|Home | About | Journals | Submit | Contact Us | Français|
IgE-mediated food allergy is a common cause of enteric disease and is responsible for approximately 100 systemic anaphylaxis deaths in the USA each year. IgG antibodies can protect against IgE-mediated systemic anaphylaxis induced by injected antigens by neutralizing antigens before they can bind to mast cell-associated IgE.
We have investigated whether IgA and IgG antibodies can similarly protect against systemic, IgE-mediated anaphylaxis induced by ingested antigens and, if so, whether IgA and IgG antibodies protect by neutralizing antigens before or after their systemic absorption.
Murine passive and active anaphylaxis models were used to study the abilities of serum vs. gut lumenal IgA antibodies and serum IgG antibodies to inhibit systemic anaphylaxis induced by ingested allergens in normal mice, mice deficient in the ability to secrete IgA into the intestines, and mice in which intestinal IL-9 overexpression has induced intestinal mastocytosis and increased intestinal permeability.
IgE-mediated systemic anaphylaxis and mast cell degranulation induced by antigen ingestion are suppressed by both serum antigen-specific IgA and IgG, but not by IgA within the gut lumen.
Systemic, rather than enteric antibodies protect against systemic anaphylaxis induced by ingested antigen. This implies that ingested antigens must be absorbed systemically to induce anaphylaxis and suggests that immunization protocols that increase serum levels of antigen-specific, non-IgE antibodies should protect against severe food allergy.
Induction of a systemic IgG or IgA antibody response against a food allergen should protect against induction of systemic anaphylaxis by ingestion of that allergen.
Systemic anaphylaxis, which can be characterized by urticaria, angioedema, bronchospasm, diarrhea, dysrhythmias and cardiovascular collapse, is responsible for approximately 150,000 emergency department visits 1, 15,000 hospital admissions 2–3 and 1,500 deaths 4–6 each year in the United States. Although parenteral allergen administration is more likely to trigger systemic anaphylaxis than ingestion of the same allergen, the high prevalence of food allergy, coupled with the much greater likelihood of eating an allergenic protein than being injected with an allergenic protein makes food allergy responsible for ~one third to one-fifth of the emergency department visits for anaphylaxis 1, 7 and 100–200 deaths annually 4, 7–8 in the U.S. The immune mechanisms that cause food-induced systemic anaphylaxis and the immune mechanisms that may protect against food-induced systemic anaphylaxis are not as well understood as those that promote and protect against parenteral allergen-induced anaphylaxis and are likely different. Systemic anaphylaxis elicited by allergen injection, for example, can be induced in mice by either an IgE/FcεRI/mast cell-dependent mechanism or an IgG/FcγRIII/basophil- or macrophage-dependent mechanism 9–10, while triggering of systemic anaphylaxis by allergen ingestion appears to be always or nearly always IgE, FcεRI and mast cell dependent 11–12. In addition, while histamine is the predominant mediator responsible for IgE-dependent anaphylaxis induced by injected allergen, PAF and serotonin appear to have a more important role in IgE-dependent anaphylaxis induced by ingested allergen 9, 11.
The evidence that Ig isotypes other than IgE have little or no role in food allergen-induced systemic anaphylaxis, when coupled with evidence that non-IgE antibodies(Abs) can protect against IgE-mediated anaphylaxis caused by injected antigens (Ags) by binding allergen epitopes before they can react with mast cell-associated, allergen-specific IgE 13, raises questions about whether these other isotypes can also protect against food allergen-induced anaphylaxis. Furthermore, if allergen-specific non-IgE Abs are protective, do they protect by binding ingested allergen in the gut lumen, before it has been absorbed, in the same way that IgA neutralizes intestinal toxins and blocks bacterial binding to epithelial receptors14–18, or by binding to allergen after it has been adsorbed systemically. This question is related to an additional issue – does ingested allergen induce systemic anaphylaxis predominantly by activating mucosal mast cells that are interspersed with mucosal epithelial cells at the interface of intestinal villi with the gut lumen, in which case systemic allergen absorption may not be necessary, or by interacting with mast cells that are associated with lymphatics and blood vessels, in which case systemic absorption is likely to be important.
These questions have clinical implications: if ingested allergens do not have to be absorbed systemically to induce systemic anaphylaxis, allergen-specific Ab, presumably of the IgA isotype, would have to be secreted into the gut lumen to intercept allergen before it could activate mast cells and induce anaphylaxis. In contrast, if induction of systemic anaphylaxis by ingested allergens requires their systemic absorption, then IgG and non-secretory IgA Abs should be able to inhibit systemic anaphylaxis induced by ingested allergens. These alternative possibilities should influence strategies for the optimal induction of Abs able to inhibit food allergy-related systemic anaphylaxis.
To address these issues, we have used both passive and active anaphylaxis models to study the ability of secreted vs. non-secreted IgA Abs and IgG Abs to inhibit systemic anaphylaxis induced by ingested allergens in three models: 1) normal mice that have been sensitized passively by injection of a TNP-specific IgE antibody; 2) IgE anti-TNP mAb passively sensitized mice in which intestinal IL-9 overexpression has induced intestinal mastocytosis; and 3) normal mice in which i.p., followed by oral immunization with ovalbumin has induced both intestinal mastocytosis and an IgE anti-ovalbumin antibody response. Our observations provide evidence for suppression of systemic anaphylaxis by both IgG and IgA Abs and for better suppression of systemic anaphylaxis by systemic rather than by enteric IgA. These observations support a requirement for systemic absorption of ingested allergens to induce systemic anaphylaxis and favor the adoption of immunization strategies capable of inducing high titers of IgG antibodies to anaphylaxis-inducing food allergens.
BALB/c mice, BALB/c background IL-9 tgn 12, PIgR deficient (Jackson Lab, Bar Harbor, ME) 19, Fcα/µR deficient 20, J-chain deficient (a gift from Dennis Metzger, Albany Medical College) 21 and FcγRIIb deficient 22 mice were bred in-house. All experimental procedures were performed with approval from the Institutional Animal Care and Use Committees of the Cincinnati Children’s Hospital Research Foundation, which follows the “Guide for the Care and Use of Laboratory Animals” prepared by the Institute of Laboratory Animal Resources, National Research Council, and published by the National Academy Press.
TNP-labeled OVA was prepared as previously described 13. OVA and propranolol were purchased from Sigma (St. Louis, MO). Serum levels of MMCP-1 were measured by ELISA according to the manufacturer’s instructions (Moredun Scientific, Midlothian, United Kingdom).
The following hybridomas were grown as ascites in Pristane-primed athymic nude mice and purified by ammonium sulfate precipitation, followed by DE-52 cation exchange chromatography when appropriate: mouse IgE anti-TNP (IGEL 2a); IgA anti-TNP (2F.11.15), IgA isotype control mAb (J558) and IgG1 anti-TNP (1B7.11) from the ATCC (Rockville, MD) and rat IgG1 anti-mouse IL-4 mAb (11B11) was purchased from Verax (Lebanon, NH). Recombinant mouse IL-4 was purchased from PeproTech (Rocky Hill, NJ).
Serum rich in IgA anti-OVA antibody was collected from J-chain deficient mice that were immunized i.p. with 50 µg of OVA adsorbed to 1 mg of alum; then, starting 2 weeks later, boosted multiple times by oral gavage with 50mg of OVA. Serum pooled from several OVA-immunized mice was heated to 56°C for 30 minutes to inactivate all IgE.
IL-4 was administered as a complex (IL-4C) of two molecules of IL-4 bound by one molecule of a monoclonal neutralizing anti-IL-4 mAb, 11B11, which were mixed 5 min or more before injection. This complex dissociates in vivo over a 3 d period, releasing IL-4. IL-4C itself is inactive because 11B11 blocks binding to IL-4Rs. IL-4C does not activate complement or bind more avidly than uncomplexed IgG to FcγRs, because it contains a single IgG molecule 23–24.
Five freshly excreted fecal pellets per mouse were collected on ice in a pre-weighed 1.5ml tube. The pellets were weighed and 25 µl of protease inhibitor cocktail (Sigma P2714, St. Louis, MO) in 225 µl PBS was added per 50 mg of feces and vigorously mixed to inhibit further protein digestion. Samples were then diluted 10-fold with PBS. Tubes were vigorously mixed until the pellet was completely suspended in solution. Samples were then centrifuged at 3000g for 10 min and supernatants transferred into fresh 1.5 ml tubes containing 10 µl of protease inhibitor cocktail. Samples were then re-centrifuged at 12,000g for 5 min and supernatant collected and stored at −80°C until analyzed.
Mice (5–6/grp unless noted otherwise) were injected i.v. with 10 µg of IgE anti-TNP mAb followed 24 hr later by o.g. challenge with TNP-OVA in 300 µl saline. Some mice were also injected i.v. with 1 µg of IL-4C (complex of 1 µg IL-4 with 5 µg 11B11) 24 hr prior and with 0.03 mg of propranolol 30 min prior to o.g. challenge. Anaphylaxis severity was determined by change in activity scores and rectal thermometry 9, 23 with a Digital Thermocouple Thermometer (Model BAT-12; Physitery Instruments, Clifton, NJ). All s tudies were repeated at least once to assure reproducibility.
Mice (10/grp) were injected i.p. with 50 µg of OVA adsorbed to 1mg of alum. Starting 14 days later, these mice were inoculated o.g. 3 times a week with 50 mg of OVA in 300 µl of saline. Mice were followed for up to 120 minutes post-OVA challenges for the development of diarrhea (intestinal anaphylaxis) and hypothermia (an indicator of systemic shock).
Standard sandwich ELISA technique was used with 96-well microtiter plates coated with anti-chicken egg albumin mAb (Sigma, St. Louis, MO) followed by anti-chicken-OVA-HRP as detection antibody (Sigma, St. Louis, MO) and SuperSignal ELISA substrate (Pierce Biotechnology, Cheshire, United Kingdom). Known quantities of TNP-OVA were used as standard.
IgA levels were determined by standard sandwich ELISA, with microtiter plate wells coated with anti-mouse IgA mAb (BD Bioscience) followed by sample and standard. After incubation for 60 minutes, anti-mouse IgA-biotin (BD Bioscience), streptavidin-HRP and SuperSignal ELISA substrate (Pierce Biotechnology, Cheshire, United Kingdom) were added sequentially. Purified mouse IgA (BD Bioscience) was used as standard.
The significance of differences in temperature, TNP-OVA, MMCP-1 and IgA concentrations between grps of mice was compared using the Mann-Whitney t test (GraphPad Prism 5.0; GraphPad software). A p value <0.05 was considered significant.
In humans systemic anaphylaxis can occur within minutes of ingestion of an allergen25 suggesting that triggering of anaphylaxis either occurs at or near the surface of the gut lumen or that the offending allergen is rapidly absorbed into the systemic circulation in sufficient amount to induce anaphylaxis. To test the hypothesis that ingested Ags must be rapidly absorbed systemically to induce systemic anaphylaxis, we first evaluated whether ingested Ag can rapidly induce systemic anaphylaxis and be systemically absorbed in sufficient quantity and with sufficient speed to account for systemic shock. BALB/c mice sensitized by i.v. injection of 10 µg of an IgE anti-TNP mAb all developed mild clinical anaphylaxis, manifested as reduced movement, 5–10 min after i.v. injection of 1 µg of TNP-OVA or oral gavage (o.g.) of 50 mg of TNP-OVA, although the hypothermia induced by the oral TNP-OVA was considerably less severe than that induced by the iv. TNP-OVA (Fig. 1A). This difference in severity probably resulted from the much higher concentration of TNP-OVA immediately after i.v. challenge with 1 µg of TNP-OVA (calculated to be ~800 ng/ml, based on a mouse plasma volume of ~1.25 ml), than that induced by oral gavage with 50 mg of TNP-OVA, which reaches ~80 ng/ml 5 minutes after gavage (Fig. 1B). At this 5 minute timepoint, plasma TNP levels in the intravenously challenged mice had declined to ~10 µg/ml. These observations demonstrate that: 1) ingested Ag can be absorbed systemically with a speed consistent with the kinetics of development of systemic anaphylaxis; and 2) the severity of systemic anaphylaxis induced in this system appears to be related more closely to the initial or th e peak plasma concentration of Ag to which mast cells are sensitized, rather than to how long the Ag concentration remains elevated.
The requirement for a high oral dose of TNP-OVA to induce anaphylaxis in our system had two disadvantages: 1) after adjusting for differences between mouse and human weight, it was disproportionate to the doses of ingested Ag that are known to induce anaphylaxis in some sensitized humans; and 2) it was too large for it to be practical for us to try to neutralize it by mixing it with an equimolar amount of anti-TNP mAb. To address both issues, we adopted a previous observation that pre-treatment with a long-acting form of IL-4 (IL-4C) decreases the dose of injected Ag required to induce anaphylaxis by making mice more sensitive to mediators released by mast cells 23. This IL-4-dependent increase in sensitivity is observed in mice that have been induced to generate a strong Th2 response and may also occur in food-allergic humans. As expected, IL-4C pre-treatment decreased the dose of ingested TNP-OVA required to induce measurable shock by a factor of ~50 (Supplementary Fig. 1). To further increase sensitivity to ingested Ag, we also treated mice with the β-adrenergic antagonist propranolol, which inhibits the ability to compensate for the decreased intravascular volume caused by vascular leak, the predominant pathophysiologic mechanism responsible for murine anaphylaxis. Similar to IL-4C, propranolol pretreatment decreased the dose of ingested TNP-OVA required to induce measurable shock around 50-fold (Supplementary Fig. 1). Together IL-4C and propranolol increased sensitivity to oral Ag challenge ~250-fold, causing mice sensitized by IgE anti-TNP mAb to develop mild hypothermic shock following o.g. challenge with 100 µg of TNP-OVA and severe hypothermic shock following o.g. challenge with 1000 µg of TNP-OVA (Supplementary Fig. 1). This amount would be closely equivalent on a weight basis to the ingestion of one averaged-sized peanut by a typical 8 year old child.
Because IgG Ab in blood can inhibit IgE-mediated anaphylaxis induced by i.v. Ag injection 9, 13, we hypothesized that IgA Ab, which can be induced by oral vaccination, might be able to similarly suppress anaphylaxis induced by oral Ag ingestion. Surprisingly, mixing TNP-OVA with IgA anti-TNP mAb prior to oral gavage had no effect on the severity of the systemic anaphylaxis (Fig. 2A). Similar negative results were observed when IgA anti-TNP mAb was administered by o.g. minutes to hours prior to o.g. challenge with TNP-OVA (Fig. 2A) and when the dose of IgA anti-TNP mAb was increased to make the molar ratio of anti-TNP mAb to TNP 1.9. (Figure 2B). In contrast, i.v. injection of a considerably lower amount of IgA anti-TNP mAb (molar mAb to Ag ratio of 0.05) in similarly sensitized mice immediately prior to TNP-OVA oral gavage significantly inhibited anaphylaxis (Figure 2A). Intravenous injection of IgA anti-TNP mAb even inhibited systemic anaphylaxis when mice that had not been pretreated with IL-4C or propranolol were inoculated with 50 mg of TNP-OVA by o.g. (Figure 2C). The greater inhibitory effect of i.v. injected IgA mAb than ingested IgA mAb makes sense – the molar ratio of i.v. injected IgA anti-TNP mAb to the small percentage of ingested TNP-OVA that is absorbed systemically is much higher than the molar ratio of ingested IgA anti-TNP mAb to ingested TNP-OVA in all of these experiments – but only if ingested Ag has to be absorbed systemically to induce systemic anaphylaxis.
These results could also be explained, however, by the trivial possibility that i.v. IgA mAb injection protected against systemic anaphylaxis by increasing plasma volume and oncotic pressure rather than by neutralizing Ag. Although this seemed unlikely, because all IgA mAb injections were controlled by injection of an equal volume of saline and the oncotic contribution of injected IgA mAb would be small relative to that of plasma albumin, an experiment was performed in which injection of an IgA mAb that lacks relevant Ag specificity was used as a control for IgA anti-TNP mAb. Results of this study (Fig. 2D) confirmed the Ag-specific protective effect of IgA anti-TNP mAb.
Although these observations provide strong evidence that IgA mAb protects against systemic anaphylaxis caused by ingested Ag by neutralizing Ag that had been absorbed systemically, it remained possible that neutralization of Ag by IgA that had been actively secreted into the gut (or was in the process of being secreted) was also important. Indeed, some i.v. injected IgA mAb appears 2–4 hr later in defecated feces, although the amount is ~100-fold less than the amount of ingested IgA that is recovered in feces within 1 hr (Fig. 3A and B). To limit the amount of i.v. injected IgA anti-TNP mAb that could be secreted into the gut lumen, we injected this mAb into mice deficient in the polymeric Ig receptor (PIgR), which is required to secrete IgA 19, 26. Only trivial quantities of IgA mAb injected into these mice could be recovered in feces, even when PIgR mice had first been induced by an active OVA immunization protocol to develop allergic diarrhea, which is accompanied by a considerable increase in intestinal permeability (Fig 3C). Despite the very limited passage of i.v. injected IgA mAb into the gut lumen in PIgR mice, i.v. injected IgA anti-TNP mAb protected PIgR-sufficient and deficient mice equally well against anaphylaxis induced by TNP-OVA ingestion (Figure 3D). Similar results were obtained using J-chain deficient mice, which are also defective in their ability to secrete IgA into the intestinal lumen18, 21 (data not shown). Thus, active secretion of IgA Ab does not contribute to its ability to protect against anaphylaxis.
To confirm this conclusion, we compared the abilities of IgA anti-TNP mAb (which can be secreted into the gut of PIgR- and J chain-sufficient mice) and IgG1 anti-TNP mAb (which cannot be secreted) to protect wild-type, IgE anti-TNP mAb-sensitized BALB/c mice against systemic anaphylaxis induced by TNP-OVA ingestion. Both IgA and IgG1 mAbs bound TNP-OVA with high affinity (not shown). The two Ig isotypes protected equally well against anaphylaxis induced by ingested TNP-OVA (Fig. 3E), confirming the lack of importance of Ab secretion in protection against oral Ag-induced systemic anaphylaxis.
Because systemic anaphylaxis in our system is mast cell-dependent9, 13, IgA mAb would be expected to protect against anaphylaxis by inhibiting mast cell degranulation. To more directly evaluate this expectation, we measured serum levels of mouse mast cell protease 1 (MMCP-1), an enzyme released by degranulating mast cells 9, 11, 12, 23, in IgE-anti-TNP mAb-sensitized mice that have been challenged orally with TNP-OVA after injection of IgA anti-TNP mAb or saline. As expected, IgA anti-TNP pretreatment considerably reduced the MMCP-1 response to oral TNP-OVA challenge (Fig. 4).
All studies up to this point had been performed with wild type mice that had been passively sensitized with IgE and, therefore, had baseline numbers of intestinal mast cells. Because humans with food allergies have increased numbers of intestinal mast cells27–28 and increased numbers of intestinal mast cells are associated in mice with increased intestinal permeability12, 29, it was possible that mastocytosis and increased intestinal permeability might eliminate the systemic absorption requirement for ingested Ag to induce systemic anaphylaxis. If so, injected IgA mAb would lose its ability to protect against systemic anaphylaxis. To test this possibility, we studied mice that express an IL-9 transgene regulated by the intestinal fatty acid binding promoter (iFABP, which induces gene expression only in small intestinal enterocytes)12. iFABP-IL-9 transgenic mice (IL-9 tgn) have increased intestinal permeability as well as a large increase in the number of intestinal mucosal mast cells; these abnormalities make it possible to induce systemic anaphylaxis in these mice by challenge with a relatively low concentration of TNP-OVA following sensitization with IgE anti-TNP mAb in the absence of pretreatment with IL-4C or propranolol. Even in this system, i.v. injection of IgA anti-TNP mAb still protected against anaphylaxis induced by oral ingestion of TNP-OVA (Fig. 5). Thus, systemic IgA protects against induction of anaphylaxis by ingested Ag even in mice that have the intestinal mastocytosis characteristic of chronic allergic inflammation. Thus, Ag must be absorbed systemically, even under these circumstances, to induce systemic anaphylaxis.
Mice express a functional receptor for both IgA and IgM (Fcα/µR) 30–31. Because it is not known if this receptor has stimulatory or inhibitory effects on cellular activation or if it is expressed on mast cells, we could not rule out the possibility that signaling through this receptor, rather than IgA interception of Ag, was involved in IgA protection against systemic anaphylaxis. To address this possibility, we evaluated the ability of IgA anti-TNP mAb to protect Fcα/µR-sufficient and deficient mice 20 against the induction of IgE-mediated anaphylaxis by ingested TNP-OVA. Results of this experiment demonstrate equal protection by IgA anti-TNP mAb in Fcα/µR-sufficient and deficient mice (Fig. 6); thus, signaling through this receptor does not contribute to IgA-mediated protection against IgE-dependent anaphylaxis. Similar studies with FcγRIIb-deficient mice demonstrated that this inhibitory receptor, which has been described to indirectly interact with IgA through galectin-3 32–34, is also not involved in IgA-mediated protection against IgE-dependent anaphylaxis (data not shown).
Although the use of a passive anaphylaxis model facilitates interpretation of experiments by allowing tight definition of the isotypes and quantities of antibodies present, it was possible that the requirement for systemic absorption of ingested Ag might differ in a more complex active anaphylaxis model, where immunized mice produce multiple isotypes of Ag-specific Ab. To evaluate the possible role of IgA as an inhibitor in an active anaphylaxis model, we sensitized WT mice to OVA by injecting them with OVA/alum i.p. followed by several oral OVA challenges until the mice developed systemic anaphylaxis to the oral OVA challenge. After 3 days rest, mice were injected i.v. with two 0.5 ml aliquots of pooled heat-inactivated serum from OVA-immunized J-chain deficient mice, which has several fold higher titers of OVA-specific IgA than serum from similarly immunized WT mice (data not shown) or with heat-inactivated serum from un-sensitized mice. Figure 7 demonstrates that the IgA-rich OVA-immune serum from J chain-deficient mice effectively inhibited systemic anaphylaxis. Al though we cannot exclude the possibility that protection is mediated by transferred serum IgG, instead of IgA, this would still be consistent with our central hypothesis that ingested Ag must be absorbed systemically to induce systemic anaphylaxis, even when mice are sensitized by active immunization.
Inspired by the demonstration that IgG Ab can suppress IgE-mediated anaphylaxis induced by i.v. injected Ag 13, we evaluated whether a non-IgE Ab isotype can also protect against IgE-mediated systemic anaphylaxis induced by an ingested Ag. These studies, which are justified by the high frequency and potential severity of IgE-mediated food allergy 3, 5–7, focused on IgA Ab. This focus was warranted by the greater production of IgA than any other isotype, with most IgA produced and secreted at mucosal surfaces 35 ; by demonstration in murine IgE-mediated food allergy models of an inverse correlation between food allergy severity and IgA titers 36–37 ; and by the association of human IgA deficiency with increased atopic disease and asthma 38. Furthermore, studies of the protective effects of IgA against enteric pathogens 14–16 and toxins 17–18 seemed to provide a model of how IgA might prevent anaphylaxis caused by ingested allergens. IgA can inhibit systemic disease caused by pathogens that gain entry to mucosal surfaces by blocking pathogen molecules required for binding to mucosal surfaces and can similarly inhibit the effects of toxins produced by intestinal pathogens by neutralizing those toxins before they can bind to host receptors. Consequently, it seemed reasonable to hypothesize that IgA Abs to allergens could inhibit food allergen-induced anaphylaxis by binding to allergens before they could be absorbed systemically.
Our experimental results, however, do not support this hypothesis and instead, point to the surprising conclusion that ingested Ag must be absorbed systemically to induce systemic anaphylaxis. Mixing 10 mg of IgA anti-TNP mAb with 300 µg of TNP-OVA prior to o.g. inoculation failed to inhibit systemic anaphylaxis in mice primed with IgE anti-TNP (Fig. 2B), while i.v. injection of a considerably lower dose of IgA anti-TNP mAb inhibited systemic anaphylaxis induced by ingestion of an even higher dose of TNP-OVA (Fig. 2A). Intravenously injected IgA anti-TNP mAb was just as effective at inhibiting TNP-OVA-induced IgE-mediated anaphylaxis in PIgR- and J chain-deficient mice (Fig. 3D), which secrete only trivial quantities of IgA even in the presence of allergic intestinal inflammation, as in non-immune wild-type mice (Fig. 3C). Similarly, IgG anti-TNP mAb, which cannot be secreted into the gut lumen even in WT mice, was effective as IgA anti-TNP mAb at suppressing IgE-mediated systemic anaphylaxis induced by ingested TNP-OVA (Fig. 3E). The trivial quantities of IgA (or IgG) that leak into the intestinal lumen in these mice (Fig. 3C) could not possibly have neutralized the relatively much larger quantities of ingested Ag, inasmuch as the much higher quantities luminal IgA in mice that had ingested this mAb had no inhibitory effect (Fig. 2B). TNP-specific IgA mAb most likely suppressed anaphylaxis induced by TNP-OVA in mice primed with IgE anti-TNP mAb by blocking TNP epitopes before TNP-OVA could bind to and crosslink mast cell-bound IgE anti-TNP rather than by activating inhibitory receptors, as demonstrated by suppression of the MMCP-1 response in our model (Fig. 4) and by equivalent suppression in wild-type, Fcµ/αR-deficient and FcγRIIb-deficient mice (Fig. 6 and data not shown). We and others have reported on similar epitope masking by non-inflammatory IgG isotypes in other systems 13, 39.
If IgA Ab indeed protects against anaphylaxis by blocking Ag binding to mast cell-associated IgE, it might seem more effective for Ab to bind ingested Ag before it is absorbed systemically, when the distance between Ag and mast cell is relatively great, rather than neutralize Ag only after it has been absorbed, when it is much closer to mast cells. This perspective, however, ignores the difference between the relatively high concentration of ingested Ag in the gut lumen and the much lower systemic concentration of intact ingested Ag; a result of proteolytic digestion of most ingested Ag. Thus, although a small percentage of ingested Ag is absorbed intact with sufficient rapidity to account for the kinetics of the systemic anaphylaxis (Fig. 1), the ratio of specific Ab to Ag is much higher systemically than it is in the gut lumen. This should greatly increase the effectiveness of systemic, as opposed to enteric, blocking of critical Ag epitopes by specific Abs. While it remains possible that intestinal Ab could sufficiently neutralize Ag in the gut lumen prior to systemic Ag absorption and block systemic anaphylaxis by this mechanism, the concentrations of specific Ab in the gut would have to be extraordinarily high, as we find that an ~2:1 molar ratio of IgA mAb to Ag has no detectable effect (Fig. 2B).
The observation that IgA blocks systemic anaphylaxis predominantly by binding to allergen that has been systemically absorbed indicates that allergen must be absorbed systemically to induce systemic anaphylaxis, as opposed to inducing systemic disease by triggering mast cells at the interface between the villus and the gut lumen. If Ag could induce systemic anaphylaxis by interacting solely with mast cells at the luminal interface, without being absorbed systemically, anaphylaxis would not have been inhibited by mAb restricted to systemic tissues. Our demonstration that this is true even in IL-9 transgenic mice (Fig. 5) and in mice actively immunized with OVA (Fig. 7), which both have high numbers of intestinal mast cells and increased intestinal permeability, suggests that this will also be true for individuals who have intestinal atopic disease, in whom similar changes develop 40–43. Most importantly, our conclusion that ingested Ags must be absorbed systemically to induce systemic anaphylaxis suggests that immunization procedures that induce systemic Ab responses should have some capacity to protect against food allergen-induced anaphylaxis and that immunization protocols that induce an IgG response should be as effective as those that induce IgA production at inhibiting systemic responses to food allergens.
This work was supported by a Merit Award from the U.S. Department of Veterans Affairs (FDF), RO1 AI072040 from the National Institutes of Health (FDF), and P30DK078392 from NIH and Cincinnati Children’s Digestive Disease Core (RTS)
We thank Dr. Dennis Metzger for providing us with J chain-deficient mice.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.