Postnatal Ozone Exposure Disrupts Alveolar Development, Exaggerates Mucoinflammatory Responses, and Suppresses Bacterial Clearance in Developing Scnn1b-Tg⁺ Mice Lungs

Ozone, a highly reactive oxidant gas, is one of the common environmental air pollutants (1). Ozone reduces lung function (2, 3), increases risks for asthma (4–6), and exacerbates pulmonary symptoms in mucoinflammatory diseases, including cystic fibrosis (CF) (7, 8), chronic obstructive pulmonary disease (9), and asthma (10). Increased ambient ozone levels pose serious health hazards, especially for young children who spend a considerable amount of time playing outdoors (11). It has been reported that young children, with their lung development still underway, are more vulnerable to environmental toxicants because of their higher respiratory rates that allow them to rapidly inhale toxicants (12, 13). However, the superimposed effect of ozone on the progression of impending mucoinflammatory lung disease in developing lungs has never been tested, thus further experimentation is warranted.

CF is a chronic, progressive respiratory disease that primarily affects the respiratory system in children and young adults. Although a number of mouse models with CFTR mutation have been developed, none of them recapitulate features of human CF-like lung disease (14). This is likely due to the dominance of calcium-activated chloride channels in murine airways (15) that compensate for the ionic imbalance caused by CFTR dysfunction or deletion. The sodium channel, non–voltage-gated 1, β subunit transgenic (Scnn1b-Tg⁺) mouse, a well-accepted model for human CF-like lung disease, overexpresses a transgene encoding Scnn1b in the club cells (16). As a result of airway epithelial-specific overexpression of the Scnn1b transgene and its protein product (i.e., β subunit, epithelial sodium ion [Na⁺] channel [βENaC]), the absorption of Na⁺ is increased in the epithelial cells. Due to increased ionic concentration of Na⁺ in the epithelial cells, the buildup of osmotic gradient drives the movement of water from the airway surface liquid (ASL) layer into the epithelial cells, which leads to ASL layer dehydration (16, 17). As a result, within the first week of postnatal life, the Scnn1b-Tg⁺ mice develop human CF-like lung disease with features such as mucostasis, mucus obstruction, airway inflammation, mucous cell metaplasia (MCM), and poor mucociliary clearance (17). As a result of poor mucociliary clearance, the Scnn1b-Tg⁺ mice also manifest spontaneous bacterial infections and impaired bacterial clearance (18, 19). The airway inflammation in Scnn1b-Tg⁺ mice is characterized by the presence of increased numbers of activated macrophages, neutrophils, eosinophils, and lymphocytes in the airspaces (16, 17, 19). Given the early initiation of mucoinflammatory pathology in the Scnn1b-Tg⁺ mouse and the close resemblance of this pathology to human-CF-like disease, the Scnn1b-Tg⁺ mouse can be used to study the initiation and progression of pediatric lung diseases.

Several studies done in the past have investigated the effects of environmental toxicants including nanoparticles, fungal spores, mainstream tobacco smoke, and secondhand smoke in Scnn1b-Tg⁺ mice (20–24). Three epidemiological studies report that exposure to ozone increases the risk of pulmonary exacerbations in patients with CF (7, 8, 25), but, to the best of our knowledge, no studies have been conducted to examine the effects of ozone exposure on the initiation and progression of airway disease in Scnn1b-Tg⁺ mice. Thus, the effect of ozone on key characteristics of Scnn1b-Tg⁺ mice (i.e., airspace inflammation, spontaneous bacterial infection, mucus hypersecretion, and MCM) remains unexplored and warrants investigation, especially in neonates in which the lung is in the developing stages.

To study the interactions between ozone exposure and mucoinflammatory disease in Scnn1b-Tg⁺ mice, we hypothesized that ozone exposure from postnatal days (PND) 3–20 will perturb normal lung homeostasis in wild-type (WT) neonates and exaggerate Scnn1b-Tg⁺ lung pathology. Accordingly, we exposed Scnn1b-Tg⁺ and WT littermates to ozone from PND3–20 and examined lung phenotypes at the weaning age of PND21. Because Scnn1b-Tg⁺ mice exhibit airway inflammation (16, 17), bacterial infections (18, 19), and profound type 2 immune response (17), we measured these attributes by assessing bronchoalveolar lavage fluid (BALF) cellularity, cytokine levels, bacterial CFU counts in BALF, mucus obstruction, MCM, mucin expression, and gene expression relevant to type 2 mucoinflammatory responses. The results from these studies provide novel insights into the effects of early postnatal ozone exposure on lung development and the manifestation of responses associated with muco-obstructive lung diseases.

Scnn1b-Tg⁺ mice [Tg (Scgb1a1-Scnn1b)6608Bouc/J] were procured from The Jackson Laboratory (Bar Harbor, ME) on a congenic C57BL/6J background and maintained on this background by selective breeding between WT (Scnn1b-Tg⁻) and Scnn1b-Tg⁺ mice. All of the mice used in this study were genotyped for the Scnn1b transgene using PCR as reported previously (24). Mice were maintained in hot-washed, individually ventilated cages on a 12-h dark/light cycle at the Division of Laboratory Animal Medicine at Louisiana State University (Baton Rouge, LA). Mice were supplied with food and water ad libitum except during filtered air (FA) or ozone exposure (4 h/d). All animal use and inhalation procedures were approved by the Institutional Animal Care and Use Committee of the Louisiana State University.

Three-day-old neonates (entire litter) were transferred to individual cages with perforated lids and exposed to either FA or 800 parts per billion (ppb) ozone (798 ± 8.561 [SEM] ppb) for 4 h. Following 4 h of exposure each evening to either FA or ozone, mice were returned to their parental cages. Ozone concentration was monitored during the entire experimental period (PND3–20) of 4 h/d exposure. Mice were exposed in the nightly conditions (high physical activity) (26, 27), which simulates real-life scenarios of the higher activity phase in humans, as detailed previously (28). Loading of animals into the light-protected chambers was coordinated with the start of the night cycle at the vivarium. The timing of exposures was maintained strictly and consistently for all of the exposures that lasted for 18 consecutive days.

FA- or ozone-exposed mice (PND21) were anesthetized via i.p. injection of 2,2,2-tribromoethanol (Millipore Sigma, Burlington, MA), and BALF was harvested aseptically as described previously (18). A total of 60 µl BALF fluid was taken for CFU enumeration, and the remaining BALF was centrifuged in a cooling centrifuge at 500 × g for 5 min. Cell-free BALF supernatant was collected and stored at −80°C for cytokine and Ig analyses. Cell pellets were resuspended in 250 µl PBS. A total of 50 µl cell suspension was used for total cell counts using a hemocytometer (Bright-Line; Hausser Scientific, Horsham, PA), and 200 µl was used to make cytospins followed by differential staining (Modified Giemsa Kit; Newcomer Supply, Middleton, WI). Cytospins were analyzed for the determination of macrophage surface area. Briefly, photographs were captured under ×20 objective of the ECLIPSE Ci-L microscope with DS-Fi2 camera attachment (Nikon, Melville, NY). Thereafter, captured images were processed using ImageJ software (National Institutes of Health) to determine the surface area of the macrophages (29). Unlavaged left lung lobes were stored in 10% neutral buffered formalin, and lavaged right lung lobes were stored at −80°C for gene expression and cytokine analyses.

Cell-free BALF was assayed for keratinocyte chemokine (KC; CXCL1), MIP-2 (CXCL2), G-CSF, MCP-1 (CCL2), MIP-1α (CCL3), MIP-1β (CCL4), IL-4, IL-5, IL-6, IL-13, IL-17, IP-10, IL-1α, IL-1β, and IL-10 using Luminex-xMAP–based assay (MCYTOMAG-70K), according to the manufacturer’s instructions (EMD Millipore, Billerica, MA). To assess the levels of eosinophilic inflammation relevant cytokines (i.e., Eotaxin, CCL7, CCL12, CCL17, CCL22, and CCL24) (Bio-Plex Pro Mouse Chemokine Panel; Bio-Rad, Hercules, CA), we used lung homogenates because of the unavailability of BALF sample. The data were normalized to the total protein concentration. Total protein contents in the BALF were determined by Bradford assay (Bio-Rad). Total dsDNA contents were determined spectrophotometrically using NanoDrop 8000 (Thermo Fisher Scientific, Waltham, MA). Ig concentrations in the cell-free BALF were determined with the MILLIPLEX MAP mouse Ig isotyping magnetic bead panel isotyping multiplex assay (MGAMMAG-300K), according to the manufacturer’s instructions (EMD Millipore).

To estimate the bacterial burden in the BALF, aseptically harvested cell-free BALF (60 µl) was serially diluted onto Columbia blood agar plates (Hardy Diagnostics, Santa Maria, CA) and incubated in anaerobic candle jars at 37°C as previously described (24). CFUs were counted at 48 h postincubation, and physical characteristics of colonies, including size, shape, color, margins, and elevation, were recorded.

Unlavaged left lung lobes were stored in 10% neutral buffered formalin for fixation. Subsequently, formalin-fixed tissues were paraffin-embedded, sectioned at 5 μm, and processed for histopathological analyses. H&E staining was performed to access the structural and morphological alterations in the left lung. Alcian blue–periodic acid-Schiff (AB-PAS) staining was performed to examine the airway mucus contents and the mucous cells. The slides were graded by a board-certified anatomic pathologist in a blinded manner using the histological semiquantitative grading strategy, as previously described (30, 31).

Immunohistochemical staining for major basic protein (MBP), lymphocyte Ag 6B.2 (Ly-6B.2), matrix metalloproteinase-12 (MMP12), mucin 5B (MUC5B), and mucin 5AC (MUC5AC) was performed using the procedure published previously (24, 28, 32, 33). Briefly, formalin-fixed, paraffin-embedded lung sections were deparaffinized with CitriSolv (2 × 5 min each) and were rehydrated with graded ethanol (100%, 95%, 70%, and 30%, distilled water; 3 min each). For MBP and Ly-6B.2, Ag retrieval was performed using Proteinase K (sections incubated for 20 min at 37°C, followed by cooling to room temperature [RT]). For MMP12, MUC5B, and MUC5AC, Ag retrieval was performed using a citrate buffer–based heat-induced Ag-retrieval method (heating slides in 10 mM sodium citrate solution (with 0.05% Tween 20 [pH 6]) at 95–100°C for 30 min, followed by cooling to RT). Quenching for endogenous peroxidases was performed with 3% hydrogen peroxide in deionized water for 10 min at RT. Sections were blocked with the blocking buffer for 20 min followed by primary Ab incubation with rat monoclonal MBP primary Ab (MT-14.7.3; Mayo Clinic, Scottsdale, AZ), Ly-6B.2 (MCA771G; clone 7/4; Bio-Rad), rabbit polyclonal to MMP12 (ab66157; Abcam, Cambridge, MA), rabbit polyclonal MUC5B primary Ab (UNC223; University of North Carolina, Chapel Hill, NC), and MUC5AC primary Ab (UNC294; University of North Carolina) at RT for 2 h. Sections were washed and incubated with diluted biotinylated secondary Abs for 1 h at RT. The sections were then rinsed in deionized water (2 × 5 min each) and processed using a VECTASTAIN Elite ABC HRP Kit (PK-6101; Vector Laboratories, Burlingame, CA), followed by chromogenic substrate conversion to insoluble colored precipitate using ImmPACT NovaRED HRP Substrate Kit (SK-4800; Vector Laboratories). After a rinse in tap water, sections were counterstained with Gill’s hematoxylin-I, rinsed in deionized water, dehydrated with graded alcohol solutions, and coverslipped with VectaMount Mounting Media (H-5000; Vector Laboratories). Mounted slides were imaged using transmitted light microscopy (Nikon Ci-L microscope). Immunostained slides were analyzed for the determination of positively stained cells. Briefly, photographs were captured under ×40 objective of the ECLIPSE Ci-L microscope with DS-Fi2 camera attachment (Nikon, Melville, NY). Thereafter, captured images were processed using the ImageJ software (National Institutes of Health) to determine the percent positively stained area (MUC5B and MUC5AC) or positively stained cells (MMP12) (29).

Gene expression analyses on mRNA harvested from right lungs were performed as previously described (32).

One-way ANOVA followed by Tukey post hoc test for multiple comparisons was used to determine significant differences among groups. Outliers were removed using the Grubbs test. Individual data points in scatter plots represent individual animals. All data were expressed as mean ± SEM. To minimize the number of horizontal bars indicating statistically significant differences between experimental groups, we used a single bar when one group was significantly different from multiple other groups and used smaller vertical bars to indicate the groups that were being compared. The p values <0.05 were considered statistically significant. Statistical analyses were performed using Prism 8.0 (GraphPad Software, La Jolla, CA).

To study the characteristics of ozone-induced lung disease in developing mouse lungs in WT and in Scnn1b-Tg⁺ (Tg⁺) mice that manifest spontaneous mucoinflammatory lung disease with neonatal onset (PND2–3), 3-d-old WT and Scnn1b-Tg⁺ neonates were exposed to FA or 800 ppb ozone until the age of 20 d and necropsied at PND21 (weaning age) (Fig. 1A). Neonates were exposed to a designated environment for 4 h/d in nightly conditions and analyzed for designated end points, including BALF immune cell composition, inflammatory mediator analysis, bacterial burden, Ig isotype determination, histopathological and immunohistochemical analyses, and gene expression analyses (Fig. 1A).

Body weight gain was comparable between FA-exposed WT (6.12 ± 0.27 g) and ozone-exposed WT (5.69 ± 0.38 g) mice (Fig. 1B). However, the body weight gain was significantly less in ozone-exposed Scnn1b-Tg⁺ (4.71 ± 0.27 g) versus FA-exposed Scnn1b-Tg⁺ (6.32 ± 0.18 g) mice (Fig. 1B). Ozone is known to cause injury to the gas exchange barrier, which is indicated by the leakage of proteins into the alveolar airspaces (34, 35). To investigate the extent of injury to the gas exchange barrier in the developing neonates with and without Scnn1b-Tg⁺ lung disease, we assayed the total protein contents in the BALF from WT and Scnn1b-Tg⁺ mice exposed to either FA or ozone. The total protein contents were higher in ozone-exposed WT (267.9 ± 22.61 μg/ml) versus FA-exposed WT (187.9 ± 17.76 μg/ml) mice. The total protein contents in FA-exposed Scnn1b-Tg⁺ mice were 246.9 ± 23.32 μg/ml. Ozone exposure significantly increased the total protein in Scnn1b-Tg⁺ mice (571.3 ± 49.81 μg/ml; ∼3-fold as compared with FA-exposed WT mice, ∼2.1-fold as compared with ozone-exposed WT mice, and ∼2.3-fold as compared with FA-exposed Scnn1b-Tg⁺ mice) (Fig. 1C).

Increased levels of dsDNA in the BALF are indicative of airway epithelial damage and inflammation (36, 37). To determine whether ozone exposure causes an increase in BALF dsDNA, we assayed DNA contents in the BALF from WT and Scnn1b-Tg⁺ mice exposed to either FA or ozone. To our surprise, despite the presence of a mucoinflammatory lung disease in the Scnn1b-Tg⁺ neonates, FA-exposed Scnn1b-Tg⁺ neonates had dsDNA contents that were comparable to FA-exposed WT neonates (Fig. 1D). However, as expected, ozone exposure resulted in a significant increase in the dsDNA content in the BALF of WT as well as Scnn1b-Tg⁺ mice (Fig. 1D).

To study immune cell recruitment in ozone-exposed neonates during active phases of postnatal lung development in WT and Scnn1b-Tg⁺ neonates with mucoinflammatory lung disease, we performed immune cell analyses of the BALF from FA- or ozone-exposed WT and Scnn1b-Tg⁺ mice. Ozone exposure resulted in a significant increase in total cell counts in ozone-exposed WT mice versus FA-exposed WT mice (Fig. 2A). This increase in total cell counts was attributable to significantly increased numbers of neutrophils, eosinophils, and lymphocytes (Fig. 2B–D). As reported previously (24), the FA-exposed Scnn1b-Tg⁺ mice had significantly higher numbers of BALF immune cells (∼3.1-fold versus FA-exposed WT mice) (Fig. 2A), which was attributable to a significant increase in macrophage, neutrophil, eosinophil, and lymphocyte counts (Fig. 2B–D). Ozone-exposed Scnn1b-Tg⁺ mice exhibited a significant increase in the total cell counts as compared with FA-exposed WT mice and ozone-exposed WT mice (Fig. 2A). Further, total cell counts were significantly higher in ozone-exposed Scnn1b-Tg⁺ mice versus FA-exposed Scnn1b-Tg⁺ mice (Fig. 2A), which were attributed to a significant increase in neutrophils, eosinophils, and lymphocytes (Fig. 2B–D). A similar trend was observed when the lung sections were immunostained for MBP, an eosinophil-specific marker (Fig. 2E, 2F), and Ly-6B.2, a neutrophil- and inflammatory monocyte-specific marker (Fig. 2E, 2F).

Although only 0.1% of BALF macrophages from FA-exposed WT mice had surface area >300 μm², ∼5.6% of macrophages had surface area >300 μm² in ozone-exposed WT mice (Supplemental Fig. 1). Further, the relative frequency of macrophages with 200-μm² surface area was ∼5% in FA-exposed WT mice versus ∼19.5% in ozone-exposed WT mice. As previously reported (19, 34), BALF macrophages from Scnn1b-Tg⁺ mice were morphologically activated. Although ∼12.6% of macrophages in FA-exposed Scnn1b-Tg⁺ mice had surface area >300 μm², ∼19.1% BALF macrophages were >300 μm² in ozone-exposed Scnn1b-Tg⁺ mice (Supplemental Fig. 1). These data suggest that early postnatal ozone exposure enhances immune cell recruitment and morphological activation of macrophages in the lung airspaces of WT mice and that this response is further exaggerated in the presence of an existing/ongoing muco-obstructive airway disease.

We reasoned that the increased numbers of immune cells in the airspaces of ozone-exposed neonates are due to the elevated levels of chemokines in the airspaces; therefore, we assessed the levels of chemokines in the BALF from FA- and ozone-exposed WT and Scnn1b-Tg⁺ mice. First, we investigated the levels of primary neutrophil chemokines, including KC/CXCL1, MIP-2/CXCL2, MCP-1/CCL2, and G-CSF. KC/CXCL1 levels trended higher in ozone-exposed WT mice as compared with FA-exposed WT mice, but the increase was not statistically significant (Fig. 3A). As previously reported (17), KC levels were significantly higher in FA-exposed Scnn1b-Tg⁺ mice as compared with FA-exposed WT mice (Fig. 3A). FA-exposed Scnn1b-Tg⁺ mice also displayed significantly higher levels of KC as compared with ozone-exposed WT mice (Fig. 3A). Contrary to elevated neutrophil counts, the KC levels in ozone-exposed Scnn1b-Tg⁺ mice were significantly lower versus FA-exposed Scnn1b-Tg⁺ mice (Fig. 3A). As previously reported (19), FA-exposed Scnn1b-Tg⁺ mice showed significantly increased levels of MIP-2/CXCL2 as compared with FA-exposed WT mice (Fig. 3B). Although there was a significant increase in MIP-2 levels in FA-exposed Scnn1b-Tg⁺ mice as compared with ozone-exposed WT mice, there was no difference in MIP-2 levels among other groups. MCP-1/CCL2, a chemokine for neutrophils as well as monocytes, was comparable among FA-exposed WT, ozone-exposed WT, and FA-exposed Scnn1b-Tg⁺ mice. The BALF levels for MCP-1 were, however, significantly elevated in ozone-exposed Scnn1b-Tg⁺ mice (∼3.6-fold versus FA-exposed Scnn1b-Tg⁺ mice, ∼5.5-fold versus ozone-exposed WT mice, and ∼5.5-fold versus FA-exposed WT) (Fig. 3C). Consistent with increased neutrophilic counts in ozone-exposed Scnn1b-Tg⁺ mice, the G-CSF levels in ozone-exposed Scnn1b-Tg⁺ mice were also significantly elevated (∼11.7-fold as compared with ozone-exposed WT mice and ∼60-fold versus FA-exposed WT mice) (Fig. 3D).

Next, we investigated the levels of primary eosinophil chemokines, including CCL5 (RANTES) (Fig. 3E), CCL7 (MCP3) (Fig. 3F), CCL11 (Eotaxin-1) (Fig. 3G), CCL12 (MCP5) (Fig. 3H), CCL17 (Fig. 3I), CCL22 (Fig. 3J), and CCL24 (Eotaxin-2) (Fig. 3K). Except for CCL17, none of these chemokines were significantly elevated in ozone-exposed WT or FA-exposed Scnn1b-Tg⁺ mice. As compared with FA-exposed WT, whereas the mean concentration of CCL17 was insignificantly increased in ozone-exposed WT mice, the FA-exposed Scnn1b-Tg⁺ had significantly higher levels of CCL17 (Fig. 3I). Except for CCL5 (Fig. 3E), all of the eosinophil chemokines (Fig. 3F–K) were significantly elevated in ozone-exposed Scnn1b-Tg⁺ mice.

Expression levels of MIP-1α/CCL3, another chemokine for macrophages, lymphocytes, eosinophils, and neutrophils, trended higher in ozone-exposed WT mice versus FA-exposed WT mice, but the difference was not significant. Consistent with the significant increase in macrophage, lymphocyte, eosinophil, and neutrophil counts, MIP-1α levels were significantly increased in FA-exposed Scnn1b-Tg⁺ mice (∼8.9-fold versus FA- exposed WT mice) (Fig. 3L). Although ozone-exposed Scnn1b-Tg⁺ and ozone-exposed WT mice showed comparable macrophage as well as eosinophil counts, the significant increase in MIP-1α in ozone-exposed Scnn1b-Tg⁺ (∼5.4-fold versus ozone-exposed WT) (Fig. 3L) might in part contribute to increased neutrophil and lymphocyte recruitment in this group (Fig. 2B, 2C). The expression level of MIP-1α was comparable among ozone-exposed Scnn1b-Tg⁺ and FA-exposed Scnn1b-Tg⁺ mice.

MIP-1β/CCL4, a chemokine for monocytes, eosinophils, and lymphocytes, was comparable among ozone-exposed WT versus FA-exposed WT mice. FA-exposed Scnn1b-Tg⁺ mice showed significantly increased MIP-1β levels as compared with FA-exposed WT mice (∼8.5-fold) (Fig. 3M), which mirrored the significant increase in macrophages, eosinophils, and lymphocytes in this group. Although MIP-1β levels trended higher in ozone-exposed Scnn1b-Tg⁺ mice as compared with FA-exposed Scnn1b-Tg⁺ mice, the difference was not statistically significant. IL-6 and IP-10 (CXCL10), both consistently found in ozone-exposed airspaces, were significantly upregulated in the ozone-exposed Scnn1b-Tg⁺ mice (Fig. 3N, 3O).

Next, the effects of ozone exposure on the parenchymal morphology and alterations associated with the inflammatory changes in WT and Scnn1b-Tg⁺ mice were evaluated. FA-exposed WT mice had no signs of the infiltration of immune cells around pulmonary vessels (perivascular inflammation) or in and around the intrapulmonary airways (peribronchiolar inflammation) (Fig. 4A, 4E, 4I), alveolar space enlargement (Fig. 4B, 4E, 4I), lymphoid hyperplasia (Fig. 4C, 4E, 4I), and increase in the thickness of the alveolar septa/infiltration of alveolar spaces with inflammatory cells (septal thickening/consolidation) (Fig. 4D, 4E, 4I). In contrast, the ozone-exposed WT mice exhibited marked signs of perivascular and peribronchiolar inflammation (Fig. 4A, 4F, 4M), alveolar space enlargement (Fig. 4B, 4F, 4J), and septal thickening/consolidation (Fig. 4D, 4F, 4M).

As previously reported (24), FA-exposed Scnn1b-Tg⁺ mice had significant perivascular and peribronchiolar inflammation (Fig. 4A, 4G, 4N), alveolar space enlargement (Fig. 4B, 4G, 4K), lymphoid hyperplasia (Fig. 4C, 4G, 4N), and septal thickening/consolidation (Fig. 4D, 4G, 4N). As compared with FA-exposed Scnn1b-Tg⁺ mice, the ozone-exposed Scnn1b-Tg⁺ mice had significantly increased perivascular and peribronchiolar inflammation (Fig. 4A, 4H, 4O, 4P), alveolar space enlargement (Fig. 4B, 4H, 4L), and septal thickening/consolidation (Fig. 4D, 4H, 4O, 4P). Despite the relatively larger size of lymphoid aggregates in ozone-exposed Scnn1b-Tg⁺ mice, the incidence of lymphoid aggregates was comparable between FA-exposed and ozone-exposed Scnn1b-Tg⁺ mice (Fig. 4C, 4P).

MMP12 is a critical protease in the pathogenesis of emphysema in various disease models, including Scnn1b-Tg⁺ mice (38–40). To determine whether the emphysematous responses in the three experimental groups are caused by the upregulated expression of MMP12, we assessed MMP12 protein contents and Mmp12 mRNA levels in the lungs of these mice. The MMP12-stained cells were scarce in FA-exposed WT mice but widespread in the lungs from the three groups with alveolar space enlargement (Fig. 5A, 5B). Among the three groups with increased MMP12⁺ cell counts, only ozone-exposed Scnn1b-Tg⁺ mice had significantly higher MMP12⁺ cell counts versus FA-exposed WT mice (Fig. 5B). Although macrophage staining intensity was comparable between FA-exposed Scnn1b-Tg⁺ and ozone-exposed Scnn1b-Tg⁺ mice (Fig. 5A), macrophages were morphologically enlarged in ozone-exposed Scnn1b-Tg⁺ versus FA-exposed Scnn1b-Tg⁺ mice (Supplemental Fig. 1). This suggests that macrophages are morphologically activated and may contribute toward relatively greater MMP12 activity and result in exaggerated alveolar space enlargement in ozone-exposed Scnn1b-Tg⁺ mice (Fig. 5A, 5B). Consistent with MMP12 immunostaining, as compared with the FA-exposed WT mice, the Mmp12 mRNA levels were increased in both ozone-exposed WT and FA-exposed Scnn1b-Tg⁺ mice, but the differences were not statistically significant (Fig. 5C). The ozone-exposed Scnn1b-Tg⁺ mice had significantly higher expression of Mmp12 mRNA as compared with FA-exposed Scnn1b-Tg⁺ mice (Fig. 5C).

Ozone inhalation causes epithelial remodeling with type 2 immune response in the airways of adult mice (26, 28, 41), and type 2 responses have been previously reported in the Scnn1b-Tg⁺ mice (24, 33). To determine the effect of ozone exposure on epithelial remodeling in developing lungs of WT and Scnn1b-Tg⁺ mice, we analyzed the levels of soluble mediators in the BALF and the expression of genes that are known to drive type 2 inflammatory responses in the lung.

First, we analyzed the mRNA expression levels of the three master regulators of type 2 immunity (i.e., Il33, Tslp, and Il25). As compared with FA-exposed WT mice, the ozone-exposed WT mice had upregulated levels of mRNAs for Il33 (p < 0.05) (Fig. 6A) as well as Tslp (p = 0.07) (Fig. 6B). Although the FA-exposed Scnn1b-Tg⁺ mice had elevated levels of mRNAs for Il33 as well as Tslp as compared with FA-exposed WT mice, the upregulation was not statistically significant. As compared with FA-exposed Scnn1b-Tg⁺ mice, the ozone-exposed Scnn1b-Tg⁺ mice had increased levels of Il33 mRNA and Tslp mRNA, but the difference was not statistically significant. Although the Il25 mRNA was not detected in all of the samples analyzed, the comparison of mean expression in those samples in which it was detected revealed higher levels in ozone-exposed WT as well as ozone-exposed Scnn1b-Tg⁺ mice (Fig. 6C). Following the Il33 mRNA trend, IL-33 protein levels were elevated in both ozone-exposed WT and ozone-exposed Scnn1b-Tg⁺ mice (Fig. 6D).

Next, we analyzed protein and mRNA levels of primary type 2 cytokines (i.e., IL-4, IL-13, and IL-5) in all four experimental groups. IL-4 contents trended higher in the BALF from ozone-exposed WT (1.13 ± 0.24 pg/ml) versus FA-exposed WT (0.55 ± 0.016 pg/ml) mice (Fig. 6E). Consistent with the cytokine protein contents, the Il4 mRNA levels also trended higher in ozone-exposed WT versus FA-exposed WT mice (Fig. 6F). As compared with FA-exposed WT mice, FA-exposed Scnn1b-Tg⁺ mice had elevated levels of both IL-4 protein (p < 0.01) (Fig. 6E) and Il4 mRNA (NS) (Fig. 6F). Although the IL-4 protein contents were comparable between FA- and ozone-exposed Scnn1b-Tg⁺ mice (Fig. 6E), the ozone-exposed Scnn1b-Tg⁺ mice had significantly elevated levels of Il4 mRNA (Fig. 6F).

IL-13 protein levels (Fig. 6G) and Il13 mRNA (Fig. 6H) were comparable among ozone-exposed WT versus FA-exposed WT mice but trended higher in FA-exposed Scnn1b-Tg⁺ mice. IL-13 protein contents were significantly elevated (Fig. 6G), whereas Il13 mRNA levels trended higher (p = 0.07) (Fig. 6H) in ozone-exposed Scnn1b-Tg⁺ versus FA-exposed Scnn1b-Tg⁺ mice.

IL-5 is a type 2 cytokine that plays a key role in the proliferation, maturation, differentiation, and migration of eosinophils (42–44). IL-5 protein levels trended higher in ozone-exposed WT versus FA-exposed WT mice and FA-exposed Scnn1b-Tg⁺ versus FA-exposed WT, but the differences were not statistically significant (Fig. 6I). Consistent with the presence of increased numbers of eosinophils in ozone-exposed Scnn1b-Tg⁺ mice (Fig. 2), IL-5 levels were significantly higher in ozone-exposed Scnn1b-Tg⁺ mice (∼3.1-fold versus FA-exposed Scnn1b-Tg⁺, ∼12.4-fold versus ozone-exposed WT mice, and ∼196.7-fold versus FA-exposed WT mice) (Fig. 6I). The Il5 mRNA levels trended higher in ozone-exposed WT, FA-exposed Scnn1b-Tg⁺, and ozone-exposed Scnn1b-Tg⁺ mice compared with FA-exposed WT mice (Fig. 6J).

Airway mucus obstruction is a consistent feature of Scnn1b-Tg⁺ lung disease (17). To investigate the role of ozone in modulating mucus obstruction, we assessed the airway luminal mucus contents in AB-PAS–stained lung sections from WT and Scnn1b-Tg⁺ mice exposed to FA or ozone. Although the FA-exposed WT mice had no signs of mucous cell metaplasia and airway mucus obstruction (Fig. 7A, 7B), the airways from ozone-exposed WT mice had a marked increase in the proportion of mucous cells and the degree of airway mucus obstruction (Fig. 7A, 7B). As previously reported (24), FA-exposed Scnn1b-Tg⁺ mice had elevated numbers of mucous cells associated with airway mucus obstruction (Fig. 7A, 7B). Ozone exposure to Scnn1b-Tg⁺ mice, however, significantly increased the degree of airway mucus obstruction (Fig. 7A, 7B).

Airway mucus obstruction in Scnn1b-Tg⁺ mice is primarily caused by two gel-forming mucins (i.e., MUC5B and MUC5AC). To determine the effect of ozone exposure on the expression of these mucins and their contribution to the mucus obstruction, we quantified the mRNA expression for Muc5b (Fig. 7C) and Muc5ac (Fig. 7F) in total mRNA collected from lungs as well as performed immunolocalization of MUC5B (Fig. 7D, 7E) and MUC5AC (Fig. 7G, 7H) in lung sections. The mRNA levels for Muc5b were minimal in FA-exposed WT mice. The Muc5b mRNA levels were significantly elevated in ozone-exposed WT as well as FA-exposed Scnn1b-Tg⁺ mice. The epithelial immunostaining intensity for MUC5B (Fig. 7D, 7E) was consistent with their mRNA levels (Fig. 7C). Importantly, although the Muc5b mRNA levels were comparable between FA-exposed and ozone-exposed Scnn1b-Tg⁺ mice, the intraluminal staining was more widespread in ozone-exposed Scnn1b-Tg⁺ mice, which is consistent with the significantly greater airway mucus obstruction in ozone-exposed Scnn1b-Tg⁺ mice (Fig. 7D). The percent area stained positively for MUC5B intracellular immunostaining were significantly increased in ozone-exposed WT versus FA-exposed WT, FA-exposed Scnn1b-Tg⁺ versus ozone-exposed WT, and ozone-exposed Scnn1b-Tg⁺ mice versus FA-exposed Scnn1b-Tg⁺ mice (Fig. 7E).

As compared with the FA-exposed WT mice, Muc5ac mRNA levels were significantly elevated in the ozone-exposed WT mice (Fig. 7F). Consistent with the mRNA levels, the MUC5AC immunostaining was widespread in the airways of ozone-exposed WT mice compared with FA-exposed WT mice (Fig. 7G, 7H). As compared with FA-exposed Scnn1b-Tg⁺ mice, the Muc5ac mRNA levels trended higher in ozone-exposed Scnn1b-Tg⁺ mice (Fig. 7F). Consistent with this, the relative proportion of MUC5AC-stained cells was higher in the ozone-exposed Scnn1b-Tg⁺ versus FA-exposed Scnn1b-Tg⁺ mice (Fig. 7H). These data suggest that MUC5AC is upregulated in the presence of ozone or ASL dehydration, but the superimposition of ozone exposure onto the ASL dehydration does not have any synergistic or additive effect.

Next, we performed gene expression analyses for genes including Spdef (Fig. 7I), Foxa2 (Fig. 7J), Agr2 (Fig. 7K), Slc26a4 (Fig. 7L), Retnla (Fig. 7M), and Chil4 (Fig. 7N) that are known to be associated with mucus production and type 2 inflammation. As compared with the FA-exposed WT mice, the ozone-exposed WT mice had a higher expression for all of the assayed genes, but a significant difference was attained only for Retnla and Chil4 (Fig. 7M, 7N). In contrast to WT mice, the effect of ozone exposure was more pronounced in Scnn1b-Tg⁺ mice. As compared with the FA-exposed Scnn1b-Tg⁺ mice, the ozone-exposed Scnn1b-Tg⁺ mice had a significantly higher expression for all of the assayed genes, except Foxa2 (Fig. 7J) and Chi4 (Fig. 7N).

Due to mucociliary clearance defect, the Scnn1b-Tg⁺ mice are prone to spontaneous bacterial infections arising from the aspiration of bacteria of oropharyngeal origin that usually clear up by 3 to 4 wk of age (19). To determine the effect of ozone exposure on the clearance of bacterial infections, we analyzed the bacterial burden in aseptically harvested BALF from FA- or ozone-exposed WT and Scnn1b-Tg⁺ mice. Only 13% of FA-exposed WT mice (3 out of 23; mean CFU ∼8.7/ml) and 20% of ozone-exposed WT mice (3 out of 15; mean CFU ∼15.6/ml) exhibited mild bacterial infection. As expected from previous studies, 34.6% (9 out of 26 mice; CFU ∼82.7/ml) of FA-exposed Scnn1b-Tg⁺ mice still had bacterial infection (mean CFU ∼82.7/ml) at PND21. In contrast, 75% (15 out of 20 mice) of ozone-exposed Scnn1b-Tg⁺ mice displayed significantly higher bacterial infection with a mean CFU ∼725/ml (Fig. 8A).

IL-10, an immunosuppressive and anti-inflammatory cytokine, was recently shown to suppress bacterial clearance in Scnn1b-Tg⁺ mice (33). To determine whether the reduced bacterial clearance in ozone-exposed Scnn1b-Tg⁺ mice is associated with IL-10 overproduction, we analyzed IL-10 protein levels in BALF and Il10 mRNA levels in lung homogenates. Il10 mRNA levels were comparable in the lungs from FA-exposed WT, ozone-exposed WT, and FA-exposed Scnn1b-Tg⁺ mice. Interestingly, the ozone-exposed Scnn1b-Tg⁺ mice had significantly increased levels of Il10 mRNA (Fig. 8B). Consistent with the mRNA levels, the BALF IL-10 levels were comparable in FA-exposed WT, ozone-exposed WT, and FA-exposed Scnn1b-Tg⁺ mice, whereas the BALF levels of IL-10 were elevated in ozone-exposed Scnn1b-Tg⁺ mice (Fig. 8C).

Ozone has been shown to compromise the Ig production by B cells in vitro (45). Accordingly, we speculated that the Ig-mediated bacterial clearance defense system may have been compromised in ozone-exposed Scnn1b-Tg⁺ mice. Therefore, we assayed the levels of Igs (IgA, IgG1, IgG2a, IgG2b, IgG3, and IgM) in the BALF from FA- and ozone-exposed WT and Scnn1b-Tg⁺ mice. IgA was found elevated only in the BALF from FA-exposed Scnn1b-Tg⁺ mice versus all of the other groups (Fig. 8D). IgG1 was not different among any of the groups (Fig. 8E). Three isotypes of IgG (i.e., IgG2a, IgG2b, and IgG3) were significantly elevated in BALF from ozone-exposed WT mice versus FA-exposed WT (Fig. 8F–H). Although the mean concentration values for IgG isotypes were higher in FA-exposed Scnn1b-Tg⁺ mice versus FA-exposed WT mice, the differences were not statistically significant (Fig. 8F–H). The concentrations of both IgG2b and IgG3 were significantly elevated in ozone-exposed Scnn1b-Tg⁺ versus FA-exposed Scnn1b-Tg⁺ mice (Fig. 8G, 8H). As compared with FA-exposed WT mice, both ozone-exposed WT and FA-exposed Scnn1b-Tg⁺ mice had elevated levels of IgM, but the differences were not statistically significant (Fig. 8I). The levels of IgM were significantly elevated in ozone-exposed Scnn1b-Tg⁺ as compared with FA-exposed Scnn1b-Tg⁺ mice (Fig. 8I). In summary, except for IgA, in which a trend for lower concentration was observed in ozone-exposed Scnn1b-Tg⁺ (10.1 ± 1.47 ng/ml) versus FA-exposed Scnn1b-Tg⁺ (16.9 ± 4.47 ng/ml) mice, all of the other Ig subtypes were present at elevated levels (Fig. 8E–I) in ozone-exposed Scnn1b-Tg⁺ mice versus FA-exposed Scnn1b-Tg⁺ mice. In particular, IgG2b, IgG3, and IgM were significantly elevated in ozone-exposed Scnn1b-Tg⁺ mice versus FA-exposed Scnn1b-Tg⁺ mice (Fig. 8G–I). These results indicate that reduced bacterial clearance in Scnn1b-Tg⁺ mice following ozone exposure is not attributed to reduced production of Igs.

A significant knowledge gap exists in our understanding of the interactions among ozone-dominated air pollution, early postnatal lung development, and pathogenesis of mucoinflammatory lung diseases of genetic origin. CF, a genetic disease caused by a gene mutation in the CFTR gene, initiates during early childhood, and mucoinflammatory lung disease in patients with CF contributes to significant morbidity and mortality (46). Our PubMed search for the interaction between ozone and CF found only three studies that reported the association between ozone exposure and pulmonary exacerbations in patients with CF (7, 8, 25). Experimental in vivo studies investigating the three-dimensional interaction among inhaled ozone, lung development, and disease pathogenesis have never been performed. Accordingly, this mouse-based study was conceived to address two key questions: 1) Does ozone affect the normal lung development in WT neonates? 2) Does ozone affect the developmental history of CF-like mucoinflammatory lung disease in mice?

In WT mice, ozone exposure resulted in a significant disruption in lung development, both morphologically and in terms of inflammatory responses. For instance: 1) the increase in the proportions of AB-PAS–positive, MUC5AC-positive, and MUC5B-positive mucous cells following ozone exposure reflected significant airway epithelial remodeling; 2) increased proportion of mucous cells possibly resulted in mucus obstruction of the airways after ozone exposure; 3) alveolar space enlargement (emphysema) was prominent in ozone-exposed WT mice; 4) granulocytic (neutrophilic and eosinophilic) inflammation in airspaces as well as in subepithelial regions was established following ozone exposure; 5) gene signatures of type 2 inflammation, including Il33, Tslp, Il5, Muc5b, Muc5ac, Retnla, and Chil4 were upregulated following ozone exposure; and 6) IgG Ig subtypes, including IgG2a, IgG2b, and IgG3, were significantly elevated in the BALF following ozone exposure. These outcomes indicate a marked impact of ozone on postnatal lung development and in the establishment of type 2 inflammatory disease. However, it remains to be tested whether these pathological features are reversible and whether these early life ozone exposures affect respiratory health during adulthood.

In infants with CF, the clinical signatures of neutrophil-dominated airway inflammation are present within the first few weeks of life (47, 48). Bacterial infections have been reported in CF infants as early as 3 mo of age (49). Structural changes (i.e., bronchiectasis and air trapping) become evident during preschool years, in particular, between 3 and 5 y of age (50, 51). During the first 3 y of life, a period critical for the alveolar maturation phase of lung development (52), most of the pathological features of the CF lung disease are well established. The Scnn1b-Tg⁺ mice, a widely accepted model of CF, exhibit clinical signatures of human CF-like lung disease, including early postnatal initiation of airway mucus obstruction, ASL dehydration, granulocytic airway inflammation, and spontaneous bacterial infection within the first 21 d of their life (17, 19). Accordingly, we exposed Scnn1b-Tg⁺ neonates to ozone beginning at PND3 until the age of 20 d, which corresponds to early preschool years (53).

Ozone exposure to the Scnn1b-Tg⁺ mice resulted in a significant worsening of most of the mucoinflammatory features that are inherent to this model (17). For instance: 1) granulocytic (neutrophilic and eosinophilic) recruitment to Scnn1b-Tg⁺ lungs was significantly increased; 2) a large number of inflammatory mediators were present at significantly higher levels; 3) bacterial clearance was compromised; 4) alveolar space enlargement was severe; and 5) mucus obstruction and associated MUC5B protein expression were significantly increased. Putting together data from ozone-exposed WT and Scnn1b-Tg⁺ mice, it appears that the ozone and ASL dehydration independently initiate type 2 inflammatory cascades. Further, the superimposition of the two stresses exaggerates the type 2 mucoinflammatory features.

MMP12, a matrix metalloproteinase, has been implicated in alveolar wall destruction leading to the alveolar space enlargement, also known as emphysema (38, 39). Ozone-exposed WT mice had widespread alveolar space enlargement. Alveolar space enlargement, a consistent feature in Scnn1b-Tg⁺ mice (17), was marked in FA-exposed Scnn1b-Tg⁺ mice. In both of these groups (i.e., ozone-exposed WT and FA-exposed Scnn1b-Tg⁺ mice), Mmp12 mRNA and protein levels were elevated. Further, the superimposition of ozone exposure on Scnn1b-Tg⁺ lung disease caused further enlargement of alveolar spaces that was associated with increased mRNA and protein expression of MMP12. Based on a previous report that demonstrated macrophage-specific MMP12 expression in Scnn1b-Tg⁺ mice is involved in alveolar space enlargement (40), ozone appears to induce MMP12 expression in activated macrophages that initiates alveolar space enlargement in WT mice and exaggerates this feature in Scnn1b-Tg⁺ mice. In addition to MMP12, neutrophil elastase (NE) has also been implicated in alveolar space enlargement in Scnn1b-Tg⁺ mice (54). Given significant neutrophilic inflammation in ozone-exposed WT mice, FA-exposed Scnn1b-Tg⁺, and ozone-exposed Scnn1b-Tg⁺ mice, the contribution of neutrophil-derived NE toward alveolar space enlargement cannot be ruled out. Whether both or one of the two proteases (i.e., macrophage-derived MMP12 or neutrophil-derived NE) contribute to emphysema, however, remains untested.

The Scnn1b-Tg⁺ mice are prone to spontaneous bacterial infections that usually clear up by 3 to 4 wk of age (19). Interestingly, whereas only ∼35% of the FA-exposed Scnn1b-Tg⁺ mice had bacterial infection (CFU ∼82.7/ml), ∼75% of the ozone-exposed Scnn1b-Tg⁺ mice still had bacterial infection (CFU ∼725/ml), indicating an effect of ozone on bacterial clearance. Delayed bacterial clearance has also been reported in Scnn1b-Tg⁺ mice that were either exposed to secondhand smoke (24) or were functionally or numerically deficient in immune cells (18, 33). In these studies, various factors, including compromised immune cell function or exaggerated mucus obstruction, were suggested as the contributors to sustained bacterial infections. In the current study, IL-10, a known suppressor of the antibacterial function of macrophages in Scnn1b-Tg⁺ mice (33), was found to be significantly elevated in ozone-exposed Scnn1b-Tg⁺ mice. Mucus obstruction, another factor favoring bacterial colonization, was also more widespread in the ozone-exposed Scnn1b-Tg⁺ mice. Therefore, it is likely that these two factors (i.e., elevated IL-10 and the static mucus, a known fertile niche for bacterial growth) contributed to the persistence of bacterial infection in ozone-exposed Scnn1b-Tg⁺ mice. Their relative contribution toward defective bacterial clearance, however, remains to be tested. Interestingly, the Ig levels as well as neutrophils were sufficient and, in fact, increased in ozone-exposed Scnn1b-Tg⁺ mice, further suggesting that the compromised bacterial clearance was a result of mucus obstruction–induced mucociliary clearance defect rather than the compromised functioning of Ig-mediated or phagocyte-mediated bacterial clearance mechanisms.

Eosinophil-dominated inflammatory responses to subchronic levels of ozone exposure have been reported in adult mice (26, 28, 41). However, the nature of ozone-induced inflammatory response during the active phase of postnatal lung development has never been explored. Eosinophils, both within the BALF and the lung tissue, were markedly increased following ozone exposure in WT mice, which was also associated with elevated levels of mRNAs for type 2 inflammation–associated cytokines (i.e., Il33, Tslp, and Il5). The superimposition of ozone exposure onto the preexisting Scnn1b-Tg⁺ lung disease further promoted type 2 responses, including BALF and tissue eosinophilia, Il4 and Il13 mRNA levels, and BALF IL-5 and IL-13 protein levels. Eosinophil chemokines, including CCL7 (MCP3), CCL11 (Eotaxin-1), CCL12 (MCP5), CCL17, CCL22, and CCL24 (Eotaxin-2), were also significantly increased in ozone-exposed Scnn1b-Tg⁺ mice. Further, type 2 inflammatory gene signatures, including Retnla (Fizz1), Chil4 (YM2), Slc26a4 (Pendrin), and Mmp12, were also significantly elevated following ozone exposure in either (WT or Scnn1b-Tg⁺) or both WT and Scnn1b-Tg⁺ mice. Based on previous reports (26, 33, 41), it could be speculated that these pronounced type 2 responses are mediated through IL-33/TSLP→ILC2 axis; however, direct experimental testing of this conceptual model (Fig. 8J) in neonates is awaited.

MUC5B and MUC5AC are prominent gel-forming mucins in the Scnn1b-Tg⁺ mucus plugs (55). The genetic ablation of MUC5B, but not MUC5AC, results in significantly reduced mucus obstruction in Scnn1b-Tg⁺ mice (55), suggesting a close association between the expression levels of MUC5B and the severity of mucus obstruction. We have previously reported reduced MUC5AC expression but sustained MUC5B expression and airway mucus obstruction in IL-33 knockout Scnn1b-Tg⁺ mice (32). Further, our recent study demonstrated that the ablation of innate lymphoid cells results in suppressed MUC5B expression and diminished airway mucus obstruction (33). Based on these reports, the significantly increased mucus obstruction in ozone-exposed Scnn1b-Tg⁺ mice is most likely caused by a significant increase in airway MUC5B expression. However, the primary driver of increased MUC5B expression remains elusive.

Although the current exposure paradigm revealed several interesting findings, some intriguing questions have also surfaced such as: 1) How would the various exposure parameters (i.e., ozone dose and exposure time [single versus repeated]) affect the outcome of various features of lung disease? 2) How would ozone exposure impact the clearance of experimentally introduced bacteria in the lungs of WT and Scnn1b-Tg⁺ mice? 3) What is the effect of ozone exposure on the phagocytic functions of macrophages and macrophage recruitment? Future studies addressing these critical questions are warranted.

In conclusion, this study revealed interesting interactions between two independent airspace stressors (i.e., ozone and ASL dehydration). Although both of these insults initiate type 2 mucoinflammatory lung responses along a common axis (i.e., IL-33/TSLP→ILC2→MCM), the degree of mucoinflammatory features is more pronounced in Scnn1b-Tg⁺ mice. The superimposition of ozone onto CF-like Scnn1b-Tg⁺ lung disease results in exaggerated pulmonary pathology, including type 2 inflammation, persisted bacterial infection, mucus obstruction, and alveolar space enlargement. Based on these key findings, we propose that ozone exposure promotes pulmonary expression and secretion of IL-33 and/or TSLP that in turn orchestrates a series of type 2 responses (Fig. 8J).

We thank Sherry Ring for histological tissue processing, Thaya Stoufflet for assistance with multiplex cytokine assays, and Dr. Camille Ehre (University of North Carolina at Chapel Hill) for providing MUC5B and MUC5AC Abs.

The authors have no financial conflicts of interest.