INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Cystic fibrosis (CF) is characterized by progressive development of chronic pulmonary infections with bacterial and viral pathogens such as Haemophilus influenzae, Staphylococcus aureus, Burkholderia cepacia, Pseudomonas aeruginosa (1), and respiratory syncytial virus (2, 3). In the great majority of CF patients, a vicious circle of recurrent pulmonary infections leads to inflammation, lung damage, bronchiectasis, respiratory failure, and death (4). Although it is well established that CF is caused by mutations in the Cftr gene, the link between a defective Cftr-mediated chloride channel and predisposition to chronic lung infections still remains unclear.

In an attempt to develop a suitable animal model to study the pathogenesis of CF lung disease, a number of investigators have generated Cftr knockout mice by gene targeting (5). Animals carrying F508 (9, 10) or G551D (11) mutations at the Cftr locus have also been produced. In most of these systems, a large proportion of animals die of intestinal obstruction within the first month of life (6). Histological examination of the lungs and trachea of Cftr knockout animals revealed only mild signs of lung disease pathology (5, 6, 11). This could reflect the pathogen-free environment in which these animals are housed and/or their very short life span. Interestingly, evidence of pathological changes in the lungs of CF patients is rarely apparent at birth, but develop progressively after a few months of life (12). Overall, it appears that Cftr knockout mice could be helpful in investigating Cftr gene defect related lung disease if the appropriate conditions of infection for these animals are well defined.

As chronic lung infection with P. aeruginosa is the major cause of morbidity and mortality in CF, we have focused our efforts on the elaboration of an animal model to study the pathogenesis of infection with this particular pathogen. As rodents are known to be generally more resistant than human to several bacteria including P. aeruginosa, it was of importance to choose a relatively susceptible strain of mice among those available. C57BL/6 strain of mice was previously shown to be more susceptible to P. aeruginosa lung infection as compared with BALB/c, A/J, and DBA/2 mice (13). In the present study, we have investigated the severity of P. aeruginosa lung infection in 3 to 4-mo-old Cftr^unc knockout mice initially developed by Snouwaert and colleagues (6) and transferred by us on homogenous C57BL/6 genetic background [B6-Cftr (/)]. The great majority of B6-Cftr knockout mice could survive for a long time over weaning period, since they were fed sterile Peptamen diet and kept under conditions previously shown to prevent intestinal obstruction (16). This allowed us to use a relatively large number (24 mice) of knockout animals for this study. In order to assess the innate ability of B6-Cftr (/) mice to resist infection with P. aeruginosa, we have infected these mice intratracheally with a single dose (10⁵) of P. aeruginosa entrapped in agar beads. We have previously shown that these conditions of infection allow the establishment of a prolonged, severe endobronchial infection lasting more than 4 wk in susceptible C57BL/6 mice (15) (Stevenson and coworkers, unpublished results). Moreover, a similar model of infection has been extensively used in several other strains of mice as well as in rats to induce chronic lung infection (13, 14, 17). In the present study, the susceptibility to P. aeruginosa infection was evaluated according to the mortality rate, the number of viable P. aeruginosa recovered from the lungs, and histopathological changes in the lungs following intratracheal infection. This is the first report that demonstrates the impaired ability of B6-Cftr^unc knockout mice to control lung infection with P. aeruginosa, resulting in a 10-fold increase in the number of viable bacteria recovered from the lungs and in a 5-fold increase in the mortality rate. The model of infection reported here could thus provide a helpful tool for the study of the pathogenesis of P. aeruginosa lung infection in CF.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Generation of Cftr Knockout Mice on C57BL/6 Background

Cftr^unc knockout mice generated at the University of North Carolina (6) by insertion of a stop codon in exon 10 of Cftr gene, were obtained from Beverly Koller (University of North Carolina). These mice had a mixed 129/J and C57BL/6 genetic background. To transfer this mutation onto a homogenous C57BL/6 background, we have backcrossed Cftr^unc mice to C57BL/6 mice. At each generation, Cftr (+/) mice with the highest level of homozygosity for C57BL/6 background were selected by microsatellite typing analysis and bred for another generation with C57BL/6 mice. The level of homozygosity for C57BL/6 effectively reached 100% within four generations according to 66 markers dispersed throughout the mouse genome. Mice used in these studies represent offsprings from B6-Cftr^unc (+/) mice backcrossed to C57BL/6 mice for 10 to 12 generations.

Housing Conditions

B6-Cftr (/) and B6-Cftr (+/+) mice were kept under specific pathogen-free conditions conforming to NIH guidelines. They were housed on corn cob bedding (Anderson, Maumee, OH) and fed sterile Peptamen (Clintec Nutrition, Deerfield, IL), a low-residue liquid diet which was previously shown to prevent intestinal obstruction in Cftr^unc knockout mice (16). One hundred milliliters of diet contains 84.7 g water, 12.7 g carbohydrates, 4 g protein, 3.9 g fat, 0.7 g ash, 0.5 g linoleic acid, 45 mg choline, 400 IU vitamin A, 28 IU vitamin D, 2.8 IU vitamin E, 0.008 mg vitamin K, 14 mg vitamin C, 2.8 mg niacin, 1.4 mg pantothenic acid, 0.4 mg vitamin B6, 0.24 mg riboflavin, 0.2 mg thiamine, 0.05 mg folacin, 0.04 mg biotin, 0.0008 mg vitamin B12; 125 mg potassium, 100 mg chloride, 80 mg calcium, 70 mg phosphorus, 50 mg sodium, 40 mg magnesium, 8 mg taurine, 8 mg carnitine, 1.4 mg zinc, 1.2 mg iron, 0.27 mg manganese, 0.14 mg copper, 0.012 mg molybdenum, 0.01 mg iodine, 0.004 mg chromium, 0.004 mg selenium. The caloric content of the diet is 420 KJ/100 ml and an adult mouse consumes approximately 15 ml/d. In order to maintain the sterility of the diet, it was changed daily. Table 1 shows details concerning age and weight of mice selected for experiments. Mice used for all experiments were males.

Cftr Genotyping

Genomic DNA was prepared from tails collected from mice at the weaning period. DNA was extracted by a modification of the salting out procedure as previously described (18, 19). Briefly, the tails were digested overnight (55° C) with proteinase K (Boerhinger Mannheim, Laval, Quebec) and proteins were precipitated with a saturated NaCl solution. DNA was then precipitated with absolute ethanol and redissolved in TE (10 mM Tris 1 mM ethylenediaminetetraacetic acid) buffer. Normal and mutated Cftr alleles were types by polymerase chain reaction (PCR) performed under standard conditions (94° C for 4 min, followed by 30 cycles of 94° C for 60 s, 59° C for 90 s, 72° C for 90 s). Primers CF1248R (5' CTT TGA TAG TAC CCG GCA TAA TC 3') and CF944 (5' TGA ACC TTA GTC CTA TGT TGC C 3') were used to detect the normal allele and primers neo-1 (5' CTT GGG TGG AGA GGC TAT TC 3') and neo-2 (5' AGG TGA GAT GAC AGG AGA TC 3') were used to detect the mutated allele (all primers were synthesized at Sheldon Biotechnology Center, Montreal, Quebec, Canada). The PCR products (300 bp for normal allele and 200 bp for mutated allele) were electrophoresed in 10% polyacrylamide gel and stained with ethidium bromide.

Infection of Mice

P. aeruginosa (strain 508) initially isolated from a CF patient (Dr. Jacqueline Lagacé, University of Montreal, Montreal, Canada) was selected for mucoid character. Pulmonary infection with this strain of bacteria entrapped in agar beads has been previously described (13- 15). Briefly, a log-phase suspension of the bacteria was diluted in warm (52° C) trypticase soy agar (TSA). The bacteria was then entrapped in agar beads by mixing with heavy mineral oil (52° C) (Fisher Scientific, Fair Lawn, NJ) and stirring vigorously followed by cooling of the mixture at 4° C. The bacteria-containing beads were washed extensively and resuspended in phosphate-buffered saline (PBS) (ICN Biomedicals Inc., Costa Mesa, CA). The size ( 150 µM) and uniformity of the beads were confirmed by microscopic examination. The number of viable P. aeruginosa entrapped in agar beads was determined by plating serial dilutions of the homogenized bead suspension on TSA plates.

Mice were anesthetized with a mixture of ketamine hydrochloride (75 mg/kg body weight; MTC Pharmaceuticals, Cambridge, Ontario, Canada) and xylazine (30 mg/kg body weight; Bayvet Division, Chemagro Limited, Etobicoke, Ontario, Canada) injected intramuscularly. A 50-µl inoculum containing 10⁵ viable P. aeruginosa entrapped in agar beads was injected into the lungs through the trachea with a 22-gauge intravenous catheter (Critikon, Tampa, FL).

Bronchoalveolar Lavages (BAL)

The alveoli of infected mice were washed 7 times with 1 ml of PBS via the cannulated trachea. The volume of BAL recovered was approximately 6 ml. Alveolar cells were centrifuged, resuspended in 1 ml PBS, stained with Turk's solution and counted. The proportions of macrophages, lymphocytes, and polymorphonuclear leukocytes (PMN) were calculated after counting 200 alveolar cells on cytospin preparations stained with Diff-Quick (American Scientific Products, McGaw Park, IL).

Measurement of Bacterial Burden

Lungs, spleen, kidneys, and livers were harvested independently from infected mice and homogenized for 60 s at high speed (homogenizer PT10135; Brinkmann Instruments Co., Mississauga, Ontario, Canada) in 10 ml of PBS. Serial 10-fold dilutions of homogenates were plated on petri dishes containing TSA and the number of colony-forming units (CFU) per organs was counted after overnight incubation at 37° C. In each experiment, the identity of the bacteria recovered from infected animals was confirmed by the Department of Microbiology (Montreal General Hospital) using a Vitek gram-negative identification system (BioMérieux Vitek, Inc., Hazelwood, MO).

Histopathological Studies

The lungs were inflated (without removing alveolar cells) with 10% buffered formalin (Sigma Chemical Co., St. Louis, MO). Each lobe was sectioned longitudinally in the proximal, medial as well as the distal regions. The three regions were then individually and systematically cut three times at 0.5-mm intervals. For each lobe, nine different sections were stained with hematoxylin and eosin or according to the periodic acid-Schiff (PAS) method and scored blind by two observers using a semiquantitative scale ranging from 0 to 3. A zero on this scale indicated that no inflammatory cells were detected and that the architecture of the lung was normal, while 3 was the highest score given for highly severe infiltration of inflammatory cells accompanied by complete disruption of the lung structure. The lung sections were also scored for PAS-positive material lining the airways. A zero on this scale indicated that no PAS-positive material was detected, while a 3 represents the most positive PAS-stained slide examined. A third observer has confirmed that the scoring was very similar between the observers. Thus, no major difference can be ascribed to interobserver variability. The kidney, spleen (half), and liver (one lobe) collected from all infected animals were individually fixed in formalin. Distal and proximal sections stained with hematoxylin and eosin were examined for any signs of infections or histopathological abnormalities.

Statistical Analysis

The log₁₀CFU detected in the lungs of B6-Cftr (/) and B6-Cftr (+/+) control mice were compared by the nonparametric Mann-Whitney U test (Figure 1). A chi-square test was used to analyze the incidence of mortality after P. aeruginosa infection (Table 2).

View larger version (11K): [in this window] [in a new window]   Figure 1.   Bacterial burden in the lungs of B6-Cftr knockout mice and their littermate controls. The number of CFU was measured in the lungs of B6-Cftr (/) (solid triangles) or B6-Cftr (+/+) (solid circles) mice 3 (A) or 6 (B) d after intratracheal infection with 105 P. aeruginosa entrapped in agar beads. Horizontal lines represent the median log10 CFU. Mice were monitored every 6 h for mortality. The numbers of CFU measured in the lungs of mice found dead at 6 d postinfection were included in this study and they are represented by crosses (). Results in A and B were obtained in three separate experiments. The results presented in B are pooled from two different experiments. The mean CFU counts detected in the lungs of B6-Cftr (/) mice were shown to be significantly different (p < 0.01 in A) (p < 0.05 in B) from that found in B6-Cftr (+/+) mice.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Incidence of Mortality

In order to evaluate the capacity of B6-Cftr knockout mice to control P. aeruginosa lung infection, they were infected intratracheally with 10⁵ P. aeruginosa entrapped in agar beads. The severity of infection was initially evaluated on the basis of mortality rate following infection and compared with that obtained in B6-Cftr (+/+) control mice. Results presented in Table 2 indicate that B6-Cftr (/) mice are clearly more susceptible to P. aeruginosa lung infection than littermate controls. Indeed, 58.4% (n = 24) of the B6-Cftr (/) mice died within the first 6 d of infection, whereas only 12.1% (n = 41) of B6-Cftr (+/+) mice died over the same period of time (p < 0.005). Interestingly, the peak of mortality in the B6-Cftr (/) group occurred between 4 and 6 d postinfection. None of B6-Cftr (/) or B6-Cftr (+/+) mice died after inoculation with agar beads that did not contain any embedded bacteria (data not shown).

Bacterial Burden in the Lungs

In order to determine whether the high susceptibility of B6- Cftr^unc knockout mice is associated with high bacterial load in the lungs, we measured the number of CFU in the lungs of these mice and their controls at 3 and 6 d postinfection. Results presented in Figure 1A show that the number of bacteria detected at 3 d of infection was clearly higher than that initially injected in the lungs, suggesting that the bacteria proliferate extensively during the first 3 d of infection. The number of CFU measured in the lungs of B6-Cftr (/) mice at 3 d postinfection was significantly (p < 0.01) higher than that detected in B6-Cftr (+/+) mice. Similarly, the median CFU found in the lungs of B6-Cftr (/) mice 6 d after infection was significantly (p < 0.05) higher than that observed in Cftr (+/+) control mice (Figure 1B). The number of bacteria found in the lungs of animals who were found dead at 6 d postinfection was also included in Figure 1B. In a preliminary experiment, we have established that the CFU counts recovered from the lungs of P. aeruginosa-infected mice at 2 to 6 h post mortem remained equivalent to that measured immediately after death (Table 3). Based on these results, we therefore measured the CFUs in the lungs of dead mice within 6 h post mortem or at 6 d postinfection. The bacteria isolated from the lungs of infected mice were confirmed to be P. aeruginosa (strain 508).

In order to ascertain that the mortality observed over the course of the infection cannot be attributed to sepsis, the spleens, livers, and the kidneys collected from animals deceased over the course of infection and from mice killed at 6 d postinfection were assessed for bacterial colony counts and for any histopathological signs of infections. No bacteria were found in the liver, spleen, or kidneys of B6-Cftr (/) or B6- Cftr (+/+) mice that were found dead over the course of infection or that were killed at 6 d postinfection (data not shown). In addition, histological examination of liver, spleen, and kidneys of infected mice revealed no evidence of abnormal inflammation or pathology (data not shown).

As previously reported by Snouwaert and colleagues (6), evidence of intestinal pathology was observed in B6-Cftr (/) mice. The major signs of pathology were eosinophilic secretions and dilated goblet cells filled with mucus in the colon and enlarged crypts in the ileum and jejunum. The severity of these pathologies remained unchanged following P. aeruginosa infection (data not shown).

In the next series of experiments, the number of bacteria found in dead B6-Cftr (/) mice was compared with that obtained in their littermate B6-Cftr (+/+) controls deceased or killed at the same time postinfection. Table 4 summarizes the results obtained for each pair of animals analyzed. The numbers of CFU measured in the lungs of dead B6-Cftr (/) mice were consistently and markedly higher (5 to 50 times) than that found in their littermate control mice (Table 4).

As the weights of the B6-Cftr (/) mice were on average 25 to 30% less than that of their normal controls and the ages were variable within the experimental groups, it was of importance to establish whether the differences in CFU counts observed between B6-Cftr (/) and B6-Cftr (+/+) mice correlated with any variations in weight or age of the animals. Results presented in Figure 2 show no correlation between the number of CFU detected in the lungs of infected mice and the weight (r = 0.03) or age (r = 0.10) of the animals.

View larger version (12K): [in this window] [in a new window]   Figure 2.   Lack of correlation between the number of viable P. aeruginosa detected in the lungs of infected animals and their weight (A) or age (B). The total number of CFUs was measured in the lungs of B6-Cftr (/) (solid triangles) and B6-Cftr (+/+) (solid circles) mice 6 d after intratracheal infection with 105 P. aeruginosa entrapped in agar beads. The number of CFU was plotted on y-axis with the weight or age of the animals on the x-axis.

Histopathology of the Lungs

There was no evidence of severe airway disease in the uninfected B6-Cftr (/) or their B6-Cftr (+/+) controls. Two of the B6-Cftr (/) mice examined displayed a small degree of peripheral dilatation when compared with control mice. In one of the two B6-Cftr (/) mice there was evidence of slight focal interstitial pneumonitis in one area (not shown). Furthermore, examination of the airways of B6-Cftr (/) and B6-Cftr (+/+) mice revealed no evidence of epithelial cell damage, inflammatory cell accumulation, secretory cell hyperplasia (Figure 3A and B), or mucus retention (not shown). In contrast, the lungs of mice infected with P. aeruginosa were characterized by severe focal endobronchial and alveolar infiltration as well as substantial obstruction of the airways with bacteria-containing beads surrounded by inflammatory cells (Figure 3C to 3F). Moreover, acidic mucopolysaccharide, as evidence by PAS-positive staining, was also found lining the airways of P. aeruginosa-infected mice (not shown). The lung pathology observed following infection was heterogeneous and characterized by foci of infection. A semiquantitative histopathological analysis of the lung sections from P. aeruginosa-infected mice showed no significant difference in the severity of pneumonitis or mucus retention between Cftr (/) and Cftr (+/+) mice (Table 5). The proportion of bronchi heavily blocked with bacterial microcolonies and inflammatory material was prominent in B6-Cftr (/) mice (Figure 3C and E). In age-matched B6-Cftr (+/+) control mice, inflammation was mainly concentrated around smaller airways or alveoli (Figure 3D), although some bacteria-containing beads could also be found in the bronchi (Figure 3F).

View larger version (149K): [in this window] [in a new window]   View larger version (142K): [in this window] [in a new window]   View larger version (143K): [in this window] [in a new window]   View larger version (145K): [in this window] [in a new window]   View larger version (143K): [in this window] [in a new window]   View larger version (129K): [in this window] [in a new window]   Figure 3.   Histological examination of the lungs of B6-Cftr knockout mice and their littermate controls. Representative hematoxylin- and eosin-stained lung sections prepared from uninfected B6-Cftr (/) (A) or B6-Cftr (+/+) (B) mice. Panels C and D show evidence of severe pneumonia characterized by extensive cell infiltration in the bronchi and alveoli of infected B6-Cftr (/) or B6-Cftr (+/+) mice, respectively (×100). Panels E and F illustrate a higher magnification of the airways of infected B6-Cftr (/) (E ) or B6- Cftr (+/+) (F ) mice blocked by microcolonies of bacteria attached to beads (×250).

No evidence of bacterial colonization was found in the lungs of B6-Cftr (/) or control mice after intratracheal injection of agar beads that did not contain any embedded bacteria (data not shown). In addition, only mild signs of inflammation were observed in the lungs of both B6-Cftr (/) and B6-Cftr (+/+) mice injected with sterile agar beads.

Recruitment of Alveolar Cells

Chronic P. aeruginosa lung infection in CF patients is generally associated with an exuberant and destructive inflammatory response characterized by massive infiltration of PMN (20). In an attempt to determine whether this would also be the case in infected B6-Cftr knockout mice, we measured the numbers and proportions of PMN, macrophages, and lymphocytes recruited to the alveoli of B6-Cftr (/) and B6-Cftr (+/+) mice following infection with P. aeruginosa. The proportion of PMN found in the alveoli of B6-Cftr (/) mice (57%) following P. aeruginosa infection was similar to that found in B6-Cftr (+/+) control mice (55%) (Figure 4). The total number of alveolar cells found in B6-Cftr (/) mice following infection was slightly, but not significantly, higher than in B6-Cftr (+/+) control mice. However, this may reflect the higher bacterial load observed in B6-Cftr (/) mice.

View larger version (49K): [in this window] [in a new window]   Figure 4.   Magnitude of the inflammatory response in the alveoli of B6-Cftr (/) and B6-Cftr (+/+) mice following P. aeruginosa infection. The total number of alveolar cells was measured in BAL collected from B6-Cftr (/) and B6-Cftr (+/+) mice 6 d after intratracheal infection with 105 P. aeruginosa enmeshed in agar beads. The absolute numbers of alveolar macrophages (solid bars), lymphocytes (open bars), and PMN (hatched bars) were calculated on the basis of the differential cell counts of Diff-Quik cytospin preparations. Numbers in parenthesis indicate the percentages of macrophages, lymphocytes, and PMN in BAL.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Despite recent advances in unraveling the basic gene defect in CF, chronic and destructive lung infection with P. aeruginosa remains the major cause of mortality in this disease. Our current understanding of the pathogenesis of this lung infection has been hampered by the lack of an appropriate animal model of infection. Although Cftr knockout mice display only mild lung pathology at birth (5, 6, 11) they remain the only system available to date to investigate the intrinsic basis of CFTR-dependent dysregulation of lung homeostasis. Davidson and colleagues (21) have previously investigated the course of repeated infections with two other CF-related pathogens, S. aureus or B. cepacia, in Cftr^m1HG knockout mice. They have reported higher bacterial loads in the lungs of Cftr (/) mice after repeated exposure to S. aureus or B. cepacia as compared with their littermate controls. In parallel with this finding, they have also observed an increased incidence of goblet cell hyperplasia, mucus retention, and bronchiolitis. Snouwaert and colleagues (22) who studied another Cftr knockout mouse model reported that Cftr^unc knockout mice do not differ from Cftr^unc (+/) control mice in their ability to clear single or repetitive intranasal infection with S. aureus. The knockout mice used for the latter study were, however, selected on the basis of their capacity to survive for more than 1 yr. It is, thus, possible that they were selected for a phenotype that could compensate for the defective gene. In addition, it remains possible that the mixed genetic background on which these mice were developed could influence their susceptibility to infection.

Although P. aeruginosa is the major cause of mortality in CF, no appropriate animal model for the study of the pathogenesis of this lung infection in CF has yet been developed. In the present study, we report an impaired ability of B6-Cftr^unc knockout mice to control P. aeruginosa lung infection. The establishment of appropriate conditions for prolonged infection appears to be critical to detect the exacerbation of the infection in B6-Cftr (/) mice. In the present study, mice were infected intratracheally with a single dose of 10⁵ P. aeruginosa entrapped in agar beads. We have previously observed that when B6-Cftr (/) mice were infected with a lower dose of bacteria which leads to a more acute infection that resolves within 7 d (14), they did not differ from B6-Cftr (+/+) control mice in their capacity to control the infection (23). However, as shown in the present study, B6-Cftr^unc knockout mice become highly susceptible to P. aeruginosa lung infection, compared with B6-Cftr (+/+) mice when they are infected with a higher dose of bacteria. As for normal healthy subjects, mice are relatively resistant to P. aeruginosa lung infection. Therefore, these conditions of infection were necessary to allow the bacteria to persist in the lungs for a period long enough to mimic the chronicity of lung infection in CF. This is an important feature in CF, as it is involved in the induction of inflammation and lung injury, which in turn impairs the host response to lung pathogens (19).

The susceptibility of B6-Cftr^unc knockout mice to P. aeruginosa infection was demonstrated by a high incidence of mortality within the first week following infection. The death of B6-Cftr (/) animals correlates with very high bacterial load in the lungs. The death is unlikely to result from a systemic infection because no bacteria as well as no inflammation were found in spleen, liver, and kidneys of infected animals. The high susceptibility of B6-Cftr (/) mice to P. aeruginosa infection was also demonstrated by high CFU counts detected in the lungs at 3 d postinfection and at 6 d postinfection among animals that survived. Our results suggest that the bacteria proliferate in the lungs of mice within the first 3 d postinfection. The B6-Cftr (/) mice allow the bacteria to proliferate more extensively than the B6-Cftr (+/+) control mice, as the CFU counts detected in the lungs of B6-Cftr (/) mice were significantly higher than those found in the lungs of control animals. The first 3 d appear to be critical for the outcome of the infection, since a high incidence of mortality was recorded in the B6-Cftr (/) group from 4 to 6 d postinfection. It thus appears that B6-Cftr (/) mice exerted a lower ability to control the infection within 3 d postinfection and this is leading to death in 58% of the cases. The B6-Cftr (/) mice that survived until 6 d postinfection could control to some extent the bacterial proliferation, but the bacterial load remains significantly higher than that found in the lungs of B6-Cftr (+/+) control animals. The higher CFU counts detected in B6-Cftr (/) mice cannot be explained by the fact that they were substantially smaller than B6-Cftr (+/+) controls, as no correlation was found between CFUs detected in the lungs and the weight of animals. Our results also indicate that the age of animals does not significantly affect the course of P. aeruginosa lung infection.

A severe pneumonia characterized by cellular inflammation of the bronchi and alveoli was found in the lungs of B6- Cftr (+/+) as well as B6-Cftr (/) mice. Similarly, the lungs of infected CF patients are also characterized by severe inflammation mainly centered in the bronchi and spreading to the alveoli (12, 24, 25). In contrast to the human situation, no sign of fibrosis was seen in infected B6-Cftr^unc knockout mice. However, it is most likely that the time point chosen for this study (6 d postinfection) is too early to detect fibrosis. Increased mucus secretion and viscosity have been reported in almost all CF patients and are accompanied by mucopurulent plugging of the airways (26). Interestingly, P. aeruginosa infection clearly induced secretion of acidic mucopolysaccharide in the airways of both B6-Cftr (/) and B6-Cftr (+/+) mice. Although severe pneumonia and mucus secretion were clearly induced in both B6-Cftr (/) and B6-Cftr (+/+) mice following P. aeruginosa infection, no striking histopathological differences were observed between B6-Cftr (/) mice and their littermate controls. It thus appears that the high susceptibility of B6-Cftr (/) mice to P. aeruginosa infection does not correlate with a more severe pneumonia compared with their controls. However, these results do not necessarily imply that pneumonitis does not play any role in the death of the animals, but that other factors, such as high bacterial load, bacterial virulence, bacterial and inflammatory side products, might also weaken B6-Cftr (/) mice and contribute to their high mortality rate.

The inflammatory process is believed to play a major role in the exacerbation of CF lung disease (4, 18). In fact, the inflammatory response, mostly mediated by PMN, not only fails to eradicate the pathogen, but there is a growing body of evidence to suggest that this response contributes to a defect in local host defenses that perpetuates the infection (18). In normal healthy individuals, neutrophils account for less than 5% of alveolar cells, but they become the predominant phagocytic cells in BAL, as CF lung disease progresses. In accordance with the study of Sapru and associates (personal communication) performed in C57BL/6 mice, we have observed that the proportion of alveolar PMN is markedly increased after pulmonary infection with P. aeruginosa. The magnitude of the increment was similar in B6-Cftr (/) versus B6-Cftr (+/+) mice, thereby suggesting that the accumulation of neutrophils is most likely a consequence of host response to P. aeruginosa infection, rather than a direct result of CFTR-mediated chloride conductance dysfunction. Although the high numbers of alveolar cells and PMN are consistent with an inflammatory response, it is possible that the level of activation as well as the proinflammatory mediators secreted by PMN are more significant parameters for determining the level of inflammation in the lungs. This question is currently under investigation in our laboratory.

Overall, our results demonstrate that a loss of CFTR-mediated chloride conductance impairs the ability of mice to control lung infection with P. aeruginosa. The precise mechanism underlying the link between the Cftr gene defect and predisposition of CF patients to chronic P. aeruginosa lung infection remains to be elucidated. Several hypotheses addressing this issue have emerged from in vitro studies. These include mucus hypersecretion and viscosity (26), impaired mucociliary clearance (4, 27), increased adherence of bacteria to epithelial cells (28), defective uptake of P. aeruginosa by epithelial cells (29), and inhibition, under high salt concentration, of the activity of a defensin-like molecule secreted by airway epithelial cells (30). Overall, these findings imply that a loss of Cftr-mediated chloride channel function would affect the natural defense system against lung pathogens. Although we did not specifically address these hypotheses, our results are consistent with this concept in that the increased susceptibility to P. aeruginosa infection in B6-Cftr knockout mice compared with littermate controls was detectable as early as 4 d postinfection. Thus, we have now a useful tool to explore, in vivo, the mechanisms underlying the pathophysiology of P. aeruginosa infection in CF in the hope of hindering the fatal outcome of the disease.