INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Asthma is an inflammatory disease of the airways characterized by reversible airflow obstruction, eosinophilic bronchitis, and bronchial hyperresponsiveness (BHR). Asthma is also associated with increased nitric oxide (NO) production by the airways. Although the relationship between asthma and NO production is strong enough that exhaled NO has been proposed as a clinically useful marker of asthmatic airway inflammation, the role of NO in the pathophysiology of asthma remains incompletely understood. Specifically, the relationship of NO to eosinophilic inflammation has not been fully elucidated.

Eosinophils are the most prominent effector cells in asthma, and their presence in sputum, bronchoalveolar lavage fluid (BALF), or bronchial biopsy specimens correlates with asthma severity (1). Eosinophils possess an arsenal of mediators capable of contributing to inflammation. For example, eosinophilic granular proteins can promote eosinophil chemotaxis (2) and induce BHR (3). Human eosinophils can produce lipid mediators including peptidoleukotrienes, as well as cytokines such as interleukin (IL)-5 (2). In addition, like other granulocytes, eosinophils produce reactive oxidant species (ROS) (4), including superoxide (O₂) and hydrogen peroxide (H₂O₂). Although there is in vitro evidence that eosinophils produce NO (5), the relative importance of eosinophils in the overall production of NO in asthma is unclear.

Given their capacity to generate ROS, the presence of large numbers of eosinophils suggests that an oxidative milieu is present in the asthmatic airway. In such a context, the production of NO can lead to the generation of highly reactive intermediates. For example, NO combines in a near diffusion-limited fashion with O₂ to produce peroxynitrite (ONOO), which can oxidize lipids and nitrate proteins (6). Recently, evidence of protein nitration, possibly mediated by ONOO, was detected in asthma. Using immunohistochemistry, Saleh and colleagues (7), examining bronchial biopsy specimens from asthmatic airways, detected 3-nitrotyrosine (3NT) that was associated with evidence of increased expression of the cytokine-inducible form of nitric oxide synthase (iNOS or NOS II) (7). Administration of glucocorticoids reduced both 3NT staining and NOS II expression (7). This observation, which has since been replicated (8), underscores the potential of NO to exert a toxic, oxidative effect in asthma, since tyrosine nitration may alter the function of both regulatory and structural proteins (9).

In the context of eosinophilic inflammation, the close relationship between eosinophilia and nitrotyrosine formation (7) suggests that the eosinophil itself is an important factor in protein nitration. To investigate this possibility, we studied the relationship among eosinophilia, NO production, and 3NT formation in mice sensitized and challenged with ovalbumin (OVA). Our results indicate that in allergic inflammation, eosinophils are a physiologically significant source of NO and play a major role in protein nitration.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animals

Male A/J mice, weighing 24 to 26 g, were obtained from a commercial supplier (Harlan Sprague-Dawley, Indianapolis, IN) and housed in a conventional animal facility at our laboratory. All methods used in the experiments done with these animals were evaluated and approved by the Animal Ethics Committee at McGill University.

Sensitization and Allergen Challenge

Animals were sensitized twice, at a 7-d interval, with 100 µg OVA (chicken egg albumin grade V; Sigma Chemical Co., St. Louis, MO) in 0.4 ml of a 4 mg/ml suspension of Al(OH)₃ (Sigma), given subcutaneously. Seven days after the second injection, mice were lightly anesthetized with halothane (MTC Pharmaceuticals, Cambridge, ON, Canada) and then challenged intranasally with either 10 µg OVA in 50 µl sterile saline or with saline alone.

Experimental Protocols

Inflammatory response. To examine the effect of specific allergen challenge on cellular influx into the airways and on NO production, we investigated groups of animals at baseline or 6, 24, and 48 h after OVA challenge. At each time point, bronchoalveolar lavage (BALF) was done for determination of the total and differential cell count and measurement of NO metabolites (NO_x = NO₂ plus NO₃) in BALF, as subsequently outlined. After BAL, the lungs were harvested for detection of NOS II and 3NT by immunohistochemistry and Western blot analysis.

Effect of NOS inhibitors. The effects of NOS inhibitors on the response to OVA challenge were studied as follows. Sensitized mice were divided into: (1) an unchallenged group; (2) an OVA-challenged group; (3) a group treated with the nonspecific NOS inhibitor, N-nitro- L-arginine methyl ester (L-NAME; 10 mg/kg in 300 µl of sterile saline) (Sigma); and (4) a group treated with the highly NOS II-specific inhibitor N-3-aminomethyl-benzyl-acetamidine-dihydrochloride (1400W; 1.0 mg/kg in 300 µl of sterile saline; Calbiochem, San Diego, CA) (10). The inhibitors were administered intraperitoneally at 0.5 h before and 8, 20, 32, and 44 h after OVA challenge. Total and differential cell numbers, as well as NO_x in BALF, were measured 48 h after OVA challenge.

Contribution of eosinophils to NO production. To investigate the contribution of eosinophils to NO production and tyrosine nitration, we pretreated a group of mice with an anti-IL-5 antibody previously shown to suppress airway eosinophilia after OVA challenge (11). This monoclonal antimouse IL-5 antibody (rat IgG, clone TRFK 5; R&D Systems, Inc., Minneapolis, MN) was given intravenously 0.5 h before OVA challenge at a dose of 0.1 mg/kg. The eosinophil count, NO_x, and superoxide production in BALF cells were examined 48 h after the challenge.

BAL Procedure

Animals were deeply anesthetized with an intraperitoneal injection of 50 mg/kg sodium pentobarbital (Somnotol; MTC Pharmaceuticals). The upper trachea was cannulated and the lungs were lavaged twice (0.6 ml/lavage) with cold (4° C), Ca²⁺- and Mg²⁺-free phosphate-buffered saline (PBS) containing 0.5 mM ethylenediaminetetraacetic acid (EDTA) (Sigma). The samples were centrifuged at 800 × g for 10 min at 4° C, and the supernatants were removed and stored at 80° C. After resuspension, BALF cells were stained with 0.4% trypan blue (Sigma) and the total cell number was counted. A total of 3 × 10⁴ cells/slide were centrifuged at 120 × g for 2 min (Cytospin 2; Shandon, Pittsburgh, PA) and stained with May-Grunwald-Giemsa (BDH Laboratory, Poole, Dorset, UK) for differential cell counting.

Tissue Preparation

After BAL, the chest was opened and the left lung was removed, rinsed with PBS, embedded in ornithyl carbamyltransferase (Tissue TEK; Sakura Finetechnical Co., Tokyo, Japan), and then slowly frozen by immersion in isopentane (Sigma) cooled in liquid nitrogen. Cryostat sections (7 µm) were cut onto glass slides (Microm HM500; Microm International GmbH, Walldorf, Germany). The slides were then fixed in acetone:methanol (6:4 [vol/vol]) (Fisher) at room temperature (RT) for 7 min, and were stored at 80° C after 1 h of air drying.

NO_x Assay

We measured NO_x levels in BALF as an index of NO production, using the colorimetric Griess reaction (Cayman Chemical Co., Ann Arbor, MI). In order to convert NO₃ into NO₂, we mixed 80 µl of BALF with 10 µl of NO₃ reductase and 10 µl of its enzyme cofactor, and incubated the mixture at RT for 3 h. We then added 100 µl of Griess reagent and incubated the mixture for 10 min before reading. Absorbance was measured with a plate reader (SLT 400 ATC; SLT Lab Instruments, Salzburg, Austria) at 540 nm. The concentration of NO₂ was determined from standard curves constructed with serial concentrations of NaNO₂.

Immunohistochemistry

We used two methods for immunohistochemical studies, as follows.

Alkaline phosphate-mediated detection. Slides were defrosted and incubated with universal blocking solution (DAKO Diagnostics Canada Inc., Mississauga, ON, Canada) for 15 min at RT to block the nonspecific binding of antibodies. Optimal concentrations of primary antibodies were diluted in antibody dilution buffer (ADB; DAKO) and slides were incubated overnight with these antibodies in a humidified chamber at 4° C. On the next day the slides were washed in 0.05 M Tris-buffered saline (TBS: 0.05 M Tris base, pH 7.6, and 0.15 M sodium chloride; Sigma) for 10 min, and were then incubated with biotinylated antirabbit swine immunoglobulin (1:300 in ADB; DAKO) at RT for 50 min. Slides were washed and incubated with streptavidin-alkaline phosphatase-(AP) (1:200 in ADB; DAKO) at RT for 50 min. After washing, the slides were developed with Fast Red in substrate solution (0.1 M Tris base, pH 8.2; 0.5 mM naphthol AS-MX phosphate; 2% dimethylformamide; 1 mM levamisole; Sigma) or Vector blue (Vector Laboratories, Inc., Burlingame, CA).

Horseradish peroxidase-mediated detection. Slides were treated with 1% H₂O₂ for 30 min at RT, and were washed three times with 0.05 M TBS solution for 15 min each, after which they were incubated in universal blocking solution for 15 min. Optimal concentrations of primary antibodies were incubated overnight in a humidified chamber at 4° C. Following this the slides were washed in 0.05 M TBS for 10 min and incubated with biotinylated antirabbit swine immunoglobulin at RT for 50 min. The slides were then washed again and incubated with Strepavidin-biotin complex and horseradish peroxidase (HRP) (DAKO) for 50 min at RT. After washing, the slides were developed with a diaminobenzidine (DAB)-substrate chromogen kit (DAKO).

Antibodies and controls. We used a rabbit polyclonal antimouse macrophage NOS II IgG antibody (1:200; Transduction Laboratories, Lexington, KY) and a rabbit polyclonal anti-3NT IgG antibody (1:100; Upstate Biotechnology, Lake Placid, NY) in staining for NOS II and 3NT, respectively. Negative controls included buffer alone or dilutions of nonspecific purified rabbit IgG in the primary layer. As an additional control to confirm the specificity of 3NT staining, we incubated some slides with 500 mM sodium hydrosulfite (dithionite) dissolved in 100 mM sodium borate (Sigma), to reduce 3NT to aminotyrosine. Immunostaining was then done as previously described. Counterstaining was done with hematoxylin or Eosin Y solution (Fisher Scientific, Pittsburgh, PA).

Double Staining

Although eosinophils were the predominant granulocyte species present in BALF after OVA challenge, neutrophils were also present. To determine the contribution of neutrophils to 3NT formation, we performed double staining as follows. Slides were pretreated as described earlier and incubated overnight at 4° C with rat monoclonal antimouse neutrophil IgG antibody (Clone 7/4; Cedarlane, Hornby, ON, Canada) and anti-3NT antibodies (see the previous discussion). The slides were then incubated with a biotinylated goat antirat IgG antibody (Vector) and 2% normal mouse serum (Cedarlane) for 50 min, with streptavidin-AP (1:200) for 50 min, and with Streptavidin- biotion complex and HRP for 30 min. The secondary antibody for anti-3NT was exposed to Vector blue for color development. The secondary antibody for antineutrophil staining was exposed to DAB for color development, with repeated application of the second and third layers.

Morphometric Analysis

The number of 3NT-positive cells was estimated morphometrically with established techniques (12). Briefly, 3NT-positive cells were counted in the airway mucosa and peribronchial region of two to three airways in each animal. Cell counts were done by applying a 0.1 × 0.1-mm counting grid around the circumference of each airway. Cells in an area of at least 0.5 mm² were counted for each airway. The counts were expressed as the number of positive cells per 0.1 mm².

Protein Extraction from Lungs

The right lung was homogenized with six volumes of ice-cold tissue lysis buffer consisting of 50 mM Tris base (pH 7.4), 1% Triton X-100, 0.25% sodium deoxycholate, 150 mM sodium chloride, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 1 µg/ml aprotinin, 1 µg/ml leupeptin, 1 µg/ml pepstatin, 1 mM sodium orthovanadate, and 1 mM sodium fluoride (Sigma). Homogenized samples were incubated for 1 h at 4° C and centrifuged for 10 min at 10,000 × g at 4° C. The protein concentration was determined by the method of Bradford (13).

Western Blot Analysis

Proteins from each study group of animals, at a concentration of 100 µg/lane, were mixed with sample buffer (63 mM Tris base, pH 6.8; 2% sodium dodecylsulfate [SDS]; 0.0025% bromophenol blue; 10% glycerol, and 5% -mercaptoethanol) and separated on 4 to 12% Tris-glycine gradient gels (NOVEX, San Diego, CA) at 125 V for 2 h. Experiments were done both with pooled proteins from three animals per lane and with protein extracted from individual animals. Proteins were then electrophoretically transferred onto nitrocellulose membranes with 7.9 mM Na₂CO₃, 3.8 mM NaHCO₃ (Sigma), and 20% methanol (Fisher) at 250 mA for 2 h. Membranes with the transfers were stained with Ponceau Red (Sigma) for 1 min to confirm uniform protein loading.

Membranes were then blocked with TTBS solution (0.02 M Tris base, pH 7.5; 0.5 M sodium chloride, and 0.1% of Tween 20; Sigma) containing 7% nonfat milk and 1% fetal bovine serum (Life Technologies, Rockville, MD), and were probed overnight at 4° C with anti-NOS II antibody (see the previous discussion) diluted 1:10,000. The membranes were subsequently rinsed in 200 ml of TTBS for 30 min and exposed to the secondary HRP-conjugated antibody (donkey antirabbit Ig; Amersham Pharmacia Biotech, Inc., Baie d'Urfé, PQ, Canada) for 1 h at RT. A chemiluminescence detection system (ECL Plus; Amersham), Hyperfilm (Amersham), and Fluorchem 8000 (Alpha Innotech Corporation, San Leandro, CA) were used to detect the binding of this antibody. The membranes in this assay were also stripped, via 30 min of incubation with 62.5 mM Tris base (pH 6.8), 100 mM -mercaptoethanol, and 2% SDS at 55° C, and were then blocked with TTBS solution containing 1% bovine serum albumin (BSA; Sigma) at RT for 1 h. The stripped membranes were then probed at 4° C overnight with anti-3NT antibody (see the previous discussion) diluted 1:5,000 in the same blocking solution and were then processed as for NOS II. As a positive control for NOS II, we used protein extracted from a lysate of murine macrophages (RAW 264.7 cells; Transduction Laboratories) stimulated with interferon (IFN)- (10 ng/ml) and lipopoylsaccharide (LPS) (1 µg/ml) for 12 h. For 3NT, the positive control consisted of a mixture of nitrated proteins (bovine superoxide dismutase and BSA; Upstate Biotechnology) prepared according to the manufacturer's instructions. Density measurements and molecular weight calculations were done with commercial software (FluorChme FC; Alpha Innotech). Each density was calculated as the sum of the pixel values after background correction.

Superoxide Measurement (Lucigenin-Enhanced Chemiluminescence)

Spontaneous superoxide release was measured by lucigenin-enhanced chemiluminescence in a luminometer (Lumat LB 9501/16; EG&G Berthold, Wildbad, Germany) with modifications of Stevens' method (14). BALF cells (1.5 × 10⁵) were resuspended in 900 µl of Hanks' balanced salt solution without calcium chloride, magnesium sulfate, or phenol red (Sigma), and were placed in a 12 µ × 75 mm glass tube at 37° C for 5 min. After incubation, bis-N-methylacridinium nitrate (lucigenin; Sigma) was added, leading to a final concentration of 0.23 mM. Light output was monitored every 30 s for up to 60 min. The background count was subtracted at each point. The peak count and the area under the curve (AUC) over a 60-min period (calculated by trapezoid integration) were taken as indices of superoxide release.

Statistical Analysis

A one-way analysis of variance (ANOVA) with post hoc Bonferroni's analysis was used to determine significant differences. Results are expressed as mean ± SEM, and values of p < 0.05 were considered statistically significant. All statistics were determined with the commercial software package InStat Mac (GraphPad Software, San Diego, CA).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Airway Inflammation and NO_x Levels after OVA Challenge

The time course of inflammatory cell accumulation in BALF is shown in Table 1. Although eosinophils were predominant by 24 h after challenge, different cell types followed different temporal patterns. Before OVA challenge, more than 97% of the BALF cells were macrophages. Six hours after OVA challenge, the number of BAL neutrophils reached a peak (0.60 ± 0.31 × 10⁴ cells/ml; 10.5%; p < 0.01 as compared with unchallenged animals), but subsequently declined over 48 h. In contrast, BALF eosinophils were increased at 24 h and peaked at 48 h after the OVA challenge (9.34 ± 1.29 × 10⁴ cells/ml; 60.2%: p < 0.001, as compared with unchallenged or saline-challenged animals). The number of lymphocytes was small, but like that of eosinophils peaked at 48 h after the challenge (0.72 ± 0.09 × 10⁴ cells/ml; 4.7%; p < 0.001 as compared with unchallenged or saline-challenged animals). Saline-challenged animals failed to mount an eosinophilic response at 48 h. NO_x increased following challenge, with a time course similar to that of eosinophils (Figure 1). NO_x levels in OVA-challenged mice at 48 h and in saline-challenged mice were 4.63 ± 0.12 µM and 3.50 ± 0.04 µM, respectively.

View larger version (18K): [in this window] [in a new window]   Figure 1.   Time course of the number of eosinophils and the change in NOx in BALF after OVA challenge. Filled bars represent the number of eosinophils and open bars show the change in NOx compared with that measured in unchallenged animals. Although there was a transient decrease of NOx at 6 h, both eosinophils and NOx increased with similar time courses. Sham challenge with vehicle (saline) did not significantly affect either eosinophilia or NOx. (n = 4 to 9 mice per group, as in Table 1; *p < 0.05 compared with saline-challenged mice, ***p < 0.001 compared with saline-challenged mice).

Immunolocalization of NOS II, and 3NT in the Lung, and of BALF Cells

In unchallenged mice, we observed minimal NOS II staining in airway epithelial cells and in the alveolar region (Figure 2A). This was also the case in two saline-challenged mice. In contrast, at 48 h after OVA challenge there was enhanced NOS II staining of both airway epithelial cells and peribronchial inflammatory cells (Figure 2B). Upregulation of NOS II was also associated with evidence of increased tissue 3NT. In unchallenged mice, there was only minimal patchy positive immunostaining of inflammatory cells in the alveolar region, and of airway epithelial cells (Figure 2C). By 48 h after challenge, inflammatory cells around the airway, as well as epithelial cells, were strongly positive (Figure 2D). Similar results were found when Fast Red staining was used (data not shown), confirming that staining did not simply result from artifactual 3NT formation after addition of H₂O₂. BALF cells harvested at 48 h also stained with anti-NOS II antibody (Figure 2E). More than 80% of eosin-positive cells were stained for NOS II. These cells also exhibited staining with anti-3NT antibody (Figure 2F). The pattern of staining varied according to cell type. In general, macrophages exhibited spotty staining in the cytoplasm, whereas eosinophils were stained primarily near the cell membrane.

View larger version (97K): [in this window] [in a new window]   Figure 2.   (A) NOS II expression in a representative airway from an unchallenged mouse. Only background staining is seen. (B) NOS II expression in an airway 48 h after OVA challenge. Positive staining is present in airway epithelial cells and some peribronchial inflammatory cells (arrows). (C ) 3NT staining in a representative airway from an unchallenged A/J mouse. Minimal patchy positive immunostaining of airway epithelial cells is seen. (D) 3NT staining 48 h after OVA challenge. There is marked staining of airway epithelial cells and some peribronchial inflammatory cells. (E ) Immunostaining for NOS II of BALF cells counterstained with eosin. More than 80% of eosin-positive cells also stained for NOS II. (F ) Immunostaining of BALF cells for 3NT. Evidence of 3NT formation is present in both eosinophils and macrophages. (G) To confirm the specificity of antibody staining, we applied nonspecific purified rabbit IgG to BALF cells and stained with the same secondary and tertiary antibodies as for 3NT and NOS II studies. No nonspecific staining was detected. (H ) Double staining for both 3NT (blue) and neutrophils (brown), demonstrating evidence of 3NT formation in some but not all neutrophils.

Double Staining

Although at 48h after OVA challenge the majority of BALF inflammatory cells were eosinophils, 3NT staining appeared to be present in neutrophils as well. To determine the relative contribution of neutrophils, we performed double staining, using an antibody directed against neutrophils and the anti-3NT antibody (Figure 2H). By 48 h after challenge, cells positive for both 3NT and the neutrophil marker were seen in the peribronchial region, indicating that neutrophils were responsible for at least some of the 3NT staining in granulocytes. Nevertheless, many 3NT-positive cells did not stain for the neutrophil marker, confirming the contribution of eosinophils to tyrosine nitration. Furthermore, the majority of cells positive for the neutrophil marker were 3NT-negative.

Effects of NOS Inhibitors on Eosinophilic Inflammation, NO Production, and 3NT Formation

As expected, administration of NOS inhibitors decreased NO_x production. A similar magnitude of suppression was found when we used high doses of L-NAME as with lower doses of the NOS II-selective inhibitor 1400W (Figure 3B), in accord with the concept that NOS II is responsible for most of the increase in NO production in this model. Pharmacologic suppression of NO production was also associated with a substantial reduction in 3NT formation (Figure 3C). Interestingly, suppression of NO production resulted in a marked reduction in BALF eosinophilia as compared with OVA challenge (61.5% reduction with 1400W, 37.2% reduction with L-NAME, Table 2 and Figure 3A), suggesting that eosinophil recruitment is at least partly driven by NO. With regard to other inflammatory cells, the influx of neutrophils tended to increase after administration of NOS inhibitors, but this did not reach statistical significance (Table 2).

View larger version (31K): [in this window] [in a new window]   Figure 3.   Effects of NOS inhibitors on number of BALF eosinophils (A), concentration of BALF NOx (B) and number of 3NT-positive cells around the airways (C ). Both the nonselective NOS inhibitor L-NAME (horizontal bars) and the highly selective NOS II inhibitor 1400W (diagonal bars) blocked increases in eosinophils, NOx, and 3NT formation after OVA challenge (solid bars) compared with unchallenged mice. Columns represent group means (SEM), n = 6 mice per group. *p < 0.05, **p < 0.01, ***p < 0.001.

Effect of Anti-IL-5 Antibody on Eosinophilic Inflammation, NO and O₂ Production, and 3NT Formation

We attempted to suppress eosinophilic inflammation at 48 h by administering an anti-IL-5 antibody previously reported to block eosinophil influx after OVA challenge (11). Administration of the anti-IL-5 antibody led to nearly complete suppression of BALF eosinophilia (90.1% reduction as compared with OVA-challenged mice, p < 0.001; Figure 4A), with a corresponding reduction of BALF NO_x by 25.0% (p < 0.05 as compared with OVA-challenged animals; Figure 4B). Although the reduction in NO production was only partial as compared with that with NOS inhibitors (Figure 7), anti-IL-5 antibody treatment almost completely suppressed 3NT staining around the airways (p < 0.01 as compared with OVA; Figure 4C). The reductions in NOS II expression and 3NT formation were also seen on Western blot analysis. The amount of NOS II protein following anti-IL-5 antibody treatment was similar to that in unchallenged mice (Figure 5A). The effect of anti-IL-5 antibody on 3NT was most marked in proteins with molecular weights near 16, 22, 32, and 47 kD (Figure 5B). When the intensity of bands was quantified densitometrically, the total intensity (calculated as the sum of the intensities of individual bands) increased by 16.6% over the baseline value in OVA-challenged mice. Administration of anti-IL-5 antibody decreased this to 9.5%. The increase in density after OVA challenge varied greatly, depending on molecular weight (Figure 5B). For example, there was a 60% increase in density of the band at 32 kD, whereas bands at 34 kD and 42 kD did not change at all. Similarly, the effect of anti-IL-5 antibody was variable, although the density of the majority of bands (16, 18, 22, 26, 32, 47, and 60 kD) was decreased as compared with that after OVA challenge.

View larger version (24K): [in this window] [in a new window]   Figure 4.   Effects of anti-IL-5 antibody treatment on BALF eosinophils (A), BALF NOx (B), and 3NT-positive cells around the airways (C ), shown as in Figure 3. Intravenous administration of 0.1 mg/kg of monoclonal antimouse IL-5 antibody 0.5 h before OVA challenge nearly completely prevented eosinophilia, while substantially reducing both NOx and 3NT immunostaining. Bars represent group means (SEM), n = 6 mice per group. *p < 0.05, **p < 0.01, ***p < 0.001. Groupwise comparisons are indicated by horizontal bars.

View larger version (22K): [in this window] [in a new window]   Figure 7.   Change in NOx in BALF, expressed as a percentage of levels in unchallenged mice. Filled bar : OVA-challenged mice at 48 h; horizontal bar : OVA-challenged and L-NAME-treated mice; diagonally shaded bar : OVA-challenged and 1400W-treated mice; crosshatched bar : anti-IL-5 antibody- and OVA-treated group. Bars represent group means (SEM), n = 6 mice per group. *p < 0.05, ***p < 0.001 compared with OVA-challenged mice.

View larger version (72K): [in this window] [in a new window]   Figure 5.   Western blots of NOS II (A) and 3NT (B), using 100 µg of protein pooled from three different animals in each group. Lane 1: unchallenged animals; lane 2: OVA-challenged animals at 48 h; lane 3: anti-IL-5 antibody treated and OVA-challenged mice at 48 h. The positive control for NOS II was macrophage lysate (2 µg). The positive control for 3NT was a mixture of nitrated SOD and BSA (2 pg). Positive controls were prepared according to the manufacturer's instructions. These results are representative of four similar experiments.

Because 3NT formation depends on granulocyte oxidant activity as well as NO, we also assessed the effect of anti-IL-5 antibody on spontaneous O₂ production as an index of the capacity of BALF cells to generate ROS. Administration of anti-IL-5 antibody substantially reduced the increase in O₂ production by BALF cells after OVA challenge as compared with that in OVA-challenged animals (p < 0.05, Figures 6A and 6B).

View larger version (21K): [in this window] [in a new window]   Figure 6.   Spontaneous superoxide production as detected by lucigenin-enhanced chemiluminescence. BALF cells (1.5 × 105) were resuspended in HBSS and then incubated with 0.23 mM lucigenin. The resulting light output was monitored every 30 s for up to 60 min. The area under the curve for 60 min was calculated by trapezoid integration. Bars represent group means (SEM), n = 6 mice per group. *p < 0.05. Groupwise comparisons are indicated by horizontal bars.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We investigated the role of eosinophils in NO production and nitrotyrosine formation in a murine model of allergic inflammation. During the 48 h after OVA challenge, NO production and eosinophilia increased in concert with evidence of increased NOS II expression and 3NT formation. Furthermore, the administration of NOS inhibitors suggested that the eosinophilia was partly dependent on NO production. Conversely, eosinophils exhibited evidence of NOS II expression, and inhibition of eosinophilia with an anti-IL-5 antibody demonstrated that eosinophils contributed to both NO production and 3NT formation. Taken together, these findings demonstrate the close relationship between NO production and eosinophilia, and underscore the importance of eosinophils in protein nitration.

As with any study involving animals, our results depended on the choice of experimental animal and the study protocol. Although our model does not fully recapitulate all the features of asthma, the OVA-challenged A/J mouse shares several features with the clinical syndrome of acute inhalational allergen exposure as seen in human asthma. For example, the time course of cellular influx into the BALF compartment, with neutrophils predominating early, followed subsequently by increased eosinophil and lymphocyte counts, is similar to that described in other models and in human segmental airway challenges (15, 16). The parallel rise in the number of eosinophils and in NO production is also similar to what is seen in human asthma. Thus, although extrapolation from animal models to human disease is always difficult, this model provides some insight into the possible relationships between NO production and eosinophilia in airway inflammation.

Both asthma (7, 17) and allergic airway inflammation in other animal models (18) are associated with the detection of increased NOS II protein expression. Constitutively expressed in human airway epithelium (19), NOS II is upregulated by proinflammatory cytokines including tumor necrosis factor (TNF)-, IL-1, and IFN- (20). In asthmatic airways, NOS II is upregulated in the epithelium (17) and in submucosal inflammatory cells (7), in accord with the concept that NO plays an important role in the inflammatory process. In this study, we used pharmacologic NOS inhibitors to help confirm the functional importance of increased NOS II protein. We found the selective NOS II inhibitor 1400W to be as effective in suppressing NO_x as a higher dose of the nonselective inhibitor L-NAME. Given that the selectivity of 1400W for NOS II is much greater than that of other relatively selective inhibitors such as aminoguanidine (10), our results strongly indicate that NOS II is the principal source of increased NO_x in this model.

In addition to their expected effects on NO, both L-NAME and 1400W reduced eosinophil influx after OVA challenge, an observation previously made in mice (21) and other species (22, 23). This finding is also consistent with a report that NOS II-deficient mice exhibit reduced airway eosinophilia after allergen challenge as compared with wild-type controls (24). In contrast, De Sanctis and colleagues failed to detect any reduction in eosinophilia in OVA-challenged NOS II-deficient mice (25), an inconsistency that may relate to differences in sensitization and challenge protocols, or perhaps to the genetic background of the experimental animals. Nevertheless, our data are consistent with the majority of reports that indicate a role for NO and NOS II in airway eosinophilia.

The mechanism by which NO promotes allergic eosinophil influx is incompletely understood. NO may regulate cytokine production, possibly via the ability of nitrogen oxides to activate G proteins, such as p21^ras, and via increased nuclear concentrations of the transcription factor nuclear factor- (26). Moreover, NO can mediate the TNF--induced activation of transcription factor-activating protein-1 through protein G kinase activation (27). In NOS II-deficient mice (24), Xiong and colleagues have demonstrated increased IFN- production by cultured lymphocytes harvested from lung and spleen, suggesting that deficiency of NOS II leads to a shift in cytokine profile. In support of this, NOS II-deficient mutant mice show enhanced T-helper type 1 (Th1) responses following infection (28), and T-cell clones treated with NOS inhibitors show an increase in IFN- secretion (29). NO has further been shown to promote a Th2 response in macrophages by blocking IL-12 secretion (30). NO may also promote eosinophilic inflammation through other mechanisms. For example, NO has been reported to alter the chemotactic activity of eosinophils in vitro (31), and may inhibit eosinophil apoptosis (32). On the other hand, it is unlikely that NO acts through its effects on adhesion molecules. NO downregulates vascular cell adhesion molecule and intracellular adhesion molecule-1, thereby inhibiting neutrophil recruitment (33), an effect opposite to that reported here.

As in bronchial biopsy specimens from asthmatic airways (7), we detected tyrosine nitration in OVA-challenged mice. Although we found 3NT staining among neutrophils and macrophages, 3NT was primarily detected in eosinophils in this model. To further explore the role of eosinophils in NO production and 3NT formation, we used an anti-IL-5 antibody to inhibit eosinophilia without directly altering NOS activity. IL-5 is a Th2 cytokine produced mainly by T lymphocytes, but also by eosinophils and basophils. Administration of anti-IL-5 antibody has been previously reported to prevent airway eosinophilia and BHR in murine models of allergic bronchoconstriction (34). In the present study, we found that administration of anti-IL-5 antibody almost completely prevented BALF eosinophilia after OVA challenge. This inhibition of eosinophil influx was associated with a reduction in NO_x in BALF. Although anti-IL-5 antibody treatment was less effective in suppressing total NO production in BALF than was L-NAME or 1400W (Figure 7), anti-IL-5 antibody reduced NO_x nearly to its baseline value. Although it is possible that eosinophils indirectly contribute to NO production by stimulating other cells, the detection of NOS II in eosinophils themselves indicates that they are a physiologically important source of NO. Furthermore, previous in vitro observations (5), and a recent report that induction of eosinophilia by exogenous eotaxin administration increased NO production in allergic rhinitis (35), also suggest that eosinophils contribute to NO production in allergic inflammation.

As might be expected from its effect on BALF NO_x, anti-IL-5 antibody also markedly reduced staining for 3NT. Thus, as in models of neutrophilic inflammation, activated granulocytes (in this case eosinophils) contribute to tyrosine nitration after specific allergen challenge. Interestingly, although the reduction in NO_x by anti-IL-5 antibody was less than that by L-NAME or 1400W, the effect on 3NT staining was as great or greater (Figure 5), suggesting that reduced NO production may not fully explain the suppression of tyrosine nitration by anti-IL-5 antibody. Detection of 3NT has generally been considered the "footprint" of NO₃, since this highly toxic metabolite is readily formed from NO and O₂, and is capable of nitrating proteins at tyrosine residues (6). Thus, decreases in NO production might be expected to decrease NO₃ availability and to thereby reduce 3NT formation. It is controversial, however, whether NO₃ alone is responsible for protein nitration in vivo. In intact organisms, tyrosine nitration appears to require the presence of granulocytes such as neutrophils or eosinophils, which can use NO₃ as a substrate for the production of additional toxic free radicals (9). It has been further suggested that peroxidases, such as neutrophil myeloperoxidase and eosinophil peroxidase, are responsible for at least some 3NT formation (36, 37). In this regard, we have recently reported preliminary findings of markedly reduced 3NT formation in eosinophil peroxidase-deficient mice (38), suggesting that the reduction in 3NT following anti-IL-5 antibody administration may be a function of decreased levels of eosinophil peroxidase.

Another mechanism by which anti-IL-5 antibody could potentially reduce 3NT formation would be by a reduction in the overall production of ROS. To address this, we measured spontaneous O₂ generation by BALF cells as an indicator of the capacity of airway inflammatory cells to generate ROS. We found that administration of anti-IL-5 antibody completely suppressed the increase in spontaneous O₂ generation by BALF cells after OVA challenge. In that the primary effect of anti-IL-5 antibody was to block eosinophil influx, the decrease in O₂ generation that we detected was almost certainly a result of a reduced number of eosinophils in BALF. We cannot be certain, however, whether decreased 3NT staining after treatment with anti-IL-5 antibody directly reflects decreased O₂ generation or reflects other aspects of eosinophil function, such as the action of eosinophil peroxidase noted earlier.

In asthma, both the identity and functional significance of nitrated proteins is unclear. In this model, we detected nitration of proteins through analysis of molecular weights. The densities of bands at 16 kD and 32 kD were increased after OVA challenge, raising the possibility that superoxide dismutase (SOD) (39) was nitrated. ONOO is known to inactivate human MnSOD protein through exclusive nitration of tyrosine 34 (39). Nitration of this dimeric protein would be predicted to produce bands at these molecular weights. Similarly, Haddad and colleagues (40), using Western blot analysis, reported that nitration of surfactant protein (SP)-A decreased its ability to aggregate lipids and was associated with 3NT-positive bands with molecular weights in the range between 31 kD and 66 kD. We detected increased densities of bands at 32 kD and 66 kD, suggesting that SP-A could account for some of the nitrated proteins in this model. Further work is required to definitively identify the proteins that undergo nitration, in order to determine the functional significance of 3NT formation.

In summary, in this model of allergic airway inflammation, NO production and airway eosinophilia were tightly coupled. Although eosinophilic inflammation is in part promoted by NO production, eosinophils themselves are a major source of NO. Furthermore, when airway eosinophilia following allergic challenge was suppressed by administration of an anti-IL-5 antibody, there was a marked reduction in tyrosine nitration. These findings underscore the importance of eosinophils as effector cells in asthma, and suggest an important role for these cells in oxidant injury in this disease.