Balb/c mice were obtained from the Jackson laboratories (Bar Harbor, ME) and bred in-house. All experiments were performed on 8–10 weeks old mice. Experiments where mice were sensitized and challenged with Af and exposed to air or O₃ were carried out as previously described (30, 39, 40). In brief, mice were sensitized with 20 μg Af and alum by intraperitoneal injection (i.p.) on days 0 and 7, then challenged with 25 μg Af by intranasal (i.n.) instillation on day 13. In Figure 1, mice were treated with vehicle (Dimethyl sulfoxide, DMSO) or budesonide (0.25 or 2.5 mg/kg) i.n. at the time of Af challenge. 48 h post challenge, lung function (enhanced pause, Penh) was measured using the Buxco^® system.

In Figures 2–5, mice followed the Af sensitization and challenge protocol as described, however 84 h post challenge/budesonide they were exposed to 3 ppm O₃ or air for 2 h. Animals were studied 96 h post Af challenge (12 h post O₃). These time points were selected to mimic O₃-induced exacerbation of allergic changes, because by 96 h post Af challenge airway inflammation subsides while O₃ exposure induced inflammation peaks 12 h post exposure (Figures 2A,B) (33, 40). That a 3 ppm inhaled dose in rodents results in O₃ concentration in the lungs relevant to human exposure levels has been experimentally validated by others, using oxygen-18-labeled O₃ (¹⁸O₃). Hatch et al. showed that exposure to ¹⁸O₃ (0.4 ppm for 2 h) caused 4–5-fold higher ¹⁸O₃ concentrations in humans than in rats, in all of the BAL constituents measured (41). Rats exposed to 2.0 ppm, had still less ¹⁸O₃ in BAL than humans exposed to 0.4 ppm. The species discrepancies between the recoverable O₃ levels in the lung are not entirely clear. It is thought however that as rodents are obligate nose breathers (while humans breathe through their nose and mouth), this reduces the delivered dose of O₃ to the lungs of rodents. Further, Slade et al. found that after exposure to O₃, mice react by a rapid decrease of core temperature, a species and strain specific characteristics (42). The recoverable ¹⁸O₃ in the lung tissue was negatively associated with the extent of hypothermia that significantly altered O₂ consumption and pulmonary ventilation, explaining at least partly, the interspecies differences seen in O₃ susceptibility. In addition, in pilot studies we also performed a careful assessment of the biological effects on a range of 0.5–6.0 ppm O₃ exposure. Doses lower than 3 ppm did not elicit a significant inflammatory response that would be commensurate with what is seen in humans, in regards to BAL or peripheral blood neutrophilia, upon O₃ inhalation for 2 h. Higher than 3 ppm doses caused observable respiratory distress especially in Balb/c mice. The O₃ dose we used here therefore represents a level of exposure that is well-tolerated by both Balb/c and C56BL/6 mice and that causes a significant airway inflammatory response.

Lung function was measured using the Flexivent^® system (Scireq, Montreal, Canada) in response to increasing concentrations of inhaled methacholine. BAL and lung cells were harvested to study inflammatory cells by flow cytometry. Following collection of BAL, 10 mL ice-cold PBS was injected into the right ventricle to perfuse the lung. The lung lobes were then carefully removed and snipped into small pieces before undergoing digestion with Liberase TL (Millipore Sigma, Burlington, MA) for 40 min at 37°C on a shaker. Digested whole lung homogenate was filtered through a 70 μm cell strainer to create a single cell suspension for flow cytometric analysis. In cell-free BAL supernatant, a custom mouse magnetic Luminex assay was utilized to study cytokines and chemokines while SP-D was measured by western blot, native gel electrophoresis (structure) or sandwich ELISA (quantity). All mouse procedures were reviewed and approved by the University of California, Davis, and University of Pennsylvania Institutional Animal Care and Use Committees.

BAL and lung cells were harvested and single cell suspensions were prepared as previously described for analysis by flow cytometry (40). Fluorescently-conjugated monoclonal antibodies were purchased from Biolegend (San Diego, CA), BD Biosciences (San Jose, CA), or eBioscience (San Diego, CA). Single cell suspensions were incubated with antibodies targeting surface markers for 20 min at 4 degrees C in the dark. In the BAL samples, the following antibodies were used: APC-Cy7-CD11c, PE-Siglec F, PE-Cy7-CD11b, PerCP-Cy5.5-Ly6G. In the lung digest suspensions, neutrophils, and eosinophils were identified using APC-Cy7-CD11c, PE-Siglec F, and PerCP-Cy5.5-Ly6G. Live/dead Aqua was used in the panels throughout the study to exclude dead cells. Flow cytometry was performed on a Fortessa (BD Biosciences, San Jose, CA) and data was analyzed using FlowJo software (Ashland, OR).

Cytokines and chemokines were assayed in the BAL via a custom Magnetic Mouse Luminex Assay (R&D System, Minneapolis, MN). C-C motif chemokine 11 (CCL11), Interleukin-23p19 (IL-23p19), IL-13, IL-6, Chemokine (CXC motif) ligand 2 (CXCL2), and CCL20 were measured in the premixed kit. BAL fluid was first concentrated using a 2 mL, 3 k Amicon Ultra Centrifugal Filter (Millipore Sigma, Burlington, MA) spun at 3,000 g for 30 min. The kit was performed following the instructions from the manufacturer.

Total protein concentration was measured by the BCA Assay (Thermo Fisher Scientific, Waltham, MA). BAL SP-D was assayed by sandwich ELISA using our in-house generated monoclonal and polyclonal antibodies as previously described (33). BAL SP-D was also measured by native gel electrophoresis (33) to assess the tertiary structure of SP-D, which is critical to maintain its anti-inflammatory functions (43, 44). Proteins were transferred to a nitrocellulose membrane (Thermo Fisher Scientific, Waltham, MA). Goat anti-mouse SP-D (1:2,000, R&D Systems, Minneapolis, MN) was the primary antibody while donkey anti-goat antibody coupled to horseradish peroxidase (1:10,000 GE Healthcare Life Sciences, Marlborough, MA) was the secondary antibody. SP-D signal was detected using the ECL Western Blotting Substrate (Thermo Fisher Scientific, Waltham, MA) on film (ECL Hyperfilm, GE Healthcare Life Sciences, Marlborough, MA). Image J (National Institutes of Health, Rockville, MD) analysis was used to determine the optical density of SP-D bands.

Human primary type II airway epithelial (hAECII) cells were acquired from normal human lung tissues from NDRI (National Disease Research Interchange). A549 cells were purchased from ATCC (Manassas, Virginia). Dexamethasone, budesonide, curcurbitacin I (Cu I), and RU486 was purchased from Millipore Sigma (Burlington, MA). A549 cells are a human type II alveolar epithelial cell line used by our laboratory and others (45–47) to model functions including expression of mRNA for SP-D. We used these readily available cells to establish conditions of SP-D mRNA expression upon treatment with ozone and budesonide (Figure 6). The budesonide effects were then recapitulated in primary hAECII cells (Figure 7). A549 cells were cultured in DMEM supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin (Thermo Fisher Scientific, Waltham, MA). Primary hAECII cells were cultured in DMEM-H21 plus F-12 Ham's (1:1) supplied with 5% fetal bovine serum, 100 U/ml penicillin, 0.1 mg/ml streptomycin, 2 mM L-glutamine. D-valine (Invitrogen) prevented growth of fibroblasts. hAECII cells were treated with budesonide/dexamethasone for 2 h, with or without RU486 and Cu I added to the culture. A549 cells were treated with budesonide and exposed to air or O₃ (300 ppb) for 2 h, which was generated as previously described (31). DMSO was used as the vehicle control in the in vitro studies. At the time points indicated in the figure, cells were harvested for analysis of sftpd mRNA (qPCR) or SP-D protein (western blot). Cells were harvested in TRIzol reagent (Thermo Fisher Scientific, Waltham, MA) for mRNA analysis by qPCR or RIPA buffer (Thermo Fisher Scientific, Waltham, MA) for protein analysis by western blot.

For the western blots, total intracellular protein was measured by the BCA assay, then 20 μg protein was loaded for each lane. The primary antibody was goat anti-SP-D (1:500). The secondary antibody was HRP anti-goat IgG (1:1,000). For control antibodies, the primary was rabbit anti-GAPDH (1:1,000) and the secondary was HRP anti-rabbit (1:5,000). All antibodies were purchased from Santa Cruz Biotechnology (Dallas, TX). Image J (National Institutes of Health, Rockville, MD) analysis was used to determine the optical density of SP-D bands. For the qPCR, RNA was extracted from the TRIzol by chloroform layering and isopropanol precipitation, then reverse transcribed into cDNA via the QuantiTect Reverse Transcription Kit (Qiagen, Hilden, Germany). qPCR was performed on the recovered cDNA using SYBR green reagents (Applied Biosystems, Foster City, CA) on a ViiA 7 Real-Time PCR system (Thermo Fisher Scientific, Waltham, MA). Fold change was calculated using the ΔΔCt method, first normalizing values to GAPDH. Human SP-D primer with sequence of 5′-ACACAGGCTGGTGGACAG-3′ (sense); 5′-TGTTGCAAGGCGGCATT-3′ (anti-sense) were used to produce 61 bp products.

All statistics were performed using Prism v7 software (GraphPad Inc., La Jolla, CA). Data are expressed as mean ± SEM and are representative of at least 2 independent experiments. A Student's t-test was used to compare vehicle vs. budesonide or air vs. O₃. A One-way ANOVA with Tukey's multiple comparison's test was used in the budesonide dose response experiment (Figure 1). A Two-way ANOVA with Tukey's multiple comparison's test was used when comparing all experimental groups. A p < 0.05 was considered statistically significant.

Inhaled glucocorticoids improve lung function and airway inflammation in allergic asthma (1, 48, 49) but their effectiveness in acute asthma exacerbations are subject of on-going investigations (49–51). We developed a model in which mice were sensitized (i.p.) and challenged with Af at the time of budesonide administration (i.n.) (Figure 1A). To establish the dose-dependent effects of budesonide on methacholine responsiveness we used Penh (enhanced pause), a non-invasive measure of airway obstruction, because it enabled us to obtain data from multiple individual animals simultaneously and complete a study using a large number of mice. These results showed that sensitization and challenge to Af significantly increased baseline Penh and methacholine responsiveness and that budesonide significantly inhibited airway hyperreactivity to methacholine at 2.5 mg/kg dose (Figure 1B). For the subsequent experiments presented in this paper we used this budesonide dose and confirmed its inhibitory effects on allergic airway hyperreactivity by the invasive FlexiVent^® system (Figures 2C,F, 3D). These data also provided the basis for the subsequent studies on the effects of O₃ on allergic airway inflammation and glucocorticoid responsiveness.

To establish the time course of O₃ induced airway inflammation, Balb/c mice were exposed to air or 3 ppm O₃ for 2 h and studied at several time points afterwards (Figure 2A). Neutrophil influx into the airways peaked 12 h after O₃ exposure (Figure 2B). O₃ exposed mice and air exposed controls were studied for methacholine responsiveness at the 12 h time point. O₃ induced a significant increase in methacholine responsiveness compared with air exposed controls (p < 0.01) (Figure 2C). To investigate the effects of O₃ on allergic airway inflammation Balb/c mice were sensitized and challenged with Af. In this model the acute inflammatory changes resolve by 96 h after allergen challenge. We therefore studied the mice at this time point, but we also exposed them to either air or O₃ 12 h before (Figure 2D). BAL neutrophils and eosinophils were quantitated by FACS analysis (the gating strategy is shown in Figure 3B). O₃ exposure clearly enhanced numbers of eosinophils and neutrophils in the airways of Af sensitized and challenged mice in comparison with air exposure (p < 0.01; Figure 2E). In addition, lung resistance to methacholine challenge was also significantly amplified in the O₃ exposed animals (p < 0.001; Figure 2F). These results confirm our previous findings (39) and strongly indicate that O₃ induces airway hyperreactivity on its own and that it enhances airway changes in allergen sensitized and challenged animals.

Experiments in dogs (19) our previous studies in healthy volunteers (52) and investigations in mild asthmatics (53) showed that glucocorticoid treatment inhibited O₃-induced inflammation in the airways. However, how would O₃ alter the inhibitory effects of budesonide on allergic airway inflammation and hyperreactivity has not been documented.

To study the hypothesis that O₃ impairs the anti-inflammatory effects of budesonide, Balb/c mice were sensitized and challenged with Af as described, and were intranasally treated with 2.5 mg/kg budesonide or vehicle. 82 h post-Af challenge mice were exposed to air or 3 ppm O₃ for 2 h, then 12 h later (96 h post-Af challenge), lung function was measured (Flexivent^®), and BAL and lungs were harvested (Figure 3A). BAL eosinophils (live Siglec-F⁺CD11c⁻ cells) and neutrophils (live Ly6G⁺CD11b⁺ cells) and lung eosinophils (live CD45⁺CD11c⁻Siglec-F⁺) and neutrophils (live CD45⁺CD11c⁻Ly6G⁺) were analyzed by FACS from single cell suspensions as shown in Figure 3B. Budesonide significantly suppressed eosinophil (not neutrophil) numbers both in the BAL and the lung in air exposed but not in O₃ exposed mice (Figure 3C). Strikingly, inhibition of lung resistance (upper panel) and tissue damping (lower panel) by budesonide seen in air exposed mice (gray plain squares) was completely abolished in O₃-exposed mice (gray hatched squares, p < 0.001, Figure 3D). These data indicated that the inhibitory effects of budesonide on Af-induced airway inflammation were attenuated and on airway hyperreactivity were completely abolished by O₃ exposure in sensitized and challenged mice.

Since O₃ mitigated the inhibitory effect of budesonide on airway eosinophilia and airway hyperreactivity, we wanted to investigate the underlying mediator profile. We measured BAL CCL11 (eotaxin, an eosinophil chemoattractant), IL-23p19, IL-6, CXCL2 (pro-neutrophilic mediators), CCL20 (a lymphocyte chemoattractant), and IL-13 (known to prime smooth muscle cells for airway hyperreactivity). O₃ strongly induced BAL IL-6, CXCL2, and CCL20 in a budesonide-independent manner (Figure 4). Meanwhile, budesonide significantly reduced BAL CCL11 showed a trend for reduction of IL-13 (p = 0.07) and IL-23p19 (p = 0.05) in the BAL of air exposed, but not O₃ exposed mice (Figure 4, lower panels). These data suggested that O₃ induced pro-neutrophilic factors regardless of budesonide treatment, and that the suppressive effects of budesonide on eosinophilia and airway hyperreactivity-inducing factors was attenuated by O_3.

We and others previously showed that SP-D plays an important role in suppressing proinflammatory mediator release in allergen or O₃-induced airway inflammation (28–30, 54) and that production of SP-D required the presence of glucocorticoids in airway epithelium (34–37). Further, we found that O₃-induced airway inflammation in allergen challenged mice resulted in abnormal oligomeric molecular forms of SP-D indicating that oxidative damage can cause conformational changes with a potential inactivation of SP-D's immunoprotective function (28, 32, 33). Here we wanted to investigate how the combination of allergen with O₃ exposure would alter the glucocorticoid effects on SP-D expression.

Assessment of total BAL protein levels showed that those were returned to normal 96 h after Af challenge, indicating inflammatory resolution in air exposed mice. In O₃ exposed mice however, BAL protein levels were significantly elevated indicating acute inflammation that was not prevented by budesonide treatment (Figure 5A). As expected on the basis of previous investigations (34–37) budesonide significantly increased BAL SP-D in air exposed mice. Importantly, this budesonide effect on SP-D expression was lost in O₃ exposed mice (Figure 5B). By native gel electrophoresis, structurally intact SP-D was found at the top of the gel and did not separate from the well, while de-oligomerized SP-D was resolved as a smear (Figure 5C). Native SP-D density was not statistically different between the groups studied (Figure 5C). O₃ caused de-oligomerization of SP-D in the BAL of mice sensitized and challenged with Af. This change was prevented by budesonide treatment (gray hatched bar, Figure 5D). Our data suggested that budesonide induction of SP-D is inhibited by O₃. We speculate that in budesonide treated mice SP-D was indirectly protected from de-oligomerization possibly as a result of inhibition of eosinophils (the main source of iNOS and nitric oxide) in the lungs of mice (Figure 3C).

We previously showed that IL-6 directly induced SP-D in type II alveolar epithelial cell cultures (28). This is interesting in light of O₃ while strongly inducing IL-6 (Figure 4), did not increase SP-D 12 h after exposure, but in fact it prevented the stimulatory effects of budesonide on SP-D release in the airways of mice (Figures 5A,B). To better understand the mechanisms of how budesonide and O₃ regulate SP-D expression we used A549 cells, a readily available human type II alveolar epithelial cell line that models functions such as expression of the SP-D gene (sftpd, Figure 6A). To confirm our findings, the budesonide effects were then recapitulated in primary human type II alveolar epithelial cells (hAECII, Figure 7). O₃ exposure of A549 cells inhibited sftpd expression 1.5 h later, but by 48 h post exposure this effect was reversed into an induction of the sftpd gene (Figure 6B). Budesonide induced sftpd mRNA in A549 cells exposed to air. O₃ completely prevented the budesonide induction of the sftpd gene 1.5 h later. However, by 48 h O₃ and budesonide synergistically increased sftpd mRNA (Figure 6C). These results are in line with our previous in vivo study on Balb/c mice (28) and suggest that O₃ acts on sftpd transcription in a bi-phasic manner with an early phase inhibition (<12 h) and a late phase activation >48 h. Based on these and our previous findings on IL-6 we speculated that budesonide and O₃ may interact on a common signaling pathway involved in SP-D transcription in airway epithelial cells.

The proximal promoter region of the human SP-D gene (sftpd) has binding elements for C/EBP, NFAT, AP1, HNF-3, and STAT3/6 that all contribute to transcription of SP-D (Figure 7A) (55). Dexamethasone induced lung SP-D in mice at the level of transcription in the absence of a full glucocorticoid response element in the proximal promoter region of sftpd. Zhang and colleagues previously reported that STAT3 can act as a co-activator in glucocorticoid receptor signaling (56). To test if the glucocorticoid receptor works in concert with the STAT3/6 binding element to induce SP-D we studied primary hAECII cells using a specific inhibitor of STAT3 (cucurbitacin Cu I) (57) and the glucocorticoid receptor (RU486). We treated human primary type II alveolar epithelial (hAECII) cells with budesonide and dexamethasone in vitro and studied SP-D mRNA (qPCR for sftpd) and protein (western blot for SP-D) expression (Figure 7B). We used dexamethasone in the in vitro experiments as a positive control because it is a well-characterized glucocorticoid that induces SP-D (35). Indeed, dexamethasone induced sftpd expression in human primary type II aleveolar epithelial cells that was abolished in the presence of Cu I or RU486 (Figure 7C). Similarly, Cu I and RU486 inhibited dexamethasone and budesonide-induced SP-D protein in hAECII cells (Figures 7D–F). Optical density analysis by Image J analysis confirmed that antagonism of the glucocorticoid receptor or STAT3 impaired dexamethasone-induced SP-D (Figures 7D,E) and that budesonide induced SP-D in a glucocorticoid receptor dependent manner (Figure 7F). These data suggested that glucocorticoid-induced SP-D synthesis was dependent on glucocorticoid receptor and STAT3 activation.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.