then embedded in paraffin and sections at center of punch biopsies were prepared on slides. Skin tissue sections were stained with 1% toluidine blue for mast cells or with 1% eosin and a 0.05% methyl green counterstain for eosinophils. Stained slides were examined for the number of mast cells and eosinophils using a microscope with 40× objective.

Statistical analyses

Data were analyzed by a one-way ANOVA followed by Wilcoxon multiple comparisons test (SigmaStat, Jandel Scientific, San Ramon, California, USA). Presented are the means ± the standard errors. The statistical analysis for the change in temperature was done using the area under the curve for the change in temperature for each pup. The data in the figures include both genders for the offspring because there were no differences in outcomes by gender as previously described.⁴⁵ Shown are representative experiments of at least 2 experiments, and all groups had 6 to 8 pups/group.

Results

Treatment with anti-IL-5 antibodies reduces eosinophils and blocks anaphylaxis in FT^+/− mice

We have shown that FT^+/− neonatal mice develop food allergy and have increased skin eosinophils when co-exposed to detergent, environmental allergens, and food allergens.⁴⁵ IL-5 mediates maturation and survival of eosinophils,⁶² and anti-IL-5 antibodies partially block eosinophil recruitment to the intestine in a model with intraperitoneal OVA/alum sensitization and intragastric OVA challenge.^53–55 Also, anti-IL-5 antibodies such as mepolizumab have been widely used for the treatment of eosinophilic asthma in humans.⁶³ To examine whether anti-IL-5 treatment can reduce the number of eosinophils and block PNE-induced anaphylaxis in FT^+/− pups, FT^+/− pups were intraperitoneally injected with anti-IL-5 or isotype control antibodies one day before each skin-sensitizations as indicated in the timeline (Fig. 1A). Upon examination of eosinophil levels, anti-IL-5, but not the isotype control, significantly reduced eosinophil recruitment in the pups’ blood (Fig. 1B) and skin (Fig. 1C) in Alt+PNE sensitized pups. Anti-IL-5 did not reduce serum PNE-specific IgE levels in pups treated with Alt+PNE and anti-IL-5 as compared to pups treated with Alt+PNE and isotype antibody (Fig. 1D). Nevertheless, oral PNE-induced anaphylaxis was blocked in Alt+PNE sensitized pups treated with anti-IL-5 but not IgG1 isotype control (Fig. 1E). The controls, saline, Alt alone and PNE alone, did not induce food allergy in these pups (Fig. 1), as we have previously published.⁴⁵ The change in temperature in the Alt&PNE group is consistent with the level of temperature decrease ^45,47 as pups have a much lower basal temperature than adult mice.

Treatment with diphtheria toxin depletes eosinophils in iPhil^+/− FT^+/− mice

Although anti-IL-5 treatment was sufficient to block anaphylaxis, receptors for IL-5 are expressed by several immune cells besides eosinophils and thus is not specific for assessment of eosinophils role in initiation of food allergy. Therefore, we used the iPhil^+/− mice with knock-in mutation in the eosinophil peroxidase (EPX) gene which expresses human diphtheria toxin (DTX) receptor.⁵⁶ The iPhil^+/− mutation enables eosinophil-specific depletion upon injection of DTX. Neonatal iPhil^+/− FT^+/− pups and FT^+/− control pups were treated as indicated in the timeline (Fig. 2A). DTX treatment ablates eosinophils during maturation.⁵⁶ This timeline of DTX treatment ablated eosinophils in the blood (Fig. 2B), skin of Alt+PNE exposed iPhil^+/− FT^+/− pups (Fig. 2C and 2D), and intestine ileum of Alt+PNE exposed iPhil^+/− FT^+/− pups (Fig. 2E).

Eosinophil depletion, during food allergen sensitization and challenge, reduces PNE-Specific antibodies and blocks anaphylaxis in iPhil^+/− FT^+/− neonatal mice

To understand the impact of eosinophil depletion on development of food allergy, we examined anaphylaxis and serum PNE-specific IgE, IgG1, and IgG2b levels in DTX-treated iPhil^+/− FT^+/− pups. Skin sensitizations with Alt+PNE induced an increase in PNE-specific IgE (Fig. 3A), IgG1 (Fig. 3B), and IgG2b (Fig. 3C) whereas this was blocked when eosinophils were depleted by DTX treatment. Furthermore, oral-PNE induced anaphylaxis was blocked in DTX-treated Alt+PNE sensitized iPhil^+/− FT^+/− pups (Fig. 3D). In contrast, anaphylaxis was not blocked in DTX-treated Alt+PNE sensitized FT^+/− pups, since the iPhil^+/− mutation is required for the eosinophil depletion upon DTX injections (Fig. 3E). This supports the impact of specific depletion of eosinophils on the development of food allergy by allergen exposures of skin with skin barrier mutations.

Eosinophil depletion reduces allergen-induced plasma IL-33 and OSM but not Areg in iPhil^+/− FT^+/− neonatal mice

Because Alt and/or PNE induced Areg, OSM and IL-33 in FT^+/− pups and these cytokines regulate development of food allergy,⁴⁷ we determined whether eosinophil depletion, during skin sensitization and challenge in FT^+/− pups (Fig. 2A), affects the levels of these cytokines by analysis of IL-33, Areg, and OSM in the plasma of pups at 4 h after the last sensitization on PND19. Alt+PNE-treated or Alt-treated iPhil^+/− FT^+/− pups without DTX have increased IL-33 levels as compared to saline treated pups without DTX, and this was blocked by depletion of eosinophils with DTX (Fig. 4A). Alt+PNE treated iPhil^+/− FT^+/− pups without DTX have increased OSM levels relative to the control groups, which was blocked by depletion of eosinophils with DTX (Fig. 4B). The Alt+PNE-induced increase in plasma Areg, however, was not affected when eosinophils were depleted by DTX treatment of the iPhil^+/− FT^+/− pups (Fig. 4C).

Depletion of eosinophils blocks mast cell recruitment and degranulation in the skin and ileum of iPhil^+/− FT^+/− neonatal mice

Tissue-resident mast cells are crucial players in eliciting food anaphylactic reactions.^64,65 These cells are primed during allergen sensitization and are activated upon allergen challenge to release mast cell mediators within minutes. To examine the effect of eosinophil ablation on skin and intestine ileum mast cells, we fixed and stained the skin and intestine ileum from allergen-sensitized and challenged iPhil^+/− FT^+/− pups with and without DTX-treatment. Our histological analysis (Fig. 5A) revealed that with DTX-treatment, there was a decrease in Alt+PNE-induced number of degranulated mast cells in the skin (Fig. 5B) and total number of skin mast cells (Fig. 5C). Because mast cell protease-1 (MCPT-1) is a measure of mast cell degranulation and food allergen-induced anaphylaxis is mediated by mast cell release of histamine,^64,65 we measured MCPT-1 and histamine levels in plasma from DTX-treated and nontreated allergen-sensitized iPhil^+/− FT^+/− pups (Fig. 5B, C). In the DTX-treated Alt+PNE sensitized iPhil^+/− FT^+/− pups at 4 h after oral PNE-challenge, there was a decrease in number of intestine ileum mast cells (Fig. 6A) and decreases in plasma histamine (Fig. 6B) and MCPT-1 (Fig. 6C), suggesting that eosinophils are important in signals upstream of mast cell recruitment and histamine release into the circulation.

Eosinophil depletion impacts food allergen sensitization but not the food allergen challenge phase in the iPhil^+/− FT^+/− neonatal mice

We demonstrated above that depletion of eosinophils during allergen sensitization and challenge phases blocked food-induced anaphylaxis. Given the advantage of the inducible features of eosinophil deficiency in iPhil^+/− mice, we sought to understand whether eosinophil depletion impacts food allergen sensitization and/or allergen challenge phases. Also, previous studies using the iPhil^+/− mice have shown that eosinophil depletion can be maintained over 2 weeks until the humoral immune responses neutralize the effect of DTX-treatment and eosinophils are recovered.⁵⁶ Thus, given the advantage of this reversible feature of eosinophil deficiency in iPhil^+/− mice, we sought to assess whether depletion of eosinophils during allergen sensitization and then recovery of eosinophils before PNE-challenge could impact anaphylaxis.

As shown in Fig. 7, iPhil^+/− FT^+/− pups were first skin-sensitized (ie, while eosinophils are present) and then their eosinophils were depleted before oral PNE-gavage. In Fig. 7, pups were sensitized during DTX-depletion of eosinophils, and then there was a week for restoration of their eosinophils before oral PNE gavage. In these pups, the delayed DTX treatment after sensitization (Fig. 7A) depleted eosinophils in the blood, skin and intestine ileum at the time of PNE-challenge in the iPhil^+/− FT^+/− pups (Fig. 7C, D, E) but did not protect the Alt+PNE sensitized pups from oral PNE-induced anaphylaxis (Fig. 7F) and did not reduce serum PNE-specific IgE levels (Fig. 7G) as compared to pups without DTX, consistent with generation of allergen-specific IgE during sensitization. Then, in the model with depletion of eosinophils during sensitization and then eosinophil restoration in the blood before oral PNE-challenge of iPhil^+/− FT⁺⁻ pups (Fig. 7B, C), there was a trend for a decrease in serum PNE-specific IgE as compared to pups without DTX but this did not reach statistical significance (Fig. 7G). At 4 h after oral PNE challenge in the mice from timeline in Fig. 7B, there was also no restoration of skin eosinophils (Fig. 7D) and there was a partial restoration of intestinal ileum eosinophils (Fig. 7E), consistent with an intestinal but not skin allergen stimulation during the oral-allergen challenge phase. Nevertheless, in these pups (Fig. 7B), the depletion of eosinophils during sensitization and recovery of eosinophils for a week before PNE-challenge, blocked oral PNE-induced anaphylaxis (Fig. 7F), indicating a role for eosinophils during allergen sensitization. The depletion of eosinophils did not decrease the plasma cytokines IL-33, Areg, or OSM (Fig. 8), suggesting that the eosinophil function in regulating sensitization is subsequent to the generation of these cytokines which we previously showed regulates food allergy in FT^+/− pups.⁴⁷

To further characterize mechanisms of eosinophil regulation of food allergen anaphylaxis in the DTX-treated iPhil^+/− FT^+/− neonatal mice, we assessed skin mast cells, intestinal ileum mast cells and plasma histamine (Fig. 9). Depletion of eosinophils during the oral peanut challenge phase (Fig. 7A) did not alter the number of skin mast cells (Fig. 9A), skin mast cell degranulation (Fig. 9B), number of intestinal ileum mast cells (Fig. 9C), ileum mast cell degranulation (Fig. 9D), or plasma histamine (Fig. 9E). In contrast, depletion of eosinophils during the skin sensitization phase (Fig. 7B) reduced the number of skin mast cells (Fig. 9A), skin mast cell degranulation (Fig. 9B), number of intestinal ileum mast cells (Fig. 9C), ileum mast cell degranulation (Fig. 9D), and plasma histamine (Fig. 9E), consistent with the reduced anaphylaxis in this group (Fig. 7F).

Discussion

Although mast cells have been the focus in studies of the mechanisms for food allergy ^64,65 and we have reported that eosinophils are recruited during food allergen sensitization with PNE,⁴⁵ the importance of eosinophils in the development of food allergy were not known. Our results indicate that in a food allergy mouse model with skin barrier dysfunction and inducible eosinophil deficiency, depletion of eosinophils during allergen sensitization blocked early life development of food allergy, whereas depletion of eosinophils after allergen sensitization but before challenge did not block PNE-induced anaphylaxis or generation of PNE-specific IgE. In this study, we demonstrate that eosinophils are important for initiation of food allergy, particularly for allergen-induced increases in tissue mast cell numbers, for the generation of allergen-specific antibodies and for release of downstream mediators that lead to anaphylaxis.

We previously reported that there are increased numbers of skin and intestine eosinophils and elevated skin expression of eotaxins (CCL11 and CCL24) in Alt+PNE sensitized FT^+/− pups.⁴⁵ Our present study reveals that anti-IL-5 treatment reduced but did not ablate skin eosinophil numbers, and blocked oral PNE-induced anaphylaxis in FT^+/− pups. This reduction in eosinophils is consistent with results from allergic airway and gastrointestinal studies using the IL-5 and eotaxin1/2 knockout mice which demonstrated reduced but not complete ablation of eosinophil numbers and activities.^53–55,57 However, even though anti-IL-5 is sufficient to block anaphylaxis, anti-IL-5 treatment is not eosinophil-specific because B cells, basophils, neutrophils, fibroblasts, and epithelium also have IL-5 receptors ^67–71 and thus anti-IL-5 may have altered the functions of these other cells. The PHIL mice and dlbGATA-1 mice with congenital deficiency of eosinophils have been the hallmark models for studying eosinophil functions in allergic respiratory inflammation. Nevertheless, these congenitally eosinophil-deficient mouse strains have developmental defects (eg basophil survival) and thus impede investigation of the kinetic role of eosinophils in allergic inflammation.⁵⁷ In contrast, the iPhil^+/− mice with the inducible feature of eosinophil depletion enables one to study whether specific depletion of eosinophils participate in different stages of life development without potential effect on development prior to their depletion. Therefore, the studies with iPhil^+/− FT^+/− pups were imperative for our studies of eosinophil-dependent function in the development of food allergy.

Consistent with studies of allergic lung diseases using iPhil^+/− mice,^56,57 here our studies of food allergy demonstrate that injections of DTX before and during skin-sensitization were sufficient to deplete eosinophils in the skin and blood. However, unlike the studies of allergic respiratory diseases that suggest eosinophils are not required during sensitization but rather have a function during airway allergen challenge,⁵⁷ our studies suggest that eosinophils are necessary in the skin allergen-sensitization phase but not in the oral challenge phase of the development of food allergy. Indeed, eosinophils are known to participate in pathways for Th2 activation and Th2 cytokine-dependent class switching to IgE antibodies.^72–74 Eosinophils could also uniquely influence DC functions for antigen presentation.^75,76 In our studies, depletion of eosinophils by DTX during sensitization attenuated PNE-specific IgE/IgG1/IgG2b levels in iPhil^+/− FT^+/− pups. This critical impact at the allergen sensitization phase was substantiated by lack of protection from anaphylaxis with delayed depletion of eosinophils after allergen sensitization. Thus, these results suggest that after establishment of increases in mast cell numbers and induction of food allergen-specific IgE which can bind to FceR1 on mast cells, the stage is set for allergen responsiveness by mast cells during food allergen challenge. Nevertheless, even though eosinophils were not necessary for the food allergen-challenge induced anaphylaxis, they may have other roles during or after allergen challenge, as eosinophils are important during injury and repair.^48,49

Eosinophils exert immunoregulatory functions through direct effects and interactions with other cells. For example, mast cells and eosinophil have been reported to be juxtaposed within biopsy specimens from patients with allergic respiratory diseases and EoE.⁷⁷ Eosinophils also produce soluble lipid mediators, cytokines, and granule proteins that when released may impact function of other cells in the tissue. In our present studies, not only the activation but also the expansion of mast cells in skin and intestine in Alt+PNE-treated pups was blunted and plasma histamine was reduced when eosinophil depletion was sustained during allergen sensitization and challenge. This eosinophil regulation of skin and intestine mast cell expansion and increases in plasma histamine was primarily an effect of eosinophils during the allergen skin sensitization phase rather than the acute effects of eosinophils during the challenge phase. These data suggest that eosinophil generation of signals or cell-cell interactions in the skin may directly or indirectly induce expansion of mast cells in Alt+PNE sensitized skin of iPhil^+/− FT^+/− pups.

IL-33 has been shown to regulate increased intestinal mast cells in an OVA food allergy model with mechanical skin injury by skin tape stripping to remove layers of epiderimal skin.^66,78 In other tissues of allergic inflammation in the airways, IL-33 has been reported to be involved in lung epithelial expression of CCL11 and CCL24 for recruitment of eosinophils into the lung following airway allergen challenge, and this is attenuated in IL-33-deficient mice.⁷⁹ In the food allergy model with skin exposure to allergen in the FT^+/− mice, we previously demonstrated that Alt and PNE co-exposure induces pathways involving skin expression of IL-33, OSM and Areg for generation of PNE-specific IgE and for anaphylaxis.⁴⁷ This OSM and IL-33 were necessary for generation of PNE-specific IgE.⁴⁷ Our previous bulk RNAseq and single cell RNAseq analyses indicate that Alt+PNE induced an increase in IL-33 in skin keratinocytes and fibroblasts in FT^+/− pups.⁴⁷ As many other cell types can express IL-33 during allergic diseases⁸⁰ and because IL-33 can be secreted or act intracellularly on gene expression, ongoing studies are investigating the impact of fibroblast and keratinocyte expression of IL-33 on development of food allergy. In addition to increasing IL-33, our previous studies also revealed that during food allergy, Alt+PNE sensitized FT⁺⁻⁻ pups exhibited increases in numbers of eosinophils in skin and intestine.⁴⁷ Our data here with eosinophil depletion, throughout allergen sensitization and challenge, reduced Alt+PNE induction of plasma IL-33 and OSM as well as serum PNE-specific IgE, which is consistent with our previous report of OSM/IL-33 dependence for generation of PNE-specific IgE. However, when eosinophils were depleted only during sensitization or only during challenge, there was no effect on Alt+PNE-induced plasma IL-33, Areg, or OSM, and a trend for a decrease in serum PNE-specific IgE. Thus, a sustained eosinophil deficiency during sensitization and challenge was necessary to reduce these cytokines in the plasma. Also when eosinophils were depleted only during the sensitization phase, it reduced the increase in ileum mast cells after the allergen challenge phase without altering IL-33, suggesting eosinophils may regulate the mast cells numbers downstream or independent of IL-33. Nevertheless, eosinophils deficiency reduced skin and intestine mast cell expansion in the neonatal mouse FT^+/− food allergy model. Eosinophil mediators may communicate with several cell types in the skin, including keratinocytes, T cells, antigen presenting cells, neurons, etc to regulate mast cell expansion, the generation of antigen-specific antibodies and the development of responsiveness to PNE-induced anaphylaxis. These are currently under study and beyond the scope of this manuscript.

In conclusion, recruitment of eosinophils into allergen-sensitized skin of iPhil^+/− FT^+/− pups was necessary for food allergen sensitization during the development of food allergy, including expansion of mast cells in the skin, generation of increased plasma IL-33, OSM, and histamine and induction of the generation of food allergen-specific IgE (Fig. 10). After food allergy sensitization in the presence of eosinophils, the eosinophils were not necessary acutely during the challenge phase for oral food-induced anaphylaxis. However, since eosinophils have many other functions including injury repair, it does not preclude the possibility of other non-anaphylaxis-related functions for eosinophils in food allergy after sensitization. The presented study here suggests that there are underappreciated roles for eosinophils in food allergy. Early prevention of food allergy initiation, particularly the involvement of eosinophils, is poorly understood but has the potential to impact the early life generation of food allergy and thus impact life-long struggles with food allergies for those susceptible to development of food allergy. Ongoing investigations will provide insights into the specific actions of eosinophils and their derived mediators during food allergy early in life.

Funding

National Institutes of Health (NIH) grant number R01AI153239 (J.M.C.-M.).

Conflicts of interest

The authors have declared that no conflict of interest exists.

Data availability

The data underlying this article are available in the article.

References