IL-33, IL-25 and TSLP contribute to development of fungal-associated protease-induced innate-type airway inflammation

Certain proteases derived from house dust mites and plants are considered to trigger initiation of allergic airway inflammation by disrupting tight junctions between epithelial cells. It is known that inhalation of proteases such as house dust mite-derived Der p1 and/or papaya-derived papain caused airway eosinophilia in naïve mice and even in Rag-deficient mice that lack acquired immune cells such as T, B and NKT cells. In contrast, little is known regarding the possible involvement of proteases derived from Aspergillus species (fungal-associated proteases; FAP), which are ubiquitous saprophytic fungi in the environment, in the development of allergic airway eosinophilia. Here, we found that inhalation of FAP by naïve mice led to airway eosinophilia that was dependent on protease-activated receptor-2 (PAR2), but not TLR2 and TLR4. Those findings suggest that the protease activity of FAP, but not endotoxins in FAP, are important in the setting. In addition, development of that eosinophilia was mediated by innate immune cells (ILCs) such as innate lymphoid cells, but not by acquired immune cells such as T, B and NKT cells. Whereas IL-33, IL-25 and thymic stromal lymphopoietin (TSLP) are involved in induction of FAP-induced ILC-mediated airway eosinophilia, IL-33—rather than IL-25 and/or TSLP—was critical for the eosinophilia in our model. Our findings improve our understanding of the molecular mechanisms involved in induction of airway inflammation by FAP.

In addition, Der p1 can induce cell death (necrosis) of epithelial cells, which then release damage-associated molecular patterns (DAMPs; also called alarmins) such as HMGB-1, IL-1α, IL-33, uric acid and ATP 2,10 . Those alarmins immediately induce inflammation by activating immune cells. Inhalation of a plant-derived cysteine protease, papain, which is homologous to HDM-derived Der p1 11 , by humans resulted in development of asthma-like airway inflammation 12 . Likewise, inhalation of papain and/or Der p1 induced airway eosinophilia in naïve mice and even in Rag-deficient mice that lack acquired immune cells such as T, B and NKT cells 13,14 . In the setting, IL-33 derived from alveolar epithelial cells damaged by papain acts as a DAMP that activates group 2 innate lymphoid cells (ILC2) and basophils, followed by induction of type 2 cytokine-dependent eosinophilia [13][14][15][16] . Spores and/or conidia from Aspergillus species, which are ubiquitous saprophytic fungi in the environment 17,18 , are thought to contribute to development of atopic asthma [19][20][21] . However, airway neutrophilia, rather than eosinophilia, was observed in mice at 18 hours after a single inhalation of proteases derived from Aspergillus species (fungal-associated proteases; FAP) 22 . We previously reported that inhalation of high-dose papain resulted in infiltration of neutrophils, whereas inhalation of the optimal dose of papain resulted in infiltration of eosinophils, into BAL fluids of naïve mice 16 . Therefore, we hypothesized that inhalation of FAP also may result in infiltration of eosinophils into BAL fluids of naïve mice. In fact, we found that over a certain concentration range FAP induced airway eosinophilia in mice, suggesting involvement of FAP in development of asthma 13,14 . Therefore, in the present study, we investigated the mechanisms underlying airway eosinophilia induced by FAP.

Results
Fungal-associated proteases (FAP) induce asthma-like airway eosinophilia in mice without prior sensitization. Others observed infiltration of neutrophils rather than eosinophils in BAL fluids of naïve mice at 18 hours after one inhalation of FAP 22,23 . We previously reported that inhalation of a high dose of papain resulted in infiltration of neutrophils, whereas inhalation of the optimal dose of papain resulted in infiltration of eosinophils, into BAL fluids of naïve mice 16 . Therefore, we hypothesized that inhalation of FAP also may lead to infiltration of eosinophils into BAL fluids of naïve mice. Based on those reports, we attempted to establish a murine model of FAP inhalation-induced airway eosinophilia. C57BL/6 wild-type mice were intranasally treated with 800 μg/ml FAP once per day for 1 to 3 days (see the box in Fig. 1). Twenty-four hours after the last FAP inhalation, the cell profiles in BAL fluids were investigated. As shown in Fig. 1a, the numbers of eosinophils, neutrophils, macrophages and lymphocytes were slightly, but not significantly, increased after a single inhalation of FAP. Macrophages and lymphocytes, but not eosinophils or neutrophils, were significantly increased after two inhalations of FAP, while macrophages, lymphocytes and eosinophils were significantly increased after three inhalations of FAP (Fig. 1). Therefore, our model for studying the mechanisms of FAP-induced airway eosinophilia used mice treated intranasally with FAP once per day for 3 days.
Next, C57BL/6 wild-type mice were treated intranasally with various concentrations of FAP (100-1600 μg/ ml), heat-inactivated FAP (100-1600 μg/ml) or PBS once per day for 3 days. In our pilot studies, we found that the numbers of leukocytes (total, macrophages, lymphocytes, neutrophils and eosinophils) in the BAL fluids of mice treated with 100, 200, 400, 800, 1200 and 1600 μg/ml heat-inactivated FAP were comparable to those of mice treated with PBS (data not shown). Therefore, we used 800 μg/ml heat-inactivated FAP in the later experiments. The numbers of various leukocytes in the BAL fluids and the severity of lung inflammation determined by histological analysis were not dramatically increased at the lowest concentrations of FAP (100 and 200 μg/ml) compared with PBS ( Fig. 2a-c). Although the numbers of leukocytes in the BAL fluids were not significantly increased at 400 μg/ml FAP, lung inflammation accompanied by epithelial hyperplasia and mucus secretion was observed ( Fig. 2a-c). Consistent with Fig. 1, a significant increase in eosinophils but not neutrophils was observed in BAL fluids from mice that inhaled 800 μg/ml FAP, while both eosinophils and neutrophils were significantly increased after inhalation of the highest FAP doses (1200 and 1600 μg/ml) (Fig. 2a). Asthma-like inflammation was seen in the lungs of mice treated with 800 μg/ml FAP, while chronic obstructive pulmonary disease-like inflammation associated with tissue destruction was observed at 1200 and 1600 μg/ml FAP (Fig. 2b,c; and data not shown). Similar observations were reported in mice treated with a cysteine protease, papain 16 . In that study, ELISA hardly detected cytokines in the BAL fluids of mice after papain inhalation because they had been cleaved by the protease activity of papain 16 . Likewise, using ELISA, we hardly detected IFN-γ, IL-4, IL-13 or IL-17A in the BAL fluids of mice after FAP inhalation (data not shown). On the other hand, we found that the levels of IL-5 were significantly increased in the BAL fluids of mice after inhalation of 800 μg/ml FAP (Fig. 2d).
Therefore, in our subsequent experiments, we used 800 μg/ml FAP to induce asthma-like airway eosinophilia in mice. In addition, inhalation of heat-inactivated FAP did not induce airway eosinophilia accompanied by eosinophilia or elevation of IL-5 in the BAL fluids ( Fig. 2a-d). Those findings suggest that the protease activity of FAP is crucial for induction of asthma-like airway eosinophilia. In support of this, mice deficient in protease-activated receptor-2 (Par2 −/− mice) showed reduced inflammatory cell counts in the BAL fluids during FAP-induced airway eosinophilia (Fig. 3a). In addition, the number of BAL cells was comparable between wild-type and Tlr2 −/− Tlr4 −/− mice during FAP-induced airway eosinophilia, suggesting that contamination by endotoxin does not influence the induction of FAP-induced airway eosinophilia (Fig. 3b).
IL-33, but not IL-25 or TSLP, is critical for FAP-induced airway eosinophilia, independently of acquired immune cells. Infiltration of neutrophils into BAL fluids was seen even in Rag1 −/− mice at 18 hours after one inhalation of FAP, indicating that acquired immune cells such as T, B and NKT cells are not essential for that response 22 . In our model, as well, FAP-induced airway inflammation was observed in Rag2 −/− mice (Fig. 4a). In particular, neutrophils, macrophages and lymphocytes, but not eosinophils, were increased in BAL fluids of Rag2 −/− mice compared with wild-type mice (Fig. 4a), suggesting that acquired immune cells such as T, B and NKT cells were not essential for FAP-induced airway eosinophilia. On the other hand, FAP-induced airway eosinophilia was markedly impaired in Rag2 −/− Il2rg −/− mice, which lack acquired immune cells, innate lymphoid cells (ILCs) and IL-2, IL-4, IL-7, IL-9, IL-15 and IL-21, compared with wild-type mice (Fig. 4b), suggesting that ILCs and/or IL-2Rγ-ligands may be important for FAP-induced airway eosinophilia. We found that the number of group 2 ILCs (ILC2s) was increased in the lungs of wild-type mice, but not Rag2 −/− Il2rg −/− mice, after FAP inhalation (Fig. 4c). IL-25, IL-33 and TSLP are known to be produced by epithelial cells in the lungs 2,24 and involved in activation of ILC2s to produce type 2 cytokines, thereby contributing to induction of type 2 cytokine-dependent inflammation such as airway eosinophilia 25,26 . We found that the mRNA expression level of Il33, but not Il25 or Tslp, was significantly increased in the lungs of wild-type mice treated intranasally with 800 μg/ml FAP once per day for 1 to 3 days as shown in Fig. 1 (Fig. 5a), or with 0-1,200 μg/ml FAP once per day for 3 days as shown in Fig. 2 (Fig. 5b), in comparison with wild-type mice treated intranasally with PBS or heat-inactivated FAP. In association with this, the IL-33 protein levels were increased in the nuclei of alveolar epithelial cells from FAP-treated wild-type mice compared with heat-inactivated FAP-treated wild-type mice (  (a) Naïve C57BL/6-wild-type mice were treated intranasally with or without 800 μg/ml FAP once per day for 1 to 3 days as in Fig. 1. Twenty-four h after the last FAP inhalation, lungs were collected. mRNA expression levels of Il25, Il33 and Tslp in the lungs were determined by quantitative PCR (naïve, n = 10; FAP, n = 10). (b) Naïve C57BL/6-wild-type mice were treated intranasally with various concentrations of FAP (100-1600 μg/ml), heat-inactivated FAP (HI-FAP; 800 μg/ml) or PBS once per day for 3 days as in Fig. 2a. Twenty-four h after the last inhalation, lungs were collected. mRNA expression levels of Il25, Il33 and Tslp in the lungs were determined by quantitative PCR (PBS, n = 9-12; HI-FAP, n = 6-11; FAP, n = 6-10). (c) Mice were treated intranasally with 800 μg/ml FAP and heat-inactivated FAP (HI-FAP) once per day for 3 days. Twenty-four h after the last inhalation, lungs were collected. IL-33 expression in sections of the lungs was detected by immunohistochemistry. Blue = Mayer's hematoxylin, and brown = anti-IL-33 Ab staining, respectively. ×10 and ×160 (inserted photos). Data show the mean ± SEM and representative results that were obtained in 2 independent experiments (a,b). *p < 0.05, **p < 0.005 and ***p < 0.0005 vs. the corresponding values for PBS-treated mice.  (Fig. 6a). On the other hand, lymphocytes were significantly reduced in the BAL fluids of FAP-treated Il33 −/− and Crlf2 −/− , but not Il25 −/− , mice (Fig. 6a). In addition, the histological score of lung inflammation was significantly lower in FAP-treated Il33 −/− , but not Il25 −/− or Crlf2 −/− , mice than in FAP-treated wild-type mice (Fig. 6b,c). These observations suggest that IL-25, IL-33 and TSLP are crucial for FAP-induced airway eosinophilia, but IL-33 is more important than IL-25 and TSLP for FAP-induced airway inflammation accompanied by infiltration of macrophages and lymphocytes as well as eosinophils. Type 2 cytokines such as IL-5 and IL-13 were shown to be important for IL-33-induced airway eosinophilia 13,27,28 . Therefore, we investigated involvement of type 2 cytokines in airway eosinophilia after FAP inhalation. The numbers of eosinophils in the BAL fluids were similar in Il4 −/− mice compared with wild-type mice after FAP inhalation (Fig. 7). On the other hand, eosinophils in the BAL fluids were greatly reduced in both Il5 −/− mice and Il13 −/− mice compared with wild-type mice (Fig. 7). These observations indicate that IL-5 and IL-13 are crucial for FAP-induced airway eosinophilia.

Discussion
Protease allergens derived from HDM and/or fungi are considered to be important risk factors for development of allergic asthma. An HDM-derived cysteine protease, Der p1, can induce necrosis of epithelial cells, leading to release of DAMPs such as HMGB-1, IL-1α, IL-33, uric acid and ATP and then induction of local inflammation via activation of innate immune cells 2,10 . Inhalation of a plant-derived cysteine protease, papain, which is homologous to HDM-derived Der p1 11 , led to development of asthma-like airway inflammation 12 in naïve mice and even Rag-deficient mice 13 . Likewise, airway inflammation was observed in naïve and Rag-deficient mice at 18 hours after a single inhalation of FAP derived from Aspergillus species 22 . In the setting, both neutrophils and eosinophils were recruited in the lungs, with neutrophils being dominant 22 . In addition, IL-17F, but not IL-17A, was shown to be important for induction of FAP-mediated airway neutrophilia 23 . On the other hand, it is unclear which cytokines are involved in FAP-induced airway eosinophilia, although STAT6, which is essential for IL-4 and IL-13 signals, and C3 were not required. To clarify this issue, we established a model of airway eosinophilia, rather than neutrophilia, by treating mice intranasally with a certain concentration range of FAP once per day for 3 days.
IL-25, IL-33 and TSLP, which are produced mainly by epithelial cells in airways, induce type 2 cytokines by various types of cells such as Th2 cells, NKT cells and/or ILC2 2,24-26 . Epithelial cell-derived IL-33 was reported to be crucial for papain-induced airway eosinophilia by production of type 2 cytokines by ILC2s and basophils [13][14][15][16] . Likewise, we found that epithelial cell-derived IL-33 was important for development of FAP-induced airway eosinophilia. Similar to Il33 −/− mice, Il25 −/− mice and Crlf2 −/− mice showed attenuated airway eosinophilia after FAP inhalation, suggesting that IL-25 and TSLP, in addition to IL-33, are involved in FAP-induced airway In addition, the expression levels of Il33, but not Il25 or Tslp, mRNA were increased in the lungs of mice after FAP inhalation. IL-33 protein was also increased in the nuclei of alveolar epithelial cells in the setting. Since FAP-induced airway eosinophilia in Par2 −/− mice was attenuated, but not completely abrogated, induction of IL-33 in alveolar epithelial cells is both Par2-dependent and -independent. We previously demonstrated that IL-25 is produced by epithelial cells in the lungs and noses of mice during OVA-induced airway inflammation and HDM-induced allergic rhinitis, respectively 29,30 . On the other hand, in the lungs of FAP-treated mice, both IL-25 and TSLP were below the limit of immunohistochemical detection. These observations suggest that, rather than IL-25 and TSLP, IL-33 is the most important cytokine for development of FAP-induced airway inflammation. Papain-induced, IL-33-mediated airway eosinophilia was reportedly mediated by ILC2s and basophils, but not by acquired immune cells such as T, B and NKT cells [13][14][15][16] . Similarly, we showed here that FAP-induced airway eosinophilia was markedly abrogated in Rag2 −/− Il2rg −/− mice, but developed normally in Rag2 −/− mice. These observations suggest that ILCs, but not T, B or NKT cells, may contribute in the setting, although we generated no direct evidence that IL-33-stimulated ILC2s induced airway eosinophilia in mice after FAP inhalation.
Taken together, we found that inhalation of FAP derived from Aspergillus resulted in induction of IL-33 in alveolar epithelial cells via the PAR-2 signaling pathway, but not the TLR2 or TLR4 pathway, followed by development of airway eosinophilia mediated by ILCs, but not T, B or NKT cells.

Bronchoalveolar lavage fluids.
Twenty-four h after the last challenge with FAP, HI-FAP or PBS, bronchoalveolar lavage (BAL) fluids were collected as described elsewhere 33,34 . The BAL fluids were centrifuged, and the BAL cells were resuspended in 200 μl of HBSS supplemented with 2% FCS. The number of each cell type among the BAL cells was counted with an automated hematology analyzer (XT-1800i; Sysmex, Hyogo, Japan), according to the manufacturer's instructions.
Quantitative PCR. Total RNA was extracted from lungs of mice 24 h after the last inhalation of FAP, HI-FAP or PBS, and cDNA was prepared as described elsewhere 34,35 . The expression levels of mRNA for cytokines in each sample were determined by the comparative Ct method after normalization with GAPDH expression using a CFX384 Touch qPCR System (Bio-Rad, Hercules, CA), as described elsewhere 34,35 . Primers, which were designed as shown in Table 1, were prepared by Eurofins Operon (Tokyo, Japan).
Histology. Twenty-four h after the last FAP inhalation, lungs were harvested and fixed in Carnoy's solution. The fixed tissues were embedded in paraffin, sliced into 4-µm sections, and subjected to hematoxylin and eosin staining or periodic acid-Schiff staining. The severity of inflammation in lungs was scored as described elsewhere 34 . Immunohistochemistry. Twenty-four h after the last FAP inhalation, the trachea was cannulated with a 22-G needle attached to a syringe, and 4% paraformaldehyde was injected. Then the lungs were harvested and fixed in 4% paraformaldehyde at 4 °C overnight. The fixed tissues were embedded in paraffin and sliced into 4-µm sections. The sections were incubated with Blocking One Histo (0634904; Nacalai, Japan). After blocking, they were incubated with goat anti-mouse IL-33 polyclonal Ab (AF3626; Abcam plc, UK) in a humidified staining box