DISCUSSION
In this paper we propose a new murine AP model induced by two antigens, OVA and SEB, for studying MCs dependent AI. OVA, a typical experimentally used antigen and allergen was administered together with SEB, that by itself is a superantigen, used as adjuvant. The SEB adjuvant-like activity was investigated before in an asthma model (18). It was shown that repeated sensitization with a combination of SEB together with OVA increased total cells and Eos number in bronchoalveolar lavage compared to sensitization with either OVA or SEB alone (18). Considering this data, we sought to use SEB to avoid the MCs independent inflammation induction by Alum (10). Moreover, we aimed to use a model that would include the typical S.aureus /exotoxins co-existence with the allergen taking place in allergy diseased patients. In our model we observed an inflammation pattern similar to what was previously reported for OVA/Alum AP (19). A difference was found though in the timing of the peak of inflammation that occurred in the OVA/Alum model at 72h- 96h after challenge in BALB/C mice (19, 20) and also in C57BL/6 (unpublished data). In the present OVA/SEB model the peak of inflammation is 48h after challenge as indicated by the increase of Eos numbers a main characteristic of AI. This is in accordance with the results of Zuany-Amorim C et al, showing that OVA/Alum AP induction in BALB/C mice displayed a peak of inflammation at 48h (21). An additional explanation to this early peak could be provided by the superantigen role of SEB, as we previously found that SEB-induced peritonitis presented with increased total cell and Eos numbers 48h after challenge (22). In our new model we also detected an increase in macrophages infiltration. This is not surprising, since it was previously reported that both SEB and OVA, respectively, increased macrophages numbers on the site of inflammation (23, 24). As expected, Eos and consequently total cell numbers decreased in the course of AP, while macrophages remained stable from 48h after challenge. This might be the effect of the contribution of macrophages in both inflammation and resolution (25, 26). Indeed, we observed that macrophages phenotype in our model shifted towards M2, as shown by Arg1 (17) expression increasing during the course of AP (Supplemental Appendix Figure 1B). To further emphasize the induction of AI by using OVA/SEB AP, we analyzed sCD48 levels in peritoneal lavage at the detected peak of inflammation. It was previously published by our group that sCD48 levels correlate with Eos numbers in SEB-induced peritonitis (22). Furthermore, sCD48 levels in serum of mild asthma patients were significantly higher than in the control group and correlated with the Eos numbers (27). Consistently, in this model we found high levels of sCD48 in the peritoneal lavage at 48h that returned to basal levels 168h after challenge.
Many allergic diseases are characterized by the involvement of a mixed Th2 and Th1 (28–31) response and, as demonstrated by more recent evidence, of a Th17 one (32). For example, an AD murine model induced by OVA/SEB displayed increased mRNA levels of the Th2 cytokines IL-4 and IL-13, together with the Th1-related cytokines INFɣ and 12p40 (12). On the other hand, Bui et al. showed, in a murine OVA/Alum-induced asthma model, induction of IL-4 but not INFɣ release in mice bronchoalveolar lavage fluid (33). In addition, Bui et al. also showed an increase in the release of the Th17-related cytokine IL-17A (33). Significant levels of IL-17A, both at mRNA and protein levels were found in asthma patients (32). IL-17A was detected in the skin of OVA/SEB-induced AD mice (31). Correspondingly with the aforementioned evidence, our AP model showed significantly elevated levels of IL-4 simultaneously with a trend of increased levels of both IL-17A and INFɣ. These findings indicate that our model might induce a mixed Th1/Th2/Th17 response skewed towards Th2. After characterizing the inflammatory features of the OVA/SEB-induced AP, our next aim was to understand the involvement of MCs. We therefore evaluated MCs degranulation by measuring the levels of the MC preformed mediators, tryptase and TNFα (34, 35), in the peritoneal lavages. Indeed, we detected significantly higher levels of both mediators shortly after challenge (30 min), indicating that MCs are directly activated by OVA/SEB, possibly via IgE-dependent activation as we detected an increase in OVA specific IgE levels in the mice serum and SEB specific IgE, as shown by TNFα release from serum-sensitized BMMCs following activation with SEB. This is in accordance with previous reports showing that both OVA and SEB induce production of specific IgE during murine AI (12). In addition, it was shown that children suffering from AD presented with high levels of SEB specific IgE in their blood (36).
MC activation in vivo has been linked to increased vascular permeability (37). Indeed, this effect was found in our model as well, since 30 min after challenge we detected a significant increase in this parameter. To bolster the important role of MCs, we performed the OVA/SEB AP in MC-deficient (Sash) mice. Due to a mutation in the ckit regulatory element, Sash mice have MC deficiency in the peritoneal cavity and in additional sites at a young age, and in the skin as they age (38). Therefore, they have been extensively employed to assess the MC role in different pathological conditions. Sash mice sensitized and challenged with OVA/SEB presented attenuated inflammatory features such decreased recruitment of Eos and reduced release of sCD48. Reconstitution with BMMCs in the peritoneum in an overshoot protocol partly restored the inflammatory phenotype, as shown by the increase of peritoneal total cells and Eos numbers. Notably, BMMCs overshoot in the peritoneal cavity of PBS-challenged WT mice did not elicit Eos recruitment, demonstrating that the injected BMMCs did not induce inflammation by themselves (unpublished data).
In conclusion, the OVA/SEB-induced AP model is a Th2-skewed, MC-dependent AI model. This model is also closer to human pathology due to the presence of the SEB together with a potential allergen, such as OVA, that is relevant to human allergic diseases. Therefore, we suggest this model as an additional, useful tool to study AI.
ACKNOWLEDGEMENTS: H.P. planned and performed experiments, analyzed the data, prepared the figures, and wrote the manuscript. P.G.P. performed macrophage related experiments, assisted with in vivo experiments, analyzed the data and edited the manuscript. F.L.S. designed and supervised the study, analyzed the data, advised, reviewed and edited the manuscript.
The authors would like to thank Dr. Micha Ben-Zimra for helpful scientific discussions, Dr. Mansour Seaf for technical assistance and Ms. Alexandra Eliassaf (The Core Research Facility, The Faculty of Medicine, The Hebrew University of Jerusalem) who provided advice on the FC analysis and technical help.
The work was supported by grants from the Israel Science Foundation (ISF, 472/15), Vigevani Foundation, Rosetrees Trust (UK), Aimwell Trust (UK) to FLS. FLS is affiliated with the Adolph and Klara Brettler Center for Molecular Pharmacology and Therapeutics at the School of Pharmacy of The Hebrew University of Jerusalem.
Conflict of Interest: The authors report no conflicts of interest.