|Home | About | Journals | Submit | Contact Us | Français|
Systemic anaphylaxis is generally recognized as a severe allergic reaction caused by IgE-mediated activation of mast cells, leading to massive release of vasoactive mediators that induce acute hypotension and shock. However, experimental evidence in mice suggests that this view is too simple. Using a variety of techniques to manipulate immune cell makeup, Jönsson et al. come to the conclusion in this issue of the JCI that recognition of IgG1 and IgG2 antibodies by FcγRIII and FcγRIV receptors on neutrophils is a major pathway for induction of anaphylaxis. These exciting results suggest that we have to reevaluate our models for anaphylaxis in humans, which will have a direct impact on our therapeutic approaches for prevention of this potential deadly hypersensitivity reaction.
I turned to Wikipedia when I was searching for a way to explain the basics of anaphylaxis and came across the following statement: “True anaphylaxis is caused by degranulation of mast cells or basophils mediated by immunoglobulin E (IgE).” This is the classic teaching, present in all the immunology textbooks, but the paper in this issue of the JCI by Jönsson et al. (1) informs us that this view is, at best, incomplete. Instead, we learn that anaphylaxis can be mediated by neutrophils recognizing IgG/antigen complexes. In addition to turning around our understanding of anaphylaxis, this paper adds to the growing list of neutrophil functions besides just bacterial killing and protease production (2). Lately we have learned that neutrophils are major sources of cytokines and chemokines (3, 4). They play a direct role in influencing the recruitment and activation of monocytes/macrophages, T cells, and NK cells during inflammation (5–7). Neutrophils have been implicated as the primary initiators of immune complex–mediated diseases (8, 9). And now the shocking news (pun intended!) that they are major players in initiating anaphylaxis.
Anaphylaxis is an acute, multisystem, severe type I hypersensitivity reaction that develops in minutes to hours following antigen exposure (10). In its mildest forms, it results in rashes (hives), wheezing, and some gastrointestinal symptoms (cramping, bloating). In its more severe forms, patients develop bronchoconstriction with hypoventilation, systemic vasodilation (leading to frank shock), cardiac dysrhythmias, and central nervous system abnormalities. Anaphylaxis most often occurs in patients with severe allergy to insect stings (specifically Hymenoptera venom), specific foods (mainly nuts, shellfish, some milk products), or in response to some medications. Amazingly, it is estimated that 1%–15% of people in the US are at risk for anaphylaxis-type reactions and that upwards of 1,500 deaths per year are attributed to acute hypersensitivity reactions (11, 12).
Experimentally, anaphylaxis is studied in two fashions. Active systemic anaphylaxis (ASA) is induced by immunizing experimental animals, then rechallenging them with the antigen in a form that induces an acute hypersensitivity response; this model closely mimics human anaphylaxis. Passive systemic anaphylaxis (PSA) involves adoptive transfer of antigen-specific Abs into naive animals followed by injection of the antigen. In both cases, use of knockout mice lacking various immune cells, receptors, or signaling molecules has allowed investigators to dissect the mechanisms of hypersensitivity reactions, as exemplified by Jönsson et al. (1).
The term anaphylaxis was coined by Charles Richet, a French physiologist, in his work with colleague Paul Portíer, trying to establish immunity in dogs to sea anemone toxin (13). Their objective was to make the animals tolerant to the toxin by first injecting them with nonlethal doses, followed by subsequent challenge doses. This protective effect of prior exposure had been demonstrated with other toxins in other animal species. To their dismay, Portíer and Richet found that their dogs developed lethal hypersensitivity reactions within minutes following a second injection of even small doses of the toxin. Portíer and Richet coined a new term, anaphylaxis, derived from the Greek words a (against) and phylaxis (protection), to describe what appears today to be a failed experiment. This work was published in 1902 (14), and Richet received the Nobel Prize in Medicine/Physiology in 1912. That is making lemonade out of lemons, indeed!
A century after Portíer and Richet’s work, our molecular understanding of anaphylaxis is that it is an IgE/mast cell/basophil-mediated event (reviewed in ref. 10). Sensitized individuals develop antigen-specific IgE, which binds to FcεR receptors on mast cells and basophils, priming the cells for robust responses. Following reexposure to the allergen, FcεRs are aggregated on the responding cells, leading to degranulation and release of “preformed” mediators, principally histamine, which in turn induces the symptoms of allergy, or in its most systemic form, anaphylaxis (Figure (Figure1).1). There is a wealth of experimental data supporting this model, including years of experiments in defining the “reagin” of allergy as the immunoglobulin IgE, then the cloning of the FcεR receptors and demonstrating their presence on mast cells, along with the recognition of mast cell products as having direct vascular effects that can lead to hypotension and shock, which characterizes severe anaphylaxis. It is also clear that many other mast cell products may contribute to allergy and anaphylaxis, including proteoglycans, serine proteases, lipid-derived mediators, cytokines, and platelet-activating factor (PAF) (10).
With more modern approaches in mouse models, it becomes clear that this simple paradigm is unlikely to be the entire story. Using the PSA model in mice, it is well established that administration of either IgE or IgG1 can cause reactions following antigen challenge, with IgE acting through FcεR and IgG1 acting through FcγRIII receptors (15). Using basophil-depleting mAbs and mast cell–deficient mice, it appears that the IgG1 responses are initiated primarily by FcγRIII receptors on basophils, which release large amounts of PAF (16). More importantly, it has long been recognized that ASA can be induced in mice lacking either mast cells or IgE, suggesting that other pathways must be operative (17, 18). ASA can also be induced in basophil-depleted animals (16), but fails to occur in mice lacking activating FcγRs (15). Moreover, blockade of the PAF receptor limits peanut-induced allergy in mouse models (19). These findings lead us to conclude that antibodies other than IgE can mediate ASA, that cells other than mast cells or basophils are involved, and that histamine is not the only mediator of anaphylaxis.
Jönsson et al. (1) address these issues using a panel of relatively new knockout mouse models as well as some novel blocking mAbs. Their basic model involves immunization of mice with BSA in Freund’s adjuvant or alum, followed by confirmation of IgE and IgG anti-BSA titers, then subsequent rechallenge with intravenously injected BSA. The animals develop a fatal ASA within minutes and can be monitored by a drop in core body temperature. Using this simple model, the authors demonstrate that ASA still develops in 5KO mice, which are derived by interbreeding of 5 different knockout strains and lack all IgE FcεRs and all IgG FcγRs except FcγRIV, which is present only on neutrophils and macrophages. Treatment of mice with anti-FcγRIV blocks the ASA in 5KO mice, and in combination with a new anti-FcγRIII–blocking mAbs, ASA is blocked in regular wild-type mice. Macrophage depletion does not protect mice in their ASA model, but neutrophil depletion (especially combined with loss of basophils) is fully protective. The type of adjuvant used in the immunization matters, with alum favoring production of IgG1 antibodies, which then results in loss of ASA in the 5KO strain. The authors conclude that BSA-specific IgG1 binding to FcγRIII on basophils (and likely other cells) and IgG2 binding to FcγRIV on neutrophils is sufficient to mediate ASA. Amazingly, the authors confirm the neutrophil dominance in their ASA model by demonstrating that adoptive transfer of human neutrophils into mice that lack all activating FcεRs and FcγRs restores normal reactions to BSA challenge. They confirm the relative specificity of IgG1 to FcγRIII versus IgG2 to FcγRIV in PSA-type experiments, using different anti-DNP mAbs. They also confirm the neutrophil dominance in PSA using a polyclonal Ab model, performed by injecting mice with preformed immune complexes. Finally, using inhibitors and cPLA2 mutant mice that fail to make PAF, they demonstrate that PAF, but not histamine, seems to be the dominant mediator in their ASA model.
Although this is an outstanding and well-controlled study, there are always caveats. There are minor issues, such as the need to test ASA and IgG2 PSA in newly developed FcγRIV KO mice (20), which will be quickly done. The studies involving PAF as a major mediator are somewhat limited at this point. There is also the general concern that all these experiments are done with substantial immunization protocols (such as Freund’s adjuvant) followed up by injection of large amounts of antigen, paradigms that do not recapitulate the development of lethal hypersensitivity in a child following a single bee sting. Moreover, the repertoire and expression pattern of FcγRs differs significantly between mouse and humans, so it remains generally unknown how translatable mouse models for anaphylaxis are to humans. Nevertheless, it is clear that physicians and scientists need to pay more attention to anti-specific IgGs and neutrophils in patients with hypersensitivity reactions. Therapeutically, this could translate into more effort put toward blocking IgG or FcγR binding or inhibiting PAF function in patients with severe allergy. Finally, one thing is for sure: this wonderful study by Jönsson et al. (1) will mean that somebody has to sit down and update Wikipedia; hopefully, by the time you have finished reading this paper, it will be done!
The author’s work is supported by the NIH grants AI65495 and AI68150.
Conflict of interest: The author has declared that no conflict of interest exists.
Citation for this article: J Clin Invest. 2011;121(4):1260–1263. doi:10.1172/JCI57296
See the related article beginning on page 1484.