Role of Cftr genotype in the response to chronic Pseudomonas aeruginosa lung infection in mice

Anna M. van Heeckeren, Mark D. Schluchter, Mitchell L. Drumm, Pamela B. Davis

Abstract

Patients with cystic fibrosis have a lesion in the cystic fibrosis transmembrane conductance regulator gene (CFTR), which is associated with abnormal regulation of other ion channels, abnormal glycosylation of secreted and cell surface molecules, and vulnerability to bacterial infection and inflammation in the lung usually leading to the death of these patients. The exact mechanism(s) by which mutation in CFTR leads to lung infection and inflammation is not clear. Mice bearing different mutations in the murine homolog to CFTR (Cftr) (R117H, S489X, Y122X, and ΔF508, all backcrossed to the C57BL/6J background) were compared with respect to growth and in their ability to respond to lung infection elicited with Pseudomonas aeruginosa-laden agarose beads. Body weights of mice bearing mutations in Cftr were significantly smaller than wild-type mice at most ages. The inflammatory responses to P. aeruginosa-laden agarose beads were comparable in mice of all four Cftr mutant genotypes with respect to absolute and relative cell counts in bronchoalveolar lavage fluid, and cytokine levels (TNF-α, IL-1β, IL-6, macrophage inflammatory protein-2, and keratinocyte chemoattractant) and eicosanoid levels (PGE2 and LTB4) in epithelial lining fluid: the few small differences observed occurred only between cystic fibrosis mice bearing the S489X mutation and those bearing the knockout mutation Y122X. Thus we cannot implicate either misprocessing of CFTR or failure of CFTR to reach the plasma membrane in the genesis of the excess inflammatory response of CF mice. Therefore, it appears that any functional defect in CFTR produces comparable inflammatory responses to lung infections with P. aeruginosa.

  • cystic fibrosis transmembrane conductance regulator gene
  • agarose beads
  • mutation
  • inflammation

the cause of cystic fibrosis is due to a recessive genetic defect in the cystic fibrosis transmembrane conductance regulator gene (CFTR), a cAMP-activated chloride channel found in secretory epithelial cells. However, cystic fibrosis is a disease manifest by chronic lung infections with mucoid Pseudomonas aeruginosa, which causes the majority of morbidity and mortality in cystic fibrosis patients (13). The exact link between the CFTR defect and the permissive nature of the cystic fibrosis lung to acquiring and maintaining chronic lung infection with P. aeruginosa is unknown. The pathophysiological cascade that leads to lung disease probably includes trapping of bacteria in abnormal airway surface fluid and secretions (1, 2), reduced ability to kill or clear the trapped bacteria (4, 17, 22, 24), and an exuberant neutrophilic inflammatory response which persists even if the stimulus is reduced and provides the agents of lung destruction in the form of proteolytic enzymes and oxidants (13, 19).

How the cystic fibrosis defect leads to these disastrous downstream consequences is not clear. Three possibilities have been proposed. First, the lack of CFTR function may somehow translate into all the downstream consequences of the cystic fibrosis lesion (30). A second possibility is that most patients with cystic fibrosis in the United States have at least one misprocessed CFTR allele, and this abnormal protein may stimulate an inflammatory response or prime the cell to respond more vigorously to an inflammatory stimulus (7). A third possibility is that most patients with cystic fibrosis in the United States have a deficit of CFTR at the apical membrane, due either to misprocessing of a CFTR allele or to the presence of a stop codon, which prevents production of normal CFTR. If the membrane connections of CFTR are important in maintaining homeostasis, these individuals may have disrupted homeostasis. Indeed, one group has provided evidence that the COOH terminus of CFTR is critical for the normal production of the cytokine regulated on activation normal T cells expressed and secreted (RANTES) (8). If this regulation extends to other cytokines as well, it could promote the excessive inflammatory response that accompanies the cystic fibrosis lung disease.

Natural animal models of cystic fibrosis do not exist, however, through genetic manipulation, mouse models of cystic fibrosis lung disease have been developed; there are mice with defects in the murine homolog to CFTR (Cftr). These mice do not typically acquire spontaneous lung infections housed in most vivariums and therefore must be experimentally infected. A well-established model of chronic infection was used, whereby the primary host defenses are circumvented and the organism is physically retained in the lung by mechanical means. This latter model has shown that the inflammatory response to these P. aeruginosa-laden agarose beads is excessive in cystic fibrosis mice bearing the S489X knockout mutation compared with wild-type littermates (9, 26, 27). However, other murine genotypes are also available. This availability affords the opportunity to test several hypotheses. First, if the misprocessing of CFTR makes an independent contribution to the inflammatory response over and above that associated with lack of CFTR, then cystic fibrosis mice bearing the ΔF508 allele should have increased inflammatory responses. On the other hand, if it is the lack of CFTR in the apical membrane that is the primary driver of the inflammatory response, cystic fibrosis mice bearing the R117H allele should have significantly less inflammation in our model compared with mice in which Cftr fails to reach the apical membrane (knockout mice and ΔF508 mice). If CFTR defect is the primary driver of disease, however, all the cystic fibrosis mice should have comparable inflammatory responses. In these experiments we tested cystic fibrosis mice bearing the ΔF508, R117H, Y122X, or S489X genotypes, all backcrossed to the common C57BL/6J genetic background, using the mucoid P. aeruginosa agarose bead model to compare their inflammatory responses. Measures of the inflammatory response to bronchopulmonary infections with mucoid P. aeruginosa included weight loss, epithelial lining fluid levels of the acute-phase cytokines tumor necrosis factor (TNF)-α, interleukin (IL)-1β and IL-6, the murine neutrophil chemokines macrophage inflammatory protein (MIP)-2 and keratinocyte chemoattractant (KC), the eicosanoids PGE2 and leukotriene (LT) B4, and relative and absolute cell counts in bronchoalveolar lavage (BAL) fluid three days after infection.

MATERIALS AND METHODS

Mice

Breeding pairs of heterozygote congenic mice (≥N10) bearing the S489X mutation (B6.129P2-Cftrtm1Unc, stock no. 2196) and C57BL/6J mice were purchased from Jackson Laboratory (Bar Harbor, ME). Breeding pairs of heterozygote mice bearing the ΔF508 Cftr mutation in a mixed genetic background were a kind gift from Dr. Kirk Thomas from the University of Utah. Breeding pairs of heterozygote mice bearing the other Cftr mutations in a mixed genetic background were a kind gift from the laboratory of Dr. Mario Capecchi created by Mark Hamilton at the University of Utah. All mice originating from the University of Utah were backcrossed into the C57BL/6J background for at least 10 generations before use. Nomenclature rules available on the Jackson Laboratory website were followed; B6.129S6-Cftrtm2Uth mice bear the R117H mutation, and B6.129S6-Cftrtm3Uth mice bear the Y122X mutation. Cystic fibrosis mice for these strains are indicated by their Cftr mutation (e.g., B6.129P2-Cftrtm1Unc are referred to as cystic fibrosis mice bearing the S489X mutation). Mice were >5 wk of age weighing >16 g at the time of inoculation. Breeding females were fed a high-fat diet rodent chow (Laboratory Autoclavable Rodent Diet 5010; Purina Mills, St. Louis, MO). Wild-type mice were fed a standard rodent chow [Harlan Teklad 9F Sterilizable Rodent Diet (W) 7960, Harlan Teklad, Madison, WI, or irradiated Prolab RMH 3000, Purina Mills] after weaning. Cystic fibrosis mice bearing the severe mutations S489X, Y122X, and ΔF508 were fed the liquid elemental diet Peptamen (Nestle Clinical Nutrition, Deerfield, IL) after weaning, whereas cystic fibrosis mice bearing the mild mutation R117H were fed a standard rodent chow until 1 wk before the start of the experiment at which point they were fed Peptamen until the termination of the experiment. Autoclaved city water was provided in polyvinyl chloride water bottles with sipper tubes in stoppers and made available at all times. Mice were housed in static isolator units (Lab Products, Seaford, DE) on corncob bedding (combination size; The Andersons, Maumee, OH). Light cycles were 12 h on, 12 h off. Case Western Reserve University's Health Sciences Animal Resource Center is an Association for Assessment and Accreditation of Laboratory Animal Care-accredited facility, and mice were maintained in specific pathogen-free conditions (see Ref. 29 for details). All procedures involving mice were reviewed and approved by the Institutional Animal Care and Use Committee.

Genotyping Mice

The targeted Cftr allele is derived from the 129/SvJ mouse strain, and this allele has been crossed onto the C57BL/6J background. Comparison of Cftr genomic sequence between these two strains revealed several nucleotide variants that could be used to track the targeted allele. Here, single nucleotide differences in exons 14a and 17a that alter RsaI and AluI restriction sites, respectively, are used for genotyping the transgene by the presence or absence of restriction sites. Tissues from toe clips were taken at 7 days of age, in accordance with current Institutional Animal Care and Use Committee guidelines, for purposes of genotyping and a permanent means of identification. Tissues were digested at 75°C for 15–20 min with 200 μl of 0.2 M sodium hydroxide and mixed vigorously. Tris (800 μl of 0.04 M Tris, pH 7.5) was added to neutralize the base and centrifuged at top speed for 15–30 s, and these DNA samples were stored at 4°C until the genotype assay was performed. A 2-μl aliquot of supernatant from a DNA preparation was added to 23 μl of PCR mix [16 μl water, 1 μl each primer, 1.8 μl 2.5 mM dNTP, 2.5 μl 10× PCR buffer, 0.5 ml MgCl2, and 0.2 μl Taq polymerase (5 U/μl Taq DNA polymerase, catalog no. 18030-042, contains 10× PCR buffer, 50 mM MgCl2, and Taq polymerase; Invitrogen Life Technologies, Carlsbad, CA) per reaction]. Primers used were mCFEx14a5′ (GAG TGT TTT CTT GAT GAT GTG) and mCFEx14a3′ (ACC TCA ACC AGA AAA ACC AG), which generate a 130-bp amplicon. After the PCR reaction, 10 μl of the restriction mix [0.5 μl of 10× RsaI buffer, 0.5 μl RsaI (10 U/μl RsaI, catalog no. 15424-013, contains RsaI and 10× buffer; Invitrogen Life Technologies) and 9 μl of water per reaction] were added to 5 μl of each PCR reaction, after which time the samples were incubated at 37°C for at least 3 h. Fragments were fractionated on 3% agarose gels (Fisher Scientific, Cleveland, OH). Fragments of 109 and 21 bp indicate a wild-type allele, and an undigested 130-bp fragment corresponds to mutant.

Weighing Mice

Male cystic fibrosis mice and their wild-type littermates were identified by genotyping, as described above, and weighed on a digital scale starting at 7 days of life and then every 7 days. At 3 wk of age mice were weighed and then weaned, at which point cystic fibrosis mice bearing the severe Cftr mutations S489X, Y122X, and ΔF508 were fed a liquid diet. All others were fed a solid rodent chow. In previous studies, diet did not play a significant role in the response to lung infections with P. aeruginosa using the agarose bead model (28). Wherever possible, mice used in this study were also used in other experiments (therefore, some data starting at 6 wk of age are censored by the elimination of mice for use in other experiments) or until they were 8 wk old. Mice that were unsuitable for use in other experiments were killed by carbon dioxide according to the 2000 American Veterinary Medical Association Panel on Euthanasia.

Nasal Potential Difference Measurements

The potential differences across the nasal epithelium of the mice were measured, as described (3). Briefly, the nasal cavities of anesthetized mice (cocktail of ketamine-acepromazine-xylazine administered intramuscularly or 2.5% Avertin administered intraperitoneally) were perfused with warmed HEPES-buffered Ringer solution. The sodium channel blocker amiloride was added to the solution followed by a chloride-depleted solution containing amiloride and forskolin, which increases cAMP production.

P. aeruginosa Strain

Dr. Michael Tosi generously provided the mucoid clinical strain PA M57-15. Bacteria were maintained in glycerol stocks and stored at −70°C.

Making P. aeruginosa-laden Agarose Beads

Agarose beads laden with PA M57-15 were created as previously described (29). Briefly, bacteria were grown to late log phase. A 5-ml aliquot of the bacterial broth was swirled with 2% agarose. The agarose-broth mixture was added to heavy mineral oil equilibrated at 50–55°C, rapidly stirred for 6 min at room temperature, and then cooled over 10 min. The agarose beads were washed once with 0.5% deoxycholic acid, sodium salt (SDC) in PBS, once with 0.25% SDC in PBS, and three to four times with PBS. The bead slurry was allowed to settle by gravity so that 75% of the final volume consisted of beads. Quantitative bacteriology was performed on an aliquot of homogenized bead slurry. Bead diameter was measured under an inverted light microscope in several fields with the software package Image ProPlus (Media Cybernetics, Baltimore, MD). P. aeruginosa-laden agarose beads were prepared the day before inoculation and stored overnight at 4°C, and a different bead preparation was used for each experiment.

Inoculating Mice

Male mice were inoculated with P. aeruginosa-laden agarose beads, as previously described (29). Briefly, mice were anesthetized with 2.5% Avertin (0.015 ml/g body wt ip). The ventral cervical region was surgically prepared, a 1-cm skin incision was made just cranial to the thoracic inlet, and the trachea was visualized by blunt dissection. A 27-G 1-in over-the-needle intravenous catheter (angiocatheter) was used to cannulate the trachea. Mice were inoculated with an angiocatheter directed into the right main stem bronchus. The original bead slurries were diluted 10-fold in sterile PBS. The average 0.05-ml bolus contained 1.0 × 105 colony-forming units (CFU)/mouse, with a range of 5.5 × 104–1.9 × 105 CFU/mouse.

Evaluation of Mice Postoperatively

Mice were observed daily for clinical signs, such as coat quality, posture, ambulation, hydration status, and body weight. Mice that were moribund (could not right themselves after being placed in lateral recumbence) were killed before termination of the experiment; this was the only clinical outcome measure used that could definitely predict death. Mice were killed 3 days after inoculation by carbon dioxide narcosis followed by exsanguinations.

BAL

After death of mice with carbon dioxide followed immediately by exsanguination by direct cardiac puncture, we performed BAL in situ via a 22-G bead-tipped feeding needle ligated to the trachea to prevent backflow. Three 1-ml aliquots of sterile PBS were used to lavage the lungs and pooled. Qualitative bacteriology was performed on a 10-μl aliquot of unprocessed BAL fluid on tryptic soy agar plates. For eicosanoid analysis, 0.1 ml of BAL fluid was treated with 0.4 ml of cold HPLC-grade methanol immediately after collection. The samples were evaporated by vacuum centrifugation and reconstituted in a 1:1 by volume ratio of water to methanol and chilled on ice. Samples were assayed by a competitive immunoassay, as previously described (14). For cytokine analysis, BAL fluid was treated with a final concentration of 0.1 mM phenylmethylsulfonyl fluoride and 5 mM EDTA and then centrifuged for 10 min at 100 g at 4°C. The supernatant was sterile filtered (22-μm syringe filters, Millex-GV; Millipore, Bedford, MA) and stored at −70°C until cytokine analysis could be performed. Pellets were resuspended in 1 ml of PBS. A cell count was performed using a hemacytometer. Cytocentrifuge preparations (Cytospin 3; Shandon, Pittsburgh, PA) were stained with hematoxylin and eosin by standard techniques, and a differential cell count was performed. Cytokines measured were the murine proinflammatory mediators TNF-α and IL-1β, the immunomodulatory cytokine murine IL-6, and the murine neutrophil chemokines MIP-2 and KC, by enzyme-linked immunosorbent assay, according to the manufacturer's recommendations (R&D Systems, Minneapolis, MN). Values that fell below the limits of detection for the assay were assigned a value equal to the lowest limit of detection for each assay. Cytokine and eicosanoid concentrations were normalized for urea dilution (21) and expressed as units/ml epithelial lining fluid.

Statistics

The statistical software packages SigmaStat version 2.03 (Jandel Scientific, SPSS Science, Chicago, IL) or SAS (Cary, NC) were used. An unpaired t-test was used to compare differences between two groups. If either the normality or equal variance test failed, a Mann-Whitney rank-sum test was performed. A one-way analysis of variance (ANOVA) was used to test differences between three or more groups. Wilcoxon rank-sum tests, stratified by (i.e., controlling for) experiment, were used to compare genotype groups with respect to cytokines, chemokines, eicosanoids, and cell counts.

RESULTS

Growth Patterns of Cystic Fibrosis Mice

Table 1 compares the absolute body weights of mice at 7, 14, and 21 days of age (before weaning when all the mice are nursing), and Fig. 1 shows weekly body weights of cystic fibrosis mice compared with their littermates from 1 to 8 wk of age. Cystic fibrosis mice bearing the severe Cftr mutations Y122X, S489X, and ΔF508 weighed significantly less (P < 0.05) than homozygote wild-type controls at 7, 14, and 21 days of life with one exception; cystic fibrosis mice with the Y122X mutation did not differ significantly from wild-type mice at 7 days of age (P > 0.05), but sample sizes were small. The weight of cystic fibrosis mice bearing the mild R117H Cftr mutation did not differ significantly from wild-type controls at 7 days of age (P > 0.05) and weighed significantly more than mice bearing the severe mutations at 7 and 14 days of age (P < 0.05); however, they weighed significantly less than wild-type controls thereafter (P < 0.05) and were not significantly different from cystic fibrosis mice bearing the severe Cftr mutations at 21 days of life (P > 0.05). All mice bearing mutations in Cftr weighed significantly less than their wild-type littermate controls starting at 2 wk of age.

Fig. 1.

Body weight of cystic fibrosis mice and their wild-type littermates. Cystic fibrosis mice (▿) and their wild-type littermates (•) were weighed weekly starting at 1 wk of age. All available data are shown and represented as means ± SE. Cystic fibrosis mice bearing the S489X mutation were removed from this study to be used in other studies starting at 6 wk of age, and data are censored by death in cystic fibrosis mice bearing the Y122X mutation starting after 6 wk of age. *Significantly different than cystic fibrosis littermates of the same age (Student's unpaired t-test; P < 0.05).

View this table:
Table 1.

Absolute body weight of cystic fibrosis mice before weaning

Nasal Potential Difference

After amiloride administration, a chloride-depleted solution with amiloride and forskolin was applied. Traces are shown from a representative mouse in each strain studied, including a wild-type C57BL/6J mouse for comparison (Fig. 2). The potential differences after the amiloride response stabilized are cited in Table 2. The responses to low chloride challenge and forskolin stimulation are also shown in Table 2. The amiloride response was significantly different between wild-type mice and S489X mice (P = 0.005). The change in nasal potential difference after the application of chloride-depleted solution containing amiloride and forskolin was significantly different between wild-type mice and all mice bearing mutations in Cftr (P < 0.001). Statistical analysis was performed using one-way ANOVA, and pair-wise comparisons were made by Tukey’s test.

Fig. 2.

Nasal potential difference (PD) assays. The nasal PD was determined after amiloride was administered and reached a baseline, at which time a chloride-depleted solution containing amiloride and forskolin was administered (arrowhead). A junction potential of ∼12 mV occurred due to switching to a low chloride solution (3). The difference in nasal PD from the start of the response to 1 min later was determined, and representative tracings are shown to the end of the tracing period from wild type (A), S489X (B), Y122X (C), ΔF508 (D), and R117H (E) mice.

View this table:
Table 2.

Nasal potential differences in wild-type mice and mice bearing mutations in Cftr

Inflammatory Responses of Cystic Fibrosis Mice

Experiments were conducted with each genotype compared in the same experiments to mice bearing the S489X mutation. Only males were used in a narrow age range, although over the range of 6–12 wk, age was not found to play an important role on the inflammatory response (27, 28), and all the mice were backcrossed to the C57BL/6J congenic background to at least 10 generations. Every effort was made to keep the experimental conditions uniform from experiment to experiment, but inevitably, small differences in bacterial inoculum or ambient conditions may occur. Nevertheless, each genotype can be compared directly with S489X under identical conditions. Seven separate experiments were conducted (Table 3). Deaths due to surgical complications and pulmonary obstruction are noted in the table. Comparisons were made between the different cystic fibrosis mice within experiments by Wilcoxon tests that stratified on experiment or by ANOVA, adjusting for experiment. That is, when comparisons were made between mice bearing the S489X mutation and those bearing the Y122X mutation, we combined data from experiments 68, 72, 80, and 94, taking into account any differences between the experiments. Similarly, comparisons between mice bearing the S489X mutation and those bearing the ΔF508 mutation were made by combining data from experiments 68, 75, 90, and 94, and by combining data from experiments 64 and 94, comparisons were made between mice bearing the S489X mutation and those bearing the R117H mutation.

View this table:
Table 3.

Starting sample sizes in each experiment

Body weight.

Cystic fibrosis mice were weighed on day 0. There were no significant differences in starting weight between cystic fibrosis mice bearing the S489X mutation and those bearing any other Cftr mutation (Fig. 3A). Mice were then inoculated preferentially into the right main stem bronchus with mucoid P. aeruginosa-laden agarose beads. Mice were weighed once daily. After differences between the experiments are taken into consideration (Fig. 3B), weight loss in cystic fibrosis mice bearing the S489X mutation is significantly greater than those bearing the Y122X mutation on days 1, 2, and 3 (P < 0.05) and significantly less than those bearing the R117H mutation on days 1 and 2 (P < 0.05). There were no significant differences in weight loss between cystic fibrosis mice bearing the S489X mutation and those bearing the ΔF508 mutation.

Fig. 3.

Weight of mice before and after lung infection with mucoid Pseudomonas aeruginosa. Data are grouped according to the experiments represented in Table 2; raw data points are shown. A: absolute body weight on day 0 when mice were inoculated with P. aeruginosa-laden agarose beads. The boundary of the box closest to 0 indicates the 25th percentile, a line within the box marks the median, and the boundary of the box farthest from zero indicates the 75th percentile. Whiskers above and below the 0 indicate the 90th and 10th percentiles. Outlying points are also graphed. There were no significant differences in initial body weight between any of the cystic fibrosis mouse strains. B: change in body weight after infection. Data are represented as means ± SE. *Significantly different from cystic fibrosis mice bearing the S489X mutation at the same time point, after differences between experiments are taken into consideration.

Inflammatory mediators.

Data were pooled from all available data, and mediator concentrations are shown in Table 4. When stratifying for experiment, we found significant differences (P < 0.05) between cystic fibrosis mice bearing the Y122X mutation and those bearing the S489X mutation with regard to TNF-α and IL-1β concentrations in the epithelial lining fluid, although there were no significant differences in cytokine levels between cystic fibrosis mice bearing the S489X mutation and those bearing the ΔF508 or R117H mutations. In direct comparisons when stratifying for experiment, we found no significant differences in epithelial lining fluid concentrations of IL-6, MIP-2, KC, LTB4, or PGE2 (P > 0.05) between mice bearing the S489X mutation and those bearing any of the other Cftr mutations.

View this table:
Table 4.

Inflammatory mediators in epithelial lining fluid

Cell counts.

Cystic fibrosis mice bearing the R117H mutation had significantly lower relative neutrophil numbers and greater absolute alveolar macrophage numbers (Table 5) compared with those bearing the S489X mutation in direct comparisons when stratifying for experiment.

View this table:
Table 5.

Cell numbers in BAL fluid

A post hoc approximate power analysis for the unstratified Wilcoxon test was carried out, using the power function for the Wilcoxon test with underlying normal distributions (16). In a comparison of S489X vs. Y122X, ΔF508, and R117H genotypes, the available sample sizes provide 80% power to detect shifts in the mean of 0.74, 0.72, and 1.02 SD, respectively, using two-sided tests with significance level of 0.05. Thus the study had high power to detect effect sizes of ∼1 SD or more for all comparisons. When comparing the S489X to Y122X or ΔF508 strains, we could detect even smaller differences with high power. These calculations do not apply to LTB4 and PGE2, however, since the sample sizes were considerably smaller for these variables.

Kruskal-Wallis tests were performed only on data from the S489X mice, comparing responses across experiments for the weight, cytokine, and cell count data. Statistically significant differences across experiments were found for almost all variables examined, which emphasizes the importance of controlling for experiment in the analysis when comparing genotypes. In particular, significant differences across experiments were seen for weight loss 1 day after infection; TNF-α, IL-1β, IL-6, MIP-2, KC, PGE2; relative neutrophil and lymphocyte numbers; and absolute leukocyte, alveolar macrophage, neutrophil, and lymphocyte numbers.

DISCUSSION

Cystic fibrosis mice respond to challenge with mucoid P. aeruginosa-laden agarose beads with a higher death rate and greater inflammatory response in the BAL fluid than mice with at least one normal cystic fibrosis allele (9, 10, 17). This has been demonstrated for cystic fibrosis mice bearing the S489X mutation both backcrossed to the C57BL/6 background (9) and on a mixed genetic background of 129P2 and C57BL/6 (10) and for cystic fibrosis mice bearing the G551D mice on a mixed genetic background of CD-1 and 129/Sv (17), in three different laboratories, each with one of two mucoid clinical strains of P. aeruginosa. In two of the laboratories, this challenge resulted in increased lung burden of bacteria as well (9, 17), although in our studies, this difference was not clear (10). Nevertheless, there is remarkable uniformity among the groups that the cystic fibrosis mice have increased inflammatory responses. Moreover, Oceandy and colleagues (20) have shown that restoring CFTR to airway epithelium by transgenic technology in cystic fibrosis mice restores the normal response, suggesting that it is the Cftr mutation and not some other factor that accounts for the increased inflammatory response. Correcting the Cftr defect in the gut of cystic fibrosis mice bearing the S489X mutation, by transgenic provision of human CFTR driven by the fatty acid binding protein promoter, results in a much more robust cystic fibrosis mouse that grows normally and does not have intestinal obstruction on a diet of normal mouse chow. Nevertheless, these mice, too, display the typical cystic fibrosis excessive inflammatory response and increased death rate when challenged with P. aeruginosa-laden agarose beads (27). Thus the deficit of Cftr specifically in airway epithelial cells is implicated in the abnormal, excessive inflammatory response.

In this model of lung infection and inflammation, four different genotypes of cystic fibrosis mice were tested: two knockout mice, Y122X and S489X; mice homozygous for the major processing mutation in cystic fibrosis, ΔF508; and mice homozygous for a channel mutant, R117H, which reaches the plasma membrane but does not function normally. None of the cystic fibrosis mice studied here grows as well as their wild-type littermates, although the cystic fibrosis mice bearing the R117H mutation maintain weight better at week 1 of life. However, by week 2, they, too, are falling behind wild-type littermates in weight, although they will tolerate a diet of normal mouse chow without intestinal obstruction, unlike the other cystic fibrosis mouse strains examined here.

Previous results indicate that cystic fibrosis mice bearing the ΔF508 mutation display the cystic fibrosis phenotype established for the S489X knockout mouse (5) with regard to the nasal potential difference (3, 12). Here we show that cystic fibrosis mice bearing the Cftr mutations S489X, ΔF508, Y122X, and R117H on the congenic C57BL/6J background also display the cystic fibrosis electrophysiological phenotype. In addition, they were indistinguishable in their inflammatory responses to lung infection with P. aeruginosa from each other. Thus there is no evidence that the ΔF508 processing mutant confers more dramatic inflammatory responses or poorer growth. It has been suggested that in cells that harbor the ΔF508 mutation in CFTR, this provides a stimulus for the unfolded protein response, which results in activation of inflammatory cascades even in unstimulated cells (6). However, this response appears not to occur in substantive form in the cystic fibrosis mice. Some of the studies of this response have been conducted in heterologous systems in which there is substantial overexpression of mutant CFTR, so overexpression, as much as the nature of the mutant, could produce the response (11, 15, 18, 25). However, this will not explain the results in IB-3 cells, which are compound heterozygous cystic fibrosis cells containing W1282X and ΔF508 (6). These cells appear to have an underlying activation of NF-κB even without stimulation that is relieved by growing the cells at reduced temperature (6). However, this reduced temperature also allows functional CFTR to reach the membrane, so this experiment does not distinguish between correction of misprocessing and correction of function. If ΔF508 and other processing mutants do confer additional, independent pathology beyond that of functional CFTR deficiency, then gene therapy or activating alternative chloride channels, strategies that have been proposed to treat cystic fibrosis, may not protect against the typical cystic fibrosis lung inflammation. The data presented here suggest that this is likely not the case.

The R117H allele, unlike those of the other genotypes studied here, is expected to reach the plasma membrane. Although the R117H animals have weight comparable to wild-type littermates at 1 wk of age, as early as 2 wk of age until at least 8 wk of age they, too, are as underweight as the cystic fibrosis knockout mice, despite the fact that they can survive on normal rodent chow after weaning. Their nasal electrophysiology is comparable to the other cystic fibrosis genotypes, and their inflammatory responses are similar. Thus a cystic fibrosis allele that reaches the plasma membrane confers no protection against the excess inflammatory response. Although this seems to be an important feature in some cell model systems for the aberrant production of the cytokine RANTES by cystic fibrosis airway epithelial cells (23), we could not demonstrate the importance of a CFTR that reaches the plasma membrane but does not function properly, in reducing the inflammatory response in intact animals. These data are concordant with those of McMorran et al. (17), who studied mice bearing the G551D allele, which reaches the plasma membrane but fails to be activated to open the channel. These mice, too, show excess inflammatory response (17). Thus, although there may be subtle differences in the responses that were not detected in our outcome measures or compensations in intact lung that do not occur in isolated epithelial cells, there was not a functionally significant difference in the genotypes.

There are several caveats in this analysis, however. First, all of the outcome measures studied here are variable, and, therefore, despite the substantial number of mice we examined, small differences in the inflammatory responses may not be detected. The variability was reduced to the greatest extent possible, however, for care was taken to study only fully backcrossed mice of comparable C57BL/6J genetic background at comparable age, body weight, and sex and to study as many genotypes as feasible on the same day with the same preparation of mucoid P. aeruginosa-laden agarose beads, to control as many variables as possible. Each genotype was studied at the same time as cystic fibrosis mice bearing the S489X mutation, so that direct comparisons are possible (same preparation of agarose beads, same circumstances in the animal facility), and there were no substantive differences between cystic fibrosis mice bearing the S489X mutation and any other genotype. Only the cystic fibrosis mice bearing the Y122X mutation differed in two cytokines from S489X, and this may represent the effect of multiple comparisons, rather than a real difference between them. Another caveat is that mice are not humans, and there may be much greater genotype-related differences in humans. Because the cystic fibrosis mouse (of any genotype) does not develop spontaneous lung infection, this possibility cannot be discounted. In addition, in the studies of the ΔF508 mice, Zeiher and colleagues (31) have noted that this allele does not express as well as normal Cftr. Thus the burden of misfolded protein may be less in this animal model than it would be for humans of this genotype, simply because less protein is made. However, together with clinical observations that patients with severe mutations all seem to have similar lung disease, whether the mutations are stop codons, processing mutants, or activation mutants that reach the membrane, such as G551D, our data suggest that lack of function of CFTR, and not the absence of CFTR at the membrane or the presence of misprocessed CFTR in the cells, is the predominant factor that accounts for the excess airway inflammation.

GRANTS

This work was funded by National Institutes of Health grants P30 DK-27651 and HL-60293 and a research development program grant from the Cystic Fibrosis Foundation.

Acknowledgments

We thank several members from Core facilities at the Case Western Reserve University Cystic Fibrosis Center: specifically, Animal Core members Alma Genta Wilson, Lisa Hogue, Christiaan van Heeckeren, James Poleman, Veronica Peck, Ebony Boyd, and Merle Fleischer for breeding and weighing mice and performing the infection studies; Molecular Core personnel Bill Marcus for genotyping the mice; and Inflammatory Mediator Core supervisor Christopher Statt for expert technical assistance in performing cytokine analyses. In addition, we thank Jessica Hoyt for weighing untreated mice and Nicole Kyle and Assem Ziady for performing the nasal potential difference assays.

Footnotes

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

REFERENCES

View Abstract