4. Discussion

Our study shows that dermal contact with CoCl2 followed by repeated airway exposure to CoCl2 leads to several hallmarks of asthma, such as airway hyper-reactivity (AHR) and lung inflammation, with characteristics of the type 2 adaptive immune response, in combination with increased cytokines linked to the innate immune system. Moreover, for the first time we show that in an asthma model, early changes in the lungs can be identified by µCT imaging in free-breathing mice.
While the link between skin sensitization and lung susceptibility has been demonstrated for several chemicals, including diisocyanates, acid anhydrides and persulfate salts [8–10], this has never been investigated for metals with sensitizing properties. Cobalt-induced asthma has been frequently reported in occupations involving inhalation of cobalt-containing materials, such as hard metal (a composite consisting mainly of tungsten carbide with cobalt) [14–16]. Previous studies of the respiratory toxicity of cobalt have focused on histological or immunological changes [17–19]. However, to our knowledge, no study has investigated the possible role of skin sensitization in the subsequent development of cobalt asthma. .
In this study, we included an imaging modality for the first time to investigate whether alterations can be identified in a mouse model of asthma. The advantage of µCT is that it can provide high-resolution information in living animals, repeatedly, without causing harm to the animal, and at relatively low cost. We have previously shown, in mouse models of lung emphysema (elastase) and fibrosis (bleomycin) [20, 21], that early alterations can be detected by µCT. Moreover, a low-dose scanning protocol was recently developed, allowing radio-safe repeated follow-up of disease progression [22]. Our data showed larger total and aerated lung volumes when mice received CoCl2 respiratory challenge after skin exposure with cobalt. We interpret these changes as reflecting early onset lung alterations associated with inflammation.
Increased airway volume and decreased lung density were observed in both groups receiving CoCl2 via the airways. This image pattern, also named paradoxical airway dilation, has been described in several animal models undergoing bronchoconstriction interventions, such as mice challenged with methacholine [23] or treated with intraperitoneal naphthalene [24]. A recent study of hamsters infected by SARS-CoV-2 also indicated a positive correlation between bronchial dilation and apoptotic bodies in the bronchial walls [25]. Since we found a correlation between µCT-derived airway enlargement and BAL airway inflammation (Figure 4A and Table S2), airway dilation could possible serve as a marker of lung inflammation.
One of the hallmarks of (occupational) asthma is non-specific AHR. Epidemiological studies show that cobalt may induce or exacerbate AHR in humans [26, 27]. In our animal model, mice having received both prior skin exposure and respiratory challenge (the Co/Co-group) responded more strongly to a methacholine provocation than the other mice. Thus, the Co/Co-group had a more pronounced decrease in FEV0.1 under methacholine challenge compared with all other groups, which is in agreement with the obstructive pattern of asthmatic murine models [9, 28, 29]. To our knowledge, this is the first time that FEV0.1 is assessed with methacholine challenge in an occupational asthma model. In the future, this may lead to more concordance in assessing lung function of mice in comparison to human.
Consistent with the increases in airway volume and AHR, mice of the Co/Co-group exhibited an influx of neutrophils and eosinophils in BAL. In contrast, mice receiving CoCl2 respiratory challenge without prior skin exposure (Veh/Co) showed only a minor neutrophilic influx, indicating that the mixed neutrophilic/eosinophilic inflammatory response in the Co/Co-group was due to the initial dermal sensitization. Admittedly, a few mice in the Veh/Co-group responded partly like the Co/Co-group, and this may have been due to the repeated instillations of Co.
In the present study, the total dosage of the oropharyngeal challenge (6 μg Co2+ given in five doses) was much lower than that in previous animal models (13 μg Co2+ for 10 days) [30], but features of a type 2 allergic inflammation still occurred, presumably due to the initial dermal exposure. The eosinophilic inflammatory response in BAL of dermally CoCl2 treated and CoCl2 challenged group has never been reported in animal models. The eosinophilic inflammation produced only in animals with prior dermal contact, even for a short time, strengthens the conclusion that dermal sensitization increases the airway responsiveness to the corresponding chemical allergen. We have argued previously that this general concept is not only of academic importance in terms of immunological mechanisms, but it has also relevance in the field of occupational hygiene and the prevention of occupational asthma [7].
Although we observed a mixed neutrophilic/eosinophilic lung inflammation, we found that only cytokines linked to neutrophilic inflammation (KC and MCP-1) were increased in the Co/Co-group at the time of autopsy. Several cytokines have previously been assessed in studies of cobalt-induced respiratory toxicity, but the results varied depending on the challenging dosage and frequency. KC, also known as CXCL1 or GROα, is a chemokine critical to neutrophil recruitment. This is the first observation in an animal asthma model that BAL KC is associated with prior skin exposure. While we did not find changes in IL-4, IL-5, or IL-13 in BAL for all groups, which are linked to type 2 eosinophilic inflammation and airway hyper-reactivity, we found higher MCP-1 in both groups receiving respiratory cobalt challenge. MCP-1, also known as CCL2, is a cytokine tightly associated with Th2 reaction. MCP-1 can induce mast cell degranulation, thereby enhancing AHR [31]. The upregulation of MCP-1 has also been frequently observed in animal models with skin exposure followed by challenge [7]. The observation in this study may imply a more important role of MCP-1 in the Th2 reaction compared with IL4, IL-5 and IL-13. Unexpectedly, the level of MIP-2 was lower in the Veh/Co-group. MIP-2 (also known as CXCL2) is critical to neutrophil attraction and secreted mainly by macrophages [32]. Though the reason for decreased MIP-2 in the Veh/Co-group remains to be verified, a possible explanation is that the immune reaction shifted away from Th1-macrophage axis, thereby leading to downregulated macrophage and MIP-2 downstream.
Dendritic cells and innate lymphoid cells have been investigated extensively in ovalbumin and house dust mite asthma models [33, 34], yet this has never been done in mouse models of asthma caused by chemical or metal exposure. The analysis of DCs in lung revealed elevated cDC1 in both groups having received cobalt skin exposure. On the other hand, pulmonary cDC2, moDC and total dendritic cells increased only in the Co/Co-group. cDC1 has been demonstrated to modulate Th17 immune responses [35], though its full role to enhance Th2 reaction remains controversial. cDC2s and moDCs are known to facilitate Th2 and Th17 cell differentiation [36, 37], which is also in accordance with the cytokine pattern from our study. Though some studies have evaluated the shifting of pulmonary DC subtypes after respiratory exposure to sensitizing agents, the shifting after solely skin exposure has never been reported. Accumulation of cDC2 and moDC in mediastinal lymph node has been revealed in mice receiving solely respiratory exposure to HDM [37]. Our team has previously examined pulmonary DC subtypes in mice, and observed elevated moDC after sequential skin and airway exposure to TDI [38], but the effect of exclusive skin exposure was not assessed.
As is the case for dendritic cells, the innate lymphoid cells have not been investigated in a skin/lung model. In this study, pulmonary ILC2 numbers increased in both groups receiving skin exposure, but not in the group receiving solely respiratory challenge. This finding might explain the presence of the eosinophilic inflammation in BAL from the Co/Co-group without an elevation of Th2-related cytokines, such as IL-5, and imply an association between skin exposure and ILC2/Th2/eosinophil axis. The role of ILC3 in asthma has rarely been studied in animal models compared with ILC2. One ILC3 subtype, NCR-ILC3, is known to produce IL-17 [39]. In our study, respiratory CoCl2 challenges irrespective of prior dermal contact showed a trend to increase the level of NCR- ILC3 in CD45+ pulmonary cells, admittedly not at a significant level. Only in the group receiving both skin and airway exposure (Co/Co), did NCR- ILC3 increase significantly. As shown in Figure 4B, the neutrophil proportion in BAL was positively associated with NCR-ILC3. Together with the observed inflammatory patterns in BAL and lung, the results imply that the respiratory exposure to cobalt is associated with ILC3/Th17/neutrophil axis. In spite of the mixed type 2/type 3 inflammatory reaction in airway, other related cytokines in BAL did not increase altogether. We did not observe increasing IL4/IL5/IL13, which are related to ILC2/Th2/eosinophil reaction. Likewise, IL-17A and IL-17F did not increase despite the elevation of KC and prominent neutrophilic reaction in BAL. Other cytokines associated with asthma, such as IL-22, IL-25, IL-33 might also serve as potential influential factors [40, 41].
Though the mice in the Co/Co-group presented a more marked inflammation than those in the other groups, the intensity of the inflammatory responses varied considerably between animals and some mice did not show inflammation at all. We hypothesize that these “non-responding” mice had not been successfully sensitized via the skin, possibly because the cobalt solution (in DMSO) did not always persist long enough on the ears (in our previous studies we always used a mixture of acetone and olive oil to dissolve the test chemicals, but CoCl2 did not dissolve in this vehicle). This may also explain why we failed to observe significant cell proliferation of the auricular lymph nodes with concordant cytokine release, and consistent increases in total serum IgE in the mice dermally sensitized and pulmonary challenged with CoCl2. The absence of these responses contrasts with our previous extensive experience with chemical sensitizers in this skin/lung model, since we always observed pronounced signs of skin and systemic sensitization after applications of TDI, trimellitic anhydride or ammonium persulfate [8–10]. The possibility that some animals were in fact not dermally sensitized to Co is borne out by our correlation analysis showing that in spite of the relatively low average changes at group level, T helper cells (CD3+CD4+) in the auricular lymph nodes and serum IgE were positively associated with numbers of cDC1 and ILC2 cells in the lung and other immunological and physiological responses in individual animals (Table 1, Figure 4C-D). The intra-individual consistency between the various responses strengthens our conclusion that dermal sensitization to cobalt enhanced the susceptibility of the airways to subsequent challenges with cobalt.
In conclusion, prior skin exposure to cobalt modulates immune cells in lung, induces airway hyper-reactivity and a mixed neutrophilic/eosinophilic inflammation upon respiratory challenge with Co, which is accompanied by dendritic cells and innate lymphoid cells of the type 2/type 3. This observation with cobalt strengthens the evidence that asthma caused by low-molecular weight agents is not only related to sensitizers reaching the lungs via airway exposure, but also to chemicals sensitizing the body via skin exposure, which increases the susceptibility of the lungs to inhaling sensitizers.