Main

Multiple sclerosis (MS) and its animal model, experimental autoimmune encephalomyelitis (EAE), are immune-mediated diseases of the central nervous system in which T helper 1 (Th1) CD4+ cells reacting to myelin are considered to have a leading role.1 Therapeutic approaches aimed to shift the immune response against myelin antigens from Th1 to Th2 by i.v. injection(s) of soluble myelin peptides or proteins have been shown to induce antigen-specific immune tolerance and amelioration of EAE (reviewed in Fontoura et al2). However, such therapeutic approaches have substantial potential risk.3 Mice with EAE can develop fatal anaphylaxis on re-exposure to certain self-peptides of the myelin sheath.4, 5, 6, 7, 8 Repeated administration of self- or altered self-peptides resulted in immediate hypersensitivity reactions in a subset of patients with MS treated with NB5788, an altered peptide ligand (APL) of myelin basic protein,9 and in 9% of MS patients treated with glatiramer acetate, a random four-amino-acid non-self-polymer widely used for the treatment of this disease.10

Anaphylaxis in humans is thought to be mediated largely (if not solely) by a pathway consisting of IgE antibodies, the high-affinity receptor for IgE (FcɛRI), mast cells and basophils, and histamine is thought to represent a major mediator of such anaphylactic reactions (reviewed in Kawakami et al11). However, in mice, anaphylaxis can be elicited either by IgE- or by IgG1-dependent mechanisms.12 The second, IgG1-dependent, pathway is IgE-independent, but requires the low-affinity receptor for IgG (FcγRIII) and, depending on the setting, can involve basophils and perhaps macrophages, and platelet-activating factor (PAF).12, 13, 14, 15 In a chronic EAE model, induced in C57BL/6 mice by immunization with the self-myelin peptide oligodendrocyte glycoprotein (MOG) 35–55, we and others found that anaphylaxis to self was abrogated in mice lacking the γ-chain common to Fc receptors (FcR),5, 8 but not in mice lacking the α chain of FcγRIII (FcγRIII α chain−/−), suggesting that the IgE-FcɛRI-mast cell/basophil–histamine pathway might contribute to the development of anaphylaxis in this model.5, 8 However, except for the findings obtained in various FcR-knockout mice, the effector cells and molecules involved in anaphylaxis to this self-myelin peptide are not fully understood. A better understanding of the mechanisms contributing to such reactions may help in the development of strategies aimed at preventing and/or treating undesired allergic reactions associated with peptide therapies for autoimmune disorders. We therefore analyzed the importance of mast cells, histamine and IL-4, key elements of IgE-dependent anaphylaxis, in the development of anaphylaxis to self-MOG35–55 in a chronic model of EAE. We also attempted to prevent anaphylaxis to this self-peptide by using CV6209, an antagonist of PAF, a key mediator in the development of IgG1-dependent anaphylaxis. Here we show that anaphylaxis to self-MOG35–55 requires neither mast cells nor histamine, but it does require IL-4. Our results also suggest that PAF can contribute to the development of anaphylaxis to this self-peptide, but PAF probably is not the only mediator involved.

MATERIALS AND METHODS

Mice

Eight- to twelve-week-old mast cell-deficient WBB6F1-KitW/Wv (W/Wv) and congenic wild-type (WBB6F1-Kit+/+) female mice (Jackson Laboratory, ME, USA), C57BL/6J-IL-4−/− and wild-type C57BL/6J mice (Jackson Laboratory), histidine decarboxylase-deficient mice (HDC−/−) backcrossed six generations onto C57BL/6 (provided by Dr H Ohtsu, Sendai, Japan)16 and wild-type C57BL/6N (Charles-Rivers, Calco, Italy) were used in this study. Experiments were conducted in the animal facilities of the Neurological Institute Foundation Carlo Besta or of Stanford University. All procedures involving animals were approved by the ethical committee of the Neurological Institute Foundation Carlo Besta and the Division of Comparative Medicine at Stanford University, and carried out according to the Principles of Laboratory Animal Care (European Communities Council Directive 86/609/EEC) and the National Institutes of Health guidelines.

Peptides

MOG35–55 (MEVGWYRSPFSRVVHLYRNGK), control peptide-1 (random-CAVPLKTQMSGSSFM), and control peptide-2 (acetylcholine receptor, AchR, 97–116; DGDFAIVKFTKVLLDYTGHI) were synthesized and purified to >95% by analytical reverse-phase HPLC.

Immunization Protocol

EAE was induced as described earlier5 by subcutaneous immunization with MOG35–55 (100 μg per mouse) in Complete Freund's Adjuvant (CFA; Difco Laboratories, Detroit, MI, USA), containing 4 mg/ml of heat-killed Mycobacterium tuberculosis H37Ra (Difco Laboratories). All mice also received two intravenous injections of Bordetella pertussis (B. pertussis) toxin (200 ng/mouse) (List Biological Laboratories, Campbell, CA, USA) on day 0 and day 2 post immunization. The clinical features of EAE, which developed in all of the mice used in this study, have already been described.6, 17, 18, 19 In another model of MOG35–55-induced EAE, mice were immunized without using B. pertussis toxin as an adjuvant (see Results section for details).

Induction of Active Systemic Anaphylaxis

We evaluated whether mice exhibited active systemic anaphylaxis by challenging them 6 weeks after induction of EAE with an i.p. injection of MOG35–55 (100 μg) dissolved in PBS (100 μl). Control mice were challenged with an i.p. injection of a control peptide at the same dose. Mice were observed for 1 h after peptide challenge for clinical signs of anaphylaxis, and for each mouse body temperature was recorded with a rectal probe (Physitemp Instruments, Clifton, NJ, USA) at baseline and at 5, 10, 20, 30, and 60 min after challenge.5 Data are shown as mean±s.e.m. for all mice in each group. In the groups in which some mice died during the observation period of 1 h, mean values were determined on the basis of data for the surviving mice. Mice were considered to have anaphylaxis when suggestive clinical signs (ie, reddening of the skin, piloerection, prostration, reduced or lack of response to stimuli, and death) were accompanied by a decrease in body temperature of at least 1°C. In some experiments, mice were treated with an i.p. injection of the PAF antagonist CV6209 (Biomol International Inc., Plymouth Meeting, PA, USA), given at the dose of 100 μg 5 min before peptide challenge, and/or with the histamine receptor 1 (H1R) antagonist triprolidine (Sigma, St Louis, MO, USA), given i.p. at the dose of 200 μg 30 min before peptide challenge. Control mice were treated i.p. with PBS. Reagents were diluted in PBS to a final volume of 200 μl.

Measurement of Serum Ig Responses

Blood was collected from the tail 6 weeks after the induction of EAE and sera were stored at −20°C until analyzed. Peptide-specific IgG1, IgG2a, and IgE antibodies were measured by enzyme-linked immunosorbent assay (ELISA) as described.5, 6, 16 Briefly, 96-well microtiter plates (Immunol, Thermo Labsystems) were coated overnight at 4°C with MOG35–55 diluted in coating buffer (0.010 mg/ml). Plates were blocked with PBS 10% FCS for 2 h. Samples were diluted in blocking buffer at 1:100 for IgG1 and IgG2a, and at 1:25 for IgE, and antibody binding was tested by the addition of peroxidase-conjugated monoclonal goat anti-mouse IgG1 or IgG2a (Southern Biotechnology Associates, Birmingham, AL, USA). Enzyme substrate was added and plates were read at 450 nm on a micro plate reader. Total IgE was measured by sandwich ELISA (BD PharMingen) following the manufacturer's instructions.16

Statistical Analysis

Differences among groups in the time course of body temperature were examined by ANOVA. Differences among groups in the number of mice exhibiting systemic allergic reactions were analyzed by the Fisher's exact test. Two-tailed student's t-test was used to compare results between two groups. In all tests, P<0.05 was considered statistically significant.

RESULTS

Expression of Anaphylaxis to MOG35–55 in Mast Cells- or Histidine Decarboxylase-Deficient Mice

WBB6F1-KitW/Wv (W/Wv) mice, which virtually lack mast cells in all tissues, have been shown to be resistant to the development of IgE-dependent passive systemic anaphylaxis,12, 15, 20 but to be susceptible to passive IgG1-mediated or active anaphylaxis.12, 15 Challenge with MOG35–55 6 weeks after peptide immunization induced anaphylaxis in both W/Wv and congenic Kit+/+ wild-type mice (Figure 1a and Table 1), indicating that mast cells are not required for the expression of anaphylaxis in this model. Consistent with earlier observations,12, 19 serum titers of total IgE and, to a lesser extent, of peptide-specific IgG1 antibodies were higher in W/Wv vs Kit+/+ mice (Figure 1b). We failed to detect antigen-specific IgE with our ELISA method in sera of these mice, as well as in those of all of the other strains used in this study.

Figure 1
figure 1

Expression of anaphylaxis to self-MOG35–55 in mast cell-deficient mice. (a) Changes in body temperature in mast cell-deficient WBB6F1-KitW/Wv (W/Wv) and congenic WBB6F1-Kit+/+ (Kit+/+) mice on i.p. challenge with MOG35–55 or control peptide-1. †One mouse dead from anaphylactic shock at that time point. The exact numbers of challenged mice in the various groups are given in Table 1. *P<0.001. (b) MOG35–55-specific IgG1 and IgG2a, and total IgE antibody titers in sera of W/Wv and congenic Kit+/+ mice (n=6–8 mice per group). *P<0.005.

Full size image
Table 1 Anaphylaxis in mast cell-, histamine- or IL-4-deficient mice and their controls on challenge with MOG35–55
Full size table

HDC−/− mice, which are profoundly histamine-deficient in all tissues,16, 21 cannot exhibit hypothermia in association with IgE-dependent passive anaphylaxis.22 A paucity of mast cells and abnormalities in mast cell cytoplasmic granules have been described among the phenotypic abnormalities in HDC−/− mice,21 whereas no major differences in basophil number have been observed. We found that FcɛRI+, B220− cells (regarded as basophils) represented 2.9±0.5% of the total blood cells in HDC−/− mice and 2.8±0.7% in HDC+/+ mice (data not shown); similarly, Dr Elke Schneider—Necker Hospital, Paris—has found that CD49+, FcɛRI+ cells (regarded as basophils) represented from 0.6 to 1.2% of bone marrow cells in HDC−/− or HDC+/+ mice (Elke Schneider, personal communication). Challenge with MOG35–55 induced anaphylaxis in both HDC−/− and wild-type mice, interestingly with HDC−/− mice developing a greater drop in body temperature than wild-type mice (Figure 2a and Table 1). As we reported earlier,16 total serum IgE titers were significantly increased in HDC−/− vs wild-type mice, whereas no significant differences were observed in titers of peptide-specific IgG1 (Figure 2b).

Figure 2
figure 2

Expression of anaphylaxis to self-MOG35–55 in histamine-deficient mice. (a) Changes in body temperature in C57BL/6-HDC−/− (HDC−/−) and control C57BL/6 mice on i.p. challenge with MOG35–55 or control peptide-2. †Three and ††four mice dead from anaphylactic shock at that time point. The exact numbers of challenged mice in the various groups are given in Table 1. *P<0.05 and **P<0.001. (b) MOG35–55-specific IgG1 and IgG2a, and total IgE antibody titers in sera of HDC−/− and control C57BL/6 mice (n=8–11 mice per group). *P<0.05.

Full size image

The results obtained in mast cell-deficient W/Wv mice and histamine-deficient HDC−/− mice indicate that neither mast cells nor histamine, key players in IgE-dependent anaphylaxis,14 are necessary for the development of anaphylaxis to MOG35–55 in mice.

Expression of Anaphylaxis to MOG35–55 in IL-4-Deficient Mice

IL-4 is required for isotype switching to IgE response and can also promote isotype switching to IgG1.23 Active IgE-dependent anaphylaxis does not develop in IL-4- or IL-4 receptor-deficient mice.13, 14 Moreover, the development of IgG1 antibodies with anaphylactic activity requires IL-4,24 suggesting that this cytokine is also involved in the development of IgG1-dependent anaphylaxis. As shown in Figure 3a and Table 1, IL-4-deficient mice exhibited little or no anaphylaxis to MOG35–55. As expected in these IL-4-deficient mice, which have a general impairment of Th2 responses and a significant reduction of IgE and IgG1 antibody production in response to nematode infection,25, 26, 27 total IgE antibody titers were significantly lower in the sera of IL-4−/− vs wild-type mice, and peptide-specific IgG1 antibodies were below the detection limit (Figure 3b). Accordingly, we hypothesize that the resistance against anaphylaxis to MOG35–55 in these IL-4-knockout mice reflects, at least in part, the absence of a peptide-specific IgG1 (or IgG1 and IgE) response. Indeed, anaphylaxis to MOG35–55 requires antibody binding to FcR, as anaphylaxis to the peptide is abrogated,5 or significantly reduced,8 in mice that lack the FcR common γ-chain, and therefore lack FcɛRI (that binds IgE) and FcγRIII (that binds IgG1 immune complexes). However, other mechanisms might also have contributed to resistance against anaphylaxis observed in IL-4−/− mice, such as reduced sensitivity to mediators of anaphylaxis (eg, histamine, PAF, etc) in the absence of this cytokine (reviewed in Finkelman et al13) or the absence of antigen binding to basophils, which has recently been described in these knockout mice.28

Figure 3
figure 3

Expression of anaphylaxis to self-MOG35–55 in IL-4-deficient mice. (a) Changes in body temperature in C57BL/6-IL-4−/− (IL-4−/−) and control C57BL/6 mice on i.p. challenge with MOG35–55 or control peptide-1. †One mouse dead at that time point. *P<0.05 and **P<0.005. The exact numbers of mice in the various groups are given in Table 1. (b) MOG35–55-specific IgG1 and IgG2a, and total IgE antibody titers in sera of IL-4−/− and control C57BL/6 mice (n=5–7 mice per group). *P<0.005.

Full size image

Effects of Blockade of PAF on the Incidence and Severity of Anaphylaxis to MOG35–55

PAF has been shown to play a major role in IgG-dependent anaphylaxis in the mouse,14 and blockade of this mediator can prevent IgG1-mediated passive anaphylaxis in mice.15 As we observed that anaphylaxis to MOG35–55 was not prevented in mice lacking histamine (Figure 2a) or mast cells (Figure 1a), we analyzed the role of PAF in the development of anaphylaxis to this self-peptide. As shown in Table 2, treatment with the PAF antagonist CV6209 slightly reduced the incidence of anaphylaxis to self-MOG35–55 (although this effect was not statistically significant in the number of mice studied), but such treatment did not affect the severity of anaphylaxis, as defined by the mean lowest body temperature developed by mice treated with CV6209 compared with mice treated with vehicle. Treatment with H1R receptor antagonist triprolidine was not effective in preventing anaphylaxis to self-MOG35–55, confirming the results obtained in HDC−/− mice, whereas treatment with CV6209 and triprolidine in combination was slightly more effective than treatment with CV6209 alone in reducing the incidence of anaphylaxis to this peptide (Table 2). Although the effects detected did not achieve statistical significance, these results suggest that PAF plays a role in anaphylaxis to MOG35–55, but is not the only mediator involved.

Table 2 Anaphylaxis in wild-type C57Bl/6 mice treated with a PAF antagonist and/or an H1R antagonist before peptide challenge (in mice immunized to MOG35–55 with Bordetella pertussis toxin)
Full size table

We also evaluated the role of PAF in another model of anaphylaxis to MOG35–55, elicited in C57BL/6 mice that were immunized with this peptide in CFA, but without B. pertussis toxin (PTX) used as an adjuvant. As reported earlier,6 under this protocol of immunization, mice developed anaphylaxis with higher doses of antigen as compared with mice that had received PTX at the time of peptide immunization (500 vs 100 μg). Also, compared with mice that received PTX at the time of immunization, these mice had significantly lower titers of total IgE in their serum (total IgE levels were 204.2±13 ng/ml in mice that did not receive PTX vs 1.0±0.25 μg/ml in mice that received PTX; P=0.007) and higher antigen-specific IgG1 titers (OD was 1.137±0.191 in mice that did not receive PTX vs 0.724±0.074 in mice that received PTX; P=0.037). In mice subjected to this immunization protocol (Table 3), treatment with the PAF antagonist was more effective in reducing the severity of anaphylaxis to MOG35–55 than in our standard protocol. Anaphylaxis occurred in 3 of 10 mice treated with CV6209 vs 6 of 10 mice treated with vehicle (P=0.353), and the mean maximum drop in body temperature was 0.8±0.4° in mice treated with CV6209 vs 2.6±0.7° in mice treated with vehicle (P=0.037). Moreover, in this immunization protocol, treatment with both CV6209 and triprolidine was more effective in preventing anaphylaxis than was treatment with CV6209 or with triprolidine alone (Table 3). Anaphylaxis occurred in 1 of 10 mice treated with CV6209 and triprolidine vs 6 of 10 mice treated with vehicle (P=0.028); the mean maximum body temperature drop was 0.3±0.2° in mice treated with CV6209 and triprolidine vs 2.6±0.7 in mice treated with vehicle (P=0.010). Taken together, our data suggest that PAF may have a more significant role in this model of anaphylaxis to self-MOG35–55 (Table 3) compared with the model in which PTX is used (Table 2). However, in both models, our data suggest that mediators other than PAF are also involved. Indeed, our results with triprolidine are consistent with the possibility that even though it is not essential for the responses, histamine may contribute to some of the changes seen in association with anaphylaxis to MOG35–55, especially in the model in which immunization was elicited with PTX.

Table 3 Anaphylactic shock in wild-type C57Bl/6 mice treated with a PAF antagonist and/or an H1R antagonist before peptide challenge (in mice immunized to MOG35–55 without Bordetella pertussi s toxin)
Full size table

DISCUSSION

Our results indicate that anaphylaxis to self-MOG35–55 can be induced in mice in the absence of mast cells or histamine, but does require IL-4. The findings in this study are consistent with the possibility that the pathway of anaphylaxis to self-MOG35–55 involves IgG1, FcγRIII, macrophages and/or basophils, and PAF.5 However, the PAF antagonist CV6209 was not fully effective in abrogating the changes in body temperature associated with anaphylaxis to MOG35–55 in either of the two models tested, indicating that other mediators are also probably involved. Prior work by us5 and others8 found that anaphylaxis to self in this model was not eliminated in mice lacking the α chain of FcγRIII (FcγRIII α chain−/−). Taken together with the results presented here, this finding suggests that, if IgG1 antibodies indeed contribute importantly to anaphylaxis in this model, such IgG1 antibodies may be functioning through mechanisms other than binding as immune complexes to FcγRIII.

However, our results do not rule out possible contributions in this model of anaphylaxis of an alternative pathway that involves IgE but that requires neither mast cells nor histamine. Like mast cells, basophils express FcɛRI and FcγRIII, and therefore might be activated in this model by either IgE- or IgG1-dependent mechanisms. Basophils are present in the mast cell-deficient W/Wv mice that we used in our experiments, although in some settings in somewhat lower numbers than in the corresponding wild-type mice, as well as in mast cell-deficient W-Sh/W-Sh mice.29, 30, 31, 32 Indeed, depletion of basophils in W-Sh/W-Sh mice in vivo with a monoclonal antibody suppressed IgG1-mediated passive systemic anaphylaxis, revealing an important role for basophils in that model of IgG1-mediated anaphylaxis.15 Thus, although we cannot exclude an involvement of other cells, such as macrophages, in the models of MOG35–55-induced anaphylaxis that we tested, basophils activated by IgE (and/or IgG1) might represent important effector cells in anaphylaxis to self-MOG, especially in mast cell-deficient mice. In future studies, selective depletion of basophils and/or macrophages in vivo might help to define the role of these cells in anaphylaxis to self-MOG35–55.

Blockade of PAF with the antagonist CV6209, which effectively blocked IgG1-dependent passive systemic anaphylaxis,15 only partially reduced the incidence of active systemic anaphylaxis to MOG35–55 in the main model of immunization to MOG35–55 that was used in our study, and this effect did not achieve statistical significance. However, CV6209 seemed to be more effective in substantially reducing the severity of anaphylaxis to this self-peptide in a model that omitted PTX as an adjuvant at the time of immunization. These results suggest that PAF plays a role in anaphylaxis to this self-peptide in the two immunization models examined, but is unlikely to be the only mediator involved. Our results also indicated that a combination of both CV6209 and the H1R antagonist triprolidine was more effective in inhibiting anaphylaxis than was treatment with CV6209 or with triprolidine alone, especially in the immunization model that omitted PTX as an adjuvant (Table 3). These findings are consistent with those reported in a model of self-anaphylaxis induced in a non-obese diabetic (NOD) mouse model of human diabetes mellitus, in which only the treatment with both PAF antagonist and H1R antagonist in combination, but not with either of them used alone, was effective in preventing anaphylaxis to self-insulin peptides.33 One way to interpret all of our data is that even though anaphylaxis to MOG35–55 in mice does not require either mast cells or histamine, and PAF is one of the mediators that contribute to the pathology in this setting, histamine (or another mediator whose effects may be reduced by treatment with triprolidine) also can contribute to the pathology.

The data presented here were derived entirely from studies in mouse models, and their implications for the clinical management of MS and other autoimmune disorders remain to be ascertained. Even though IgE is thought to be the main (or only) Ig isotype that contributes significantly to anaphylactic and allergic responses in humans (reviewed in Finkelman et al13), anaphylaxis has been reported in people in whom there was no evidence of antigen-specific IgE antibodies or mast cell degranulation (reviewed in Finkelman et al13). Some of such cases might reflect the existence of IgG-dependent, but IgE-independent, pathways of anaphylaxis. It is worth mentioning here that MS patients who developed allergic reactions to NB5788, which usually occurred in those receiving the highest doses of this APL, presented with elevated titers of peptide-specific IgG1 antibodies, but not of peptide-specific IgE antibody levels.9 Also, despite the frequent appearance of allergic reactions in glatiramer acetate-treated patients, glatiramer acetate-specific IgE was detected only in one patient.34 It is possible that the allergic reactions to peptide therapies occurring in such individuals might have components of IgG-, but not IgE-, dependent pathways of anaphylaxis. However, the relationship of these clinical findings to the data reported herein from our studies in mice remains to be ascertained. Indeed, in humans, the antibody isotype thought to be most similar to IgG1 in mice is not IgG1, but IgG4.35, 36

In conclusion, we have shown here that anaphylaxis to self-MOG35–55 in mice can occur in the absence of mast cells and histamine, key elements of the classical IgE-, mast cell-, and histamine-dependent pathway of anaphylaxis. Our pharmacological data suggest that PAF can contribute to the pathology associated with anaphylaxis to MOG35–55, especially in a model in which mice are immunized to the peptide without the use of PTX as an adjuvant, but that other mediators are likely also to be involved. Although one must always be cautious in extrapolating the results of mouse studies to humans, our data raise the possibility that drugs that block effector mechanisms of the classical IgE-dependent pathway of anaphylaxis, such as anti-histamines or inhibitors of mast cell degranulation, might be ineffective in preventing certain forms of peptide-induced anaphylaxis in human subjects.