IgE antibodies increase honeybee venom responsiveness and detoxification efficiency of mast cells

Background: In contrast to their clearly defined roles in allergic diseases, the physiologic functions of Immunoglobulin E antibodies (IgEs) and mast cells (MCs) remain enigmatic. Recent research supports the toxin hypothesis, showing that MCs and IgE-related type 2 immune responses can enhance host defense against certain noxious substances, including honeybee venom (BV). However, the mechanisms by which MCs can interfere with BV toxicity are unknown. In this study, we assessed the role of IgE and certain MC products in MC-mediated BV detoxification. Methods: We applied in vitro and in vivo fluorescence microscopyimaging, and flow cytometry, fibroblast-based toxicity assays and mass spectrometry to investigate IgE-mediated detoxification of BV cytotoxicity by mouse and human MCs in vitro. Pharmacologic strategies to interfere with MC-derived heparin and proteases helped to define the importance of specific detoxification mechanisms. Results: Venom-specific IgE increased the degranulation and cytokine responses of MCs to BVin vitro. Passive serum sensitization enhanced MC degranulation in vivo. IgE-activated mouse or human MCs exhibited enhanced potential for detoxifying BV by both proteolytic degradation and heparin-related interference with toxicity. Mediators released by IgE-activated human MCs efficiently degraded multiple BV toxins. Conclusions: Our results both reveal that IgE sensitization enhances the MC’s ability to detoxify BV and also assign efficient toxin-neutralizing activity to MC-derived heparin and proteases. Our study thus highlights the potential importance of IgE, MCs, and particular MC products in defense against BV.


| BACKG ROU N D
Mast cells (MCs) and immunoglobulin E antibodies (IgEs) are generally best known for their adverse functions in allergic diseases. 1 Allergic patients elaborate IgEs recognizing epitopes on environmental antigens, which typically represent no obvious danger. However, the exposure of sensitized hosts to antigens recognized by IgEs bound to MCs via the high-affinity IgE receptor, FcεRI, result in rapid MC activation with the release of many bioactive compounds, including MC-associated proteases and proteoglycans(eg, heparin), as well as diverse additional stored and newly synthesized mediators. 1,2 In concert, such MC-derived mediators cause immediate allergic symptoms and inflammation, 1,2 in extreme cases leading to potentially fatal anaphylaxis. 3 heparin and proteases. Our study thus highlights the potential importance of IgE, MCs, and particular MC products in defense against BV.

K E Y W O R D S
honeybee venom, host defense, IgE, mast cells, toxin hypothesis

G R A P H I C A L A B S T R A C T
Immunoglobulin E antibodies increase mast cell responsiveness to honeybee venom in vitro and in vivo. Compounds released by IgE-activated mouse and human mast cells can efficiently detoxify honeybee venom. Mast cell-derived heparin and proteases contribute to prevention of honeybee venom toxicity F I G U R E 1 Immunoglobulin E sensitization modulates FSCMC degranulation, signaling, and gene expression responses to honeybee venom. FSCMCs were sensitized overnight with serum derived either from mock (PBS)-immunized or BV-immunized mice (PBS serum or BV serum, respectively) that was either untreated or pre-treated (before cell incubation) with (A, B, D) an anti-IgE (or isotype control) antibody or (D) by heating as indicated. (A) FSCMCs were unstimulated or exposed to BV at 1, 5, or 10 µg/ml for 1 h and analyzed for ß hexosaminidase (ß-hex) release (data are from one of three independent experiments, each of which gave similar results). (B) FSCMCs were stimulated with 10 µg/ml BV and fluorescent signal in live cells due to heparin/avidin AF488 interaction was recorded over 20 min by flow cytometry. The avidin AF488 mean fluorescence intensity (MFI) kinetics are shown. The right panel represents a y-axis magnification of the data as indicated by the dashed lines in the left graph in order to allow identification of minor fluorescence differences (data are from one of three independent experiments, each of which gave similar results). (C) Phosphorylation of AKT, MAPK/ERK and PLCγ1 and membranelocalized Lamp-1 was analyzed in FSCMCs without stimulation (0 min) and after exposure to 10 µg BV for 1, 5, 10, or 30 min by (phospho-) flow cytometry (representative of three independent experiments). (D) Heatmap representing gene expression data (mean of triplicates) from realtime PCR analysis of cytokines and chemokines in FSCMCs after 1 h exposure to 10 µg/ml BV (fold change compared to nonstimulated cells; data are from one of two independent experiments, each of which gave similar results).(A and B) Two-way ANOVA with Tukey's test for multiple comparisons; (A) * indicates comparisons to the respective unstimulated ctrl in the group; ° indicates comparisons with the same concentration of the BV serum group; p values are adjusted for multiple testing; graphs represent mean +SD; (D) * (or n.s.-not significant) indicates comparisons of the fold changes of PBS serum/isotype-treated vs. BV serum/isotype-treated groups by t test (neither BV serum/anti-IgE-treated nor BV serum/heat-treated groups were significantly different compared to the PBS serum/isotype-treated samples for any of the investigated genes); ns: not significant;*/°p ≤ .05; **/°°p ≤ .01; ***/°°°p ≤ .001; numbers in (A) show the actual p value  Traditionally, IgE and MCs have been assigned a protective role in host defense against parasites. [4][5][6] However, the importance of such protective functions seems to depend on the respective parasite and model system. 5 In contrast to parasites, toxins and venoms represent acute and highly dangerous threats that require immediate counter-measures in order to avoid extensive damage, 7 and it has been suggested that allergies developed initially as protective responses to noxious substances. [8][9][10] Notably, several venoms either contain or are themselves allergens. 8,9 One example of a highly allergenic venom is honeybee (Apis mellifera) venom (BV). 11

| ME THODS
Additional methodological details regarding statistics, cellular models, analyses of gene expression, signal transduction, degranulation, viability, degradation of BV, proteomics, PLA 2 activity, imaging, and venom cytotoxicity, as well as more information regarding the mice, reagents, and antibodies used in this study, can be found in the online Appendix S1.

| Serum generation
For generation of mouse sera, C57BL/6 wtmice were shaved on the back skin and received two subcutaneous injections of 50 µl F I G U R E 2 Serum sensitization increases FSCMC activation at the single-cell level. (A-C) FSCMCs were incubated overnight with isotype control-pre-treated PBS serumor isotype control-or anti-IgE-pre-treated BV serum. Next day, cells were seeded in chamber slides and exposed to 10 µg/ml BV. The avidin SR -mediated signal of degranulating MCs was recorded over 60 min by realtime confocal fluorescence microscopy. ( Material) or PBS in the shaved area. 18 Three weeks later, mice were sacrificed and sera were collected.

| Statistics
Statistical analyses were conducted using GraphPad Prism (GraphPad Software), and tests were applied as indicated in the figure legends.
Differences with p values equal or below 0.05 were considered significant. All experiments were performed at least twice.

| IgE sensitization increases BV-mediated MC degranulation and signaling
We first tested the ability of BV to cause degranulation and cytotoxicity in mouse fetal skin-derived cultured MCs (FSCMCs).
Assessing degranulation by detecting release of the granule enzyme ß-hexosaminidase, and monitoring viability by flow cytometry, we found significant BV-related degranulation of FSCMCs at 1 or 10 µg/ ml BV without detectable effects on cell viability ( Figure S1A We then used flow cytometry to monitor FSCMC degranulation in real-time, assessing degranulation by fluorescence emitted upon binding of Alexa Fluor 488-tagged avidin [avidin AF488 ] to granular heparin exposed on the cell surface upon granule exocytosis. 21 We found that degranulation of BVserum-sensitizedFSCMCs, which began less than 5 min after addition of BV, reached a peak plateau after 10-15 min ( Figure 1B). Anti-IgE pre-treatment again abrogated degranulation to a level indistinguishable to that of cells mocksensitized with PBS serum ( Figure 1B).  Figure 1C). Interestingly, sensitization with BV serum markedly increased BV-mediated ERK phosphorylation ( Figure 1C).
The mediator release response of MCs can be tuned depending on the stimulation signals. 23,24 For instance, IgE/antigen-mediated activation via FcεRI leads to degranulation and extensive cytokine production, whereas stimulation viaG protein-coupled receptors, such as Mrgprb2, induces predominantly degranulation with minor de novo cytokine production. 24,25 We therefore wanted to test whether IgE sensitization affected mediator synthesis of BV-stimulated MCs.
We observed that BV exposure induced increased transcription of il1b, il4, il5, il6, il13, mip1a, ccl1, ccl2, and ccl4 ( Figure 1D). However, many elements of this innate response were profoundly amplified by sensitization with untreated BV serum, reflected by increased levels of il1b, il6, il13, tnf, mip1a, ccl1, and ccl2. Interference with IgE function, either by serum pre-treatment with anti-IgE or heating the serum to 56C°, 18 Figure 3C) and a greater effect on ERK phosphorylation than stimulation with BV alone at 1 min ( Figure 3D) and 5 min ( Figure S2A) after exposure. We conclude from these results that simultaneous stimulation with BV and IgE and antigen can induce a higher activation response in mast cells than either of the individual stimuli on its own.

| IgE-mediated activation increases the detoxification potential of mouse and human MCs
Having observed the effects of IgE (directed against either BV or DNP-HSA) sensitization on MC responses to BV, we next wanted to address the potential biologic relevance of this phenomenon. We performed PAGE (polyacrylamide gel electrophoresis) to assess the potential proteolytic effects of FSCMC compounds on the BV proteome. Exposure of BV to supernatants of IgE-activated F I G U R E 4 Immunoglobulin E-mediated activation increases the detoxification potential of mouse FSCMCs. (A and B) BV (12.5 µg at 500 µg/ml final concentration, resulting in a PLA 2 concentration of approximately 60 µg/ml) was either mixed with (A) indicated amounts of heparin or buffer or (B) supernatant of DNP-HSA-exposed non-sensitized or anti-DNP IgE-sensitized MCs for 10 min. After mixing with PLA 2 substrate and incubation, substrate cleavage (as indicator of PLA 2 activity) was measured. (C) BV (10 µg) was incubated (or not) with supernatant (MC SN) of unstimulated (unstim.) or BV-exposed (10 µg/ml; for 1 h) BV serum-sensitized FSCMCs for 1 h. In some cases, supernatant of activated FSCMCs was pre-incubated with protease inhibitor. After incubation of BV and supernatants, the mixtures were separated by PAGE, followed by Coomassie blue stain. (D-G) 3 T3 fibroblasts were seeded in chamber slides and stained with Fluo-4 (green). BV was pre-treated (or not) with protamine (100 µg)-or protease inhibitor (1x final concentration)-treatedMC SN of PBS serum or BV serum-sensitized FSMCs that were exposed to 10 µg/ml BV for 1 h. The pre-treated BV was then transferred onto the fibroblasts and monitoring of cell death, FSCMCs resulted in a shift in the low molecular weight venom fraction (below 10 kDa), presumably caused by degradation of toxins into smaller fragments ( Figure 4C). This effect could be partially rescued by pre-treatment of the FSCMC supernatant with a broadspectrum protease inhibitor ( Figure 4C). The predominant peptide in this low molecular weight BV fraction that was visibly modified by MC proteases is melittin (which has a molecular weight of approximately 2.8 kDa; Figure S3A), the main BV toxin with high cytolytic activity. 13 To assess the effects of the FSCMC supernatant on BV cytotoxicity, we exposed mouse 3 T3 fibroblasts to BV and monitored the cell   To test the contributions of released heparin and proteases to the observed BV detoxification, we pre-treated FSCMC supernatant with protamine, a heparin-antagonist, 28 or with a cocktail of protease inhibitors which do not mediate cytotoxicity ( Figure S4A,B).
Interestingly, while either treatment delayed the kinetics of BVinduced cytotoxicity, the fibroblasts eventually died. This suggests that both heparin-and protease-mediated mechanisms can contribute to efficient BV detoxification in this model ( Figure 4D-G and Videos S6 and S7).
We next addressed translational aspects of our findings, using human peripheral blood-derived cultured MCs (hu PBCMCs). Like FSCMCs, hu PBCMCs also exhibited enhanced degranulation when challenged with 1 or 10 µg/ml BV (reflected by increased supernatant ß-hexosaminidase and tryptase activity as well as avidin AF488 binding) without detectable cytotoxicity ( Figure 5A-B and Figure S5A-B). We then used PAGE to assess the potential of hu PBCMC-released mediators to degrade BV and observed a profound effect of supernatants from hu PBCMCs stimulated for 1 h with either BV (10 µg/ml) or IgE and anti-IgE ( Figure 5C and Figure S5C). Importantly, such treatment substantially decreased the immunorecognition of BV, reflected by a strongly diminished signal mediated by BV-specific serum IgG antibodies in Western blots ( Figure 5D).
Mass spectrometry analysis revealed that treatment of BV with supernatant collected from IgE-activated hu PBCMCs significantly decreased the abundance of BV toxins, including the major allergens 29,30 melittin (also known as Api m 4) and venom dipeptidyl peptidase 4 (DPP4, also known as Api m 5), as well as carboxypeptidase (also known as Api m 9) and vitellogenin (also known as Api m 12) (Figure 5E,F; Table S1). Of note, while not reaching statistical significance, many other BV components, such as Icarapin-like (also known as Api m 10) and venom carboxylesterase-6 (also known as Api m 8), also showed lower abundance after supernatant exposure ( Figure S5D; Table S1). Finally, detoxification experiments with supernatant collected from IgE and anti-IgE activated hu PBCMCs recapitulated our observations with FSCMCs: hu PBCMC supernatant counteracted BV cytotoxicity in a predominantly heparin but also partially protease-dependent manner ( Figure 6A-D and Videos S8-S11).
We conclude from this series of experiments that IgE-activated human PBCMCs and mouse FSCMCs share the potential of detoxifying BV, and that this reflects, at least in part, their release of both heparin and proteases.

| DISCUSS ION
MCs are found in most vascularized tissues of mammals, as well as in fish, reptiles, and urochordates, but their biological function has remained unclear. 4 As one of just a few cell types, MCs express the high-affinity IgE receptor, FcεRI, in a hetero-tetrameric configuration consisting of one IgE-binding α-chain, one transmembraneßchain, and two signal-transducing γ-chains (the γ-chain is also shared with several IgG Fc receptors). 31 Table S1 for full designations). Of the 39 proteins with lower abundance (FDR 0.1), the characterized compounds are indicated as white circles with blue border, the uncharacterized ones are indicated as white circles with gray border. Known allergens are annotated. (F) Heatmap of BV protein abundance of compounds with a statistically significantly different abundance (FDR 0.1). Raw and adjusted p values and fold changes are shown in Table S1. Raw abundances were log 2 and z-score transformed prior to visualization. (A-D) data are from one of at least two independent experiments each of which gave similar results. (A and B) graphs represent mean +SD; one-way ANOVA with Dunnett's test for multiple comparisons (p values are adjusted for multiplicity analysis); * indicates comparisons with the untreated (0 µg/ml BV) group; p values are adjusted for multiple testing; **p ≤ .01; ***p ≤ .001(E and F) Mass spec analysis of sample triplicates was performed once F I G U R E 6 Heparin and proteases released by IgE-activated human PBCMCs interfere with BV toxicity. (A) 3 T3 fibroblasts were seeded in chamber slides and stained with Fluo-4 (green). Untreated BV (20 µg/ml final concentration) or BV pre-treated with supernatants of hu PBCMCs (stimulated for 1 h by IgE/anti-IgE; hu MC SN) that were either untreated or pre-treated with protamine (100 µg) or protease inhibitor (1x final concentration) were transferred onto the fibroblasts, and monitoring by confocal fluorescence microscopy imaging over 60 min was started immediately.  In the current study, we found that while certain concentrations of BV can trigger active degranulation (independent of cell lysis) of non- However, treatment of MC supernatants with a pan protease inhibitor seemed to not fully restore the PAGE migration pattern of untreated BV, possibly indicating incomplete interference with protease activity. Also, the partial BV digestion due to limited protease inhibition was not sufficient to fully detoxify the venom. In summary, our findings suggest that both heparin and proteases can have important roles in BV detoxification.
In humans, population studies show that healthy individuals (ie, without diagnosed atopic disease) can express IgE specific for hymenoptera venom toxins. 36,37 This may indicate that toxin-specific IgE hassome non-pathologic function(s) in such people. While occasional bee stings are common in the general population, 11 beekeepers represent a subpopulation that is regularly exposed to BV. Up to ~30% of beekeepers respond with large local or systemic reactions to bee stings and up to 60% exhibit positive skin tests and detectable BV-specific IgE. 38 Remarkably, systemic allergic reactions seem to be more frequent in beekeepers who are infrequently stung. 38 This might be due to the higher levels of venom-specific IgG in such frequently exposed individuals, which outcompetes MC-bound IgE with the same epitope specificity. However, the precise function of BV-specific IgE in the increased venom tolerance of beekeepers 38 is still not fully understood.
Overall, our study has identified the contribution of two classes of preformed MC compounds, the proteoglycan heparin, as well as proteases, to resistance against BV toxicity. Importantly, we also show that specific serum IgE can substantially increase expression of the detoxifying potential of MCs. These findings thus reveal the significant benefit of this specialized anti-venom module of the adaptive immune system.