Frozen TIB-194 hybridoma cells (ATCC, Manassas, VA) were thawed according to manufacturer’s instructions, transferred and diluted with warmed RPMI-1640 (10% FBS, 100U/ml penicillin). Cells were centrifuged at 300 x g for 5 minutes, and the diluted freezing additive was removed and replaced with new media. Cells were propagated by the addition of fresh media approximately every 2-3 days. As cells multiplied, they were transferred into bigger flasks. Culture supernatants were tested by a sandwich ELISA for presence of anti-TNP IgA. Supernatants were collected and run over a protein L column (ThermoFisher Scientific, Waltham, MA). IgA was then concentrated and dialyzed with Amicon Ultra Centrifugal Filter Devices carrying a 100 kDa cut-off (Millipore Sigma, Darmstadt, Germany) and filter-sterilized with 0.2μM syringe filters (Millex, EMD Millipore, Billerica, MA).

BALB/cJ mouse bone marrow was cultured in RPMI-1640 media supplemented with 10% fetal bovine serum (Atlanta Biologicals, Lawrenceville, GA), 10mM HEPES buffer, 100 μg/ml streptomycin, 100U/ml penicillin 10μg/ml gentamicin, 1% Minimum Essential Medium non-essential amino acids, 1 mM sodium pyruvate, and 55μm 2-mercaptoethanol, (all from ThermoFisher Scientific, Waltham, MA) in the presence of both 10-20 ng/ml of IL-3 and SCF (Shenandoah Biotechnology, Warwick, PA) for the differentiation into BMMCs. BMMCs were cultured for 4-6 weeks until cultures reached 90% purity. The purity of BMMCs were assessed by flow cytometry (c-Kit⁺ FcϵRIα⁺).

For the lysosomal-associated membrane protein 1 (LAMP-1) assay, mouse BMMCs were incubated overnight with anti-TNP IgE (50 ng/ml, BD Biosciences, Franklin Lakes, NJ), and anti-TNP IgA (100 μg/ml, TIB-194 hybridoma, ATCC, Manassas, VA). Cells are then washed with RPMI to remove unbound antibody, and then stimulated for 10 mins at 37°C with TNP BSA (50 ng/ml, Biosearch Technologies, Teddington, UK) while simultaneously being stained with, BV605 anti-c-Kit (Biolegend, San Diego,CA), fixable viability dye eFluor 780 (eBioscience, San Diego, CA), and PE anti-LAMP-1. For antigen specificity experiments, anti-OVA IgA (gifted by Dr. Duane Wesemann), was used alongside anti-TNP IgA.

For imaging experiments, BMMCs were IgE/IgA sensitized as per the conditions above, and stained with NucBlue Live Cell Stain (Thermofisher Scientific, Carlsbad, CA), and anti-mouse Fc block and FITC anti-IgA (BD Biosciences, Franklin Lakes, NJ), or isotype control. Cells were then cytospun onto a slide. Slides were visualized and imaged using an EVOS M7000 microscope (Thermofisher Scientific, Bothell, WA).

For measurements of phospho-Syk, mouse BMMCs were incubated with anti-TNP IgE, and anti-TNP IgA as above, and stimulated at 37°C for zero, one and two minutes. Cells were fixed and stained using Thermofisher Scientific Protocol C: Two-step Protocol for Fixation/Methanol. Cells were stained with mouse PE anti-phospho-Syk (Cell Signaling Technologies, Danvers, MA), in the presence of mouse Fc block and assessed on an LSR Fortessa.

For phospho-Syk/Syk western blots, 10⁶ BMMCs were sensitized overnight with anti-TNP IgE, and anti-TNP IgA as described in conditions above. After sensitization, cells were stimulated with 100 ng/mL of TNP-OVA for 0–3 minutes. Cold FACS buffer was used to stop stimulation, and cells were pelleted and lysed with radioimmunoprecipitation assay buffer (Sigma-Aldrich) containing a cocktail of protease and phosphatase inhibitors (MilliporeSigma, Burlington, Mass).Lysed cells were then spun down to collect protein supernatant. Twenty microliters of lysate were resolved on a 4-15% MiniProtean TGX precast gel (Biorad, Hercules, Calif) and electrophoretically transferred to a polyvinylidene fluoride (PVDF) membrane (MilliporeSigma). Membranes were blocked with 5% BSA in tris-buffered saline with 0.05% tween 20 (TBST) for 1 hour. Membranes were probed with primary antibodies overnight, washed with TBST, and then probed with appropriate peroxidase-conjugated secondary antibodies for 1 hour. Membranes were developed with Supersignal chemiluminescent substrate reagent and imaged on an iBright Western Blot scanner (Thermo Fisher Scientific, Waltham, Mass).

For the cytokine release assay, BMMCs were stimulated for 6 hours. Supernatants were collected and cytokines were measured using the Cytometric Bead Array (CBA) Mouse/Rat Soluble Protein Master Buffer kit, along with the cytokine flex sets for IL-13, IL-6, and TNF-α according to manufacturer’s instructions (BD Biosciences, Franklin Lakes, NJ).

T cells were purified from a BALB/cJ spleen by MACS sorting using the CD3ε Microbead kit according to manufacturers instructions (Miltenyi Biotec, Bergisch Gladbach, Germany). BMMCs were incubated with 100 μg/ml anti-TNP IgA or anti-TNP OVA in 1X HBSS for 30 minutes at 4°C. Purified T cells were also incubated with 100 μg/ml anti-TNP IgA in 1X HBSS for 30 minutes at 4°C. Mast cells were stained for BV605 anti-c-Kit, PE anti-FcϵRI, fixable viability dye eFluor 780 (eBioscience, San Diego, CA), and FITC anti-IgA (BD Biosciences, Franklin Lakes, NJ) in the presence of anti-mouse Fc block (Biolegend, San Diego, CA), in FACS buffer (1X PBS, 10% FBS) at 4°C for 30 minutes. T cells were stained for PE anti-CD3 (eBioscience, San Diego, CA), BV605 anti-CD4 (Biolegend, San Diego, CA), FITC anti-IgA (BD Biosciences, Franklin Lakes, NJ), and fixable viability dye eFluor 780 (eBioscience, San Diego, CA), in FACS buffer.

Peritoneal lavage was isolated from BALB/cJ by injection of 5 ml of 1X HBSS into the peritoneal cavity to dislodge attached cells. Cells were pelleted and resuspended in 1X HBSS containing 100 μg/ml of anti-TNP IgA – incubated for 30 minutes at 4°C. Cells were then washed in FACS buffer and stained with mouse AF700 anti-CD45 (Biolegend, San Diego, CA), APC anti-CD11b (eBioscience, San Diego, CA), BV605 anti-c-Kit (Biolegend, San Diego, CA), PE anti-FcϵRIα (Biolegend, San Diego, CA), and FITC anti-IgA (BD Biosciences, Franklin Lakes, NJ.

For removal of N-linked glycans, anti-TNP IgA was incubated with PNGase F (New England Bioscience, Ipswich, MA) at 37°C for 72 hours according to manufacturer’s instructions. For desialylation, anti-TNP IgA was digested with neuraminidase from Vibrio cholerae (Sigma, Saint Louis, MO) at 37°C for 24 hours according to manufacturer’s instructions. The reactions (control IgA in glycobuffer, and digested IgA in glycobuffer and neuraminidase) were then incubated at 65°C for 10 min. to deactivate the enzymes. Digestion efficacy was assessed by lectin blot analysis.

As OIT induces a strong food allergen-specific IgA response, we used sera obtained from patients who had undergone peanut OIT as a source of peanut-specific IgA (25). Post OIT sera were pooled from several patients and IgA was purified using an anti-human IgA CaptureSelect affinity column as per manufacturer’s instructions (ThermoFisher Scientific, Carlsbad, CA). The IgA-enriched eluate from this column was then run through a Protein G column as per manufacturer’s instructions to further deplete it of IgG (ThermoFisher Scientific, Carlsbad, CA). The IgA was concentrated to a volume equivalent to that of the original input serum and dialyzed with Amicon Ultra Centrifugal Filter Devices carrying a 100 kDa cut-off (Millipore Sigma, Darmstadt, Germany) and filter-sterilized with 0.2 μm syringe filters (Millex, EMD Millipore, Billerica, MA). Total IgA and IgG was quantified using the Human IgA and IgG ELISA kits (ThermoFisher Scientific, Carlsbad, CA).

Basophil activation tests (BAT) were performed using the Flow CAST Basophil Activation Test kit (Bühlmann Laboratories, Schönenbuch, Switzerland). 50 μl aliquots of whole blood from peanut allergic patients were pre incubated with 125-1000 μg/ml of purified IgA for 2-4 hours at 37°C in 100 μl of basophil stimulation buffer. Samples were incubated for l5 min at 37°C with 3.6x10^-4 μg/ml of crude peanut extract (CPE) and staining antibodies for human PE anti-CCR3 (Biolegend, San Diego, CA) and FITC anti-CD63 (BD Biosciences, Franklin Lakes, NJ). Anti-FcϵRI stimulation was used as a positive control for basophil activation. After red blood cell lysis, cells were washed and assessed for activation by flow cytometry. Basophils were identified as SSC^lowCCR3⁺ and activated cells identified based on CD63 expression. Around 200 basophils were evaluated for each sample. Peripheral blood from peanut allergic donors was obtained with informed consent under a protocol approved by the Institutional Review Board of Boston Children’s Hospital.

Human IgA purified from post OIT sera was fluorescently labeled using the Pierce FITC antibody labeling kit (Thermofisher Scientific, Rockford, IL) according to the manufacturer’s instructions. For IgA binding experiments, donor whole blood was stained with labeled IgA, along with antibodies to human leukocyte markers PE anti-CCR3, APC anti-FcεR1, AF700 anti-CD3, PE-Cy7 anti-CD16, BV510 anti-CD14 all from (Biolegend, San Diego, CA) for 30 minutes at 37°C.

In order to evaluate the effects of IgA antibodies on IgE-mediated mast cell activation, we took advantage of the availability of monoclonal hapten-specific IgA antibodies as well as the well-characterized experimental model system of cultured BMMCs in which murine bone marrow cells are incubated with IL-3 and SCF to differentiate to c-Kit⁺ FcϵRI⁺ mast cells. Our own investigations as well as others have shown that IgG antibodies can inhibit IgE-mediated BMMC activation in a manner that is antigen specific and FcγR2b -dependent (e.g. the effect is not observed in FcγR2b^-/- BMMCs) (14, 15). To test the effects of IgA on IgE-mediated mast cell activation, BMMCs were incubated with IgE ± IgA monoclonal antibodies specific for the hapten, trinitrophenyl (TNP) prior to antigen exposure. Cells were then washed to remove unbound antibody, and then stimulated for 10-minutes with (TNP-BSA). LAMP-1, which is extruded to the plasma membrane surface during granule fusion, was measured by flow cytometry as an indicator of mast cell degranulation.

In the absence of anti-TNP IgE, TNP-BSA induced no detectable BMMC degranulation, with LAMP-1 expression only 0.1% ± 0.02% ( Figure 1 ). Anti-TNP IgE-sensitized BMMCs exposed to TNP-BSA exhibited a robust induction of surface LAMP-1 on the cell surface (16.9% ± 0.5%). Addition of anti-TNP IgA to these cultures resulted in a marked suppression of IgE-induced LAMP-1 expression (3.5% ± 0.2%) ( Figure 1 ). In contrast, incubating anti-TNP IgE sensitized mast cells with IgA directed against an irrelevant antigen, ovalbumin (OVA), did not impair IgE mediated activation (20.5% ± 0.7%), indicating that IgA mediated suppression is antigen specific. While anti-OVA IgA did not suppress TNP-IgE responses, we confirmed that it did still bind to BMMCs in a similar manner to anti-TNP IgA ( Figure S1 ).

The observation that IgA antibodies blocked IgE-induced activation of allergen-exposed BMMCs, led us to consider the possibility that IgA might be physically interacting with these mast cells. Surface expression of IgA on c-Kit⁺ FcϵRI⁺ cells was measured by flow cytometry following incubation of BMMCs with anti-TNP IgA for 30 minutes in HBSS. In the presence of IgA, BMMCs uniformly stained with anti-IgA FITC in a dose dependent manner ( Figure 2A , upper and lower panel), however no IgA binding was observed with purified splenic T cells. Immunofluorescence imaging of BMMCs, incubated with or without anti-TNP IgA similarly revealed the presence of staining (green) on mast cells incubated with IgA ( Figure 2B ). In order to evaluate IgA binding to primary mast cells, we examined peritoneal lavage cells. Staining for surface IgA after incubation with anti-TNP IgA, revealed IgA binding to CD11b^- c-Kit⁺ FcϵRI⁺ mast cells isolated from the peritoneum ( Figure 2C ).

As mast cells do not express any known IgA receptors, we sought to characterize the physical properties of IgA binding to BMMCs. To test calcium dependence, exogenous calcium was chelated by addition of EGTA to BMMC suspensions prior to incubation with IgA. Calcium removal dramatically impaired IgA binding with a decrease in MFI from 3060 ± 13.4 to 961 ± 79.8 ( Figure 3A ), but not IgE binding ( Figure S2 ). We similarly observed that BMMCs suspended in calcium-free PBS failed to bind IgA (data not shown).

The abrogated IgA binding after calcium chelation suggested interaction with a calcium dependent receptor. Some such receptors are lectins, leading us to consider that the carbohydrate moieties on IgA might be mediating its interaction with BMMCs. To test this, we treated anti-TNP IgA with either PNGase F to remove all N-linked sugars, or neuraminidase to remove terminal sialic acid residues. Either N-deglycosylation of anti-TNP IgA or the removal of sialic acid rendered the antibody completely unable to bind to BMMCs ( Figure 3C ) without affecting its ability to bind to TNP-OVA ( Figure 3B ). To corroborate this sialic acid dependence of IgA binding to BMMCs, we incubated cells with sialic acid for 30 minutes prior to addition of IgA and observed suppression of IgA binding to undetectable levels ( Figure 3D ). The suppression of this response is dose dependent (data not shown). Using desialylated IgA, we also confirmed a functional requirement for sialic acid in the ability of IgA to suppress IgE-mediated mast cell activation. We observed that while untreated IgA inhibited LAMP-1 induction (of 8.8% ± 0.06% in treated compared to 16.2% ± 1.2% in controls), desialylated anti-TNP IgA was incapable of suppressing IgE-mediated BMMC degranulation (18.3% ± 2.3%) ( Figure 3E ). We could not similarly establish the effects of calcium depletion on the inhibitory function of IgA antibodies since it is not possible to perform mast cell activation assays in the absence of calcium. These experiments reveal that the binding of IgA to BMMCs is dependent both on ambient calcium concentration and on sialylation of the IgA antibodies and that sialyation is also critical for functional inhibition of mast cell activation by IgA.

IgE-antigen receptor crosslinking on BMMCs results in the phosphorylation of the protein tyrosine kinase (PTK) Syk. Activation of a PTK signaling cascade downstream of Syk drives many of the phenotypes of activated mast cells, including degranulation, synthesis of arachidonate-derived lipid mediators and induction of cytokine transcription. In order to establish if IgA antibodies block this critical signaling pathway, we evaluated their effects on phosphorylated-Syk (p-Syk), the active signaling form of this PTK. BMMCs loaded with anti-TNP IgE and stimulated with TNP-BSA exhibited rapid increases in p-Syk levels, peaking one minute after stimulation (MFI 298 ± 7.6) and then quickly decreasing ( Figure 4A ). In contrast, allergen-exposed BMMCs sensitized with anti-TNP IgE and also incubated with anti-TNP IgA exhibited attenuated p-Syk induction (176.6 ± 4.7) 1 minute after stimulation. To further validate these results as well as to normalize p-Syk to total Syk, we performed phospho-western blot analysis. This similarly showed Syk phosphorylation detectable as early as 1 minute after stimulation ( Figure 4A ). Quantification of the ratio of signal intensities of phospho-Syk to total Syk revealed that phosphorylation induced by IgE was less at all time points in BMMCs that were also incubated with IgA.

Cytokine production is an important property of activated mast cells and is critical for the generation and tissue recruitment of the effector cells of inflammation as well as for the expansion of type 2 adaptive immune responses. We evaluated the effects of antigen-specific IgA on IgE-induced cytokine production by BMMCs. As expected, IgE-sensitized BMMCs exposed to antigen produce IL-13, TNF-α, and IL-6. Addition of IgA to these IgE-sensitized BMMCs prior to antigen stimulation resulted in a complete suppression of the production of these cytokines ( Figure 4B ). These results indicate that IgA acts early in the FcεRI signaling cascade, inhibiting formation of the most proximal signaling intermediate, p-Syk and that IgA-mediated blockade of mast cell activation extends beyond degranulation to also affect cytokine production.

Peanut-specific IgG and IgA antibodies are both present in the plasma of subjects with peanut allergy and both of their levels increase after OIT. We have observed that IgE-mediated activation of basophils is potently suppressed by IgG signaling via FcγR2b (14). We took advantage of post OIT sera as a source of PN specific IgA to test whether IgA exerts similar protective effects as PN-IgG in peanut allergic patients. IgA was enriched from these patient sera by affinity chromatography and further purified by IgG depletion using Protein G Sepharose, yielding a preparation containing 2.5 mg/ml IgA and no detectable IgG. Whole blood samples from a peanut allergic donor (with confirmed IgE mediated allergy to peanut) were pre-incubated for two to four hours with varying concentrations of purified IgA (125-1000 μg/ml). The samples were then stimulated with CPE and expression of CD63 was used to as a marker of basophil activation. Basophil activation was markedly suppressed in the presence of IgA ( Figure 5A ). This inhibitory effect of IgA was dose dependent over a range of IgA amounts from(250-1000 μg/ml) ( Figure 5A ). While IgA has been found to interact with several human cell types, including neutrophils and eosinophils (26, 27), its interaction with mast cells and basophils is incompletely characterized. At least one report has suggested an activating effect of IgA antibodies on basophils (28). We sought to determine if IgA might bind to basophils. We stained whole blood with fluorescently labeled human IgA along with antibodies defining the key leukocyte subsets. As expected, IgA was bound by neutrophils but not by T cells ( Figure 5B ). Basophils also exhibited IgA binding with a unimodal fluorescence shift relative to isotype control. These observations indicate that IgA antibodies, like IgG antibodies, bind to and are capable of inhibiting the activation of antigen stimulated basophils from peanut-allergic subjects.