Scnn1b-Transgenic BALB/c Mice as a Model of Pseudomonas aeruginosa Infections of the Cystic Fibrosis Lung

The opportunistic pathogen Pseudomonas aeruginosa is responsible for much of the morbidity and mortality associated with cystic fibrosis (CF), a condition that predisposes patients to chronic lung infections. P. aeruginosa lung infections are difficult to treat because P. aeruginosa adapts to the CF lung, can develop multidrug resistance, and can form biofilms. Despite the clinical significance of P. aeruginosa, modeling P. aeruginosa infections in CF has been challenging.

ABSTRACT

The opportunistic pathogen Pseudomonas aeruginosa is responsible for much of the morbidity and mortality associated with cystic fibrosis (CF), a condition that predisposes patients to chronic lung infections. P. aeruginosa lung infections are difficult to treat because P. aeruginosa adapts to the CF lung, can develop multidrug resistance, and can form biofilms. Despite the clinical significance of P. aeruginosa, modeling P. aeruginosa infections in CF has been challenging. Here, we characterize Scnn1b-transgenic (Tg) BALB/c mice as P. aeruginosa lung infection models. Scnn1b-Tg mice overexpress the epithelial Na⁺ channel (ENaC) in their lungs, driving increased sodium absorption that causes lung pathology similar to CF. We intranasally infected Scnn1b-Tg mice and wild-type littermates with the laboratory P. aeruginosa strain PAO1 and CF clinical isolates and then assessed differences in bacterial clearance, cytokine responses, and histological features up to 12 days postinfection. Scnn1b-Tg mice carried higher bacterial burdens when infected with biofilm-grown rather than planktonic PAO1; Scnn1b-Tg mice also cleared infections more slowly than their wild-type littermates. Infection with PAO1 elicited significant increases in proinflammatory and Th17-linked cytokines on day 3. Scnn1b-Tg mice infected with nonmucoid early CF isolates maintained bacterial burdens and mounted immune responses similar to those of PAO1-infected Scnn1b-Tg mice. In contrast, Scnn1b-Tg mice infected with a mucoid CF isolate carried high bacterial burdens, produced significantly more interleukin 1β (IL-1β), IL-13, IL-17, IL-22, and KC, and showed severe immune cell infiltration into the bronchioles. Taken together, these results show the promise of Scnn1b-Tg mice as models of early P. aeruginosa colonization in the CF lung.

INTRODUCTION

Pseudomonas aeruginosa is an opportunistic pathogen and important causative agent of serious lung infections in hospitalized and immunocompromised patients. These infections are difficult to treat, because P. aeruginosa can develop multidrug resistance and is capable of forming biofilms. Biofilms are microbial communities encased in a matrix of extracellular proteins, nucleic acids, and polysaccharides (1). Bacteria growing as a biofilm can tolerate antibiotic exposure and resist clearance by the host immune system (2), resulting in the establishment of chronic infections. P. aeruginosa is also metabolically adaptable and capable of producing a variety of virulence factors, enabling it to survive in the environment as well as human hosts.

Mice infected with planktonic P. aeruginosa develop acute pneumonia, resulting in sepsis or rapid clearance of the bacteria (3, 4). Acute pneumonia models have been valuable for the identification of P. aeruginosa virulence factors, including flagella (5) and pili (6), a type III secretion system, and secreted exotoxins (7), proteases (8), and phenazines (9, 10). To mimic a chronic lung infection, P. aeruginosa has been embedded in an alginate/agarose bead (11, 12) or a fibrin plug (13, 14) that serves as an artificial biofilm and a long-term nidus of infection. Chronic P. aeruginosa lung infection models, which can prolong infection for up to 3 months (15), have been useful for testing anti-P. aeruginosa therapeutics (16, 17) and contributed to the understanding of the immune responses to chronic infection. For example, agar bead models have suggested that Th2-skewed immune responses increase susceptibility to infection (18), raising the possibility that immunomodulatory therapy could be beneficial.

P. aeruginosa is responsible for much of the lung function deterioration associated with cystic fibrosis (CF) (19), and by age 20, 60 to 70% of CF patients are intermittently colonized by P. aeruginosa (20). CF is an autosomal recessive genetic disease caused by defects in the production and function of the cyclic AMP-regulated cystic fibrosis transmembrane conductance regulator (CFTR) channel, resulting in disruption of chloride and bicarbonate transport across epithelial cells. Without proper ion transport, CF patients develop thick, dehydrated mucus and impaired mucociliary clearance in their airways, predisposing patients to recurrent and chronic lung infections. Early in CF patients’ lives, P. aeruginosa infections are treatable with aggressive antibiotic regimens (21), but over time, P. aeruginosa adapts to the CF lung environment, typically through losing specific virulence factors while becoming increasingly antibiotic resistant (22, 23). Once a chronic infection is established, it is nearly impossible to eliminate.

Mucoid strains of P. aeruginosa hyperproduce the polysaccharide alginate and form thick biofilms. Mucoid P. aeruginosa is rarely observed in nature (24) but is selected through environmental adaptation within the CF lung (25, 26). The emergence of mucoid P. aeruginosa is linked to pulmonary exacerbations (27), coinfection with Staphylococcus aureus (28), increased antibiotic tolerance (29), resistance to opsonization and phagocytosis (30, 31), and worsened lung function deterioration (32). Although much progress has been made in treating the multisystem complications of CF, preventing and treating lung colonization with P. aeruginosa remains a major challenge.

A significant impediment to the study of CF lung infections is the lack of a CFTR mutant mouse that reliably develops the mucus plugging, neutrophil recruitment, and bronchiectasis observed in CF patients (33, 34). CFTR regulates the function of the epithelial Na⁺ channel (ENaC), and these channels together largely control the flux of sodium and chloride across epithelium (35). When CFTR is defective in CF airways, ENaC is overactive, resulting in increased sodium absorption (36). To attempt to mimic airway mucus obstruction and test the hypothesis that increased sodium absorption would produce CF-like lung pathology in mice, Mall et al. (37) developed transgenic mice that overexpress the beta subunit of ENaC (gene Scnn1b) under the control of the club cell secretory protein promoter. These Scnn1b-transgenic (Tg) C57BL/6 and BALB/c mice show increased sodium absorption in the airways, which drives airway surface liquid dehydration, mucus obstruction, and neutrophilic inflammation, much like pathological observations in CF patients (37, 38). To date, Scnn1b-Tg mice have primarily been used to study the pathophysiology and treatment of CF-like muco-obstructive lung disease (39–42), but they have also shown promise as models for acute infections in CF-like lungs. Neonatal Scnn1b-Tg C3H × C57BL/6 mice have been shown to clear intratracheal infections with Haemophilus influenzae or P. aeruginosa more slowly than wild-type mice, but the infections have not been studied for longer than 3 days (37). Scnn1b-Tg C57BL/6 mice have also been infected with P. aeruginosa using a fibrin plug model to study mechanisms by which treatment with nebulized alpha-1 antitrypsin improves lung function and decreases P. aeruginosa burden in CF patients (14). That study was limited to the first 3 days of infection, so it did not address the chronic nature of CF lung infections, but it demonstrated the potential utility of Scnn1b-Tg mice for testing therapeutics against infections in a CF-like lung.

Here, we developed a model of CF lung infection with P. aeruginosa using Scnn1b-Tg BALB/c mice. This model uses an intranasal route of infection and does not embed bacteria in foreign substances to prevent clearance, similar to natural infections. An advantage of this model is its ease of use, permitting the study of P. aeruginosa in a mouse with CF-like lung disease for up to 12 days postinfection. The utility of this model was confirmed by infecting mice with P. aeruginosa PAO1 as well as clinical isolates from young children with CF and evaluating clearance rates, bacterial burdens, and the immune response.

RESULTS

PAO1 consistently colonizes Scnn1b-Tg mouse lungs when it is grown as a biofilm.

To determine whether planktonic or biofilm-grown PAO1 was important in the colonization of the lung, Scnn1b-Tg mice and their wild-type littermates were intranasally infected with 2 × 10⁶ to 4 × 10⁶ CFU of planktonic or 3 × 10⁶ to 5 × 10⁶ CFU of biofilm-grown PAO1 on day 0. The biofilm inoculum was homogenized prior to plating, which was confirmed to break up bacterial aggregates using microscopy (see Fig. S1B in the supplemental material). On day 7 postinfection, more planktonically grown bacteria were recovered from Scnn1b-Tg mice than from wild-type mice (P = 0.055) (Fig. 1); however, only 41% (11/27) of surviving Scnn1b-Tg mice maintained the infection (geometric mean bacterial burden was below the limit of detection). In contrast, when PAO1 was grown as a biofilm, 90% (9/10 mice) of surviving Scnn1b-Tg mice remained infected on day 7, and they carried significantly more bacteria (geometric mean of 1,008 CFU/g of lung; P = 0.006). Among wild-type mice, only 25% (6/24 mice) and 27% (3/11 mice) remained infected on day 7 when infected with planktonic or biofilm-grown PAO1, respectively, demonstrating that the mucus-obstructed Scnn1b-Tg mice are more susceptible to infection with PAO1.

FIG 1.

Planktonic PAO1 is less infectious than PAO1 grown as a biofilm. PAO1 consistently colonizes the Scnn1b-Tg mouse lung when it is grown as a biofilm prior to infection rather than grown planktonically. Scnn1b-Tg mice and wild-type littermates were intranasally infected with 2 × 10⁶ to 4 × 10⁶ CFU planktonic PAO1 or 3 × 10⁶ to 5 × 10⁶ CFU biofilm-grown PAO1. Bars show geometric means with 95% confidence intervals. CFU data were log transformed and then analyzed using 2-tailed t tests. The limit of detection is 100 CFU/g of lung. Results are from independent planktonic infections (data from 4 infections; total of 24 wild-type and 27 Scnn1b-Tg mice) and biofilm infections (data from 3 infections; total of 11 wild-type and 10 Scnn1b-Tg mice). **, P < 0.01; ***, P < 0.001.

Scnn1b-Tg mice clear PAO1 infections more slowly than wild-type littermates.

To assess clearance rates of biofilm-grown PAO1, Scnn1b-Tg mice and their wild-type littermates were intranasally infected with 3 × 10⁶ to 5 × 10⁶ CFU. Mice were sacrificed on days 3, 7, and 12 postinfection. The bacterial load in the lungs declined over time for both wild-type and Scnn1b-Tg mice; however, Scnn1b-Tg mice cleared the infections more slowly than wild-type littermates (Fig. 2). On day 3, 100% of surviving Scnn1b-Tg and wild-type mice were infected: Scnn1b-Tg mice carried a geometric mean of 4,235 CFU/g of lung, whereas wild-type mice carried only 705 CFU (P = 0.034). Between days 3 and 7, wild-type mice cleared 99.5% of the bacteria in their lungs, compared to 76% in Scnn1b-Tg mice. By day 12, only 18% (2/11) of wild-type mice remained infected, compared to 50% (5/10) of Scnn1b-Tg mice. The mortality rates for the Scnn1b-Tg and wild-type mice were 21% and 3%, respectively (P = 0.020) (Fig. S2A). Mortality was limited to the first 3 days postinfection, and preliminary experiments suggested that mortality was a result of sepsis (determined by dissemination to the kidneys; data not shown). Both Scnn1b-Tg and wild-type mice lost weight following infection and then rebounded (Fig. S2B). No significant differences were found between male and female mice.

FIG 2.

Scnn1b-Tg mice clear PAO1 more slowly than wild-type littermates. Biofilm-grown PAO1 is cleared from Scnn1b-Tg lungs over 12 days of infection. Scnn1b-Tg and wild-type BALB/c mice were infected with 3 × 10⁶ to 5 × 10⁶ CFU PAO1 and sacrificed on days 3, 7, and 12 postinfection. Values are log-transformed numbers of CFU with standard errors of the means. CFU data were log transformed and compared using unpaired, two-tailed t tests. The limit of detection is 100 CFU/g of lung. Results are from independent infections (data from 3 infections; total of 34 wild-type mice and 38 Scnn1b-Tg mice). *, P < 0.05; ***, P < 0.001.

Cytokine analysis.

Sputum samples from CF patients have been reported to contain large amounts of interleukin 17 (IL-17) during pulmonary exacerbations (43). To determine if Scnn1b-Tg BALB/c mice generate Th1-, Th2-, or Th17-skewed responses to P. aeruginosa infection and compare the response to that observed in CF patients, cytokine production in the lungs of Scnn1b-Tg and wild-type mice was compared on days 3 and 7 postinfection. Compared to wild-type mice, Scnn1b-Tg BALB/c mice produced increased levels of the Th1 cytokine gamma interferon (IFN-γ) and the neutrophil chemoattractant KC at baseline (P = 0.037 and 0.040), as well as on day 3 (postinfection (P = 0.013 and 0.049) (Fig. 3; Table S1). On day 3 postinfection, both Scnn1b-Tg and wild-type mice showed significantly increased production of IFN-γ over mock-infected controls (P = 0.003 and 0.005); however, production of IL-12 p70, another Th1-associated cytokine, was unaffected. The Th2-associated cytokines IL-4 and IL-13 were not significantly increased by P. aeruginosa infection. In contrast, the proinflammatory cytokines IL-1β and IL-6, as well as the Th17-associated cytokine IL-22, were significantly induced by P. aeruginosa infection on day 3 in both Scnn1b-Tg and wild-type mice (IL-1β, P = 0.011 and 0.008; IL-6, P = 0.040 and 0.024; IL-22, P = 0.024 and 0.041). Scnn1b-Tg mice also produced increased levels of IL-17 on day 3 postinfection (P = 0.025). Production of TNF-α and TGF-β was not significantly affected by P. aeruginosa infection and did not differ between Scnn1b-Tg and wild-type mice. By day 7 postinfection, cytokine levels were similar to baseline, with the exception of wild-type mice continuing to produce slightly elevated levels of IL-1β (P = 0.018).

FIG 3.

PAO1 infection elicits an early Th17-type response. The cytokine response to PAO1 infection was measured on days 3 and 7 postinfection in lung homogenates using Luminex multianalyte assays. Data are expressed as fold change relative to wild-type mock-infected mice. Bars represent means with standard deviations. Differences between groups were analyzed with two-tailed unpaired t tests (4 to 11 mice per group). *, P < 0.05; **, P < 0.01.

Infection with P. aeruginosa CF clinical isolates.

As PAO1 is a laboratory-adapted strain of P. aeruginosa, Scnn1b-Tg and wild-type mice were infected using P. aeruginosa isolates from CF patients to determine whether Scnn1b-Tg mice were more susceptible than their wild-type littermates. To model the earliest stages of P. aeruginosa infections in the CF lung, Scnn1b-Tg and wild-type mice were infected with the nonmucoid CF isolates CF001 and CF002, both grown as biofilms. CF001 and CF002, collected as part of previous studies that characterized the early natural history of CF lung disease (44, 45), were from individual 3-month-old and 4-month-old CF patients, respectively. Both isolates were collected via bronchoalveolar lavage. Similar to PAO1, both CF001 and CF002 exhibited colonization of the airways of Scnn1b-Tg mice on day 7, whereas they were cleared from the lungs of wild-type mice (P = 0.001 and 0.068) (Fig. 4). The bacterial burdens recovered from mice infected with PAO1 were not significantly different from those in mice infected with CF001 or CF002 (P = 0.107 and P = 0.142, respectively).

FIG 4.

Scnn1b-Tg mice carry higher bacterial burdens when infected with CF P. aeruginosa isolates. Scnn1b-Tg and wild-type mice were intranasally infected with the CF isolate CF001, CF002, or CF1188 and sacrificed on day 7 postinfection. Bars represent geometric means with 95% confidence intervals. CFU values were log transformed and analyzed using two-tailed unpaired t tests (CF001 data are from 2 independent infections with 16 Scnn1b-Tg mice and 16 wild-type mice; CF002 data are from 1 infection with 9 Scnn1b-Tg mice and 7 wild-type mice; CF1188 data are from 1 infection with 6 Scnn1b-Tg mice and 8 wild-type mice). The limit of detection is 100 CFU/g of lung. ***, P < 0.001.

To model CF lung infection with a P. aeruginosa isolate that displayed specific CF lung adaptations, including mucoidy, mice were infected with 1 × 10⁵ CFU of the CF isolate CF1188, a mucoid isolate from an 18-month-old CF patient. In contrast to infections with PAO1 and the nonmucoid early CF isolates CF001 and CF002, mice infected with CF1188 displayed decreased signs of disease in the first 3 days after infection (less ruffled fur, no hunched posture); however, CF1188-infected Scnn1b-Tg mice gradually lost weight during the course of the infection. Wild-type BALB/c mice did not show significant weight loss (data not shown). On day 7, a geometric mean of 1.5 × 10⁸ CFU/g of lung was recovered from Scnn1b-Tg mice, whereas the geometric mean in BALB/c mice was below the limit of detection (P < 0.0001) (Fig. 4). In two of six Scnn1b-Tg mice, bacteria were recovered from the kidneys, indicating dissemination of CF1188 from the lungs of Scnn1b-Tg mice, which did not occur in wild-type mice. Repeated infections with CF1188 yielded similar results (data not shown).

Cytokine production in the lungs of mice infected with all three CF clinical isolates was compared to that in mice infected with PAO1 on day 7 postinfection. Generally, mice infected with the nonmucoid early CF isolates, CF001 and CF002, mounted cytokine responses similar to those of mice infected with PAO1, while mice infected with the mucoid CF isolate CF1188 generated increased cytokine responses. As observed in PAO1-infected Scnn1b-Tg mice on day 7, no significant increases in tumor necrosis factor alpha (TNF-α), IL-1β, IFN-γ, IL-13, IL-17, or IL-22 were observed in Scnn1b-Tg mice infected with CF001 or CF002. In contrast, Scnn1b-Tg mice infected with either CF001, CF002, or CF1188 produced significantly more KC than mock-infected controls (P = 0.013, 0.035, and 0.006, respectively). Scnn1b-Tg mice infected with the mucoid isolate CF1188 produced significantly more IL-1β, IL-13, IL-17, IL-22, and KC than mice infected with PAO1, CF001, or CF002 (P = 0.005, 0.023, 0.010, 0.010, and 0.001, respectively) (Fig. 5). In wild-type mice, infection with CF isolates resulted in a significant increase in IL-13, compared to infection with PAO1 (P = 0.007), and decreases in the production of IL-1β, TNF-α, IL-17, and KC (P = 0.004, 0.006, 0.016, and 0.021, respectively).

FIG 5.

Scnn1b-Tg mice generate elevated cytokine responses after infection with CF1188. Bars represent means with standard deviations. Differences between groups were analyzed with Kruskal-Wallis tests and Dunn’s multiple-comparison post hoc tests. Horizontal bars and asterisks represent the results of Dunn’s multiple-comparison tests (5 to 6 mice per group). *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Histology.

As previously reported (37), Scnn1b-Tg mouse lungs were characterized by mucus accumulation in the bronchioles, neutrophilic infiltration, and enlarged alveoli, compared to those of wild-type littermates, even in the absence of infection (Fig. S3). On day 7 postinfection, H&E-stained Scnn1b-Tg mouse lungs showed slightly increased immune infiltration into the bronchioles (Fig. 6). Both wild-type and Scnn1b-Tg mice displayed alveolar septal thickening and cellular debris in the alveoli in response to PAO1 infection. Scnn1b-Tg mice infected with CF001 and CF002 showed similar signs of inflammation on day 7 (Fig. S4). Both neutrophils and alveolar macrophages were present in the mucus in Scnn1b-Tg mice, and individual bacteria and small bacterial aggregates were located in the bronchioles and alveoli using immunohistochemistry (Fig. S5). In contrast, Scnn1b-Tg mice infected with CF1188 showed extensive immune cell infiltration into the bronchioles and some alveoli. Wild-type mice, on the other hand, showed minimal pathology (Fig. 7). In order to locate CF1188 in the airway, lung sections were stained with anti-P. aeruginosa antibodies. Aggregates of bacteria were identified both in bronchioles and in alveoli, extending to the edge of the lung parenchyma (Fig. 8; Fig. S5). Although individual bacteria were observed, CF1188 was primarily found in aggregates that ranged in size from approximately 2 to 20 μm in diameter.

FIG 6.

PAO1 infection elicits increased immune infiltration in the bronchioles of Scnn1b-Tg mice (C and D) on day 7 postinfection. Wild-type mice (A and B) display signs of inflammation but lack excessive neutrophil recruitment. Bars, 400 μm (A and C) and 100 μm (B and D).

FIG 7.

Scnn1b-Tg mice show pronounced immune cell infiltration into the bronchioles on day 7 following infection with CF1188 (C and D). Wild-type mice show minimal pathology (A and B). Bars, 400 μm (A and C) and 100 μm (B and D).

FIG 8.

(A) CF1188 (red) formed aggregates in the lungs of Scnn1b-Tg mice, visualized using immunohistochemistry. Bacteria were localized using an anti-P. aeruginosa antibody with a DAPI counterstain. (B) Negative control. Bar, 10 μm.

DISCUSSION

Here, we showed the applicability of Scnn1b-Tg BALB/c mice as a model of P. aeruginosa colonization of the CF lung. In these studies, Scnn1b-Tg mice remained infected longer and showed increased neutrophilic infiltration into the bronchioles, compared to wild-type mice. In addition, Scnn1b-Tg mice generated Th17-type cytokines in response to infection, which is consistent with observed responses in CF patients.

We found that PAO1 consistently colonized the lungs of Scnn1b-Tg mice when it was grown as a biofilm prior to infection rather than planktonically. The formation of bacterial biofilms in the lungs of CF patients plays an important role in the persistence of lung infections, and bacteria have been observed in microcolonies and aggregates in sputum (46, 47). Biofilm growth offers multiple advantages to a pathogen, such as the ability to better evade the host immune response and tolerate antibiotic exposure. Although aggressive antibiotic treatment can be beneficial in treating CF patients’ airway disease, chronic P. aeruginosa infections are not eradicated by antibiotic treatment. Biofilm-related antibiotic tolerance has been attributed to reduced penetration into the biofilm and decreased metabolic activity of bacteria within the biofilm (48). Although the role of biofilms in maintaining chronic disease is well established, a potential role for bacterial aggregates in the initiation of disease is not as well understood. Vibrio cholerae, for example, is more infectious as a biofilm than it is as a planktonic organism (49, 50), but the physical structure of the aggregate has been found to be dispensable (51). Future studies will address the features of the PAO1 biofilm that enhance the colonization of Scnn1b-Tg mouse lungs.

We extended these studies by using P. aeruginosa isolates (CF001, CF002, and CF1188) collected from young patients with CF. When P. aeruginosa colonizes the CF lung, it encounters a stressful environment and must evade the immune response and survive exposure to antibiotics (52) and osmotic and oxidative stresses (53, 54). To survive and establish a chronic infection, P. aeruginosa must adapt to the CF lung environment. Several recurring adaptations have been identified from genotypic and phenotypic studies of CF isolates. For example, P. aeruginosa becomes nonmotile (55) and hypermutable, loses quorum sensing (56), and becomes increasingly antibiotic resistant (57). The development of mucoidy is especially important in the transition to a chronic infection in CF (58).

All three clinical isolates used in this study were obtained from individual patients 18 months old or younger; however, CF1188 is mucoid and CF001 and CF002 are nonmucoid. Infections in Scnn1b-Tg mice with PAO1, CF001, and CF002 were similar with regard to the bacterial numbers recovered from the lungs and the cytokine response at day 7 postinfection. Scnn1b-Tg mice carried higher bacterial burdens than wild-type mice when infected with P. aeruginosa CF isolates. Notably, infection with the mucoid isolate CF1188 resulted in the highest bacterial burdens (>10⁸ CFU) in Scnn1b-Tg mice on day 7. Correspondingly, Scnn1b-Tg mice infected with CF1188 mounted increased cytokine responses, compared to mice infected with PAO1, CF001, or CF002. On histological examination of hematoxylin-and-eosin (H&E)-stained lungs, Scnn1b-Tg mice infected with PAO1, CF001, or CF002 all displayed signs of inflammation, including neutrophil infiltration, lymphoid hyperplasia, the appearance of foamy alveolar macrophages, and thickening of the alveolar septa. Scnn1b-Tg mice infected with CF1188 additionally showed extensive immune cell infiltration into the bronchioles. Although CF1188 proliferated in the Scnn1b-Tg lung and caused dramatic cytokine responses, wild-type mice cleared CF1188 infections and showed minimal pathology, a dichotomy that illustrates the significance of CF-like pathology in permitting mucoid P. aeruginosa to establish lung infections.

In Scnn1b-Tg mice infected with PAO1, IL-17 was elevated in lung homogenate on day 3 but had returned to baseline by day 7, suggesting that innate-like cells could be responsible for much of the production of IL-17. Invariant natural killer cells, type III innate lymphoid cells, and γδ T cells have all been shown to secrete IL-17 and are present at mucosal surfaces (59, 60). During infection with CF1188, IL-17, IL-22, and KC remained elevated at day 7, likely due to the high bacterial load in the lung. IL-17-family cytokines have been found in the sputum of CF patients during exacerbations (43), and neutrophils from P. aeruginosa-infected CF patients produce IL-17 during infections (61). As previously shown in other murine models, the IL-17–IL-22 axis plays an important role in both the control of infections and the repair of epithelium following inflammation. Mice deficient in IL-17 or IL-23 production are more vulnerable to respiratory infection with bacterial (59, 60) and fungal (62) pathogens; however, in CFTR knockout mice, the administration of IL-17-blocking antibodies prior to infection with P. aeruginosa-laden agarose beads resulted in decreased lung pathology and less weight loss (63). In the highly inflamed CF lung, blocking the actions of IL-17 to decrease neutrophil infiltration and control inflammation is a potential therapeutic strategy.

In Scnn1b-Tg mice infected with PAO1, P. aeruginosa was primarily identified in the bronchioles as individual organisms. In contrast, in the lungs of Scnn1b-Tg mice infected with the mucoid isolate CF1188, aggregates of CF1188 were found in both bronchioles and alveoli. The observed CF1188 aggregates were approximately 20 μm across their longest axis. Infecting mice with a mucoid CF isolate appears to produce bacterial aggregates similar in size to those observed in both the alginate bead model and in CF patients. In the alginate bead model of P. aeruginosa lung infections, aggregates are approximately 23 to 342 μm² (64). P. aeruginosa aggregates have been reported to range in size from 4 to 3,227 μm² in explanted CF patient lungs (65), and in a review comparing characteristics of biofilms in human infections and in vitro models, aggregates in CF patients were reported to be in the range of 5 to 100 μm long (66).

This study had several limitations. Although Scnn1b-Tg mice maintained P. aeruginosa infections for 7 to 12 days, we did not assess whether an infection would continue for a longer duration. By day 12 postinfection, half of the infected Scnn1b-Tg mice had cleared the infection, but it is possible that some of the remaining infected mice would have maintained the infection longer. Because Scnn1b-Tg mice do not carry CFTR mutations, this model is also limited in its utility for the study of CFTR-targeting therapies.

Advantages of utilizing Scnn1b-Tg mice as models of CF lung infections include the ability to study microbial pathogenesis and the host immune response in a mucus-obstructed lung that recapitulates many of the features of CF lung pathology. This model can potentially be adapted to evaluate antibiotics or immunomodulatory treatments or to assess virulence of isolates in a CF-like lung. Future studies will expand this model to characterize other important CF-associated pathogens and study bacterial interactions in polymicrobial infections.

MATERIALS AND METHODS

Animals.

Wanda O’Neal and the Marsico Lung Institute Mouse Models Core at the University of North Carolina School of Medicine generously provided male Scnn1b-Tg (also known as β-ENaC) (37) BALB/c mice for the initiation of the colony. Scnn1b-Tg mice are reviewed in references 67 and 68. Female BALB/c mice were purchased from Taconic Biosciences (Rensselaer, NY) as breeders. Scnn1b-Tg mice were bred as hemizygotes, and their wild-type BALB/c littermates served as controls. Pups were genotyped by PCR of genomic tail DNA as previously described (37). DNA was extracted using a DNeasy blood and tissue kit (Qiagen, Hilden, Germany) per the manufacturer’s instructions. All mice were housed in the animal facility at the University of Maryland School of Dentistry. The Institutional Animal Care and Use Committee at the University of Maryland, Baltimore, MD, approved the experiments.

Bacterial strains and cultures.

Bacteria were stored at −80°C in 20% glycerol (vol/vol). Thomas Bjarnsholt (University of Copenhagen) provided the isolate of PAO1 (69) used in this study. CF isolates were from young CF patients enrolled in a study at the Children’s Hospital and Regional Medical Center in Seattle, Washington (44). The children were younger than 15 months old at enrollment and were followed until age 3. P. aeruginosa isolates were obtained from bronchoalveolar lavage fluid or oropharyngeal cultures. Strain information for all P. aeruginosa isolates used for the study is listed in Table 1.

TABLE 1.

Characteristics of strains used in this study

Strain	Origin	Patient no.	Mucoidy	Reference(s)
PAO1	Wound; Melbourne, Australia, 1954	NAa	No	69
CF001	CF patient; 3 months	003	No	44, 45
CF002	CF patient; 4 months	001	No	44, 45
CF1188	CF patient; 18 months	008	Yes	44, 45

^aNA, not applicable.

Bacteria were maintained on tryptic soy agar (TSA; Sigma-Aldrich Corporation, St. Louis, MO) plates and cultured overnight in tryptic soy broth (TSB; Sigma-Aldrich) at 37°C (shaking at 200 rpm). To generate biofilm-grown cultures, the overnight cultures were diluted 1:100 and incubated for an additional 3 h to reach mid-log phase, and then 10 μl of the subculture was spiked into wells of a 6-well polystyrene tissue culture plate (Corning Inc., Corning, NY) with 3 ml TSB. The plate was cultured overnight at 37°C without shaking. The resulting biofilm and planktonic populations were harvested together from each well of the plate; thus, the biofilm inoculum was a mixed population containing both planktonic bacteria and biofilm aggregates. The bacteria were centrifuged (4,200 × g for 10 min), washed with 10 ml phosphate-buffered saline (PBS; Sigma-Aldrich), and then centrifuged again. The pellet was resuspended in PBS to an optical density at 600 nm (OD₆₀₀) of 0.065 ± 0.003 (for PAO1, CF001, and CF002), which had been previously determined to correspond to a bacterial concentration of approximately 7 × 10⁷ CFU/ml. CF1188 was capable of establishing an infection with a lower-concentration inoculum, so the inoculum was diluted an additional 30-fold from an OD₆₀₀ of 0.065. The biofilm inoculum was vortexed vigorously for 1 min to disrupt visible bacterial aggregates. Following infection, the remaining inoculum was homogenized for 1 min with a Kinematica Polytron PT1200E tissue homogenizer (Kinematic Inc., Lucerne, Switzerland). The inoculum was serially diluted and plated in triplicate to confirm the bacterial concentration. The planktonic inoculum was prepared by growing an overnight culture in TSB (shaking at 200 rpm), diluting it 1:100, and incubating it for an additional 3 h. The subcultures were centrifuged and washed with PBS in the same manner as the biofilm inocula and then diluted to the same OD₆₀₀. The preparation of the biofilm and planktonic inocula is summarized in Fig. S1A.

Lung infection model.

Seven to 10-week-old male and female mice were anesthetized with inhaled isoflurane (VetOne, Boise, ID), intraperitoneal ketamine (Putney Inc., Portland, ME), and xylazine (VetOne) (100 to 150 mg/kg of body weight and 10 to 16 mg/kg, respectively) and then infected intranasally with P. aeruginosa suspended in 50 μl PBS. Mice were euthanized by CO₂ narcosis and cervical dislocation on days 3, 7, and 12 postinfection. Mock-infected mice were anesthetized in the same manner, intranasally inoculated with 50 μl sterile PBS, and sacrificed on day 3 after mock infection. Mice were monitored daily for the duration of the infection, and moribund mice were euthanized. Only surviving mice were included in CFU calculations, calculations of percentage of mice infected, and cytokine analysis.

Quantitative bacteriology.

Right lungs were excised aseptically, weighed, and homogenized in PBS (3 ml PBS/g lung tissue) using a Kinematica Polytron PT1200E tissue homogenizer. Lung homogenate was serially diluted and plated in triplicate on TSA and on Pseudomonas-selective agar (CHROMagar, Paris, France) to determine bacterial load. Plates were incubated at 37°C overnight. Colonies were counted and presented as CFU per gram of lung to standardize values across different mouse sizes. The limit of detection was 100 CFU/g.

Cytokine production.

Right lungs were weighed and homogenized in PBS (3 ml/g lung tissue) containing a cOmplete protease inhibitor cocktail (Roche, Basel, Switzerland) and 2% bovine serum albumin (BSA; AmericanBio Inc., Canton, MA). The lung homogenate was centrifuged at 4,200 × g for 10 min at 4°C, and the supernatant was stored at −20°C until analysis. The University of Maryland Cytokine Core Laboratory performed all cytokine assays using a Luminex multianalyte system.

Histology.

Left lungs were removed aseptically and harvested directly into 10% (wt/vol) neutral buffered formalin (Sigma-Aldrich). After a minimum of 48 h in formalin at room temperature, the lungs were embedded in paraffin wax and cut into 5-μm-thick sections, followed by hematoxylin-and-eosin (H&E) staining. Sectioning and H&E staining were performed by the Pathology, EM, and Histology Laboratory at the University of Maryland. Slides were analyzed by workers who were blind to the mouse genotype and infection status.

Immunohistochemistry was used to locate P. aeruginosa in the airways of infected mice. In brief, paraffin-embedded lung sections were deparaffinized and rehydrated in water before undergoing antigen retrieval (sodium citrate buffer, pH 6.0) at 60°C overnight. Nonspecific staining was blocked with 2% BSA and 0.1% Triton X-100, and slides were stained overnight with a rabbit anti-P. aeruginosa antibody (Abcam 68538; Abcam, Cambridge, United Kingdom) at a 1:400 dilution. The slides were washed, stained with a goat anti-rabbit secondary antibody conjugated to Texas Red (Abcam 6719) at a 1:400 dilution, and then counterstained with DAPI (4′,6-diamidino-2-phenylindole) Prolong Diamond (Invitrogen, Carlsbad, CA). Mock-infected mice and slides from infected mice with the primary antibody replaced with an additional blocking step served as negative controls. Slides were visualized using a Zeiss Axio Imager fluorescence microscope equipped with an ApoTome module.

Statistical analysis.

GraphPad Prism 5.0 (GraphPad Software, Inc., San Diego, CA) was used for graph creation and statistical analysis. t tests were used to analyze normally distributed data. Nonnormally distributed data were log transformed prior to performance of t tests. Kruskal-Wallis tests with Dunn’s multiple comparison post hoc tests were used to assess differences between more than two groups when variances were not equal. Differences in mortality were assessed with log-rank tests. Fold changes in cytokine expression were calculated for each mouse compared to mock-infected wild-type mice. A P value less than 0.05 was considered statistically significant.