Herein we describe a role for allergen-derived proteases and PAR-2 in the release of CCL20 and GM-CSF into the airways of mice following a single exposure to GC frass. GC frass is complex and contains a number of components including active serine proteases [23
], endotoxin [23
], a TLR2 agonist [24
], and a number of unknown components. In this report, our goal was to assess the effect of the protease component(s) in GC frass on their ability to induce chemokine production in vivo. To do this, we used a variety of reagents, including protease-depleted GC frass, protease-enriched GC frass and PAR-2-deficient mice. We found that the kinetics of GM-CSF and CCL20 release are different following in vivo exposure to GC frass, in that GM-CSF peaks early while CCL20 has more of a delayed release. While the mechanism responsible for differences in the kinetics of GM-CSF and CCL20 release are unclear, we speculate that the early burst of GM-CSF may be to facilitate DC activation/maturation. As mature DCs will then migrate from the lung to the lung-draining lymph nodes, the subsequent burst of CCL20 release may be responsible for replacing the mature DCs with immature DCs recruited directly from the circulation. In support of this, it has been demonstrated that allergen-bearing, activated DCs can be observed in the lung-draining lymph nodes as rapidly as 6 h after allergen exposure [25
We confirmed that these chemokines are being synthesized by the airway epithelium, and are likely first responders to allergen exposure. It is not surprising that allergen-derived proteases can act directly on airway epithelium as we have previously shown the role of proteases in regulating signaling pathways following PAR-2 activation [23
]. Other reports have also investigated the importance of the allergen-associated protease in regulating CCL20 production in vitro. Pichavant et al. [28
] showed that the proteolytically active Der p1 induced CCL20 while the inactive form, pro-Der p1, failed to regulate CCL20 production in BEAS-2B cells. In addition, they showed that chemical inhibition of Der p1 protease activity also inhibited CCL20 production, suggesting that its action is linked to its protease activity. Trypsin has been shown to induce PAR-2-mediated CCL20 production in human gingival epithelial cells [29
]. However, Nathan et al. [14
] recently showed that HDM-induced CCL20 production in the human airway epithelial cell line 16HBE14o- cells was independent of protease activity. The protease activity in the HDM preparation was not discussed in that study, so it is unclear how much protease activity was initially in the HDM preparation. In the current study, we enriched the protease in GC frass using column chromatography methods. Our data indicate that the enriched protease can induce GM-CSF and CCL20 production at a level similar to that of GC frass at 3 h after instillation. Interestingly though, at 18 h after instillation, the levels of CCL20 production from the enriched protease were substantially lower than those from GC frass. One interpretation of these data is that enrichment of the protease may remove a component in GC frass that aids in the stabilization of the protease or in protection from degradation. It is likely that endogenous pulmonary antiproteases (i.e. serine leukocyte protease inhibitor or α1
antitrypsin) could inactivate the enriched protease at an increased rate, or with increased potency.
Interestingly, a consequence of the enrichment of the protease was the removal of endotoxin. We have previously attempted to remove the endotoxin using a commercially available endotoxin-removal column; however, we were unsuccessful in removing more than 50% of the endotoxin [K. Page, unpublished observation]. It is important to note that endotoxin was not the only component removed from GC frass during the enrichment procedure and at this point it is unclear what role these unknown components have on chemokine production. Complete removal of endotoxin often includes very harsh conditions (i.e. exposure to acids or alkalis at concentrations equal to or higher than 0.1 M or temperatures of 250°C for 30 min) which would also alter the activity of a serine protease. Since we were unable to selectively remove endotoxin from GC frass while retaining the presence of proteases and other proteins, we cannot conclude in the present study that endotoxin plays a major role in regulating CCL20 and GM-CSF production. In addition, treatment with commercially available ‘purified’ endotoxin may not be similar to the endotoxin found in GC frass. Thus, while it is possible that endotoxin plays a role in mediating these effects, the overall goal of this study was to examine the ability of the active serine protease in GC frass to regulate chemokine expression in vivo.
At this point, we have no evidence to suggest that PAR-2-deficient mice are unable to mount a normal immune response because of an abnormality in the DC population. While Fields et al. [30
] showed that DCs do not spontaneously develop from the bone marrow of PAR-2-deficient mice, a subsequent study failed to find a direct role for PAR-2 in the differentiation of bone marrow from wild-type and PAR-2-deficient mice into bone marrow DCs when cultured in the presence of GM-CSF [31
]. In addition, there was no difference between DC subset frequencies in the lymph nodes of PAR-2-deficient mice compared to wild-type mice [31
]. In our study, there was no difference in the numbers of mDCs in the PBS-stimulated PAR-2-deficient mice compared to wild-type mice. If PAR-2 deficiency was responsible for the development of DCs, we would expect a lack of DCs in the lung even in the unchallenged state. One difference between our work and that of Ramelli et al. is that they found that selective PAR-2 activation increased maturation of bone marrow DCs as evidenced by increased MHC class II and CD86 expression on DCs and that PAR-2-deficient mice demonstrated a reduced frequency of FITC+ DCs following FITC painting [31
]. While we did not study the maturation of bone marrow DCs in the presence of GC frass proteases in this study, our results did not show any differences in the uptake of Alexa-Fluor 405-labeled GCs by pulmonary DCs between wild-type and PAR-2-deficient mice. The reasons for this are unclear, but may reflect differences in mechanisms of allergen uptake between small molecules (FITC) and complex antigens (GC frass) or differences in PAR-2 biology between DCs at epidermal versus mucosal sites. Further study is needed to determine the root cause of these differences.
GM-CSF is clearly a proallergic signal, as overexpression of GM-CSF was shown to induce allergic airway inflammation to ovalbumin exposure compared to ovalbumin exposure alone [18
]. Our data demonstrate that the complex allergen GC frass can directly induce GM-CSF expression from epithelial cells in a mechanism at least partially dependent on PAR-2 activation. In support of this, BEAS-2B cells treated with Der p1 or Der p9 were shown to produce GM-CSF [32
], and Der p3 and Der p9 were shown to activate PAR-2 and induce the release of GM-CSF [33
]. Another study confirmed that recombinant allergens Der f 1 and Der p 1 stimulated the production of GM-CSF in normal human keratinocytes and this could be inhibited by the addition of cystatin A (a cysteine proteinase inhibitor) [34
]. Normal and asthmatic bronchial epithelial cells have been shown to release GM-CSF following Der p exposure [35
] and, recently, nasal biopsies from patients with chronic rhinosinusitis without nasal polyps had increased PAR-2 expression [36
]. Collectively, these studies suggest that PAR-2 activation by protein allergens can induce GM-CSF production by epithelial cells, suggesting a possible mechanism by which allergen exposure could initiate allergic responses.
Overall, the findings in this study show that the protease-PAR-2 plays a role in GM-CSF and CCL20 production in the airways of mice. The subsequent infiltration and/or differentiation of mDCs could be sufficient to initiate allergic airways in the presence of allergen. It is important to consider that the relatively small, but statistically significant change in the percentage of mDCs in the airways of wild-type and PAR-2-deficient mice following GC frass exposure was performed on whole lung. It is possible that if we looked directly at the actual differences in mDC populations in a local microenvironment, these changes might be greater. We anticipate that in the upper airways where the instillation of GC frass is likely to be the most concentrated, there may be even higher levels of mDC recruitment. In our study, it does not appear that PAR-2 plays a role in antigen uptake, but it is possible that an additional role for PAR-2 could be in the regulation of the overall maturation state of the DCs, the ability of DCs to process and present antigen, or in the migration of activated DCs to the draining lymph nodes, where T cell activation can occur. Elucidation of the mechanism by which proteases associated with allergens can initiate and/or augment the early innate immune response may ultimately lead to a better understanding of how an allergen is able to elicit a shift from tolerance to disease.