Introduction

Abstract Introduction Materials and Methods Results Discussion References

Pseudomonas aeruginosa is an opportunistic pathogen that colonizes the respiratory tract of patients with cystic fibrosis (CF), bronchiectasis and severe chronic obstructive airways disease (1). Once established, it is difficult to eradicate even with intravenous and/or continuous antibiotic therapy (2). P. aeruginosa produces several toxins that impair the structure and function of the respiratory mucosa in different ways, and infection stimulates an exuberant host inflammatory response, which has also been shown to damage the epithelium (3). P. aeruginosa infection is associated with increased morbidity and mortality, and impaired quality of life (1, 9, 10).

Nitric oxide (NO) is a widely distributed signaling molecule and is one of the most abundant free radicals in the body. It is a multifunctional molecule that at low concentrations is thought to play a key role in neurotransmission, vasodilation, bronchodilation, and the inflammatory response (11, 12). NO also increases ciliary beat frequency and has antimicrobial activity (13, 14). It is produced endogenously in the human lung by the action of nitric oxide synthase (NOS), which exists in constitutive (cNOS) and inducible (iNOS) forms, on the terminal guanidine nitrogen of L-arginine (15). The source of NO is uncertain, but it seems likely that it is derived from cells in both the upper and lower respiratory tracts. NO has been detected in the expired air of animals and normal individuals (16). The increased exhaled NO that occurs in asthma is mainly derived from the lower respiratory tract (17), whereas in normal subjects, exhaled NO may largely originate from the nose (18). Human bronchial epithelial cells express iNOS in response to interleukin-1 (IL-1), interferon- (IFN-), and tumor necrosis factor- (TNF-) (19). NO can also be a cytotoxic molecule, because it reversibly inhibits respiratory enzymes of mitochondria (20) and reacts at near diffusion-limited rates with other free radicals to produce the far more damaging oxidant species peroxynitrite.

Asymmetric dimethyl arginine (ADMA) is an analogue of L-arginine and acts as a false substrate for NOS, thereby inhibiting endogenous NO biosynthesis. Using scanning electron microscopy (SEM), we investigated the effect of ADMA and the inactive enantiomer N^G-methyl-D-arginine (D-NMMA) on the interaction of P. aeruginosa with the respiratory mucosa of an organ-culture model with an air-mucosal interface, and measured nitrite levels in the medium feeding organ cultures. We also investigated the effect of P. aeruginosa culture filtrates (24 h and 36 h) and of several P. aeruginosa toxins on expression of the iNOS gene in an epithelial cell line.

	Materials and Methods

Abstract Introduction Materials and Methods Results Discussion References

Bacteriology

P. aeruginosa strain P455 is a clinical isolate that has previously been studied in our laboratory (21, 22). P455 is a piliated, nonmucoid strain that produces alkaline protease, elastase, pyocyanin, 1-hydroxyphenazine (1-HP), lipase, deoxyribonuclease (DNase), and rhamnolipid. It was stored at 70°C in a brain-heart infusion broth (BHI; Oxoid, Basingstoke, UK) and glycerol mixture (80:20 [vol/vol]), and then retrieved onto BHI agar (Oxoid). After overnight culture, two or three colonies were dispersed in 5 ml of BHI broth and incubated overnight at 37°C. The culture was diluted with BHI to give an optical density (OD) of 0.365, which previous experiments had shown corresponded to approximately 1.0 × 10⁸ cfu/ml. The culture was centrifuged and washed twice with 10 ml of phosphate-buffered saline (PBS; Oxoid). The bacterial pellet was resuspended in 1 ml of PBS, and counts of viable cells were made.

P. aeruginosa (P455) culture filtrate was prepared by incubating two or three colonies of the organism in 5 ml of BHI broth for either 24 h or 36 h. The culture was then centrifuged and filtered through a sterile Millipore filter (0.45 µm) to produce a cell-free culture filtrate.

Preparation of ADMA and D-NMMA

ADMA or D-NMMA (Sigma, Poole, UK) were dissolved in minimal essential medium (MEM; Gibco, Paisley, UK) to yield the final concentrations used in experiments.

Preparation of Reagents for Nitrite Assay

The following reagents (all from Sigma) were prepared at the specified concentrations: sodium nitrite (10 mg/ml), sulfanilic acid (37.5 mM), N-(1-naphthyl)-ethylenediamine (dihydrochloride) (12.5 mM), and concentrated HCl (6.5 M).

Organ Cultures

The method used for organ culture has been described previously (21). Briefly, human turbinate tissue was resected from patients undergoing surgery for nasal obstruction, and was transported to the laboratory in MEM containing antibiotics (50 µg/ml streptomycin, 50 IU/ml penicillin, and 50 µg/ml gentamicin). Dissection was performed in antibiotic medium to yield small squares of tissue approximately 3 mm² in area and 2 to 3 mm thick. The tissue was screened for ciliary beating with a Dialux 20 phase-contrast microscope (Leitz, Wetzlaar, Germany) with a warm stage at 37°C. Only tissue squares with at least one fully ciliated edge were selected. The tissue was immersed in antibiotic medium for at least 4 h to remove commensal bacteria, and was then immersed in non-antibiotic-containing medium for at least 1 h to remove antibiotics.

A sterile 3.5-cm-diameter petri dish (Sterilin, Stone, UK) was placed aseptically within a sterile 6-cm-diameter petri dish. A strip of sterile filter paper (Whatman No. 1, Maidstone, UK) with dimensions approximately 5 mm by 70 mm was soaked in MEM without antibiotics and positioned aseptically across the diameter of the inner petri dish. The filter-paper strip was manipulated with sterile forceps such that its middle portion adhered to the base of the inner petri dish and each of its moistened ends adhered to the base of the outer petri dish. A single tissue square was placed, ciliated surface facing upward, on the center of the filter-paper strip in the inner petri dish. Thirty microliters of molten agar (Oxoid No. 1) at approximately 40°C were carefully pipetted around the organ culture. The agar set as it cooled to 37°C, and formed a seal around the tissue edges. Care was taken to avoid the formation of a rim of agar above the level of the tissue surface. Four milliliters of MEM without antibiotics or other pharmacologic agents were pipetted into the outer petri dish. The filter-paper strip acted as a wick to draw medium from the outer petri dish to feed the underside of the tissue.

Experimental Design

Three series of experiments were performed (each n = 6). In the first, eight organ cultures were assembled in each experiment: control, tissue preincubated with either 1 × 10⁴ M, 2 × 10⁴ M, or 4 × 10⁴ M ADMA only, tissue infected with P. aeruginosa only, and tissue preincubated with 1 × 10⁴ M, 2 × 10⁴ M, or 4 × 10⁴ M ADMA and subsequently infected with P. aeruginosa. In the second series of experiments, lower concentrations of ADMA were investigated (0.1 × 10⁴ M, 0.5 × 10⁴ M, and 1 × 10⁴ M). In the third series, five organ cultures were assembled to compare the effect of the inactive enantiomer D-NMMA (2 × 10⁴ M) with that of ADMA (2 × 10⁴ M): control, tissue preincubated with D-NMMA only (2 × 10⁴ M), tissue infected with P. aeruginosa only, and tissue preincubated with either ADMA (2 × 10⁴ M) or D-NMMA (2 × 10⁴ M) prior to infection with P. aeruginosa. Tissue squares were preincubated with 4 ml of ADMA (0.1 to 4 × 10⁴ M), D-NMMA (2 × 10⁴ M), or MEM alone for 30 min prior to construction of the organ cultures. Twenty microliters of washed bacteria suspended in PBS were gently pipetted onto the surface of the appropriate tissue. The other organ cultures were inoculated with 20 µl of PBS.

All organ cultures were incubated in a humidified atmosphere at 37°C in 5% CO₂ for 8 h. At the end of an experiment, each of the four edges of the organ culture were touched gently with a sterile loop and plated onto BHI agar to assess the sterility of the control, ADMA-only-, and D-NMMA-only-treated organ cultures, and the purity of P. aeruginosa growth in the infected organ cultures. The filter-paper strips were then cut near the tissue with a sterile blade, removed with tissue attached, and fixed for SEM.

Assessment of Tissue with SEM

Tissue was fixed and processed for SEM as previously described (22). All specimens were coded and randomized prior to analysis. Each tissue square was examined with a Hitachi S-4000 scanning electron microscope (Katsuta-shi, Ibaraki-Ken, Japan) by the same observer, who was unaware of the experimental protocol. The tissue was initially viewed at ×50 magnification. A transparent acetate sheet with 100 equal squares was placed over the screen of the visual display unit, and a predetermined pattern of 40 grid squares was selected for further viewing and analysis at ×3,000 magnification (22). This pattern involved the horizontal, vertical, and both diagonal axes, thus giving a representative survey of the mucosal surface measuring 1.42 × 10⁴ µm². Each square was assessed for percentage of the surface area occupied by four mucosal features: mucus, damaged epithelium, ciliated cells, and unciliated cells. Extruding cells, cell debris, dead cells, and loss of epithelium were scored together in the category of damaged epithelium. Unciliated areas were defined as areas not covered by cilia, with or without microvilli. Summation of the scores allowed an assessment to be made of the percentage of the organ-culture surface that was occupied by each mucosal feature.

The number of bacteria associated with each of the four mucosal features was counted. An approximation was made when large numbers of bacteria were present in sheets. In these instances it was difficult to determine to which mucosal component(s) the bacteria were adhering, but observation of the tissue surrounding the bacteria enabled an estimate to be made. The density of bacteria adhering to each mucosal component was calculated in order to overcome the difficulty caused by different proportions of the surface of each organ culture being occupied by the mucosal features. The number of bacteria adhering to a mucosal feature was divided by the proportion of the surface of the organ culture occupied by that feature to give the number of bacteria adhering per unit area (3.55 × 10² µm²), at ×3,000 magnification, to each mucosal component (22). The total number of bacteria adhering to an organ culture in the area surveyed at ×3,000 magnification (1.42 × 10⁴ µm²) was also calculated.

Nitrite Assay

At the end of experiments with ADMA, the 4 ml of MEM in the outer petri dish of each organ culture were saved and stored at 70°C. Nitrite concentrations were determined via the Greiss reaction (24). To determine whether P. aeruginosa itself produced NO, nitrite levels were measured after overnight incubation in BHI broth and MEM.

Expression of iNOS

Cell culture. For each experiment (n = 5), A549 cells were grown to confluence in Dulbecco's MEM (Gibco, Paisley, UK) with 10% fetal calf serum (FCS), 100 IU/ml penicillin, 100 µg/ml streptomycin, and 2 mM L-glutamine (Gibco). These cells were used instead of nasal turbinate tissue to extract an adequate amount of mRNA for subsequent probing. Cells were incubated for 24 h with 2 ml of serum-free medium, and were then transferred to 1 ml of serum-free medium containing P. aeruginosa 24-h culture filtrate (3:1 or 1:1 with cell culture medium), 36 h culture filtrate (1:1), BHI broth alone, IL-1 (10 ng/ml) (Sigma), or one of the P. aeruginosa toxins lipopolysaccharide (LPS) (10 µg/ml) (Sigma), elastase (2 µg/ml) (Nagase Biochemicals Ltd., Fukuchiyama, Japan), or 1-HP (12 µg/ml) (prepared by photolysis of phenazine metasulphate), and incubated for 24 h.

Isolation of RNA and Reverse Transcription-Polymerase Chain Reaction

Cells were harvested and RNA was extracted according to the method of Gough (25). Reverse transcription (RT) reactions were done as previously described (26). A 1/20 volume was used for polymerase chain reaction (PCR), with 0.5 units Taq polymerase (Boehringher Mannheim, Lewes, UK), according to the manufacturer's specifications, and 0.125 µg of each primer in a 25-µl reaction volume, using a Hybaid OmniGene thermocycler (Hybaid, Teddington, UK). PCR primers (5' to 3') were: iNOS, 5'-GAG CTT CTA CCT CAA GCT ATC-3' (sense) and 5'-CCT GAT GTT GCC ATT GTT GGT GGT-3' (antisense); and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) 5'-CCA CCC ATG GCA AAT TCC ATG GCA-3' (sense) and 5'-TCT AGA CGG CAG GTC AGG TCC ACC-3' (antisense). Annealing temperatures were 62°C and 58°C, and product sizes were 312 bp and 598 bp, respectively. Cycling parameters were 94°C for 30 s, specific annealing temperature for 30 s, and extension at 72°C for 30 s. The number of PCR cycles used was that necessary to achieve exponential amplification as subsequently described, followed by a 10-min extension at 72°C. In each case PCR products were cloned and their identity verified by double-stranded sequencing with Sequennase II (Amersham International, Buckinghamshire, UK).

PCR Cycle Profile for Establishing the Exponential Phase of Amplification

A cycle profile was created to determine the exponential phase of DNA amplification, in which product formation is related to the starting template (27). An "average" sample was made by mixing equal amounts of complementary DNA (cDNA) from all RT reactions in a batch to be tested. PCR was then performed on this average sample at various cycle numbers, with positive and negative controls at the maximum number of cycles tested. Aliquots (10 µl) were run on 2% agarose gels stained with ethidium bromide, and were analyzed visually. The cycle number needed to just visualize a product in the average sample was then used for analysis of the entire batch.

Semiquantization, Southern Blotting, and Cerenkov Counting

PCR was performed in duplicate for 35 and 24 cycles for iNOS and GAPDH, respectively, for each sample. Aliquots (10 µl) were size-fractioned on 1.5% agarose gels. Southern hybridization with the appropriate cloned cDNA was performed according to standard procedures (28) to confirm the identity of the products generated, and allowed the exclusion of possible genomic DNA contamination (data not shown). In addition, 5 µl of each PCR reaction mixture was dot-blotted onto Hybond-N (Amersham International, Buckinghamshire, UK), and the PCR product was hybridized to the appropriate cDNA probe (28). Dot blots were excised and their radioactivity measured by Cerenkov counting. We have previously found that this technique yields a linear relationship between starting cDNA and product formation (cpm) for a range of 2 to 3 log units. Data were expressed as the ratio of iNOS to GAPDH as a percentage of IL-1.

Statistical Analyses

Values are described as means ± SE. Comparisons of the mean percentage surface area occupied by each of the four mucosal features in the organ-culture model were made with the Mann-Whitney U test. Comparisons of bacterial densities adherent to each mucosal feature and of total bacterial numbers adherent to the respiratory mucosa were made with Wilcoxon's matched signed-ranks test. Nitrite levels were compared by means of the Mann-Whitney U test, and iNOS/GAPDH mRNA levels were analyzed with a two-sample unpaired t test. Values of P  0.05 were judged to be significant.

	Results

Abstract Introduction Materials and Methods Results Discussion References

Bacteria

The mean inocula of P. aeruginosa in 20 µl of PBS for the infected organ cultures were 5.7 ± 2.7 × 10⁶ cfu, 2.4 ± 1.3 × 10⁶ cfu, and 9.5 ± 0.5 × 10⁶ cfu for the three series of experiments (higher concentrations of ADMA, lower concentrations of ADMA, and comparisons with D-NMMA, respectively). The inocula for the third series of experiments were significantly (P  0.05) higher than those for the first two. However, there were no significant differences within each experimental protocol, and so comparisons within each protocol could be made. At 8 h, all control and ADMA-only or D-NMMA-only organ cultures were sterile, and all P. aeruginosa-infected organ cultures gave a pure growth.

SEM

Control organ cultures at 8 h exhibited very little mucosal damage and were well ciliated. ADMA alone at all concentrations used (0.1 to 4 × 10⁴ M), and D-NMMA (2 × 10⁴ M), had no effect on the mucosal features. P. aeruginosa infection of organ cultures caused a significant (P  0.05) increase in mucosal damage and a significant (P  0.01) decrease in the number of both ciliated and unciliated cells as compared with control cultures (Table 1 and Figure 1A).

View larger version (151K): [in this window] [in a new window]   View larger version (152K): [in this window] [in a new window]   Figure 1.   (A) Scanning electron micrograph of nasal turbinate tissue infected with P. aeruginosa. Bacteria are seen adhering to damaged cells and mucus. 1 cm = 4.0 µm. (B) Scanning electron micrograph of nasal turbinate tissue preincubated with ADMA (2 × 104 M) and infected with P. aeruginosa. There is less damage to the epithelium, but bacteria still adhere to the damaged areas and to mucus. 1 cm = 4.0 µm.

Tissue preincubated with ADMA (1 × 10⁴ M, 2 × 10⁴ M, or 4 × 10⁴ M) prior to P. aeruginosa infection had significantly less mucosal damage (P  0.01) and more ciliated cells (P  0.05) than did infected tissue not exposed to ADMA (Table 1 and Figure 1B). There were some concentration-dependent effects of ADMA, because tissue preincubated with 1 × 10⁴ M ADMA exhibited significant (P  0.05) differences in damaged epithelium and ciliated cells, compared with control tissue, which were not present with the higher concentrations. In addition, tissue preincubated with 2 × 10⁴ M ADMA prior to P. aeruginosa infection had significantly (P  0.05) more ciliated cells than did tissue preincubated with 1 × 10⁴ M ADMA. In the second series of experiments, tissue preincubated with either 0.1 × 10⁴ M or 0.5 × 10⁴ M ADMA prior to P. aeruginosa infection had more mucosal damage and less ciliated cells than tissue preincubated with 1 × 10⁴ M (Table 2). Tissue treated with the lowest concentration of ADMA (0.1 × 10⁴ M) prior to P. aeruginosa infection was not significantly different from P. aeruginosa-infected tissue with respect to mucosal damage and number of ciliated cells.

In the third series of experiments, tissue preincubated with D-NMMA (2 × 10⁴ M) prior to P. aeruginosa infection had significantly (P  0.01) more mucosal damage than tissue preincubated with ADMA (2 × 10⁴ M), and did not differ significantly from tissue infected with P. aeruginosa alone (Table 3). Tissue preincubated with ADMA (2 × 10⁴ M) prior to bacterial infection again showed significantly (P  0.05) reduced P. aeruginosa-induced epithelial damage and loss of ciliated cells. However, in this third series of experiments, tissue preincubated with ADMA (2 × 10⁴ M) prior to bacterial infection had significantly (P  0.01) more mucosal damage than did control tissue. This may have been due to the significantly (P  0.05) larger number of bacteria inoculated onto the tissue (see the previous discussion).

P. aeruginosa Adherence to Organ Cultures

The interactions of P. aeruginosa with the organ cultures were similar to those previously reported (21, 22, 29). Bacteria were commonly seen adhering to mucus and damaged epithelium, particularly in the gaps between separated epithelial cells. There was no significant difference in the density of bacteria adhering to each individual mucosal feature in the presence or absence of ADMA or D-NMMA at any of the concentrations investigated (Tables 4, 5 and 6). However, ADMA at all concentrations analyzed, but not D-NMMA, significantly (P  0.04) reduced the total number of bacteria adherent to organ cultures. The difference between the total bacterial numbers in the three protocols is probably explained by the size of the inocula, which was greatest in the third series of experiments. Tissue preincubated with 0.1 × 10⁴ M ADMA in the second series of experiments had significantly (P  0.04) more bacteria adhering to the respiratory mucosa than did tissue preincubated with 1 × 10⁴ M ADMA, although the numbers were still significantly (P  0.01) smaller than those for untreated tissue. In separate experiments we showed that ADMA did not significantly affect P. aeruginosa growth in vitro (data not shown).

Nitrite Assay

Tissue infected with P. aeruginosa alone had significantly (P  0.02) greater nitrite levels in surrounding medium than did control tissue (Figure 2). Preincubation of tissue with ADMA (1 × 10⁴ M, 2 × 10⁴ M, and 4 × 10⁴ M) prior to bacterial infection did not significantly reduce the level of nitrite in the surrounding medium as compared with tissue exposed to P. aeruginosa alone, but there was a trend in reduction of the nitrite level (Figure 2). Tissue preincubated with the lower concentrations of ADMA had nitrite levels that were similar to those obtained with P. aeruginosa infection alone (data not shown). P. aeruginosa cultured in BHI broth or MEM in the absence of tissue had no effect on nitrite concentrations (data not shown).

View larger version (61K): [in this window] [in a new window]   Figure 2.   Effect of ADMA on P. aeruginosa-induced increase in nitrite levels. *P  0.02 versus control. P.a = P. aeruginosa; P.a & 1 = P. aeruginosa and 1 × 104 M ADMA; P.a & 2 = P. aeruginosa and 2 × 104 M ADMA; P.a & 4 = P. aeruginosa and 4 × 104 M ADMA.

Induction of iNOS mRNA in A549 Cells

Incubation of A549 cells with IL-1 (10 ng/ml) for 24 h resulted in an average 13.6-fold increase in iNOS/GAPDH in the series of experiments in which cells were exposed to P. aeruginosa culture filtrates (Figure 3A), and an average 9.2-fold increase when cells were exposed to individual P. aeruginosa toxins (Figure 3B), as compared with basal levels. iNOS mRNA levels were expressed as a ratio with GAPDH, and were then expressed as a percentage of IL-1-induced iNOS mRNA levels. P. aeruginosa culture filtrates at 24 h (1:1 and 3:1 with cell culture medium) and 36 h (1:1) caused a significant (P = 0.02) increase in iNOS/GAPDH mRNA levels over basal levels (Figure 3A). However, the individual P. aeruginosa toxins LPS, elastase, and 1-HP had no effect at the concentrations tested (Figure 3B).

View larger version (50K): [in this window] [in a new window]   Figure 3.   Effect of P. aeruginosa culture filtrates and individual toxins on iNOS expression in an epithelial cell line. (A) IL-1 (10 ng/ml), P. aeruginosa culture filtrate 24 h (1:1), P. aeruginosa culture filtrate 24 h (3:1), P. aeruginosa culture filtrate 36 h (1:1), and BHI broth. (B) IL-1 (10 ng/ml), LPS (10 µg/ml), elastase (2 µg/ml), and 1-HP (12 µg/ml). *P = 0.02 versus control; P = 0.01 versus BHI broth; P = 0.001 versus IL-1.

	Discussion

Abstract Introduction Materials and Methods Results Discussion References

NO is a highly reactive free radical that is involved in a wide range of physiologic processes, and has been implicated in the pathophysiology of airways disease. In animals, iNOS is expressed in various respiratory cells, such as alveolar macrophages, lung fibroblasts, and bronchial epithelial cells (30). Proinflammatory cytokines (e.g., IL-1, TNF-, and IFN-) have previously been shown to upregulate the expression of iNOS in human and murine lung epithelial cells (19), and Escherichia coli LPS (50 µg/ ml) has been shown to cause a slight increase in nitrate/ nitrite concentrations in medium overlying rat pleural mesothelial cells both on its own and in combination with various cytokines (e.g., IFN, TNF, and IL-1) (33, 34). Inducible NOS activity, measured as the ability of tissue homogenates to convert L-arginine to L-citrulline, has also been shown to be increased in inflammatory lung diseases such as asthma and CF (11).

In the present study, we used an organ-culture model with an air-mucosal interface to investigate the effect of ADMA and D-NMMA on P. aeruginosa interactions with the respiratory mucosa in vitro. This organ-culture model is an endeavor to mimic the physiologic conditions present in vivo (23). The model retains a fully differentiated epithelium and mucus blanket, and we have previously shown that it is sufficiently sensitive to discriminate between closely related bacterial strains (35). In the present study, we used human nasal turbinate tissue resected from patients undergoing surgery for nasal obstruction. Although this tissue cannot be considered normal, the relative effects on it of P. aeruginosa infection and prior incubation with ADMA and D-NMMA remain valid.

P. aeruginosa interactions with the mucosa were similar to those previously reported (21, 22, 29). Infection of the respiratory mucosa with P. aeruginosa induced a reduction in the number of ciliated cells and caused extensive epithelial damage, as evidenced by the presence of cell extrusion with tight-junction separation, dead cells, and cell debris. The epithelium was frequently stripped away, exposing basement membrane and the collagen layer (22). Bacteria were seen to associate most frequently with damaged cells, particularly in the gaps between separated tight junctions, and with mucus, on which they formed a biofilm.

P. aeruginosa infection of organ cultures was accompanied by increased levels of nitrite in the surrounding medium, which were reduced (but not significantly) when the tissue was preincubated with ADMA (Figure 2). Epithelial damage and loss of ciliated cells caused by infection were significantly (P  0.05) reduced by preincubation of tissue with the NOS inhibitor in a concentration-dependent manner (Figure 1B). Conversely, D-NMMA, an inactive enantiomer, had no effect. In addition, P. aeruginosa culture filtrates induced a significant (P  0.02) increase in iNOS/GAPDH mRNA levels in epithelial cells of the A549 line as compared with basal levels (Figure 3A). However, the identity of the P. aeruginosa toxin(s) responsible for upregulating iNOS remains to be elucidated. Taken together, these data suggest that NO is a mediator of P. aeruginosa-induced epithelial damage. The results also suggest that relatively small changes in the mean level of NO production are associated with epithelial protection. However, the relative insensitivity of the Greiss reaction, which we used to assay nitrite levels, and the fact that in the organ-culture model the tissue is not directly in contact with the medium that we assayed may account for the large SEs achieved and hence for the nonsignificance of the nitrite data. In addition, we cannot exclude the possibility that ADMA has a cytoprotective effect independent of NO.

The classification of NO as a "good" or "bad" biomolecule is the subject of much debate. It is generally agreed that NO produced in small amounts from the constitutive form of the enzyme (cNOS) can produce beneficial physiologic effects. For example, NO may play a host-defense role by killing pathogens such as Leishmania, Mycobacterium tuberculosis, and malaria parasites, and is also toxic to tumor cells (36). At low concentrations NO is a bronchodilator, since it mediates the nonadrenergic, noncholinergic neural inhibitory responses in human airways (37). NO may play an important role in regulating mucociliary clearance by increasing ciliary beat frequency (13).

However, NO produced by iNOS may have cytotoxic and even genotoxic effects in the airways, depending on its concentration and the chemical changes it undergoes in a given biologic microenvironment. NO may be responsible for the epithelial shedding frequently observed in asthmatic individuals, and it has been shown that NO is capable of rapidly and reversibly inhibiting cytochrome C oxidase, the terminal enzyme in the mitochondrial respiratory chain (20). NO is a potent vasodilator in the bronchial circulation, and may potentiate airways inflammation by increasing blood flow to leaky postcapillary venules, resulting in increased exudation of plasma (11, 38).

NO has been implicated in the epithelial damage caused by another respiratory pathogen, Bordetella pertussis. B. pertussis tracheal cytotoxin (TCT), a muramyl peptide fragment secreted during bacterial growth, damages ciliated epithelial cells and induces IL-1 production by hamster tracheal epithelium. Exogenous IL-1 reproduced the cytopathology caused by TCT (39). Both TCT and IL-1 induced high levels of NO production by epithelial cells, and inhibition of NOS prevented the destruction of ciliated cells in hamster tracheal organ cultures (40). These observations suggest that TCT triggers the production of IL-1, which in turn stimulates NO production leading to epithelial cell damage.

NO has the potential to combine with other reactive species to produce far more damaging cytotoxic molecules. In the microenvironment of the lung, both NO and superoxide are often produced simultaneously from inflammatory cells, and may combine to produce peroxynitrite, a strong oxidant that has been implicated in diverse forms of free-radical-induced tissue injury (41, 42). Peroxynitrite can attack many types of biologic molecules and produce a wide variety of effects. It has been shown to induce cell injury involving DNA strand breakage. This then activates polyadenosine diphosphate ribosyl synthetase, triggering a futile repair cycle and leading to cellular energy depletion (43). Peroxynitrite also causes the oxidation of tryptophan and cysteine, and the nitration of tyrosine, resulting in increased protein fragmentation (44), and it has been implicated in oxidative injury to isolated rat lung from ischemia-reperfusion (42).

NOS inhibitors have anti-inflammatory activity. L-N^G-monomethyl-arginine (L-NMMA) reduced inflammation, downregulated inflammatory cytokines, and enhanced IL-10 production in carrageenin-induced edema in mice (45). Analogues of L-arginine protected against immune complex-induced vascular injury brought about by activated macrophages (46). Hence, ADMA may protect the respiratory epithelium against damage caused by P. aeruginosa infection in a variety of ways. It may downregulate the production of inflammatory cytokines, reduce NO-mediated inhibition of mitochondrial respiration and DNA damage, and reduce the production of damaging oxidant species.

If the observations in the present study are confirmed in vivo, iNOS inhibitors may be used to prevent tissue damage caused by P. aeruginosa infection. This therapeutic approach is attractive, since it is usually impossible to eradicate P. aeruginosa despite prolonged courses of antibiotics to which the bacterium is sensitive in vitro (2, 8- 10). Total bacterial numbers adherent to the mucosa of the organ culture in our study were reduced, despite ADMA having no antibacterial activity. This may be explained by ADMA reducing the amount of epithelial damage and thus decreasing the number of sites on the epithelium to which P. aeruginosa can adhere. ADMA may also decrease the availability of nutrients for bacterial growth that are released by damaged cells. The effect of ADMA on bacterial numbers in the present study was concentration-dependent, but even at low concentrations there was a significant reduction in bacterial numbers. More research is needed to elucidate the P. aeruginosa toxin(s) that upregulate iNOS levels, and to differentiate between the beneficial and adverse effects of NO before results obtained in an organ culture can be extrapolated to the in vivo situation.