Introduction

Abstract Introduction Materials and Methods Results Discussion References

Cystic fibrosis (CF) is the most common autosomal recessive disease in the Caucasian population. In 1944, Farber offered a unifying explanation for the complex pathology associated with CF when he noted that changes in the various organs of CF patients were all secondary to the obstruction of mucus-secreting glands (1), including those in the respiratory tract, the biliary tract, the duodenum, and the sublingual glands. As meconium ileus and pancreatic insufficiency have been treated with increasing success through surgical intervention and enzyme replacement therapy, respectively, respiratory failure has become the most life-threatening clinical manifestation of CF.

CF is caused by mutations in the gene encoding the cystic fibrosis transmembrane conductance regulator (CFTR), a cAMP-regulated chloride channel (2, 3). When this gene was cloned in 1989, it became possible for the first time to create animal models of CF by gene targeting in mouse embryonic stem (ES) cells. We used this technique to create a mouse line (Cftr^m1Unc) in which the Cftr gene was inactivated by introduction of an in-frame stop codon (termed allele S489X) that resulted in the production of a truncated gene product similar to that seen in some human CF patients (4). In humans this type of mutation results in a severe phenotype, similar to that caused by the more prevalent F508 mutation. Several other groups have now reported the generation of similar mouse lines, including one carrying the F508 mutation (5, 6).

Because of the complex phenotype associated with CF, it is not surprising that the various symptoms associated with the human disease vary in the extent to which they have been reproduced in mice homozygous for the targeted Cftr locus. The most obvious pathology in Cftr^/ animals is meconium ileus, which results in the death of pups shortly after birth (4). Obstruction of the intestinal tract followed by rupture and peritonitis is also seen, especially at weaning. Histologic examination of the obstructed intestinal tract in CFTR-deficient mice has revealed altered glandular secretions, which leads to blockage and eventual destruction of the glands.

Although Cftr^/ mice show some changes in the upper airways similar to those seen in human CF patients, they do not develop the pathologic changes in the lower airways that characterize the human disease. In addition, pathogens similar to those routinely isolated from the respiratory tract of humans with CF are not found in Cftr^/ mice. One possible reason for the lack of airway disease in Cftr^/ mice is the difference between human and mouse in the ability of the airway epithelium to produce mucus.

Mucus is believed to be an important component of the fluid barrier that coats the surface of the airway epithelium and protects the airways against damage by airborne irritants, particles, and microorganisms (7, 8). Foreign particles are trapped in this fluid barrier and removed from the respiratory tract by mucociliary transport. The efficiency of this process is determined primarily by the viscosity and elasticity of the fluid barrier, and these properties are conferred on the barrier by the high-molecular-weight glycoconjugates referred to as mucus glycoproteins or mucins. Mucins are believed to be produced primarily by specialized cells in the epithelium, the most prominent of which are goblet cells. After secretion, these molecules mix with water, ions, and other components to form mucus. In humans, airways respond to the presence of many foreign bodies through the production of mucus by both submucosal glands and goblet cells, and prolonged exposure to the irritant results in an increase in the number of goblet cells in the airway epithelium (7, 8).

Exposure of airways to noxious substances results in an inflammatory response and increase in mucus production in both normal and CF patients. It is possible, however, that because of the absence of CFTR in the airway epithelia of the CF patient, a concomitant increase in fluid secretion necessary for proper hydration of the mucus may not occur. Rather than facilitating the clearance of pathogens, the resulting dehydrated mucus provides an environment for their growth and protects them from the immune system. The airways respond to the continued presence of foreign substances with a prolonged inflammatory response, goblet cell proliferation, and mucus hypersecretion that eventually extends to the small airways. In the CF patient, this mucus production may therefore be central to continuing cycles of infection, primarily by Staphylococcus aureus and Pseudomonas aeruginosa. This in turn results in airway disease characterized by obstruction of small airways with mucus plugs and inflammatory cells, dilation of bronchi and bronchioles, and eventually the replacement of cells lining the lower airways with fibrous tissue. These changes lead to a gradual decline in pulmonary function and ultimately to death.

In the normal mouse airway, mucus-producing glands are less numerous than in humans and do not extend distal to the first few cartilaginous rings of the trachea. Furthermore, goblet cells, although prevalent in the nasal passage, are virtually absent from the trachea and relatively sparse in the proximal bronchi. Differentiation of precursor cells into goblet cells is also rare in the mouse, even after irritation by exposure to infectious agents (9).

To test the hypothesis that absence of airway disease in Cftr^/ mice is due to differences in the capacity of the human and mouse airway epithelium to produce mucus, we have induced allergic airway disease in these animals. To do this, we have modified existing protocols to allow rapid and simultaneous treatment of large groups of mice. Extensive goblet cell metaplasia and hyperplasia extending into secondary airways characterizes the airway disease induced by this method. The increase in goblet cell numbers begins proximally and extends distally in a pattern similar to that observed in most human airway diseases, including CF. Airways obstructed by mucus are also common in mice subjected to this protocol, and these changes are relatively consistent among the treated animals. In this report we describe experiments in which this technique is applied to Cftr^/ mice to determine whether increased mucus production (1) leads to CF-type respiratory pathology in Cftr^/ mice, or (2) alters the clearance of bacteria in Cftr^/ mice relative to normal controls.

	Materials and Methods

Abstract Introduction Materials and Methods Results Discussion References

Animal Husbandry

BALB/c, DBA/2, C57BL/6, and B6D2 mice used for morphometric quantitation of goblet cell hyperplasia were purchased from Jackson Laboratories (Bar Harbor, ME). Cftr^m1UNC mice generated in the University of North Carolina (UNC) animal facility are a hybrid strain containing genetic material from C57BL/6, 129/SvEv, Balb/c, and DBA/2 mice (4). All mice are housed in a barrier unit in sterile microisolator cages on autoclaved, recycled paper bedding. The health of the colony is routinely monitored, and to date the mice in this colony have tested negative for all mouse pathogens for which tests are available. This has included serologic analysis for titers to mouse hepatitis virus (MHV), minute virus of mice (MVM), Mycoplasma pulmonis, mouse orphan parvovirus (MOPV), pneumonia virus of mice (PVM), Sendai virus, Theiler's mouse encephalomyelitis virus (GDVII), epizootic diarrhea of infant mice (EDIM), and lymphocytic choriomeningitis (LCM). In addition, mice are monitored three times a year for the presence of pinworms and mites.

The mouse diet consisted of solid irradiated Prolab rodent chow (RHM2000; Agway, Syracuse, NY), and Colyte (Schwarz Pharma, Inc., Mequon, WI) to prevent intestinal obstruction. Tail-clip samples were obtained from 12-d-old mice and processed for genotyping by polymerase chain reaction (PCR), as described previously (4). Cftr^+/ and Cftr^/ mice used for aerosolization and deposition of bacteria were between 2 and 4 mo of age. Each Cftr^/ mouse was age-matched with a Cftr^+/ control, and study groups included both male and female animals. The number of mice included in each experiment is indicated in the figure legends.

Induction of Mucus in Mouse Lower Airways

On Day 0, mice were sensitized with intraperitoneal injections of chicken ovalbumin (100 µg ovalbumin and 1 mg alum in 0.5 ml sterile saline per mouse). Control mice received intraperitoneal injections of saline. Starting on Day 14, the mice were challenged by exposure to aerosolized saline or 1% ovalbumin (using a modified mouse cage and a compressor/nebulizer [Model 5650]; Devilbiss, Somerset, PA) for 30 min daily for 7 or 14 d. After daily aerosol treatment for 14 d, Cftr^+/ and Cftr^/ mice used in the chronic aerosol study were aerosolized three times per week for up to 6 mo.

Growth and Deposition of Bacteria

A 200-ml culture of P. aeruginosa (strain PA01) was grown overnight, and 3-ml aliquots were stored at 80°C. On the day of deposition, an aliquot was thawed and grown in 50 ml of yeast/tryptone medium until the OD₆₀₀ = ~ 0.50 (about 4 h). The cells were spun down and resuspended in phosphate-buffered saline (PBS) such that the OD₆₀₀ = 0.50. This stock preparation was then diluted as indicated for deposition. The final dilution for deposition was plated out in duplicate on sheep blood agar plates to determine the concentration of the stock. Five isolates of S. aureus were individually grown in tryptone soy broth, harvested during the logarithmic growth phase, and resuspended in PBS at an OD₆₅₀ = ~ 1.4. Equal volumes of each isolate were combined for deposition into the mouse airways. The numbers of bacteria were enumerated by plating on mannitol salt agar. For deposition, mice were anesthetized by intraperitoneal injection with 0.015-0.017 ml of 2.5% avertin per gram of body weight (100% avertin is made by mixing 10 g of tribromoethyl alcohol in 10 ml of tertiary amyl alcohol, and is diluted in isotonic saline to a 2.5% stock). Fifty microliters of the P. aeruginosa or S. aureus dilution was deposited tracheally.

Experiments in which mice were exposed to bacteria during the induction of airway disease were done in the following manner. On the day following the initial aerosol exposure, the second aerosol exposure was omitted, and instead a low dose of P. aeruginosa was deposited into the trachea. To estimate the number of organisms deposited, four animals were sacrificed immediately following deposition, and the right lobes of the lung of each animal were homogenized and plated on sheep red blood agar as described previously. This revealed that on average, 2.25 × 10⁴ cfu were deposited into the right lobes of the lung of each animal. To increase the chances of establishing lower-respiratory infections, the fifth aerosol exposure was also replaced by administration of a low dose of bacteria. The number of bacteria deposited was again quantified, and in this case it was determined that each animal received an average of 8.5 × 10³ cfu in the right lobes. On the day following the final aerosol exposure, a higher dose of bacteria was administered; 1.45 × 10⁵ cfu were recovered on average from the right lobes, and lungs were harvested 0, 24, and 48 h after this final deposition.

Lung Analysis

Mice were exsanguinated after receiving an intraperitoneal injection of a lethal dose of chloral hydrate (1 ml of a 20 mg/ml solution) by severing the aorta. The left lobe of the lung was removed and inflated with 10% phosphate-buffered neutral formalin (pH 7.3), and further immersed in formalin for at least 24 h. Lungs were embedded in paraffin wax and longitudinally sectioned (3 µm) for histologic analysis with hematoxylin and eosin (H&E) and periodic acid-Schiff/alcian blue (PAS/AB) (pH 2.5) staining. The remaining four right lobes were homogenized in 2 ml sterile PBS and diluted in PBS over a range of 1:10 through 1:10⁵, and 100 µl of each dilution were plated on sheep blood agar (for P. aeruginosa analysis) or mannitol salt agar (for S. aureus analysis). Plating was done in duplicate over a range of dilutions such that colonies could be accurately counted on a plate. For all bacterial clearance measurements, bacteria counted were recovered from the right lobes. The left lobe was fixed for histologic examination of goblet cell hyperplasia and the inflammatory response induced by a given protocol.

Morphometric Analysis

Lungs were perfused by in situ inflation with 10% phosphate-buffered neutral formalin, pH 7.0, at a fixative pressure of 20 mm. The trachea was clamped and the lungs and the trachea were excised and immediately immersed in 10% phosphate-buffered neutral formalin for at least 24 h. Serial sections of the left lobes of the lungs were cut, and those that yield maximum longitudinal visualization of the intrapulmonary main axial airway were chosen and stained with PAS/AB. To avoid bias for a certain region, and to consistently view the identical region in all slides, a 2-mm length of airway, located midway along the length of the main axial airway, was examined in all animals. Morphometric studies were completed for one section per mouse.

The morphometry system used is based on a Nikon FXA microscope (Nikon Inc., Garden City, NJ) equipped with an Optronics TEC-470 CCD Video Camera System (Optronics Engineering, Goleta, CA). Images were captured on a Macintosh 840AV computer (Apple Computer, Cupertino, CA), using a Scion LG-3 capture card (Apple). Image processing, analysis, and measurements were done with the public domain NIH Image program (U.S. National Institutes of Health, National Technical Information Service, Springfield, VA). For each magnification used in our study, NIH Image uses a spatial calibration based on a digitized image of a stage micrometer. For a ×10 objective lens (used in this study) the magnification on the computer screen was ×950. Using NIH Image, we measured the area and length of the PAS/AB-stained region in the section. The data are expressed as the mean volume density (V_s = nl/mm² basal lamina) ± SEM of PAS/AB-stained material within the epithelium, as described in Harkema and associates (10).

Detection of Mouse Gastric Mucin Gene by Reverse Transcription-Polymerase Chain Reaction and In Situ Analysis

For detection of the mouse gastric mucin gene (Mgm), RNA was isolated from the lungs of Cftr^+/ and Cftr^/ mice that had been sensitized and challenged with either ovalbumin or saline, using RNAzolB (TelTest, Inc., Friendswood, TX) according to the protocol supplied by the manufacturer. RNA was also isolated from mouse stomach as a positive control. Reverse transcription-polymerase chain reaction (RT-PCR) was completed using the Invitrogen (San Diego, CA) cDNA Cycle Kit and the primers MGM1A, 5'-CAA TTG GCT AGA TGG CAG TTA CCC-3'; and MGM1B, 5'-CTC TCC GCT CCT CTC AAT GTT-3'. The resulting product was cloned into the pCRII vector and sequenced. Using this construct, ³⁵S-labeled sense and antisense RNA (Maxiscript SP6/T7 kit; Ambion, Austin, TX) were made and hybridized to sections of inflated lungs isolated from saline-treated and ovalbumin-treated mice. The methods used for in situ hybridization were similar to those previously described (11, 12). Lungs were inflated with 50% Tissue-Tek (Sakura Finetek U.S.A. Inc., Torrance, CA) and immersed in a block containing Tissue-Tek, and the block was immersed in liquid nitrogen. Cryostat sections (8 µm) were mounted on slides and stored at 80°C until use. Sections were fixed in 4% paraformaldehyde and then dehydrated in a graded series of ethanol solutions before being digested with proteinase K (10 µg/ml) at 30°C for 30 min. Proteinase K was inactivated with 4% paraformaldehyde and rinsed in triethanolamine (TEA) prior to acetylation with 0.25% acetic anhydride for 10 min. Sections were then rinsed in 0.2× standard saline citrate (SSC) and dehydrated. A quantity of Mgm sense or antisense probe producing 1 × 10⁷ cpm was hybridized overnight at 54°C in 50% formamide; 1× Denhardt's solution; 0.6 M NaCl; 10 mM Tris, pH 8.0; 1 mM ethylenediamine tetraacetic acid (EDTA); 0.1% sodium dodecyl sulfate (SDS); 10 mM dithiothreitol (DTT); 1 mg/ml yeast transfer RNA (tRNA); and 10% dextran sulfate. Slides were prewashed with 4× SSC and then treated with 20 µg/ml ribonuclease (RNase) at 37°C, followed by four washes with 2× SSC/1 mM DTT and three washes at 54°C in 0.5× SSC/1 mM DTT. Following dehydration, slides were hand-dipped in NTB2 emulsion (Eastman Kodak, Rochester, NY), exposed for two weeks at 4°C, developed, and stained with H&E.

	Results

Abstract Introduction Materials and Methods Results Discussion References

Induction of Goblet Cell Hyperplasia in BALB/c, C57BL/6, DBA/2, and 129/SvEv Strains

The genetic background on which the S489X mutation is carried consists of a mixture of the BALB/c, C57BL/6, DBA/2, and 129/SvEv mouse strains. It was therefore necessary to rule out the possibility that any difference between control and experimental animals in the induction of allergic airway disease was due to genetic differences other than loss of expression of the Cftr gene. To accomplish this, we conducted experiments to establish that the induction of allergic airway disease and subsequent induction of goblet cell hyperplasia was similar in each of these strains.

Allergic airway disease was induced in mice from the BALB/c, C57BL/6, DBA/2, and 129/SvEv strains, as well as in B6D2 mice derived from a cross between the C57BL/6 and DBA/2 strains. Animals were killed, and the right bronchus was clamped to allow fixation without inflation of the right lobes in order to visualize mucus present in the airways. The left lobe was inflated at 20 mm of fixative pressure and subjected to quantitative analysis. Morphometric analysis of PAS/AB-stained slides of the main airway of the left lobe from mice exposed to aerosolized ovalbumin revealed a dramatic but similar increase in the volume of stored mucosubstances within the epithelium relative to that in saline-treated controls in all strains tested (Figure 1). The increase in stored mucosubstances correlated with an increase in the number of PAS/AB-staining goblet cells seen in this region of the airway. After 7 d of aerosol treatment, most of the cells in the bronchiole contained PAS/ AB-positive granules, and goblet cells were also present in the secondary airways, often in large numbers. Again no differences were observed between strains in the extent of this goblet cell hyperplasia. Interestingly, goblet cells remained sparse in the tracheal epithelium, especially in the proximal regions, which appear in the mouse to consist predominantly of ciliated cells. Examination of the airways of the uninflated right lobes revealed the presence of goblet cells as well as numerous airways obstructed with material, presumably mucus, that stained with PAS/AB. Staining of lung tissue with H&E in all mouse strains tested revealed the presence of myelocytic and lymphocytic infiltrates similar to those previously described for this model of allergic airway disease (13, 14).

View larger version (25K): [in this window] [in a new window]   Figure 1.   Quantitative analysis of changes in the amount of stored mucosubstance in the surface epithelium of the main bronchiole of mice of various genetic backgrounds after sensitization and challenge with ovalbumin. Sections through the main bronchiole of the left lobe of the lung were stained with PAS/AB, and the volume of PAS/AB-stained mucosubstance per square millimeter of basal lamina (Vs) was determined, as described in MATERIALS AND METHODS; n indicates number of animals examined in each group; bar indicates the SEM.

Induction of Goblet Cell Hyperplasia in Cftr^+/ and Cftr^/ Mice

Allergic airway disease was induced in Cftr^+/ and Cftr^/ mice. Ovalbumin treatment resulted in the appearance of goblet cells and mucin production in the airways of both the Cftr^+/ and Cftr^/ mice (Figure 2). Interestingly, however, no difference in the volume of stored mucosubstances between Cftr^+/ and Cftr^/ mice could be detected by morphometric examination of the main bronchiole of the inflated left lobe of the two groups of animals (Figure 3). In addition, histologic examination showed that the distribution and number of goblet cells in the secondary airways appeared similar in the two groups (Figure 2). Histologic examination suggested that this increase in number of goblet cells was paralleled by a decrease in the number of ciliated cells present in the airway epithelium. We have not addressed this observation with quantitative analysis of cell types present in treated and untreated animals, nor have we examined the mechanism underlying these morphologic changes. However, exfoliated epithelial cells, both goblet cells and ciliated cells, were seen occasionally in the airway lumen, usually entrapped in the mucus. Examination of the uninflated lobes of the right lung revealed numerous airways in which the lumen was partly obstructed by mucus, although again no difference was seen between the Cftr^/ mice and control animals. No goblet cells were seen in saline-treated controls of either genotype.

View larger version (86K): [in this window] [in a new window]   Figure 2.   Representative lung sections showing pathologic changes characteristic of allergic airway disease in ovalbumin-sensitized and aerosolized-challenged Cftr/ and control animals. In PAS/AB-stained lung sections from mice sensitized and challenged with saline (A and B), no goblet cells are detected. In contrast, in similar sections from both the control Cftr+/ (E and F) and Cftr/ (I and J) animals sensitized and challenged with ovalbumin, many goblet cells containing both neutral (magenta) and acid (purple) mucins are seen (arrows in F and J). In the saline-treated airway (B), the airway epithelium is composed primarily of ciliated cells. Although ciliated cells are easily observed in both the Cftr/ mice (J) and littermate controls (F), their numbers appear decreased in the ovalbumin-treated airways. Examination of low-power magnifications of PAS/AB-stained, uninflated lungs from both the Cftr/ mice (I) and control littermates (E) shows airways occluded by mucus plugs. H&E-stained sections of the ovalbumin-treated airways from both Cftr/ mice (K) and control littermates (G) reveal extensive inflammatory-cell infiltration into both perivascular and peribronchial sites (solid arrowheads in G and K). No inflammation was noted in the saline-treated animals (C). Eosinophils were easily detected in H&E-stained sections from lungs of Cftr/ mice (L) and littermate controls (H) on examination under high-power magnification (open arrowheads in H and L). Only the occasional alveolar macrophage was detected in similar sections from saline-treated animals (D). In A, C, E, G, I, K: bar = 200 µm; in B, D, F, H, J, L: bar = 50 µm.

View larger version (20K): [in this window] [in a new window]   Figure 3.   Quantitative analysis of changes in amount of stored mucosubstance in surface epithelium of the main bronchiole of CFTR-deficient and control mice after sensitization and challenge with ovalbumin. Sections through the main bronchiole of the left lobe of the lung were stained with PAS/ AB, and the volume of PAS/AB-stained mucosubstance per square millimeter of basal lamina (Vs) was determined, as described in MATERIALS AND METHODS; n indicates number of animals examined in each group; bar indicates the SEM.

In both Cftr^+/ and Cftr^/ mice, the presence of abundant secretions in the airways did not compromise the integrity of the adjacent epithelium. Sections of lung tissue from each group of animals were stained with H&E to determine whether treatment with ovalbumin invoked an inflammatory response of similar magnitude in the CFTR-deficient and control animals (Figure 2). Perivascular and peribronchial cuffing were evident in both control and Cftr^/ mice. This consisted of large numbers of eosinophils and lymphocytes, including plasma cells, monocytes, and some neutrophils. Inflammatory cells were also seen in the airways, and consisted primarily of eosinophils.

Mucin Expression in Ovalbumin-Treated Lungs

MUC5AC is one of the mucins produced in human airways. Recently, a gene isolated from a mouse stomach complmentary DNA (cDNA) library, referred to as Mgm, was found to have a very similar nonrepeat segment to MUC5AC (15). To determine whether induction of goblet cell hyperplasia and the presence of PAS/AB-staining material in the airways of Cftr^+/ and Cftr^/ mice is correlated with induction of Mgm expression in mice, and to determine whether induction of Mgm expression is similar in Cftr^/ and control animals, we performed RT-PCR analysis, using primers specific for the Mgm transcript. No PCR product was obtained when RNA from lungs of saline-treated mice was used as a template (data not shown). However, a PCR product was obtained with RT-PCR of ovalbumin-treated lungs, and sequence analysis confirmed that it was derived from Mgm messenger RNA (mRNA).

These results indicate that the appearance of PAS/AB-positive granules in airway epithelial cells correlates with induction of expression of at least one of the murine mucin genes. Although quantitation of the level of expression of the mRNA in normal and Cftr^/ mice was not done, the similar intensity of the PCR product amplified from these mice does not support any difference between these two groups of animals in their ability to induce mucin gene expression.

To verify that Mgm is expressed by airway epithelium and that its pattern of expression coincides with regions of the airway that by PAS/AB staining are shown to produce mucins, in situ analysis of sections of lungs from saline- and ovalbumin-treated airways were examined. Expression of Mgm in the lungs of ovalbumin-treated mice was easily detected. Surprisingly, expression of the gene could also be detected in the lungs of saline-treated animals that were shown by PAS/AB staining to contain virtually no goblet cells (Figure 4). The pattern of expression of Mgm paralleled that of the goblet cell hyperplasia, with highest expression seen in the primary airways and decreased expression in the secondary airways (Figure 4). Again no difference in Mgm expression was seen in the lungs of CFTR-deficient mice and their littermate controls. In situ analysis of sections of the mouse stomach indicates that expression of Mgm in the airway epithelial cells does not approach the high levels of expression seen in the mucus-secreting cells of the mouse stomach.

View larger version (135K): [in this window] [in a new window]   Figure 4.   Expression of the mucin gene Mgm in mouse airways after sensitization and challenge with ovalbumin. Sections through the main bronchiole of the left lobe of mouse lungs were prepared from mice that had been treated with either saline or ovalbumin, and were subjected to hybridization with 35S-labeled Mgm antisense probe. Expression of Mgm is easily detected in the epithelium of both the main bronchiole and secondary bronchioles of the saline-treated animals (A). Hybridization of the Mgm probe is more intense in the airway epithelium from ovalbumin-treated animals (C). In both groups, airway expression of Mgm is not seen in the terminal bronchioles or the alveolar regions. Brightfield microscopy of the sections stained with H&E is shown on the right. Levels of nonspecific hybridization were established by hybridization of a lung section with a 35S-labeled Mgm sense probe (E). Bar = 200 µm.

Goblet Cell Hyperplasia and Inflammation in Chronically Aerosolized Mice

To determine whether the failure to observe differences between Cftr^+/ and Cftr^/ mice reflects the relatively short time course of the allergic airway disease induced in our study, CFTR-deficient mice and their littermate controls were exposed to antigen over a period of 6 mo. Histopathology was monitored throughout this period. Sections were examined under conditions of blinding to genotype, and were scored on a scale of 1 to 3 for the presence and severity of pathologic changes, including focal and diffuse leukocyte infiltrates and goblet cell hyperplasia. Following 1 mo of treatment, lymphocytic and myelocytic infiltrate was mildly decreased, whereas the goblet cell hyperplasia remained unchanged. The reduction in inflammation as well as the number of goblet cells and the amount of mucus was similar in the airways of both groups of mice. In addition, the mucus in the airways of Cftr^/ mice did not appear to form the dehydrated concretions that are often observed in the airways of CF patients and in the intestinal crypts of the CFTR-deficient mice. Lungs examined 3 and 6 mo after the initiation of exposure of the sensitized mice to antigen indicated that the inflammatory-cell infiltrate and perivascular and peribronchial cuffing were markedly decreased. In addition, only a smattering of goblet cells was seen in the main bronchiole (Figure 5). The resolution of the inflammatory response was similar in both control and CFTR-deficient animals.

View larger version (71K): [in this window] [in a new window]   Figure 5.   Allergic airway disease resolves in a similar manner in sensitized Cftr/ mice and control littermates after chronic exposure to aerosolized ovalbumin. Representative lung sections stained with PAS/AB or H&E from a Cftr/ mouse (E through H) and a control littermate (I through L) exposed to aerosolized ovalbumin for 6 mo following sensitization, as compared with sections from lungs of animals exposed to aerosolized saline for a similar period (A through D). Although goblet cells (see arrows and F and J) are still more numerous in the PAS/AB-stained sections of ovalbumin-treated airways (E and I) than in the airways of saline-treated animals (A), the number of goblet cells present has been dramatically and equally reduced in the Cftr/ mice and control littermates as a result of chronic exposure to the antigen (compare with Figures 2E and 2I). This is confirmed by high-power magnification of the airway epithelial cells of both the Cftr/ mice (J) and littermate controls (F), and by comparison with similar sections from saline-treated animals (B). High-power magnification of similar sections stained with H&E indicates that after chronic exposure to allergen, the majority of airway epithelia of both the Cftr/ mice (L) and control animals (H), like those of the saline-treated animals (D), are composed of ciliated cells (open arrowheads in H and L). Low-power magnification of H&E-stained sections from Cftr/ mouse lungs (K) and lungs of control littermates (G) reveals a similar decrease in the inflammatory infiltrate (solid arrowheads in G and K) in both groups of animals (compare with Figures 2G and 2K). In airways of both Cftr/ mice and control littermates, small numbers of inflammatory cells were occasionally still present in some perivascular and peribronchial sites by comparison with saline-treated animals (C). In A, C, E, G, I, K: bar = 200 µm; in B, D, F, H, J, L: bar = 50 µm.

P. aeruginosa Clearance and Goblet Cell Hyperplasia

We next sought to determine whether the presence of mucus in the airways would reveal any difference between Cftr^+/ and Cftr^/ mice in susceptibility to bacterial infections. We also wished to determine whether earlier failures to see differences between CFTR-deficient and normal mice in clearance of S. aureus extended to other strains of bacteria (9). P. aeruginosa was chosen for these experiments because it is associated with respiratory disease in almost all CF patients, and the PAO1 strain was chosen because it is capable of colonizing mouse airways (16). In addition, the PAO1 strain also expresses many P. aeruginosa-associated pathogenicity factors, and mucoid transformation of this strain has been reported in a chemostat system under conditions designed to reflect those likely to be present during chronic infection in the lower airways of CF patients (17). An LD₅₀ for the P. aeruginosa strain PAO1 was established in mice of similar mixed genetic background to our Cftr^/ population. This value of ~ 3 × 10⁶ cfu/mouse is similar to that reported previously for the PAO1 strain in other strains of mice.

Goblet cell hyperplasia was induced in Cftr^+/ and Cftr^/ mice, and to ensure that mucus production was induced in all animals, exposure to the aerosolized antigen was increased to 14 d. P. aeruginosa was then administered to all animals by tracheal deposition (9 × 10⁴ cfu were deposited on average per mouse). In all experiments, numbers of bacteria remaining were determined for the right lungs only. These numbers are not extrapolated to those expected for the whole lung based on relative volumes, since deposition and clearance rates for the left lung may differ. Lungs were harvested at 0, 7, 24, and 48 h after deposition. The cultured homogenate revealed that bacteria were still present in the lungs at 7 h, but had been almost completely cleared by 48 h (Figure 6A). The rate of clearance in mice treated with ovalbumin was similar to that in the saline-treated controls, and did not differ between Cftr^+/ and Cftr^/ animals. Histopathologic changes were assessed by scoring of the sections obtained from the left lobe of the same group of animals. Again, scoring was done on a scale of 1-3, and evaluated both chronic airway disease resulting from exposure to antigen and acute histopathologic changes, such as edema and accumulation of neutrophils resulting from bacterial infection. Histologic examination revealed goblet cell hyperplasia and mucus production in all of the mice that were immunized and aerosolized with ovalbumin. Only a few scattered goblet cells were present in the mice that received bacteria after saline treatment. Histologic examination also revealed the presence of neutrophils at 7 and 24 h after exposure to bacteria, and of macrophages/monocytes at 48 h after exposure, as would be expected for an acute and resolving bacterial infection. No difference was noted in the scores obtained for these parameters of acute inflammation in ovalbumin-treated mice versus saline-treated controls, or in Cftr^+/ versus Cftr^/ mice. Mice in all treatment groups appeared lively and healthy within 24 h after bacterial deposition. In a second set of experiments, an identical protocol was followed, with the exception that the number of bacteria deposited was increased by 10-fold, to 10⁶ cfu/ mouse. Again, pathologic changes as determined by blinded scoring and clearance of bacteria were similar among all four experimental groups (Figure 6B).

View larger version (34K): [in this window] [in a new window]   Figure 6.   Effect of mucus on clearance of P. aeruginosa from airways of Cftr/ and Cftr+/ mice. A bolus of P. aeruginosa was inoculated into the tracheas of Cftr/ and Cftr+/ mice in which allergic airway disease and mucus production had been induced by treatment with ovalbumin. Cftr/ and Cftr+/ mice treated with saline received an identical bolus of bacteria. At the time noted, animals were killed, the right lobes of the lung were removed and homogenized, and bacteria were enumerated by plating. Each bar indicates the number of bacteria present in the right lobes of the lung from a single animal. In the experiment shown in A, analysis of animals immediately after deposition indicated that an average of 5.1 × 104 cfu had been deposited into the right lobes (0 h). The genotype and treatment of the animals did not markedly affect the initial deposition of bacteria. In the second experiment (B), the number of bacteria deposited (0 h) into the right lobes (average: 2.36 × 105 cfu) was established with normal, untreated animals. No difference in the rate of bacterial clearance was observed among the four groups of animals.

We considered the possibility that chronic infections of the lower airways were not observed in the experiments described here, despite the presence of mucus, because bacteria were prevented from reaching the lower airways by the presence of mucus plugs, and that perhaps repeated exposure to bacteria was required to establish a chronic infection in Cftr^/ mice. To examine this, we performed additional experiments in which P. aeruginosa was deposited prior to and throughout the development of goblet cell hyperplasia, which preliminary experiments had shown begins as early as 3 d after challenge with antigen. As in our initial experiments, this procedure did not reveal a significant difference in the rate of bacterial clearance between Cftr^+/ and Cftr^/ mice or between mice treated with saline or ovalbumin (Figure 7). In all cases, from 0 to 3 bacteria were detected on the bacterial plates plated with lung extracts at 48 h after infection. This corresponds to less than 50 bacteria in the right lungs of the mice at 48 h after deposition. Although these low residual numbers of bacteria were found in one of three normal mice and three of three Cftr^/ animals, the small numbers of bacteria recovered from the plates made it unlikely that a significant difference existed between the groups. Although we believe it to be unlikely, we cannot rule out the possibility that the analysis of large groups of animals may result in significant differences between Cftr^/ and control animals in the numbers of residual bacteria present in the airways. At this point, however, the data do not support the colonization of Cftr^/ animals with the bacteria used in this experiment.

View larger version (41K): [in this window] [in a new window]   Figure 7.   Clearance of P. aeruginosa from Cftr/ and Cftr+/ airways after multiple inoculations throughout the induction of allergic airway disease. Fourteen days after sensitization, Cftr/ and Cftr+/ animals were challenged by exposure to aerosolized antigen. Parallel experiments were performed using Cftr/ and Cftr+/ animals treated with saline. On the following day (Day 15), a bolus of bacteria was inoculated into the tracheas of the animals in the four groups. Animals were then exposed to aerosolized antigen or saline for two consecutive days (Days 16 and 17), and this was again followed by tracheal inoculation in all groups with P. aeruginosa (Day 18). This treatment was repeated for a final cycle. At the times noted, animals were killed, the right lobes removed and homogenized, and bacteria enumerated by plating. Each bar indicates the number of bacteria present in the right lobes of a single animal. The number of bacteria deposited into the right lobes of animals on Days 15, 18, and 21 was determined experimentally by the inoculation of an identical bolus of bacteria into the airways of untreated wild-type mice. On Day 21 (0 h), the number of bacteria in the lungs after deposition was also established for each of the experimental groups of animals, to determine the numbers of P. aeruginosa remaining from previous depositions (average number of bacteria recovered from the right lobes: 6.4 × 104 cfu). After 48 h, lungs from all four groups of animals had fewer than 200 bacteria.

Interestingly, by Day 21, ovalbumin-treated mice appeared to have cleared from their lungs all of the P. aeruginosa from prior depositions, and the amount deposited mirrored the amount deposited in previously uninfected mice. In addition, all groups of mice appeared to clear the bacteria from their lungs more quickly than the mice from prior experiments in which similar amounts of bacteria were deposited, suggesting that prior depositions of P. aeruginosa "primed" the mice to clear P. aeruginosa more effectively. Notably, the multiple depositions of bacteria did not induce goblet cell hyperplasia in saline-treated animals.

S. aureus Clearance and Goblet Cell Hyperplasia

To address further the question of whether increased mucus production would allow the establishment of chronic respiratory infections in Cftr^/ mice, we simultaneously infected mice with five isolates of S. aureus. These isolates were obtained from CF patients at the UNC Hospitals. Preliminary experiments in our laboratory with one of these isolates indicated that S. aureus does not normally colonize mouse airways. However, it is possible that this species difference from P. aeruginosa depends on factors such as the presence of large amounts of mucus in the airways.

Cftr^+/ and Cftr^/ mice were sensitized and challenged daily for 2 wk with saline or ovalbumin. We anticipated that prior bacterial infections might increase susceptibility to S. aureus, and therefore, on the day after the final aerosol exposure, we deposited P. aeruginosa (9 × 10⁴ cfu) into the tracheas of all mice. Two weeks were allowed for recovery from exposure to P. aeruginosa, and the animals were then subjected to another week of daily aerosol exposure. At the conclusion of this aerosolization procedure, the five strains of S. aureus were grown, combined, and inoculated (8 × 10⁷ cfu/mouse) into the tracheas of all mice. An average of 3.7 × 10⁷ bacteria was recovered from the right lobes of animals examined immediately after deposition. Lungs were harvested at 0, 7, and 24 h, and clearance was monitored as described for P. aeruginosa in the previous section. Clearance was rapid, with a 3-log decrease in bacterial numbers within 48 h, and the rate of clearance was similar in Cftr^+/ and Cftr^/ mice and was not affected by the presence of mucus in the airways (Figure 8).

View larger version (42K): [in this window] [in a new window]   Figure 8.   Clearance of S. aureus from Cftr/ and Cftr+/ airways after induction of allergic airway disease and mucus production. Cftr/ and Cftr+/ mice were sensitized to ovalbumin and challenged for a week by exposure to aerosolized antigen. Similar groups of animals were treated with saline. A bolus of P. aeruginosa (5 × 104 cfu recovered from the right lobes) was inoculated into the tracheas of all animals. Mice were allowed to recover for 1 wk. Mice that had been exposed to ovalbumin were again challenged by exposure to aerosolized antigen for 1 wk. After this period, all animals were inoculated with a bolus of S. aureus (8 × 107 cfu) composed of equal numbers of five independent S. aureus isolates. This resulted in the deposition of an average of 3.7 × 107 cfu into the right lobes. The animals were killed at the time indicated, and the number of bacteria remaining in the right lobe was determined. Each bar represents the bacteria recovered from a single mouse.

	Discussion

Abstract Introduction Materials and Methods Results Discussion References

The mechanism by which loss of CFTR function leads to lung disease has not been established, although it has been suggested that increased Na⁺ absorption and/or decreased Cl secretion reduces the volume of airway surface liquid (18, 19). If this is the case, the loss of CFTR function may not affect the health of the individual until mucus production is stimulated by exposure to noxious substances. Once this occurs, the decreased surface-liquid volume, and perhaps the inability to increase fluid secretion, could result in an inability to hydrate properly mucus produced in response to the irritant. The resulting abnormally viscous mucus may adversely affect airway function in a number of ways. Decreased mucociliary clearance rates might result in the inability to clear bacteria properly from the airways, and continued growth of bacteria in lower airways could stimulate a prolonged inflammatory response. The increased viscosity of the mucus could also directly damage the epithelial layer, which could in turn increase adherence of bacteria. Viscous mucus might also interfere with opsonization and phagocytosis of bacteria by neutrophils and macrophages. These factors could lead to the repeated cycles of infection and extended inflammatory responses seen in human CF patients. In these individuals, infections become established in increasingly distal regions of the lung with each cycle, leading to goblet cell hyperplasia in the lower airways and ultimately to loss of lung function.

Previous work has indicated that mice deficient in the Cftr gene fail to develop the airway disease characteristic of human CF patients. The possibility has been raised that the failure of CFTR-deficient mice to develop airway disease reflects the lack of airway submucosal glands and goblet cells in this species. To test this hypothesis, we have induced allergic airway disease in CFTR-deficient and control animals by sensitization and repeated exposure to ovalbumin.

We report that the induction of goblet cell hyperplasia and mucus secretion in Cftr^/ mice occurs in a pattern that is indistinguishable from that seen in normal control mice. In both groups of mice, more than half of the epithelial cells in the main bronchiole after 7 or 14 d of exposure to ovalbumin consisted of goblet cells. The origin of these mucus-producing cells has not been ascertained. Although the anatomic location of Clara cells in the respiratory tree suggests that goblet cells may arise from Clara cell metaplasia, it is also clear that the increase in goblet cells is paralleled by a decrease in the number of ciliated cells in these regions.

Surprisingly, we found that expression of the mucin gene, Mgm, occurred in mice treated with saline, which displayed no evidence of an ongoing inflammatory response and in which histologic examination using PAS/AB staining failed to reveal the presence of goblet cells. It therefore appears likely that even in normal airways, mucins are produced by epithelial cells, and that these mucins contribute to the properties of the airway surface liquid that facilitate the clearance of noxious substances. Alternatively, it is possible that although these mice express the Mgm gene, the mucin mRNA or polypeptide encoded by the gene is not processed into mucin. The mucin polypeptide itself constitutes only 10 to 30% of the total weight of the final mucin molecule. Through specific glycosyltransferase, sialytransferases, and/or sulfotransferases, a variety of oligosaccharides are covalently linked to the amino acid chain. These oligosaccharides account for the remaining 65 to 90% of the final mucin weight. The PAS/AB staining reflects the presence of the saccharides and not of the mucin polypeptide. With the development of antibodies specific for mouse mucin, it should be possible in the future to distinguish between these two possibilities.

In both Cftr^/ mice and normal controls, the basal level of Mgm gene expression was similar. Moreover, expression of the mucin gene was increased to a similar extent in both groups of animals with the onset of allergic airway disease. However, even in diseased airways the level of expression of Mgm remained much lower than observed in cells of the mouse stomach by in situ hybridization. Therefore, although Mgm is expressed in mouse airways, its induction alone does not appear to account for the dramatic increase in mucin production associated with allergic airway disease. It seems likely that induction of numerous other mucin genes, only some of which may not be expressed in the untreated mouse airway, contribute to the alterations seen in the ovalbumin-treated airways in our study. However, further studies, using methods that measure mucin directly, will be required to determine whether once again, differential processing of the mucin polypeptide in saline- and ovalbumin-treated airways contributes to these observations.

Neither gross examination of airways nor culture of lung homogenates revealed the presence of any bacterial pathogens in the airways of Cftr^/ mice. Similarly, histologic analysis did not reveal any differences between the airways of Cftr^/ mice and those of normal controls, even in animals over 2 yr of age. Furthermore, we show here that the increased production of mucus associated with allergic airway disease did not render CFTR-deficient mice more susceptible than normal controls to airway disease. Also, once an inflammatory response is established in Cftr^/ mice, the time course for resolution of the histopathologic changes characteristic of this airway disease was similar to that of normal mice treated in a similar fashion. We cannot rule out the possibility that differences exist between the two groups of animals in the production of specific cytokines or growth factors, or in other elements that do not lead to morphologic changes in the airways. Our findings contrast with those made on examination of lungs from mice with reduced Cftr expression (20). These mice have been reported to have inflamed airways at a young age (20).

Prior studies have shown that introduction of high numbers of S. aureus does not result in airway disease in CFTR-deficient mice (9). In the present study we extended this finding by showing that failure of S. aureus to colonize the mouse airway was not limited to the isolate used in the original report. Each of the five isolates we tested was cleared rapidly from the airways of both Cftr^/ mice and normal controls, precluding the establishment of an LD₅₀. In addition, histologic examination of the airways of both groups of mice indicated that S. aureus did not induce an inflammatory response characterized by neutrophil infiltration in either group of animals. These findings suggest that S. aureus cannot colonize the airways of normal or CFTR-deficient mice. In contrast, administration of high doses of P. aeruginosa resulted in lethal respiratory infections in both Cftr^/ mice and normal controls. However, the LD₅₀ and clearance rate for P. aeruginosa were similar in both groups of animals. In addition, scoring done with blinding to genotype indicated that the histologic changes in response to bacterial infection were similar in the two groups of animals. We again, however, cannot rule out the existence of a difference in the inflammatory response between the two groups of animals that does not result in altered rates of bacterial clearance or in morphologic changes detectable by histologic examination of the lungs. Our results contrast with those of Davidson and associates, who reported a decreased clearance of S. aureus from the airways of other mice expressing only low levels of CFTR (21).

Deposition of bacteria into the airways of the mice in which allergic airway disease had been induced failed to produce chronic infections in either Cftr^/ mice or normal controls. The presence of mucus associated with allergic airway disease does not allow S. aureus to colonize mouse airways, nor does it reduce the rate at which P. aeruginosa is cleared from the airways of CFTR-deficient mice or normal controls. These results suggest that the failure to produce large amounts of mucus in response to bacterial infection cannot account by itself for the difference in airway phenotype between the Cftr^/ mouse and the CF patient. However, because high levels of mucus production cannot be maintained in mouse airways for periods beyond 1 mo, we cannot yet determine whether, under conditions of prolonged mucus secretion, differences in colonization of the Cftr^/ and normal mouse airways would be seen.

A possible explanation for the lack of respiratory pathology in CFTR-deficient mice is the presence of an alternative pathway for Cl secretion by mouse airway epithelium. This hypothesis is supported by electrophysiologic studies indicating that a non-CFTR chloride channel is present on the apical surface of mouse tracheal epithelium (22). In the absence of CFTR, this pathway may provide a mechanism for fluid secretion adequate for the maintenance of airway surface-liquid volume and normal hydration of mucus.

The induction of allergic airway disease in mice by sensitization with and repeated exposure to ovalbumin indicates that the progression and resolution of this condition are similar in CFTR-deficient mice and normal controls. In addition, we found that the presence of allergic airway disease did not increase the susceptibility of either group of mice to the establishment of respiratory infections with organisms known to cause such infections in human CF patients. This work suggests that factors other than increased mucus production play a critical role in the pathogenesis of respiratory disease in humans with CF, and that determination of the relative importance of the various factors that contribute to the development of respiratory disease in human CF patients will require further modification of the airways of CFTR-deficient mice.