INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Keywords: cystic fibrosis; interleukin-10; inflammation; Pseudomonas aeruginosa; lung

Most patients with cystic fibrosis (CF) succumb to a lung disease that is characterized by chronic bacterial infection and neutrophilic inflammation (1). Although various bacterial infections may wax and wane early in life, most CF patients eventually become infected with a dominant organism, usually Pseudomonas aeruginosa, which they can no longer eradicate. In response to the bacterial infection, chemoattractants are produced that increase neutrophil (PMN) influx into the airways. PMNs release elastase and other mediators that help contain the infection, but excess quantities of these mediators damage host tissues, ultimately resulting in the death of the patient. The kinetics of the establishment of infection and their relationship with the subsequent inflammatory response are poorly understood. Although bronchoalveolar lavage (BAL) fluid from infected CF infants contains greatly elevated PMNs, interleukin (IL)-8, and PMN elastase, these mediators are also present at lower concentrations in BAL fluid from uninfected CF infants (2). We speculate that these latter infants may have been transiently infected and then had a prolonged inflammatory response that persisted after the infection had been cleared, on the basis of similar findings in CF infants who underwent bronchoscopy with BAL before and after intravenous antibiotic therapy for acute infection (3).

Until now there have been few studies evaluating the kinetics of inflammation in the CF lung. However, there are data to suggest that this inflammatory response is excessive relative to the burden of infection (4). BAL fluid from CF infants that was positive for Haemophilus influenzae contained more PMNs and IL-8 than did BAL fluid from non-CF infants with similar burdens of H. influenzae (5). Defects in the CF transmembrane conductance regulator (CFTR) may lead to increased production of IL-8 in response to P. aeruginosa and other stimuli and may be due to excessive activation of transcription factor NF-B (6). Recent evidence also suggests that deficient production of IL-10 in the lung may be associated with CFTR defects (3, 7-9). In addition, stimulated T cells from CF patients produced less IL-10 in vitro than did stimulated T cells from healthy volunteers (10). Furthermore, the concentration of IL-10 in BAL fluid from uninfected CFTR knockout (KO) mice contained little IL-10 compared with BAL fluid from uninfected wild-type (WT) mice (11). Likewise, concanavalin-A- stimulated lymphocytes from CFTR KO mice produced less IL-10 than did similarly stimulated lymphocytes from WT mice (11). Decreased IL-10 is important because it has potent anti-inflammatory and immunoregulatory activities. It is now recognized that in addition to inhibiting cytokine production by T lymphocytes, IL-10 has multiple other effects including inhibiting production of the PMN chemoattractant IL-8 and other NF-B-dependent cytokines including tumor necrosis factor alpha (TNF-) and IL-1 (12-16). By inhibiting proinflammatory cytokine production, IL-10 would tend to quell inflammatory reactions and restore homeostasis (12-14). At least some of the effect of IL-10 may involve increasing the production of IB, the inhibitor of NF-B activation (17, 18). IL-10 may also facilitate resolution of lung inflammation by promoting apoptosis of PMNs (19). We propose that decreased IL-10 in the CF airway may be a preexisting immunoregulatory abnormality that allows inflammation to persist even after acute infectious stimuli have been cleared. Recurrent episodes of abnormally persistent inflammatory responses may then lead to a vicious cycle of chronic inflammation that could cause lung damage. Understanding the kinetics of inflammatory responses to infection in CF is thus important in developing new anti-inflammatory therapeutics.

We hypothesize that IL-10 is important in terminating the inflammatory response and that in its absence, inflammation persists even after the stimulus has been removed. To evaluate this hypothesis, we studied intra- and extracellular regulators of the inflammatory response to acute pulmonary P. aeruginosa infection in IL-10 KO and WT mice. Our results demonstrate that mice deficient in IL-10 had prolonged and excessive proinflammatory cytokine production and neutrophil infiltration in the airways that persisted after the bacteria had been eradicated. This was associated with impaired replacement of IB and prolonged activation of NF-B in the IL-10 KO mice. These findings provide new information on the mechanisms underlying the resolution of the inflammatory response to acute infection in normal individuals. Furthermore, they suggest that IL-10 deficiency leads to a protracted as well as excessively intense inflammatory response in CF and that altered kinetics of resolution of inflammatory responses may contribute to the observations of inflammation in the absence of infection in infants with CF.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Experimental Model

To generate a transient infection, seventy-five 7 to 10-week old male IL-10 KO mice (20) on a C57Bl/10J background and 62 age-, sex-, and weight-matched WT (C57Bl/10J) mice (Jackson Laboratories, Bar Harbor, ME) received right mainstem bronchus inoculations (21) of 5 × 10⁶ CFU of mucoid P. aeruginosa, strain M5715 (a clinical isolate), suspended in 20 µl of phosphate-buffered saline (PBS). Mice were randomized at the time of inoculation as to which day (Day 2, 4, 6, or 8) postinoculation they were to be killed and then randomly assigned to one of two outcome measures: removal and homogenization of lungs for quantitative bacteriology or BAL for analysis of inflammatory cells and mediators, followed by preparation of lung homogenates for Western blot analysis for IB and electrophoretic mobility shift assay (EMSA) for activated NF-B. Five uninfected IL-10 KO and 5 uninfected WT mice were also killed to determine baseline parameters. The protocol was approved by the Institutional Animal Care Committee of Case Western Reserve University.

Quantitative Bacteriology

Quantitative bacteriology was determined from whole lung homogenates (21). To determine the presence of bacteremia, quantitative cultures were performed on spleen homogenates from mice that died spontaneously and from mice whose lungs were designated for quantitative bacteriology.

BAL Fluid Analysis

BAL was performed and processed as previously described (21). BAL fluid was assayed for cell count with differential and for TNF-, IL-1, IL-6, KC/N51 (KC), and MIP-2 by enzyme-linked immunosorbent assay (R&D Systems, Minneapolis, MN).

Western Blot Analysis for IB

After completion of the BAL, the pulmonary vasculature was flushed with 1 ml of ice-cold sterile PBS. Lungs were removed and homogenized, and extracts were prepared as described previously (17). Equal amounts of extracts were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to nitrocellulose. Blocking was performed in 25 mM Tris-HCl, pH 8.0, 125 mM NaCl with 0.1% Tween-20 and 5% nonfat dry milk (TBST-M). Primary and secondary antibodies (Cell Signaling Technology, Inc., Santa Cruz, CA; Amersham Pharmacia Biotech, Inc., Piscataway, NJ) were diluted in TBST-M and detected by enhanced chemiluminescence (Amersham Pharmacia Biotech).

EMSA for Activated NF-B

Nuclear extracts of whole lung tissues were prepared (22), and EMSA (17) was performed with a  [³²P] ATP (Amersham Pharmacia Biotech) end-labeled double-stranded NF-B consensus oligonucleotide (Promega, Madison, WI).

Statistical Analysis

Two approaches were employed for quantitative bacteriology data. The first assigned a log₁₀ value of 0.50 to all undetectable values, and the Wilcoxon rank sum test tested the hypothesis of equal medians between mouse groups at each day postinoculation. The second dichotomized the data into detectable versus nondetectable values, and the Fisher's exact test tested the hypothesis of equal proportions of detectable values between mouse groups. Two-factor analyses of variance were used with mouse group and day for log₁₀-transformed BAL fluid cell counts. For the cytokine data the minimum detectable limit was assigned to those values below the limit of detection, and the Wilcoxon rank sum test was used. Data are expressed as mean ± SEM.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Because our major goal was to determine the effect of IL-10 deficiency on the clearance of an endobronchial P. aeruginosa infection and the resolution of the resulting inflammatory response, we studied acute infection. In this acute model, P. aeruginosa is suspended evenly in sterile PBS. The dose of bacteria for these studies, 5 × 10⁶ CFU in 20 µl of suspension, was chosen on the basis of preliminary studies that demonstrated that at higher doses, both IL-10 KO and WT mice developed increased mortality. Mice were killed at 2-day intervals over 8 days. Because mice were killed according to an a priori schedule throughout the time course, overall survival was not a primary outcome measure. Of the 75 IL-10 KO mice inoculated with P. aeruginosa, nine animals died before their designated day of killing (88% survival), and of the 62 WT mice inoculated with P. aeruginosa, five animals died before their designated day of killing (92% survival). These values were not significantly different.

Quantitative Bacteriology

To determine if there was an effect of the absence of IL-10 on the clearance of the bacteria, quantitative bacteriology was performed on lung homogenates from mice killed on Days 2 (n = 10, IL-10 KO; n = 7, WT), 4 (n = 8, IL-10 KO; n = 7, WT), 6 (n = 7, IL-10 KO; n = 9, WT), and 8 (n = 6, IL-10 KO; n = 5, WT). There was no difference in the ability of IL-10 KO and WT mice to clear the P. aeruginosa infection (Figure 1). Both groups of mice cleared the bacteria promptly. No bacteria were detected in any IL-10 KO mice after Day 4. However, cultures of lung homogenates from one WT mouse killed on Day 6 (1.9 × 10² CFU) and one WT mouse killed on Day 8 (1.3 × 10² CFU) still had detectable P. aeruginosa. When the log₁₀ CFU/mouse lung was compared for each individual time point, there was no significant difference between IL-10 KO and WT mice (Day 2, p = 0.56; Day 4, p = 0.95; Day 6, p = 0.38; Day 8, p = 0.29). Results from Fisher's exact test concluded that the mouse groups were similar in proportion of detectable values at all days. Of the 31 IL-10 KO and 28 WT mice killed for quantitative bacteriology, cultures of spleen homogenates were positive for P. aeruginosa from one mouse in each group. Both of these mice were killed 2 days after inoculation. The culture of the spleen homogenate from the IL-10 KO mouse that was positive for P. aeruginosa grew 4 × 10¹ CFU, whereas the culture of the spleen homogenate from the WT mouse that was positive for P. aeruginosa grew 2.9 × 10³ CFU. Of the nine IL-10 KO mice that died before their designated day of killing, cultures of spleen homogenates were positive for P. aeruginosa from only one mouse (2.1 × 10³ CFU, mouse died 2 days postinoculation), and of the five WT mice that died before their designated day of killing, cultures of spleen homogenates were also positive for P. aeruginosa from only one mouse (2 × 10⁴ CFU, mouse died 4 days postinoculation).

View larger version (16K): [in this window] [in a new window]   Figure 1.   Quantitative bacteriology in IL-10 KO (open squares, n = 31) and WT (closed circles, n = 28) mice inoculated with P. aeruginosa. Values represent the mean ± SEM of log10 of CFU per mouse lung 2, 4, 6, and 8 days after inoculation. The CFU at Day 0 indicate the inoculating dose. There was no difference between IL-10 KO and WT mice in the number of bacteria isolated at any time point as determined by the Wilcoxon rank sum procedure.

BAL Fluid Cell Counts and Differentials

Thirty-five infected IL-10 KO mice and 29 infected WT mice were randomized to undergo BAL. Five additional uninfected mice in each group were killed to determine baseline BAL fluid cell counts and differentials on Day 0. The infected mice were further randomized a priori to be killed on Days 2 (n = 9, IL-10 KO; n = 6, WT), 4 (n = 8, IL-10 KO; n = 7, WT), 6 (n = 10, IL-10 KO; n = 8, WT), and 8 (n = 8, IL-10 KO; n = 8, WT). The total number of leukocytes in BAL fluid from both groups was greatest 2 days after inoculation and decreased thereafter. However, the decrease toward baseline was more rapid in WT mice. IL-10 KO mice had significantly more total leukocytes than did WT mice (log₁₀ leukocytes/ml BAL) on Days 4 (5.89 ± 0.12 versus 5.30 ± 0.08, p < 0.0001) and 6 (5.80 ± 0.11 versus 5.31 ± 0.06, p = 0.0002).

Because PMNs may play a particularly important role in the pathogenesis of CF lung disease, the relationship between IL-10 and PMNs may be of special interest. Infected IL-10 KO mice had a significantly greater percentage of PMNs in BAL fluid compared with infected WT mice on Days 2 (p = 0.0038), 4 (p < 0.0001), and 6 (p < 0.0001) (Figure 2A). Similar results were seen for the absolute number of PMNs in BAL: IL-10 KO mice had significantly more PMNs than did WT mice (log₁₀ PMN/ml BAL, Figure 2B) on Days 4 (p < 0.0001), 6 (p < 0.0001), and 8 (p = 0.042).

View larger version (18K): [in this window] [in a new window]   Figure 2.   Percent PMNs and absolute numbers of PMNs in BAL fluid in IL-10 KO (squares, n = 35) and WT (circles, n = 29) mice inoculated with P. aeruginosa. (A) Values represent the mean ± SEM of the percent PMNs in BAL fluid 2, 4, 6, and 8 days after inoculation. IL-10 KO mice had a significantly larger percentage of PMNs in BAL fluid on Days 2 (p = 0.0038), 4 (p < 0.0001), and 6 (p < 0.0001) after inoculation. (B) Values represent the mean ± SEM of log10 of PMN per milliliter of BAL fluid on Days 2, 4, 6, and 8 after inoculation. IL-10 KO mice had more PMNs in BAL fluid than did WT mice on Days 4 (p < 0.0001), 6 (p < 0.0001), and 8 (p = 0.042) after inoculation. Statistically significant differences as determined by two-factor analysis of variance are denoted by an asterisk.

BAL Fluid Cytokine Concentrations

In addition to cell counts and differentials, BAL fluid was also analyzed for the presence of the proinflammatory cytokines TNF-, IL-1, and IL-6, and the neutrophil chemokines KC and MIP-2, which are analogous to IL-8 in humans. The concentrations of all of these cytokines were below the limits of detection in BAL fluid from all uninfected mice. Of the days in which cytokines were measured, the greatest concentrations of all proinflammatory cytokines were found 2 days postinoculation and were again undetectable by 4 days postinoculation in infected WT mice. The largest BAL fluid concentrations of all of these proinflammatory cytokines in infected IL-10 KO mice were detected 2 to 4 days postinoculation. All infected IL-10 KO mice had measurable concentrations of all cytokines on Day 2 and TNF- and IL-6 on Day 4. Infected IL-10 KO mice, as compared with infected WT mice, had significantly more TNF- (Figure 3A) on Day 4 and significantly more IL-1 (Figure 3B) and IL-6 (see online data supplement, Figure E1) on Days 2 and 4. Similarly, infected IL-10 KO mice had significantly more of the neutrophil chemoattractants KC (Figure 3C) and MIP-2 (Figure 3D) on Days 2 and 4. Seventy-five percent of infected IL-10 KO mice had measurable concentrations of IL-1 on Day 4, and 87.5% of infected IL-10 KO had measurable concentrations of the neutrophil chemoattractants KC and MIP-2 on Day 4, when these were undetectable in the WT mice. In fact, some BAL fluid cytokines were still measurable in several (1-6) IL-10 KO mice 6 days postinoculation. Two IL-10 KO mice continued to have measurable concentrations of IL-6 and MIP-2 8 days postinoculation.

View larger version (21K): [in this window] [in a new window]   Figure 3.   BAL fluid proinflammatory cytokines in IL-10 KO (squares, n = 35) and WT (circles, n = 29) mice inoculated with P. aeruginosa. Values represent the mean ± SEM of the concentration (pg/ml) of TNF-, IL-1, KC, and MIP-2 in BAL fluid 2, 4, 6, and 8 days after inoculation. Note that as compared with IL-10 KO mice, none of these cytokines was detectable in WT mice after Day 2. All of these proinflammatory cytokines were detectable in some IL-10 KO mice on Day 6. IL-10 KO had significantly more TNF- (A) on Day 4 (p < 0.001), IL-1 (B) on Days 2 (p < 0.001) and 4 (p = 0.007), KC (C) on Days 2 (p < 0.001) and 4 (p = 0.001), and MIP-2 (D) on Days 2 (p = 0.002) and 4 (p = 0.001) as determined by the Wilcoxon rank sum procedure. Statistically significant differences are denoted by an asterisk.

Western Blot Analysis for IB and EMSA for Activated NF-B

Because increasing synthesis of IB has been proposed as being one of the mechanisms by which IL-10 exerts its anti-inflammatory effects (17, 18), homogenates of perfused lungs from mice that had undergone BAL were prepared for Western blot analysis of this important intracellular inhibitor of NF-B activation. A typical result is seen in Figure 4A. In uninfected mice from both groups (Day 0), approximately equivalent amounts of IB were detected by Western blot analysis, consistent with the lack of spontaneous inflammation (similarly low PMNs and undetectable proinflammatory cytokines) at baseline. Two days after inoculation with P. aeruginosa, IB decreased in both WT and IL-10 KO mice. By Days 4 and 6, IB was again observed in WT mice, but not in IL-10 KO mice. Western blot analysis was repeated on several different animals in both groups with identical results. IB was again detectable in IL-10 KO mice by Day 8 (data not shown). The results of Western blot analysis for IB correspond with EMSA results for activated NF-B. No activated NF-B was isolated from lung homogenates from uninfected (Day 0) mice in either group. However, 2 days after P. aeruginosa inoculation, activated NF-B increased in both groups. Activated NF-B was not detected after Day 2 in lung homogenates from WT mice but was still detectable on Days 4 and 6 in lung homogenates from IL-10 KO mice. A typical result is seen in Figure 4B.

View larger version (38K): [in this window] [in a new window]   Figure 4.   A representative Western blot for IB (A) and EMSA for activated NF-B (B) in homogenates of perfused whole lungs from IL-10 KO and WT mice that had undergone BAL and were killed 2, 4, 6, and 8 days after inoculation with P. aeruginosa. Equal protein application was verified by equal staining for -actin. Note that in (B), nonspecific binding, as determined in control animals with unlabeled oligonucleotide (not shown), is denoted by n.s.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The pathway from gene defect to clinical manifestation is incompletely characterized in CF lung disease. However, it is becoming increasingly apparent that the inflammatory response occupies a central pathologic position. Patients with CF are born with pristine lungs, but infection begins early in life and is accompanied by an exuberant, PMN-dominated inflammatory response (4, 23). Although infection triggers the inflammatory cascade, it appears that the local inflammatory response is excessive relative to the bacterial burden and overwhelms local inhibitors, such as antiproteases. Recent evidence suggests that an underlying abnormality in immunoregulation may be present. One potential contributor to this underlying immunoregulatory defect in the lungs of patients with CF may be decreased IL-10 (3, 7-9). Previous studies from our group have shown that IL-10 is constitutively present in normal airways but that it is absent, or its concentration is reduced in CF airways (3, 8, 11). We hypothesized that persistent inflammation following acute infection in conjunction with deficiency of IL-10 might explain observations in CF infants of inflammation in the absence of culturable organisms (2). Serial BAL studies in a few CF infants suggest that they were transiently infected with microorganisms that were cleared, but the inflammatory response persisted after (3). IL-10 may be necessary for applying the brakes to this response in normal lungs. In the absence of IL-10, inflammation persists, even after a transient infectious stimulus has been removed. The inability to appropriately terminate the acute inflammatory response may contribute to establishment of a vicious cycle of obstruction, infection, and chronic inflammation that ultimately spirals out of control (4). We propose that decreased IL-10 in the CF lung may account not only for excessive inflammation during active infection but also for prolonged inflammatory responses that persist after eradication of acute infection in those patients still able to clear their lungs of bacteria.

We previously demonstrated the importance of IL-10 in the quantitative regulation of the ongoing inflammatory response in a murine model of chronic endobronchial P. aeruginosa infection (24). Mice deficient in IL-10 had more drastic weight loss, greater PMN infiltration, larger amounts of lung occupied by inflammation, and higher concentrations of proinflammatory cytokines but no alterations in the bacterial burden. Comparable results were seen for CFTR KO mice (25). We further demonstrated the role of IL-10 in regulating the response to chronic infection by showing that exogenous IL-10 reversed many of the aforementioned findings in normal mice with chronic endobronchial P. aeruginosa infection, again without alterations in the bacterial burden (24).

Because the kinetics of the inflammatory response to transient P. aeruginosa challenge in the presence of a preexisting abnormality of immunoregulation had not been adequately evaluated, we employed a model of acute infection in IL-10 KO mice. The acute infection model contrasts with our previous studies that employed P. aeruginosa embedded in agarose beads to produce chronic infection (24). In the chronic infection model, the agarose beads hold the bacteria in the airway and markedly decrease clearance of the infection; so chronic inflammation ensues (24-27). In the acute model used in these studies, P. aeruginosa is suspended in PBS, and the organisms are killed within the first few days after inoculation; hence, the inciting stimulus is no longer present. Therefore, the kinetics of the resolution of the inflammatory response can be studied once the stimulus for that response has been removed.

In these studies of acute infection, only a few animals of either genotype died before their designated day of killing. All had some degree of lung disease at the time of necropsy. However, on the basis of cultures of spleen homogenates, the incidence of coexistent bacteremia was low. There was no significant difference in overall survival between IL-10 KO and WT mice, but this was not a primary outcome measure because the inoculum was selected to minimize mortality. As mice were killed throughout the 8-day time course, it is still possible, but unlikely, that the survival data may have been censored. We minimized the risk of this by the a priori randomization of mice to day of killing. It is possible that if all of the mice were followed for 8 days, there may have been differences in survival, but this is unlikely because markers of lung pathology were improving. Thus, we designed this study to allow evaluation of the kinetics of the host response to an acute, transient infection. Other studies in different models, some with gram-positive and some with gram-negative bacteria, detected effects of IL-10 on survival (24, 28-30). The repeated aerosol administration of P. aeruginosa to IL-10-deficient mice was associated with increased mortality compared with IL-10-sufficient mice (28). We and others have demonstrated that mice with a P. aeruginosa respiratory tract infection that received exogenous IL-10 had improved survival compared with placebo-treated mice (24, 29), but others found the opposite effect of IL-10 on survival in mice with acute pneumococcal pneumonia (30). We speculate that this disparity is due to inherent differences in the models and differences in the bacteria. Due to the design of this study, we were unable to assess an association between IL-10 and survival.

In order to determine the effect of IL-10 deficiency on the bacterial burden, cultures of lung homogenates were performed at 2-day intervals. Essentially, both groups of mice eradicated the infection before Day 6, with no difference in the kinetics of bacterial clearance in the two groups of mice. Note that beyond Day 4, none of the IL-10 KO mice had culturable P. aeruginosa, whereas one WT mouse killed on Day 6 and one WT mouse killed on Day 8 did. These mice may represent outliers. Importantly, there was neither an increase nor a decrease in the bacterial burden in IL-10 KO. Thus, it is unlikely that alterations in IL-10 affect the ability of host defense mechanisms to contain this infection.

However, when the association between IL-10 and the kinetics of resolution of inflammation was evaluated, marked differences between IL-10 KO and WT mice were appreciable. IL-10 KO mice had more BAL PMNs than did WT mice, as evidenced by both the larger percentage and greater numbers of PMNs. The cytokine data suggest that at least part of the persistence of PMNs in BAL fluid was due to continued production of PMN chemoattractants in the lung. However, as IL-10 may play a role in PMN apoptosis (19), failure of the PMN to undergo programmed cell death may have contributed to their continued presence, but we did not evaluate this directly. Despite this possibility, we speculate that the elevated concentrations of chemokines in IL-10 KO mice, particularly on Days 4 through 8, suggest that these PMN chemoattractants are primarily responsible for the large numbers of PMNs that continued to be present in the airway. It is also possible that IL-10 KO mice responded differently to bacteria-derived chemotactic factors than did WT mice. For WT mice, all of the cytokines we studied in BAL fluid were detectable 2 days postinoculation but undetectable thereafter. However, for IL-10 KO mice, the cytokine concentrations were highest 2 to 4 days after inoculation and then they decreased more slowly. Moreover, some cytokines were still detectable 6 and 8 days after inoculation. Therefore, these mice were similar to those CF infants who were found on serial BAL to have persistently large concentrations of PMNs and IL-8, even after culturable organisms had been eradicated (3). Similar phenomena may be responsible for others' observations of increased PMNs and IL-8 in CF infants with negative BAL fluid cultures (2).

As the production of many of the cytokines evaluated is promoted by NF-B and IL-10 has been postulated to play a major role in upregulation of its inhibitor, IB (17, 18), we thought it prudent to study the amount of IB and activated NF-B in tissue extracts of lavaged, perfused lungs from these mice. Similar amounts of IB were present in both groups of mice before inoculation, and they decreased markedly 2 days after inoculation in both groups. Interestingly, whereas IB was again detectable in WT mice 4, 6, and 8 days after inoculation, it was not detectable again in IL-10 KO until 8 days after inoculation. Changes in IB corresponded with reciprocal changes in activated NF-B, which increased in both groups of mice 2 days after inoculation. Activated NF-B remained detectable in lung homogenates from IL-10 KO mice but not WT mice 4 and 6 days after inoculation. Others reported similar results in a model with a noninfectious inflammatory stimulus (17). IL-10 has been reported to act by other mechanisms as well, and the importance of different mechanisms in natural infection in vivo remains incompletely understood.

Our results support the hypothesis that IL-10 deficiency leads to the failure of appropriate termination of the inflammatory response once it has been initiated by infection. Despite the presence or absence of IL-10, we found no difference in the ability of mice to eradicate an acute bacterial challenge. However, the absence of IL-10 appears to be associated with a prolonged and dysregulated local cytokine response and the persistence of PMNs in the lung after acute infection with P. aeruginosa. One contributor to this protracted inflammatory response may be the failure of IL-10 KO mice to promptly regenerate IB, which was degraded during the initial response to the infection. This allows prolonged and excessive activation of NF-B, which then stimulates transcription of proinflammatory cytokines, including the PMN chemoattractants that mediate the excessive and prolonged inflammatory response. We speculate that continued cycles of prolonged inflammatory responses eventually lead to chronic inflammation and a vicious cycle of unremitting infection that likely results in airway destruction and lung damage. We demonstrated previously that modulation of the inflammatory response by inhibitors of NF-B (24, 27) might hold important therapeutic potential for the attenuation of excessive inflammation during the chronic phase of CF lung disease. The results of the current study suggest that these anti-inflammatory therapies may be even more important in the attenuation of inflammation early in the course of the patient with CF, when they might prevent or delay establishment of the vicious cycle of inflammation, infection, and obstruction.