In this study, we identified the eosinophil subsets in lung tissue of HDM + DEP exposed mice. Inhibition of IL-5 signalling in HDM + DEP exposed mice reduced both pulmonary eosinophil subsets and peribronchial neutrophilic inflammation, but not NET formation and BAL neutrophilia, mucus production and AHR. We demonstrated that neutrophils accumulate rapidly in BALF upon combined HDM + DEP exposure concomitant with higher expression of neutrophil-recruiting chemokines and NET formation. Therapeutic inhibition of NE (sivelestat) only tended to decrease inflammation in the pollutant-aggravated allergic asthma model. Moreover, neutrophil depletion did not reduce eosinophilic inflammation upon HDM + DEP exposure. None of the neutrophil-interfering treatments affected the eosinophil subsets (Supplemental Fig. 6).
Two subsets of eosinophils -hEOS with homeostatic functions and inflammatory iEOS- have previously been identified by flow cytometry in the mouse lung after allergen exposure (7) and in human samples including blood and induced sputum of asthma patients (18, 19). Mesnil et al. provided evidence that lung hEOS and iEOS represent distinct terminally differentiated eosinophils (7), while another study supported the concept that hEOS can become iEOS during inflammatory processes (20). In our experiments, both hEOS and iEOS were present in lung tissue of HDM + DEP exposed mice and increased significantly upon combined HDM + DEP exposure. Notably, after the last HDM + DEP exposure, the percentage hEOS in lung tissue declined over time while the percentage iEOS increased. Moreover, in blood, nearly all eosinophils were hEOS and their numbers also increased with time after the last HDM + DEP exposure. Together, these data suggest that hEOS may transform into iEOS under inflammatory conditions. Anti-IL-5 treatment clearly affected both eosinophil subsets in lung tissue of HDM + DEP exposed mice which is not in accordance with Mesnil et al., who reported that hEOS are not dependent on IL-5 for their presence in lungs and blood after HDM exposure (7). Differences in experimental design may explain these contrasting data. Our findings however may be of clinical importance since it has been suggested that benralizumab (anti-IL-5R) could be more harmful because both eosinophils subsets are depleted while mepolizumab (anti-IL-5) would only affect the iEOS subset (21). In blood of severe eosinophilic asthma patients, mepolizumab treatment induced a marked reduction of iEOS while the proportion of hEOS increased. Moreover, they also assumed that hEOS and iEOS could be the same cells in different activation states, depending on the cytokine release in the environment (22).
Interestingly, anti-IL-5 therapy significantly decreased peribronchial neutrophils in HDM + DEP exposed mice, without affecting BAL neutrophils and NET formation. The expression of CD125 (IL-5 receptor alpha) was thought to be restricted to eosinophils and basophils in mice, but was also recently found on murine lung neutrophils (7, 23, 24) suggesting that neutrophils may also be influenced by IL-5. Moreover, a recent study has shown that CD125 is also widely expressed on human blood and airway neutrophils (25), suggesting that IL-5 and IL-5R targeting may also affect neutrophilic inflammation. Although the specificity of neutrophil staining with some anti-CD125 clones is debated (26), it is noteworthy that treatment with anti-IL-5R (MEDI-563) in mild atopic asthma patients led to a slight decrease in blood neutrophil number (27). However, results are controversial since another study observed concurrent increased neutrophil sputum counts in severe asthma patients treated with benralizumab (28).
Notably, in our model also type 2-associated asthma features including mucus production, AHR, chemokine and cytokine production were not significantly affected by anti-IL-5 treatment upon HDM + DEP exposure. These results are in accordance with a recent study in which IL-4/IL-13 blockade (anti-IL-4Rα) and IL-5 inhibition were compared in a HDM-induced asthma murine model. IL-4Rα blockade improved lung function decreased chemokine production and prevented mucus production. IL-5 neutralization however did not significantly impact these changes meaning that reducing eosinophil numbers alone does not influence other inflammatory features (29). Also studies with eosinophil-deficient mice have shown that type 2 inflammation remodelling and lung function are not dependent on eosinophils (30, 31). Many biologics targeting type 2 inflammation have been approved for the treatment of asthma leading to reduced asthma exacerbations. However, studies with anti-IL-5, mepolizumab and reslizumab, have reported inconsistent improvements in other secondary endpoints (FeNO and FEV1, measures of type 2 inflammation and lung function respectively) despite decreased eosinophil levels and exacerbation rate (32, 33).
Upon subacute HDM + DEP exposure, neutrophils were highly present 24 hours after the exposure and declined thereafter. This early neutrophilic inflammation associated with increased expression of neutrophil-attracting chemokines and NET formation. These data, together with the increase in eosinophils with time, suggested that neutrophils may play a role in the development of eosinophilic responses. To investigate this, we administered anti-Ly6G antibodies to deplete the neutrophils in BAL, lung and blood (Supplemental Fig. 8). This neutrophil depletion did not induce differences in eosinophilic inflammation upon HDM + DEP exposure. Patel et al. demonstrated that neutrophil depletion in a murine HDM asthma model resulted in increased type 2 inflammation, airway remodelling and hyperresponsiveness (34), whereas neutrophil depletion in an Alternaria alternata asthma model led to a significant reduction of eosinophils in BALF (35). Of note, HDM + DEP exposure in our model induces a very strong eosinophilic inflammation compared to other murine models, which may explain the limited effects of neutrophil depletion. Moreover, anti-Ly6G can induce rapid renewal of neutrophils in the bone marrow which have lower Ly6G membrane expression and thus less susceptible to anti-Ly6G-mediated depletion (36).
NE is a neutrophil-associated protease with a very important role in NETosis since it translocates to the nucleus and degrades histones, promoting chromatin decondensation and further NET release (37). In an OVA-asthma mouse model, treatment with the NE inhibitor sivelestat significantly attenuated allergic airway responses including type 2 cytokine levels and eosinophilia leading to reduced AHR and goblet cell metaplasia (38). Therefore, we aimed to target NE in our HDM + DEP asthma model. Therapeutic inhibition of NE only tended to decrease BAL inflammation. Of note, BAL neutrophil numbers were higher in the OVA asthma model compared to our HDM + DEP- model, which may explain their better response towards sivelestat. In future studies, it could therefore be of interest to test sivelestat in our previously published chronic HDM + DEP model in which neutrophils are more prominent (16). Since sivelestat has poor pharmacokinetics, alternative delivery methods could also be considered. Nanocarrier delivery of sivelestat to neutrophils can improve biodistribution and thus its efficacy. For example, in an LPS- mouse model, free sivelestat was not effective, however, vesicles incorporating sivelestat were successfully taken up by neutrophils and prevented NET formation leading to less pulmonary inflammation (39). Still, it should be pointed out that despite potential advantages of these strategies in the treatment of certain NET-mediated diseases, systemic inhibition of NETosis in animal models increases the susceptibility towards infections (39). Sivelestat is clinically available in Japan and South Korea for acute lung injury (ALI), however, efforts to use sivelestat in other countries have failed since a multinational clinical trial on ALI patients was unsuccessful (40). Other challenges in the use of NE inhibitors include the fact that NE bound to extracellular DNA in NETs is resistant to the activity of inhibitors (39). Moreover, sivelestat functions extracellularly thus only inhibiting NE released into the extracellular space (39, 41). These challenges could explain the partial decrease of BAL inflammation in our asthma mouse model.