Pathways of immediate hypothermia and leukocyte infiltration in an adjuvant-free mouse model of anaphylaxis

Abstract

Background

Conflicting results have been obtained regarding roles of Fc receptors and effector cells in models of active systemic anaphylaxis (ASA). In part, this might reflect the choice of adjuvant used during sensitization, as various adjuvants might differentially influence the production of particular antibody isotypes.

Objective

We developed an ‘adjuvant-free’ mouse model of ASA and assessed the contributions of components of the ‘classical’ and ‘alternative’ pathways in this model.

Methods

Mice were sensitized intra-peritoneally (i.p.) with ovalbumin at weekly intervals for 6 weeks and challenged i.p. with ovalbumin two weeks later.

Results

Wild-type animals developed immediate hypothermia and late-phase intra-peritoneal inflammation in this model. These features were reduced in mice lacking the IgE receptor FcεRI, the IgG receptor FcγRIII or the common γ-chain FcRγ. FcγRIV blockade resulted in a partial reduction of inflammation without any effect on hypothermia. Depletion of monocytes/macrophages with clodronate liposomes significantly reduced the hypothermia response. By contrast, depletion of neutrophils or basophils had no significant effects in this ASA model. Both the hypothermia and inflammation were dependent on platelet-activating factor (PAF) and histamine and were reduced in two types of mast cell (MC)-deficient mice. Finally, engraftment of MC-deficient mice with bone marrow-derived cultured MCs significantly exacerbated the hypothermia response, and restored inflammation to levels similar to those observed in wild-type mice.

Conclusion

Components of the classical and alternative pathways contribute to anaphylaxis in this adjuvant-free model, with key roles for mast cells and monocytes/macrophages.

Capsule summary

FcεRI, FcγRIII, mast cells, histamine, and PAF are required for full development of hypothermia and numbers of infiltrating leukocytes in an adjuvant-free ASA model. Monocytes/macrophages also contribute to hypothermia in this model.

Introduction

Anaphylaxis is an acute, life-threatening systemic allergic reaction with a lifetime prevalence of 0.05% to 2.0% in developed countries^1–3. In humans, it is largely accepted that anaphylaxis can be triggered by histamine and other mediators released in response to antigen cross-linking of IgE bound to its high-affinity receptor, FcεRI, on mast cells (MCs)^1–3. However, IgG might also contribute to anaphylaxis in some patients, especially those treated with infused drugs such as therapeutic monoclonal antibodies^{4, 5}.

Several mouse models of anaphylaxis have been developed and used to assess the contributions of IgE and IgG antibodies, and the roles of various effector cells and mediators. The analysis of passive local or systemic anaphylaxis (PSA) models has allowed identification of two major pathways of anaphylaxis in mice: a ‘classical’ pathway consisting of IgE, FcεRI, MCs and histamine^6–8; and an ‘alternative’ pathway consisting of IgG, FcγRIII and/or FcγRIV^{5, 9–11}, platelet-activating factor (PAF)^{12, 13} and, depending on the exact model used, either macrophages^{5, 14}, basophils^{5, 13} and/or neutrophils^{5, 11, 15}.

Active systemic anaphylaxis (ASA) models, which are arguably more reflective of the clinical situation, have generated more conflicting results. Some ASA models critically depend on IgE, FcεRI and MCs^16–19, while some others can develop, at least with respect to the features analyzed, in IgE^−/−, FcεRI^−/− and/or MC-deficient mice^{9, 13, 20}. Depending on the mouse strain and ASA model used, basophils have been shown either to contribute to the systemic response^{13, 19} or to play little to no significant role^{17, 21}. Similarly, depletion of monocytes/macrophages or antibody-mediated neutrophil depletion reduces anaphylaxis in some but not all ASA models^{11, 17, 19, 21}.

We hypothesize that such conflicting results in ASA models might reflect, at least in part, the choice of adjuvant used during sensitization, as various adjuvants might differentially influence the production of individual antibody isotypes. Moreover, in humans who develop anaphylactic reactivity, sensitization to antigen generally occurs in the absence of an artificial adjuvant. We therefore developed an ‘adjuvant-free’ mouse model of ASA and assessed the contributions of components of the ‘classical’ and ‘alternative’ pathways of anaphylaxis in this model.

Methods

Mice

C57BL/6J (WT) mice were purchased from Jackson Laboratories (Bar Harbor, Me) or Charles River (France). FcγRIII^−/− mice (B6.129P2-Fcgr3^tm1Sjv/J mice, backcrossed to C57BL/6 for 12 generations) were purchased from Jackson Laboratories. FcRγ^−/− mice (B6.129P2 Fcer1g^tm1Rav mice backcrossed to C57BL/6 for 12 generations) were from Taconic (New York, NY). C57BL/6-Kit^W-sh/W-sh mice were originally provided by Peter Besmer (Memorial Sloan-Kettering Cancer Center, NY, USA); we then backcrossed these mice to C57BL/6J mice for more than 11 generations²². FcεRI^−/− mice⁷ (FcεRI alpha chain-deficient mice backcrossed to C57BL/6 for more than 8 generations and kindly provided by Jean-Pierre Kinet, Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, MA), Mcpt8^DTR mice (backcrossed to C57BL/6 for at least 10 generations)²³, and Cpa3-Cre; Mcl-1^fl/fl mice (backcrossed to C57BL/6 for at least 9 generations)²⁴ were bred and maintained at the Stanford University Research Animal Facility. Aged-matched male mice were used in all experiments. Experiments in Fig 2 used C57BL/6J WT mice as controls. We used littermate controls in all other experiments. All animal care and experimentation was conducted in compliance with the guidelines of the National Institutes of Health and with the specific approval of the Institutional Animal Care and Use Committee of Stanford University and of the Animal Ethics committee CETEA (Institut Pasteur, Paris, France) registered under #C2EA-89.

Figure 2. Roles of Fc receptors in OVA-induced ASA.

(A–B) OVA-induced hypothermia in OVA-sensitized WT, FcεRI^−/−, FcγRIII^−/− and FcRγ^−/− mice (A), or WT mice treated with an anti-FcγRIV antibody or an isotype control (B). (C–H) Numbers of leukocytes in the PLF 3 days after challenge. Data are pooled from two (‘Isotype’ and ‘α-FcγRIV’; n=7/group) or three (all other groups; n=9–15/group) independent experiments. *, ** or *** = P < 0.05, 0.01 or 0.001.

OVA-induced adjuvant-free model of active anaphylaxis

Six to 8-week-old mice were sensitized intraperitoneally (i.p.) with 10 µg of endotoxin-free ovalbumin (Endograde OVA; BioVendor; < 0.01 EU endotoxin per injection) in 100 µL of PBS once a week for 6 weeks. Two weeks after the last i.p. sensitization, mice were challenged i.p. with 500 µg of OVA. Rectal measurements of body temperature were performed immediately before (time 0) and at different time points for up to 2 h after challenge. Mice were sacrificed at various time points after challenge (as indicated) for assessment of inflammatory cell numbers in the peritoneal cavity and histology.

Other methods

Please see this article’s Online Repository at www.jacionline.org for the methods for flow cytometry, peritoneal lavage and differential cell counts, depletion of basophils, monocytes/macrophages and neutrophils, blockade of FcγRIV, histologic analysis, measurement of serum OVA-specific IgG₁ and IgG_2c antibodies, IgE-mediated PSA, ASA with adjuvant, treatment with an H₁ anti-histamine, a PAF receptor antagonist, quantification of histamine and PAF-AH, and generation and adoptive transfer of bone marrow-derived cultured mast cells (BMCMCs).

Statistical analyses

Results represent mean ± SEM or mean + SEM, with values for individual mice represented for quantifications of histamine, PAF-AH and leukocytes. We used an unpaired Student’s t test (body temperature) or an unpaired Mann-Whitney U test (all other data) to assess the significance of differences between two sets of data. P values < 0.05 are considered statistically significant.

Results

Development of an ‘adjuvant-free’ model of ASA in C57BL/6 mice

We developed an ‘adjuvant-free’ mouse model of ASA consisting of performing i.p. sensitizations with endotoxin-free OVA once a week for 6 weeks, and i.p. challenge with OVA two weeks after the last sensitization (Fig 1, A). OVA-sensitized C57BL/6 mice developed hypothermia following OVA challenge that was maximal at 30 min and decreased thereafter (Fig 1, B). Sensitized and challenged mice also developed a ‘late-phase’ inflammatory response with increased numbers of total cells, eosinophils, macrophages, and lymphocytes in the peritoneal cavity, with highest numbers of leukocytes on day 3 after challenge (Fig 1, C-H). Consistent with the increase in eosinophil numbers, we detected significant amounts of IL-5 in the plasma of 3 out of 5 OVA-sensitized mice 18 h (but not 72 h) after challenge, while levels of IL-4, IL-6, IL-10, TNF-α and IFN-γ were all below the detection limit of standard ELISAs at both time points (data not shown).

Figure 1. Development of hypothermia and intra-peritoneal inflammation in mice sensitized and challenged with OVA in the absence of adjuvant.

(A) Experimental outline. (B) Changes in body temperature following challenge with OVA in OVA-sensitized (‘OVA/OVA’) or PBS-treated (‘PBS/OVA’) mice. (C–H) Numbers of leukocytes in the peritoneal lavage fluid (PLF) at the indicated time points. Data are pooled from two or three independent experiments (n=6–10/group). *, ** or *** = P < 0.05, 0.01 or 0.001.

OVA-induced hypothermia and inflammation depend on the high-affinity IgE receptor FcεRI, the IgG receptor FcγRIII and their common activating subunit FcRγ

OVA sensitization induced significant elevation of OVA-specific IgG₁ and IgG_2c antibodies in C57BL/6 mice (see Fig E1, A & B in the Online Repository). We did not detect significant levels of OVA-specific IgE in sera by standard ELISA (data not shown). However, we observed degranulation of peritoneal cell-derived MCs (PCMCs) incubated in vitro with serum from OVA-sensitized mice followed by stimulation with OVA (see Fig E1, C & D in the Online Repository). Such degranulation was not observed with PCMCs incubated with serum from PBS-treated mice, or with serum from OVA-sensitized mice which had been pre-incubated with anti-IgE antibodies (see Fig E1, C & D in the Online Repository), demonstrating the presence of functional OVA-specific IgE in the serum of sensitized mice.

We then assessed responses of C57BL/6 mice lacking the IgG receptor FcγRIII, the high-affinity IgE receptor FcεRI, or their common activating subunit FcRγ. OVA-induced hypothermia was abolished in FcRγ^−/− mice and FcγRIII^−/− mice, and partially reduced in FcεRI^−/− mice, as compared to WT mice (Fig 2, A). Antibody-mediated blockade of FcγRIV had no effect on immediate hypothermia (Fig 2, B). WT, FcεRI^−/− and FcγRIII^−/− mice had similar levels of intra-peritoneal leukocytes at baseline (except for a small increase in macrophages and lymphocytes in FcγRIII^−/− mice) (see Fig E2 in the Online Repository). However, we observed decreased numbers of total cells, macrophages, eosinophils, and lymphocytes in the peritoneal cavity of FcεRI^−/− and FcγRIII^−/− mice, as compared to WT mice 3 days after challenge (Fig 2, C-G). These numbers were further decreased in mice lacking the common FcRγ (Fig 2, C-G). Antibody-mediated blockade of FcγRIV induced only a moderate decrease of peritoneal leukocyte numbers at day 3, with reduced numbers of total cells and eosinophils as compared with mice treated with an isotype control antibody (Fig 2, C-G).

Monocyte/Macrophage depletion with clodronate liposomes decreases immediate hypothermia and enhances late-phase intra-peritoneal inflammation

We injected OVA-sensitized mice with clodronate liposomes (or PBS liposomes, as a control) 24 h before challenge with OVA to deplete monocytes/macrophages and assess their potential contribution to anaphylaxis via the alternative pathway. Clodronate liposome-treated mice exhibited depletion of circulating CD11b^high monocytes at the time of OVA challenge, with no effect on blood Ly6G⁺ neutrophils (see Fig E3, A, C & D in the Online Repository). OVA challenge led to a decrease in CD11b expression in blood monocytes in PBS liposome-treated mice at day 3 (see Fig E3, A in the Online Repository). Consistent with a previous report²⁵, we found similar percentages of blood monocytes in both PBS liposome- and clodronate liposome-treated mice at day 3 (i.e. 4 days after the injection of liposomes) (see Fig E3, A, D & E in the Online Repository). However, F4/80⁺ macrophages were depleted in the spleen of clodronate liposome-treated mice at this time point, with no effect on neutrophils (see Fig E3, B, F & G in the Online Repository). Confirming the efficiency of depletion, we also found that levels of peritoneal macrophages were reduced by 73% in clodronate liposome-treated mice as compared to control mice at day 3 (Fig 3, E).

Figure 3. Assessment of the effect of monocyte/macrophage depletion on OVA-induced ASA.

Mice were treated with clodronate liposomes (‘Clodronate lipo.’) or PBS liposomes (‘PBS lipo.’) 24 h before challenge. (A) OVA-induced hypothermia. (B) Levels of histamine in the plasma 20 min after challenge. (C) PAF-AH activity in the spleen 20 min after challenge. (D–I) Numbers of leukocytes in the PLF 3 days after challenge in non-sensitized mice (‘-/OVA’) or OVA-sensitized mice (‘OVA/OVA’). Data are pooled from three independent experiments for all OVA-sensitized groups (total n=8–9/group) and from one experiment for non-sensitized controls (n=4–5/group). *, ** or *** = P < 0.05, 0.01 or 0.001.

In line with previous reports^{17, 26}, treatment with clodronate liposomes reduced features of anaphylaxis in an ASA model using sensitization with OVA together with the adjuvant alum and Bordetella pertussis toxin (see Fig E4, B & C in the Online Repository). Treatment with clodronate liposomes also significantly diminished the hypothermia response in OVA-sensitized and challenged mice in the adjuvant-free ASA model (Fig 3, A). Clodronate liposome-treated mice had lower levels of histamine in the plasma 20 min after challenge, as compared to PBS liposomes-treated mice (Fig 3, B). Since such a reduction in histamine levels could reflect a toxic effect of the liposomes on MCs, we also assessed responses of mice treated with clodronate or PBS liposomes in a MC-dependent IgE-mediated PSA model. Hypothermia was slightly but not significantly reduced (except at 10 min after challenge) in clodronate liposomes-treated mice, which suggests that the clodronate liposomes have only modest effects on MC activity in this model (see Fig E3, H in the Online Repository). The activity of PAF acetylhydrolase (PAF-AH, an enzyme which degrades PAF) has been shown to inversely correlate with the severity of anaphylaxis in human²⁷. However, we found similar PAF-AH activity in the spleen of mice treated with clodronate or PBS liposomes 20 min after challenge (Fig 3, C). By contrast, clodronate liposome-treated mice displayed significantly higher levels of eosinophils, lymphocytes and MCs than PBS liposome-treated mice in the peritoneal cavity 3 days after challenge (Fig. 3, F-I). Levels of blood neutrophils were also higher in such clodronate liposome-treated, macrophage-depleted mice as compared to control mice at day 3 (see Fig E3, A & C in the Online Repository). However, we also found a small but significant increase in lymphocyte and MC numbers 3 days after OVA challenge in naïve mice treated with clodronate liposomes, which could reflect some pro-inflammatory effects of the clodronate liposomes (Fig 3, D-I).

Neutrophils are not required for OVA-induced hypothermia and inflammation

We injected OVA-sensitized mice with an anti-Gr-1 antibody (or an isotype control antibody) 40 h before and 24 h after challenge with OVA to deplete neutrophils. Such treatment lead to complete ablation of circulating neutrophils at the time of challenge (see Fig E5, A & C in the Online Repository), and neutrophils remained absent in the blood and spleen of anti-Gr-1-treated mice 3 days after challenge (see Fig E5, A-C & F in the Online Repository). By contrast, treatment with anti-Gr-1 antibodies did not deplete circulating monocytes or spleen macrophages (and the percentages of these cells were even slightly higher than those in mice treated with the isotype control antibody) (see Fig E5, A, B, D, E & G in the Online Repository).

In agreement with findings obtained in IgG-mediated PSA models or ASA models using adjuvants for the sensitization^{5, 11}, we found that anti-Gr-1-treated mice developed markedly reduced features of anaphylaxis as compared to isotype control-treated mice in an ASA model using sensitization with OVA together with the adjuvants alum and Bordetella pertussis toxin (see Fig E4, D & E in the Online Repository). By contrast, isotype control-treated mice and anti-Gr-1-treated mice developed similar levels of immediate hypothermia following challenge with OVA (Fig 4, A). Both groups also had similar levels of leukocytes in the peritoneal cavity 3 days after challenge (although eosinophil and lymphocyte numbers were slightly decreased in the anti-Gr-1-treated group as compared to isotype control-treated mice, this difference did not reach statistical significance) (Fig 4, B-G).

Figure 4. Assessment of antibody-mediated depletion of neutrophils on OVA-induced ASA.

OVA-sensitized mice were treated with an anti-Gr-1 neutrophil-depleting antibody (‘α-Gr-1’) or an isotype control (‘Isotype’) 40 h before and 24 h after challenge with OVA. (A) OVA-induced hypothermia (B–G) Numbers of leukocytes in the PLF 3 days after challenge. Data are pooled from three independent experiments (total n=9/group).

Cpa3-Cre⁺; Mcl-1^fl/fl mice develop less OVA-induced hypothermia and inflammation

We next assessed responses of MC-deficient, basophil-depleted Cpa3-Cre⁺; Mcl-1^fl/fl mice²⁴ in this ‘adjuvant-free’ ASA model. Cpa3-Cre⁺; Mcl-1^fl/fl mice remained highly deficient in MCs and had reduced numbers of basophils even after repeated sensitizations and challenge with OVA (Fig 5, D-I). These mice developed similar levels of OVA-specific IgG₁ and IgG_2c antibodies as compared to Cpa3-Cre⁺; Mcl-1^+/+ littermate control mice (see Fig E6, A & B in the Online Repository), and displayed similar amounts of antigen specific-IgE activity in the serum (see Fig E6 C & D in the Online Repository), suggesting that MCs and basophils do not contribute substantially to the sensitization phase in this model. However, sensitized Cpa3-Cre⁺; Mcl-1^fl/fl mice developed significantly reduced levels of hypothermia following challenge with OVA, suggesting an important role for MCs and/or basophils in this feature of the model (Fig 5, A). Cpa3-Cre⁺; Mcl-1^fl/fl mice had markedly reduced levels of histamine in the plasma but similar PAF-AH activity in the spleen 20 min after challenge (Fig 5, B & C). Cpa3-Cre⁺; Mcl-1^fl/fl mice also had diminished intra-peritoneal inflammation, with decreased numbers of total leukocytes, macrophages, eosinophils, and lymphocytes, as compared to Cpa3-Cre⁺; Mcl-1^+/+ mice, indicating that MCs and/or basophils can also importantly contribute to the late phase leukocyte numbers in this model (Fig 5, J-N).

Figure 5. Assessment of OVA-induced ASA in genetically MC-deficient and basophil-depleted Cpa3-Cre; Mcl-1^fl/fl mice.

(A) OVA-induced hypothermia in OVA-sensitized Cpa3-Cre⁺; Mcl-1^+/+ and Cpa3-Cre⁺; Mcl-1^fl/fl mice. (B) Levels of histamine in the plasma 20 min after challenge. (C) PAF-AH activity in the spleen 20 min after challenge. (D–E) Representative FACS profile (D) and percentage (E) of blood basophils (CD49b⁺; IgE⁺) 24 h before challenge. (F) Numbers of MCs in the PLF 3 days after challenge. (G–I) Toluidine blue staining for MCs (G) and MC numbers (H, I) in sections of back skin and ear pinna. (J-N) Numbers of leukocytes in the PLF 3 days after challenge. Data are pooled from three independent experiments (total n=9–14/group). *, ** or *** = P < 0.05, 0.01 or 0.001. The crossbones symbol indicates death of one mouse. Scale bars: 100 µm.

Basophils are not required for OVA-induced hypothermia and inflammation

We assessed the potential role of basophils using Mcpt8^DTR mice, which express the diphtheria toxin (DT) receptor in basophils only, and in which basophils can be selectively ablated by injection of DT²³. We first confirmed that treatment with DT induces ablation of basophils, but not MCs, in OVA-sensitized Mcpt8^DTR mice (Fig 6, B-D). We ruled out a significant role for basophils in this model by showing that DT-mediated depletion of basophils in OVA-sensitized Mcpt8^DTR mice does not affect OVA-induced hypothermia (Fig 6, A) or late phase intra-peritoneal leukocyte numbers (Fig 6, E-I).

Figure 6. Assessment of the effect of diphtheria toxin-mediated basophil depletion on OVA-induced ASA in Mcpt8^DTR mice.

OVA-sensitized Mcpt8^+/+ and Mcpt8^DTR mice were treated with diphtheria toxin 48 h before and 24 h after challenge with OVA. (A) OVA-induced hypothermia. (B-C) Representative FACS profile (B) and percentage (C) of blood basophils (CD49b⁺; IgE⁺) 2 h before challenge. (D–I) Numbers of leukocytes in the PLF 3 days after challenge. Data are pooled from two independent experiments (total n=7–9/group). ** = P < 0.01. Each crossbones symbol indicates death of one mouse.

MCs exacerbate OVA-induced hypothermia and inflammation

We used Kit mutant MC-deficient Kit^W-sh/W-sh mice to assess further the contributions of MCs in this ‘adjuvant-free’ ASA model. Since Kit^W-sh-W-sh mice have many KIT related phenotypic abnormalities beside their MC deficiency ^28–33 we also included a group of Kit^W-sh/W-sh mice that had been engrafted with bone marrow-derived cultured MCs (BMCMCs➔Kit^W-sh/W-sh mice) both i.p. and i.v. 12 weeks before the first sensitization with OVA.

We found no significant difference in the severity of anaphylaxis between Kit^+/+ mice and MC-deficient Kit ^W-sh/W-sh mice in an ASA model using sensitization with OVA together with the adjuvants alum and Bordetella pertussis toxin (see Fig E4, F & G in the Online Repository), confirming previous findings obtained using MC-deficient Kit^W/W-v mice⁹. By contrast, MC-deficient Kit^W-sh/W-sh mice developed significantly less hypothermia following challenge with OVA as compared to Kit^+/+ mice (Fig 7, A). BMCMCs➔Kit^W-sh/W-sh mice developed significantly more hypothermia than did Kit^W-sh/W-sh mice following challenge with OVA, although levels of hypothermia in these mice did not reach those of Kit^+/+ mice (Fig 7, A). The intermediate body temperature response in MC-engrafted Kit^W-sh/W-sh mice compared with WT or MC-deficient Kit^W-sh/W-sh mice is consistent with our previously reported data in a peanut-induced ASA model¹⁹. We think it very likely that the technical limitations of such systemic MC engraftment experiments contributed to the intermediate temperature response seen in MC-engrafted Kit^W-sh/W-sh mice. Specifically, compared to the corresponding WT mice, MC-engrafted Kit^W-sh/W-sh mice had similar levels of MCs in the peritoneal cavity and mesenteric windows, and even higher numbers of MCs in the spleen. However, these mice had no detectable MCs in the skin, thus eliminating the numerically large population of skin MCs as a potential source of mediators in such mice (Fig 7, and see Fig E7 in the Online Repository).

Figure 7. Responses of WT mice, MC-deficient Kit^W-sh/W-sh mice, and Kit^W-sh/W-sh mice engrafted with bone marrow-derived cultured MCs in OVA-induced ASA.

(A) OVA-induced hypothermia in Kit^+/+ mice, MC-deficient Kit^W-sh/W-sh mice and Kit^W-sh/W-sh mice engrafted with WT BMCMCs. (B-G) Numbers of leukocytes in the PLF 3 days after challenge. Data are pooled from three to five independent experiments (total n=12–25/group). *, ** or *** = P < 0.05, 0.01 or 0.001 vs. Kit^+/+ group (A) or indicated group (B-G).^# or = ^## = P < 0.05 or 0.01 vs. BMCMCs➔Kit^W-sh/W-sh group (A). The crossbones symbol indicates death of one mouse.

Finally, MC-deficient Kit^W-sh/W-sh mice developed significantly diminished numbers of leukocytes in the late phase intra-peritoneal inflammation, as compared to Kit^+/+ mice (Fig 7, B-F). Confirming the role of MCs in this feature of our ASA model, engraftment of Kit^W-sh/W-sh mice with BMCMCs restored levels of total peritoneal leukocytes, neutrophils, eosinophils, macrophages and lymphocytes quantified 3 days after challenge to those observed in Kit^+/+ mice (Fig 7, B-F). Altogether, these results show that MCs can amplify both the immediate hypothermia and the late phase inflammatory reaction in Kit^W-sh/W-sh mice in this ‘adjuvant-free’ ASA model.

Histamine and PAF contribute to OVA-induced hypothermia and inflammation

Pre-treatment of mice with the platelet-activating factor (PAF) receptor antagonist, CV-6209, significantly decreased both the immediate hypothermia and the numbers of leukocytes in late phase inflammatory response (Fig 8). Pre-treatment with the H₁ anti-histamine, triprolidine, slightly reduced the immediate hypothermia and numbers of leukocytes in the late phase response (Fig 8), although differences did not reach statistical significance except for eosinophil numbers (Fig 8, D). However, combined treatment with the anti-histamine and PAF receptor antagonist almost entirely blocked both the hypothermia and the increased numbers of leukocytes observed in the inflammatory responses (Fig 8). By contrast, treatment with triprolidine markedly reduced both the hypothermia and numbers of intra-peritoneal leukocytes in mice pre-treated with clodronate liposomes, while CV-6209 had no significant effects on these features (see Fig E8 in the Online Repository). Collectively, these data demonstrate the involvement of both histamine and PAF in this ASA model in WT mice, and suggest that monocytes/macrophages represent the main source of PAF in this model.

Figure 8. Roles of histamine and PAF in OVA-induced ASA.

OVA-sensitized mice were treated with the H₁ anti-histamine triprolidine or with the PAF receptor antagonist CV-6209 alone or in combination 30 min before and 1 day after challenge with OVA. Control mice were injected with vehicle (saline) only. (A) OVA-induced hypothermia. (B-G) Numbers of leukocytes in the PLF 3 days after OVA challenge. Data are pooled from two independent experiments (total n=8/group). *, ** or *** = P < 0.05, 0.01 or 0.001 vs. vehicle group.

Discussion

Most ASA models employ adjuvants during the sensitization phase, and such methods typically prime the animals to exhibit allergic reactions that require little or no contributions from MCs and IgE. We hypothesized that the use of adjuvants might boost the production of certain IgG isotypes, favoring the activation of ‘alternative’ pathways of anaphylaxis, which might mask contributions from the ‘classical’ IgE- and MC-dependent pathway. We therefore designed a new ‘adjuvant-free’ model of ASA, and assessed the potential roles of various components of the ‘classical’ and ‘alternative’ pathways in that model.

Our sensitization protocol resulted in production of OVA-specific IgG₁ and IgG_2c antibodies (see Fig E1 & E4 in the Online Repository), two isotypes that can induce anaphylaxis through FcγRIII and/or FcγRIV^{5, 10, 11}. Indeed, we found that FcγRIII^−/− mice developed markedly reduced hypothermia and reduced numbers of peritoneal leukocytes as compared to WT mice in this model (Fig 2). We used an anti-FcγRIV blocking antibody to assess the contribution of this receptor, and found no role for FcγRIV in the hypothermia response, and a relatively minor contribution to numbers of intra-peritoneal leukocytes (Fig 2). Altogether, our results confirm the important involvement of FcγRIII in allergic shock^{10, 11, 17, 26, 34} and reveal that this receptor also has an important role in the development of antigen-induced allergic inflammation.

Although we did not detect significant levels of OVA-specific IgE in the sera of most mice by classical ELISA (data not shown), we demonstrated the presence of functionally active specific IgE in serum from OVA-sensitized mice using an ex vivo MC activation test (see Fig E1 & E6 in the Online Repository). These levels of antigen-specific IgE were sufficient to contribute to anaphylaxis and allergic inflammation, since mice lacking the high-affinity IgE receptor FcεRI displayed significantly diminished hypothermia and numbers of intra-peritoneal leukocytes as compared to WT mice (Fig 2). These results are in agreement with previous findings showing less hypothermia in FcεRI^−/− or IgE^−/− mice in models of peanut-induced ASA^{16, 17, 35}. However, our results are in sharp contrast with reports showing no evidence for involvement of IgE or FcεRI in models of OVA-induced ASA which employ alum as an adjuvant during the sensitization phase^{9, 10, 20}. In line with the strong contributions of both FcγRIII and FcεRI in our model, we found that FcRγ^−/− mice, which lack the common activating subunit of Fcε and Fcγ receptors³⁶, were completely protected from hypothermia, and developed even lower numbers of intra-peritoneal leukocytes than FcεRI^−/− or FcγRIII^−/− mice (Fig 2).

In mice, FcγRIII is expressed on all myeloid cells³⁷ while FcεRI is mainly expressed on MCs and basophils³⁸. Depending on the model used, MCs, basophils, monocytes/macrophages and/or neutrophils have been reported to contribute to ASA^{11, 13, 14, 16–19, 21, 26, 35}. Several studies using passive or active models of anaphylaxis have reported an important role for monocytes/macrophages in the immediate hypothermia response following antigen challenge^{5, 14, 17, 21, 26, 39–41}. We confirmed these findings in an ASA model using the adjuvants alum and Bordetella pertussis toxin for the sensitization (see Fig E4 in the Online Repository). Our depletion experiments also demonstrated that monocytes/macrophages contributed to the hypothermia response in our adjuvant-free ASA model (Fig 2). By contrast, little is known about the function of monocytes/macrophages in the late phase allergic inflammation following anaphylaxis. We found that mice treated with clodronate liposomes to deplete monocytes/macrophages exhibited significantly increased numbers of intra-peritoneal leukocytes (Fig 2). This raises the possibility that monocytes/macrophages might contribute to the resolution phase in this ASA model. However, care should be taken in the interpretation of these findings, since such increased leukocyte numbers might also reflect, at least in part, pro-inflammatory effects of the clodronate liposomes. Indeed, we also found slightly increased numbers of intra-peritoneal MCs and lymphocytes in non-sensitized mice treated with clodronate liposomes.

The importance of neutrophils and basophils in anaphylaxis is debated¹⁵. Some reports indicate that antibody-mediated depletion of neutrophils can reduce IgG₂-mediated passive systemic anaphylaxis (PSA)^{5, 11} and ASA¹¹ in mice, while others found no role for neutrophils in the hypothermia reaction following antigen challenge in IgG-mediated PSA^{13, 41} or ASA^{19, 21} models. In the present study, we used anti-Gr-1 antibodies to deplete neutrophils. We found that such treatment markedly reduced anaphylaxis in an ASA model using the adjuvants alum and Bordetella pertussis toxin for the sensitization (see Fig E4 in the Online Repository). By contrast, we found no significant contribution for neutrophils in either the immediate hypothermia reaction or the late phase accumulation of intra-peritoneal leukocytes in the adjuvant-free ASA model (Fig 4). Similarly, some reports indicate a contribution of basophils to IgG-mediated PSA^{5, 13, 34} or ASA^{11, 17, 19}, while others found no significant role for basophils in anaphylaxis models^{21, 41, 42}. Here, we used conditional basophil-deficient Mcpt8^DTR mice²³, and found no roles for basophils in either the immediate hypothermia reaction or the late phase allergic inflammation in our adjuvant-free ASA model (Fig 6).

It is largely agreed that MCs play important roles in food allergy and anaphylaxis^{14, 43–45}. While most of the literature on the roles of MCs in experimental anaphylaxis is based on data obtained using Kit mutant genetically MC-deficient mice, several new models have been developed in which the MC deficiency is not dependent on mutations affecting c-kit structure or expression^{24, 30–32, 46–49}. While discordant findings have been reported in some disease models in newer versus older MC-deficient strains^{30, 46, 47, 50, 51}, the importance of MCs to both IgE-mediated PSA^{24, 47, 52} and peanut-induced ASA¹⁹ has been confirmed using multiple MC-deficient mouse strains. In agreement with these findings, we demonstrate here that both Kit^W-sh/W-sh mice and Kit-independent Cpa3-Cre; Mcl-1^fl/fl MC-deficient mice developed reduced immediate hypothermia reactions and diminished numbers of late phase intra-peritoneal leukocytes in our adjuvant-free ASA model (Fig 5 & 7). We obtained additional evidence for an important contribution of MCs in this model by showing that engraftment of genetically MC-deficient Kit^W-sh/W-sh mice with bone marrow–derived cultured MCs partially restored the immediate hypothermia and completely restored the ‘late phase’ allergic inflammation induced by antigen challenge (Fig 7).

These results stand in marked contrast with previous reports^{9, 13} and our own data (see Fig E4 in the Online Repository) showing that anaphylaxis can fully develop following challenge with OVA in MC-deficient mice sensitized with OVA together with the adjuvant alum. One potential explanation for such results is that MCs are particularly potent at promoting allergic reactions at low levels of antibodies, and/or that increased levels of antibodies (promoted by the use of adjuvants during the sensitization phase) lead to a greater contribution of alternative pathways that mask or render redundant the role of MCs. Indeed, previous reports show that the contribution of MCs to some ASA models is greater when using low doses of adjuvant and/or antigen for the sensitization and low doses of antigen for the challenge^{17, 19}. Our results are also in line with previous reports demonstrating that genetically MC-deficient mice exhibit significantly diminished OVA-induced allergic airway inflammation when sensitized without adjuvant^53–56, but can fully develop airway inflammation when alum is employed as an adjuvant during the sensitization phase^{53, 57}.

Finally, consistent with previous reports, we found that both histamine and PAF contributed to the immediate hypothermia reaction in WT mice^{14, 20, 31, 58–60}. These two mediators also contributed to the late phase inflammatory reaction, with a more pronounced effect of PAF (Fig 8). By contrast, treatment with a PAF receptor antagonist had no effect in clodronate liposome-treated mice, suggesting that monocytes/macrophages represent the major source of PAF in this adjuvant-free ASA model. As expected, we found that plasma histamine levels were markedly reduced in MC- and basophil-deficient Cpa3-Cre⁺; Mcl-1^fl/fl mice. More surprisingly however, we also found reduced plasma histamine levels in clodronate liposome-treated mice (although the reduction was not as substantial as that found in Cpa3-Cre⁺; Mcl-1^fl/fl mice), suggesting that monocytes/macrophages might also directly or indirectly contribute to histamine release in this model.

In conclusion, we demonstrate here that FcεRI- and FcγRIII-dependent signaling, histamine, and PAF are required for the full development of hypothermia and intra-peritoneal leukocyte accumulation in an ‘adjuvant-free’ model of ASA. In this model, both MCs and monocytes/macrophages are critically involved in the immediate hypothermia reaction. In addition, MCs are also required for the full development of intra-peritoneal inflammation, as assessed as numbers of intra-peritoneal leukocytes. Our data thus strongly support the hypothesis that MCs and monocytes/macrophages are the main effector cells of anaphylaxis in this setting.

Key messages.

• The IgE receptor FcεRI, the IgG receptor FcγRIII, histamine, and PAF contribute to hypothermia and leukocyte numbers in an adjuvant-free active systemic anaphylaxis (ASA) model.

• Mast cells are required for full development of hypothermia and leukocyte numbers in this ASA model, with no significant role for basophils or neutrophils.

• Monocytes/macrophages contribute to hypothermia in this ASA model.

Abbreviations used

ASA

Active systemic anaphylaxis

BMCMC

Bone marrow-derived cultured mast cell

Cpa3

Carboxypeptidase A3

DT

Diphtheria toxin

DTR

Diphtheria toxin receptor

MC

Mast cell

Mcpt

Mast cell protease

OVA

Ovalbumin

PAF

Platelet-activating factor

PAF-AH

Platelet-activating factor acetylhydrolase

WT

Wild-type